Leeds Thesis Template - White Rose eTheses Online

Loading...
1

The classification and analysis of spirals in decorative designs Alice Edith Humphrey

Submitted in accordance with the requirements for the degree of Doctor of Philosophy

The University of Leeds School of Design

August 2013

2 The candidate confirms that the work submitted is her own, except where work which has formed part of jointly-authored publications has been included. The contribution of the candidate and the other authors to this work has been explicitly indicated below. The candidate confirms that appropriate credit has been given within the thesis where reference has been made to the work of others.

Chapter 3 – Analysis and Classification Method includes material published in: Humphrey, A. and Hann M.A. (2012) Steps Towards the Analysis of Geometric Decorative Motifs Using Shape-matching Techniques. in R.Bosch, D.McKenna and R.Sarhangi (eds) Proceedings of Bridges 2012: Mathematics, Music, Art, Architecture, Culture. Phoenix, Arizona: Tessellations Publishing. pp.495-498.

The method proposed and examples presented in this paper were the candidate's work. Prof. M.A.Hann presented the paper at the conference and provided comments and advice as to the phrasing of the published paper.

This copy has been supplied on the understanding that it is copyright material and that no quotation from the thesis may be published without proper acknowledgement.

The right of Alice Humphrey to be identified as Author of this work has been asserted by her in accordance with the Copyright, Designs and Patents Act 1988.

© 2013 The University of Leeds and Alice Humphrey

3

Acknowledgements I am grateful for the help and advice of my supervisor Prof. Michael Hann throughout the course of the research and to Dr. Briony Thomas for her help during the early stages of the work. Antonia Lovelace at Leeds City Museum; Dr. Ashley Cooke and Lynne Heidi Stumpe at Liverpool World Museum; Rachel Grocke at the Oriental Museum, University of Durham and Andrew Parkin at the Great North Museum provided generous access to their collections which was much appreciated. I am grateful to the survey participants for their time and effort. I would also like to thank Tom Humphrey for his help with writing the analysis applet and Prof. Charles Taylor for advice on statistical analysis. This research was funded by a School of Design scholarship from the University of Leeds which was much appreciated. I am very grateful for the support and patience shown by my family throughout the writing of this thesis.

4

Abstract Structure in decorative patterning has been used to identify perceptual preferences and broader social and cultural patterning of a society. This thesis proposes a new method of analysis and classification of decorative spirals as a test of whether the analysis of an individual type of geometric decorative motif in isolation can reflect social and cultural trends in a similar manner. The proposed analysis method, which takes a quantitative, bottom-up approach to classification, was compared with a topdown classification method used in previous studies based on mathematical spiral forms. Similarities were found across the two classification methods, both are clearly strongly influenced by the expansion rate of the spiral but the newly proposed method is felt to offer a more flexible approach to the classification of motifs which accommodates gradual variation and the decorative properties of the motifs beyond a simple mathematical curve. The proposed method was tested on four archaeological case studies: Egyptian scarabs of the Middle Kingdom and First and Second Intermediate Periods, Cypriot Bronze Age painted pottery and gold-work, Shang and Western Zhou Dynasty bronze vessels, and Japanese Jōmon Period pottery. Quantification and cluster analysis of spirals in these studies identified variation in motif shape that could be attributed to the production method but the chief finding was one of homogeneity across the four cultures and a limited range of forms of decorative expression which suggests an influence of natural spiral forms on decorative examples across the cultures sampled. An analysis of the possible influence of spiral optical effects arising from disturbances in the functioning of the visual cortex on decorative spirals did not yield significant results.

5

Table of contents Acknowledgements ...................................................................................................... 3 Abstract ........................................................................................................................ 4 List of tables ................................................................................................................. 7 List of figures ................................................................................................................ 8 Chapter 1 Introduction............................................................................................... 13 1.1 The analysis of design ................................................................................... 14 1.2 Spirals in nature ............................................................................................ 17 1.3 Symbolic meanings of spirals in decorative arts ........................................... 24 1.4 The choice of spiral motifs as a research focus ............................................ 25 1.5 Overview of the thesis .................................................................................. 28 1.5.1 Research questions ............................................................................. 30 1.5.2 Thesis structure................................................................................... 34 Chapter 2 Definition of terminology and variables and discussion of spirals as subjective visual phenomena..................................................................................... 36 2.1 The definition of a spiral ............................................................................... 37 2.1.1 Mathematical spiral forms .................................................................. 41 2.2 Decorative spiral classification...................................................................... 42 2.2.1 Form .................................................................................................... 45 2.2.2 Stylistic variations in decorative spirals .............................................. 46 2.3 Perception of spiral forms............................................................................. 56 2.4 Endogenous visual phenomena and their occurrence in decoration ........... 62 2.5 Construction techniques ............................................................................... 67 2.6 Summary ....................................................................................................... 70 Chapter 3 Development of the classification and analysis method .......................... 71 3.1 Data collection and processing ..................................................................... 71 3.1.1 Sources of data ................................................................................... 71 3.1.2 Data collection .................................................................................... 73 3.1.3 Data processing................................................................................... 75 3.1.4 Identification of the centre ................................................................. 92 3.1.5 Precision .............................................................................................. 93 3.2 Identification of form .................................................................................... 94 3.2.1 Creation of the primary path .............................................................. 94

6 3.2.2 Normalisation...................................................................................... 96 3.2.3 Quantification of form ........................................................................ 96 3.3 Quantification of style................................................................................... 97 3.3.1 Orientation .......................................................................................... 97 3.3.2 Rotational length ................................................................................ 97 3.3.3 Curvilinearity ....................................................................................... 98 3.3.4 Irregularity .......................................................................................... 98 3.3.5 Width of paths and spacing .............................................................. 100 3.4 Methods of comparison .............................................................................. 101 3.4.1 Summarisation of spiral form data ................................................... 101 3.4.2 Variable weighting ............................................................................ 104 3.4.3 Clustering .......................................................................................... 105 3.5 Summary ..................................................................................................... 106 Chapter 4 Background to the case studies .............................................................. 107 4.1 Choice of case studies ................................................................................. 107 4.2 Culturally dependent interpretation of decorative spirals ......................... 108 4.3 Background to the case studied.................................................................. 114 4.3.1 East Asia ............................................................................................ 114 4.3.2 The Eastern Mediterranean and Egyptian Area ............................... 126 4.4 Summary ..................................................................................................... 149 Chapter 5 Testing and evaluation of the classification method .............................. 150 5.1 Comparison of the method with fitting to mathematical spirals as a classification technique ............................................................................................ 150 5.2 Results of the application of the method to the test case studies ............. 156 5.2.1 Jōmon Period case study .................................................................. 156 5.2.2 Shang Dynasty and Western Zhou Period case study ...................... 160 5.2.3 Middle Kingdom Egyptian case study ............................................... 164 5.2.4 Bronze Age Cypriot case study ......................................................... 167 5.2.5 Cross-cultural comparison ................................................................ 172 5.3 Endogenous visual phenomena .................................................................. 178 5.3.1 Comparison with decorative spirals……………………………………………….182 5.4 Inspiration for spiral decoration ................................................................. 185 5.5 Evaluation of the method ........................................................................... 191

7 5.5.1 Digitisation ........................................................................................ 194 5.5.2 Irregularity ........................................................................................ 195 5.5.3 Curvilinearity ..................................................................................... 195 5.5.4 Orientation ........................................................................................ 196 5.5.5 Path width and spacing ..................................................................... 196 5.5.6 Interpretation of the spectral density plots ..................................... 197 5.5.7 Context of the proposed method with other classification systems 201 5.6 Summary ..................................................................................................... 203 Chapter 6 Summary and conclusions ....................................................................... 205 6.1 Summary of the method ............................................................................. 206 6.2 Potential for future development of the method ...................................... 208 6.2.1 Surface curvature and three-dimensionality .................................... 208 6.2.2 Expanding the range of motifs .......................................................... 210 6.3 Comparison with mathematical spirals ...................................................... 211 6.4 Results of the case studies .......................................................................... 212 6.5 Spiral source ................................................................................................ 214 6.5.1 Cortically generated spiral forms ...................................................... 214 6.5.2 Naturally occurring real spirals as sources of inspiration ................. 216 6.6 Potential for expanding the application of the analysis method ............... 216 6.7 Conclusion ................................................................................................... 217 References

............................................................................................................ 218

Appendix 1 Perceptual survey ................................................................................. 240 A1.1 Participants ............................................................................................... 240 A1.2 Apparatus and stimuli ............................................................................... 242 A1.3 Procedure ................................................................................................. 242 A1.4 Results ....................................................................................................... 243 A1.4.1 Similarity ......................................................................................... 243 A1.4.2 Placement of centre and end points .............................................. 246 A1.4.3 Motif descriptions .......................................................................... 248 A1.5 Potential expansion of the perceptual survey.......................................... 249 Appendix 2 Results tables ........................................................................................ 275

8

List of tables 3.1 3.2 4.1 4.2 4.3 4.4 4.5 4.6 A1.1 A1.2 A1.3 A1.4 A1.5 A1.6 A2.1 A2.2 A2.3 A2.4 A2.5 A2.6 A2.7 A2.8 A2.9 A2.10 A2.11 A2.12 A2.13 A2.14 A2.15 A2.16

Types of contextual data collected for the spirals ........................................... 75 Image rectification tests average error/mm ..................................................... 77 Jōmon Period divisions and numbers of spirals sampled from each period ..115 Types of Shang Dynasty vessel from which spirals were sampled. ................125 Timeline of Egyptian kingdoms and dynasties and numbers of spirals sampled in each period .................................................................................127 Timeline of Cypriot periods and the spirals sampled from each. ...................137 Places of origin of the Cypriot vessels and gold-work from which spirals were sampled. ........................................................................................................139 Cypriot vessel types from which spirals were sampled. .................................148 Similarity values by test motif ........................................................................244 Overview of similarity ratings by variable ...................................................... 245 Free text descriptors used for motif shape. ...................................................250 Free text keywords for motif description. Participants were asked to enter up to five keywords. ..................................................................................... 256 Frequency of choice of predefined keywords. ..............................................261 Perceptual survey test motifs .........................................................................262 Early examples of spiral decoration and the distribution of prehistoric and early historic cultures in which spirals were commonly occurring motifs ..274 Contextual and style data for the Jōmon Period case study .......................... 275 Contextual and style data for the Shang Dynasty and Western Zhou Period case study .....................................................................................................279 Contextual and style data for the Egyptian Middle Kingdom case study ......284 Contextual and style data for the Bronze Age Cypriot case study ..................289 Contextual and style data for the EVP examples ............................................295 Classification of Jōmon period spirals according to correlation with mathematical spiral types. ............................................................................300 Classification of Shang/Western Zhou Dynasty spirals according to correlation with mathematical spiral types. ................................................301 Classification of Egyptian scarab spirals according to correlation with mathematical spiral types. ...........................................................................302 Classification of Bronze Age Cypriot spirals according to correlation with mathematical spiral types. ...........................................................................303 Clustering of the Jōmon Period spirals. ......................................................... 304 Clustering of the Shang and Western Zhou Dynasty spirals. ......................... 305 Clustering of the Egyptian Middle Kingdom spirals. ......................................306 Clustering of the Bronze Age Cypriot spirals. .................................................307 EVP type groups and corresponding decorative spirals. ................................ 308 Clustering of spirals by representational meaning. ....................................... 309

9

List of figures 1.1 1.2

1.3 1.4

1.5 1.6 1.7 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11

2.12 2.13 2.14 2.15 3.1 3.2 3.3

Natural spiral examples: a. vetch tendril. b. burnet moth proboscis. c. Medicago sp. seed. d. cactus leaf rosette ......................................................... 19 Decoration depicting or inspired by naturally occurring biological spirals a. fern shoot, b. Maori spiral carving. c. regal ringneck snakea d. faience snake ornament from Late Period Egyptian coffinb. e. Indian C19th-20th conch shell modified for use as a trumpetb f. Edward Lear illustrationc. g. Aztec printing stamp depicting a spider monkeyb. h. Asanti gold weight depicting crocodiles with a curled tails. ............................................................................ 21 Spiral forms on representations of human figures. ............................................ 22 Physical phenomena represented as spirals: a. Qing Dynasty panel with spirals in clouds and wavesd. b. Japanese c.20th century textile printing stencil with conventionalised lightning depictiond. .......................................... 23 Early spiral decoration. ........................................................................................ 26 Representationally or geometrically similar motifs: a. Iznik, Turkeyb. b. Britishb. c. Peruvian CEb. .................................................................................... 27 Mathematical spiral types. .................................................................................. 31 Variety in decorative spirals. ............................................................................... 36 High spiral recognition motif (test motif 100). .................................................... 38 Pictorial test motifs nos 154, 155, 156, 157, 157b. ............................................. 38 Similar spiral and non-spiral motifs: a. Papuan drum, linked spiral and nonspiral curves. b. scarab with spiral and non-spiral linkse. ................................. 39 Test motifs 112 and 113 representing low rotation key patterns. ..................... 39 Whirl motif of low rotational length. .................................................................. 40 a. False spirals on a scarab. b. A spiral resembling concentric circles on a Cypriot barrel jug............................................................................................... 41 Shang Dynasty spirals with a high form correlation but low style match. .......... 47 Angular spiral meander pattern on a ninth-century cross. ................................. 48 Celtic or Anglo Saxon hanging bowl escutcheon with trumpet spirals showing path width variation.b ....................................................................................... 51 Patterned or random irregularity. a. Plate with a freehand spiral 1950’s-60’s designed Teddy Millington Drakec; b. Detail of an octoput tentacle from a jar from Crete, 1450-1400BCE. ......................................................................... 53 Test motif 88 showing surface patterning........................................................... 54 EVP class groupings. ............................................................................................ 58 Spirals and spiraloids within a hexagram construction. Detail on a Mamluk incense burner.b ................................................................................................ 67 San Sebastian ceramic figurine, 2nd century CE.f ................................................. 69 The specimen spiral CSBM17.9 from a Western or Eastern Zhou Dynasty you (ritual vessel).b ................................................................................................... 71 Scarab showing spirals damaged beyond analysis. ............................................. 74 Sample image rectification test images............................................................... 76

10 3.4 3.5 3.6 3.7

3.8 3.9 3.10 3.11

3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 3.23 3.24 3.25 3.26 3.27 3.28 3.29 3.30 3.31 3.32

Damage and distortion to spiral decoration. a. Cypriot gold mouthpiece with warpingb. b. Scarab with cracking. .................................................................... 78 The spiral line segments obtained by the best automatic edge detection result for the specimen spiral (indicated by change of colour). ....................... 79 Tracing of the specimen spiral in two directions of the spiral outline in green (anticlockwise) and red (clockwise) and the averaged lines in black. .............. 80 Spirals with end embellishments and central eyes. a. Mycenaean flask. b. Uzbek tile panel c.1359CEc. c. feeding bottle from Crete (1300-1190BCE). d. Zuni olla, New Mexico. ...................................................................................... 81 a-f. Placement of end points by survey participants. g. Placement of centre point in a spiral with an eye. ............................................................................. 82 Examples of spiral ends with the tracing lines and end points marked. ............. 83 Spiral end point determination. ......................................................................... 84 Spirals joining a line or sharing part of their curve. a. Detail from a textile design by Guido Marcinid. b. Detail from an Asante wrap, Nigerian. c. Jar fragment, Crete. ................................................................................................ 85 Tests of path continuity with end point estimations marked. ............................ 86 a. Shang bronze vessel with ambiguous figure/ground configuration. b. Mimbres pottery with ambiguous decoration. ................................................. 87 Figure/ground differentiation. ............................................................................ 87 Peruvian textile with two balanced fields of colour meeting. ............................ 88 Decorative context used to identify the foreground spiral line of the sampled spiral line as the wider line. .............................................................................. 88 a. Cypriot painted spiral requiring interpolation due to damage.g b. Cypriot mouthpiece with the path broken by another decorative element.b .............. 89 Test motifs for interpolation of the spiral path. .................................................. 90 Test motifs for centre placement of an unclosed shape ..................................... 91 Spiral centre within the incircle of the specimen spiral. ..................................... 92 Digitally created spiral and manual trace line. .................................................... 93 Specimen spiral with the primary path marked. ................................................. 95 Calculation of the primary path around the eye of a spiral. ............................... 95 Maximum diameter of the specimen spiral marked, orientation 1.03 radians from the positive x axis. .................................................................................... 97 Calculation of curvilinearity. ................................................................................ 98 Calculation of irregularity. ................................................................................... 99 Areas of the specimen spiral with irregularity detail lost due to corrosion ........ 99 Measurement lines used for calculation of spacing on the specimen spiral. ...100 Specimen spiral - a. Path width measurements (red) and trendline. b. Spacing measurements and trendline ............................................................ 101 Down-sampled sine wave of the specimen spiral. ............................................102 Spectral density plot of the specimen spiral. ....................................................103 Spectral density plots. ....................................................................................... 104

11 4.1 4.2 4.3 4.4 4.5

4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13

4.14 4.15 4.16 4.17 4.18 4.19 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12

Spirals in representational context in the eyes and tail of a frog on a 1 st century bell from Yunnan province, China...................................................... 109 Perceptual survey test motifs 97, 100b and 100c .............................................111 a. Test motif 157 - ironwork scroll in railings. b. Test motif 157b - modern ceramic plate from Granada with spiraloids. ..................................................112 a. Possible snake motif (indicated by arrow) on a Middle Jōmon pot, Tsunan Town Board of Education. b. Bird motif drawn from a Middle Jōmon pot. ...118 Shang/Western Zhou Dynasty animalistic decoration. a. A taotie mask detail from a late Shang Dynasty jia.b b. A long, one of several types of dragon depicted in the period.b ..................................................................................123 Consistency in use of spirals in depictions of birds' wings in different materials. a. Late Shang jade goose.b b. On a Shang bronze vessel.b .............123 Background leiwen patterning on a Shang vessel.b ...........................................124 Ninth Dynasty button seal from Qaw.e.............................................................. 131 a. Middle Kingdom spiral border.e b. New Kingdom spiral border.e .................132 Crossed spiral decoration on a New Kingom scarab.e .......................................132 Spiral representing a lotus or reed.e..................................................................133 a. Hathor horns.b b. An udjat-eye on the top of a scarab. c. An udjat-eye on a 26th Dynasty amulet.b ...................................................................................... 133 a. Snake as border decoration on a Middle Kingdom scarab.b b. Endless knot design on a Middle Kingdom scarab.e c. Snake in an endless knot form on an animal mummy case.e ................................................................................134 Base Ring ware juglet.b ...................................................................................... 144 White Painted Wheelmade III bowl.b ................................................................ 144 a. Black-on-red ware jug.g b. Bichrome ware amphora.g ..................................145 End and profile view of a base ring ware jug with spiral decoration over a highly curved surface. ..................................................................................... 146 a. Stamped motifs on a Cypriot diadem.b b and c. Possible freehand motifs on a mouthpiece and a diadem.b ...................................................................147 Detail from vessels showing spiral plant representations.g .............................. 147 Mathematical template spirals and their spectral density plots .................. 151-2 Concave triangular spiral CSBM15.2. ................................................................ 154 Form correlation spirals and spectra. a. JKAN13.1. b. JKAN2.1. ....................... 157 Discriminant scatter plot for the Jōmon case study. .........................................157 Spiral JKAE19.1 and its density spectrum. ......................................................... 159 Jōmon chronological period discrimination scatter plot...................................159 Spiral CSBM15.1 and its spectrum density plot. ...............................................161 Differentiating spirals for the Shang case study ............................................ 161-2 Discriminant scatter plot for the Shang case study ...........................................162 Shang Dynasty/Western Zhou Period orientation bimodality histogram. ........164 Differentiating spirals for the Egyptian case study ...........................................165 Discriminant scatter plot for the Egyptian case study ......................................166

12 5.13 5.14 5.15 5.16 5.17 5.18 5.19 5.20 5.21 5.22 5.23 5.24 5.25 5.26 5.27 5.28 5.29 5.30 5.31 5.32 5.33 6.1

Differentiating spirals for the Cypriot case study .......................................... 167-8 Discriminant scatter plot for the Cypriot case study .........................................168 Spiral CYBM39.4 and its spectral density plot ...................................................170 Distribution of Cypriot spirals according to manufacturing technique .............171 Spiral EGPET8.2 and its density spectrum. ........................................................ 172 Cross-cultural comparison discriminant scatter plot ........................................173 Distribution of rotational lengths by case study. ..............................................175 Median path width and spacing distributions between cultures. ....................176 Distribution of curvilinearity values by culture. ................................................178 Distribution of irregularity of EVP and decorative spirals .................................180 Form correlation spirals and density spectra for EVP assessment ...................181 Discriminant scatter plot for EVP spirals ........................................................... 182 Distribution of rotational length for EVP and representational spirals ............184 Spiral JKAE15.2 and its spectrum density. ......................................................... 186 Discriminant analysis of representational and non-representational spirals ...187 Discriminant variables for construction technique ...........................................190 Discriminant scatter plot for spiral construction technique ............................. 191 Variability in the interpretation of path width ..................................................193 Ambiguities in recording curvature. ..................................................................196 Path width trend assessment for spiral EGLIV7.4 .............................................197 Contrast of angular and curvilinear spiral signal density profiles. ....................200 Spiral and spiraloid motifs on vessels with high curvature: a. Naqada II period vessel. b. Banshan culture burial urn. ............................................................. 209 6.2 Distortion of a logarithmic spiral on cylinders and spheres of increasing curvature. ........................................................................................................210 A1.1 Symmetrical polygons: a. test motif 166, b. test motif 168, c. test motif 167, d. test motif 56, e. test motif 169. Unclosed shapes: f. test motif 54, g. test motif 55, h. test motif 29. ...............................................................................246 a – © Mark Herr b – © The Trustees of the British Museum c – © Victoria and Albert Museum, London d – © ULITA, an Archive of International Textiles e – © Petrie Museum of Egyptian Archaeology, University College London f – © Metropolitan Museum of Art (www.metmuseum.org) g – © Medelhavsmuseet, Stockholm. Photo Ove Kaneberg and Medelhavsmuseet. h – © National Museums Scotland i – © Royal Albert Memorial Museum, Exeter

13

1 Introduction The creation of decoration requires effort beyond what is essential for function at the most basic level and, over a prolonged period of time, social groups have been willing to support those engaged in the production of decoration (Eyre 1997, p.99). It must be assumed, therefore that the creation of decoration has an important purpose within human society. The ornamentation by humans of themselves and their environment may date back 350,000-400,000 years; scratch marks on an elephant tibia from Bilzingsleben, Germany of this date do not resemble butchery marks although their meaning cannot be known (Scarre 2005, p.118). Decoration can be applied to functional objects and cover a broad range of expressions from those which take minimal skill and little or no material cost and may have been intended to be ephemeral to creations using expensive or difficult to obtain materials; requiring substantial skill, specialisation and time investment; and intended as lasting statements, for example glorifying a deity or leader. Because of this diversity, the spread of decoration throughout society and the chances of its survival and retrieval are liable to be greater than for works of art.

The deliberate application of design rules underlying the production of decoration can be identified in many cultures (Washburn 1983a, p.4). Examples of these design rules can be seen in the non-random distributions of symmetrical transformations that have been identified in Japanese textiles (Hann 1992); in the standardised arrangement of design elements in relation to each other in Hopi pottery (Bishop, Canouts, De Atley, Qöyawayma & Aikins 1988) and in the set of transformational rules of symmetry, scaling, curvature and figure-ground configuration that govern the structure of kapkap motifs in New Ireland (Were 2003). Structure and patterning present in art and decoration can be replicated in other aspects of the culture such as settlement and house layout and language Washburn (1983a, p.4).

If the principles governing the structure of decoration, whether applied consciously or unconsciously, can be determined, they have the potential to shed light on the producing culture. They may, for example, highlight important symbolic meanings reflecting religious beliefs (Washburn 1983a, p.4); social structure (Fischer 1961);

14 perceptual and aesthetic preferences (Nisbett & Miyamoto 2005) and the phenomena which the society considered worthy of representation (Fischer 1961). Jablan (2002, pp.292-294), in his study of modularity in decoration, observed that some simple prototiles are very common across unrelated cultures. Although noting that the mathematical description of a pattern does not necessarily reflect how it was constructed or explain the origin of a design, Jablan suggested that the occurrence of the same prototile in different cultures could be used to study: the development of geometric designs; cultural interaction; and mathematical and technological knowledge. Decoration sets out a clear and static presentation of the regularities and patterning which may apply across more ephemeral aspects of a culture. As such, these patterns warrant identification and the analysis of their properties in geometric terms offers a useful tool for identifying and recording these patterns.

1.1 The analysis of design The analysis of decoration has been put to a diverse range of uses including: the replication of historic designs (e.g. Bain 1951) encouraging the development of new designs (e.g. Jones 1856); in archaeological and ethnological studies (e.g. Leroi-Gourhan 1968; Washburn & Crowe 2004); and perceptual psychological studies (e.g. Boselie 1984; Vallentin & Nieder 2008). Informal design appraisal, in the sense of assessment and improvement or innovation, must have occurred for as long as decoration has been applied to objects. Sieveking (1971) cites examples of the transfer of decorative styles between cultures and the modification and incorporation of new motifs into existing design styles at least as far back as 11,000BCE.

The formalised recording of design analysis is more recent. To take European examples, the earliest known treatise that includes discussion of decorative techniques as part of architectural analysis is Vitruvius’ work De architectura libri decem written around 33-22BCE. Systematic design analysis was established by the second half of the fifteenth century (Frings 2002; Andrey & Galli 2004). At this period, the deconstruction of decoration into simpler geometric forms was mainly found in guides for creating decoration, usually in the context of architecture or engineering. Dürer devoted one volume of his 1525 work Die Underweisung der Messung to the construction of curves including logarithmic and Archimedes' spirals (Strauss 1977, p.7). It is unclear to what

15 extent publications relating to abstract geometry may have had an influence on decorative work. It seems very probable that many formalised systems for setting out or constructing decoration were unrecorded due to: the illiteracy of those creating the decoration; apprenticeship systems providing no need to record working methods which could be taught directly; and a desire to preserve trade secrets. Equally, many decorative productions may have been most strongly influenced in their structure by the nature of the materials and tools. Although the decorative product may be susceptible to retrospective geometric description, its production could be purely pragmatic.

The study of decorative arts to identify cultural change and transmission can be dated back to the late-nineteenth century anthropological study of 'primitive' societies with a view to proposing qualitatively assessed sequential developments of decorative forms within a culture. The concept proposed being that these societies represented early stages in an evolutionary perception of art running through Classical Antiquity to the Renaissance then to a highest state of development in contemporary Western society (Morphy & Perkins 2009, pp.2-4). Similar principles of the creation of sequences of design development were used to identify commonalities of decoration between cultures and propose routes of contact and cultural transmission (Morphy & Perkins 2009, p.5).

Early quantitative analyses of decorative design, which aimed to typify its characteristics from a psychological, ethnographic or theoretical aesthetic point of view focussed particularly on the underlying structure of the decorative design. Brainerd (1942) adopted the approach of using symmetry classification as a means of differentiating between the typical design styles of cultures and of identifying the copying of the motifs of one culture by another. The symmetry classification groups created by Brainerd allowed him to show a clear distinction between pottery types, from Chichen Itza and Rainbow Bridge Monument Valley, Arizona. Many experimentally-based studies have been carried out using sets of controlled motifs to assess cross-cultural differences in preferences for proportioning systems. Pittard, Ewing and Jevons (2007), for example, studied preferences for golden proportions in logos in Singapore, South Africa and Australia and found no significant difference

16 between inclinations to favour golden proportions. In contrast to lab-based crosscultural comparison of proportioning systems, studies of proportions of pre-existing artwork or ornament have tended to take the decorative schemes of individual cultures in isolation, or groups of cultures where a strong cross-cultural link is believed to have existed, to look for evidence of consistency in proportioning systems. Golden section proportions have been a particular focus of this type of study (e.g. Hambidge 1920; Tons Brunes (discussed by Kappraff 2000); Lynch & Hathaway 1993).

In design analyses, geometric constructs such as symmetry which affect the whole composition tend to be considered rather than specific motifs (e.g. Washburn & Crowe 1988). In the study of rock art, the relative positions of motifs to each other and their placement on the rock surface have been studied but these have often derived the significance of arrangement from the contours of the rock surface and the larger structure of the monument (O'Sullivan 1986) rather than an apparent formalised geometric scheme. Analyses of motifs have tended to focus on examples that are linked to particular cultures such as Islamic star tilings (Lee 1987), Celtic knot patterns (Kaplan & Cohen 2003) or Japanese abstracted fern patterning (Fuchigami 2001). In these examples, the decoration has very clear mathematical analogues and characterisation was more closely linked to re-creation or the potential for producing computer-generated variation on traditional designs rather than finding commonalities or differences between cultures.

Several evolutionary studies of decoration used spiral motifs as examples in qualitative studies of cultural interaction seen in decorative arts (e.g. Goodyear 1893; Mackenzie 1918, pp.248-250). These evolutionary studies used aesthetic judgements of the whole motif to identify commonalities and developments. The concept of analysing and characterising spiral forms in isolation from an architectural or design class and according to a set of geometric properties was influenced by the work of Theodore Cook (1914, Ch.15). Cook was primarily interested in spirals in biological forms but also noted commonalities between natural and designed forms. Cook's study was unusual in regarding mathematical analysis and characterisation of spirals as an end in itself as opposed to functioning as a guide to the synthesis of new designs. More recent studies of spiral motifs in decoration have drawn links to mathematical forms.

17 Work by Zhushchikhovskaya and Danilova (2008) used the arrangement of motifs and analogy to mathematical forms to assess three case studies: the Jōmon culture in Japan (13,600-900BCE), the Yangshao culture in eastern China (4000-2000BCE), and the Neolithic cultures of the Lower Amur River basin in far eastern Russia (mid 6th to mid 2nd millennium BCE) to identify characteristic regional patterns in the spirals. In the Jōmon case study, variation over time was also identified. Mamaní and Fernández Distel (2010, p.41) also drew parallels between geometric spirals and pre-Colombian petroglyphs and painted pottery decoration although they did not create class groups on this basis. Andrey and Galli (2004) and Takaki and Ueda (2007) used analogy to mathematical spirals most directly in measuring radii or curvature to fit mathematical spiral forms to decorative spirals.

1.2 Spirals in nature The frequency of the occurrence of spirals in natural forms suggests that they convey an evolutionary advantage. There are common themes in the occurrence of natural spirals in grasping, protection and in efficient packing.

Animal horns show particularly complex spiralling structures, exhibiting logarithmic expansion of the horn and the spacing between twists as well as showing a spiral twist along the length of the horn (Cook 1914, ch.12). In plants, tendrils and stems of climbing plants exhibit helical and spiral forms (Figure 1.1a). In animals, spiralling grip is observed in cephalopod tentacles; in seahorses as a means of anchoring themselves; in the tails of climbing animals such as howler monkeys and chameleons; and in the coiling of chameleon tongues and elephant trunks as a means of gathering food.

Spiralling as protection allows linear structures such as butterfly and moth probosces (Figure 1.1b) to be retracted when not in use to minimise exposure to damage. In snakes and arthropods, spiral coiling as a response to threats allows the vulnerable underside of the animal to be protected. Similarly, the coiling of plant shoots (Figure 1.2a) allows the protection of the vulnerable shoot tip whilst it is growing. The coiling seed pods of some species of Medicago form spherical spirals (Figure 1.1c) which are believed to help in protecting the seeds from being eaten by insects (Small & Brookes 1984).

18

Spirals as a function of close packing occur in molluscs, spiral shells provide an efficient packing structure in a manner which can expand with the growth of the organism (Amodeo, Oliverio, Versacci & Marino 2012). Exponential expansion in logarithmic spirals as a growth patterns in flowers spiral whorls allows large numbers of seeds to be contained in a small area. Similar spiral whorl structures can also be seen in the leaf arrangement of some succulent plants (Figure 1.1d).

Other organic spiral forms provide mechanical advantage for example, to the distribution of wind-blown seeds (Cook 1914, pp.137-140) and the mechanical propulsion of seeds. Some spirals constructions in organisms appear to convey no evolutionary advantage that is dependent on the shape of the spiral such as the spirals in the tail feathers of mallard drakes. Possibly as intimidation or warning, some ringneck snake species coil their tails when threatened (Figure 1.2c) (Greene 1973).

Because many natural forms have evolved due to functional advantage, spirals occurring in the same natural context are often very similar in structure; horns and shells have expansive forms with or without spacing between the coils; most plant spirals show logarithmic growth both in the width of the spiral and the gap between coils. Spirals in contexts evolved for gripping tend to show a lower expansion rate and lower rotational length than solid protective spirals. These properties allow greater flexibility than would be possible with a broad base structure and maximise efficiency by keeping the need for coiling and uncoiling to a minimum. With reference to efficiency, plant tendrils which remain fixed once an attachment has been formed tend to have a longer rotational length than grasping spirals in animals which remain active. Consistency can even be seen in organic spirals with no functional purpose insofar as that these structures often conform to logarithmic growth patterns.

Naturally occurring inorganic spiral forms tend to have a low rotational length and so appear less obviously spiral than many biological spiral forms. Some of the most pronounced examples of inorganic spirals occur at a microscopic or telescopic level which demonstrate their existence in principle but would not have been accessible as an inspiration for decoration within the period of the case studies used in this

19 research. Physical spiral forms that would have been visible occur in cracking during desiccation (Lazarus & Pauchard 2011) and vortices in fluid, clouds, dust devils and cyclones which would have been recognisable as very substantial spiral forms (Kappraff 1992, pp. 8-16).

a. b. c. d. Figure 1.1 Natural spiral examples: a. vetch tendril. b. burnet moth proboscis. c. Medicago sp. seed. d. cactus leaf rosette The occurrence of spirals in natural forms has undoubtedly influenced its representation in the decorative arts (Takaki and Ueda 2007; Coutil 1916, p.385). Shells have been modified into decorative or functional ceremonial items (Figure 1.2e) (Starzecka, Neich & Pendergrast 2010, p.67) or have been represented as sculpted forms (Masayuki 2009. p.150). The identification of conventionalised planar spiral representations of shells is less common in prehistoric art but can certainly be seen in more recent decoration (Figure 1.2f). Herbivores with spiralling horns would long have been of interest to humans in hunting or through domestication and the spiralling horns are a very striking feature of these animals which would be likely to remain in conventionalised depiction. The killing of the animals would also have made the horns available for use as functional items and there is evidence for their association with ritual both as vessels and musical instruments (Splitter 1952; Adler 1893) although the practical requirements of these generally means that these examples show little or no spiral rotation; depictions of horns as spirals tend to show greater rotational length (see examples in Table A2.16, Appendix 2). This extension of spirals is seen in other decorative instances for example in printing stamps and weaving from South America depicting monkey tails (Figure 1.2g). Snakes are a very common spiral representation (Figure 1.2d); they are known, or their presence suggested, as representational motifs in all of the case studies used in this research. Although they are conventionally represented as spiral, relatively few species of snake appear to assume organised coiled spiral forms so it is noteworthy that they are apparently more often represented

20 as spirals than in forms that more closely resemble natural sinuous configurations. Other decorative representational examples are found that do not recreate the actual structure of the animal. Crocodiles are depicted with extended coiled tails on Asanti goldweights (Figure 1.2h). In both cases the coiling of an extended form in decoration could be an aesthetic decision or a practical consideration to make it more noticeable or avoid damage.

Maori woodcarving and tattooing feature spirals prominently. The most frequent Maori spirals in painting are associated with uncurling fern fronds and forms inspired by fern shoots are seen in Maori wood carving (Figure 1.2a and b) (Te Rangi Hiroa 1949. pp.316, 319-20). The other common representation of plants with spirals is in the grouping of spirals to represent the stem structure itself. Clusters of spirals provide an effective way of representing the fractal forms often seen in plant growth. Coiling plant forms in particular, are often represented as helices due to their spiralling growth structure.

21

a.

b.

c.

d.

e.

f.

g. h. Figure 1.2 Decoration depicting or inspired by naturally occurring biological spirals a. fern shoot, b. Maori spiral carving, Leeds City Museum, acc.no.FL.1985.1.294. c. regal ringneck snake d. faience snake ornament from Late Period Egyptian coffin, BM, acc. no. EA30453. e. Indian C19th-20th conch shell modified for use as a trumpet, BM acc.no. AS1972, Q.135 f. Edward Lear illustration, V&A, acc.no. E.871-1951. g. Aztec printing stamp depicting a spider monkey BM acc. no. AM1946,04.1. h. Asanti gold weight depicting crocodiles with a curled tails, Liverpool World Museum. Spirals in Maori decorative arts also appear on figurines depicting tattoos and highlighting joints and the contours of the body (Figure 1.3a) (Te Rangi Hiroa pp.315-

22 316). Spirals on joints also appear on Japanese prehistoric human figurines and Shang dynasty animal and bird figures. The depiction of spirals on the abdomens of pregnant figurines and on breasts and testicles could also be linked to contour representation although Gimbutas (1989, pp. 143, 282-283) links the use of spirals in these areas to an association between spirals and fertility (see section 1.3 below). Spirals on human figures have also been used to represent ears and hair (Figures 1.3b and c) (Coutil 1916, p.474) where the derivation from the natural form is more direct. Few elliptical spirals occur in natural forms but fingerprints do exhibit these forms and may have been noticed in a decorative context in the creation of handprints in pigment, although equally, elliptical spirals are uncommon as decorative forms.

a. c. b. Figure 1.3 Spiral forms on humans: a. Maori figure, Leeds City Museum. b. South American figure, Cambridge Museum of Archaeology and Anthropology. c. Babylonian relief figure, Ashmolean Museum. The interpretation of early prehistoric spiral decoration as having been inspired by inorganic natural forms is much less common than suggestions of a biological source. However, more recent art and decoration shows that spirals have been derived from or used in the depiction of physical phenomena. Historic Chinese decoration uses spirals to depict water and cloud effects (Figure 1.4a). Kappraff (1992, p.15) suggests that some examples of running spirals on Papuan shields are inspired by natural vortex forms. Whirlpools and, more recently, spiral galaxies have been represented in logos. As with organic representations, some representational depictions of natural inorganic forms have been conventionalised as spirals although their natural form is not spiralling. Figure 1.4b shows a conventional depiction of lightning in early-twentiethcentury Japanese decoration.

23

a.

b. Figure 1.4 Physical phenomena represented as spirals: a. Qing Dynasty panel with spirals in clouds and waves, ULITA, acc. no. 137. b. Japanese c.20th century textile printing stencil with conventionalised lightning depiction, ULITA, acc.no.189. Two common themes emerge in natural phenomena depicted as spirals - plant shoots and new growth, and representations of threats. Bando and Hirabayashi (2004) suggested that the human visual system is particularly tuned to detect spirals because they are associated with natural forms which are potentially threatening to humans such as snakes and whirlpools. Spirals in spider webs might be added to these threat stimuli. These dangerous stimuli will provoke a strong response and may make their depiction in decoration more likely, as can be seen in the invocation of snakes as protective symbols discussed below.

An additional type of natural spiral which is believed by some to have been depicted in decoration is those generated within the visual system without objective external stimulus. Geometric motifs as internally generated structures can arise from illness, drug induced disturbance of brain function or even simply semi-wakefulness (Bressloff, Cowan, Golubitsky, Thomas & Wiener 2001, Ermentrout & Cowan 1979). Depictions of these forms have been proposed in Palaeolithic rock art (e.g. Bednarik (1986); LewisWilliams and Dowson (1988 and 1993)) and contemporary examples of the

24 representation of hallucination seen in trance states have been cited in ethnographic studies (Siegel and Jarvik 1975, p.138; Reichel-Dolmatoff 1978). The possible origins and mechanisms of geometric visions are discussed further in Chapter 2.

1.3 Symbolic meanings of spirals in decorative arts The widespread nature of spirals in the decorative arts has led to a wide range of symbolic meanings being attributed to them; many of these are unique to a particular group or culture. The known symbolic meanings of spirals associated with the individual case studies are discussed in Chapter 4. The symbolic interpretations below, as representations of the earth, fertility and passing time, have been assigned to several cultures with no apparent link between them.

Two recurring, and apparently linked, interpretations are of spirals as symbols of the earth and of fertility and rebirth. A link might be drawn between spirals as symbols of fertility and their occurrence in the shoots of plants. The Hopi people use spirals to represent ‘Mother Earth’, Tapu’at (mother and child) or, rebirth (Doczi 2005 p.25). Doczi (2005 p.27) also suggested that spirals with two centrally linked coils, such as those in the Neolithic passage tomb at Newgrange, Ireland (3200BCE) represent death and rebirth – one spiral winding into the centre leading into another which winds out again. Petrie (1920, p.17) was highly critical of a similar interpretation applied to Egyptian spiral decoration, particularly in relation to linked spirals. Gimbutas (1989, pp.58-59, 121) linked spirals in early European art to snakes as a manifestation of a creative force.

The conceptual interpretation of a spiral as a sun symbol representing the path of the sun over the course of a day or year has been suggested as an alternative explanation for the carvings at Newgrange (Cook date, p.275) and for Southwest American petroglyphs produced by the Zia although Zeilik (1985) notes that this interpretation of spirals is unusual. Other interpretations of Zia spiral petroglyphs are as symbols of water or wind (Zeilik 1985). Spiral motifs on Postclassic Aztec spindle whorls have been interpreted as representing the passage of the sun during the day or night depending on their angular or curvilinear form or to seasonal transition (Brumfiel

25 2007). On a similar principle, Gimbutas (1989, p.283) linked spirals to lunar symbols on prehistoric Balkan pottery. In a broader sense of the passing of time, Frolov (1983) suggested that numbers of dots in the Mal'ta spiral plaque represent a division of the disk into winter and summer seasons and draws a comparison with nineteenth-century spiral calendars used in Siberia.

These contradictory interpretations of motifs, when applied to cultures for which no written record exists have been much disputed. Even when the producers of the motifs can be consulted, it can be difficult to identify symbolic meanings. Neich (1993, p.36) observed that although the representational origins of Maori decorative motifs could be identified with relative ease by speaking to the producers and consumers of the motifs, symbolic meanings were often understood unconsciously within the culture and so required a greater understanding of the culture to grasp them. It would be of great interest if similar abstract conceptual meanings could be shown to have been assigned to spirals in cultures which had no contact with each other. It may be that common interpretations between cultures owe more to the cultural understanding of the interpreter than to real commonalities between cultures.

This thesis considers the spiral motifs in isolation from their context in an attempt to identify aesthetic commonalities between cultures which are isolated from each other. Because of the diversity of known meanings applied to spirals, attempting to identify symbolic meaning of the spiral in a prehistoric culture, in isolation from its context, would be inappropriate and the classification method proposed in this thesis considers only the physical shape of the spiral and representational interpretations where they can be securely identified. If applied as a classification system within a single culture it would be most appropriate to use the method alongside separate consideration of symbolic meanings either to discover if shape-based discrimination supported conceptual divisions or as a secondary level of analysis to attempt to subdivide class groups identified on a conceptual or representational level.

1.4 The choice of spiral motifs as a research focus The use of decorative spirals covers a large time span. The earliest example of decoration which includes spirals are shells pierced with holes, probably to be strung,

26 found in a 75,000 year-old layer in Blombos Cave in South Africa (Henshilwood et al. 2004). The earliest securely dated examples of created spirals come from Mal'ta in Siberia (Figure 1.5a) and Mezin in Ukraine (Fig 1.5b) both of which date back to the Upper Palaeolithic (State Hermitage Museum 2003; Iakovleva 2009). Spiral motifs appear to have become much more prolific with the start of the Neolithic period in a broad geographic spread of cultures (Table A2.1, Appendix 2). Coutil’s (1916) extensive survey of spiral ornament focusing mainly on Europe from the Palaeolithic period to the Iron Age shows how frequent a motif spirals were across this area and highlights the range of contexts in which they appear.

b. a. Figure 1.5 Early spiral decoration: a. Pierced plaque with spiral motifs from Mal’ta, Siberia c.21000-17000BCE, State Hermitage Museum. b spiral patterning on a tusk and an ivory bracelet from Mezin, Ukraine (adapted from Iakovleva (2009) figs. 12 and 37) Spiral motifs were originally chosen as the focus of this thesis on the assumption that spirals could be precisely described using their mathematical equation as a means of creating a quantitative comparison of shape but, as review of the variety of spiral decoration progressed, it was found that mathematical descriptions were inadequate to fully, or accurately describe the shapes of decorative spirals. This is not necessarily a drawback as combining geometric and non-geometric properties in a single system has proved one of the most interesting aspects of this research.

It has been suggested that some stages of visual processing may be especially well suited to the detection of spiral shapes (Bando & Hirabayashi 2004; Ermentrout & Cowan 1979; Tass 1997) and that spiral forms have an innate aesthetic appeal. Hogarth (1753) attributed the appeal of spiral forms to their capacity to depict the volume of a form. Spehar, Clifford, Newell and Taylor (2003) suggested that logarithmic spirals occurring in the decorative arts of several cultures may be linked to a human liking for patterns containing fractals, and possibly a particular preference for fractal patterns with similar fractal dimension to those commonly found in nature –

27 within the range 1.3 to 1.5. More broadly, Pittard, Ewing and Jevons (2003), in their study of logo preferences, suggested that humans exhibit an innate preference for natural organic forms over non-natural shapes.

An advantage in considering geometric motifs in general in cross-cultural studies is their consistency between cultures. Washburn (1983a, p.2) noted that representational depictions are not exact depictions of the object under study but of the perception of the object and that the cultural background of the person creating the image may encourage certain aspects to be emphasised or minimised. It can therefore be difficult to find a set of variables which apply universally in the analysis of motifs according to their representational meaning. In terms of representation or symbolism, apparently dissimilar objects may have similar intent (Grabar 1992, pp.2401) so the focus on decoration as geometric constructs about which certain fundamental assumptions can be made makes it more readily comparable than motifs linked by an abstract concept. Figure 1.6 shows examples of birds depicted in three cultures with contemporaneous spiral motifs from each culture highlighting the greater consistency between spirals than between the bird representations.

a.

b.

c.

Figure 1.6 Representationally or geometrically similar motifs: a. Iznik, Turkey 16th century CE, BM, acc.no. G.165. b. British 7th century CE, BM, acc. nos. 1908,0217.1 and 1939,1010.110. c. Peruvian 1st century BCE-6th century CE, BM, Am1954,05.464 and Am 1954,05.568. Geometric patterning plays an important or dominant role in the decorative schemes of some cultures (Roe 2004, p.239). Despite this, geometric decorative motifs have

28 been the focus of less research than representational decoration. Washburn (1995, p.117) commented that symbolic meanings tend to be more often ascribed to, and more often studied in, representational decoration than geometric motifs. It was apparent during the literature review and data collection for this thesis that, where geometric decoration with no known representational meaning appeared on an object alongside representational decoration, geometric motifs were often not mentioned in catalogue descriptions of museum objects. Where spirals are a prominent motif in the decorative canon of a culture, for example in south-western Pueblo pottery decoration and the Bronze Age Aegean and Egypt, their development and transmission as a motif has been the focus of research (e.g. Barber 1998) and spirals have played a prominent role in qualitative classification systems of decoration within these cultures (e.g. Tufnell 1975; Jernigan 1986). Quantitative assessment of geometric motifs has, however been minimal even for cultures with a significant presence of spiral decoration and so there was potential and value in developing a new quantitative analysis method for spirals.

It is clear in a number of substantially separated cultures that spirals were, and continue to be, important decorative motifs even when allowance is made for the potentially subjective limits of what constitutes a spiral in decoration (discussed in Chapter 2). Of importance in this research, engaging in cross-cultural comparison, is the range of possible inspirations for the presence of spirals in decoration from natural objective and subjective phenomena. Also of the practical constraints imposed by the properties of the tools and materials used to construct the decoration. The variety of possible explanations for the production of spiral decoration offers the potential for a broad range of expressions of decorative variables within spiral motifs and also for identifiable consistencies running across cultural boundaries. These factors make spiral motifs well suited to testing whether the quantitative comparison of geometric motifs in isolation from their decorative context can be a useful tool in cross-cultural studies.

1.5 Overview of the thesis The methodological basis of this thesis focusses on producing detailed quantitative descriptions of a single type of geometric decorative motif and using these as a basis

29 for comparisons within and between cultures. Decorative spirals are, regardless of modification and elaboration, underlain by a universal geometric framework which provides potential for analysis within a more generic system. The method takes into account decorative as well as geometric aspects of spirals. Such a classification method has applications, as a non-culture specific method.

The principle that a generic classification system based on geometric properties could be a useful method of cultural comparison draws on Dorothy Washburn's statement that “it should be possible to define art systems by describing the consistencies in their formal organization. However, too often the classifications are too culturally specific to allow cross-cultural comparisons or their units cannot be easily used for diachronic studies of change...” Washburn (1983a, p.3) Taking this principle, Washburn and Crowe (1988) proposed a symmetry classification system which has been applied to archaeological and ethnographic studies including temporal change and regional variation in the Neolithic Aegean (Washburn 1983b), the aesthetic properties governing Bakuba raffia cloth type groups (Washburn 1990). The method of analysis and classification proposed in this research is intended to be generic and tries as far as possible to ignore any culturally individual meanings. Complete disregard for cultural meaning is, in itself impractical or impossible to implement completely since ‘geometric decoration’, and even geometry itself, is not a culturally neutral concept. Fukagawa and Rothman (2008, p.2) note that, although the fundamental principles of Japanese geometry and that developed from a Classical Greek origin are the same, the manner of expressing them and the types of problems posed are substantially different.

In many ways, the analysis method proposed in this thesis takes a similar approach to that of Zhushchikhovskaya and Danilova (2008) in applying a quantitative analysis to spiral motifs. It differs in trying to draw away from the application of named geometrical shapes which may not have been used or known by the culture in question and, instead, using statistical grouping methods to attempt to build class groups from the measured variation of the decorative spirals themselves. This method of assigning groups does not necessarily identify the type groups which would be

30 assigned by the makers of the decoration any better than mathematical spiral classes. However, it may be more flexible than externally defined classes in allowing real aesthetic differences in the decoration to guide the number and nature of the classes generated and provides a tool for consistent comparison rather than a means of creating classes which can reliably be assumed to have cultural meaning.

A positivist approach has been chosen to allow the broad scope of potential applications to different cultures although it is clear that this will come at the expense of some depth of interpretation. In the nature of analysing motifs primarily by their physical shape, symbolic, emotive and, to some extent, representational interpretations are excluded. To suggest the reasons for the production of the decoration under analysis and to use the decorative motifs as a means of studying social structure of societies, a broader consideration of the context of the decoration would be needed. For the purposes of this study the decontextualized approach is felt to be appropriate to the aim of designing a generic analysis method which, as with Washburn and Crowe's applications, may be given a more nuanced interpretation in the light of context provided by individual case studies to which it is applied.

1.5.1 Research questions The view taken in this research is that analysis of decoration may provide a means of identifying perceptual and aesthetic similarities within and between cultures which reflect an origin of the motifs in naturally occurring spiral forms. A culturally invariant, if not neutral, analysis system was required to test the following questions.

1.5.1.1 Are comparisons of decorative spirals with mathematical spiral types useful as a classification tool? The first research question posed relates to the functioning of the classification method proposed in this thesis. As discussed above, existing classification systems tend to use mathematical spiral type classes. Decoration analysed by this approach often appeared, subjectively, to show poor conformity to the mathematical class to which it was assigned or to be equally well classifiable to two or more classes.

31

0

a

2

4

6

8 10 12

0

b

10

20

30

c

e d f Figure 1.7 Mathematical spiral types. a. Archimedes’ spiral, b. logarithmic spiral, c. hyperbolic spiral, d. parabolic spirals, e. lituus, f. Euler spiral. Andrey and Galli (2004) used radial measurements as a function of cumulative spiral length to test ruler-and-compass based construction methods proposed in the Renaissance for constructing Ionic volutes for conformity to mathematical spirals. They found that three of the four spiral constructions they analysed which were composed of segments of a circle did not conform sufficiently to any mathematical spiral curve to make any valid identification of a spiral type. The fourth spiral construction showed the gradual change which is typical of a spiral so could be tested for conformity to a mathematical type. However, attempts to fit a spiral equation to this curve showed that it conformed equally well to several mathematical spiral types. In contrast, an exploratory study Takaki and Ueda (2008) used cumulative path length and the radius of curvature to compare decorative and mathematical spirals. They attempted to match the decorative spirals to a general mathematical type rather than identifying the equation of the curve. Takaki and Ueda felt that there was sufficient conformity to mathematical spiral types to propose that decorative spirals tended to fit into one of three mathematical classes although they noted that some spirals fell between type classes and that further analysis with a larger data sample might suggest an alternative division of type classes. The natural occurrence of spirals which closely conform to logarithmic curves (Figure 1.7b) as a possible source of inspiration for decoration (Kappraff 1992) and the potential for some methods of decorative spiral construction to show close conformity

32 to Archimedes' spirals (Figure 1.7a) made the hypothesis that conformity to these mathematical forms would be found in decorative forms worth testing. Besides Archimedes' and logarithmic spirals, four other types of spiral curve - Euler, hyperbolic and parabolic spirals and lituus (Figure 1.7c-f) have also been identified by visual assessment of decorative spirals. Euler, hyperbolic and parabolic spirals were also included as test templates. The reversal of curvature of the lituus is excluded from consideration in this method and, without this feature the lituus was felt to be too similar to the Euler spiral to be included as a separate form. The comparison method employed used the same standard of conformity as Takaki and Ueda, searching for trends of conformity to mathematical spiral types rather than exact or very close match of individual curves.

1.5.1.2 Does the quantification of decorative spiral motifs in isolation create meaningful intra-cultural type groups? It is not the intention of this thesis to suggest that any single geometric decorative motif type can reliably be used in isolation from its context to assign an object to a particular regional or temporal cultural group. There is no need to make such an attempt when further useful contextual data will always be available, at the very least in the form of material and technique. Also, the potential for coincidental similarities between the motifs on a single object and the range of decoration within several different cultures is too great to make this type of classification worth attempting.

It was however, considered to be worth testing whether spiral motifs considered as groups displayed consistent characteristics which allowed a sample to be subdivided accurately by period within a culture or into separate cultural groups from a crosscultural sample. If divisions of this sort could be obtained there might be potential to identify perceptual preferences in a culture, although the effect of medium and technique on spiral shape would also need to be considered. If strong differentiation between cultures were found it would also reflect on research questions which assess the presence of cross-cultural similarities.

33

1.5.1.3 Is there consistency across cultures in the shape of spirals according to representational subject? Some spiral decoration can be securely traced back to a representational origin and there is clear evidence of close observation of natural forms in prehistoric representational decoration. Because of this, it is hypothesised, firstly that spiral decoration from different cultures with a similar representational subject may show cross-cultural similarities based on the objective spiral source. Additionally, it may in some cases be possible to suggest an objective spiral source for decorative spirals without a known representational origin. This second hypothesis would rely on strong class groupings and unusual characteristics to be effective. As Layton (1988) commented, in reference to identification of subjective and hallucinated origins for decoration, it is necessary to be able to assess the probability of alternative sources having been the inspiration for the motifs.

1.5.1.4 Is there evidence for the influence of subjective, internally generated, spiral forms on decorative spirals? This question considers the influence of subjective spiral visual effects arising from the structure of the visual cortex on spiral decoration. There are no decorative spirals which are sufficiently well accepted as having been influenced by subjective imagery to make a cross-cultural assessment in the manner employed for objective representational spirals. Before an assessment could be made of influence on decoration the question was posed of whether one, or more, typical subjective spiral shape can be identified based on the analysis of contemporary drawings of these phenomena.

Dominant spiral types identified were then tested against the cultural case study spirals to measure conformity to the subjective types. The analysis of a single motif on its own could not be used to confirm the depiction of subjective spiral forms in decoration. However, a strong conformity to a subjective spiral template could be used to support the suggestion of the depiction of subjective visual phenomena in decoration. Especially so if other geometric decoration which also resembles cortically generated visual effects were present within the culture.

34

1.5.2 Thesis structure The following chapter discusses the definition of spirals and the perceptual basis for the variables used in the description of the motifs within the classification proposed in this thesis. The second theme of Chapter 2 provides a background to the occurrence of geometric visual disturbances arising from the structure of the visual cortex and discusses how these visual effects have been studied and classified as well as briefly reviewing existing studies which have proposed or identified the decorative or artistic representation of these phenomena in ethnological or archaeological contexts.

The method, developed in this research, of classifying and analysing spirals is detailed in Chapter 3. This chapter discusses the problems associated with initial data processing in ensuring that it could cope with the variation found in decorative spirals in a consistent and aesthetically appropriate manner. As far as possible, rules needed to be set down for the identification of the start and end points of the spiral and which part of the spiral to analyse when potential confusion arose between lined spirals and figure-ground configuration. Analysis variables described the basic shape of the spiral, the decorative elaboration of this curve. The measurement of individual variables is described. Finally, this chapter covers the conversion of the raw data into a format suitable for collation and the creation of the class groups using cluster analysis.

To assess the perceptual importance of the variables employed in the classification system a survey was conducted on perceptions of a range of motifs. Participants were asked to mark centre and end points on motifs, to judge the similarity of motifs when one, or more, perceptual variables was altered and to describe spirals and other geometric motifs so that the terminology used could be evaluated. Results from the survey were also used to guide the development of the analysis method and the limits of identification of a spiral. The results of this survey are presented in Appendix 1 and discussed in Chapters 2, 3 and 4 in relation to relevant perceptual variables and considerations in the method.

The classification method proposed in this thesis was tested against two pairs of case studies. The first pair, in East Asia, focusses on the Shang Dynasty further to the east of China and the Neolithic/pre-Neolithic Jōmon culture in Japan. The second pair is

35 centred on the Eastern Mediterranean and focuses on Middle Kingdom Egypt and the Cypriot Bronze Age. Chapter 4 introduces these studies giving a brief background to the structure of the cultures; the main sites which were the source of the material analysed and the occurrence and symbolism of decorative spirals in the culture.

In Chapter 5 the analysis and classification method is evaluated. Improvements that might be made to precision and to the measurement of the irregularity and curvature of spirals are proposed. Alternative methods of classification and statistical analysis are also considered. The findings of the application of the classification system to the case studies and the comparison of the case study data with mathematical and subjective spiral forms are presented. Following this, the results of the four research questions are discussed.

Chapter 6 summarises this research, covering the development of a new analysis method and the results of its application to the case study data. Possible future developments of the method which might allow it to analyse spirals in three dimensions and cover a broader range of unclosed geometric motifs are suggested. An overview is provided of the links that were drawn between classes of spiral motifs based on their originating culture and representational origin. Future work in extending the scope of the case studies and the potential to re-evaluate the presence of subjective spiral influence with an expanded dataset is discussed.

36

2 Definition of terminology and variables and discussion of spirals as subjective visual phenomena. The potential scope for defining a decorative motif as a spiral is broad; as even the few examples shown in Figure 2.1 demonstrate and clear defining boundaries need to be drawn. This chapter discusses existing mathematical definitions of spirals and sets out the criteria used in this thesis for identifying a decorative spiral. Since the contention of this thesis is that mathematical definitions are inappropriate to the description of decorative spirals, the second part of this chapter introduces an alternative set of aesthetically based variables and uses the results of the survey of perceptions of geometric motifs carried out during this research (Appendix 1) alongside literature review to provide a perceptual basis for their inclusion within the classification system as valid measures of difference in decorative forms.

a.

b.

c.

d.

e.

f.

Figure 2.1 Variety in decorative spirals: a. Neolithic Chinese vessel, 2500BCE, British Museum, acc. no. 1966,0223.1. b. lintel board, Trobriand Islands, Cambridge Museum of Archaeology and Anthropology, acc.no. Z10036. c. Greek kylix, c.490BCE, British Museum, acc.no. 1836.0224.25. d. North American jar, Pitt Rivers Museum, acc.no. 1928.9.2. e. jar, Ban Chiang (N.Thailand), 2000-1BCE, British Museum, acc. no. 1972,0919.1. f. Chinese Shang Dynasty bronze vessel, British Museum, acc. no. 1947,0712.429. The second part of the chapter discusses the possibility that there is a basis in the structure of the human visual cortex for spirals to be a preferentially noticed form in the natural environment and the occurrence of spiral visual effects generated within the brain as a potential inspiration for decorative spirals.

37

2.1 The definition of a spiral Definitions of spirals found in the literature tend to operate in terms of the rotation of a line or point around a fixed point (e.g. Thompson 1942, p.748; Sharp 2002, p.59). These definitions exclude angular spirals and key patterns which, in a decorative context, have been accepted as legitimate spiral forms (e.g. Hull 2003, p.102; Dunham 2004). The results of the perceptual survey also supported this definition with 12 participants describing angular motifs as spiral.

In order to include the broad range of spirals found in decorative design, the following definition is proposed for what will be considered a spiral in this research: “A spiral is a line on a plane which winds or bends in a coiling form towards or away from a point without crossing over itself and in such a way that the line would entirely surround the point after travelling through 360 degrees and that the distance from the centre shows a decreasing trend.”.

Generally, the participants in the perceptual survey assessing aesthetic responses to geometric motifs (Appendix 1) showed a low use of the term 'spiral'. Participants were asked to describe the shape of 18 motifs, nine of which would conform to the definition of a spiral given above. For each 'spiral' motif, at least one participant gave a description which included the word spiral or a term felt to be synonymous (terms included 'angular spiral', 'rudimentary spiral', 'coil', 'helix', 'irregular spiral', 'squiggle spiral', 'jagged spiral') but only one motif (Fig. 2.2) which was the most regular of the spirals and the only curvilinear example attracted a high number of descriptions of its shape as a spiral (92%). All the others had spiral identifications of 50% or lower. As the intention of this test was to identify the limits of what could be perceived as spiral, other motifs tested were highly irregular so the low rate of identification as spiral is not surprising.

38

Figure 2.2 High spiral recognition motif (test motif 100). In a second experiment, when asked to give a free description of a different set of motifs, four out of the five motifs tested scored much more highly for spiral identification with 78-100% of participants identifying the motifs as spiral. These motifs (Fig. 2.3a-d) might be considered more 'traditionally' spiral than the test motifs discussed above in that the rotation is greater and the expansion rate is regular. The fifth motif (Fig 2.3e) , which only one participant identified as spiral, has a low rotational length and a high expansion rate, both factors which mitigated against identification of curves as spirals.

a.

b.

c.

d.

e.

Figure 2.3 Pictorial test motifs nos 154, 155, 156, 157, 157b. Theoretically, a curve of any arc length could be tested for conformity to the above definition of a spiral, studies of golden proportioning have identified very short lengths of curve as spiral (e.g. Doczi 2005) but, when asked to identify motif shape by choosing amongst four supplied keywords in the perceptual survey, there was a significant drop (p=0.01) in the number of participants who identify motifs as spirals when less than one full rotation was completed. No significant difference was found in the identification of motifs as spiral between rotations of less than 270 degrees and 270360 degrees or those of 360-450 degrees and >450 degrees. These findings suggest a minimum curvature of 360 degrees to count a motif as spiral.

There is a strong indication in the chronological development of decorative motifs in a number of cultures that spirals occur on a continuum with non-spiral motifs (Fig. 2.4)

39 (Bagley 1996, p.19). The trend to find spirals of greater than and less than one full rotation in apparently similar decorative contexts appears particularly where two or more spiralling forms meet as running spirals or whirls (Figures 2.5 and 2.6) possibly due to the implied rotational motion increasing a sense of rotation length (Lidwell, Holden & Butler 2003, p.34).

a.

b.

Figure 2.4 Similar spiral and non-spiral motifs: a. Papuan drum, linked spiral and nonspiral curves, Leeds City Museum, acc.no. FL.1987.1.1. b. scarab with spiral and nonspiral links, (© Petrie Museum of Egyptology, University College London), acc.no. UC11151. Petrie (1920, pp.20-21) identified spirals in the context of running patterns with linked pairs of spirals where each spiral typically only completes around 270 degrees rotation although this identification was not well supported by the results of the perceptual survey in which the angular non-spiral key pattern motif (test motif 112, Appendix 1) was only identified as spiral by two out of 22 respondents and the curvilinear equivalent motif (test motif 113) by only eight out of 22 respondents (Fig.2.5).

Figure 2.5 Test motifs 112 and 113 representing low-rotation key patterns. Some cultures appear to show evidence of development from non-spiral forms into spirals (Bagley 1996, p.15) so the inclusion of these earlier forms is relevant in considering the development of the motif. The initial intent of this research, to study spirals in complete isolation from other geometric motifs, was therefore shown to be inappropriate and this highlights one of the difficulties of applying constructs from outside a culture which is discussed further in Chapter 4.

40

For the purposes of this study, to allow for the continuum of change, motifs which fulfil the criteria of spirals but are of 270-360 degrees rotation have been included and are described by the term 'spiraloid'1 (below 270 degrees rotation, only one motif (Figure.2.6) was identified as spiral and its perceived rotation may have been enhanced by being part of a whirl structure).

Figure. 2.6 Whirl motif of low rotational length (test motif 110). Although this restriction on rotation will exclude some examples of decoration and of natural forms (such as flower heads) where spirals might be perceived, some limit on rotation is needed within the definition of a spiral for the term to have meaning.

When concentric circles are tightly packed and numerous, or when spirals have a low and constant expansion rate and high rotation, an optical illusion is created which makes the two forms visually readily interchangeable. Finally, in defining the spiral, mention must be made of ‘false spirals’ – groups of circles or concentric circles joined by tangential lines in a manner resembling interlinked spirals (Figure 2.7). To limit the scope of this research, false spirals have been excluded from analysis.

1

Unless otherwise stated, the use of the term spiral in this thesis includes qualifying spiraloids.

41

Figure 2.7a. False spirals on a scarab, Liverpool World Museum acc. no. 56.20.297. b. A spiral resembling concentric circles on a Cypriot barrel jug, British Museum acc. no. 1876,0909.74.

2.1.1 Mathematical spiral forms Mathematical spiral definitions have provided a basis for classification of spiral types in a number of previous studies of decorative spirals (e.g. Takaki & Ueda 2007; Zhushchikhovskaya & Danilova 2008; Banakh, Verbitsky & Vorobets 2010). Key tenets of a mathematical spiral are that the distance of the path from the pole can be related to rotation and that a single equation can be used to describe the full extent of the curve. Six mathematical spirals have been identified for which close equivalents have been found in the decorative arts. Archimedes’ spiral, logarithmic spiral, hyperbolic spiral, parabolic spiral, lituus, and Euler spiral (Figure 1.7, p.31). With the exception of the Archimedes’ spiral, for which the equation was known by the end of the third century BCE (Pickover 2009, p.66), the knowledge of these spirals as mathematical forms is more recent, only dating back to the seventeenth and eighteenth centuries CE (Sharp 2002, Wassenaar 2009, Weisstein 2010).

