Archaeological Paleography: A Proposal for Tracing the Role of Interaction in Mayan Script Innovation via Material Remains 9781784912406, 1784912409

This volume explores the development of the Maya writing system in Middle-Late Formative and Early Classic period (700 B

140 75 8MB

English Pages 212 [218] Year 2016

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Revised proposal for encoding Naxi Dongba pictographic script

Yunnan Minority Language Commission, Lijiang Dongba Culture Institute. ISO/IECTC1/SC2/WG2 N4043, 2010. — 186 p.As “the o

401 99 1MB Read more

Proposal to encode Arabic-script letters for African languages

600 103 4MB Read more

Proposal to encode additional Arabic-script characters

716 71 8MB Read more

Innovation and Collaboration in the Digital Era: The Role of Emotional Intelligence for Innovation Leadership and Collaborative Innovation 9783110665383, 9783110665116

Innovation and Collaboration in the Digital Era provides a holistic approach to collaborative innovation, innovation man

299 49 3MB Read more

Innovation and Collaboration in the Digital Era: The Role of Emotional Intelligence for Innovation Leadership and Collaborative Innovation 9783110665383, 9783110665116

Innovation and Collaboration in the Digital Era provides a holistic approach to collaborative innovation, innovation man

105 36 14MB Read more

The Role of Advance Composite material In Contemporary Buildings

Composite materials have been used from the earliest times, from wood, which is a naturally occurring composite of lignin and cellulose, through straw reinforced clay bricks to reinforced concrete. In the 20th century, a new breed of composite materials was developed using polymer matrices with high performance reinforcement fibres. The great effect and uncompromising properties of advance composite materials has enabled the emergence of composites cut across all fields of application and all areas or work, just to mention a few aeronautic engineering, automobile engineering, and medicine, military and building construction. Therefore, with emphasis on building construction, advance composite material has played a vital role in today’s contemporary building construction method, by presenting its self as an alternative building construction material, its application has made the contemporary building construction much more flexible and achievable, compare to traditional building materials and its methods of construction. It further offers the building construction industry the technical know-how of having new possibilities of design styles, shapes and forms. Therefore, advance composite material proves it’s self to be a better and a new alternative building construction material that remains construction friendly and flexible based on its properties. This study therefore tends to provide an overview on advance composite material, its application as well as its role in today’s contemporary building. JOURNAL OF CONTEMPORARY URBAN AFFAIRS (2018), 2(3), 95-101. https://doi.org/10.25034/ijcua.2018.4723

0 0 1008KB Read more

The Role of Innovation and Entrepreneurship in Economic Growth 9780226810645

This volume presents studies from experts in twelve industries, providing insights into the future role of innovation an

277 36 18MB Read more

A Grammar of the Yucatecan Mayan Language

767 104 32MB Read more

A Grammar of the Yucatecan Mayan Language

1,829 297 2MB Read more

Roman Soldiers and the Roman Army: A study of military life from archaeological remains 9781407310008, 9781407322391

This study combines archaeological material from Romano-British forts located in northern Britain with concepts and meth

364 147 77MB Read more

Archaeological Paleography: A Proposal for Tracing the Role of Interaction in Mayan Script Innovation via Material Remains
9781784912406, 1784912409

Author / Uploaded
Joshua D. Englehardt

Table of contents :
Cover
Title Page
Contents
Preface
Acknowledgements
Chapter 1
Introduction
Research Objectives
Overview of the Investigation
Regional Context of Dataset
Analytic Units and Comparative Methodology
Interpretive Synthesis: Evaluating the Model
Statistical Methods of Quantitative Analysis
Organization of this Volume
Conclusions
Figure 1.1. Map of Mesoamerica. The micro–regional study area is outlined in the cross–hatched box.
Figure 1.2. The northwest Maya lowlands, detailing the micro–regional study area. The light grey overlay demarcates the northwest Maya lowlands of the Middle and Lower Usumacinta River basin in Tabasco, south–eastern Mexico (after Hernández Ayala 1981: 68
Figure 1.3. Detail of micro–regional study area and location of sites which provide ceramic evidence discussed in the text.
Chapter 2
Theoretical Framework and Methodological Premises
Modeling Interaction and Innovation in Ancient Societies
Systems and Complexity Theories
Theories of Network and Social Exchange
Examining Interaction, Integration, and Variability through Material Culture
Boundary Areas and Material Innovation
Interaction and The Development of Writing Systems in Mesoamerica
Writing and the Development of Writing Systems
Shared Features, Linguistic Encoding, and the Development of Mesoamerican Scripts
The Emergence and Nature of the Mayan Script
Evaluating The Relationship between the Development of Writing and Material Interaction in Formative Period Mesoamerica
Recontextualization
Interpretive Framework: Correlating Script Diversification and Material Change
Conclusions
Figure 2.1. A hypothetical lattice model of Middle Preclassic period scale–free interregional interaction networks, showing nodes of interaction (after Demarest 1989: 337, fig. 13.2).
Figure 2.2. An analytic classification of writing systems based on types of signs and symbols employed (adapted from Gelb 1963: 14, fig. 2). In the typology detailed above, ideography and pictography/iconography are classified as semasiographic scripts, w
Figure 2.3. Classification of Mesoamerican scripts (after Justeson et al. 1985; see also Coe 1976: fig. 1; Justeson 1986; Justeson and Matthews 1990; Marcus 1992a; Mora–Marín 2001: 444–46, figs. 1.7–1.9).
Figure 2.4. Acrophany and reformulation in Maya writing. a: T740 hu, hu, ‘iguana;’ phonetic sign; represents the upended head of a lizard or other reptile; b: T740:121.126 hu–li–ya, huliiy, intransitive verb, ‘arrived;’ c: T740.23 hu–na, hun, ‘paper,’ ‘bo
Figure 2.5. a: k’u (k’u) (T604) ‘nest;’ phonetic sign; b: k’u–xa–ja (k’uxaj) (T604:114.181) passive verb; ‘was eaten;’ ‘was ground;’ ‘was hurt.’ Drawings by Pearl Lau.
Figure 2.6. Lazy–S / cloud / T632 substitution set (drawing by Pearl Lau after Reilly 1996: 414, fig. 3).
Chapter 3
Figure 3.1. Map of the central and northwest Maya lowlands, showing sites included in this study and their location in relation to other Classic period Maya centers.
The Northwest Maya Lowlands: Site Selection and Regional Background
Regional Context
Location and Environment
Previous Investigations
Site Selection and Background
San Claudio
Tiradero
Mirador
Revancha
Conclusions
Figure 3.2. Distribution of early Mesoamerican script groups overlying distribution of Early Formative ceramic traditions (1150–850 BC). The bold black lines separate the Oaxacan, Southeastern, and Mayan script traditions. The light dashed lines indicate
Figure 3.3. Map of Mesoamerica. The extent of the Classic Maya area is roughly outlined in the light grey overlay, with traditional internal highland–lowland divisions noted. The central Maya core area of the Petén is highlighted in the grey cross–hatched
Figure 3.4. Relief map of the site of San Claudio, 1m contour. (González Moreno 2006: 32, fig. 30). Map by Mario Retíz.
Figure 3.5. Relief map of the site of Tiradero, 1m contour. Areas of excavation outlined in black cross hatched boxes. Map by Mario Retíz after Hernández Ayala 1981: 49, fig. 49.
Figure 3.6. Map of the site center of Mirador. Map by Mario Retíz after Hernández Ayala 1981: 54, fig. 55.
Figure 3.7. Relief map of the site of Revancha, 0.5m contour. Map by Mario Retíz after Hernández Ayala 1981: 58, fig. 60.
Ceramic Sample and Analytic Methods
Ceramic Sample
Archaeological Contexts of the Ceramic Sample
Sorting and Typing
Chronology and Phasing
Ceramic Sequence of the Lower San Pedro Mártir Basin
Middle Formative Period
Late Formative Period
Early Classic Period
Variables, Scale, and Analytic Units
Type–Variety
Form and Shape Class
Techno–Stylistic Attributes and Dimensions
Distribution
Comparative Analysis of Attribute Variability
Quantitative Analyses of Similarity and Diversity
ANOVA Cluster Analysis of Mean Attribute Similarity and Distance
H Score Measures of Diversity
Conclusions
Chapter 4
Figure 4.1. Middle Formative period ceramic type–varieties present in sample (n ≥ 10), showing quantities and group and ware associations (González Moreno 2006; Hernandez Ayala 1981).
Figure 4.2. Late Formative period ceramic type–varieties present in sample (n ≥ 10), showing quantities and group and ware associations (González Moreno 2006; Hernandez Ayala 1981).
Figure 4.3. Early Classic period ceramic type–varieties present in sample (n ≥ 10), showing quantities and group and ware associations (González Moreno 2006; Hernández Ayala 1981).
Figure 4.5. Plan of San Claudio Structure 1. Illustration by Mario Retíz after González Moreno 2006: 34, fig. 32.
Figure 4.6. Plan of San Claudio Structure 4. Illustration by Mario Retíz after González Moreno 2006: 35, fig. 33.
Figure 4.8. Plan of San Claudio Structure 12. Illustration by Mario Retíz after González Moreno 2006: 35, fig. 34.
Figure 4.9. Map detailing excavated areas at House 1, Tiradero. Illustration by Mario Retíz after Hernández Ayala 1981: 50, fig. 51.
Figure 4.10. Detail of excavated areas at the Tiradero ballcourt. Illustration by Mario Retíz after Hernández Ayala 1981: 52, fig. 54.
Figure 4.11. Floor plans of the three houses at Mirador in which explorations were undertaken and ceramic materials recovered. Illustration by Mario Retíz after Hernández Ayala 1981: 55, fig. 57.
Figure 4.12. Detail of excavations at the Mirador ballcourt. Illustration by Mario Retíz after Hernández Ayala 1981: 56, fig. 59.
Figure 4.13. Regional ceramic sequences and correlations for the Maya lowlands, with relative and absolute chronological correlation. (Adams 1971: 136, table 23; Hernández Pons 1984: fig. 5; Hernández Ayala 1981: 77; Holley 1987; Lee 1972; Muñoz 2004; Ran
Figure 4.14. Breakdown of quantities and percentages of five most common type–varieties present in sample at each site in the Middle Preclassic period. Percentages indicate proportions of selected type–varieties and totals in relation to the respective Mi
Figure 4.15. Breakdown of quantities and percentages of five most common type–varieties present in sample at each site in the Late Preclassic period. Percentages indicate proportions of selected type–varieties and totals in relation to the respective Late
Figure 4.16. Breakdown of quantities and percentages of five most common type–varieties present in sample at each site (a–d) in the Early Classic period. Percentages indicate proportions of selected type–varieties and totals in relation to the respective
Figure 4.17. The variable stylistic attributes and categories of those attributes that were observed and recorded on diagnostic artifacts within the sample.
Figure 4.18. Comparative interpretation of stylistic attributes. *At the regional scale, units refer to distinct ceramic traditions (or cultural groups). At the micro–regional scale, units refer to the four sites as a clustered whole (when compared with s
Ceramic Analyses
Results of Ceramic Analyses
Statistical Analyses of the Ceramic Sample
ANOVA Analysis of Middle Formative Period Ceramics
ANOVA Analysis of Late Formative Period Ceramics
ANOVA Analysis of Early Classic Period Ceramics
The H Score Diversity Measure
Summary of Statistical Analyses
Comparative Assessment between Assemblages at the Regional Level
Middle Formative Period
Late Formative Period
Early Classic Period
Summary of Comparative Analysis
Patterns of Interaction and Innovation Revealed through the Ceramic Analyses
Variability over Time
Variability through Space
Conclusions
Chapter 5
Interpreting the Results of the Comparative and Statistical
Figure 5.1. Results of ANOVA statistical analysis on Middle Formative period ceramic sample. (1)
Figure 5.2. Pie chart illustrating occurrences of specific type–varieties within the Middle Formative period ceramic sample.
Figure 5.3. Frequencies of Middle Formative period type–varieties within the sample.
Figure 5.4. Cross tabulation of form/shape by site, Middle Formative period.
Figure 5.5. Tukey HSD test for between site variability in Middle Formative period formal attributes.
Figure 5.6. Percentage of occurrence of specific form or shape class within the Middle Formative period assemblages.
Figure 5.7. Results of ANOVA statistical analysis on Late Formative period ceramic sample. (1)
Figure 5.8. Cross tabulation of ware by site, Late Formative period.
Figure 5.9. Bar graph illustrating site specific percentages of ceramic wares within the Late Formative period sample.
_GoBack
Figure 5.11. Tukey HSD test for between site variability in Late Formative period ceramic wares.
Figure 5.12. Percentage of occurrence of specific slip within the Late Formative period assemblages.
Figure 5.13. Percentage of occurrence of specific decoration types within the Late Formative period assemblages.
Figure 5.14. Percentage of occurrence of surface treatments within the Late Formative period assemblages.
Figure 5.15. Clustered boxplot displaying ranges of classificatory attribute variability during the Late Formative period.
Figure 5.16. Cross tabulation of form/shape by site, Late Formative period.
Figure 5.17. Bar graph illustrating site specific percentages of ceramic forms and shapes within the Late Formative period ceramic sample.
Figure 5.18. Clustered boxplot displaying ranges of formal attribute variability during the Late Formative period.
Figure 5.19. Percentage of occurrence of specific paste colors within the Late Formative period assemblages.
Figure 5.20. Percentage of occurrence of specific paste textures within the Late Formative period assemblages.
Figure 5.21. Clustered boxplot displaying ranges of paste attribute variability during the Late Formative period.
Figure 5.22. Clustered boxplot displaying ranges of morphological attribute variability during the Late Formative period.
Figure 5.23. Results of ANOVA statistical analysis on Early Classic period ceramic sample. (1)
Figure 5.24. Bar graph illustrating site specific percentages of ceramic wares within the Early Classic period sample.
Figure 5.25. Bar graph illustrating general breakdown and site specific percentages of individual type–varieties within the Early Classic period ceramic sample.
Figure 5.26. Cross tabulation of surface treatment by site, Early Classic period.
Figure 5.27. Percentage of occurrence of specific surface treatments within the Early Classic period assemblages.
Figure 5.28. Cross tabulation of decoration type by site, Early Classic period.
Figure 5.29. Percentage of occurrence of specific decoration types within the Early Classic period assemblages.
Figure 5.30. Clustered boxplot displaying ranges of classificatory attribute variability during the Early Classic period.
Figure 5.31. Tukey HSD test for between site variability in Early Classic period classificatory attributes. (1)
Figure 5.32. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along classificatory parameters of ware and type–variety.
Figure 5.33. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along classificatory parameters of slip and surface treatment.
Figure 5.37. Percentage of occurrence of specific paste textures within the Early Classic period assemblages.
Figure 5.38. Clustered boxplot displaying ranges of paste attribute variability during the Early Classic period.
Figure 5.39. Tukey HSD test for between site variability in Early Classic period paste attributes.
Figure 5.40. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along paste parameters of color and texture.
Figure 5.41. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along paste parameter of temper content.
Figure 5.42. Cross tabulation of form/shape by site, Early Classic period.
Figure 5.44. Tukey HSD test for between site variability in Early Classic period formal attributes.
Figure 5.45. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along formal parameters of form/shape and angles/flanges.
Figure 5.46. Clustered boxplot displaying ranges of morphological attribute variability during the Early Classic period.
Figure 5.47. Tukey HSD test for between site variability in Early Classic period morphological attributes.
Figure 5.48. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along morphological parameters of base and support type.
Figure 5.49. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along morphological parameter of lip type.
Figure 5.50. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along morphological parameters of wall and rim/neck type.
Figure 5.51. Cross tabulation of wall thickness by site, Early Classic period.
Figure 5.52. Percentage of occurrence of specific wall thickness within the Early Classic period assemblages.
Figure 5.53. Cross tabulation of vessel height by site, Early Classic period.
Figure 5.54. Cross tabulation of neck length by site, Early Classic period.
Figure 5.55. Clustered boxplot displaying ranges of dimension attribute variability during the Early Classic period.
Figure 5.56. Tukey HSD test for between site variability in Early Classic period dimension attributes.
Figure 5.57. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along dimension parameters of wall thickness and vessel height.
Figure 5.58. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along dimension parameter of neck length.
Figure 5.59. H score heterogeneity measures of assemblage diversity over time.
Figure 5.60. Comparative assessment between assemblages at the regional level, Middle Formative period (Mamom sphere), illustrating correspondence and discrepancies between sampled ceramics and adjacent sites and regions.
Figure 5.61. Comparative assessment between assemblages at the regional level, Late Formative period (Chicanel sphere), illustrating correspondence and discrepancies between sampled ceramics and adjacent sites and regions.
Figure 5.62. Comparative assessment of San Claudio assemblage at the regional level, Early Classic period (Tzakol sphere), illustrating correspondence and discrepancies between sampled ceramics from San Claudio and adjacent sites and regions.
Figure 5.63. Comparative assessment of Tiradero, Mirador, and Revancha assemblages at the regional level, Early Classic period (Tzakol sphere), illustrating correspondence and discrepancies between sampled ceramics from Tiradero, Mirador, and Revancha and
Figure 5.70. The spatial distribution of common Late Formative period ceramic attributes (e.g., thick waxy red, cream, and black slips, groove–incised, fluted, and striated decoration, flat–based shallow dishes and deep bowls with thick everted rims and d
Figure 5.71. The spatial distribution of divergent central (Tzakol; dark grey overlay) and developing northwestern (stippled light grey overlay) Maya lowland ceramic spheres in the Early Classic period. The study area is outlined in the black cross–hatche
Figure 5.72. Detail of the central Petén and northwestern Maya lowlands, illustrating the geographic extent of the Early Classic period Tzakol sphere (dark grey overlay) and the developing divergent ceramic tradition of the northwestern Maya lowlands (cro
Figure 5.73. Interpretation of patterns of interaction within and between units deduced from ceramic analyses for the Middle and Late Formative periods. *At the regional scale, units refer to distinct ceramic traditions (or cultural groups). At the micro–
Chapter 6
Comparative Analysis of Iconographic and Linguistic Evidence
Methodological Considerations: Tracing the Role of Iconographic Innovation in Script Development
Linguistic Data: Framing Interaction and Innovation
Linguistic Affiliation of the Ancestral System and the Impact of Mixe–Zoque on GLM Languages
Temporal Contexts of Interaction: GLM Linguistic Diversification and the Influence of Cholan–Tzotzilan
Archaeological–Linguistic Correlations
Textual Evidence Indicative of Linguistic Diffusion
Summary of Linguistic Evidence
The Visual Dataset: Iconographic Transformations and Mayan Script Development
Contextualizing the Evidence
Earth Bands and T23 na
Vegetal Bundles or ‘Torches’
Hand Motifs
The Lazy–S
Calendrical Notations
Human Foot Motif
Summary of Iconographic Evidence
Conclusions
Figure 6.1. The development of Mesoamerican writing systems (after Lacadena 2010).
Figure 6.2. Phonological aspects of early writing or ancestral script adopted by the Maya as compared with four Mesoamerican language families (after Lacadena 2010:36, table 3).
Figure 6.3. Mixe–Zoque loans into Greater Lowland Mayan languages. These loans are widespread in Mayan and other Mesoamerican languages, and probably reflect contact with the Olmec (Justeson et al. 1985: 23). This table provides only those loans into Lowl
Figure 6.4. Extent of Mixe–Zoque language area (L) and probable movement of Olmec ethnic groups/Mixe–Zoque language groups and Olmec artistic styles and ceramic technologies during the Middle and Late Formative periods (R); based on linguistic data and ar
Figure 6.5. Phylogenetic grouping of Mayan languages detailing glottochronological estimates for divergence (cf. Campbell 1984: 2–3, figs. 1–2; Justeson et al. 1985: 3, fig. 1).
Figure 6.6. Sign reformulation to reflect the m to b’ linguistic shift and problems of adaptation. The glyphs are identical except for the two circles with infixed cross–hatched patterns (T741) affixed to the protuberance in T19 and T21 (Englehardt 2005:
Figure 6.7. MS130, T548, and T528.
Figure 6.8. Epi–Olmec sign MS44 and Maya signs T23, T526, and T529. a: MS44 in different contexts; b: down–turning ground motif in Izapan art and Olmec iconography; c: Epi–Olmec ‘sun–at–horizon’ glyph collocation and its Maya equivalent; d: early examples
Figure 6.9. Frozen uses and continuing visual associations of T23. a: Palenque Tablet of the 96 Glyphs; b: Palenque Palace Tablet; c: unidentified text from Tikal; d: Ahuelicán greenstone tablet; e: Dos Pilas Stela 8 (after Mora–Marín 2001: 680, Fig. 7.13
Figure 6.10. Examples within the Mayan script of the reformulation of established signs to reflect new or alternate linguistic values. These examples show the versatility of single signs, and the range in which they can be reused or reformulated to reflec
Figure 6.11. Objects with Olmec–style iconography found in the study area. a: Unprovenanced Olmec low relief, currently in the Museo Municipal de Tenosique; b: Olmec–style incised lápida from Balancán, Tabasco; c: Olmec–style incised lápida from Emiliano
Figure 6.12. Spatial distribution of Mesoamerican down–turning ground or ‘basal band’ motif related to Mayan sign T23, Formative–Early Classic period, detailing iconographic or scribal affiliation.
Figure 6.13. Rough temporal distribution of the down–turning ground or ‘basal band’ motif related to Mayan sign T23 in distinct Mesoamerican iconographic and scribal systems.
Figure 6.14. Spatial distribution of Olmec vegetal bundle or ‘torch’ motif, Middle Formative period.
Figure 6.15. Subsequent iterations of the Olmec iconographic bound vegetal bundle or ‘torch’ motif in distinct Mesoamerican iconographic and scribal systems. a: CS 29 bundle element (L) and CS 12 torch element (R) in glyphic contexts on the Cascajal Block
Figure 6.16. Spatial distribution of Mesoamerican disembodied hand motifs, Formative–Early Classic period, detailing iconographic or scribal affiliation. Only securely provenienced examples are shown.
Figure 6.17. Rough temporal distribution of the outstretched, ‘thumbs up’ hand motif in distinct Mesoamerican iconographic and scribal systems.
Figure 6.19. Rough temporal distribution of the flat, outstretched hand motif in distinct Mesoamerican iconographic and scribal systems.
Figure 6.20. Rough temporal distribution of the ‘casting’ hand motif in distinct Mesoamerican iconographic and scribal systems.
Figure 6.21. Spatial distribution of Mesoamerican Lazy–S motif, Formative–Postclassic period, detailing iconographic or scribal affiliation.
Figure 6.22. Rough temporal distribution of the Lazy–S motif in distinct Mesoamerican iconographic and scribal systems.
Figure 6.23. Spatial distribution of Mesoamerican disembodied foot motif, Formative–Postclassic period, detailing iconographic or scribal affiliation.
Figure 6.24. Rough temporal distribution of the disembodied foot motif in distinct Mesoamerican iconographic and scribal systems.
Figure 6.25. Spatial distribution of possible Mesoamerican calendric count sign related to Maya Initial Series Introductory Glyph (ISIG), Formative–Early Classic period, detailing iconographic or scribal affiliation.
Figure 6.26. Rough temporal distribution of possible Mesoamerican calendric count sign related to Initial Series Introductory Glyph (ISIG) in distinct Mesoamerican iconographic and scribal systems (cf. Englehardt 2005: 478, figs. A4.38 and A4.39).
Figure 6.27. Rough temporal distribution of a possible ‘foliated ajaw’ motif in distinct Mesoamerican iconographic and scribal systems.
Chapter 7
Interpretation and Discussion: The Relationship Between Material Interaction, Innovation, and Script Development
Interpreting Variability: A Multi–Scalar Correlational Approach
Synthesis of Data
The Middle Formative Period
The Late Formative Period
The Early Classic Period
Summary
Correlations with the Developing Mayan Script
Evaluation and Implications
Conclusions
Chapter 8
Conclusions
Results and Implications of this Investigation
Assessment of the Proposed Model and Future Directions
Final Thoughts
Bibliography

Citation preview

Archaeological Paleography A Proposal for Tracing the Role of Interaction in Mayan Script Innovation via Material Remains

Joshua D. Englehardt

Archaeopress Pre-Columbian Archaeology 6

Archaeological Paleography A Proposal for Tracing the Role of Interaction in Mayan Script Innovation via Material Remains

Joshua D. Englehardt

Archaeopress Pre-Columbian Archaeology 6

Archaeopress Publishing Ltd Gordon House 276 Banbury Road Oxford OX2 7ED

www.archaeopress.com

ISBN 978 1 78491 239 0 ISBN 978 1 78491 240 6 (e-Pdf)

© Archaeopress and J D Englehardt 2015 Front and back covers: Mayan Glyphs, drawings by Pearl Lau.

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the copyright owners.

This book is available direct from Archaeopress or from our website www.archaeopress.com

Contents Preface�� vii Acknowledgements�� ix Chapter 1 Introduction��1 Research Objectives��2 Overview of the Investigation��4 Regional Context of Dataset�� 4 Analytic Units and Comparative Methodology��7 Statistical Methods of Quantitative Analysis��8 Interpretive Synthesis: Evaluating the Model��8 Organization of this Volume�� 9 Conclusions��10 Chapter 2 Theoretical Framework and Methodological Premises��12 Modeling Interaction and Innovation in Ancient Societies��12 Systems and Complexity Theories��12 Theories of Network and Social Exchange��14 Examining Interaction, Integration, and Variability through Material Culture��15 Boundary Areas and Material Innovation��18 Interaction and The Development of Writing Systems in Mesoamerica��19 Writing and the Development of Writing Systems��19 Shared Features, Linguistic Encoding, and the Development of Mesoamerican Scripts��22 The Emergence and Nature of the Mayan Script��24 Evaluating The Relationship between the Development of Writing and Material Interaction in Formative Period Mesoamerica��26 Recontextualization�� 28 Interpretive Framework: Correlating Script Diversification and Material Change��29 Conclusions��30 Chapter 3 The Northwest Maya Lowlands: Site Selection and Regional Background��32 Regional Context��33 Location and Environment��36 Previous Investigations�� 37 Site Selection and Background��39 San Claudio�� 39 Tiradero�� 41 Mirador�� 42 Revancha�� 43 Conclusions��45 Chapter 4 Ceramic Sample and Analytic Methods��46 Ceramic Sample��46 Archaeological Contexts of the Ceramic Sample��47 Sorting and Typing�� 54 Chronology and Phasing�� 55 Ceramic Sequence of the Lower San Pedro Mártir Basin��56 Middle Formative Period�� 59 Late Formative Period�� 60 Early Classic Period�� 61 Variables, Scale, and Analytic Units��64 Type–Variety�� 65 Techno–Stylistic Attributes and Dimensions��66 Form and Shape Class�� 66 Distribution�� 68 Comparative Analysis of Attribute Variability��68 Quantitative Analyses of Similarity and Diversity��70 i

ANOVA Cluster Analysis of Mean Attribute Similarity and Distance��71 H Score Measures of Diversity�� 72 Conclusions��74 Chapter 5 Interpreting the Results of the Comparative and Statistical Ceramic Analyses��75 Results of Ceramic Analyses��76 Statistical Analyses of the Ceramic Sample��76 ANOVA Analysis of Middle Formative Period Ceramics��76 ANOVA Analysis of Late Formative Period Ceramics�� 78 ANOVA Analysis of Early Classic Period Ceramics�� 87 The H Score Diversity Measure�� 97 Summary of Statistical Analyses��97 Comparative Assessment between Assemblages at the Regional Level��98 Middle Formative Period�� 100 Late Formative Period�� 105 Early Classic Period�� 111 Summary of Comparative Analysis��116 Patterns of Interaction and Innovation Revealed through the Ceramic Analyses��116 Variability over Time�� 117 Variability through Space��117 Conclusions��124 Chapter 6 Comparative Analysis of Iconographic and Linguistic Evidence�� 128 Methodological Considerations: Tracing the Role of Iconographic Innovation in Script Development��129 Linguistic Data: Framing Interaction and Innovation��130 Linguistic Affiliation of the Ancestral System and the Impact of Mixe–Zoque on GLM Languages��131 Temporal Contexts of Interaction: GLM Linguistic Diversification and the Influence of Cholan–Tzotzilan��133 Archaeological–Linguistic Correlations��136 Textual Evidence Indicative of Linguistic Diffusion��138 Summary of Linguistic Evidence��142 The Visual Dataset: Iconographic Transformations and Mayan Script Development��143 Contextualizing the Evidence��146 Earth Bands and T23 na�� 148 Vegetal Bundles or ‘Torches’��150 Hand Motifs�� 152 The Lazy–S�� 156 Human Foot Motif�� 158 Calendrical Notations�� 158 Summary of Iconographic Evidence��162 Conclusions��163 Chapter 7 Interpretation and Discussion: The Relationship Between Material Interaction, Innovation, and Script Development�� 165 Interpreting Variability: A Multi–Scalar Correlational Approach��165 Synthesis of Data��168 The Middle Formative Period��168 The Late Formative Period��169 The Early Classic Period��171 Summary�� 172 Correlations with the Developing Mayan Script��174 Evaluation and Implications��178 Conclusions��181 Chapter 8 Conclusions�� 183 Results and Implications of this Investigation ��184 Assessment of the Proposed Model and Future Directions��187 Final Thoughts��189 Bibliography�� 190

ii

List of Figures Figure 1.1. Map of Mesoamerica�� 2 Figure 1.2. The northwest Maya lowlands, detailing the micro–regional study area.�� 5 Figure 1.3. Detail of micro–regional study area and location of sites which provide ceramic evidence discussed in the text.�� 6 Figure 2.1. A hypothetical lattice model of Middle Preclassic period scale–free interregional interaction networks�� 14 Figure 2.2. An analytic classification of writing systems based on types of signs and symbols employed�� 20 Figure 2.3. Classification of Mesoamerican scripts�� 22 Figure 2.4. Acrophany and reformulation in Maya writing.�� 25 Figure 2.5. a: k’u (k’u) (T604) ‘nest;’ phonetic sign; b: k’u–xa–ja (k’uxaj) (T604:114.181) passive verb; ‘was eaten;’ ‘was ground;’ ‘was hurt.’.� 25 Figure 2.6. Lazy–S / cloud / T632 substitution set�� 25 Figure 3.1. Map of the central and northwest Maya lowlands�� 32 Figure 3.2. Distribution of early Mesoamerican script groups overlying distribution of Early Formative ceramic traditions�� 33 Figure 3.3. Map of Mesoamerica.�� 34 Figure 3.4. Relief map of the site of San Claudio�� 40 Figure 3.5. Relief map of the site of Tiradero�� 41 Figure 3.6. Map of the site center of Mirador�� 43 Figure 3.7. Relief map of the site of Revancha�� 44 Figure 4.1. Middle Formative period ceramic type–varieties present in sample�� 47 Figure 4.2. Late Formative period ceramic type–varieties present in sample�� 47 Figure 4.3. Early Classic period ceramic type–varieties present in sample�� 48 Figure 4.4. Sketch map of San Claudio Group II�� 49 Figure 4.5. Plan of San Claudio Structure 1�� 49 Figure 4.6. Plan of San Claudio Structure 4�� 49 Figure 4.7. Sketch map of San Claudio Group III..�� 50 Figure 4.8. Plan of San Claudio Structure 12�� 50 Figure 4.9. Map detailing excavated areas at House 1, Tiradero.�� 51 Figure 4.10. Detail of excavated areas at the Tiradero ballcourt�� 52 Figure 4.11. Floor plans of the three houses at Mirador in which explorations were undertaken and ceramic materials recovered.�� 53 Figure 4.12. Detail of excavations at the Mirador ballcourt�� 53 Figure 4.13. Regional ceramic sequences and correlations for the Maya lowlands�� 57 Figure 4.14. Breakdown of quantities and percentages of five most common type–varieties present in sample�� 59 Figure 4.15. Breakdown of quantities and percentages of five most common type–varieties present in sample�� 61 Figure 4.16. Breakdown of quantities and percentages of five most common type–varieties present in sample�� 63 Figure 4.17. The variable stylistic attributes and categories of those attributes that were observed and recorded�� 67 Figure 4.18. Comparative interpretation of stylistic attributes.�� 69 Figure 5.1. Results of ANOVA statistical analysis on Middle Formative period ceramic sample.�� 77 Figure 5.2. Pie chart illustrating occurrences of specific type–varieties within the Middle Formative period ceramic sample.�� 78 Figure 5.3. Frequencies of Middle Formative period type–varieties within the sample.�� 78 Figure 5.4. Cross tabulation of form/shape by site, Middle Formative period.�� 79 Figure 5.5. Tukey HSD test for between site variability in Middle Formative period formal attributes. �� 79 Figure 5.6. Percentage of occurrence of specific form or shape class within the Middle Formative period assemblages.�� 79 Figure 5.7. Results of ANOVA statistical analysis on Late Formative period ceramic sample�� 80 Figure 5.8. Cross tabulation of ware by site, Late Formative period.�� 81 Figure 5.9. Bar graph illustrating site specific percentages of ceramic wares within the Late Formative period sample.�� 81 Figure 5.10. Pie chart and bar graph illustrating general breakdown and site specific percentages of individual type–varieties.�� 82 Figure 5.11. Tukey HSD test for between site variability in Late Formative period ceramic wares.�� 82 Figure 5.12. Percentage of occurrence of specific slip within the Late Formative period assemblages.�� 83 Figure 5.13. Percentage of occurrence of specific decoration types within the Late Formative period assemblages.�� 83 Figure 5.14. Percentage of occurrence of surface treatments within the Late Formative period assemblages.�� 84 Figure 5.15. Clustered boxplot displaying ranges of classificatory attribute variability during the Late Formative period.�� 84 Figure 5.16. Cross tabulation of form/shape by site, Late Formative period.�� 85 Figure 5.17. Bar graph illustrating site specific percentages of ceramic forms and shapes within the Late Formative period.�� 85 Figure 5.18. Clustered boxplot displaying ranges of formal attribute variability during the Late Formative period.�� 86 Figure 5.19. Percentage of occurrence of specific paste colors within the Late Formative period assemblages.�� 86 Figure 5.20. Percentage of occurrence of specific paste textures within the Late Formative period assemblages.�� 87 Figure 5.21. Clustered boxplot displaying ranges of paste attribute variability during the Late Formative period.�� 87 Figure 5.22. Clustered boxplot displaying ranges of morphological attribute variability during the Late Formative period.�� 88 Figure 5.23. Results of ANOVA statistical analysis on Early Classic period ceramic sample�� 89 Figure 5.24. Bar graph illustrating site specific percentages of ceramic wares within the Early Classic period sample.�� 90 Figure 5.25. Bar graph illustrating general breakdown and site specific percentages of individual type–varieties�� 91

iii

Figure 5.26. Cross tabulation of surface treatment by site, Early Classic period.�� 91 Figure 5.27. Percentage of occurrence of specific surface treatments within the Early Classic period assemblages.�� 91 Figure 5.28. Cross tabulation of decoration type by site, Early Classic period.�� 92 Figure 5.29. Percentage of occurrence of specific decoration types within the Early Classic period assemblages.�� 92 Figure 5.30. Clustered boxplot displaying ranges of classificatory attribute variability during the Early Classic period.�� 93 Figure 5.31. Tukey HSD test for between site variability in Early Classic period classificatory attributes�� 94 Figure 5.32. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 95 Figure 5.33. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 95 Figure 5.34. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 96 Figure 5.35. Cross tabulation of temper content by site, Early Classic period.�� 96 Figure 5.36. Percentage of occurrence of specific paste colors within the Early Classic period assemblages.�� 97 Figure 5.37. Percentage of occurrence of specific paste textures within the Early Classic period assemblages.�� 98 Figure 5.38. Clustered boxplot displaying ranges of paste attribute variability during the Early Classic period.�� 98 Figure 5.39. Tukey HSD test for between site variability in Early Classic period paste attributes.�� 99 Figure 5.40. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along paste parameters of color and texture�� 100 Figure 5.41. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along paste parameter of temper content�� 100 Figure 5.42. Cross tabulation of form/shape by site, Early Classic period.�� 100 Figure 5.43. Bar graph illustrating site specific percentages of ceramic forms and shapes within the Early Classic period�� 101 Figure 5.44. Tukey HSD test for between site variability in Early Classic period formal attributes.�� 101 Figure 5.45. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 102 Figure 5.46. Clustered boxplot displaying ranges of morphological attribute variability during the Early Classic period.�� 102 Figure 5.47. Tukey HSD test for between site variability in Early Classic period morphological attributes.�� 103 Figure 5.48. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 104 Figure 5.49. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 104 Figure 5.50. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 104 Figure 5.51. Cross tabulation of wall thickness by site, Early Classic period.�� 105 Figure 5.52. Percentage of occurrence of specific wall thickness within the Early Classic period assemblages.�� 105 Figure 5.53. Cross tabulation of vessel height by site, Early Classic period.�� 105 Figure 5.54. Cross tabulation of neck length by site, Early Classic period.�� 106 Figure 5.55. Clustered boxplot displaying ranges of dimension attribute variability during the Early Classic period.�� 106 Figure 5.56. Tukey HSD test for between site variability in Early Classic period dimension attributes.�� 107 Figure 5.57. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 107 Figure 5.58. Extrapolated Tukey test illustrating means for groups in homogeneous subsets�� 108 Figure 5.59. H score heterogeneity measures of assemblage diversity over time.�� 108 Figure 5.60. Comparative assessment between assemblages at the regional level, Middle Formative period�� 109 Figure 5.61. Comparative assessment between assemblages at the regional level, Late Formative period�� 109 Figure 5.62. Comparative assessment of San Claudio assemblage at the regional level, Early Classic period�� 112 Figure 5.63. Comparative assessment of Tiradero, Mirador, and Revancha assemblages at the regional level, Early Classic period�� 112 Figure 5.64. Clustered boxplot displaying ranges of classificatory attribute variability over time.�� 118 Figure 5.65. Clustered boxplot displaying ranges of paste attribute variability over time.�� 118 Figure 5.66. Clustered boxplot displaying ranges of formal attribute variability over time.�� 119 Figure 5.67. Clustered boxplot displaying ranges of morphological attribute variability over time.�� 119 Figure 5.68. Diachronic attribute means plot, illustrating increasing variability over time.�� 121 Figure 5.69. The spatial distribution of common Middle Formative period ceramic attributes.�� 121 Figure 5.70. The spatial distribution of common Late Formative period ceramic attributes�� 122 Figure 5.71. The spatial distribution of divergent central�� 123 Figure 5.72. Detail of the central Petén and northwestern Maya lowlands�� 124 Figure 5.73. Interpretation of patterns of interaction within and between units deduced from ceramic analyses.�� 125 Figure 5.74. Interpretation of patterns of interaction within and between units deduced from ceramic analyses�� 126 Figure 6.1. The development of Mesoamerican writing systems.�� 129 Figure 6.2. Phonological aspects of early writing or ancestral script adopted by the Maya�� 132 Figure 6.3. Mixe–Zoque loans into Greater Lowland Mayan languages�� 133 Figure 6.4. Probable movement of Olmec ethnic groups/Mixe–Zoque language groups and Olmec artistic styles and ceramic technologies�� 134 Figure 6.5. Phylogenetic grouping of Mayan languages detailing glottochronological estimates for divergence �� 135 Figure 6.6. Sign reformulation to reflect the m to b’ linguistic shift and problems of adaptation�� 139 Figure 6.7. MS130, T548, and T528.�� 139 Figure 6.8. Epi–Olmec sign MS44 and Maya signs T23, T526, and T529�� 141 Figure 6.9. Frozen uses and continuing visual associations of T23�� 141 Figure 6.10. Examples within the Mayan script of the reformulation of established signs to reflect new or alternate linguistic values�� 146 Figure 6.11. Objects with Olmec–style iconography found in the study area�� 148 Figure 6.12. Spatial distribution of Mesoamerican down–turning ground or ‘basal band’ motif�� 149 Figure 6.13. Rough temporal distribution of the down–turning ground or ‘basal band’ motif�� 149 Figure 6.14. Spatial distribution of Olmec vegetal bundle or ‘torch’ motif, Middle Formative period.�� 151 Figure 6.15. Subsequent iterations of the Olmec iconographic bound vegetal bundle or ‘torch’ motif�� 151 Figure 6.16. Spatial distribution of Mesoamerican disembodied hand motifs, Formative–Early Classic period�� 152 Figure 6.17. Rough temporal distribution of the outstretched, ‘thumbs up’ hand motif�� 153 Figure 6.18. Rough temporal distribution of the flat, outstretched hand motif in distinct Mesoamerican iconographic and scribal systems.�� 153

iv

Figure 6.19. Rough temporal distribution of the ‘grasping’ hand motif in distinct Mesoamerican iconographic and scribal systems�� 154 Figure 6.20. Rough temporal distribution of the ‘casting’ hand motif in distinct Mesoamerican iconographic and scribal systems.�� 154 Figure 6.21. Spatial distribution of Mesoamerican Lazy–S motif, Formative–Postclassic period�� 157 Figure 6.22. Rough temporal distribution of the Lazy–S motif in distinct Mesoamerican iconographic and scribal systems�� 157 Figure 6.23. Spatial distribution of Mesoamerican disembodied foot motif, Formative–Postclassic period.�� 159 Figure 6.24. Rough temporal distribution of the disembodied foot motif in distinct Mesoamerican iconographic and scribal systems.�� 159 Figure 6.25. Spatial distribution of possible Mesoamerican calendric count sign�� 160 Figure 6.26. Rough temporal distribution of possible Mesoamerican calendric count sign.�� 160 Figure 6.27. Rough temporal distribution of a possible ‘foliated ajaw’ motif in distinct Mesoamerican iconographic and scribal systems�� 162

v

Preface

This research explores the development of the Maya writing system in Middle–Late Formative and Early Classic period (700 BC–AD 450) Mesoamerica. It seeks to correlate script development with interregional interaction and diachronic changes in material culture, and proposes a new methodological template for examining script development via material remains. In doing so, it contributes to anthropological debate regarding the role and effects of interregional interaction in processes of development and change of material and symbolic culture. This investigation posits that Maya writing developed in late Middle Formative through Early Classic period Mesoamerica as a correlate of interregional sociopolitical and economic interaction. Scholars working in many areas of the world have long claimed that interaction is central to cultural innovation, especially in relation to the development of writing. If the emergence of the Mayan script is a correlate of systemic interaction, then its developmental process should be traceable archaeologically through artifactual evidence. This hypothesis is tested by exploring archaeological indicators of interaction against a backdrop of previously documented transformations in the emerging Mayan script. The methodological model proposed here builds on current models of the development of Mesoamerican writing systems and models of interregional interaction and cultural development to associate archaeological remains with the development of the Mayan script. A significant revelation of this research is that the contextual framework in which material and symbolic goods were used and exchanged in past societies is equally as important as the formal qualities of the artifacts themselves in achieving a more complex understanding of their developmental histories and specific cultural meanings. This research represents a rare instance of investigation at the nexus of epigraphy, archaeology, and linguistic anthropology. Examining the development of writing in relation to stylistically defined zones of interaction permits more nuanced questions about the relationship between writing, other aspects of material culture, and cultural meaning. Archaeologists can infer cultural logics from artifactual exchange to create clearer links between material artifacts and symbolic concepts. The investigation shows how combining epigraphic, linguistic, and archaeological data can illuminate wider questions related to the development of sociopolitical complexity, cultural innovation, and long–term processes of linguistic and socio–cultural change, furthering anthropological debate in each sub–discipline. The primary merit of this work is that it adds to the increasingly nuanced understanding of emerging complexity in the Late Formative period of Southeastern Mesoamerica. This study underscores the effects of shifting networks of interregional interaction on lowland Maya material culture, linguistics, and scribal traditions. By examining the relationship between such transformations and material variability against a backdrop of changing sociopolitical organization at a crucial moment in Mesoamerican history, the model proposed here elucidates more complex understandings of larger archaeological questions related to boundary formation, emergent hierarchies, the development of specialized systems of material production, and the functional uses of ostensibly ‘elite’ material culture. The model is evaluated with ceramic data recovered from the archaeological sites of San Claudio, Tiradero, Mirador, and Revancha, located in southeastern Tabasco State, Mexico. This area is a boundary region between Mesoamerican interaction spheres. At such boundaries, ideas entwine with material goods in generative ways through interaction. In trans–regional contexts, differential interpretive principles prompt the emergence of innovative recontextualizations of artifactual elements, forms, and functions. This investigation analyzes stylistic, functional, and distributional variability in material markers of interaction to evaluate the relationship between material variability, interaction, and script development. The ceramic sample includes approximately 22,000 total sherds from the four sites dating from the late Middle Formative through the Early Classic period. Ceramics are an excellent variable by which to measure interaction and its relation to the development of the Mayan script. This is because ceramic materials readily display marked changes in style, form, and function that lend themselves well to comparative and quantitative evaluation against the morphological and functional changes is iconographic and linguistic evidence involved in the emergence of writing. Statistical measures of similarity and diversity within and between a sample dataset and regional sequences reveal quantitative patterns of change in ceramic materials. Patterns of continuity and disjunction in the formal stylistic characteristics and functions of material artifacts are compared to the distribution and recontextualization of shared iconographic elements across space and through time. Such patterns of exchange in the material evidence should correspond spatially and temporally to iconographic and linguistic data, suggesting both the centrality of interaction to cultural innovation and a correlation between the developmental processes. Stylistic and functional variability in the distribution of material artifacts indicates both degrees of interaction and differential use of homologous material culture in discrete functional and stylistic contexts,

vii

paralleling the emergent transformations of diffused icons into written signs. If the hypothesis is correct, an analysis of the data should reveal a greater degree of interaction and less relative variability between the ceramic sample and regional sequences in earlier temporal contexts, followed by a decrease in interaction and increase in material variation in the Early Classic as localized imposition of cultural meaning on icons and artifacts intensified in the Maya lowlands. Statistical measures of the archaeological evidence should indicate the significantly different functional and formal attributes of stylistically similar artifacts that parallel divergence in the localized use of diffused iconographic symbols. The emergence of Maya writing and the differential use of material culture would thus express the same processes of interregional interaction and innovation. Quantitative and comparative analyses suggest that interregional interaction was intimately involved with both material and scribal innovation in the contexts of the study area. Ceramic, linguistic, and iconographic data indicate that a great degree of interaction occurred within and across the study area in the Formative period and that interaction decreased significantly in the Early Classic period. The iconographic and linguistic data also suggest that the innovations involved in the development of the Mayan script occurred in earlier temporal contexts than the inventive and locally specific changes in regional material and ceramic traditions. The interaction evident in the Middle Formative data appears intimately connected to subsequent innovations in ceramic, linguistic, iconographic, and scribal traditions across Mesoamerica. The timing of such innovation, at least in the case of ceramic materials and the Mayan script, is variable. The three datasets suggest that patterns of interregional interaction shifted at slightly different times within the Late Formative period. Thus, the data hint that changes in broad patterns of interaction occurred in stages throughout the Late Formative period and consequently were reflected in the evidence at different points in time. The results also suggest that distinct types of data were linked to specific types of interaction whose patterns shifted in discrete spatial and temporal contexts. The first significant conclusion suggested by this research is the implication that large–scale changes in cultural processes within southeastern Mesoamerica may have occurred earlier than has previously been thought, closer to the Middle–Late Formative period transition. A second is the suggestion that subtle transformations in contextual frameworks may prove equally as integral to understanding processes of long–term cultural change as diachronic variation in the formal characteristics of material data. The results are complicated by the fact that so little evidence for the early history of Mesoamerican writing systems, including the Mayan script, is available. Further investigation may reveal new data, or suggest alternate lines of evidence that may be more profitably applied to an attempted correlation of developing scripts with material goods in Mesoamerican contexts. Alternately, the model proposed in this work may be applied in other spatial, temporal, or cultural contexts to elucidate the significance of the results suggested through this research.

viii

Acknowledgements This volume grew out of dissertation research conducted through the Department of Anthropology at Florida State University. First, I would like to offer my sincere gratitude to the members of my dissertation committee—Mary Pohl, Bill Parkinson, Joseph Hellweg, Daniel Pullen, and Michael Carrasco—for their support, encouragement, and patience. I am deeply indebted to all of these individuals, and I offer my most profound thanks for their guidance, encouragement, and diligent care. Also at FSU, I thank the Eisele Foundation for their generous support of my research and for funding fieldwork in Mexico during 2009 and 2010. I also offer thanks to the members of the FSU Department of Anthropology for their support and feedback, particularly to Dan Seinfeld and Ian Pawn. Thanks also go to Barb Speck and the Congress of Graduate Students for their support, and to the FSU Dissertation Research Grant program for providing funds to purchase statistical analysis software. I am also indebted to Kelly McGinnity, Felicia Gray, and Kunle Olumide and the FSU Department of Statistics Consulting Center for their advice and assistance with statistical analyses. While conducting my research in Mexico, I owe a debt of gratitude to many individuals. In Mexico City, thanks go to Lety Juarez Hernández, Joshua Balcells, and the family of Hector Muñoz Hernández. At the INAH Consejo de Arqueología, I extend thanks to Dolorez Juarez and Rach Cobos for their help and guidance in securing permission to conduct this research. At the UNAM Instituto de Investigaciones Antropológicas, my heartfelt gratitude goes to the late Dr Lorenzo Ochoa for his tremendous assistance and support of this investigation, and for opening the ceramic collection of the Proyecto Tierras Bajas Noroccidentales. Thanks also go to the faculty and staff of UNAM–IIA, particularly Rodrigo Liendo and Paty Peláez. In Villahermosa, Tabasco, I offer thanks to José Luis Romero and Juan Antonio Ferrer Aguilar for receiving me so warmly at the Centro INAH Tabasco. Also at the Centro INAH, thanks to the investigators and associated staff, particularly Paty Islas for her assistance with administrative matters. Special thanks to Ariadna, Ito, and Roger for making my stay in Villahermosa infinitely more bearable. I offer my sincere gratitude to Angela González Moreno for all of her help and encouragement in my work with the archived ceramic materials. Also in Villahermosa, many thanks to Rebeca Perales and her staff at the Museo Regional de Antropología Carlos Pellicer Cámara and the Instituto Estatal de Cultura del Estado de Tabasco for their support, guidance, and for allowing me access to museum collections during the renovations of summer 2009. Special thanks go to Javier Ruíz Torrecilla and the family of Don Hugo Ruíz for their hospitality. Finally, I am grateful to Pearl Lau and Mario Retiz for their considerable help in the preparation of images. All errors of fact or omission are the sole responsibility of the author.

ix

x

Chapter 1 Introduction This research explores the development of the Maya writing system in Middle–Late Formative and Early Classic period (700 BC–AD 450) Mesoamerica. It seeks to correlate script development in these contexts with interregional interaction and diachronic changes in material culture, and proposes a new methodological template for examining script development via material remains. In doing so, it contributes to anthropological debate regarding the effects of interaction on material and symbolic culture, and demonstrates how an investigation of writing that utilizes a combination of archaeological, linguistic, and epigraphic data may contribute to larger anthropological debate regarding the development of increasingly complex material culture, cultural symbolism, and sociopolitical configurations. In a broader sense, this study of script development offers useful parallels that may be applied to various questions of wider anthropological interest in multiple contexts, furthering knowledge across subfields and disciplines.

systems of material production, and the functional uses of ostensibly ‘elite’ material culture. The conjunctive study of the relationship between script development, material culture, and interregional exchange complements investigations of social innovation informed by cultural or linguistic anthropology. A close reading of the epigraphic and visual records, when examined in tandem with archaeological evidence, produces a richer picture of Maya society than one garnered from but a single source. This interdisciplinary approach illuminates wider questions related to social structure and change in both ancient and modern societies. If the emergence of Maya writing is a correlate of systemic regional interaction, then its developmental process should be archaeologically traceable through artifactual evidence and observable in material remains. Ceramic materials are examined in their capacity as archaeological indicators of interaction. Ceramics are an excellent variable by which to measure interaction and its relation to the developmental dynamics of the Mayan script. This investigation builds on current models of the development of Mesoamerican writing systems (Houston 2004a; Justeson 1986; Justeson et al. 1985; Justeson and Matthews 1990; Marcus 1992a; Mora–Marín 1997, 2001) and models of interregional interaction and cultural development in Mesoamerica (Blanton et al. 1996; Marcus 1992b, 1998; Willey 1991) to correlate evidence of incipient Maya writing with Mesoamerican regional interaction networks or ‘spheres’ (Freidel 1979). It is suggested that innovative reformulations of material, iconographic, and linguistic elements will occur at the boundaries of interacting regional systems. At such boundaries, users of a common iconographic system challenge symbols’ respective meanings through recontextualization, a process reflected in differing material forms and functions. If the model holds, archaeological evidence should indicate the differential functional and formal attributes of stylistically similar artifacts that parallel divergence in the localized use of shared iconographic symbols in similar temporal and spatial contexts. The emergence of writing and the differential use of material culture would thus parallel the same processes of interregional interaction.

Writing, that ensemble of scripts that textually represents language through codified signs and symbols, is a uniquely intriguing human technology. Writing developed within a horizon that witnessed the appearance of increasingly complex material culture in many cultural and spatial–temporal contexts (Bagley 2004; Baines 2004; Cooper 2004; Fischer 1989; Grube 1994; Keightly 1989; Lamberg–Karlovsky 1986; Schoep 1999). Yet in most contexts, the emergence of writing eludes adequate understanding, since there is scant evidence that speaks directly to incipient script technologies, and what little exists does not fit neatly into current models of the development of writing (e.g., Christen 2002; Houston 2004a, 2006). This study proposes a new methodological model for the investigation of the development of Maya writing, one that correlates script development with interregional sociopolitical and economic interaction. This model is tested by exploring the relationship between archaeological indicators of interaction and the emergence of Maya writing. Variation in the sociopolitical context of material change is examined against a backdrop of previously documented transformations in the emerging Maya writing system in order to associate archaeological remains with the development of the Mayan script.