Assigning decoration to predetermined classes has less relevance than in the classification of objects where the form is related to physical function in the way that, for example, the shape of a ceramic vessel might be constrained by ease of pouring. Function may exist in decoration – invoking gods or indicating status but these functions are not determined by extra-cultural laws. Where function is affected by physical laws, types of object made for a particular purpose are more likely to have defined limits of shape beyond which they cease to function. These would have clearer class divisions than in decoration, which could be on a smooth continuum. There is

42 also less reason to expect commonality in type groups of decoration between cultures due to the lack of restraints; so where commonalities between cultures do exist this point may be used to strengthen arguments for cultural transmission when commonality is detected since an analogous link cannot be argued.

Since some natural spiral forms show apparent close conformity to mathematical spiral types, the use of these categories to classify decorative spirals has a reasonable basis in the possible origins of decorative forms but raises problems with the potential relevance of applying mathematically based classes which may not have been deemed significant. If mathematical classes are taken to accommodate the full range of decoration, there is a risk of pushing decorative spirals into mathematical classes to which they do not conform to such an extent that the use of mathematical terms becomes meaningless and implies a precision which is not present. Assessed from a purely visual standpoint without measurement there are many decorative and natural spirals which might be reasonably assigned to more than one mathematical class and even with measured values, Andrey and Galli (2004) found that construction systems for some volutes might equally well be said to approximate logarithmic or parabolic spirals. Mathematically a straight line could be described using the equation of a logarithmic spiral with an infinite expansion rate. Using ruler and compass-based construction techniques, there is almost bound to be ambiguity since it is not possible to draw a perfect spiral using these tools. The ambiguities that arise from attempting to match irregular decorative curves to mathematical spirals are felt to be one of the key drawbacks in attempting to use mathematical spirals to describe decoration. The variables of the alternative analysis system proposed in this thesis are discussed below.

2.2 Decorative spiral classification An effective classification system needs to take into account the range of variation in forms and styles of decorative spirals or, at the very least, acknowledge and justify any variables which are excluded. An initial set of variables considered for inclusion within the classification system was identified from a review of decorative spirals in museum collections, attempting to cover a diverse range of cultures and subjectively noting the ways in which the spirals encountered appeared to differ.

43 Nine potential variables were identified: overall shape (form); style o curvilinearity; o direction of turn; o orientation; o rotational length; o diameter; o width of lines and gap between lines (path width and spacing); o irregularity; o colour and patterning applied to wide lines. The proposed variables were divided into two types - form and style. The term form is used to describe the gross curvature of the spiral regarded as a notionally infinitely thin path shorn of all decorative attributes. The decorative variables relating to a single spiral were considered to be variables of style. Style is taken to describe all aspects of a single spiral coil which are independent of the underlying curvature.

For the identified potential variables, evidence was sought from the perceptual survey and from literature review of studies of perceived motif similarity and grouping to establish whether they were appropriate for inclusion within the classification system. Evidence that an attribute made a significant difference to assessment of similarity was taken as indicating that the variable was valid in creation of class groupings.

Some criteria for perceptual grouping employed by humans have been observed in other primates (Kayaert, Biederman and Vogels (2005) identified that macaques could differentiate according to irregularity and curvilinearity and rhesus monkeys, tested under laboratory conditions, showed the ability to distinguish proportions of rectangles at an average accuracy of 85.56% (Vallentin & Nieder 2008)). It was believed that the creation of perceptual groups was a low level process so relatively unlikely to be influenced by cultural variation, more recent work has, however, suggested that perceptual grouping is a higher level visual process (Palmer, Neff & Beck 1996). There is contradictory evidence as to whether assessment of constancy (invariance relative to viewing conditions) occurs before or after grouping (Palmer & Rock 1994; Beck 1975). Studies of uniform connectedness (physically linked motifs), and amodal completion (filling in of missing elements) suggest that these processes

44 precede the formation of groups (Palmer & Rock 1994; Palmer et al. 1996). The interaction of perceptual variables and the strength of their expression can affect their perceived importance in grouping (Kayaert & Wagemans 2010). Prior knowledge and learning can also influence the strength of grouping effects (Beck & Palmer 2002). This last aspect, in particular, creates problems with suggesting that a classification based on perceptual grouping can be considered culturally independent. The further a set of classificatory criteria are removed from a primitive physiologically-based interpretation of a visual stimulus, the greater the risk of cross-cultural non-equivalence.

That perceptual grouping judgements do not appear to operate fully at a pre-attentive level and may be influenced by learning does not necessarily imply that they cannot attain wide cross-cultural relevance particularly when dealing with the properties of geometric constructs which are found widely in the natural environment and can be perceived from the structure of the human visual system. It is suggested that underlying principles governing perception are learnt at an early age to an extent where they can operate at a subconscious level (Kunst-Wilson & Zajonc 1980; Pittard et al. 2007). However, Pittard et al. (2007) suggest that some preferences are learnt from natural forms, and that the exposure of different cultures to similar natural forms (such as the proportions of the human body) produces universal learnt preferences although the significance of golden section as a universal preference, which was the focus of Pittard et al.'s study, has been contradicted by other research (van Damme 2000, p.261). Attempts to find universal aesthetic preference values have produced mixed results and the idea of such universal values existing tends now to be considered unlikely in reference to specific variables (van Damme 2000, p.258) There have, however, been sufficient studies which show cross-cultural agreement in preferences (e.g. Berlyne 1975; Pittard et al.2007) to suggest that some basis for agreement might be identified. Van Damme (2000) suggested that universal values might be found in the incentives for preferences, such as variable expressions deemed to indicate good health, but not necessarily in the variables which are considered to exemplify these values. Some contexts in which spirals naturally occur may acquire associations with danger and power (snakes, cracking patterns, aberrant mental imagery) or with new life and potential food sources (plant shoots). Both of these concepts would have had substantial evolutionary importance and there is reasonable

45 evidence to suggest particular sensitivity to them within the visual cortex so it seems credible that spirals may be amongst the natural constructs which have a fundamental influence in early development on decorative preferences. If the basis of criteria for grouping is derived from a common natural input then the variables remain valid even if their expression is changed by other cultural influences as Pittard et al. (2007) results appear to suggest in a preference for 1:1 proportions in abstract logo designs in contrast to a preference for golden proportions in organic based designs which they attribute to a learnt preference from exposure to common brand logos with 1:1 proportions.

The evidence of this review shows that it is necessary to treat the proposed variables as an artificial and incomplete system. Even when a variable can be shown to be significant in perceptual grouping in cross-cultural studies, it is not possible to know their relative importance for the cultures studied or to know if variables, which the cultures in question would have considered more perceptually important, have been missed. The claim for the validity of the variables used in this classification system is, therefore, that they provide a fuller description than one based solely on curve shape and that they measure variation in consistent and quantifiable terms not that they create the same type groupings that the members of the cultures studied would have created. One aim of the perceptual survey was to test whether the potential variables identified for inclusion in the classification system were seen to describe aesthetically important aspects of spiral motifs and whether they were valid differentiators to be used as a basis for creating type classes of motifs. To test variables as a means of differentiating spirals, survey participants were asked to rate the similarity of 53 pairs of spirals on a scale of one to ten (Appendix 1 Tables A1.1 and 2; test motifs 131-153, 158-165, Table A1.6).

2.2.1 Form Of the identified variables, form most closely approximates the spiral classifications based on mathematical groups but because of the deviations in the form of decorative spirals from smooth mathematical curves form and the potential for limitless variation in form, the identification of those variations of form which are expressed in decoration comes to be of aesthetic importance.

46

In the perceptual survey test of similarity, nine of the motif pairs tested form differentiation. Of these, only one pair yielded a significant result. As discussed in Appendix 1, all significant similarity test results returned results indicating high similarity and there are methodological reasons to suggest that the structure of the survey may have skewed results towards high similarity scores. The form similarity test registered as the most important differentiating variable in the perceptual survey with a dissimilarity score of two but, with only a single significant result, the relevance of this result needs further checking. The experiment asking for free description of motifs supported more strongly the aesthetic importance of form - all participants provided at least one form-related descriptor for four of the five test motifs (85% gave a form description for the fifth, Figure 2.3e, p.38).

Washburn (1983a, p.2) asserts the primacy of shape in the identification of objects and the ability of humans to abstract an underlying shape from complex objects. Studies of constancy in the ability to identify objects under rotational transformations have shown that form can be detected in the face of considerable distortion of the retinal image (Jolicoeur & Humphrey 1998) and have suggested that the extraction of an object's form occurs at an early stage in the processing of a visual image. The survey results and the evidence from literature review of the facility of the visual processing system to identify form both support the inclusion of this variable within the classification system.

Two of the variables of style discussed below might, in principle, be considered to fall within the scope of the form variable: irregularity and curvilinearity. These two variables were separated out because of a potential loss of resolution in these variables if considered as part of form arising from abstracting the decorative spiral down to a single line.

2.2.2 Stylistic variations in decorative spirals Stylistic variables are included in the proposed analysis system because decorative spirals, exhibit aesthetic traits which render the description of the spiral form, by itself, insufficient to characterise and differentiate them. The examples shown in Figure 2.8

47 have a form correlation of 0.997 but the stylistic variables create a marked difference between them and in analysis they were eventually grouped in separate type classes on the strength of these differences. The potential variables for the description of style are discussed below.

Figure 2.8 Shang Dynasty spirals with a high form correlation but low style match.

2.2.2.1 Curvilinearity As a stylistic variable, curvilinearity is determined by the number of line segments making up the spiral – a spiral composed of only one line segment is necessarily curvilinear. The group of angular spirals covers those formed from straight line segments meeting at angles rather than continuous curves (Wassenaar 2009). Irregular spirals may combine angular and curvilinear elements as can be seen in the Shang Dynasty examples above.

In the decorative arts, angular spirals are rarely found as individual motifs. They usually form elements of border and all-over patterns such as key (or meander) patterns (Figure 2.9). Their linear structures make them well suited to reproduction of woven textiles and they are a common element of Central and South American woven decoration (Figure.3.15, p.81). The known use of spiral key patterns as a decorative motif dates back 14-15000 years at Mezin in Ukraine (Iakovleva 2009).

48

Figure 2.9 Angular spiral meander pattern on a ninth-century cross, St.Illtyd’s Church, Llantwit Major. Pittard et al. (2007) suggested that angular shapes may have a symbolic significance as representing conflict and that this may affect preferences for such forms across a culture with societies more inclined to conflict-avoidance more likely to favour smoothly curving forms. It seems likely though that they are most often deployed because they can create an all-over covering of a surface without large gaps between either motifs or individual lines. It is only in this context that they appear to be used preferentially to curvilinear forms.

In the perceptual survey, four similarity pairings tested curvilinearity as a differentiator, none of these yielded significant results. Curvilinearity featured more prominently in tests of the aesthetic importance of the variable. In an experiment asking participants to describe the shape of 23 motifs (Appendix 1, Table A1.3), nine of the 31 participants gave responses which included the words 'angular', 'jagged' or 'curvy'. In the free description of motifs, three participants used the terms 'curvy', 'curly' or 'loopy'. These descriptions cannot be given a strength rating as a measure of importance but indicates that the degree of curvilinearity of the motifs under examination was noted and considered to be a valid descriptor.

Kayaert et al. (2005) found that irregular shapes in which contours were composed of entirely curved or entirely straight lines provided a very strong stimulus for differentiation. Regular shapes, some of which were composed of curved and straight sections were less strongly differentiated from the angular set than from the entirely curvilinear set suggesting that the ability to find straight lines within a motif may be a significant grouping factor.

49

Although no conclusive results were obtained from the perceptual survey, the results of literature review were felt to indicate that curvilinearity should be considered a significant differentiating variable and included in the analysis.

2.2.2.2 Orientation Zhushchikhovskaya and Danilova's (2008) study also supports the inclusion of orientation as a variable on cultural grounds since they found cultural commonalities in the vertical or horizontal orientation of double linked spirals.

Orientation was found to be the second most important differentiating variable within the perception survey. Surprisingly, of the three significant test results, spirals with markedly different perpendicular axis lengths (test motif 158A) showed the lowest differentiation based on orientation. The highest differentiation was between groups of spirals with ordered and disordered orientation (test motif 164A) and may relate more to the potential to find symmetrical associations within the clustered elements than to the value of orientation as a differentiator. Considered only on the comparison of individual spirals, orientation was the least important differentiator. The majority of the motifs were presented singly and in a neutral context where the impact of absolute or relative orientation would be likely to be minimised. The only instance in which orientation related terminology was used in description was in relation to test motif 90 which was described by six participants as an upside-down L. Seven participants, however described the motif as having an L-shape without commenting on it being upside-down so had 'corrected' the orientation without noting it as important.

Beck (1966a) found that orientation was a key differentiator in grouping when there was a difference in the angle of the primary axis. Indeed, this registered as a stronger grouping variable than form. However, the value placed on form and orientation as differentiators conflicted according to context and orientation variation which did not show difference in the primary axis was sometimes subsidiary to form. This variable was included in the analysis although, for some objects analysed in the case studies, it was not possible to record orientation since the object itself had no single correct

50 orientation. This meant that approximated values had to be substituted for some objects in order to include them in the analysis as discussed in Chapter 3.

2.2.2.3 Rotational length Beyond the minimum rotational length limit discussed above, the rotational length of spirals showed some of the highest extremes of variation of the style variables in the chosen case studies. Mamaní and Fernández Distel (2003, p.39) suggest that spirals in pre-Colombian decorative arts show sufficient consistency in the number of rotations that this must have had some symbolic significance but no evidence for a known meaning assigned to the number of turns in a decorative spiral has been found. Evidence has been found in concentric circle motifs of consistency in the number of circles produced when representing a particular phenomenon (the use of four of five concentric rings to represent a sacred water hole in Australian Aboriginal groundpaintings (Nicholson & Firnhaber 2003, p.61) though it is not clear that the number of circles is seen as significant.

Within the perceptual survey, this variable did not rank highly as a differentiating variable (Appendix 1, Tables A1.1 and 2). Significance of rotational length was assessed in eight tests, of these only two yielded significant results both with a modal similarity score of nine. No descriptive terms were used in survey responses which had a bearing on the rotational length of the spiral.

A study of the similarity of gestures, represented by the animation of the creation of various linear patterns, found that line length had a low bearing on perceptions of similarity (Long, Landay, Rowe & Michiels 2000). The gestures used to test line length significance in the experiment by Long et al. would, in the terminology of this study, be regarded as spirals and spiraloids so this study offers a relevant comparison to the results of the perceptual survey. It is interesting to note that Dronfield (1996b) used the number of full rotations of a spiral as a measure of its 'intensity' in the analysis of Irish prehistoric tomb rock carvings. In this context, he derived the intensity measure from an estimation of the amount of material which would have needed to be removed to carve the motif as an index of the time taken over the creation of the motif. Although this measure would apply much less strongly for the production of

51 spirals in less labour intensive media, it might reflect the time available for the production of motifs and therefore be used as a means of assessing mass production of objects and their decoration in contexts where each motif has to be produced afresh.

Although rated as a variable of low significance through literature review and survey, the extremes of variation within the case study data was felt to warrant the inclusion of rotational length as a potentially relevant variable in classification.

2.2.2.4 Path width and spacing The width of lines and the gap between them was also found to be highly variable within and between cultures used as case studies. Evidence from literature review supported the view that path width and spacing show considerable variability in decorative expression. Several of the different methods of drawing Ionic volutes vary according to the width of line and gap between lines (Andrey & Galli 2004). As well as recording average widths, these variables measured changes in the trend of width and included decorative embellishments to the ends of spirals such as the ‘trumpet spiral’ (Figure 2.10) (Hull 2003, p.113).

Figure 2.10 Celtic or Anglo Saxon hanging bowl escutcheon with trumpet spirals showing path width variation, British Museum, acc.no.1870,1013.16 Within the perceptual survey, only one out of 13 tests, which included spacing as a variable, yielded significant results. The significant test indicated that spacing did not rank highly as a differentiating variable but both spacing and rotational length were varied in this test so reliable results could not be obtained for the spacing variable alone. The spacing variable is specific in application to spirals whereas the other variables used in the classification system might be relevantly applied to a broad range

52 of unclosed motifs. No significant results were obtained for path width as a differentiator in similarity tests. Three survey participants used descriptive terms for spiral motifs which reflected on path width and spacing. Across the free description motifs they used the terms 'tight', 'evenly spaced' and 'concentric' for Figure 2.6b and contrasting descriptions of 2.6c as 'varied in width and 2.6e as 'expanding'.

Beck (1966b) found that increasing line width produced a greater perceived contrast with dark on light patterns but no significant result with light on dark patterns. Similarly, variation in spacing affected figure-ground contrast but no evidence was found to suggest how important these variables are as differentiating factors in perceptions. Studies of target identification have found that line width variation can be a significant factor in the rate of rejection of potentially matching motifs (e.g. Becker 2011) suggesting that in these studies, line width is playing a role in creating an impression of similarity. Pinna (2004), incidentally to her main study focus, also found that line width commonality was relevant to the creation of similarity groups. In contrast with the spirals drawn to represent endogenous visual phenomena (section 2.4), few of the decorative spirals encountered in this research showed overlapping lines. This implies, particularly with tightly packed spirals, that care was taken in the creation of the motifs to regulate the expression of this pair of related variables. Because of the variability seen in decoration and the findings from literature review that line width could be a significant grouping variable, these variables were included within the analysis.

2.2.2.5 Irregularity of coil Irregularity is taken to refer to deviation from the path of the identified primary curve due to unsystematic ‘wobble’ or decorative patterning. In many cases, apparently random wobble will be an unintentional product of freehand drawing but there may be instances, such as Figure 2.11a below where the apparently random irregularity represents a deliberate decorative effect as much as more obviously regular patterning (Figure 2.11b). For this reason patterning which affected the outline of the spiral was included in the analysis of irregularity.

53

Figure 2.11 Patterned or random irregularity. a. Plate with freehand spiral 1950’s-60’s designed Teddy Millington Drake (Victoria and Albert Museum acc.no. C.54-1991); b. Detail of an octopus tentacle from a jar from Crete, 1450-1400BCE, Ashmolean Museum, AN1911.608. The influence of irregularity on the perception of similarity was tested in nine motif pairings in the survey of which three yielded significant results. These placed irregularity behind form and orientation and of equal importance with all other included variables as a significant variable within the perceptual survey. As with the other similarity results from the survey, consensus tended to be greatest on those motifs that might be deemed to be most similar – those comparing similar magnitudes of irregularity that only differed in exact expression. Comparisons between smooth and irregular motifs and between systematic and unsystematic irregularity, did not yield any significant results. Nine participants used 'regular' or 'irregular' as shape descriptor or in free description terms for motifs. However, these responses related to smooth near-circular spirals and to motifs with changes in direction of curve on a large scale which were covered by the form variable within this method. Minor irregularity due to wear seen in Figures 2.6a and b was not apparently registered as important, both of these spirals were only described as regular.

In contrast to the similarity results of the survey, Kayaert et al. (2005) found that the strength of separation between regular and irregular shapes was high and similar to that between irregular curvilinear and angular shapes as mentioned above. Because the survey results tended to register similarity more strongly, possibly unduly strongly (see Appendix 1), priority was given to the literature review results and irregularity was included within the analysis.

54

2.2.2.6 Colour and patterning Colour and patterning are of clear aesthetic importance in decoration. Phillipps (1949) used patterning on the spiral to distinguish type classes of Maori spirals, prioritising it above other style variables and overall form. Herner (1989), studying spirals on Bronze Age Scandinavian brooches, and Zhushchikhovskaya and Danilova (2008), studying East Asian Neolithic pottery spirals, also noted culturally significant differences in spiral patterning. An example of the importance of colour as a distinguishing variable can be seen in the emotive and symbolic terminology used by Yörük weavers to describe both the choices of colour and colour harmony in their work (Daugherty 2004, p.331).

None of the perceptual tests of patterning as a differentiator yielded significant results but in describing the two photographic images that included patterning, five participants made explicit reference to the patterning of test motif 88 and six to that of test motif 157b (Figures 2.12 and 2.3e, p.38). Three participants made explicit reference to the texturing of the stone and bone spirals (Figure 2.6a-c). In reference to both patterning and texturing several other descriptors were used which may indirectly relate to the surface patterning and texturing; in reference to motif 88 descriptions of it as a fossil or sun and in reference to the stone and bone spirals, descriptors relating to technique or material must have made reference to texture.

Figure 2.12 Test motif 88 showing surface patterning. All three tests of similarity by colour in the survey yielded significant results with a high level of similarity being reported (modal value of nine). This result may have been influenced by the emphasis on shape throughout the rest of the survey. In the free description questions, eight participants used colour as a descriptor at least once.

Quinlan and Wilton (1998) found that in most tests colour was more important grouping variable than shape in consideration although some grouping combinations

55 could counter this effect. Cultural variation in the preferences for and the importance placed on colour may also vary. Hirschfeld (1977) studying the aesthetic preferences amongst the Kuna2 found that embroiderers preferences for applique designs found that properties of form – complexity and symmetry ahead of colour were more important than colour in determining preferences.

Although both colour and pattern are important descriptors the decision was made to exclude both variables from the classification system because neither attribute is inherent to the nature of the spiral as a motif – similar patterning or colouration might appear on any planar surface.

2.2.2.7 Diameter Diameter was felt to be the most appropriate variable to represent the absolute size of spiral motifs. The size of the decorated object is clearly an important limiting factor on motif size so measurements of diameter would need to be made in proportion to object size.

In the perceptual survey, the size difference of pairs of, otherwise identical, motifs with no background context was ranked as equally important as direction of turn, rotational length, irregularity, spacing and colour as a significant differentiator so the exclusion of this variable is not ideal. The key reason for excluding this variable was one of practical difficulty in a lack of known measurements for analysed examples drawn from publications or online museum catalogues.

2.2.2.8 Direction of turn The cultural meaning of spirals can be dependent on how a culture interprets their physical shape. Perceptions of direction of turn depend on whether the spiral is regarded as coiling inwards or outwards; on which way up the object is; and on whether the spiral is regarded as being static or revolving.

There may be a physiological element to the direction of turn of spirals in the decorative arts; due to the arrangement of the muscles in the arm it is easier to rotate

2

Indigenous group of the San Blas Islands, Panama.

56 the right hand in a clockwise direction and the left hand in an anticlockwise direction (McManus 2002 pp.46-48). Depending on the handedness of the artisan a natural tendency to draw spirals in one direction or the other might be expected although this would also assume a consistent tendency for people to draw spirals from the outside inwards or the inside outwards. It appears that the influence of the handedness of the sculptor on the direction of spirals is evident in a statue on which the curls of hair are carved in opposite directions on either side of the head (Cook 1914 p.241).

One of the four tests of direction of turn as a similarity measure yielded a significant result. with a modal similarity score of nine although three participants used direction of turn as either a shape descriptor or as a free description turn.

Although it can be shown that symbolic significance is sometimes placed on the direction of turn (Cook 1914, p. 30), the lack of perceptual importance placed on this variable and the likelihood of a physiological influence on its expression led to this variable being excluded from further analysis.

2.3 Perception of spiral forms Two questions posed in this research aimed to identify possible motivating factors in the production of spiral decorative motifs by comparison with naturally occurring spirals as possible sources of inspiration. As noted in Chapter 1, spirals can be found in a wide range of naturally occurring structures and varied studies have attributed the use of spirals in decoration to the reproduction of natural forms (e.g. Takaki & Ueda 2007; Lewis-Williams & Dowson 1988) but the fact that spirals are frequently copied from natural forms does not explain the reason for the particular selection of this structure for reproduction. Bando and Hirabayashi (2004) found that spiral motifs were given preferential attention when compared with concentric circle and radiating line patterns both in terms of which motifs first received attention and in the length of attention devoted to them. An explanation for this may be found in the structure of the human visual cortex. Graziano, Andersen and Snowden (1994 cited in Bando and Hirabayashi 2004) identified that the V5 region of the visual cortex which is associated with the detection of movement is structured to respond most strongly to particular patterns of motion and found that spiralling movement is one of four fundamental

57 motion types which are detected by neurons in the V5 region. An innate tendency to recognise spiral shapes and their association with natural forms which are important either as a potential food source or as a threat may go some way to explain the depiction of natural real spiral forms.

A second possible naturally occurring inspiration for spiral decoration considered in this research addressed spiral visual effects generated within the brain. Since the nineteenth century, optical effects which occur during periods of unusual brain activity such as migraines and before epileptic seizures have been studied and recorded (Ermentrout & Cowan 1979). It was observed that, in addition to hallucinations of real objects, a recurring set of geometric light patterns could be identified in these visual effects. Psychoactive drugs can induce similar effects and the availability of these for experimental study under controlled conditions allowed attempts to be made to record the nature of the geometric patterns across a number of subjects. The concept of the 'form constant' was introduced in 1928 by Heinrich Klüver (1967 cited in Ermentrout & Cowan 1979) to describe the geometric visual phenomena generated within the optical system independently of external visual stimuli. Klüver divided the visual phenomena into four classes (tunnels, webs, lattices and spirals (Figure 2.13)).

58

Klüver (mescaline) Kellogg, Knoll & Kugler (electrical stimulation)

fingers

cherries

Horowitz (multiple sources)

Becker (flickering light) Ermentrout and Cowan (theoretical model)

lines and angular figures

zig-zags, waves

Figure 2.13 Class groupings of geometric visual phenomena.

tunnels, funnels, cones

pinwheels springs

kaleidoscope mandala

58

Siegel and Jarvik (cannabis derivative) text description only

59 Ethnographers studying the use of psychoactive drugs in ritual observed that the decorative styles of the cultures they were studying included geometric patterning which resembled the classes of geometric visual phenomena observed by Klüver and others. On some occasions, informants expressedly linked the decorative style to representation of visual effects seen during trance states. Using ethnographic analogy, later studies have drawn a link between the presence of motifs resembling form constant types in rock art of prehistoric cultures and the use of trance states in religious practise. Lewis-Williams and Dowson (1988), in particular have suggested that the presence of these motifs is indicative of shamanistic practises within the cultures that produced them. This association is controversial as others have proposed alternative explanations for the use of abstract geometric motifs or have suggested that these motifs do not appear in contemporary shamanistic societies with sufficient consistency to make a link (see comments on Lewis-Williams and Dowson (1988); Bednarik (1990); Bradley (1989)). The use of the spiral classification system proposed in this research aimed to test whether the concept of such a link could be supported by measured analysis of spiral shape and to search for these forms in a wider range of cultures than those with explicit links to shamanistic practise on the assumption that if these physiologically engendered effects were represented in decoration, their presence would be likely to be widespread (Hodgson 2000). This section discusses the mechanisms of production of the geometric visual effects; types of trigger and the shape classes of visual phenomena identified.

Internally generated visual phenomena have variously been termed ‘phosphenes’ ‘entoptic phenomena’ and ‘subjective visual phenomena’ and have been subdivided into two classes: hallucinations, which partially draw on memory to create non-real imagery, and endogenous phenomena, arising from the structure of the visual system (Dronfield 1996a). This research is solely considering the second class which will be termed 'endogenous visual phenomena' (EVP). Although some of the spirals analysed as representations of subjective visual phenomena were given representational interpretations by those experiencing them, the drawings produced appear to accord more closely with forms typically identified as endogenous, Horowitz (1964) noted a marked contrast between the representational language used in describing visual phenomena and drawings of the same phenomena.

60

Klüver’s (1967) study induced EVP using the psychedelic drug mescaline. Similar types of geometrically-patterned EVP can be triggered by deliberate or unintentional sensory deprivation (La Barre 1975, p.11); they may occur before epileptic fits, in the transitional stage of falling asleep, during migraines and can be induced by other psychedelic drugs (Ermentrout & Cowan 1979).

The concept of classifying the patterns identified in EVP has been developed from Klüver's original set of four form constant patterns in subsequent studies employing a wider range of natural and drug induced occurrences of EVP. Knoll and Kugler (1959) used electrical stimulation to induce EVP and found that stimulation at differing frequencies could create differing typical pattern types (Fig.2.13). This suggested that the type of EVP represented by form constants originate in the optical cortex rather than the eye itself and arise as a result of neuronal activation or inhibition patterns usually established when an external visual stimulus has been removed (by a darkened environment or closing the eyes) (Siegel & Jarvik 1975, p.143); the occurrence of EVP is attributed to a disinhibitory effect on the brain stem permitting increased levels of cortical excitation causing instances of hypersynchrony in neuronal firing to arise from the normal low level excitation state (Ermentrout & Cowan 1979). The exact location of generation of EVP within the visual cortex remains uncertain – the striate (V1) visual cortex has been suggested as an origin point. Bressloff et al. (2001) suggest a cellular structure of the V1 cortex which could account for all the form constant patterns identified by Klüver with the various form constant structures being translated from a radially symmetrical form occurring in the retina to a translationally symmetrical form in the cortex causing circular tunnel effects to be represented in the cortex as vertical lines and spiral whorls as diagonal lines. However, since spiral and tunnel EVP are often associated with movement, the V5 visual cortex (which receives input from the V1 region) may be either the source or a necessary additional structure for the perception of these types of EVP (Dronfield 1996b) as well as the locus of detection of external spiral structures.

Causatory factors of EVP have been divided into three types: toxically induced - arising from exposure to psychedelic substances, endogenous – arising from an internal

61 disturbance of the nervous system and exogenous – arising from an external stimulus to the nervous system involving, for example, electrical impulses applied to the temples or pressure on the eyeballs (Becker 2005, p.4). In discussing studies of spiral EVP it must be noted that there is not necessarily consistency between studies in what is meant by 'spiral' – some use whorls of low rotational length and others single spirals of greater rotation and it is not always clear what participants have described, what terminology participants used and what assumptions were made by the researchers. Siegel and Jarvik (1975, p.112) assigned nine groups based on verbal descriptions of imagery seen by subjects taking cannabis derivatives (Figure 2.13); this categorisation included complex representational imagery as a single class. Once an initial set of constants had been identified Siegel and Jarvik trained participants in subsequent tests to use these groups and so may have influenced the variety and identification of descriptions.

Spiral EVP groups have been identified using all three causatory factors (e.g. Siegel and Jarvik 1975 (drugs), Horowitz (1964) (schizophrenia) and Becker (2005) (high frequency flickering light)). Siegel and Jarvik's results from studying toxically induced subjective visual phenomena suggest that spirals and webs were the least frequently reported geometric EVP both in controlled tests with trained subjects and in reported experience of Huichol participants in rituals (Siegel & Jarvik, p.129 Fig 12, p.139 Table 4). Classifications were ambiguous and spirals may have been reported within the much more frequently occurring class of tunnel imagery compare Siegel and Jarvik (1975) Figs. 8 and 21; Nicholson and Firnhaber (2003, p.58)). Kellogg, Knoll and Kugler (1965) only identified 14 spirals amongst 520 electrically induced geometric EVP patterns, ranking them tenth in frequency in a class system of 15 EVP types. In contrast, Becker's (2005, p.47) experiments recorded a higher frequency of spiral observation with 22% of visual effects classed as spiral. The classes identified within any of these systems are not clearly differentiated, EVP can merge, fragment, overlap and multiply in a manner which makes absolute boundaries between classes difficult or impossible to establish (Lewis-Williams & Dowson 1988). The presence of a static spiral EVP class has been doubted by some analysts; Horowitz (1964) does not identify spirals as a specific geometric EVP class but associates them instead with the tendency to see circular shapes and parallel lines. Lewis-Williams and Dowson (1988) exclude

62 spirals from consideration on the grounds that they are a ‘special case’, but present spirals in Coso rock art (of the California Great Basin) within the class of web patterns (Figure 2.13).

Siegel and Jarvik (1975, p.110) suggest that drug induced EVP of the geometric type identified by Klüver as form constants are more intense than those produced by other stimuli and may take different forms. It may be the case that geometric EVP induced by events other than drug taking feature spirals more frequently; Feinberg (1970, p.126) identified that the presentation of dream imagery, schizophrenic hallucinations and those induced by mescaline differed in their typical presentations. Tass (1997) modelled different types of neuronal activation or inhibition pattern to show that if these were varied by the agent causing the cortical excitation different types of EVP might be expected. Also, there might be variations in whether spiral EVP are static or rotating, their direction of turn, number of arms and curvilinearity. This variation must be seen as speculative because studies of variation in EVP according to trigger type have not been found. Siegel and Jarvik (1959, pp.134-136) cite two examples of frequent spiral imagery experienced by people with schizophrenia. Horowitz (1964) studied the subjective visual phenomena associated with schizophrenia and recorded some spiral effects being drawn either as individual images or as components of larger drawings. Panayiotopoulos (1994) and Schott (2007) identify typical migraine and epilepsy related imagery spiral EVP do not appear to be characteristic of these disorders. The limited and consistent forms of migraine-related EVP may indirectly give support to the idea of variation according to stimulus type. Oster (1970) noted that light or harder pressure on the eyeballs also produced differing and consistent patterning.

2.4 Endogenous visual phenomena and their occurrence in decoration. The physiological origin of EVP patterns should provide a sound basis for cultural comparison but Bahn (1988) raises the question of whether there is yet sufficient evidence to suggest that geometric EVP can reliably be considered universal and consistent over time. Hodgson (2000) suggested that the necessary neural pathways to generate EVP's would have evolved by the time of the creation of the earliest marks

63 resembling them in the Lower Palaeolithic. The likelihood of physiological consistency is supported by the generation of EVP proposed by Ermentrout and Cowan (1979) occurring in the part of the visual cortex first reached by signals from the retina with minimal higher level visual processing having taken place; inter-species consistency has been shown in the basic process of mapping a received visual image onto the V1 cortex as an array (Stewart 2011, p.173). Butterworth (1967 cited in Siegel and Jarvik 1975) found notable consistency even in the occurrence of representational imagery between individuals in some studies which suggests similar underlying mechanisms in their production. Laboratory tests have not, however, covered a broad ethnic span and study of change over a relevant timescale would not be possible, so this assumption of physiological consistency has not been tested.

There are several factors which might be expected to introduce inconsistency in decoration inspired by EVP, the manner of expressing of the patterns in an outward form will be culturally variable (Siegel & Jarvik 1975, pp.137-8) and reactions to psychedelic drugs can differ between cultures (Siegel & Jarvik 1975, p.136). A further complication is introduced in that two stages of subjective visual phenomena may be observed, the first involving geometric EVP and the second developing into hallucinations involving more complex representational images (Siegel and Jarvik 1975, p.111). Lewis-Williams and Dowson (1988) added a middle stage marked by a transition between rapidly changing EVP which those experiencing them have difficulty describing and the attempt to assign representational meaning to the EVP being experienced. This middle stage could be a property of the subject gaining experience and learning to observe and describe recurring images or of the interference of hallucinatory imagery from the last stage encouraging a representational interpretation of the abstract EVP structures. Representational language is frequently used in the description of geometric EVP by those who have experienced them (Horowitz 1964) and this interpretation will necessarily be restricted to the cultural experience of the individual (Siegel & Jarvik 1975, p.137) and may affect their interpretation and recollection of EVP observed in the first stage especially given the gradual process of change from abstract/geometric to representational imagery (Maclay & Guttmann 1941).

64 The perceptual survey responses highlighted the problem of language constraints on interpretation of motifs. Most participants felt able to answer tests that did not require a high degree of reference to an external set of concepts (e.g. mark the centre) but response rates were lower when asked to describe a motif and some participants responded that they did not know what shape the motif was. This suggests that EVP forms for which the language and culture do not provide a convenient descriptive term are more likely to go unreported. External stimuli can, in some cases, influence the hallucinated imagery (Siegel & Jarvik 1975, p. 130) which might, again, lead to a culture-specific effect. The potential for cross-cultural inconsistency in the reporting of EVP lowers the likelihood of their identification from analysis of the decoration itself. From the approach taken in this research, starting from the motifs rather than their causatory factors it also limits comparability if two cultures influenced by EVP produce different sets of motifs. However, if an underlying physiologically derived form can be consistently observed, the variation around these forms which makes their identification difficult is likely to be driven by culturally specific factors and so to be potentially informative and to provide a basis for comparison.

The presence of EVP-related decoration has been identified in ethnographic studies and historical sources from a wide range of cultures. Mural and pottery decoration produced by the Tukano people of Colombia uses geometric imagery resembling EVP and explicitly linked by the local populous to visual effects associated with the consumption of the psychedelic drug yagé, obtained from Banisteriopsis vines, during divination and healing rites (Reichel-Dolmatoff 1978, p.297). Reichel-Dolmatoff (p.301) also notes that the Colombian Chocó people who are geographically separated and, in many ways, culturally distinct from the Tukano also consume yagé and have a similar geometric decorative style. Similarly, the Huichol people of Mexico recreate geometric EVP (nieríka) observed in the first stage of trances (induced by the consumption of peyote cactus) in woven and embroidered textiles (Schaefer 1997, pp.156-157). The Huichol produced stylised representations of the peyote cactus in forms that are typical of geometric EVP, usually as a tunnel or web motif but depictions as spirals are also found (Lumholtz 1900, pp.199-200; Siegel & Jarvik 1975, p.137). Lewis-Williams and Dowson (1988) identified depictions of trance dances of medicine men associated with EVP in southern African San rock art and that the San

65 medicine men identified themselves in EVP they saw in a trance state which included spiral forms. Delluc and Delluc (1976) observed that a case has been made for the influence of EVP in the work of Paul Klee and Joan Miro and Keeler (1970, p.206) studying Max Ernst although the example proposed to illustrate spirals Young man intrigued by the flight of a non-euclidean fly does not, in fact include any spirals that complete a full rotation and better illustrates the confusing illusory effect created by close concentric curves as the eye jumps between lines. Contemporary usage of spirals to indicate hallucination or disorientation can be seen as a convention of comics and graphic novels.

Links have been suggested between prehistoric rock art and geometric EVP in European Palaeolithic and Neolithic rock art (e.g. Lewis-Williams and Dowson (1988); Dronfield (1996a and b; Hodgson 2000). Bednarik (1986) made the same association in Australian rock art and Mamaní and Fernández Distel (2010) suggest that EVP may be visible in pre-Colombian petroglyphs and ceramics. Lewis-Williams and Dowson (1988) proposed a model, derived from ethnographic studies, for the identification of EVPlinked imagery in prehistoric rock art. Owing to the physiological origin of EVP, they argued that the model can be applied to the archaeological record with a greater confidence than socially or culturally-based ethnographic observations. The model proposed for the identification of imagery arising from trance states in rock art requires:- the presence of a range of types of EVP (they identify seven classes from literature review); imagery linked to the later stages of visions as well as EVP; and evidence of seven types of interaction of motifs including merging, splitting and repetition of EVP which characterise subjective visual phenomena.

The idea that EVP inspired rock art motifs has been disputed or caveats raised as to the scope of its application suggesting that it could only be used to account for some Palaeolithic art (e.g. Bahn 1988; Martindale 1988; Mirimanov 1988) and suggesting that the collation of EVP derived motifs from the full range of Palaeolithic imagery may be drawing together motifs which only have a coincidental association (Layton 1988). Bahn (1988) argued that only a limited number of basic shapes of the type attributed to EVP exist and questions whether these could not arise widely simply from doodling. It is true that some of the supposed EVP could arise from basic hand movements –

66 back and forth to create zig-zags and lines, circling and, up and down to create dots (Mirimanov 1988). This link to mechanical movement cannot be ruled out as an explanation for the motifs but they often accompany sophisticated and controlled drawings and were produced by cultures capable of considerable dexterity in tool production. Turner (1988) suggested that simple rock art motifs may be the work of children. Kellogg et al. (1965) also noted a strong similarity between the types of drawing produced by children around the age of two to four and the form constant classes identified in subjective visual phenomena in adults. Hodgson (2000) also noted the similarity to children's drawings of geometric patterns in Lower and Middle Palaeolithic marks but suggested that this represented a stage in learning to create representational art that resembled that seen in young children. If, as Martindale (1988) suggests, the drawings produced by young children are influenced by a focus on different stages of image processing from the reproduction of replicas of real objects, the drawings would be likely to reflect the same structures as EVP experienced by adults so could probably not be reliably differentiated by the end product alone but might still be considered valid representations of EVP. The overlap with motifs that might be generated from non-EVP stimuli and the possibility, noted by Lewis-Williams and Dowson, that the representations of EVP, if they are such, may not have been produced by those who experienced the EVP creates considerable potential for ambiguity in the interpretation of the origins of these motifs.

Lewis-Williams and Dowson (1988) state that the presence of a single type of motif which is associated with EVP is not sufficient to argue that EVP influenced the artwork of a particular culture. The intention in this research is not to dispute that statement but rather to take the assertion that EVP have influenced the decorative schema of numerous cultures as a plausible explanation for at least some geometric decorative motifs and to focus on a single type to establish whether a detailed study of the manner in which it is produced can aid in identification of possible EVP influence or the influence of objective sources on the decoration. Bahn (1988) noted that one of the challenges in the work of Lewis-Williams and Dowson is to define the limits of a motif which can be considered to be related to an EVP on a continuous scale of decorative variation.

67

2.5 Construction techniques The manner of constructing decorative spirals can have a considerable impact on the form and style of the outcome. Geometric layouts can be used to construct regular spirals or approximations to them but there is little evidence for complex geometric constructions being used in the layout of decorative spirals. The earliest evidence for compasses, necessary for the majority of geometric spiral constructions comes from the eleventh to tenth centuries BCE, their origin is thought to be in Attica (Lemos 2002, p.81) but the use of string compasses would leave little archaeological evidence (Herner 1989, p.135) so could predate this time. Herner (1989, pp.133-4), in her study of Scandinavian bronze spiral engraving, found evidence of compass point marks in the centre of spirals suggesting the use of concentric circles as layout guides. This might be expected to produce spiral forms with a low and constant expansion rate to match best the circular guides.

The number of spirals typically linked together may reflect geometric frameworks underlying the decoration for example, particular grids or polygon layouts lend themselves to the linking of particular numbers of spiraloid arms (Fig 2.14). In spirals laid out by geometric techniques, imperfections in the spiral form might indicate the underlying framework used to construct the spiral (Hull 2003, p.110). Analysis of spiral forms therefore has potential to highlight both the mathematical and technological knowledge of the culture although with very close approximations such as the whirling rectangles golden spiral (Sharp 2002) minor deviations from an exact spiral form might be masked within the margin of error in analysis.

Figure 2.14 Spirals and spiraloids within a hexagram construction. Detail on a Mamluk incense burner, British Museum acc. no. 1878,1230.682. The nature of the decoration can be similarly affected by the medium and the technique employed. The path width (or spacing if a negative impression is created by casting) necessarily has a minimum value of a single stroke or cut of the implement

68 used to create it and some techniques naturally lend themselves to particular patterns of spacing – wire coils tend to have spacing of zero or low and irregular variation whereas patterning created by engraving or painting tends to have higher spacing to differentiate the coils. In consideration of line width and gap, spirals created by physically coiling an element such as wire, clay or fibre are more easily created with no gap between the coils (and will therefore tend to have a similar form with a constant expansion rate) whereas painted spirals, will tend to have a gap between coils to make the shape clear. Form variations can be introduced by the type of tool used. Wrought iron spirals created around a former will often have a straight segment at the centre of the spiral where the rod was gripped and cold bending with a jig tends to require a thinner metal strip than hot-wrought spirals; iron spirals curved freely on an anvil can curve smoothly all the way to the centre (examples of the three techniques can be seen at Artisans of the Anvil (2005); Vega (2008) and Shop Outfitters (2010)).

Straight sided spirals have few natural analogues, one of the theoretical models of cortically generated spirals produced by Tass (1997) gives rise to a rotating angular spiral form with rounded corners but reports of these visual phenomena suggest that they are difficult to focus on clearly (Horowitz 1964; Maclay & Guttman 1941) and as a rotating form, the detailed shape of the modelled form might be difficult to register even assuming that the modelled form arises in reality. It seems likely that highly angular spirals in decoration may arise largely from the manufacturing process. Woven textiles, in particular, are better suited to the use of angular rather than curvilinear forms. A contrast of the effect of medium and technique on spiral form can be seen in Mexican San Sebastián Culture ceramic figurines (Figure 2.15) which have depictions of angular spirals, which would be easier to weave than curvilinear spirals, painted on the representation of the fabric loincloth whereas those over the breasts believed to represent painted skin patterning (British Museum gallery object label: Am,Hn.121) are often curvilinear. The type of linkage may also be influenced by the technique for example, key patterns lend themselves better to weaving than whorls.

69

Figure 2.15 San Sebastian ceramic figurine, 2nd century CE, Metropolitan Museum of Art, acc.no. 2005.91.14. However, the majority of angular spirals found during this research were either filling spaces around other decoration or appeared as linear or two dimensional key patterns on media that do not inherently appear to lend themselves to the use of straight over curved lines so in some cases, the production of tessellating patterning that fully and evenly covers the ground appears to be a stronger motivation in the creation of angular spirals that the constraints of technique.

The idea that medium and technique may be significant in the decision to create spiral decoration is supported by the common restriction, even in cultures in which spiral decoration is frequent, to restrict it to certain types of material or technologies. Shang Dynasty jade ornaments rarely feature spirals although they are commonplace on bronzes and in Middle Kingdom Egypt spirals were common on scarabs made of soft materials such as steatite or faience but rare on scarabs carved from hard gemstones. In both these cases, the difficulty of carving fine detail into hard stone may be offered as an explanation for the occurrence of spiral decoration. Other choices of technique, although clearly delineated, are less easy to explain. Examples of painted spiral decoration on pottery are common in pre-Dynastic Naqada II period Egypt and again in New Kingdom mural painting but painted spirals are not found in Middle Kingdom contexts.

70

2.6 Summary In the development of the analysis system used in this study, the concept of the basic form of a spiral as a universally known, if not necessarily formally defined, construct was expanded to consider other aspects of the decorative embellishment of the underlying construct. An initial group of eight possible variables of style were derived from a review of spiral decoration on museum objects. The set of five variables of style added to this analysis were chosen on the basis of their apparent significance as means of differentiating and classifying decoration assessed from literature review of studies of perception; from evidence for their cultural or symbolic significance and from the results of the perceptual survey of spiral motifs carried out for this study. The set of variables discussed in this chapter forms the basis of the proposed spiral classification system and of its application in testing for evidence of a recurring form constant structure across disparate cultures.

Spirals have been identified as a type of EVP which can arise, independently of culture, from the structure of the optical system. The prevalence of spiral decoration across a wide range of cultures has been attributed, particularly in reference to early rock art, to the occurrence of EVP spirals although these assertions are not universally accepted.

71

3 Development of the classification and analysis method The analysis method proposed in this research needed to provide a quantitative description which represented the visual attributes of the spiral motifs. Analysis took three stages: digitisation of the motif; quantification of the variables; and comparison and the creation of class groups. In initial recording the spiral motifs, the spiral definition as applied needed to be clarified to ensure that equivalent motifs were being compared. Requirements of the quantification stage were that it should describe each motif on the same terms and in a manner which allowed descriptions to be stored and comparisons to be made. The classification system needed to be objective in its operation and to accommodate the high number of variables used in form and style description. This chapter sets out the method of data collection, characterisation and analysis of spirals which was developed for this research using the example of a single spiral (Figure 3.1), shown through each stage of the method.

Figure 3.1 The specimen spiral CSBM17.9 from a Western or Eastern Zhou Dynasty you (ritual vessel), British Museum, acc. no. 1983,0202.1.

3.1 Data collection and processing 3.1.1 Sources of data This thesis focuses on decorative design - designs applied to the surface of a functional item which are not essential to its practical function. The identification of what is functionally necessary does not take into account what the maker of the object may have considered necessary for function in religious or symbolic senses and, from this

72 point of view, the interpretation of functionality is an external imposition but a division can be drawn between culturally independent and culturally dependent functions which allows consistency of selection of sample material. Spirals serving a primarily functional purpose themselves such as springs in brooch pins were excluded from analysis. Even when it appeared possible that they also played a decorative role it was felt that the main criterion governing their shape would be likely to be functional rather than decorative requirements. Data for the cultural case studies were obtained from direct access to objects in museum collections; and from online and hard-copy catalogues of museum collections. In using data obtained from published records, priority was given to the analysis of spirals on objects where distortion due to camera angle or angle of the object was lowest. Data used in the creation of template EVP spirals were obtained from illustrations in publications and academic papers. The identification of a typical geometric EVP shape is clearly hampered by having to rely on the description of the shapes rather than direct access to the imagery being seen. Siegel and Jarvik (1975, p.115) noted that participants in their experiments had difficulty in describing what they were seeing. Schott (2007), in discussing the identification of the causes of migraine aura felt that illustrations provided a much more powerful tool than written description in that they highlight a consistency in the aura seen which cannot be identified so effectively from written description. He distinguished between the types of illustration in their value to analysis. The most effective illustrations were those made by conscious observation whilst the subject was experiencing the aura. However these could be constrained by other effects of the EVP cause such as lack of concentration. Useful information could also be obtained from drawings made after the fact provided that the intention of these was to record the aura objectively. Artistic works influenced by the experience of migraines, including aura, were deemed to be of less analytical use in identifying the causes of migraine aura but works displayed in an exhibition of migraine art did include characteristic hallucinatory effects and showed consistency in these effects between pictures (Schott 2007).

73 To attempt to approximate form constants, artwork and diagrammatic representations stated to be directly inspired by subjective visual phenomena, along with theoretical models of EVP, were analysed as a sample set. Both technical depictions and artists’ representations from the reports of others clearly introduce greater potential for their own subjective interpretation to influence the manner in which EVP are represented. This secondary representation may, however, reflect more accurately the process of production of decorative representation of geometric EVP. The participants in a study by Becker (2005) of EVP induced by flickering light made up the majority of spiral EVP source images. Participants were asked to look for and draw geometric forms derived from earlier free descriptions by research participants which had been classified by the researchers into nine form groups (Becker 2005, p.46). All participants were asked to look for a specific EVP type at the start of each test run and three of the participants had been shown template images (Becker 2005, pp.107,154). The use of verbal and visual stimuli would have had a training effect both in causing the subject to be more ready to see the form in question and in the way in which it was reproduced. When asked to draw spirals, only participants who had not seen stimulus cards drew helices (highlighting the ambiguity in the interpretation of the term spiral). Two of the six participants drew spirals in response to non-spiral verbal stimuli and there was considerable elaboration in decoration and line width away from the basic stimulus card motifs present (Becker 2005, Appendix H). Since Becker's subjects did not conform exactly to reproduction of the stimulus cards, it is likely that the drawn responses represent the geometric imagery that the subjects saw so can be considered usable data. Given the shortage of EVP data available, the drawings by Becker's subjects were felt to be acceptable although not ideal. Specimen mathematical spiral forms for comparison were obtained within the analysis software by plotting the equations of the spiral and tracing them using a point which had its trajectory fixed to the curve so removing human tracing error from these comparison examples. 3.1.2 Data collection Initial identification of spirals during data collection was based on visual assessment employing a minimum rotation estimate of 225 degrees at this stage to allow some spiraloids to be rejected during analysis and to reduce the possibility of failing to

74 record spiraloids which should have been included according to the minimum rotation criteria. Where it was possible to take photographs directly from objects, the spiral or the middle of a field of spirals was centred in the field of view, to minimise distortion, using a guide grid on the camera screen. Photographs were taken as jpeg images at a resolution of 9 mega-pixels. For each culture studied, a minimum sample size of 100 spirals was aimed at although for the Jomon case study, only 94 spirals were obtained. Two spirals were sampled from each object (where two or more were present). Where 25 or fewer spirals were present, the sample was chosen at random by assigning individual numbers to the spirals. For objects with more than 25 spirals, the sample was chosen by overlaying a grid and selecting the spirals closest to the two chosen grid square numbers. Spirals that were subjectively assessed to be too damaged for analysis (Figure 3.2) or which were not fully visible due to the angle of a published photograph were excluded from the sample.

Figure 3.2 Scarab showing spirals damaged beyond analysis (indicated by arrows), World Museum, Liverpool acc. no. 44.19.357).

75 The initial stage of data collection focussed on contextual data which was used in cross-cultural comparison of the analysed spirals. The essential contextual information is summarised below: Contextual variables Assigned number Current location / publication Museum's accession number / publication page and figure number Date of object

Data for the specimen spiral CSBM17.9 British Museum 1983,0202.1 mid-late Western Zhou/early Eastern Zhou 800-700BCE south/south-west China n/a you (ritual vessel) bronze cast snake

Place made (or found) Associated culture/ethnic group Type of object Medium Technique Symbolic or representational meaning of the spiral Table 3.1 Contextual data collected for the spirals

The assigned number consisted of a two letter code identifying the case study; two letters identifying the museum or publication; and a running number for the object with a decimal for the number of the spiral sampled. For several objects, some aspect of the contextual information was unknown. In these cases, the object was recorded to be used in analysis as far as was appropriate. 3.1.3 Data processing Data processing involved the digitisation of the spirals to permit analysis; and the recording of contextual data and measurements in a suitable format for comparison. Contextual data and measurements were recorded in separate database tables linked on the primary keys of object and spiral number. 3.1.3.1 Image distortion The first issue arising in the digitisation of the spiral was perspective and lens distortion in the photographs. The effect of lens distortion was tested by digitising photographs of a ruler from heights of 100, 125 and 200mm (Figure 3.3). The images were divided into vertical and horizontal quadrants and the ruler photographed vertically in each of four horizontal quadrants vertical error measurements were made

76 by comparing the apparent length of the ruler within each vertical quadrant with the length measured from the digitised image. This process was then repeated with the ruler placed horizontally in each vertical quadrant. The average error per millimetre was calculated for each vertical and horizontal offset (Table 3.2). The maximum error encountered was ±0.008/mm this was used as the margin of error for calculation (see section 3.1.5, p.86). Average vertical error was ±0.003/mm and average horizontal error was ±0.001/mm. Distance from the camera did not have a large effect with average error of ±0.002/mm at 100 and 200mm and ±0.001/mm at 125mm.

Figure 3.3 Sample image rectification test images.

77

vertical quadrant V0 (centre)

V 1/4

V 3/4

V1

100 100 125 200 200 100 125 200 200 100 125 200 100 125 200 125

H frame 1/4 0.002 -0.001 0.002 0.001 0.001 0.002 0.006 0.002 0.002 0.004 0.008 0.003 0.006 0.008 0.004 0.004

H frame 1/2 0.002 0.000 0.002 0.001 0.001 0.002 0.004 0.002 0.002 0.002 0.003 0.002 0.005 0.005 0.003 0.003

H frame 3/4

H frame 1

0.003 0.001 0.002 0.002

0.003 0.001 0.003 0.002

0.002 0.002 0.003 0.002 0.002 0.002 0.003 0.002 0.004 0.003 0.002 0.003

not measured 0.002 0.003 0.002 0.002 0.003 0.003 0.002 0.003 0.003 0.002 0.002

200

0.006 0.004 0.003 0.002 V1 values could not be calculated at 100 100mm

Table 3.2 Image rectification tests average error/mm

horizontal quadrant H0 (centre)

H 1/4

H 1/2

H 3/4

H1

V frame 1/4 100 100 125 200 125 125 200 100 125 200 200 100 125 200 125 200

0.001 0.002 -0.001 0.000

V frame 1/2 0.000 0.001 -0.001 0.000

V frame 3/4 0.001 0.001 0.000 0.001

V frame 1 0.002 0.002 0.001 0.001

-0.001 -0.001 0.000 0.001 -0.001 -0.001 0.001 0.001 0.000 0.000 0.000 0.001 0.001 0.003 0.002 0.002 -0.001 -0.002 0.000 0.001 0.000 0.001 0.001 0.001 0.000 0.001 0.001 0.001 0.001 0.005 0.004 0.004 -0.001 0.001 0.001 0.001 0.002 0.002 0.002 0.001 -0.001 -0.002 -0.001 0.000 0.000 0.003 0.002 0.002 H1 values could not be calculated at 100 100mm

77

V 1/2

Vertical camera height

78 A second set of pilot studies tested image rectification software and mathematical correction methods to see whether improvement could be achieved on the best image produced as a photograph using Adobe Photoshop CS3 and Corel PhotoPaint 12. The best result using image manipulation was obtained from Adobe Photoshop’s lens correction filter. A good, but not perfect, correction of rectilinear grids was achieved but these results were achieved by visual assessment of regularity which would be much more difficult to apply to the artefacts themselves with fewer straight line or right angular visual cues, so correction would be likely to be less successful for these. Since error of the distorted images was slight the images were left uncorrected so that error could be estimated rather than attempting image correction with the available software which would produce variable results with unknown distortion errors. Spirals on curved surfaces caused considerably greater distortion in the image. Excluding any spiral that is not completely flat from classification would have been extremely limiting but if the spiral is at all three-dimensional, its cross sectional shape becomes an additional variable and difficulties may be introduced into the assessment of its form. A compromise was reached in which spirals where the increase or decrease in height between any two successive coils is less than half the change in the width of the coil were included in the study. Given that these spirals were being regarded as notionally flat and that, in the majority of cases, the third dimension was introduced by vessel curvature, the height and cross-sectional shape of the spirals were not used as variables in this analysis system. Spirals with obvious warping or distortion (Figure 3.4a and b) were excluded from analysis.

a.

b.

Figure 3.4 Damage and distortion to spiral decoration. a. Cypriot gold mouthpiece with warping, British Museum, acc. no. 1897,0401.448. b. Scarab with cracking, Liverpool World Museum, acc. no. 56.20.423.

79 3.1.3.2 Software implementation of the method One of the objectives of this research was to develop a method which was easy to apply. With this in mind, the measurement of form and style variables was implemented using an applet written in JavaScript within the dynamic geometry software Geogebra. This software was chosen as it is freely available and so does not limit accessibility. The method functions independently of the software implementation and could be automated with other mainstream geometry software which allows scripting. 3.1.3.3 Creation of a digital tracing To create a digital copy of the spiral, both sides of the line of the spiral were traced with the tracing lines ending at the point where the edge of the spiral line turned to finish the line or reached a decorative end point. Spirals were digitised by manual tracing of printed spirals using a digitising tablet and stylus (Wacom Intuos GD-0608-R). This method was chosen in favour of automated edge detection because the tracing of the motifs required an understanding of the definition of the shape - it was insufficient to simply follow the line to its end when the spiral formed part of or joined a non-spiral path (see p.80). Also, available edge detection software was not effective in identifying low contrast spiral boundaries (Figure 3.5).

Figure 3.5 The spiral line segments obtained by the best automatic edge detection result for the specimen spiral (indicated by change of colour). Tracing of printed spirals yielded lower tracing error from tremor than using the tablet to follow the outline of the image on screen. The printed spiral image was aligned to one edge of the tablet tracing area and taped to the tablet's surface rather than being placed under the film covering the tracing area of the tablet as the stylus point was

80 less prone to slipping on the paper than on the film surface. This stage of the analysis had the most variation in precision since the degree to which irregularity in the spiral curve was recorded depended on the resolution of the image and on the perception of the analyst as to what constituted significant irregularity. Following the method of Liu, Carroll, Wang, Zajicek and Bain (2005), each side of the spiral line was traced in both clockwise and anti-clockwise directions and the average line calculated for both sides of the spiral path to minimise tracing error (Figure 3.6). Subsequent analyses were based on the results of tracing returned as Cartesian coordinates for each recorded data point.

Figure 3.6 Tracing of the specimen spiral in two directions of the spiral outline in green (anticlockwise) and red (clockwise) and the averaged lines in black. One of the main difficulties encountered in the development of the analysis method was to find consistent and appropriate means of identifying the spiral path which should be traced. Identifying start and end points of tracing Decorative spirals often have embellishments of the start or end of the spiral line for example, 'eyes' at the centre of the spiral seen in Maori and Bronze Age and Classical Mediterranean examples and foliate decoration (Figure 3.7a-b) and outer end decoration (Figure 3.7c-d) which is often characteristic of a particular cultural style as with Celtic 'trumpet' ends and the extended flaring of spiral ends in some Southwestern Pueblo cultures (Figure 3.7d). These additions to the basic spiral line can cause ambiguity over the start and end points of the spiral line.

81

a.

c.

b.

d.

Figure 3.7 Spirals with end embellishments and central eyes. a. Mycenaean flask, Ashmolean Museum AN1938.196. b. Uzbek tile panel c.1359CE, Victoria and Albert Museum, acc. no. 574 to B-1900. c. feeding bottle from Crete (1300-1190BCE), Ashmolean Museum AN1967.528. d. Zuni olla, New Mexico, Cambridge Museum of Archaeology and Anthropology, acc. no. Z43115. To make an objective assessment of whether these decorative ends should be traced as part of the spiral curve, participants in the perceptual survey were asked to mark the centre points and outer ends of spirals with decorative embellishments. Where changes at the end of a spiral involved the sudden broadening of the spiral path (motifs 62-65), a significant majority of participants (p=<0.001) felt that these elements should be included within the spiral (Figure 3.8a-d). Where the decoration involved a sudden change in direction of an otherwise unchanged path, there was less consensus and the test motifs did not yield significant results (Figure 3.8e-f).

82 Figure 3.8a-f. Placement of end points by survey participants. g. Placement of centre point in a spiral with an eye. (Figure runs over two pages.)

a.

b.

c.

d.

e.

f.

83

g. Figure 3.8g yielded a slight tendency to place the centre in the centre of the eye suggesting that the spiral path around the eye was perceived to be a continuation of the spiral. Only one participant clearly placed the centre of the spiral at the junction between the eye and the spiral path which might imply that they were excluding the eye. Internal and external end points for analysis were placed between the two traced lines to mark the maximum extent of the spiral at the halfway point between the two tracing lines. For spirals with eyes, the internal end point was placed at the point at which the spiral path joined the eye (Figure 3.9).

Figure 3.9 Examples of spiral ends with the tracing lines and end points marked in red. When two spirals were linked together at the outermost end, the end point was taken to be halfway along the line that linked them (Figure 3.10a). The approach chosen for determining the maximum extent of the spiral path of a single spiral was to work from the maximum possible extent of the spiral in both directions. The end of the spiral was taken to be the first point where either: a line perpendicular to the path runs through

84 the incircle1 of the spiral (Figure 3.10b) or the osculating circle (a circle with radius equal to the radius of curvature) of the path encompasses the centre of a spiral which is contiguous with the path, discounting minor breaks in the path as detailed below. Straight line sections were not deemed to have an osculating circle (Figure 3.10c). A second requirement was that the spiral ran in an unchanging direction of turn around the incircle between these two points. If this were not the case, the end point was adjusted inward accordingly (Figure 3.10d). To allow for irregularity, it was not required that either of the first two criteria should pertain along the whole of the spiral path.

End point of the spiral

End point of the spiral

incircle

a.

b. osculating circle

End point of the spiral

End point of the spiral

c.

d.

Figure 3.10 Spiral end point determination. a. End point at midpoint of two linked spirals. b. End point determined by a line through the incircle. c. End point determined by the osculating circle. d. End point determined by the osculating circle and the direction of turn. Further ambiguity occurred with regard to the end point where a spiral joined an unclosed line, other than at its start or end point (Figure 3.11a) or where two or more spirals had sections of their paths in common (Figure 3.11b and c).

1

Although not strictly an incircle, this term was used to refer to the area of the centre of the spiral defined by a line drawn from the innermost tip of the spiral through the end of the inside tracing line (see 3.1.4, p.92).

85

b. c. a. Figure 3.11 Spirals joining a line or sharing part of their curve. a. Detail from a textile design by Guido Marcini, ULITA, acc.no. 2010.295. b. Detail from an Asante wrap, Nigerian, Leeds City Museum. c. Jar fragment, Crete, Ashmolean Museum acc.no. AE.788.

Five tests were used in the perceptual survey to assess whether a consistent choice of path end could be made. The strongest result was shown in the selection of a spiral which had good continuity of curvature over one with a sharp angle in an otherwise curved line (Figure 3.12e). Segregation by curvature or by curve length (Figure 3.12ad) only yielded statistically significant results in the test with the most extreme absolute line length difference (Figure 3.12d) where the longer and more curved path was preferred (p<0.05). However, the longer path end was favoured in all four tests and a gradient of preference was seen according to path length difference which was not apparent in curvature so path length was used as the identifier of the spiral path to be traced.

86

a.

b. c.

d.

e.

Figure 3.12 Tests of path continuity with end point estimations marked. In the light of the perceptual survey results, decorative features which did not follow the spiral path such as the snake's head on the specimen spiral (Figure 3.1, p.71) were excluded from analysis. In spirals with an ambiguous path due to splitting or joining a line, priority was given to the curve which showed best continuity then to the longest spiral curve. Spirals joining a non-spiral line were deemed to continue along that line provided that they later broke away to continue the spiral path and subject to the definitions of end points described above. Identifying the spiral line Two issues arose in the identification of the spiral line, the discrimination of figure from ground and the interpolation of missing sections of the spiral path. Several examples of decoration were found during analysis where there was ambiguity over which line should be traced as the dominant spiral figure (Figure 3.13a). There is no inherent reason to suppose that figure/ground differentiation is culturally independent indeed, experimental results suggest that there is cultural variation in the interpretation of this variable (Parush, Sharoni, Hahn-Markowitz & Katz 2000). Bagley (1996, p.12) noted that there may be deliberate intent to create ambiguity between

87 figure and ground. He cites the spirals on Mimbres pottery as examples of this phenomenon (Figure 3.13b).

a.

b.

Figure 3.13a. Shang bronze vessel with ambiguous figure/ground configuration, Ashmolean Museum acc.no. b7,LI1301.6. b. Mimbres pottery with ambiguous decoration, British Museum, acc. no. Am1979,08.23. Three possible methods were considered for differentiating figure and ground. The first took the spiral which was distinct in colour or patterning from its surroundings as the foreground curve to be traced (Figure 3.14 a-d). This would not always be effective as, on some objects, the ground colour as the dominant spiral form (e.g. Figure 3.14e)

a. Three spirals are counted as distinct from the background.

b. Four spirals are distinguished from the ground by patterning or colour.

d. Four spirals are counted – there is no sense that the dark grey is a ground colour.

Figure 3.14 Figure/ground differentiation.

c. Two foreground spirals

e. Atchana ware goblet, Ashmolean Museum, acc. no. AN1938.201

88 Secondly, where two fields of colour or patterning met in a way that formed a spiral, the intersection point of the two fields could be taken as a single linear spiral (rather than two planar spirals) (Figure 3.15).

Figure 3.15 Peruvian textile with two balanced fields of colour meeting, Manchester Museum acc. no. 0.4036. Finally, when there was no means of identifying a single spiral path by distinction of patterning or colour, the path followed by the tool was taken to be the foreground path. This would mean that the raised line in Figure 3.13a would be traced as the vessel is cast so the raised lines represent incisions into the clay. This approach could be problematic where, what would be taken as the ground of the image was worked leaving the spiral shape to show through from the underlying layer as is found with Mimbres pottery decoration (Bagley 1996, p.13) and may also be seen in some carving techniques. Since no culturally consistent formula could be determined for identifying the foreground spiral path, decisions on this point were made individually taking into account the three methods proposed above and the context provided by other decoration on the object (for example, whether identifiably representational imagery was present which could indicate the foreground decoration) (Figure 3.16).

Figure 3.16 Decorative context used to identify the foreground spiral line of the sampled spiral line (indicated with arrow) as the wider line. Late Shang/Western Zhou hu (ritual vessel), British Museum, 1956,1016.1.

89 Spirals with broken paths due to minor damage; to cutting by the edge of the object or cutting by another decorative element (Figure 3.17) posed two problems in analysis – their identification as spirals and, subsequently the issue of whether and how to interpolate the missing curve section.

b. a. Figure 3.17a. Cypriot painted spiral requiring interpolation due to damage, Medelhavsmuseet, acc.no. E.003:210. b. Cypriot mouthpiece with the path broken by another decorative element, BM acc.no. 1897,0401.376. Ideally, a base data set would include no motifs with any degree of interpolation but this would rarely be practical. Studies of contour completion have shown a greater than 50% identification of broken or interpolated objects when twenty per cent of the outline is present (distributed in fragments around the outline) (Panis, De Winter, Vandekerckhove & Wagemans 2008). Kellman and Shipley (1991) use the concept of "relatability" to describe the ability to join two sections by a smooth monotonic curve including mathematical spirals. This offers an explanation for the ability to interpolate spirals but possibly also suggests that such an interpolated curve is likely to be more inaccurate in decorative spirals where a smooth curve cannot always be relied upon. Also, when two or more potential curve paths are close together assessment of which curve elements should be linked requires considerably more of the curve to be present. This need to consider the context of the missing section means that interpolation by the analyst is more appropriate than attempting to automate assessment at the first stage of identifying the path as a spiral. Tests of curve continuity within the perceptual survey which assessed breakage of the curve with blocks and lines; with sharp or smooth changes in curvature under the obscured section; and with multiple possible continuations of the spiral path showed that there was a significant inclination to continue the spiral across the break point (p=0.003) (Figure 3.18).

90

a.

b.

c.

d.

e.

f.

g.

Figure 3.18 Test motifs for interpolation of the spiral path.

91 The only instance where the majority of participants ended the spiral at the break point (Figure 3.18g) was when a sharp change in curvature was obscured under the block which seems more likely to reflect a desire for good continuity in the spiral than an inability to conceive that the spiral path could continue past the obstruction. When asked to place the centre of an unclosed motif which did not complete more than one rotation, the majority of participants marked the centre of a planar shape rather than the line indicating that they were interpolating a closure for even an irregular shape (Figure 3.19) or even extrapolating a circle from a quadrant given a curve with good continuity of curvature.

Figure 3.19 Test motifs for centre placement of an unclosed shape. To allow for a poorer quality of interpretation in a non-smooth decorative spiral the maximum allowable interpolated section was taken to be 30% by rotation of a single rotation. Spirals with longer contiguous missing sections were excluded from the study. Where the decorative technique employed resulted in a broken path (e.g. Figure 1.5a, p.26), the spiral was deemed to continue as long as the length of the each missing section measured as a straight line did not exceed the maximum spacing between consecutive rotations. In this case, interpolation of sections between decorative points was not recorded. Where longer sections were interpolated these were noted. Kernel regression (Teknomo, 2007) was tested as a means of standardising interpolation during the quantification of the spiral path. Due to the irregularity of decorative spiral curves, this approach showed a deterioration in estimation quality after a few points to a proposed curve which differed so substantially from the spiral path of similar complete spiral motifs that this interpolation method was deemed to

92 be inappropriate to use. For this reason, estimation by the analyst was used to interpolate missing sections of curves. This introduced a greater level of subjectivity into the method than was desirable but allowed the interpolation to be guided by the context of the individual spiral curve using the curvature of the sections preceding and following the missing section and that of coils immediately inside or outside the missing section. Where interpolation arose because of damage, fragmentary traces of the spiral path within the interpolated section could also act as guides. More importantly in respect of decorative spirals, the analyst could make comparison with other spirals on the object and similar spirals on different objects to guide interpolation. 3.1.4 Identification of the centre The method chosen for the identification of the spiral centre was to use the centroid of the polygon created by intersecting the curve with a ray drawn from the start point through the beginning of the innermost tracing line (Figure 3.20). The centroid is the centre of gravity of the central polygon (Kempton 1981, p.155).

Figure 3.20 Spiral centre within the incircle of the specimen spiral. Hel-Or, Peleg and Avnir (1991) demonstrated that the chiral centre of a spiral (the point at which the highest proportion of the curve shows either clockwise or anticlockwise increase in distance from the centre) approximated its mathematical centre better than the centroid, however, the concept of a centre based on the mathematical equation of the spiral is irrelevant to many decorative spirals. For the application of this method it was only necessary that the placement of the centre was consistent and repeatable and, for the functioning of the applet as written, that it was inside the innermost rotation of the spiral. The calculation of the centroid fulfilled both these criteria and could be carried out within the applet more rapidly than the

93 calculation of the chiral centre. For spiraloids the area of the innermost rotation was created by joining the start and end points of the spiraloid. 3.1.5 Precision Tracing sampled the spirals to an average precision of one recorded data point for every 0.038mm of the real curve length. Some longer spirals, mostly from the Cypriot case study, had to be sampled at a lower resolution to prevent the applet crashing. These are noted in the raw data tables in Appendix 2 (Tables A2.2-A2.5). The minimum recording resolution was one point for every 0.14mm of the curve. To ensure maximal comparability of form measurements, each spiral ideally needed the same number of sampling points. The problems with achieving this are discussed in section 3.4.1 (p.101). The effect on precision of the chosen solution of reducing the number of sampled points was to reduce resolution to a minimum of one point for every 0.78mm of the curve. Style variables were measured using the highest resolution available rather than the reduced form resolution. Image distortion tests showed that the contribution of parallax to the margin of error was 0.008mm/1mm at worst. The contribution of erratic hand movement to error was tested on three spirals and an average was found of 0.34mm deviation either side of the true line (Figure 3.21).

Figure 3.21 Digitally created spiral in black and manual trace line shown in red (to scale).

94 Calibration of the digitising tablet within the analysis software environment yielded an error margin of 0.02mm which is the maximum recording resolution of the software. The digitising tablet was therefore not taken to contribute to the margin of error of the analysis. Not including photographic distortion, which would have to be calculated separately for each spiral based on its diameter, the total error margin of the measurements was deemed to be ±0.4mm for the high resolution tracings rising to ±0.5mm for low resolution tracings and to ±1.1mm for form measurements. In general, quantification of variables should operate at a level which accords with the limit of human perception as discrimination beyond this point may introduce spurious differentiation (Wu & Dalal 2005). Size discrimination is a relative judgement which has been found to be equivalent to two percent of the dimension being compared (Stuart, Bossomaier & Johnson 1993) but with many contextual factors affecting perceptions of difference (Smith 1982) and with variation between cultures in relative size judgement (Doherty, Tsuji & Phillips 2008). Because of this it is not possible to make a general statement that this margin of error is not perceptually important. However, it was felt likely that such a difference would only be noticed under close scrutiny and direct comparison and that this behaviour would not be likely to represent the typical use even of meaningful decoration so the margin of error was perceptually acceptable.

3.2 Identification of form Analysis of form was based on sampling points on the primary path – a calculated line running along the centre of the traced spiral. Working from the primary path, the centre of each spiral was identified and the scale and orientation of the spiral were normalised for comparison. 3.2.1 Creation of the primary path The primary path was used to remove from consideration the decorative aspects of the spiral, described as style variables, leaving only its basic shape. Tests were used in the perceptual survey to establish the effect of tracing of the inside, outside or middle line for the assessment of form. Seven motifs tested in the perceptual survey had lines wide enough to assess whether participants were calculating from the inside, outside or centre line when placing end points (Figures 3.8b-d p.82 and 3.12a-d p.86). These

95 tests showed no strong trend in outline preference. Outside was the numerically dominant placing in four tests, middle line twice (once tied with outside) and inside twice. Without clear differentiation, the primary path was deemed to run from the innermost end, through the midpoints between the sampling points of the inner trace line and the closest point on the outer trace line, to the outermost end (Figure 3.22).

Figure 3.22 Specimen spiral with the primary path marked in red. For spirals with central eyes, the primary path was extrapolated around the inside of the eye measuring from the outline half the distance between the two first points of the spiral line (Figure 3.23). This approach did not allow for an increasing or decreasing trend in path width to be reflected in the course of the primary path around the eye. An attempt to extend a statistical interpolation tool – Klasson splines (Klasson 2008; Klasson 2011) - to extrapolation resulted, as with regression based interpolation, in a rapid diminution of prediction quality. Any attempt at statistical extrapolation is liable to be hampered by the irregularity of decorative spirals.

Figure 3.23 Calculation of the primary path around the eye of a spiral.

96 3.2.2 Normalisation Spiral tracing data were normalised prior to analysis. The spirals were aligned along their maximum diameter with the innermost end of the spiral lying on the right of the maximum axis. Where spirals had an identical maximum diameter at more than one location, the diameter which placed the start point closest to the top of the spiral was used. The spirals were reflected, where necessary, to run anticlockwise from the centre. All measurements of dimensions were proportional to the maximum diameter of the spiral. 3.2.3 Quantification of form An advantage of the quantitative recording of decoration is the potential to carry the descriptions obtained across more than one study and to augment a bank of descriptions in future work. To achieve this, an absolute description of those attributes which had been analysed is needed. Pattern matching techniques which provide a measure of relative difference within the current data sample were therefore not appropriate as a means of comparing spiral form. The chosen method of description of spiral form was derived from the ψ-s curve which represents an irregular curve or closed shape as a sequence of tangent angles (ψ) measured at points of known distance along the path (s) (McConnell, Kwok, Curlander, Kober & Pang 1991). The method used in this research differs from that of McConnell et al. (1991) in that a constant path length between tangent angle measurements was assumed. This change was made to accommodate the production of spectral density plots (discussed in section 3.4.1, p.101) as a means of summarising the form data prior to the creation of class groups. Actual rotational distance between measurements showed a standard deviation of 0.009 radians so no substantial inaccuracy was introduced by this assumption of constancy. Since the spiral path segments are plotted as straight lines within the analysis applet, the tangent at the generating points of the primary path was deemed to be a line perpendicular to the bisector of the angle made by the two segments meeting at each point.

97

3.3 Quantification of style Style variables were measured as interval variables to make them compatible with the summarised form data. Orientation, rotational length, curvilinearity and irregularity were all described by a single variable each. Spacing and path width were described with four variables each. 3.3.1 Orientation In choosing the basis for measuring orientation, the most strikingly ‘orientated’ aspect was felt to be the axis of the maximum diameter (Figure 3.24).

Figure 3.24 Maximum diameter of the specimen spiral marked, orientation 1.03 radians from the positive x axis. There were instances where the orientation could not be determined if a single ‘correct’ orientation for the artefact could not be identified, for example, when the spiral was in the centre of a circular dish. Cluster analysis (see section 3.4.3, p.105) requires a measurement to be present for every variable. Accommodating missing values by making orientation a categorical field would limit the types of statistical classification system which could be used. To avoid losing analytical power, two analyses were run for each test which included spirals with missing orientation data. The first excluded the orientation variable. From this, the median orientation value of the classes to which the spirals with missing data had been assigned were taken. These were used as estimated values in the second analysis which included the orientation variable. The estimated values are highlighted in the case study data tables (Appendix 2, Tables A2.2-A2.5). 3.3.2 Rotational length The maximum diameter of the spiral, where available, provides the most appropriate sense of the overall size of the spiral. Calculation of the length of the spiral path is also likely to reflect the size of the spiral so it is more relevant as a perceptual variable to

98 measure rotational length as an angle. Rotational length was measured from the primary path in radians. 3.3.3 Curvilinearity Perceptions of curvilinearity may vary according to viewing distance. For example embroidered spirals which, seen at a likely viewing distance for the garments or furnishing fabrics in use, appeared curvilinear may, when viewed close up, appear angular due to the way in which they are stitched. For consistency and accuracy the spiral was assessed according to its actual shape rather than approximating to a curve since the natural viewing distance is subjective and variable between different types of object. Extrapolating a curve, rather than closely tracing the edges of the spiral path, would introduce error. In assessing the effect of medium or technique on spiral appearance, recording of actual, rather than perceived, curvilinearity would also be important. Curvilinearity was quantified as the proportional length of straight (s) to curved (c) segments given by s/(s+c) using the average values of the inside and outside tracings of the spiral. A sequence of tracing points was deemed to form a single straight segment if the angle between the first pair of points and at least eight subsequent points was less than 0.035 radians (Figure 3.25).

angle less than 0.035rad. from origin straight segment angle greater than 0.035rad. from origin curvilinear

Figure 3.25 Calculation of curvilinearity. 3.3.4 Irregularity Irregularity covered minor variation in the path of the spiral which did not have significant impact on the spiral form. Irregularity was determined by measuring the chirality of inflections in the traced spiral outline. Chirality was calculated using Hel-

99 Or, Peleg and Avnir's (1991) measure which divided the spiral path into segments and measured whether each segment, taken separately, would gather or deflect particles when rotated about the centre by subtracting r1 (the distance from the proposed centre to the start of the segment) from r2 (the distance from centre to the segment end) (Figure 3.26). The count of negative distances was subtracted from the count of positive distances and divided by the total count giving a chirality score. A chirality of one indicated a regular spiral and deviation towards zero indicated increasing irregularity. The average score was found from the tracings of both sides of the spiral path. Interpolated sections of the traced spiral were excluded from assessment of irregularity. r+ve r+ve r-ve r-ve r+ve

centre

Figure 3.26 Calculation of irregularity. Minor damage which does not obscure form may still have a substantial effect on assessment of irregularity. It is therefore necessary to consider if local irregularity was present at the time of making due to decorative intent; or the technology employed; or if it is a consequence of subsequent wear and tear. Areas of obvious minor damage (Figure 3.27) which was not significant enough to require the spiral path to be interpolated were discounted in tracing and represented as far as possible as a smooth curve.

Figure 3.27 Areas of the specimen spiral with irregularity detail lost due to corrosion.

100 This means that irregularity which was present in the original curve will have been lost in these sections. There are probably also some decorative spirals where damage is not identified or artefacts of the manufacturing process are wrongly identified as damage. Analysing only perfect examples of decoration would greatly reduce the sample size so these flaws in the assessment of irregularity are difficult to avoid and the significance of the effect is difficult to quantify. Comparison of the automated mathematical traces, which should be free from irregularity error, with manual traces of the same spirals showed that the process of tracing decreased the irregularity score by an average of 0.34. 3.3.5 Width of paths and spacing between paths Path width and spacing (the distance between the outer edge of one rotation and the nearest inner edge point of the next rotation) were quantified using two criteria – median and trend, derived from normalised measurement of path and spacing distances taken from each tracing point of the inside curve line to the nearest point on the next curve line (Figure 3.28).

Figure 3.28 Measurement lines used for calculation of spacing on the specimen spiral marked in green. The same principle was used for calculation of the primary path. If two or more consecutive rotations of an irregular spiral overlapped, by less than fifty per cent of narrowest path width in the overlapping section, the spacing was regarded as zero and the full path width was measured including the overlapping section. Dividing the path width up would otherwise have disproportionately reduced the average path width. If the paths overlapped by more than fifty per cent of the

101 width, the definition of a spiral was deemed not to have been fulfilled and the curve was excluded from analysis or the end points determined accordingly. The median path and spacing widths gave a sense of the values of path width and spacing as a proportion of spiral size whereas trend indicated change over the course of the spiral. The trend in path width and spacing (i.e. unchanging, increasing or decreasing) was assessed by fitting a third degree polynomial curve to the path width or spacing measurements plotted against the rotational length of the spiral (Figure 3.29). A third degree polynomial was chosen as it accommodated the three most common types of path width or spacing trend – a constant value along the full length of the path; a gradual increase or decrease along the spiral path; and a sudden increase at the end of the spiral path following constant values or gradual change.

a.

b.

Figure 3.29 Specimen spiral - a. Path width measurements (red) and trendline. b. Spacing measurements and trendline. The specimen spiral had path width trend parameters of 0,0,0 and spacing trend parameters of 0,0,0.01. Due to the need to have a value for each variable, median spacing for spiraloids was calculated as the distance between the start and end points of the spiraloid. Trend parameters for spiraloids were recorded as zero.

3.4 Methods of comparison The raw data obtained from the analysis applet were exported into Microsoft Excel, collated and down-sampled for form quantification. Statistical tests were carried out using PASW Statistics 20 (SPSS). Cluster analysis was used to test for significant type groupings within cultures and the results of these analyses used to highlight significant variables and groupings for cross-cultural comparisons. 3.4.1 Summarisation of spiral form data For best comparability, the spirals should be the same rotational length with the same number of sampling points. It was felt better to reduce the longer spirals (down-

102 sampling) rather than to interpolate points into the shorter spirals to raise the number of sampling points closer to that of the longer spirals. Up-sampling (interpolating sampling points into) the shortest spirals sufficiently to match the longer ones would have required that most of the sampling points were interpolated. Down-sampling was achieved by substituting in a single mean value for equal consecutive groups of points to bring the number of sampling points to the nearest integer below 256. The significance of the smoothing effect of down-sampling the longer spirals was offset by calculating irregularity separately using the full tracing data.

1.00 0.50 0.00 -0.50

1 13 25 37 49 61 73 85 97 109 121 133 145 157 169 181 193 205 217 229 241

Sine values of tangents

1.50

-1.00 -1.50

Figure 3.30 Down-sampled sine wave of the specimen spiral. The sine of the tangent angle was used in the analysis of the form data to correct for 360˚ modulation (Figure 3.30). To compare waves directly it needed to be established that waves relating to equivalent sections of the curve, in terms of rotational distance from the centre of the spiral, were being compared. Extrapolation of an irregular decorative curve to its centre would rapidly become highly inaccurate due to the irregularity of the curve. To avoid the need to attempt this, spectral density plots were used in comparison. This analysis technique decomposes the wave into periodic sine and cosine waves (Pennsylvania State University 2013). The spectral density plot shows the component wave frequencies present in the spiral wave and the amplitudes of the waves at each frequency (Figure 3.31) (Pennsylvania State University 2013).

103

Figure 3.31 Spectral density plot of the specimen spiral. A moving average transformation (using a Tukey-Hamming window with a span of three) has been applied to the calculated frequency values in Figure 3.31 creating a smooth density plot and removing high variance which makes the identification of significant frequency peaks difficult (Meko 2013). The window model and span were chosen to allow the frequency of the highest density peak to be precisely defined which was necessary to characterise the expansion rate effectively (Figure 3.32). The decomposition of the wave form allows the frequency of the highest amplitude signals which best represent the gross form of the spiral to be identified.

104

a.

b.

c. Figure 3.32 Spectral density plots. a. The automatically traced hyperbolic spiral at a span of five. b. Manually traced hyperbolic spiral at span of seven for noise minimisation. c. A quadrilateral spiral at a span of seven with higher frequency variance retained. Comparison of the form of the spirals was carried out using the Pearson correlation coefficient. The spectral density plot of each spiral was correlated with every other spiral in the analysis group and these sets of correlation values for each spiral were fed into a cluster analysis to create class groups along with the style data. 3.4.2 Variable weighting It is apparent from studies comparing the influence of pairs of grouping principles that some provide a more powerful grouping stimulus than others (e.g. Ben-Av & Sagi 1995; Palmer & Rock 1994). This evidence suggests that the variables used within the analysis should be weighted according to their influence on the creation of groups. The limited results of the perceptual survey suggest that form is the strongest differentiating factor followed by orientation with the other variables used all having the same modal value (Appendix 1, Table A1.2). However, all significant similarity

105 scores obtained from the survey were high so do not provide a strong basis for segregation. Perceptual studies have also shown that the importance of variables as differentiators depends on the combinations of differing variables; the nature of the motif; and the expression of the variable (e.g. Quinlan & Wilton 1998; Beck 1966a). A further problem in assigning weights to variables is the difference between cultures in the perceived importance of variables. As discussed in Appendix 1, the results of the perceptual survey did not allow cultural differentiation and the majority of identified studies measuring similarity have drawn from North American or European populations with a sufficiently undifferentiated visual culture that they cannot be taken as indicative of any universal norms that might be applied to the weighting of variables. Because it is neither possible to determine the relative perceptual values of the cultures being studied nor to assume that there was commonality between cultures in this respect, it was decided that weighting would not be imposed on the variables. To minimise the effect of differing magnitudes of variation in different measurements, the Z-scores of the variables (representation of the values as a proportion of the standard deviation from the mean (Shennan 1988, p.75)) were used in the creation of cluster groups. 3.4.3 Clustering Hierarchical clustering was used to create class groups of spirals. This is an agglomerative clustering method which merges the spirals into groups until they are all contained within a single group. Of the three types of clustering available in SPSS, this method had advantages over two-step clustering of not needing the number of clusters to be specified in advance and of indicating the amount of difference between clusters in different levels. Also, it does not assume that all classes contain a similar number of members as is the case with k-means clustering (Cleophas 2012). Testing on the Shang-Zhou case study determined the choice of furthest neighbour (CLINK) clustering using squared Euclidean distance as a measure. These methods yielded classes which accorded best with subjective assessment of difference. Hierarchical clustering is dependent on the order in which the cases are arranged. To allow for this, each test was run four times with the spirals arranged in a different random order each time. Clusters which were robust for each run were accepted as valid. Where comparisons were being made, for example between cultures or media,

106 the clustering was carried out with the whole sample randomised and with samples randomised within separate sub-groups. The reordering of the sample rarely changed the resulting classes. Hierarchical clustering imposes a grouping structure on datasets when it may not be appropriate to do so. For this reason subjective judgement of the aesthetic properties of the spirals and consideration of the distance values between groups need to be used to assess whether the clustering is appropriate and to determine the appropriate number of meaningful clusters. Hierarchical clustering is a descriptive method which does not indicate the factors which cause the clusters to be created or necessarily imply that clusters arise from the same underlying cause. Within identified spiral classes discriminant analysis was used to identify which variables were important in creating the group.

3.5 Summary The method proposed in this chapter for classification and analysis of spiral motifs has been designed to cope with the wide variety of spiral shapes encountered in decorative contexts. In attempting to encompass this variation, reference has been made to studies of human perception of form as well as to existing quantitative methods of analysing spirals. Raw data, in the form of images, collected from a range of source types were assessed according to three classes of measurement variable – form, style and interrelationship. The results of multivariate statistical tests applied to these data to identify groupings of types of spiral motif were set alongside contextual information to offer a quantitative means of studying change and diffusion in spiral motifs within and between the two case study regions. The results of these case studies are the focus of Chapter 5.

107

4 Background to the case studies This chapter discusses the choice of case studies and provides context for the studies looking at the structure of the society, the religious practises, known connections with surrounding cultures and the range of techniques and media associated with decorative spirals. Evidence for the existence of representational or symbolic meanings for spirals within the culture is considered and existing typologies of decoration and means of classifying spirals within the decorative canon of the cultures in question are discussed.

4.1 Choice of case studies A problem in the analysis of a specific geometric motif in cross-cultural comparative studies is that it is required to be present in more than one culture whereas the underlying structure of decoration is always measurable. In assessing the movement of decorative elements between cultures, Grabar (2006) highlights the need to consider how significant the foreign influence is and how it is identified – either by consistencies with known external types or by inconsistencies with the internal type. He states that to establish whether outside influence is, firstly, present and, secondly, of more than one-off significance it is necessary to have good typologies for artefacts or designs from the cultures being studied. A similar principle of identifying typing significance would ideally need to be applied in considering the evolution of a motif within a single culture – evolution requires a frequently used motif. The final choice of case studies was driven by several essential requirements for inclusion in the research: spirals must be a commonly used motif in at least one area of the decorative arts of the culture. Serpell (1979) found that cross-cultural comparisons could be distorted by a difference in skill and familiarity with materials and techniques in reproducing designs and the same might be said for motifs - that valid comparison can only be made between cultures where there is a common view that a spiral constitutes a valid and defined decorative motif rather than an occasional and incidental extension of a non-spiral curve. This does not necessarily require an understanding of a spiral as a geometric construct;

108 there must be sufficient data available to make a valid comparison (this requires the artefacts used to have a date and place of origin sufficiently precise to make a comparison of change over time and between cultures at the same time); and a sufficient available sample size must be available for study. Within these constraints, the aim was to cover a range of geographical areas and media and to choose case studies which were independent of each other. The chosen studies are of East Asia focussing on: Japan during the Early-Late Jōmon periods (c.3000-10001), the Shang to Western Zhou Dynasties (c.1800 to 771), and the Eastern Mediterranean and Egyptian area represented by examples centred on: the Dynastic Egyptian Middle Kingdom and First and Second Intermediate Periods (c.2181-1550), Bronze Age Cyprus (c.1600-600).

Peregrine (2001) emphasises the need for random sampling from the population of potentially appropriate cultures. The primary limiting factors in not taking a random sampling approach to the choice of case study were the failure with some potential study areas to access a sufficiently large sample of provenanced material e.g. Maori wood carving, Chavin (Peruvian) textiles and the painted pottery of a number of Balkan Neolithic cultures. The inclusion of a greater range of case study areas would improve the geographical and temporal scope of this study and it is hoped that the range may be expanded with future work.

4.2 Culturally dependent interpretation of decorative spirals In some ways, this study harks back to early approaches of taking objects and, specifically, particular elements of their decoration out of their cultural context. There is a considerable body of literature that argues against the study of individual design elements in isolation from their context. Washburn (1983a, p.3) states that to isolate a motif from its cultural context is to cause it to lose meaning. Such an approach may 1

All dates relating to the case studies are BCE unless otherwise stated.

109 also cause potentially irrelevant meanings to be imposed by an outside observer (Casrnir 1999). Indeed, there is no way of knowing if the motif identified by the analyst was perceived as a valid decorative unit in isolation by its creator (Jernigan 1986) and this problem may be especially significant in the consideration of geometric constructs found within representational contexts (e.g. Figure 4.1) or where possible symbolic meanings are unknown.

Figure 4.1 Spirals in representational context in the eyes and tail of a frog on a 1st century bell from Yunnan province, China, British Museum, acc. no. 1948.1013.15. Both O'Sullivan (1986) and Bradley (1997, p.5), in their discussions of North Atlantic rock art, note that the composition of the decoration can be equally or more important than the motifs represented. O’Sullivan (1986) expands on the problems of considering motifs in isolation in observing that this approach can cause discrete motifs which are readily describable to be favoured over apparently nebulous background decoration. A risk inherent in all classification systems which focus on a single construct or element of a design is giving the aspect under consideration undue prominence to motifs which may only be a minor element of the design aesthetic under consideration. This risk applies particularly to motifs as the single motif may only be a small element of a larger design. It must also be remembered that even apparently objective classification techniques may still be subject to cultural variation in their application. In discussing the appliqué textiles of the Cuna, Hirschfeld (1977) noted that most of the grossly symmetrical designs had some minor elements of asymmetry. Having identified a preference for complex asymmetrical designs within the culture, Hirschfeld believed that these asymmetries should be regarded as the conscious production of complex noise rather than errors so to assess only the gross symmetrical structure of the design may be to miss relevant detail.

110 In spite of the problems, raised above, in considering motifs in isolation, the decision was made to asses an isolated motif in this research because the focus was on developing a method of analysis and a system of classification to test if and what useful information could be extracted from a single motif type, not to suggest that this should be considered in isolation from all other available information in practical application. The classification approach used in this research is put forward as an alternative to the externally structured classification system commonly applied to decorative spiral motifs. It is suggested that more appropriate groups can be built from the properties of the decoration itself but not that meaning can be assigned by this method. The variables used in analysis are, in themselves, external constructs which may or may not have guided the production of the motif and the relevance of which may vary from culture to culture. Variables for testing were sought that have been shown to be valid perceptual differentiators to attempt to mitigate this risk but the limited temporal and geographical approach to this validation has not eliminated it. Within the scope of this research, the limitations raised above in isolating a single motif type for analysis will vary for the different research questions. In making comparison with EVP or naturally occurring spiral forms, the potential overvaluing of spiral motifs and their isolation from a decorative context does not pose a significant problem since it is the spiral in isolation which is being considered. However, attempts to create meaningful type classes based solely on spiral motifs may be hampered by analysing only a single component of a design scheme and it would be remarkable if they created a typology which reflected classifications drawn up by analysing the object and its decoration as a whole. The question raised by Jernigan of the validity of the motif as a decorative unit is addressed by requiring its presence as a common element of the decorative scheme of the cultures in question although this does limit the scope of applicability of any single motif. An advantage of breaking down the description of a motif is that it is possible to identify aspects which are consistently similar within a culture or between cultures. If the expression of a measured variable of geometric decoration can be shown to be associated with patterning consistent with another social or cultural aspect of a

111 society, this can be used as a basis for broader suppositions that individual perceptual values which guide the decoration production are associated with the social or cultural concept which is the focus of study provided that this commonality can be demonstrated across a sufficient range of societies. It had initially been intended that this study should focus only on spirals in nonrepresentational decorative contexts. It has, however, become apparent that assumptions of the presence or absence of pictorial or symbolic meaning in decoration (particularly for prehistoric cultures lacking additional sources of interpretive information) placed arbitrary and potentially biased limits on the inclusion or exclusion of data. The strength of representational response to the survey motifs was striking. Even though the focus of the survey was on motifs as geometric rather than representational constructs, only two of the 22 motifs which participants were asked to describe did not attract at least one representational interpretation. This shows a strong inclination amongst survey participants to assign representational meaning even to highly distorted and abstracted shapes. Test motif 97 (Figure 4.2a), for example was described as a kite/broken kite (three times), sleeping penguin (once), suit (once), coffin/broken coffin (twice) and a destroyed paperclip (once). The assigned meanings included some culturally nuanced responses such as the description of test motifs 100b and 100c as Pac-Man (by six participants) or 100b as the Utility Clothing Scheme mark by one participant (Figure 4.2b and c).

c. b. Figure 4.2 Perceptual survey test motifs 97, 100b and 100c. a.

112 The inclusion of representative motifs is also appropriate in searching for the presence of EVP-inspired decoration since this may have been interpreted in a representational manner in explaining the visual effects. Representational decoration may introduce non-aesthetic, factors into the estimation of similarity and the creation of groups (Goldstone 1995). Vessel and Rubin (2010) found that aesthetic preferences varied less between individuals when they were asked to judge representational imagery than with abstract images and hypothesised that this was an effect of common experiential and emotional associations overriding a purely aesthetic reaction. All spiral decoration found within the cultures and periods in question was included in the sample group with the exception of spirals which appeared in the context of a known pictogram or writing system. There is potential for overlap in interpretation between writing and decorative symbols. In giving a free response to describing the shape of a motif in the perceptual survey (Appendix 1, Table A1.3), eight participants identified abstract motifs as letters, numbers or musical notation. No other symbolic meanings were attributed to motifs. However, the inclusion of writing within the study introduces such a an large additional field that writing or equivalent notation was excluded from the scope of this research. The identification of writing raises potential problems of bias but the risk of ambiguity was felt to be less than that between abstract and representational imagery. Emotional responses also vary between cultures; Pittard Ewing and Jevons (2007) commented that angular shapes could be viewed as hostile or aggressive forms societies whereas Fischer (1961) assigned a male and female symbolism to angular and curvilinear shapes and therefore associated them with the expression of sexuality. When asked to provide free description of test motifs in the perceptual survey (Appendix 1, Table A1.4), four participants used emotive language relating to the perceived attractiveness of the objects. The clearly emotionally charged terms were all positive 'beautiful', 'pleasing', 'elegant' and 'lovely'. All four participants gave positive responses to Figure 4.3a and one also to Figure 4.3b.

113

a. b. Figure 4.3a. Test motif 157 - ironwork scroll in railings. b. Test motif 157b - modern ceramic plate from Morocco with spiraloids. It is not clear why the two motifs above were described using emotive language. They were the only two of the five free description motifs to attract descriptions relating them to flower or plant forms but one of the participants who found Figure 4.3a attractive described it as a wave rather than a plant form and Figure 4.3b received most responses describing it as plant or flower. There is no indication whether cultural familiarity affected the nature of emotional responses. The ironwork spiral might be expected to be more familiar to the, participants (three of whom were British, the fourth did not state their nationality) than the spiraloid whorl with a Moroccan origin. An emotive response in description of the motifs had not been expected and would be an interesting area to explore further. An additional consideration raised in the analysis of geometric motifs in isolation is that the cultural significance placed on a geometric motif may be highly variable. In some cases the spiral may represent a key symbolic element of a design whereas in other instances it may be no more than a decorative means of ending a trailing line with no special significance placed on it. To compare both within the same classification system cannot, in any culturally meaningful way, be considered a comparison of like with like. This difference could lead to spurious comparisons being drawn and care would need to be taken in choosing comparative studies to subjectively assess context and judge equivalence. For other questions the potential to assess the ornamental and the meaningful might be turned to an advantage by allowing consideration of the differences between the two forms. One might hypothesise that decoration would be executed with less care and therefore greater variation than meaningful symbols or alternatively that meaningful symbols would

114 occur more often and therefore be subject to greater elaboration and variation. A common quantification system could test this or consider whether movement of motifs between cultures affected their presentation when in one instance the motif had meaning and in the other a purely ornamental function. The inclusion of spirals identified in any context almost certainly incurs the problem of identifying as spirals a greater range of spiral types than the artisans may be aware of producing or intending to produce but it was felt less likely to introduce distortions into the sample to treat the spirals as meaningless for the purposes of sampling than to attempt to edit the sample on the basis of incomplete or absent interpretive information. It must therefore be remembered that the application of geometric classification methods cannot necessarily be taken as an insight into the mind of the designer but this does not make the application of geometric criteria in analysis necessarily invalid. The producers of the decoration may still conform to a consistent set of characteristics in creating a symbol which is coincidentally spiral so still allow consistent type classes to be created from spirals and in the abstraction of a symbol or natural form to represent it in decoration it is still as probable that the same perceptual values which guide the creation of a pleasing decoration will apply as when creating decoration with no meaning attached.

4.3 Background to the case studies For each group of case studies, the social and political structure of the societies and major changes in these within the period of study have been briefly summarised. Known external trade and cultural contacts are also noted. The focus of these case study introductions is on the contexts in which spiral decoration occurs within the culture looking at the range of media; the use of spirals in representational contexts; and the techniques used to produce spiral motifs. Where appropriate, existing typologies which make reference to spirals are recorded. 4.3.1 East Asia This group of case studies analyses spiral decoration on pottery vessels from Neolithic Japan and on Shang Dynasty bronze vessels. These studies are rather more disparate in location and in temporal continuity than is ideal. Habu (2004 pp.34-35) draws

115 cultural links, predating the period of this research from South China to the Amur River Valley to Japan which do not encompass the more inland regions of China included in this study so not directly linking Jōmon culture to the Chinese study areas. However, given the proximity of Japan to the Chinese coast it would be extremely surprising if there were not some form of contact between the Chinese east coast and Japan within the case study period and cowrie shells have been found in large numbers as grave goods in the Shang capital of Anyang (Chang 1980, p.88). These indicate that contact and the movement of desirable goods across the full geographical area of the case studies would have been possible. 4.3.1.1 Ceramic vessels and figurines of Jōmon Japan The Jōmon period of Japan is the earliest period studied in this thesis. The earliest identified evidence of Jōmon culture dates back to 12500BCE (Table 4.1), having developed from an indigenous Palaeolithic culture (Habu 2004, p.3). This study focusses on the Early to Late Jōmon periods from 5000-1000BCE. Period Incipient Jōmon Initial Jōmon Early Jōmon Middle Jōmon Late Jōmon Final Jōmon

Date 12500-8000 8000-5000 5000-2500 2500-1500 1500-1000 1000-300

Spirals sampled no spiral decoration found no spiral decoration found 3 spirals 80 spirals 11 spirals no spiral decoration analysed Table 4.1 Jōmon Period divisions (Kaner 2009, p.169) and numbers of spirals sampled from each period. The Jōmon employed Neolithic technology in pottery and reached population densities typically associated with agricultural communities. However, foraging retained significant economic importance and sedentism was not generally established (Kaner 2009, p.15; Habu 2004, pp.132-3). Marine resources were important in the Jōmon economy and a number of the known Jōmon sites are represented by shell middens which appear to have begun as specialist processing sites in temporary seasonal occupation but later grew into permanent settlements (Matsui 2001, p.120; Habu 2004, p.58). The Early and Middle Jōmon periods appear to have coincided with a period of relative warming with a subsequent period of cooling into the Late Jōmon (Habu 2004, pp.42-43). The warmer climatic period allowed the exploitation of an

116 increased range of wild food plants, nuts being of greatest importance (Habu 2004, p.45). Sika deer and boar were the most commonly hunted mammals (Habu 2004, pp.45, 59). Poorer climatic conditions at the start of the Late Jōmon period were associated with a reduction in population and a move of settlement to higher ground (Kaner 2009, p.44). From the Early Jōmon period, there is evidence of some plant cultivation including gourd, beans, millet, buckwheat, burdock and rice (Habu 2004, p.59). However, the adoption of agriculture as the main substitute method did not occur until the start of the subsequent Yayoi period when both bronze and iron working were introduced from Korea (Habu 2004, p.50). Ceramics were the main medium for decoration in the Jōmon culture although carved bone, horn and stone have been found in small quantities (Kidder 1964, p.14). This case study analysed spirals found on vessels and on dogū – ceramic figurines. Jōmon ceramic decoration is entirely textural with no evidence of painting having been found. The earliest decorated Jōmon pottery had clay applied in strips or beads (Habu 2004, p.28; Kidder & Esaka 1968, p.150). This technique continued to be employed on vessels throughout the Jōmon period with the addition of decoration created by finger-working (Kidder & Esaka 1968, p.154); deep ridging; or point marks created with a bamboo tools (Kidder & Esaka 1968, p.190); twisted cord impressions in closely packed parallel lines (the decorative technique after which the Jōmon culture is named); and the use of sculptural elements (see Kaner 2009, cat.no.41 for sculpted example). Dogū decorative styles followed those found on pottery and some overlap between the two object types is found in the representation of human figures on vessels (Kaner 2009). During the Middle Jōmon period, dogū became more technologically complex. They developed from flat slabs forms designed to be held up with props or strings to being self-supporting. Some of the more complex figures were constructed with hollow heads (Kaner 2009, pp.41-44). The disruption to the settled way of life at the end of the Middle Jōmon period arising from climatic change coincided with a reduction in the production of dogū with new designs being introduced at the end of the Late Jōmon and into the Final Jōmon (Kaner 2009, p.44). Vessels developed into highly elaborate forms with an emphasis on decoration over function by the Middle Jōmon period possibly reflecting a change from functional cooking vessels to ritual objects (Foster McCarter 2007, p.104). In this context, it is

117 suggested that much of the decoration on the pots is of representational or symbolic significance and possibly relates to ritual celebrating the transformation of food from raw to cooked, although the meanings of the representations can only be speculated on (British Museum 2013). Ritual dancing is believed to be depicted on some vessels and a possible interpretation of a bulbous vessel of the Middle Jōmon period is as a drum with pierced holes around the top for attaching a skin and a dancer depicted on the side (Takashi 2009, pp.136, 140). Kaner (2009, p.32) cites ceramic models of mushrooms as possible evidence for the consumption of psychoactive drugs in rituals. The primary use of Dogū was in ritual. The majority of dogū are female and may represent an earth mother figure (Kaner 2009, p.52). Ritual practise involving dogū appears to have involved breaking and scattering the figures, depositing concentrated broken fragments in a cemetery or pit burial of dogū accompanying a body or in isolation (Kaner 2009, pp.50-51). Both vessels and dogū show considerable regional variation; Habu (2004, pp.39, 21011) provides a chart indicating some of the regional and temporal variation in Jōmon pottery styles. The internal regional variation is not indicative of cultural isolation. Movement of jade and obsidian provides evidence of extensive trade networks having been established across the Japanese archipelago by the Middle Jōmon period transporting raw materials and finished goods (Habu, 2004, pp.3, 223-225). Contacts outside the Japanese archipelago from a very early period have also been proposed. Early lithic assemblages in northern Japan resemble those found in the Amur River Valley Basin in south-eastern Russia along with pottery which may predate Jōmon examples and it is argued that ceramic technology may have come from this region into the north of Japan (Habu 2004, pp.32-34). South-eastern Russian contact has been proposed at the start of the Jōmon period and the change at the end of the Jōmon period to the Yayoi culture appears to have been stimulated by the arrival of Korean migrants (Noma 1966, p.20). However, there appears to be little evidence of foreign influence during the Jōmon period (Noma 1966, p.286). Most Early and Middle Jōmon pottery has been found in central Japan in the mountainous region of Honshū (Noma 1966, p.11). Noma suggests that this would have created cultural isolation away from possible foreign contacts. A study of the diet of Honshū communities indicated little input of marine resources (Habu 2004, p.75).

118 During the Late Jōmon and, more noticeably, during the Final Jōmon, a Korean influence seems to have returned and can be seen in pottery styles with a resulting simplification of decoration (Habu 2004, p.208). Given the long distance contact at either end of the Jōmon period, the isolation from contact outside the archipelago for much of the Jōmon period seems surprising but does seem indicated by the cultural record. Spirals in Jōmon decorative culture Spirals appear on both dogū and vessels at the end of the Early Jōmon period. On dogū, spiral motifs continue to be found on Final Jōmon figurines which are outside the scope of this study. On pottery spiral decoration is dominant in the Middle Jōmon period although it must be noted in this regard that the vessels analysed in the study were less precisely dated than the dogū. The spiral decoration found on vessels and dogū is markedly different. On dogū, spirals tend to be densely coiled with a narrow path width and to have greater rotational length than the large spirals of high expansion rate which dominate on pottery vessels. Only one spiral motif seems to overlap both object types, a spiral of low rotational length seen on the head of dogū and, around the very top of vessels. On pottery vessels, snakes are thought to be an early motif and it has been suggested that many early curvilinear forms represent snakes (Figure 4.4a) (Kidder & Esaka 1968, p.94). Some examples of spiral motifs have snake heads and there also appears to be a split tail form characteristic of snake representations but many similar forms are highly abstracted with no apparent representational features. These snake forms were later transformed into other representational motifs including highly abstracted birds with spirals present at the tips of the wings and forming the head or eye (Figure 4.4b) (Kidder & Esaka 1968, p.150) and, as mentioned above, dancing humans with linked hands. One example of a pottery drum is particularly interesting as a possible link between trance states and EVP in that it depicts a discrete spiral alongside a dancing figure (Takashi 2009, p.140).

119

b. a. Figure 4.4a. Possible snake motif (indicated by arrow) on a Middle Jōmon pot, Tsunan Town Board of Education. b. Bird motif drawn from a Middle Jōmon pot. Spirals on dogū have been used in several representational contexts. A recurring example is in the depiction of a spiral as hair or a headpiece. An alternative interpretation of these spirals is, again, as a snake (Kaner 2009, p.55). It has been conjectured that the snake figure represents an earth mother and that this motif was associated with the nurturing of food explaining its occurrence on vessels (Kaner 2009, p.55), although that would make their placement on the heads of humanoid representations of a goddess figure incongruous. Besides placement on the forehead of figures, spirals on dogū are particularly placed in the centre or the back of the figure usually as a single motif (e.g. Kaner 2009 cat. nos. 7, 21) but, in one instance, as a running line of spirals (Kaner 2009, cat. no. 61). A number of spiral motifs are depicted over limb joints possibly as a depiction of decorated clothing (Kaner 2009, p.30). The objects included in this case study were well provenanced with only two lacking a place of origin and all having dates assigned within a cultural period (Appendix 2, Table A2.2). Final Jōmon Period spiral decoration was not included within this research. This did not affect the decoration analysed on pots which was concentrated in the Middle Jōmon period but did exclude some dogū with spiral decoration. The spirals sampled in this case study were drawn from 48 vessels, one lamp and 13 dogū. Most analysed spirals were in a good state of preservation with little chipping or degradation of the edges; 24 spirals had interpolated sections up to a maximum of 28% of the total length. Some large coiled snake motifs were the only, notable, exclusion from analysis; since they often consisted of multiple curves running in

120 parallel from a single centre (Figure 4.4a), they only inconsistently fulfil the minimum rotation criterion of spirals/spiraloids so some examples were excluded that could be seen to form part of an aesthetic continuum with the analysed examples. 4.3.1.2 The Shang Dynasty and Western Zhou Period The main body of data for this case study was drawn from Shang Dynasty (c.17661027) bronze vessels, running on into the subsequent Western Zhou Dynasty (1046771). The Shang Dynasty was centred on the lower Yellow River Valley. The dynasty is divided into three political periods – Erligang, Late Erligang and Anyang (Chang 1980, p.322). In this region, millet and rice were the principal crops with dogs, cattle, sheep, pigs, water buffalo, Pere David's deer and (imported) horses being the main economically valuable animals (Lu 2005, p.51; Chang 1980 Ch.2). The settled agricultural economy of the Yellow River Valley allowed the development from segregated villages into clustered village groups and later the development of a state system with the rise of the Xia Dynasty (Zhimin 1988, Chang 1980, pp.287,361). The Shang Dynasty represents a continuation of a ruling line, the Tzu clan, from the predynastic period in a hierarchical structure with a capital and local rulers in walled towns (Chang 1980, pp.1-5, 158-162). Mutual influence between contemporaneous Yellow River Valley cultures can be seen in the similar forms of some bronze vessels (Zhimin 1988). Potential copper ore sources and war raids to Kung Fang in northern Shaanxi suggest contact within a c.200 mile radius was viable and high status cowrie shells must have been brought inland at least a similar distance (Chang 1980, pp.151-5). Salt was also traded outside the Shang area (Chang 1980, p.258). By this period there is evidence of the exchange from Central Asia to China along the jade route (Wood 2002, p.26). Shang cultural influence can be identified as far north as Beijing to border the Lower Xiajiadian culture during the Erligang period (Chang 1980, p.294). Allan (2007) argues that Erlitou pottery motifs may have influenced Lower Xiajiadian painted pottery decoration. To the south, the early Shang settlement area was bounded by the Yangtze River with evidence of cultural contacts again being present in the bordering Jiangxi Culture. Some Shang vessel forms spread southwards and glazed, stamped geometric pottery entered Shang areas from cultures to the south (Chang 1980, pp.305-6). Stamped

121 geometric pottery included spirals amongst its characteristic motifs in later phases (Chui-Mei 1984) which very strongly resemble some of the leiwen spirals seen on Shang bronze vessels. The Western Zhou Dynasty began with the conquest of Shang around 1122. The temporal overlap and evidence of cultural contact between the Shang and Zhou permitted significant continuity of bronze vessel styles between the Shang and Western Zhou cultures and late Shang and early Western Zhou bronzes can be difficult to differentiate stylistically (Chang 1980, p.21). The Zhou state structure developed from that of the Shang but was made more systematic during the Zhou Dynasty. A system of local administration was formalised with rulers being given recorded mandates to rule by the central authority (Allan 2007). Bronze working was known in eastern China by the end of the third millennium (Allan 2007). Bronze was used in the construction of weaponry, harness and vehicle fittings and vessels. The earliest evidence for the production of bronze ritual vessels comes from the Erlitou period, which immediately preceded the Shang Dynasty (Allan 2007). The Shang Dynasty shows an increase in the amount and detail of decoration on bronze vessels (Chang 1980, p.233) and an expansion in the range of types of vessel from wine vessels in the Erlitou period to the inclusion of food and water vessels by the Late Shang (Allan 2007). Bronze casting centres have been found at Hsiao-t'un and in the region of Miao-p'u-pei-ti/Hsüeh-chia-chuang. It is suggested that these centres developed contrasting styles (Chang 1980, pp.28, 233). Shang and Western Zhou bronzes were cast in clay piece moulds decorated by carving an initial stage of decoration into a model of the vessel then creating an impression of the model in a mould and finishing the decoration by carving into the inside of the mould (Chang 1980, p.233; White & Otsuka 1993, p.15). Pottery of the Shang Dynasty continued the wheel technology developed in the earlier Longshan Culture (Chang 1980, pp.234-5, 340-1). New developments in the Shang period, found on high-status White Pottery, were the introduction of glaze and of finely engraved decoration. Cord marked and stamped decoration continued in use from earlier cultures on lower-status pottery (Chang 1980, pp.234-5, 340-1).

122 Shang Dynasty bronze vessels are accepted to be associated with ritual practises (e.g. Allan 2007) so the decoration on them might be expected to reflect religious beliefs and practises. Keightley (1978) stresses the ordered nature of Shang religion and suggests that a similar desire for order guides the highly geometric and abstracted decorative style. The structure of Shang religion placed worship of the god Ti at the highest level, through the intercession by royal ancestors and the living king who took on a magician or shaman role in reading oracles (Keightley 1978, Tong 2002). Evidence for Shang Dynasty religious practises from historical and archaeological sources shows the use of divination and oracle bones, a practice involving the heating of bone and studying cracking patterns and of requests for divine intervention made by animal sacrifice (Keightley 1978; Tong 2002). The Guo yu text (compiled in the 4th century BCE) refers to individuals performing religious rites for themselves but the Shang Dynasty provides the first written evidence of magicians or shamans and suggests that people assuming these roles would have to be susceptible to being possessed by a divinity (Tong 2002). Similarly, Keightley (1978) describes Pin ceremonies in which the king received an ancestral spirit 'as a guest'. The Shang/Zhou case study offers the strongest evidence of all the cultures considered in this research of the possible use of psychoactive drugs capable of inducing hallucinations. Hemp is frequently referred to as a textile and food crop in the Shih Ching, written in the Zhou Dynasty (Li 1974). References to cannabis as a psychoactive drug or medicine post-date the period of this case study. An Eastern Zhou period text refers to hemp in a negative context which is attributed to its depressant properties and the Shen-nung pen-ts'ao ching, a 5th-7th century CE herbal compiled from earlier sources, states that cannabis "if taken in excess will produce hallucinations. If taken over a long term, it makes one communicate with spirits and lightens one's body." (Li 1974, pp.445-446; Prioreschi 1995, p.154). Texts relating to the Zhou Dynasty (the Tongdian compiled in 735-812 and the Zhou li, 4th-3rd centuries BCE) describe ritual dancing and drumming as a continuation of older practises which retained religious significance in that period (Tong 2002). Spirals in Shang Dynasty and Western Zhou Period decorative culture Spiral decoration in this period shows continuity from Erlitou period woodcarving and lacquer work through Erligang cast bronze vessels. Spirals are found on a wide range

123 of types of object and in multiple media including bone plaques, ceramic vessels of the White Pottery style, stone figures, wood (surviving as impressions in clay) and on bronze goods (Loehr 1953; Chang 1980, pp.114-117). White Pottery has been found in storage pits at Hsiao-t'un (Chang 1980, p.98) implying day-to-day usage whereas bronze vessels, in the large majority of cases, have been found as grave goods from high status sites (Chang 1980, pp.132, 205, Allan 2007). Loehr (1953) suggests a direct link through White Pottery between decorative motifs of Machang painted vessels and Shang bronzes. However, Zhimin (1988) suggests that the origins of the Shang cultural tradition were more localised and ran back through the Longshan, Yangshao and Peiligang Cultures within the middle Yellow River Valley area. Chang (1980, p.284) also proposes a local cultural succession through the Yangshao, Longshan, Xia and Shang cultures. Typical Shang Dynasty decoration is an animal style which, by the Late Shang, showed a limited range of regional variation but a consistent set of motif types across the Shang cultural area, prominently featuring taotie, dragons and birds (Figures 4.5 and 4.6) (Allan, 2007).

a. b. Figure 4.5 Shang/Western Zhou Dynasty animalistic decoration. a. A taotie mask detail from a late Shang Dynasty jia, British Museum, acc.no. 1945,1017.191. b. A long, one of several types of dragon depicted in the period, British Museum, acc.no. 1952,1216.1. Spirals motifs are used in a small group of roles in representational depictions – as horns, ears, tails, joints (at the tops of legs and wings), and as surface patterning across the bodies of the animals. Depending on its use, the spiral form appears remarkably consistent in each type of use even when the overall decorative style of vessels differs and when transferring across different media (Figure 4.6). The ground 'thunder cloud'

124 (leiwen) spiral (Fig 4.7) which is characteristic of Late Shang (Yinxi Period) bronze decoration is found on Erligang vessels as a border motif and, earlier, as a pottery motif during the Erlitou I period (Allan 2007, Loehr 1953).

b. a. Figure 4.6 Consistency in use of spirals in depictions of birds' wings in different materials: a. Late Shang jade goose, British Museum, acc.no.1945,1017.40. b. On a Shang bronze vessel, British Museum, acc.no. 1936,1118.4

Figure 4.7 Background leiwen patterning on a Shang Dynasty vessel, British Museum acc. no. 1957,0221.1. A typology by Max Loehr of the decoration of bronze vessels which is based on the shape of the vessels and the type and arrangement of the decoration in an evolutionary type series identifies identifies spiral shape as an important differentiator, particularly between the final three styles III, IV and V. In style III, patterns of flattened spirals on a plane with the rest of the decoration begin to emerge (Loehr 1953). In style IV leiwen appear for the first time and spirals are separated from other motifs as a background pattern (Loehr 1953). In style V, non-spiral motifs are raised in relief above the spiral ground which is sometimes omitted (Loehr 1968 cited in Chang 1980 pp.28-30). Chang (1980, pp.81, 264-267) proposed that Loehr's styles IV and V

125 developed around the time of the establishment of the Shang capital near Anyang and that earlier styles may have been pre-dynastic in origin; as they are already present in the earliest phase Shang site at Erligang. Karlgren also identified spirals as amongst the motifs characterising different decorative styles on Shang bronzes (Chang 1980, p.27). This case study used only vessels as examples. The spirals sampled in this case study were drawn from 51 vessels (Appendix 2, Table A2.3). 41 dated to the Shang Dynasty, seven to the Western Zhou and two were undetermined. 18 types of vessels were sampled (Table 4.2). vessel type

number sampled 2

pou (wine vessel) one example described as pou/weng gu (goblet) 3 jue (spouted wine vessel) 4 li ding (tripod vessel) 1 li (3-lobed tripod vessel) 3 zun (wine jar) 5 ding (food/cooking vessel) 10 fangzun (wine vessel, rectangular cross 1 section with flared lip) grain vessel unspecified type 1 jia (tripod wine vessel with round mouth) 2 fang yi (wine vessel with rectangular cross 2 section and pointed lid) he (3-lobed tripod wine vessel) 2 gui (food vessel) 6 hu (wine vessel) 2 you (wine vessel) 4 pan (shallow water basin with a foot) 1 jiao (wine vessel) 1 zhi (wine vessel) 1 Table 4.2 Types of Shang Dynasty vessel from which spirals were sampled. Although corrosion had damaged some decoration, the spirals sampled were generally in a good state of preservation. There were very few vessels from which no spirals could be analysed. 23 of the analysed spirals included some degree of interpolation of the path up to a maximum of 26% of the length. The provenance of the vessels studied was very poorly resolved in terms both of date and place of origin with only five having an, even approximate, place of origin and only 21 having a time resolution

126 less than the extent of a full dynasty. Three vessels in the sample came from Anyang. Yinxu, the final Shang Dynasty capital and the site of the royal mausoleums is located in the Anyang region (Chang 1980, p.70; Tang, Jing & Rapp 2000). It is not clear if the vessels were found in the excavation of the city itself or the broader Anyang region which includes other major urban sites and is possibly the location of some earlier Shang capital cities (Tang et al. 2000). This lack of provenance poses a considerable restriction on the study, creating type clusters and in the conclusions that can be drawn from the results. Regional style differences cannot be segregated. 4.3.2 The Eastern Mediterranean and Egyptian area From the Neolithic into the Classical Greek period, spirals were a frequently occurring motif in a large area centred on the Eastern Mediterranean and Egypt. A further link has been suggested to the spiral decoration of Danubian Neolithic pottery (Crowley 1989, p.105) through the Cyclades and from there throughout the Aegean/Eastern Mediterranean (Shaw 1970). The number of cultures within this area with spirals as a notable element of their decorative arts creates a difficulty in determining the origins of spiral motifs and a number of theories have been put forward in this extensively studied area (e.g. Crowley 1989; Tufnell & Ward 1966; Ward 1902). Although the Cypriot sample in this study is of a later period than the Egyptian case study, both cultures have had a common source for spiral decoration suggested in transmission of textiles and pottery from Crete and mainland Egypt (Barber 1998; Steel 2004, pp.17, 154; David 2002, p.153). Shaw (1970) noted links in spirals as pottery decoration between the Minoan ivy leaf motif, which first appears in the Early Minoan III period, and Egyptian heart-shaped spiral pairs. Shaw (1970) argues that this spread from Crete to Egypt in or slightly before the Middle Minoan I period in woven patterning of cloth. Along with Marinatos (1951), Shaw suggests that the later naturalisation of the heart-shaped spiral motif to a more leaf-like form combined with a central flower shoot in Aegean decoration towards the end of the Middle Minoan period reflects an Egyptian influence from lotus or papyrus representations. By the Egyptian Middle Kingdom and Cypriot Middle Bronze Age, Cyprus was well established as a trading post for contacts between the Near East, the Western Aegean and Egypt (Steel 2004, p.170). These close links over a long period suggest that a commonality in spiral forms

127 might be expected between the two case studies providing a contrast with the Shang/Western Zhou and Jōmon case studies. 4.3.2.1 Egyptian scarabs and scaraboids The Middle Kingdom (Table 4.3) represented a time of reunification of Egypt under a single king following the fragmentation of the First Intermediate Period into autonomous regions under a purely notional monarch (Grajetzki 2006, pp.3, 7). The new ruling group arose from Thebes whose rulers gradually assumed control of the whole country during the Eleventh Dynasty (Grajetzki 2006, p.10) although regional rulers remained powerful outside parts of southern Egypt in the early Middle Kingdom (Grajetzki 2006, pp.97-99). Social stratification is shown in both formal divisions of ranks amongst the ruling group and evidence from settlements of the organisation of towns into areas of small (manual labourers') and large (officials') houses and there is little evidence for social mobility (Grajetzki 2006, pp. 139-142). However, social divisions between professions in the mass of the population are less clear (Grajetzki 2006, p.143). The Middle Kingdom ended with fragmentation of central rule at the end of the thirteenth dynasty giving rise to the Second Intermediate Period (Grajetzki 2006, p.63). Thirteenth Dynasty kings appear to have only had strong control over Upper Egypt with Lower Egypt coming under the control of an incoming population from southern Palestine (Grajetzki 2006, pp.64, 73). An advance of the Nubian state from the south further reduced the kings' control (Murnane 1997, p.30). Over the following century the eastern invaders became established in the Delta settlement of Avaris as the Hyksos rulers and gained functional rule over Upper Egypt until the renewed advance of the Upper Kingdom vassal states established the first New Kingdom dynasty (Murnane 1997, p.31).

128 Date

Period

2181

First Intermediate Period

21602025 20251995 19761794

Dynasty no. of analysed spirals VII 2 VIII IX 3 X XI 1

Middle Kingdom

33 XII

2 4

XIII

1

1700Second Intermediate 1 1550 Period 1550New Kingdom XVIII 5 1292 1292IXX 1185 1186XX 1069 Table 4.3 Timeline of Egyptian kingdoms and dynasties (adapted from http://www.digitalegypt.ucl.ac.uk/chronology/index.html) and numbers of spirals sampled in each period The Middle Kingdom population was organised into urban settlements with temples and administrative buildings with facilities for defence and food storage (Grajetzki 2006, pp.81-83). This was supported by a large rural population underlying an economy based on agriculture and, in some parts of the country, cattle farming (Grajetzki 2006, p.78). Evidence exists for both free trade in private sales at markets and for centralised collection and distribution of resources for state employees (Grajetzki 2006, p.84). Contacts outside Egypt had been limited in the Old Kingdom although there were steady military assaults on Nubia (Redford 1997, p.40). During the Middle Kingdom, there is evidence of trade with Nubia and of the construction of fortresses at Egyptian land borders to the south and northeast. Continued military forays from the Eleventh Dynasty led to the conquest of Nubia in the Twelfth Dynasty and a resident Egyptian population there in the Thirteenth Dynasty (Grajetzki 2006, p.134; Murnane 1997, p.28). Nubia and sub-Saharan Africa, reached through Nubia, had important luxury

129 raw materials such as gold and ivory (Redford 1997, p.40). Archives at Uronarati on the southern Egyptian border record trade with Nubians and edicts controlling the movement of Nubian merchants into Egypt (Smith 1990). The Kerma Culture of southern Nubia became powerful during the Second Intermediate Period and conquered most of northern Nubia as well as venturing into Upper Egypt (Grajetzki 2006, p.135). Initial trade contact with Palestine was to obtain access to Syrian and Mesopotamian raw materials – turquoise and copper from the Sinai Peninsula and cedar wood from Lebanon through Byblos; excavations at both Byblos and Ugarit to the north have revealed large numbers of Egyptian finds (Grajetzki 2006, p.146; Redford 1997, p.43). A resident Palestinian population became established in the Egyptian Delta some of whom occupied high administrative posts (Grajetzki 2006, p.136). Minoan pottery first occurs in Egypt during the Middle Kingdom and it is suggested that the presence of this material is the source for the rise in popularity of the spiral in Egyptian decoration (Grajetzki 2006, p.136). There is, however, little evidence of direct trade with Crete. Minoan material may have been traded via Lebanon during the Middle Kingdom and direct contact established only during the New Kingdom (Grajetzki 2006, p.136; Redford 1997, p.44). The original use of scarabs reflects the administration of trade, communications and storage of supplies. They derived from stamp seals, used to seal documents, containers or doors, which developed from cylinder seals during the First Dynasty, initially as button seals mounted on rings (Reisner & Wheeler 1930). The oval seal form with an amulet carving on the reverse generally in the form of a scarab beetle developed into its final form between the Sixth and Ninth Dynasties (Reisner & Wheeler 1930). Different scarab types are identified with ‘royal name’, ‘personal name’ and ‘decorative’ scarabs being stylistically distinct (Reisner and Wheeler 1930). Of the analysed sample for this case study 44 of the scarabs had pattern inscriptions, seven had name inscriptions, four were uncertain and one of the spirals in the sample was from a carving on the reverse of the scarab. A study of Middle Kingdom and Hyksos scarabs and seal impressions from Semna and Uronarti from the first half of the Thirteenth Dynasty identified two main types of seal – the first to develop bore the name of the king, an official or an administrative department as the primary inscription (Reisner & Wheeler 1930; Smith 1990). Later,

130 private seals of individual officials with either a name or an all-over geometric pattern were used as a secondary seal alongside official administrative seals (Reisner & Wheeler 1930; Smith 1990). Administrative name seals have little decoration but often include spirals or spiraloids as border decoration (Reisner & Wheeler 1930). Pattern seals have spiral, concentric circle or knot decoration as the main imagery sometimes combined with auspicious symbols (Reisner & Wheeler 1930); Petrie (1920, p.25) suggests that these developed in simple form around the Eighth Dynasty. Smith (1990) suggests that the private seals with names belonged to more senior officials than those with only pattern or pattern and symbol decoration. Reisner and Wheeler (1930) distinguish between high quality carving which, they suggest, was carried out by carvers employed by the royal court and cruder carving techniques which they suggest were locally made; they link the quality and detail of the carving to the status of its owner. This study has focussed on scarabs rather than seal impressions. Alongside their use as administrative seals, scarabs appear to have served a decorative and auspicious function as amulets, particularly associated with regeneration as well as being a declaration of status possibly passed down several generations of a family before being deposited in a tomb (Grajetzki 2006, pp.103, 143; David 2002, p.175). In Middle Kingdom Egypt, a change in the focus of religion occurred, transferring away from the leader as an absolute divine authority to a belief in the control and authority of the gods over the whole population (David 2002, p.150). The practise of religion focussed on purification, recitation and offerings of food to icons of a deity (Teeter 1997, pp.148-151). Texts from Dynastic Egypt mention requests to ancestors to appear in the dreams of the supplicant “Become an akh for me before my eyes so that I can see you in a dream fighting on my behalf” and the ritual of the early New Kingdom cult of Amenhotep I includes drum beating and chanting (Teeter 1997, p.153) which has, in some ethnographic examples, been associated with trance states (Neher 1962). However, there is no strong evidence in either the written record or archaeological finds to suggest that psychotropic drugs or trance states which might have induced EVP were a part of Middle Kingdom religious practise. Petrie (1920) identified four categories of Egyptian decoration: geometric, representational, structural and symbolic. Geometric decoration included angular forms – arrangements of zig-zags, triangles, hexagons and squares, which Petrie

131 associates with origins in woven structures (Nelson 1936; Petrie 1920, p.14), particularly in earlier periods but by the Middle Kingdom appears to have been dominated by more curvilinear forms, predominantly arrangements of knots and spirals. Spirals in Middle Kingdom decorative culture Spirals on scarabs were uncommon before the Middle Kingdom, but were more frequently found on button seals (Figure 4.8) in a different character from that seen later on scarabs – the number of rotations tends to be greater and sharply angular forms are found which are not seen on later scarabs. The origins of spiral decoration on Old Kingdom and Middle Kingdom seals and scarabs is unresolved. Throughout the period spiral decoration was used in Crete, the Cyclades and the Near East and, by the Middle Kingdom also on Mycenaean pottery (Ward 1970). Reisner and Wheeler (1930) and Ben-Tor (2007, p.12) suggested a development of Middle Kingdom spirals from earlier Egyptian motifs – either from Naqada II pottery or Old Kingdom motifs or gold wire jewellery. Alternatively, both Ward (1964) and Reisner and Wheeler (1930) raise the possibility of multiple sites of independent evolution for relatively simple geometric motifs.