To test the model, this research investigates the boundaries of Formative period Mesoamerican regional interaction networks for evidence of a relationship between incipient Maya writing and material interaction. The development of new styles and functions in the archaeological remains of interregional exchange should correspond spatially and temporally to iconographic and linguistic transformations that reflect cognitive

This attempt to model the development of writing on stylistically defined artifactual zones of interaction permits more pertinent questions about the relationship between material and ideological aspects of past societies, as well as the elucidation of more complex understandings of larger processes related to boundary formation, emergent hierarchies, the development of specialized 1

Archaeological Paleography changes in interpretation involved in and propelled by the transition from iconography to writing. A quantitative analysis of the distribution of the material evidence permits an assessment of the extent, temporality, and direction of interregional interaction. If the methodological template is sound, patterns of exchange in the material evidence should spatially and temporally correspond to iconographic and linguistic data indicative of interaction. Further, continuity and disjunction in the formal stylistic characteristics and functions of material artifacts is qualitatively examined and compared to the distribution and recontextualization of iconographic and linguistic elements across the same region. Temporal and spatial parallels between these two patterns of stylistic and functional distribution should take the form of differential use of homologous material culture in distinct, locally specific functional and stylistic contexts. Such patterns would suggest both the centrality of interaction to cultural innovation in this case and a correlation between the dynamics of the developmental processes.

presenting a subtler view of the myriad processes that brought about innovations and changes in these traditions at a critical transitionary period in Mesoamerican history (cf. Guernsey 2012; Rosenswig 2010). By examining the relationship between such transformations and material variability against a backdrop of changing sociopolitical organization, this investigation elucidates more complex understandings of the broader archaeological questions detailed above. This research therefore proposes a new way of thinking about and investigating potential connections between these larger investigative foci. Finally, it is suggested that the methodological template employed in this investigation may be profitably applied to a variety of anthropological questions in many different contexts. Research Objectives Building on received scholarship regarding the development of the Mayan script (Houston 2004a; Justeson 1986; Justeson and Mathews 1990; Justeson et al. 1985), it is suggested that the Maya writing system emerged in late Middle and Late Formative period Mesoamerica (Figure 1.1) as a correlate of interregional sociopolitical and economic interaction. Archaeologists have long recognized that interregional interaction sparks cultural innovation in many contexts (Cherry 1984; Demarest 1989; Flannery 1968a; Renfrew 1972), including the development of writing systems (Bagley 2004; Green 1989; Grube 1999; Justeson 1986). Likewise, researchers have effectively employed systemic complexity theory to elucidate correlations between innovations in material

In addition to testing the proposed theoretical model, a significant contribution of this research is the presentation of a new methodological template for the archaeological consideration of material interaction and its effects on long–term processes of sociocultural change. The primary merit of this work is that it adds to the increasingly nuanced understanding of emerging material, symbolic, and sociopolitical complexity in the Late Formative period of Southeastern Mesoamerica. This investigation underscores the effects of shifting networks of interregional interaction on lowland Maya material culture, linguistics, and scribal traditions by

Figure 1.1. Map of Mesoamerica. The micro–regional study area is outlined in the grey box.

2

Introduction culture and increasing interregional interaction between regional groups in a variety of contexts (e.g., Allen and Strathern 2005; Binford 1965; Cherry 1986; Demarest 2002; Flannery 1968b; O’Sullivan et al. 2006; Walby 2007). Models based on systemic complexity theory have thus proven strong frameworks from which to approach questions concerning innovation and emergent cultural technologies. Previous research has measured and evaluated the archaeological correlates of interaction at the level of the regional interaction sphere (Neitzel 2000; Peterson and Drennan 2003; Possehl 2007; Stein 2002). The term ‘interaction sphere’ describes a set of sites within a geographic region that, through some form of communication or interchange, share an assemblage of cultural traits evidenced in artifactual remains (Freidel 1979). To investigate variability across the boundaries of such interaction spheres, archaeologists have employed stylistic and distributional analyses of material variables, particularly ceramic assemblages (Hirth 1998; Hodder 1972, 1982, 1985; O’Shea and Milner 2002; Parkinson 2006: 34; Rice 1996a, 1996b; Skibo et al. 1989).

Similarly, current models of script emergence in Mesoamerica suggest that developmental processes are dependent on systemic interaction (Houston 2004a; Justeson 1986; Justeson and Mathews 1990; Justeson et al. 1985; Mora–Marín 2001). In these contexts, writing appears to have emerged as regional iconographic systems assumed different meanings through interaction in multi– lingual, trans–regional contexts. Previous research has offered interaction–based models for the development of writing in Mesoamerica (Chinchilla Mazariegos 1999; Fahsen Ortega 1999; Houston 2004a; Justeson 1986; Stross 1990). These models explain the emergence of Mesoamerican scripts in terms of a recontextualization of shared visual symbols within new linguistic contexts as a result of interregional iconographic interaction (Gelb 1963; Harris 1986; Salomon 2001; Senner 1989). Sustained interaction between groups employing a common iconographic complex juxtaposes linguistic values in unprecedented ways, leading to the linguistic codification of iconographic elements, usually as logographic, syllabic, or alphabetic elements. Fixing variable meanings to icons in the context of interregional interaction detached them from previous interpretations to facilitate their new use as written signs. Current scholarship thus suggests that Mesoamerican writing emerged from a ‘crisis of meaning’ as regional iconographic scripts assumed distinct and increasingly specific significative values through interaction in trans–regional contexts.

The association of style with interaction is complex and highly contextualized, since material culture systems are historically situated phenomena (Stark 1998: 8–9). To address this dilemma, it is useful to approach the relationship between material culture and boundary from a dynamic perspective (Barth 1969; Kowalewski et al. 1983). Interregional variability in material culture is a function of multiple factors, including the relative degrees of integration, interaction, and interdependence of the cultural group or groups that occupy the landscape on either side of a social boundary. Measuring stylistic variability in material objects thus speaks to the relative permeability and social maintenance of the boundary itself. In turn, comparing patterns of variability over time yields clues regarding the temporal and spatial contexts of interaction and material innovation. These patterns may also be compared to diachronic changes in sociopolitical contexts. Such methods have been applied to questions of interaction and the relationship between material traditions with success in Mesoamerica (e.g., Cheetham 2007; Demarest and Sharer 1982; Neff et al. 1999) and elsewhere (e.g., O’Shea and Milner 2002; Parkinson 2006). Greater material and distributional uniformity across a border area suggests extended interaction on a wider scale and more relaxed structural integration, resulting in more fluid social boundaries. Conversely, increased variation in material assemblages indicates narrower interaction on a more localized scale, a greater degree of integration within larger cultural systems, and a less permeable, more established boundary. The patterns of stylistic and distributional variability observed in material data can be related to social processes of interaction and material innovation by focusing on similarities or changes in artifacts over space and through time (Green and Perlman 1985: 6; Parkinson 2006: 36; Skibo et al. 1989).

Building on this observation, it is suggested that in Late Formative and Early Classic Mesoamerica, the developmental dynamics of writing can be correlated with sociopolitical and economic interaction. This study does not suggest that writing always emerges from the crucible of interregional interaction, or that the identification of interaction in itself accounts for the development of writing. Rather, in this specific case it is suggested that documented linguistic and iconographic shifts in the developing Maya writing system (Englehardt 2005; Lacadena 1995) should correspond to transformations in the sociopolitical context within which they occurred (i.e., changes in the extent, direction, or intensity of interaction or the degree of integration in larger regional networks). Although tracing script differentiation differs from tracing linguistic divergence, the two processes can be comparatively examined and correlated. Historical linguistics shows that boundary areas are associated with innovation. The correlational approach suggested here proves a useful new avenue of investigation, but not one that will measure linguistic affiliation or differentiation in any direct way. Archaeological indicators of interaction and sociocultural change, however, may be compared with linguistic data to yield insights into the nature and processual dynamics of interaction and cultural innovation (Josserand and Hopkins 1999). Quantitatively, spatial and temporal patterns of exchange evident in the distribution of material remains should 3

Archaeological Paleography match similar patterns in the iconographic and linguistic data indicative of interaction. Qualitative variation in stylistic attributes and functions of material culture should correspond to similar formal and functional iconographic and linguistic variability. Localized innovative forms or functions of broadly distributed material culture should parallel the recontextualization of visual signs involved in the transition from iconography to writing. If a correspondence between the two patterns is evident, then it would follow that interregional interaction is at the root of both phenomena, and that a relationship exists between script development and material innovation in this context. The crisis of meaning behind the emergence of Maya writing does not represent interaction; rather, it stems from it. Continuity and disjunction in material remains suggest the operation of similar processual dynamics in both cases. In effect, qualitative stylistic and functional variability and quantitative distributional variation reflected materially should parallel documented iconographic and linguistic transformations suggestive of interregional interaction in the emergence of Maya writing.

Models of innovation and cultural development in Mesoamerica suggest that ideas entwine with material goods in generative ways through interaction. All artifacts, including glyphs as well as other material goods, represent ideas and cultural categories. In transregional contexts, differential interpretive principles prompt artifacts to assume meanings previously unassociated with them (Clark 2004). Innovative artifactual and iconographic forms and functions thus emerge. The qualitative and quantitative analysis of stylistic, functional, and distributional variability (Hodder 1978; Hurt and Rakita 2001; Stark 1998) in the emerging Mayan script and material markers of interaction permits a more nuanced evaluation of the developmental dynamics of the symbolic categories that simultaneously underlie material exchange and writing. The model thus proposes that stylistic, functional, and distributional variability in archaeological remains will indicate both degrees of interaction and differential use of material culture. These variables, in turn, will parallel the emergent transformations of diffused icons into written signs.

This attempt to synthesize information from distinct sources relies on the implicit assumption that script development, like linguistic history, is intimately connected with material cultural history (Josserand 1975: 501)—an underexplored avenue of investigation in Mesoamerican contexts. Texts and inscriptions may be understood in two ways: ‘as purveyors of content and as archaeological artifacts’ (Houston 2004c: 229). As Trigger (2004: 67) explains, all material culture is a product of knowledge systems inherited from the past and transformed as a result of the pursuit of greater technical efficiency or particular sociocultural interests. This investigation therefore examines the development of Maya writing in Formative Mesoamerica in a new light. Tracing the emergence of writing in archaeological and linguistic terms simultaneously will offer a more detailed understanding of this complex cultural technology. Just as writing develops out of a ‘crisis of meaning,’ material cultural forms and usages exhibit diachronic changes that indicate a similar process of localized imposition of unique meaning or function. In archaeological terms, the stylistic attributes, functions, and forms of shared or shared or ‘diffused’ material culture do not remain static over time.

In sum, this study seeks to resolve the following critical queries: 1) Where are regional boundaries located in late Middle Formative through Early Classic Mesoamerica?; 2) How do patterns of degree, extent, and directionality of interregional interaction reflected in material, iconographic, and linguistic data vary over time?; 3) Do patterns of interregional interaction correlate with script development or the emergence of increasingly complex cultural technologies?; and 4) Do diachronic changes in forms and functions of material culture spatially and temporally correspond to previously documented linguistic and iconographic transformations involved in the development of the Mayan script? In short, can the development of scripts—in general and in this specific case—be traced via material remains? Overview of the Investigation Regional Context of Dataset To test the proposed model and evaluate this methodology, a database of ceramic artifacts was assembled from collections previously gathered at sites in the Northwest Maya lowland region of southeast Tabasco State, Mexico, a boundary area between interacting regions in Formative Mesoamerica (Figures 1.2 and 1.3). If systemic interaction at the regional level is a catalyst for cultural innovation, then changes or advances in cultural technologies should be most evident at the boundaries of the interacting units within the regional system. The investigation considers data collected from four sites in the Middle Usumacinta and San Pedro Mártir river basins: Tiradero, San Claudio, Mirador, and Revancha. The sites selected for study are second and third tier centers that exhibit evidence of occupation dating from the Middle

This investigation will move beyond traditional studies of the development of the Mayan script by showing that in this case writing served as a ‘symbolic intensification’ (Hanks 1989) that both paralleled and encouraged material forms of intensification and production. The interaction of distinct, interdependent ethnic, linguistic, or cultural groups emerged as both a material and an ideational phenomenon, a symbolic exchange of artifacts and information. This study simultaneously draws on and refines current and previous scholarship to further anthropological understanding of the nature and development of writing, as well as its relationship with material culture and interregional interaction. 4

Introduction

Figure 1.2. The northwest Maya lowlands, detailing the micro–regional study area. The light grey overlay demarcates the northwest Maya lowlands of the Middle and Lower Usumacinta River basin in Tabasco, south–eastern Mexico (after Hernández Ayala 1981: 68, fig. 1). The micro–regional study area is outlined in the cross–hatched black rectangle.

Moreover, regional analyses of Classic period ceramics (e.g., González Moreno 2006; Hernández Ayala 1981; Hernández Pons 1984; Ochoa and Casasola 1991; Sánchez Caero 1979) suggest that the region in which the selected sites are situated was a nexus of interaction located on a frontier between Classic period Maya ceramic spheres. Such a boundary likely developed and existed in the earlier Formative period as well.

Formative through the Classic period (González Moreno 2006; Hernandez Ayala 1981; Ochoa 1983). These sites were chosen because they are situated along established trade routes at a boundary between distinct Mesoamerican sub–regions, including the lowland Maya area, that exhibit evidence of sustained, intensive interaction throughout the Formative and Classic periods (Clark and Cheetham 2002; Golden and Scherer 2006; Lee 1978, 1989; Price 1978). Previously identified regional interaction spheres in Mesoamerica include interior Chiapas, the western Maya area, the Gulf coast Olmec heartland of Tabasco and southeast Veracruz, and the Southern Maya Lowlands of the Petén (Demarest 1989: 337, fig. 13.2). These regions surround the study area to the southwest, south, west, and east, respectively.

Documented regional evidence suggests widespread iconographic and linguistic diffusion throughout the Formative period (Campbell et al. 1986; Fields 1991; Joyce et al. 1991; Justeson et al. 1985; Kaufman 1976; Lacadena and Wichmann 1999; Martin 2002; Reilly 1991; Schele 1999). Formative Mesoamerica saw the development of a number of distinct scripts (Houston 2004a; Justeson 1986; Stross 1990). Formal and functional similarities, as well as temporal overlap in usage and homogeneity of motifs, techniques, and materials across Mesoamerica suggest a relation among scripts. If interaction correlates with script development, then evidence of incipient script technologies should be encountered at the peripheries of cultural interaction spheres. The region in which the selected sites are located abuts areas in which early Mesoamerican scripts have been found, including San Andrés in the coastal plain of northwest Tabasco immediately to the northwest of the study area, and San Bartolo, in the lowlands of the Petén, approximately 100km to the east (e.g., Pohl et al. 2002; Saturno et al. 2006).

Previous investigations have revealed various homologies in Mesoamerican traditions of material culture, as well as shared aesthetic traditions, between artifacts from southeastern Tabasco and adjacent sub–regions (Ochoa 1983; Ochoa and Casasola 1991; Parsons 1986, 1988; Proskouriakoff 1968, 1971). Researchers often conceive of these homologies as the results of processes of interaction between independent yet interdependent entities, although the nature, character, duration, and dynamics of the exchange process are unclear (Flannery and Marcus 2000). Prior analyses of regional evidence intimate the presence of Mesoamerican exchange networks stretching far back into the Early Formative period (Braswell 2003; Ochoa 1983; Rosenswig 2010). 5

Archaeological Paleography

Figure 1.3. Detail of micro–regional study area and location of sites which provide ceramic evidence discussed in the text.

the broad temporal and spatial coverage necessary to address these issues. In those cases in which extensive regional survey has been undertaken, the data is not always readily available. The lowlands of southeastern Tabasco, on the other hand, have been investigated at such scales, and data from a long–term regional survey was made available for this investigation. Therefore, due to the location of the study area at a key historical crossroads of Mesoamerican trade and interchange, and the fact that a critical mass of material evidence from a prior regional investigation was accessible for study, the northwest Maya lowlands of southeastern Tabasco are an excellent test case in which to examine the material expression of larger, long–term patterns of interaction and the relationship between interregional interaction and the broad processes involved in the development of the Mayan script.

Insofar as early Mesoamerican writing likely developed out of a preexisting iconographic system, it is therefore noteworthy that many examples of Middle Formative period ‘Olmec–style’ iconography have been encountered within the study area (Ochoa 1978, 1983; Ochoa and Hernández 1975). Such finds indicate that, although evidence of early Maya writing is thus far absent in southeastern Tabasco, the region is far from a marginal area. This region was an integral part of a broader pattern of long–term interaction, situated on well–established riverine trade routes that played a vital role in interregional exchange (Navarrete 1978; Ochoa 1983; Ochoa and Vargas 1983; Vargas and Ochoa 1982). Thus, the focus area of this investigation is a test case in which to discern the broader patterns of interaction and exchange that occurred across this particularly well– situated region. Script development is itself a long–term process. To determine its association with material innovation and variability, it is necessary to study an area that will allow the identification of long–term processes at the regional scale in wide–ranging discrete temporal contexts. Despite their significance in wider processes of interaction, very few border areas in Mesoamerica have been investigated at a regional scale that provides

Data from both published sources and archaeological collections is utilized. This research focuses on one specific material variable, ceramics, as a correlate of the complex interaction that led to the emergence of Maya writing. Ceramics were chosen as the primary variable due to their extensive presence in the archaeological record. Ceramics are also an excellent base by which to 6

Introduction measure interaction (Demarest and Sharer 1982; Holley 1983, 1987; Neff et al. 1999; Pring 1977; Rice 1996a, 1996b; Skibo et al. 1989). A dataset of approximately 22,000 ceramic artifacts from the region was complied. The sampled artifacts date from the Middle Formative through the Early Classic and were recovered from systematic surface collection, securely identified archaeological contexts (i.e., burials, domestic middens, ritual deposits, construction fill) or stratified test pits (see González Moreno 2006: 87–95; Hernández Ayala 1981: 82–84). The ceramic sample includes examples of all group and type–variety classifications encountered at each site in quantities greater than ten.

at these sites in differing temporal contexts allowed for a characterization of the interaction that occurred across the boundary on which they are situated in both synchronic and diachronic terms, and on multiple scales. Variation in the material data permits an appreciation of the permeability and social maintenance of boundaries, adding resolution to inferences regarding regional interaction. Such stylistically defined zones of social interaction, and overlaps in archaeologically identified regional exchange networks, may parallel incipient or emerging linguistic or ethnic divisions, the presence of which offers a useful comparative baseline (e. g., Barth 1969; Josserand and Hopkins 1992; Justeson 1986; Lacadena and Wichmann 1999).

Analytic Units and Comparative Methodology

Formal variability in the ceramic sample was recorded and assessed along two parameters: the traditional type– variety system (Smith et al. 1960) and classification of vessel form and shape (Culbert and Rands 2007). Type–variety taxonomies are most useful for recording and evaluating variation in surface finish. Measuring subtle variability in the evolution of forms over time yields detectable changes in different vessel shapes across the sample. Such changes refine placement of a specific ceramic artifact within a given chronological sequence, since forms evolve more rapidly than types, and demonstrate a greater range of variation. Classes of forms identified in my sample include plates, jars, basins, ollas (common earthen cooking pots), tecomates (common spheric storage vessels), apaxtles (shallow flat–bottomed serving vessels with flaring sides), cajetes (flat earthen serving dishes), and cazuelas (pots or bowls with a nonrestricted opening, no neck, and no handles). Such categories generally address the function and the size of the vessels, as shape classes are closely related to function (Culbert and Rands 2007: 185; San Román Martín 2009). In recording temporal variation in form, shape classes are related to corresponding ceramic phases based on modifications over time, such as rim orientation, thickness of the walls, or occurrence of specific decorative techniques. To connect shape class and phase within the same analytic category, a concise description of the diagnostic characteristics that warrant the specific class–phase association was developed.

Measurable analytic variables in the sample include vessel types and forms, paste texture and content, color and content of tempers and slips, and stylistic attributes such as surface treatment, decoration and decorative technique, vessel diameter and thickness, and elaboration and standardization of shape and dimensions. These variable formal characteristics permit interpretation of functions, contexts of use and deposition, affiliations with specific ceramic groups, wares, complexes, and spheres, and differences in specific processes of manufacture (Rice 1996a, 1996b; Stark 1998). In order to address the issues of interaction and innovation, the sample database was compared against regional diagnostic collections to evaluate the relationship between material variability, interaction, and script development. Distributional variability in the stylistic and functional attributes of these material markers of interregional interaction was quantitatively evaluated across multiple analytic units: sites, micro–region (the area in which the sites are located), and region (by comparing micro–regional distribution with distributional patterns in adjacent areas). Diachronic variability in the distributional patterns of the data over time was then compared across the same analytic units. Simultaneously, documented formal stylistic and functional variability in the ceramic evidence were comparatively assessed by evaluating the sample against regional diagnostic and typological collections. Finally, the quantitative statistical and comparative analyses of changes in ceramic materials were related to iconographic and linguistic transformations involved in the emergence of writing across space and through time.

Subsequently, regional variation in the evidence is interpreted by comparing the recorded functional and stylistic attributes observable in the sample with established Mesoamerican ceramic sequences and typologies. The relative percentages and occurrences of each stylistic variable within the sample during each chronological period (Middle and Late Formative, Early Classic) were compared with the formal attributes of ceramics in adjacent sites and regions in corresponding temporal contexts. This preliminary comparison illustrates relative levels of sample richness (Kintigh 1984; Leonard and Jones 1989). Patterns of synchronic and diachronic correspondence in ceramic styles, types,

Stylistic and functional analyses have proven valuable tools in discerning the nature and extent of interaction observed in material markers (Cheetham 2007; Hodder 1978; O’Shea and Milner 2002; Parkinson 2006; Stark 1998). The diachronic regional stylistic evaluation of the formal and functional attributes of material evidence found at the sites permitted an evaluation of the degree and quality of variation observable in the collected data. Measuring grades of similarity and diversity within, among, and between ceramic assemblages 7

Archaeological Paleography and forms between the sample and adjacent sequences allow for an initial characterization of the degree and direction of interaction that occurred in and across the region in distinct chronological periods. The presence of widely shared stylistic attributes implies extended interaction, whereas formal or functional discrepancies in the material data suggest innovation in terms of differential manufacturing processes, the development of localized styles, or the introduction of new material technologies (Stark 1998).

derived from the sample were then compared with observed distributional proportions of type–varieties and forms (idealized variable composition; Garraty 2009: 160) across analytical units (site, micro–region, and region) and over time. Including statistical measures of both inter–assemblage similarity (cluster analysis) and intra–sample diversity (H Scores) nuances interpretations regarding interregional interaction presented here.

Statistical Methods of Quantitative Analysis

Following the formal classification, the comparative assessment permits an elucidation of patterns of stylistic similarity and variability in the material data over time. The stylistic analysis allows for a preliminary evaluation of the proposed model, in terms of identifying or confirming the boundaries of regional interaction spheres, the relative permeability of these boundaries, and the degrees of interaction that occurred across them. Comparing recorded formal variation in the sample with diagnostic typological collections of ceramics in adjacent areas illustrates vertical and horizontal linkages across archaeological remains through time and space. Functional variability and material innovation may be deduced from discrepancies in shape classes and site–specific distribution. Patterns of formal and functional homologies and variability reflected in the material data may thus be deduced. The presence of interregional interaction would be suggested by synchronic stylistic homologies observed in the majority of defined variables, reflected in formal and functional characteristics shared within both the sample and regional sequences. Conversely, fewer shared attributes would suggest less interaction and a greater degree of localized innovation. Such variations allow for the determination of the direction, intensity, and temporality of interaction on both regional and micro–regional scales. Finally, if interaction is behind the process of material innovation, intermediate forms in the sample should be observable, that is, artifacts that represent a blend of the stylistic characteristics observed in adjacent areal typological sequences.

Interpretive Synthesis: Evaluating the Model

For the quantitative study of attribute similarity and diversity, two statistical measures of the recorded stylistic variables outlined above were employed. In considering similarities in ceramic assemblages between sites, sub–regions, and regions, cluster analysis of variance (ANOVA) of the distributional patterns of individual analytic variables was used. Based on statistical measures of mean distance and similarity among specific recorded attributes, cluster analyses demonstrate similarity and variability between the formal attributes observed in the sampled assemblages and characteristics of ceramic materials in neighboring sites and regions (Kintigh 2002; Shennan 1988). Greater similarity suggests a greater degree of interaction between the sites or areas that exhibit such correspondences. Comparing differential cluster patterns of the analytic variables over time and at different scales elucidates the relationships between the sampled material assemblages and adjacent sites and regions. These patterns illustrate the nature, extent, and direction of interregional interaction reflected in the ceramic materials. At the same time, levels of internal multivariate diversity within sites, sub–regions, and regions were quantified and compared. The investigation adapts Garraty’s (2009) operationalization of the distributional approach to identifying interregional interaction (Hirth 1998). This method establishes a relative measure of diversity among the dataset and is well–suited for comparing collections within sites and regions (Garraty 2009: 161). A heterogeneity measure (H Score) was determined to quantify diversity and compare variability in the distribution of suites of distinct ceramic attributes within the sample (Garraty 2009: 160). The H Score calculation yields a single, positive score for each collection, which illustrates distance between the sample collections and overall ‘representative’ type percentages (Garraty 2009: 160; Kintigh 1984, 2002). The idealized base set of proportions stems from the percentages of the same categories present in typological collections diagnostic of ceramic sequences at sites and regions adjacent to the study area. Perfect heterogeneity, the global proportions of each category of ceramic type or form among the total of all collected samples, is the baseline against which distributional variability will be evaluated. H Scores

The quantitative analysis complements and adds resolution to the comparative stylistic assessment by elucidating the nature, temporal contexts, and extent of interaction deduced from the stylistic analysis. Statistical measures of similarity and diversity serve to quantitatively evaluate relative variability between and among samples and datasets and complement the comparative analysis. Similar cluster patterns of attribute similarity across assemblages and lower intra–assemblage diversity reflected in H scores would suggest greater distributional uniformity between and within material complexes, and therefore a relatively greater degree of interaction on a regional scale. These measures of similarity and diversity also speak to material innovation, insofar as greater inter–assemblage variability and increased diversity within complexes suggest the emergence of new forms and functions. Comparing clustered patterns of similarity 8

Introduction and H scores in distinct spatial and temporal contexts and on different scales thus illuminates variability in the quantity, intensity, and directionality of expanding interregional exchange networks. If interaction is behind observed variation, patterns of interaction suggested by the comparative and statistical analyses should closely correspond in temporal and spatial contexts.

interaction to cultural innovation in this case, and would also suggest a correlation between the dynamics of the developmental processes. Organization of this Volume Chapter Two presents the theoretical framework and methodological premises of this investigation. The current state of scholarship regarding models of sociocultural organization and change is discussed, as well as theories of interaction, innovation, and material variability, specifically those that focus on systems and complexity perspectives. The chapter also presents a discussion of current theoretical models regarding the development of writing and Mesoamerican scripts, outlining the basic argument for a relationship between material interaction, innovation, emergent cultural technologies, and the development of writing. The current consensus on the emergence of writing in Mesoamerica, and the Mayan script, is briefly presented, as is a discussion of recontextualization and how this concept may be related to processes of material interaction and innovation. The chapter concludes with a basic outline of the methodological framework of the investigation.

Patterns of diachronic variation in the material data derived through the comparative and statistical analyses are then compared against previously documented iconographic and linguistic transformations (e.g., changes in the form, syntax, distribution, and textual recurrence of widely shared iconographic motifs and identifiable signs within discrete script systems; see e.g., Englehardt 2005; Fields 1991; Josserand and Hopkins 1999; Justeson et al. 1985; Lacadena 1995; Reilly 1996; Schele 1999). If the model holds, patterns indicative of a systematic transition towards greater technical complexity, formal innovation, and functional variability over time in the material data should correlate with a greater use and complexity of the emerging Mayan script. Increased variability in and localized adaption of material forms should correspond to formal and functional iconographic and linguistic variations as icons transformed into written symbols and acquired new meaning. Variability would thus indicate innovation in terms of localized adaptations of shared cultural technologies or site–specific singularities. Stylistic and distributional similarity in the material data reflects the interaction that drives the developmental process, whereas discrepancies in form, function, or distribution suggest localized innovation, paralleling dynamics of diffusion and divergence evident in the development of Maya writing. Variability in material, iconographic, and linguistic data should exhibit strong spatial and temporal correspondence if in fact interaction drives both developmental processes.

Chapter Three outlines the process of site selection and discusses the regional background and context of the sites included in this study, with attention to their environmental, geological, and geomorphological settings, as well as a discussion of previous investigations in the study region and at the four specific sites that provide the ceramic sample under consideration, as well as the results of those studies as they relate to the current investigation. In Chapter Four, the ceramic sample and analytic methods employed in this study are presented. The chapter outlines the regional ceramic sequences considered in this study and discusses the in–site archaeological contexts of the micro–regional sample, as well as the procedures and methods surrounding its classification. Measurable analytic variables are presented, and chronology and phasing, type–variety classification, and form and shape class classification of the dataset as a whole is specified. The qualitative comparative and quantitative statistical methods—including ANOVA and H Score—employed in the analysis are then presented and explained.

The advantage of the interpretive approach employed here is that it allows for the reconciliation of the changing meanings of symbols with the analytical need to proceed from known symbols and meanings back to unknown ones, and to compare these to natural possibilities along the way (Clark 2004: 215–16). The comparative and quantitative analyses allow for a determination of the extent to which regional material interaction corresponds with the development of Maya writing. If a relationship exists between interaction, script development, and material innovation, patterns in the data within the regional and micro–regional analytic units will indicate a greater degree of interaction and less relative material variability between my sample and regional sequences in earlier Middle and Late Formative temporal contexts, followed by a decrease in interaction and increase in material variation in the Early Classic period as localized imposition of cultural meaning on icons and material artifacts intensified in the Maya lowlands. If such a pattern exists, it would confirm the centrality of

Chapter Five encompasses the interpretation and discussion of the results of ceramic analyses, including both comparative and quantitative methods, with a specific focus on the statistical analyses, detailing the spatial and temporal patterns suggestive of interaction and innovation that emerged over time and through space in the material data. Chapter Six, in turn, presents and considers a selection of linguistic and iconographic evidence that speaks to the relationship between interregional interaction and script development. 9

Archaeological Paleography Chapter Seven synthesizes the distinct datasets by comparing the results of the ceramic analyses presented in Chapter Five with the iconographic and linguistic shifts involved in the emergence of the Mayan script discussed in Chapter Six in order to evaluate possible correlations between script development, interregional interaction, cultural–technological innovation, and material variation. Departing from a multi–scalar correlational approach, this chapter encompasses an in–depth interpretation and discussion of the results of both the comparative assessment and statistical analyses of both datasets. The chapter concludes with a brief discussion of the implications of these results for the study of Maya writing in general.

regarding the origins of sociopolitical complexity. Finally, this investigation illuminates the dynamics inherent in processes of Mayan script formation and increasing interregional interaction on a more general level, and elucidates the functional relationships between these two phenomena. Both are topics that have long interested archaeologists. This research stands at the nexus of linguistic anthropology and archaeology. Its unique situation therefore offers a rare chance to correlate material and linguistic data and to evaluate the temporal depth of interregional interaction exhibited in distributions of material remains (Dahlin et al. 1987; Dixon 1997). The attempt to model the development of Maya writing on stylistically defined artifactual zones of interaction permits investigators to ask more pertinent questions about the relationship between writing, aspects of material culture, and cultural meaning, enhancing archaeological comprehension of the interplay between material and ideological aspects of past societies. Archaeologists may thus infer cultural logics from artifactual exchange to attain a more holistic understanding of the link between material artifacts and symbolic concepts in antiquity. Further, the correlational method employed here has the advantage of offering an alternative means of interpreting pre–colonial linguistic data in Mesoamerica—to determine the temporal depth of loan words, for example—that moves beyond glottochronological reconstruction (cf. Dixon 1994).

Chapter Eight concludes the study with an evaluation of the proposed model and methodology, a brief summary of the results of this investigation and their potential implications on a broad scale, and suggestions for future research and expansion of the present study. Briefly, the results of this research carry several potentially significant implications. First, the results suggest that large–scale changes in cultural processes may have occurred earlier in Mesoamerican history than has previously been thought, closer to the Middle–Late Formative period transition. Further, this study illustrates that subtle transformations in contextual frameworks may prove equally as integral to understanding processes of long–term cultural change as diachronic variation in the formal characteristics of material data. Finally, this study provides a novel methodological template that may prove useful for investigations of script development and diversification in both Mesoamerica and beyond. This research thus opens several potentially productive lines of theoretical and methodological inquiry that may be pursued and tested with further research.

In general terms, this study has the potential to clarify correlations between the development of writing and the emergence of increasingly complex material culture in both general and specifically Mesoamerican contexts. Many scholars have traced the emergence of complex archaeological phenomena to interregional interaction (Houston 2004b: 8–10). This research reveals how writing may serve as a heuristic device for anticipating other features of social or cultural complexity. In addition, this investigation adds to the increasingly nuanced understanding of emerging complexity in the Late Formative period of Southeastern Mesoamerica (cf. Guernsey 2012). This investigation thus suggests that the study of writing may serve as a proxy for wider investigations of sociocultural innovation, demonstrating how a focus on a combination of epigraphic and archaeological data can illuminate broader questions related to the development of complex cultural and material technologies (Josserand and Hopkins 1999; Trigger 2004).

Conclusions This investigation underscores the effects of shifting networks of interregional interaction on lowland Maya material culture, linguistics, and scribal traditions by presenting a subtler view of the processes that brought about innovations and changes in these traditions at a crucial period of transition in Mesoamerican history. The study seeks to trace the dynamics of the emerging Maya writing system in Formative and Early Classic period Mesoamerica, exploring the role of interregional interaction in this developmental process and the relationship between the emergence of writing systems, interaction, and material cultural complexity. This study demonstrates how investigations of the development of script systems offer useful analogies for wider archaeological research. For example, examining the emergence of writing may serve as a proxy for the study of material changes, thus informing debate regarding the emergence of sociocultural, material, and technological innovations in antiquity. This project thereby contributes to larger anthropological debate

More specifically, this investigation may help determine the kinds of spatial contexts in which archaeologists are likely to encounter evidence of early writing. This work also informs studies of material remains related to the emergence of writing and may assist investigators who encounter such evidence in the field (Campbell 1984; Houston 2004c; Josserand 1975; Kaufman 1976). Finally, this research also has the potential to clarify and refine the regional ceramic chronology of its primary study area.

10

Introduction interaction and its effects on long–term processes of sociocultural change. This investigation demonstrates how a combination of archaeological, linguistic, and epigraphic data produces a richer picture of Maya society than one constructed from a single body of data and may contribute to larger anthropological debate regarding the origins of sociopolitical complexity. The conjunctive study of the relationship between script development, material culture, and interregional exchange complements investigations of social innovation informed by other anthropological subfields. The interdisciplinary approach employed here illuminates wider questions related to social structure and change and elucidates more complex understandings of broader archaeological processes, such as boundary formation, the emergence of hierarchy, or incipient systems of specialized craft production, among others. The novel methodological study of potential connections between larger investigative foci that emerges from this conjunctive research may prove useful to future investigations in Mesoamerica and beyond.

A significant merit of this work is that it quantitatively and qualitatively tests a methodological proposal for examining the emergence of the Mayan script by correlating a theoretical model with material evidence. The value of such a model to large regional systems such as Formative period Mesoamerica is clear. In those contexts, researchers do not fully understand the emergence of writing, since available evidence is scarce (Pohl et al. 2002; Rodriguez Martínez et al. 2006; Saturno et al. 2006), and what little exists does not fit neatly into current models of the development of writing (e.g., Christen 2002; Houston 2004a, 2006). The outcomes of this investigation may suggest new directions for evaluating the relationship between the development of writing and complexes of material culture in various contexts, thus contributing to a broader understanding of the development of writing, in both general and specific terms. In broader contexts, a primary contribution of this research is the presentation of a new methodological template for the archaeological consideration of material

11

Chapter 2 Theoretical Framework and Methodological Premises change in archaic societies are founded on conceptions of adaptive responses to external stimuli, which effect transformation in societies that would otherwise remain in a hypothetical state of equilibrium. Individuals and their communities, however, do not live and act in a vacuum, nor are all societal decisions determined solely by rational cost–benefit analyses. Rather, theories that explain emergence, change, and interaction in social systems must take into account internal social structure, innovation, and decisions or choices based on a variety of mutable cultural beliefs and practices. It is therefore now widely held that a variety of both internally generated and externally reactive causes and influences exists that have the potential to bring about broad societal changes (Brumfiel and Earle 1987; Clark and Cheetham 2002; Goldstein 2000; Marcus 1992b, 1998; Parkinson and Galaty 2007; Parkinson 2002; Stein 2002).

Modeling Interaction and Innovation in Ancient Societies Describing and explaining the role of interaction in processes of sociocultural and technological change in ancient societies has long been a central goal of archaeology. Complicating this goal is the fact that societies are organized in a variety of forms and display relative degrees of integration and developmental complexity (Blanton et al. 1996; Marcus 1992b, 1998; Parkinson 2002, 2006; Parkinson and Galaty 2007, 2010). Such middle–range societies are flexible and opportunistic. These societies expand and contract, sometimes developing new sociocultural institutions and technologies while at the same time actively seeking to continue past lifeways and belief systems—or preserve specific material culture and technologies—in the face of cultural contact and interaction with other groups. Thus, as societies develop and change, they simultaneously make concerted efforts to maintain links to ancestral practices and to develop ties with any new peoples or cultures with which they come into contact. Systems of exchange and interaction, specialized production, and forms of sociopolitical organization in ancient societies will vary according to a number of factors, including available resources, the individual histories of each group, and the continued reinforcement of social structure through practice while, at the same time, integrating new experiences, ideas, beliefs, strategies, and technologies brought about in part through interaction and cultural contact. Describing such societal changes and transformations is often the easier part of the equation; it is considerably more problematic to explain the processes and mechanisms behind such changes.

A ‘middle ground’ between these opposing views is often overlooked. Elsewhere (Englehardt 2005), I have argued in favor of just such a conjunctive approach, especially as applied to the emergence of the Classic Maya writing system. The theoretical and methodological precepts that underpin this investigation therefore draw on a number of closely related theoretical frameworks, including systems and complexity theories, as well as theories of network and social exchange. Below, the fundamental precepts of these theories, are detailed as they relate to both the contention that interaction drives sociocultural change and the discussion of the development of Mesoamerican scripts and Maya writing. The combination of these theories allows for a deeper understanding of the multiple factors involved in material innovation and social change, underlines the centrality of interaction to cultural development, and permits a more detailed examination of potential correlations between interaction and material cultural variability.

There are a number of models that attempt to explain the relationships between interregional interaction and sociocultural development, material variability and change, and the introduction of new cultural practices, technologies, and institutions, as well as the role that interaction plays in any number of other archaeologically-observable phenomena (e.g., the transition from social and economic autonomy to centralization). The theoretical arguments that underlay these models have been fully summarized and discussed numerous times by many other scholars (e.g., Brumfiel 2003; Brumfiel and Earle 1987; Clark and Blake 1994; Fash and Sharer 1991; Freidel 1979; Marcus 1998; Parkinson 2002; Willey 1991; Yoffee 1995). The majority of these models fall into two basic camps, which posit that change occurs as the result of either external or internal influences, respectively. Many theoretical models of socio–organizational and material

Systems and Complexity Theories Originally a theory of open systems, as first described by von Bertalanffy (1972: 411), general systems theory holds that all entities behave as systems comprised of several subsystems that are governed by rules. Changes in the subsystems, and their rules, can be measured, as can their effects on other subsystems. Thus, a change in one part of the system would result in consequent, related changes in the other parts. A variety of approaches to anthropological inquiry have been developed that utilize systems theory, including network and graph theory, game theory, cybernetics, and information theory (Bertalanffy 1972: 416; Pickel 2007; cf. Trigger 1989). With roots in the British school of social anthropology, 12

Theoretical Framework and Methodological Premises particularly within the structural–functionalist paradigm, systems theory developed as a reaction against the culture–historical paradigm, which explained all sociocultural change as the result of external, not internal, forces (e.g., migration and diffusion). Emphasis was initially placed on identifying the internal processes that allowed a system to remain stable, although the focus turned toward the identification and description of change within and among systems following World War II, influenced by the growing field of cybernetics (Trigger 1989).

explanatory dimension lacking (or overly simplistic) in ‘pure’ systems theory frameworks. Like ecosystems, human societies and systems of material production are complex and adaptive. Macroscopic patterns of structure and interaction emerge, develop, and transform as a result of interactions at the local level, although these interactions are themselves linked to other systemic components via feedback loops (Anderson 1999; Lansing 2003; Levin 1998, 2002; Railsback 2001). Thus, the complex adaptive systems (CAS) proposed in complexity theory are comprised of agents, whether individuals, groups, villages, or regional networks, that are capable of adapting as they interact with each other through time (Railsback 2001: 49). From this perspective, one may begin to assess how the characteristics of individual subsystems affect responses and changes at larger systemic levels, moving beyond the mechanistic determinism originally posited in systems theory.

In systems theory, a system may be defined as a grouping of a variety of interacting elements that, as a whole, can be regarded as a single large unit. The subsystems that comprise the system are related and linked to each other and the behavior in or of one subsystem can be understood through the links and between that and other subsystems (Hodder 1992: 25). External impacts or interventions have only temporary effects on the system, since internal processes and rules ultimately dictate its behavior. As Bertalanffy (1972: 411) explains it: ‘In order to understand an organized whole we must know both the parts and the relations between them.’ Systems theory does not explain how systemic change occurs; it only describes changes in the relations and structures of dependent subsystems and thus the system as a complex whole. As applied in anthropology, a systems perspective conceives of societies as being ‘comprised of integrated systems, with interrelated institutions’ (Trigger 1989: 245).

An investigation of sociocultural change based on systems and complexity theories relies on three main assumptions: 1) that the diversity and individuality of components is maintained, 2) that interactions between the components are localized, and 3) that subsets of the components are selected for reproduction and alteration based on these local interactions through autonomous processes (Levin 1998: 432). CAS are thus dynamic, self–organizing, and nonlinear. Further, they maintain diversity as they co–evolve. The property of self– organization refers to the fact that patterns observed on smaller scales are replicated at larger scales throughout the system, since CAS are nested hierarchies that may contain other CAS (Anderson 1999). Such intra– systemic replication may be brought about or influenced through interaction.

Based primarily on systems theory, complexity theory emerged to counter the inadequacy of general systems theory as an explanatory framework. Rather than considering variable relations between subsystems as a primary causal factor, complexity theory incorporates the notion that systemic changes are based on a multiplicity of causes rather than a single ‘prime mover,’ and rejects the premise that systems can be explained by a reduction to their component parts (Bentley and Maschner 2003: 1). In anthropological terms, whereas a strict systems perspective views human society as a closed system that strives to maintain stability, complexity theory reverses that conception, holding that societies are open and dynamic systems that are fundamentally incapable of maintaining equilibrium.

Emergent network patterns of interaction in a CAS self– organize hierarchically by power–law distributions. That is, relatively few key nodes within the network are well– connected and the majority is loosely connected through these key nodes. ‘Egalitarian’ (or ‘random’) networks are comprised of nodes connected via approximately the same number of links; conversely, those networks in which a few nodes control numerous links between components connected to a lesser degree are described as ‘aristocratic’ (Buchanan 2002: 119; cf. Bentley 2003). Further, a power–law distribution results in scale invariance, implying that changes occur on all scales, no matter how small or large. Such CAS networks are therefore labeled ‘scale–free’ (Bentley 2003: 17; Bentley and Maschner 2001: 41). Scale–free networks are susceptible to collapse if well–connected nodes are removed from the system. Conversely, the removal of nodes that are minimally connected to the system will not result in network failure (Barabási and Albert 1999; Wang and Chen 2003: 14–15). Figure 2.1 illustrates an example of a scale–free aristocratic network in Mesoamerican contexts.

In such open systems, all subsystems, whether as individuals, groups, or larger regional networks, are interconnected so that changes at one scale affect others. In this context, complexity may be defined as ‘the capacity of nonlinear interactions within clusters of activities and processes... to generate emergent structuring through self–organization’ (McGlade 2003: 115). Complexity theory effectively adds an adaptive component to its definition of a system in order to explain diachronic changes within the system and its components—an 13

Archaeological Paleography NCV – North Central Veracruz; OH – Olmec heartland; GH – Guatemalan highlands; CM – Central Mexico; MG – Middle Grijalva; NML – Northern Maya lowlands; Mor – Morelos; CC – Chiapas coast; SML – Southern Maya lowlands; WM – West Mexico; GC – Guatemala coast; WS – Western El Salvador; Ox – Oaxaca; IC – Interior Chiapas; CA – Central America Figure 2.1. A hypothetical lattice model of Middle Preclassic period scale–free interregional interaction networks, showing nodes of interaction (after Demarest 1989: 337, fig. 13.2).

CAS models thus integrate the interactions between individual agents or subsystems into explanations the development of system–wide patterns of organization and adaptation, particularly as expressed in material culture. In this case, interaction is systemic in that groups of phenomena act in concert to produce a more complex result through feedback within the system. Since the system is open, these patterns of interaction are dynamic and variable, and new components are constantly introduced to the system (Anderson 1999; Lansing 2003; Levin 1998, 2002). Over time, as various systemic components interact and adapt, the system reaches a tipping point. At such a point, a threshold is crossed and changes sweep through the entire system (Bentley and Maschner 2003: 67). Both social and natural systems are constantly emerging from a variety of sources, as are systems or suites of material culture. In such circumstances, structures may appear superficially stable, but may be drastically altered by seemingly minor influences.

whole. Thus, from the perspective of complexity theory, a driving force behind intra–systemic interaction is the degree of centralization of those well–connected nodes. Equally significant is the degree to which smaller, localized networks of exchange are integrated into the larger macro–regional system. As Parkinson (2002: 398) suggests, these theoretical assumptions imply that the structural configuration of a given ancient society is predisposed to accommodating a variety of distinct social processes. This is not to say that the actions of social agents are wholly prescribed by external structural forces beyond individual control. Individuals, through social actions (including interaction with others), create and modify their own social structure (Giddens 1979, 1984). Nevertheless, through active social practice, individuals do in fact produce and reproduce the overarching social structures that to varying degrees both constrain and enable their behavior within socially acceptable bounds (see also Bourdieu 1977, 1990). Complexity theory is therefore capable of accommodating variable considerations of the internal and external factors that act in tandem to drive and shape social, organizational, material, or technological changes in ancient societies.

Patterns of networks within CAS, such as ‘small world’ networks, aid in understanding how specific artifactual distributions are achieved, or how material cultural traditions are maintained or transformed (Bentley 2003: 27). Bentley and Maschner (2003: 69) describe small world networks as those in which relatively smaller communities are integrated in a larger network via a small number of links connecting those ‘nested’ communities to larger nodes of interaction. That is, the bulk of systemic components are connected to each other along short paths, while a few well–connected hubs develop through interaction and feedback. Closer neighbors may interact more frequently through short but relatively weak connections, whereas larger, well–connected nodes are joined via relatively longer, stronger, and less frequent links that bind the system together as a complex

Theories of Network and Social Exchange Two subsets of systems and complexity theory that are of further relevance to this investigation are the theories of network and social exchange. Network theory assumes that the structure of a given society is comprised of patterns of ties between localized communities or nodes of exchange, and that variation throughout the network is based on the relative strength of these meaningful ties (i.e., degrees of integration; Cook and Whitmeyer 1992: 114–18). Network theory is thus empirical, holding that

14

Theoretical Framework and Methodological Premises patterns of interaction between actors can be measured and graphed. Approaches based on network theory attempt to understand how societies are structured internally through the interactions and networks that connect individual communities (or smaller, nested networks) to each other. In turn, one may gauge how systemic structure influences and affects interaction (as well as the effects of that interaction) at and across smaller scales.

these nodes may also be centers of specialized production or distribution, and thus their physical location within the systemic landscape may not be central to other nodes within the network. In fact, as I argue below, such centers of innovation are actually more likely to be located between interaction spheres (Altschul 1978; Caldwell 1964; Freidel 1979; Possehl 2007; Struever 1972) or on the frontiers of or borders between developing systems or exchange networks.

In contrast, social exchange theory conceptualizes social interactions as series of exchanges of valued items, whether tangible or intangible. From this perspective, material or symbolic exchange within a system is driven by perceived rewards and the individual self– interest of social actors, who are assumed to be rational beings who undertake cost–benefit analyses of any potential exchange (Zafirovsky 2005: 1). Changes at the level of the individual, in terms of the choices they make regarding particularly desirable traits or characteristics of specific material cultural items, result in larger systemic consequences, particularly in terms of the development of power relationships between systemic components. Social exchange theory focuses primarily on the long–term life cycle of social relations—emergence, maintenance, and disappearance—and the factors that influence those relations, as well as their material manifestations over time. Social exchange theory does not, however, typically account for symbolic exchanges, and usually does not take into account variable cultural contexts (Cook and Whitmeyer 1992; Zafirovsky 2005: 12; cf. Bird and Smith 2005).

Systems–based and complexity theories can thus aid archaeological research by describing and suggesting hypothetical explanations of the spatial patterns of distribution of multiple classes of material correlates of systemic interaction, at both micro– and macro– levels (Manson 2006; O’Sullivan et al. 2006; Portugali 2006). Local, regional, and macro–regional systems of interaction, exchange, and communication can be hypothesized and mapped on the basis of such spatial patterns of artifact distribution. These interaction spheres will contain evidence of social shifts as manifested in tangible changes in material goods, and can potentially demonstrate how variation in the qualities, styles, and composition of material culture is indicative of broader changes across the entire system. Examining Interaction, Integration, and Variability through Material Culture Archaeologists working in many areas of the world have suggested that interregional interaction sparks cultural innovation (Cherry 1984; Demarest 1989; Flannery 1968a; Renfrew 1972). Models based on systems and complexity theories have proven strong frameworks from which to approach questions concerning innovation and emergent cultural technologies (see, e.g., Binford 1965; Cherry 1986; Demarest 2002; Flannery 1968b). Researchers have employed complexity theory to elucidate correlations between innovations in material culture and increasing interregional interaction in a variety of contexts. Previous applications suggest that complexity theory as a heuristic device has great potential in archaeological inquiry, insofar as it is capable of describing and potentially aiding in the explanation of the spatial patterns of distribution of multiple classes of material correlates of systemic interaction, at both micro– and macro– levels (Manson 2006; O’Sullivan et al. 2006; Portugali 2006).

What these two ‘daughter’ theories have in common is a particular conceptualization of the central role of the social actor as a potential agent of systemic change at both micro– and macro– scales (Bourdieu 1977, 1990; Giddens 1979, 1984). Nevertheless, although the actor may be motivated by rewards and cost–benefit analyses, neither network nor social exchange theory is capable of explaining all patterns of interaction. Systems and complexity theory posit that the structure of the system in which the actor exists (and, by extension, in which interaction and exchange occurs) necessarily shapes and influence individual choices at the local level, which in turn reverberate through the system. In this investigation, I am more concerned with these broad distributional macro–regional patterns than with the individual choices or motivations of social agents operating at the local level.

Complexity theory suggests that variable degrees of integration of the nodes that bind the larger macro– regional networks act as a catalyst for interregional interaction. From this perspective, and due primarily to the decentralized and segmentary nature of ancient middle–range societies, archaeologists have often approached issues of socio–organizational and material change by examining diachronically variable patterns of integration and interaction, two complex, potentially

In the archaeological terms of artifactual distribution, these theories suggest that material indicators of long– distance exchange between well–connected nodes (e.g., stylistically or functionally similar material culture) should be encountered at other well–connected systemic hubs—particularly along the borders of contiguous interacting systems or spheres of exchange. Of course, 15

Archaeological Paleography ambiguous, and overlapping concepts (e.g., Caldwell 1964; Flannery 1968a, 1968b; Hegmon 1989; Struever 1972; Parkinson 2002, 2006). Traditionally, ‘integration’ is conceived as a group–level phenomenon; the term usually refers to processes that bring individuals together into more–or–less formalized units at (potentially) variable scales (Parkinson 2002: 394; Parkinson and Galaty 2010: 10). The term ‘interaction,’ in turn, refers to the more general processes that may occur at either the group or individual level (Caldwell 1964; Parkinson 2002). Although the concept of interaction generally indicates some kind of direct or indirect social contact, as a rule it suggests nothing specific about the nature of the contact or the particular relationships between those in contact (Parkinson 2002: 394).

of a single flexible tradition. In order to operationalize this concept for studying material variability in ancient societies, one must define the temporal and spatial scales of analysis independently, and establish a methodology for modeling interaction within those bounds (Parkinson 2002: 394, 398–99, 2006; Parkinson and Galaty 2010). Previous research has suggested that the scale on which archaeological correlates of interaction and integration may be measured and evaluated is at the level of the regional interaction sphere (Altschul 1978; Caldwell 1964; Freidel 1979; Neitzel 2000; Peterson and Drennan 2003; Possehl 2007; Stein 2002). The term ‘interaction sphere’ describes a set of sites within a geographic region that share an assemblage of cultural traits evidenced in artifactual remains (Altschul 1978: 111; Freidel 1979; Struever 1972: 315). In this case, the interaction is systemic in the sense that groups of phenomena act in concert to produce a more complex result through feedback. Interaction indicates that some form (or forms) of communication, intercourse, or articulation existed that enabled groups to share an assemblage of cultural traits.

Following Parkinson (Parkinson 2002: 394; Parkinson and Galaty 2010: 10–11), in this study the term integration refers to ‘social processes that incorporate individuals into specific organizational (i.e., decision–making) units.’ Under this definition, societies are composed of various integrative units: individual households, settlements, micro–regional settlement clusters, macro–regional interaction spheres, and so on. Interaction, on the other hand, refers to more diffuse social processes that operate between individuals or groups. Although individuals and groups may interact, they are not necessarily incorporated into centralized or necessarily well–defined integrative units. In the context of this research, the dimension of integration refers to the size, scale, and organization of the basic social segments within larger regional systems, and the dimension of interaction refers to the ways in which these social segments, and their members, interact amongst themselves and with segments in differing regional or systemic contexts.

One method that has proven particularly effective for modeling interaction in ancient contexts and elucidating variability across the boundaries of such interaction spheres is the stylistic and distributional analysis of material variables, particularly ceramic assemblages, since ceramics are excellent variables through which to measure interaction and material innovation (e.g., Parkinson 2006: 34; Skibo et al. 1989; see also Hirth 1998; Hodder 1972, 1982, 1985; O’Shea and Milner 2002). By articulating such stylistic analyses of ceramic materials within the context of integrative units, it is possible to more accurately describe the degree of interaction that occurred at different levels within a social system over a given amount of time. This study follows these prior investigations by focusing on the stylistic (and technological) variability exhibited both within and between integrative units and considering material innovation, in terms of stylistic and distributional similarity and variability in material culture, as the result of differing degrees of interaction between cultural groups or regions (Hegmon 1995; Plog 1995).

Although the intertwined concepts of interaction and integration are by no means mutually exclusive, it is critical to distinguish between them methodologically, insofar as different lines of archaeological evidence tend to speak more to one process than the other (Parkinson 2002: 394, 2006). Treating interaction and integration as discrete analytical dimensions and independent lines of evidence allows archaeologists to measure specific archaeological variables that relate predominantly to one dimension or the other, to more precisely distinguish between the differing processes of integration and interaction, and to generate models that more accurately represent the size, scale, and organization of particular elements of social structure (Parkinson 2002: 394, 398– 99, 2006; Parkinson and Galaty 2010). These models then can be used to track diachronic changes within a distinct suite (or suites) of material culture that reflect structural variability, and to compare the various trajectories followed by different material cultural traditions over time. Utilizing a comparative database allows for cross– cultural or interregional analysis of differing traditions of material culture, or of the variable manifestations

The association of style with interaction and integration is complex and highly contextualized, since material culture systems are historically situated phenomena (Stark 1998: 8–9). In any given region, interaction, like integration, may occur at a variety of levels. Diachronic patterns of interaction may also operate at varying degrees of intensity between and within integrative units, since some units assume focal or primary roles (the ‘well–connected nodes’ described above). Any attempt to model social interaction in prehistoric contexts therefore must account for this dynamic quality (Feinman 1998; Marcus 1992b, 1998; Parkinson 2002: 398–99; Parkinson and Galaty 2007; Yoffee 2005). 16

Theoretical Framework and Methodological Premises To address this dilemma, it is productive to approach the relationships between interaction, integration, and innovation in material culture from a dynamic perspective (Barth 1969; Hegmon 1989; 1992, 1998; Kowalewski et al. 1983). Interregional variability in material culture is a function of multiple factors, including the relative degrees of integration, interaction, and interdependence of the cultural group(s) that occupy the landscape on either side of a sociocultural or geographic boundary, as well as the degree of centralization of the larger regional systems into which such cultural groups are potentially integrated (Galaty et al. 2010; Kowalewski et al. 1983; Neitzel 2000; Parkinson 2006; Parkinson and Galaty 2007, 2010; Peterson and Drennan 2003).

in discerning the nature and extent of interaction and innovation in distinct sociocultural, temporal, and spatial contexts (Carr 1995; Carr and Neitzel 1995; Cheetham 2007; Hegmon 1992, 1998; Hodder 1978; O’Shea and Milner 2002; Parkinson 2006; Stark 1998; Stark et al. 2000; Voss and Young 1995). In this investigation, I undertake a diachronic regional stylistic evaluation of the formal and functional attributes of material evidence in order to characterize the degree and quality of variation observable in the collected data, as well as the likely role that a particular stylistic attribute(s) would have played in different contexts (Carr 1995; Voss and Young 1995). Measuring grades of similarity and diversity within, among, and between ceramic assemblages in differing temporal contexts characterizes the nature and extent of interaction that occurred across the boundary on which they are situated in both synchronic and diachronic terms and on multiple scales.1 Variation in the material data permits an appreciation of the relative permeability and social maintenance of boundaries, adding resolution to inferences regarding regional interaction (Parkinson 2006; Stein 2002).

Measuring stylistic variability (or lack thereof) in material objects thus potentially speaks to the relative extent of interaction that occurs across a macro– or micro– regional boundary between given variably centralized or integrated social systems, as well as the amount of material innovation that occurs within those systems. In turn, comparing patterns of variability over time and at different scales yields clues regarding the temporal and spatial contexts of interaction and material innovation. These patterns may also be compared to diachronic changes in sociopolitical contexts or the material assemblages of adjacent sites and regions. Such methods have been applied to questions of interaction and the relationship between material traditions with success in Mesoamerica (e.g., Cheetham 2007; Demarest and Sharer 1982; Neff et al. 1999). Greater material and distributional uniformity across a region suggests extended interaction on a wider scale by groups whose suites of material culture exhibit such strong correspondences, more relaxed structural integration, and a relatively lower level of localized innovation.

A systems perspective also allows for the identification of elements of material culture that may be considered prestige goods. For example, ceramic materials may be identified as valued commodities due to their scarcity within material assemblages, their origins, or symbolic connotations in lower–level peripheral sites situated on cultural boundaries (Bird and Smith 2005; Goldstein 2000). Ceramic materials may be considered central to processes of social reproduction, serving as status markers that are potentially indicative of relationships with larger nearby centers and nodes of interaction. Analyzing stylistic and functional variability and innovation (Carr 1995; Carr and Neitzel 1995; Hurt and Rakita 2001; Plog 1995; Stark 1998; Voss and Young 1995) in material markers of interaction thus permits evaluation of the processual dynamics of the symbolic categories that underlie both material exchange and writing. Expressive behavior, including style, is essentially an act of ‘re–creation’ of social structure that is contingent on a number of factors, including general sociocultural and specific interpersonal reference relationships, and the abstract values and cultural ideals that guide behavior (Parkinson 2002: 398–99; Voss and Young 1995: 91–2). In this investigation a sample database of ceramic data is compared against regional diagnostic collections to evaluate the relationship between material variability, interaction, and script development. Stylistic

Conversely, increased variation in material assemblages indicates narrower interaction on a more localized scale, a greater degree of integration within larger cultural systems, and greater potential for the emergence of localized styles, forms, and functions. The analysis and interpretation of stylistic and distributional patterns observed in material data can be related to social processes of interaction and material innovation by focusing on similarities or changes in artifacts over space and through time (Green and Perlman 1985: 6; Parkinson 2006: 36; Skibo et al. 1989). Following Stark (1998), I consider variation as the result of processes reflected in the material data; it is also paramount to consider variability in the specific manufacturing processes, which is reflected in diachronic technological and stylistic discrepancies across archaeological remains at the regional scale (Hegmon 1998; Stark et al. 2000).

For example, as Carr (1995; Carr and Neitzel 1995) argues, more visible attributes generally tend to correspond to a wider range of more active social processes (e.g., of active interaction and group–boundary maintenance), whereas less visible (more obscure) attributes relate to more passive processes (e.g., of enculturation; Parkinson 2002, 2006).

1

Stylistic and functional analyses of variability observed in material markers thus have proven valuable tools 17

Archaeological Paleography and functional variability in the distribution of material artifacts should indicate both degrees of interaction and integration, as well as differential use of material culture in distinct contexts, paralleling the emergence of writing in Mesoamerica. The theoretical premises behind the proposed association between material interaction and the development of Maya writing will be addressed shortly. First, however, the discussion turns to the question of where innovation and other effects of interaction occur, and are most visible, on the archaeological landscape.

culture become visible. Material culture is crucial to processes of social reproduction, with particular formal configurations serving as potential indicators of relationships with larger nearby centers and nodes of interaction. These assumptions simultaneously rest on, in part, and contradict another ‘daughter’ of general systems theory: world systems theory (Wallerstein 1974). The world systems approach postulates the existence of inter– social networks in which interaction is fundamental to reproduction of the internal social structures. In this class of systems the behaviors associated with the production and exchange of goods and ideas between individuals effect an interactive impact on a given locality (Hall and Chase–Dunn 1996). As a heuristic device, the world systems perspective permits the study of networks of interaction at different levels, including the social, political, and economic relationships that are established between a core area and its periphery or boundaries, as well as the relationships between areas or micro–regions centered on a particular core node. World systems hold that as centralized network focal points and core areas expand, they incorporate micro–regional ‘mini–systems’ into their own growing system and socio–economic sphere of influence. From this perspective, developing cores, supported by an expanding periphery, are the focal points of technological and cultural innovation and manufacture. Systemic changes are thus caused by the economic and political concerns of the core that spread throughout the system, with little input from peripheral areas or borderlands; ‘higher levels affect lower levels’ (Brumfiel 1995: 125).

Boundary Areas and Material Innovation If systemic interaction at the regional level is a catalyst for cultural innovation, then innovations and advances in cultural technologies should be especially visible at the boundaries of the interacting units within the regional system. Therefore, if interaction correlates with script development, one would expect to encounter the correlative evidence of emergent script technologies at the peripheries of cultural interaction spheres. The boundaries of Formative period regional interaction spheres are therefore investigated for material evidence that correlates with the interaction–driven developmental processes involved in both material innovation and the emergence of Maya writing. Coincidentally, these stylistically defined zones of sociopolitical or economic interaction, and overlaps in archaeologically identified regional exchange networks, may parallel preexisting linguistic or ethnic divisions, the presence of which offers a useful comparative baseline (e. g., Josserand and Hopkins 1992; Justeson 1986; Justeson et al. 1985; Lacadena and Wichmann 1999). Boundary areas are particularly fertile zones for detecting emergent material forms and cultural technologies, particularly in Mesoamerica (e.g., Golden and Scherer 2006; Green and Perlman 1985; cf. Goldstein 2000; Hodder 1978, 1985; Stark 1998; Stein 2002). Comparing patterns of formal and functional variation in suites of material culture from adjacent regions on either side of a sociocultural or geographic frontier illustrates vertical and horizontal linkages across archaeological remains through time and space. Synchronic homologies and shared stylistic characteristics in material evidence are indicative of interaction across a regional border. Conversely, fewer shared attributes suggests a greater degree of localized innovation, as do intermediate forms—artifacts that represent a blend of the stylistic characteristics observed in adjacent areal assemblages and typological sequences. As centralization and integration increase in a cultural system, boundaries close and the symbolic connotations of material culture are potentially more open to reinterpretation by lower–level peripheral sites—removed from well–connected regional nodes—situated on such cultural boundaries. Thus, emergent formal and functional attributes in material

Since its introduction in the 1970s, many applications of world systems theory in archaeology have concentrated on the relationships between the core and the peripheral ‘other.’ Until recently, interaction and social change on the frontier itself has not been a subject of extensive inquiry. Recently, archaeologists have modified world systems theory in order to better apply the concept in non–Western, non–capitalist contexts (e.g., Hall and Chase–Dunn 1996; Trigger 1989). Although as originally conceived world systems theory relies on many questionable assumptions regarding prehistory and the organization of non–Western societies, some of its primary tenets—namely, that societies should be studied or viewed as interacting systems comprised of many dynamic components rather than isolated entities—hold true. Recent examinations of innovation in regional systems (e.g., Stein 2002) illustrate the critical role played by the geographical proximity and local institutional conditions of peripheral or border areas (i.e., factors not dictated by central core concerns) for the production of new knowledge and its sociocultural, political, or economic exploitation. In any case, causation is a secondary concern: whether cultural and technological 18

Theoretical Framework and Methodological Premises innovations were driven ‘down–the–line’ by the core or locally motivated at the periphery, the material effects and temporal contexts of those innovations will be seen most clearly at the borders of regional systems.

Interaction and The Development of Writing Systems in Mesoamerica A variety of prior research has offered systemic interaction–based models for the development of writing in Mesoamerica (Chinchilla Mazariegos 1999; Englehardt, 2005; Fahsen Ortega 1999; Houston 2004a; Justeson 1986; Justeson and Matthews 1990; Justeson et al. 1985; Stross 1990). These models explain the emergence of Mesoamerican scripts in terms of a recontextualization of iconographic symbols within new linguistic contexts as a result of interregional interaction (Gelb 1963; Harris 1986; Justeson 1986; Justeson et al. 1985; Senner 1989; Trigger 2004). Sustained interaction between distinct groups employing a common iconographic complex carries the potential for radical reinterpretations or reconfigurations of shared iconographic elements. Subsequently, new values may be juxtaposed in unprecedented ways, leading to the grammatical or linguistic codification of iconic motifs, at which points writing emerges. In short, fixing variable semantic, grammatical, or linguistic meanings to shared icons in the context of interregional interaction detached them from previous interpretations to facilitate their new use as written signs.

Developing sociopolitical and material complexity in the lowland Maya world, reflected in increasing variability in material culture, thus could have occurred within the framework of interregional interaction between regions that included well–connected nodes or central ‘cores’ (see, e.g., Cherry 1984, 1986; Parkinson and Galaty 2007, 2010; Sabloff 1987). Insofar as well–connected network nodes were successful in increasing the flow of commercial exchange or ‘tribute,’ dominating production, and controlling ritual activity, it is possible that interregional rivalries resulted in a greater degree of centralization and integration within a larger, more unified pan–regional lowland system during and after the Late Formative–Early Classic period transition (c. 250 BC–AD 250). A central problem in archaeology is the placement of the boundaries investigators use to circumscribe and define archaeological ‘cultures’ (Green and Perlman 1985: 6–9; Kowalewski et al. 1983; Parkinson 2006: 33–4; Stark 1998). Although stylistic and distributional analyses of variation in material data may be used to define a boundary, the boundary does not necessarily enclose a particular suite of material culture (Stark 1998: 8–9). Moreover, drastic changes in material culture are not always explained by coeval transformations in sociocultural organization or a concurrent redefinition of social boundaries. This is true since, from a dynamic perspective, variability or continuity in material culture across social boundaries is a function of multiple factors.

Writing and the Development of Writing Systems The idea of writing is difficult to define. The term tends to be used vaguely, complicating identification and classification of Mesoamerican scripts. Although a ‘broad definition of writing is not inherently objectionable,’ as Trigger (2004: 44) notes, ‘it must account for differences between various categories of writing’ (as does Boone 2004). This investigation adopts the typology of writing systems proposed by Sampson (1985: fig. 3), which distinguishes between two main types of writing systems: glottographic and semasiographic. In glottographic systems, language is the primary means of encoding a message, whereas in semasiographic systems, language is not the primary means of encoding (Mora–Marín 2001: 11). In turn, this typology necessitates a definition of script or writing system as a graphic, conventionalized, patterned, and potentially permanently rendered set of marks used to communicate ideas in a relatively specific way (Sampson 1985: 19; cf. Mora–Marín 2001: 11).

Despite the inherent difficulty of drawing boundaries, stylistic analyses have the potential to elucidate social processes related to the formation, permeability, and maintenance of boundaries (Green and Perlman 1985; Kowalewski et al. 1983; Parkinson 2006; Skibo et al. 1989). Therefore, a detailed comparison and analysis of the stylistic attributes and functions of both textual evidence and material culture found at sites in border areas should indicate the degree and nature of interaction that occurred across the sociocultural and geographic boundaries on which they are situated over space and through time (O’Shea and Milner 2002; Parkinson 2006). Variability in form and function allows for the definition of micro–regional units of analysis (Hodder 1978; Stark 1998). The temporal scale of the analysis is crucial, since Maya society, material culture, and script underwent radical transformations in the transition between the Late Formative and Early Classic periods (Fahsen Ortega 1999; Grube 1999). In the context of this investigation, the spatial scale of the analytic units (and their boundaries) and the temporal scale of analysis is discussed at greater length in the following chapters.

For the purposes of this study, writing is defined as a system of communication by means of standardized visible marks, that is, as the representation of language in a textual medium through the use of signs or symbols (Gelb 1963: 253; Justeson 1976; Salomon 2001; Sampson 1985).2 Writing systematically represents speech. The This discussion alternates between the terminology employed in the Peircean trichotomy of icon, index, and symbol and de Saussure’s sign vs. symbol dichotomy. Semantically, it is more instructive to employ the term ‘sign’ when discussing glyphic elements and ‘symbol’ or ‘icon’ when discussing a particular iconic or visual motif. However, the

2

19

Archaeological Paleography Signified Linguistic Unit

Written Symbol

System of Signs

Single sound (phoneme)

Letter or Alphabetic Sign

Alphabet

Syllable

Syllabogram

Syllabary

Word

Logogram

Logography

Idea/abstract concept

Ideogram or Icon

Ideography

Object

Pictogram, Icon, or Drawing

Pictography/Iconography

Figure 2.2. An analytic classification of writing systems based on types of signs and symbols employed (adapted from Gelb 1963: 14, fig. 2). In the typology detailed above, ideography and pictography/iconography are classified as semasiographic scripts, whereas logographies, syllabaries, and alphabets possess a direct correlation with specified linguistic values unique to a particular spoken language.

written symbol represents the spoken one and symbolizes morphemes of a particular, spoken language (Robertson 2004: 20; Sampson 1985: 149). Writing therefore differs from pictography, ‘picture writing,’ iconography, or ideography in that such ‘semasiographic’ signs are para– linguistic; they have no one–to–one correlation with any particular or specified grammatical and/or linguistic value (Figure 2.2).3 In the sense that all visual signs are to some extent iconic in nature, the distinction between ‘iconography’ and ‘writing’ lies more in the relationship between signs: in writing, this relationship is, at least to a greater extent, based on a linguistic structure rather than a mimetic or visual resemblance (Carrasco and Englehardt 2015, n.d.). In centering on the Mayan script, this investigation is more focused on what Sampson labels glottographic writing, in which there is a direct association between graphemes and basic linguistic units (Justeson 1976). The linguistic units involved may be of different levels (e.g., phonetic, phonemic, syllabic, morphemic, or polymorphemic; see Mora–Marín 2001: 11), but at its root, in the same way that language may be regarded as a ‘system of relationships between meaning and speech–sound’ (Sampson 1985: 28), a glottographic script is used to maintain those relationships through visual representation and conventionalization.

graphic communication are brought together into a single format, or when a single system of symbols exists in multiple formats, in which the principles of no one preexisting system remain sufficient to interpret the relations of all symbols to one another, while grammatical or linguistic principles do suffice to determine such relationships (Justeson et al. 1985: 34; cf. Houston 2004a; Justeson 2012). Writing creates a specific and often arbitrary link between the written word and its spoken object of reference that does not depend solely on visual principles for interpretation, in contrast to iconography or pictographs (Robertson 2004: 16–19). As I have argued previously regarding the development of writing in Mesoamerican contexts: ‘Initially, signs came to be more contextualized within a linguistic matrix and/or the reader understood or read the sign within the system as a whole rather than on the basis of its resemblance to a specific or natural referent… As writing develops, however, iconic elements become ‘divorced’ from the iconographic conventions and canonical compositional patterns in art. Subsequently, a grammatical, linguistic framework and the conventions of the writing system become the primary organizational principles for the ordering of signs and their interpretation. That is, as iconography moves towards the spectrum of writing, the organization of signs conforms to the structure of language as opposed to the representational schemes of which a given sign was originally an iconographic element. Such re–contextualization allows the scribe or reader to derive the meaning of a sign or sign sequence via the phonetic or word (i.e., linguistic) values associated with the sign(s)’ (Carrasco and Englehardt 2015: 4).

Current models suggest that phonetic writing develops when symbols from different systems of Peircean notion of symbol (or sign) differs greatly from that proposed by de Saussure. When the term ‘sign’ is used in this chapter and throughout, it is employed it in the Peircean sense of ‘icon + symbol.’ Likewise, the term ‘icon’ is employed in the Peircean semiotic sense of a sign that denotes its referent via formal resemblance or by virtue of shared qualities. 3 Not all signs in all writing systems or scripts fit neatly into the categories represented in Figure 2.2. One example is the rebus, in which a secondary sign, or a transferred pictogram is re–employed as the sign for a word or syllable which by chance happens to be close or identical in pronunciation to the word which originally motivated its pictographic form (e.g., a pictograph of ‘sun’ might be used to represent the word for ‘son;’ Harris 1986:32, 34). Nor do writing systems and scripts exclusively employ only one type of sign. For example, alphabetic English writing often uses the logograph ‘$’ to represent the word ‘dollar,’ or the same sign in an ideographic context to represent the abstract idea of ‘wealth.’

It is at precisely the point when meaning may be derived on a basis other than open ideational or pictographic representational conventions that a movement toward linguistic codification and potential phoneticism begins in visual notational systems (Carrasco and Englehardt 2015: 6). 20

Theoretical Framework and Methodological Premises In more abstract terms, Jakobson (1971a, 1971b) suggests that visual signs are time transcendent or atemporal, imitative, and immediate. Visual perception is iconic, and linked to the cultural–symbolic, because it exhibits a clear similarity of analogy between the sign and its object (Robertson 2004). In contrast, the auditory stimuli to which writing corresponds are temporal (vanishing in the absence of the referent), imputed (taking on meaning when assigned to an object), and sequential (achieving holism through differentially arranged patterns in particular words; Robertson 2004: 16). Writing lies at the nexus of visual and auditory perception because it simultaneously includes the holistic character of visual perception and the sequential character of auditory perception (Jakobson 1971). Nonetheless, the basis of writing is the picture: ‘Just as speech developed out of imitation of sound, so writing developed out of the forms of real objects or beings’ (Gelb 1963: 27).

is transformed into a written symbol. The symbol may then be used in purely phonetic contexts. In essence, the process is one of increasing simplification, specialization, and direct connection with a spoken language, with symbols developing greater formal complexity through experimentation from purely visual to phonetic signs. Linguistically speaking, writing may develop when the symbols of an iconographic system exist within a single spatial context in which multiple languages, ethnic groups, or sets of interpretive principles exist. Over time, interaction between languages, groups, or conventions encourages human agents to modify the iconographic system to represent more closely the particular linguistic values of a specific language. In such a situation, visual interpretive principles fail to sustain meaning adequately solely on the basis of familiarity with established iconographic conventions. The resulting tension requires readers to recodify the original iconic system in a more discretely organized manner, infusing icons with linguistic meaning, in order to determine more precisely the relations of the now written symbols to one another (Hopkins 1997; Justeson et al. 1985: 34; Robinson 2003; Rogers 2005).4

Of course, even in the context of pictorial systems icons possess latent phonetic values, insofar as they depict objects that can be identified in language, even if these values are not tied to a specific phonetic articulation. In glottographic scripts, however:

In this sense, writing develops from a crisis of meaning. In this crisis, new semantic meanings or grammatical– linguistic values adhere to a sign that previously depended on visual conventions or iconographic principles to interpret both the visual message and the relationship between constituent signs. In effect, an iconographic system previously open to interpretation or visual exploitation among and between multiple audiences or languages becomes ‘closed’ within a single one (Houston 2004b). Signs derived through this process of visual exploitation leading to linguistic codification (acrophony) often retain something of their previous iconic nature; that is, they have roots in the iconicity of visual perception (Miller 1989; Robertson 2004: 30; Robinson 2003). This is particularly true in the vast majority of Mesoamerican scripts, which maintained a more fluid interaction with iconography. Although the written symbol assumes a new linguistic value, it retains a visual connection to its former iconic significance. This process is not uniform in all cases, and phonetic values may just as easily develop from contextual associations or semantic classification, not just iconicity. In either case, the meaning or value of a specific iconographic sign still becomes separated from a strictly visual interpretation, and a trend toward increasing specificity in the visual message is observed.

‘…the interpretation of a sign’s meaning relies more heavily on the phonetic realization of the iconic representation—even when the phonetic reading of the sign remains tethered to its mimetic value… Thus, the value of iconic signs is determined primarily by how they are used. The difference between their use in a pictorial versus a writing system lies not in their phonetic value—potential or realized—but rather in their function within particular contexts’ (Carrasco and Englehardt 2015: 4). In other words, within pictorial contexts, icons are not necessarily intended to be read off as words. Within a writing system, in contrast, icons are designed to be read as words within a linguistic syntax and with a more precisely defined phonetic value (Carrasco and Englehardt 2015: 3). The development of writing implies a shift from the Peircean notions of icon to symbol (Peirce 1931– 1966). Over time, new visual representational forms may abstract iconic elements, or single elements may assume multiple meanings (a phenomenon known as polyvalence). Consequently, it becomes increasingly difficult to make a specific connection between an icon and its particular or intended meaning. Because an iconographic sign may assume multiple meanings, readers may exploit the emerging tension between visual and auditory perceptions, assigning a new linguistic value to the sign, particular to a single spoken language, in order to identify its meaning. Subsequently, the icon is excised from a specifically visual interpretation and

Alternately, scripts users may recognize the potential to include additional information or more effectively communicate a message through increasing abstraction of iconic elements, as appears to have been the case with the development of writing in the Near East (Cooper 2004; Green 1989; Schmandt–Besserat 1977, 1992).

4

21

Archaeological Paleography The development of early scripts and glottographic writing is a systemic process. This fact does not imply an evolutionary directionality between distinct types of scripts (e.g., pictographic to ideographic, followed by logographic, which invariably leads to syllabic, etc.; cf. Gelb 1963). Nonetheless, certain universal tendencies in the development of scripts have been noted. For example, in those cases in which logographic scripts are borrowed by speakers of a language distinct from that of the script’s original users, the phonetic component of the script tends to expand, making possible and potentially giving rise to syllabographic and alphabetic systems (Justeson and Stephens 1993: 3; Mora–Marín 2001: 12; Trigger 2004). Of course, such tendencies are by no means unidirectional or absolute. A syllabic script may result from a formerly alphabetic script, and some scripts may be detached from purely linguistic principles of encoding (Justeson 1986; Justeson and Matthews 1990; Justeson and Stephens 1993).

the modification of specific visual elements in a more discretely organized fashion, tending toward increasing complexity and specificity, in order to more precisely determine meaning in such circumstances. Shared Features, Linguistic Encoding, and the Development of Mesoamerican Scripts The Mayan script is one of over at least a dozen pre– Hispanic scripts indigenous to Mesoamerica (Houston 2004a; Justeson 1986; Lacadena 2010). A variety of detailed classifications of these scripts have been presented previously (e.g., Coe 1976; Justeson 1986; Justeson and Matthews 1990; Justeson et al. 1985; Marcus 1992a; Mora–Marín 2001: 14–16, 444, fig. 1.7, 445, fig. 1.8), but most may be subsumed under two major traditions: Southeastern and Oaxacan (Figure 2.3; cf. Coe 1965: 769, fig. 57; Justeson 1986: 438, fig. 1). Following Mora–Marín (2001: 13), This investigation is less concerned with the status of the Mayan script within these varying typological classifications than with the factors underlying the spread of various script traditions, and their iconographic precursors, throughout Mesoamerica.

A further quality of nascent writing systems that suggests a systemic component is the function and social context of writing as it is being developed. In many cases, writing develops under specific universal circumstances, and generally in response to common functional needs. These functional necessities include a need for more elaborate and efficient administrative, economic, or clerical record keeping, as appears to be the case in Mesopotamia with Sumerian and Cuneiform (Cooper 2004; Green 1989; Schmandt–Besserat 1977, 1992). Alternately, writing may emerge as a functional adaptive response to religious or socio–ideological stimuli in order to promote order or legitimacy, as Fischer suggests is the case with the development of Egyptian hieroglyphs (Baines 2004; Baines and Yoffee 1998; Fischer 1989), or a combination of both (Englehardt 2005: 56). As Mora–Marín (2001: 13) concludes, the most general possible remark is that the sociolinguistic and cultural processes surrounding the innovation or adoption of a system of writing are often the most cogent determining factors in the development of that script.

There have been three main types of proposals for the origins of Mesoamerican writing systems: monogenesis, polygenesis, and multiregional co–development. Elsewhere (Englehardt 2005), I have argued in favor of a modified multiregional co–development model for the emergence of the Mayan script. Building on previous scholarship (e.g., Houston 2004a; Justeson 1986, 2012; Justeson and Matthews 1990; Justeson et al. 1985), I have suggested the existence of an ancestral Olmec script between c. 1100–400 BC (late Early Formative through the end of the Middle Formative).5 At this

Finally, although the development of writing systems is not necessarily predicated upon interaction in all contexts, it is logical that interaction has the potential to act as a catalyst or accelerant for the sort of structural redeployment and codification involved in script development. Indeed, in order for canonical compositional conventions to fail as an interpretive framework for discerning the meaning of a specific visual element (and its relationship to others), some confusion must exist regarding the representation itself and that for which it stands (i.e., the message it is intended to convey). It is reasonable, therefore, to hypothesize that such interpretive confusion is a potential by–product (or consequence) of long–term interaction between audiences that employ variable languages or representational– interpretive conventions—even as applied to a shared iconographic corpus. Systemic feedback would dictate

Figure 2.3. Classification of Mesoamerican scripts (after Justeson et al. 1985; see also Coe 1976: fig. 1; Justeson 1986; Justeson and Matthews 1990; Marcus 1992a; Mora– Marín 2001: 444–46, figs. 1.7–1.9). 5 Many scholars have rightly noted the problematic nature of the term “Olmec”—or “Olmec style,” “Olmecoid,” etc. (see, e.g., Flannery and Marcus 2000; Grove 1989; Lesure 2004; Rosenswig 2010). In this investigation, I employ the term “Olmec” to refer to an art style prevalent in the Middle Formative period in various regions of Mesoamerica and, following Rosenswig (2010:49), the use of this term does not imply primacy for the Formative period archaeological culture of the Gulf Coast. Thus, I do not suggest that the Olmec culture “invented” writing in Mesoamerica, or that all regional scripts developed directly from an ancestral, specifically Olmec writing system.

22

Theoretical Framework and Methodological Premises time, the ancestral script diverged into two branches, the Southeastern and Oaxacan traditions. Subsequent script systems in Oaxaca and Central Mexico were derived from the Oaxacan tradition, whereas the Southeastern tradition fostered the development of two sub–traditions: Isthmian (also referred to as Epi–Olmec; Justeson and Kaufman 1993, 1997) and Maya–Izapan. In this scenario, an advanced iconographic sign system (or symbol set) was shared throughout Formative Mesoamerica and inherited by various groups who subsequently employed its forms and conventions as the basis for the other representational and script traditions, including the Mayan script. Subsequent independently developed elements were added to inherited visual representational conventions, facilitating the process of linguistic codification (see, e.g., Lacadena 2010).

Such shared iconographic elements can only be appreciated when considered in its functional context within the developmental processes of writing itself. Insofar as writing develops through the concurrent usage of a graphic system within contexts in which canonical representational conventions fail as the primary framework through which to decipher symbolic relationships and content (Justeson 1986: 439), Maya writing emerged when an ancestral representational system encountered itself in distinct and specific contexts in which the insertion of grammatical and/or linguistic principles was necessary to render clear the semantic content of the message itself (Justeson et al. 1985: 34). As Justeson (1986: 439) notes, however, incipient writing only emerges in the presence of a ‘particular necessity’ that demands the reinterpretation of artistic units as linguistic (cf. Justeson 2012).

Justeson et al. (1985: 41, table 16) summarize the degree of agreement in several different formal traits among various scripts from the Southeastern tradition (Mora–Marín 2001: 355–58, tables 1.1–1.4). Justeson and colleagues distinguish between shared formal traits that develop from independent invention and those that result from inherited or diffused innovations. This study follows their assumptions that a greater degree of arbitrariness of a shared feature indicates greater likelihood of descent from a common ancestral iconography or script (cf. Proskouriakoff 1968, 1971). Numerous formal, ideological, ideographic, and representational–conventional associations exist between Late Formative imagery, scripts of the Southeastern tradition, and an antecedent Olmec script or Olmec– style Middle Formative iconographic complex (Carrasco and Englehardt 2015; Coe 1965, 1977; Fields 1991; Freidel and Suhler 1999; Joyce et al. 1991; Martin 2002; Reilly 1991, Riese 1988). These connections imply historical exchange relationships between differing representational systems that employed shared elements and motifs.

The origin of linguistic encoding may lie in the representation of sequential actions, such as that observed in the Middle Formative period Humboldt Celt (Justeson 1986; Justeson et al. 1985; Mora–Marín 2001: 23, 451, fig. 1.14). On this object, a ‘rudimentary logography’ is created by the representation of a series of ritual actions through segmented body parts and isolated implements which refer back to related actions (Carrasco and Englehardt 2015: 4–5; Justeson and Matthews 1990: 91). Mora–Marín (2001: 23, 452, fig. 1.15) suggests that the juxtaposition of numerals with icons referring to names may also be at the root of linguistic encoding, as seen in the Olmec–style Painting 3 from the Oxtotitlán Cave, Guerrero, c. 900–700 BC. Each type of sign represented in Painting 3 may have originated in a separate graphic system, and once juxtaposed in the correct linguistic order they became interpretable in terms of linguistic principles specific to a particular language—most likely a Mixe–Zoquean language, in this case (Justeson et al. 1985; Stross 1982, 1990; cf. Lacadena 2010). According to Sampson’s (1985) classification, the conventions of the Middle Formative iconographic complex were the conventions of a true script (Carrasco and Englehardt 2015, n.d.; Mora–Marín 2001: 23). The fact that so many of these conventions were shared by subsequent scripts in the region (Justeson et al. 1985: 41, table 16) suggests that the Middle Formative Olmec script and/or iconographic complex gave rise to the similar representational frameworks observed in other Mesoamerican scripts. For example, many of the earliest Middle and Late Formative period examples of Mesoamerican scripts, including early examples of the Mayan script, evidence shared sign inventories and orthographic conventions. Such commonalities are unlikely to have developed independently in more than one script and strongly suggest an historical relationship among scripts stemming from intense contact and interaction among scribal traditions (Alfonso Lacadena, personal communication, 2008; Mora–Marín 2001:

From the perspective of social exchange theory, this evidence suggests the spread and sharing of a Middle Formative Olmec representational system that functionally served as a legitimizing device for budding elites within the context of emerging sociopolitical complexity in Formative period Mesoamerica (Flannery 1968a; Houston 2004a; Mora–Marín 2001: 33–6; Reilly 1995). Thus, Olmec–style artifactual symbols and rituals of power and authority (e.g., Maya royal diadems, the Jester God headdress, bloodletting; see Fields 1991; Freidel and Suhler 1999; Joyce et al. 1991; Reilly 1991, 1996) were adopted and subsequently deployed in differing spatial, temporal, and cultural contexts (cf. Carrasco and Englehardt n.d.; Stuart 2015). In other words, Mesoamerican scripts and artistic systems shared a common iconographic vocabulary, in terms of a collective pictography, ideography, and—arguably— underlying ideological base. 23

Archaeological Paleography 25–6, 359–60, table 1.5; Pohl et al. 2002: 1985, fig. 3; Pohl et al. 2008; Rodríguez Martínez et al. 2006; Saturno et al. 2006; this relationship is not necessarily genetic; see Proskouriakoff 1968, 1971). Additionally, the themes of political legitimation and name–tagging frequent in early exemplars of Mesoamerican writing are likewise common in subsequent Mesoamerican scribal traditions. In the specific case of the Isthmian (or Epi– Olmec) and Mayan scripts, a close historical relationship is evident (Justeson and Kaufman 1993; Lacadena 1995; Macri 1991; Méluzin 1992, 1995). These scripts share numerous signs, orthographic conventions, and calendrical components. Although some shared signs have differing values and functions in each script, the majority of signs common to both scripts have similar forms in the antecedent script and Middle Formative period iconography, suggesting common descent (Mora– Marín 2001: 27, note 20).

established in Mesoamerica since the Early Formative period (Flannery 1968a). The exchange of objects that contained the stylistic and functional predecessors of the Mayan script was predicated upon several centuries of prior interregional exchange of such preciosities and other objects of various types. These exchanged objects may have constituted a significant medium of political–economical exchange, given their political and ideological themes and functions, serving to integrate Mesoamerica as a cultural and economic whole (e.g., Blanton et al. 1996; Flannery 1968a; Mora–Marín 2001: 25). Potential confusion regarding the meaning of the messages conveyed by such objects may have constituted the ‘particular necessity’ that demanded the reinterpretation of artistic units in a new linguistic framework, as Justeson (1986: 439) suggests. Current models of the development of Mesoamerican writing thus suggest that its developmental processes are linked to systemic interaction (Chinchilla Mazariegos 1999; Fahsen Ortega 1999; Houston 2004a; Justeson 1986; Justeson et al. 1985; Justeson and Matthews 1990; Marcus 1992a; Mora–Marín 1997, 2001; Stross 1990; Urcid 2001). These models explain the emergence of writing in terms of the re–situation of visual symbols shared widely across Mesoamerica within new, locally specific grammatical–linguistic) frameworks as a corollary of interregional interaction (Salomon 2001; Senner 1989). Sustained, long–term interaction between groups employing a common iconographic complex juxtaposed new meanings and linguistic values in unprecedented ways, making possible the redeployment and phonetic codification of iconic elements.

Of course, scholars agree that these scripts represent different languages and were used by distinct cultural groups. Nonetheless, in this case linguistic data indicate that the majority of internal linguistic influence (i.e., diffusion into Lowland Mayan languages) derives from Cholan, a language group centered in interior Chiapas, which is also the language represented in and tied to the Classic Maya writing system (Campbell 1984: 12–13; Justeson et al. 1985: 10, table 2a; cf. Houston et al. 2000). The majority of external linguistic influence (i.e., diffusion into Cholan) derives from the Mixe–Zoquean language family—the likely language of both Olmec writing and the Isthmian script (Campbell 1984: 7, 9–10, 12–13; Hruby and Child 2004; Justeson et al. 1985: 4; Kaufman and Norman 1984). When combined with the iconographic data indicative of shared symbolic–representational and scribal traditions, these linguistic data strongly suggest direct Olmec representational and Mixe–Zoquean linguistic influence on the historical development of the Classic Mayan script (Campbell 1976; Campbell et al. 1986; Englehardt 2005: 151, fig. 4.5; Justeson 1985; Justeson et al. 1985; Lacadena 1995, 2010; Lowe 1977, 1989). Thus, the interplay between a shared symbolic–iconographic complex and Mixe–Zoquean linguistic influences resulted in the reformulation of the ancestral Middle Formative Olmec iconographic complex that ultimately gave rise to the Classic Maya writing system.

The Emergence and Nature of the Mayan Script Previous scholarship suggests that the Classic Maya writing system emerged in the manner detailed above from a shared iconographic system or precursor script (Fields 1991; Houston et al. 2000: 328; Joyce et al. 1991; Justeson 1989; Martin 2002; Reilly 1991, 1996; Schele 1999). In the Mayan script, linguistic codification resulted in a logo–syllabic writing system, in which signs could code individual morphemes or polymorphemic stems or words. In the Mayan script both types of logograph were used (Mora–Marín 2001: 12; Stross 1990). Like Chinese writing, in the Mayan script logographs are frequently accompanied by a phonetic indicator to determine the specific semantic meaning and pronunciation of polyvalent signs (Bagley 2004; Justeson 1976; Keightly 1989). In principle, any logograph could be used phonetically for near homophonous sequences within the system (Mora–Marín 2001: 12). Thus, Maya writing is essentially phonographic, employing signs with both semantic and phonological values.

By the end of the Late Formative period, hieroglyphic writing with single– and double–column format was in use throughout the lowland Maya region (Mora–Marín 2001: 465–69, figs. 1.28–1.32, 359–60, table 1.5). Previous research has suggested that the origin and spread of the Middle Formative iconographic complex— or ancestral script—that gave rise to subsequent regional scripts was likely facilitated by long–distance exchange, primarily of inscribed luxury or prestige display items intended for ritual use (Justeson 1986; Mora– Marín 2001: 24–5). This exchange system had been

Individual signs in the Mayan script remained in continuous dialogue with their iconic precursors. As discussed above, 24

Theoretical Framework and Methodological Premises

Figure 2.4. Acrophany and reformulation in Maya writing. a: T740 hu, hu, ‘iguana;’ phonetic sign; represents the upended head of a lizard or other reptile; b: T740:121.126 hu–li–ya, huliiy, intransitive verb, ‘arrived;’ c: T740.23 hu–na, hun, ‘paper,’ ‘book,’ ‘paper headband;’ d: T740:23:178 hu–na–la, hunal, ‘headband.’ Drawings by Pearl Lau.

and auditory perception of an iconic sign. Iconically, the sign depicts the upended head of an iguana (Figure 2.4a). Readers were able to exploit the tension between the icon’s mimetic visual value (hu as iguana) and its auditory significance (hu or ju as morphemes within Mayan languages) to assign it purely linguistic meanings: the logograph ‘HU’ as the word for iguana, or the syllabogram ‘hu’ to represent the morpheme. Due to its polyvalency, the sign could then be used in phonetic contexts to construct syllabic spellings of words specific to Mayan languages, such as hul, a verb meaning ‘to arrive’ (Figures 2.4b, c, d). In these cases, the iconic element is divorced from a specifically iconographic interpretation, despite the maintenance of representational continuity. Maya sign T604 k’u (nest; Figure 2.5) is another example that illustrates this concept.

Figure 2.5. a: k’u (k’u) (T604) ‘nest;’ phonetic sign; b: k’u– xa–ja (k’uxaj) (T604:114.181) passive verb; ‘was eaten;’ ‘was ground;’ ‘was hurt.’ Drawings by Pearl Lau.

signs derived through the process of acrophany will retain a degree of iconicity, remaining rooted (to varying degrees) in the iconicity of visual perception and moored to their previous iconic significance(s) regardless of their script–specific meaning or codified grammatical/linguistic value (Miller 1989; Robertson 2004: 30). For example, the Maya sign T740 hu (‘iguana’) illustrates the multiple latent potentialities that result from the interplay between visual

Similarly, T632 muyal (‘cloud’) illustrates the redeployment of a widely–shared iconographic element (Reilly 1996; Figure 2.6). The icon appears in Early Formative period contexts at Chalcatzingo (Monuments 1 and 31) in the central Mexican highlands. Over time,

Figure 2.6. Lazy–S / cloud / T632 substitution set (drawing by Pearl Lau after Reilly 1996: 414, fig. 3).

25

Archaeological Paleography possibly as a result of sustained interaction and systemic feedback, new representational forms abstracted the shared visual element that represented ‘cloud.’ When readers were no longer able to make the specific connection between the icon and its intended semantic meaning, they imposed a linguistic value on the icon, creating a written symbol. In this case, given that T632 retains a great deal of continuity with its iconic predecessor, it is likely that the imposition of a particular linguistic value on the sign in Maya writing was intended to impart further specificity in the visual message—not unlike a phonetic indicator that aids in the decipherment of semantic meaning. T632 is also instructive in that it illustrates the re–situation of an ‘ancestral’ visual motif in new contexts to reflect locally specific linguistic values—although in this case the specific linguistic value is identical to the semantic value of the iconic precursor.

Evaluating The Relationship between the Development of Writing and Material Interaction in Formative Period Mesoamerica The principal aim of this investigation is twofold: to evaluate the degree to which processes of interaction are involved in the development of the Maya script and to determine how the processes of interaction involved in script development may be traceable archaeologically, or correlate with specifically archaeological data. Insofar as interregional interaction is fundamentally connected to the development of Mesoamerican and Maya writing, the relation between the emergence of writing and interregional material interaction is no coincidence (cf. Cherry 1986: 32). Indeed, writing itself may be considered an archaeological object, especially given that the objects on which texts were inscribed are themselves material goods (Alfonso Lacadena, personal communication, 2004). In Mesoamerican contexts, Mora–Marín (1997, 2001) argues that the interregional exchange of portable objects was linked to the development of the Mayan script. Thus, texts and inscriptions may be understood in two ways: ‘as purveyors of content and as archaeological artifacts’ (Houston 2004c: 229). All material culture is a product of knowledge systems inherited from the past and transformed as a result of the pursuit of greater technical efficiency or particular sociocultural interests (Trigger 2004: 67). Just as writing develops out of a ‘crisis of meaning,’ material cultural forms and usages exhibit diachronic changes that indicate a similar process of localized imposition of unique meaning or function. In archaeological terms, the functions and forms of shared material culture vary over time, rather than remaining static.

The Classic Maya writing system therefore appears to have emerged via the tension that stemmed from the attempt to impart culturally–specific values on a precursor visual complex—most likely what Reilly (1995: 29–30) has labeled the Middle Formative symbolic–ceremonial complex. Insofar as writing systems develop from the interplay that results when the visual elements of a shared system of graphic communication encountered in multiple contexts are reformulated and redeployed to reflect locally– or script–specific values, it is clear that interregional interaction was intimately involved in the development of all Mesoamerican scripts—including the Mayan—as numerous scholars have suggested (Englehardt and Carrasco 2015; Houston 2004a; Justeson 1986, 2012; Justeson and Matthews 1990; Justeson et al. 1985; Lacadena 1995, 2010; Mora–Marín 1997, 2001; Stross 1982, 1990).

Many models of innovation and cultural development in Mesoamerica (Blanton et al. 1996; Marcus 1992b, 1998; Willey 1991) suggest that ideas entwine with material goods in generative ways through interaction. As with writing systems, a variety of models explain sociocultural and material transformations as a correlative function of systemic interaction among autonomous units at the interregional level (Cherry 1984, 1986; Goldstein 2000; Parkinson and Galaty 2007, 2010; Renfrew 1972). Such models posit that cultural changes of both kind and degree result from processes of long–term interaction (Cherry 1986: 24–7, 43; Renfrew 1972: 495), which is as evident in the distribution of material remains as in the distribution of reconfigured iconographic symbols. Both phenomena rely on sustained interaction between discrete social groups that come together in specific areas distributed across a bounded landscape. Evidence of such interaction thus takes the form of similarities and differences within an identifiable complex of material culture widely distributed across the boundaries of the interacting groups. In the Mesoamerican Formative period, prior research has indicated the existence of historical relationships, though not necessarily ‘genetic’ in nature, between and across cultural groups and, by

The ancestral Middle Formative iconographic complex was widely shared throughout Mesoamerica through interaction (see, e. g., Flannery 1968a), providing the visual base on which subsequent scribal traditions developed. Material and linguistic interaction between groups that developed and employed these scribal traditions potentially created the crisis of meaning that resulted in the particular necessity for increasing specificity in systems of graphic communication, resulting in the linguistic codification that, in this case, gave rise to Maya writing. Such systemic interaction is evident in numerous formal and functional features, orthographic conventions, semantic and contextual parallels, and linguistic aspects shared between all Mesoamerican scripts, particularly those in the Southeastern group, including Maya writing (Englehardt 2005; Justeson et al. 1985: 41; Lacadena 2010; Marcus 1992a; Mora–Marín 2001: 355–58). Thus, this investigation departs from the well–founded premise that Maya writing developed out of a preexisting Middle Formative Olmec sign system shared throughout Mesoamerica (Houston 2004a: 286; Justeson 2012: 838). 26

Theoretical Framework and Methodological Premises extension, specific cultural technologies, such as ceramic technologies, iconographic and artistic conventions, or writing systems or scribal traditions (Parsons 1979, 1986, 1988; Proskouriakoff 1968, 1971; Quirarte 1977).

In functional terms, Mesoamerican scripts and their iconographic precursors exhibit a variety of uses. Their thematic content is normally ideological in nature. Consequently, scripts and iconography are strongly linked to ritual–religious processes of validation and legitimization of emerging elites (Fields 1991; Martin 2002; Reilly 1995; Schele 1999). The fact that many early examples of writing in Mesoamerica occur on portable luxury items (i.e., elite goods such as celts, jewelry, or prestige objects) supports this contention (Mora–Marín 1997, 2001). It is likely that the possession of script technology was itself just as much a status symbol as the possession of the exotic material goods (Alfonso Lacadena, personal communication, 2008). The spatial distribution of such items may also indicate the extent and directionality of regional exchange networks.

In trans–regional contexts, differential interpretive principles prompt all artifacts and material goods to assume meanings previously unassociated with them (Clark 2004; Renfrew 2001). Innovative artifactual and iconographic forms and functions thus emerge. For example, many suggest (e.g., Flannery 1968a: 111; Graham 1989; Lowe 1989; Sharer 1989) that emerging Mesoamerican elites borrowed an Olmec material–symbolic complex to justify their developing authority through association with Olmec prestige (Clark 2004; Clark and Blake 1994). Over time, that complex became independent, divorced from its initial reference to the Olmec, and functioned purely within locally explicit contexts. Because formerly ‘exotic’ objects no longer functioned to illustrate a connection with the prestige of the Olmec as they initially did, a change in meaning must have occurred. Over time those objects acquired reference to a local system. Likewise, in functional terms, recent investigations of chocolate pots (e.g., Henderson and Joyce 2006; Hurst 2006), have demonstrated diachronic changes in the form and function of these ceramic vessels, which correspond to changes in the preparation of cacao beverages and their increased ritual significance, as well as a greater number of textual references to cacao and variation in cacao symbolization. This example illustrates how variation in ceramic materials may be linked to emerging script technologies.

In terms of this investigation, systems–based theories of network and social exchange allow for the identification and measurement of relationships between interacting units within larger regional systems. Extending this theoretical framework, it is possible to identify and measure the relationships between the distinct material correlates of that interaction. Comparative analyses of material data allow for a preliminary illustration of the nature, extent, and temporal depth of interregional interaction, the boundaries of interacting units, and the identification of functional and technological innovation through the development of localized styles (Hegmon 1998; Stark 1998). Such measures also speak to relative degrees of integration within developing regional systems. By extension, these measures may also illuminate processes of increasing centralization, given the correlation between centralization and integration in regional systems, as discussed above.

Interaction thus has the potential to encourage the stylistic and functional attributes of material culture to change over time, just as it seems to encourage emerging scripts to rework the stylistic and functional attributes of shared iconography in order to express particular and contextually specific meanings or linguistic values. Stylistically similar material culture that appears in functionally distinct contexts indicates both interaction and a local imposition of meaning, paralleling the developmental process of writing in Mesoamerica. In both cases, the recontextualization of style and function in distinct situations effects a structural transformation that gives rise to new meaning (Hodder 1987). In short, both icons and material objects carry specific yet potentially mutable latent meanings and values (whether realized or not) that may be mixed and matched to convey distinct, culturally–specific connotations (Clark 2004: 218; Renfrew 2001). In ceramic materials, the presence of widely–shared stylistic attributes implies extended interaction. Formal or functional discrepancies in such material data suggest innovation in terms of differential manufacturing processes, the development of localized styles, or the introduction of new material technologies (Stark 1998; Stark et al. 2000).

Methodologically, to evaluate the degree to which material innovation corresponds to script development, it is necessary to add a quantitative measure of variation in the archaeological materials. Statistical measures of similarity and diversity serve to quantitatively evaluate relative variability between and among samples and datasets and complement a more general comparative analysis (Leonard and Jones 1989; Shennan 1988). Similar cluster patterns of attribute similarity across assemblages and lower intra–assemblage diversity would suggest greater distributional uniformity between and within material complexes, and therefore a relatively greater degree of interaction on a regional scale (Kintigh 1989; Parkinson 2002, 2006). These measures of similarity and diversity also speak to material innovation, insofar as greater inter–assemblage variability and increased diversity within complexes suggest the emergence of new forms and functions (Hegmon 1998; Stark 1998). Comparing clustered patterns of similarity with heterogeneity measures in distinct spatial and temporal contexts and on different scales thus may illuminate variability in the quantity, intensity, and directionality of 27

Archaeological Paleography expanding Formative interregional exchange networks. If interaction is related to observed variation, the patterns of interaction suggested by the comparative and statistical analyses should closely correspond in temporal and spatial contexts.

this recontextualization led to a sort of interpretative breakdown (the ‘crisis of meaning’ discussed above), in which the audience could no longer determine the relations of icons to one another, or their intended meanings or semantic values. Readers had to modify the iconographic script, infusing icons with grammatical and linguistic properties, thus establishing new values within a culturally–specific framework to more precisely determine meaning and to understand the relationships among signs in a given text.

Further, if interregional interaction may be correlated across developmental processes, changes in material culture should parallel iconographic reconfiguration as icons transform into written signs. Stylistic similarity in distinct material and iconic/scribal forms would indicate the interaction that drives innovative development, whereas discrepancies in form, function, or distribution may suggest local imposition of meaning. In terms of the central hypothesis of this investigation, stylistic evidence of a systematic transition towards greater formal and functional variability in ceramics should correspond to a greater formal diversity, functional complexity, and semantic specificity among signs in the emerging Mayan script. Increasing abstraction in and localized adaption of material forms should positively correlate with relative degrees of distance from original iconic referents as icons transform into written symbols and acquire new meanings and values.

Modifications and variability in the formal characteristics of the script reflect this recodification stemming from recontextualization. Often, composite signs were established by making very slight modifications or adding small visual elements to preexisting icons (Lacadena 1995). The new features drew on the iconicity of the original while adding culturally specific visual or linguistic markers in order to aid in identification of the sign and its meaning. Such composite signs often incorporated semantic mnemonic information, or rely on the rebus principle for interpretation. Although new composite signs continued to be introduced, some other distinctive sign forms were reshaped and simplified. With further linguistic codification, sign combinations increasingly employed phonetic markers. Eventually, the linking of signs in semantic or phonetic units began to approximate written phrases of a language and thereafter, with the addition of grammatical indicators, approached sentence structure (Green 1989: 45–6; Hanks 1989a; Justeson 1989; Miller 1989).

Recontextualization The examples of script development, Oaxacan appropriation of Olmec iconography, and functional variation in chocolate pots outlined above illustrate the concept of recontextualization. Mesoamerican writing systems developed out of a sophisticated ancestral iconographic complex. Many Mesoamerican scripts retained their artistic–representational quality to a great extent, remaining essentially iconic on the level of the individual sign and linguistic on the level of the relationship among those signs (Carrasco and Englehardt 2015: 3). Since the Mayan script also developed out of this widely shared Middle Formative iconographic complex, the visual elements and signs that comprise the writing system often exhibit a high degree of formal similarity with other Mesoamerican scribal traditions, particularly those of the Southeastern group (Coe 1977; Justeson 1986; Justeson and Matthews 1990; Justeson et al. 1985; Lacadena 2010; Mora–Marín 2001; Pohl et al. 2008; Riese 1988; Stross 1990). In the developmental dynamics of writing, such formally similar icons are modified or recontextualized to represent more closely the particular grammatical or linguistic values of a specific language, thus transforming into written signs.

The same process may potentially occur with any material object. When an artifact is excised from its original context or interpretative framework, new principles are applied to its interpretation, and new locally–specific meanings may be assigned to the object (Bird and Smith 2005; Clark 2004; Renfrew 2001; Wobst 1977). As with scripts, such changing meanings may be reflected by variable formal or functional characteristics of the object (e.g., changes in the formal properties of an initially Olmec–style material–symbolic complex, or the functional contexts of chocolate pots, as discussed above). Thus, the recontextualization of the artifact potentially fosters diachronic stylistic or functional variability of the material object itself. Paralleling iconic recontextualization, such emergent material forms should exhibit slight modification of the original material template and possess a great deal of similarity to original forms.

As detailed above, such modification and transformation occurs when similar icons are encountered in new contexts, in which visual interpretive principles can no longer sustain meaning adequately solely on the basis of established interpretive conventions (Coe 1976; Hopkins 1997; Justeson et al. 1985: 34; Robinson 2003; Rogers 2005). In the case of the Mayan script,

It is suggested that interregional interaction itself drives the process of recontextualization (cf. Englehardt and Carrasco 2015). Through interaction, objects and ideas are shared in multiple contexts across geographic regions or cultural groups. In such a situation, the originally intended meaning or function of an artifact is potentially open for reinterpretation by multiple audiences. 28

Theoretical Framework and Methodological Premises Continuing and expanding Houston’s (2004b) analogy, the object is originally closed within a single established interpretative framework, becomes opened to multiple interpretations through interaction, which places the object in variable contexts, and subsequently becomes closed within a single interpretative system again through recontextualization. The changing meaning of the object as it is closed within its new significant framework is reflected in formal and functional changes. Interaction thus fosters recontextualization and material change by creating the conditions for potential changes in meaning to occur, whether in material artifacts or iconic elements. Interpretive Framework: Correlating Diversification and Material Change

spatial and temporal patterns of exchange evident in the distribution of material remains should match similar patterns in the iconographic and linguistic data indicative of interaction. Variation in stylistic attributes and functions of material culture should correspond to similar formal and functional iconographic and linguistic variability. Localized innovative forms or functions of broadly distributed material culture should parallel the recontextualization of visual signs involved in the transition from iconography to writing. If a correspondence between the two patterns is evident, then it would follow that interregional interaction is at the root of both phenomena, and that a relationship exists between script development and material innovation in this context. The crisis of meaning behind the emergence of Maya writing does not represent interaction; rather, it stems from it. Continuity and disjunction in material remains would suggest the operation of similar processual dynamics in both cases. In effect, stylistic and functional variability and quantitative distributional variation reflected materially should parallel iconographic and linguistic transformations suggestive of interregional interaction in the emergence of Maya writing.

Script

It is not suggested that writing always emerges from the crucible of interregional interaction. Nor does the identification of interaction in itself account for the development of writing. Rather, it is suggested that in the specific contexts of Late Formative and Early Classic period Mesoamerica, the developmental dynamics of writing can be correlated with sociopolitical and economic interaction. Documented iconographic and linguistic shifts in the developing Maya writing system (Englehardt 2005; Lacadena 1995, 2010) should correspond spatially and temporally to material variability as well as transformations in the sociopolitical context within which they occurred, such as changes in the extent, direction, or intensity of interaction or the degree of integration in larger regional networks.

Thus, comparing diachronic variation in material data against iconographic and linguistic transformations should yield similar patterns of variability in identical spatial and temporal contexts. Available evidence indicates that Maya writing emerged in the Late Formative period (Houston 2004a; Justeson 2012; Saturno et al. 2005, 2006; Taube et al. 2010). If a relationship exists between interaction, script development, and material innovation, one would expect to encounter patterns within micro–regional analytic units that indicate a greater degree of interregional interaction and less relative variability between the sample and regional phases and sequences in Middle and Late Formative temporal contexts, followed by a decrease in interaction and increase in material variation in the Early Classic period as localized imposition of cultural meaning on icons and material artifacts intensified in the Maya lowlands. If such a pattern exists, it would confirm the centrality of interaction to cultural innovation in this case, and would also suggest a correlation between the dynamics of the developmental processes.