Figure 4.8 Ninth Dynasty button seal from Qaw, (©Petrie Museum of Egyptology, University College London acc. no.UC20752). The first occurrence found of spirals on scarabs dates to the Fifth Dynasty (Petrie 1920, p.19). These earliest designs tend more towards hooks than spirals and tend not to be linked in complex interrelationships (Petrie 1920, p.21). Nevertheless true spirals are found in small numbers on Old Kingdom and First Intermediate Period scarabs. Spirals only became a common motif on scarabs from the start of the Middle Kingdom, possibly through transference of spiral decoration from Minoan or Cycladic decoration to Eleventh Dynasty Egypt via Syria (Crowley 1989, pp.182-183, 194). By the Twelfth to

132 Thirteenth Dynasties, the use of linked spiral or concentric circle motifs became a characteristic feature of Egyptian scarabs (Ben-Tor 1997). Spiral decoration on scarabs declines in the late Middle Kingdom. In the Second Intermediate Period and New Kingdom scarabs with spirals are found but are less frequent and designs which prioritise spirals over inscriptions in terms of surface area also reduce in number. Whereas, on Middle Kingdom scarabs, the area of the scarab available for inscription often seemed compressed by the desire to render a smoothly curving border of spirals (Fig 4.9a), in the Second Intermediate Period, there seems to be a tendency, as in early examples, to relegate spiral decoration to a compressed hook forms (many of which were excluded from this research since they did not have sufficient rotation length) and are longitudinally compressed to allow a larger area for inscription (Fig 4.9b). The beginning of this trend to reduction appears to be visible in the late Middle Kingdom scarabs from Lahun although examples with fuller spirals are not precisely dated within the Middle Kingdom period so the point at which they begin to decline cannot be determined.

b. a. Figure 4.9a. Middle Kingdom spiral border, (©Petrie Museum of Egyptology, University College London, acc.no.UC29073). b. New Kingdom spiral border, (©Petrie Museum of Egyptology, University College London, acc.no.UC11897). Only one spiral decorative style on scarabs appears to have become more common in the New Kingdom – crossed spirals (Figure 4.10). They are found on Middle Kingdom scarabs but make up a much greater proportion of the spiral decoration found on New Kingdom examples. These may be two dimensionally symmetrical abstractions of plant representations.

133

Figure 4.10 Crossed spiral decoration on a New Kingom scarab, (©Petrie Museum of Egyptology, University College London, acc. no.UC60391). Spirals were also found as architectural decoration; examples of painted spiral tomb decoration were found in the Middle Kingdom, for example, in the tomb of Hepzefa at Assiut (completed between 1971 and 1928) (Shaw 1970) but reached a peak later than the spiral decoration on scarabs with most examples dating to the New Kingdom (Petrie 1920. p.31). Petrie (1920, p.28) differentiates the spiral decoration on scarabs from spirals on architecture by the nature of the space they were intended to fill with large wall or ceiling surfaces allowing regular designs expanding in all directions rather than spirals on scarabs which he suggests were initially distorted to fit around name inscriptions. In representational contexts, spirals on scarabs have mainly been associated with plant forms identified by a broadening of the outer end of the coil to represent a lotus seed pod or a reed head (Figure 4.11). On pre-Middle Kingdom scarabs, the large majority of spirals seem to have formed part of representational depictions of plants. This is a curious abstraction since the plants with which they are most commonly identified – papyrus (a symbol of the Lower Egyptian kingdom) and lotus (Ward 1902, p.102; Petrie 1920, p.42; Marinatos 1951) do not naturally assume any planar spiral growth forms of the types depicted; the use of spirals seems more likely to be an imposition of the spiral motif onto the traditional three stemmed depiction of water plants than a derivation of spiral decoration from a plant source.

134

Figure 4.11 Spiral representing a lotus or reed, (©Petrie Museum of Egyptology, University College London, acc.no.UC18125). In the majority of cases, spirals cannot be identified as representational although there is consistent representation of the horns of Hathor as spirals (Figure 4.12a) and a single instance on a scarab of an udjat-eye including a spiral (Figure 4.12b). Representations of an udjat-eye in other contexts not infrequently include spirals and this usage continued into the later Dynastic period (Figure 4.12c).

c. b. a. Figure 4.12a. Hathor horns, British Museum, acc.no.1930,0711.202. b. An udjat-eye on the top of a scarab, Oriental Museum, University of Durham (Wellcome Collection), acc.no.101585/17. c. An udjat-eye on a 26th Dynasty amulet, British Museum, acc. no. EA63381. On the scarabs used in the study, there is a single instance in which a snake's head has been added to spiral decoration (Figure 4.13a). Although this explicit representation as a snake appears to be unusual, there may be a case for arguing that a number of the spiral structures on Middle Kingdom scarabs relate to snakes. The basic structure of the running spiral border seen in Figure 4.13a is common on many scarabs of this period so it is possible that less readily identifiable snakes are represented in other examples. In other examples where spirals form the main decoration of the scarab, similarities of layout can be seen between linked spirals (Figure 4.10) and endless knot patterns (e.g. Figure 4.13b) which have also been used to represent snakes (Figure 4.13c).

135

a. b. c. Figure 4.13a. Snake as border decoration on a Middle Kingdom scarab, British Museum, acc. no. 1960,0514.53. b. Endless knot design on a Middle Kingdom scarab, (©Petrie Museum of Egyptology, University College London, acc. no. UC60359). c. Snake in an endless knot form on an animal mummy case, (©Petrie Museum of Egyptology, University College London, acc. no. UC8220). Scarabs have long been used to make categorical distinctions in typology of scarab decoration. The majority of the classification systems use the arrangement and interrelationship of spirals as the basis for type classes rather than the shape of the spirals themselves e.g. Petrie (1920, pp.20-21). Petrie (1920, p.23) also noted variation over time in the way these features occurred in the placement of spirals on the scarab – initially around the sides of a name inscription but later sometimes fully encircling it and also in the nature of linkages with long links being more typical of pattern scarabs than name scarabs. Both Shaw (1970) and Grenfell (1915) segregate the different interrelationships of spirals as being important in identifying different origins and meanings for motifs which include spirals as a main component with Shaw, on the basis of brush strokes in tomb paintings, separating pairs of spirals joined to form a heart-shape as leaf forms from S spiral links although she suggests that the two forms may have merged in copying. Although the focus of this study was on Middle Kingdom scarabs when spiral decoration was at its height, a small sample of earlier and later scarab spiral decoration was included since, as discussed above, some types of spiral decoration appeared to show change over time. The Old Kingdom and First Intermediate Period sample was too small to draw any conclusions but more examples were available from the Second Intermediate Period and New Kingdom.

136 The majority of scarabs sampled in this study (Appendix 2, Table A2.4) are made of faience or steatite. Only two examples were carved in a hard stone (obsidian and jasper) and one incised into gold sheet. A number of the scarabs had spirals which were too damaged to analyse. Some which could be analysed had minor damage and chipping. 15 of the spirals sampled had interpolated sections to a maximum of 30% of the length. Provenance of the scarabs was poor, only 36 of the 53 scarabs in the study were identified to an excavation site. The scarabs and scaraboids in this study of known provenance were found at nine sites from both Upper and Lower Egypt. The earliest scarabs in the sample from the 9th Dynasty came from tombs at Qaw, (Brunton 1927, p.40). The main evidence for the 9th Dynasty comes from chamber tombs representing a moderately well off population (Brunton 1927, p.74). At the time of excavation, scarabs from Qaw represented the earliest examples of scarabs used as amulets rather than seals (Brunton 1927, p.74). The majority of scarabs in this sample were from Lahun which began as a workers' village associated with the construction of the pyramid of Senusret II but became an important, planned, settlement in the late Twelfth Dynasty with a population of around 3000 (Grajetzki 2006, pp. 57,92,116-8,141; Shaw 1997, p.71). There is evidence of tombs and houses of officials as well as a large number of smaller houses (Grajetzki 2006, pp.57, 92, 116-8, 141). The Upper Egyptian sites from which scarabs have been sampled were placed on important trade or communication routes. Tell el Yehudijeh a settlement founded in the Middle Kingdom or Second Intermediate Period (David 1998, p.80) operated as a control point for the land route from the Sinai Peninsula to Egypt; there is evidence of a Hyksos fortress at this site (Bietak 1999, pp.791-2). Nubt was an major settlement centre, a religious site for the worship of Seth (Grajetzki 2006 p.93) and an important communication point with access to navigable stretches of the river from which gold from nearby mines could be transported (Grajetzki 2006 pp.92-93; David 2002, p.40). Koptos, likewise, was associated with the start of the shortest access route to the Red Sea and an access route to silt stone quarries and gold mines (Grajetzki 2006 pp.92-93). A greater range of sites from Lower and Middle Egypt was sampled, although sample size from each site was small, in several cases only a single scarab. Meydum was an important royal burial ground in the Old Kingdom with a pyramid built by Snefru and

137 substantial surrounding tombs.(David 2002, pp.79-80, 100-1). The scarab analysed in this case study was from a deposit dating from the New Kingdom reuse of some of the tombs as Meydum although the scarab itself dated from the Hyksos period (Petrie, Mackay & Wainwright 1910, pp.36-7). A single scarab was analysed from Hawara, a long-standing cemetery site, established in the 12th Dynasty around the pyramid of the deified Amenhat III and continuing into the Roman period (Uytterhoeven & Blom-Böer 2002). One scarab from Sedment, a cemetery site near the town of Herakleopolis, and an important First Intermediate Period settlement was included in the sample. Tombs in the Sedment cemeteries do not reflect a very wealthy population and may have been those of local village populations (UCL 2001). The Mostagedda cemetery was, similarly, probably used by smaller settlements in the area and does not include burials of members of the ruling group or of officials (Grajetzki 2006, pp.103, 144)). 4.3.2.2 Cyprus The period of the Eastern Mediterranean case study covers prehistoric Chalcolithic society to the end of the, protohistoric, Bronze Age on Cyprus (Steel 2004, p.1). However spirals only appear as a common motif in the Late Cypriot period (Table 4.4). Period

Late Cypriot

CyproGeometric

CyproArchaic

Date

IA

1650-1550

IB IIA IIB IIC IIIA IIIB IA

1550-1450 1450-1375 1375-1300 1340/1315-1200 1200-1100 1100-1050 1050-1000

IB II III I

1050-950 950-900 900-750 750-600

No. sampled spirals (gold-work spirals in brackets)

(26) (2)

5, (2)

(1)

3

1

2

1 10

2 (1)

1 1

II 600-475 Table 4.4 Timeline of Cypriot periods (adapted from Kiely n.d.) and the spirals sampled from each.

138 Two major population movements to Cyprus have been posited within the time period of this study, from Anatolia in the third millennium and from the Aegean in the second millennium, both associated with times of significant cultural change on Cyprus (Steel 2004, p.20). The Bronze Age on Cyprus represented a period of major population increase, a settled economy and increased arable production with wheat and barley forming the staple crops (Steel 2004, pp.128, 131). By the Late Cypriot period the three cornerstones of the Cypriot economy were agriculture (sheep, goats and cattle predominating), copper production and international trade (Steel 2004, p.158). By the thirteenth century, a hierarchical social structure was well established with the development from entirely village settlements to larger fort settlements (as possible centres for internal agricultural distribution (Steel 2004, p.152)) and satellite villages and settlement structure; by the Late Cypriot period there was a tendency for settlement sites to move towards the coast and major urban settlements had developed (Steel 2004, pp.129, 149-50, 152, 156). Three lower tiers of settlement developed inland - secondary administrative settlements tertiary settlements and sites showing evidence of craft specialisation (Steel 2004, p.156). The provenanced material used in this case study came from twelve sites summarised in Table 4.5.

139

Table 4.5 Places of origin of the Cypriot vessels and gold-work from which spirals were sampled. Site Type Period2 Economy External influence Enkomi

Kouris Valley

Ayios Iakovos

2

settled MiddleLate Cypriot transition

settled MiddleLate Cypriot transition (LCIIIB)

Settled beginning C16th occupied through to the end of the case study period

religious sanctuaries BA-IA, LC (BA-IA) and cemeteries (MBALBA)

Commercial centre and trading port; secondary copper processing – dominated the copper industry before 1300 Commercial centres and trading ports, secondary copper processing copper and later ironworking

?copper extraction

mainly Mycenaean imports and local copies of high status pottery often with representational scenes

Direct access to imported goods; settled Mycenaean population producing pottery in local styles but with Mycenaean decoration indirect probable contact with Enkomi though Hadjicosti suggests possibly a Mycenaean settlement following 13th-12th century migrations indirect probable contact with Enkomi

Abbreviations used in this table: LC – Late Cypriot, BA – (Early/Middle/Late)Bronze Age, IA – Iron Age

References

1 pottery vessel

Steel (2004, p.156); Karageorgis (1969); Reyes (1994); Muhly (1989); Iacovou (1989) SCE website (n.d.); Hadjicosti (1997)

3 pottery vessels

1 pottery vessel

Steel (2004, pp.152,156); Collard (2011 p.289); passim; SCE website (n.d.); Muhly (1989)

Steel 2004, p.157); Collard (2011, p.217); SCE website (n.d.)

139

Idalion

urban walled settlement with wealthy cemetery and religious sanctuaries (one of the earliest urban sites on Cyprus) encompasses major coastal trading ports including Kourion and EpiskopiBamboula fortified settlement, religious sanctuaries, cemeteries

Objects analysed 7 pottery vessels, 29 gold diadems or mouthpieces

140

Type

Period

Economy

External influence

Palaepaphos and PalaepaphosSkales

Palaepahos urban settlement by 14th13th century, Palaepaphos-Skales wealthy cemetery

trading port, most important Cypriot trading port by the Cypro-Geometric period

Near Eastern imported pottery, notably little Aegean imported pottery

Maroni

Wealthy urban settlement and cemetery by 14th13th century. Influence diminished due to centralisation of power to coastal cities in LCIII Tomb near Klavdia village

LC/LBA; occupied in C11th; extant by LCIIA continued in occupation throughout the study period Occupied by C14th; LCII

Regional centre controlling food production Secondary copper processing. Control of Kalavassos copper mine from LCIICLCIII Tomb CAI, Klavdia Klavdia trading settlement Late port and access Cypriot, route from coast to continued in centre of island occupation throughout study period

Aradhippou

Objects analysed 2 pottery vessels

References

Staging post for traders from Euboia travelling to the Near East; imports of Aegean and Near Eastern pottery

2 gold mouthpieces

Steel (2004, p.157); Muhly (1989); Reyes (1994); Cadogan (1989)

Egyptian and Aegean imports found from MBA

1 pottery vessel

Kiely (n.d.); Bailey (1969)

Steel (2004, p.157); Karageorghis (1983 cited in Popham 1985); Coldstream (1989); Iacovou (1989)

140

Site

141

Site

Type

Economy

EpiskopiSettlement site Bamboula (within Kouris valley)

Late Cypriot (occupation ended in LCIIIA)

Amathus

necropolis

C13th. IA, Cypro Archaic

Possibly on the transport route for copper from inland mines to the coastal cities Trading port

Marion

settlement and cemetery, some wealthy tombs

Stylli

cemetery

Lapithos

wealthy cemetery

established by Chalcolithic, most archaeological remains date to 7th-4th centuries Geometric and Archaic periods cemetery EBA-IA; also Neolithic remains

Trading port

n/a Copper extraction and trade in region

External influence

imports of Proto-Geometric and Geometric pottery; Staging post for traders from Euboia travelling to the Near East; possible Assyrian imports in C13th; later Egyptian and Phoenician trade contact minimal within period of study but Attic and Cycladic imports in significant quantities in late IA; Mycenaean type tombs Phoenician influence Mycenaean type tombs; possible Phoenician influence at end of case study period

Objects analysed 1 pottery vessel

References

4 pottery vessels

Childs (1997); SCE website (n.d.); Reyes (1994)

2 pottery vessels

Childs (1997); Coldstream (1989)

1 pottery vessel 2 pottery vessels

SCE website (n.d.); Winbladh (2007) SCE website (n.d.); Coldstream (1989); Karageorgis (1969); Reyes (1994); Kiely (n.d.)

Hadjisavvas (1989); Iacovou (1989); Collard (2011, p.289 passim)

141

Period

142 Although at a key interface between the Near East and Mediterranean Europe, Cyprus remained considerably isolated until the Bronze Age when contacts were initially made with Crete, then Syria, Palestine and southern Anatolia during the Early Cypriot period (Tatton-Brown 1987, p.26). It was only in the early second millennium BCE that Cyprus came to be significantly influenced by contacts with the Near East and Egypt with a variety of prestige goods including pottery imported and exported but the main trade interest is believed to have been in the copper sources on Cyprus which were traded for silver and finished luxury goods (Steel 2004, pp.2, 143-144; Tatton-Brown 1987, p.19). At the end of the Middle Cypriot period, trade in pottery indicates that important contacts with Lower Egypt and the Levant had been established (Steel 2004, pp.17, 154). Trade continued with Egypt and the Levant into the Late Cypriot period and, by the fourteenth to thirteenth centuries Cyprus had established a significant role in Aegean trade networks and was importing and imitating Mycenaean pottery styles (Steel 2004, p.170). Collard (2011) suggests use of psychoactive drugs in Late Cypriot ritual in the form of opiates also the consumption of alcohol with the intention of inducing hallucinatory experiences for ritual or social purposes. On the basis of vessel shape representing its contents, Collard (p.156) also speculates on the use of mandragora root although this cannot yet be supported by evidence of organic residues. Opium residues have been found in Base Ring ware pottery (Collard 2011, p.137) and the shape of some vessels has been conjectured to resemble poppy heads (Collard 2011, p.138). The typical decoration on Base Ring ware of spirals, wavy lines and concentric arcs (Collard 2011, p.133) could accord with typical EVP forms. Although both alcohol and opiates, consumed in large enough quantities, are hallucinogens, it is not clear, from descriptions by those experiencing the hallucinations, that they induce the type of EVP which are the focus of this research (Collard 2011, p.161; Manford & Andermann 1998). As well as possible hallucinatory effects, as a religious stimulus for the representation of spirals, there is evidence of the worship of horned gods and sacrifice of horned animals. However representation of horns on figurines of gods are not presented as spirals (Steel 2004, p. 175). Pottery production was introduced in Cyprus in the late Neolithic. Early decorative styles on vessels consisted of incised or painted geometric linear patterning (Tatton-

143 Brown 1987, p.24). Although floral and representational decoration of vessels was briefly introduced around 1600 in imitation of Near Eastern styles, geometric decoration created using combs or multi-headed brushes remained the most common decorative style on Cypriot pottery until the reintroduction of representational decoration on pottery vessels in the Iron Age (c.1050) again under Near Eastern influence (Tatton-Brown 1987, pp.26, 29). Production of pottery on Cyprus changed over the period of study from a domestic base in the Middle Cypriot period with regional variation identifiable across the island to specialist centres in the Late Cypriot period which allowed the development of more technically skilled forms of hand-made pottery such as thin-walled Base Ring ware in the Late Cypriot period (Steel 2004, p.163; Tatton-Brown 1987, p.26). Wheel-made pottery was introduced by 1600 and until about 1250 was used to produce imitations of Near Eastern styles which co-existed with local hand-made pottery styles (TattonBrown 1987, p.26). In the late thirteenth and early twelfth centuries, refugees or attacking parties from mainland Greece settled in Cyprus following the collapse of the Mycenaean civilization and caused substantial disruption to the previous way of life on Cyprus (Collard 2011, p.32). Following the collapse of Mycenaean production centres, regional demand for Mycenaean pottery styles persisted (Collard 2011, p.33). In response to this, a change to greater mass production occurred around the Late Cypriot IIC period along with the disappearance of traditional styles such as Base Ring ware and of elaborate vessel shapes; standardisation of sizes; poorer quality surface decoration and the gradual dominance of fast-wheel made pottery, particularly White Painted Wheelmade III (Steel 2004, pp.163-165; Collard 2011, p.33; Tatton-Brown 1987, p.27). Spirals in Ancient Cypriot decorative culture Imported goods were used as a means of asserting status and foreign motifs copied in locally produced goods (Steel 2004, p.169). The earliest examples of Cypriot spiral decoration found in this study are, however, on Base Ring pottery (Figure 4.14), a style developed locally at the start of the Late Cypriot period which continued into the Late Cypriot IIIA (Steel 2004, p.150, Kling 1991, p.182).

144

Figure 4.14 Base Ring ware juglet, British Museum, acc. no. 1897,0401.1282. Although initially relatively uncommon, spiral motifs rapidly became more frequent reflecting the introduction or development of different pottery styles. The appearance of spirals as a common motif may relate to the increasing import of Mycenaean pottery and the status value placed on Aegean iconography around this time (Steel 2004, p.165) although the highest status and most controlled imported pottery types were those featuring pictorial rather than geometric decoration (Steel 2004, p.170-1). A particularly common Mycenaean import was the stirrup jar (Steel 2004, p.170). Locally produced examples of these frequently included spiral motifs. Aegean influence cannot be said to explain fully the use of spiral decoration. Between approximately the fifteenth and ninth centuries most spiral decoration is found on White Painted Wheelmade III pottery (Figure 4.15). This style group of locally-made pottery is based on the technical structure of the vessels rather that the cultural influences on them. Its creation acknowledged the difficulties in Cyprus of creating type groups in the face of merging local, Aegean and Near Eastern influences in both vessel shape and decoration (Kling 1991, p.182).

Figure 4.15 White Painted Wheelmade III bowl, British Museum acc. no. 1896,0201.3.

145 The majority of analysed examples of spiral decoration on pottery were from the Cypro-Geometric and early Cypro-Archaic period appeared on Black-on-red ware (Figure 4.16a) with a few examples also found on Bichrome ware (Figure 4.16b), which had a Syro-Palestinian origin (Tatton-Brown 1987, p.29) (Appendix 2, Table A2.5). Black-on-red ware was constructed using a technique of Phoenician origin although Phoenician vessels tended to use only concentric circle decoration (Coldstream 2003, p.44).

a. b. Figure 4.16a. Black-on-red ware jug, Medelhavsmuseet, acc.no. MM1997:003. b. Bichrome ware amphora, Medelhavsmuseet, acc.no. MM acc 0699. The Cypro-Geometric decorative style provides prime examples of the ambiguous concentric circle/spiral motifs discussed in Chapter 2 and the occurrence of spirals on pottery decoration from this period may reflect a mechanically convenient means of creating decoration which was meant to be interpreted as the, more common, concentric circle motif. It is possible that the multi-headed compass which is believed to have been used in the creation of smaller concentric circle decoration (Eiteljorg 1980) slipped during the drawing of lines and that their spiral nature is coincidental. Without surviving compasses having been discovered it is not possible to tell whether slipping in this manner would have been possible or likely. None of the instances of Cypro-Geometric spiral decoration seen in this research show signs of a break in curve or sudden change in curvature which might indicate a slip or a painter making the best of a mistake. From this period, spirals are most common as large motifs on opposite sides of stirrup jugs or around the centre of dishes and bowls. Multi-headed compasses would be difficult to operate on the highly curved sides of some jugs and

146 flasks (Figure 4.17). If spirals were created whilst the vessel was turned on a wheel, as concentric circle motifs were on Greek Geometric vessels (Murray 1978, p.83), a spiral form would be easier to create using the natural tendency for the brush to drift down the curved side. Similarly, pottery decoration in the centre of dishes and bowls might have been created more easily on a wheel by moving the brush radially outwards as the wheel turned than by holding steady a multi-headed compass which would have required a high number of brush heads and a relatively large diameter.

Figure 4.17 End and profile view of a base ring ware jug with spiral decoration over a highly curved surface, British Museum acc.no.1876,0909.72. A second set of data used in this case study was drawn from gold diadems and mouthpieces from tombs at Enkomi and Maroni. In the Late Bronze Age, a reliable source of gold was established through trade with Syria and Egypt (Tatton-Brown 1987, p.42). Gold was worked into rolled strip wire or with repoussé or stamped decoration on gold sheet (Tatton-Brown 1987, p.43). The spirals on Cypriot goldwork show great uniformity. The large majority are dense coils linked in pairs most of which were created as stamped motifs. The similarity between motifs suggests that each single punch head depicted a pair of spirals (Figure 4.18a). Only two examples show sufficient variability in motifs to suggest that they may have been created freehand (Figure 4.18b and c).

147

a.

b.

c. Figure 4.18a. Stamped motifs on a diadem 1897,0401.272. b and c. Possible freehand motifs on a mouthpiece and a diadem 1897,0401.517 and 1897,0401.269. All from Enkomi. British Museum. The majority of spiral decoration on the Cypriot pottery vessels has no identified representational or symbolic significance. The few exceptions of known representational function depicted a lotus flower (Figure 4.19b), and an unidentified, presumed land plant (Figure 4.19a). Only one item of goldwork had an apparently representational meaning (Figure 4.18b) showing a probable land plant.

b. a. Figure 4.19 Detail from vessels showing spiral plant representations, Medelhavsmuseet, acc. nos. E.003:261 and NM Ant 1719.

148 An upper date limit of 600BCE was imposed on the collection of data because the inclusion of a larger geographical range of case studies had initially been envisaged and, in most cases this date made a convenient cultural break. Within the Cypriot case study, this choice of date was less appropriate as it cut across the Cypro-Archaic period and the style of regular spiral decoration on pottery continued in use to a significant extent after this point. The spiral motifs on pottery sampled in this case study were drawn from 12 types of vessels with distributions given below. krater 4 plate/dish 3 bowl 11 kylix 1 barrel jug 2 amphora 2 jug 3 flask 1 cup 1 carinated cup 1 bottle 1 pyxis 1 Table 4.6 Cypriot vessel types from which spirals were sampled. The spirals sampled from pottery were generally in a good state of preservation although the earlier material dating to Late Cypriot I and II tended to be in poorer condition; 13 spirals included sections of interpolation with a maximum interpolation length for a single spiral being 19%. Of the sampled pottery vessels, six lacked a precise place of origin and two could only be broadly dated to within the study period. The distortion caused by curvature of vessel walls prevented the analysis of spirals on 21 pottery vessels and particularly affected the earlier examples of spiral decoration from the Late Cypriot II period. The diadems and mouthpieces showed generally poorer preservation, with 37 of the sampled spirals having some degree of interpolation to a maximum of 31%. The mouthpieces and diadems sampled in this case study show a lack of diversity in sampling, with all but one example having been excavated at Enkomi. This lack of diversity reflects the data available from the museum collections surveyed. Of the sampled mouthpieces and diadems, one had no precise provenance and a second example lacked a defined place of origin.

149

4.4 Summary The case studies chosen for this research were selected as examples of regions which showed a sequence of cultures with high prevalence of spirals running from the Neolithic into the Bronze Age in the eastern Mediterranean and East Asia. This allowed for some assessment of internal continuity and potential cultural transmission within each of the case study clusters. The eastern Mediterranean and East Asian study areas, appear to have had, at most, minimal and indirect contact. Comparisons were made between the two groups to assess whether cultural distinctions can be drawn and whether similarities can be found supporting the idea of a common response to a common natural stimulus when the cultural input to the development of decorative styles might be expected to be quite different.

150

5 Testing and evaluation of the classification method The first stage in considering the results of this research compared the proposed method with an existing classification system based on mathematical spiral forms, assessing the principles which underlie the creation of the class groups and the visual coherence of the class groups created by correlation to mathematical spirals and cluster analysis of the form and style variables used in the newly proposed method. The effectiveness of the method in application is considered in addressing the final three research questions. Application to case study data tested whether it is possible to use this method to distinguish between the spiral decoration of different cultures. The case study data were then turned to address the opposite questions of whether commonalities could be identified which crossed cultural boundaries and which might be attributed to a known natural origin in either substantive forms or disturbances of the visual system. As a final stage, in assessing the results of this research, an evaluation of the proposed analysis method is required; whether it is effective in discriminating between spiral motifs; whether there are systematic problems in the current implementation of the method which may have affected its application to the test case studies; and how precision and accuracy might be improved. Beyond the basic working of the method, it needs to be shown that such a method is of practical use and is reliable and could be used to classify the full breadth of variation seen in decorative spiral motifs.

5.1 Comparison of method with fitting to mathematical spirals as a classification technique As an initial test of the viability of the established approach of making reference to mathematical spiral types in classifying decorative spirals, the spectrum density plots of the five template spirals (Figure 5.1) were correlated with each other, using twotailed Pearson correlation, to test whether there was sufficient difference between the types to make the distinction of separate decorative classes viable given that the majority of decorative spirals do not match perfectly any mathematical spiral curve. Two pairs of template spirals showed significant positive cross correlations – parabolic and Archimedes' spirals had a correlation of 0.912 and logarithmic and hyperbolic spirals had a correlation of 0.997. The high values of both these correlations suggest

151 that in the case studies used in this research, and probably in the majority of historic or prehistoric cases, many decorative spirals will share a high correlation with both spirals in these pairs. Division of parabolic from Archimedes' spiral and, particularly of logarithmic and hyperbolic spirals may not be appropriate. For comparison tests, the five spiral types were retained as separate divisions since that has been the norm in previous studies using mathematical spirals as classification groups. Figure 5.1 Mathematical template spirals and their spectral density plots: a. parabolic b. Archimedes'. c. Euler. d. logarithmic. e. hyperbolic. Orange circles indicate the characteristic secondary signal peaks characteristic of the mathematical spiral forms (see section 5.5.6). (Figure runs over two pages.)

2

4

4

0 0

1

2

3

4

a.

4

2

2

4

4

0 0

4

b.

2

4

6

8 10 12

4

2

152

0

c. 2

4

4

0 0

10

4

20

30

4

d.

2

2

4

4

0 0

4

e.

4

2

All decorative spirals had a significant correlation of at least 0.5 with at least one of the mathematical template spirals. Spirals which showed the lowest correlations tended to register similar correlations with two template spirals suggesting that they include characteristics of more than one mathematical form. Spirals were placed into mathematical classes according to the template type with which they showed the highest correlation (Appendix 2, Tables A2.7-10).

153 Although the logarithmic and hyperbolic type groups could not be reliably separated, it is interesting that a significant majority (p<0.00 for each case study) of the spirals in this group tended to resemble most strongly the hyperbolic examples. This may reflect a genuine tendency to intentionally reproduce hyperbolic forms but it may, alternatively, reflect a difficulty in drawing a regular logarithmic spiral. The idea that the spirals in the hyperbolic/logarithmic group may have been intended to replicate logarithmic rather than a hyperbolic form is supported by the presence of all the known representational plant, horn and reptile tail depictions which, in real examples, all approximate logarithmic forms. In attempting to draw a logarithmic spiral, there is a tendency to use the previous rotation as a guide to drawing the next. This creates a spiral with a relatively low expansion rate in the centre, then a relatively sudden increase to resemble a logarithmic form, but in a mathematical sense, more closely fitting a hyperbolic form. All of the decorative spirals known to depict logarithmic forms fell into the logarithmic/hyperbolic group, so this classification can be said to be valid. However, the size of the hyperbolic/logarithmic group raises difficulties with using this type as a decorative class. Among the set classified as hyperbolic, those with a lower correlation end would not be assessed visually as hyperbolic, so assigning decorative groups on this basis would only be effective if a minimum correlation value were applied. This approach would leave some spirals unclassified. Decorative spirals which visually resemble Archimedes' spirals mostly fell into the mathematical spiral class of Euler spirals. None of the Shang/Western Zhou Dynasty spirals which appear to resemble an Archimedes' form were classified as such. The only examples which were classified as Archimedes' types were those from the Cypriot case study with greatest rotational length. The gradual decline in spectrum density frequency peaks of some of the angular Shang/Western Zhou spirals more resembled Archimedes' than Euler forms. However, the failure to match the position of the maximum peak, which is raised with an increasing number of rotations, exerted a strong countering influence and caused the spirals to be classified as Euler. The Cypriot Archimedes' spiral-resembling motifs assigned to both the Euler and Archimedes' type groups were similarly defined by the frequency of the maximum peak rather than the shape of the subsequent signal. Spirals in the Cypriot study which

154 were linked to other spirals in S or C-coils also fell into the Euler group. This does match the visual impression of these spirals relatively accurately. A problem in the use of mathematical definitions is highlighted in the correlation of the triangular spiral CSBM15.2 (Figure 5.2) to mathematical spiral forms in that the aspects of spiral form which are mathematically important do not necessarily accord with those that are visually important. The property emphasised in the classification of a regular expansion rate caused the spiral to have a correlation to an Euler spiral of 0.645 setting aside the more perceptually important property of its triangular shape which causes it to bear little visual resemblance to any mathematical spiral.

Figure 5.2 Concave triangular spiral CSBM15.2. No spirals from any of the case studies correlated most strongly with parabolic spirals. This agrees with the visual impression of the spirals, none of which showed a consistently decreasing trend in expansion rate. Within each case study culture, both grouping based on mathematical spirals and by cluster analysis were effective at bringing together the spirals known to be of representational origin. The cluster based method missed more representational spirals from the classes in which representational spirals dominated but the mathematical grouping drew more non-representational spirals into the same group with the representational spirals. The same tendency of over-separating or overgrouping is found with visual distinctions of spirals. For example, the separation in clustering of the, visually similar, classes 4 and 5 of the Shang/Western Zhou Dynasty spirals (Appendix 2 Table A2.12). The grouping based on mathematical spirals links these but places the majority of angular spirals used as background decoration on vessels in the same class as the representational motifs whereas clustering separates most of these into class 1. For most case studies there is a high degree of

155 homogeneity and variation is continuous so it is difficult to make class divisions. This raises the question of whether is appropriate to attempt to assign groups at all. Zhushchikhovskaya and Danilova (2008) who use mathematical spirals as a means of grouping note that some spirals fall between groups and it may be more appropriate to regard many of the motifs as being on a continuum of change. Although clustering and discriminant analysis rely on the creation of separated classes, they are better suited to the assessment of continuous variation than the use of mathematical spiral classes since they assign a rank order of similarity which runs across the class divisions but still also allow spiral shape classes which are truly separated to be identified. Across cultures where no clear link or contact can be established, high numbers of spirals correlate with a hyperbolic/logarithmic form. The common natural occurrence of logarithmic forms (Kappraff 1992) suggests that reference to spirals found in the natural environment may have guided these cultures and that abstraction of these falls within limited form boundaries. To establish clearly the consistent link between known representational spirals and a form resembling a logarithmic or hyperbolic spiral, measurement of resemblance to the mathematical spiral type by correlation of spectral densities or by another method which provides a quantified indicator is preferable to assigning spirals to mathematical classes based on visual impression. A quantified similarity strengthens the argument for an objective link between representational forms and logarithmic spirals and makes identifying a common underlying principle between cultures easier. The objection in this thesis to the use of mathematical spirals as classification groups is largely methodological. Such classifications assume absolute boundaries rather than a continuum of variation and, as such, will be compelled to include forms that bear little or no visual resemblance to the mathematical spiral form or an arbitrary cut off of correlation must be imposed so that these mathematical classes do not encompass all the spirals present in the culture and would have to be combined with nonmathematically based class groups. Also mathematical spiral definitions do not necessarily reflect visually important properties of spirals. When visual assessment of the mathematical type of spirals is inaccurate, as it is bound to be, given the close correlation of some spiral forms, the use of mathematical terminology misrepresents the form of the spiral. In contemporary culture where digital arts allow much more

156 precise construction of decorative spirals or where a mechanical, or a ruler and compass based means of laying out a decorative spiral were frequently used in the decorative scheme of a culture a closer accord might be found to a mathematical curve and the use of mathematical terminology might provide a relevant description of spiral form. Without clear evidence of a mathematical basis having been employed in the construction of decorative spirals it is argued that mathematical classes are inappropriate.

5.2 Results of the application of the method to test case studies For each case study, the spirals were clustered based on their visual properties to identify whether significantly differentiated type classes of spiral motif could be identified within each culture. Discriminant analysis was carried out on the clusters to identify the variables which were important in dividing them, taking a minimum factor loading of magnitude 0.3 to identify significant variables. Finally, the data from all four case studies were pooled and a differentiation analysis carried out to see if the individual cultures could be distinguished from each other. The relevance to grouping of the contextual attributes across the entire sample was also tested. 5.2.1 Jōmon Period case study The Jōmon spirals were clustered into five classes (Appendix 2, Table A2.11). The fifth class only contained a single spiral which was clustered with the remaining spirals at the highest level of differentiation and was shown to be highly separated from the other groups in discrimination analysis. This spiral had the highest rotational length of any in the sample and it was probably this that caused its strong separation because of the effect of rotational length on the form variables. For this reason, the spiral was treated as an outlier and excluded from discrimination analysis although it was included in the assessment of the distribution of individual style variables. Discriminant analysis identified the form correlation with four spirals, JKAN13.11 and JKAN2.1 (Figure 5.3) as required variables for separating the classes although the eigenvalues of both factors were relatively low (9.485 on function 1 and 3.485 on function 2) so these correlations did not fully explain variation. 1

Spirals from the four case studies were assigned codes with the prefix J for the Jōmon study, CS for the Shang/Western Zhou study, EG for the Egyptian study and CY for the Cypriot study.

157

a.

b. Figure 5.3 Form correlation spirals and spectra. a. JKAN13.1. b. JKAN2.1.

JKAN13.1 JKAN2.1

Function 1 0.445 0.74

Function 2 0.896 0.997

Figure 5.4 Distribution of clustering classes and variable loadings on the first two discriminant functions for the Jōmon case study.

158 Discrimination showed poor separation of the classes identified (Figure 5.4). Function one is most strongly identified by positive correlation with JKAN2.1 with JKAN13.1 operating more weakly. The two spiral form spectra are similar, JKAN13.1 has a more pronounced secondary peak associated with an increasing expansion rate at the outermost end whereas JKAN2.1 shows a gradual decline in signal peaks more associated with a constant expansion rate. Visually, the strongest division is between the pair of classes 3 with 4, and 1 with 2 (see Table A2.5). Classes 3 and 4 are visually dissimilar from the spirals used in discrimination and show a negative value on function one. These classes consisted of spirals of very low rotation or those with a very high increase in expansion rate. All spirals of known representational form from vessels are included in these two classes. These were all representations of snakes. However, the headdress spirals on dogū, which have also been suggested to represent snake forms, more frequently fell into the first two classes. On the strength of this, discriminant analysis was used to test whether spiral motifs on vessels differed as a group from those on dogū. This did not yield a significantly different mean on any variable. Most of the spirals in the first two classes were of relatively high rotational length with a low spacing and constant path width and expansion rate. This group tends to encompass small decorative motifs on the vessels and the majority of the dogū motifs. Function two (Figure 5.4) separates spirals in classes 1 and 4 from classes 2 and 3. The spirals with a high positive value on the second function in classes 1 and 4 showed the sudden drop in density signal to almost zero, which is characteristic of rotational lengths below two full rotations. Given the lack of separation amongst the spiral groups it seems more appropriate to view these spirals as being on a continuum of variation across the four classes. The order of the classes in the discriminant analysis was changed from that in clustering with the low rotation spirals in classes 1 and 4 drawn together at the centre of the continuum; short smooth spirals of low expansion rate at one end of the range; and the shortest spirals with a very high expansion rate, at the other. On the basis of this continuum, the single class 5 spiral was placed as an outlier beyond the class 1 spirals. Discriminant analysis based on the separation of the motifs into Early, Middle and Late Jōmon periods operated on the form correlations of spiral JKAE19.1 (Figure 5.5) and

159 the first coefficient of spacing trend. With only three objects, the Early Jōmon sample was too small to draw any effective conclusions about grouping. Eigenvalues of both discriminant functions were very low 0.564 for function one and 0.040 for function two. On neither function were any of the periods separated effectively (Figure 5.6).

Figure 5.5 Spiral JKAE19.1 and its density spectrum.

JKAE19.1 Spacing trend 1

Function 1 0.474 0.643

Function 2 0.881 -.766

Figure 5.6 Jōmon chronological period discrimination and loading on the discriminant functions.

160 Discriminant analysis applied to spirals to distinguish between modelled and incised decoration did not yield a significantly different mean value on any variable. Both Middle and Late Jōmon clusters showed similar variance on both functions and Early Jōmon spirals fell within the range covered by both groups. Both Middle and Late Jōmon groups included outliers on function two. Differences are marked between the Jōmon chronological periods in pottery decorative styles and techniques (Kaner 2009, p.18; Kidder & Esaka 1968) but there is no indication that it is possible to separate the spirals according to cultural period. Zhushchikhovskaya and Danilova (2008) analysed Jōmon pottery. Using visual assessment they identified both logarithmic and Archimedes' spirals amongst the Middle Jōmon examples the opposing loadings on the spacing variable would not contradict this indicating the presence of spirals with high and low expansion rates. Zhushchikhovskaya and Danilova found organised orientation in Middle Jōmon vessels with horizontal banded decoration. This characteristic could not be identified in the Middle Jōmon sample used in this case study either taking the sample as a whole or assessing only those vessels with horizontal banded decoration. Sample size was very small, only eight spirals from vessels with horizontal banded decoration were included in this case study. The Late Jōmon sample were not compared with the findings of Zhushchikhovskaya and Danilova because no vessels were sampled from that period in this study so like with like comparison could not be made. 5.2.2 Shang Dynasty and Western Zhou Period case study Of the four case studies, the Shang/Western Zhou study had the greatest number of spirals with known representational meaning. This was the only study in which the classes generated in clustering could be seen to differentiate between representational forms. The spirals were clustered into five classes (Appendix 2, Table A2.12). Classes 2 and 3 contained 22 of the 30 representations associated with mythical creatures in the form of tails, horns, whiskers, ears, nostrils and decoration on the face of taotie; the remainder fell into class 1. Classes 4 and 5 separated out all but one of the snake representations; the exception, CSBM15.1 (Figure 5.7) fell into class 3. Curiously, this exception resembled the Jōmon snake depictions in having a divided tail, a feature not present in any of the other Shang/Western Zhou snake depictions. Clustering of this spiral with the other snake representations put it with the Jōmon

161 and Egyptian snakes separately from the two Western Zhou Period snakes but, as is discussed below, these two snakes are outliers from the whole case study sample.

Figure 5.7 Spiral CSBM15.1 and its spectrum density plot. Discriminant analysis identified the form correlations of three spirals as being suitable to separate the clusters – CSVAM8.1, CSBM12.84 and CSVAM3.1 (Figure 5.8). Figure 5.8 Below. Differentiating spirals and their density spectra. a. CSVAM8.1. b. CSBM12.84. c. CSVAM3.1

a.

b.

162

c.

The first two discriminant functions encompassed 87.6% of variance. The spirals of class 1 clustered with classes 2 and 3 on function one (eigenvalue 20.307) (Figure 5.9).

CSVAM8.1 CSBM12.84 CSVAM3.1

Function 1 -0.846 -0.819 -0.200

Function 2 0.232 -0.263 -0.912

Figure 5.9 Distribution of the clustering classes and variable loadings on the first two discriminant functions, Shang Dynasty and Western Zhou Period. This function was characterised by two representational spirals, depicting a horn and a tail. Both spirals showed an increasing expansion rate and represent the two types of expansion typically found in depictions of objective natural forms – dense low expansion curvature followed by a sudden expansion in CSBM12.84 and a more

163 constant increase in CSVAM8.1. These are the typical form patterns of the class 2 and 3 spirals. The similarity of class 1 to classes 2 and 3 on this function is more difficult to explain given the constant and low expansion rate of most of the spirals in this class, it may be a result of a common frequency peak around 0.02 arising from the angular nature of CSVAM8.1 and the angular spirals in the class. This effect also explains the inclusion of especially oblong or rhombic angular spirals which do not have a representational interpretation in class 3. Classes 4 and 5, consisting of spirals with a constant expansion rate and high rotational length, are the most strongly segregated by function one. Most of the spirals they contain are Late Shang Dynasty or Western Zhou Period. This clustering result may have been skewed by the inclusion of the Western Zhou Period vessel CSBM17, which represents an unusual regional style variation (Rawson 1987, cat. no.27). Cluster formation may have drawn spirals into this group which would otherwise fit relatively well into the main groups. Class 1 is best characterised visually as containing angular spirals, most of which have a constant expansion rate although some spirals of increasing expansion rate explain the wide spread of values over function two (eigenvalue of 6.297). Class 4 overlaps with class 1 on this function. Spirals in this class match the quadrilateral shape of CSVAM3.1 and have a relatively lower rotational length than those in class 5, explaining the distinction on this function. Classes 2 and 5, which overlap, and class 3 are increasingly distinct from CSVAM3.1 according to their increasing expansion rate and decreasing rotational length. Examination of individual variables of style showed that orientation had a significantly non-random (p<.000), bimodal distribution (Figure 5.10). It is likely that this arises from the higher number of triangular and quadrilateral spirals in the Shang Dynasty sample. These tend to be aligned parallel to the top of the vessel but to have a slight rhombic skew causing the two peaks in the maximum diameter depending on whether they are skewed to the right or left. Loehr (1953) noted that the leiwen groundcovering spirals which make up the majority of the angular spiral sample were an important feature of his style types four and five. The high numbers of these spirals on vessels means that they are likely to be strongly represented in a random sample.

164

Figure 5.10 Shang Dynasty/Western Zhou Period orientation bimodality. The Western Zhou Period sample group did not form a significant separation from the Shang Dynasty vessels sampled on the mean value of any variable. There is considerable stylistic continuity between the two periods (Chang 1980, p.21) and the spirals from each period cannot be visually separated. 5.2.3 Middle Kingdom Egyptian case study This case study was separated by clustering into five classes (Appendix 2, Table A2.13) but the spiral decoration does not show strong separation into class groups. Only classes 4 and 5 appear visually distinct in containing more rounded spiral forms with a greater rotational length than those in the other classes. The same lack of separation was reflected in discriminant analysis. The spirals were divided according to form correlations with spirals EGBM3.3 and EGPET1.5 (Figure 5.11) on the first two functions which cover 91% of the variance. Function one, with an eigenvalue of 18.604, was strongly loaded by positive correlations to the spiral forms (Figure 5.12). Function two with an eigenvalue of 7.392 much less so.

165

a.

b.

Figure 5.11 Differentiating spirals with their density spectra. a. EGBM3.3 (marked with a red dot) British Museum acc.no. 1960, 0514.53. b.EGPET1.5 ©Petrie Museum of Egyptology University College London, acc.no. UC11131. The third variable, median path width, is characterised by the third function (eigenvalue 0.228) and separates class 3 spirals from the other groups which show little variation on the median path width. All five clusters show separation on the first function, the second differentiates the spirals into three clusters made up of class 4; classes 5 and 3 together; and classes 1 and 2 together.

166

EGBM3.3 EGPET1.5 path width median

Function 1 0.928 0.939 0.058

Function 2 0.373 0.343 -0.080

Figure 5.12 Scatter plot of spiral classes and variable loadings on the first two discriminating functions, Middle Kingdom Egypt. One extreme of difference on the first function is represented by class 2 which is made up of spiraloid or barely spiral hooks. Most of these have a relatively high expansion rate and spacing because they form a centrally linked interrelationship with another spiral so the expansion needs to accommodate a second coil. At the opposite extreme classes 4 and 5 are those with the strongest visual difference from the other classes. Classes 4 and 5 include spirals from representational depictions of plants and those which form final motifs in running lines of spirals decorating the entire scarab face. Since these fall at the end of the chain they do not link with any other spiral. They are mostly single spirals with a longer rotation and a low expansion rate. Representational spirals are spread across the class groups. The spectrum density frequency which represents an elliptical tendency in spirals is closely merged into the main density peak of elliptical spirals. Because of this, the visual difference between the elliptical spirals which form running borders and the more circular spirals or spiraloids which are more commonly part of all over linked spiral decorative patterns on scarabs is not identified in the form correlations.

167 Most scarabs in the sample were made of steatite or faience. 12 scarabs were excluded from the sample as their material, and therefore construction technique, was unknown. Gemstones were represented in too small quantities in the sample to assess whether the hardness of the stone affected the nature of the carving. Faience scarabs could not be assessed on difference in technique since both moulding and engraving techniques were used in the working of faience (Nicholson & Peltenburg 2000, p.188) and the techniques could not be securely identified in the sample. Assessment according to material type did not show significant difference between the groups on the mean values of the discriminating variables. 5.2.4 Bronze Age Cypriot case study Spirals in this case study were divided into three classes over nine levels of clustering (Appendix 2, Table A2.14). Stepwise discriminant analysis revealed that the key dividing variables were the form correlations to spirals CYBM4.1 and CYBM27.2 (Figure 5.13) which operated on function one which had an eigenvalue of 15.127 and CYBM30.1 which operated mainly on function 2 which had an eigenvalue of 4.087. A fourth form correlation to CYBM15.2 acted negatively on both functions, slightly more strongly on function 1 (Figure 5.14). Figure 5.13 Below. Discriminating spirals with their corresponding density spectra underneath. a. CYBM4.1. b. CYBM27.2. c. CYBM30.1. d.CYBM15.2

a

b

168

c

d

Function 1 Function 2 CYBM4.1 0.561 -0.576 CYBM27.2 0.258 0.845 CYBM30.1 0.567 -0.315 CYBM15.2 0.554 -0.446 Figure 5.14 Scatterplot of Cypriot spirals on the first two discriminant functions and variable loading on the functions

169 Class 1 was the largest group encompassing three quarters of the goldwork spirals and half the pottery motifs. On the second function, correlation with CYBM27.2 loaded most strongly as a factor of both the low rotational length and the low frequency of the maximum density peak, indicative of a rapidly increasing expansion rate, of the spirals in class 2. Class 2 was most clearly characterised by representational function – all of the identified plant representations fell into this cluster. Only one pottery spiral featured in this group which was not a plant representation. A number of goldwork spirals in this group resembled the style of spirals depicting plants with a constant, narrow path width suggesting that there may be a broadly plant inspired source for some of the goldwork decoration as well. Class 1 predominantly consisted of spirals with a tight central core and a sudden decrease in curvature to a straight line. In most cases, this arose from the motif being a component of a linked pair of spirals. The cluster divisions for classes 1 and 3 were not entirely satisfactory from a visual perspective. One visually coherent group of tightly coiled, regular spirals with a constant expansion rate was split between classes 1 and 3. Since these groups were most strongly divided on form variables, it is likely that the visually inappropriate split of this group of spirals arises from the problem of the maximum spectrum density peak being affected by rotational length. Class 3 spirals, the most strongly segregated of the classes, show a negative relationship to the forms of all the characterising spirals. With spirals CYBM30.1 and CYBM15.2 this can be related to the increase in expansion rate at the end of the spiral. With CYBM4.1 the negative relationship seems obscure and arises from the impact of rotational length on form description. Visual differentiation can be made on the presence of a large open centre of the spirals in class 3. All the class 3 spirals are those presumed to have been painted whilst the vessel was being turned on a wheel. This class is the most strongly differentiated in form and rotational length and is typical of the stark contrast between the Late Bronze Age/Early Iron Age Geometric decorative style in the Aegean which appears to have had an origin in Athens (Murray 1978, p. 83) and earlier more flowing and less regular decorative styles. Discrimination of spirals according to their construction technique initially identified irregularity as a key distinguishing variable. This may be attributed to digitisation error

170 (see section 5.5.1) in the case of guided painted spirals, and damage in the case of stamped spirals, so was excluded from subsequent analysis. Without this variable, discrimination operated on two form correlations with one spiral, CYBM39.4, and six variables of style – the first coefficient of path width, median path width, the first and second coefficients of spacing, median spacing and irregularity (Figures 5.15 and 5.16). Of these, no individual variable loaded strongly on function one and only the path width variables had a loading above 0.3 on function two. Eigenvalues for both functions were also low 4.436 on one and 1.352 on two.

Figure 5.15 Spiral CYBM39.4 and its spectral density plot. Four technique classes were assigned – freehand painted decoration; 'guided painted' decoration, presumed to have been carried out on a wheel; stamped decoration on metal; and decoration where a recurring stamp shape could not be identified which was presumed to be freely incised into the gold sheet. Segregation of technique groups was poor (Figure 5.16).

171

Spacing trend 2 CYBM39.4 Median path width Path width trend 1 Median spacing Spacing trend 1

Function 1 0.229 0.16 0.120 -0.024 0.245 0.085

Function 2 -0.41 -0.53 0.477 0.787 0.208 0.145

Figure 5.16 Distribution of Cypriot spirals according to manufacturing technique and loadings on the first two discriminant functions. The separation on function one of the freely painted and impressed spirals from stamped and guided examples is most strongly affected by spacing. Painted spirals are likely to have a higher spacing than the stamped spirals for which three-dimensionality can distinguish the coils with little or no spacing needed. The low spacing values of the guided spirals support the supposition that they were meant to resemble concentric circles. A large spacing value and in particular variation in spacing would make this illusion less effective as the eye would be able to follow the line better and nonparallel lines arising from curvature variation would appear incongruous. Function two reflects the regularity of the guided spirals. Median path width is lower and tends to be constant. This is reflected in the loading of the first path width trend coefficient indicating a tapering towards the centre of the stamped and free impressed spirals. There is a broad homogeneity across the spirals in the Cypriot study. The segregation of classes 2 and 3 appears to have arisen from representational significance and from

172 variation in technique respectively although neither of these divisions was absolute. There was no clear distinction between spirals on the goldwork and on pottery vessels except for the tight, regular spirals of group 3 which were exclusively found on pottery. The inclusion of the maximum diameter variable would certainly have allowed a distinction to be made but this is most likely to be a function of the size of the decorated object so is not a perceptually significant difference. 5.2.5 Cross-cultural comparison The scope of decoration covered in this analysis is too narrow to be likely to reflect fully the variety of decoration in a culture. The second research question asked whether spiral motifs considered in isolation from context showed characteristic differences or whether homogeneity and limits on variation could be identified across cultural boundaries. The results obtained from discriminant analysis carried out on the full set of spirals from all four cultural case studies differentiated the cultures on form correlation with spiral EGPET8.2 (Figure 5.17) and with the median spacing variable.

Figure 5.17 Spiral EGPET8.2 (triangle marking the relevant spiral, detail from UC29073, ©Petrie Museum of Egyptology University College London) and its density spectrum.

173

EGPET8.2 Spacing median

Function 1 0.892 0.827

Function 2 -0.452 0.561

Figure 5.18 Cross-cultural spiral distributions on the first two discriminant functions and loadings of variables on the two discriminant functions. Although Cypriot and Egyptian spirals are slightly differentiated on both functions in their extreme values (Figure 5.18), the scale of difference is too low to identify a clear distinction between these cultures. The extremes seen in the Cypriot and Egyptian examples reflect rotational length most strongly; with the long Cypro-Geometric spirals which are assumed to have been produced with some form of mechanical aid contrasting with the spiraloid or barely spiral motifs on scarabs which make up interlinked running spiral patterns so have a low rotational length to accommodate two or more spiral paths wrapping around each other in a small space. The median spacing value is similarly affected. Both the Cypro-Geometric painted spirals and the spirals punched into gold sheet have an exceptionally low median spacing value. In the case of the pottery, this may be to create the impression of concentric circles. The low spacing of the goldwork may be a purely aesthetic choice or may imitate coiled wire. Proportionally to their overall diameter, the running spirals on scarabs have a particularly high spacing, again to accommodate the linking of several spirals. The Jōmon , Shang and Western Zhou distributions overlap each other and the

174 Mediterranean case studies. Assessment of the distribution of individual style variables in isolation also showed no significant cultural differentiation with the exception of curvilinearity discussed below (p.177). The two regionally paired case studies were chosen to study whether commonalites could be identified within the pairs that might suggest contact and a regional style which contrasted with the other case study pair. In no instance could a significant difference be identified on the basis of this pairing either taking the discrimination results or the distributions of individual style variables. The most striking aspect of the cross-cultural comparisons was the apparent similarity between cultures and lack of detectable variation in the expression of individual variables. The distribution of spirals shown in Figure 5.18 is most influenced by form variance. Both this distribution and the division of spirals according to their mathematical form resemblance show that data are skewed to a relatively small range of form expressions. As discriminatory measures, style variables were largely redundant with only variables relating to path width and spacing registering as relevant variables for discrimination. Scatter plots of the distribution of individual variables by culture show why the style variables did not register in discrimination – there is a great homogeneity between cultures on all variable expressions. The distribution of the style variables was not, however, always random suggesting that there are some constraints on the typical shapes of spirals which cross cultural boundaries. The Cypriot case study included some spirals of exceptionally high rotational length and similarly long spiral paths are typical of other cultures such as painted motifs on pre-Dynastic Egyptian Naqada II period pottery vessels. However, the large majority of spirals in each case study were of no more than 3.6 rotations length and, with the exception of the Cypriot study, the mean value was around 2.5 rotations (Figure 5.19).

175

Figure 5.19 Distribution of rotational lengths by case study. There are a number of possible explanations for this cross-cultural consistency. An aesthetically acceptable cut off point for maximal length may be implied across cultures, possibly where the motif can be readily perceived as a spiral and further rotations are redundant and possibly confuse the interpretation of the form by encouraging the eye to skip between consecutive rotations. It may also be relevant to consider the finding during the digitisation of spirals for this research that, in tracing spirals of high rotational length fatigue and therefore error increased (section 5.5.1, p.194). Depending on the tightness of the spacing and the desire for regularity, it may be that at a physiological level, around 3.6 rotations represents a suitable maximum length for avoiding the onset of increased irregularity particularly in mass production where the time available for the production of each motif will be limited. The form of the spirals will also impact on maximum rotational length; spirals with a constantly increasing expansion rate grow very rapidly in size over a few rotations so spirals approximating natural forms with logarithmic characteristics will be limited in the number of rotations by the available space. It is likely that expansion rate and spacing are reflected in the rotational length of the Egyptian sample which has the lowest range of all the cultures. Spirals which link together at a central point, as most of the scarab spirals do, require a relatively high expansion rate and have little room for a high rotational length before the paths merge into each other.

176 Irregularity could not be shown to be non-random in any of the case studies except the Cypriot spirals where the non-random indicating high irregularity distribution may be attributable to tracing error (see section 5.5.2, p.195). Trends in path width and spacing along the course of the spiral path show undesirably high regression error from a third order polynomial curve from a methodological point of view (see section 5.5.5, p.196). However, from a cultural comparative regard, they show remarkably high agreement with a trend line which accommodates only two changes in magnitude of curvature along the length of the spiral. Median path width and spacing values (Figure 5.20), which were proportional to a spiral maximum diameter of 10, showed significant differences amongst the four case studies although all within a limited range of variation which suggests a consistent preference for a balance of equal path width and spacing though with greater variance in spacing.

a. median path width

Cyprus

b. Egypt

Jōmon

Cyprus Egypt Jōmon Shang/Western Zhou

<0.001 <0.001 <0.001

<0.001 0.004 ns

<0.001 0.004 ns

median spacing

Cyprus

Egypt

Jōmon

Shang/ Western Zhou <0.001 ns ns -

Shang/ Western Zhou Cyprus <0.001 <0.001 <0.001 Egypt <0.001 <0.001 <0.001 Jōmon <0.001 <0.001 ns Shang/Western Zhou <0.001 <0.001 ns Figure 5.20 Median path width (a) and spacing (b) and significant differences in distributions between cultures.

177 The clearest differences which can be seen between cultures are in the tendency towards high spacing in the Egyptian spirals reflecting the interlinking of spirals at the centre and the higher number of spiraloids in this study than the others generating a relatively higher spacing and lower path width. The Cypriot spirals have the lowest range, accounted for by the guided spirals with a particularly high rotational length so low proportional path width and spacing values. With the exception of the Shang/Western Zhou data discussed in section 5.2.2 (p.163), the distribution of orientation was not significant in any of the case studies. The quadrilateral spirals of the Shang were also represented on the curvilinearity variable (Figure 5.21). All case studies showed a significant distribution (p=.012) for the combined Shang/Western Zhou data, p<.000 for the other three). In each case though, the trend was towards curvilinearity and the Egyptian, Cypriot and Jōmon cultures did not show significant differences in distribution but all three differed significantly from the Shang/Western Zhou study (p<0.01). This agrees with Bar and Neta's (2006) finding of a preference for curvilinearity in a laboratory study of preference for curvilinearity in abstract shapes and real objects. They attribute this to an aversion to angular objects as potentially injurious. The unusually high occurrence of angular spirals in the Shang/Western Zhou study may be a particular attribute of that culture or it may reflect the context in which the spirals are used. In none of the other case studies are spirals used as a two-dimensional background pattern (Figure 4.7, p.124) (the two-dimensional spiral patterning on the Cypriot goldwork is the only decoration so should be considered a foreground pattern) and in this context, angular spirals allow complete coverage of the ground whereas highly curvilinear spirals would only provide incomplete coverage. The possibility that the use of angular spirals is an adaptation to a specific decorative need is raised because of the trend running across all cultures for a high degree of curvilinearity suggesting that more circular forms are usually found desirable.

178

Figure 5.21 Distribution of curvilinearity values by culture. That four distinct culture groups should show such commonality in motif structure suggests that some underlying principles guiding the shapes of the spirals do run across the culture studies.

5.3 Endogenous visual phenomena The clustering of EVP spirals to identify type spirals for comparison with the spirals in the cultural case studies yielded five type groups (Appendix 2, Table A2.15). The large majority of the spirals sampled fell into two groups (named Type 1 and Type 3). The three remaining types consisted of a small group of six spirals (Type 4) and two individual spirals which were sufficiently differentiated in the clustering that it was necessary to keep them as separate types (2 and 5). These two spirals, FC12.4 and FC1.1, were excluded from further analysis due to the small sample size of the clusters. Type 1 spirals include all the theoretical EVP examples modelled by Tass (1997). The strongest style characteristic of this group is a low rotational length ranging from 4.8610.79 radians. The majority of this type are computer generated models so regularity is generally high although the total range of regularity is large. Path width and spacing median values were highest in this type group. Type 1 also showed greatest variance in path width and spacing trend values. This can be accounted for by the majority of the computer generated models having a high expansion rate of both path width and spacing and the small number of drawn spirals in this group having much lower values. The few drawn spirals in the type 1 group tended to be those of low rotational length and relatively high median path width.

179 The majority of the type 3 spirals were longer, but overlapped in rotational length with the type 1 spirals. Spirals in all three main type groups showed generally high curvilinearity but, within this group, curvilinearity showed a bimodal distribution. This difference was more apparent in measurement than visually and reflects the presence of short straight sections within largely curvilinear spirals. Type 3 and 4 spirals both had a very low mean median path width, although the range of type 3 path width median values overlaps almost completely with type 1 examples. This reflects the spirals sampled from Becker's (2005) test which were drawn by most subjects as a single line, the width of the drawing implement, with no width represented. As with the drawn spirals in the type 1 group, the type 3 and 4 spirals tended to have a low spacing median and trend values for path width and spacing are clustered around zero, reflecting little change in values along the length of the drawn spirals. The spirals drawn as representations of EVP are notably more irregular than those found in decorative contexts (Figure 5.22). There may be a number of explanations for this – the subjects are being asked to represent what they see rather than produce a work of art so aesthetic considerations will be reduced; most of the subjects were not artists and were not selected on the basis of perceived artistic talent or draughtsmanship ability; EVP are difficult to focus on and to draw (Maclay & Guttmann 1941) so this may encourage tentative and sketchy reproduction and the apparent association of spirals with rotating images may exacerbate this difficulty. The tidying and smoothing of imagery in artistic representation may also arise from the artists drawing on the secondary experience of those who had entered a trance state (LewisWilliams and Dowson 1988). Knoll et al. (1963 cited in Lewis Williams & Dowson 1988) found that the passage of time after experiencing subjective visual phenomena caused people to generalise what they had seen and to create more discrete and defined representations.

180

Figure 5.22a. Irregularity of EVP spirals. Types 3 and 4 represent the majority of drawn examples. b. Irregularity of the decorative case study spirals. Higher values indicate greater regularity. Type 4 spirals exhibited the highest rotational length and irregularity and exhibited the greatest homogeneity in path width and spacing. Amongst the drawn spirals there was no apparent difference according to the nature of the stimulus. Representations of EVP spirals by those who have experienced them differ markedly from those in Tass's theoretical models. The modelled spirals more closely resemble Klüver’s (1967 cited in Ermentrout & Cowan 1979) classification which uses the term spiral in reference to high-expansion-rate whorls. There may be a poor match between the modelled and actually experienced spirals. Alternatively, it could be that the drawn spirals represent an attempt on the part of those experiencing them to convey rotational movement rather than a perceived spiral form. One drawn example includes motion arrows around the spiral form. The impact of training cards in the experimental examples cannot be ruled out. Discriminant analysis used the form correlation of three spirals to separate the type classes (Figure 5.23).

181

a.

b.

c. Figure 5.23 Form correlation spirals and density spectra for spirals a. FC20.8, modelled by Tass (1997); b. FC1.2, drawn by a person with schizophrenia (Horowitz 1964); and c. FC14.2 modelled by Tass (1997). The form variables were loaded on two discriminating factors. The first, with an eigenvalue of 18.177, dividing types 1 and 3 from type 4 (Figure 5.24). Type 4 spirals tend to have the greatest rotational length and a low expansion rate with little fluctuation in spacing and path width. Types 1 and 3 are not visually similar although they are grouped closely together on both functions. The contrast between the visual assessment and the clustered results probably arises from the tendency for shorter

182 rotation to create a sudden drop in signal density from the peak so creating a resemblance to the spirals and spiraloids of type 1 with the very high expansion rate. The second factor had an eigenvalue of 6.044 and primarily discriminates type 3 spirals from types 1 and 4 although the distinction is not strong and for any type on this function. That the most separated group is type 4 is visually inappropriate and difficult to explain. It appears to show up a flaw in the method which has not arisen in the case study analysis. The discriminant plot showed a high variance around the centroid in each type cluster. This lack of coherence within each type weakens them as a basis for identifying a single type spiral for comparison with the decorative examples.