Understanding the recontextualization of iconic elements shared among various representational systems and scripts in discrete contexts is complex and highly contextualized. The individual and collective significance, semantic meaning, and potential phonetic value(s) of given icons and signs are intertwined in the two complementary historical processes of linguistic change and script development. Although tracing script differentiation differs from tracing linguistic divergence, the two processes can be comparatively examined and correlated. Historical linguistics shows that boundaries and frontiers are certainly associated with linguistic innovation. The correlational approach I suggest should prove a useful new line of evidence, but not one that will measure linguistic affiliation or differentiation in any direct way. Nonetheless, archaeological indicators of interaction and sociocultural change may be compared with iconographic and linguistic data to yield insights into the nature and processual dynamics of interaction and cultural innovation (Hruby and Child 2004; Josserand 1975; Josserand and Hopkins 1999; Justeson et al. 1985).

To that end, a sample database of ceramic materials, themselves excellent variables through which to measure interaction (Demarest and Sharer 1982; Holley 1983, 1987; Neff et al. 1999; Pring 1977; Rice 1996a, 1996b; Skibo et al. 1989) is compared with observable changes in the form, function, syntax, distribution, and textual recurrence of certain widely–shared Mesoamerican iconographic motifs and identifiable signs within discrete script systems. If interaction and recontextualization are behind material and scribal innovation and variability in this case, one would expect iconographic and scribal variation to correspond to formal and functional variation

In this specific case, variability in material, iconographic, and linguistic data should exhibit strong spatial and temporal correspondence if in fact interaction drives both developmental processes. Quantitatively, 29

Archaeological Paleography in ceramics between Late Formative and Early Classic period regional ceramic phases. Patterns indicative of a systematic transition towards greater technical complexity, formal innovation, and functional variability over time should be evident in the material data, which would correlate with increasing use and greater complexity of the emerging Mayan script. Increased variability in and localized adaption of material forms should spatially and temporally correspond to formal and functional iconographic and linguistic variation as icons transformed into written symbols and acquired new meaning. Insofar as new cultural meanings may be reflected in new material forms, changes in the material data thus parallel iconographic reconfiguration as icons become written signs. Material variability indicates innovation in terms of localized adaptations of diffused cultural technologies or site–specific singularities. Stylistic and distributional similarity in the material data reflect the interaction that drives the developmental process, whereas discrepancies in form, function, or distribution suggest localized innovation driven by recontextualization, paralleling dynamics of diffusion and divergence (i.e., external vs. internal influence) evident in the development of Maya writing.

usages exhibit diachronic changes that indicate a similar process of localized imposition of unique meaning and/or function. It is suggested that evidence of such processes should be most visible at the boundaries of discrete spheres of interaction. In Mesoamerican contexts, current models suggest that writing emerged as regional iconographic systems assumed different meanings through interaction in trans– regional contexts. Fixing variable meanings to icons in the context of interregional interaction detached them from previous interpretations to facilitate their new use as written signs. The model proposed in this investigation builds on and complements previous research in attempting to associate archaeological remains with the emergence of writing. The development of new styles and functions in the material remains of interregional exchange should temporally and spatially correspond to linguistic and iconographic changes involved in the transition from iconography to writing.

Conclusions

This research therefore examines the development of Maya writing in Formative period Mesoamerica in a new light by exploring the material expression of larger, long–term patterns of interaction and the relationship between interregional exchange and the broad processes involved in the development of the Mayan script. In doing so, the investigation draws on and refines current scholarship to further anthropological understanding of the nature and development of writing, as well as its relationship with material culture and interregional interaction. The attempt to synthesize information from distinct sources relies on the implicit assumption that script development, like linguistic history, is intimately connected with material cultural history (Josserand 1975: 501). In this sense, writing served as a ‘symbolic intensification’ (Hanks 1989) that both paralleled and encouraged material forms of intensification and production. The interaction of distinct, interdependent groups emerged as both a material and an ideational phenomenon, a symbolic exchange of artifacts and information. Tracing the emergence of writing in archaeological and linguistic terms will simultaneously offer a more detailed understanding of this complex cultural technology.

This chapter has outlined the theoretical and methodological premises by which this study seeks to associate material innovation and variation with iconographic and linguistic data involved in the development of the Mayan script. Archaeologists have long claimed that interaction is central to cultural innovation, particularly in terms of the development of writing. Previous archaeological research suggests that systems and complexity theories are productive frameworks through which to elucidate links between cultural innovation and interaction within and between regional groups. Just as writing develops out of a ‘crisis of meaning,’ material cultural forms and

This investigation departs from the methodological premise that an analysis of stylistic and functional variability (Hurt and Rakita 2001; Stark 1998) in material markers of interaction can elucidate the developmental dynamics of the symbolic categories that underlie material exchange and writing. If the development of the Mayan script is related to systemic material exchange, then stylistic and functional variability in the distribution of material artifacts should indicate both degrees of interaction, integration, and differential use of material culture, paralleling the emergent transformations of diffused icons into written signs. To evaluate this

To evaluate this premise, and to determine the association of script development with material innovation and variability, it is necessary to study an area that will allow the identification of these long–term processes at the regional scale in wide–ranging discrete temporal contexts. The region in which the sites selected for study are located abuts areas in which examples of Middle Formative iconography, early Mesoamerican scripts, and Maya writing have been found (e.g., Ochoa 1978, 1983; Ochoa and Hernández 1975; Pohl et al. 2002; Saturno et al. 2006). Due to its situation as an integral part of a broader pattern of Mesoamerican interaction and interregional exchange (Navarrete 1978; Ochoa 1983; Vargas and Ochoa 1982), the focus area of my investigation is an excellent test case to examine the material expression of long–term patterns of interaction and the relationship between interregional interaction and the development of the Mayan script.

30

Theoretical Framework and Methodological Premises proposition, it is necessary to investigate an area that can potentially illuminate the broader patterns of interaction and exchange involved in the processes of script development. Such an area should present evidence of long–term, regional processes of interaction in broad temporal and spatial contexts. The northwest Maya

lowlands present just such a test case. This investigation focuses on the region of southeast Tabasco State, Mexico, a key historical crossroads of Formative period trade and interchange on the border of early Mesoamerican interaction spheres, to which we turn in the following chapter.

31

Chapter 3 The Northwest Maya Lowlands: Site Selection and Regional Background The focus area of this investigation is a test case in which it may be possible to discern the archaeological correlates of the broader patterns of interaction and exchange that occurred across a particularly well– situated region. In terms of script development, evidence of incipient script technologies should be encountered at the peripheries of cultural interaction spheres. The northwest Maya lowland region of southeastern Tabasco (Figures 1.2, 1.3, and 3.1) abuts areas in which evidence for developing Mesoamerican scripts, dating to at least the Late Formative period, has been found, such as the southeastern Gulf coast lowlands of northwestern Tabasco and the central Maya lowlands of the Petén, Guatemala (e.g., Pohl et al. 2002; Saturno et al. 2006). Documented iconographic and linguistic evidence suggests widespread iconographic and linguistic interaction across the northwest Maya lowlands throughout the Formative period (Fields 1991; Joyce et al. 1991; Justeson et al. 1985; Martin 2002; Reilly 1991; Schele 1999). Prior linguistic analyses (Justeson et al. 1985; Kaufman 1976; Lacadena and Wichmann

1999) confirm that this area was indeed a nexus of interaction located in a frontier zone of the greater Maya lowlands. Therefore, due to the location of the four sites selected for study within a frontier region on a boundary between both script and ceramic traditions (Figure 3.2) that exhibits evidence of material, iconographic, and linguistic interaction throughout the Formative period, the ceramic data from these sites are appropriate for the analyses proposed, and are well suited to answer the kinds of questions this research poses. This study employs ceramic evidence form four sites in the San Pedro Mártir River basin of southeastern Tabasco, Mexico: Tiradero, El Mirador, Revancha, and San Claudio (Figures 1.3 and 3.1). These sites are second and third tier centers that exhibit evidence of occupation dating from the Middle Formative through the Classic period (Fournier 1983; González Lauck 1993; González Moreno 2006; Hernandez Ayala 1981; Ochoa 1978, 1983). Previous investigations have revealed evidence of various material homologies in the ceramic materials derived from this

Figure 3.1. Map of the central and northwest Maya lowlands, showing sites included in this study and their location in relation to other Classic period Maya centers.

32

The Northwest Maya Lowlands: Site Selection and Regional Background

Figure 3.2. Distribution of early Mesoamerican script groups overlying distribution of Early Formative ceramic traditions (1150–850 BC). The bold black lines separate the Oaxacan, Southeastern, and Mayan script traditions. The light dashed lines indicate the extent of the Early Formative Locona ceramic tradition (light gray stippled area), with red–on–buff to the west (horizontal hatches). Sites included in this study are shown in cross–hatched gray box. (after Flannery and Marcus 2000: 9, fig. 3; Justeson 1986: 438, fig. 1; Rice 2007: 168, fig. 8.2).

northwestern frontier of the lowland Maya core area of the Petén (Lee 1978; Navarrete 1978; Price 1978). These sub– regions include interior Chiapas, the western Maya area, the Gulf coast Olmec heartland, and the Southern Maya Lowlands of the Petén (Demarest 1989: 337, fig. 13.2).

area, as well as shared aesthetic traditions and similarities with adjacent regional sequences. This documented correspondence strongly suggests processes of sustained, intensive interregional exchange between southeastern Tabasco and adjacent sub–regions throughout the Formative and Classic periods (e.g., Clark and Cheetham 2002; Golden and Scherer 2006; Lee 1978; Navarrete 1978; Ochoa 1983; Ochoa and Casasola 1991; Parsons 1986, 1988; Price 1978; Proskouriakoff 1968, 1971; West et al. 1985). Prior surveys of ceramic materials from the northwest Maya lowland region intimate the presence of exchange networks stretching far back into the Early and Middle Formative periods (Braswell 2003; Ochoa 1983; Price 1978). Regional analyses of Late Classic period ceramics (e.g., García Moll 2005; González Moreno 2006; Hernández Ayala 1981; Hernández Pons 1984; Ochoa and Casasola 1991; Sanchez Caero 1979) suggest that the region in which the selected sites are situated was a nexus of interaction located on a frontier between Classic period Maya ceramic spheres.

Regional Context Classic Maya society developed in southeastern Mesoamerica, in what today are the countries of Mexico, Guatemala, and Belize, as well as the western portions of the nations of Honduras and El Salvador. This territory is usually divided into three broad zones: the northern lowlands of the northern half of the Yucatán peninsula, the central lowlands of the southern Yucatán peninsula, Belize, and the Petén department of Guatemala, and the southern highlands of Guatemala and central Chiapas (Figure 3.3). The highland areas are fairly well–defined, but there is no clear demarcation between the northern and central lowlands, and these are usually considered one area, called the greater Maya lowlands. In general, the term ‘lowlands’ is applicable to those areas that do not exceed 800m above sea level and that are demarcated by the traditional limit between the hot, humid tropical coastal plains and the cooler temperate regions of the highland interior.

The materials recovered from these sites are particularly apt for identifying the nature, extent, and direction of interregional interaction precisely because they are situated along established trade routes at a boundary between distinct Mesoamerican sub–regions, on the 33

Archaeological Paleography

Figure 3.3. Map of Mesoamerica. The extent of the Classic Maya area is roughly outlined in the light grey overlay, with traditional internal highland–lowland divisions noted. The central Maya core area of the Petén is highlighted in the grey cross–hatched box. The study area is shown as the small black cross–hatched box. Sites and Mesoamerican regions discussed in the text are also detailed.

area reveal a preliminary regional sequence that stretches back at least to the later part of the Early Formative period (Ochoa 1977, 1978; Ochoa and Casasola 1983, 1991; Rodrigo Liendo, personal communication, 2014). Habitation of the region was relatively sparse during the Early and Middle Formative periods. Nevertheless, previous investigations of the area revealed a significant amount of Olmec–style objects in the northwest lowlands (Ochoa 1977, 1978, 1983; Ochoa and Casasola 1983, 1991; Ochoa and Hernández 1975). During the Late Formative period, Chicanel phase ceramics begin to appear in the region, followed by phases affiliated stylistically with the Early Classic Tzakol and Tepeu spheres (although in fewer numbers than in the core Maya area of the Petén). From these scant data, it appears that southeast Tabasco was located on a border area that was intimately involved in processes of long–term interaction between distinct Mesoamerican regions.

The study area is located within the geographic area known as the northwest Maya lowlands, a region that extends north and west from the middle Usumacinta and upper San Pedro Mártir river basins in the extreme southeastern corner of Tabasco along the Guatemalan border, and comprises the majority of the modern state of Tabasco, as well as southwestern Campeche and northern Chiapas (Ochoa 1977, 1978; see Figure 1.2). This micro– region sits near the natural physiographic division between the humid Gulf coastal plain and the cooler highlands of Chiapas. All sites included in this study lie within this lowland area on the northwest frontier of the Maya world. The site with the highest elevation is San Claudio, at just under 90m (González Moreno 2006: 24, fig. 23). The lower San Pedro Mártir basin lacks exaggerated topographic features, and the river itself functioned as a vital route of communication and trade, facilitating transport and commerce between sites within the region and both the Gulf coastal plain and centers further inland (Hernández Ayala 1981: 16; Navarrete 1978; Ochoa 1978: 65, 69; Vargas and Ochoa 1982). Its situation at the junction of highland and lowland zones made this area particularly open to influence from adjacent regions and cultures throughout Mesoamerican history.

In the later Early and Middle Formative periods, the San Pedro Mártir basin of Tabasco was situated astride the easternmost limit of the distribution of the Locona ceramic tradition, ubiquitous in southeastern Mesoamerica (see Figure 3.2; see also Flannery and Marcus 2000: 9, fig. 3; Rice 2007: 168, fig. 8.2). Ochoa (1983) suggests that the Middle Usumacinta region in general, and specifically the intermediate plains between the Usumacinta and San Pedro Mártir Rivers, were a nexus of interaction, serving as

Prior investigations within the region indicate that the area was occupied from at least the Early Formative period. The few ceramic analyses that have been realized in the 34

The Northwest Maya Lowlands: Site Selection and Regional Background a link between the earlier Olmec culture on the Gulf coast and the lowland Maya world centered in the Petén (see also Ochoa and Casasola 1983). Many investigators have suggested the presence of Olmec or Mixe–Zoque groups in southeastern Tabasco and northwestern Chiapas, or at the least the movement of such groups across the regional landscape during the Early and Middle Formative periods (e.g., Lowe 1977: 220, fig. 9.3, 1989). Not coincidentally, given the role that linguistic and iconographic interaction play in script development, the region also sits near Justeson’s (1986: 438, fig. 1) demarcation of the division between the Isthmian and greater lowland Mayan script groups (Figure 3.2).

and Calakmul. These centers maintained political links with the powerful central Mexican state of Teotihuacán, and local elites consolidated power through the control of extensive networks of long–distance exchange throughout the Maya lowlands and greater southeastern Mesoamerica (Braswell 2003a; Navarrete 1978; Vargas and Ochoa 1982). The pan–regional hegemony of Tikal and Calakmul was based primarily on expansionist warfare, the maintenance of links with contemporary Mesoamerican centers, and the building of coalitions with smaller, less powerful dependent sites in order to control trade and the exchange of exotic, luxury goods and utilitarian resources. During the Formative–Classic period transition, the large centers of the central Maya lowlands continued to establish political power and control through the formation and administration of multi–site territories, especially by integrating smaller sites such as those within their spheres of influence. The increasing integration of smaller sites, in turn, led to the demarcation of more well–defined boundaries in the Early Classic period (Englehardt 2010).

Finally, southeast Tabasco lies immediately beyond the northwestern corner of the central lowlands of the Petén (Figure 3.3). In general terms, during the Formative period the central lowland Petén region of Guatemala was the stage on which Maya civilization first developed in the Middle and Late Formative periods (c. 1000 BC–AD 150). The sites of El Mirador and Nakbe in the central northern Petén are among the first Maya centers at which evidence has been found that indicates marked shifts towards increasing sociopolitical complexity, including monumental architecture and control over an extensive hinterland (Hansen 1991). Locating the place of the northwest lowlands in the emerging Maya system of political organization in the Late Formative is complicated by the fact that the lowland Petén region suffered a decline in the Terminal Formative period (c. AD 150), perhaps as the result of a prolonged drought (González Moreno 2006: 12). At this time, there was a sharp decrease in population and a cessation in the construction of large–scale ritual–religious and monumental public architecture.

Considering long–term processes of interaction in Mesoamerica since the Archaic–Formative period transition c. 2000 BC, the emerging hegemony within the lowland system during the Early Classic period must have been sustained by macro–regional interaction between a series of multi–center interregional spheres linked through pre–existing exchange networks. Since the power of first–tier centers varied over time, the control that such centers were able to exert over the smaller, subordinate sites in the interior and on the frontiers of their domains likewise varied. Certain sites, often lower– level centers within regional settlement hierarchies— such as those selected for study—were situated as nodal points on the borders of overlapping micro– and macro– regional exchange networks. The distribution patterns of material markers at such sites are particularly well–suited for elucidating diachronic variability in interregional interaction, as well as the degree to which these sites were integrated into both historical exchange networks and the emerging lowland Maya system during the Late Formative–Early Classic transition. A more detailed comprehension of how such smaller, subordinate centers were organized and integrated into larger sociopolitical spheres thus illuminates how lowland political entities interacted with one another, as well as the nature and extent of interaction between the greater Maya lowlands and other regions within Mesoamerica (Golden et al. 1998: 436; Golden et al. 2004).

The historical developments in the Petén region, such as the collapse of the Middle Formative states in the Mirador basin, likely played a significant role in the development and expansion of the sites currently under investigation, given their propitious location and potential for control over routes of riverine commerce and interchange along the San Pedro Mártir River (González Moreno 2006: 70; Román et al. 2004). The emergence of San Claudio, Tiradero, Revancha, and Mirador as population centers is framed within a series of global processes underway during the Middle and Late Formative periods. The transitional period between the Late Formative and Early Classic was characterized by the emergence of new patterns of sociopolitical organization in the Maya area, tending towards increased centralization, nucleation of settlements, and integration within larger regional systems. At the same time, population increased and increasingly marked wealth, power, and status differentiation emerged.

Previous research has established a complex understanding of the settlement hierarchy in southeastern Tabasco and the western Petén during the Middle and Late Classic period (e.g., Anaya Hernández 2002; García Moll 2005; Golden and Scherer 2006; Golden et al. 1998; Golden et al. 2004; Juárez Cossío 2003; Martin 2003; Perales Vela 2002). The smaller sites of Tiradero,

By the Early Classic period, two large, centralized states had developed in the southern lowlands, at Tikal 35

Archaeological Paleography Revancha, Mirador, and San Claudio must have been second and third order sites within the regional settlement hierarchy, subordinate to larger centers such as Piedras Negras, Santa Elena, or Moral–Reforma. Interaction between diverse lowland Maya political entities took various forms, although sociopolitical power and prestige were expanded primarily through warfare and conquest, military alliance, or interdynastic marriages between noble families and ruling groups. During the Classic period, a prolonged history of conflict and conquest allowed large military powers such as Calakmul and Tikal to unify ample territories, although no single site ever managed to achieve hegemonic power (Braswell 2003a; González Moreno 2006; Vázquez et al. 2005). Nonetheless, these large kingdoms in the central lowlands orchestrated expansionistic strategies that led to the formation of a complex mosaic during the Classic period, in most cases built on the base of foundations of previous Formative period developments.

1). Also at this time, higher–ranked neighboring sites within the regional hierarchy, such as Piedras Negras, Moral–Reforma, and Pomoná (Figure 3.1), underwent a marked expansion and reached their apogee as centers of power (Anaya Hernández 2002; García Moll 2005; Juárez Cossío 2003; Martin 2003; Muñoz 2002, 2004). Location and Environment The lowlands of southeastern Tabasco lie within the humid, tropical coastal plain of the Gulf of Mexico in southeastern Mexico and comprise part of the richest alluvial plain in the country (Hernández Ayala 1981: 16; West et al. 1985: 15). This coastal plain is formed by the deltas of several major rivers and includes the vast majority of Tabasco State, the southwestern portion of Campeche, and parts of the extreme northern lowlands of Chiapas. It is bounded on the west by the Tertiary epoch hills of southeastern Veracruz, on the south by the foothills of the Sierra Madre de Chiapas and Sierra del Lacandón that form the Chiapas highlands, on the east by the limestone platform of the Yucatán peninsula, and on the north by the Gulf of Mexico itself (González Moreno 2006: 26–7; West et al. 1985: 15). This coastal plain is a lush riverine and lacustrine environment with an abundance of rivers, lakes, streams, swamps, and estuaries. The terrain is predominantly flat, and the copious summer rains, coupled with insufficient drainage, lead to frequent flooding and standing water, which complicates agricultural production in the region. The recurrent flooding necessitates building structures on earthen platforms, a pre– Hispanic practice which continues today.

In this intricate geopolitical mosaic, the role of smaller sites such as those under investigation is difficult to determine, especially considering the lack of inscriptions and a textual record to aid in the reconstruction of specific site histories (Golden and Scherer 2006). Understanding the development of these sites depends exclusively on material correlates. These smaller sites have largely been ignored in models of Classic Maya sociopolitical organization based on epigraphy or elite ritual activity (González Moreno 2006: 15). Nevertheless, the study of smaller settlements such as these can contribute to a better understanding of the role of regional exchange in both diachronic changes in material culture and its effects on political, social, and economic development in the lowlands. In particular, smaller sites illuminate the mechanisms and dynamics of intra– and interregional interaction, especially in terms of defining the degree to which smaller sociopolitical units were integrated within the larger regional systems of the Maya world.

In actuality, the Usumacinta River and its tributaries form part of the frontier between the nations of Mexico and Guatemala, as well as internal boundaries between the Mexican states of Tabasco, Chiapas, and Campeche. In antiquity, the wide alluvial floodplains of the mid– lower Usumacinta and San Pedro Mártir River basins comprised a region that was situated along established trade routes and has been traditionally considered a physiographic division and boundary area between distinct Mesoamerican interaction spheres. This area was witness to a strong human presence throughout the Formative period, which increased substantially in the Classic period (Ochoa 1978: 21).

The sites under investigation lie within a zone that appears to have been a cultural and physiographic border between the Gulf coast plain and the greater Maya lowlands, as well as an internal division between the northwestern and central Maya lowlands, throughout Mesoamerican history. The later situation of these sites as satellites of larger Classic period centers could be a function of their location on the borders of Formative period interaction spheres and at the frontiers of the emerging regional political entities that were beginning to integrate into an expanding lowland Maya system of political organization during the Late Formative–Early Classic period transition (Englehardt 2010; Fahsen Ortega 1999). Although evidence for Middle Formative (Mamom sphere) occupation of these sites is limited, they expanded rapidly in the Late Formative and blossomed in the Early Classic period, becoming integrated in the Chicanel ceramic sphere (González Moreno 2006: 70–

Within this zone of rivers, lagoons, plains, and swamps, this study focuses on the far western fringe of the coastal plain in the basin of the lower San Pedro Mártir River system, one of the primary tributaries to the Usumacinta River. The San Pedro Mártir River originates from a source in the central Petén, northern Guatemala, near the Lago de Flores basin, just northwest of the present–day Hacienda Papactun and near the ancient site of Motul de San José, at an altitude of 175 m (Figure 3.1). It flows for approximately 280 km, running for 185 km east to west in Guatemala and crossing into Mexico at the northeast tip of the Sierra de Lacandón foothills near the village of 36

The Northwest Maya Lowlands: Site Selection and Regional Background m3/s at San Pedro (www.inegi.org.mx, accessed August 14, 2010; Hernández Ayala 1981: 21). The river itself is considered younger than the Usumacinta, due to the absence of sudden changes in the course of the riverbed, vertical cuts in the land, or beaches along its banks during the dry season. The water level in the river is relatively constant throughout the year, and during the rainy season the river floods the low–lying areas along the river banks, turning these into swamps (Hernández Ayala 1981: 21). The river water itself is odorless and colorless, with a relatively high content of calcium and sodium sulfates and bicarbonates, owing to the abundance of limestone in the region.

El Ceibo. The river follows the ridgeline northwest for 20 km before turning due north and running for 50 km, at which point it curves sharply to the west near the village of San Pedro before emptying into the Usumacinta River 25 km downstream. Numerous tributary streams and small rivers feed into the San Pedro Mártir on its course through Mexican territory. The study area occupies a roughly rectangular area of approximately 1800 square kilometers in the extreme southeast of the state of Tabasco, within the Municipalities of Balancán and Tenosique. This region is located between 17°15′ and 17°50′ north latitude and 91° and 91°15′ west longitude, bordered on the north by the low Usumacinta floodplain and the border of Tabasco and Campeche States, on the east by the Guatemalan border, on the south by the Guatemalan border and the low sierra foothills of the Chiapas highlands, and on the west by the parallel of 91°15′, approximately midway between the San Pedro Mártir and Usumacinta Rivers in the intermediate plains between these rivers. This is generally a low, flat area, with an elevation that ranges from 27m above sea level in the north of the study at Tiradero to a maximum of 90m in the south as the San Pedro Mártir River approaches the low sierra foothills near San Claudio (see Figure 1.3). The slightly undulating terrain of the intermediate plains consists of low terraces and hills of no more than 10m in height interspersed with the lower–lying bajos and aguadas (Hernández Ayala 1981: 18).

The prevailing climate in the state of Tabasco is tropical humid, with an average temperature of 26.4° C and median annual precipitation of approximately 1500mm (Hernández Ayala 1981: 18; González Moreno 2006). The rainy season spans June to November, while the rest of the year is relatively dry and subject to periods of prolonged drought. The region is covered by deciduous and semi– deciduous broadleaf tropical dry forest, with flora and fauna typical of tropical lowlands environments.1 In the extreme south of the study area within Tabasco State there is a small area of low sierra hills on the northern and western edges of the Chiapas and Petén highlands, respectively, in which evergreen and low montane forest flora and fauna are found (Hernández Ayala 1981: 16; Ochoa and Hernández 1975; Ochoa 1978). The rich luvisols and alluvial sediments support medium–scale agriculture, although the frequent flooding that occurs along the river complicates agricultural intensification. The most common cultigens in the present day are maize, beans, and chilies, along with tropical fruits such as mamey, papaya, plantain, and coconut. The land in the region is principally dedicated to pasture, and animal husbandry and livestock comprise the primary economic base of the current inhabitants of the area (González Moreno 2006). The region remains relatively isolated, with only two paved all–weather roads in the whole of the southeast of Tabasco, although riverine communication and transit routes remain as important today as they were in Prehispanic times, connecting the region with the rest of Tabasco and Guatemala.

Geologically, the area is divided into two zones: the recently formed broad alluvial floodplain of the Usumacinta River delta to the north, and, inland to the south, the intermediate plains between the Usumacinta and San Pedro Mártir Rivers, comprised of superficial deposits of eroded Pleistocene sediments washed downstream by river currents and deposited in low terraces, bajos, or aguadas (Hernández Ayala 1981: 17; González Moreno 2006: 39; West et al. 1985: 15). Eocene deposits of limestone and flint, characteristic of marine basins, are prevalent throughout the region. Limestone formations occasionally break through the surface, but in general limestone and a porous, decomposed, or pulverized limestone bedrock (sascab) lies under an approximately 20 cm strata of flint, especially in higher areas and existing hills or natural terraces. In the relatively higher zone in the south of the study area, deposits of highland colluvial materials are common (Ochoa 1978: 15). The soils of the region are generally between 50–100 cm in thickness, vary between low and moderate drainage, and are formed by a10–20 cm layer of dark brown and black alluvial material that rests atop white colluvial sascab (Hernández Ayala 1981: 17).

Previous Investigations Despite its promising location, southeastern Tabasco has been long underrepresented in Mesoamerican archaeology, with only a handful of systematic, regional archaeological investigations directed at the area. Even fewer studies have explicitly focused on the sites included in this study. In the late nineteenth century, Désiré Charnay elaborated a regional map during his tour of the Usumacinta and San Pedro rivers (Charnay 1884: 437;

The basin of the San Pedro Mártir River drains an area of approximately 25,000 square kilometers with an average annual volume of 2.819 million mm3 and an average annual discharge of 52.9 m3/s at El Ceibo and 89.3

Hernández Ayala (1981: 19–21) provides a detailed list of floral and faunal species encountered in the region.

1

37

Archaeological Paleography González Moreno 2006: 28, fig. 27). In 1898, Teobert Maler created a more comprehensive map of the middle Usumacinta basin as part of his explorations of Tabasco. Maler described several sites in detail, including Moral– Reforma, Pomoná, and Chinikihá, as well as Yaxchilán, Piedras Negras, and Altar de Sacrificios (Fernández Tejedo 1988: 28; González Moreno 2006: 29, fig. 28). These investigations were of a general nature, producing little more than sketch maps of the region, and may be considered part of the larger nineteenth century fascination with voyages of discovery in Mesoamerica.

fundamental goal was to ‘obtain a general knowledge of the area’ (Sánchez Caero 1979: 1). Over the course of several years (1974–1979), the TBN project produced a number of valuable reports and detailed the results of investigations on settlement patterns and history, trade routes, linguistic studies, the identification of production and commercial centers, and regional ceramic sequences, among others (Hernández Ayala 1981; Hernández Pons 1984; Ochoa 1977, 1978; Ochoa and Hernández 1975; Vargas and Hernández 1979; Vargas and Ochoa 1982). Shortly before and immediately after the TBN project, several investigations were conducted in areas immediately adjacent to and on the regional boundaries of the northwest Maya lowlands. Beginning in the 1950s, Robert Rands began his exhaustive study of the ceramic sequence of Palenque, in the far northwest corner of the Maya lowlands (Figures 1.2 and 3.1). This work spanned decades of research at the site and has been continued by his students (Rands 1972, 1987; Rands and Rands 1957; San Román Martín 2009). Even further to the northwest, Hernández Pons (1984) conducted an initial survey of the Tulija river valley micro–region, identifying a preliminary Classic period settlement history and rudimentary ceramic sequence. In the early 1980s, George Holley initiated his investigations of the ceramic sequence of Piedras Negras, Guatemala, on the southern border of the northwest lowlands (Holley 1983, 1987). Holley’s work on the ceramics and regional sequence at Piedras Negras has recently been revisited and expanded by Rene Muñoz (2002, 2004). In addition, Patricia Fournier (1983) expanded on the settlement history of the lower San Pedro Mártir basin undertaken by Hernández Ayala (1981) to include the middle Usumacinta region and the sites of Panhalé, El Cayo, and La Pasadita, adding needed resolution to the original regional chronology. In a final offshoot of the TBN Project, Ernesto Vargas (1985) published a report on the site of El Arenal, on the western bank of the middle Usumacinta River in central southeast Tabasco.

During the first half of the twentieth century, more scientifically–oriented studies began to be conducted. In 1924 Frans Blom was selected to direct one of the first scientific expeditions that had as its objective the study of the customs and dialects of Maya indigenous groups in the northwest lowlands (González Moreno 2006: 29–31; Ochoa 1978, 1983). Blom’s study focused primarily on the Chontal speaking area in the northeast of Tabasco and did not include information on the archaeological sites or material remains of the intermediate plains and low sierras of southeastern Tabasco. Later, in 1943, Alberto Ruz Lhullier undertook a survey of the Campeche coast down to the Río Candelaria and San Pedro y San Pablo River with the objective of collecting archaeological materials, especially ceramics (Berlin 1955; González Moreno 2006: 31; Sánchez Caero 1979: 1). These and other data facilitated the first anthropological study of the cultures of the Gulf coast of Campeche and Tabasco. Ruz Lhullier focused on three sites of historical importance: Xicalango, Tixchel, and Champotón. In each of these sites he reported the presence of ceramic materials corresponding to the Formative, Classic, and Postclassic periods, establishing the first regional chronology of the northwest Maya lowlands (González Moreno 2006: 31). In 1953, Heinrich Berlin undertook the first systematic archaeological investigation of the lower Usumacinta basin, focusing on a detailed description of the site of Atasta, as well as various other sites near the confluence of the Usumacinta and San Pedro Mártir (Berlin 1955: 129, 1956; González Moreno 2006: 30, fig. 29). Following up on this work in 1961, Lizardi investigated the settlement at Balancán, presenting a detailed description of that site’s stelae 4 and 5 (Sánchez Caero 1979: 1).

Following the successes of the TBN project and the burgeoning interest in the northwest lowlands, the Archaeological Atlas of the State of Tabasco Project conducted an extensive investigation of southeastern Tabasco in 1986, identifying approximately 1545 archaeological sites. The Centro INAH Tabasco has since reported a much greater number of sites that were not registered during the Atlas project (Fernández Tejedo 1988; González Moreno 2006: 27). As part of the Atlas project, Rebeca Perales and Jacobo Mugarte realized a study of the settlement patterns around the sites of Resaca and Santa Elena, located approximately 30 km north of San Claudio, midway between that site and Mirador (Perales Vela and Mugarte 1995). Based on analyses of settlements patterns, Atlas

By far the most extensive investigation of the northwest Maya lowland was the Proyecto Tierras Bajas Noroccidentales (TBN), conducted in the late 1970s though the Centro de Estudios Mayas (CEM) of the Universidad Nacional Autónoma de México (UNAM) under the direction of Lorenzo Ochoa. The TBN project organized a series of investigations of the region and at that point was the only study that had systematically investigated the intermediate plains of the middle Usumacinta and lower San Pedro Mártir river basins, including the majority of sites included in this study. The TBN project was a total regional survey whose 38

The Northwest Maya Lowlands: Site Selection and Regional Background project investigators considered these settlements on the San Pedro Mártir to be third and fourth tier sites in the regional settlement hierarchy, although more recent scholarship suggests that Santa Elena was at the least a second tier center and possibly a first–order site (Anaya Hernández et al. 2003; García Moll 2005; Golden 2003; Golden and Scherer 2006; Juárez Cossío 2003; Martin 2003; Perales Vela 2002). Unfortunately, the vast majority of the valuable data derived from the work of Perales Vela and Mugarte at Santa Elena remains unpublished.

Classic period, as well as their location within the study area, in the San Pedro Mártir valley approximately 35 km to the east of the Usumacinta River; a flat area of intermediate plains between the low foothills of the Sierra del Lacandón to the south, and the wider Usumacinta floodplain to the north (González Moreno 2006; Hernández Ayala 1981; Ochoa 1983; Ochoa and Casasola 1991; Rands 1987: 204). The site of Tiradero is located in the north of the study area near the floodplain, and San Claudio is situated approximately 60 km to the south, adjacent to the low sierras. Revancha is slightly to the south of Tiradero on the west bank of the San Pedro Mártir River, and Mirador lies at the head of a tributary stream to the east of the river on the intermediate plains between Revancha and San Claudio. Smaller second and third order sites were selected in order to ensure a sample that represented more than the remnants of elite interaction, which are often most visible at larger, primary centers. It is likely that elite tastes and preferences filtered through to the smaller sites under the control of large centers, but I also hoped to encounter a greater proportion of utilitarian and non–sumptuary ceramic materials at these sites, which in turn would provide a clearer image of quotidian or non–elite interaction.

In the 1990s, the Centro INAH Tabasco conducted a variety of investigations throughout southeastern Tabasco (e.g., Cuevas Reyes 1996; González Lauck 1993; Juárez Cossío 1992, 1994, 1999). The majority of these investigations were of an exploratory nature and directed simply at identifying sites for future study (José Luis Romero Rivera, personal communication, 2009). Anaya Hernández (2002), López Varela (1991), and García Moll (2005) undertook extensive investigations of the site of Pomoná and its hinterlands, including Panhalé. These studies detailed the settlement history and chronology of the area along the Usumacinta River just north of Tenosique, and to the south of Vargas’ previous investigations at El Arenal. Immediately to the west of the study area, Rodrigo Liendo of UNAM has been conducting a regional survey of the adjacent Chiapas lowlands surrounding the sites of Palenque and Chinikihá, primarily focusing on settlement patterns (Liendo Stuardo 2005, 2007). Directly to the south of the study area, the Sierra del Lacandón Regional Archaeology Project (Golden 2003; Golden and Scherer 2006) has in recent years undertaken a massive survey of sites along the middle Usumacinta River, re–investigating several sites in both Mexico and Guatemala identified by Fournier (1983), including El Cayo and La Pasadita. Finally, the Centro INAH Tabasco’s Proyecto Arqueológico San Claudio, under the direction of José Luis Romero Rivera (1997, 1998, 2000), has been investigating the site of San Claudio and its environs for the last two decades. To date these excavations, along with the TBN Project and specifically the investigation of Hernández Ayala (1981), remain the only archaeological study of the sites under consideration.

San Claudio San Claudio is an archaeological site located along the Tenosique-El Ceibo highway at kilometer 33+600 in the extreme southeast of Tabasco, within the Municipality of Tenosique. The site is located approximately one kilometer from the Agricultural Colony of El Xotal, near the Laguna de San Claudio (Figure 3.4), which sits approximately 5 km southwest of the western bank of the San Pedro Mártir River. San Claudio is cataloged in the Public Registry of Monuments and Archaeological Zones with the identification key E15D3627011. The site comprises approximately 94 structures distributed over a series of low hills and covers around 70 hectares. The regional plan elaborated by Charnay (1884) during his journey through the Usumacinta and San Pedro Mártir basins mentions the Xotal ranch (‘Chotal’ on the Charnay map), located approximately 1000m from San Claudio, but Charnay apparently did not notice the ruins of the settlement, and he does not describe any ruins or archaeological sites in the vicinity (González Moreno 2006: 27, fig. 27). San Claudio appears on the Archaeological Atlas of the state of Tabasco project map, but the site is not described in any detail. The data primarily regard geographic location. In 1996 and 1997, INAH realized brief surveys of the extreme southeast of Tabasco State along the border with Guatemala as part of a project to safeguard the archaeological heritage and patrimony of the region (Cuevas Reyes 1996). Prior to the construction of the Tenosique–El Ceibo highway, INAH investigators located approximately 22 Prehispanic settlements along the proposed route of the highway,

Site Selection and Background In order to explore processes of interregional interaction and early script formation, this investigation culls information from both published sources and archaeological collections of ceramic materials from four sites on the eastern edge of the focus area within the northwest Maya lowlands: Tiradero, Revancha, Mirador, and San Claudio (Figures 1.3 and 3.1). The sites were selected on the basis of their long occupational histories dating from the Middle Formative through to the Late 39

Archaeological Paleography

Figure 3.4. Relief map of the site of San Claudio, 1m contour. Map by Mario Retíz after González Moreno 2006: 32, fig. 30

the site and created a topographic map, identifying approximately 94 structures in six architectonic groups over an area of 70 hectares. Investigators concentrated on Structures 1, 4, and 12 in Groups II and III. Over 40 stratigraphic test pits were also dug, although these pits were ultimately not very useful in determining site chronology, due to an exceedingly low water table surrounding the Laguna de San Claudio (José Luis Romero Rivera, personal communication, 2009).

including San Claudio, indicating a dense Prehispanic population dating back to the Middle Formative period. It was through the survey of the proposed highway route that archaeologists first noted the dimensions of the settlement at San Claudio and determined that the highway project presented serious risks to the site and could potentially adversely affect future investigations of the area. INAH intervened and the highway was rerouted around San Claudio and other sites through an archaeological rescue project from 1996 through 1998 under the direction of Francisco Cuevas of the Centro INAH Tabasco (González Moreno 2006: 31; Cuevas Reyes 1996).

San Claudio extends over a roughly rectangular 2 km strip of land and is oriented on an east–west axis (González Moreno 2006: 36). The defining geographic feature of the site is the Laguna de San Claudio, which gives the site its name. The architectural groups of the site straddle the low sierra to the immediate south and are distributed along a series of small hills, as evident in the topographic map (González Moreno 2006: figs. 23 and 36). The majority of the structures at the site are rectangular or L–shaped, and are arranged around large open plazas (González Moreno 2006: figs. 31, 35, 38–9). Although the architectural groups are relatively widely dispersed within the site, with the exception of Groups II and III that are located near the lagoon, when

In 1997, José Luis Romero Rivera of the Centro INAH Tabasco developed a project to investigate the site of San Claudio. The project intended to rescue and preserve archaeological remains threatened by the encroaching highway project, to elaborate a topographic plan of the site, and to determine the settlement history of San Claudio and its place in the regional chronology (González Moreno 2006: 31–4; Romero Rivera 1997, 1998). Over the course of three field seasons in San Claudio (1997, 2000, 2001), investigators surveyed 40

The Northwest Maya Lowlands: Site Selection and Regional Background compared with the settlement pattern data derived from recent studies conducted in the Sierra del Lacandón and in the Guatemalan Petén, particularly from the site of Tecolote (Golden 2003; Golden and Scherer 2006), the place of San Claudio in the regional hierarchy was lower than larger, first tier sites such as Santa Elena, and it was likely a second or third tier site like Resaca, subordinate to Santa Elena or, more likely, Piedras Negras (Anaya Hernández et al. 2003; González Moreno 2006: 40–1; Martin 2003; Perales Vela 2002).

Tiradero The site of Tiradero (Figure 3.5) is located on the banks of the San Pedro Mártir River in the far north of the study area, sitting astride the confluence of the intermediate plains between the Usumacinta and San Pedro Mártir basins to the south and the wider Usumacinta floodplain to the north. The site lies within the Municipality of Balancán, approximately 1.5 km northwest of the small village of San Pedro (Hernández Ayala 1981: 47, Plate 49, 48, fig. 48, 69, fig. 2, 71, fig. 8). It extends over an area of approximately 300 hectares, although only a portion of its total extent has been mapped. The approximately 20 mounds that comprise the site center are distributed over a series of terraces that rise from the river banks immediately to the northwest of a peninsula that juts southward into a horseshoe bend of the San Pedro Mártir River and on which were constructed various building complexes. The site was cataloged by the TBN project with the identification key BA16.

San Claudio occupies an area officially denominated as the ‘sub–province of plains and swamps of Tabasco’ (González Moreno 2006: 35; Romero Rivera 2000). Geologically, the site sits astride a northwest–southeast strip of recent alluvial soils that are contiguous with the intermediate plains to the north between the Usumacinta and San Pedro Mártir Rivers, and abuts the low sierra foothills that form the Sierra Madre de Chiapas Mountains to the south and the low hills of the northwest Petén, Guatemala to the east (González Moreno 2006: 37, 39, fig. 36; Ochoa 1978). The area contains sedimentary rocks and volcanic sediments characteristic of the Miocene period of the Upper Tertiary, with a predominance of arenaceous stone. Flint is abundant in the region, and objects of worked flint or blocks are readily visible throughout the site in moderate frequency (González Moreno 2006: 37).

Tiradero was noticed by Berlin and included in his map of the region (Berlin 1953: 129; González Moreno 2006: 30), but he did not provide any detailed description of the site. The TBN project investigation of Hernández Ayala (1981) remains the only dedicated archaeological study of Tiradero. That said, the TBN excavations were extremely limited in scope. Their primary aims were to

Figure 3.5. Relief map of the site of Tiradero, 1m contour. Areas of excavation outlined in black cross hatched boxes. Map by Mario Retíz after Hernández Ayala 1981: 49, fig. 49.

41

Archaeological Paleography create a topographic map of the site in order to delimit potential areas of activity, to determine the maximum extent and occupational history of the site, and to excavate exploratory trenches and stratigraphic test pits. Despite these modest goals, lack of resources and time dictated by the excavation permissions granted to TBN severely limited the scope of the fieldwork that investigators were able to conduct.2 Hernández Ayala (1981: 47) describes the site as a major ceremonial center on the basis of the presence of pyramidal mounds and temple platforms, stelae and altars, large buildings with potentially vaulted architecture, and a ballcourt. In the absence of intensive excavation and detailed evidence to support this claim, it appears more likely that Tiradero was a secondary or tertiary site, subordinate to the larger center of Moral– Reforma to the west in the regional settlement hierarchy (Hernández Ayala 1981: 46).3 Its strategic location on a bend in the river would likely have facilitated a function as a commercial center for riverine trade, although the absence of defensive fortifications suggests that its role may have been relatively minor.

fill in excavations in both Complex A and the residential group indicate a settlement history that stretches back into at least the Middle Formative period, although investigators recognized the necessity to conduct further excavations to determine a more precise occupational history (Hernández Ayala 1981: 48).

The site is comprised of approximately 20 mounds of no more than 10 m in height, with the exception of a slightly taller mound at the far west of the site center directly to the east of the aguada. The majority of the mounds correspond to residential groups distributed over the two square kilometers of the peninsula, and are oriented along an east–west axis. Somewhat incongruously, the entire site occupies a low–lying area that is subject to flooding during the rainy season, with buildings constructed primarily on natural bluffs of 20 x 40m in area and three to four meters in height (Hernández Ayala 1981: 47–8; Ochoa and Hernández 1975). Investigators concentrated on the ceremonial center of the site, principally on Complex A, in which excavators encountered the greatest concentration of structures. On the western edge of a group of approximately 10 irregularly distributed mounds, four pyramidal mounds that form a small plaza stand out (Hernández Ayala 1981: 47). The southwestern mound in this group appeared to be an L–shaped building that investigators named the palace. Directly to the north excavators found two parallel platforms that were identified as a ballcourt.

Mirador

To the south of Complex A, investigators also explored a residential group, in which a house footprint was identified, as well as a possible domestic midden that contained several large burnished ollas (Hernández Ayala 1981: 50, fig. 51). Ceramics recovered through surface collection throughout the site and found as construction

The primary aims of the TBN investigation were to determine the extent of inhabited area at the site, to create a map of the structures in the site center, to excavate test trenches and stratigraphic pits in both the residential area and site center, to collect surface materials between Structures II, X, XI, and XII (Hernández Ayala 1981: 53). Mirador is a compact settlement; structures are sparse at the site, and those that do exists are of relatively small dimensions. The residential groups surround the site center to the east, south, and southwest, with groups and structures much more dispersed to the south (Hernández Ayala 1981: 47, 53; Ochoa and Hernández 1975). Like

Like all sites included in this study, geologically, Tiradero lies within the sub–province of Tabascan plains and swamps. The site sits on the southern border of the broader floodplain formed by the Usumacinta delta and the lower reaches of the Río Candelaria in southern Campeche to the north. The peninsula on which the site sits is located atop a large natural porous decomposed limestone (sascab) platform oriented north to south which lies between 50cm and 1.2m below sea level. The frequent flooding in and around the site has built up rich deposits of alluvial sediment and luvisols along the banks of the river upstream and downstream from Tiradero. Limestone and flint rocks, washed downstream, are intermittently found at the site.

Mirador is located at the head of the Naranjito stream, a small tributary to the San Pedro Mártir River, approximately 25 km upstream from the village of San Pedro, 3 km inland to the east of the river itself, and 5 km southwest of the village of Apatzingán within the Ejido Apatzingán in the Municipality of Balancán, Tabasco. The site sits squarely within the intermediate plains between the Usumacinta and San Pedro Mártir basins, approximately midway between Tiradero and San Claudio. As a whole, Mirador is significantly smaller than Tiradero, covering 125 hectares, although the site center (Figure 3.6) comprises 17 identified structures that cover approximately four hectares and are oriented on a north–south axis (Hernández Ayala 1981: 53, 70–1, fig. 8). Mirador went unnoticed by prior archaeological reconnaissance and the TBN project is to date the only archaeological study of Mirador. The site lies at the southern limit of the TBN project’s study area, and was cataloged with the TBN identification key BA17.

The low quality relief map produced by Hernández Ayala (1981: 49, fig. 49) from which Figure 3.5 is adapted is the only available map of the site. 3 Hernández Ayala (1981: 47) indicates that Tiradero was connected to the larger center of (Moral–) Reforma via a communicating ‘road,’ although she does not describe the road, and no roads or raised causeways (sacbe’o’ob) appear on her map of the site. 2

42

The Northwest Maya Lowlands: Site Selection and Regional Background Formative period (Hernández Ayala 1981: 46–47, 53, 55, fig. 57). Investigators also discovered an Olmec– style stela on the east side of Structure V in the site center, which may indicate an even earlier occupation of the site (Hernández Ayala 1981: 53; Ochoa and Hernández 1975; Ochoa 1983: 165, fig. II.4.4). Geologically, Mirador sits atop a platform of porous decomposed limestone (sascab) which lies between 1.5 and 2m below sea level. Several stratigraphic test pits revealed that certain sections of the limestone floor had been opened by the site’s occupants to form irregular pits. The functions of these pits are unclear, but they were filled with sandy soil and were found to contain human and animal remains, shell, ceramics, and lithics (Hernández Ayala 1981: 53). The surface of the site and its surroundings is covered in a relatively thick layer of rich, black alluvial soil of good quality. Underneath this topsoil were found light brown and black clay soils, sitting atop the sascab platform. The area around the site consists of thick tropical forest. Revancha The site of Revancha lies on the western bank of the San Pedro Mártir River approximately 10 km upstream from Tiradero, and 8 km downstream from Mirador on intermediate plains between the Usumacinta and San Pedro Mártir basins. It is located on the land of a cattle ranch known as Rancho la Revancha within the Municipality of Balancán, Tabasco. Revancha is a small site oriented on a roughly northwest–southeast axis, whose center and residential areas combined comprise 20 low mounds of less than three meters in height and between five and twelve meters in length and seven and ten meters in width that cover just less than 100 hectares (Hernández Ayala 1981: 57). Like many other sites investigated through the TBN project, there had been no previous archaeological study of the site, and the TBN excavations directed by Ramón and Paula Kroster remain the only investigation of Revancha to date (Hernández Ayala 1981: 57; Ochoa and Hernández 1975). The TBN project cataloged the site of Revancha with the identification key BA33.

Figure 3.6. Map of the site center of Mirador. Map by Mario Retíz after Hernández Ayala 1981: 54, fig. 55.

Tiradero, Hernández Ayala (1981: 47) classifies the site as a major ceremonial center, due to the presence of pyramidal structures, temple platforms, altars, and a ballcourt (Structure VIII in Figure 3.6). Also like Tiradero it is likely that the site was a secondary center subordinate to either Moral–Reforma to the northwest or Santa Elena to the south, or possibly even a tertiary site under the control of Revancha or Resaca (González Moreno 2006; Perales Vela 2002). Its location on a tributary stream of the San Pedro Mártir River suggests that it could have been an outpost or way center on the riverine trading route.

Investigators initially sought to create a topographic map of the site, to determine the extent of its occupation, and to repair damage caused by previous looting of the site, which had opened the cores of various structures. Attention was directed on the site center of Revancha, which is comprised of five groups of structures of variable dimensions (Hernández Ayala 1981: 57). Three of these groups (Figure 3.7) are located atop a large, artificially leveled terrace constructed of rubble and limestone rocks, separated by an aguada. To the west of the aguada stands a complex of 13 structures whose dimensions and distribution suggested an administrative function distinct from the other

Excavations and surface collection rendered a great deal of lithic materials, including an extremely well– worked obsidian knife, as well as numerous faunal remains and even an infant cranium from House 2 in the eastern residential group. In addition to these materials, investigators recovered a variety of ceramics comparable to those found by the TBN project at other sites throughout the region, whose analysis indicated an occupational history dating to at least the Middle 43

Archaeological Paleography

Figure 3.7. Relief map of the site of Revancha, 0.5m contour. Map by Mario Retíz after Hernández Ayala 1981: 58, fig. 60.

the control of either Moral–Reforma of Santa Elena. Ceramic materials recovered from excavations of the three mounds in the site center, including the Palace, and in the residential area indicate an occupational history that began in at least the later part of the Middle Formative period.

architectural groups at the site. Investigators named this complex the Palace. To the east and northeast of the Palace are two groups of three and seven mounds that also stand atop the artificial terrace. The final mound groups lie to the west and southwest of the site center or natural bluffs that rise along the western bank of the San Pedro Mártir River. The southwest group, consisting of seven mounds, appears to have been the location of several lithic workshops (Hernández Ayala 1981: 57). Also to the west of the site center is the residential area of the site.

Revancha is situated along a series of low bluffs that rise naturally along the western bank of the San Pedro Mártir River that slope downstream. The eastern edges of the site nearest the river were subject to severe erosion caused by periodic flooding. This erosion prevented the accumulation of alluvial sediment along the banks and exposed deposits of flint on the surface, rendering poor quality soils in the site center (Hernández Ayala 1981: 57). The lower–lying residential areas to the west of the site center were subject to more extreme flooding, yielding richer alluvial soils more suitable for agriculture. Test pits in the largest mound of the residential area revealed a relatively thick (20cm) layer of alluvial and vegetal topsoil atop black clay soils produced by sedimentation (Hernández Ayala 1981: 57). The site as a whole sits atop the porous decomposed limestone bedrock (sascab) ubiquitous in the region.

Hernández Ayala (1981: 47) classified Revancha as a minor ceremonial center on the basis of the platforms and structures built on the artificially leveled terrace, as well as the presence of numerous lithic workshops to the west of the site center near the residential area. Its location atop artificial terraces on the banks of the river suggests a possible function, like Mirador, as an outpost or way center on the riverine trading route. Due to its relatively small size and location between what appear to be larger secondary centers (Tiradero and Mirador or Resaca), it is likely that Revancha was a third–tier center subordinate to secondary centers under 44

The Northwest Maya Lowlands: Site Selection and Regional Background

The northwest Maya lowlands have long been recognized as a nexus of interaction in Mesoamerica. The Usumacinta and San Pedro Mártir Rivers served as vital routes of communication and exchange between diverse sub–regions throughout the Early and Middle Formative periods. Later, these rivers served as communicative arteries between the core Maya area of the central Petén and the frontier of the northwest lowlands during the Late Formative and Early Classic periods as Maya civilization developed and expanded (Navarrete 1978; Ochoa 1983; Vargas and Ochoa 1982). The region selected for study thus is located at a crucial boundary area that played a vital role in interregional exchange throughout Mesoamerican history.

early Mesoamerican scripts have been found, including San Andrés in the coastal plain of northwest Tabasco immediately to the northwest of the study area, and San Bartolo, in the lowlands of the Petén, approximately 100 km to the east (e.g., Pohl et al. 2002; Saturno et al. 2006). Given that the Mayan script developed from a Middle Formative period (‘Olmec’) iconographic base, the presence of Olmec–style iconography at an early date in the northwest lowlands (Ochoa 1978, 1983; Ochoa and Hernández 1975) indicate that southeastern Tabasco was far from a marginal area in the consideration of the development of the Mayan script. Rather, such advanced iconographic systems are intimately connected to the development of writing, especially in areas in which long–term interaction occurred between discrete cultural groups, such as the northwest lowlands.

Its situation at an intersection of interregional interaction makes this area a particularly appropriate test case in which to evaluate broader patterns of large–scale interaction and exchange in Mesoamerica. The presence of Olmec and Olmec–style artifacts in the intermediate plains of the San Pedro Mártir and wider Usumacinta basin attest to the long history of occupation and cultural exchange in and across the area. Although little evidence of early writing has been encountered within the study area, the region in which the selected sites are located abuts areas in which

In sum, due to the location of the study area at a key historical crossroads of Mesoamerican trade and interchange, and the fact that a critical mass of material evidence from prior regional investigations was accessible for study, the northwest Maya lowlands of southeastern Tabasco are an excellent test case in which to examine the material expression of larger, long–term patterns of interaction and the relationship between interregional interaction and the broad processes involved in the development of the Mayan script.

Conclusions

45

Chapter 4 Ceramic Sample and Analytic Methods and Sharer 1982; Holley 1983, 1987; Neff et al. 1999; Pring 1977; Rice 1996a, 1996b; Skibo et al. 1989), and previous research has recovered ceramic materials in great quantities in southeastern Tabasco (Ochoa 1983; Ochoa and Casasola 1991). Finally, identifying formal changes in ceramic materials is a relatively straightforward process. In turn, functional variation may be deduced from such stylistic variability, or through changes in the depositional contexts of ceramic materials within and among sites, micro–regions, and regions.

This chapter introduces the ceramic sample employed in this investigation. It provides an overview of the regional and site–specific sequences of the study area from which the global ceramic dataset was compiled, as well as brief contextualizing comparison of the regional sequence with both a general pan–lowland Maya ceramic tradition and the ceramic sequences of neighboring sites and regions. The chapter details the archaeological contexts of the ceramic materials under investigation and the selection of an analytic sample from the global dataset. Further, it explains the methodological procedures that underpin the regional ceramic sequence and by which the ceramic materials were sorted, typed, and classified. The general classes of variables appraised, the definition of discrete analytic units, and the specific variables measured within the sample are then discussed. The chapter concludes with a discussion of the qualitative and quantitative methods—including statistical analysis of variance (ANOVA) of mean attribute variability and the H Score measure of intra–sample diversity—employed in this study to evaluate the nature and extent of the relationship between interaction and material variability as evident in the ceramic materials under investigation.

The sample itself includes data from both published sources and collections of archaeological materials from the lower San Pedro Mártir River basin, specifically from the sites presented and discussed in the preceding chapter. The bulk of ceramic data examined in this investigation was culled from the ceramic archives of the Proyecto Tierras Bajas Noroccidentales, housed at the Instituto de Investigaciones Antropológicas (IIA) of the Universidad Nacional Autónoma de Mexico (UNAM). Dr Lorenzo Ochoa kindly granted access to the ceramic collections from the sites of Tiradero, Mirador, and Revancha, as well as the regional ceramic data from other sites throughout the San Pedro Mártir River basin collected through the TBN project. The ceramic materials from San Claudio are housed at the Instituto Nacional de Antropología e Historia (INAH) Centro Tabasco in Villahermosa, Tabasco. José Luis Romero Rivera, Subdirector of Archaeology at the Centro INAH Tabasco and the principal investigator of the site of San Claudio, allowed me to examine a large quantity of the ceramic materials recovered through three seasons of fieldwork at the site.

If the development of the Mayan script is related to systemic material exchange, stylistic and functional variability in the distribution of material artifacts should indicate degrees of interaction, integration, innovation, and differential use of material culture, paralleling the emergent transformations of diffused icons into written signs. Thus, comparing the results of qualitative and statistical analyses of synchronic and diachronic variability in the ceramic sample against diagnostic linguistic transformations and innovative variations in shared iconography involved in script development may illustrate the relationship between material variability, interaction, and script development. In the context of this research, the analysis of stylistic and functional variability in material markers of interaction permits a more robust and nuanced evaluation of the developmental dynamics and the symbolic categories that underlie both material exchange and writing.

Through investigations of these collections conducted during 2009 and 2010, I compiled a dataset of approximately 22,000 ceramic artifacts from the region dating from the Middle Formative through the Early Classic periods (c.700 BC–AD 450). The sample includes all group and type–variety classifications encountered at any site in quantities greater than ten. Artifacts in the sample were recovered from systematic surface collection, securely identified archaeological contexts (i.e., burials, domestic middens, ritual deposits, construction fill, etc.), or stratified test pits (González Moreno 2006: 87–95; Hernández Ayala 1981: 82–4). Within the sample, there are approximately 1000 sherds from four distinct ceramic groups and two wares that date to the Middle Formative period (Figure 4.1). Approximately 6000 sherds date to the Late Formative period in the San Pedro Mártir basin that represent 15 type–varieties from eight ceramic groups and three wares (Figure 4.2). In the Early Classic period greater

Ceramic Sample In order to explore potential correlations between script development and changes in material culture, this research focuses on one specific material variable: ceramics. Ceramics were chosen as the primary variable due to their extensive presence in the archaeological record of the Northwest lowlands. Ceramics are also an excellent base by which to measure interaction (Demarest 46

Middle Formative

Period

Ceramic Sample and Analyltic Methods

Ware

Group Joventud

Paso Caballos Waxy

Pital Tierra Mojada

Uaxactún Unslipped

Achiotes

Type: variety

Sample Size

Joventud Red: VUa

820

Pital Cream: VU Pital Cream: pital Tierra Mojada: VU Achiotes Unslipped: VU Achiotes Unslipped: achiotes

25 43 20 34 25

In this and following tables, ‘VU’ denotes undetermined variety.

a

Total 967

Period

Figure 4.1. Middle Formative period ceramic type–varieties present in sample (n ≥ 10), showing quantities and group and ware associations (González Moreno 2006; Hernandez Ayala 1981).

Ware

Group

Sierra Paso Caballos Waxy Late Formative

Flor Pital Polvero

Sample Size

Altamira Fluted: altamira

68

Correlo Incised Dichrome: correlo

13

Laguna Verde Incised: laguna

118

Repasto Black on Red: repasto

32

Sierra Red: sierra

2910

Flor Cream: flor

984

Mateo Red on Cream: flor

17

Pital Cream: pital

135

Polvero Black: polvero

360

Centenario Fluted: centenario

143

Setok Fluted: VU

178

Chunchinta Black: chunchinta

77

Desprecio Incised: VU

96

Achiotes

Achiotes Unslipped: achiotes

459

Sapote

Sapote striated: sapote

Setok Flores Waxy Monochrome Black Uaxactún Unslipped

Type: variety

254 Total 5844

Figure 4.2. Late Formative period ceramic type–varieties present in sample (n ≥ 10), showing quantities and group and ware associations (González Moreno 2006; Hernandez Ayala 1981).

variation becomes evident. Within the sample, roughly 15,500 sherds date to the Early Classic period, which are broken down into 25 type–varieties of 13 groups and five distinct wares (Figure 4.3).

makes inferences regarding the social processes that resulted in the deposition of a specific artifact at a particular location on the landscape (Schiffer 1987). Thus, multiple processes of context formation are viewed as the result of cultural deposit, that is, as a phenomenon that stems from the actions of social agents across space and through time. This approach facilitates a more nuanced understanding of how elements of material culture pass from a systemic context of use or consumption to an archaeological context of deposition. Once discarded, artifacts lose a purely utilitarian functionality, insofar as they deposited at the place in which they were utilized and secondary when an object was deposited in a place distinct from its context of primary use (Schiffer 1987).

Archaeological Contexts of the Ceramic Sample The formation of an archaeological context is due to a totality of social processes, whose diverse temporal and spatial aspects form activity areas in which certain classes of artifact may be encountered. In addition, processes of formation and transformation of archaeological contexts may be interpreted through the precepts of middle–range theories (see, e. g., Binford 1965), which consequently allows investigators to 47

Period

Archaeological Paleography

Ware Uaxactún Unslipped

Paso Caballos Waxy

Group

Type: variety

Sample Size

Sapote

Sapote Striated: sapote

437

Triunfo

Triunfo Striated: triunfo

6041

Sierra

Laguna Verde Incised: laguna

874

Sierra Red: sierra

696

Playa Dull

San Martín

San Martín Variegated Brown: san martín

1035

Holmul Orange

Ixcanrio

Ixcanrio Orange Polychrome: VU

103

Actuncan Orange Polychrome: VU

397

Aguila Orange: aguila

422

Pita Incised: VU

60

Dos Arroyos Orange Polychrome: dos arroyos

746

Balanza Black: balanza

119

Lucha Incised: bolocantal

71

Lucha Incised: lucha

1479

Bolonchac Orange Polychrome: bolonchac

451

Santa Rosa Cream Polychrome: santa rosa

210

Suktan Cream Polychrome: suktan

48

Mataculebra Cream Polychrome: mataculebra

182

Moro Orange Polychrome: moro

346

Anaité Red: anaité

284

Cameron Incised: cameron

388

Tinaja Red: aduana

602

Carmelita Incised: maculis

52

Infierno Black: bolocantal

328

Saxché Orange Polychrome: saxché

125

Zacatel Cream Polychrome: zacatel

35

Aguila

Dos Arroyos

Early Classic

Balanza

Santa Rosa Petén Gloss

Saxché

Tinaja

Infierno

Palmar

Total 15,531 Figure 4.3. Early Classic period ceramic type–varieties present in sample (n ≥ 10), showing quantities and group and ware associations (González Moreno 2006; Hernández Ayala 1981).

At San Claudio, exploration was undertaken on Buildings 1 and 4 in Group II, and Building 12 in Group III (Figures 4.4–4.8). An exploratory trench transected Building 12 in order to define the structure’s form and function. 43 stratigraphic test pits were also sunk (González Moreno 2006: 34, 84–5, 92–5, appendix 1 for details on the stratigraphic test pits; see also Romero Rivera 1998: 6–27, 2000: 2–10.). Over 5000 bags of

ceramic materials were recovered over the course of three field seasons. Ceramics dating to various periods were found in secondary contexts within the rubble fill of the structures and platforms in these architectural groups. Chronological inferences on the basis of architectural contexts are difficult in this case, due primarily to the fact that ceramic materials dating to various periods were encountered in the fill of all four successive construction 48

Ceramic Sample and Analyltic Methods

Figure 4.4. Sketch map of San Claudio Group II. Illustration by Mario Retíz after González Moreno 2006: 36, fig. 35.

Figure 4.5. Plan of San Claudio Structure 1. Illustration by Mario Retíz after González Moreno 2006: 34, fig. 32.

Figure 4.6. Plan of San Claudio Structure 4. Illustration by Mario Retíz after González Moreno 2006: 35, fig. 33.

49

Archaeological Paleography

Figure 4.7. Sketch map of San Claudio Group III. Illustration by Mario Retíz after González Moreno 2006: 39, fig. 38.

phases (González Moreno 2006: 41–4). Ceramic materials were also recovered in secondary contexts from only four of the 43 test pits sunk at various sectors within the site (González Moreno 2006: 93–4). Unfortunately, very few ceramics were encountered in the test pits, and in any case, the pits themselves were shallow and thus of little practical use in defining a precise site–specific ceramic sequence or chronology on the basis of stratigraphy (González Moreno 2006: 42, 93). The only ceramic materials recovered in primary contexts were from the various burials encountered in Structure 1 (Figure 4.5) and in various rooms within Structure 4 (Figure 4.6), all of which corresponded typologically to the Late Classic period (González Moreno 2006: 44). A great quantity of Late Classic period ceramics was also recovered through systematic surface collection. The vast majority of ceramics dating to the Formative and Early Classic periods was found in the rubble fill of the structures excavated in Group II. In many cases, the ceramic materials were encountered in a very poor state of conservation, which made typological identification difficult. The fact that so few ceramics were encountered in primary contexts limits their utility in determining activity areas and deducing functional contexts.

Figure 4.8. Plan of San Claudio Structure 12. Illustration by Mario Retíz after González Moreno 2006: 35, fig. 34.

50

Ceramic Sample and Analyltic Methods

Figure 4.9. Map detailing excavated areas at House 1, Tiradero. Illustration by Mario Retíz after Hernández Ayala 1981: 50, fig. 51.

Likewise, the relatively poorly defined archaeological contexts in which ceramic materials were encountered precluded relative stratigraphic dating or the definition of an occupational sequence for the site (González Moreno 2006: 42). The chronological sequence and possible functional contexts of the Formative and Early Classic period ceramic materials from San Claudio can, however, be established through other means (Culbert and Rands 2007; San Román Martín 2009).

then undertaken on two low mounds to the southwest of the site center, which turned out to be a habitational area (Figure 4.9). The area was parceled into 1m squares, and two trenches were opened, roughly oriented north–south and east–west. The trenches were excavated to a depth of 1.2 m, revealing three layers before hitting the porous, decomposed limestone sascab bedrock (Hernández Ayala 1981: 48). Once investigators had cleaned the house floor, they recovered a variety of ceramic materials, both from primary contexts directly from the floor and secondary contexts in the construction fill from the west wall of the house structure. The materials encountered on the house floor included several large and particularly well–preserved Early Classic period ollas (earthen cooking pots).

The ceramic materials from the other sites under investigation were recovered from systematic surface collection and excavation of 16 stratigraphic test pits, four house lots, two ballcourts, and 23 excavation units at 29 different sites, yielding a total of over 40,000 sherds, of which the vast majority1 were recovered from Tiradero, Mirador, and Revancha (Hernández Ayala 1981: 74, 78, 80–5, fig. 71). At Tiradero, investigators first undertook a systematic surface collection, which yielded 15,394 sherds, 95% of which dated to the Late Classic period (Hernández Ayala 1981: 82, fig. 71). Explorations were

Investigators sunk four 1m2 stratigraphic test pits in the site center. One test pit revealed a clear stratigraphic sequence consisting of a vegetal layer followed by a layer of small stones (at 30 cm depth) that covered a small cache of three intact ceramic vessels (Hernández Ayala 1981: 52, fig. 53, 82, fig. 71; Ochoa and Hernández 1975). Below this offering investigators encountered a dark clay soil with ashes and a large number (4152) of fragmented ceramic materials that were determined to be of Late Formative or Early Classic period origin

A total of 38,870 sherds were recovered from the three TBN sites under investigation, 91.84% of the total of 42,214 ceramic materials recovered by TBN investigations at all 29 sites in the lower San Pedro Mártir basin (Hernández Ayala 1981: 81).

1

51

Archaeological Paleography investigators encountered approximately 3300 ceramic fragments in secondary contexts dating from the Early Formative through the Late Classic periods (Hernández Ayala 1981: 53, fig. 71). Investigators then turned to the exploration of the supposed residential area to the southwest of the site center. Two trenches were opened that revealed three house units (Figure 4.11). The stratigraphy was particularly clear in this area of the site, which facilitated the establishment of a chronology and occupational history of the site, and permitted relative dating of the materials encountered in the residential group (Hernández Ayala 1981: 53). 4300 ceramic materials were recovered from two of the houses, in both primary contexts on the floors and directly in front of the hearth in house two and in the patio area of house 1, and in secondary contexts in the rubble fill of the house walls (Hernández Ayala 1981: 82–3, fig. 71). Hernández Ayala (1981: 53) reports that these materials were of particular diagnostic value not only in determining the ceramic sequence of the site but also in terms of intra–regional comparison. The stratigraphy of the residential units revealed that certain sections of the sascab bedrock opened into large, irregularly shaped holes, in which human remains, shell, lithics, and a great variety of ceramic materials were encountered, possibly indicating a primary archaeological context of burial or ritual deposit. Investigators also directed excavations

(Hernández Ayala 1981: 48). Immediately below these materials a rock formation was found, which investigators suggested could have been a small altar. The ceramic materials encountered in this test pit therefore appear to have constituted a cached offering of some sort, and thus may be considered to have been found in primary contexts. Finally, investigators turned to the Tiradero ballcourt (Figure 4.10), located on the west fringes of the site center. Three trenches were opened, two on the internal corners of the two structures that comprised the ballcourt, and one on the western exterior of the west structure. A large quantity of Formative and Early Classic period ceramic materials were recovered from secondary contexts in the rubble fill of these structures (Hernández Ayala 1981: 48–9). Absolute dates for these materials could not be determined at the time of excavation, but the ceramics were later dated stylistically and typologically and assigned to the Middle or Late Formative, or Early Classic period on that basis. At Mirador, after completing a systematic surface collection which yielded primarily Late Classic period ceramic materials, investigators sank eight 2m2 stratigraphic test pits and exploratory trenches at various points throughout the site center. Four test pits and a single exploratory trench were sunk to a depth of 2m on the east side of Structure V, in which

Figure 4.10. Detail of excavated areas at the Tiradero ballcourt. Illustration by Mario Retíz after Hernández Ayala 1981: 52, fig. 54.

52

Ceramic Sample and Analyltic Methods

Figure 4.11. Floor plans of the three houses at Mirador in which explorations were undertaken and ceramic materials recovered. Illustration by Mario Retíz after Hernández Ayala 1981: 55, fig. 57.

at Structure VIII, the ballcourt complex (Figure 4.12), in which two units were opened. The trenches were shallow—investigators hit a floor at 20cm—but a small amount of ceramics were found in secondary contexts in the fill associated with the ballcourt structures. Finally, at the site of Revancha investigators discovered a damaged site that had suffered a great deal of looting (Hernández Ayala 1981: 57). Consequently, the nature of the archaeological contexts of materials encountered at this site could not be definitively determined. Systematic surface collection recovered a modest amount of ceramic materials (586 fragments; Hernández Ayala 1981: 84–5, fig. 71). Three excavation units were opened: along the northeast corner walls of the palace, in the residential area to the west of the site center, and along the south mound of complex four (the northeast group; see Figure 3.7). The stratigraphy of the operations was clear in all cases, revealing four natural layers, which facilitated relative stratigraphic dating of the materials encountered. The majority of ceramic materials—approximately 1800 out of a total of 2900—were recovered from the lowest level of the palace trench (Hernández Ayala 1981: 84–5, fig. 71). Most of these were typologically similar to ceramic materials encountered at other sites that were dated to the Formative and Early Classic periods. Significantly fewer

Figure 4.12. Detail of excavations at the Mirador ballcourt. Illustration by Mario Retíz after Hernández Ayala 1981: 56, fig. 59.

53

Archaeological Paleography materials were encountered in the other two excavation units in the northeast group and the residential area (Hernández Ayala 1981: 57, 84–5).

products of human culture and permits an appreciation of the relationship between the individual and society (Gifford 1960).

Sorting and Typing

The type–variety system has proven a flexible heuristic tool and conceptual framework that facilitates the creation of comparable spatial and temporal analytic units and consequently has assisted in intra– and intersite interpretations of variable yet interrelated formal attributes in ceramic materials (Forsyth 1989: 229–41). By establishing a unified classificatory terminology, the type–variety system allows for the presentation of clear, objective ordering and description of ceramic assemblages, avoiding arbitrary or subjective interpretations. The systematic application of the type– variety concept therefore permits the establishment of analytical ceramic units that are comparable across the whole of the lowland Maya region (Smith et al. 1960; Willey et al. 1967). Investigators may thus realize areal chronological studies and use ceramic materials as an interpretive device for deducing specific cultural histories, technological phases, and developmental trajectories, which is especially useful for those minor sites that are spatially removed from ceremonial centers (Sabloff and Smith 1972)—sites such as those within the study area.

At San Claudio, the 5000 bags of ceramic materials and sherds encountered through surface collection and excavation were washed and marked. The initial classification of ceramic materials was undertaken during the third field season (2002) by crew member Roberto Aguirre Cadena. The ceramics were later re–examined and written up by Angela González Moreno as part of her MA thesis investigation (González Moreno 2006). The materials are currently in curation at the facilities of the Centro INAH Tabasco in Villahermosa and at the bodega of the site of San Claudio. The ceramics from Tiradero, Mirador, and Revancha formed the majority of the ceramic materials recovered during three field seasons of the TBN Project from 1975– 1979 (Hernández Ayala 1981: 74; Ochoa 1977, 1978; Ochoa and Hernández 1975). The ceramic materials were washed and marked in the field, and preliminary sketches and line drawings of ceramics were produced. Once the materials had been returned to the Centro de Estudios Mayas at UNAM, classification of the entire sample was undertaken by Martha Ivon Hernández Ayala, who wrote up the results of the ceramic classification for her MA thesis investigation (Hernández Ayala 1981). The materials remain in curation at the archives of the TBN Project, currently housed at UNAM’s Instituto de Investigaciones Antropológicas.

The classification of a ceramic collection in the type– variety system is based primarily on surface treatment, decoration, and form. This last criterion is not always discernable, due to the fragmented nature of many ceramic materials (i.e., sherds) recovered from archaeological contexts. The type is the fundamental unit of analysis in this case. A type is ceramic unit that is generally differentiated according to recognizably distinct visible characteristics or tangible attributes (Gifford 1960). Types represent a complex of distinct attributes that is indicative of a particular ceramic category produced during a particular temporal interval or region. Ideally, types are designated with a name that references a geographic point or area in which the ceramics were distributed, although naming is often arbitrary. This first term is complemented by a second—generally a descriptive attribute, such as color or surface finish—so that a binomial designation is generated (e.g., Joventud red, Pital cream, Balanza black, Sapote striated, etc.; Smith and Gifford 1966). The taxonomic designation of the type thus describes the ceramic entity and represents a complex whole of distinct characteristics and often implies specific cultural affiliations, geographic distribution, or temporal connotations (Gifford 1960). A variety is a subdivision of a ceramic type that is based on small yet significant differences within type classes and interrelated attributes, such as the presence or absence of minor techno–stylistic characteristics or restricted geographic or temporal distribution. The first identified variety is designated with the first type name (e.g., Pital cream: pital; Robles Castellanos 1990: 26),

Investigators originally applied the type–variety method of classification to the ceramic materials recovered from all four sites. The type–variety system is a method of ceramic typology in which vessels are grouped into successively finer hierarchical categories on the basis of some combination of variables, including paste, decorative treatment, or surface finish. The type–variety method traditionally has been preferred by a number of scholars of the ceramics of the Maya region (e.g., Adams 1971; Culbert and Rands 2007; Forsyth 1989; Holley 1983; Sabloff 1975; Sabloff and Smith 1969, 1972; Smith 1955; Smith and Gifford 1966; Smith et al. 1960; Willey 1970; Willey et al. 1967, among various others). The type–variety system developed within the theoretical framework of historical particularism and was originally conceived as a Linnaean taxonomic scheme (Fournier 1987; Shepard 1965). In practice the type–variety method moves beyond a simple classificatory scheme and carries particular sociocultural implications. The type–variety system approaches archaeological objects and artifacts in terms of their mutual interrelation. Thus, as a methodological approximation the system may be considered a cultural classification insofar as it is concerned with the analysis of synthetic, non–biological 54

Ceramic Sample and Analyltic Methods and subsequent varieties are usually arbitrary or refer to identifiable stylistic attributes (e.g., Tinaja red: aduana, Lucha incised: lucha vs. Lucha incised: bolocantal).

method, her designations are irregular, substituting her own terminology and classificatory categories in the place of the established conventions of the type–variety system (e.g., Hernández Ayala 1981: 86–98). As such, after having examined, digitally photographed, and taken measurements of the collection, a re–classification of the TBN ceramic materials was undertaken in order to create a typology that more closely conforms to the classifications of ceramic materials from other lowland Maya sites, and more closely follows the general rules of the type–variety method. These results are reflected in Figures 4.1, 4.2, and 4.3.

A ceramic group is comprised of a combination of several similar or closely related types that display homogeneity in form, paste color, and technological and stylistic attributes independent of decorative variability. That said, the principal attributes employed in determining ceramic groups are ranges of variation in form, color, technology, surface finish and paste (Rye 1981; Smith et al. 1960). In those cases in which ceramic materials are very fragmented or eroded, limited group–level identification may be realized without defining specific type–varieties, which often assists those investigations in which detailed ceramic analyses are unnecessary. A ceramic group is essentially a ‘super–type’ (Smith and Gifford 1966). Group names are designated in the same manner as types, often adopting the type name when a specific type is used to first refer to the ceramic group to which it corresponds (e.g., Joventud, Sierra, Sapote, Infierno, etc.). The ceramic types that comprise a particular group invariably belong to the same ceramic ware.

Chronology and Phasing A complex is the typological and analytic unit at which one may begin to address the chronology and spatial contexts of ceramic assemblages.2 A ceramic complex is the sum total of all types, varieties, and groups that as a whole constitute an identifiable temporal sequence that spans a specific chronological interval at a particular archaeological unit or site. Complexes may be separated into phases, temporal subdivisions that denote changes within a complex, for example, the appearance of new types, varieties, or forms (Robles Castellanos 1990; Smith and Gifford 1966; Willey et al. 1967). The content of a ceramic complex therefore represents the whole of the ceramic materials utilized by an archaeological culture in a determined geographic area during a specific temporal interval.

Wares are classes of ceramic materials characterized by similar technology, material, and surface treatment and defined by attributes of paste composition and characteristics of surface finish. Pastes are classified according to texture, temper, color, and porosity. Classifications of surface finish consider whether the finish is course or polished, if it is lustrous or matte, the presence or absence of a slip, or burnishing, and so on. A ware is therefore a collection of ceramic types which displays uniformity in the various technological attributes of paste composition and surface finish (Robles Castellanos 1990; Sabloff and Smith 1969). Wares are designated with a first name that is usually geographic, suggesting the origin of the ware or the area in which it was first encountered or identified, and a second term that refers to an identifiable attribute of the paste or surface finish (e.g., Flores Waxy, Petén Gloss, Uaxactún Unslipped, etc.). Although wares possess direct spatial and temporal implications, further analytic units are necessary to more fully address issues of chronology in ceramic assemblages.

Temporally, complexes and phases are often conceived as separated (or subdivided) by horizons, which reflect spatial trajectories of ceramic complexes bounded in a short term temporal context. Although characteristically short in duration, horizontal elements are often dispersed beyond their sources over a wide geographic space. Horizons are identified by the presence of a ‘horizon marker’ in two or more ceramic complexes. A horizon marker is an easily identifiable ceramic type that is approximately contemporary within several ceramic complexes in distinct spatial contexts or at different sites. Once recognized, ceramic complexes of various sites can be related to each other in time, thus allowing for the definition of a ceramic horizon. Horizon markers are thus a powerful tool that permits the association of ceramic complexes between sites, although the presence of a horizon marker at disparate sites does not imply that there existed an identical trajectory of techno–cultural development at the sites that share a horizon marker, nor that the distinct ceramic complexes themselves are

In the case of the ceramic materials from San Claudio, this study adopts the typological classifications of González Moreno (2006: 105–62). The ceramics from Tiradero, Mirador, and Revancha presented considerably more difficulty, as the typology developed by Hernández Ayala appears to have been created arbitrarily and without direct reference to ceramic materials, type–varieties, groups, or wares present in the collections of other lowland Maya sites (Lorenzo Ochoa, personal communication, 2009). Although Hernández Ayala did classify the ceramics according to the general precepts of the type–variety

‘Assemblage’ is a term not often adequately defined in the literature. Here, assemblage refers to the conglomeration of ceramic materials encountered or recovered at a specific archaeological unit (e.g., site, operation, excavation lot, etc.). Although the term assemblage is not directly related to type–variety classification, assemblages are particularly relevant as heuristic devices for the interpretation of the physical space or locale in which ceramic materials were encountered and recovered.

2

55

Archaeological Paleography similar or related (Neff et al. 1999; Robles Castellanos 1990: 28–9; San Román Martín 2009).

defining site–specific occupational histories, dating derived through ceramic analysis is particularly helpful in the case of smaller sites that lack detailed architectural and epigraphic evidence or materials appropriate for the application of absolute dating techniques. Such is the case for the sites of the lower San Pedro Mártir River basin currently under investigation, at which limited exploration and excavation has taken place.

Beyond the horizon, when two or more ceramic complexes share a high percentage of their most common ceramic types, implying a high degree of material, technological, and typological similarity, they form a ceramic sphere (Robles Castellanos 1990: 29). The content of a sphere is the sum total of all types and groups represented in its constituent complexes and allows for the interrelation of multiple complexes in which the majority of types is common and the distribution of specific stylistic attributes is relatively homogeneous (Willey et al. 1967). The diagnostic elements of a ceramic sphere are all those techno–stylistic characteristics that are shared across complexes through space and over time. The high degree of similarity between ceramic materials within a sphere implies extensive cultural contact or interaction at a technological level (Willey et al. 1967).

In the absence of extremely secure archaeological contexts or absolute dates for the materials within the sample, the chronology of the ceramic sequence presented below is necessarily based on relative dating by seriation, through the comparative relation of the types, groups, complexes, and spheres in the sample with the more established sequences of larger, more extensively investigated lowland Maya sites at which a greater amount of evidence was available and the determination of more absolute dates was possible (Adams 1971; Bryant et al. 2005; Eppich 2004; Forsyth 1989; Fournier 1987; García Moll 2005; Hernández Pons 1984; Holley 1983, 1987; Lee 1972; López Varela 1991; Matheny 1970; Muñoz 2002, 2004; Rands 1972, 1987; Rands and Rands 1957; Román et al. 2004; Sabloff 1970; Sánchez Caero 1979; Smith 1955; Smith and Gifford 1966; Willey 1970).

The final and perhaps most relevant category in establishing ceramic chronologies is the sequence, which both unifies and transcends all other typological, spatial, and chronological considerations. A ceramic sphere is comprised of a multitude of ceramic types, groups, and complexes that display similarity in decoration, techno– stylistic elements, surface finish, and any number of attributes that may be demonstrated to have developed simultaneously or ‘evolved’ over time and through space (Orton et al. 1993; Rice 1987; Rye 1981; Shepard 1965; Sinopoli 1991). These diagnostic elements of a ceramic sequence form a complex whole and function as indicators of developmental continuity over considerable spans of time. As a result, sequences may include typological elements from any number of distinct ceramic types, groups, wares, complexes, or spheres in distinct spatial and temporal contexts. While it may be considered similar or related to a ware, a sequence supersedes this category by encompassing characteristics from multiple, possibly dissimilar categories, and includes the crucial element of demonstrable developmental relationships between its various typological components. Hence, in many ways the sequence is the most useful category for comparing multifaceted ceramic assemblages between (or within) cultures, regions, and temporal periods.

Ceramic Sequence of the Lower San Pedro Mártir Basin After completing the reclassification of ceramic materials, the sample was compared with the established ceramic sequences of neighboring lowland regions and Maya sites3 in order to establish a sequence specific to the whole of the lower San Pedro Mártir basin. In general, this study follows the complex, phase, horizon, and sphere terminology proposed for the lower San Pedro Mártir basin by Hernández Ayala (1981: 77) and incorporates the ceramic materials recovered from San Claudio. The creation of a ceramic sequence was complicated by several factors. First, the convoluted and overly specific typological classification developed by Hernández Ayala for the sites of Tiradero, Mirador, and Revancha made direct comparisons more difficult. Additionally, although the sample itself is large, the materials themselves come from limited excavations and systematic investigation at restricted areas within the individual sites and thus do not necessarily represent the entire range of typological variability that may have been present during the distinct occupational phases at each site.

In essence, these categories are theoretical constructs that essentially represent periods of ceramic production, which are themselves basically equivalent to the periods of use of that particular ceramic complex, horizon, or sphere. Ceramics are thus useful in establishing dates for architectural sequences (and therefore occupational histories), especially in the Maya lowlands, since in pre–Columbian traditions the use of sherds and ceramic materials as construction fill was common and widespread (González Moreno 2006: 48). Although when possible it is preferable to combine architectural, epigraphic, artistic, and available absolute dating techniques when

Thus, the sequence employed in this investigation is based almost exclusively on the comparison of the sample with ceramic materials recovered from adjacent regions and sites. Nevertheless, contextual archaeological (i.e., stratigraphic) data were also utilized wherever possible. In particular, the sequence of Uaxactún (Smith 1955; Smith and Gifford 1966), which is the oldest established lowland ceramic sequence and is generally held to be the standard against which other assemblages are compared.

3

56

Ceramic Sample and Analyltic Methods

? ?

500 600

Bari early

400

Middle Formative

300

(late) Ayn

(late)

Isthmo

Plancha

Horcones

(early)

1

Pom

Chicanel

Abal

San Felix

Mamom

(early)

Xot

Cascada

Motiepa Picota

(late)

Francesa

Palenque

Middle Grijalva

Chacibcan

Otolum

2

(early) Salinas

Kundapi

3

Naba

Juspano

Jiquipilas

Piedras Negras

Altar de Sacrificios

Uaxactún

Balche

Veremos

Guanacaste

200

Tepeu 1

Ipsan

Classic Horizon)

Chixoy

Guañoma

late

100

Late Formative

100

BC

Laguna

By way of explanation, and to lay groundwork for the comparative analysis of ceramic materials discussed below, a generalized lowland ceramic sequence, as established by others, is detailed in order to illustrate the general baseline against which I compared the sample to create the sequence. The Early Formative Xe and Real ceramic complexes that developed in the Pasión River region of western Guatemala are the earliest southern lowland traditions (Adams 1971; Forsyth 1989; Fournier 1987; González Moreno 2006: 48;

Hol (early facet Mamom)

Felisa

Pichi

200

AD

Kaxabyuc (Early

300

Taxinchan

based on the comparative analysis of lowland Maya ceramic collections and contextualized within the established ceramic chronologies of neighboring sites and regions, is presented in Figure 4.13.

Tzakol

400

Early Classic

500

Caoba   (early)

600

Chiapa de Corzo

Lower Usumacinta

Lower San Pedro Mártir

Period

Date

Diachronic typological similarities and differences between the sample and ceramic materials known from elsewhere in the Maya lowlands therefore provides the most reliable means for dating the ceramic sample. This approach is a generally accepted methodology, and is especially useful for determining the chronology of types and groups that span multiple chronological periods and ceramic spheres (e.g., achiotes, sierra, sapote, etc.). The approximation this comparative method yields creates a relevant framework for the discussion and comparison of the chronological contexts and development of ceramic traditions in the lower San Pedro Mártir basin. These developments were framed within general regional and macro–regional processes of interaction. The sequence,

(Late Waxy Horizon)

Misolha (Early Waxy Horizon) (Pre– Waxy Horizon)

Figure 4.13. Regional ceramic sequences and correlations for the Maya lowlands, with relative and absolute chronological correlation. (Adams 1971: 136, table 23; Hernández Pons 1984: fig. 5; Hernández Ayala 1981: 77; Holley 1987; Lee 1972; Muñoz 2004; Rands 1972, 1987; Smith 1955; Smith and Gifford 1966).

57

Archaeological Paleography Willey 1970). The most common forms in these complexes are short–necked jars and tecomates (common spheric vessels), some with white, red, and black slips (Adams 1971; Fournier 1987; Sabloff 1970). In the Middle Formative period, Mamom horizon ceramics (c. 700–200 BC) were widely shared across lowland Mesoamerica, although slight regional stylistic differences developed, possibly as ethnic, linguistic, or cultural groups began to separate. Mamom sphere ceramics are generally unslipped monochrome, with red being the most common color, followed by black and cream. Fluting is the most common decorative technique in Mamom ceramics. Forms are mostly simple, with short– necked jars, tecomates, hemispheric plates, and shallow basins predominant. Vessels generally lack supports, with vertical walls and very slightly everted or beveled rims (Fournier 1987; González Moreno 2006: 49).

Guatemala during the Late Formative period (Demarest and Sharer 1982; Neff et al. 1999). Moving into the Classic Period, the lowland Maya area experienced intensive development and expanding sociopolitical complexity. At this time, greater specialization in processes of ceramic production becomes increasingly apparent, and the ceramic traditions of the lowland Maya area begin to show marked differentiation when compared to other areas of Mesoamerica (Bryant et al. 2005; Lee 1972). Localized and micro–regional styles also begin to appear within the Maya lowlands at this time. Early Classic period ceramics of the Tzakol sphere (c. AD 300–600) display increasing technological and stylistic complexity (e.g., Smith 1955; Smith and Gifford 1966). Tzakol ceramics are characterized by glossy, lustrous slips (e.g., Petén Gloss ware), primarily orange and black in color. Polychrome red or black on cream and black on orange are common, decorated with stylized zoomorphic and anthropomorphic elements and, in some cases, glyphs. Striation and fluting remain common decorative techniques, but increasingly elaborate decorations emerge that imply alteration of vessel surfaces, such as precision incising, relief carving, and painting (Adams 1971; López Varela 1991; Sabloff 1970; Smith and Gifford 1966). Characteristic Early Classic period forms include rounded plates and cajetes with Z–angles, basal flanges, and tripod mammiform supports, as well as cylindrical jars and glasses (Fournier 1987; González Moreno 2006: 52).

Late Formative period ceramics generally remained standardized and uniform across the lowlands, and indeed, throughout Mesoamerica. The distribution of ceramic materials suggests an increasing degree of economic integration on local, regional, and macro– regional scales, although the design, production, and consumption of ceramic materials, and the development of new styles, forms, and techniques was unimpeded by developing sociopolitical hierarchies or increasingly demarcated sociocultural boundaries (Englehardt 2010; Fournier 1987; Golden and Scherer 2006; González Moreno 2006: 51–2). The Late Formative Chicanel sphere (c. 200 BC– AD 300) ceramics of the central Maya lowlands are characteristically red, black, or cream monochrome, with some bichromes, such as red on orange or black on red, evident. Common decorative techniques included striation, fluting, and negative resist. Short–necked jars remain a characteristic form, although new forms begin to be introduced, such as cajetes (flat earthen serving dishes) with hemispheric walls, plates with divergent straight walls and everted rims (often pierced and reinforced), and vessels with medial or labial flanges and solid, hollow supports, often mammiform (e.g., Smith 1955; Smith and Gifford 1966).

Although this study avoids a detailed consideration of Late or Terminal Classic period ceramics, it is necessary to detail the continuation of the general lowland sequence briefly, if only in terms of differentiating Early from Late Classic period ceramics. The Late Classic Tepeu 1 and 2 spheres (c. AD 600–800) represent the culmination of the Maya polychrome ceramic tradition. Maya ceramics exhibit maximum complexity and elaboration in design, technological attributes, and stylistic decorative elements. Increasingly distinctive and divergent localized traditions emerge (vis à vis Tzakol horizon divergence). Late Classic polychrome ceramics are decorated with abstract geometric designs and painted with figures, narrative scenes, glyphic bands and texts, and characteristic Classic Maya iconographic motifs (e.g., Kan cross, sky bands, etc.). The decorative template is generally orange in color, although cream bases are also common in some ceramic types and groups. Red and black monochrome traditions continue, although multiple additional decorative techniques are applied to vessel surfaces. Forms also become increasingly elaborate, with planar based cylindrical vases, tripod plates with notched basal flanges, and high–necked drums and jars (Fournier 1987; González Moreno 2006: 53–4; Smith 1955; Smith and Gifford 1966). In the Terminal Classic period Tepeu 3 (c. AD 800–900) divergent regional traditions predominate

At the same time, dynamic, localized traditions begin to develop in the Late Formative period. For example, the Floral Park ceramic sphere, identified in Belize and prevalent in the far eastern Maya lowlands, is dominated by orange polychrome vessels decorated with stylized geometric motifs and unslipped striated vessels. Unique forms, such as cajetes with Z–angles, basal flanges, and tetrapod mammiform supports and open basins with fluted walls and rims are evident in Floral Park assemblages (Adams 1971; Fournier 1987; González Moreno 2006: 52; Smith 1955; Smith and Gifford 1966). In addition, the Usulutan ceramic tradition, with its distinctive cream paste polychrome pottery and singular forms, emerges along the extreme southeastern boundary of the Maya lowlands in western Honduras and El Salvador and the southern highlands and Pacific coast of 58

Ceramic Sample and Analyltic Methods and the popularity of Classic polychrome ceramics wanes with the emergence of orange and gray fine paste ceramics. These highly localized styles display a diversity of designs, forms, technological attributes, and decorative techniques (Fournier 1987; González Moreno 2006: 55).

Within the sample six type–varieties from four ceramic groups and two wares are represented (Figure 4.1). Slips tend to be thick and waxy. Decorated red monochromes dominate the assemblage with primarily undecorated cream and black monochrome vessels making up a smaller percentage. Joventud Red is by far the predominant type encountered at all sites, comprising nearly 85% of the total sample (Figure 4.14). Joventud Red is ubiquitous throughout Middle Formative period lowland Mesoamerica, and is the primary ceramic group represented in the collections of the Hol phase at Piedras Negras, the Monos complex at El Mirador, Petén, and the San Felix phase at Altar de Sacrificios (Adams 1971: 120; Forsyth 1989; Muñoz 2002, 2004). All type components in the Middle Formative sample are evident in the ceramic sequences of neighboring sites and regions.

Certain utilitarian forms are common throughout both this generalized sequence and the specific San Pedro Mártir sequence. These include ollas, cazuelas (pots or bowls with a nonrestricted opening, no neck, and no handles), shallow basins, flat plates, recurved cajetes, and apaxtles (shallow flat–bottomed vessels with flaring sides). Such common vessels are often technologically simple (e.g., unslipped), and display little decoration or elaboration of features (e.g., rims, supports, etc.). Below, a summary of the San Pedro Mártir sequence from the Middle Formative through the Early Classic period is presented, detailing the chronology and ceramic spheres to which the materials within the sample have been assigned.

In general, incising and fluting are the most common decorative modes. Resist decoration (e.g., Tierra Mojada) is also present, but infrequent. Parallel rows of incised lines on the interior of plate rims are the most common decorative motif. Fluting on the exteriors of bowls is also common. There is a noticeable lack of variety in vessel forms. This fact may be related to the relatively small and fragmented sample, which makes it difficult to describe the complete range of forms represented. The most common forms appear to be thick–walled bowls or plates with divergent to out–curving walls, slipped jars with short, nearly vertical necks, and cajetes and apaxtles with thickened direct, or thickened everted rims. Ollas appear more frequently toward the end of the Bari phase.

Middle Formative Period Although an extremely small quantity of ceramic materials that date to the Early Formative period has been found in the study area, the early Bari phase ceramic complex represents the first major occupation of sites along the lower San Pedro Mártir River. The total sample of diagnostic early Bari complex ceramic material is small and fragmented, containing less than 1000 diagnostic sherds. Almost all of the Middle Formative period ceramics were encountered in construction fill mixed with later Formative materials at all sites, although the trenches excavated in the residential area at Mirador revealed a variety of Middle Formative ceramic materials in clear stratigraphic contexts. Based on comparisons with the ceramic sequences of other sites, this phase is estimated to have lasted from about 600 BC to approximately 200 BC.

In almost all respects, the Middle Formative ceramics in the sample resemble contemporary Mamom sphere materials from elsewhere in lowland Mesoamerica, particularly the Pasión River and Petén regions of Guatemala, the Grijalva River basin, and the Gulf and Pacific coasts (Adams 1971; Bryant et al. 2005; Cheetham 2007; Forsyth 1989; Lee 1972; Muñoz 2004;

Type: variety

Tiradero

Mirador

Revancha

San Claudio

Total

Joventud Red: VU

245 81.40%

423 86.50%

119 82.63%

33 100%

820 84.79%

Pital Cream: pital/VU

20 6.64%

37 7.57%

11 7.64%

0 -

68 7.03%

Achiotes Unslipped: achiotes/ VU

26 8.64%

19 3.89%

14 9.72%

0 -

59 6.10%

Tierra Mojada: VU

10 3.32%

10 2.04%

0 -

0 -

20 2.07%

Total

301 100%

489 100%

144 100%

33 100%

967 100%

Figure 4.14. Breakdown of quantities and percentages of five most common type–varieties present in sample at each site in the Middle Preclassic period. Percentages indicate proportions of selected type–varieties and totals in relation to the respective Middle Preclassic period ceramic assemblage as a whole.

59

Archaeological Paleography Ochoa and Casasola 1991: 9–10; Smith 1955; Smith and Gifford 1966). Compared with complexes outside of the greater Maya lowland area, although dissimilar in many ways, early Bari ceramics seem close to Cerro de las Mesas Lower I and part of the La Venta sequence (Adams 1971: 123; Cheetham 2007; Lee 1972). This loose association, coupled with the presence of Olmec– style objects in the region (e.g., the Olmec–style stela at Mirador; Ochoa 1983: 165, fig. II.4.4, 169, fig. II.4.8) suggests that several contemporaneous ceramic traditions existed in the middle Usumacinta and lower San Pedro Mártir region in the Middle Formative period (Ochoa 1983; Ochoa and Casasola 1991: 10).

from eight ceramic groups and three wares are evident (Figure 4.2). In most cases, Late Formative materials were found mixed with earlier, early Bari phase, ceramics. Nevertheless, a large deposit of unmixed Late Formative ceramics was found stratified above Middle Formative ceramics in one of the test pits in the site center of Tiradero and in the excavations undertaken in the residential area at Mirador. These deposits proved vital to the identification and definition of the late Bari phase. Due to the overlap in the majority of materials, the beginning of the late Bari phase is placed slightly earlier than the Mamom–Chicanel transition, contemporaneous with the Hol–Abal phase transition at neighboring Piedras Negras, the early–late Plancha phase transition at Altar de Sacrificios, and the early–late waxy horizon in the Misolha complex at Palenque (Figure 4.13; Adams 1971; Holley 1983, 1987; Muñoz 2002, 2004; Rands 1972, 1987). Late Bari phase ceramics are basically a continuation of the early Bari tradition (e.g., Achiotes types), and correspond more closely to the Mamom sphere. The following Pichi phase (c. AD 1–300) signals the discontinuation of Mamom traditions and the full incorporation into the subsequent Chicanel sphere (e.g., Sierra group ceramics, Polvero Negro; Smith 1955; Smith and Gifford 1966).

Early Bari phase ceramics are likely a regional variant of the Mamom horizon, in which Joventud Red is also the dominant type. The materials in the sample are certainly affiliated with the Mamom ceramic sphere, given the presence of diagnostic Mamom types that likely originated in the central Maya lowlands of the Petén or Pasión River region (e.g., Sierra, Sapote, Achiotes) and Mamom–style decorative techniques such as horizontal grooved–incised lines on vessel exteriors (Adams 1971: 120). Nonetheless, the presence of Olmec–style materials in the region, particularly towards the early Middle Formative (c. 600 BC), suggests the coexistence of groups, ideas, or ceramic technologies that originated in the Gulf coast Olmec heartland or Grijalva depression and were incorporated into the autochthonous population of the lower San Pedro Mártir region (Ochoa and Casasola 1991: 10). In the Late Formative period, the emergence of more diagnostic late Mamom and Chicanel types suggests that the region as a whole had begun to orient itself toward the central Maya lowlands as a center of ceramic innovation, adopting and continuing the development of a distinct and separate lowland Maya tradition that had arisen with the Early Formative and early Middle Formative Xe ceramic tradition (Adams 1971: 123; Ochoa and Casasola 1991: 10; Willey 1970).4 Nevertheless, references to Olmec material culture continued in the Usumacinta basin well into the Late Formative period.

The late Bari and Pichi ceramic complexes are dominated by monochrome reds, blacks, and creams. As at other nearby sites such as Piedras Negras and El Perú–Waka’, the Late Formative period late Bari and Pichi ceramic complexes are dominated by Sierra Red and other types of the Sierra group (e.g., Altamira Fluted and Laguna Verde Incised; see Figure 4.2; Eppich 2004; Forsyth 1989; Muñoz 2004; Román et al. 2004). Flor Cream, Achiotes Unslipped, Sapote Striated, and Polvero Negro are also heavily represented (Figure 4.15). The five most common Late Formative type–varieties are uniform at each of the sites and occur in roughly the same proportions. The Late Formative period assemblages are dominated by waxy wares, specifically Sierra Red ceramics, at roughly 50% of the total Late Formative sample. The Sierra Red ubiquitous in the late Bari and early Pichi phases of the lower San Pedro Mártir basin is virtually identical to that found in the Chicanel sphere throughout the Late Formative period Maya lowlands. Specific occurrences include during the Abal phase at Piedras Negras, the Plancha phase at Altar de Sacrificios, the Chicanel phase at Uaxactún, the Cantutse phase at Seibal, and even as far afield as the Guanacaste and Horcones phases at Chiapa de Corzo and the Guañoma phase of the Middle Grijalva region (Adams 1971: 21; Lee 1972: 11; Muñoz 2004; Smith 1955). Sierra Red is also the dominant type at both Altar and Uaxactún in the Plancha and early Chicanel phases, respectively (Adams 1971: 123).

Late Formative Period The late Bari phase (c. 200 BC–AD 1) is defined by the emergence of types more clearly and directly associated with the later Mamom and Chicanel ceramic spheres, particularly those of the Sierra group. Within the Late Formative period ceramic materials, 15 type–varieties This ‘re–orientation’ is attested to by the rapid appearance of diagnostic Mamon types throughout the middle Usumacinta and lower San Pedro Mártir region in and around the Middle–Late Formative period transition (i.e., early–late Bari phase; see below), at sites such as Tierra Blanca, Pocvicuc, Santa Elena, Chablé, Trinidad, El Arenal, and Pomoca, among others (Ochoa and Casasola 1991: 10; Perales Vela 2002; Perales Vela and Mugarte 1995; Vargas 1985; Lorenzo Ochoa, personal communication, 2009; Charles Golden, personal communication, 2010).

4

Of the 15 major Late Formative type components in the sample, 12 have equivalents at Piedras Negras (out of a 60

Ceramic Sample and Analyltic Methods

Type: variety

Tiradero

Mirador

Revancha

San Claudio

Total

Sierra Red: sierra

1085 45.51%

640 45.42%

325 49.46%

860 61.47%

2910 49.75%

Flor Cream: flor

485 20.34%

176 12.49%

82 12.48%

241 17.22%

984 16.82%

Achiotes Unslipped: achiotes

235 9.85%

83 5.89%

25 3.81%

116 8.29%

459 7.84%

Polvero Black: polvero

139 5.83%

75 5.32%

69 10.50%

77 5.50%

360 6.15%

Sapote Striated: sapote

127 5.32%

71 5.03%

13 1.98%

43 3.07%

254 4.34%

2071 86.87%

1045 74.16%

514 78.23%

1337 95.56%

4967 84.92%

Total

Figure 4.15. Breakdown of quantities and percentages of five most common type–varieties present in sample at each site in the Late Preclassic period. Percentages indicate proportions of selected type–varieties and totals in relation to the respective Late Preclassic period ceramic assemblage as a whole.

total of 32 type components present at both sites), and 14 are evident at Altar (out of 35 type–varieties present at both sites). The sampled sites thus share approximately 37% of their Late Formative ceramic types with both Piedras Negras and Altar. The predominance of monochrome red, black, and cream ceramics with thick, waxy slips in the sample corresponds closely to evidence from Abal phase Piedras Negras and the late waxy horizon at Palenque and throughout the lower and middle Usumacinta basin (Muñoz 2004; Ochoa and Casasola 1991: 10; Rands 1987). As elsewhere, broad– line incising and fluting are the most common decorative modes evident within the sample during Late Formative times (Lee 1972: 11; Muñoz 2004; Rands 1987: 210). In terms of shape classes, the most common forms in the sample are wide, shallow serving dishes, such as plates, apaxtles, and cajetes, with thickened, slightly everted rims, globular ollas, straight–bodied glasses, and wide–mouthed, thick–walled jars with short, out– curving necks. These forms are common during the Late Formative period in the lower Usumacinta basin and in the Petén lowlands, and vessels with similar forms and surface finish are also evident in the Guanacaste and Horcones phases at Chiapa de Corzo and the Middle Grijalva Guañoma phase (Lee 1972: 12; Muñoz 2004; Rands 1987: 212).

and widespread diffusion of the preceding Mamom horizon style, whose influence is extensively noted in Middle Formative period ceramic phases throughout southeastern Mesoamerica (Willey et al. 1967). Most wares and types in the subsequent Late Formative Chicanel sphere, including those in the sample, are essentially modified continuations of widely shared earlier ceramic traditions. Their formal similarity is thus somewhat unsurprising. The medial or labial flanges typical of Late Formative ceramics found elsewhere, however, are rare in the lower San Pedro Mártir basin. Early Classic Period The Caoba ceramic complex (c. AD 300–600) corresponds chronologically with the Early Classic period. As with the Middle–Late Formative period transition, it is difficult to definitively separate the Late Formative materials from those of the Early Classic period, due to the fact that many materials were not excavated from sealed primary contexts. Thus, Chicanel sphere types such as Sierra Red and Polvero Negro could have coexisted with Early Classic Tzakol sphere types such as Aguila Orange or Dos Arroyos Orange Polychrome (Tables 4.2 and 4.3). In general, however, the Early Classic Caoba types in the sample appear to be a continuation of, and correspond heavily to the preceding Pichi complex and wider Chicanel sphere. Vessel forms and decorative techniques assisted in separating the Late Formative and Early Classic materials, insofar as certain forms and decoration (e.g., basal flanged bowls, tripod mammiform supports, precision incising) are particularly diagnostic of Tzakol sphere ceramics. In many cases within the lower San Pedro Mártir basin, Caoba ceramics are found immediately above the decomposed limestone sascab bedrock and mark the initial construction of a number of structures that have been dated to the Classic period (González Moreno 2006; Hernández Ayala 1981).

In most respects, the late Bari and Pichi phase ceramics in the sample are formally and stylistically close to Late Formative period materials in adjacent regions, manifesting differences primarily of emphasis as opposed to striking departures. Although some varietal variation exists between the sample and neighboring sequences, the types are essentially the same, and ranges of formal variation and decorative technique are virtually indistinguishable. The internal consistency and standardization of Late Formative period lowland Maya pottery complexes likely stems from the development 61

Archaeological Paleography The Pichi–early Caoba horizon is thus roughly contemporaneous with the Late Formative–Early Classic period transition c. AD 250.5 The initial date of this phase is based on comparisons with other sites, and corresponds to the Salinas–early Ayn transition at Altar de Sacrificios, the Early Classic Kaxabyuc horizon in the lower Usumacinta, the Chicanel–Tzakol horizon at Uaxactún, and the Picota–Motiepa transition at Palenque (Adams 1971; Hernández Pons 1984; Ochoa and Hernández 1975; Rands 1972, 1987; Smith 1955; Smith and Gifford 1966). The great typological similarity between the Early Classic ceramics in the sample and Early Classic materials known elsewhere in the Maya lowlands provides the principle means for dating the Caoba ceramic phase. The appearance of Petén Gloss Wares and distinctive vessel modes (e.g. basal flanges, hollow conical supports, and bolstered rims) indicate that Caoba is roughly equivalent in time to Tzakol assemblages known from other sites.

five distinct wares (Figure 4.3). There is more variation in terms of an increased number of wares, groups, and type–varieties present within the sample (relative to the Middle and Late Formative period materials). Additionally, the five most common Early Classic type components are no longer uniform. Instead, type– varieties of Petén Gloss ware stand out in the assemblage at San Claudio, whereas unslipped and waxy wares continue to dominate at the other sites (Figure 4.16). The gloss ware, polychrome, and orange–slipped pottery characteristic of Early Classic period lowland Maya ceramics of the Floral Park and Tzakol spheres, while present in significant quantities at San Claudio (almost 55% of the Early Classic sample at the site), are lacking at Mirador, Revancha, and Tiradero, just 50 km to the north. Although gloss ware and polychrome types represent only 15.8% of the total sample from the three northern sites, it is interesting that there appears to be a clinal distribution, with gloss and polychrome ceramics present at 25.6% of the sample at Mirador, 15.7% at Revancha, and only eight percent at Tiradero.