Function 1 Function 2 FC20.8 0.725 0.164 FC1.2 0.705 0.142 FC14.2 0.571 0.474 Figure 5.24 Scatter of EVP spiral types and variable loadings on the discriminant functions. 5.3.1 Comparison with decorative spirals In testing the EVP type spirals against the decorative spirals within the case studies, spirals which fell within one standard deviation of the mean for each of the diagnostic variables were deemed to be a potential match to an EVP form (Appendix 2 Table A2.15). The Cypriot and Shang/Western Zhou spirals had the highest number of spirals which accorded with the parameters of type 3. Types 1 and 4 did not have a high representation in any of the culture groups. Two Cypriot spirals fell within type 4

183 boundaries and one each into type one within the Shang/Western Zhou and Egyptian studies. The type 1 Shang example is part of a representational image – a taotie horn. This illustrates the risk of overlap of potential EVP inspired spirals with natural real forms. Within the range of spirals in each case study the chance of a random match to any of the type groups was less than 0.000. This suggests that the numbers of spirals which accord with type 3 are significant. The variance on each of the discriminating variables in each of the cultural case studies overlaps with the EVP spiral types with the exception of type 4 spirals. These had a low overlap on the Egyptian case study with only five of the nine variables having a range which overlaps with the range of expressions of the Egyptian spirals. The occurrence of the type 3 matches across these three cultures suggests that there may be a culturally independent common source, such as EVP but other exogenous factors cannot be discounted by this study. The stamped gold spirals make up all but one of the Cypriot type 3 group and, as noted above (section 5.2.4), there is a possibility that these were created to resemble coiled wire decoration rather than referencing any natural form. An alternative explanation for the prevalence of this type group can be offered in a fit to a natural spiralling hand movement. The spirals sampled from Becker (2005), in particular, have a tendency towards an elliptical form which is not seen in any of the modelled spirals or those published as type examples of EVP spirals. An elliptical tendency would accord with a natural hand movement in which lateral hand movement is freer than the perpendicular axis. The majority of type 3 Shang Dynasty spirals are slightly angular. As rounded angular forms register similarly to ellipses in their spectral density, this would account for the matches with the ellipses in the EVP group. It is not possible to separate out the influence of the physiology of hand movement from a possible perceptual influence arising from EVP. Both the Cypriot and the Shang Dynasty type 3 spirals form part of two-dimensional ground covering patterns. It may be that the appeal of this spiral type's dense shape and constant expansion rate lies in its potential to create a regular covering pattern without substantial gaps and so is unlikely to be of EVP origin. Although not sufficient evidence on its own, the distribution of rotational length across the case studies might be used to support an interpretation of decorative spirals as

184 having been inspired by EVP since a distinction can be drawn on this variable from the decorative representational forms. The range of distributions seen in Figure 5.19 (p.175) is not significantly different from the rotational range seen in the EVP forms (the modelled forms are represented by the spike at seven to ten radians) whereas the distribution of rotational length found in spirals of known representational origin is significantly different (p=0.006) (Figure 5.25). This suggests that it is not imitation of natural real examples governing this distribution, although it might be mechanical or visual economy.

a. b. Figure 5.25 Distribution of rotational length (radians) for: a. EVP spirals; b. representational decorative spirals The necessary use of secondary imagery in drawings inevitably reduces the ability to identify clear EVP type spirals. In the sample used in this research, some spirals were drawn by experimental subjects who had been guided in what to look for. Examples derived from pathological abnormal visual function had often been assigned representational interpretation by the subject which may have influenced the presentation of the drawings (Appendix 2, Table A2.6). The first factor could be controlled for in experimental conditions but the inclination to interpret visual effects (also seen in the perceptual survey) is always likely to affect their presentation. It is likely that type groups derived from drawings will always be subject to uncontrollable external influences. It was not possible within this study to obtain evidence of a clearly defined EVP spiral type. As a result of this, although the form and rotation variables of the sampled spirals were not inconsistent with an EVP influence, other influencing factors could not be excluded and no influence of EVP spirals could be identified.

185

5.4 Inspiration for spiral decoration The question of whether natural representational inspiration for decorative spirals affected the form and style of the spirals in a cross-culturally consistent manner was difficult to address with the data from the case studies as there was relatively little overlap in representational types between case studies and representational meanings may have been missed restricting the data set even further. The identification of representational meaning in itself can be subjective. Hirschfeld (1977) noted that there was not consistency of attribution of representational meaning by the producers of embroidered designs. This implies both that representational meaning is flexible and that a strong contextual basis is needed for attempting to assign a representational meaning to abstracted spirals motifs. The meanings attributed in this study were used on the basis of consistent references across a number of publications and in most cases there was sufficient contextual information in the decoration to identify representations at their most basic level such as tail or horn. Spirals with a known representational meaning from all four case studies were collated in a single clustering test to see if the spirals separated according to subject (Appendix 2, Table A2.16). Shang/Western Zhou spirals made up most of the data set and these depictions tended to be elements of mythical creatures for which there were no comparable equivalents in other culture samples. There was a very high form correlation between spirals in this clustering group, with the exception of the three spirals separated out in the cluster analysis, the lowest mean correlation was 0.957 and the three outliers were all above 0.7. The highest mean correlation within the full data set was 0.858. In clustering, the majority of spirals were placed in a single group and only four clustering steps were required to bring together all the spirals. Three spirals were strongly segregated from the main group. One of these (CSBM30.46) was part of a conventionalised depiction of a cicada. The other two spirals separated from the main group were both from the same Western Zhou Dynasty vessel and depicted snakes with a constant path width and a low and constant expansion rate in contrast to most of the larger group. The three separated spirals are most strikingly different from the rest of the group in their rotational length and it is probably this that caused the separation. Discriminant analysis was not carried out on this study because of the small size of classes 2 and 3. There is neither a significantly

186 high correlation nor an apparent strong visual similarity between the form of Western Zhou Dynasty, Jōmon and Egyptian scarab snake representations even though the Jōmon and scarab examples were placed in the same group. Snake depictions might have been expected to differ from the rest of the natural spirals in their expansion rate, in the manner of the Western Zhou examples but the Jōmon snake depictions tend to fall at the opposite end of the spectrum with some of the highest expansion rates. The only other representational theme with sufficient cross-cultural representation to make a comparison was depictions of plants on scarabs and on Cypriot vessels. These all fell into the first large cluster which drew together several different representational types. All of the plant spirals had a high degree of overlap with the hyperbolic/logarithmic mathematical spiral group. Unsurprisingly, given the high path width of some of the plant depictions (e.g. CYMM12.2), path width median showed a significantly higher value (p=0.033) for the plant depictions than for the other types of representational spiral but form correlations were not significantly higher within the plant group than with the rest of the sample. The mix of representational types and the split of snake depictions between clusters both suggest that it is not possible to differentiate spirals clearly by subject between cultures. Discriminant analysis on the full, cross-cultural data set to divide representational and non-representational spirals identified three discriminating variables: form correlation to JKAE15.2; median spacing value; and irregularity (Figure 5.26) operating on a single function with an eigenvalue of 0.145. Of the identified variables, median spacing did not load strongly.

Figure 5.26 Spiral JKAE15.2 and its spectrum density.

187 The representational and non-representational spirals showed a very high degree of overlap (Figure 5.27) although the variable means of JKAE15.2 and irregularity both showed a significant difference between the two groups. The slight positive skew in the representational motifs reflects agreement with the spectrum of JKAE15.2 which resembles a logarithmic spiral with higher frequency peaks, characteristic of angular spirals. This possibly reflects the high proportion of Shang/Western Zhou Dynasty spirals in the representational group. There is no clear reason for the representational group to be more regular than the non-representational group so the higher regularity in the representational group may reflect their shorter length and be a factor of digitisation. Excluding irregularity from the analysis did not have a large impact on the results.

a.

b. Function 1 JKAE15.2 0.612 median spacing -0.078 irregularity 0.523 Figure 5.27 Distribution of representational (a) and non-representational (b) spirals and variable loadings on the discriminant function. The difficulty in discriminating representational spirals from the larger nonrepresentational group does not imply that natural forms are not important as a basis for the shape of decorative spirals. It could imply the use of shapes associated with natural forms in decoration even when the intention is not to represent the natural form. It is also probable that some of the spirals in the non-representational group did have representational meanings to the cultures producing them which have now been lost.

188 The preference for high curvilinearity seen across the case studies is consistent with a natural inspiration for spiral forms which are rarely highly angular although Gestalt principles of good continuity and the nature of the object on which the spirals are drawn might also account for a preference for curvilinearity. When creating a large motif on a two-dimensionally curved surface an angular motif would be liable to appear more distorted than a curvilinear example. The angular Shang/ Western Zhou spirals could avoid this problem as they are smaller than the decoration on vessels from the other case studies so apparent distortions of curvature would not be obvious. Within each case study culture, both mathematical spiral grouping and cluster analysis proved effective at bringing together spirals that are of representational origin. Differentiation analysis was also carried out on classes defined by the technique used to create the spiral. This was intended to identify both whether the motifs split into markedly separated groups on these lines and, if so, the variables which were affected by variation in contextual attributes. Imitation of technical artefacts and the method of construction may explain some of the consistencies between spiral shapes. A distinct technological basis for angular spirals from curvilinear forms may have arisen in woven textile patterning or possibly more distantly to basketry although this latter is more speculative. Although angular spiral forms are highly associated with woven textiles, their earliest known occurrence at Mezin (Figure 1.5, p.26) predates woven textiles and in the context of the other decorative forms found at Mezin (examples shown in Gimbutas 1989, p.5; Iakovleva 2009) suggests an alternative possible origin as an elaboration of chevron patterns. As with woven spirals, the early creation of decoration with straight lines may be a product of the type of decoration which the technology available best lent itself to – carving straight lines with the edge of a stone tool would be easier than carving curves, particularly on a non-flat surface, where using a smaller cutting edge to enable the creation of a curve would be more likely to cause the tool to slip across irregularities in the surface. Tight coiling with no spacing between coils is widely found in decorative spirals created using wire. There are clear practical advantages in tight coiling in this medium: the wire guides itself; the decoration becomes more visible than separated strands of wire would be; and the consecutive coils support each other and are less susceptible to damage than an open structure.

189 Discriminant analysis of the method of production was run on the cross-cultural data. Eight methods of production were classified: free painted, found only in the Cypriot study; guided painted, also only in the Cypriot study; carved and carved/moulded found in the Egyptian study; stamped, found in the Cypriot study; modelled, found in the Jōmon study; incised, found in the Jōmon study, two examples of Cypriot goldwork and a single scarab; and cast, making up the whole of the Shang/Western Zhou sample. Since the Shang/Western Zhou sample were cast into incised clay moulds similarity might be expected between these last two groups if technique is an important influence on spiral shape. Four functions were identified, the first two accounted for 87% of variance and the third for 9.3%. Eigenvalues of the first two functions were, however, low at 1.147 for function one and 0.399 for function two. Three form correlation variables and median spacing were identified as significant (Figure 5.28).

190

a.

b.

c.

Figure 5.28 Discriminant variables for construction technique: a. JKAE10.2. b. CYMM24.1. c. JBM4.1 and spectral density plots. The distribution of technique groups on functions one and two separated the guided painted spirals most strongly on function one (Figure 5.29). This matched the opposite loadings of the extended spiral of low expansion rate and the short spiral of high expansion rate on this function. Although no group was entirely distinct the guided spirals were most different, also having the greatest extreme of positive variation on function two. The two strongest loadings on this function would have opposite impacts on the guided painted spiral group with form correlation increasing the values and the positive association with median spacing lowering them. Median spacing also separated out the stamped spirals, the only other group which showed some distinction from the general cluster. That both these Cypriot groups are distinct probably does imply a distinction by technique rather than culture because the two

191 spiral types are separated by 300-800 hundred years so cannot be considered an idiosyncrasy of a single culture. No other technique showed any strong distinction from the majority of the data which clustered around zero with a slightly positive trend on both functions. Although the Cypriot guided painted spirals were somewhat separated, the extent of variation in the whole sample does not suggest that differences in the manner of production have a strong impact on spiral shape.

JKAE10.2 CYMM24.1 JBM4.1 median spacing

Function 1 0.922 -0.714 0.954 0.527

Function 2 0.149 0.497 0.022 0.646

Function 3 0.073 0.430 0.076 -0.461

Figure 5.29 Distribution of spirals by production technique and loadings of the variables on the first three functions. The chart of cultures in which spirals were commonly used motifs (Appendix 2, Table A2.1) was produced to assess whether there was an apparent link between the introduction of a new type of material or a technological development and the introduction of spiral motifs. No pattern emerged from this assessment.

5.5 Evaluation of the method In the development of the method used in this research, the aim was to devise a system which provided an objective quantitative description of spiral motifs which could be used to differentiate the motifs and to assign them to classes according to

192 visually important properties. If the method were to be practical in building up a set of data for cross-cultural comparison, it needed to be accessible and simple to apply. Application of the method to the case studies showed that the variables included within the method could be used to create groupings of spirals for which plausible justifications could be offered based on visual assessment and that these groups differed from each other at a statistically significant level. This suggests that the variables chosen for measurement are appropriate for judging visual divisions, at least according to the cultural perspective of the author. How much these classes reflected the cultural choices of the producers and consumers of the decoration is, of course, uncertain. This caveat must be applied to any cross-cultural comparison. Having quantified the decoration, it is possible to carry out discriminant analysis to identify the key variables which cause one set of motifs to be considered different from another. This is more readily achieved with a set of quantified data than with subjective assessment where the impact of individual variables may be difficult to separate from the cumulative effect on the viewer. Subjective classification of the spiral motifs in isolation would avoid some of the obviously aberrant outliers which arose in the quantified classification but attempts to create subjectively assessed class groups based solely on visual criteria prior to the spiral clustering showed that many class divisions were unclear. This method allows decoration to be considered along a continuum of variation and measures of similarity and of the variability within a group to be made. Where subjective judgment does have an advantage over this classification system is in the ability to assign subtle weight gradations to variables according to the perceived importance of the variable within a single motif. This involves a complex interaction of the expression of each variable within the motif. Provided that their judgement is culturally valid, it is likely that a human’s assessment of visual difference will be better than a quantitatively based system but, where an assessment is being made without a means of identifying local perceptual judgment, a stronger case can be made for using a quantitative description.

193 One of the main difficulties in developing the method was in ensuring that the method was robust across the full range of decorative spirals. Many of these problems arose in the accommodation of factors where the appropriate manner of measuring a variable was visually clear but computationally ambiguous so liable to create aberrant results in the automation of the method. As an example, the measurement of path width at the end of spirals could not consistently measure to the end of both sides of the spiral to produce results which gave an accurate impression of the actual path width (Figure 5.30). The difficulty shown below of calculating path width appropriately was resolved by stopping measurement of path width if the same point of one line was measured repeatedly as the closest so causing measurement to stop in Figure 5.1b when the lines are no longer parallel

a.

b.

Figure 5.30 Variability in the appropriate interpretation of path width at the end of a spiral. Calculated path width to the closest point on the outside line at the outer end shown in green. A learning process has operated within the method as it was applied to the case study data allowing its implementation to be refined as previously un-accommodated spirals were encountered. The analysis method in its current implementation is believed to cover a wide range of decorative spirals and the addition of further data can be used to continue the learning process as new forms and stylistic expressions are encountered. Although it was intended that the classification system should be objective, subjective judgements were employed at several stages of image processing and the degree of class division accepted within the clustering results was also based on the visual inspection and the subjective assessment of similarity. To devise logical rules for identifying, for example, the distinction between figure and ground, a common understanding of decoration between cultures would have to be assumed. This would require a substantial quantity of contextual data be included within the analysis

194 extending to the surrounding decoration on the object and to the decorative styles seen on other objects within that culture. Although some variables could be tested according to objective rules which allowed consistency, it proved impossible to implement the analysis method as an entirely objective system. The reliability of the analysis may be affected by the subjective placement of the start point of the spiral and the consequent placement of the centre point. The measurements of irregularity, path width and spacing were dependent on the placement of these points. Subjectivity in assessment of type clusters would have a less negative impact on reliability than variation at the stage of data collection since clustering results will be specific to the research question posed. It is of greater fundamental concern that the raw data can be said be reliable. The implementation of a classification flowchart in the manner of Washburn and Crowe’s (1988) symmetry classification was inappropriate to a construct which shows continuous variation. This factor and the number of sampling points needed to produce a precise description of shape made automation of the measurement of variables the most appropriate approach. In its final form as an applet within Geogebra, the method is easy to apply and could be made easily accessible over the internet. Given a detailed specification, the method could also be implemented in other software. In each of the case studies quantitative outliers occurred which did not have a good visual match with the class in which they were placed. This suggests that there is scope for refining the method and improving some of the ways of measurement. During the application of the method to the case study data, some possible improvements which could be made to the digitisation of motifs, the measurement of style variables and the quantification of spiral form were identified. These are detailed below. 5.5.1 Digitisation Tracing error is highest at the start and end of the spiral and at the start and end of interpolated sections when the stylus is lifted from the tablet. Error at the start and end cannot be eliminated. Interpolation error can be reduced by changing the method

195 of recording the start and end of interpolated sections to a key press without needing to lift the stylus. Further errors in tracing are introduced where the diameter of the spiral reached a point where hand movement and change in the angle of the stylus are required to continue tracing. At such points, the stylus could be kept in contact with the tablet but some erratic movement arising from the adjustment inappropriately increased measurement of irregularity. This was particularly evident in the Cypriot case study. With longer spirals, fatigue and increased tremor from gripping the stylus for a prolonged period may have had some negative effect on the accuracy of tracing and measured irregularity. This problem is more difficult to resolve. The ideal solution would be an effective means of automatic edge detection which could be manually edited to control interpolation and location of the ends of the spiral. Failing this, a rig which allowed the stylus to be held upright and provided a larger gripping area might reduce, though not eliminate, these errors. 5.5.2 Irregularity The problem of minor damage affecting the assessment of irregularity might be resolved by introducing two levels of interpolation into the tracing of the spirals. Interpolated parts of the spiral which are clear enough to be included within the assessment of other variables could be excluded from assessment of irregularity. This would, however risk making irregularity unmeasurable for some spirals. As noted in 5.5.1, there is a cost in tracing accuracy in increasing the number of times the stylus needs to be lifted and re-placed. 5.5.3 Curvilinearity In application, the measurement of curvilinearity proved to have a drawback in that the appropriate maximum angle of curvature and minimum number of path segments that fell within that range to determine a straight section within the spiral varied according to the average curvature of each individual rotation (Figure. 5.31). This problem was manifest in two ways. The first produced an overestimation of the number of straight segments when irregularities in spirals with low curvature registered wrongly as separate segments within what would be visually assessed as a single straight line (as might be considered the case in the right hand side of the outermost rotation of Figure 5.31 which would record as three straight segments).

196 The second caused an underestimation of the number of straight segments when spirals featured high resolution irregularity or patterning which prevented segments which appeared straight at a lower resolution being recorded as such (the left hand side of Figure 5.31 would not record as a straight segment).

Figure 5.31 Ambiguities in recording curvature. The applet identifies, a straight segment where there is a sequence of at least eight segments in which the change in angle from the first segment is less than 0.035 radians. A future method to test for improving the measurement of curvature would be whether calculation of the average curvature of each rotation of each spiral was more appropriate although the calculation of curvature of irregular shapes is complex and might be difficult to accommodate as a small number of comparable values within the classification system. 5.5.4 Orientation Orientation is given the same importance regardless of the spiral form. From a visual point of view it would seem to make sense to weight the importance of orientation according to how far the spiral deviates from a circular shape. Orientation is likely to be considered perceptually more important in spirals with strongly differing maximum and minimum axis lengths than in those where the difference is negligible. This would also permit the isolation of spiral motifs where orientation is more likely to be a discernible variable for separate consideration of this variable. 5.5.5 Path width and spacing The fitting of polynomial curves to describe changes in the trend of path width and spacing imposes an external order on measurement in a manner which this method is trying to avoid. Any form of parameterisation of trend in path width and spacing to make it comparable in a quantitative system will require some degree of imposed

197 artificial structure. The mean square error (MSE) of the trend line varied quite considerably, with some spirals being poorly described. The spiral EGLIV7.4 for example had a path width trend MSE of 235.759 due to a fluctuating path width (Figure 5.32). In this instance a ninth degree polynomial would provide a better fit but increasing the degree of polynomial without good reason can result in overfitting – including random noise within the trend and reducing its predictive value (Blöschl & Grayson 2000). Where fluctuating variation was not adequately accounted for by a third degree polynomial in multiple spirals, the fitting of a higher polynomial curve of a higher degree might be appropriate but the full scarab case study, of which EGLIV7.4 was a part, showed an average MSE for path width of 5.843 so a third order polynomial was felt to be acceptable.

Figure 5.32 Path width trend for spiral EGLIV7.4 (left, primary path in yellow), right: actual path width measurements in black, third degree polynomial shown in blue, ninth degree polynomial in red. Non-parametric regression such as spline fitting which takes a bottom up approach to extracting trend (Klasson 2008) and would appear more appropriate than predetermined curves does not yield results in a form which could be included within cluster analysis. A polynomial curve was therefore felt to be the best compromise in giving an impression of the perceptually important variables of path width and spacing trend. 5.5.6 Interpretation of the spectral density plots The form of the spirals is represented by a spectral density plot of the tangent angles of the spirals transformed into a sine wave (section 3.4.1). By identifying frequencies of signal with particular characteristics of decorative spirals they can be used to identify features of the spiral form. The frequency value of density peaks in the plot

198 represents the length of rotation needed for the tangent angle of the spiral to return to its starting value, calculated as (rotational length in radians/number of sampling points)x(1/frequency). For a circle, which is represented by a single sine wave, a single density spike would appear at a frequency of 0.0039 so the calculation would be (6.283/256)x(1/0.0039)=6.293 showing that (allowing for rounding error) the tangent angle of the circle takes one rotation to return to its starting angle. For spirals, the movement of the path away from the centre causes the sine transformation of the tangent angles to a wave plot composed of a high number of sine waves of differing frequencies. The average expansion rate of the spiral is therefore represented by a main frequency peak and peaks of lower magnitude following this representing the manner in which the spiral expands. For the purposes of describing the effect of expansion rate on the shape of the spectral density plot, mathematical spiral types with clearly defined characteristics are used as exemplars, decorative spirals may be expected to have much noisier density signals than mathematical forms due to small or large scale irregularities of the curve. The tangent angle of logarithmic and hyperbolic spirals (Figure 5.1d and e, p.152), with an increasing expansion rate, takes more than one full rotation to return to its starting value. The average value of these curves results in the highest magnitude peak falling at a frequency close to zero due to the reducing curvature at the outermost end of the spiral. The constant expansion rate of Archimedes' spirals (Figure 5.1b), causes the tangent angle to return to its starting value in less than one rotation so the signal peak falls at a higher frequency than that of a circle at around 0.047. Parabolic spirals (Figure 5.1a) with a decreasing expansion rate have the highest maximum peak frequency. Euler spirals (Figure 5.1c) are the only type of spiral which can link a circle to a straight line without a cusp (sharp change in angle of the line). Although at their outermost end they have a low curvature, tending towards that of logarithmic spirals, expansion rate is much lower in the centre of the spiral so the average signal peak across the spiral falls between that of a logarithmic spiral and an Archimedes' spiral. The change in expansion rate is reflected in the shape of the spectral density plot around the maximum peak. A more rapid change in curvature is represented by a more rapid decline in density values so the Euler spiral, changing from a circle to a line shows a rapid fall following the density peak to almost zero. Logarithmic and

199 hyperbolic spirals show the next most rapid decline after the peak and Archimedes' spirals the most gradual. Parabolic spirals strongly resemble Archimedes spirals in their decline following the maximum peak but have a discrete frequency peak preceding the maximum peak which is not seen in Archimedes' spirals. Characteristic discrete frequency peaks or changes in the gradient of decline are also seen in logarithmic, hyperbolic and Archimedes' spirals (marked on Figure 5.1d, e and b). Average expansion rate is reflected in the magnitude of the maximum density peak with a higher expansion rate producing a peak of greater magnitude. The magnitude and frequency of the maximum density peak are also affected by the rotational length of the spiral with a greater rotational length tending to increase both values. The noise seen in decorative spiral density plots that reflects small scale irregularity measured by the irregularity variable is represented by low magnitude peaks at high frequencies and has little impact on the measurement of similarity. More substantial variations in spiral shape found, for example, in elliptical or quadrilateral spirals are represented by spikes in the density plot. The jagged frequency peaks created by angular spirals can be seen in the contrast between the spectra of spirals CSVAM3.2 and EGPET12.1 (Figure 5.33). The profile of CSVAM3.2 differs from that of an Archimedes' spiral in not being a smooth decline in frequency.

200

a.

b.

Figure 5.33 Contrast of angular and curvilinear spiral signal density profiles. a. CSVAM3.2. b. EGPET12.1 (marked with a red dot), (detail from UC61187 ©Petrie Museum of Egyptology, University College London). The effect of rotational length on the density spectrum created problems in classification. The highest variation in rotational length was in the Cypriot case study resulting in visually inappropriate groupings being made. It is anticipated that the method would be improved by transforming the density spectra to eliminate the effect of rotational length, possibly by dividing the spectrum values by the rotational length of the spiral. Spiraloids and spirals of very short rotational length tend to fall rapidly to zero after the maximum peak. This short rotational length created little higher frequency noise. The single rotation is more equivalent to a circle in its restricted signal range and can mimic the sudden frequency change in spirals with a rapid expansion rate causing confusion between the two forms. The variation that is a factor of rotational angle in non-stationary data is masked by an averaging effect in spectral density plots, preventing individual sections of the spiral

201 being compared. Discrete wavelet transform is more sensitive to variation and may allow individual sections of spirals to be compared and the form of the spiral to be represented more precisely. Both the variables of curvature and rotational length appear to be encompassed within the form spectrum density signal so future development of the method might be used to extract these values from the form data, particularly if wavelet transform, rather than spectrum density, were used to create a more clearly defined signal. 5.5.7 Context of the proposed method. Studies focussing on the overall structure of decoration or design in an ethnographic and archaeological context either using geometric terminology or perceptual impressions of the overall design structure are not directly comparable to the study of individual motifs in terms of their potential applications. Analyses of design structure have been extended to make comparisons between the structure of decoration and other aspects of cultural life for example social structure in addressing whether egalitarian or hierarchical social structures are reflected in symmetry repetition and the extent of ground covering within a design as a whole (Fischer 1961). Peregrine (2007) extended Fischer's research to the archaeological record to identify further correlations between assessed design complexity and social structure in political autonomy and social stratification. These assessments relied on the analysis of the full set of design or decoration. Analyses of symmetry, in isolation or as a variable in a broader scheme take the geometric concept as a means of classifying since it has repeatedly been shown to be a valid concept in decoration even where the manner of measurement has varied (e.g. Brainerd 1942; Washburn 1990; Grünbaum 2004). With geometric constructs such as spirals which represent a physical construction rather than an organisational principle, reference back to mathematical principle directly has not proved appropriate in the context of decoration. In these cases the value of a geometric structure as a basis for analysis lies in providing a set of fundamental values to define a shape across the cultures in which it appears. The geometric construct forms the basis for deciding to classify the motif rather than the means of classifying it. These two approaches provide breadth of analysis in different manners. Symmetry and similar structural concepts such as modularity and proportions can be analysed in a wider range of

202 cultures than is permitted by the study of a single geometric shape where cultures must be identified where its use it sufficiently common to argue that its presence is not coincidental. However, within these cultures, a broader range of perceptual variables can be assessed within the expression of the motif than in the identification of the presence and manner of expression of a structural concept. It is true, as Douglass and Lindauer (1988) note, that some properties of design are not applicable to individual motifs and can only be applied to the structure of the whole design. However, the simplicity of a single geometric motif makes it easier to define that a more complex design composed of multiple elements and the possibility of assigning a limited set of variables to its detailed description may allow perceptual properties to be assessed in a single motif that could not readily be addressed in the full design. As has been discussed in Chapter Four, the isolation of motifs has been criticised. Bradley (1997, pp.5, 8) especially, opposes recording motifs in isolation as losing meaning which requires the full design for interpretation. He points out that simple geometric motifs such as crosses may have a diverse range of meanings in different societies (1997, p.11). Analysis of motifs in isolation may not be effective in identifying meaning. Even within the well-defined scope of the natural representational forms studied in this thesis, it was not possible to subdivide the spirals by what they depict so to go further and interpret abstract meaning would be fraught with difficulty. Perceptual preferences occurring at a much earlier level of visual processing may be assessed independently of assigning conscious meaning and it is suggested that these may be drawn from the analysis of geometric motifs rather than meaning in the motif itself. Some of the variables analysed in this research are motif specific (for example rotational length) and could not be used easily in broader cultural analogy but preferences, for example, for regularity or curvilinearity/angularity might well be expected to reflect or be reflected in the broader environment and Fischer (1961), Takashi (2009, p.140) and Pittard et al. (2007) have suggested social constructions attached to curvilinearity in masculine/feminine associations or symbolic links to aggression. Preferences in regularity; the openness of design and ground covering could similarly have broader relevant analogy. Filatova (2008) used similar principles

203 drawn from the broader design scheme to classify Neolithic pottery of the Lower Amur region. The large majority of studies of perceptions of individual motifs have been carried out in laboratory conditions with a necessarily constrained sample group. Studies carried out in these conditions were used to support the choice of variables used in this method. The measurement of perceptual grouping values in decorative motifs might be used to assess whether similar observations of preferences could be seen in created decoration. Prior studies of spiral decoration have used two main approaches to their analysis. The approach usually taken in classification of decoration and the study of cultural transmission is qualitative assessment of their shape either as a class of motif within a broader decorative typology (e.g. Loehr 1953) or in relation to their conformity to mathematical classes. The second approach to analysis, uses quantified assessment either, again, in relation to mathematical spiral forms (e.g. Andrey & Galli 2004) or in terms of measured values which would equate to those of style used in this thesis (e.g. Herner 1989). The method proposed in this thesis tends strongly towards a quantitative approach. This has not filled the function of cultural individuation where qualitative assessments of spirals as one decorative class amongst several have usually been used. The approach taken is most similar to that of Herner but with the addition of a more detailed description of form (Herner (1989, pp.22-37) uses three shape classes to describe form). Since form was consistently the strongest differentiator of the spirals, this aspect of the method is felt to be a useful addition that might be used alongside studies of broader decorative schema to enhance the information that can be obtained from them.

5.6 Summary The analysis method proposed in this thesis proved capable of identifying perceptually coherent differences between decorative spirals. However, application to the studies in this thesis has highlighted several areas in which improvements could be made to the method particularly in the discriminatory power of the form variable for which a wavelet transform summarisation approach is suggested in future. In comparison with mathematical spiral classification, both systems were shown to have some advantages

204 and disadvantages, but the classification method applied in this thesis is felt to offer greater flexibility and make fewer assumptions about the data under consideration than the use of mathematical spiral classes. In all of the cultures analysed it was possible to differentiate between spirals using clustering to create a gradient of variation across the sample and to identify some strongly outlying classes of decoration. This illustrates that a graduated approach to the description of the motifs was more appropriate than assigning them to separate classes. By far the strongest theme to come out of cross-cultural analysis was that of restriction on the range of variation. In assessing possible reasons for this restriction, depiction of substantive natural spirals appears the most likely determinants of spiral shape. Some forms of technical production process also produce characteristic shapes but there is substantial overlap in the shapes of spirals produced by most technical methods suggesting that this is not a strong controlling factor on spiral shape. On the current analysis no case can be made for the presence of a type of EVP inspired spiral decoration that can be distinguished from the influence of mechanical, representational or hand movement constraints.

205

6

Summary and conclusions

Ethnographic and archaeological studies of the structure of decorative patterns have established that they show cultural variation and that the identification of characteristic employment of decorative structure can be used to highlight similarities between cultures or to separate those using fundamentally different decorative structures (Brandmüller, Hrouda & Wickede 1986; Filatova 2008). Within a culture, the geometric structures underpinning decorative patterning have been found to reflect architecture and the structure of settlements and even the social organisation and values of a society (Washburn 1983a, p.4). Studies in this area have tended to focus on the structure underlying the organisation of a pattern, considering, for example, symmetry (Washburn & Crowe 1988) and modularity (Jablan 2002). The analysis and classification method proposed in this thesis is intended as a test of the hypothesis that the measured properties of individual geometric motifs can be similarly used. Spiral motifs are found across a wide geographical and temporal spread of cultures. Their identification is based on a single defining property – a line winding around a central point at an increasing radius. This definition forms the basis for comparison. The manner of coiling of the spiral and any decorative embellishments to the line can be taken as variables whose expression is controlled by factors within or outside the conscious control of the creator of the decoration. It was anticipated that measurement and comparison of these variables would aid understanding of some of the processes governing their expression. The method applied in this thesis was first compared with one of the more commonly used classification systems for decorative spirals which assigns them to classes according to their resemblance to mathematical spiral forms. Following this, it was tested on four cultural case studies to establish whether differences could be identified within and between cultures on the overall variable expression and shape of the motifs. Commonalities of variable expression were searched for within the full data set and an explanation for their occurrence sought in naturally occurring spirals either as real forms such as coiling plants and animals or in relation to spiral visual

206

effects arising from abnormal functioning of the visual cortex and in the technological process used to construct the spirals.

6.1 Summary of the method The development of the method used to analyse the spiral motifs was underpinned by published studies of perceptual grouping and by the survey of the perceptions of spiral motifs carried out during the development of the method. These two sources were used to provide evidence for the validity of choice of variables and to determine constraints on the definitions of spiral motifs. It had been hoped that these two sources of data would provide a basis on which to weight perceptual variables in the application of the method but both sources were too culturally restricted to provide a basis for weighting. A further difficulty arose in that literature review highlighted the variation in weight placed on perceptual grouping variables depending on their absolute values and how they interrelated demonstrating that a single set of constant weighting values would not be appropriate. The method uses seven variables to describe and classify spirals, one describing the basic form of the motif considered as a notional infinitely thin line and the others, classed as variables of style, describing decorative embellishment applied to that line in the form of irregularity, the path width and spacing between coils and the curvilinearity, rotational length and orientation of the motif. Key considerations in the development of the method were that it should, as far as possible avoid subjective description and the application of prescriptive top-down mathematical constructs in the description of the spirals. These influences were minimised but the method was not entirely successful in either of these objectives. It may be possible to find an alternative to the use of parametric curves to describe spacing and path width but it seems unlikely that it will be possible to entirely remove subjective judgement from the analysis process and the analyst would need to be aware of this factor. At the level of practical implementation, considerations were to create a method which was accessible and simple to run. These objectives are largely felt to have been achieved; a low level of computer programming skill would be needed to implement the method as an applet and once that is done operation is simple and there is facility

207

to freely distribute the applet within some geometry software applications. Statistical analysis could have been carried out using freely available software and demands for unusual computer hardware were not high requiring only a graphics tablet for digitisation. The form data were summarised for the purpose of comparison as density spectra of the tangent sine values. Correlation matrices were used to make these values comparable between spirals. Classification of spirals within and between cultural groups was carried out by hierarchical cluster analysis and discriminant analysis was used to identify the key characteristics of the identified classes and to assess how much they differed. Problems arising in the application of the classification method to the case study data discussed in Chapter 5 highlight potential to improve the discriminatory power of the method particularly in the interpretation of the spiral form which is too strongly influenced by rotational length and where spectrum density peaks at low frequency allow confusion between spirals with sharply angular turns and those of a constant expansion rate. Improvements to the measurement of the spiral path width and spacing; the curvilinearity of the spiral, should it remain as a separate variable; and orientation could also increase the ability of the classification system to capture finer detail of variation in spiral shape and give the orientation greater weighting when appropriate. It is possible that the power of the method to distinguish spirals of different cultures could be improved by a different choice of perceptual variables to better reflect the shape and decorative properties of the motifs since visual assessment prior to analysis of the traced spiral outlines resulted in a better differentiation of cultures. The chosen analytical variables might be improved by the application to a larger sample size and broader range of case studies by discarding variables currently being measured if they show consistently non-significant distributions and by identifying more important visual variables with reference to a greater range of data. Despite the problems with the method identified during its application to the case studies, hierarchical clustering on the recorded data was able to segregate some visually coherent groupings in each case study and to place the motifs which made up

208

the main, and fairly homogeneous, clusters of each study on a scale of variation which broadly reflected changes in form of the spirals. In the creation of decorative typologies within a culture clearer groupings can be proposed by assessing material, object form and the full decorative scheme found on the objects. Although some of the clear divisions in the shape of the spirals could have been identified by visual analysis without measurement, and some aberrations of the statistical classification system avoided, it would be difficult to create a consistent graduated scale of variation on visual assessment alone because of the number of variables affecting the decorative appearance of the spirals.

6.2 Potential for future development of the method Beyond the improvements to the measurement of individual variables which may improve the precision and accuracy of measurement, there are two possible areas in which the method could be expanded. The first would improve the range of objects from which decoration could be analysed and the second could increase the range of motif types which could be analysed. 6.2.1

Surface curvature and three-dimensionality

The potential to expand the analysis method to encompass motifs on highly curved surfaces would improve the accuracy of results and allow the inclusion of more material. Two cultures which would have added interesting comparative material to those included in this research Spirals on pre-dynastic Egyptian Naqada II period pottery and spirals and spiraloids on Neolithic Chinese pottery of the Banshan and Machang cultures would have added interesting comparative data to the case studies included in this research but both studies had to be excluded because the size of many of the spiral examples in relation to vessel curvature (Figure 6.1) caused distortion in digitisation that made their inclusion within the current set of case studies impractical.

209

Figure 6.1 Spiral and spiraloid motifs on vessels with high curvature: a.Naqada II period vessel, Petrie Museum b. Banshan culture burial urn, Victoria and Albert Museum, C.286-1938. Figure 6.2 shows the distortion of a regular logarithmic spiral caused by surface curvature. Solutions to this problem might be found in the use of rectified photography or by reconstructing an approximation of the three dimensional profile of the object and projecting the traced image onto this surface so allowing measurements of distance and angle to take curvature into account. Accuracy in this approach would depend on the accuracy of the profile. Laser scanning now allows highly accurate reconstructions to be made and, where the decoration is three dimensional could, in some instances create a direct reconstruction on the modelled object surface. A lower precision method of creating object profiles by a sequence of diameter measurements and reconstructing these in a three dimensional design programme would create a more accessible means of compensating for a curved object profile. Although this could clearly only be used with objects with a simple curvature profile, its implementation would considerably improve both the accuracy of measurement and the range of objects from which motifs could be analysed.

210

Figure 6.2 Distortion of a logarithmic spiral on cylinders and spheres of increasing curvature. 6.2.2

Expanding the range of motifs

This research focuses solely on single spirals without consideration of their relationships with other decorative elements on the object or the common decorative phenomenon of two linked spirals. This approach was chosen because it allowed comparisons to be made in terms of a single, clearly defined construct. As the scope of the application is broadened to consider relational context, so definitions would broaden beyond the properties of single motifs towards examining the overall structure of decoration which has been the focus of a large body of research particularly in terms of proportioning systems and symmetry. However, it was clear during data collection that it was not always appropriate on a visual level to consider spirals in isolation from other motifs. There is a clear visual continuum between spiral and false spiral (linked concentric circle) motifs (Figure 2.7, p.41) and both the survey of perceptions of spiral motifs and

211

a number of decorative typologies separated spirals which were linked together as single motif types in themselves rather than collections of spiral motifs (e.g. Petrie 1920, p.20; Jernigan 1986) suggesting that, at least at the level of direct contact relationships, links between spirals need to be considered as an important perceptual factor. The arguments above support the extension of the method to consider a broader range of motifs under the same set of terms so that they can be placed on the same continuum of change for measurement of similarity. In practical terms, the current implementation of the analysis applet might be adapted to include non-spiral unclosed motifs composed of a single line without great difficulty although the inclusion of closed or self-intersecting motif forms would be more complex. Some motifs which are visually similar such as running spirals and linked concentric circles are very different in terms of the way they would be described. The false spirals would have to be described as a set of separate shapes which link to form a single motif and at that stage problems arise with identifying the limits of what constitutes a single motif. If more types of motif were analysed, it seems probable that the range of variables would have to increase and some variables included for the purposes of analysing spirals would be irrelevant to some other motif types.

6.3 Comparison with mathematical spirals Studies of decorative spiral motifs have tended to use the terminology of mathematical spirals to either describe decorative spiral shape or as the basis for classification (Zhushchikhovskaya & Danilova 2008; Clegg 2010, p.51; Takaki & Ueda 2007). Andrey and Galli's (2004) study of construction methods for volutes from Renaissance architectural treatises found that matching decorative spirals to mathematical spiral equations was either impossible or produced ambiguous results with equally good matches to more than one mathematical type arising. This, and the findings from two of the classification studies that decorative spirals fell into hybrid classes between mathematical types (Zhushchikhovskaya & Danilova 2008; Takaki & Ueda 2007), should mitigate against the use of mathematical spiral types as class descriptors. There are, however, reasons relating to the production method and source of inspiration for spiral motifs that suggest that conformity to Archimedes' or logarithmic spirals might be expected in decorative spirals. Visual assessment of

212

decorative spirals in the early stages of this research suggested that some spirals appeared to show good agreement to both these forms and also to hyperbolic spirals but that others could be assigned to more than one mathematical class. On the strength of these observations, conformity to mathematical spirals was tested, at a less precise level than that used by Andrey and Galli, by measuring correlation to the density spectra of mathematical template spirals to see if general trends towards particular types of mathematical spiral could be said to exist. A strong trend towards a logarithmic or hyperbolic spiral form was identified in the decorative spirals and correlation of the spirals to the mathematical type groups was high. There may be a case for making reference to logarithmic spiral forms in studying associations with real natural spirals which are known to have a logarithmic form but, ideally a sample would be taken of spiral shapes from the natural forms themselves as a more accurate description. However, as a more general system it was felt that reference to mathematical spiral forms for description or classification was not ideal because it imposes an arbitrary external class system to which the spirals with lower correlation scores showed a poor visual match and encourages assumptions about shape and precision of measurement which may not be appropriate for decorative spirals. Assigning spirals to mathematical classes as the sole means of classification would also miss the stylistic data which, by its non-random distribution across all four case studies has been shown to be an important set of data to consider in the assessment of spirals as decorative constructions.

6.4 Results of case studies The intention at the start of this research was to use spiral motifs to differentiate between cultures and to study cultural exchange. From this initial approach, the results of measuring the properties of geometric motifs have shown little potential to use these to characterise culture groups. Cluster groupings within each of the cultural case studies showed low differentiation in most cases. The Shang/Western Zhou and Egyptian studies had small outlying classes which were visually distinct from the rest of the sample but these contained few spirals with the other classes showing little or no separation in discriminant analysis. The Jōmon study showed high levels of continuity of variation across all identified classes. Only the Cypriot case study had a separated class of reasonable size in the

213

Class 3 guided painted spirals which also stood out as the most distinctive group in the comparison between manufacturing techniques across the four case studies. Although, to a large extent forming into single groups, a continuum of change in form, rotational length and, in the Shang/Western Zhou study, curvilinearity was detected across the spirals when ranked in their clustering order suggesting that in cases where motif variation were stronger, viewing motifs as displaying graduated change in shape rather than assigning class groups would be the most appropriate way of viewing them. Cross-cultural analyses of the four case studies using discriminant analysis of the full range of variables also showed a very high degree of similarity. Although the Cypriot and Egyptian studies had extreme values which separated from the main body of data, there was no basis for separating any of the cultures. Similarly, analysis of distribution according to production technique showed no clear separations apart from the Cypriot guided painted spirals. In Chapter Two, the difficulties of identifying which aesthetic variables are important in cross cultural studies when perceptions are culturally varied and identifying the interaction of grouping variables with each other were discussed. It is inevitable when considering two very separated pairs of cultures viewed from the perspective of a third culture that the chosen variables for analysis will not reflect the full range of factors governing the production of spiral motifs and that, at best, only some of these variables will have been consciously considered in production. In favour of the method proposed in this research, it can be argued that, for the purposes of cross-cultural comparison, the ability to apply a consistent assessment of similarities and differences is important even though it does not necessarily represent the conscious design processes of those constructing the decoration. Taking a single motif in isolation allows variables to be analysed which would be too complex to address when considering the full range of decorative forms found in a culture. Where these variables can be shown to have a significantly non-random distribution of expression, there begins to be a basis for suggesting coherent preferences, which must be assumed to reflect either deliberate or unconscious perceptual choice or technologically driven constraints on decoration. The findings of

214

the four case studies highlight most strongly the homogeneity of the dataset as a whole. In the analysis of separate variables this was expressed in the commonality of form tending towards a high expansion rate and in low rotational length and high curvilinearity amongst the case studies. The perceptual values of prehistoric and early historic cultures must be largely or completely assessed through the material record. If an expanded data set and analyses of a broader range of motifs could be used to consistently identify common expressions of form and style variables then these might be used with caution to shed light on the perceptual judgements or preferences of the makers of the objects.

6.5 Spiral source Two possible sources of inspiration for decorative spirals were considered in the research questions of this thesis arising from naturally occurring forms. In the course of running the case study analyses, the distinctive nature of mechanically assisted methods of constructing spirals in the Cypriot study became apparent so variation in production method was considered as another potential influence on the shape of decorative spirals. 6.5.1

Cortically generated spiral forms

Spirals are amongst the geometric motifs identified as visual effects (endogenous visual phenomena – EVP) generated by the abnormal functioning of the optical cortex. Lewis-Williams and Dowson (1988) amongst others have suggested that these appear in art work in cultures which practice religious trance states. This has been attested to by informants in ethnographic studies (e.g. Reichel-Dolmatoff 1978). On this account, the third research question sought to establish whether a 'typical' EVP spiral shape could be identified and, if so, whether there was evidence for this shape influencing decorative spirals. The type spirals were obtained from external representation of internal visual phenomena produced by those who had experienced them and from theoretical models of visual phenomena so their accuracy must be assessed cautiously. This is especially so given the limitations on the available data in the number of subjects whose experiences were being measured, the homogeneity of the culture group of these subjects and in the range of causatory factors of these stimuli. The theoretically modelled spirals which made up 28% of the data for generating the type spirals

215

showed a low similarity with the drawn spirals so may not be representative of the spirals experienced. Most studies of the artistic expression of EVP have linked them to trance states (e.g. Lewis-Williams & Dowson 1988; Bednarik 1986; Reichel-Dolmatoff 1978; Dronfield 1994a and b). Although there are indications that states of abnormal consciousness could be associated with religious practices in the case study cultures, in no instance is there sufficient evidence to suggest that EVP were a relevant part of religious practices in any of the cultures studied. The Jōmon case study comes closest to a possible link in the use of spiral decoration in association with figures interpreted as trance dancers (Takashi 2009, pp.136, 140). Any suggestion that EVP were represented in decoration would therefore depend on a strong correlation with EVP spiral test forms and a differentiation from other likely sources and would need to be supported by contextual evidence of other geometric EVP structures featuring in decoration. Three classes of EVP spirals created by clustering were used as the basis for type spirals but there is a high degree of variance in all three groups so no clearly defined types emerged. Type 3 EVP spirals were the group with the largest number of similar spirals from the decorative case studies and significant numbers of the Cypriot, Shang, and Jōmon spirals correlated with this type. The low average rotational length of the decorative spirals could also fit with the typical characteristics of EVP spirals. These results suggest that an EVP influence on the decorative spirals cannot be ruled out but the variance within the type classes and the possible alternative explanations for the shape of the correlating Cypriot spirals in a technological source; of one of the Shang Dynasty spirals in a representational source; or in the natural trend of human hand movement means that there is not sufficient basis to identify an EVP influence with certainty. Future work could consider expanding the EVP data set to see if clearer type groups could be identified. A larger sample might allow the division of the generated type spirals according to the generating stimulus. As it stands, characteristic forms arising from individual stimuli may have been lost in the merging of types of stimulus. Culturally the time gap between the studied decorative spirals and the available data for creation of EVP type spirals cannot be overcome but, as Bahn (1988) noted it would

216

be useful to broaden the contemporary geographical scope of studies of EVP as one test of whether inter-cultural consistency pertains. However, it is likely that spirals generated by the human visual system will show the same logarithmic tendency that is characteristic of many natural real spirals and the problem of a potential overlap may mean that the analysis of a single motif type is insufficient to identify the influence of EVP. 6.5.2

Naturally occurring real spirals as sources of inspiration

Although some interpretations of decorative spirals which represent real natural forms are speculative, in most cases, context or supplementary decoration of the spiral could be used to identify a representational source. These spirals were used as a basis for assessing whether cross cultural consistency could be seen in the shape of representational spirals and whether influence from different natural forms could be separated. As a whole, the representational spiral group showed a very high intrasample correlation suggesting that there is commonality in the manner of representing decorative spirals across the four studied cultures. These spirals tended towards an increasing expansion rate which would be expected from the characteristics of natural spiral forms. Although two of the snake representations were separated from the main cluster, it was not possible to segregate spirals across cultures on the basis of their specific representation. In many ways it is unsurprising that spirals intended to represent natural forms should resemble them in their shape characteristics but there is some evidence in high average curvilinearity values and the prevalence of high expansion rate forms that an influence of EVP or real natural spiral forms can be identified extending into apparently non-representational decorative spirals.

6.6 Potential for expanding the application of the analysis method Analytical methods based on quantitative description of a set of attributes of a motif have the benefit of largely being able to develop organically to encompass new forms as they are encountered without revision of the stored description of previously analysed motifs whereas descriptions which assign named types to motifs may force re-evaluation of the basic description in the light of new data being received. The quantification of motif shape permits the descriptions to be stored as a dataset which could be added to in future work to increase sample sizes and therefore the robustness of analysis and the range of comparative material available.

217

It would be interesting to expand the scope of the chosen case studies to include more recent cultural developments. Spiral decoration continued to form a significant part of the decorative canon from the Jōmon into the subsequent Yayoi period; in the later Cypro-Archaic and Classical Aegean periods and into the Eastern Zhou period following the Western Zhou so there would be scope within three of the case studies to continue the analysis forwards and to consider, particularly how the introduction of new technologies and mathematical knowledge may have affected the more recent spiral decorative styles. Low numbers of spirals in some of the clustering groups within the case studies examined limited the power of analysis in some cases. Expansion of the data sets might strengthen results and aid in establishing whether some outlying spirals which were consistently clustered on their own were genuine outliers or whether a larger dataset could associate them as part of a class.

6.7 Conclusion This research covered a small sample of objects from four cultures working on a single motif as a test of a principle. The proposed method was able to operate across the breadth of variation in spiral motifs in the four studies and to segregate decorative spirals into visually coherent orders reflecting a graduation of change according to the form and certain style properties of the motifs. In the application of the analysis and classification system, the range of overall shape expressions found in spiral decoration indicates that certain expressions of the measured variables are preferentially expressed across the cultures. Evidence for the presence of EVP spirals could not be identified but the range of decoration found across all four cultures was consistent with the influence of natural real spirals and it was shown that some technological variation has characteristic shape expressions which can be identified from this approach. This suggests that there is similar potential value in the analysis of geometric motifs to that found in pattern structure to assess perceptual preferences and change and consistency in decorative expression.

218

References ADLER, C. 1893. The Shofar – its use and origin. Proceedings of the National Museum. 16:936:287-301 ALLAN, S. 2007. Erlitou and the formation of Chinese civilization: toward a new paradigm. The Journal of Asian Studies. 66:2:461-496 AMODEO, A., M. OLIVERIO, P. VERSACCI and B. MARINO. 2012. Spiral shapes in heart and shells: when form and function do matter. European Journal of Cardio-Thoracic Surgery. 41:473-475. ANDREY, D and M. GALLI. 2004. Geometric Methods of the 1500s for Laying Out the Ionic Volute. Nexus Network Journal [online] 6:2:31-48, [accessed 5.1.2010]. Available from: http://www.nexusjournal.com/ ARTISANS OF THE ANVIL. 2005. Me scrolling a small bar [online video], [accessed 16.5.2010]. Available from: http://www.artisansoftheanvil.com/ BAGLEY, R.W. 1996. Il Neolitico. in M.Pirazzoli-t'Serstevens (ed.) La Cina. Turin: Unione Tipografico-Editrice Torinese BAHN, P. 1988. Comment on: J.D.Lewis-Williams and T.A.Dowson. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201245 BAILEY, D.M. 1969. The village priest’s tomb at Aradippou in Cyprus. The British Museum Quarterly. 34:1/2:36-58 BANAKH, T., O. VERBITSKY and Y. VOROBETS. 2010. Fermat’s Spiral and the Line Between Yin and Yang. American Mathematical Monthly [online] 117:9:786-800, [accessed 10/12/2010] pp.786-800. Available from: doi:10.4169/000298910X521652. BANDO, T. and D. HIRABAYASHI. 2004. Attractiveness of Spiral Pattern in Human Vision. Forma [online] 19:223-232, [accessed 12.5.2012]. Available from: http://www.scipress.org/journals/forma/ BAR, M. and NETA, M. 2006. Humans prefer curved visual objects. Psychological Science. 17:8:645-648

219 BARBER, E. J. W. 1998. Aegean ornaments and design in Egypt. in E.H.Cline and D.Harris-Cline (eds.) Aegeum 18: The Aegean and the Orient in the second millennium. Proceedings of the 50th Anniversary Symposium, University of Cincinnati, 18-20th April 1997. BECK, D.M. and S.E.PALMER. 2002. Top-down influences on perceptual grouping. Journal of Experimental Psychology: Human Perception and Performance. 28:5:10711084. BECK, J. 1966a. Effect of orientation and of shape similarity on perceptual grouping. Perception and Psychophysics. 1:300-302 ______. 1966b. Contrast and assimilation in lightness judgements. Perception and Psychophysics. 1:342-344 ______. 1975. The relation between similarity grouping and perceptual constancy. The American Journal of Psychology. 88:3:397-409 BECKER, C. 2005. Subjective visual experiences of colour and form induced by temporally modulated light. Phd thesis [online], [accessed 7.6.2013]. Available from: http://edoc.ub.uni-muenchen.de/4179/. Munich: Ludwig-Maximilians-Universität. BECKER, S.I. 2011. Determinants of dwell time in visual search: similarity or perceptual difficulty? PLoS ONE [online] 6:3:1-5, [accessed 28.5.2013]. Available from: doi:10.1371/journal.pone.0017740 BEDNARIK, R.G. 1986. Parietal finger markings in Europe and Australia. Rock Art Research. 3:1:30-61 BEN-AV, M.B. and D.SAGI. 1995. Perceptual grouping by similarity and proximity: experimental results can be predicted by intensity auto-correlations. Vision Research. 35:6:853-866 BEN-TOR, D. 1997. The Relations Between Egypt and Palestine in the Middle Kingdom as Reflected by Contemporary Canaanite Scarabs. Israel Exploration Journal. 47:3/4:162-189

220 __________. 2007. Scarabs, chronology and interconnections: Egypt and Palestine in the Second Intermediate Period. Fribourg: Academic Press Fribourg. BERLYNE, D.E. 1975. Extension to Indian subjects of a study of exploratory and verbal responses to visual patterns. Journal of Cross-Cultural Psychology. 6:316-330 BIETAK, M. 1999. Tell el Yahudiya. in K.A.Bard (ed.) Encyclopedia of the Archaeology of Ancient Egypt. London: Routledge, pp.791-2 BISHOP, R.L., V. CANOUTS, S. P. DE ATLEY, A. QÖYAWAYMA and C.W. AIKINS. 1988. The formation of ceramic analytical groups: Hopi pottery production and exchange, a.c.1300-1600. Journal of Field Archaeology. 15:3:317-337. BLÖSCHL, G. and R.GRAYSON. 2000 Spatial observations and interpolation. in R.Grayson and G.Blöschl (eds.) Spatial Patterns in Catchment Hydrology: Observations and Modelling. Cambridge: Cambridge University Press, pp.17-50 BOSELIE, F. 1984. The aesthetic attractivity of the golden section. Psychological Research [online] 45:4:367-375, [accessed 16/11/09]. Available from http://www.springerlink.com/content/r7mqt7573022j435/ BRADLEY, R. 1997. Rock Art and the Prehistory of Atlantic Europe: Signing the Land. London: Routledge BRAINERD, G.W. 1942. Symmetry in primitive conventional design. American Antiquity. 8:2:164-66. BRANDMÜLLER, J., B.HROUDA and A.V.WICKEDE. 1986. Symmetry in archaeology. Computing and Mathematics with Applications. 12B:3/4:783-787 BRESSLOFF, P.C., J.D.COWAN, M.GOLUBITSKY, P.J.THOMAS and M.C.WIENER. 2001. Geometric visual hallucinations, Euclidean symmetry and the functional architecture of striate cortex. Philosophical Transactions of the Royal Society, London. B. 356:299-330 BRITISH MUSEUM. 2013. Flame and water pots: prehistoric ceramic art from Japan. Exhibition, 4th October 2012-20th January 2013. BRUMFIEL, E. 2007. Solar disks and solar cycles: spindle whorls and the dawn of solar art in Postclassic Mexico. Treballs d'Arqueologia 13:91-113

221 BRUNTON, G. 1923. Qau and Badari I. British School of Archaeology in Egypt and Egyptian Research Account 29th Year, London: BSAE BUTTERWORTH, A.T. 1967. The psychotomic effect: a discussion of its unique nature and character. Existential Psychiatry. Winter, pp.489-498 CADOGAN, G. 1989. Maroni and the monuments. in E.Peltenburg (ed.) Early society in Cyprus. Edinburgh: Edinburgh University Press, pp.43-51 CASRNIR, F.L. 1999. Foundations for the study of intercultural communication based on a third-culture building model. International Journal of Intercultural Relations. 23:1:91-116 CHANG, K.-C. 1980. Shang Civilization. London: Yale University Press CHILDS, W.A.P. 1997. The Iron Age Kingdom of Marion. Bulletin of the American Schools of Oriental Research. The City-Kingdoms of early Iron Age Cyprus in their eastern Mediterranean context. 308:37-48 CHUA, H.F., J.E.BOLAND and R.E.NISBETT. 2005. Cultural variation in eye movements during scene perception. Proceedings of the National Academy of Sciences. 102:35:12629-12633. CHUI-MEI, H. 1984. Pottery in South China: Xijiang and Upper Red River Basins. World Archaeology. 15:3:294-325 CLEGG, J. 2010. Spirals at Sturt's Meadows. in C. Fidalgo, L. Oosterbeek, J. Kolber, J. Clegg and A. Distel (eds.) Session C81: Spirals and Circular Forms: the Most Common Rock Art in the World? Proceedings of the XV UISPP World Congress, Lisbon, 4-9 September 2006 BAR Int S2108, Oxford: Archaeopress. pp. 49-54 CLEOPHAS, T.J. 2012. Machine learning in pharmaceutical research: data clustering, why so and how so. Journal of Pharmaceutical Science and Research. 4:11:1964-1969. COLDSTREAM, J.N. 1989. Status symbols in the 11th century BC. in E.Peltenburg (ed.) Early society in Cyprus. Edinburgh: Edinburgh University Press, pp. 325-335 . 2003 (2nd ed.). Geometric Greece: 90-700BC. London: Routledge

222 COLLARD, D. 2011. Altered states of consciousness and ritual in Late Bronze Age Cyprus. [online], [accessed 4.7.2013]. Available from: http://etheses.nottingham.ac.uk/2322/. PhD thesis, University of Nottingham. COOK, T.A. 1914 reprinted 1979. The Curves of Life. New York: Dover Publications COUTIL, L. 1916. L'ornementation spiraliforme Périodes Paléolithique et Néolithique. Ages du Bronze et du Fer. Bulletin de la Société Préhistorique française. 13:7:385-484. CROWLEY, J. 1989. The Aegean and the East: an investigation into the transference of artistic motifs between the Aegean, Egypt and the Near East in the Bronze Age. Studies in Mediterranean Archaeology Pocketbook 51. Jonsered: Åströms DAUGHERTY, P. 2004. Cognitive structures and body metaphors in the aesthetic and everyday discourse of Turkish-Yörük weavers. in D.K.Washburn and D.W.Crowe (eds.) Symmetry comes of age: the role of pattern in culture. Seattle: University of Washington Press DAVID, R. 1998. Handbook to life in ancient Egypt. Oxford: Oxford University Press _______. 2002. Religion and Magic in Ancient Egypt. London: Penguin Books DELLUC, B. and DELLUC, G. 1988. Comment on: J.D.Lewis-Williams and T.A.Dowson. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201-245 DOCZI, G. 2005. The Power of Limits: Proportional Harmonies in Nature, Art and Architecture. Boston: Shambhala Publications DOHERTY, M.J., H.TSUJI and W.A.PHILLIPS. 2008. The context sensitivity of visual size perception varies across cultures. Perception. 37:1426-1433 DOUGLASS, A.A. and O.LINDAUER. 1988. Hierarchical and nonhierarchical approaches to ceramic design analysis: a response to Jernigan. American Antiquity. 53:3:620-626 DRONFIELD, J.1996a. The vision thing: diagnosis of endogenous derivation in abstract arts. Current Anthropology. 37:2:373-391

223 ___________1996b. Entering alternative realities: cognition, art and architecture in Irish passage-tombs. Cambridge Journal of Archaeology. 6:1:37-72 EITELJORG II, H. 1980. The fast wheel, the Multiple-Brush Compass and Athens as home of the Protogeometric style. American Journal of Archaeology. 84:4:445-452 ERMENTROUT, G.B. and J.D.COWAN. 1979. A Mathematical Theory of Visual Hallucination Patterns. Biological Cybernetics. 34:137-150. EVANS, A. 1935. The Palace of Minos at Knossos. Vol. IV, Part 1, London: Macmillan and Co. EYRE, C. 1997. The boundaries of knowledge. in D.P.Silverman (ed.) Ancient Egypt. London: Duncan Baird Publishers. pp. 90-103 FAUBERT, J. 2002. Visual perception and aging. Canadian Journal of Experimental Psychology. 56:3:164-176 FEINBERG, I. 1970. Hallucinations, dreaming and REM sleep. in W.Keup (ed.) Origin and Mechanisms of Hallucinations. Proceedings of the 14th Annual Meeting of the Eastern Psychiatric Research Association, New York City, 14-15 Nov. 1969. New York: Springer Science and Business Media. pp.125-132 FILATOVA, I.V. 2008. Ornamental Traditions of the Lower Amur Neolithic. Archaeology, Ethnology and Anthropology of Eurasia. 34:2:88-95 FISCHER, 1961. Art styles as cultural cognitive maps. American Anthropologist. 63:7993 FOSTER McCARTER, S. 2007. Neolithic. New York: Routledge FRINGS, M. 2002. The Golden Section in Architectural Theory. Nexus Network Journal [online] 4:1:9-32, [accessed 16/8/2013]. Available from: http://www.nexusjournal.com/volume-4.html FROLOV, B.A. 1983. On Southern San Rock Art. Current Anthropology. 24:2:237-239 FUCHIGAMI, K. 2001. Analysis of the spiral pattern karakusa. Journal for Geometry and Graphics. 5:1:35-43.

224 FUKUGAWA, H. and ROTHMAN, T. 2008. Sacred mathematics: Japanese temple geometry. Princeton: Princeton University Press. GIBBS, L.L. and R.ANDERSON. 1977. Effect of weight and sex on perception of line length. Perceptual and Motor Skills. 44:783-786 GIMBUTAS, M. 1989 (rpt 2001). The Language of the Goddess. London: Thames and Hudson GOLDSTONE, R.L. 1995. Mainstream and avant-garde similarity. Psychologica Belgica. [online] 35:145-165, [accessed 24.8.2013]. Available from: http://cognitrn.psych.indiana.edu/papers.html GOODYEAR, W.H. 1893. Are Conventional Patterns Spontaneously Generated? Architectural Record. 2:4:391-418 GRABAR, O. 2006. Trade with the East and the Influence of Islamic Art on the "Luxury Arts" in the West. Constructing the Study of Islamic Art Vol.2, Islamic Visual Culture, 1100-1800. Hampshire : Ashgate Publishing. GRAJETZKI, W. 2006. The Middle Kingdom of Ancient Egypt. London: Gerald Duckworth & Co. GRAZIANO, M.S., R.A.ANDERSEN and R.J.SNOWDEN. 1994. Tuning of MST neurons to spiral motions. Journal of Neuroscience. 14:54-67 GREENE, H.W. 1973. Defensive tail display by snakes and amphisbaenians. Journal of Herpetology. 7:3:143-161 GRENFELL, A. 1915. The scarab collection of Queen's College, Oxford. The Journal of the Egyptian Society. 2:4:217-228 GRÜNBAUM, B. 2004. Periodic Ornamentation of the Fabric Plane: Lessons from Peruvian Fabrics. in D.K.Washburn and D.W.Crowe (eds) Symmetry Comes of Age: The Role of Pattern in Culture. Seattle: University of Washington Press GUTHERIE, A.H., P.W.SEELY, L.A.BEACHAM, R.A.SCHUCHARD, W.A.DE L'AUNE and A.BACON MOORE. 2010. Age effects on visual-perceptual processing and confrontation

225 naming. Aging, Neuropsychology and Cognition: a Journal on Normal and Dysfunctional Development. 17:2:160-190. HABU, J. 2004. Ancient Jomon of Japan. Cambridge: Cambridge University Press. HADJICOSTI, M. 1997. The kingdom of Idalion in the light of new evidence. Bulletin of the American Schools of Oriental Research, The City-Kingdoms of Early Iron Age Cyprus in their Eastern Mediterranean Context. 308:49-63 HADJISAVVAS, S. 1989. A Late Cypriot community at Alassa. in E.Peltenburg (ed.) Early society in Cyprus. Edinburgh: Edinburgh University Press, pp. 32-42 HAMBIDGE, J. 1920. Dynamic Symmetry: The Greek Vase. Connecticut: Yale University Press HANN, M. A. 1992. Symmetry in regular repeating patterns: case studies from various cultural settings. Journal of the Textile Institute. 83:4:579-590. HEL-OR, PELEG and AVNIR 1991. Characterization of right-handed and left-handed objects. Computer Vision Image Processing (CVIP): Image Understanding. 53:297-302 HENSHILWOOD, C., F. D’ERRICO, M.VANHAEREN, K.VAN NIEKERK and Z.JACOBS. 2004. Middle Stone Age Shell Beads from South Africa. Science. 304:5669:404 HERNER, E. (trans. A. CROZIER) 1989. Spiral Decoration in Early Bronze Age Scandinavia: A Technical and Qualitative Analysis and Study of Production. British Archaeological Reports: International Series 552. HIRSCHFELD, L.A. 1977. Cuna aesthetics: a quantitative analysis. Ethnology. 16:2:147166 HODGSON, D. 2000. Shamanism, phosphenes and early art: an alternative synthesis. Current Anthropology. 41:5:866-873 HOGARTH, W. 1753 (1997 edn.). The Analysis of Beauty. Edited and annotated by R. Paulson. London: Yale University Press HOROWITZ, M.J. 1964. The imagery of visual hallucinations. Journal of Nervous and Mental Disease. 138:6:513-523

226 HULL, D. 2003. Celtic and Anglo-Saxon Art: Geometric Aspects. Liverpool: Liverpool University Press IACOVOU, M. 1989. Society and settlements in Late Cypriot III. in E.Peltenburg (ed.) Early society in Cyprus. Edinburgh: Edinburgh University Press, pp.52-59 IAKOVLEVA, L. 2009. L’art mézinien en Europe orientale dans son contexte chronologique, culturel et spirituel. L’Anthropologie [online] 113:(5:1):691-752, [accessed 20/4/10]. Available from: http://www.sciencedirect.com/ (doi:10.1016/j.anthro.2009.09.015). JABLAN, S.V. 2002. Symmetry, Ornament and Modularity. London: World Scientific Publishing JERNIGAN, E.W. 1986. A non-hierarchical approach to ceramic decoration analysis: a southwestern example. American Antiquity. [online] January, 51:1:3-20, [accessed: 10/12/2010]. Available from: http://www.jstor.org/stable/280390 JOLICOEUR, P. and G.K.HUMPHREY. 1998. Perception of rotated two-dimensional and three-dimensional objects and visual shapes. in V.Walsh and J.Kulikowsky (eds.) Perceptual constancy: why things look as they do. Cambridge: Cambridge University Press. pp.69-123 JONES, O. 1856 (rpt. 1986). The Grammar of Ornament. London: Studio Editions KANER, S. 2009. The power of dogu: ceramic figures from ancient Japan. London: British Museum Press KAPLAN, M and COHEN, E. 2003. Computer generated Celtic design. in P.Christensen, and D.Cohen-Or (eds.) Eurographics Symposium on Rendering. Leuven, Belgium June 25-27. Leuven: Computer Graphics Group KAPPRAFF, J. 1992. The Spiral in Nature, Myth and Mathematics. in I.Hargittai and C.A.Pickover (eds.) Spiral Symmetry. Singapore: World Scientific Publishing Company. pp. 1-46 KARAGEORGHIS, V. 1969. The ancient civilisation of Cyprus. New York: Cowles Education Corp.