Thus, Early Classic period Caoba complex ceramics appear to fall into the Tzakol sphere, insofar as the complexes contain types common throughout the contemporaneous southern and central Maya lowlands, such as Triunfo Striated, Lucha Incised, Balanza Black, and Aguila Orange (Figure 4.3). Decorated materials are relatively poorly represented, although cream and orange polychrome types are evident. These polychrome decorated ceramics are more evident at San Claudio than at the other three sites. The presence of polychrome ceramics also helps situate the Caoba complex within the Tzakol sphere, since polychrome decoration is characteristic of Tzakol, but not Chicanel, ceramics. The types that came to define the Caoba complex (and Tzakol sphere) appear after c. AD 250 at sites elsewhere in the southern central Maya lowlands. Their appearance in the lower San Pedro Mártir basin is placed later, at about AD 300 in order to accommodate for Piedras Negras’ peripheral location. The end of the Caoba complex is established by the appearance of fine paste ceramics and Tepeu 1 and Otulum modes and types at about AD 600 and the beginning of the Altar Orange tradition in the northwest Maya lowlands (Muñoz 2004; Ochoa and Casasola 1991: 10; Rands 1972, 1987; Sánchez Caero 1979).

Comparing the sampled ceramics with adjacent sequences, a similar pattern of divergence is evident. The major feature differentiating the Caoba assemblage from Early Classic materials elsewhere is the absence of ring base, basal flange bowls.6 Of the 25 type–varieties in the Early Classic sample, seven are evident at Piedras Negras, and just nine find correspondence with materials at Altar de Sacrificios. In contrast to the Late Formative period, during the Early Classic the sampled sites shared an average of only 10.5% of their ceramics with these neighboring centers. Four of these seven shared type– varieties are found only at San Claudio. The orange monochromes and polychromes (e.g., Bolonchac Orange Polychrome, Moro Orange Polychrome, Aguila Orange) present in great quantities at San Claudio also dominate the Pom and Naba phase assemblages at Piedras Negras, and monochrome blacks (e.g., Infierno Black, Balanza Black) are also evident at both San Claudio and Piedras Negras (Muñoz 2002, 2004). The Usulutan– style decoration and mammiform supports diagnostic of the Late Formative–Early Classic period transition in the central Petén and found in Isthmo and Jiquipilas phase ceramics at Chiapa de Corzo, although evident at Piedras Negras, Altar de Sacrificios, and Uaxactún, are lacking within the sample. Moreover, the medial– and basal–flanged bowls characteristic of the Tzakol ceramic sphere are absent within the sample, as they are at Piedras Negras and Altar de Sacrificios until well into the latter half of the Early Classic, in the Naba and late Ayn and Veremos phases, respectively (Adams 1971: 127; Muñoz 2004).

In the Early Classic period greater variation within the sample is evident. The Early Classic period materials are comprised of 25 type–varieties of 13 groups and As with other sites in the region (e.g., El Ocotlán, Paraíso, San Joaquín, San Marcos, La Carmelita, Nueva Esperanza, and Jonuta), it was not possible to establish a phase or complex exclusively affiliated with the Terminal Formative or Protoclassic, due primarily to the scarcity and lack of diagnostic value of the available materials and the absence of modes and forms diagnostic of the Terminal Formative elsewhere in the lowlands (e.g., Usulutan style decoration on hooked–grooved rim plates, the use of mammiform supports, and the initial appearance of polychrome decoration; Muñoz 2004; Ochoa and Casasola 1991: 10; Pring 1977; Sánchez Caero 1979).

5

Such modes and forms are also absent at Piedras Negras (Muñoz 2002, 2004).

6

62

Ceramic Sample and Analyltic Methods

Tiradero, Early Classic ceramics Sample

b

Type: variety

Sample

Triunfo Striated: triunfo

4392 48.67%

Triunfo Striated: triunfo

1060 33.91%

San Martín Variegated Brown: san martín

910 10.09%

Lucha Incised: lucha

497 15.90%

Lucha Incised: lucha

869 9.63%

Dos Arroyos Orange Polychrome: dos arroyos

347 11.10 %

Laguna Verde Incised: laguna

455 5.04%

Laguna Verde Incised: laguna

145 4.63%

Dos Arroyos Orange Polychrome: dos arroyos

267 2.96%

San Martín Variegated Brown: san martín

56 1.79%

a

Type: variety

Mirador, Early Classic ceramics

Total

6893 76.39%

Total

Revancha, Early Classic ceramics

2105 67.34%

San Claudio, Early Classic ceramics

Type: variety

Sample

d

Triunfo Striated: triunfo

519 34.58%

Bolonchac Orange Polychrome: black

451 23.98%

Laguna Verde Incised: laguna

265 17.65%

Moro Orange Polychrome: moro

346 18.39%

Dos Arroyos Orange Polychrome: dos arroyos

119 7.92%

Aguila Orange: aguila

233 12.39%

Lucha Incised: lucha

102 6.80%

Infierno Black: infierno

224 11.91%

San Martín Variegated Brown: san martín

69 4.60%

Anaité Red: anaité

217 11.54%

c

Total

1074 71.55%

Type: variety

Sample

Total

1471 78.20%

Figure 4.16. Breakdown of quantities and percentages of five most common type–varieties present in sample at each site (a–d) in the Early Classic period. Percentages indicate proportions of selected type–varieties and totals in relation to the respective Early Classic period ceramic assemblage as a whole.

(e.g., Pucte Brown) are common at Piedras Negras, but not at San Claudio.

On the other hand, the majority of ceramics from Mirador, Revancha, and Tiradero in many ways appear only loosely affiliated with (and in some cases typologically unrelated to) the Tzakol sphere, where Uaxactún Unslipped and Paso Caballos Waxy wares continue to be heavily represented well after their disappearance at San Claudio and, in general, throughout the Maya lowlands. Instead, the ceramic materials from these sites seem more closely aligned with the Picota phase at Palenque or the site of Pomoná, where gloss ware is likewise absent and brown paste and Playa Dull ware ceramics such as the San Martín Variegated Brown (present at the three northern sites, but not at San Claudio) are unusually well represented in Early Classic materials (García Moll 2005; Ochoa and Casasola 1991: 10–11; Rands 1972, 1987: 214). Somewhat curiously, the only orange polychrome present at Mirador, Revancha, and Tiradero, Dos Arroyos, is exceedingly rare at both San Claudio and Piedras Negras (Muñoz 2004). Dos Arroyos also happens to be the type most commonly associated with characteristic Tzakol basal flange bowls, which are also absent from the materials at San Claudio, and indeed, throughout the sample. Moreover, brown paste ceramics

There is a much wider range of vessel forms evident in the Early Classic Caoba complex, although shape classes demonstrate similar distributional variability. At San Claudio and Piedras Negras, ollas with incurving walls and thickened rims, bolstered rim unslipped basins and utility jars, thin–walled moulded rim bowls, shallow plates and cajetes with hollow, conical tripod supports and a basal ridge, and cazuelas and dishes with composite profiles and thick, nearly vertical rims are common (Holley 1987: 189–90; Muñoz 2004). Such vessel forms are rare elsewhere in the sample and in the Petén in general and could represent an incipient localized style resulting from the peripheral location of the sites under investigation and their relative isolation from trends evident in the central Petén (Munoz 2004; Rands 1977).7 In contrast, the forms evident at Mirador, A possibility also suggested by the appearance of distinct, locally produced types at San Claudio and Piedras Negras (e.g., Otatal Orange

7

63

Archaeological Paleography more closely affiliated with the Tzakol sphere (Golden et al. 1998; Golden et al. 2004, Perales Vela 2002; Perales Vela and Mugarte 1995). Ceramics at sites to the north and west of the study area (e.g., Tierra Blanca, El Arenal, Pomoná, Panhalé, Trinidad, and the lower Usumacinta basin) appear only loosely related to the Tzakol sphere, and more closely aligned with the emerging distinctive northwest Maya lowland tradition most evident at Palenque (Anaya Hernández 2002; García Moll 2005; Hernández Pons 1984; López Varela 1991; Ochoa and Casasola 1983, 1991; Rands 1972, 1987; Sánchez Caero 1979; Vargas 1985).8 At these other sites within the middle Usumacinta and lower San Pedro Mártir basin, Early Classic period ceramic materials are scarce, which some investigators suggest may be due to a decrease in population or movement of groups inhabiting the region towards other areas (Ochoa and Casasola 1991: 10–11; Rands 1977). These processes, and the general variability evident in the Caoba complex, may be related to the peripheral location of the lower San Pedro Mártir region at the northwest frontier of the greater Maya lowlands, or a function of larger sociopolitical processes begun in and around the Late Formative–Early Classic period transition, such as increasing centralization, integration within the emerging lowland Maya cultural sphere, or the demarcation of sociocultural boundaries and frontiers (Englehardt 2011; Fahsen Ortega 1999; Grube 1999; Rands 1977).

Revancha, and Tiradero are primarily deep bowls with everted rims, large, steep–walled basins, short–necked, roughly finished jars, and thin–walled, shallow, direct rim dishes. Fluting and incising remain common decorative modes, though some examples of carved or gouged decoration are evident. In general, these are indistinguishable from analogous types (e.g., Lucha Incised) found elsewhere. These forms and decorative techniques again correspond closely to those found in Picota phase Palenque and throughout the Early Classic horizon lower Usumacinta basin and appear more firmly rooted in earlier, Formative period traditions (Hernandez Pons 1984; Rands 1987: 214). Striated jars, however, occur frequently throughout the sample, as they do at Piedras Negras, but not at Palenque (Holley 1987: 189). To recapitulate the ceramic data and sequence, the archaeological evidence indicates that the sites that provide the ceramic materials under investigation were continuously occupied for more than a millennium. Overall, the ceramic materials from the Middle Formative early Bari complex fall clearly within the Mamom sphere, but the presence of Olmec–style objects and the loose association of the Bari complex with ceramic materials found outside of the greater Maya lowlands suggest the coexistence of several traditions at this time. During the Late Formative period, late Bari and Pichi complex ceramics illustrate a firm orientation towards the Maya core area of the Petén lowlands. These materials display strong affinities with the Chicanel sphere ceramics encountered in neighboring sites and regions, for example Altar de Sacrificios, Piedras Negras, Yaxchilán, Moral–Reforma, and El Perú–Waka’ (Adams 1971; Eppich 2004; Fournier 1987; Holley 1983, 1987; Juárez Cossío 1992, 1994, 1999; Muñoz 2002, 2004; Román et al. 2004; Vargas and Hernández 1979).

Variables, Scale, and Analytic Units The first task necessary to evaluate the nature of interaction in prehistoric contexts is to establish the temporal and spatial scales and units of analysis (Parkinson 2002: 398). Within a given geographically defined region, interaction may occur at a variety of levels within ancient societies. Although certain integrative units may come to assume a primary role in interaction, the patterns of that interaction may vary in degree and intensity within, among, and between those units over time (Parkinson 2006). As Parkinson (2002: 399) notes, any attempt to model social interaction in prehistoric contexts must address this dynamic quality.

The Early Classic Caoba complex represented in the sample can only be partially ascribed to the Tzakol ceramic sphere, despite being chronologically contemporaneous. Caoba ceramics from San Claudio strongly resemble Tzakol materials and appear to continue a relational orientation with those from adjacent sites that display influences from the central Maya lowlands. The materials from Tiradero, Revancha, and Mirador, on the other hand, contain unrelated types otherwise unassociated with the Tzakol sphere (Ochoa and Casasola 1991: 10–11). Intra– assemblage and intra–complex variability thus increases. A similar pattern is observed in the ceramics of other sites within the lower San Pedro Mártir basin. Materials from sites to the south and east in the region (e.g., Santa Elena or La Pasadita in the Sierra del Lacandón) are

In this case, the temporal scale of analysis is determined by the nature of the data available. Since the dating of the ceramic sample utilized in this investigation is determined by stratigraphic sequences or ceramic phases than span much longer units of time, it is only possible to consider and discuss interaction as a generalized process that occurred over a broad temporal scale. This fact does not present any major obstacle to the central goal of this study, insofar as the development of writing was itself a long–term process. Patterns in the material correlates of its developmental processes should thus be most visible

Polychrome at Piedras Negras and Bolonchac and Moro Orange Polychromes at San Claudio; Muñoz 2004). These types are usually decorated parallel red and black lines painted over an Aguila–like orange slip. This slip and design combination is restricted to this period and is similar to the slip and decoration found on Terminal Formative vessels from Altar de Sacrificios (Adams 1971: fig. 26b).

The ceramics found at the site of Moral–Reforma being an obvious and notable exception (González Moreno 2006; Juárez Cossío 1992, 1994, 1999; Vargas and Hernández 1979).

8

64

Ceramic Sample and Analyltic Methods when considered over the longue–durée, keeping in mind that patterns of interaction refer ‘to the operation of a palimpsest of interactive spheres that operated sequentially or simultaneously over long periods of time’ (Parkinson 2002: 399). Additionally, due to the diachronically fluid nature of social boundaries, and since the configurations of structural patterns can vary over the landscape (e.g., Barth 1969; Englehardt 2010; Kowalewski et al. 1983; Parkinson 2006; Peterson and Drennan 2003), the spatial scale of analysis is defined in geographic, rather than social terms. Following Parkinson (2002: 399), this study holds that by defining the spatial scale in this independent manner it is possible to generate models of interaction that then can be compared to each other to identify trajectories of sociocultural and technological change within a region.

form and shape (Culbert and Rands 2007; San Roman Martín 2009), and distribution (Hirth 1998). Type– varieties, and affiliations with specific ceramic groups, wares, complexes, and spheres, were determined during the re–classification and elaboration of the ceramic sequence. Following formal re–classification, various techno–stylistic attributes (e.g., surface treatment, slip color, decoration and decorative technique, neck/ support/base/rim/lip type, etc.; see below) and formal– morphological characteristics (e.g., vessel diameter, height, and wall thickness, form and shape class; see below) observable on artifacts within the sample were measured and recorded. From these variable stylistic and formal characteristics it was possible, in some cases, to deduce functions, contexts of use and deposition, and differences in specific processes of manufacture (Rice 1987: 209, fig. 7.1; 1996a, 1996b; Stark 1998).

Thus, in order to address the issues of interaction and innovation, the following analyses examine stylistic and technological patterning at three different social and geographic scales. A comparative and quantitative assessement of the occurrence and frequency of variable typological, techno–stylistic, formal, and deduced functional attributes (cf. Rice 1987: 209, fig. 7.1) was undertaken across multiple analytic units: sites, micro– region (the area in which the sites are located), and region (by comparing micro–regional distribution with distributional patterns in adjacent areas). Diachronic variability in the distributional patterns of the data over time was compared across the same analytic units. At the regional scale of the Maya lowlands, patterns of typological and techno–stylistic variability in ceramic form and decoration during the Middle Formative through Early Classic periods were compared between the sample and adjacent sites and regions. At the micro– regional scale, stylistic and technological variability between Middle Formative, Late Formative, and Early Classic period ceramic assemblages within the study area were analyzed in greater detail. Finally, at the local level of the site, variability in the discrete ceramic assemblages between the individual sites over time was explored. Thus, at the regional scale, analytic units refer to distinct ceramic traditions (or cultural groups). At the micro–regional scale, units refer to the four sites within the study area as a clustered whole (when compared with data from adjacent regions). At the local scale, units refer to the ceramic assemblages of individual sites themselves as discrete archaeological entities.

To elucidate the nature and extent of cross–boundary interaction and integration, These variables, and their distribution within the sample, were then compared with the frequency at which the same characteristics have been observed and reported in established ceramic sequences and typologies at sites in immediately adjacent areas (Figure 4.13), including Palenque and the lower Usumacinta basin to the west, Chiapa de Corzo and the middle Grijalva basin to the southwest, and Piedras Negras, El Perú–Waka’, Altar de Sacrificios, and Uaxactún to the south and east within the Maya lowlands of the middle Usumacinta basin, the Mirador basin, and central Petén (Adams 1971; Bryant et al. 2005; Eppich 2004; Forsyth 1989; Hernández Pons 1984; Holley 1983, 1987; Lee 1972; Muñoz 2002, 2004; Ochoa and Casasola 1991; Rands 1972, 1987; Román et al. 2004; Sabloff 1970; Sánchez Caero 1979; Smith 1955; Smith and Gifford 1966). Each variable class is discussed in turn below. Type–Variety The type–variety system is a method of ceramic typology in which vessels are grouped into successively finer categories on basis of some combination of variables, including paste, decorative treatment, surface finish, and decoration. Type–variety taxonomies are most useful for recording and evaluating variation in surface finish. The classification of type–varieties undertaken to determine a unified ceramic sequence for the lower San Pedro Mártir region yielded clues as to the ceramic affiliations and chronological placement of the various ceramic materials. The appearance or disappearance of ceramic types and varieties at different points in time potentially illuminates the operation of sociocultural processes. For example, if a certain type is widely distributed across a particularly wide geographic area, appearing and disappearing in roughly the same temporal contexts, then it may be reasonably assumed that processes of interaction were involved in the development and

In order to address issues of interaction and innovation as revealed through diachronic variation in the ceramic materials, stylistic, formal, and functional variability evident within the sample were recorded and assessed along four main classes of variables: the traditional type–variety system (Gifford 1960; Sabloff 1969, 1972; Smith et al. 1960; Willey et al. 1967), stylistic attributes (Carr and Neitzel 1995; Hegmon 1989, 1992, 1998; Hurt and Rakita 2001; Plog 1995; Skibo et al. 1989), vessel 65

Archaeological Paleography evolution of that type, precisely because it was so broadly shared. Alternately, wide sharing of ceramic types could speak to the degree to which groups that produced or consumed that type were integrated in larger ceramic spheres (or cultural systems), or the degree to which those larger spheres or cultural systems sought to clearly demarcate their boundaries (e.g., Englehardt 2010; Hodder 1978, 1985; Parkinson 2006; Stark et al. 2000). Additionally, the emergence of distinct types (or varieties within a specific type class) in discrete temporal contexts could indicate a localized adaptation of a widely diffused ceramic template, or the introduction of new technologies or manufacturing techniques (Hegmon 1998; Rice 1996b; Stark 1998). In the comparative analysis the variable occurrence of specific wares and type–varieties within the sample as a whole, was associated with their frequency in the ceramic assemblages reported at sites in neighboring regions. For the quantitative analysis each individual ware and type–variety from successive chronological periods was assigned a number, considered an attribute value, and entered into an Excel spreadsheet to facilitate statistical comparison.9 The statistical analysis of attribute values and the quantitative methods employed in that analysis are discussed at greater length below.

As with type–varieties, in many contexts techno–stylistic attributes may have become standardized or changed at different times during the developmental history or a particular region. In this respect, diachronic variability in the techno–stylistic attributes of ceramic materials may have served as social or ethnic markers, or (as with type– varieties) may yield clues as to the existence and dynamics of particular sociopolitical or cultural processes that occurred in discrete temporal contexts (e.g., Carr and Neitzel 1995; Hegmon 1992, 1998; Hodder 1982; Parkinson 2006; Plog 1995; Stark 1998; Stark et al. 2000; Voss and Young 1995; Wobst 1977). At the same time, techno–stylistic variability observed within types, sequences, and ceramic traditions may be used to identify and describe changes in the context of sociocultural and political organization. New formal and morphological characteristics suggest the introduction of new styles and techniques, possibly facilitated by interaction between ceramic traditions or the integration of a ceramic tradition within a larger cultural system or sphere of material exchange. Using standard procedures (see Rice 1987), the variable techno–stylistic, morphological, and dimensional attributes of each artifact were recorded, and assigned an attribute state—a numerical value that refers to the occurrence of a particular category of trait or characteristic—to the individual artifact on which it occurs. Because not each individual artifact was diagnostic (or indicative of a particular type– variety classification), modes, which are individual traits such as paste, profile, border or lip type, surface color, or any other criteria observable that would be properly labeled an attribute if observed in a specific type or variety, were also considered. Finally, individual vessel dimensions were considered, which were either measured or extrapolated from rim or wall profiles, including diameter, wall thickness, vessel height, and neck length. All assigned attribute values were added to the Excel spreadsheet. In the statistical analyses, the relative distribution and occurrence of specific attributes and attribute combinations within the sample was compared against the frequency at which the same attribute combinations were observed and reported in the ceramic materials of other sites and regions. These categorical combinations are evident in the output of statistical analyses. They include classificatory attributes (ware, type–variety, slip color, surface treatment, and decoration type), paste attributes (color, texture, and temper content), morphological attributes (base type, wall type, rim type, support type, and lip type), formal attributes (form or shape class and body angles or flanges), and dimensional attributes (wall thickness, vessel height, and neck length), although not all attributes or classes were always clearly visible or identifiable in all materials.

Techno–Stylistic Attributes and Dimensions Techno–stylistic attributes are those elements of construction, production, or fabrication that are combined in the formation of an artifact and specifically related to the process of manufacture of ceramic vessels. At its most basic level, an attribute is a variable property, characteristic, or feature of an entity. These are observable data that represent the most rudimentary aspects of manufacture of which the producer of a particular vessel could have been conscious or aware. These attributes thus encompass characteristics or traits particular to the design and technologies employed in the creation of a ceramic artifact. Such stylistic or symbolic characteristics possess a singular importance for the type and a meaning beyond any purely descriptive trait (i.e., beyond the heuristic value of the type–variety description; cf. Hodder 1987; Sabloff and Smith 1969). The most common stylistic attributes considered in ceramic analyses are paste, slip, temper, surface treatment, color, decorative technique, and vessel morphology (for more detailed definitions, see, e.g., González Moreno 2006; Orton et al. 1993; Rice 1987; Rye 1981; Shepard 1965; Sinopoli 1991). The specific techno–stylistic attributes recorded and assessed within the sample include decoration type, surface treatment, slip, paste, temper, and morphological traits related to vessel form and shape, such as neck, base, support, lip, body angle, and rim type (Figure 4.17).

Form and Shape Class Slight variation in the evolution of forms over time was also measured, yielding subtle yet detectable changes in different vessel shapes across the sample. Morphological changes in form refine the placement of

For the specific numerical values assigned to measured attributes within the sample, please refer to appendix A in Englehardt 2011.

9

66

Ceramic Sample and Analyltic Methods

Decoration type

incised grooved fluted striated painted

Surface treatment

Slip

smooth

color

Neck type

polished Lip Type

orange red color black cream presence/absence

Paste

fine grainy/included orange red black cream

flat pointed external bevel internal bevel fluted medial flange basal flange basal angle Z base angle external rounded projection

horizontal everted projection   inflated interior

vertical straight everted

Rim profile/ collar type

flat convex rounded

Body angles/ flanges

content

curved everted

Base type

Support type

monochrome polychrome burnished

rough

texture

Temper

pedestal rounded conical cylindrical hemispheric globular mammiform rounded

direct inflated exterior curved everted projection internal rounded projection T shaped straight everted projection

Figure 4.17. The variable stylistic attributes and categories of those attributes that were observed and recorded on diagnostic artifacts within the sample.

a specific ceramic artifact within a given chronological sequence, since forms evolve more rapidly than types, and demonstrate a greater range of variation (Culbert and Rands 2007: 185; Sabloff and Smith 1972). Form classes identified in the sample include plates, jars, cups, ollas, tecomates, apaxtles, cajetes, and cazuelas. Each form class encompasses several categories of a particular form or shape (e.g., lightly incurved cazuela, curved everted jar, rounded cajete, etc.). Such categories generally address the function, reflecting context of use, and the size of the vessels, as shape classes are closely related to function (Culbert and Rands 2007: 185; Sabloff and Smith 1969; San Román Martín 2009; Rice 1987: 209, fig. 7.1).10 In addition, the development and emergence of new forms or shape classes often corresponds to the appearance of new ceramic types or stylistic attributes, or

the introduction of new production technologies (Culbert and Rands 2007; Stark 1998). Thus, form classes may be related to both stylistic and functional innovation. In determining form and shape classes, it bears mention that not all sherds and materials within the sample are diagnostic, or necessarily indicative of particular shape or form classes or categories. In general, this study follows established ceramic procedures in deducing form class from sherd morphology (e.g. Orton et al. 1993; Rice 1987; Rye 1981; Shepard 1965; Sinopoli 1991). In the simplest terms, sherd thickness and curvature are often good indicators of overall vessel form: basal sherds are usually thickest, body sherds typically lack morphological attributes and occur as either flat or slightly curved, collars usually consist of thickened areas that provide extra relief to the vessel wall profile and are occasionally decorated, neck sherds are relatively thin compared to sherd thickness on other parts of the vessel, and so on. Sherd thickness was measured and recorded, from which it was possible

For example, the volume of a particular vessel is often a product of its function, since different activities (i.e., storage, cooking, processing, etc.) usually require different kinds of containers. 10

67

Archaeological Paleography to deductively approximate vessel diameter and height. These characterizations also rely on the preliminary shape classifications undertaken by González Moreno (2006) and Hernández Ayala (1981).

although the quantitative statistical analysis of variation in the distribution of ceramic types, attributes, and form classes inherently adds a higher degree of resolution to interpretation than a purely comparative evaluation.

To record temporal variation in form, shape classes and their sub–categories were related to corresponding ceramic phases based on specific techno–stylistic and morphological modifications over time, such as rim orientation, thickness of the walls, or occurrence of specific decorative techniques. Following González Moreno (2006: 98–103) and San Román Martín (2009), to connect shape class and phase within the same analytic category, a concise description of the diagnostic characteristics that warrant the specific form and class–phase associations was developed. Diachronic in–category variability allows for inferences regarding relative degrees of standardization of vessel shapes and dimensions over time within the sample.

Comparative Analysis of Attribute Variability The goal of the comparative assessment was to elucidate spatial and temporal patterns of stylistic similarity and variability in the material data over time. Style is an extremely powerful construct that can help to bridge the gap between archaeological materials and the groups, peoples, or traditions that created and produced such objects (Hegmon, 1995: 7; Parkinson 2006: 36). The comparative analysis and interpretation of stylistic, technological, and formal change in material culture is one of the most essential methods available to the archaeologist for reconstructing patterns of interaction in prehistoric contexts. Stylistic, formal, and functional analyses of attribute variability have thus proven valuable tools in discerning the nature and extent of interaction observed in material markers (e.g., Cheetham 2007; Hodder 1978; O’Shea and Milner 2002; Parkinson 2002, 2006; Stark 1998).

After measuring the sampled materials, the form and shape classes were recorded in the Excel spreadsheet, assigning each specific form class and category a unique numerical attribute value. Again, formal class variability was compared within the sample as well as between the sample and the sequences and materials encountered in adjacent areas. Raw data regarding the occurrence of specific forms in adjacent areal sequences was unavailable. It was only possible to perform a cursory comparative evaluation between the sample and adjacent sites and regions, although statistical analyses of form classes present within and among the sample was undertaken.

Following the formal classification, a diachronic comparative evaluation of the techno–stylistic, formal, and functional attributes of material evidence found at the sites under investigation was undertaken in order to characterize the degree and quality of variation observable in the collected data. Measuring grades of similarity and diversity within and among the sample, and between the sample and ceramic assemblages from neighboring sites and regions in differing temporal contexts allows me to identify and characterize the nature and extent of interaction that occurred across the boundary on which they are situated in both synchronic and diachronic terms, and on multiple scales (i.e., in terms of the analytic units of site, micro–region, and region). Variation in the material data permits an appreciation of the permeability and social maintenance of boundaries (i.e., the relative ease with which ceramic styles, forms, and technologies were shared or transferred between regions), adding resolution to inferences regarding regional interaction.

Distribution The final variable considered is distribution. Whereas other variables address the visual aspects of stylistic and morphological variation, distribution allows for consideration of the patterned spatial dimensions of that variability, in both synchronic and diachronic terms (Voss and Young 1995: 92). In this case, distribution refers to the relative occurrence of different type– varieties, techno–stylistic attributes, and form classes evident in the sample from different sites within the study area. The relative distribution of these variable categories in the sample was also compared with their occurrence, frequency, and distribution at sites and regions outside of the study area. By assessing variation in the distribution of a battery of different ceramic variables within, among, and between the different sites within the study area, as well as between the sample and adjacent regions, it is possible to assess the relative kinds and degree of interaction that occurred within and among the sites in the study area, both as clusters and individually, as well as between the sites in the study area and adjacent sites and regions. Distribution may be assessed both comparatively and quantitatively,

In addressing the relationships between attribute variability and sociocultural processes, this investigation follows several recent comprehensive treatments of this complex topic (e.g., Carr and Neitzel 1995; Hegmon 1992, 1995, 1998; Parkinson 2002, 2006; Plog 1995; Stark 1998). These investigators generally agree that it is necessary to consider various distinct kinds of stylistic variation. The data is therefore approached at several different levels to assess variability in a variety of discrete attributes, which may themselves have functioned differentially within diverse sociocultural, spatial, or temporal contexts (Parkinson 2002: 399). To determine which aspects of variability indicate different dimensions of social 68

Ceramic Sample and Analyltic Methods regional scale, ceramic traditions (or cultural groups), at the micro–regional scale, the sample sequence as a representative whole (when compared to the ceramic sequences and assemblages at neighboring sites), and, at the local scale, between the ceramic assemblages from each individual site within the study area (Parkinson 2002: 400, table 1, 2006: 37, table 1). The table includes only those interpretations that are applicable in discussing the patterns in ceramic assemblages dealt with in this comparative analysis. The variable attribute distributions (i.e., uniform, discrete, clinal, and random) relate to the relative occurrence of different attributes within, among, and between distinct ceramic traditions, assemblages, and site– or region–specific sequences at different scales and in differing temporal contexts. Underlying the various relationships drawn in the table is the fundamental assumption that uniformity (i.e., homogeneity), within an analytic unit is most likely indicative of a high degree of interaction between units, regardless of the visibility of the attribute (Parkinson 2002, 2006; Wobst 1977). As Parkinson (2006: 38) cautions, the interpretations

interaction, this study follows Carr (1995) and Voss and Young (1995) in concentrating on two comparative qualities that are helpful in determining the role a specific attribute could have played within a particular context: 1) the relative visibility of the attribute, and 2) its geographic distribution (Parkinson 2002, 2006). These authors suggest that by focusing upon the variation in the relative visibility and distribution of particular stylistic elements, one may generally deduce the likely role (or roles) that a specific technological, stylistic, or formal attribute could have played in variable sociocultural, geographic, or temporal contexts (Parkinson 2002: 399, 2006: 36). Although these authors are concerned primarily with style as a general social phenomenon, the tenets they outline provide a general guideline for the potential identification of interaction in material remains. Figure 4.18 synthesizes the interpretations presented by Carr (1995) and Voss and Young (1995: 92–3) for the distribution of high and low visibility characteristics given different spatial distributions within, at the

Attribute visibility

Distribution within unit*

Degree of Distribution interaction * between units within units*

High

Uniform

Discrete

High

High

Uniform

Clinal

High

High

Uniform

Uniform

High

High

Clinal

Clinal

Moderate

High

Discrete

Discrete

Low

High

Random

Random

?

Low

Uniform

Discrete

High

Low

Uniform

Clinal

High

Low

Uniform

Uniform

High

Low

Clinal

Clinal

Moderate

Low

Discrete

Discrete

Low

Low

Random

Random

?

Degree of interaction Potential interpretation of patterns between units* Marking well–maintained group Low boundaries between units, little trade between units Marking permeable group boundary Moderate near or within units; some trade between units Marking group unity or cooperation High across area; extensive trade between groups Marking very permeable group boundaries between and within units; Moderate continuous trade throughout area Marking settlements as group Low boundary; little or no trade between units ? Marking individual creativity? No shared learning or passive Low interaction between units? No trade between units? Some shared learning between units? Moderate or Some trade between Low units? Much shared learning between units; High trade between clusters Shared learning between units; trade Moderate between units No shared learning between units; No Low trade between units? ? Individual competency?

Figure 4.18. Comparative interpretation of stylistic attributes. *At the regional scale, units refer to distinct ceramic traditions (or cultural groups). At the micro–regional scale, units refer to the four sites as a clustered whole (when compared with site clusters from adjacent regions). At the local scale, units refer to the individual sites. After Parkinson 2002: 400, table 1, 2006: 37, table 1 (Carr 1995; Voss and Young 1995).

69

Archaeological Paleography overlap a great deal, and they are susceptible to alternate explanations in different sociocultural, geographic, or temporal contexts. Generally more visible attributes tend to correspond to a wider range of more active social processes (e.g., of active interaction and group– boundary maintenance), and less visible (more obscure) attributes primarily relate to more passive processes (e.g., of enculturation). For example, the distribution of high–visibility attributes is more likely to indicate active interaction between groups, boundary–marking, and to a lesser extent, innovation. Conversely, the distribution of low–visibility attributes is more likely to indicate passive interaction through learning, intermarriage, or enculturation. These potential interpretations provide a basic framework for illustrating which ceramic attributes reveal different patterns of active and passive interaction at several different levels throughout and across the study area. To facilitate the comparison of analytic results, measured ceramic attributes were grouped into five distinct classes—classificatory, paste, formal, morphological, and dimension—and into high and low visibility attributes, following Carr (1995), Parkinson (2002: 400, table 1), and Voss and Young (1995; see also Parkinson 2006: 37, table 1).

characteristics shared within both the sample and regional sequences. On the other hand, fewer shared attributes, reflected in greater formal or functional discrepancies in the material data, suggest less interaction and a greater degree of localized innovation in terms of differential manufacturing processes, the development of localized styles, or the introduction of new material technologies (Stark 1998). If interaction plays a major role in processes of material innovation in this case, one would expect stylistic, formal, technological, and functional variation in ceramic design attributes to occur in roughly the same temporal contexts across analytic units and between the sample and adjacent areal typological sequences. Quantitative Analyses of Similarity and Diversity Quantitative analysis adds resolution to the comparative assessment by elucidating the nature, temporal contexts, and extent of interaction deduced from the stylistic analysis. Statistical measures of similarity and diversity serve to quantitatively evaluate relative variability within and among the sample dataset and complement the comparative analysis. For the quantitative study of attribute similarity and diversity, two statistical measures of the recorded variables outlined above were employed. In considering similarities and differences in ceramic assemblages between sites and among the sample as a whole, a cluster analysis of the distributional patterns of individual analytic variables, was undertaken based on analyses of variance (ANOVA) of mean attribute similarity and distance, derived from the attribute values assigned to the ceramic data. Variability within the sample was also considered, for which a relative heterogeneity measure (H score; see Garraty 2009) was calculated. Including statistical measures of both inter– assemblage similarity (ANOVA cluster analysis) and intra–sample diversity (H scores) adds further detail to interpretations regarding interregional interaction.

Variation in the evidence is interpreted by comparing the relative percentages and occurrences of each stylistic variable within the sample during each chronological period (Middle and Late Formative, Early Classic) against the formal attributes evident in the ceramic sequences of adjacent sites and regions (and established Mesoamerican ceramic typologies) in corresponding temporal contexts. The identification of variability allow for the determination of the direction, intensity, and temporality of interaction on both regional and micro– regional scales. This comparison also illustrates relative levels of sample richness (Kintigh 1984; Leonard and Jones 1989) as well as vertical and horizontal linkages across archaeological remains through time and space. Functional variability and material innovation are deduced through a consideration of discrepancies in design characteristics, technological attributes, shape classes, or site–specific distribution. The comparative analysis allows for the identification of patterns of formal and functional homologies and variability reflected in the material data.

Essentially, these statistical measures illustrate the distribution of specific typological, stylistic, and formal variables within and across the sample. Comparing these variables in terms of clustered patterns of similarity and H scores in distinct temporal contexts and on different scales illuminates variability in the quantity, intensity, and directionality of dynamic interregional exchange networks, as expressed at the micro–regional level within the sample. For example, similar cluster patterns of attribute similarity across assemblages and lower intra–assemblage diversity reflected in H scores would suggest greater distributional uniformity between and within the sampled material complex, and therefore a relatively greater degree of interaction on a regional scale. These measures of similarity and diversity also speak to material innovation, insofar as greater inter– assemblage variability and increased diversity within assemblages form distinct sites suggest the emergence of new forms and functions over time. If interaction is

In turn, these patterns of synchronic and diachronic correspondence or variation in ceramic styles, types, and forms allow for an initial characterization of the degree and direction of interaction that occurred in and across the San Pedro Mártir region in distinct chronological periods. Although the interpretive value and implications of these patterns is duscussed in greater detail below and in following chapters, briefly, the presence of widely–shared synchronic stylistic homologies observed in the majority of defined variables implies extended interaction, reflected in formal and functional 70

Ceramic Sample and Analyltic Methods behind observed variation, one would expect the patterns of interaction suggested by the statistical analyses to closely correspond in temporal and spatial contexts.

and structures in the sampled data, and subsequently attempt to correlate those patterns with the results of the comparative analysis.

For the quantitative analyses, a ten percent stratified random sample of the ceramic materials from the four sites in the study area was selected. The raw data during fieldwork was recorded and entered specific measurements into an Excel spreadsheet. Attribute values were later assigned to the raw data. Whereas for the comparative analysis it was possible to consider materials and assemblages from neighboring sites, for the quantitative analysis a consideration of specific evidence from other sites was impossible, as raw quantitative data regarding ceramic collections from adjacent areas was unavailable. The statistical analyses therefore consider only those materials within the sample, and the distribution of specific type– varieties, techno–stylistic attributes, and form classes within the study area. It was, however, possible to compare the results of the quantitative measures with the reported frequencies and occurrences of the same variables in adjacent assemblages. Insofar as the statistical analyses employed primarily speak to the distribution of ceramic types, forms, and attributes within the sample, the results of the quantitative analyses (in terms of their relationship with processes of interaction and integration) were interpreted along the lines outlined in Figure 4.18 above. Below, the particulars of the two statistical measures employed are discussed. The results of these statistical analyses and their implications are presented and discussed in the following chapter.

The creation of a series of two–step hierarchical tree clusters derived from statistical distance between measured attributes in the sample was initially planned. This process proved impractical, due to discrepancies in sample sizes between chronological periods and the high degree of similarity in all attributes among all sites in the Middle and Late Formative periods. Although the tree plots produced were valid and interpretable, they were unwieldy and overly difficult to display in graphic form. Since a separate database of assigned numerical attribute values had been previously created, an analysis of variance (ANOVA) was employed to cluster the data. ANOVA is essentially a collection of statistical models, and their associated procedures, in which the observed variance in a particular variable is partitioned into components attributable to different sources of variation. In its simplest form, ANOVA provides a statistical test of whether or not the means of several groups are all equal. Thus, by analyzing mean attribute values it was possible to determine similarity and distance between and among the attributes present in the individual assemblages from each site within the sample. Although based on numerical means derived from the assigned attribute values rather than the attributes themselves, ANOVA nonetheless allowed for the identification of statistically significant differences (p < 0.05) between both individual variables and suites of attributes, effectively illustrating ‘clusters’ of similar attributes (i.e., those in which no significant difference in means was observed).

ANOVA Cluster Analysis of Mean Attribute Similarity and Distance

The primary advantage of ANOVA is that it permits for comparison of the components of the total deviation, as well as their distribution, since ANOVA is based on comparing the F–test statistic11 with an F– (or continuous probability, i.e., null) distribution. ANOVA also permitted the creation of clustered boxplots of attribute ranges that effectively illustrated how the data was joined into individual clusters, with sites as the dependent variable. Although clusters derived through ANOVA are based on mean attribute values derived from arbitrarily assigned attribute value numbers, the presence, absence, or relative occurrence of certain attributes, or suites of attributes, at particular sites within the sample is expressed by relatively larger ranges or differences in the means themselves. A further advantage of ANOVA was that it allowed me to perform a post–hoc Tukey analysis of honestly significant differences (HSD). Tukey’s test compares the means of every treatment to the means of every other treatment and identifies where the difference between two means is greater than the standard error would be expected to allow. Tukey’s test illustrates the

The methodological term ‘cluster analysis’ does not identify a particular statistical method or model. Rather, cluster analyses encompass a number of different algorithms and methods for grouping objects of similar kind into respective categories. The general purpose of cluster analysis is to discover a system of organizing observations into groups whose constituents share a common property (or properties). In essence, cluster analysis classifies a set of observations into groups based on combinations of interval variables, assigning sets of observations into clusters so that observations in the same cluster are similar (Leonard and Jones 1989; Shennan 1988). In other words, cluster analysis is tool for data analysis whose goal is to sort different objects into groups in a way that the degree of association between two objects is maximal if they belong to the same group and minimal otherwise. In this case, based on statistical measures of distance and similarity, cluster analysis has the potential demonstrate similarity and variability between the formal attributes observed in the sampled assemblages and characteristics of ceramic materials in neighboring sites and regions (Kintigh 2002; Shennan 1988). Cluster analysis was applied to discover patterns

F = variance between items/variance within items, also expressed as F = MSTR/MSE where MSTR = SSTR/I–1 (I = number of treatments) and MSE = SSE/nt– I (nt= total number of cases).

11

71

Archaeological Paleography presence or absence of statistically significant degrees of difference between the attributes evident at a specific site in a particular temporal context. The results of the post– hoc Tukey analyses greatly assisted in associating the results of the statistical analyses with the comparative evaluation by illustrating which sites exhibited attributes that were relatively more alike or dissimilar in comparison to other sites in the sample, as well as in relation to the attributes reported in ceramic sequences and assemblages from neighboring regions and sites.

on objects that are further apart, and thus to illustrate more clearly distances and similarities in the clustered boxplots that the PASW program produces. A successful joining analysis allows for the detection of distinct clusters, represented as the ranges and mean values of a boxplot graph, which may then be interpreted. In interpreting clustered boxplots, each range in the graph represents a grouping of objects that display less variability between distance criteria (i.e., the objects in that cluster are essentially more similar than those in other clusters). When the data contain a clear structure in terms of clusters of objects that are similar to each other, then this structure will often be reflected in the boxplot through different mean attribute values, or larger ranges of variability in attribute values. Evaluating the means of each cluster on each dimension through the Tukey test permits appreciation of how distinct those clusters actually are. Ideally, one would hope to obtain very different means for most, if not all, dimensions used in the analysis as the number of distinct clusters and statistically significant differences between and among attribute values increases. Essentially, fewer clusters indicate greater intra–assemblage similarity (i.e., less distance between specific dimensional variables). Clusters of specific variable elements were then compared with the relative frequency of occurrence of the same variables in adjacent areas to deduce variability between the sample and regional ceramic sequences. Greater similarity (i.e., fewer distinct clusters; less variability) would suggest a greater degree of interaction between the sites or areas (and their immediate neighbors) that exhibit such correspondences. Comparing differential cluster patterns of the variables over time and at different scales elucidates the relationships between the sampled material assemblages and adjacent sites and regions. These patterns allow for inferences regarding the extent and direction of interaction that occurred within the study area, and also speak to degrees of integration within cultural systems. Cluster analysis thus permits the graphical illustration of the nature of interregional interaction reflected in the ceramic materials.

Utilizing the PASW(SPSS) Statistics 18 software suite, it was first determined whether or not significant differences existed between the mean attribute values of the ceramic materials from each site in each successive chronological period. The data were then joined by creating a series of clustered boxplots based on dissimilarities or distances between mean attribute values. In this case, the distances are based on the degrees of similarity between each variable (i.e., type– variety, techno–stylistic and dimensional attributes, and form class) within the sample. The rule by which objects were grouped is based on multiple dimensions. First, the univariate distribution of each attribute (or variable) within the sample was considered, creating clusters based on the occurrence of specific elements, types, forms, and suites of attribute categories at each site in the study area. Those variables for which specific metric measurements were available (e.g., wall thickness, vessel height, neck length) were also considered, attempting to group those materials that display similarity in height, wall thickness, etc. Additionally, the sheer number of distinct attributes present on a given ceramic artifact, or within a site–specific assemblage, were considered as a crude measure of relative morphological complexity.12 Materials were grouped based on the number of techno– stylistic attributes that they display. Finally an attempt was made to group materials based on the frequency of their occurrence in the sample at different times, in order to visually demonstrate how the essential makeup of the sample as a whole changed diachronically. Clustered boxplots for the range of variance in each attribute value as well as in the larger attribute categories outlined above for each successive chronological period (i.e., Middle Formative, Late Formative, and Early Classic) were also created. The PASW program illustrates distances between objects and their attribute values in multidimensional space by computing Euclidean distances, the most common and methodologically accepted measure of distance. Euclidean distance is simply the geometric distance in the multidimensional space, computed as: distance(x,y) = {∑i (xi - yi)2}½. I also computed squared Euclidean distance (distance(x,y) = ∑i (xi - yi)2) in order to place progressively greater weight

H Score Measures of Diversity Levels of internal multivariate diversity within the sample were also quantified and compared. Diversity measures provide a simple and concise way of quantifying and comparing differences among a set of analytic units using a multivariate dataset (e.g., Garraty 2009: 160; Kintigh 1984; Leonard and Jones 1989). In this case, the variables refer to the distribution of distinct ceramic types and their respective attributes as a collective whole. This study adapts the recent operationalization of the distributional approach to identifying interregional interaction (Hirth 1998) developed by Garraty (2009). The H score diversity measure follows the same basic principle as measures of evenness by quantifying

12 E.g., a thick–walled, unslipped monochrome olla would be less morphologically complex than a thin–walled, orange slipped polychrome incised cajete with an everted rim.

72

Ceramic Sample and Analyltic Methods to perform descriptive statistical analyses of the sample as a whole at different points in time to determine both expected and observed percentages. These data were then input in Microsoft Excel, and a formula was programmed to calculate and tally the squared distances according to the H score formulaic equation.

deviation between observed variable proportions for a given case and a hypothetical even distribution of variables (Garraty 2009: 160; Leonard and Jones 1989; Shennan 1988). Rather than calculating distances from perfect evenness, ‘heterogeneity is defined as the differences in the variable composition between each analytical unit and a hypothetical standard’ (Garraty 2009: 160). This baseline standard of ‘representative’ variable composition forms the ‘expected’ proportions against which observed and measured proportions are compared, calculated as the variable percentages over all cases combined (Boone 1987: 340; Garraty 2009: 160).

The calculation yielded a single, positive score for each collection, illustrating distance between the sample collections and overall ‘representative’ type percentages (Garraty 2009: 160; Kintigh 1984, 2002). The score was then ‘weighted’ to calculate distance between individual collections and overall type percentages using Brainerd– Robinson coefficients of similarity (Brainerd 1951; Garraty 2009: 160; Kintigh 2002; Robinson 1951). The Brainerd–Robinson (BR) coefficient, widely used in archaeology, is a similarity measure that ranges from 0 (dissimilar) to 200 (similar). The statistic totals the absolute value of the differences of the type percentages between defined categories for pairs of assemblages. The sum of the differences is subtracted from 200, because the maximum possible ‘distance’ between two collections, based on percentages, is 200. By subtracting any calculated difference from 200, an equivalent measure of similarity is obtained. The formula is BRAB = 200 - S (i = 1 to N) |PiA - PiB|.13 A BR value of 200 represents the highest possible similarity, whereas zero represents the lowest possible similarity. For more intuitive results, the BR coefficient can be scaled by dividing the statistic by 200; thus a scaled BR value of 1 represents identical assemblages, whereas a scaled BR value of zero represents totally different assemblages.14

The measure illustrates the cumulative differences between observed percentages and perfect evenness across all variables. This method establishes a relative measure of diversity among the dataset and is well– suited for comparing collections within or among sites, regions, and sub–regions (Garraty 2009: 161). An attempt was made to expand and modify the original method to evaluate its utility on the regional scale and in wider temporal contexts, as Garraty (2009: 159) suggested. Following Garraty, and utilizing both PASW and Excel software, a heterogeneity measure (H score) was determined to quantify diversity and compare variability in the distribution of distinct ceramic types and the variables defined above within the sample (Garraty 2009: 160). In this case, heterogeneity refers to differences in variable composition between each unit and a hypothetical ‘representative’ variable composition, against which observed proportions of type–varieties, stylistic attributes, and forms in the sample were compared (Garraty 2009: 160). In other words, the cases are the individual assemblages from each site selected for study, and the variables include the various type– variety categories, form and shape classes, and stylistic attributes observed and measured within the sample.

H scores and scaled BR coefficients derived from the sample were then compared with observed distributional proportions of type–varieties and forms (idealized variable composition; Garraty 2009: 160) across analytical units (site, micro–region, and region) and over time. Generally, lower scores signify that observed percentages closely resemble the expected pattern, indicating greater intercollection uniformity, thus suggesting greater interaction across a larger area and relatively integration in a centralized regional system. On the other hand, higher scores indicate a larger difference between observed and expected percentages, suggesting greater heterogeneity among the collections, implying a lesser degree of interaction and greater integration in a localized system (Garraty 2009: 160; Kintigh 2002). In the absence of raw data with which to perform H score diversity analysis of ceramic assemblages at neighboring sites, the results of these diachronic diversity measures within the sample are evaluated against those derived through both the comparative and cluster analyses.

‘Perfect heterogeneity’ is the baseline against which distributional variability is evaluated. After Garraty, perfect heterogeneity is defined as the overall global proportions of each category of ceramic type or form (j) among the total of all collected samples within the selected region (i) (Boone 1987; Garraty 2009: 160). The reference dataset of idealized global type proportions comprise the ‘expected’ values (Pij) against which individual sample collections (pij; Garraty 2009: 160) were compared. In the case of type–varieties, the idealized base set of proportions stems from external datasets, namely, the percentages of the same categories present in typological collections diagnostic of ceramic sequences at sites and regions adjacent to the study area. For formal and stylistic attributes, a hypothetical standard based on the variable percentages of specific attributes over all cases within the sample combined was employed. The H score is the sum of the squared deviations between the observed per–unit percentages and the expected, ‘idealized’ pattern across all formal and typological categories (∑[Pij – pij]2; Garraty 2009: 160). With PASW, it was possible

Where PiA is the percentage representation of attribute or type i in assemblage A, and PiB is the percentage representation of attribute or type i in assemblage B. 14 Alternately, a scaled BR coefficient can be calculated as 1.0 minus the half the sum of the absolute value of the differences in the proportions across all variables. 13

73

Archaeological Paleography the result of processes reflected in the material data. Variability in ceramic materials thus speaks to a variety of sociocultural and technological processes, including the permeability and social maintenance of boundaries, the degree of interaction that occurs across such boundaries, and degrees of integration within cultural systems and traditions. Innovation and variability in specific manufacturing processes is reflected in technological and stylistic discrepancies across archaeological remains. Following that, the comparative analysis of stylistic and functional variation in the ceramic data was discussed. By comparing patterns of variability in the sample with established sequences in neighboring sites and regions across differing temporal contexts it is possible to characterize interaction in both synchronic and diachronic terms, and on multiple scales. Finally, the chapter detailed the quantitative measures of similarity (derived through cluster analysis) and diversity (inferred from the H score heterogeneity measure) at the site and micro–regional scale that I employ to nuance my interpretations of varying degrees of interaction and integration. The following chapter presents and discusses the results of these comparative and statistical analyses, as well as their interpretation.

Conclusions This chapter has presented and discussed the ceramic sample that is employed in this investigation. It also detailed the archaeological contexts of the approximately 22,000 ceramic materials that comprise the sample, and reviewed the methods utilized in the creation of a holistic ceramic sequence for the San Pedro Mártir basin. This process involved the typological and formal re–classification of the ceramic materials examined and documented during fieldwork, as well as previously published materials (González Moreno 2006; Hernández Ayala 1981). The ways in which the sample sequence is related to broader regional ceramic typologies and sequences was then discussed, briefly presenting some preliminary similarities and differences between the ceramic materials evident in the sample and those present at neighboring sites in adjacent areas and regions over time. The analytic methods employed in this study were then explored, outlining the scales, analytic units, and variables observed and measured within the sample, including type–variety, specific techno–stylistic attributes, form class, and distribution. Variation is considered as

74

Chapter 5 Interpreting the Results of the Comparative and Statistical Ceramic Analyses In this chapter, the results of the comparative and statistical analyses performed on the ceramic data are presented. The patterns of interaction and material innovation suggested by these results are discussed. Based on the results, the general patterns of variability over time and through space that are evident in the ceramic sample are detailed. Synchronic and diachronic variability both within the sample and between assemblages at the regional level is examined. The chapter concludes with a discussion of the implications of those patterns for issues of interaction and material innovation. In short, the ceramic data do support the contention that interregional interaction played a vital role in the development, maintenance, and innovation of material traditions in the temporal and spatial contexts of this study. Results suggest that diachronically variable degrees of interaction within and across the study area positively correlated with a number of sociocultural processes, including maintenance of material traditions, material innovation, boundary formation, and variable degrees of integration of the study area as a whole, and the sites therein, within larger, pan–lowland material traditions and broader sociopolitical configurations. To contextualize these results, it is useful to revisit ‘established’ conceptions regarding pan–lowland regional chronology and complexity. Prior to the Middle Formative period, evidence for complex societies in the lowland Maya region is limited (González Moreno 2006: 48–51; Hansen 1991; cf. Rosenswig 2010). The traditional view of the Maya lowlands at this time is that the region was a loosely organized amalgamation of comparatively homogeneous small–scale societies that lacked status differences or centralized political leadership and were relatively less integrated into networks of regional interaction and exchange (vis à vis the Gulf coast and the central Mexican highlands, for example; see, e.g., Fournier 1987; González Moreno 2006: 49). Since there is no data in my sample that predates the Middle Formative—and very little Early Formative evidence from the study area in general— the ceramic analyses offer no direct insight regarding the veracity of these claims. However, the patterns that emerged from the ceramic analyses—particularly when considered in tandem with the iconographic and linguistic evidence presented in the following chapter—do speak, albeit indirectly, to these issues, especially regarding the ‘established’ pan–Mesoamerican chronology (i.e., Formative, Classic, Postclassic). A fuller treatment of those implications is reserved for the synthetic discussion in Chapter Seven.

75

Following the Early–Middle Formative period transition, however, c. 900–800 BC, interaction between the Maya lowlands and other Mesoamerican regions began to increase. The ceramic evidence in the sample suggests that a great degree of interaction occurred within and across the study area in the Middle Formative. The clearest indicator of this increase in interregional communication and exchange is the Middle Formative tendency toward homogeneity in ceramic production and materials that had, at the end of the Early Formative period, begun to demonstrate regional variability. The data suggest that this trend continued into the early Late Formative. During the later part of the Late Formative period, and certainly in the Terminal Formative or ‘Protoclassic’ (c. AD 100–250) and Early Classic periods, interaction across the study area appears to have decreased significantly, as reflected by increased variability in lowland ceramic traditions, the innovation of new ceramic forms and modes specific to the Maya lowlands at certain sites, and greater diversity within and between the sample relative to adjacent sequences. This development is roughly contemporaneous with general trends toward greater complexity in lowland sociopolitical configurations, more defined settlement hierarchies, and a tighter control over sociocultural and political boundaries associated with the Formative– Classic period transition (Englehardt 2010; Fahsen Ortega 1999; Grube 1999; cf. Guernsey 2012). At this point, lowland Maya ceramic production became a more specialized activity; an incipient lowland Maya style developed that was distinctive from contemporary traditions elsewhere in Mesoamerica. The homogeneity in ceramic design and production that characterized the Middle and early Late Formative periods was replaced by regional variability and increasing heterogeneity in ceramic materials, both across regions and within the lowlands. Changes in the design and production of finished articles do not appear to have been impeded by developing political hierarchies or the burgeoning political power of emergent elites, but their distribution does seem to have been affected by the stricter control over borders that developed in the Early Classic period (Englehardt 2010). Thus, the data reflect less variability in earlier temporal contexts, indicating more interaction, followed by greater variation that implies a greater degree of innovation, decreasing interaction, and increasing integration into more localized cultural systems in the later temporal contexts of the crucial Late Formative–Early Classic

Archaeological Paleography period transition. Perhaps unsurprisingly, the distinctive characteristics of Classic Maya society also appear at and around the Formative–Classic transition, such as polychrome ceramics, roads and infrastructure, and a greater amount of monumental architecture, as well as the Classic Maya writing system. Given the timing of these developments, it would appear logical to assume that interregional interaction was intimately involved with both material and scribal innovation in the spatial contexts of both the lowland Maya region as a whole, as well as the study area. Again, however, a fuller consideration of this possibility is reserved for Chapter Seven. For the moment, this discussion must turn to the presentation and discussion of the statistical analyses performed on the ceramic sample

ANOVA Analysis of Middle Formative Period Ceramics ANOVA analyses of the Middle Formative period ceramics revealed no statistically significant (p < 0.05) differences between the ceramics encountered at each site (Figure 5.1). Thus, as suggested by the San Pedro Mártir basin regional ceramic sequence, during the Middle Formative period there is a great deal of uniformity in the ceramic materials among sites and across all measured parameters. There is also a relatively even distribution of all attributes across all sites. Such uniformity between sites across variables is likely due to the fact that a single type–variety, Joventud Red, dominates the assemblages, accounting for 82.8% of the ten percent stratified random sample (Figures 5.2 and 5.3).1

Results of Ceramic Analyses

In terms of classificatory attributes, the prevalence of Joventud Red ceramics also accounts for predominance of red slips (at 85.9% of the sample), as well as the fact that 93% of Middle Formative period ceramics are Paso Caballos Waxy ware, 97% are undecorated, and 77% have a smooth or polished surface.2 For the paste attributes during the Middle Formative period, 90% of all sampled ceramics are calcite tempered. Fine to medium pastes dominate the assemblages, at 75% of the total, and 91.9% of the sample illustrates the red, grey, or brown pastes typical of Joventud Red and Paso Caballos Waxy ware ceramics. Within the morphological attribute category, 90% of all ceramics have a flat base, and 74% lack vessel supports—although 12% exhibit pedestal supports, the most common Middle Formative period vessel support. The majority (83.8%) of ceramics have direct or everted rims, with rounded (64.7%) and flat (24.2) lips dominant. Wall and rim/ neck types do not display significant differences across sites, although vessel walls do appear slightly straighter within the assemblage at Mirador. This fact is likely due to a relatively larger sample from Mirador, owing to the slightly greater quantity of Middle Formative period ceramics at that site. The dimensional attributes reveal primarily shorter vessels with no great elaboration and seemingly standardized dimensions. No significant variation in vessel height is evident, and vessel walls are relatively thicker when compared with ceramics in subsequent temporal contexts. On the small number of vessels with necks (12%), the overwhelming majority are short–necked.

Statistical Analyses of the Ceramic Sample To recapitulate, two statistical measures of the recorded variables in the sampled ceramic materials were employed. The raw data had been previously recorded, specific measurements entered into an Excel spreadsheet, and attribute values assigned to the data. A ten percent stratified random sample of the ceramic materials from four sites in the San Pedro Mártir basin was then selected. Using PASW(SPSS) 18 statistics software, cluster analysis of the distributional patterns of individual analytic variables was then performed, based on analyses of variance (ANOVA) of mean attribute similarity and distance, derived from the attribute values assigned to the data. At the same time, a Tukey post–hoc test of honestly significant differences (HSD) was also undertaken. These analyses allowed for the determination of degrees of similarity and difference across and between specific ceramic attributes, and suites of attributes, among sites within the sample in both synchronic and diachronic terms. A series of tables, cross–tabulation charts, and clustered boxplot graphs illustrates the results of these analyses. A relative H Score heterogeneity measure was also calculated, weighted through Brainerd–Robinson coefficients of similarity, and scaled to one (see Garraty 2009). This statistical measure, calculated using both PASW 18 software and Excel, complements the Tukey HSD test by illustrating the degree of heterogeneity within the sample as a whole during the three chronological periods under consideration. The results of this analysis are presented as a single table. Finally, the results of these statistical analyses were compared with the distributions and proportions of those attribute categories observed in neighboring sites and regional ceramic sequences. The relationship between regional interaction, innovation, and stylistic variability, as reflected in these statistical and comparative evaluations, was interpreted according to the precepts previously outlined (see Figure 4.18).

Joventud Red is also the only type–variety encountered at San Claudio during the Middle Formative period, which likely explains the slightly higher significance of difference (p = 0.054) within the sample along this variable. 2 These percentages derive from descriptive cross–tabulated statistics produced through ANOVA analysis. Each set of descriptive statistics for each chronological period is not reproduced here in entirety, although selected cross–tabulations are presented and discussed below. 1

76

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Sum of Squares Ware

TV

Slip

Surface treatment Paste texture

Paste color

Temper content

Decoration type

Form/shape

Base type

Support type

Wall type

Rim/neck type

Lip type

Angles/flanges

Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total

0.129 6.376 6.505 12.668 152.019 164.687 6.028 178.719 184.747 4.024 61.148 65.172 62.882 761.805 824.687 12.374 176.737 189.111 0.481 36.933 37.414 .905 25.600 26.505 42.626 524.000 566.626 1.987 180.457 182.444 38.947 545.376 584.323 6.004 171.632 177.636 10.418 467.905 478.323 7.254 233.251 240.505 0.000 0.000 0.000

df 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98 3 95 98

Mean Square

F

Sig.

0.043 0.067

0.640

0.591

4.223 1.600

2.639

0.054

2.009 1.881

1.068

0.366

1.341 .644

2.084

0.107

20.961 8.019

2.614

0.056

4.125 1.860

2.217

0.091

0.160 0.389

0.412

0.745

0.302 0.269

1.120

0.345

14.209 5.516

2.576

0.058

0.662 1.900

0.349

0.790

12.982 5.741

2.261

0.086

2.001 1.807

1.108

0.350

3.473 4.925

0.705

0.551

2.418 2.455

0.985

0.403

0.000 0.000

.

.

Figure 5.1. Results of ANOVA statistical analysis on Middle Formative period ceramic sample. (1)

There are some slight statistically significant differences in ceramic forms between sites (Figures 5.4 and 5.5).3

These differences, however, are likely only the product of the extremely low number of Middle Formative period ceramics encountered at San Claudio (n1 = 8). Nonetheless, the same ceramic forms are generally encountered at the other three sites in roughly the same relative

The numerical site numbering scheme used in Figure 5.5 (and all figures throughout this chapter, unless otherwise indicated), is as follows: 1 = San Claudio; 2 = Tiradero; 3 = Mirador; 4 = Revancha.

3

77

Archaeological Paleography

Diameter

Wall thickness

Height

Neck length

Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total

Sum of Squares 244.587 7631.918 7876.505 0.787 31.233 32.020 1.844 45.994 47.838 2.166 92.743 94.909

df 3 95 98 3 95 98 3 95 98 3 95 98

Mean Square 81.529 80.336

F

Sig.

1.015

0.390

0.262 0.329

0.798

0.498

0.615 0.484

1.270

0.289

0.722 0.976

0.740

0.531

Figure 5.1. Results of ANOVA statistical analysis on Middle Formative period ceramic sample. (2) Frequency Pital Cream:VU Pital Cream:pital Tierra Mojada:VU Achiotes:VU Achiotes:achiotes Joventud Red:VU Total

3 5 2 4 3 82 99

Percent 3.0 5.1 2.0 4.0 3.0 82.8 100.0

Figure 5.3. Frequencies of Middle Formative period type– varieties within the sample.

tables displaying the results of Tukey HSD tests for any other measured variable are not included here. Although these analyses were performed, their inclusion would be redundant: since no variability was detected across attributes within the sample, no variation between sites is evident. The ANOVA analysis indicates that ranges of variability and mean attribute values are virtually identical across all attribute categories and between all sites. Such uniformity in the ranges and means of attribute values at each site suggests a general lack of variability in the Middle Formative period ceramic sample across all attribute categories. What little variation is evident is likely again due to the relatively small sample of Middle Formative period ceramics from San Claudio.

Figure 5.2. Pie chart illustrating occurrences of specific type– varieties within the Middle Formative period ceramic sample.

ANOVA Analysis of Late Formative Period Ceramics

percentages and frequencies (Figure 5.6).4 Because no other discernible differences were detected through the ANOVA analysis in any additional attribute category,

ANOVA analysis revealed little dissimilarity between ceramic materials among sites within the sample during the Late Formative period (Figure 5.7). No statistically significant differences are evident across any measured attribute within the sample. There is, however, a greater range of variability in the Late Formative period ceramic attributes, due in large part to the relatively greater quantity of Late Formative ceramic materials. That is, more distinct attribute types are evident within each particular attribute category. For example, within

Figure 5.6 and other percentage bar graphs throughout this chapter express percentages of attributes observed in the assemblages of specific sites relative to the sample as a whole. Site–specific percentages (i.e., proportions of specific attributes relative to the assemblage at a specific site) can be appreciated through the cross–tabulation tables accompanying selected figures. The advantage of presenting the data in this way is that one can gauge proportions both relative to the sample as a whole as well as within the assemblages at each site, both in the figures and by comparing the figures to the cross–tabulations.

4

78

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Site San Claudio

Form/ shape

Tiradero

Mirador

Total

Revancha

plate

4

6

8

2

20 (20.2%)

cajete

4

8

12

3

27 (27.3%)

apaxtle

0

5

5

4

14 (14.1%)

basin

0

1

3

3

7 (7.1%)

cazuela

0

7

6

1

14 (14.1%)

olla

0

8

8

1

17 (17.2%)

8

35

42

14

99

Total

Figure 5.4. Cross tabulation of form/shape by site, Middle Formative period.

Mean Difference (I-J) 1 2 -2.500* 3 -2.143 4 -1.714 2 1 2.500* 3 0.357 4 0.786 Form/shape 3 1 2.143 2 -0.357 4 0.429 4 1 1.714 2 -0.786 3 -0.429 * The mean difference is significant at the 0.05 level. Dependent Variable

(I) Site*

(J) Site

Std. Error 0.920 0.906 1.041 0.920 0.538 0.743 0.906 0.538 0.725 1.041 0.743 0.725

Sig.* 0.039 0.091 0.358 0.039 0.910 0.716 0.091 0.910 0.934 0.358 0.716 0.934

95% Confidence Interval Lower Bound Upper Bound -4.91 -4.51 -4.44 .09 -1.05 -1.16 -.23 -1.76 -1.47 -1.01 -2.73 -2.32

-0.09 0.23 1.01 4.91 1.76 2.73 4.51 1.05 2.32 4.44 1.16 1.47

Figure 5.5. Tukey HSD test for between site variability in Middle Formative period formal attributes.

Figure 5.6. Percentage of occurrence of specific form or shape class within the Middle Formative period assemblages.

79

Archaeological Paleography

Sum of Squares Ware

Between Groups Within Groups Total TV Between Groups Within Groups Total Slip Between Groups Within Groups Total Surface treatment Between Groups Within Groups Total Paste texture Between Groups Within Groups Total Paste color Between Groups Within Groups Total Temper content Between Groups Within Groups Total Decoration type Between Groups Within Groups Total Form/shape Between Groups Within Groups Total Base type Between Groups Within Groups Total Support type Between Groups Within Groups Total Wall type Between Groups Within Groups Total Rim/neck type Between Groups Within Groups Total Lip type Between Groups Within Groups Total Angles/flanges Between Groups Within Groups Total

4.289 220.396 224.685 45.844 5035.214 5081.058 9.221 2249.133 2258.354 3.657 604.208 607.864 35.360 4039.081 4074.441 3.966 1750.388 1754.354 0.740 481.015 481.756 17.705 4587.869 4605.575 72.392 5945.818 6018.210 2.579 1362.040 1364.619 19.754 3084.450 3104.203 3.811 1031.120 1034.931 16.780 2619.159 2635.939 3.926 3418.631 3422.558 7.324 1141.498 1148.822

df 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589 3 586 589

Mean Square

F

Sig.

1.430 0.376

3.801

0.010

15.281 8.593

1.778

0.150

3.074 3.838

0.801

0.494

1.219 1.031

1.182

0.316

11.787 6.893

1.710

0.164

1.322 2.987

0.443

0.723

0.247 0.821

0.301

0.825

5.902 7.829

0.754

0.520

24.131 10.146

2.378

0.069

0.860 2.324

0.370

0.775

6.585 5.264

1.251

0.290

1.270 1.760

0.722

0.539

5.593 4.470

1.251

0.290

1.309 5.834

0.224

0.879

2.441 1.948

1.253

0.290

Figure 5.7. Results of ANOVA statistical analysis on Late Formative period ceramic sample. (1)

the Late Formative period sample, 15 type–varieties from eight ceramic groups and three wares are evident, compared with the six type–varieties from four ceramic groups and two wares represented in the Middle Formative period sample. Slight differences in certain attributes between sites are apparent, particularly in

terms of classificatory and formal attributes. As a whole, the data indicate a continuation of the similarity in ceramic materials across all measured parameters that marked the Middle Formative period, as well as relative uniformity in the distribution of ceramic attributes across sites. 80

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Sum of Squares Diameter

Wall thickness

Height

Neck length

Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total

df

475.086 57104.209 57579.295 0.892 124.830 125.722 4.871 446.120 450.992 8.245 836.130 844.375

Mean Square

3 586 589 3 586 589 3 586 589 3 586 589

F

Sig.

158.362 97.447

1.625

0.182

0.297 0.213

1.395

0.243

1.624 0.761

2.133

0.095

2.748 1.427

1.926

0.124

Figure 5.7. Results of ANOVA statistical analysis on Late Formative period ceramic sample. (2) Site

Ware Total

Paso Caballos Waxy Uaxactún Unslipped Flores Waxy

San Claudio 105 17 0 122

Tiradero 153 17 23 193

Mirador 109 23 17 149

Revancha 100 15 11 126

Total 467 72 51 590

Figure 5.8. Cross tabulation of ware by site, Late Formative period.

Within the classificatory attributes, Paso Caballos Waxy is again the dominant ceramic ware, at 79.2% of the total, occurring in roughly the same relative proportions at all sites (Figure 5.8; and relative to the sample as a whole, Figure 5.9). A small quantity of Flores Waxy ware also appears at the three northern sites, but not at San Claudio. The five most common Late Formative type–varieties are uniform at each of the sites and occur in roughly the same proportions (cf. Figure 4.15). Sierra

Red is the dominant type–variety, at 49.3% of the total sample, although greater diversity in specific types and varieties is evident—and certain type–varieties only occur at particular sites (Figure 5.10). Consequently, Tukey analysis revealed statistically significant differences between San Claudio and both Tiradero and Mirador (Figure 5.11). Nevertheless, as with the Middle Formative period ceramics, since the ANOVA analysis revealed no significant variation along any

Figure 5.9. Bar graph illustrating site specific percentages of ceramic wares within the Late Formative period sample.

81

Archaeological Paleography

Figure 5.10. Pie chart and bar graph illustrating general breakdown and site specific percentages of individual type–varieties within the Late Formative period ceramic sample.

Dependent Variable

(I) Site 1

(J) Site

2 3 4 2 1 3 4 Ware 3 1 2 4 4 1 2 3 *The mean difference is significant at the 0.05 level.

Mean Difference (I-J) -0.187* -0.243* -0.154 0.187* -0.056 0.033 0.243* 0.056 0.089 0.154 -0.033 -0.089

Std. Error 0.071 0.075 0.078 0.071 0.067 0.070 0.075 0.067 0.074 0.078 0.070 0.074

Sig.* 0.043 0.007 0.196 0.043 0.836 0.966 0.007 0.836 0.629 0.196 0.966 0.629

95% Confidence Interval Lower Bound Upper Bound -0.37 -0.44 -0.35 0.00 -0.23 -0.15 0.05 -0.12 -0.10 -0.05 -0.21 -0.28

0.00 -0.05 0.05 0.37 0.12 0.21 0.44 0.23 0.28 0.35 0.15 0.10

Figure 5.11. Tukey HSD test for between site variability in Late Formative period ceramic wares.

82

Interpreting the Results of the Comparative and Statistical Ceramic Analyses attribute parameter in the sample as a whole, any other Tukey results are not prsented here. In any case, the post–hoc Tukey analyses illustrated no further statistically significant between–site variation in additional ceramic attributes. Despite the slight variation between San Claudio, Tiradero, and Mirador in terms of ware and type–variety, ranges of variability and mean attribute values across all categories are essentially uniform. Owing to the prevalence of Sierra Red in the Late Formative period sample, red slips dominate the assemblage at 52%, and occur in roughly the same percentage and frequency across sites (Figure 5.12). Cream (18.3%) and black (10.0%) slips are also evident, due to the presence of the Flor Cream Figure 5.12. Percentage of occurrence of specific slip within the Late Formative period assemblages. type–variety (16.8%). These results confirm the regional sequence of the San Pedro Mártir basin outlined in Chapter 4 (see also Hernández Ayala 1981), and also reflect the dominance of monochrome red, black, and cream ceramics with thick, waxy slips in the assemblages and sequences of adjacent sites and regions, such as Abal phase Piedras Negras, the late waxy horizon at Palenque, and throughout the lower and middle Usumacinta basin (see Adams 1971; Fournier 1983; Hernández Pons 1984; Holley 1987; Muñoz 2002, 2004; Ochoa and Casasola 1991; Pérez Robles 2006; Rands 1972, 1987). Although the majority (68.3%) of Late Formative period ceramics in the sample was undecorated, incising (8.5%) and fluting (16.6%) are the most evident decorative techniques, as elsewhere (Figure 5.13). The predominance Figure 5.13. Percentage of occurrence of specific decoration types within the Late Formative period assemblages. of Sierra Red and Paso Caballos Waxy ceramics also account for the dominance of waxy (56.1%) and smooth (20.2%) surface of attribute variability within the assemblages at each treatments, common at all sites, although 21.9% of the site. Although based on numerical means derived from sample exhibited rough surface treatment, likely due to the assigned nominal attribute values rather than the the significant amount of Achiotes Unslipped ceramics attributes themselves, ANOVA nonetheless allows for across assemblages (Figure 5.14). the identification of ranges of variation and statistically significant differences between the means of both Another way of visualizing the uniformity or lack of individual variables and suites of attributes. As Figure variability between specific attributes—or suites of 5.7 illustrates, ANOVA and chi–square means analysis attributes—in the Late Formative period ceramic sample revealed no statistically significant variation in the mean is through a clustered boxplot that displays the ranges attribute values of any ceramic attribute between the 83

Archaeological Paleography

Figure 5.14. Percentage of occurrence of surface treatments within the Late Formative period assemblages.

and other similar boxplot graphs throughout this chapter, illustrates simply the range of distinct attribute values observed and recorded within the ceramic materials from each site. No relationship between specific attribute values or their means within the larger classes is implied or should be inferred from these charts. The wider, pattern–coded ranges illustrate the 1–sigma range of standard deviation from the mean individual attribute value, thus displaying the most common range of attributes present in the assemblage at each site. The T–bar whiskers illustrate the full range of variable attribute values encountered in discrete assemblages. Outliers that do not fall within the 2–sigma range of standard deviation are expressed as points or stars outside of either extreme of the T–bar whiskers, with the individual sample numbers beside them. A similar range of variability across individual attributes implies little variation in the standard values of ceramic attributes between sites; that is, common ceramic characteristics are uniformly encountered at each site. Thus, Figure 5.15 illustrates essentially uniform ranges for values across all variable categories within the suite of classificatory attributes, suggesting little between– site variation the Late Formative period ceramics within the sample.

In terms of shape classes, the most common forms in the sample are wide, shallow serving dishes, such as plates, apaxtles, and cajetes with thickened, slightly everted rims, globular ollas, and thick– Figure 5.15. Clustered boxplot displaying ranges of classificatory attribute variability walled jars with short, out– during the Late Formative period. curving necks (Figures 5.16 and 5.17). Such forms are common during the Late Formative period throughout assemblages from each site. Figure 5.15 displays the the lower Usumacinta basin and the Maya lowlands in ranges of attribute variability for classificatory attributes general. Vessels with similar forms are also evident in within the ceramic materials at each site.5 This figure, the Guanacaste and Horcones phases at Chiapa de Corzo and the Middle Grijalva Guañoma phase (Adams 1971; 5 In Figure 5.15 and other clustered boxplots illustrating ranges of Forsyth 1989; Holley 1987; Lee 1972; Muñoz 2004; Pérez variability within specific attribute suites, no relationship between the individual values—or the ANOVA–derived means—of specific attributes within the larger attribute class is implied. These nominal values were assigned arbitrarily and do not relate to each other across attributes. These figures only display relative similarity in the ranges of variability of individual attributes within classes across sites. In

the case of the statistical analysis of the Early Classic period sample presented below, the clustered boxplots conversely illustrate the comparative diversity in ranges of variability of individual attributes within classes across sites.

84

Interpreting the Results of the Comparative and Statistical Ceramic Analyses Site San Claudio

Form/shape

plate cajete apaxtle basin cazuela cup jar olla tecomate bowl unidentified

Total

24 29 8 5 7 2 17 11 1 10 8 122

Tiradero 19 48 20 12 18 4 24 24 4 15 5 193

Mirador 14 25 13 8 12 2 33 16 4 13 9 149

Revancha 12 28 9 10 7 3 18 12 6 15 6 126

Total 69 (11.7%) 130 (22%) 50 (8.5%) 35 (5.9%) 44 (7.5%) 11 (1.9%) 92 (15.6%) 63 (10.7%) 15 (2.5%) 53 (9.0%) 28 (4.7%) 590

Figure 5.16. Cross tabulation of form/shape by site, Late Formative period.

Robles 2006; Rands 1972, 1987; Smith 1955; Smith and Gifford 1966). There does appear to be some slight variation between San Claudio and the three sites to the north, with utility ware such as basins, jars, cazuelas, and ollas relatively more common at Tiradero, Mirador, and Revancha. Nevertheless, post–hoc Tukey analysis revealed no statistically significant between–site variability in terms of shape class or formal attributes. Body angles and flanges on vessels remain relatively uncommon throughout the Late Formative period sample, although basal angles (6.9%) and medial flanges (6.3%) are evident. Mean attribute values were relatively identical across Figure 5.17. Bar graph illustrating site specific percentages of ceramic forms and sites. The clustered boxplot graph shapes within the Late Formative period ceramic sample. indicates essentially uniform attribute value ranges among all between sites. Red (62.7%) and brown (23.7%) pastes ceramic materials within the sample (Figure 5.18).6 The predominate, primarily of medium texture (65.6%), minor variability evident in the mean attribute value for which is characteristic of Paso Caballos Waxy ware form at Mirador and, to a lesser extent, Revancha, is likely ceramics, and reflects the prevalence of the Sierra Red due to the relatively lower percentage of serving ware— type–variety during the Late Formative period. Calcite and higher percentage of utility ware—at these sites. (37.1%) and sand (25.9%) tempers remain common, For the remaining attribute categories (paste, dimensions, although limestone temper (35.3%) is also evident. and morphology), ANOVA analysis revealed little Paste color and texture values occur in roughly the variability within the Late Formative period sample, or same proportions between sites (Figures 5.19 and 5.20). The clustered boxplot (Figure 5.21) illustrates the 6 In Figure 5.18 there is no range for angles/flanges because these consistency of ranges for all paste attribute values across attributes were not encountered in statistically significant quantities on sites. The absence of statistically significant between– Late Formative period ceramics from any site. Their near total absence site variability also indicates a uniform distribution of at all sites nonetheless again illustrates the absence of significant variability between assemblages. In the figure, the bold line indicates these attributes across all sites. Analysis of dimensional a mean attribute value of 6, corresponding to the absence of angles or attributes illustrated identical ranges and mean attribute flanges, although the few examples that are present are indicated by values for wall thickness, height, and neck length in the outliers. 85

Archaeological Paleography all assemblages. The comparatively larger range in vessel neck length at Mirador is a function of the relatively greater percentage of vessels with necks (e.g., jars) at that site. As with the Middle Formative period ceramics, the Late Formative dimensional attributes indicate an abundance of short vessels with short necks, no great vessel or dimensional elaboration, and seemingly standardized dimensions. No significant variation in vessel height is evident. Vessel walls are slightly thinner than in the Middle Formative period, although they remain relatively thicker when compared with ceramics in subsequent temporal contexts, with 78% of vessels exhibiting walls medium to thick walls (0.7–1.3cm thick). Morphological ceramic attributes Figure 5.18. Clustered boxplot displaying ranges of formal attribute variability remain relatively evenly distributed during the Late Formative period. among the Late Formative sample, displaying no significant between– site variability. Ranges of variability in morphological attributes are uniform across the sample, with no specific attribute exhibiting statistically significant variability (Figure 5.22). As in the Middle Formative period, flat bases, at 62.7% of all materials in the sample, remain predominant in the Late Formative period, although rounded (17.6%), convex (7.6%), and concave (5.4%) bases also appear. The majority (57.8%) of vessels continues to lack supports, but cylindrical (16.7%) and hemispheric (9.0%) supports are evident, appearing primarily on utility ware. Annular ring supports are also relatively abundant, found on 10% of Late Formative period ceramics. Incurved vessel walls (51.5%) are most common, with straight walled Figure 5.19. Percentage of occurrence of specific paste colors within the Late Formative period assemblages. vessels comprising a large minority at 25.9% of the Late Formative across categories. That is, there are a greater number sample. Everted (62.2%) and direct (24.9%) rims of wares, type–varieties, slips, surface treatments, remain prevalent, displaying no significant differences forms, and morphological variables evident within the across sites. Rounded (65.6%) and flat (11.7%) lips sampled materials. Nevertheless, the ranges, means, are widespread throughout the Late Formative period and distribution of those specific attributes generally sample. remain uniform, continuing the consistency across sites In sum, during the Late Formative period, ceramic and variables exhibited in the Middle Formative period materials display increasing diversity, in terms of ceramics. What little statistically significant variability a wider number of attributes and attribute values does exist (e.g., in terms of type–varieties, wares, and 86

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.20. Percentage of occurrence of specific paste textures within the Late Formative period assemblages.

distinct wares, 13 ceramic groups, and 25 individual type–varieties are present in the Early Classic period sample. Additionally, marked between–site variability appears at this time, visible through post–hoc Tukey HSD tests. All attribute categories, with the exception of rim/neck type and diameter, display statistically significant differences (Figure 5.23). In terms of the classificatory attributes, Petén Gloss ware dominates the assemblage at San Claudio, comprising 91.2% of the total Early Classic period ceramics at that site. In contrast, Petén Gloss ware is represented in only 30.4% of the remaining sample—29.7% at Tiradero and 30.9% of the materials at both Mirador and Revancha. Instead, Uaxactún Unslipped ceramics continue to comprise the majority of the ceramic materials encountered at these sites (49.7%), whereas Uaxactún Unslipped virtually disappears Figure 5.21. Clustered boxplot displaying ranges of paste attribute at San Claudio during the Early Classic variability during the Late Formative period. (Figure 5.24). The five most common Early forms) is between the assemblage at San Claudio Classic type components are no longer uniform, and and the materials encountered at the other three sites several type–varieties are not represented at all sites. under consideration. This divide becomes much more Instead, glossy, polychrome ceramics, such as Bolonchac pronounced in the subsequent Early Classic period. and Moro Orange polychrome, come to the fore at San Claudio, while the rough, unslipped Triunfo Striated ANOVA Analysis of Early Classic Period Ceramics is the dominant type at the three northern sites (Figure 5.25). Figures 5.24 and 5.25 illustrate what appears to be In the Early Classic period, greater variability within the a clinal distribution for wares and certain type–varieties. sample across all measured attributes becomes evident Given these ware and type–variety distributions, it is through ANOVA analysis. A greater range of variability unsurprising that orange and red–orange slipped pottery and diversity in terms of attributes and attribute values is prevalent at San Claudio (57.6%), whereas unslipped across categories is also manifest.7 For example, five ceramics comprise a full 50.7% of the materials present in the combined assemblages of Tiradero, Mirador, and Revancha. In turn, surface treatments reflect similar 7 These trends are also likely due in part to the larger size of the Early variability (Figures 5.26 and 5.27), with glossy finishes Classic period sample. 87

Archaeological Paleography

Figure 5.22. Clustered boxplot displaying ranges of morphological attribute variability during the Late Formative period.

common at San Claudio and rough surfaces widespread in the assemblages to the north. Although smooth and glossy finishes are present on some materials from Tiradero, Mirador, and Revancha, variability in surface treatment across sites is evident. Similarly, decoration type is variable, with polychrome decoration dominant in the ceramics at San Claudio, whereas striated and incised decoration is evident at the other sites (Figures 5.28 and 5.29).

greater amount of statistically significant variation is evident in the Early Classic period ceramic sample along classificatory attribute parameters within the sample as a whole. Additionally, distinct clusters now appear, illustrating a greater degree of between–site variability. Certain attributes suggest a clinal distribution, and it is noteworthy that, although distinct from the ceramic of San Claudio, the materials from Tiradero, Mirador, and Revancha display no statistically significant differences along classificatory parameters between the three northern sites.

The clustered boxplot of classificatory attribute ranges (Figure 5.30) illustrates the increased diversity and range of variability in the Early Classic period ceramic materials. Attribute values now show a clear distinction between the ceramics from San Claudio and those present at the sites to the north of the study area, suggesting the existence of at least two separate clusters of ceramic attributes throughout the study area. Tukey post–hoc analysis also illustrates a great degree of between–site variability in ceramics during the Early Classic period (Figure 5.31). Breaking down the Tukey test even further, the means for groups in homogeneous subsets indicate the existence of discrete clusters of ceramic attributes, one centered on the assemblage at San Claudio, and the other comprised of the materials encountered at Tiradero, Mirador, and Revancha (Figures 5.32–5.34).8 Thus, a

The cross–tabulation of temper content (Figure 5.35) reveals that calcite tempers remain common, but also that more types of temper are evident in the sample as a whole. Calcite tempers dominate the assemblage at San Claudio, whereas sand tempers are most common in the materials from other sites. Shell tempers are also conspicuously lacking at San Claudio. Paste attributes demonstrate similarity variability during the Early Classic period, with red, brown, and pink pastes common at San Claudio, and grey pastes dominant at the other sites (Figure 5.36). Paste texture is generally thicker at San Claudio, whereas pastes at Tiradero, Mirador, and Revancha are fine to medium, and often included (Figure unequal (i.e., there are discrepancies in the gross number of ceramic materials present at each site) the PASW program used a harmonic mean sample size of 355.111.

These and following tables derived from Tukey analyses display means for groups in homogeneous subsets. Since group sizes are

8

88

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Sum of Squares Ware

TV

Slip

Surface treatment

Paste texture

Paste color

Temper content

Decoration type

Form/shape

Base type

Support type

Wall type

Rim/neck type

Lip type

Angles/flanges

Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total Between Groups Within Groups Total

230.816 2206.251 2437.067 18136.237 28372.715 46508.952 1959.544 9972.555 11932.100 61.113 1670.350 1731.463 349.654 10926.810 11276.464 179.508 7243.467 7422.975 74.224 2415.121 2489.346 1140.430 8438.385 9578.815 1244.141 12707.396 13951.537 111.349 3514.919 3626.268 661.381 8510.680 9172.061 36.877 2665.918 2702.795 21.818 9587.828 9609.646 214.255 14517.166 14731.421 136.974 3399.317 3536.291

df 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564

Mean Square

F

76.939 1.413

54.437

0.000

6045.412 18.176

332.604

0.000

653.181 6.389

102.242

0.000

20.371 1.070

19.037

0.000

116.551 7.000

16.650

0.000

59.836 4.640

12.895

0.000

24.741 1.547

15.991

0.000

380.143 5.406

70.322

0.000

414.714 8.141

50.944

0.000

37.116 2.252

16.484

0.000

220.460 5.452

40.436

0.000

12.292 1.708

7.198

0.000

7.273 6.142

1.184

0.314

71.418 9.300

7.679

0.000

45.658 2.178

20.967

0.000

Figure 5.23. Results of ANOVA statistical analysis on Early Classic period ceramic sample. (1)

89

Sig.*

Archaeological Paleography

Sum of Squares Between Groups Within Groups Total Between Groups Wall thickness Within Groups Total Between Groups Height Within Groups Total Between Groups Neck length Within Groups Total * differences are significant at the 0.05 level

Diameter

1175.636 326617.376 327793.012 120.096 1015.127 1135.223 163.687 1110.860 1274.547 66.157 2068.733 2134.891

df 3 1561 1564 3 1561 1564 3 1561 1564 3 1561 1564

Mean Square

F

Sig.*

391.879 209.236

1.873

0.132

40.032 0.650

61.559

0.000

54.562 0.712

76.672

0.000

22.052 1.325

16.640

0.000

Figure 5.23. Results of ANOVA statistical analysis on Early Classic period ceramic sample. (2)

5.37). The clustered boxplot graph for paste attributes (Figure 5.38) does not indicate as much variability in paste and temper as in the classificatory attributes— the ranges are essentially identical, but the mean attribute values are distinct across sites. The Tukey tests, on the other hand, show statistically significant between–site variation along these paste parameters, and again suggest the existence of two discrete clusters of ceramic attributes (Figures 5.39–5.41). Formal attributes display similar statistical and distributional variability (Figures 5.42 and 5.43).9 At Figure 5.24. Bar graph illustrating site specific percentages of ceramic wares within the Early Classic period sample. San Claudio, serving dishes, particularly shallow cajetes (39.1% of the site total), are widespread throughout the sites. These discrepancies in form and shape class may assemblage. These forms are relatively less evident at stem from issues of sampling or preservation. Due to the Mirador, Revancha, and Tiradero, where utility wares, lack of extensive, systematic excavation at any of the especially ollas, jars, and basins are more common. sites under consideration, it is possible that certain forms Deep bowls are also abundant at the three northern were more likely to be encountered in particular contexts. Given the fact that relatively more survey and excavation 9 Curiously, a small percentage of tecomate forms were identified in the work at house mounds and ostensibly residential areas was Early Classic period sample at Tiradero, Mirador, and Revancha. This undertaken at Tiradero, Mirador, and Revancha than at San fact is odd given the relative rarity of this ostensibly Early and Middle Claudio, it is unsurprising that a larger number of utilitarian Formative shape class in other Early Classic period contexts. I believe forms are evident at these sites. Nevertheless, ANOVA that their presence is likely the result of vessel re–use, particularly given their appearance primarily in construction fill. Alternately, these analysis revealed marked differences in the mean attribute forms could simply be misidentified ollas, or the result of an imprecise values of form and shape classes between sites within stratigraphic sequence during the initial excavations.

90

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.25. Bar graph illustrating general breakdown and site specific percentages of individual type–varieties within the Early Classic period ceramic sample. Site San Claudio Surface treatment Total

Tiradero

Mirador

Total

Revancha

Smooth

29

72

59

30

190 (12.1%)

Waxy

36

80

48

39

203 (13%)

Glossy

218

69

49

38

374 (23.9%)

Rough

14

374

251

159

798 (51%)

297

595

407

266

1565

Figure 5.26. Cross tabulation of surface treatment by site, Early Classic period.

Figure 5.27. Percentage of occurrence of specific surface treatments within the Early Classic period assemblages.

91

Archaeological Paleography Site San Claudio

Decoration type

Total

Incised Fluted Excavated Striated Polychrome None Multiple Punctated

20 12 0 10 139 78 38 0 297

Tiradero 118 8 10 306 45 86 18 4 595

Mirador 95 4 12 202 25 54 11 4 407

Revancha 69 4 11 114 21 38 4 5 266

Total 302 (19.3%) 28 (1.8%) 33 (2.1%) 632 (40.4%) 230 (14.7%) 256 (16.7%) 71 (4.5%) 13 (0.8%) 1565

Figure 5.28. Cross tabulation of decoration type by site, Early Classic period.

The morphological attributes present on Early Classic period ceramics within the sample continue the trend of increasing diversity, as well as significant statistical and distributional variability. Attributes within this class display slightly less statistical and between–site variation. For example, analysis of rim and neck types revealed no statistically significant differentiation within the sample (Figure 5.23), although a greater range of rim and neck types is evident at Tiradero, Mirador, and Revancha, as illustrated in the clustered boxplot graph Figure 5.29. Percentage of occurrence of specific decoration types within the Early of morphological attribute Classic period assemblages. variability ranges (Figure 5.46). Direct (32.3%) the sample. As with pastes, the ranges of formal attribute and everted (29.6%) rims appear most frequently values are essentially identical, since examples of all shape throughout the sample, although bolstered (18.1%) classes were found at each site, despite the fact that some and flared (6.39%) rims are relatively common as well. forms are more common at a certain site, or group of sites. The same general lack of variability is also evident in Nonetheless, mean attribute values do differ across sites, the attribute of lip type. As in previous chronological indicating increased variability along these parameters. periods, rounded lips remain most abundant throughout the sample, comprising 52.3% of the total. Once again, Body angles and flanges on vessels remain relatively a greater range of lip types is evident at the northern uncommon during the Early Classic period, appearing sites. Rounded lips are definitely more abundant at on only 18.5% of the sampled ceramic materials. These San Claudio, evident on 64.6% of ceramics at that site, formal attributes are decidedly more common at San than at the other sites, where flared (16.6%) and T– Claudio, where basal (14.1%) and Z–angles (11.8%) shaped (7.97%) lips comprise a significant minority, are comparatively frequent. In contrast, body angles and in contrast to San Claudio, where these lip types are flanges occur on only 14.7% of all ceramics from Tiradero, noticeably less evident. As Figure 5.46 illustrates, Mirador, and Revancha. The Tukey HSD analyses on ranges for lip types at Tiradero, Mirador, and Revancha formal attributes illustrated significant between–site are greater, although mean values differ only slightly variability along formal attribute parameters, and the throughout the sample. Differentiation in wall types presence of two distinct clusters within the Early Classic is similar. Again, ranges within this attribute category period sample (Figures 5.44 and 5.45). are essentially uniform, and a slightly greater range 92

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.30. Clustered boxplot displaying ranges of classificatory attribute variability during the Early Classic period.

although this again may be simply a function of the kinds of ceramic forms that are most prevalent at those sites. Support types more clearly demonstrate significant variability within the sample. Vessel supports are relatively more abundant during the early Classic period, occurring on almost 40% of materials within the sample. The assemblage at San Claudio demonstrates a greater range of variability in support type, and mean attribute values for this category differ significantly across sites. Vessel supports are much abundant at San Claudio, evident on 67% of the Early Classic period materials from that site, whereas only 31.8% of the combined total of ceramics from the other sites displays supports. Annular ring supports remain most common, occurring on 47.9% of all vessels with supports and an average of 18.9% of ceramics at all sites. Rounded conical supports are decidedly more pronounced at San Claudio, evident as the most frequent support type at that site on 27.9% of all vessels with supports from San Claudio. Conical supports occur on only 4.57% of vessels with supports from Tiradero, Mirador, and Revancha. Variation in support types among sites within the sample is reflected in Figure 5.46 by the greater range of attribute values at San Claudio.

of types present in the sample demonstrates increased diversity compared to the preceding Late Formative period. Mean attribute values for wall types do differ very slightly across sites. Incurved walls appear more frequently (on 44.2% of the sampled materials) at Tiradero, Mirador, and Revancha—perhaps due to the prevalence of specific utilitarian vessel forms on which this wall type is most common (e.g., ollas and jars) at these sites—but vessels with incurved walls are also relatively abundant at San Claudio, appearing on 38% of vessels at that site. Vertical walls are more common at the sites to the north, evident on 20.3% of ceramic materials at those sites, and only 7.74% of vessels from San Claudio. Straight divergent walls, on the other hand, are relatively more abundant at San Claudio (31.6%) when compared to the other sites (16.9% of the combined total). Base types also illustrate greater range and minor variability in mean attribute values. Flat based vessels remain most common at all sites, although flat bases occur relatively more frequently at San Claudio, on 74.1% of ceramics, than at the other sites, where flat bases are evident on 54.1% of the combined materials form Tiradero, Mirador, and Revancha. At these northern sites, convex (13.9%), rounded (9.93%), and reinforced (9.07%) bases are relatively more common when compared to the assemblage from San Claudio,

The Tukey analyses illustrate statistically significant variability between sites and across attribute categories within the morphological class, with the exception of rim 93

Archaeological Paleography

Dependant Variable

(I) Site

(J) Site

Mean Difference (I-J)

1

2 0.991* 3 0.984* 4 0.942* 2 1 -0.991* 3 -0.008 4 -0.049 Ware 3 1 -0.984* 2 0.008 4 -0.042 4 1 -0.942* 2 0.049 3 0.042 1 2 -8.397* 3 -8.968* 4 -8.783* 2 1 8.397* 3 -0.571 4 -0.386 TV 3 1 8.968* 2 0.571 4 0.186 4 1 8.783* 2 0.386 3 -0.186 1 2 -2.852* 3 -2.958* 4 -2.639* 2 1 2.852* 3 -0.106 4 0.213 Slip 3 1 2.958* 2 0.106 4 0.319 4 1 2.639* 2 -0.213 3 -0.319 * The mean difference is significant at the 0.05 level.

Std. Error 0.084 0.091 0.100 0.084 0.076 0.088 0.091 0.076 0.094 0.100 0.088 0.094 0.303 0.325 0.360 0.303 0.274 0.314 0.325 0.274 0.336 0.360 0.314 0.336 0.180 0.193 0.213 0.180 0.163 0.186 0.193 0.163 0.199 0.213 0.186 0.199

Sig.* 0.000 0.000 0.000 0.000 1.000 0.943 0.000 1.000 0.970 0.000 0.943 0.970 0.000 0.000 0.000 0.000 0.159 0.610 0.000 0.159 0.946 0.000 0.610 0.946 0.000 0.000 0.000 0.000 0.915 0.663 0.000 0.915 0.378 0.000 0.663 0.378

95% Confidence Interval Lower Bound Upper Bound 0.77 1.21 0.75 1.22 0.68 1.20 -1.21 -0.77 -0.20 0.19 -0.27 0.18 -1.22 -0.75 -0.19 0.20 -0.28 0.20 -1.20 -0.68 -0.18 0.27 -0.20 0.28 -9.18 -7.62 -9.81 -8.13 -9.71 -7.86 7.62 9.18 -1.28 0.13 -1.19 0.42 8.13 9.81 -0.13 1.28 -0.68 1.05 7.86 9.71 -0.42 1.19 -1.05 0.68 -3.31 -2.39 -3.45 -2.46 -3.19 -2.09 2.39 3.31 -0.52 0.31 -0.27 0.69 2.46 3.45 -0.31 0.52 -0.19 0.83 2.09 3.19 -0.69 0.27 -0.83 0.19

Figure 5.31. Tukey HSD test for between site variability in Early Classic period classificatory attributes. (1)

and neck type (Figure 5.47).10 Once again, statistically significant variation is most pronounced between San Claudio and the three northern sites; differences between the assemblages at Tiradero, Mirador, and Revancha are often not statistically significant. As with the classificatory, paste, and formal attribute classes, the extrapolated Tukey data for the morphological attributes of base, support, and lip type demonstrate the existence of two discrete clusters (Figures 5.48 and 5.49).

Extrapolated Tukey analysis does not suggest clustering around rim and neck type (Figure 5.50), which is unsurprising given the lack of within–sample variability along this parameter in general. The very slight variation in wall types between sites intimates the existence of three slightly overlapping clusters. This fact is likely due to the small degree of between–site variation along the parameter of wall type, statistically significant in only a few cases. Alternately, it may be a function of variation in vessel forms among sites.

10 Because ANOVA analysis revealed no statistically significant variability in rim and neck types, the results of the post–hoc Tukey test for this attribute category are not included in Figure 5.47.

Finally, the ANOVA analysis revealed considerable statistical and distributional variability along the 94

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Dependent Variable

(I) Site

(J) Site

Mean Difference (I-J)

1

2 -0.521* 3 -0.478* 4 -0.495* 2 1 0.521* 3 0.043 4 0.027 Surface treatment 3 1 0.478* 2 -0.043 4 -0.017 4 1 0.495* 2 -0.027 3 0.017 1 2 2.061* 3 2.245* 4 2.276* 2 1 -2.061* 3 0.184 4 0.215 Decoration type 3 1 -2.245* 2 -0.184 4 0.031 4 1 -2.276* 2 -0.215 3 -0.031 * The mean difference is significant at the 0.05 level.

Std. Error 0.073 0.079 0.087 0.073 0.067 0.076 0.079 0.067 0.082 0.087 0.076 0.082 0.165 0.177 0.196 0.165 0.150 0.171 0.177 0.150 0.183 0.196 0.171 0.183

Sig.* 0.000 0.000 0.000 0.000 0.916 0.986 0.000 0.916 0.997 0.000 0.986 0.997 0.000 0.000 0.000 0.000 0.607 0.594 0.000 0.607 0.998 0.000 0.594 0.998

95% Confidence Interval Lower Bound Upper Bound -0.71 -0.33 -0.68 -0.28 -0.72 -0.27 0.33 0.71 -0.13 0.21 -0.17 0.22 0.28 0.68 -0.21 0.13 -0.23 0.19 0.27 0.72 -0.22 0.17 -0.19 0.23 1.64 2.49 1.79 2.70 1.77 2.78 -2.49 -1.64 -0.20 0.57 -0.23 0.66 -2.70 -1.79 -0.57 0.20 -0.44 0.50 -2.78 -1.77 -0.66 0.23 -0.50 0.44

Figure 5.31. Tukey HSD test for between site variability in Early Classic period classificatory attributes. (2) Ware Site 2 3 4 1 Sig.

N 595 407 266 297

Subset for alpha = 0.05 1 2 2.77 2.77 2.82 3.76 0.946 1.000

Site 1 2 4 3 Sig.

297 595 266 407

Type–Variety Subset for alpha = 0.05 N 1 2 13.72 22.11 22.50 22.69 1.000 .281

Figure 5.32. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along classificatory parameters of ware and type–variety.

Slip Site N 1 4 2 3 Sig.

297 266 595 407

Subset for alpha = 0.05 1 2 2.94 5.58 5.79 5.89 1.000 .333

Site 1 3 4 2 Sig.

Surface treatment Subset for alpha = 0.05 N 1 2 297 2.73 407 3.21 266 3.23 595 3.25 1.000 .945

Figure 5.33. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along classificatory parameters of slip and surface treatment.

95

Archaeological Paleography parameter of dimension attributes during the Early Classic period. The thickness of vessel walls is decidedly thinner at Tiradero, Mirador, and Revancha, in the north of the study area (Figures 5.51 and 5.52). The majority (45.8%) of vessel walls in the assemblage from San Claudio are 0.7–1.0cm thick, with an additional 32.7% slightly thicker at 1.0–1.3cm thick. Vessel walls at Revancha are also predominantly 0.7–1.0cm thick, with 44.7% of the sampled materials from that site measuring to that thickness. In contrast to San Claudio, the majority of the remaining vessel walls at Revancha measure 0.4–0.7cm thick (64.6% of remaining ceramic materials, or 35.7% of the total assemblage). Although many vessel walls at Tiradero and Mirador (41.5% and 38.1%, respectively) measure within the 0.7–1.0cm range as well, the preponderance, 42.4% at Tiradero and 47.9% at Mirador, are significantly thinner, at 0.4–0.7cm thick. Moreover, only 8.28% of the vessels at Tiradero, Mirador, and Revancha have walls 1.0–1.3cm thick.

majority—77.3% of vessels with necks—are short– necked vessels that occur principally in the three northern sites of Tiradero, Mirador, and Revancha (Figure 5.54). As with the morphological attributes, the between–site variability in dimension attributes may stem from the distinct types of ceramic vessels prevalent at each site. The utility ware (e.g., jars, ollas, and basins) most common at Tiradero and Mirador is relatively larger overall than the plates and cajetes frequent in the assemblage at San Claudio. Taller vessels, as well as vessels on which necks are most common, are thus expected at those sites. The clustered boxplot of dimension attribute variability ranges illustrates similarity in ranges and uniformity in means across the assemblages at Tiradero, Mirador, and Revancha (Figure 5.55). The graph reveals that the assemblage at San Claudio has distinct ranges of attribute types and mean attribute values, again suggesting between–site variability. This inter–assemblage disparity is confirmed by the results of the Tukey HSD test (Figure 5.56), which indicates a statistically significant degree of variation between San Claudio and the other sites, which do not differ significantly between themselves. The extrapolated Tukey results (Figures 5.57 and 5.58) again suggest the existence of two separate clusters of dimension attributes during the Early Classic period, with the ceramic assemblage at San Claudio forming a unique homogeneous subset when compared to the materials encountered at Tiradero, Mirador, and Revancha.

Similarly, vessels are significantly taller at the northern sites (Figure 5.53). 47.7% of ceramic materials at Tiradero, 50.6% of those at Mirador, and 47.3% of the assemblage at Revancha measure over 18cm in height, with an additional 29.2% on average measuring 9–18cm tall. In contrast, the majority (56.9%) of vessels from San Claudio measure less than 9cm in height. Although relatively few (22.5%) of Early Classic period ceramic materials in the sample have necks, the overwhelming

Site 4 3 2 1 Sig.

266 407 595 297

The results of the ANOVA and Tukey analyses of the Early Classic period ceramic sample suggest a significantly higher degree of inter–assemblage variability and diversity across all general attributes classes and all but two individual attribute categories at this time. This statistical and distributional variation stands in stark contrast to the uniformity and homogeneity in ceramic materials observed through the ANOVA analyses of the Middle and Late Formative period samples. Not only is a greater number of specific variable attribute types evident during the Early Classic period, but the ranges and means of those specific attributes begin to vary

Decoration type Subset for alpha = 0.05 N 1 2 4.07 4.10 4.28 6.34 0.608 1.000

Figure 5.34. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along classificatory parameter of decoration type.

Site San Claudio Calcite Temper content

Total

198

Tiradero 183

Mirador 135

Total

Revancha 76

592 (37.8%)

Sand

64

279

210

136

689 (44%)

Limestone

20

38

27

14

99 (6.3%)

Shell

3

74

23

29

129 (8.2%)

Quartz

7

2

0

4

13 (0.8%)

Mica

5

19

12

7

43 (2.7%)

297

595

407

266

1565

Figure 5.35. Cross tabulation of temper content by site, Early Classic period.

96

Interpreting the Results of the Comparative and Statistical Ceramic Analyses intra–assemblage uniformity during that time. These results support the results of the ANOVA and Tukey analyses, which indicate a marked lack of variability in Middle Formative period ceramic materials within the sample. The Late Formative H Score and BR coefficient are low and high, respectively, demonstrating the same relative intra–assemblage homogeneity suggested by the ANOVA and Tukey analyses, which revealed little statistically significant variation between sites within the sample. The Late Formative H Score is slightly higher, perhaps reflecting the same minor increase in intra–assemblage diversity reflected in the greater ranges of attribute categories across sites noticeable in the clustered boxplot graphs for this period. Given the low degree of heterogeneity throughout the Middle Figure 5.36. Percentage of occurrence of specific paste colors within the and Late Formative period samples, it is Early Classic period assemblages. unsurprising that no clusters of attribute values were detected across any parameter in either of these periods. considerably across attribute classes and between sites. During the Early Classic period the H Score increases Consequently, discrete clusters of particular attribute by over 350% compared to the Late Formative period, categories emerge, suggesting a significant degree of indicating a drastic increase in intra–assemblage difference between the assemblage at San Claudio and diversity. The BR coefficient decreases dramatically at those ceramic materials present at the other three sites this time, suggesting a similar decrease in uniformity. within the sample. These results reflect the marked rise in between– site variability revealed through the Tukey post–hoc The H Score Diversity Measure analyses and extrapolated Tukey results. The increasing intra–assemblage heterogeneity during the Early Classic The outcomes of the H Score heterogeneity test of period as revealed through the H Score analysis also intra–assemblage diversity complement the results of resulted in the formation of the discrete attribute clusters the ANOVA and Tukey analyses. Because the H Score suggested by the extrapolated Tukey analyses. measure is a multivariate evaluation of diversity within the sample as a whole across all attribute categories, the Summary of Statistical Analyses results of these calculations may be neatly presented in a single table. Figure 5.59 presents the products of All statistical analyses performed on the sampled the H Score tests across the three distinct chronological ceramic materials indicate a similar pattern among sites periods under consideration.11 The results confirm the within the sample itself. The Middle Formative period general trend, observed in the ANOVA analyses through ceramics are highly uniform along all measured attribute greater ranges of attribute values, of increasing diversity parameters, displaying little variability, similar ranges over time in the ceramic sample as a whole. For the and mean attribute values, no diversity, and a lack of Middle Formative period, a relatively low H Score, and clustering. The Late Formative period ceramics follow comparatively high Brainerd–Robinson (BR) coefficient an identical pattern, although a slightly higher degree of are evident, both of which indicate a high degree of between–site variability and intra–assemblage diversity is noted. In contrast, during the Early Classic period, 11 See Garraty (2009) for further detail on the H score heterogeneity diversity and variation increase dramatically within the test and the specific calculations involved in this statistical measure. sample, and discrete clusters become evident along all A brief caveat on the Middle Formative period H score: due to the comparatively small sample size of ceramics from this period, the but a handful of attribute parameters. I will address the results are potentially problematic, since a sample of at least 20 implications of these patterns in the ceramic data for specimens from each site is desirable (Garraty 2009; less than 20 the issues of interaction and innovation shortly. First, total ceramic materials were encountered in Middle Formative period the comparison of the sampled ceramic materials with contexts at San Claudio). Therefore, although these Middle Formative period results generally confirm my initial suspicions, and agree the assemblages and sequences of neighboring sites and with my interpretations, I do not place great weight on them in the regions is briefly discussed. discussion below. 97

Archaeological Paleography

Figure 5.37. Percentage of occurrence of specific paste textures within the Early Classic period assemblages.

Figure 5.38. Clustered boxplot displaying ranges of paste attribute variability during the Early Classic period.

lowlands,12 Tierra Blanca and the Lower Usumacinta region, Uaxactún, El Mirador, Chiapa de Corzo, and the Middle Grijalva region (Acuña Smith 2005; Adams

Comparative Assessment between Assemblages at the Regional Level A diachronic comparative assessment of the sample with the ceramic materials from certain, well–documented sequences of selected neighboring sites was undertaken. These sites and regions include Altar de Sacrificios, Piedras Negras, Seibal, Palenque and the Northwest Maya

The limited data from smaller sites such as Tortuguero (Hernández Pons 1984), Panhalé (Anaya Hernández 2002), El Arenal (Vargas 1985), Pomoná (García Moll 2005; López Varela 1991), and Trinidad (Rands 1987, 2007b) is included in this general ‘northwest’ regional sequence.