227 KAYAERT, G., I. BIEDERMAN and R.VOGELS. 2005. Representation of Regular and Irregular Shapes in Macaque Inferotemporal Cortex. Cerebral Cortex. 15:1308-1321. KEELER, M.H. 1970. Klüver's mechanisms of hallucinations as illustrated by the paintings of Max Ernst. in W.Keup (ed.) Origin and Mechanisms of Hallucinations. Proceedings of the 14th Annual Meeting of the Eastern Psychiatric Research Association, New York City, 14-15 Nov. 1969. New York: Springer Science and Business Media. pp.205-208 KEIGHTLEY, D.N. 1983. The late Shang state: when, where and what? in D.N.Keightley (ed.) The Origins of Chinese Civilization. Berkeley: University of California Press, pp.523-564 . 2000. A Secret of Ancient Geometry. in C.Gorini (ed.) Geometry at Work. Mathematics Association of American Geometry MAA Notes No. 53 KELLMAN, P.J. and T.F.SHIPLEY. 1991. A theory of visual interpolation on object perception. Cognitive Psychology. 23:141-221. KELLOGG, R., M.KNOLL and J.KUGLER. 1965. Form-similarity between phosphenes of adults and pre-school children's scribblings. Nature. 208:1129-1130 KEMPTON, W. 1981. The folk classification of ceramics: a study of cognitive prototypes. London: Academic Press KIDDER, J.E. 1964. Early Japanese Art. London: Thames and Hudson KIDDER, J.E. and T. ESAKA. 1968. Prehistoric Japanese arts: Jomon pottery. Tokyo: Kodansha International KIELY, T. n.d. Ancient Cyprus in the British Museum. [online], [accessed17/7/2013]. Available from: http://www.britishmuseum.org/research/publications/online_research_catalogues/an cient_cyprus_british_museum.aspx KLASSON, K.T. 2008. Construction of spline functions in spreadsheets to smooth experimental data. Advances in Engineering Software. 39:422-429.

228 . 2011. Addendum to "Construction of spline functions in spreadsheets to smooth experimental data". Advances in Engineering Software. 42:621. KLING, B. 1991. A terminology for the matte-painted, wheelmade pottery of Late Cypriot IIC-IIIA. In J.A.Barlow, D.L.Bolger and B.Kling (eds.) Cypriot ceramics: reading the prehistoric record. Philadelphia: The University Museum of Archaeology and Anthropology, University of Pennsylvania KLÜVER, H. 1967. Mescal and the mechanisms of hallucination. Chicago: University of Chicago Press KNOLL, M., J.KUGLER, O.HÖFER and S.D.LAWDER. 1963. Effects of chemical stimulation of electrically induced phosphenes on their bandwidth, shape, number and intensity. Confinia Neurologica. 23:201-26. KNOLL, M. and J.KUGLER. 1959. Subjective light pattern spectroscopy in the encephalographic frequency range. Nature. 184:1823-1824 KUNST-WILSON, W.R. and R.B.ZAJONC. 1980. Affective discrimination of stimuli that cannot be recognized. Science. New Series 207:4430:557-558 LA BARRE, W. 1975. Anthropological perspectives on hallucination and hallucinogens. in R.K.Siegel and L.J.West (eds.) Hallucinations: behaviour, experience and theory. London: John Wiley and Sons. LAYTON, R. 1988. Comment on: J.D.Lewis-Williams and T.A.Dowson. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201245. LAZARUS, V. and PAUCHARD, L. 2011. From craquelures to spiral crack patterns: influence of layer thickness on the crack patterns induced by dessication. Soft Matter. 7:2552-2559. LEE, A.J. 1987. Islamic star patterns. Muqarnas IV: An Annual on Islamic Art and Architecture [online], [accessed 30.10.2009]. Available from: http://archnet.org/library/documents/one-document.jsp?document_id=3591

229 LEMOS, I.S. 2002. The Protogeometric Aegean: the archaeology of the late eleventh to tenth centuries BC. Oxford University Press LEROI-GOURHAN, A. 1968. The Art of Prehistoric Man in Western Europe. London: Thames and Hudson LEWIS-WILLIAMS, J.D. and T.A.DOWSON. 1988. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201-245 _________________________________. 1993. On vision and power in the Neolithic: evidence from the decorated monuments. Current Anthropology. 34:1:55-65 LI, H.-L. 1974. An Archaeological and historical account of cannabis in China. Economic Botany. 28:437-448 LIDWELL, W., K. HOLDEN and J.BUTLER. 2003. Universal Principles of Design. Gloucester, Massachusetts: Rockport Publishers LIU, X., C.B.CARROLL, S.-Y. WANG, J.ZAJICEK and P.G.BAIN. 2005. Quantifying druginduced dyskinesias in the arms using spiral-drawing tasks. Journal of Neuroscience Methods. 144:47-52. LOEHR, M. 1953. The bronze styles of the Anyang Period (1300-1028 B.C.). Archives of the Chinese Art Society of America. 7:42-53. _________. 1968. Ritual vessels of Bronze Age China. New York: Asia House LONG, A.C.Jr., J.A.LANDAY, L.A.ROWE and J.MICHIELS. 2000. Visual Similarity of Pen Gestures. Proceedings of ACM CHI 2000 Human Factors in Computing Systems Conference. The Hague, The Netherlands, April 1-6 2000, pp.360-367. LU, T.L.-D. 2005. The origin and dispersal of agriculture and human diaspora in East Asia. in L.Sagart, R.M.Blench and A.Sanchez-Mazas (eds.) The Peopling of East Asia: Putting Together Archaeology, Linguistics and Genetics. Abingdon: RoutledgeCurzon. pp.51-62. LUMHOLTZ, C. 1900. Symbolism of the Huichol Indians, Part 1. Memoirs of the American Museum of Natural History. Vol.3 (Anthropology Vol.2), pp.1-228

230 LYNCH, J.W. and R.J.HATHAWAY. 1993. Projectile point morphology: a note on the golden section. Journal of Quantitative Anthropology. 4:211-220. MACKENZIE, D.A. 1918. Myths of Crete and Pre-Hellenic Europe. London: Gresham Publishing MACLAY, W.S. and E.GUTTMANN. 1941 Mescaline hallucinations in artists. Archives of Neurology and Psychiatry. 45:1:130-137. MAMANÍ, J. L. and A. A. FERNÁNDEZ DISTEL 2010. Espirales en Humahuaca y en el noroeste de Argentina (Sudamérica) in C. Fidalgo, L. Oosterbeek, J. Kolber, J. Clegg and A. Distel (eds.) Session C81: Spirals and Circular Forms: the Most Common Rock Art in the World? Proceedings of the XV UISPP World Congress, Lisbon, 4-9 September 2006 BAR Int S2108, Oxford: Archaeopress. pp. 39-47. MANFORD, M. and ANDERMANN, F. 1998. Complex visual hallucinations: clinical and neurobiological insights. Brain. 121:1819-1840 MARINATOS, SP. 1951. 'Numerous years of joyful life' from Mycenae. The Annual of the British School at Athens. Papers Presented to Professor Alan Wace to Commemorate Fifty Years of Work in Archaeology. 46:102-116 MARTINDALE, C. 1988. Comment on: J.D.Lewis-Williams and T.A.Dowson. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201-245 MASAYUKI, H. 2009. Catalogue entry 53 – Clay model of roll shell. in S.Kaner (ed.) The power of dogu: ceramic figures from ancient Japan. London: British Museum Press MATSUI, A. 2001. Jomon. in P.N.Peregrine and M.Ember (eds.) Encyclopedia of Prehistory, Volume 3: East Asia and Oceania. New York: Springer Science and Business Media. MCCONNELL, R., R.KWOK, J.C.CURLANDER, W.KOBER and S.S.PANG. 1991. ψ-S Correlation and Dynamic Time Warping: Two Methods for Tracking Ice Floes in SAR Images. IEEE Transactions on Geoscience and Remote Sensing [online] 29:6:1004-1012, [accessed 15.8.2011]. Available from: ieeexplore.ieee.org/iel1/36/3161/00101377.pdf

231 McMANUS, C. 2002. Right Hand, Left Hand: The Origins of Asymmetry in Brains, Bodies, Atoms and Cultures. London: Weidenfeld and Nicholson MEKO, D. 2013. Spectral analysis: smoothed periodogram method. [online], [accessed 22.8.2013]. Available from: http://www.ltrr.arizona.edu/~dmeko/notes_6.pdf MIRIMANOV, V. 1988. Comment on: J.D.Lewis-Williams and T.A.Dowson. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201-245 MORPHY, H. and PERKINS, M. 2009. The anthropology of art: a reader. London: Wiley. MUHLY, J. 1989. The organisation of the copper industry in Late Bronze Age Cyprus. in E.Peltenburg (ed.) Early society in Cyprus. Edinburgh: Edinburgh University Press, pp.298-314 MURNANE, W.J. 1997. Three kingdoms and thirty-four dynasties. in D.P.Silverman (ed.) Ancient Egypt. London: Duncan Baird Publishers. pp. 20-39. MURRAY, O. 1978 (rpt 2001). Early Greece. Massachusetts: Harvard University Press NEHER, A. 1962. A physiological explanation of unusual behaviour in ceremonies involving drums. Human Biology. 34:2:151-160. NEICH, R. 1993. Painted histories: early Maori figurative painting. Auckland: Auckland University Press NELSON, G. 1936. A faience rhyton from Abydos in the Boston Museum of Fine Arts. American Journal of Archaeology. 40:4:501-506 NICHOLSON, P.T. and R.P.FIRNHABER. 2003. Autohypnotic induction of sleep rhythms generates visions of light with form-constant patterns. in A.Leete and R.P.Firnhaber (eds.) Shamanism in the Interdisciplinary Context. Boca Raton, Florida: BrownWalker Press, pp. 56-83. NICHOLSON, P.T. and E. PELTENBURG. 2000. Egyptian faience. in P.T. Nicholson and I.Shaw (eds.) Ancient Egyptian Materials and Technology. Cambridge: Cambridge University Press.

232 NOMA, S. 1966. The arts of Japan: vol.1 ancient and medieval. Tokyo: Kodansha International OSTER, G. 1970. Phosphenes. Scientific American. February, 222:2:82-87 O’SULLIVAN, M. 1986. Approaches to Passage Tomb Art. The Journal of the Royal Society of Antiquaries of Ireland [online] 116:68-83, [accessed: 10/12/2010]. Available from: http://www.jstor.org/stable/25508907 PALMER, S. and I.ROCK. 1994. Rethinking perceptual organization: the role of uniform connectedness. Psychonomic Bulletin and Review. 1:1:29-55. PALMER, S.E., J.NEFF and D.BECK. 1996. Late influences on perceptual grouping: amodal completion. Psychonomic Bulletin and Review. 3:1:75-80 PANAYIOTOPOULOS, C.P. Elementary visual hallucinations in migraine and epilepsy. Journal of Neurology, Neurosurgery and Psychiatry. 57:1371-1374 PANIS, S., J. DE WINTER, J.VANDEKERCKHOVE and J.WAGEMANS. 2008. Identification of Everyday Objects on the Basis of Fragmented Outline Versions. Perception [online], 37:271-289 [accessed 30.8.2011]. doi:10.1068/p5516 PARUSH, S., C.SHARONI, J.HAHN-MARKOWITZ and N.KATZ. 2000. Perceptual, motor and cognitive performance components of Bedouin children in Israel. Occupational Therapy International. 7:4:216-231. PENNSYLVANIA STATE UNIVERSITY. 2013. STAT 510 – Applied time series analysis. Lesson 12: Spectral analysis. [online], [accessed 22.8.2013]. Available from: http://onlinecourses.science.psu.edu/stat510/?q=node/57 PEREGRINE, P.N. 2001. Cross-cultural comparative approaches in archaeology. Annual Review in Anthropology. 30:1-18. _____________. 2007. Cultural Correlates of Ceramic Styles. Cross-Cultural Research. 41:3:223-235 PETRIE, W.F. 1920 (2nd ed.) Egyptian Decorative Art. London: Methuen and Co.

233 PETRIE, W.M.F., E. MACKAY and G.WAINWRIGHT. 1910. Meydum and Memphis III. British School of Archaeology in Egypt and Egyptian Research Account 16 th Year, London: BSAE PHILLIPPS, W.J. 1948. Maori Spirals. The Journal of the Polynesian Society. 57:1:30-40. PICKOVER, C.A. 2009. The Math Book: from Pythagoras to the 57th dimension, 250 milestones in the history of mathematics. New York: Sterling Publishing Company PINNA, B. 2005. The role of the Gestalt principle of similarity in the watercolour illusion. Spatial Vision. 18:2:185-207. PITTARD, N., M.EWING and C.JEVONS. 2007. Aesthetic theory and logo design: examining consumer response to proportion across cultures. International Marketing Review. 24:4:457-473. POPHAM, M. 1985. Review of: A Cypriot cemetery Palaepaphos-Skales, an Iron Age cemetery in Cyprus by Vassos Karageorghis. The Classical Review. 35:2:337-339 PRIORESCHI, P. 1995. A history of medicine: vol. 1 primitive and ancient medicine. Omaha: Horatius Press QUINLAN, P.T. and R.N.WILTON. 1998. Grouping by proximity or similarity? Competition between the Gestalt principles in vision. Perception. 27:417-430 RAWSON, J. 1987. Chinese bronzes: art and ritual. London: Sainsbury Centre for Visual Art of East Anglia REDFORD, D.B. 1997. Egypt and the world beyond. in D.P.Silverman (ed.) Ancient Egypt. London: Duncan Baird Publishers. pp. 40-57. REICHEL-DOLMATOFF, G. 1978.Drug-induced optical sensations and their relationship to applied art among some Colombian Indians. in V.Greenhalgh and V.Megaw (eds.) Art in society: studies in style, culture and aesthetics. London: Gerald Duckworth and Co. REISNER, G.A. and N.F.WHEELER. 1930. The art of seal carving in Egypt in the Middle Kingdom. Bulletin of the Museum of Fine Arts. 28:167:47-55.

234 REYES, A.T. 1994. Archaic Cyprus: a study of the textual and archaeological evidence. Oxford: Clarendon Press. ROE, P.G. 2004. At Play in the fields of symmetry: design structure and shamanic therapy in the Upper Amazon. in D.K.Washburn and D.W.Crowe (eds.) Symmetry Comes of Age: The Role of Pattern in Culture. Seattle: University of Washington Press SCARRE, F. 2005. The human past: world prehistory and the development of human societies. London: Thames and Hudson SCE WEBSITE. n.d. Swedish Cyprus Expedition 1927-1931. [online], [accessed 25.8.2013]. Available from: swedishcyprusexpedition.se/Om_Svenska_Cypernexpeditionen_1927_-_1931.html SCHAEFER, S.B. 1997. The crossing of the souls: peyote, perception and meaning among the Huichol Indians in S.B.Schaefer and P.T.Furst (eds.) People of the peyote: Huichol Indian history, religion and survival. Albuquerque:UNM Press SCHOTT, G.D. 2007. Exploring the visual hallucinations of migraine aura: the tacit contribution of illustration. Brain. 130:6:1690-1703 SERPELL, R. 1979. How specific are perceptual skills? A cross-cultural study of pattern reproduction. British Journal of Psychology. 70:365-380. SHARP, J. 2002. Spirals and the Golden Section Nexus Network Journal [online] 4:1:5982, [accessed 5/1/10]. Available from: http://www.nexusjournal.com/ SHAW, I. 1997. The Settled World. in D.P.Silverman (ed.) Ancient Egypt. London: Duncan Baird Publishers. pp. 68-79. SHAW, M.C. 1970. Ceiling patterns from the Tomb of Hepzefa. American Journal of Archaeology. 74:1:25-30. SHENNAN, S. 1997. Quantifying Archaeology. Edinburgh: Edinburgh University Press SHOP OUTFITTERS. 2010. Ornamental attachments [online], [accessed 16.5.2010]. Available from: http://shopoutfitters.com/ornamental-attachments/

235 SIEGEL, R.K. and M.E.JARVIK. 1975. Drug-induced hallucinations in animals and Man. in R.K.Siegel and L.J.West (eds.) Hallucinations: behaviour, experience and theory. London: John Wiley and Sons. SIEVEKING, A. 1971. Palaeolithic decorated bone discs. The British Museum Quarterly 35:1:206-229. SMALL, E and B. BROOKES 1984 Coiling of alfalfa pods in relation to resistance against seed chalcids: additional observations. Canadian Journal of Plant Science 64:659-665 SMITH Jr., R.A. 1982. Size discrimination with low spatial frequencies. Perception. 11:707-720. SMITH, S.T. 1990. Administration at the Egyptian Middle Kingdom frontier: sealings from Uronarti and Askut. in T.G.Palaima (ed) Aegean seals, sealings and administration. Aegaeum 5, Université de Liège, Liège [online] pp.197-216, [accessed 6/5/2013]. Available from: http://www2.ulg.ac.be/archgrec/IMG/aegeum/Aegaeum5(pdf)/Smith.pdf SPEHAR, B., C.W.G.CLIFFORD, B.R.NEWELL, R.P.TAYLOR. 2003. Universal Aesthetic of Fractals Computer and Graphics [online] 27:813-820, [accessed 17.2.2010]. Available from: doi:10.1016/S0097-8493(03)00154-7 SPLITTER, H.W. 1952. The Relation of Germanic Folk Custom and Ritual to Ealuscerwen (Beowulf 769). Modern Language Notes 67:4:255-258 STARZECKA, D.C., R.NEICH and M.PENDERGRAST. 2010. The Maori Collections of the British Museum. London: British Museum Press STATE HERMITAGE MUSEUM. 2003. Plate with a Hole in the Centre [online], [accessed 29.3.10]. Available from: http://www.hermitagemuseum.org/ STEEL, L. 2004. Cyprus before history: from the earliest history to the end of the Bronze Age. London: Gerald Duckworth & Co. STEWART, I. 2011. The Mathematics of Life. Philadelphia: Basic Books

236 STOKROCKI, M. 2001. Children’s Ethno-Aesthetic Responses to a Turkish Carpet. Journal of Art and Design Education [online] 20:3:320-331, [accessed 1/2/10]. Available from: http://www3.interscience.wiley.com STRAUSS, W.L. 1977. Introduction to a facsimile copy of A. Dürer 1525 Die Unterweisung der Messung. New York: Abaris Books STUART, G.W., T.R.J.BOSSOMAIER and S.JOHNSON. 1993. Preattentive processing of object size: implications for theories of size perception. Perception. 22:1175-1193 TAKAKI, R. and N.UEDA. 2007. Analysis of Spiral Curves in Traditional Cultures. Forma [online] 22:133-139, [accessed 28/1/2010]. Available from: http://www.scipress.org/journals/forma/pdf/2201/22010133.pdf TAKASHI, D. 2009. catalogue entries in S.Kaner (ed.) The power of dogu: ceramic figures from ancient Japan. London: British Museum Press TANG, J., Z.JING and G.RAPP. 2000. The largest walled Shang City located in Anyang, China. Antiquity. 74:479-480 TASS, P. 1997. Oscillatory cortical activity during visual hallucinations. Journal of Biological Physics. 23:21-66. TATTON-BROWN, V. 1987. Ancient Cyprus. London: British Museum Publications. TEETER, E. 1997. The life of ritual. in D.P.Silverman (ed.) Ancient Egypt. London: Duncan Baird Publishers. pp. 148-165. TE RANGI HIROA (P.H.BUCK). 1949 (rpt. 1958). The Coming of the Maori. Wellington: Maori Purposes Fund Board/Whitcombe and Tombs Ltd TEKNOMO, K. 2007. Kernel Regression. [online], [accessed 7.2.2013]. Available from: http://people.revoledu.com/kardi/tutorial/Regression/KernelRegression/ TONG, E. 2002. Magicians, magic and shamanism in Ancient China. Journal of East Asian Archaeology. 4:27-73 TUFNELL, O. 1975. Seal impressions from Kahûn town and Uronarti fort: a comparison. The Journal of Egyptian Archaeology. 61:67-101

237 TUFNELL, O. and W.A.WARD. 1966. Relations between Byblos, Egypt and Mesopotamia at the end of the third millennium B.C. a study of the Montet jar. Syria. 43:3/4:165-241 TURNER (II), C.G. 1988. Comment on: J.D.Lewis-Williams and T.A.Dowson. The signs of all times: entoptic phenomena in Upper Palaeolithic art. Current Anthropology. 29:2:201-245 UCL. 2001. Sedment. [online], [accessed 24.8.2013]. Available from: http://www.digitalegypt.ucl.ac.uk/sedment/index.html UYTTERHOEVEN, I. and I. BLOM-BÖER. 2002. New light on the Egyptian labyrinth: evidence from a survey at Hawara. The Journal of Egyptian Archaeology. 88:111-120. VALLENTIN, D. and A.NIEDER. 2008. Behavioural and Prefrontal Representation of Spatial Proportions in the Monkey. Current Biology [online] 18:1420-1425, [accessed 16.11.2009]. Available from: doi:10.1016/j.cub.2008.08.042 VAN DAMME, W. 2000. Universality and Cultural Particularity in Visual Aesthetics. in N.Roughley (ed.) Being humans: anthropological universality and particularity in transdisciplinary perspectives. Berlin: Walter de Gruyter. pp. 258-284 VEGA, E. 2008. The Marsi Chandelier – Hot Forging the Chandelier Arms [online video], [accessed 16.5.2010]. Available from: http://enrique.artmetal.com/metal_art/video/Archive.html VESSEL, E.A. and RUBIN, N. 2010. Beauty and the beholder: highly individual taste for abstract, but not real-world images. Journal of Vision. 10:2:1-14 WARD, J. 1902 (rpt. 2005) The sacred beetle: a popular treatise on Egyptian scarabs in art and history. New York: Elibron Classics/Adegi Graphics WARD, W.A. 1964. Relations between Egypt and Mesopotamia from prehistoric times to the end of the Middle Kingdom. Journal of Economic and Social History of the Orient. 7:1:1-45 __________. 1970. The origin of Egyptian design-amulets ('button seals'). The Journal of Egyptian Archaeology. 56:65-80

238 WASHBURN, D.K. 1983a. Towards a Theory of Structural Style in Art. in D.K.Washburn (ed.) Structure and Cognition in Art. Cambridge: Cambridge University Press, pp. 1-7 . 1983b. Symmetry analysis of ceramic design: two tests of the method on Neolithic material from Greece and the Aegean. in D.K.Washburn (ed.) Structure and Cognition in Art. Cambridge: Cambridge University Press, pp. 138-164 . 1990. Style, classification and ethnicity: design categories on Bakuba raffia cloth. Transactions of the American Philosophical Society. Vol. 80, part 3. Philadelphia: The American Philosophical Society. . 1995. Style, perception and geometry. in C.Carr and J.E. Neitzel (eds.) Style, society and person: archaeological and ethnological perspectives. New York: Plenum Press. pp.101-122 WASHBURN, D.K. and D.W. CROWE. 1988. Symmetries of Culture: Theory and Practise of Plane Pattern Analysis. Seattle: University of Washington Press . 2004. Symmetry Comes of Age: The Role of Pattern in Culture. Seattle: University of Washington Press WASSENAAR, J. Spiral. [online], [accessed 22/1/10]. Available from: http://www.2dcurves.com/spiral/spiral.html WEISSTEIN, E.W.2010 Wolfram MathWorld: Spirals [online], [accessed 22.1.2010]. Available from: http://mathworld.wolfram.com/topics/Spirals.html WERE, G. 2003. Objects of learning: an anthropological approach to mathematics education. Journal of Material Culture. 8:1:25-44 WHITE, J.M. and R.Y. OTSUKA. 1993. Pathways to the afterlife: early Chinese art from the Sze Hong Collection. Denver: Denver Art Museum. WINBLADH, M.-L. 2007. Four wonderful years in Cyprus: the Swedish Cyprus Expedition 1927-1931 [online], [accessed 24.8.2013]. Available from: http://cypernochkreta.dinstudio.se/empty_48.html WOOD, F. 2002. The Silk Road: two thousand years in the heart of Asia. Berkeley: University of California Press

239 WU, W. and E.N.DALAL. 2005. Perception-based line quality measurement. Proceedings SPIE 5668, Image Quality and System Performance II [online], [accessed 27.5.2013]. Available from: http://dx.doi.org/10.1117/12.585968, SPIE YAMASAKI, K. 1985. On development of drawing a square from a cross-cultural point of view. Perceptual and Motor Skills. 61:755-760 ZEILIK, M. 1985. Sun shrines and sun symbols in the U.S. Southwest. Archaeoastronomy (JHA, xvi) 9:S86-S96. ZHIMIN, A. 1988. Archaeological Research on Neolithic China. Current Anthropology. 29:5:753-559 ZHUSHCHIKHOVSKAYA, I and O. DANILOVA. 2008. Spiral Patterns on the Neolithic Pottery of East Asia and the Far East. Documenta Praehistorica [online] XXXV:215-226, [accessed 18/3/10]. Available from: http://arheologija.ff.uni-lj.si/documenta/

240

Appendix 1: Perceptual Survey The perceptual survey was designed to test perceptions of spiral motifs with four aims: to guide the definition of a spiral; to provide a basis for identifying start and end points; to guide the choice of appropriate variables; and to provide a basis for applying a weighting to variables during cluster analysis. To make this assessment, three types of test were used. Similarity measures were used to judge whether possible variables were perceptually important and were intended to suggest the weighting that should be placed on variables. Motif descriptions using keywords entered by the survey participant, or chosen from a list presented to participants, were used to support the assessment of perceptions of spiral form, orientation and curvilinearity in addition to similarity tests. Free choice and list-selected descriptors were used to identify the limits of what was considered to be a spiral to help to define the term. Free descriptors were also used to highlight the symbolic and emotional associations participants made with motifs and to highlight the potential for culturally variant interpretation. The final type of test asked participants to mark the centre or end points of motifs. These tests were mainly used in the development of the method. The centre point tests were used to assess whether any consistent means of identifying the centre of a decorative spiral could be identified. End point tests were used to judge what types of deviation from a smooth curve could be perceptually accommodated as part of the spiral and to identify tracing lines and the effect of disruption to the spiral path.

A1.1 Participants 33 participants with normal or corrected to normal vision were recruited for the survey. 26 of the participants studied design or participated in leisure activities which might imply an interest in design. Participants were informed that the purpose of the test was to identify their perceptions of motifs. They were not told that the main focus of the study was on spirals. Participants were asked their sex, nationality and age as these variables have been shown to influence perceptual judgement. Gibbs and Anderson (1977), for example, found that contrast variation affected women more than men. Faubert (2002) and

241

Gutherie, Seely, Beacham et al. (2009) found age variation, not necessarily diminution with increasing age, in perceptual discrimination. The influence of cultural background has been shown to have a strong influence on some areas of perceptual judgement (e.g. Chua et al. 2005; Yamasaki 1985; Stokrocki 2001). The influence of cross-cultural variation on the assessment of perceptual values is discussed in greater detail in Chapter 2. Of those survey participants who answered the nationality question, 23 identified as British or English with one participant each from Greece, Syria, Germany, South Korea and Australia and one participant identifying as European. The sample of participants who were not British was too small and too varied to assess the influence of nationality on perceptions. The only instance in which a culturally attributable difference from the British sample could readily be identified was in the description, by the Greek participant, of test motif 90 as the Greek letter gamma when the majority description of this motif was as an L-shape. Two participants did not state their sex, 25 were female and six male. The influence of the participants' sex could not be measured in individual tests due to the low numbers of male participants. There was no significant difference in the aggregate scores for assessment of similarity between men and women. The survey participant's ages showed two frequency peaks in the 28-37 (10 participants) and 58-67 (8 participants) groups. To create a clear division in the assessment of age variation, data from participants within the two peak frequency groups were compared. Faubert's (2002) research suggested that the ability to detect bilateral symmetry decrease with age. If the perception of rotational symmetry were also affected this might be expected to affect the placement of centre points within spiral motifs and tests of orientation similarity. Gutherie et al. (2009) found that younger participants showed a greater ability to identify shapes from their low frequency structure than older participants. No significant differences were found in identification using higher frequency signals. Within this survey, this might have been expected to affect perception of similarity in irregularity. The placement of the centres of spirals showed no clearly differentiated clustering by age. Equally, the distribution of similarities for irregularity and radial symmetry were not significantly different

242

between the age groups. These results show that age of participants did not have a significant influence on the survey results. A1.2 Apparatus and stimuli The survey was conducted online. The motifs were presented in random order on 480x480 pixel squares using a colour matched background to highlight only the intended motif rather than the card shape itself. 196 stimulus cards were used in the survey (Table A1.6, p.263). Nine cards used photographs of decorative spirals, the rest of the cards showed line drawings designed for the purposes of the survey. With one exception, all cards were designed as abstract motifs. The exception, showing a line drawing of a flower (test motif 51, Table A1.6) was intended to test whether the identification of a representational meaning skewed geometric assessments in placing centre point. A1.3 Procedure Participants were asked to adjust the screen to ensure that the motifs were fully visible on the screen. There was no limit on the length of time the motifs were displayed and participants chose when to move on to the next test motif. In tests of similarity, pairs of motifs were shown together and participants entered a similarity rating from one to ten. Motif descriptions were tested by free entry text fields for which participants were either asked to enter a short description of the shape of the motif or to describe the motif using up to five keywords. A second test of descriptors asked participants to select from four predefined keywords all those which they felt described the shape accurately. To identify centre and end points of motifs, participants were presented with two copies of each motif and asked to click to place the centre on the right hand motif causing a dot marking the point placed to appear on the left hand motif. The survey participants were allowed to adjust the motif as many times as they wanted. The use of two motifs in placing the centre and end points prevented placement of the point affecting subsequent adjustments. Alongside the test questions, twelve control questions using named non-spiral motif shapes were added asking for binary responses to motif description or placement of a marker. These were intended to reduce the sense that the survey was investigating only spiral motifs as this could have biased participants answers. However, since the majority of test motifs are spiral, this effect may not have been eliminated. The first motif each participant was presented

243

with was used as a practise motif for the placement of the centre point. Participants did not have to answer all questions. A1.4 Results The results of the perceptual tests are discussed in detail within the main body of the thesis where appropriate. This section presents the aggregated results of the tests; describes how responses were recorded; and discusses problems and alternative testing methods that might have been employed. Because participants were not obliged to answer all questions, some tests had a relatively low response rate. The lowest number of responses to any question was 18. This figure includes both missing responses due to not having completing the survey and therefore not having been presented with the test question and those participants who chose not to answer a question. Numbers of participants who chose not to answer a question were low when asked to assess similarity or asked only to place a centre or end point. A maximum of two voluntary non-responses were obtained for these types of questions. Higher non-response rates to other questions are discussed in the relevant sections below. A1.4.1 Similarity The aggregated similarity results show a skew towards rating motif pairs as similar. Since the question asked participants to assess 'On a scale of one to ten, with one being least similar, how similar are these [two motifs/descriptive motif name]?' it is possible that this trend towards similarity arises from an agreement bias with the participants seeing a response towards the 'similar' end of the scale as more desirable. Possibly a more appropriate approach would have been to use the method employed in a number of published similarity tests where participants are asked to separate clusters of motifs into groups (Beck 1966) since that approach would be more likely to encourage a neutral response. Significant results particularly tended towards higher similarity which meant that differences between the modal scores of variables were minimal and made differentiation for weighting difficult. Because similarity results were high and because the sample group was culturally homogeneous, similarity results were not used to assign weightings to variables in the creation of type clusters. Some similarity pairs of motifs tested more than one variable, for example when testing path width it is necessary to also vary spacing or diameter. It had been hoped

244

that by pairing each variable under consideration with a range of other variables, the impact of having multiple uncontrolled variables could be balanced and a meaningful aggregate similarity score obtained. The very low numbers of significant results hampered this aim and made the assessment of the value of individual variables difficult. Table A1.1 Similarity values by test motif motif number

comparing

TG131 TG132 TG133 TG134 TG135 TG136 TG137 TG138 TG139 TG140 TG141 TG142 TG143 TG144 TG145 TG146 TG147 TG148 TG149 TG150

irregularity irregularity orientation, irregularity curvilinearity spiraloids vs. concentric circles spiraloids vs. circles curvilinearity irregularity (low vs. high) irregularity (none vs. low) irregularity (low), colour irregularity (low vs. medium) patterning (type) direction of turn patterning vs. solid fill patterning/local irregularity/curvilinearity spacing, rotational length diameter, spacing diameter, rotational length diameter, spacing spacing (none vs. constant present), path width, eye presence form, path width, spacing (none vs. increasing) path width (constant vs. increasing), form spacing (none vs. increasing), path width (constant vs. increasing), form orientation rotational length (more than 360˚ vs. less than 360˚) rotational length (more than 360˚ vs. less than 360˚) orientation spacing (constant vs. fluctuating), position of centre form, diameter interrelationship (linked vs. unlinked) interrelationship (radiating vs. scatter)

TG151 TG152 TG153 TG158 TG159 TG160 TG161 TG162 TG163 TG164 TG165

significant (chi square) 0.038 0.003 ns ns ns ns ns ns ns <0.001 ns ns <0.001 ns ns ns ns <0.001 ns ns

modal value (10 represents most similar) 8 10

9

9

9

ns ns ns ns ns ns 0.03 ns ns ns ns

9

245 TG158A TG158B TG158C TG158D TG163A

orientation rotational length, centre diameter rotational length, spacing rotational length, spacing form, spacing (constant vs. fluctuating)

0.001 ns ns 0.016 ns

TG163B

form, spacing (constant vs. fluctuating)

ns

TG163C TG164A

form interrelationship, orientation (consistency within group) orientation interrelationship three-stemmed spiral vs. three-stemmed wave interrelationship three-stemmed spiral vs. three-stemmed circle interrelationship three-stemmed spiral vs. three-stemmed arc interrelationship centre linked vs. border linked

<0.001 0.018

8 8

0.001 ns

8.5

interrelationship (varying direction of turn), form interrelationship (centre and border linked vs. centre linked), rotational length curvilinearity colour colour direction of turn, local irregularity direction of turn irregularity (low), spacing (fluctuating low vs. fluctuating higher) form, spacing (none vs. increasing), eye presence

ns

TG164B TG164C TG164D TG164E TG164F TG164G TG164H TG137B TG139B TG139C TG142B TG142C TG147B TG151B

9

9

ns ns ns

ns ns 0.003 0.004 ns ns ns

10 9

ns

Table A1.2 Overview of similarity ratings by variable variable number of significant significant times similar (mode dissimilar (mode tested above 5.5) below 5.5) diameter 4 1 0 direction of 4 1 0 turn rotation 8 2 0 orientation 5 4 0 irregularity 9 3 0 path width 4 0 0 spacing 13 1 0 form 9 1 0 colour 3 3 0 curvilinearity 4 0 0

median score across all significant tests 9 9 9 8.75 9 n/a 9 8 9 n/a

246

A1.4.2 Placement of centre and end points Five polygonal test motifs (Figure A1.1a-e) were used to assess how precisely participants placed points and to test the extent of centre agreement between participants with simple shapes which exhibited a high degree of symmetry. These results showed a high level of agreement for the regular heptagon and the parallelogram. A similar result was obtained for the regular hexagon, although it is possible that the three participants who placed points above the main group were abstracting a cube. In both cases clustering is tight indicating a precise placing of points. With any degree of irregularity, however variation in the placement of centre point increased. For unclosed shapes (Figure A1.1f-h), there was a further increase in variation in the placement of centre points. A regression to the axis of symmetry is visible in unclosed symmetrical or near-symmetrical shapes but the spread of points along the axis of symmetry suggests that either the centre of gravity or the centre of the bounding rectangle could have been used as a basis for assessment of the centre and no consistent basis for preferring any one centre determination method emerged. Figure A1.1 Symmetrical polygons: a. test motif 166, b. test motif 168, c. test motif 167, d. test motif 56, e. test motif 169. Unclosed shapes: f. test motif 54, g. test motif 55, h. test motif 29. (Continues onto next page.)

a

b

247

c

d

e

f

g

h

Some results for the tests of placement of centre or end points were excluded where participants were believed to have misread the question (for example marking the outermost end point of the spiral on the innermost end point). These results were excluded from analysis as were points which were placed too far from the motif to draw accurate conclusions about the point they were trying to indicate. In tests to

248

identify the outermost end point, points which were a distance of more than six times the line width from the termination of the line and more than two point diameters from any part of the motif path were discounted. Points which were far from the black or coloured line of the motif might indicate that the participant was identifying the ground as the dominant motif but there was too little consistency in the placement of off-motif points to suggest any common interpretation of the ground as the dominant spiral. In tests to mark the centre of the motif, a greater latitude had to be allowed, only points which fell outside the bounding rectangle of the motif were excluded. In 18 tests (test motif nos. 92-101, 100a-c, 166-169), participants were asked to both describe the motif shape and mark the centre point to test whether interpretation of the shape had an influence on the perception of its geometric properties as indicated by the centre point location. This was assessed by coherence of grouping of point placement amongst participants who assigned representational or non-representational meanings. There was a markedly lower response to the placement of centre points for these questions than for questions which only asked for placement of a centre marker. This is probably due to participants forgetting to answer both parts of the question; it would have been appropriate to introduce a prompt to encourage answering both parts before continuing. A1.4.3 Motif descriptions The use of descriptors in defining a spiral assessed both free text descriptions of shapes to see how frequently the term 'spiral' or synonyms were used and to identify which motifs were readily identified as spiral. Table A1.3 presents the results of motif shape descriptions. Table A1.4 shows the terms used when participants were asked to describe the motif without an emphasis on its shape. These values were used to assess culturally variant responses to motifs in Chapter 4. Where keywords were difficult to interpret or could be assigned to more than one conflicting category (representational or geometric) they were recorded as ambiguous. The predetermined keyword test was used to identify the rotational length needed to define a motif as a spiral (Table A1.5). Questions asking for description had the lowest response rates including some responses stated as 'don't know'. As noted in Chapter Two, this non-response has implications for cross-cultural validity.

249

A1.5 Potential expansion of the perceptual survey To refine the method it would be useful to expand the perceptual survey to a more varied sample group and to explore how appealing participants find different spiral motifs – this could be used to add to the assessment of whether spirals which conform to subjective or objective natural forms would be likely to be represented in decoration by assessing whether they appeared to be consistently preferred as spiral types. It would also be interesting to introduce some tests in which participants are asked to draw spirals to look for commonalities and to see what attributes are assigned to spirals when subjects are given a free choice to draw spirals. Asking participants to copy or draw a natural spiral form with and without a visual cue could be used to assess how far the drawings differed from actual spiral shape in the representation of natural forms and what the key characteristics of the form appeared to be.

250

Table A1.3 Free text descriptors used for motif shape by test motif number and survey participant number. Test motif number

86

87

88

89

90

91

92

P1 P2 P3 P4 P5 P6 wave

iron cloths An iron rhombus wavey spiral

upside down L

hook

extended spring

spiral with ripples from centre

triangular key pattern

rectangle with rectangular shape removed from bottom right

angular spiral

twisted wire

ammonite

celtic carving

irregular hexagon

spiral

spiral bend

P10 wave

random curly

segmented spiral

triangular spiral

L shaped

geometric spiral

random

P11 wavy line P12 P13 wavy

swirl

starburst

irregular

spiral

L shaped L shape rectangular

key

spring

triangular key Triangular triangle

angular

angular

P14 wave

wiggle

waves

box

trapezoid spiral

P7 wave P8 wave P9 wave

P15 wave

P16

looped wave

spiral

bent spiral

double anti clockwise spiral

triangular spiral

triangular key

upside down L

rectangular maze

L shaped

angular square clockwise spiral

broken angular rectangle irregular broken pentagon

250

loops

Sleeve unsure

251

Test motif number

86 P17 P18 P19 P20

wave wave wave wave

87

89

90

91

don’t know L hook L shape

P21 wave

a curvey line

wavey sunlike spiral shell

triangle with in a triangle

upside down L

a maze

P22 wave P23 wave P24 P25 zig zag P26 wave P27

Squiggle scribble loops

Upside down L shaped box

ammonite

Triangular Spiral triangle

domestic iron shap[e] iron

coil

squiggle

P31 wavelength

scribble

P32 wave

looped line

P33

spiral

Triangle

Coil sandy sun

triangular coil Maze triangle spiral

well aligned spiral coil with radial waves from centre point

stylized helix angular coil

92 spiral kite stylized curl angular spiral

kite

open pentagon

L-square Upside down L. broken rectangle

Rhombus broken spiral

collapsed pentagon paper clip

trianlged spiral

Greek letter gamma

rectangular spiral

almost pentagon

triangular double spiral traingulation repeating motif

closed rectilinear shape. inverted L

trapezoid spiral

angular line

251

spiral maze triangular curl knot

spiral

freehand loops loops

spiral fossil helix sunburst

P28 P29 W P30

looped

88

252

Test motif number

93

P1

line or kite

P2 P3 P4 P5 P6

P10 P11 P12 P13 P14

95

Spaceship

arrow many uncomplete triangles Mountain

Spiral

zig zag

jagged spiral

96

97

bowl

sleeping pinguin Suit

Coffin

smile Bowl

jagged spiral

jagged spiral

jagged spiral angular spiral angled spiral bent geometric spiral

angular spiral

angular scroll

angular spiral

angular scroll

spiral bend geometric spiral

spiral bent geometric spiral

squiggle spiral

spiral

random spiral

random spiral

spiral geometric spiral

key

asymmetrical key

distorted key

angular

angular

triangular

irregular key Irregular angular

broken angular rectangle

P16 P17

irregular fractured pentagon spiral

99

kite

angular spiral

P15

98

abstract rectangle

extending right angled triangle

abstract rectangle

spiral

fractured triangle spiral

irregular angular anti clockwise spiral spiral

asymmetrical key angular

abstract rectangle

spiral

broken irregular hexagon spiral

tornado Spiral

orange spiral semicircle cup

spiral spiral spiral

semicircle

spiral

arc

bowl

spiral Spiral spiral spiral

arc

this is a spiral

bowl semi-circle

anti clockwise irregular spiral spiral

angular

exploding rectangle

100

252

P7 P8 P9

Kite

94

253

Test motif number P18 P19 P20

93 stylized curl angular spiral

94 maze stylized helix zig-zag coil

95 triangle stylized helix hook

mis shaped box

broken triangle sharp e triangular

nearly a box e paperclip

P26 P27

open pentagon

meander

Broken pentagon

collapsed heptagon Broken rhombus

P28 P29 P30 P31

P32 P33

wanna be pentagon

triangled spiral

irregular angular spiral

irregular angular spiral

spiral geometric

stylized spiral angular coil several straight lines at different angles jagged spiral diamond

geometric spiral

97 coffin stylized helix hook

a broken kite upside down e kite

98 house stylized helix zig-zag hook triangle within opensided rectangle e shape kite

geometric spiral

line geometric spiral

99 smile half circle arc

100

crescent 0 smile

spiral Spiral coil

A spiral spiral

spiral 253

P21 P22 P23 P24 P25

96

spiral bent triangle open triangle

Diamond broken legs lines

abstract and unbalanced

squared spiral

angular line

irregular angular spiral

broken coffin broken diamond destroyed paperclip irregular angular anticlockwise spiral

paper clip

wanna be spiral irregular angular anticlockwise spiral

smile

spiral

half circle

spiral

half circle

spiral

semicircle

circular anticlockwise spiral

254

Test motif number P1 P2 P3 P4 P5 P6

101 line Saucepan

P10

random spiral

169

100a

7-sided 50 pence piece

rectangle Window

hexaedr Hexagonal

triangle Hill triangle

bowl pie chart Orange segment Pram cheese

quadrangle heptagon

parallelogram

heptagon

heptagon

parallellogram parallelogram rhomboid quadrilateral

hook

no idea

rhombus

hexagon

irregular

heptagon

parallelogram

heptagon

trapezoid

P14

P16

168 Hexagon

bent hook

hexagon hexagon hexagon

hexagon

100b

100c pie chart Pacman

semi-circle

pacman

triangle

semicircle

segment of circle

segment of a circle

triangle

bowl

packman

packman

irregular triangle

semicircle

segment of circle

254

P7 P8 P9

P15

167

crooked line rudimentary spiral random squiggle

P11 P12 P13

166

semicircle

hexagonal

triangle Scalene triangle triangular

circle segment no idea of its name - utility clothing mark

semicircle

part of a circle

segment of a circle

hexagon

triangle

semi-circle

pie

pie

half a circle

three segments of a circle

three quarter segments of a circle

half

circle with radial wedge missing

circle with radial wedge missing

septagon

rhombus

hexagon

isosceles triangle

regular heptagon

parallelogram

equilateral hexagon

scalene triangle

part circle

255

Test motif number P17 P18

101 loop

166 heptagon honeycomb

167 parallelogram rectangle

168 hexagon

169 triangle

100a semi-circle

P19 P20

Freeform loose coil

heptagon heptagon

rhombus rhombus

hexagon hexagon

scalene triangle triangle

half moon semi-circle

100b pacman cheese circle with a wedge removed circle segment

P21

a worm

7sided penny

rectangle squew wif rectangle

hexagon

triangle

half a circle

a circle with a 6th cut out

100c pacman pacman circle with a wedge removed circle segment looks like an orange with 2 segments cut out

hexagon hexagonal

Triangle triangle

Semi Circular semicircle

Pie circle

pie cheese

hexagon Hexagon

triangle Triangle

half circle

hexagon hexagon

triangle

P28 P29 P30 P31 P32 P33

heptigon loop

255

P22 P23 P24 P25 P26 P27

heptagon

Lump

heptagon

dreamt to be a circle irregular curled line

heptagon regular heptagon

diamond skew rectangle skewed rectangular hexagon regular quadrilateral hexagon

non equilateral triangle scalene triangle

pie

semi-circle half close circle

part of a circle

part of circle

half closed circle pac man two-thirds of a circle closed two thirds 2 thirds circular semicircle circular segment segment

256

Table A1.4 Free text keywords for motif description. Participants were asked to enter up to five keywords. Test motif number 154 P1 P2 P3 P4 P5 P6

snail Spiral

P7

spiral

155

fossil

clockwise

stone

spiral

carved

carving spiral bas-relief

ethnic

ancient

weathered

spiral spiral

thick snail

complete shell

perfect

regular

spiral

relief

carved

P16

spiral

two-dimensional clockwise

Snake

spiral

tight

stone

ancient

spiral

coil

spring

concentric

celtic spiral coil

spiral regular

carving organic

textured

ethnic

spiral spiral

cicular

concave

evenly spaced

clockwise

concave

open

spiral

carved

clockwise

spiral

relief two dimensional

regular

256

spiral open spiral

P15

spiral Rope

spiral

P8 P9 P10 P11 P12 P13 P14

snake Spiral

257

Test motif number P17 P18 P19 P20

154 spiral shell helix coil

fossil spirl spiral

spiral curl loop

wave regular

P21 P22

spiral Spiral

rough Carved

stone stone

grey Old

P23 P24 P25 P26 P27 P28

curly

coil

waves

ammonite

P29 P30

Coil Spiral

spiral Deep

snail shell circular

paisley Ston

P31 P32 P33

historic spiral

spiral coil

stone clockwise

timeworn

155 spiral curl loop

naive Snail

spiral Fossill

blue grey stonework spiral snail

regular deep carving stone

spiral

coil

fossil

ammonite

spiral spiral

pattern

stone architect

spiral

mosquit o coil

curl

twirl

spiral spiral

stone coil

homogeneous anti-clockwise

balanced

twirl brown

helix coil

eye like old

257

curlecue

spiral spiral

paisley

258

Test motif number 156

157

P1

spiral

tribal

wave

horn

P2

bone

six

curl

hook

P3

curl music clef

spirals

flowers

Loop

Curly

Elegant

Intricate

P4 P5 P6

spiral

spiral

iron

P7 P8

spiral

branching

spirals

swirling

series

scrolls

interlocking

spiral

wave

anticlockwise bone

curvatures

thick

258

P9

spiral

pendant

hook

morphed shape

P10 P11 P12 P13 P14

spiral hook

wave curve

irregular

curlicue

ethnic

scroll scroll

curlicue organic

floral water

apiral spiral

curved

thick

white

necklace

spiral spirals

irregular

flowery

P15

spiral

white

concave

varied in width

hole for hanging

scroll

spiral

metal

P16 P17

spiral spiral

flat

oval

clockwise

spiral spiral

asymmetric swirl

art deco fronds

pleasing growing

ornate

double centre

259

Test motif number 156

157

P18 P19 P20

spiral spiral

helix coil

curl irregular

organic

scroll curl spiral

music note scroll curlecue

spiral spiral coils

ornate

organic

P21 P22

spiral Spiral

delicate Curved

bone like Loopy

smooth Number

curves curly

smooth spiraly

flows musical

wave like scroll

dramatic beautiful

P23 P24 P25 P26 P27 P28

curls

waves

shell

ammonite

curly

scrolls

loopy

hook

maori

spiral

P29 P30

spiral white ancient Greek spiral

curl spiral

paisley light spin

curl black growing plant scroll

coil metal

abstract

spiral spiral multiple spirals spirals

259

P31 P32 P33

wave like White

pendant

twirl

metallic curlicue

swirl asymetric

260

(Table A1.4 continued) Test motif number P1 P2 P3 P4 P5 P6

157b flower Star green plate

Flower

pattern flower overlapping petals

expanding

central

patterned complex

arabic

P7 P8

flower

P9 P10 P11 P12 P13 P14

regular flower geometric floral radiating Fibonacci

dodecagon regular Moorish

flowery

symmetrical hexagonal

P15

circular flower star shaped

dahlia

interlocking daisy centre diamonds

P16 P17 P18 P19 P20

flower flower hills flower floral

mandelbrot stones sun regular

trees spirograph organic

peacock

P21 P22

beautiful Window

colourful Green

lovely centred

dramatic flower

interesting plate

P23 P24 P25 P26 P27

flower

mosaic

star

flower Flower circular

plate mosaique green

geometrical

leaf

dotted

symmetric

P28

flower

dots

green

P29 P30

flower Pattern

religious stained glass dots

repeated geometrical curvy twelvesided plant Symetric

curves green

symetrical

P31 P32 P33

circular radial

flowers spiral

central elipses

dotted

green

extending

interweaving

261

Table A1.5 Frequency of choice of predefined keywords. The division of the 'spiral' keyword by rotational length was carried out during analysis and not presented to survey participants. loop spiral <270 270360 13 13 14 13

hook 361- >450 450 2 6 2 5 6 6 2 7 3 5

0 0 2

coil figure double zig-zag of spiral eight 8 10 4 11 10 9 22 6 1

5 4 2

0 5 8 13 5 10

0 0 0 2 0

2 1 3

key wave tile pattern pattern

0 1 5 4 4 16 6 10 16

5 21 20

18 16 15 12 10

4 2 2 1 1

14

13

20 4 15 12 1

3 2 4 3 9 12

whirl meander scroll curlicue

261

test motif number 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121

15

6

262

Table A1.6 The set of stimulus cards identified by test motif number

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

263

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

264

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

265

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

266

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

100a

267

100b

100c

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

268

119

120

121

122

123

124

125

126

127

128

129

130

131

132

134

133

135

269

136

137

137b

138

139

139b

139c

140

141

142

270

142b

142c

143

144

145

146

147

147b

148

149

271

150

151

151b

152

153

156

154

155

157

158

157b

158a

272

158b

158c

158d

159

160

161

162

163

163a

163b

163c

164

273

164a

164b

164c

164d

164e

164f

164g

164h

165

168

166

167

169

Egypt (Dynastic (Lower)) Egypt (Naqada (Upper)) Nigeria (Yoruba/Hausa) Ghana (Akan/Ashanti) Atlantic rock art N.European (late-Roman-Anglo N.European (Celtic) Aegean (Cyprus) Aegean (Crete) Aegean (mainland Greece) Aegean (Cyclades) Italian (Iron Age-Italic) Italian (Etruscan) Macedonia Malta Hungary, Germany Romania (MBA)

29000-25000

Balkan pottery

29000-25000 29000-25000 China (Yangshao/Majiayao, Shang-Zhou) 18000 China (Malyshevo, Voznesenovka) 18000 Japan (Jomon) 15000-14000/10000 Cucuteni and Trypilian

Thailand (Ban Chiang) Vietnam (Dong Son) Peru

possibly around 8000 but see later for surviving date

Mexico (Aztec) Mexico (Olmec) Brazil (Marajoara) Maori

carved stone spirals

metal spirals

woven textile spirals non-woven textile spirals

copper working horizon

basketry spirals

bronze spirals cast bronze spirals wrought bronze spirals iron spirals

pottery spirals modelled/incised pottery spirals painted pottery spirals carved wood spirals

gold spirals cast gold spirals wrought gold spirals

carved stone horizon carved stone known woven textile horizon woven textile known coil basketry horizon coil basketry known pottery horizon pottery known carved wood known

metalworking horizon metalworking known copper working horizon copper working known bronze working horizon bronze working known iron working horizon iron working known gold working horizon gold working known

Table A2.1 Early examples of spiral decoration and the distribution of prehistoric and early historic cultures in which spirals were commonly occurring motifs approximate technology horizons indicated reflecting dates of introduction/discovery or earliest known examples

25 0 50 0 75 0 10 00 12 50 15 00 17 50 20 00

-7 00 0 -6 75 0 -6 50 0 -6 25 0 -6 00 0 -5 75 0 -5 50 0 -5 25 0 -5 00 0 -4 75 0 -4 50 0 -4 25 0 -4 00 0 -3 75 0 -3 50 0 -3 25 0 -3 00 0 -2 75 0 -2 50 0 -2 25 0 -2 00 0 -1 75 0 -1 50 0 -1 25 0 -1 00 0 -7 50 -5 00 -2 50 0

274

275

Table A2.2 Contextual and style data for the Jōmon case study Context data

JBM4

crown pot, representational snakes

MJ

Iwanohara site, Nagaoka City

flame pot, prob. same source as Jōmon loan 1

MJ

2

1

Kidder & Esaka Kidder & Esaka Kidder & Esaka Kidder & Esaka

pl.63

JKAE1

1

vessel, Fukura type, h.19.7cm

late EJ

?Iwanohara site, Nagaoka City (see notes). Kaner 2009, p.168 associated flame pots with the Shinano River area Yamagata

pl.VI

JKAE2

1

EJ

Nagano

pl.VIII

JKAE3

1

vessel, Tōnai type, Sori, h.27.3cm vessel, Toyama, h.37.4

EJ

Asahi

pl.105

JKAE4

Kidder & Esaka

pl.121

JKAE5

Kidder & Esaka

pl.122

Kidder & Esaka

pl.125

2

1

vessel, Nashikubo type, Nakahara, h.29.5

MJ

Nagano

1

vessel, Katsusaka 3 type, Takikubo, h.36.8

MJ

Tokyo

JKAE6

1

vessel, Tōnai type, Iwakubo, h.64.5

MJ

Yamanashi

JKAE7

1

vessel, Tōnai type, representational - snake, h.37 vessel, Tōnai type, h.37

MJ

Nagano

vessel, Katsusaka 2 type, Nakahara, h.37.2

MJ

Tokyo

vessel, Katsusaka 2 type, Nakahara, h.35.5

MJ

Tokyo

2

2 Kidder & Esaka

pl.130

Kidder & Esaka

pl.131

JKAE8

1 2

JKAE9

1 2

sampling resolution

temporary exhibition loan from the Tsunan Town Board of Education, Japan

British Museum

1

spacing trend error

JBM3

2

spacing median

British Museum

temporary exhibition loan from the Tsunan Town Board of Education, Japan

Hokkaido

path width trend error

LJ

path width median

pot with spiraloids

JBM2

irregularity

1

OA+.654

curvilinearity

MJ

place

orientation

1

Museum ID/ plate no. 1998,1218.1

rotation

spiral numbers analysed

JBM1

Object notes pot with spirals/spiraloids

Museum number/ publication British Museum British Museum

period*

Object number

Style data

7.99

0.30

0.65

0.27

1.27

43.22

1.02

1.86

0.04

0.03

-0.37

1.21

0.05

-0.71

1.38

6.63

0.36

0.35

0.28

0.74

15.87

2.44

0.01

0.05

0.03

-0.24

0.45

33.88

-24.20

1.52

6.80

2.71

0.75

0.28

0.77

57.14

2.38

0.02

0.05

0.03

-0.24

0.64

-7.82

2.08

-0.11

8.50

2.63

0.79

0.41

1.74

17.58

1.55

6.52

0.04

0.01

-0.08

0.31

0.25

0.33

-1.23

11.37

2.59

0.65

0.40

0.91

19.88

1.38

8.46

0.04

0.00

-0.02

0.10

0.08

-0.43

0.80

9.70

2.74

0.84

0.16

1.62

73.27

0.84

1.46

0.04

0.01

-0.07

0.11

0.00

0.23

-1.18

10.68

0.30

0.90

0.31

0.84

21.83

2.11

23.52

0.04

0.00

-0.02

0.13

0.11

-0.60

1.04

23.81

0.09

0.80

0.13

0.46

10.62

0.72

144.10

0.05

0.00

0.00

0.01

0.00

0.01

-0.09

16.70

1.85

0.82

0.12

1.06

75.07

0.37

28.98

0.04

0.00

-0.01

-0.01

0.00

0.00

0.00

13.99

1.88

0.91

0.15

2.34

147.08

1.99

0.29

0.03

0.01

-0.17

0.93

1.28

-3.43

1.88

10.89

1.09

0.93

0.24

0.91

5.85

1.93

0.15

0.04

0.00

-0.05

0.27

-0.05

0.33

-0.23

10.78

0.84

0.89

0.25

1.53

6.78

1.78

0.36

0.03

0.01

-0.08

0.28

-0.01

-0.02

0.39

13.42

2.24

0.87

0.24

1.48

63.36

0.32

3.57

0.04

0.00

-0.02

0.08

0.01

-0.04

-0.02

12.11

0.57

0.88

0.10

0.45

0.93

1.07

0.01

0.04

0.00

-0.03

0.17

0.00

0.00

0.12

6.43

2.07

0.72

0.26

0.84

2.61

2.45

0.01

0.05

0.01

-0.10

0.27

71.95

-10.46

9.92

12.40

2.08

0.90

0.17

1.43

43.58

0.63

1.04

0.04

0.00

-0.05

0.25

-0.01

0.06

-0.14

11.83

0.18

0.89

0.17

1.37

102.42

0.76

1.64

0.04

0.00

-0.04

0.17

-0.03

0.25

-0.54

15.03

2.41

0.86

0.12

1.35

67.39

0.37

5.88

0.04

0.00

0.02

-0.09

-0.01

0.10

-0.42

8.16

0.79

0.86

0.60

0.88

8.20

3.08

1.43

0.04

-0.02

0.21

-0.53

-0.10

1.58

0.62

7.55

1.16

0.89

0.65

1.21

4.19

3.80

0.49

0.04

-0.01

0.08

-0.04

0.59

-0.58

3.87

path width trend coefficients

spacing trend coefficients

276 Context data

1

Kidder & Esaka

pl.142

JKAE13

Kidder & Esaka

pl.156

JKAE14

1 3 1

Tokyo

MJ

Nagano

3 1

vessel, Katsusaka 3 type, Tsutano, h.59

MJ

Yamanashi

MJ

Nagano

MJ

Tokyo

2

vessel, Idojiri type, ?representational - snakes, h.27.5 vessel, Katsusaka type, Nakahara, h.36.6

1

vessel, Sori type, 42.7cm

MJ

Nagano

vessel, Katsusaka type, Sakai, h.21

MJ

Yamanashi

2

sampling resolution

JKAE12

MJ

spacing trend error

pl.140

vessel, Katsusaka type, Takikubo vessel, Tōnai type, h.47.6

spacing median

1

path width trend error

JKAE11

path width median

pl.137

irregularity

Kidder & Esaka Kidder & Esaka

curvilinearity

2

MJ

orientation

1

place Tokyo

rotation

JKAE10

Object notes vessel, Katsusaka type, Narahara, h.34.5 representational - snake

period*

spiral numbers analysed

Museum ID/ plate no. pl.136

Object number

Museum number/ publication Kidder & Esaka

Style data

7.67

2.46

0.36

0.17

2.25

2.37

0.91

0.18

0.04

0.01

-0.10

0.39

-0.17

1.23

-1.91

7.27

1.82

0.87

0.30

0.87

19.13

2.97

0.12

0.04

0.01

-0.09

0.15

-6.13

6.69

1.81

15.80

0.67

0.79

0.14

1.27

140.01

0.21

12.89

0.04

0.00

-0.01

0.01

0.00

-0.01

0.07

12.02

0.11

0.85

0.27

1.24

80.11

0.61

16.38

0.04

0.01

-0.12

0.60

-0.01

0.05

0.00

9.19

1.92

0.81

0.28

1.15

16.25

0.30

6.15

0.04

0.00

-0.03

0.20

0.26

-0.17

-0.29

9.61

0.64

0.83

0.11

1.82

81.75

0.55

1.86

0.04

0.01

-0.10

0.40

-0.01

0.08

-0.24

11.60

0.97

0.81

0.23

1.73

29.38

0.56

1.05

0.04

0.00

-0.08

0.40

-0.01

0.14

-0.35

7.97 7.01

1.34 1.79

0.88 0.86

0.25 0.30

2.31 1.54

138.80 11.00

0.44 1.39

0.34 0.02

0.03 0.04

0.01 0.01

-0.12 -0.06

0.61 0.36

0.15 -0.11

-0.01 0.93

-0.49 -0.58

11.91

0.79

0.87

0.22

1.41

35.95

0.74

10.36

0.04

0.01

-0.09

0.30

-0.05

0.51

-1.26

7.25

0.14

0.75

0.29

2.47

36.48

0.95

0.18

0.04

0.02

-0.28

1.17

0.05

0.26

-0.96

7.21

2.00

0.82

0.27

2.26

38.81

1.05

0.10

0.04

-0.01

0.13

-0.42

0.07

0.98

-1.25

7.77

2.86

0.87

0.28

2.92

22.24

0.60

0.30

0.04

0.01

-0.14

0.82

0.27

0.84

-0.92

7.00

2.95

0.91

0.32

1.22

42.49

1.68

0.01

0.04

0.03

-0.23

0.33

0.96

-1.75

0.69

7.11

2.21

0.89

0.40

2.07

39.95

1.64

0.02

0.04

0.00

-0.03

0.15

1.54

-0.05

-0.98

22.75

2.56

0.33

0.20

0.89

46.16

0.34

15.86

0.04

0.00

-0.01

0.09

0.00

0.00

0.04

40.54

0.66

0.90

0.05

0.62

106.81

0.06

12.59

0.05

0.00

0.00

-0.04

0.00

0.00

-0.01

16.48

2.25

0.85

0.18

1.20

58.73

0.22

12.55

0.04

0.00

-0.02

0.14

0.00

-0.02

0.02

path width trend coefficients

spacing trend coefficients

Kidder & Esaka

pl.159

JKAE15

Kidder & Esaka

pl.160,162

JKAE16

Kidder & Esaka

pl.170

Kidder & Esaka

pl.172

JKAE18

1

vessel, Sori type, ?representational, 29.5cm

MJ

Nagano

Kidder & Esaka

pl.176

JKAE19

1

vessel, Sori I type, 61.2cm

MJ

Yamanashi

Kidder & Esaka Kidder & Esaka

pl.180

JKAE20

1

MJ

Nagano 10.78

0.32

0.81

0.17

1.08

56.45

1.28

6.09

0.04

0.00

0.00

0.04

0.03

-0.28

0.63

pl.182

JKAE21

1

vessel, Kasori E (Ubayama) type, Miyanome, 59.5cm vessel, Umataka I type, h.29.7

MJ

Niigata

8.99

2.25

0.75

0.36

1.15

21.98

0.92

1.71

0.04

0.01

-0.19

0.66

0.28

-0.38

-0.14

8.24

1.49

0.87

0.19

1.26

61.43

0.59

0.83

0.04

0.00

-0.04

0.20

0.62

-1.19

0.57

Kidder & Esaka

pl.157

JKAE22

vessel, Katsusaka I type, Minebata, h.21.6

MJ

Nagano

10.35

2.43

0.80

0.13

1.54

50.46

1.16

8.07

0.04

0.00

0.02

-0.06

-0.12

0.66

-0.67

7.38

2.56

0.89

0.28

1.23

10.58

3.30

0.20

0.04

0.02

-0.22

0.66

1.05

-2.15

2.96

Kidder & Esaka

pl.187

JKAE23

1

vessel, Umataka 1 type, h.37.5

MJ

Niigata 10.64

2.56

0.87

0.24

1.55

34.61

0.80

11.03

0.04

0.00

-0.02

0.09

-0.05

0.42

-0.75

Kidder & Esaka

pl.192

JKAE24

1

vessel, Umtaka type, some Late Jōmon traits, h.32.5

MJ

Umtaka, Niigata

10.56

2.92

0.35

0.20

1.78

34.13

0.80

1.22

0.04

0.00

0.03

0.07

-0.01

0.14

-0.52

Kidder & Esaka

pl.196

bowl, Umtaka type, h.11 cm

late MJ

Umtaka, Niigata

10.02

2.58

0.77

0.27

0.91

57.37

0.70

14.61

0.04

0.00

0.03

-0.25

0.24

-0.65

0.21

8.09

2.80

0.89

0.26

2.58

13.89

0.83

0.31

0.03

0.00

-0.08

0.68

0.99

-2.72

2.46

Kidder & Esaka

pl.199

jar, Umtaka 3 type, h.32 cm

(late?) MJ

Umtaka, Niigata

9.10

2.56

0.85

0.21

1.31

19.09

1.04

2.43

0.04

-0.01

0.13

-0.55

0.30

-0.97

0.36

7.55

2.19

0.73

0.45

1.44

21.89

1.64

0.45

0.04

0.02

-0.28

0.85

-1.64

4.82

-2.31

Kidder & Esaka

pl.200

JKAE27

1

jar, Tochikura 2 type, h.23.2

MJ

Tokushōji, Niigata

8.50

2.43

0.80

0.23

1.36

25.39

1.15

0.66

0.04

0.02

-0.16

0.28

-0.07

-0.01

0.10

Kidder & Esaka

pl.201

JKAE28

1

pot, Kasori E 1 type, h.21, representational, bird's wings

MJ

Sakai, Yamanashi

8.89

2.93

0.94

0.14

1.50

46.20

1.13

1.32

0.04

0.00

-0.06

0.27

0.32

-1.10

0.93

10.69

0.06

0.84

0.17

1.24

31.50

0.92

3.62

0.04

0.00

-0.05

0.16

0.11

-0.71

1.20

2 JKAE17

1 2

2

2 1 2

JKAE25

1 2

JKAE26

1 2

2

277 Context data

1

cooking pot (evidence of use), Monzen type analogous to Daigi 8b, pot shape resembles Late Jōmon, h.26.5

MJ

Monzen, Iwate

jar, Daigi 8a type, h.75cm

late MJ

Ochiai, Yamagata

jar, Kasori E 3 (Ubayama) type

MJ

Yosukeone, Nagano

?burial jar (base removed), comparable to Kasori E 2 (Ubayama) type, ?representational - bird's wings, h.48

MJ

Miyaji, Gifu

vessel, Kasori E 3 (Ubayama) type, h.34

MJ

vessel, Kasori E 3 (Ubayama) type, h.78

late MJ

Kidder & Esaka

pl.208

JKAE33

Kidder & Esaka

pl.215

JKAE34

Kidder & Esaka

pl.217

1 2 1 2

JKAE35

1

2 Kidder & Esaka

pl.219

JKAE36

1

Kidder & Esaka

pl.220

Kidder & Esaka

pl.221

JKAE38

1

vessel, Kasori E 1 (Ubayama) type, h.64

MJ

Ōne, Gumma

Kidder & Esaka

pl.223

JKAE39

1

vessel, Kasori E 2 (Ubayama) type, h.25.7

MJ

Narahara, Tokyo

Kidder & Esaka

pl.XV

JKAE40

1

MJ

Nusamai, Hokkaido

Kaner 2009

39 (pp.136, 132)

JKAN1

1

boat-shaped vessel, Nusamai type, typical of Kushiro, E.Hokkaido, h. 17.1 pot, incised decoration, spirals on lower part said to form the heads of human figures, this may be true of the other pots. dogū, incised spiral (part of crown),

MJ

Dōkunmae, Gunma

MJ

Tanabatake, Nagano

MJ

Imojiya, Yamanashi

2 JKAE37

1 2

Kaner 2009

3 (pp.92-93)

JKAN2

1

Kaner 2009

7 (p.99)

JKAN3

1

dogū, impressed spiral on back, ?snake representation

2

MJ MJ

Kaner 2009

8 (p.100)

JKAN4

1

dogū, impressed spiral on side ?representational dogū, incised spiral on hip

Kaner 2009

10 (pp.1023)

JKAN5

1

dogū, incised spiral on hip,

Tsuru High School, Yamanashi Ōne, Gumma

Mekiri, Nagano Nishinomae, Yamagata

sampling resolution

JKAE31

spacing trend error

Shirasuka, Yamagata

spacing median

MJ

path width trend error

jug, Daigi 8b type, h.14

path width median

1

irregularity

pl.204

JKAE30

curvilinearity

Kidder & Esaka

MJ

orientation

pl.203

1

place Shibahashi, Yamagata

rotation

Kidder & Esaka

JKAE29

Object notes bowl, Daigi 9 type, h.14.8

period*

Museum ID/ plate no. pl.202

spiral numbers analysed

Museum number/ publication Kidder & Esaka

Object number

Style data

8.94

0.99

0.87

0.19

2.14

144.03

0.39

0.46

0.04

-0.02

0.22

-0.48

0.34

-0.86

0.36

12.75

1.96

0.89

0.21

0.97

8.20

0.95

3.26

0.04

0.00

-0.02

0.13

0.00

0.05

-0.25

6.18

1.22

0.50

0.52

1.54

16.79

1.82

0.04

0.05

-0.45

1.31

0.00

0.00

0.00

7.79

2.64

0.81

0.32

0.70

2.80

1.76

0.13

0.04

0.00

0.01

-0.06

0.12

0.65

0.87

8.55

0.69

0.84

0.28

0.72

28.67

1.23

1.75

0.04

0.02

-0.20

0.61

-0.51

2.63

-2.98

10.98

1.03

0.82

0.17

1.21

74.62

0.63

2.20

0.04

0.00

-0.04

-0.02

-0.02

0.14

-0.13

8.66

0.01

0.65

0.34

1.28

30.90

0.73

0.20

0.03

0.03

-0.26

0.78

-0.13

0.63

-0.80

12.37

2.39

0.82

0.23

1.13

131.41

0.86

15.62

0.04

0.00

0.00

0.04

-0.02

0.18

-0.36

10.22

0.03

0.89

0.16

1.05

23.64

1.18

1.40

0.04

0.00

-0.04

0.18

-0.02

0.41

-1.36

7.86

0.32

0.87

0.08

1.77

131.62

1.39

0.77

0.04

-0.01

0.03

0.13

-0.05

-0.11

-0.21

9.92

2.22

0.29

0.20

1.29

77.69

1.67

17.44

0.04

0.01

-0.16

0.68

0.03

-0.23

0.24

8.14

1.69

0.85

0.25

1.69

47.26

1.87

0.17

0.04

0.01

-0.11

0.02

-0.35

0.58

0.53

6.71

1.46

0.82

0.20

1.10

11.69

2.37

0.01

0.04

0.00

0.03

-0.19

14.19

-9.12

0.80

9.52

2.56

0.87

0.24

1.11

9.40

1.06

4.45

0.04

0.00

-0.02

0.23

0.12

-0.06

-0.35

7.92

0.13

0.89

0.33

1.05

1.76

1.56

0.62

0.04

0.01

-0.11

0.22

0.03

1.15

-0.93

18.18

1.22

0.70

0.14

1.19

208.29

0.27

2.99

0.04

0.00

0.02

-0.18

0.00

0.01

-0.02

11.93

1.83

0.92

0.32

1.12

18.82

2.15

34.31

0.04

0.00

-0.03

0.25

0.08

-0.64

1.69

13.03

2.34

0.84

0.21

1.75

96.80

0.33

0.99

0.03

0.00

0.03

0.10

0.00

0.01

-0.04

9.21

1.43

0.80

0.59

0.72

9.92

2.77

4.63

0.04

0.00

0.02

-0.01

-0.45

2.78

-1.53

8.77

2.51

0.34

0.33

1.40

22.79

1.53

23.61

0.04

0.01

-0.12

0.54

-0.21

1.76

-1.80

6.83

0.44

0.82

0.32

2.49

34.00

3.94

0.04

0.01

-0.17

0.81

0.00

0.00

0.00

6.86

1.95

0.81

0.37

2.02

28.43

1.21

0.04

0.02

-0.28

1.01

-4.11

5.04

0.64

n/a

n/a 0.02

path width trend coefficients

spacing trend coefficients

278 Context data

1 (p.88)

JKAN8

1

dogū, incised spiral on back

LJ

Kaner 2009

21 (p.115)

JKAN9

1

dogū, incised spiral on back

LJ

Kaner 2009

22 (pp.1167)

JKAN10

Kaner 2009

25 (p.120)

JKAN11

Kaner 2009

61 (p.159)

JKAN12

1

dogū, incised spirals on arm ends

LJ

Nakappara, Nagano

1

dogū, spirals on chin described as tattoos

LJ

Toriitai 4, Aomori

1

heart-shaped dogū, spirals down back

LJ

Mother and child dogū, spirals on knees of mother, suggested to represent trousers (Kaner 2009 p.30) dogū carrying a child on its back, spiral roughly on back

MJ

Arakōji, Fukushima prefecture Miyata, Tokyo

pot with anthropomorphic decoration, spiral in head piece or hair, also spiral in isolation (not analysable) described as 'combining male and female imagery' though not clear if that is the spiral itself or the decoration on it. vessel for brewing/grain storage or more probably drum with figure engaging in ritual dance, spiral as ear of figure and as part of border around figure

MJ

cooking pot, spiral decoration around top, ?representational lamp, spiral on hair of face surmounting lamp and on front ?representational

MJ

2

2 Kaner 2009

57 (p.154)

JKAN13

1

Kaner 2009

55 (p.152)

JKAN14

1

Kaner 2009

43 (p.140)

JKAN15

1

Kaner 2009

42 (p.139)

JKAN16

1

2

Kaner 2009

Kaner 2009

41 (p.138)

40 (p.137)

Chobonaino, Hokkaido Ushiroda, Yamanashi

JKAN17

1

JKAN18

2 1 2

*EK – Early Jōmon, MJ – Middle Jōmon, LJ – Late Jōmon

MJ

MJ

MJ

Kamiyamada shell mound, Ishikawa prefecture Hayashi Ōji, Kanagawa prefecture

Imojiya, Yamanashi prefecture

Ōhinohara, Kanagawa prefecture Sori, Nagano prefecture

sampling resolution

Kaner 2009

Gōhara, Gunma

spacing trend error

LJ

2

dogū, incised spirals on shoulder spiral over sternum

spacing median

1

path width trend error

JKAN7

path width median

17 (pp.110111)

irregularity

Kaner 2009

curvilinearity

2

MJ

orientation

1

place Sakai, Yamanashi

rotation

JKAN6

Object notes dogū, incised spirals on hair/headdress dogū, incised spiral on hip

period*

spiral numbers analysed

Museum ID/ plate no. 11 (p.104)

Object number

Museum number/ publication Kaner 2009

Style data

12.19

0.03

0.32

0.05

1.93

151.57

0.69

1.23

0.04

-0.01

0.21

-0.85

0.02

-0.18

0.41

13.97

1.63

0.89

0.26

1.07

90.12

0.56

24.18

0.04

0.00

-0.04

0.18

0.01

-0.15

0.46

19.01

1.11

0.84

0.16

1.00

48.37

0.63

15.22

0.04

0.00

0.04

-0.32

0.00

0.01

-0.01

13.94

0.75

0.52

0.15

1.31

57.93

0.56

4.77

0.04

0.00

0.06

-0.31

0.00

0.04

-0.14

8.72

0.57

0.66

0.62

0.37

2.88

3.52

2.44

0.05

0.00

0.02

-0.01

-0.50

2.61

0.12

21.02

2.16

0.88

0.14

0.81

20.55

0.48

10.91

0.04

0.00

0.00

0.02

0.00

0.06

-0.34

22.52

1.78

0.84

0.10

0.77

24.14

0.38

9.95

0.04

0.00

0.01

-0.06

0.00

0.01

-0.09

22.99

1.18

0.85

0.09

0.67

77.25

0.44

12.30

0.04

0.00

0.02

-0.28

0.00

0.00

0.01

6.80

1.26

0.92

0.25

1.09

16.35

2.81

0.01

0.04

0.00

0.06

-0.22

7.97

-0.29

-2.10

20.25

0.99

0.83

0.18

0.62

28.91

0.76

47.20

0.05

0.00

0.00

-0.02

0.00

0.00

0.10

7.52

2.89

0.91

0.35

1.12

14.53

2.54

0.43

0.04

-0.01

0.12

-0.42

-3.18

7.10

-2.57

8.71

1.93

0.84

0.15

1.37

14.20

1.35

0.30

0.04

-0.01

0.10

-0.27

-0.23

0.92

-1.21

18.06

2.74

0.88

0.24

0.61

8.42

0.85

113.95

0.04

0.00

0.01

-0.04

0.01

-0.14

0.56

11.06

0.31

0.81

0.14

2.03

142.70

0.35

1.80

0.04

0.00

-0.01

0.29

-0.03

0.21

-0.40

11.66

0.74

0.88

0.23

1.45

223.82

0.46

1.48

0.04

0.01

-0.11

0.52

0.02

-0.12

0.13

7.54

1.87

0.88

0.30

2.25

70.13

0.65

4.46

0.04

-0.01

0.23

-0.88

4.63

-6.63

2.79

10.76 11.03

1.08 2.04

0.82 0.88

0.17 0.22

1.59 1.78

73.28 29.54

1.10 0.43

2.03 3.41

0.04 0.04

0.00 0.01

-0.05 -0.10

0.23 0.47

0.01 0.01

-0.07 -0.11

0.20 0.30

10.25 6.51

0.02 2.39

0.83 0.31

0.11 0.38

1.71 2.58

80.62 49.50

0.80 6.49

1.15 n/a

0.04 0.04

0.00 0.13

0.09 -1.13

-0.70 3.06

-0.02 0.00

0.13 0.00

-0.13 0.00

path width trend coefficients

spacing trend coefficients

279

Table A2.3 Contextual and style data for the Shang Dynasty/Western Zhou Period case study

1

5

CSASH

2

3 13

CSASH

3

59

CSASH

4

4

British Museum 1973,0726.13

food vessel (gui) food vessel (gui), representational (2.2 ?creature horn 2.3 creature head) food vessel (ding), 3.13 representational (creature tail) cooking vessel (ding) 4.4 representational (creature ?taotie whisker)

68 CSASH

5

6

CSASH

8

19

wine vessel (zhi)

S, Anyang

1

wine vessel (jue)

S, Anyang

23 CSASH

9

54 4

CSASH

10

CSASH

CSBM

11

1

S, Anyang

wine vessel (you)

4

7

WZ

S, Anyang

1 CSASH

WZ

wine vessel (jiao)

80 1

CSASH

WZ

S, Anyang

wine vessel (jue)

1

cooking vessel (ding) cooking vessel (ding), representational (11.1 taotie horn, 11.62 decoration on the horn) pou (ritual wine vessel), representational (taotie whisker)

2

representational (taotie horn)

9 1

Ashmolean Museum EA1956.855

62

S, Anyang S, Anyang

S, Anyang

S

1050

1050

1100

1200

1200

1200

1200

1200

1200

1200

1200

1300/ 1200

900

850

950

1050

1050

1050

1050

1050

1050

1050

1050

1100/ 1050

spacing median

path width trend error

image source**

1.03

19.13

1.12

4.53

0.04

0.00

0.08

-0.52

-0.01

0.11

-0.32

14.08

0.57

0.43

0.20

0.83

27.76

0.61

3.69

0.04

0.00

-0.02

0.08

0.00

0.02

-0.08

7.19

2.24

0.32

0.47

1.33

11.71

2.52

0.06

0.04

0.00

0.00

0.23

-1.77

4.04

0.25

8.80

1.37

0.76

0.31

1.48

21.48

0.63

1.34

0.04

0.01

-0.07

0.39

0.19

-0.51

0.28

10.47

2.39

0.47

0.28

1.11

16.30

1.53

11.06

0.04

0.00

-0.09

0.63

0.08

-0.50

0.87

8.67

2.58

0.63

0.24

1.82

17.67

1.11

0.57

0.04

0.00

-0.05

0.20

-0.19

0.60

-0.35

11.40

0.78

0.91

0.26

1.33

29.56

0.94

1.43

0.04

0.00

-0.01

0.06

0.08

-0.37

0.17

6.95

1.31

0.88

0.33

1.12

15.01

3.68

0.06

0.04

0.01

-0.17

0.74

4.58

-9.44

6.03

20.19

2.22

0.53

0.18

0.44

9.23

0.62

17.22

0.04

0.00

0.00

-0.04

0.00

0.02

-0.15

10.85

0.92

0.75

0.23

1.09

21.90

1.67

1.52

0.04

0.00

0.08

-0.34

0.03

-0.33

1.04

17.22

0.04

0.95

0.11

0.55

13.72

0.60

21.15

0.04

0.00

0.01

-0.02

0.00

-0.02

0.09

12.32

3.12

0.75

0.21

0.50

29.41

0.47

70.58

0.04

0.00

0.03

-0.13

0.09

-0.86

2.16

19.26

2.34

0.82

0.19

0.68

27.38

0.43

50.07

0.04

0.00

0.00

0.04

0.00

0.04

-0.14

China

11.46

2.68

0.80

0.30

1.24

76.75

0.60

2.18

0.03

0.00

-0.06

0.36

0.01

-0.08

0.10

China

14.29

2.30

0.82

0.19

0.84

11.38

0.86

11.66

0.04

0.00

0.00

0.03

0.00

0.00

0.01

7.24

2.32

0.83

0.29

2.07

9.39

1.99

0.08

0.03

0.03

-0.37

1.12

2.26

-3.28

0.78

22.28

2.56

0.85

0.16

0.48

13.75

0.50

74.26

0.04

0.00

0.00

0.03

0.00

-0.01

0.03

11.12

0.90

0.78

0.16

1.21

12.10

1.06

6.53

0.04

0.00

0.01

-0.01

-0.01

0.07

-0.04

12.35

2.09

0.42

0.27

1.28

10.34

0.96

3.73

0.04

0.00

-0.04

0.19

0.02

-0.16

0.42

8.15

2.20

0.36

0.51

0.60

3.44

3.02

0.69

0.05

0.01

-0.07

0.22

0.28

0.34

0.05

8.34

1.89

0.80

0.26

2.11

30.14

0.99

0.34

0.03

0.01

-0.15

0.97

0.17

-0.41

0.28

8.71

0.01

0.85

0.14

0.84

6.78

2.36

3.51

0.04

0.00

0.02

-0.02

0.49

-1.73

1.07

8.69

2.75

0.89

0.24

2.19

121.34

1.11

0.34

0.04

0.02

-0.23

1.11

-0.03

-0.02

0.54

China

China

China

China

China

China

China

China

China

China

b

path width median

0.17

irregularity

0.60

curvilinearity

2.36

orientation

11.02

rotation

end date

start date

sampling resolution

CSASH

3 Ashmolean Museum EA1956.3516 Ashmolean Museum EA1956.3517 Ashmolean Museum EA1956.872 Ashmolean Museum EA1956.838 Ashmolean Museum EA1956.833 Ashmolean Museum EA1956.3515 Ashmolean Museum EA1956.832

spacing trend error

Ashmolean Museum b7,LI1301.6

path width trend coefficients

place

1

2 Ashmolean Museum EA1956.830

Object notes

Style data

period*

Spiral number

Object number

Museum number Ashmolean Museum EA1956.854

Museum ID

Context data

spacing trend coefficients

280

British Museum 1940,1214.267

CSBM

2

2

British Museum 1980,0512.1

CSBM

3

62

CSBM

4

56 1

CSBM

5

3 1

British Museum 1968,0422.2

CSBM

6

2 25

British Museum 1954,0220.4 British Museum 1973,0726.5

CSBM

7

90 1

CSBM

8

2 46

British Museum 1953,1214.1

CSBM

9

69

1

British Museum 1977,0404.1 British Museum 1990,0807.1

CSBM

10

3

CSBM

11

1 14

British Museum 1983,0318.1 British Museum 1954,0511.1

1200

China

b

CSBM

12

84 62

CSBM

13

64

bronze ding bronze carved[?] ding (vessel). 4.25 representational (part of taotie horn), 10.5" high

S

bronze ding (vessel). bronze fang yi (container), [probably representational (taotie whiskers)], bronze zun (wine vessel). Representational (7.25 decoration on taotie face)

S, early Anyang

bronze he (vessel), W11.5cm at lip

S, late Anyang S, Transiti onal Period

bronze li (vessel), spirals are highly abstracted taotie, bronze gui (vessel) 'Kang Hou gui', reign of Cheng Wang, representational (10.1 part of bird, 10.3 animal ears)

S

S, Anyang

S

1200

1200 1300

1200

1200 1200

1400

1050

1100 1100

1050

1100 1050

1200

China

Anyang, Henan China

China

China

1000

China

bronze gu (vessel) bronze hu (vessel), 12.84 representational (taotie nostril)

early S

1500

1300

China

bronze ding (vessel)

late S, Anyang

1200

1100 1000

b

China

1100

1200

b

China

early WZ

S, Anyang

b

b

b

China China

b

spacing median

path width trend error

0.30

2.28

18.81

0.47

0.04

0.03

-0.03

0.38

-0.67

0.87

-1.37

0.43

9.75

0.47

0.68

0.45

1.82

17.98

2.38

8.18

0.04

0.00

-0.03

0.46

-0.26

1.94

-1.99

31.62

0.61

0.58

0.08

0.40

11.19

1.23

98.00

0.04

0.00

0.00

0.01

0.00

0.01

-0.08

26.77

0.82

0.69

0.16

0.41

7.26

0.42

26.98

0.04

0.00

0.00

-0.03

0.00

-0.02

0.13

16.37

0.55

0.73

0.24

1.45

65.02

0.42

3.71

0.04

0.00

-0.05

0.41

0.00

0.01

-0.09

18.39

2.71

0.45

0.20

0.32

2031.2 6

0.24

41.81

0.04

0.00

-0.02

0.40

0.00

-0.05

0.23

9.10

2.31

0.83

0.23

1.23

28.10

0.79

0.31

0.03

0.01

-0.11

0.42

-0.04

0.35

-0.69

10.40

0.78

0.81

0.28

1.54

13.70

0.73

3.02

0.03

0.01

-0.09

0.42

0.02

-0.18

0.31

8.90

0.53

0.81

0.34

2.25

69.32

0.79

1.23

0.03

0.00

-0.06

0.66

0.16

-0.65

0.83

8.91

2.54

0.73

0.32

2.09

131.31

0.76

2.52

0.03

0.00

0.00

0.36

0.00

-0.18

0.74

8.48

2.03

0.75

0.27

2.59

18.51

0.69

0.86

0.03

0.00

0.03

0.28

0.25

-1.17

1.41

9.81

0.44

0.62

0.22

1.01

53.96

1.45

9.78

0.04

0.00

-0.04

0.25

0.04

0.00

-0.29

16.29

0.95

0.10

0.21

2.35

93.90

0.66

18.31

0.03

0.00

0.00

0.39

0.00

0.04

0.01

14.36

2.41

0.44

0.14

0.74

12.73

0.68

25.47

0.04

0.00

-0.01

0.07

0.00

0.00

-0.06

13.40

2.17

0.77

0.16

1.26

113.64

0.57

12.30

0.04

0.00

-0.03

0.13

0.00

-0.01

-0.07

7.47

2.29

0.87

0.13

1.51

52.50

1.38

0.74

0.04

0.00

-0.05

0.24

-0.53

2.33

-2.49

13.18

1.54

0.83

0.42

0.77

26.94

1.16

15.05

0.04

0.00

-0.02

0.14

0.00

0.18

-0.34

8.80

0.68

0.81

0.26

2.10

117.50

1.07

1.05

0.04

0.00

-0.04

0.64

0.26

-0.84

0.76

8.18

2.63

0.81

0.14

0.70

5.60

2.80

0.11

0.04

-0.01

0.08

-0.20

-0.68

1.41

0.39

7.52

0.85

0.87

0.19

1.58

16.71

0.88

0.25

0.04

0.00

-0.05

0.21

0.04

0.57

-0.80

14.15

2.31

0.61

0.32

0.93

23.23

0.42

3.72

0.04

0.01

-0.09

0.39

0.01

-0.14

0.38

12.83

1.24

0.94

0.01

0.42

12.78

0.34

3.61

0.04

0.00

0.00

-0.01

0.00

0.04

-0.13

9.01

1.96

0.96

0.01

0.43

11.59

0.53

6.84

0.04

0.00

-0.02

0.12

-0.24

0.95

-1.18

path width median

0.79

irregularity

0.30

curvilinearity

7.48

orientation

path width trend coefficients

rotation

image source**

end date

start date 1500

18 25

British Museum 1947,0712.419

S

sampling resolution

British Museum 1959,1020.1

place

spacing trend error

1

Object notes bronze jia (vessel), prob representational (part of a ?dragon), representational (creature ?dragon tail)

Style data

period*

Spiral number

Object number

Museum number

Museum ID

Context data

spacing trend coefficients

281

CSBM

14

41

1 British Museum 1936,1118.2

CSBM

15

2 1

British Museum 1936,1118.4

CSBM

16

79

2

British Museum 1983,0202.1 British Museum 1973,0726.1

CSBM

17

9 24

CSBM

18

56

3 British Museum 1956,1016.1 British Museum 1947,0712.429

CSBM

19

22 1

CSBM

20

2

1

British Museum 1936,1118.1

CSBM

21

37

gui 'Xing Hou gui', ?representational - snake - based on resemblance to Jomon 'snake' motifs] bronze you (vessel), Leiwen, 16.1 representational (bird wing) bronze you, (spirals quite different from earlier vessels). Representational (snakes), sampled at low resolution bronze fang yi (vessel) hu (vessel), H11.5" (29.2cm), ?representational (taotie facial decoration) ?representational (decoration on taotie horn) cast bronze he (alcohol vessel), H19cm cast bronze zun, double ram vessel. Representational (.1 - ram's shoulders, .37 taotie facial decoration 'faces being executed in thick scrolling reminiscent of style II')

S, early Anyang

WZ (earlymid)

S, Anyang

EZ or midlate WZ S, Anyang

S early WZ

S

1050

1200

800 1200

1200

1050

1300

1200

771

1050

700 1100

1100

900

1100

China

China

China

b

South or SW China China

b

China

China

Hunan Province

b

spacing median

path width trend error

0.47

1.91

47.47

0.63

21.52

0.04

0.00

0.05

0.08

-0.02

0.64

-1.38

11.72

2.61

0.84

0.37

2.35

69.59

0.31

2.78

0.03

0.00

0.04

0.20

0.05

-0.33

0.52

7.76

0.60

0.56

0.36

0.42

7.38

3.84

1.62

0.04

0.00

0.00

0.01

2.87

-6.85

4.58

20.56

1.46

0.74

0.13

0.41

153.01

0.27

52.26

0.04

0.00

0.00

0.00

0.00

0.00

0.02

9.34

0.53

0.75

0.52

2.08

63.45

0.16

35.28

0.04

0.01

-0.16

0.89

0.00

0.05

-0.32

31.33

0.86

0.59

0.11

0.50

10.68

0.21

6.47

0.04

0.00

0.00

-0.02

0.00

0.00

-0.03

23.12

0.39

0.83

0.15

0.98

47.04

0.27

8.21

0.04

0.00

-0.01

0.10

0.00

-0.01

0.07

33.79

1.03

0.85

0.10

0.32

6.59

0.60

43.65

0.04

0.00

0.00

0.00

0.00

0.00

0.01

16.91

2.33

0.70

0.15

0.95

16.39

0.67

12.45

0.04

0.00

0.00

-0.01

0.00

0.01

-0.06

8.79

2.40

0.53

0.34

2.55

32.08

0.56

0.30

0.03

-0.01

0.06

0.26

-0.05

0.06

0.43

13.04

2.87

0.84

0.19

1.18

362.91

0.36

2.31

0.04

0.00

-0.03

0.12

0.00

0.03

-0.08

10.98

2.60

0.79

0.17

1.56

83.13

0.66

3.16

0.04

0.00

-0.03

0.19

0.03

-0.23

0.42

9.24

0.02

0.83

0.18

1.31

13.88

1.59

1.51

0.04

-0.01

0.06

-0.09

-0.46

2.02

-1.88

8.79

2.69

0.67

0.01

0.80

23.78

0.95

0.26

0.04

-0.01

0.06

-0.08

-0.01

0.19

-0.61

12.31

1.15

0.03

0.32

1.62

57.67

0.41

11.29

0.03

0.00

0.00

0.18

0.07

-0.40

0.65

11.02

2.34

0.82

0.19

2.30

170.38

0.31

2.11

0.04

0.01

-0.12

0.72

0.00

-0.03

0.18

path width median

0.50

irregularity

0.68

curvilinearity

11.19

orientation

path width trend coefficients

rotation

image source**

end date

start date 1300

sampling resolution

British Museum 1986,0417.1

bronze zun (vessel) representational (taotie horn) representational (taotie ear)

place

spacing trend error

16

Object notes

Style data

period*

Spiral number

Object number

Museum number

Museum ID

Context data

spacing trend coefficients

282

British Museum 1973,0726.9 British Museum 1973,0726.7 British Museum 1959,0512.1

CSBM

22

3 1

CSBM

23

2

British Museum 1980,0512.2

CSBM CSBM

24 25

38

bronze gu, H25.7

1

bronze jue (vessel), H19.5cm

7

CSBM

26

78 1

CSBM

27

2 14

British Museum 1957,0221.1

CSBM

28

99 14

British Museum 1953,0714.1

CSBM

29

37

46

British Museum 1945, 1017.191 National Museums Scotland A.1934.616 Victoria and Albert Museum FE.156-1988 Victoria and Albert Museum M.3-1950

S/WZ

1200

771 1050

China China

b b

CSBM

30

72

bronze gui (vessel), representational (.17 cicada, .78 creature ?dragon tail) bronze zun (vessel) bronze gui (ritual vessel), .14 ?representational (decoration on taotie) bronze you (wine vessel), .14 ?representational (decoration on taotie) bronze jia (vessel), .46 ?representational (cicada), .72 representational (taotie facial decoration)

S S/WZ

1200 1100

S

1200

late S

1200 (c.)

S

late S

late S

1200

1200

1200

1050 1001

1050

China China

Anyang, Henan China

1100

1000

1050

China

China

China

b b

b b

b

b

b

11 CSNMS

1

12 1

CSVAM

1

2 3

CSVAM

2

4

bronze ding goblet (gu?), 32cm high bronze jue, taotie, 21cm high

late S

S

S

1200

1150

1100

1000

1050

1000

Anyang, Henan

China

China

h

c

c

spacing median

path width trend error

0.39

2.31

10.36

1.28

0.13

0.04

0.01

-0.09

0.73

-0.21

1.41

-0.63

9.72

0.54

0.92

0.10

1.12

32.26

2.25

1.97

0.04

-0.01

0.12

-0.61

-0.02

-0.21

0.90

14.14

2.32

0.74

0.14

0.56

10.45

1.33

16.83

0.04

0.00

-0.02

0.15

0.01

-0.09

0.34

13.95

2.29

0.47

0.18

0.64

5.50

1.03

11.98

0.04

0.00

0.02

-0.08

0.02

-0.19

0.63

14.76

2.53

0.45

0.24

0.85

67.01

0.65

3.11

0.04

0.00

-0.04

0.23

0.00

-0.02

0.10

18.31

2.52

0.70

0.23

0.43

7.53

0.80

16.99

0.04

0.00

-0.01

0.08

0.00

-0.01

0.03

9.32

1.12

0.83

0.03

0.55

130.50

0.32

5.28

0.04

0.01

-0.17

0.64

-0.08

0.53

-1.04

26.62

1.49

0.72

0.15

0.49

9.60

0.48

17.65

0.04

0.00

0.00

0.01

0.00

-0.01

0.04

12.96

2.41

0.83

0.22

1.08

15.86

0.69

7.37

0.04

0.00

0.01

-0.05

0.01

-0.08

0.25

10.94

2.32

0.63

0.21

1.20

51.80

1.69

33.72

0.04

0.00

-0.01

0.30

0.06

-0.37

0.61

9.23

0.69

0.77

0.19

0.80

7.84

1.40

8.02

0.04

0.00

0.02

-0.03

0.14

-0.29

0.04

12.23

0.67

0.33

0.19

0.75

25.99

1.16

24.42

0.04

0.00

-0.01

0.07

0.00

0.06

-0.26

12.69

2.24

0.88

0.16

1.02

74.50

0.47

11.11

0.04

0.00

0.01

-0.08

-0.03

0.28

-0.91

13.87

0.60

0.67

0.13

0.88

13.01

0.76

5.81

0.04

0.00

0.00

0.00

0.01

-0.10

0.30

7.32

2.25

0.82

0.28

1.35

116.16

2.57

0.11

0.03

-0.05

0.52

-1.19

1.74

-2.09

0.32

13.25

0.83

0.76

0.13

1.19

89.88

0.54

2.28

0.04

0.00

0.01

-0.08

0.00

0.03

0.02

26.80

0.60

0.89

0.06

0.56

18.75

0.31

13.37

0.04

0.00

0.00

-0.05

0.00

0.00

-0.01

9.79

0.95

0.78

0.36

1.29

28.82

0.96

0.46

0.03

0.01

-0.12

0.38

0.01

-0.16

0.59

17.31

2.33

0.38

0.17

0.77

56.79

0.61

9.78

0.04

0.00

0.00

0.04

0.00

0.00

0.02

10.27

2.40

0.29

0.14

1.73

71.97

1.13

3.33

0.03

0.00

-0.08

0.59

0.05

-0.52

1.30

36.04

0.79

0.10

0.06

0.42

18.28

0.18

5.64

0.04

0.00

0.00

-0.04

0.00

0.00

0.00

20.83

0.86

0.51

0.15

0.76

14.28

1.76

41.18

0.04

0.00

0.01

-0.09

0.01

-0.08

0.18

10.84

2.60

0.38

0.11

0.90

24.18

1.33

19.55

0.04

0.00

-0.04

0.21

-0.09

0.64

-0.98

15.65

2.49

0.71

0.15

0.67

7.24

0.86

30.41

0.04

0.00

0.01

-0.03

0.00

-0.01

0.07

path width median

0.88

irregularity

0.86

curvilinearity

7.78

orientation

path width trend coefficients

rotation

image source**

end date

start date 1050

24

17 British Museum 1947, 0712.420

bronze li (vessel), H17.5cm

early WZ

sampling resolution

British Museum 1952,1216.1

place

spacing trend error

1

Object notes bronze pan (vessel), representational (.1 tail of water creature, .3 ear of long - dragon)

Style data

period*

Spiral number

Object number

Museum number

Museum ID

Context data

spacing trend coefficients

283

CSVAM

3

1 CSVAM

4

2 4

CSVAM

5

5 1

CSVAM

6

2 4

CSVAM

7

8 1

CSVAM

8

2 16

Victoria and Albert Museum M.136-1956

CSVAM

9

32

bronze li, 26cm high bronze pou/weng?, representational (decoration on a taotie head) bronze zun (wine jar), 28.7cm high bronze ding, 20.7cm high bronze fangzun (wine vessel) 45.7cm high bronze grain vessel, 9.16 representational (creature ?dragon ear)

S

S

S

S

S

S

S

1300

1400

1450

1300

1100

1100

1100

1100

China

1200

?Shaanx i region

1350

1050

1000

1000

1000

*S – Shang Dynasty, WZ – Western Zhou Period, EZ – Eastern Zhou Period ** image source codes refer to those used on the List of Figures

China

China

China

China

China

c

c

c

c

c

c

c

spacing median

path width trend error

0.17

0.59

9.68

1.24

27.09

0.04

0.00

-0.01

0.03

0.00

0.02

-0.14

9.14

0.78

0.47

0.21

1.48

17.04

1.12

1.99

0.04

0.00

0.00

-0.03

0.18

-0.89

1.03

11.39

2.29

0.77

0.31

1.60

18.96

0.37

5.94

0.04

0.00

0.04

-0.04

0.02

-0.11

0.26

11.30

0.86

0.49

0.18

1.47

24.16

0.43

1.81

0.04

0.00

0.01

0.11

0.00

0.07

-0.30

10.65

2.03

0.79

0.27

0.90

10.96

1.68

5.07

0.04

0.00

0.06

-0.26

-0.06

0.54

-0.83

7.77

0.74

0.85

0.15

1.14

8.89

2.15

0.53

0.04

-0.01

0.07

-0.12

0.77

-0.44

-1.98

13.39

2.45

0.87

0.17

0.92

42.54

2.10

0.01

0.04

0.00

0.00

0.07

-1.60

3.63

-1.80

8.97

2.50

0.91

0.24

1.29

10.48

1.58

0.51

0.04

0.00

0.09

-0.31

0.03

0.12

-0.52

22.69

2.54

0.80

0.15

0.72

7.10

1.31

15.50

0.04

0.00

0.00

-0.04

0.00

-0.04

0.13

21.40

0.50

0.84

0.17

0.71

9.89

0.47

6.12

0.04

0.00

0.00

0.01

0.00

0.00

-0.01

9.95

2.56

0.74

0.28

1.69

35.09

1.31

3.48

0.03

0.00

0.01

0.48

0.08

-0.43

0.80

10.82

2.38

0.70

0.16

1.16

10.32

1.56

1.12

0.04

0.00

-0.04

0.17

0.01

-0.12

0.11

9.57

2.47

0.56

0.39

1.78

26.30

1.05

10.81

0.04

0.00

0.08

-0.14

0.38

-1.32

1.28

6.21

0.65

0.89

0.36

2.15

4.55

2.92

n/a

0.04

-0.01

0.12

-0.38

0.00

0.00

0.00

path width median

0.70

irregularity

2.38

curvilinearity

16.95

orientation

path width trend coefficients

rotation

image source**

end date

start date

period*

Spiral number 2

place

sampling resolution

Victoria and Albert Museum M.1163:1, 21926 Victoria and Albert Museum M.4-1935 Victoria and Albert Museum M.2696-1931 Victoria and Albert Museum 195-1876

1

Object notes li ding, 3 taotie and cicadas on spiral ground, 19cm

Style data spacing trend error

Victoria and Albert Museum M.60-1953 Victoria and Albert Museum M.3-1935

Object number

Museum number

Museum ID

Context data

spacing trend coefficients

284

Table A2.4 Context and style data for the Egyptian case study

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

material

spacing trend coefficients

curvilinearity

place

path width trend coefficients

orientation***

date*

image source**

spiral numbers analysed

Object notes

Style data

rotation

British Museum E69522 British Museum 30551/ 1899, 0311.57 British Museum E66159/ 1960, 0514.53 British Museum 63182/ 1930, 0711.202 British Museum 63185/ 1930, 0711.205 British Museum 63150/ 1930, 0711.170 Oriental Museum, Durham EG2939 Oriental Museum, Durham EG3047 Oriental Museum, Durham EG6007 Oriental Museum, Durham EG6007 Oriental Museum, Durham EG6007

Object number

Museum number

Museum ID

Context data

8.82

1.36

0.40

0.11

0.77

6.02

1.90

0.48

0.04

0.00

0.02

-0.11

-0.04

0.45

-1.12

8.64

1.75

0.75

0.12

0.74

17.01

2.08

0.87

0.04

0.01

-0.10

0.29

-0.17

1.05

-2.39

7.53

1.39

0.46

0.32

1.30

43.57

2.99

0.51

0.04

-0.01

0.05

0.02

-0.03

0.84

-0.26

6.36

1.79

0.84

0.25

1.76

23.79

1.63

0.00

0.04

-0.03

0.18

-0.04

0.00

0.00

0.00

8.17

1.85

0.65

0.25

1.34

38.51

1.71

0.22

0.04

0.00

-0.05

0.16

0.59

-0.37

-1.46

8.88

1.48

0.66

0.17

1.20

23.30

1.22

1.11

0.04

0.00

-0.03

0.07

-0.03

0.74

-2.07

12.07

1.44

0.85

0.36

0.84

25.69

0.59

53.18

0.04

0.00

-0.04

0.30

0.02

0.13

-0.61

12.53

1.51

0.83

0.38

0.91

22.56

0.46

50.62

0.04

0.00

-0.02

0.16

0.02

0.10

-0.46

6.36

1.80

0.61

0.52

0.91

7.65

7.34

n/a

0.04

-0.01

0.15

-0.38

0.00

0.00

0.00

8.44

0.57

0.75

0.32

0.85

9.33

1.97

2.89

0.04

0.00

-0.02

0.06

0.67

-0.91

-0.18

12.48

1.69

0.39

0.23

0.92

26.54

0.89

8.52

0.04

0.00

-0.04

0.24

0.01

-0.12

0.46

13.28

1.71

0.89

0.26

1.05

22.37

0.67

18.05

0.04

0.00

0.01

-0.02

0.04

-0.36

0.92

11.25

0.82

0.73

0.17

0.86

34.98

2.23

31.82

0.04

0.00

0.00

-0.02

-0.09

0.73

-1.78

7.12

0.04

0.90

0.07

1.16

25.21

3.36

0.25

0.04

0.00

0.04

-0.21

4.40

-5.40

-0.28

13.26

3.08

0.88

0.25

0.65

67.62

0.64

106.06

0.04

0.00

-0.01

0.06

-0.01

0.27

-0.78

11.88

2.69

0.87

0.14

1.31

66.93

0.85

29.88

0.04

0.00

0.01

-0.14

0.05

-0.46

1.19

11.46

1.55

0.40

0.34

0.82

11.38

0.80

17.22

0.04

0.00

0.01

-0.04

0.06

-0.17

0.03

12.08

1.55

0.94

0.32

0.90

30.97

1.10

17.90

0.04

0.00

-0.06

0.34

0.07

-0.38

0.32

6.28

2.26

0.94

0.47

2.02

12.55

4.14

0.01

0.04

0.01

-0.14

0.61

55.89

-18.81

5.34

6.08

0.58

0.97

0.24

2.16

48.03

3.08

n/a

0.03

0.00

-0.04

0.28

0.00

0.00

0.00

6.87

1.81

0.62

0.43

1.10

8.73

3.33

0.08

0.04

0.01

-0.01

-0.03

6.99

-5.73

3.27

6.69

0.71

0.86

0.52

1.14

5.46

3.79

0.07

0.05

0.02

-0.21

0.66

13.07

-5.40

4.18

1 EGBM

1

10

2.29cm long

12D

Egypt

glazed steatite

b

6

EGBM

2

9 3

EGBM

3

8

25mm long appears to be representational (loops within snakes tails) 2.03cm long

13D

late 12D

Egypt

Egypt

steatite

glazed steatite

b

b

1

EGBM

4

2

representational (Hathor head), 15.1mm long.

1I/ early MK

Mostagedda

glazed steatite

b

2

EGBM

5

4

1.4cm long

1I

Mostagedda

glazed steatite

b

1

EGBM

6

2

1.15 cm long

1I

Mostagedda

glazed steatite

b

2 EGDUR

1

4

MK

Egypt

glazed steatite

1 EGDUR

2

2 1

EGDUR

3

2 6

EGDUR

4

9 7

EGDUR

6

11

Wellcome Collection Wellcome Collection, necklace of scarabs Wellcome Collection, necklace of scarabs Wellcome Collection, necklace of scarabs

MK

MK

MK

MK

Egypt

Egypt

Egypt

Egypt

steatite

steatite

steatite

steatite

285

curvilinearity

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

material

spacing trend coefficients

orientation***

place

image source**

spiral numbers analysed

date*

path width trend coefficients

11.75

1.92

0.88

0.30

0.81

34.56

0.44

11.29

0.04

0.00

-0.02

0.11

0.09

-0.48

0.60

15.73

2.30

0.51

0.29

0.96

38.83

0.49

33.79

0.04

0.00

0.01

-0.05

0.02

-0.20

0.54

5.67

1.27

0.91

0.54

1.19

4.74

6.14

n/a

0.04

0.02

-0.17

0.48

0.00

0.00

0.00

5.86

1.17

0.96

0.43

1.25

5.78

4.69

n/a

0.04

0.02

-0.14

0.31

0.00

0.00

0.00

8.69

0.47

0.44

0.26

0.88

32.58

2.19

2.70

0.04

0.00

0.02

-0.06

0.64

-1.59

0.67

6.73

0.60

0.33

0.37

1.18

13.70

3.39

0.03

0.04

0.01

-0.12

0.38

1.98

-0.21

2.47

9.46

2.30

0.84

0.06

1.31

27.26

1.40

2.70

0.04

-0.01

0.15

-0.73

0.17

-0.89

1.10

7.52

1.12

0.93

0.31

1.03

23.39

2.62

0.24

0.04

0.00

0.04

-0.10

1.89

-3.14

2.00

7.33

1.52

0.39

0.25

1.31

7.06

2.38

0.11

0.04

0.00

-0.01

0.01

1.12

-0.47

-0.50

6.48

2.40

0.80

0.59

1.56

4.82

4.27

0.01

0.04

0.03

-0.33

1.05

22.39

-7.87

3.89

7.12

1.88

0.32

0.37

1.70

17.03

2.30

0.03

0.04

0.01

-0.19

0.83

0.73

-0.18

0.23

6.38

2.38

0.83

0.34

1.69

104.96

3.08

0.00

0.04

0.01

-0.10

0.12

-67.26

20.44

-0.61

5.35

1.59

0.94

0.39

1.79

17.35

5.35

n/a

0.04

0.06

-0.49

1.27

0.00

0.00

0.00

16.75

1.88

0.90

0.11

1.15

35.65

0.48

15.05

0.04

0.00

0.00

0.04

0.00

-0.01

0.00

15.31

1.06

0.93

0.14

1.13

57.67

0.72

88.09

0.04

0.00

0.01

0.01

-0.01

0.09

-0.43

13.91

2.87

0.53

0.35

0.74

22.74

0.82

9.35

0.04

0.00

-0.01

0.04

0.03

-0.24

0.39

6.09

0.54

0.82

0.35

0.93

14.11

4.24

n/a

0.04

0.02

-0.20

0.46

0.00

0.00

0.00

1 EGLIV

21

3

Spurrell

18D on

Egypt

2 EGLIV

22

4

Spurrell

18D on

Egypt

3 EGLIV

EGLIV

24

26

16

1 1

EGLIV

27

2 4

EGLIV

28

11

Spurrell

Spurrell Meyer Collection published Petrie plate 20AE, cartouche Kheperkzie *assumed from 'MK' on catalogue card.

18D on

18D on

2I

MK*

Egypt

Egypt

Egypt

[bone/ ivory?]

[?steatite]

[glazed probably faience]

Egypt

4 EGLIV

29

7 1

World Museum, Liverpool 23.8.10.1 National Museums Scotland A.1965. 146

Object notes

Style data

rotation

World Museum, Liverpool M13737 World Museum, Liverpool M12474 World Museum, Liverpool 56.20.417

Object number

Museum number World Museum, Liverpool 56.20.428 World Museum, Liverpool 56.20.431 World Museum, Liverpool 56.20.423 World Museum, Liverpool 56.20.433

Museum ID

Context data

EGLIV

33

4

Spurrell ring, found reused in New Kingdom tomb. Petrie & Mackay 1910, pl. XXVII, p.36) Meydum and Memphis (III). London: British School of Archaeology in Egypt & Egyptian Research Account 16th Year.

18D on

Hyksos period (Petrie & Mackay 1910)

Egypt

Meydum

glazed faience

?

7

EGNMS

1

8

0.81" long (20.6mm)

MK

Egypt

glazed steatite

h

286

Petrie Museum UC11165 Petrie Museum UC11154

image source**

spiral numbers analysed

sampling resolution

Petrie Museum UC11173

spacing trend error

Petrie Museum UC29073

spacing median

Petrie Museum UC29070

path width trend error

Petrie Museum UC29061

path width median

Petrie Museum UC26118

material

irregularity

Petrie Museum UC18125

place

spacing trend coefficients

curvilinearity

Petrie Museum UC28372

date*

path width trend coefficients

orientation***

Petrie Museum UC11150

Object notes

Style data

rotation

Petrie Museum UC11131

Object number

Museum number National Museums Scotland A.1921. 1092

Museum ID

Context data

10.77

1.59

0.63

0.24

0.87

35.11

1.50

23.51

0.04

0.00

-0.02

0.12

0.12

-0.70

1.30

11.58

1.30

0.36

0.31

1.06

29.31

1.27

10.63

0.04

0.00

-0.01

0.05

0.04

-0.27

0.62

5.23

1.92

0.89

0.34

1.71

78.06

6.11

n/a

0.04

-0.06

0.48

-0.90

0.00

0.00

0.00

7.81

2.45

0.72

0.46

0.85

26.68

3.15

3.09

0.04

0.01

-0.06

0.20

-0.03

0.74

1.83

7.87

1.55

0.92

0.21

1.37

49.14

2.34

2.54

0.04

0.01

-0.10

0.34

0.44

-0.64

-0.12

7.65

2.93

0.61

0.05

1.06

23.63

3.13

0.12

0.04

0.00

0.04

-0.21

1.58

-2.85

-0.68

8.87

0.01

0.90

0.40

1.31

13.64

1.52

8.99

0.04

0.00

-0.08

0.38

0.02

0.85

-0.65

8.52

0.32

0.94

0.42

1.28

23.96

1.34

7.44

0.04

0.00

-0.04

0.27

0.16

0.67

-0.13

8.37

2.07

0.83

0.11

1.09

46.64

1.07

9.98

0.04

-0.01

0.18

-0.84

0.57

-1.39

0.55

8.13

0.92

0.73

0.11

0.95

12.81

1.52

3.78

0.04

-0.02

0.29

-1.26

-0.65

1.48

-0.46

8.86

1.56

0.85

0.24

0.82

18.46

1.90

1.76

0.04

0.00

0.00

-0.04

-0.27

1.99

-3.33

8.89

1.67

0.82

0.22

0.78

4.53

1.84

0.99

0.04

0.00

-0.05

0.14

-0.09

1.16

-2.77

11.96

2.44

0.81

0.27

0.95

21.26

0.89

6.49

0.04

0.00

0.01

-0.11

0.07

-0.39

0.48

11.10

0.77

0.69

0.30

1.13

18.69

0.58

10.71

0.04

0.00

0.00

-0.04

0.05

-0.08

-0.08

12.62

1.81

0.93

0.25

1.09

26.68

0.76

7.14

0.04

0.00

-0.04

0.26

0.04

-0.29

0.55

6.27

1.80

0.53

0.33

1.66

21.25

4.01

n/a

0.04

0.05

-0.50

1.41

0.00

0.00

0.00

6.49

1.83

0.89

0.29

1.10

21.19

2.82

0.00

0.04

0.02

-0.31

1.07

22.04

-2.47

0.32

6.89

0.03

0.34

0.29

1.45

16.07

2.96

0.02

0.04

0.00

-0.09

0.45

6.65

-2.51

-0.09

11.45

1.66

0.87

0.28

1.41

18.02

0.78

13.88

0.04

0.00

-0.01

0.08

0.06

-0.41

0.92

8.20

1.42

0.56

0.28

1.42

21.73

1.74

1.67

0.04

0.00

-0.03

0.18

1.34

-3.20

2.51

9.84

1.53

0.37

0.40

0.37

6.96

5.52

11.11

0.04

0.00

0.02

-0.17

-0.96

4.19

-1.69

10.64

1.51

0.78

0.54

0.39

2.76

3.68

14.08

0.04

0.00

0.00

-0.03

-0.37

2.53

-2.51

7.80

0.82

0.81

0.37

1.08

3.88

2.22

0.26

0.04

0.01

-0.06

0.22

1.69

-2.78

1.25

8.46

2.34

0.76

0.33

1.21

23.13

1.92

0.39

0.04

0.01

-0.08

0.30

0.77

-1.84

1.05

1

EGNMS

2

2

representational (Hathor head), 0.5"long (12.7mm)

MK

Egypt

glazed steatite

h

5 EGPET

1

6

Tomb 1653, 1.5cm long

MK

Sedment

glazed steatite

e

19 EGPET

2

25

2.1cm long

late MK

Lahun

faience

e

1 EGPET

3

2

1.8cm long

12D

Lahun

not stated

e

1 EGPET

4

2

?representational (papyrus)

9D

Qau

not stated

e

2 EGPET

5

3 1

EGPET

6

2

2.7cm long, 0.9cm wide 1.35x0.9cm, ?representational ('lotus and scroll')

MK

MK

Koptos

Nubt LXXX, 71

obsidian

glazed steatite

e

e

5 EGPET

7

6

1.3x0.9cm

MK

Nubt LXXXI, 82

steatite

e

2 EGPET

8

3

2.2x1.6cm

MK

Nubt LXXXI, 85

steatite

e

1 EGPET

9

2

1.8x1.3cm

late MK

Lahun

steatite

e

1 EGPET

10

2

1.7x1.2cm

late MK

Lahun

glazed steatite

e

3 EGPET

11

9

1.8x1.2cm

late MK

Lahun

steatite

e

287

Petrie Museum UC11158 Petrie Museum UC11161 Petrie Museum UC11164

image source**

spiral numbers analysed

sampling resolution

Petrie Museum UC7550

spacing trend error

Petrie Museum UC11354

spacing median

Petrie Museum UC11172

path width trend error

Petrie Museum UC11170

path width median

Petrie Museum UC11151

material

irregularity

Petrie Museum UC11166

place

spacing trend coefficients

curvilinearity

Petrie Museum UC11167

date*

path width trend coefficients

orientation***

Petrie Museum UC11171

Object notes

Style data

rotation

Petrie Museum UC61187

Object number

Museum number

Museum ID

Context data

9.97

0.81

0.65

0.25

1.13

28.41

1.63

9.69

0.04

0.00

-0.02

0.21

0.04

-0.05

-0.06

7.91

0.76

0.84

0.35

1.32

39.39

1.78

0.60

0.04

0.00

-0.05

0.19

-0.20

1.99

-1.62

10.56

0.26

0.90

0.25

1.26

29.69

0.98

4.06

0.04

0.01

-0.08

0.26

0.16

-0.98

1.49

7.52

0.14

0.92

0.30

1.67

33.46

1.01

0.11

0.04

0.02

-0.22

0.57

0.75

0.64

-1.02

7.05

1.22

0.71

0.46

0.70

4.36

4.44

0.08

0.04

0.00

-0.02

0.26

2.16

-2.29

4.27

5.68

1.01

0.88

0.50

0.92

9.57

4.36

n/a

0.04

-0.03

0.24

-0.52

0.00

0.00

0.00

6.70

1.92

0.96

0.29

1.12

36.53

3.34

0.07

0.04

0.02

-0.17

0.28

-1.26

1.73

4.00

7.93

2.00

0.84

0.39

0.89

16.99

2.49

0.61

0.04

0.01

-0.09

0.31

1.32

-2.50

2.30

7.25

2.65

0.91

0.25

1.05

72.13

3.01

0.07

0.04

-0.03

0.31

-0.60

3.74

-8.08

5.34

6.92

2.83

0.96

0.30

1.63

43.16

2.48

0.18

0.04

0.01

-0.09

0.22

-4.75

4.88

-1.09

14.90

1.76

0.85

0.19

1.18

37.27

0.40

17.40

0.04

0.00

0.00

-0.07

0.01

-0.07

0.18

16.35

2.24

0.85

0.20

0.93

20.08

0.39

25.03

0.04

0.00

0.02

-0.17

0.01

-0.14

0.45

7.06

0.82

0.93

0.40

1.37

7.35

2.63

0.04

0.04

-0.01

0.05

0.11

-0.47

3.33

-0.20

11.34

0.42

0.88

0.19

1.80

24.02

0.37

3.47

0.03

0.00

-0.01

0.25

0.05

-0.37

0.72

8.07

2.16

0.83

0.16

0.89

5.09

2.66

0.24

0.04

0.00

0.05

-0.16

1.01

-2.12

-0.62

8.48

1.73

0.74

0.17

0.81

2.87

2.25

4.40

0.04

0.00

-0.03

0.00

0.07

0.46

-2.03

17.42

1.72

0.85

0.21

0.91

19.30

0.79

16.16

0.04

0.00

-0.01

0.04

0.00

-0.01

0.08

17.71

1.71

0.88

0.18

0.82

34.33

0.72

30.75

0.05

0.00

0.00

0.02

0.00

-0.01

0.05

17.01

2.24

0.83

0.18

0.89

10.33

0.62

24.23

0.04

0.00

0.01

-0.07

0.01

-0.09

0.35

7.40

0.91

0.82

0.39

0.91

5.29

2.42

0.08

0.04

0.00

-0.02

0.04

1.34

-1.42

1.36

7.59

0.25

0.93

0.44

1.00

6.74

2.65

0.41

0.04

0.01

-0.07

0.27

-0.27

2.73

-1.57

6.33

2.64

0.81

0.39

1.18

34.23

4.88

n/a

0.04

-0.05

0.37

-0.45

0.00

0.00

0.00

7.05

2.57

0.94

0.41

1.21

57.40

3.03

0.20

0.04

0.00

0.03

-0.17

6.46

-8.43

5.41

6.62

2.02

0.84

0.47

1.70

13.67

3.41

0.02

0.04

0.02

-0.21

0.76

-6.20

7.12

0.14

1 EGPET

12

6

1.4cm long

MK

Tell el Yehudijeh

steatite

e

3 EGPET

13

4

2x1.4cm

late MK

Lahun

glazed steatite

e

10 EGPET

14

12

1.8x1.2cm

late MK

Lahun

faience

e

1 EGPET

15

2

1.8x1.3cm

late MK

Lahun

faience

e

4 EGPET

16

9

2.1x1.6cm

late MK

Lahun

steatite

e

4 EGPET

17

5

1.7x1cm

late MK

Lahun

glazed steatite

e

1 EGPET

18

2

2x1.4cm

late MK

Lahun

steatite

e

3 EGPET

19

8

L2.2cm

late MK

Lahun

glazed steatite

e

5 EGPET

20

6

late MK

Lahun

glazed steatite

e

6 EGPET

21

10

late MK

Lahun

glazed steatite

e

1 EGPET

22

4

late MK

Lahun

glazed steatite

e

2 EGPET

23

3

late MK

Lahun

glazed steatite

e

288

sampling resolution

image source**

spiral numbers analysed

8.29

2.54

0.73

0.24

1.01

82.61

2.12

1.49

0.04

0.00

-0.02

0.22

0.31

-0.91

1.17

9.14

1.92

0.60

0.49

0.54

7.84

1.83

9.30

0.04

0.00

-0.06

0.25

0.02

0.99

-1.11

5.78

1.73

0.74

0.23

0.95

5.39

3.43

n/a

0.04

0.01

-0.08

0.17

0.00

0.00

0.00

5.66

1.41

0.83

0.14

1.16

10.54

3.44

n/a

0.04

-0.01

0.02

-0.01

0.00

0.00

0.00

8.29

1.81

0.82

0.26

1.04

18.80

2.00

1.31

0.04

0.00

0.04

-0.22

-0.48

1.83

-1.43

7.87

2.44

0.37

0.34

0.90

6.19

2.42

1.33

0.04

0.00

-0.01

-0.03

0.78

0.11

-1.11

5.35

1.18

0.71

0.18

1.39

17.62

4.53

n/a

0.04

0.02

-0.16

0.30

0.00

0.00

0.00

5.66

1.50

0.74

0.27

1.02

6.01

4.45

n/a

0.04

0.01

-0.09

0.24

0.00

0.00

0.00

8.81

1.25

0.64

0.20

0.92

11.14

2.10

0.18

0.04

0.00

0.01

-0.05

0.03

0.48

-1.61

8.51

1.92

0.80

0.13

0.76

8.37

2.98

1.62

0.04

0.00

0.03

-0.15

-0.53

2.12

-2.59

5.88

1.76

0.73

0.21

1.19

28.63

3.52

n/a

0.05

-0.01

0.02

0.20

0.00

0.00

0.00

5.86

1.65

0.66

0.15

1.04

17.00

3.44

n/a

0.04

0.04

-0.30

0.65

0.00

0.00

0.00

10.01

2.92

0.48

0.18

1.44

16.24

0.94

3.55

0.04

0.00

-0.02

0.02

0.18

-0.97

1.23

10.46

2.57

0.78

0.19

1.24

9.66

1.08

11.71

0.04

0.00

-0.03

0.09

0.17

-1.16

1.94

8.80

1.00

0.38

0.23

0.69

28.47

2.18

1.56

0.04

0.01

-0.07

0.21

-0.69

2.32

-1.24

8.72

1.68

0.83

0.23

1.38

69.63

0.63

1.07

0.04

0.03

-0.41

1.37

0.71

-2.17

1.13

10.02

2.14

0.80

0.27

1.20

82.28

0.88

7.59

0.04

0.01

-0.13

0.52

0.36

-1.75

2.30

10.59

0.83

0.78

0.32

1.03

48.21

0.83

17.75

0.04

0.00

-0.05

0.22

0.19

-0.88

0.99

5 EGPET

24

8

late MK

Lahun

glazed steatite

e

6 EGPET

25

8

late MK

Lahun

glazed steatite

e

5 EGPET

26

11

name scarab - title deputy secretary

late MK

Lahun

glazed steatite

e

1 EGPET

27

3 3

EGPET

28

6 2

EGPET

29

3

personal name scarab personal name scarab - title inspector of guards personal name scarab - title inspector of guards

late MK

late MK

late MK

Naqada

Lahun

Koptos

gold

brown jasper

glazed steatite

e

e

e

2 EGPET

30

3

12D

Hawara

glazed steatite

e

1 EGPET

31

5 1

Petrie Museum UC20857

spacing trend error

Petrie Museum UC20732

spacing median

Petrie Museum UC16142

path width trend error

Petrie Museum UC11385

path width median

Petrie Museum UC11366

material

irregularity

Petrie Museum UC11353

place

spacing trend coefficients

curvilinearity

Petrie Museum UC11349

date*

path width trend coefficients

orientation***

Petrie Museum UC11348

Object notes

Style data

rotation

Petrie Museum UC11169

Object number

Museum number

Museum ID

Context data

EGPET

32

2

tomb 1595 tomb 1735, Found in shaft tomb probably strung as part of a necklace originally (Brunton 1927, p.40). representational context - papyrus.

9D

9D

Qau

Qau

steatite

steatite

e

e

*1I – First Intermediate, MK – Middle Kingdom, 2I – Second Intermediate, D - Dynasty ** image source codes refer to those used on the List of Figures ***underlined orientation assumed from cluster group

289

Table A2.5 Contextual and style data for the Bronze Age Cypriot case study

orientation***

curvilinearity

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

Object notes

rotation

place

16.57

2.66

0.78

0.15

0.70

55.80

0.89

77.47

0.05

0.00

0.03

-0.26

0.00

-0.07

0.34

14.80

2.85

0.86

0.15

0.89

19.48

0.76

48.35

0.04

0.00

-0.03

0.15

0.00

0.00

-0.22

b

20.10

0.24

0.82

0.42

0.38

11.70

0.33

97.11

0.04

0.00

0.00

-0.01

0.01

-0.09

0.25

b

25.07

2.54

0.86

0.10

0.57

9.07

0.35

7.22

0.04

0.00

0.00

0.03

0.00

0.00

0.03

b

18.58

1.71

0.86

0.08

0.66

4.61

0.31

7.73

0.04

0.00

0.00

-0.05

0.00

0.01

-0.11

b

32.99

2.54

0.84

0.03

0.26

9.66

0.13

8.17

0.05

0.00

0.00

0.01

0.00

0.00

0.02

b

50.19

1.5

0.81

0.07

0.38

36.27

0.23

68.55

0.05

0.00

0.00

0.00

0.00

0.00

0.01

6.73

1.93

0.88

0.26

2.22

9.68

0.78

0.04

0.04

0.04

-0.43

1.44

-0.02

0.28

-0.50

g

6.95

1.05

0.88

0.24

2.44

27.01

0.84

0.08

0.04

0.04

-0.43

1.53

-0.06

0.48

-0.57

g

19.99

1.71

0.88

0.08

0.60

9.93

0.27

5.44

0.05

0.00

0.00

0.03

0.00

0.00

-0.02

g

28.33

1.5

0.81

0.05

0.26

8.01

0.27

10.79

0.05

0.00

0.00

0.02

0.00

0.00

0.06

g

14.73

2.13

0.63

0.08

0.46

12.81

0.37

27.82

0.05

0.00

-0.01

0.03

0.00

-0.03

0.11

g

37.13

2.80

0.84

0.02

0.17

6.98

0.07

7.12

0.05

0.00

0.00

0.00

0.00

0.00

0.00

image source**

end date

period*

Museum ID

start date

Style data Object number spiral numbers analysed

Context data Museum number

path width trend coefficients

spacing trend coefficients

Pottery Ashmolean Museum AN1958.62

1375

1300

French tomb 5, Ayios Iakovos, Enkomi

1250

1200

Enkomi

1 CYASH

1

2

British Museum 1938,1120.5

CYBM

1

1

?late LC IIC, ?LC III

British Museum 1873,0320.187

CYBM

3

1

CG IIIII

950

750

Idalion

British Museum 1873,0320.191

CYBM

4

1

CG I-II

1050

900

Idalion

British Museum 1880,0710.9

CYBM

5

1

CA I

700

600

Aradhippou tomb, Larnaka

1

LC IIIA, LC IIC

1150

EpiskopiBamboula Site D tomb 96, Kourion

British Museum 1896,0201.333

CYBM

7

1

Medelhavsmuseet S.011:006 or St.011:006

CYMM

1

2

Medelhavsmuseet A. 019:012

CYMM

2

1

CYMM

3

CYMM

CYMM

Medelhavsmuseet SHM 17946:048 Medelhavsmuseet SHM 17946:0013 or SHM 17946:0013 Medelhavsmuseet SHM 17946:0018 or SHM 17946:018

1250

c.725

600

CG I-II

1050

850

Amathus tomb 19

1

CA I

700

600

?Marion

4

1

CG III

900

750

?Idalion, ?Amathus

5

1

CG III

900

750

?Paphos [Palaepaphos]

cup, 'Mycenaean style' krater, Pastoral style, White Painted Wheelmade III ware, ?representational plate, White Painted II style, W33cm plate, White Painted I style, diam 30.5cm dish, Black-onRed IV ware, wheel-made, spiral on interior. Sampled at low resolution. bowl, White Painted Wheelmade III ware, spiral on inside. Sampled at low resolution. kylix, painted, representational (plant), diam. 19.8 bowl, painted, chamber, level 15. bowl, painted, spiral on base of foot. Sampled at low resolution.

barrel jug barrel jug, White Painted III ware. Sampled at low resolution.

290

2

CA I

700

600

?Stylli

Medelhavsmuseet MM 1997:003

CYMM

8

1

?CA I

700

600

??Kouris Valley

Medelhavsmuseet SHM 17946:038

CYMM

9

1

CA I

700

600

?PalaepaphosSkales

not stated

not stated

not stated

Medelhavsmuseet NM Ant 0466 Medelhavsmuseet MM Acc 0898

1 CYMM

10

2

not stated

1 CYMM

11

2

CA I

700

600

not stated

1 Medelhavsmuseet MM Acc 0699 Medelhavsmuseet MM 1956:010 Medelhavsmuseet MM Acc 0443 Medelhavsmuseet MM Acc 0472

CYMM

12

2

CA I

700

600

not stated

CYMM

13

1

14

1

950 not stated

850 not stated

not stated

CYMM

CG II not stated

CYMM

15

1

CA I

700

600

Lapithos

not stated

1 Medelhavsmuseet E. 003:276

CYMM

16

2

LC II

1450/ 1400

1200

Enkomi, tomb 3 chamber

Medelhavsmuseet A.J. B.A.S. 030

CYMM

17

1

not stated

not stated

not stated

Ajios Jakovos (BA sanctuary)

LC II

1450/ 1400

1200

Enkomi, Tomb 3, Chamber, Floor

Medelhavsmuseet E. 003:044

CYMM

18

1

flask bowl, White Painted IV ware, rim diam 19.4 krater, representational (plant), Mycenaean IIIA:2 ware, H39.6 jug, Mycenaean style, H11 carinated cup, White Painted Wheelmade III ware, diam 6.4cm

sampling resolution

7

spacing trend error

CYMM

spacing median

Medelhavsmuseet NM Ant 1719

path width trend error

1

22.94

2.54

0.79

0.03

0.31

5.38

0.10

19.12

0.06

0.00

0.00

0.00

0.00

0.00

-0.01

6.92

1.37

0.62

0.40

2.27

16.86

2.26

0.10

0.04

0.03

-0.33

1.45

0.72

-1.84

3.73

g

7.23

1.87

0.86

0.40

1.46

19.31

3.27

0.26

0.04

0.00

0.02

0.18

-0.09

-1.27

4.43

g

62.07

0.31

0.73

0.05

0.14

3.69

0.14

5.53

0.09

0.00

0.00

0.00

0.00

0.00

0.01

g

16.26

2.14

0.69

0.07

0.47

3.47

0.34

8.64

0.04

0.00

-0.02

0.14

0.00

-0.01

0.09

54.90

0.31

0.74

0.03

0.08

1.98

0.11

6.54

0.09

0.00

0.00

0.00

0.00

0.00

0.00

38.47

1.5

0.86

0.01

0.13

5.01

0.12

9.63

0.05

0.00

0.00

0.02

0.00

0.00

0.01

68.13

0.31

0.57

0.03

0.07

1.57

0.07

3.12

0.14

0.00

0.00

0.00

0.00

0.00

0.00

63.63

0.31

0.53

0.04

0.10

11.96

0.09

26.72

0.14

0.00

0.00

0.00

0.00

0.00

0.00

5.83

1.38

0.93

0.29

2.23

5.76

1.85

0.01

0.04

0.02

-0.21

1.03

20.40

-8.15

3.62

g

5.76

2.02

0.88

0.38

1.95

3.94

1.04

0.01

0.04

0.08

-0.68

1.89

-7.80

4.96

0.96

g

13.02

1.32

0.80

0.06

0.49

9.68

0.28

13.31

0.04

0.00

0.02

-0.15

0.00

-0.03

0.14

g

15.32

2.60

0.85

0.13

0.64

9.60

0.53

31.92

0.05

0.00

0.00

0.04

0.00

-0.02

0.11

g

18.62

0.96

0.34

0.07

0.55

15.23

0.43

48.22

0.05

0.00

0.01

-0.08

0.00

-0.03

0.20

8.08

1.02

0.67

0.39

0.24

1.22

2.53

0.81

0.05

0.00

0.00

0.03

0.12

1.41

0.34

g

7.64

0.66

0.68

0.22

0.80

2.35

1.45

0.08

0.04

0.00

0.05

-0.12

-0.72

3.68

-2.34

g

25.70

0.33

0.82

0.12

0.61

48.29

0.50

74.14

0.04

0.00

0.00

-0.01

0.00

-0.01

0.07

g

17.05

2.55

0.84

0.18

0.90

55.36

1.05

34.69

0.05

0.00

0.00

-0.06

0.01

-0.21

0.82

image source**

end date 600

path width median

700

irregularity

CA I

bowl, Black on Red II (IV) ware. Sampled at low resolution. amphora, representational (lotus), Bichrome IV ware. jug, Black on Red II (IV) ware , H35. Sampled at low resolution. bowl, Black on Red II (IV) ware, rim diam 16.5 bowl, Bichrome IV ware. Sampled at low resolution. bowl, Bichrome IV ware. Sampled at low resolution. amphora, representational (lily), Bichrome IV ware, H76 bowl, Bichrome II ware.

curvilinearity

1

Object notes

orientation***

6

place Amathus tomb 2 Chamber, Stratum 3, Level 88

rotation

CYMM

period*

Museum ID Medelhavsmuseet A. 007:021

start date

Style data Object number spiral numbers analysed

Context data Museum number

g

g

g

path width trend coefficients

spacing trend coefficients

291

Medelhavsmuseet M. 068:012 Medelhavsmuseet E. 003:165 Royal Albert Memorial Museum 917/1910 Royal Albert Memorial Museum 31/1918/47 Goldwork Ashmolean Museum AN1962.244

22

Enkomi

1

LC II/LH IIIB1

Enkomi, Tomb 18, side chamber

1

CG I

1

CG IIICA I

1050

600

Lapithos, Tomb 428, floor Amathus, Tomb 23, Chamber, stratum 1, floor

950

CYMM

24

1

CA I

750

600

Marion, Tomb 68, chamber, stratum 2, level 20

CYMM

25

1

LC II

1450/ 1400

1200

Enkomi, Tomb 3, chamber

CYRAM

1

1

IA

850

600

Paphos

CYRAM

2

1

IA

850

600

Cyprus

bottle, White Painted II ware, H11.5 bowl, Black on Red II (IV) ware, diam 19.5. Sampled at low resolution. pyxis, Mycenaean IIIA ware, H13.5 barrel jug, Archaic ware, Black on Red ware jug (barrel), Archaic ware, Black on Red ware. Sampled at low resolution.

1200

probably Enkomi

embossed funerary diadem

CYMM

23

900

1 CYASH

2

6

1450

sampling resolution

Medelhavsmuseet A. 023:075

CYMM

21

2

spacing trend error

Medelhavsmuseet L. 428:080

CYMM

20

spacing median

Medelhavsmuseet E. 018 Sk:046

CYMM

LC II/LH IIIA2

path width trend error

Medelhavsmuseet E. 003:261

path width median

1

21.02

2.55

0.84

0.09

0.48

12.06

0.31

7.94

0.04

0.00

0.00

0.03

0.00

0.00

0.01

9.42

2.10

0.63

0.29

0.52

3.43

1.30

1.42

0.04

0.00

0.00

0.05

0.32

-0.98

1.20

g

10.86

0.78

0.82

0.32

0.37

1.62

1.45

4.45

0.04

0.00

-0.02

0.06

0.00

0.21

-0.03

g

16.96

0.50

0.88

0.25

0.28

5.18

1.58

206.65

0.05

0.00

-0.01

0.11

-0.01

0.17

-0.86

g

32.38

2.53

0.84

0.04

0.39

12.22

0.23

36.46

0.04

0.00

0.00

0.00

0.00

0.00

0.00

g

17.60

2.50

0.87

0.05

0.44

5.46

0.21

11.43

0.05

0.00

0.00

-0.01

0.00

0.00

-0.01

g

50.51

2.94

0.68

0.03

0.15

4.16

0.13

7.65

0.09

0.00

0.00

0.00

0.00

0.00

0.01

g

32.91

3.02

0.45

0.11

0.54

44.74

0.44

52.96

0.05

0.00

0.00

0.00

0.00

0.01

-0.04

i

76.62

0.12

0.50

0.09

0.13

2.47

0.14

18.40

0.14

0.00

0.00

0.00

0.00

0.00

0.00

i

47.98

0.20

0.83

0.04

0.25

9.05

0.22

23.06

0.05

0.00

0.00

-0.01

0.00

0.00

0.00

14.62

2.12

0.51

0.17

0.95

13.78

0.37

23.38

0.04

0.00

-0.01

0.10

0.02

-0.17

0.45

17.38

1.24

0.86

0.15

1.08

29.17

0.45

14.08

0.04

0.00

-0.01

0.13

0.00

-0.04

0.16

image source**

end date 1100

irregularity

1340

curvilinearity

1

Enkomi, Tomb 13, Chamber, Level 51

Object notes bowl, White Painted Wheelmade III ware, local imitation of Mycenaean pottery krater, Mycenaean III ware, representational (plant), H43.8, diam 25.1 krater, representational (plant), Mycenaean IIIB ware. jug, White Painted I ware, H12.6cm. Sampled at low resolution.

orientation***

19

LC IICIIIA

place

rotation

CYMM

period*

Museum ID Medelhavsmuseet E. 013:211

start date

Style data Object number spiral numbers analysed

Context data Museum number

g

path width trend coefficients

spacing trend coefficients

292

1 CYBM

17

12 1

CYBM

18

11 1

CYBM

19

2

from LC IB from LC IB

1550

1550

1550

1050

1050

Enkomi (tomb 69)

1050

Enkomi (tomb 56)

1050

Enkomi (tomb 93)

7 British Museum 1897,0401.649 British Museum 1897,0401.471 British Museum 1897,0401.448 British Museum 1900,0615.18 British Museum 1897,0401.523 British Museum 1897,0401.525

CYBM

20

11 1

CYBM

21

7 7

CYBM

22

14 9

CYBM

23

10 7

CYBM

24

9 16

CYBM

25

22

from LC IB from LC IB from LC IB from LC IB from LC IB from LC IB

1550

1550

1550

1550

1550

1550

1050

Enkomi

1050

Enkomi (tomb 88)

1050

Enkomi (tomb 79)

1050

Enkomi (tomb ?93)

1050

Enkomi (tomb 93)

1050

Enkomi (tomb 93)

1 British Museum 1897,0401.269

CYBM

26

2

from LC IB

1550

1050

Enkomi (tomb 66)

1 British Museum 1897,0401.522

CYBM

27

2

from LC IB

1550

1050

Enkomi (tomb 93)

1 British Museum 1897,0401.740

CYBM

28

2

from LC IB

1550

1050

Enkomi (tomb 90?)

image source**

end date

start date

period* from LC IB

1550

b

b

b

b

b

b

b

b

b

b

b

b

b

sampling resolution

7

Enkomi (tomb 72)

b

spacing trend error

16

Site A, Maroni

spacing median

CYBM

from LC IB

1200

path width trend error

2

LC II

path width median

British Museum 1897,0401.528

2

irregularity

British Museum 1897,0401.196

15

repoussé mouthpiece. L10 impressed mouth-piece. L9.9 impressed mouth-piece. L9.9 impressed mouth-piece. L10.1 impressed and hammered mouth-piece. impressed and hammered ?mouth-piece. L10 impressed and hammered mouth-piece. impressed and hammered mouth-piece. impressed and hammered mouth-piece impressed and hammered mouth-piece impressed and hammered mouth-piece impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment. impressed and hammered mouthpiece/diadem fragment

curvilinearity

British Museum 1897,0401.376

CYBM

c. 1400

Object notes

orientation***

British Museum 1897,0401.400.*

1

place

rotation

British Museum 1898,1201.74

Object number spiral numbers analysed

Style data

Museum ID

Context data Museum number

21.04

1.37

0.85

0.13

0.85

20.68

0.47

5.43

0.04

0.00

0.00

0.05

0.00

0.00

0.03

23.98

1.41

0.85

0.17

0.95

68.19

0.39

13.96

0.04

0.00

-0.01

0.08

0.00

-0.01

0.05

14.35

1.21

0.51

0.10

1.10

51.25

0.53

22.88

0.04

0.00

-0.01

0.23

-0.02

0.32

-0.97

12.62

0.07

0.86

0.24

0.89

42.79

0.65

5.30

0.04

0.00

0.00

0.00

0.01

-0.04

-0.02

23.80

0.85

0.87

0.23

0.78

41.74

0.29

50.54

0.04

0.00

0.00

0.02

0.00

-0.04

0.21

17.00

0.16

0.84

0.12

0.62

32.01

0.99

26.39

0.04

0.00

-0.01

0.14

0.00

-0.02

0.18

19.67

1.20

0.88

0.24

0.96

15.12

0.31

25.74

0.04

0.00

-0.02

0.18

0.00

-0.01

0.06

21.82

0.44

0.87

0.19

0.71

13.07

0.39

67.93

0.04

0.00

0.01

0.00

0.00

-0.07

0.34

17.72

1.51

0.88

0.13

1.14

28.71

0.32

6.84

0.04

0.00

0.00

0.08

0.00

-0.03

0.13

19.82

0.23

0.92

0.14

1.25

23.23

0.28

16.02

0.04

0.00

0.00

0.10

0.00

0.01

-0.09

15.21

0.25

0.87

0.12

0.98

15.00

0.78

14.33

0.04

0.00

0.02

-0.08

-0.01

0.14

-0.45

14.41

2.23

0.87

0.11

0.69

52.57

0.82

17.93

0.04

0.00

-0.02

0.19

-0.01

0.15

-0.42

19.04

0.92

0.88

0.12

0.90

23.64

0.26

11.78

0.04

0.00

-0.01

0.04

0.00

0.02

-0.11

18.85

1.98

0.92

0.11

1.04

37.65

0.17

18.17

0.04

0.00

-0.02

0.17

0.00

0.03

-0.25

17.25

2.59

0.74

0.27

0.70

17.72

0.10

65.78

0.04

0.00

-0.02

0.14

0.02

-0.16

0.20

17.70

0.52

0.83

0.21

0.71

12.31

0.51

26.84

0.04

0.00

0.01

0.03

0.01

-0.09

0.23

17.96

2.51

0.53

0.18

1.07

17.49

0.20

13.93

0.04

0.00

-0.01

0.09

0.00

-0.04

0.10

18.19

2.40

0.82

0.21

1.06

56.82

0.05

6.79

0.04

0.00

0.00

0.06

0.00

-0.05

0.14

14.65

0.56

0.88

0.11

1.35

58.63

0.31

11.59

0.04

0.00

-0.01

0.08

0.00

0.05

-0.19

15.50

0.34

0.87

0.20

1.45

69.41

0.29

12.04

0.04

0.00

-0.04

0.42

0.00

-0.04

0.09

16.85

0.51

0.91

0.07

1.17

35.91

0.45

39.98

0.04

0.00

0.00

0.00

0.00

-0.03

0.22

14.40

0.45

0.87

0.10

1.41

26.42

0.26

19.26

0.04

0.00

-0.06

0.39

0.00

0.03

-0.08

24.53

2.34

0.81

0.12

0.82

51.34

0.27

37.61

0.05

0.00

-0.01

0.17

0.00

0.01

-0.06

24.67

3.05

0.82

0.13

0.79

32.38

0.30

18.94

0.05

0.00

0.00

0.01

0.00

0.00

0.09

7.13

2.00

0.92

0.13

1.26

22.43

0.50

0.04

0.04

0.01

-0.14

0.68

-0.69

3.57

-1.56

6.55

1.60

0.76

0.08

1.01

14.61

0.26

0.00

0.04

-0.03

0.20

-0.28

7.49

4.76

-0.34

9.76

2.09

0.85

0.14

1.31

41.25

0.48

1.64

0.04

0.01

-0.11

0.54

0.12

-0.63

0.86

9.83

1.74

0.90

0.09

1.52

63.68

0.95

2.71

0.04

0.00

0.03

0.09

0.05

-0.35

0.81

path width trend coefficients

spacing trend coefficients

293

Enkomi (tomb 66)

1 British Museum 1897,0401.444

CYBM

30

4

from LC IB

1550

1050

Enkomi (tomb 79)

3 British Museum 1897,0401.520

CYBM

31

8

from LC IB

1550

1050

Enkomi (tomb 93)

1 British Museum 1897,0401.11

CYBM

32

8

from LC IB

1550

1050

Enkomi (tomb 15)

6 British Museum 1897,0401.272

CYBM

33

13

from LC IB

1550

1050

Enkomi (tomb 66)

3 British Museum 1897,0401.65

CYBM

34

5

from LC IB

1550

1050

Enkomi (tomb 19)

4 British Museum 1897,0401.64

CYBM

35

6

from LC IB

1550

1050

Enkomi (tomb 19)

6 British Museum 1897,0401.63

CYBM

36

12

from LC IB

1550

1050

Enkomi (tomb 19)

39 British Museum 1898,1201.18

CYBM

37

77

LC II/III

1400

1100

Maroni (tomb 4), Larnaka

8 British Museum 1897,0401.66

CYBM

38

12

from LC IB

1550

1051

Enkomi (tomb 19)

b

b

b

b

b

b

b

b

b

b

sampling resolution

1050

spacing trend error

1550

spacing median

from LC IB

path width trend error

4

path width median

29

irregularity

CYBM

curvilinearity

British Museum 1897,0401.270

orientation***

3

Object notes impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment

rotation

place

image source**

end date

start date

period*

Object number spiral numbers analysed

Style data

Museum ID

Context data Museum number

17.97

0.11

0.86

0.13

1.16

43.41

0.23

21.91

0.04

0.00

-0.01

0.07

0.00

0.07

-0.42

16.84

2.92

0.90

0.11

0.98

33.87

0.41

10.25

0.04

0.00

0.01

-0.08

0.00

0.08

-0.39

18.92

0.00

0.35

0.18

0.99

36.86

0.39

17.90

0.04

0.00

0.00

0.00

0.00

-0.09

0.48

14.63

0.33

0.79

0.13

1.04

32.75

0.34

9.36

0.04

0.00

0.01

-0.07

0.00

-0.05

0.22

7.50

1.84

0.85

0.16

1.19

6.82

0.81

0.12

0.04

0.00

-0.08

0.37

0.79

-1.21

0.90

9.25

0.43

0.87

0.04

1.18

20.30

1.18

16.54

0.04

-0.01

0.07

-0.17

0.11

-0.96

2.20

24.06

1.59

0.88

0.08

0.82

19.85

0.03

15.59

0.04

0.00

0.00

0.01

0.00

0.00

-0.03

23.13

2.13

0.84

0.09

0.79

29.30

0.02

23.09

0.04

0.00

-0.01

0.08

0.00

-0.02

0.09

23.15

2.06

0.83

0.11

0.90

50.34

0.13

24.71

0.04

0.00

-0.01

0.07

0.00

0.00

0.00

23.26

1.55

0.87

0.13

0.73

29.39

0.21

23.78

0.04

0.00

0.00

-0.01

0.00

-0.02

0.08

24.51

1.47

0.87

0.10

0.77

38.31

0.03

32.86

0.04

0.00

0.00

0.01

0.00

0.00

-0.01

21.80

0.06

0.86

0.06

0.81

53.40

0.03

6.79

0.04

0.00

0.00

0.06

0.00

0.00

-0.03

17.90

1.28

0.85

0.14

0.85

25.43

0.04

33.77

0.04

0.00

0.00

0.00

0.00

-0.01

-0.13

24.53

1.48

0.83

0.10

0.78

26.12

0.02

20.42

0.04

0.00

-0.01

0.07

0.00

-0.01

0.07

22.48

0.20

0.78

0.11

0.92

46.30

0.04

8.86

0.04

0.00

-0.02

0.15

0.00

0.00

-0.05

24.77

1.32

0.00

0.10

0.75

41.62

0.22

9.82

0.04

0.00

0.00

-0.03

0.00

0.01

-0.01

13.30

1.87

0.91

0.20

0.96

32.41

0.50

22.50

0.04

0.00

0.01

0.01

-0.01

0.26

-1.40

14.51

1.76

0.83

0.21

1.76

27.99

0.49

12.88

0.04

0.00

0.03

-0.06

0.03

-0.33

0.89

14.37

2.14

0.88

0.16

0.89

25.96

1.08

19.36

0.04

0.00

-0.01

0.12

0.00

-0.06

0.19

14.09

0.58

0.82

0.18

1.07

32.30

1.01

16.83

0.04

0.00

0.00

0.00

-0.01

0.17

-0.60

path width trend coefficients

spacing trend coefficients

294

Enkomi (tomb unknown)

1 British Museum 1900,0615.9

CYBM

40

8

from LC IB

1550

1051

Enkomi (tomb 93 probably)

2 British Museum 1897,0401.517

CYBM

41

6

from LC IB

1550

1051

Enkomi (tomb 93)

7 British Museum 1897,0401.518 Fitzwilliam Museum GR.155b.1909 Medelhavsmuseet E. 011:198

CYBM

42

10 18

CYFM

1

22

from LC IB LH IIIA-B

1400

LC II

1450/ 1400

LC II

1450/ 1400

CA

not stated

1 CYMM

26

3 11

Medelhavsmuseet E. 011:036

CYMM

27

Medelhavsmuseet BEK 1963.30

CYMM

28

16 4 13

1550

1051

1201

Enkomi (tomb 93)

b

b

b

b

diadem

1200

Cyprus Enkomi, Tomb 11, chamber, floor Enkomi, Tomb 11, chamber, level 96

diadem

g

not stated

not stated

diadem

g

1200

mouth-piece

g

*LC – Late Cypriot, LH – Late Helladic, CG – Cypro-Geometric, CA – Cypro-Archaic, IA – Iron Age ** image source codes refer to those used on the List of Figures *** underlining indicates orientation estimated from clustering

sampling resolution

1051

spacing trend error

1550

spacing median

from LC IB

path width trend error

4

path width median

39

irregularity

CYBM

curvilinearity

British Museum 1897,0401.651

orientation***

3

Object notes impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment impressed and hammered mouthpiece/diadem fragment

rotation

place

image source**

end date

period*

Museum ID

start date

Style data Object number spiral numbers analysed

Context data Museum number

15.56

0.00

0.89

0.16

0.94

53.88

0.05

10.59

0.04

0.00

-0.04

0.22

0.01

-0.08

0.19

13.51

0.09

0.84

0.00

0.75

57.75

0.45

8.57

0.04

0.00

-0.05

0.34

0.00

0.01

-0.07

23.67

1.73

0.87

0.13

0.76

20.74

0.23

25.15

0.04

0.00

0.00

0.02

0.00

-0.01

0.08

23.37

2.54

0.92

0.10

0.84

22.12

0.19

11.37

0.04

0.00

0.00

-0.04

0.00

-0.01

0.06

10.64

0.84

0.66

0.37

0.81

16.50

1.56

2.25

0.05

0.00

-0.02

0.16

0.11

-0.19

-0.82

10.03

0.11

0.81

0.40

1.12

48.76

2.88

6.98

0.04

0.00

-0.02

0.21

-0.12

0.62

0.73

19.46

1.01

0.83

0.13

1.36

64.65

0.03

2.62

0.04

0.00

-0.01

0.16

0.00

0.00

0.00

17.72

0.43

0.86

0.13

1.39

102.22

0.00

0.91

0.04

0.00

-0.01

0.25

0.00

0.00

-0.01

10.73

2.86

0.83

0.18

0.90

6.89

0.74

2.92

0.04

0.00

-0.02

0.15

0.05

-0.17

-0.20

12.43

2.68

0.52

0.10

0.80

6.84

0.75

4.37

0.04

0.00

-0.01

0.11

0.03

-0.25

0.68

18.38

2.96

0.85

0.14

1.09

23.38

0.24

26.84

0.04

0.00

0.00

-0.01

0.00

-0.01

0.10

18.29

3.11

0.88

0.14

1.03

41.71

0.32

6.29

0.04

0.00

0.00

0.06

0.00

0.01

-0.07

18.80

1.60

0.87

0.15

1.14

44.00

0.29

29.72

0.04

0.00

0.01

-0.02

0.00

0.00

-0.08

18.34

2.76

0.82

0.13

1.23

34.24

0.17

6.93

0.04

0.00

0.00

-0.02

0.00

-0.01

0.04

11.29

2.96

0.86

0.20

1.23

44.95

1.00

6.88

0.04

0.00

0.04

-0.19

0.03

-0.14

0.15

15.53

0.12

0.79

0.33

0.47

10.80

0.79

14.53

0.05

0.00

0.00

0.03

0.01

-0.10

0.19

path width trend coefficients

spacing trend coefficients

295

Table A2.6 Context and style data for the endogenous visual phenomena examples

orientation

curvilinearity

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

cause

Style data

rotation

spiral numbers

Page/figure

Source

Object number

Context data

34.12

0.68

0.35

0.03

0.37

6.23

0.18

72.82

0.05

0.00

0.00

0.01

0.00

0.00

-0.01

10.30

0.41

0.79

0.12

0.62

1.75

0.44

12.25

0.04

0.00

0.00

-0.02

0.06

-0.43

0.76

13.14

1.64

0.89

0.13

0.38

1.49

1.01

50.79

0.04

0.00

0.00

0.00

0.00

-0.09

0.56

24.15

0.35

0.92

0.16

0.60

16.20

1.15

232.72

0.04

0.00

0.00

0.06

0.00

0.05

-0.16

6.77

0.29

0.82

0.48

0.70

7.07

4.27

0.04

0.04

0.01

-0.19

0.75

0.60

-1.17

5.91

39.84

1.25

0.80

0.08

0.08

2.48

0.66

66.57

0.05

0.00

0.00

0.00

0.00

0.00

-0.01

30.01

0.08

0.89

0.07

0.59

35.72

0.44

53.09

0.04

0.00

0.00

-0.02

0.00

0.00

0.07

after

response to a concentric circle stimulus card

16.60

2.46

0.87

0.16

0.15

1.74

1.93

79.24

0.04

0.00

0.00

0.00

0.02

-0.38

1.59

after

response to a concentric circle stimulus card

14.74

2.71

0.39

0.02

0.15

2.61

0.96

60.07

0.04

0.00

0.00

-0.01

0.00

0.01

0.09

after

response to a concentric circle stimulus card

21.38

1.31

0.48

0.12

0.24

4.26

1.53

158.68

0.04

0.00

0.00

-0.03

0.00

-0.08

0.39

after

response to a concentric circle stimulus card

21.91

2.08

0.88

0.10

0.26

6.30

0.55

28.54

0.04

0.00

0.01

-0.05

0.00

-0.02

0.17

drawn during/ after experience

probably after

Object notes interpreted as a pebble dropped in a pond may have been redrawn from the original drawing interpreted as 'hate' may have been redrawn from the original drawing

after

interpreted as a dragon subject was an artist

Horowitz 1964

p.516, fig 2b

FC1

1

Horowitz 1964

p.516, fig 2b

FC1

2

Horowitz 1964 Maclay and Guttmann 1941

p.517, fig 4

FC2

1

schizophrenia, hallucinations exacerbated by medication, spirals not apparently seen prior to medication schizoid hallucinations for several weeks after taking LSD

p.133, fig. 4

FC3

1

mescaline

unknown

Siegel and Jarvik (1975)

p.135, fig.19

FC4

1

ketamine

unknown

subject was an artist described as multicoloured patterns' but presented within a representational scene of houses and sea flashback image, presented as a rotating image flashback image, diagnosed as schizophrenic

probably after

Siegel and Jarvik (1975)

p.135, fig.20

FC5

1

LSD/ schizophrenia

unknown

Siegel and Jarvik

p.136, fig.21

FC6

1

schizophrenia

unknown

1

light flicker 11Hz

2

light flicker 12Hz

3

light flicker 12Hz

4

light flicker 13Hz

Becker (2005)

Becker (2005)

Becker (2005)

Becker (2005)

p.202

p.202

p.202

p.203

FC7

FC7

FC7

FC7

path width trend coefficients

spacing trend coefficients

296

Object notes

curvilinearity

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

FC7

drawn during/ after experience

orientation

p.206

FC7

cause

Style data

rotation

Becker (2005)

p.205

spiral numbers

Page/figure Becker (2005)

Object number

Context data Source

after

response to a concentric circle stimulus card

19.54

0.39

0.86

0.04

0.20

3.68

1.19

46.45

0.05

0.00

0.00

0.00

0.00

0.05

-0.11

after

response to a concentric circle stimulus card

19.92

1.79

0.42

0.14

0.36

5.62

1.17

187.99

0.04

0.00

0.01

-0.10

0.00

-0.04

0.29

8.98

2.87

0.47

0.15

0.15

1.20

1.74

159.65

0.05

0.00

-0.01

0.02

-0.01

0.02

0.22

5

light flicker 16Hz

6

light flicker 18Hz

after

response to a concentric circle stimulus card

path width trend coefficients

spacing trend coefficients

Becker (2005)

p.206

FC7

7

light flicker 19Hz

Becker (2005)

p.212

FC7

8

light flicker 14Hz

after

response to a line stimulus card

20.72

0.63

0.92

0.14

0.30

8.04

1.06

12.03

0.04

0.00

0.00

-0.02

0.00

-0.02

0.15

Becker (2005)

p.217

FC7

9

light flicker 13Hz

after

response to a point stimulus card

25.79

1.05

0.86

0.03

0.13

4.30

0.89

223.90

0.04

0.00

0.00

-0.02

0.00

-0.02

0.13

Becker (2005)

p.221

FC7

10

light flicker 12Hz

after

response to a radial pattern stimulus card

19.24

1.92

0.82

0.14

0.21

6.09

1.15

53.61

0.05

0.00

0.00

0.02

0.00

-0.08

0.38

Becker (2005)

p.233

FC7

13

light flicker 11Hz

after

response to a wave stimulus card

18.15

2.87

0.85

0.09

0.15

1.81

1.34

136.55

0.04

0.00

0.00

-0.01

-0.01

0.12

-0.61

Becker (2005)

p.235

FC7

14

light flicker 15Hz

after

response to a wave stimulus card

15.54

3.04

0.87

0.06

0.36

25.56

1.42

203.53

0.04

0.00

0.00

-0.08

-0.01

0.08

-0.01

Becker (2005)

p.238

FC7

15

light flicker 14Hz

after

response to a zig-zag stimulus card

27.15

2.91

0.85

0.11

0.32

5.14

0.84

60.45

0.04

0.00

0.00

-0.02

0.00

0.00

-0.05

Becker (2005)

p.223

FC8

1

light flicker 16Hz

after

asked to look for radial patterns

8.24

1.90

0.82

0.15

1.54

33.98

1.79

3.47

0.04

0.00

0.00

0.29

0.62

-2.82

3.94

Becker (2005)

p.229

FC8

2

light flicker 12Hz

after

asked to look for radial patterns

23.71

1.13

0.84

0.13

0.08

1.84

1.58

301.49

0.05

0.00

0.00

-0.01

0.00

0.01

0.17

Becker (2005)

p.230

FC8

3

light flicker 15Hz

after

asked to look for spirals

17.43

0.82

0.82

0.15

1.43

58.20

0.12

15.21

0.04

0.00

0.01

0.09

0.00

-0.01

-0.03

Becker (2005)

p.230

FC8

4

light flicker 15Hz

after

asked to look for spirals

9.07

1.80

0.89

0.16

1.55

29.33

1.51

9.13

0.04

0.01

-0.15

0.87

0.34

-1.22

0.53

297 Context data

rotation

orientation

curvilinearity

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

Becker (2005)

p.231

FC8

5

light flicker 16Hz

after

asked to look for spirals

13.79

2.88

0.87

0.11

2.27

524.79

0.04

1.95

0.03

0.00

-0.04

0.41

0.00

-0.02

0.08

Becker (2005)

p.231

FC8

6

light flicker 17Hz

after

asked to look for spirals

18.78

1.86

0.81

0.14

1.32

110.24

0.02

6.17

0.04

0.00

-0.03

0.19

0.00

0.00

-0.01

Becker (2005)

p.231

FC8

7

light flicker 17Hz

after

asked to look for spirals

7.22

2.30

0.82

0.21

2.91

109.95

0.01

4.78

0.04

0.00

-0.21

1.27

0.01

-0.08

0.19

Becker (2005)

p.232

FC8

8

light flicker 18Hz

after

asked to look for spirals

10.25

2.89

0.89

0.24

2.58

78.68

0.07

0.56

0.04

0.02

-0.30

1.46

0.01

-0.04

0.09

Becker (2005)

p.232

FC8

9

light flicker 18Hz

after

asked to look for spirals

16.12

0.41

0.43

0.19

1.98

118.76

0.31

12.82

0.04

0.00

-0.02

0.32

0.00

0.05

-0.14

Becker (2005)

p.232

FC8

10

light flicker 19Hz

after

asked to look for spirals

11.68

1.15

0.83

0.33

2.34

61.96

0.00

0.05

0.03

0.01

-0.27

1.44

0.00

-0.01

0.00

Becker (2005)

p.229

FC9

1

light flicker 12Hz

after

response to a spiral stimulus card

14.95

0.89

0.45

0.13

1.07

147.14

0.75

43.93

0.04

0.00

0.00

-0.02

0.00

0.02

-0.12

Becker (2005)

p.229

FC9

2

light flicker 13Hz

after

response to a spiral stimulus card

19.85

0.96

0.84

0.12

0.15

3.27

1.56

279.41

0.04

0.00

0.00

0.03

-0.01

0.16

-0.65

Becker (2005)

p.231

FC9

3

light flicker 16Hz

after

response to a spiral stimulus card

19.08

1.31

0.90

0.06

0.08

2.38

1.27

94.52

0.04

0.00

0.00

0.01

-0.01

0.12

-0.35

Becker (2005)

p.232

FC9

4

light flicker 18Hz

after

response to a spiral stimulus card

16.61

2.28

0.84

0.13

2.01

403.18

0.15

0.52

0.04

0.00

0.00

0.09

0.00

0.00

-0.02

Becker (2005)

p.232

FC9

5

light flicker 18Hz

after

response to a spiral stimulus card

19.75

0.60

0.85

0.08

0.09

3.41

1.12

281.26

0.05

0.00

0.00

-0.02

0.00

0.02

-0.22

Becker (2005)

p.232

FC9

6

light flicker 19Hz

after

response to a spiral stimulus card

18.73

2.41

0.89

0.10

0.19

3.27

1.34

308.41

0.04

0.00

0.00

0.00

0.01

-0.17

0.57

Becker (2005)

p.239

FC9

7

light flicker 16Hz

after

response to a zig-zag stimulus card

6.94

2.42

0.82

0.14

1.28

86.34

0.50

0.04

0.04

0.02

-0.22

0.70

-4.21

4.01

-1.80

Becker (2005)

p.230

FC10

1

light flicker 15Hz

after

asked to look for spirals

14.61

1.24

0.83

0.08

0.20

2.66

2.14

45.26

0.04

0.00

0.01

-0.06

-0.02

0.13

0.02

Becker (2005)

p.230

FC11

1

light flicker 15Hz

after

asked to look for spirals

9.12

2.16

0.92

0.28

2.36

91.42

0.31

5.22

0.03

0.01

-0.17

1.02

0.47

-1.49

0.70

Becker (2005)

p.231

FC11

2

light flicker 16Hz

after

asked to look for spirals

15.65

1.95

0.90

0.08

0.23

3.55

1.15

161.36

0.04

0.00

0.00

-0.02

0.02

-0.23

0.67

Becker (2005)

p.231

FC11

3

light flicker 17Hz

after

asked to look for spirals

27.42

0.74

0.45

0.10

0.05

1.10

1.11

515.99

0.04

0.00

0.00

0.00

0.00

-0.04

0.34

Page/figure

spiral numbers

Style data

Object number

Source

cause

drawn during/ after experience

Object notes

path width trend coefficients

spacing trend coefficients

298 Context data

rotation

orientation

curvilinearity

irregularity

path width median

path width trend error

spacing median

spacing trend error

sampling resolution

Becker (2005)

p.231

FC11

4

light flicker 17Hz

after

asked to look for spirals

8.40

0.53

0.96

0.40

1.91

90.58

1.20

1.00

0.03

-0.02

0.12

0.04

0.05

0.73

-0.17

Becker (2005)

p.232

FC11

5

light flicker 19Hz

after

asked to look for spirals

9.96

0.40

0.67

0.29

1.22

26.19

2.07

12.60

0.04

0.00

0.00

0.01

-0.03

0.19

0.09

Becker (2005)

p.230

FC12

1

light flicker 15Hz

after

response to a spiral stimulus card

8.92

0.06

0.43

0.11

0.15

1.72

4.11

2.64

0.04

0.00

-0.02

0.06

-0.34

1.09

-0.51

Becker (2005)

p.230

FC12

2

light flicker 15Hz

after

response to a spiral stimulus card

9.07

2.15

0.87

0.04

0.18

2.16

1.07

1.71

0.04

0.00

-0.03

0.11

0.02

-0.24

0.37

Becker (2005)

p.231

FC12

3

light flicker 16Hz

after

response to a spiral stimulus card

19.73

0.82

0.82

0.17

0.10

1.53

1.70

99.79

0.04

0.00

0.00

0.01

0.00

-0.07

0.47

Becker (2005)

p.231

FC12

4

light flicker 17Hz

after

response to a spiral stimulus card

6.45

0.66

0.87

0.53

0.16

0.46

6.34

0.01

0.04

0.01

-0.06

0.13

-174.60

37.70

1.03

Becker (2005)

p.232

FC12

5

light flicker 18Hz

after

response to a spiral stimulus card

7.22

0.56

0.91

0.25

0.40

1.73

5.01

0.53

0.04

0.00

-0.01

-0.05

2.50

-4.85

2.69

Becker (2005)

p.232

FC12

6

light flicker 18Hz

after

response to a spiral stimulus card

9.03

1.07

0.43

0.11

0.18

1.48

3.72

2.23

0.04

0.00

-0.02

0.06

-0.37

1.43

-1.34

Becker (2005)

p.232

FC12

7

light flicker 19Hz

after

response to a spiral stimulus card

8.96

0.25

0.83

0.11

0.16

1.10

3.48

1.00

0.04

0.00

0.00

0.01

-0.29

1.13

-1.12

Ermentrout & Cowan (1979)

fig.1c

FC13

1

n/a

5.92

3.01

0.84

0.42

0.70

2.38

5.19

n/a

0.04

0.00

0.01

0.02

0.00

0.00

0.00

Tass (1997)

fig.6c

FC14

1

n/a

6.91

2.05

0.85

0.87

1.95

1.55

9.31

n/a

0.04

0.01

-0.10

0.78

0.00

0.00

0.00

Tass (1997)

FC14

2

n/a

4.70

1.56

0.85

0.67

2.83

9.48

9.17

n/a

0.05

0.14

-0.76

2.17

0.00

0.00

0.00

Tass (1997)

fig.6d fig.11, col.1

FC15

2

theoretical model theoretical model theoretical model theoretical model

n/a

10.79

1.64

0.80

0.78

9.28

6.80

14.31

n/a

0.06

0.00

-0.03

1.85

0.00

0.00

0.00

Tass (1997)

fig.11, col.4

FC16

3

theoretical model

n/a

4.86

0.90

0.89

0.70

2.60

22.92

6.68

n/a

0.04

-0.04

0.27

0.37

0.00

0.00

0.00

FC18

2

n/a

8.49

2.96

0.84

0.56

2.62

7.29

4.48

0.99

0.05

-0.01

0.08

0.33

0.04

-0.22

2.40

FC18

3

7.72

0.16

0.87

0.53

2.95

16.41

3.71

10.08

0.04

-0.05

0.55

-0.78

-0.37

1.63

0.37

FC18

4

n/a

7.18

0.30

0.92

0.68

2.86

14.72

5.71

3.26

0.04

-0.01

0.11

0.34

-0.28

0.63

3.27

FC18

5

n/a

7.77

1.65

0.93

0.53

2.90

37.94

5.26

9.71

0.04

-0.02

0.16

0.43

0.79

-2.69

4.57

Page/figure

spiral numbers

Style data

Object number

Source

Tass (1997) Tass (1997) Tass (1997) Tass (1997)

fig.24 c1,r3 fig.24 c2,r3 fig.24 c4,r2 fig.24 c4,r4

cause

drawn during/ after experience

Object notes

theoretical model theoretical model theoretical model theoretical model

n/a

series of plots includes spiraloids

path width trend coefficients

spacing trend coefficients

299

Tass (1997) Tass (1997)

4

FC19

5

FC20

1

FC20

2

FC20

3

FC20

4

FC20

6

FC20

8

FC20

9

theoretical model theoretical model theoretical model theoretical model theoretical model theoretical model theoretical model theoretical model theoretical model

sampling resolution

Tass (1997)

FC19

spacing trend error

Tass (1997)

theoretical model

spacing median

Tass (1997)

3

path width trend error

Tass (1997)

FC19

path width median

Tass (1997)

2

irregularity

Tass (1997)

fig.26 c2,r3 fig.26 c4,r2 fig.28 c1,r1 fig.28 c1,r2 fig.28 c1,r3 fig.28 c1,r4 fig.28 c2,r2 fig.28 c2,r4 fig.28 c4,r3

FC19

curvilinearity

Tass (1997)

1

orientation

fig.26 c2,r1

FC19

cause theoretical model theoretical model

Style data

rotation

Tass (1997)

spiral numbers

Page/figure Tass (1997)

fig.26 c1,r1 fig.26 c1,r3

Tass (1997)

Object number

Context data Source

n/a

7.35

1.87

0.83

0.64

2.94

1.35

5.08

1.21

0.06

-0.01

0.12

0.43

-0.89

1.81

2.32

n/a

6.99

0.31

0.84

0.62

2.72

0.96

4.98

0.96

0.05

-0.01

0.10

0.40

-0.42

0.83

2.26

n/a

7.50

1.75

0.89

0.58

2.45

12.42

4.93

3.31

0.05

-0.02

0.25

-0.03

-1.55

4.64

-0.64

n/a

7.01

0.34

0.81

0.61

2.75

15.02

4.69

0.29

0.05

0.02

-0.23

1.17

1.68

-3.51

3.97

n/a

6.09

0.28

0.91

0.59

3.06

2.11

4.62

0.13

0.04

0.00

-0.01

0.97

-0.89

-0.82

4.64

7.25

1.83

0.83

0.67

2.52

1.54

4.80

2.04

0.05

-0.01

0.17

0.14

-0.10

0.79

1.34

n/a

7.06

2.58

0.74

0.66

3.03

5.65

5.02

0.39

0.04

-0.02

0.20

0.26

-0.07

0.20

2.58

n/a

6.77

0.24

0.85

0.68

2.74

1.24

4.48

0.23

0.04

0.02

-0.24

1.24

-0.27

0.99

1.48

n/a

6.87

0.91

0.40

0.61

2.46

3.39

4.12

1.94

0.04

-0.01

0.11

0.34

0.23

-0.41

1.93

n/a

7.70

2.10

0.45

0.63

2.21

5.61

4.93

3.97

0.04

0.02

-0.17

0.90

0.44

-1.41

4.14

n/a

7.67

0.27

0.89

0.61

2.57

15.21

5.51

2.80

0.05

-0.01

0.16

0.20

-0.34

1.10

2.57

n/a

7.31

0.91

0.78

0.69

2.65

14.38

4.99

5.06

0.04

-0.02

0.17

0.21

-0.09

0.40

2.59

drawn during/ after experience

n/a

Object notes

series of plots includes spiraloids

path width trend coefficients

spacing trend coefficients

300

Euler spirals

JKAE19.1

parabolic

Archimedes’

logarithmic

JKAE14.3 JKAE15.2

hyperbolic

JBM2.1

JBM1.1

JKAE21.1

JKAN12.2

JKAE3.1

JKAE9.2 JBM4.1

JKAE26.2 JKAE16.2 JKAE25.2

JKAE22.2

JKAN10.1 JKAE16.1

JBM2.2

JKAE31.1 JKAE39.1

JKAE1.1 JKAN5.1

JKAE37.2

JKAE17.1

JKAE12.3 JKAE14.1 JKAE17.2 JKAN18.2

JKAN16.2 JKAN3.1 JKAE7.1 JKAE4.2 JBM4.2

JKAE25.1

JKAN10.2

JKAE10.2 JKAN4.1

JKAN11.1 JKAN1.1

JKAE24.1 JKAN3.2

JKAE33.1

JKAE27.1

JKAE34.2 JKAE28.1

JKAE37.1

JKAE18.1

JKAN9.1

JKAE40.1

JBM3.1

JKAE9.1 JKAE20.1

JKAN8.1

JKAE26.1

JKAN6.2

JKAE29.1 JKAE36.1 JKAE21.2 JKAN13.1 JKAE4.1 JKAE6.1 JKAE13.1

JKAE33.2 JKAE38.1 JKAN16.1

JKAE7.2 JKAE15.1

JKAE22.1

Table A2.7 Classification of Jōmon period spirals according to correlation with mathematical spiral types.

JKAN2.1

JBM3.2

JKAN18.1 JKAE36.2 JKAE35.2

JKAE8.1

JKAE34.1 JKAE13.2 JKAE35.1 JKAE10.1 JKAN17.2

JKAN14.1

JKAN7.2

JKAE8.2

JKAN12.1 JKAE11.1

JKAN17.1 JKAE12.1

JKAN15.1

JKAN6.1

JKAE19.2

JKAE30.1

JKAE28.2

JKAE5.1 JKAE23.1

JKAN7.1

JKAE2.1

301

Euler spirals

CSVAM1.1 CSVAM7.4

parabolic Archimedes’ logarithmic

CSBM16.79

CSBM10.3 CSASH9.23

hyperbolic

CSBM2.2 CSBM26.78

CSBM17.9

CSBM30.46

CSASH9.54

CSBM3.18

CSBM3.62

CSBM15.1

CSASH4.4

CSBM6.2

CSASH11.1

CSBM9.69 CSASH11.62

CSBM1.2

CSBM12.14

CSBM11.1

CSBM16.1

CSVAM8.1

CSBM13.64 CSBM20.1 CSBM5.1

CSASH3.13

CSASH2.3

CSBM29.14

CSBM22.1 CSASH2.2

CSBM1.1

CSBM6.1

CSBM2.1

CSVAM9.16 CSBM30.72

CSBM25.7 CSBM17.2 CSBM14.41 CSVAM5.5

CSBM7.25 CSBM10.1

CSBM21.1 CSBM8.1

CSVAM1.2 CSBM15.2

CSASH7.1

CSVAM7.8

CSBM7.90 CSASH3.59

CSASH5.68

CSBM25.1 CSBM24.38

CSASH5.80

CSBM14.16

CSBM12.84 CSBM27.1 CSASH10.9

CSBM26.17

CSASH7.19

CSBM19.22 CSVAM9.32

CSVAM8.2

CSBM24.24

CSVAM6.1

CSBM29.37

CSASH6.4

CSASH4.3

CSBM21.37

CSVAM5.4 CSBM20.2

CSBM23.2 CSBM8.2

CSBM4.25

CSASH1.5

CSNMS1.11

CSVAM3.1 CSBM13.62 CSVAM3.2

CSASH10.4

CSBM18.56

CSBM9.46 CSVAM2.4

CSBM28.14

CSVAM6.2

CSBM28.99 CSBM22.3

CSBM4.56

CSBM5.3 CSBM27.2 CSNMS1.12

CSVAM4.2

Table A2.8 Classification of Shang/Western Zhou Dynasty spirals according to correlation with mathematical spiral types.

CSASH1.1

CSBM18.24 CSVAM2.3 CSVAM4.1

CSASH8.1

CSBM23.1

CSBM19.3

CSASH6.1

302

Euler spirals

EGPET21.10

parabolic Archimedes’

hyperbolic

logarithmic

EGPET22.1

EGPET7.6

EGDUR 1.4

EGBM1.1 EGLIV 22.2

EGLIV 28.4

EGLIV 27.2

EGPET17.5

EGPET14.10 EGPET2.25

EGPET24.8

EGPET26.11 EGPET19.8

EGPET27.1

EGPET21.6

EGPET22.4

EGPET25.6

EGPET29.3

EGLIV 29.7

EGBM5.2

EGPET15.1

EGDUR 6.11

EGPET16.4

EGPET27.3

EGPET23.3

EGPET29.2

EGPET8.3

EGLIV 29.4

EGPET25.8

EGLIV 27.1 EGPET13.4

EGPET5.3

EGPET16.9

EGPET8.2

EGBM2.9

EGPET5.2

EGBM3.8

EGPET12.6

EGPET28.6

EGBM1.10 EGPET10.2

EGDUR 1.1

EGPET26.5

EGPET28.3

EGPET14.12

EGBM2.6 EGLIV 24.16

EGPET3.2

EGDUR 4.6

EGPET4.1

EGPET18.1

EGPET32.2

EGBM4.1 EGNMS2.1

EGPET4.2

EGLIV 24.3

EGPET10.1

EGBM5.4 EGPET1.6

EGPET31.5

EGPET15.2

EGPET9.2

EGPET19.3

EGPET11.9

EGDUR 3.2

EGPET23.2

EGPET24.5

EGDUR 3.1

EGPET2.19

EGPET1.5

EGBM4.2 EGLIV 26.1 EGPET6.1

EGDUR2.2

EGDUR 6.7

EGPET12.1

EGPET30.2

EGPET31.1

EGDUR 2.1

EGPET3.1

EGNMS2.2 EGPET13.3

EGNMS1.8 EGBM6.1 EGBM6.2 EGPET18.2

EGLIV 21.1

Table A2.9 Classification of Egyptian scarab spirals according to correlation with mathematical spiral types. Spiral detail images with EGPET code copyright Petrie Museum of Egyptology University College London.

EGPET11.3

EGBM3.3 EGLIV 28.11

EGPET9.1 EGDUR 4.9

EGLIV 22.4

EGPET30.3

EGPET6.2

EGPET32.1

EGLIV 21.3

EGLIV 33.4

EGPET17.4

EGPET20.6

EGPET7.5

EGLIV 33.1

EGNMS1.7

EGPET20.5

303

Euler spirals

CYBM7.1

CYMM3.1

parabolic

Archimedes’

CYMM22.1

CYMM16.2 CYRAM1.1

CYBM5.1 CYMM25.1

logarithmic

CYBM36.6

CYMM10.2

CYBM34.3

CYMM8.1

CYMM6.1

CYMM16.1

CYMM1.2

CYMM10.1 CYBM27.1

CYRAM2.1

CYBM22.14 CYBM38.12 CYFM1.22

CYBM30.4

CYBM17.1

CYBM16.2 CYMM13.1 CYBM1.1

CYBM40.1 CYBM33.13

CYBM32.8

CYMM18.1 CYBM30.1 CYBM18.1 CYBM23.10

CYBM39.4 CYBM25.22

CYBM38.8 CYASH1.1

CYBM26.1

CYASH2.1

CYBM19.1 CYBM39.3

CYMM27.11 CYBM15.1 CYMM26.3 CYBM29.3 CYBM29.4 CYMM17.1

CYASH2.6

CYMM5.1

CYBM18.11 CYMM2.1 CYBM21.1

CYBM34.5

CYBM37.39 CYBM20.11

CYMM12.1

CYBM36.12

CYBM33.6

CYBM37.77

CYBM24.9 CYBM24.7 CYBM22.7 CYBM20.7 CYBM41.6 CYBM25.16

CYMM7.2

CYBM40.8

CYFM1.18 CYBM28.2

CYMM21.1 CYMM28.13 CYMM20.1 CYBM31.8 CYMM20.2

CYBM28.1 CYMM28.4 CYBM16.7 CYASH1.2

CYBM35.6 CYMM24.1

CYMM9.1 CYBM32.1

CYMM7.1

CYMM11.1 CYMM1.1

CYMM19.1

CYBM31.3 CYBM41.2

CYMM11.2

CYMM12.2

CYBM3.1

CYBM27.2

hyperbolic

CYBM23.9

CYBM26.2

Table A2.10 Classification of Bronze Age Cypriot spirals according to correlation with mathematical spiral types.

CYMM26.1 CYBM17.12 CYBM4.1 CYMM27.16 CYBM15.2 CYBM42.10

CYBM19.2

CYBM42.7 CYBM35.4 CYMM14.1 CYBM21.7

CYMM15.1

CYMM23.1

CYMM4.1

304

Class 1

Class 2

JKAE22.1 JKAE23.1 JKAE5.1

JKAN9.1 JKAN12.1 JKAE19.2

Class 3

JKAN10.2 JKAE1.1

JKAE16.1

JKAE2.1

JBM2.1 JBM1.1 JKAE31.1JKAE14.3 JKAE39.1

JKAE10.2 JKAE22.2

JKAE37.2 JKAN17.2 JKAE13.2 JKAE35.1

JKAE11.1 JKAN14.1 JKAE8.2

JKAE40.1 JKAN7.1 JKAE18.1

JKAE15.2 JKAE3.1 JKAE25.2 JKAE16.2 JBM2.2

JKAE10.1JKAN6.1 JKAN10.1

JKAN7.2

JKAE17.1

JKAE9.2

JKAN3.1

JBM4.1 JKAE12.3 JKAE14.1 JKAE17.2 JKAN5.1

JKAE26.2 JKAE21.1

JKAN4.1

Class 4

JKAN12.2 JKAN16.2

Class 5

JKAE19.1 JKAN8.1

JKAE9.1 JKAE25.1 JKAE33.1

JKAN11.1 JKAN3.2

JKAE27.1

JKAE28.1 JBM4.2

JKAE24.1 JKAE34.2 JKAE30.1 JKAE7.2 JKAN2.1 JKAE38.1 JBM3.2

JKAE28.2 JKAE8.1

JKAN15.1

JKAE7.1

JKAN1.1 JKAE4.2

JKAE15.1 JKAE34.1

JKAN16.1 JKAE35.2 JKAE33.2 JKAE4.1 JKAN6.2 JKAE12.1 JKAE6.1 JKAE20.1

JKAN17.1 JKAE21.2 JKAE13.1 JKAE36.1 JKAE29.1 JKAN18.1 JKAN13.1 JKAE26.1 JBM3.1

JKAE37.1

JKAE36.2

Table A2.11 Clustering of Jōmon Period spirals. Increasing numbers of frames show more distant clustering.

JKAN18.2

305

Class 1

CSASH8.1 CSVAM3.1 CSBM18.24 CSNMS1.11

CSBM4.25 CSBM28.99 CSBM8.2

CSBM23.2 CSVAM2.4 CSBM23.1

CSVAM3.2 CSVAM4.1 CSASH6.1 CSBM13.62

CSVAM9.32

CSASH1.5 CSVAM4.2 CSBM29.37 CSASH10.4 CSBM9.46 CSBM28.14

CSBM22.3 CSBM21.37 CSNMS1.12 CSVAM5.4

CSBM19.3 CSASH4.3 CSBM5.3 CSBM4.56

CSBM27.2

CSASH5.68CSBM24.38 CSVAM8.2 CSBM15.2

CSASH1.1 CSVAM6.2 CSBM20.2 CSBM26.17 CSBM19.22 CSASH7.19 CSVAM2.3 CSVAM6.1

CSASH7.1 CSVAM7.8

CSBM12.84 CSBM24.24 CSASH6.4

CSASH3.59 CSASH10.9 CSBM18.56

Class 2

CSBM12.14 CSBM22.1

CSBM1.1

CSASH2.2

CSBM16.1 CSBM30.72 CSBM7.25 CSVAM9.16 CSBM8.1 CSBM21.1 CSBM14.16 CSBM25.1

CSBM11.1 CSASH3.13 CSBM1.2

CSBM14.41 CSVAM8.1 CSBM10.1

CSBM5.1 CSBM13.64 CSBM27.1 CSBM7.90

CSASH5.80 CSVAM5.5 CSBM20.1

Class 3

CSBM15.1

Class 4

CSASH4.4

CSASH9.23

CSBM29.14

Class 5

CSVAM1.1

CSBM6.2 CSBM26.78 CSASH2.3

CSBM17.2 CSBM25.7

CSASH11.62 CSBM6.1 CSBM9.69

CSBM30.46 CSASH9.54

CSBM3.62 CSBM3.18

CSBM2.2 CSASH11.1 CSBM10.3

CSVAM1.2

CSVAM7.4

CSBM17.9 CSBM16.79

CSBM2.1

Table A2.12 Clustering of Shang and Western Zhou Dynasty spirals. Increasing numbers of frames show more distant clustering.

306

Class 1

Class 2

EGPET9.1 EGBM4.2 EGBM4.1 EGDUR 3.2

EGPET31.1

EGPET12.1

EGPET3.1

EGBM6.1

EGPET4.2 EGPET24.8

EGPET22.1

EGPET12.6

EGPET4.1

EGLIV 22.2 EGPET3.2

EGDUR2.2 EGNMS2.1 EGPET32.2

EGDUR 3.1

EGLIV 22.4

EGPET6.1 EGNMS2.2

EGPET18.1

EGPET14.10 EGPET18.2

EGPET24.5

EGPET9.2

EGPET11.3

EGLIV 27.1

EGPET31.5

EGPET13.4

EGPET10.1

EGPET23.2

EGPET25.8

EGPET32.1 EGPET2.19

EGDUR 2.1

EGPET5.3

EGLIV 26.1

EGPET10.2

EGPET5.2

EGPET28.6

EGBM1.10 EGBM3.8

EGPET19.3

EGBM5.4 EGPET1.6

EGPET15.2

EGLIV 24.3

EGPET19.8

EGBM6.2 EGPET26.5

EGPET14.12

EGPET13.3

EGDUR 1.1 EGPET28.3 EGPET30.2

EGPET11.9

EGDUR 4.6

EGPET26.11 EGPET25.6

EGPET2.25

EGLIV24.16

EGBM2.6 EGBM1.1 EGPET8.3

Class 3

EGPET22.4

Class 4

EGPET1.5

EGLIV 29.4 EGLIV 29.7

EGLIV 27.2 EGBM3.3

EGLIV 28.4

EGPET21.10

EGBM2.9

Class 5

EGPET7.5

EGLIV 21.1

EGPET20.5

EGPET6.2

EGPET30.3

EGPET20.6

EGPET17.5 EGBM5.2

EGDUR 1.4

EGPET16.4

EGPET7.6

EGLIV 28.11

EGPET16.9

EGNMS1.8 EGPET29.2

EGPET23.3

EGLIV 33.1 EGDUR 4.9 EGPET21.6

EGDUR 6.11

EGPET15.1

EGLIV 21.3 EGPET17.4

EGDUR 6.7

EGLIV 33.4 EGPET27.1

EGPET27.3

EGPET29.3

EGPET8.2

Table A2.13 Clustering of the Egyptian scarab spirals. Increasing numbers of frames show more distant clustering. Spiral detail images with EGPET code copyright Petrie Museum of Egyptology University College London.

EGNMS1.7

307

Class 1

CYMM27.11 CYBM19.1 CYASH2.6 CYBM18.1 CYBM23.9 CYMM26.1 CYMM26.3

CYBM24.7 CYBM25.22 CYBM24.9

CYMM27.16 CYBM29.4 CYMM14.1 CYMM4.1 CYBM42.7 CYBM42.10

CYBM38.12 CYBM38.8 CYFM1.22

CYBM19.2 CYBM29.3 CYBM15.1 CYBM15.2 CYBM35.4 CYBM39.3 CYBM17.12

CYBM16.2 CYBM25.16 CYBM30.4

CYASH2.1 CYBM23.10 CYBM30.1 CYMM15.1 CYMM17.1 CYBM34.5 CYBM18.11

CYBM22.7 CYMM13.1 CYMM18.1

CYBM4.1 CYBM21.7 CYBM26.2 CYMM2.1 CYBM21.1 CYBM26.1 CYMM23.1

CYMM22.1

CYBM5.1

CYBM34.3 CYBM32.1 CYBM35.6

CYMM19.1

CYMM25.1

CYBM40.8 CYBM33.13 CYBM32.8

CYBM3.1

CYBM36.6

CYBM40.1 CYBM33.6 CYMM9.1

CYMM6.1

CYBM36.12 CYBM17.1

CYASH1.1 CYBM22.14 CYBM1.1 CYMM28.4 CYBM41.6

Class 2

CYMM16.2 CYBM27.2

CYMM16.1 CYBM41.2

CYBM20.11 CYBM16.7 CYASH1.2 CYBM20.7

CYMM21.1

CYMM1.1 CYMM1.2

CYBM27.1 CYFM1.18 CYBM31.3

CYBM28.2

CYBM37.39 CYBM39.4 CYMM20.1 CYMM20.2

CYBM37.77 CYMM28.13 CYBM28.1 CYBM31.8

Class 3

CYMM11.1

CYRAM1.1

CYMM11.2

CYMM8.1

CYMM10.1

CYMM24.1

CYRAM2.1

CYMM5.1

CYMM3.1

CYBM7.1

CYMM10.2

Table A2.14 Clustering of the Bronze Age Cypriot spirals. Increasing numbers of frames show more distant clustering.

CYMM12.2

CYMM7.1

CYMM7.2

CYMM12.1

308

Type 1

Type 2

FC19.1 FC20.9 FC12.7 FC20.3

FC13.1 FC20.2 FC19.2

FC19.5 FC19.4 FC14.1 FC20.4 FC16.3

FC8.4 FC20.8 FC4.1

FC8.1

FC8.7

FC18.5 FC20.1

FC19.3 FC18.2 FC18.3

FC20.6 FC18.4 FC11.5 FC9.7

FC15.2

FC12.4

FC11.4 FC11.1

Type 4

Type 3

FC7.6

FC9.3

FC11.3

FC9.5

FC7.1

FC10.1

FC12.5

FC3.1

FC7.2

FC9.6

FC8.2

FC7.7

FC12.6

FC8.10

FC8.8

FC9.1

FC7.5

FC7.8

FC8.6

FC8.5 FC7.10

FC2.1

FC8.9

FC12.1 FC1.2

FC12.2

FC7.3

FC8.3

FC7.14

FC7.13

FC9.2

FC12.3

FC11.2

Type 5

FC7.9

FC7.4

FC7.15

FC6.1

FC5.1

FC9.4

FC14.2

Shang Dynasty

Cypriot CYMM18.1

CSASH11.1

Cypriot CYBM42.7

CYBM38.12

CYBM39.3

CYBM30.1

CYBM25.22

Egyptian

CYMM27.11 CYASH2.1

CYFM1.22

CYBM25.16

CYASH2.6

CYMM27.16

EGPET11.3

CYBM29.3

CYBM15.1

CYBM23.10

CYMM26.1

CYBM17.12

CYBM24.7

CYBM30.4

CYBM19.1

CYBM19.2

CYMM26.3

CYBM35.4

Shang Dynasty CSASH6.1

CSBM18.24

CSBM23.1 CSBM23.2 CSBM27.2

CSVAM2.4

CSBM8.2

Egyptian

EGLIV 33.1

EGLIV 33.4

EGPET20.6

Jōmon JKAE11.1 JKAE8.2 JKAE19.2

Figure A2.15 EVP type groups and corresponding decorative spirals. Spiral detail images with EGPET code copyright Petrie Museum of Egyptology University College London.

CSNMS1.11 CSBM28.99

CSVAM3.1

EGPET21.10

CSBM29.37

CYBM40.8 CYBM36.6

FC1.1

309

Class 1

Class 2

JKAE10.2

CSBM14.16

JKAE14.1

CSBM19.3

CSBM21.37

JBM3.1

CSBM6.1

CSBM30.46

CSASH2.2 JKAE7.1

CYMM1.1

CYMM7.1

CYMM1.2

CYMM12.2

Class 3

CSBM17.9

CSBM17.2 JKAN3.1

CSBM12.84

CSBM30.72 CYMM16.2

CSBM10.3

JKAN15.1

CYMM7.2

CSBM7.25

EGPET4.1

JKAE28.2

EGNMS2.2

EGPET4.2

CSBM28.14CYMM16.1

CSVAM9.16

CSBM29.14

CSASH11.1

EGPET6.2

CSBM15.1

CSBM1.1

JKAN2.1

CSBM22.1 CYMM21.1

CSBM21.1

CSBM26.78

CSBM22.3

CSASH2.3

JKAE35.2

CSBM2.1

CSBM4.25

JKAN6.1

CSBM6.2

CSBM16.1

CSBM1.2

CSASH11.62 CSBM2.2

CSBM19.22

EGPET32.2

CSBM26.17 JKAE35.1

CSBM10.1

CYMM20.2 CSBM14.41

CYMM20.1

EGBM4.1

EGBM4.2

CSASH3.13 JBM3.2

EGBM3.3

EGPET32.1

CSVAM5.4

EGBM3.8

EGPET6.1

JKAE14.3

plant Hathor head snake hair/headdress

EGNMS2.1

cicada bird’s wing whisker (of a mythical creature)

CYMM12.1

JKAE28.1

CSASH4.4

CSVAM5.5

Table A2.16 Clustering of spirals by representational meaning. Unfilled representational type markers indicate an uncertain attribution. Spiral detail images with EGPET code copyright Petrie Museum of Egyptology University College London.

horn taotie face decoration animal ear creature head tail (reptile/mythical creature) bird nostril shoulder joint (animal)

Loading...

Leeds Thesis Template - White Rose eTheses Online

1 The classification and analysis of spirals in decorative designs Alice Edith Humphrey Submitted in accordance with the requirements for the degree...

20MB Sizes 1 Downloads 0 Views

Recommend Documents

l - White Rose eTheses Online
THE ORIGINS OF THE SOCIAL WAR IN SOUTH YORKSHIRE -. A STUDY OF ..... revolutionary activity May-June 1817. 245. Table 5:

Augmented intimacies - White Rose eTheses Online
Beginning with Philip K. Dick's genre-defining Do Androids Dream of Electric Sheep? (1968) and ending ..... 22 H. Bruce

Susan Mary Giblin - White Rose eTheses Online
Jun 11, 1998 - Sukarno encouraged the leaders to accept the five principles of Pancasila as the. "philosophical basis" o

November 1996. - White Rose eTheses Online
thus making English readers familiar with Oriental modes of thought which were woven into the texture of the ...... Vers

SEEDS OF TRUTH - White Rose eTheses Online
Jan 21, 1987 - Alan Hunter. Submitted in accordance with the requirements for the degree of Ph.D., The University of Lee

nursing process - White Rose eTheses Online
process approach. 10. Additional advantages relating to the implementation of. 131 the nursing process approach as suppl

Download (20Mb) - White Rose eTheses Online [PDF] - Online free
6 vi CHAPTER SIX. SEF, FOUNDATION LEGEND OF FES 128 1. Legends of Foundation 129 1:1 The Legend of Fes 131 1:2 Other Fou

Fiction and the Theory of Action - White Rose eTheses Online
Table of Contents ii. Introduction. 1. Chapter One. Character Intention: Anna Kavan. 41. The Concept of Intention in The

Chhatrapati Shivaji Terminus, Mumbai - White Rose eTheses Online
with trunk lines to Calcutta, Madras and Delhi. Our architectural analysis places CST in the context of Victorian Gothic

housing policy in malaysia: conditions - White Rose eTheses Online
Land administration. 143. 4.3.9.1. Principle of "ihya-al-mawat". 145. 4.3.9.2. Principle of "al-igta"'. 146. 4.3.9.3. Pr