12

98

Interpreting the Results of the Comparative and Statistical Ceramic Analyses Mean Difference (I-J) 1 2 -1.081* 3 -1.326* 4 -1.179* 2 1 1.081* 3 -0.246 4 -0.098 Paste texture 3 1 1.326* 2 0.246 4 0.147 4 1 1.179* 2 0.098 3 -0.147 1 2 -0.816* 3 -0.956* 4 -0.640* 2 1 .816* 3 -0.140 4 0.176 Paste color 3 1 0.956* 2 0.140 4 0.316 4 1 0.640* 2 -0.176 3 -0.316 1 2 -0.579* 3 -0.396* 4 -0.577* 2 1 0.579* 3 0.183 4 0.001 Temper content 3 1 0.396* 2 -0.183 4 -0.182 4 1 0.577* 2 -0.001 3 0.182 * The mean difference is significant at the 0.05 level. Dependent Variable

(I) Site

(J) Site

Std. Error 0.188 0.202 0.223 0.188 0.170 0.195 0.202 0.170 0.209 0.223 0.195 0.209 0.153 0.164 0.182 0.153 0.139 0.159 0.164 0.139 0.170 0.182 0.159 0.170 0.088 0.095 0.105 0.088 0.080 0.092 0.095 0.080 0.098 0.105 0.092 0.098

Sig.* 0.000 0.000 0.000 0.000 0.473 0.958 0.000 0.473 0.895 0.000 0.958 0.895 0.000 0.000 0.003 0.000 0.742 0.686 0.000 0.742 0.245 0.003 0.686 0.245 0.000 0.000 0.000 0.000 0.101 1.000 0.000 0.101 0.250 0.000 1.000 0.250

95% Confidence Interval Lower Bound Upper Bound -1.56 -1.85 -1.75 0.60 -0.68 -0.60 0.81 -0.19 -0.39 0.60 -0.40 -0.68 -1.21 -1.38 -1.11 0.42 -0.50 -0.23 0.53 -0.22 -0.12 0.17 -0.58 -0.75 -0.81 -0.64 -0.85 0.35 -0.02 -0.23 0.15 -0.39 -0.43 0.31 -0.24 -0.07

-0.60 -0.81 -0.60 1.56 0.19 0.40 1.85 0.68 0.68 1.75 0.60 0.39 -0.42 -0.53 -0.17 1.21 0.22 0.58 1.38 0.50 0.75 1.11 0.23 0.12 -0.35 -0.15 -0.31 0.81 0.39 0.24 0.64 0.02 0.07 0.85 0.23 0.43

Figure 5.39. Tukey HSD test for between site variability in Early Classic period paste attributes.

1971; Forsyth 1989; Hernández Pons 1984; Holley 1987; Lee 1972; Muñoz 2002, 2004; Ochoa and Casasola 1991; Ochoa and Hernández 1975; Pérez Robles 2006; Rands 1972, 1987, 2007a, 2007b; Sabloff 1975; Smith 1955; Vargas and Hernández 1979). Comparisons with other sites not listed above are included, where applicable. This basic comparative analysis allowed for the evaluation of variability at the regional level. Due to the fact that the absence of available raw data precluded an in–depth statistical comparison, these results and observations are of a much more general nature. Additionally, a comparison between all variables measured within the sample was not practical, since detailed information for all attributes or attribute classes defined in this

investigation is not always reported in the literature. Instead, this study focused on those ceramic attributes that are reported in great detail, including surface finish, slip, decoration, form, vessel morphology, paste, and temper. This study explicitly does not compare type– variety and ware in the tables, or as high or low visibility attributes, since these are etic categories. Where possible, however, specific percentages of correspondence across all attributes are provided, including ware and type– variety. Additionally, those variables that are compared are subdivided into high and low visibility attributes (cf. Figure 4.18), which is crucial to interpreting the results of this comparative evaluation in relation to issues of interaction and innovation. 99

Archaeological Paleography

Site 1 2 4 3 Sig.

297 595 266 407

Paste texture Subset for alpha = 0.05 N 1 2 5.08 6.16 6.26 6.41 1.000 0.603

Site 1 4 2 3 Sig.

297 266 595 407

Paste color Subset for alpha = 0.05 N 1 2 3.65 4.29 4.47 4.61 1.000 0.206

Figure 5.40. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along paste parameters of color and texture.

Site 1 3 4 2 Sig.

297 407 266 595

represented in the sample by Joventud Red ceramics, is dominant at a number of other sites. At Hol phase Piedras Negras, 84% of the assemblage is comprised of waxy wares, and Joventud Red accounts for 57.5% of the ceramic materials encountered (Pérez Robles 2006: 128). All other type–varieties in the Middle Formative period sample find correspondence at Piedras Negras, where Achiotes Unslipped (16%), Tierra Mojada (11%), and Pital Cream (10%) comprise significant minorities, as they do in the sample (see Figure 4.14). Pastes and tempers do not differ, and similar forms and vessel morphology are reported at Piedras Negras during the Middle Formative period (Holley 1987; Muñoz 2004; Pérez Robles 2006).

Temper content Subset for alpha = 0.05 N 1 2 1.60 2.00 2.18 2.18 1.000 0.203

Figure 5.41. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along paste parameter of temper content.

Middle Formative Period

Similarly, other sites in the Northwest Maya lowlands, including Palenque during the Misolha/Early Waxy horizon, Kub phase Tierra Blanca, and the Xot and Chacibcan phases at Trinidad, Paso Nuevo, and Zapatillo, are also dominated by waxy, red slipped ceramics (Ochoa and Casasola 1991: 10; Rands 1972, 1987, 2007b). As in the sample, groove–incision and fluting are the primary modes of decoration, and cajetes and plates with thickened, everted rims and divergent walls are abundant. In contrast to the sample, carbonate tempers are relatively more common at these sites, although calcite tempers are

Despite the relatively small sample size of Middle Formative period ceramics from the study area, these materials may be effectively compared to ceramics from adjacent sites and regions (Figure 5.60). Because in almost all respects the Middle Formative ceramics in the sample resemble contemporary Mamom sphere materials from elsewhere in lowland Mesoamerica, there is a striking similarity between the sample and the ceramics encountered at other sites across all attributes considered in this study. The waxy red tradition,

Site San Claudio

Form/ shape

Total

plate cajete apaxtle basin cazuela cup jar olla tecomate bowl unidentified

28 116 12 24 9 13 27 28 0 37 3 297

Tiradero 17 70 9 75 59 31 111 93 17 66 47 595

Mirador 9 43 4 36 21 5 84 107 10 65 23 407

Revancha 9 24 4 24 18 9 58 49 9 38 24 266

Figure 5.42. Cross tabulation of form/shape by site, Early Classic period.

100

Total 63 (4.0%) 253 (16.2%) 29 (1.9%) 159 (10.2%) 107 (6.8%) 58 (3.7%) 280 (17.9%) 277 (17.7%) 36 (2.3%) 206 (13.2%) 97 (6.2%) 1565

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.43. Bar graph illustrating site specific percentages of ceramic forms and shapes within the Early Classic period ceramic sample.

Dependent Variable

(I) Site 1

(J) Site

2 3 4 2 1 3 4 Form/shape 3 1 2 4 4 1 2 3 1 2 3 4 2 1 3 4 Angles/flanges 3 1 2 4 4 1 2 3 * The mean difference is significant at the 0.05 level.

Mean Difference (I-J) -1.932* -2.455* -2.406* 1.932* -0.523* -0.474 2.455* 0.523* 0.049 2.406* 0.474 -0.049 -0.784* -0.661* -0.778* 0.784* 0.122 0.005 0.661* -0.122 -0.117 0.778* -0.005 0.117

Std. Error 0.203 0.218 0.241 0.203 0.184 0.210 0.218 0.184 0.225 0.241 0.210 0.225 0.105 0.113 0.125 0.105 0.095 0.109 0.113 0.095 0.116 0.125 0.109 0.116

Sig.* 0.000 0.000 0.000 0.000 0.023 0.110 0.000 0.023 0.996 0.000 0.110 0.996 0.000 0.000 0.000 0.000 0.570 1.000 0.000 0.570 0.745 0.000 1.000 0.745

95% Confidence Interval Lower Upper Bound Bound -2.45 -1.41 -3.01 -1.89 -3.03 -1.79 1.41 2.45 -0.99 -0.05 -1.01 0.07 1.89 3.01 0.05 0.99 -0.53 0.63 1.79 3.03 -0.07 1.01 -0.63 0.53 -1.05 -0.51 -0.95 -0.37 -1.10 -0.46 0.51 1.05 -0.12 0.37 -0.27 0.29 0.37 0.95 -0.37 0.12 -0.42 0.18 0.46 1.10 -0.29 0.27 -0.18 0.42

Figure 5.44. Tukey HSD test for between site variability in Early Classic period formal attributes.

101

Archaeological Paleography

Site 1 2 4 3 Sig.

297 595 266 407

Form/ shape Subset for alpha = 0.05 N 1 2 4.48 6.41 6.89 6.94 1.000 0.070

Site 1 3 4 2 Sig.

297 407 266 595

Angles/ flanges Subset for alpha = 0.05 N 1 2 4.74 5.41 5.52 5.53 1.000 0.687

Figure 5.45. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along formal parameters of form/shape and angles/flanges.

Figure 5.46. Clustered boxplot displaying ranges of morphological attribute variability during the Early Classic period.

reported.13 Tempers at sites in the central Petén, such as Monos phase El Mirador and Mamom phase Uaxactún, illustrate slight variability when compared to the sample, although these discrepancies are likely attributable to geological differentiation between regions. In all other respects, the ceramics from the sample appear

analogous to those from El Mirador and Uaxactún. At El Mirador, Paso Caballos Waxy is the dominant ware, and Joventud Red forms 56% of the Monos assemblage, with Achiotes and Pital types also present in significant quantities (Forsyth 1989: 125). Incised and fluted vessels with simple plate and dish forms, flat bases, and direct, everted rims are likewise prevalent on the relatively small portion of decorated ceramics in the assemblage. Thick and waxy red slips are also diagnostic of the Mamom assemblage at Uaxactún, where groove–incising on flat based, divergent walled vessels that lack supports are as common as they are in the sample (Smith 1955: 21).

13 Ochoa (1983; Ochoa and Casasola 1991: 10) suggests that these and other sites in the northwest lowlands were looking toward the central Petén for inspiration in a developing regional ceramic tradition that was rooted in technologies and conventions that originated in the Gulf coast Olmec heartland or Grijalva depression and were subsequently incorporated into the autochthonous population of the lower Usumacinta region. This conclusion is based on the numerous references to Olmec material culture prevalent throughout the lower Usumacinta and San Pedro Mártir basins, which continued well into the Late Formative period.

In the Pasión and Chixoy River region of the southwest Petén, the ceramic assemblages at the sites of Altar de 102

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Dependent Variable

(I) Site

(J) Site

Mean Difference (I-J)

1

2 -0.544* 3 -0.794* 4 -0.558* 2 1 0.544* 3 -0.250* 4 -0.014 Base type 3 1 0.794* 2 0.250* 4 0.236 4 1 0.558* 2 0.014 3 -0.236 1 2 -1.625* 3 -1.762* 4 -1.507* 2 1 1.625* 3 -0.137 4 0.117 Support type 3 1 1.762* 2 0.137 4 0.255 4 1 1.507* 2 -0.117 3 -0.255 1 2 0.196 3 0.449* 4 0.296* 2 1 -0.196 3 0.253* 4 0.101 Wall type 3 1 -0.449* 2 -0.253* 4 -0.152 4 1 -0.296* 2 -0.101 3 0.152 * The mean difference is significant at the 0.05 level.

Std. Error 0.107 0.115 0.127 0.107 0.097 0.111 0.115 0.097 0.118 0.127 0.111 0.118 0.166 0.178 0.197 0.166 0.150 0.172 0.178 0.150 0.184 0.197 0.172 0.184 0.093 0.100 0.110 0.093 0.084 0.096 0.100 0.084 0.103 0.110 0.096 0.103

Sig.* 0.000 0.000 0.000 0.000 0.047 0.999 0.000 0.047 0.190 0.000 0.999 0.190 0.000 0.000 0.000 0.000 0.797 0.904 0.000 0.797 0.510 0.000 0.904 0.510 0.151 0.000 0.037 0.151 0.014 0.723 0.000 0.014 0.450 0.037 0.723 0.450

95% Confidence Interval Lower Bound Upper Bound -0.82 -0.27 -1.09 -0.50 -0.88 -0.23 0.27 0.82 -0.50 0.00 -0.30 0.27 0.50 1.09 0.00 0.50 -0.07 0.54 0.23 0.88 -0.27 0.30 -0.54 0.07 -2.05 -1.20 -2.22 -1.30 -2.01 -1.00 1.20 2.05 -0.52 0.25 -0.33 0.56 1.30 2.22 -0.25 0.52 -0.22 0.73 1.00 2.01 -0.56 0.33 -0.73 0.22 -0.04 0.43 0.19 0.71 0.01 0.58 -0.43 0.04 0.04 0.47 -0.15 0.35 -0.71 -0.19 -0.47 -0.04 -0.42 0.11 -0.58 -0.01 -0.35 0.15 -0.11 0.42

Figure 5.47. Tukey HSD test for between site variability in Early Classic period morphological attributes.

Sacrificios and Seibal—in the San Felix and Escoba phases, respectively—also demonstrate considerable correspondence with the sample. At Altar, monochrome red, waxy–slipped ceramics are prevalent, with Joventud Red comprising 63.7% of the Middle Formative period assemblage (Adams 1971: 4, 85). Achiotes Unslipped again forms a significant minority of the ceramic materials at 22.2% of the San Felix phase collection; Pital Cream is also present. Flat based plates with everted rims are most common, often displaying groove incision and resist decoration, although necked jars and the decorative techniques of champfering and punctuation are encountered with greater frequency at Altar than

within the sample. Champfering is also the dominant decoration at Seibal, although incisions, fluting, and resist painting are relatively common as well. Decoration is present on only a small percentage of the assemblages at Altar and Seibal, as in the sample. Like the sampled Middle Formative period ceramics from the study area, waxy red and cream slips on simple forms such as plates and cajetes are prevalent in the Escoba phase at Seibal, with Joventud Red again the most common type–variety, with Pital Cream and Tierra Mojada also present. Vessel morphology differs somewhat between the sample, Altar, and Seibal, with incurved walls and medial and basal angles more evident at the latter two sites. 103

Archaeological Paleography

Dependent Variable

(I) Site

(J) Site

1

2 3 4 2 1 3 4 Lip type 3 1 2 4 4 1 2 3 * The mean difference is significant at the 0.05 level.

Mean Difference (I-J) -1.009* -0.848* -0.861* 1.009* 0.161 0.148 0.848* -0.161 -0.012 0.861* -0.148 0.012

Std. Error 0.217 0.233 0.257 0.217 0.196 0.225 0.233 0.196 0.240 0.257 0.225 0.240

Sig.* 0.000 0.002 0.005 0.000 0.845 0.912 0.002 0.845 1.000 0.005 0.912 1.000

95% Confidence Interval Lower Upper Bound Bound -1.57 -0.45 -1.45 -0.25 -1.52 -0.20 0.45 1.57 -0.34 0.67 -0.43 0.73 0.25 1.45 -0.67 0.34 -0.63 0.61 0.20 1.52 -0.73 0.43 -0.61 0.63

Figure 5.48. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along morphological parameters of base and support type.

Site 1 3 4 2 Sig.

N 297 407 266 595

ceramics, and all type–varieties within the sample are also found at Cuello. In the Felisa and Francesa phases of the Middle Grijalva region and Chiapa de Corzo, respectively, a variety of waxy, polished red slipped monochrome pottery, Hama Red, is dominant. Lee (1972: 10) describes this tradition as directly related to the Joventud Red prevalent in Mamom sphere ceramics throughout lowland Mesoamerica. Incising and grooving on large, everted– rim dishes and simple vessel forms is common in Felisa and Francesa phase ceramics, although the presence of tall–necked jars and cylinders, as well as white–slipped and polished brown pottery, represent a discrepancy between the Middle Formative period ceramics of central and southwestern Chiapas and the sample.

Lip type Subset for alpha = 0.05 1 2 2.52 3.36 3.38 3.52 1.000 0.896

Figure 5.49. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along morphological parameter of lip type.

Finally, the Middle Formative period ceramic sample from the study area shows marked similarities with sites and regions even further afield, such as Chiapa de Corzo and the Middle Grijalva basin to the west, and Cuello, Belize, far to the east. In the Lopez phase at Cuello, simple forms of monochrome, waxy red slipped ceramics are again dominant, with Joventud Red comprising 40% of the Middle Formative period assemblage (Kosakowsky 1987: 23). Achiotes Unslipped and Pital Cream form a significant minority of the Lopez phase

In sum, the distribution of ceramic attributes is strikingly uniform throughout lowland Mesoamerica during the Middle Formative period. Monochrome red, waxy slipped ceramics, particularly Paso Caballos Waxy ware and Joventud Red, are dominant at assemblages across the region as a whole. Decoration is generally rare, but when present is usually formed by groove–incision and resist painting, although champfering are fluting are evident in some assemblages. Vessel forms and

Wall type Site 3 4 2 1 Sig.

Subset for alpha = 0.05

N 407 266 595 297

Rim/ neck type

1 2.04 2.19

.405

2 2.19 2.29 0.734

3

2.29 2.49 0.190

Site 1 2 3 4 Sig.

N 297 595 407 266

Subset for alpha = 0.05 1 3.18 3.21 3.43 3.44 0.483

Figure 5.50. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along morphological parameters of wall and rim/neck type.

104

Interpreting the Results of the Comparative and Statistical Ceramic Analyses Site San Claudio

Wall thickness

< 0.4cm 0.4–0.7cm 0.7–1.0cm 1.0–1.3cm > 1.3cm

Total

7 40 136 97 17 297

Tiradero 37 252 247 48 11 595

Mirador 24 195 155 27 6 407

Total

Revancha 18 95 119 30 4 266

86 (5.5%) 582 (37.2%) 657 (42%) 202 (12.9%) 38 (2.4%) 1565

Figure 5.51. Cross tabulation of wall thickness by site, Early Classic period.

such as the Middle Formative portion of the La Venta sequence, Cerro de las Mesas Lower I, and the Palangana phase at San Lorenzo (Adams 1971: 123; Cheetham 2007; Lee 1972: 10; Rands 2007b: 32). Like Mamom sphere ceramics, these complexes appear to develop out of and continue earlier traditions rooted in the Xe and Locona sequences, respectively. Uniformity in attribute distribution also suggests a significant degree of relation between ceramic traditions during the Middle Formative period, a point to which I will return shortly. Late Formative Period Late Formative period ceramics also display striking uniformity in attribute distribution and correspondences (Figure 5.61). Although somewhat more limited in spatial scope, these parallels reflect the general range of the Chicanel ceramic sphere, itself displaying continuity with the preceding Mamom sphere and marked by increasing standardization in ceramic forms and typologies. At this time, monochrome reds, blacks, and creams dominate regional assemblages across the Maya lowlands, with an especially strong concentration of Paso Caballos Waxy and Sierra group ceramics. The emergence of diagnostic Chicanel types across the lowlands suggests a burgeoning orientation toward the Petén lowlands as a center of ceramic innovation, adopting and continuing the development of a distinct and separate lowland Maya tradition that had arisen with the Early Formative and early Middle Formative Xe ceramic tradition (Adams 1971: 123; Ochoa

Figure 5.52. Percentage of occurrence of specific wall thickness within the Early Classic period assemblages.

morphology also display significant correspondences across assemblages, and even lower visibility attributes such as paste color and texture are surprisingly similar. Such uniformity illustrates the wide spatial range of the Mamom sphere, given the ample distribution of diagnostic Mamom types that likely originated in the central Petén lowlands or Pasión River region and Mamom–style decorative techniques such as horizontal grooved–incised lines on vessel exteriors (Adams 1971: 120). Consistencies across ceramic traditions justify the inclusion of the lower San Pedro Mártir ceramic sequence firmly within the Mamom horizon. Significantly, many prior studies have noted typological and formal analogies between the Mamom sphere and ceramic complexes outside of the greater Maya lowland area,

Site

Height Total

< 9cm 9–18cm > 18cm unknown

San Claudio 169 71 45 12 297

Tiradero 94 170 284 47 595

Mirador 57 120 206 24 407

Revancha 37 79 126 24 266

Figure 5.53. Cross tabulation of vessel height by site, Early Classic period.

105

Total 357 (22.8%) 440 (28.1%) 661 (42.2%) 107 (6.8%) 1565

Archaeological Paleography Site

Height Total

< 9cm 9–18cm > 18cm unknown

San Claudio 169 71 45 12 297

Tiradero 94 170 284 47 595

Mirador 57 120 206 24 407

Total

Revancha 37 79 126 24 266

357 (22.8%) 440 (28.1%) 661 (42.2%) 107 (6.8%) 1565

Figure 5.54. Cross tabulation of neck length by site, Early Classic period.

Figure 5.55. Clustered boxplot displaying ranges of dimension attribute variability during the Early Classic period.

and Casasola 1991: 10; Willey 1970). The ubiquitous Chicanel horizon marker Sierra Red comprises almost 50% of the Late Formative period sample from the study area, where distribution of Late Formative ceramics, and ceramic attributes, is essentially uniform, thus firmly ensconcing the sample within the Chicanel sphere.

15 type–varieties present in the sample also appear at Piedras Negras during this time, illustrating a very strong correspondence between the sampled materials and those at Piedras Negras. Similarly, incision and fluting are the most common decorative modes encountered in both assemblages, and vessel forms, morphology, and paste attributes also display strong similarities.

Like the sample, the Late Formative Abal phase at Piedras Negras is dominated by waxy red ceramics, specifically Paso Caballos Waxy ware and Sierra Red, at 70% and 48% of the assemblage, respectively (Pérez Robles 2006: 131). Flor cream (10%), Achiotes Unslipped (10%), and Polvero Black (5%) comprise a significant minority of the Abal phase materials. Thus, the same wares and type–varieties occur in the sample and at Piedras Negras in roughly the same proportions. Additionally, 12 of the

At sites in the northwest Maya lowlands, including Palenque, dark red slips and waxy surfaces are also common on Late Formative period ceramics. Although ceramic evidence from the Misolha and late waxy horizon at Palenque, Trinidad, Zapatillo, Paso Nuevo, and other northwest lowland sites is limited, Rands (1972: 212, 2007b: 32) recognizes a general affiliation with Sierra Red and a relative abundance of other 106

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Dependent Variable

(I) Site

(J) Site

Mean Difference (I-J)

1

2 0.428* 3 0.575* 4 0.547* 2 1 -0.428* 3 0.147 4 0.119 Neck length 3 1 -0.575* 2 -0.147 4 -0.028 4 1 -0.547* 2 -0.119 3 0.028 1 2 0.690* 3 0.760* 4 0.609* 2 1 -0.690* 3 0.071 4 -0.081 Wall thickness 3 1 -0.760* 2 -0.071 4 -0.152 4 1 -0.609* 2 0.081 3 0.152 1 2 -0.814* 3 -0.821* 4 -0.852* 2 1 0.814* 3 -0.007 4 -0.038 Height 3 1 0.821* 2 0.007 4 -0.031 4 1 0.852* 2 0.038 3 0.031 * The mean difference is significant at the 0.05 level.

Std. Error 0.082 0.088 0.097 0.082 0.074 0.085 0.088 0.074 0.091 0.097 0.085 0.091 0.057 0.062 0.068 0.057 0.052 0.059 0.062 0.052 0.064 0.068 0.059 0.064 0.060 0.064 0.071 0.060 0.054 0.062 0.064 0.054 0.067 0.071 0.062 0.067

95% Confidence Interval Lower Bound Upper Bound 0.22 0.64 0.35 0.80 0.30 0.80 -0.64 -0.22 -0.04 0.34 -0.10 0.34 -0.80 -0.35 -0.34 0.04 -0.26 0.21 -0.80 -0.30 -0.34 0.10 -0.21 0.26 0.54 0.84 0.60 0.92 0.43 0.78 -0.84 -0.54 -0.06 0.20 -0.23 0.07 -0.92 -0.60 -0.20 0.06 -0.32 0.01 -0.78 -0.43 -0.07 0.23 -0.01 0.32 -0.97 -0.66 -0.99 -0.66 -1.03 -0.67 0.66 0.97 -0.15 0.13 -0.20 0.12 0.66 0.99 -0.13 0.15 -0.20 0.14 0.67 1.03 -0.12 0.20 -0.14 0.20

Sig.* 0.000 0.000 0.000 0.000 0.195 0.501 0.000 0.195 0.990 0.000 0.501 0.990 0.000 0.000 0.000 0.000 0.520 0.528 0.000 0.520 0.080 0.000 0.528 0.080 0.000 0.000 0.000 0.000 0.999 0.930 0.000 0.999 0.966 0.000 0.930 0.966

Figure 5.56. Tukey HSD test for between site variability in Early Classic period dimension attributes.

Site 3 2 4 1 Sig.

407 595 266 297

Wall thickness Subset for alpha = 0.05 N 1 2 2.50 2.57 2.65 3.26 0.060 1.000

Site 1 2 3 4 Sig.

N 297 595 407 266

Height Subset for alpha = 0.05 1 2 1.66 2.48 2.48 2.52 1.000 0.466

Figure 5.57. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along dimension parameters of wall thickness and vessel height.

107

Archaeological Paleography

Site 3 4 2 1 Sig.

Neck length Subset for alpha = 0.05 N 1 2 407 3.22 266 3.24 595 3.36 297 3.79 0.324 1.000

Diversity Measure

Temporal Context Middle Formative

Late Formative

Early Classic

Raw H score

21.1489

24.1809

113.3650

BR Coefficienta

178.8510

175.8190

86.6346

BR scaled to 1b

0.8942

0.8791

0.4332

BR = 200 - S (i = 1 to N) |PiA - PiB|. A score of 200 indicates maximum similarity, whereas a score of 0 indicates maximum diversity. b To scale BR to 1, divide BR by 200. A score of 1 indicates identical assemblages, whereas a score of 0 indicates maximum diversity.

Figure 5.58. Extrapolated Tukey test illustrating means for groups in homogeneous subsets along dimension parameter of neck length.

a

markers of the Chicanel horizon, such as unslipped, striated jars and the red and brown pastes characteristic of Sierra group ceramics. He also notes the possible emergence of a localized tradition, evident through departures from the Chicanel tradition in northwest assemblages, such as the relative scarcity of flanged bowls (also uncommon in the lower San Pedro Mártir region during the Late Formative period) and the frequent occurrence of carbonate temper—itself striking due to the inaccessibility of limestone in the northwest lowlands (Rands 2007b: 28). On the basis of these discrepancies, Rands (2007b: 28), echoing Ochoa (1983; Ochoa and Casasola 1991: 10), concludes that the ‘Mayanization’ of ceramic materials appears to have been underway throughout the northwest lowlands from Chacibcan times and continuing into the Late Formative period, but that this process of standardization and affiliation with the late Mamom and Chicanel spheres was cut short, reflected in a ‘near–hiatus’ and the paucity of Late Formative ceramics. Nevertheless, some correspondences are marked, and the limited material does suggest a significant, if not absolute, association with the Chicanel–style attributes present both in the sample and throughout the lowlands.

Figure 5.59. H score heterogeneity measures of assemblage diversity over time.

straight divergent walls and thick everted rims, striated jars, and globular ollas. As with ceramic materials at Palenque and other sites in the northwestern lowlands, Ochoa and Casasola (1991: 10) note a trend toward standardization of ostensibly Maya–Chicanel forms and typologies beginning in the Middle Formative and continuing up to the middle Late Formative period (c. AD 1), at which point ceramics become markedly less abundant, reemerging with striking departures from Classic period lowland traditions in the Early Formative period some two to three centuries later.14 At sites in the central Petén, particularly strong parallels in the Late Formative period ceramic materials are also evident. In the Cascabel phase at El Mirador, for instance, Sierra Red is equally prevalent, comprising 52.2% of the assemblage. The waxy surfaces and red, black, and cream slips of the Paso Caballos tradition dominate, although Uaxactún Unslipped ceramics are slightly more abundant here than in the sample, forming 31.5% of the Cascabel complex. Ten of the 18 ceramic types present at El Mirador during this time are also found in the sample, albeit in different proportions, including significant minorities of Sapote Striated (26.5%), Polvero Black (8.28%), Achiotes Unslipped (1.9%), Flor Cream, Laguna Verde Incised, Correlo Incised Dichrome, Altamira Fluted, and Repasto Black on Red (Forsyth 1989). Incision and fluting are likewise the most common decorative modes, and the red and brown pastes of the Sierra group are equally abundant at El Mirador. Typical Chicanel forms and morphologies are evident in the Cascabel complex, including the everted–

At Payol phase Tierra Blanca, on the other hand, stronger correspondences with the sampled materials are noted. There, Sierra Red is the dominant type, as it is at other neighboring sites such as La Carmelita, Montebello, San Marcos, Las Mercedes, La Soledad, and Jolosinal, among others (Ochoa and Casasola 1991: 10). Likewise, Laguna Verde Incised, Pital Cream, Polvero Black are all well represented minorities at Tierra Blanca, as are Uaxactún Unslipped wares such as Sapote Striated and Achiotes Unslipped, just as in the sample. All type– varieties present in the Late Formative assemblage at Tierra Blanca are represented in the sample as well. Fine to medium red and brown pastes, calcite tempers, and the decorative techniques of incision and fluting are virtually identical between the sample and the assemblage at Tierra Blanca. Additionally, the same forms and morphologies are present in both sequences, dominated by flat–based shallow plates and cajetes with

A similar pattern is also evident at Pomoná, where limited Late Formative period ceramics are dominated by a waxy red ware, labeled Minanhá Red—likely a localized equivalent of Sierra Red—that comprises 51.1% of the Late Formative Pomontik phase assemblage. A lacuna is noted during the later part of the Late Formative period, whereupon the Early Classic Pomonaab phase assemblage at Pomoná departs from traditional lowland Maya conventions (García Moll 2005; López Varela 1991). 14

108

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Site

Attribute Decoration (HV)

Surface finish (HV)

Slip (HV)

Form/shape Morphology Paste (HV) (LV) (LV)

Temper (LV)

divergent walls; flat base; no red; grey; waxy; plates; supports; brown; calcite smooth cajetes everted/ medium thick direct rims Palenque/NW lowlands + + +/– – (?) + + +/– Tierra Blanca + + + + + + + Piedras Negras ++ ++ + + + + ++ Altar de Sacrificios + ++ + +/– +/– + +/– Seibal + + ++ + +/– + +/– El Mirador + ++ ++ + + + +/– Uaxactún + + + +/– +/– ? ? Middle Grijalva + + +/– +/– + +/– ? Chiapa de Corzo + + +/– +/– + +/– ? KEY: ++ = strong correspondence; + = correspondence; +/– = slight correspondence coupled with slight discrepancy; – = discrepancy; – – = strong discrepancy; ? = insufficient data/unreported HV = high visibility attribute; LV = low visibility attribute Based on data reported in Acuña Smith 2005; Adams 1971; Forsyth 1989; Hernández Pons 1984; Holley 1987; Lee 1972; Muñoz 2002, 2004; Ochoa and Casasola 1991; Ochoa and Hernández 1975; Pérez Robles 2006; Rands 1972, 1987, 2007a, 2007b; Sabloff 1975; Smith 1955; Vargas and Hernández 1979. plain undecorated; Sample groove– incised; resist

red/red– orange; cream; unslipped

Figure 5.60. Comparative assessment between assemblages at the regional level, Middle Formative period (Mamom sphere), illustrating correspondence and discrepancies between sampled ceramics and adjacent sites and regions.

Attribute Site

Decoration (HV)

Surface finish (HV)

Slip (HV)

Form/shape Morphology (HV) (LV)

Paste (LV)

Temper (LV)

shallow dishes; deep bowls; thick, everted groove– red; brown; waxy; red; cream; cajetes; jars; rims; flat incised; fine to calcite; sand rough; black; thick; plates; ollas; Sample base; no fluted; medium smooth unslipped bowls supports; striated divergent, incurving walls Palenque/NW lowlands +/– ++ +/– + +/– +/– +/– Tierra Blanca ++ ++ ++ ++ + + ++ Piedras Negras ++ ++ ++ ++ ++ + + Altar de Sacrificios + ++ ++ +/– +/– + ++ Seibal + + + + + + ? El Mirador ++ ++ ++ + + + +/– Uaxactún ++ + ++ + + ? ? Middle Grijalva +/– +/– +/– +/– +/– +/– ? Chiapa de Corzo – +/– +/– +/– ? +/– ? KEY: ++ = strong correspondence; + = correspondence; +/– = slight correspondence coupled with slight discrepancy; – = discrepancy; – – = strong discrepancy; ? = insufficient data/unreported HV = high visibility attribute; LV = low visibility attribute Based on data reported in Acuña Smith 2005; Adams 1971; Forsyth 1989; Hernández Pons 1984; Holley 1987; Lee 1972; Muñoz 2002, 2004; Ochoa and Casasola 1991; Ochoa and Hernández 1975; Pérez Robles 2006; Rands 1972, 1987, 2007a, 2007b; Sabloff 1975; Smith 1955; Vargas and Hernández 1979. Figure 5.61. Comparative assessment between assemblages at the regional level, Late Formative period (Chicanel sphere), illustrating correspondence and discrepancies between sampled ceramics and adjacent sites and regions.

109

Archaeological Paleography Incised, Altamira Fluted, Flor Cream, Polvero Black, and Sapote Striated. Decorative techniques are similar as well, with incision and fluting common, although striation is the dominant decorative mode, and a greater range of techniques is evident during the Cantutse phase, including punctuation, appliqué, and impression (paralleling the range of decorative techniques evident at Uaxactún and Cuello; Kosakowsky 1987; Sabloff 1975:11; Smith 1955). Divergent–walled plates and serving dishes, incurved bowls, and thick–walled jars with short, out–curving necks are common at Seibal during this time, illustrating a significant correspondence with the sample.

rim dishes and incurved–walled bowls frequently encountered in the sample. Vessel supports are similarly rare, although flanged vessels are decidedly more common at El Mirador than in the sample. Waxy wares and monochrome red, black, and cream ceramics are also abundant at Chicanel phase Uaxactún—unsurprisingly, since the diagnostic Chicanel characteristics were first identified at that site (Smith 1955). Horizontal groove– incising is the dominant decorative technique in the Late Formative assemblage at Uaxactún, occurring on 73.7% of all decorated vessels, whereas fluting is evident on 10.4% of the assemblage. In addition, the wide– mouth jars, shallow everted rim plates and cajetes, and divergent–walled serving dishes prevalent in the sample are also exceedingly common during the Late Formative period at Uaxactún, accounting for 25.8%, 20.7%, and 10%, respectively, of the Chicanel phase ceramics.

The Late Formative ceramic materials encountered at sites and regions further afield, such as Chiapa de Corzo, the Middle Grijalva basin, and Cuello, demonstrate an interesting pattern. At Cocos phase Cuello, for example, ten of 15 identified type–varieties find correspondence in the sample. The Sierra group is exceedingly prevalent at Cuello during this time, comprising an even greater proportion (73%) of the total assemblage at that site than within the sample (Kosakowsky 1987: 25). Sapote Striated, Polvero Black, Flor Cream, and Laguna Verde Incised are also represented in the Cocos complex. Vessel decoration is similar, although a wider range of techniques is evident, and medial and labial flanged vessels are much more common at Cuello than in the sample. As at Altar de Sacrificios, increasing standardization throughout the Formative period is noted, followed by greater experimentation in ceramic production techniques, forms, and characteristics is evident at the Late Formative–Early Classic period transition (Kosakowsky 1987: 28). In the Guañoma phase of the Middle Grijalva region and the Guanacaste and Horcones phases at Chiapa de Corzo (c.300 BC–AD 1), waxy red slip and red paste ceramics are abundant, with monochrome red ceramics, including imported Sierra Red types, accounting for 62.5% of the Late Formative period materials (Lee 1972: 12). Similar shallow, everted rim plates and serving dishes are also evident at this time. Nevertheless, the white slip and polished brown traditions of the preceding Felisa and Francesa phases continue to be well represented through the end of the Formative period. In the later Late Formative Ipsan and Istmo phases (c. AD 1–200), correspondences disappear entirely, and marked variability between the sample and the ceramics of central and southwestern Chiapas becomes evident. During the Ipsan phase of the Middle Grijalva region, in particular, the prevalence of intrusive types such as Yahama Rough and Yatsipo Sandy indicate a re–orientation of the ceramic sequence toward the south and west, and a burgeoning affiliation with the ceramic traditions of Oaxaca that would continue into the Early Classic period.

Parallels and strong correspondences are also evident in the Late Formative period at Altar de Sacrificios, during the late San Felix and Plancha phases, as well as in the Cantutse complex at Seibal. As at other lowland sites, the Late Formative period ceramic sequence at Altar displays a marked trend toward standardization and uniformity in the form, finish, and technical aspects of ceramic materials, as well as increasing experimentation nearing the Late Formative–Early Classic period horizon (Adams 1971: 5, 93). The Plancha ceramic phase essentially represents a continuation of earlier waxy ware and monochrome red San Felix traditions. As in the sample, Sierra Red is the predominant type at Altar during this time, and the red and brown pastes typical of the Sierra group are prevalent. Sierra Red accounts for an average of 45% of all ceramic materials encountered in all operations at the site, followed closely by Polvero Black (17.1%), Sapote Striated (6.10%), and Achiotes Unslipped (4.43%). Fourteen of the 15 type– varieties evident in the sample are also found in the Plancha complex, and, as at Piedras Negras, they occur in roughly the same proportions. Adams (1971: 93) notes an abundance of calcite tempers throughout the Plancha phase, illustrating a close parallel with the sample. Vessel forms and morphology at Altar are slightly more varied, and flanged vessels are considerably more common, although significant formal and morphological parallels are evident (e.g., wide–mouthed jars and shallow, everted–rim plates). Decoration is relatively rare, but groove–incision, fluting, and striation are common techniques. At Seibal, waxy reds, creams, and blacks, as well as a significant minority of unslipped Uaxactún ware, are also ubiquitous. Five type–varieties of the Sierra group dominate the assemblage, with Flor, Polvero, and Sapote groups well represented in the Late Formative Cantutse complex (Sabloff 1975: 11). Of the nine definitively identified type–varieties at Seibal, six are also present in the sample, including Sierra Red, Laguna Verde

Overall, the uniform distribution of ceramic attributes noted in the Middle Formative period continues into 110

Interpreting the Results of the Comparative and Statistical Ceramic Analyses discrepancies reflect the relatively more limited spatial scope of the Tzakol sphere, and intimate the emergence of a distinct ceramic tradition in the northwest Maya lowlands. Developing regional differentiation in ceramic traditions also suggests a shift in patterns of interaction, perhaps following from or resulting in the establishment of a more rigidly defined and maintained—and significantly less permeable—sociopolitical or cultural boundary that emerged in or near the northwest lowlands of the San Pedro Mártir basin around the Late Formative– Early Classic period horizon (Englehardt 2010; Holley 1987; Rands 1977).

the Late Formative period. Such internal consistency and regional standardization of Late Formative period lowland pottery likely resulted from the development and widespread diffusion of the Mamom horizon style, whose influence is extensively noted in Middle Formative period ceramic phases throughout southeastern Mesoamerica (Willey et al. 1967). Modified continuity of widely shared earlier ceramic traditions through the subsequent Late Formative Chicanel sphere is thus unsurprising. In all respects, the ceramic materials in the sample are typologically, formally, technically, and morphologically similar to Late Formative period assemblages in adjacent regions, manifesting differences primarily of emphasis as opposed to striking departures. Although some varietal variation exists between the sample and neighboring sequences, the types are essentially the same, and ranges of formal variation and decorative technique are virtually indistinguishable.15 Throughout the lowlands standardization in ceramic traditions is evident, and a wider range of attributes emerges during the Late Formative period. Such standardization and uniformity in attribute distribution suggests that lowland ceramic traditions continued to be closely related during the Late Formative period. Approaching the Early Classic period horizon, a tendency toward experimentation becomes evident at several sites, perhaps indicating incipient localized styles or traditions, or emerging regionalization. Such trends toward variability become much more pronounced in the Early Classic period.

At San Claudio, the Early Classic period assemblage is marked by the emergence of glossy orange, cream, and red slips characteristic of Petén Gloss Ware and elaborate polychrome painting as the dominant decorative technique. Typical Tzakol vessel forms and morphology, such as Z–angles and annular ring bases become progressively more common at San Claudio through the Early Classic period. Lower visibility attributes such as paste color, texture, and tempers also reflect the prevalence of Petén Gloss Ware and Tzakol sphere conventions at the site. These new ceramic standards appear to have resulted from the experimentation noted at many sites during the later part of the Late Formative period, during which the orange polychrome pottery that would eventually become a Tzakol horizon type emerged, principally at sites in the central Petén lowlands. In contrast, the ceramic materials at Tiradero, Mirador, and Revancha appear to continue the earlier waxy and unslipped traditions of the Middle and Late Formative period Mamom and Chicanel spheres well into the Early Classic period, and are typologically unrelated to contemporaneous Tzakol developments elsewhere in the lowlands. The gloss ware, polychrome, and orange– slipped pottery characteristic of Early Classic period lowland Maya ceramics of the Floral Park and Tzakol spheres, while present in significant quantities at San Claudio (almost 55% of the Early Classic sample at the site), are decidedly lacking at Mirador, Revancha, and Tiradero just 50km to the north. At these sites, gloss ware is conspicuously scarce and polychrome ceramics are almost entirely absent.16 Rather, rough, dull, and waxy surface treatments are prevalent, striation and incision remain dominant decorative techniques, and relatively simpler vessel forms and morphology persist.

Early Classic Period As indicated by the statistical analyses of the sampled ceramics presented above, variability within the sample increases significantly in the Early Classic period. Further, the Tukey analyses suggest the existence of two discrete clusters of ceramic attributes—one centered at San Claudio and another at Tiradero, Mirador, and Revancha to the north (Figures 62 and 63). Such variable attribute distribution meshes well with the lower San Pedro Mártir regional ceramic sequence, in which Early Classic period materials from San Claudio appear relatively more related to the Tzakol sphere, whereas ceramics from the other sites are only loosely affiliated with this tradition. Given this intra–assemblage variation, one would expect the sample as a whole to appear at once related and unrelated to regional sequences. The Early Classic period comparative assessment has thus been sub-divided, separately evaluating the materials from San Claudio and those from the Tiradero, Mirador, and Revancha against the ceramic assemblages of neighboring sites and regions. In general, the comparative analysis of Early Classic period materials parallels the discrete distribution of ceramic attributes suggested by both the regional sequence and the statistical analyses. These

Comparing the sampled ceramics with adjacent sequences, a similar pattern of divergence is evident. In the Early Classic period Picota and Motiepa phases at Palenque, the ceramic materials are characterized by the absence of polychrome gloss ware and the emergence of It is noteworthy that although gloss ware and polychrome types represent only 15.8% of the total sample from the three northern sites, there appears to be a clinal distribution, with gloss and polychrome ceramics present at 25.6% of the sample at Mirador, 15.7% at Revancha, and only 8.59% at Tiradero. 16

15 A notable exception to this pattern being the lack of medial or labial flanged vessels—typical of Late Formative Chicanel sphere ceramics—within the sample.

111

Archaeological Paleography

Attribute

Decoration (HV)

Site

Surface finish (HV)

Slip (HV)

Form/shape Morphology (HV) (LV)

Paste (LV)

Temper (LV)

flat base; ring/ conical supports; red; brown; cajete; bowl; polychrome calcite; rounded orange; pink; plate; San Claudio painted; glossy limestone lips; straight cream; red medium Z–angles undecorated divergent walls; everted rims Palenque/NW lowlands – – –– –– – – – –– Tierra Blanca – – – – – – +/– Piedras Negras ++ ++ ++ ++ ++ + ++ Altar de Sacrificios + ++ ++ +/– +/– + ++ Seibal +/– + + +/– +/– + ? El Mirador + + + + +/– + + Uaxactún + + + +/– + ? ? Middle Grijalva –– –– –– –– –– ? ? Chiapa de Corzo – – –– –– +/– ? ? KEY: ++ = strong correspondence; + = correspondence; +/– = slight correspondence coupled with slight discrepancy; – = discrepancy; – – = strong discrepancy; ? = insufficient data/unreported HV = high visibility attribute; LV = low visibility attribute Based on data reported in Acuña Smith 2005; Adams 1971; Forsyth 1989; Hernández Pons 1984; Holley 1987; Lee 1972; Muñoz 2002, 2004; Ochoa and Casasola 1991; Ochoa and Hernández 1975; Pérez Robles 2006; Rands 1972, 1987, 2007a, 2007b; Sabloff 1975; Smith 1955; Vargas and Hernández 1979. Figure 5.62. Comparative assessment of San Claudio assemblage at the regional level, Early Classic period (Tzakol sphere), illustrating correspondence and discrepancies between sampled ceramics from San Claudio and adjacent sites and regions.

Attribute

Decoration (HV)

Site

Tiradero, Mirador, striated; Revancha incised

Surface finish (HV)

Slip (HV)

Form/shape Morphology (HV) (LV)

rough; dull; waxy

unslipped; waxy matte reds and browns

jar; olla; basin; bowl

incurved/ vertical walls; flat base; no supports; bolstered/ flared rims

Paste (LV)

Temper (LV)

grey; sand; brown; fine calcite; to medium; quartz coarse

Palenque/NW lowlands

+/–

+

+

+

+

+

+/–

Tierra Blanca

++

++

++

+

+

++

+

Piedras Negras

––

––

–

––

–

–

––

Altar de Sacrificios

–

––

–

–

–

+/–

––

Seibal

+/–

–

+/–

+/–

+/–

+/–

?

El Mirador

––

+/–

–

+/–

–

–

–

Uaxactún

–

––

–

+/–

–

?

?

Middle Grijalva

–

+/–

––

––

––

?

?

Chiapa de Corzo

–

–

––

––

––

?

?

KEY: ++ = strong correspondence; + = correspondence; +/– = slight correspondence coupled with slight discrepancy; – = discrepancy; – – = strong discrepancy; ? = insufficient data/unreported HV = high visibility attribute; LV = low visibility attribute Based on data reported in Acuña Smith 2005; Adams 1971; Forsyth 1989; Hernández Pons 1984; Holley 1987; Lee 1972; Muñoz 2002, 2004; Ochoa and Casasola 1991; Ochoa and Hernández 1975; Pérez Robles 2006; Rands 1972, 1987, 2007a, 2007b; Sabloff 1975; Smith 1955; Vargas and Hernández 1979. Figure 5.63. Comparative assessment of Tiradero, Mirador, and Revancha assemblages at the regional level, Early Classic period (Tzakol sphere), illustrating correspondence and discrepancies between sampled ceramics from Tiradero, Mirador, and Revancha and adjacent sites and regions.

112

Interpreting the Results of the Comparative and Statistical Ceramic Analyses a regionally–specific and unique brown paste tradition (Rands 1987: 214–16, 2007b: 34). These developments are paralleled in the similar lack of polychrome ceramics and the prevalence of Playa Dull Ware and brown paste ceramics in the assemblages of Tiradero, Mirador, and Revancha at this time. Likewise, at both Palenque and the three northern sites within the study area, slips, when present, are usually matte red, and vertical groove– incision is the most common decorative technique (e.g., Laguna Verde Incised). The Pita and Lucha Incised types found only at the northern sites within the sample are also relatively common at Palenque. Rough finished jars and deep dishes with concave walls, interior beveled rims, and solid slab feet are distinctive in the Picota complex (Rands 2007a: 19), as they are at Tiradero, Mirador, and Revancha. Although the Formative period tradition of everted rim vessels continues well into the Early Classic period at Palenque, these forms merge with larger, steep walled utility ware and shallow, direct rim dishes. The basal flanges characteristic of Tzakol sphere ceramics are absent at Palenque and throughout the northwest lowlands.

(1987: 219) notes the presence of strong eastern (i.e., widespread lowland Maya–Tzakol sphere) affiliations, standing sharply apart from the local Picota and Motiepa pottery at Palenque and throughout the northwest lowlands. Nonetheless, Triunfo Striated jars, ubiquitous at Tiradero, Mirador, and Revancha, are widespread at these sites as well. Rands concludes that the differences in the Early Classic period ceramics of these smaller low sierra sites indicates the presence of important boundary phenomena, echoing an argument for the development of a more clearly defined and less porous boundary in or near the middle Usumacinta and lower San Pedro Mártir basin in the Early Classic period that I have made elsewhere (Englehardt 2010). Significantly, the mixed pattern of distribution of certain high visibility ceramic characteristics among these low sierra sites, San Claudio, and the sites to the north of the study area suggests a potentially clinal distribution of attributes across the study area and stretching into the greater northwest lowlands. In sum, the comparison of the sampled ceramic materials indicates a schism between the assemblages at San Claudio and those of Tiradero, Mirador, and Revancha. These latter sites seem typologically unrelated to the Tzakol sphere and appear more closely related to a developing northwest lowland tradition exemplified in the Picota and Motiepa phases at Palenque, more firmly rooted in earlier, Formative traditions and marked by a trajectory toward regionalism (Rands 1987: 214; 2007a: 19).

Moving into the later Early Classic period Motiepa phase, some Petén styles appear at Palenque, although in extremely limited quantities (Rands 2007a: 19). The locally produced brown paste tradition continues, and the everted rim vessels frequently found in Tzakol sphere assemblages, such as that at San Claudio, all but disappear (Rands 1987: 216). Flat based vessels with markedly thinner walls, bolstered or flared rims, vertical–walled short–necked jars, and fine textured, coarse pastes with carbonate–sand tempers become increasingly common at Palenque during Motiepa times, as they do in the materials from Tiradero, Mirador, and Revancha. The Early Classic period ceramics from these sites thus stand appear to take a direction contrastive to the Tzakol developments evident at San Claudio. Instead, these sites seem more closely aligned with the Picota and Motiepa phases at Palenque, the Taxinchán complex at Trinidad, or the Pomonaab phase at Pomoná, where gloss ware is equally uncommon (at only 5.56% of the assemblage) and monochrome red (68.7%) and fine brown paste ceramics are unusually well represented in Early Classic materials (García Moll 2005: 79; López Varela 1991; Rands 1972, 1987: 214). As at Tiradero, Mirador, and Revancha, Triunfo Striated is the dominant type at Pomoná, as well as the neighboring sites of Panhalé and El Arenal, comprising 82.5% of the Pomonaab materials (Anaya Hernández 2002; García Moll 2005: 79; López Varela 1991; Vargas 1985). Unslipped, striated brown and grey paste ceramics are also encountered frequently in the Early Classic period ceramics materials from the northwestern site of Tortuguero and throughout the Tulija River valley to the north of Palenque (Hernández Pons 1984).

The Early Classic period assemblage at Tierra Blanca is only partially incorporated into the Tzakol sphere (Ochoa and Casasola 1991: 11). During the Bakat phase at that site, and at other sites in the extreme lower Usumacinta during the Early Classic period (e.g., El Ocotlán, Paraíso, San Joaquín, and Jonuta), a fine paste tradition emerged, paralleling developments at Palenque and marked by the absence of polychrome ceramics and the abundance of monochrome red, brown, and unslipped pottery (Ochoa and Casasola 1991: 10–11; Sánchez Caero 1979). Although like Palenque, Mirador, Tiradero, and Revancha—the latter of which are located approximately 25km upstream from Tierra Blanca on the Río San Pedro Mártir—some Tzakol sphere types and attributes are noted, the vast majority of the Bakat complex appears unrelated to the burgeoning lowland Maya and Tzakol traditions, in stark contrast to San Claudio. Rough jars and utility ware such as ollas and basins are the dominant forms at Tierra Blanca, and Triunfo Striated is ubiquitous, just as at Tiradero, Mirador, and Revancha. Formal, morphological, and low visibility paste attributes are also similar between the Bakat complex and the sites to the north of the study area, which are themselves distinct from contemporaneous materials at San Claudio.

At several smaller sites in the low sierras 40km to the south and east of Palenque (e.g., Chancalá, Yoxihá), closer to San Claudio and Piedras Negras, Rands

In contrast, Pom (c. AD175–300) and Naba (c. AD 350– 560) phase ceramics at Piedras Negras are characterized by modes and forms diagnostic of the Tzakol sphere 113

Archaeological Paleography elsewhere in the Petén (Acuña Smith 2006; Muñoz 2004). Like San Claudio, the Early Classic period assemblage at Piedras Negras is marked by the appearance of polychrome ceramics and Petén Gloss Ware, which comprises 37% of the Pom complex and almost 75% of the Naba assemblage. The orange polychromes and Aguila Orange present in great quantities at San Claudio also dominate the Pom and Naba phases at Piedras Negras. Pastes, slips, tempers, and dominant modes of decoration evident in the Early Classic period ceramics from Piedras Negras are virtually indistinguishable from those at San Claudio at this time, and differ significantly from the same attributes in the northern assemblages. In contrast to the forms most common in the northern sites, at San Claudio and Piedras Negras, ollas with incurving walls and thickened rims, bolstered rim unslipped basins and utility jars, thin–walled molded rim bowls, shallow cajetes with hollow, conical tripod supports, and cazuelas and dishes with composite profiles and thick, nearly vertical rims occur frequently (Holley 1987: 189– 90; Muñoz 2004). Such vessel forms are rare elsewhere in the sample and in the Petén, and could represent an incipient localized style. Moreover, the medial– and basal–flanged bowls characteristic of the Tzakol ceramic sphere are rare within the sample and only appear at Piedras Negras—and Altar de Sacrificios—well into the latter half of the Early Classic, in the Naba and late Ayn– early Veremos phases, respectively (Adams 1971: 127; Muñoz 2004). This may again suggest a down–the–line lag or clinal distribution of certain diagnostic Tzakol morphological attributes in the Early Classic period.17

closer affiliations with those encountered in the Paixbancito and Acropolis complexes at El Mirador and the Tzakol phase at Uaxactún. At El Mirador, orange slips and polychrome painting emerge in the Early Classic period, and Petén Gloss Ware dominates the assemblage at 56% of the total materials (Forsyth 1989: 129).18 As at San Claudio, a re–orientation in ceramic types and modes is noted in the Paixbancito and Acropolis phases, with gloss ware replacing the waxy monochromes of the Late Formative period. Seven of ten ceramic types from the Early Classic assemblage are found exclusively at San Claudio in the sample, although Aguila Orange is relatively more common at El Mirador than at San Claudio, forming 28.1% of the Acropolis complex materials. Aguila Orange is virtually unknown at Tiradero, Mirador, and Revancha, and the type–varieties that the assemblages from these sites share with El Mirador (e.g., Balanza Black, Lucha Incised) are exceedingly rare in the Acropolis complex. Curiously, two types—Tinaja Red and Triunfo Striated—are shared extensively between the northern sites and El Mirador at this time. In fact, Uaxactún Unslipped Ware comprises 42.8% of the Acropolis phase materials, dominated almost entirely by Triunfo Striated, which is the most common type at El Mirador at 40.1% of the total assemblage. This somewhat odd mixture of Late Formative and Early Classic ceramics may stem from the fact that the Early Classic period materials at El Mirador rarely occur in pure deposits and are most often found mixed with earlier Late Formative Cascabel phase ceramics (Forsyth 1989: 129).19 Nevertheless, formally and morphologically, the Acropolis complex much more closely resembles the assemblage at San Claudio, particularly in the frequent occurrence of Z–angle vessels with annular ring supports and everted rims as well as tall–necked, smoothly– finished jars. Even the Triunfo striated materials are formally distinct from those at Tiradero, Mirador, and Revancha, with those at El Mirador consisting almost entirely of tall–necked jars with globular bodies and exterior beveled rims, in contrast to those found in the sample from the northern sites (Forsyth 1989: 130–1). Characteristic Tzakol sphere basal flanged bowls and dishes are also very common at El Mirador during the Early Classic period. Thus, although somewhat similar to many materials encountered at Tiradero, Mirador, and Revancha, the presence of these diagnostic Tzakol horizon markers, the fact that painting is the dominant decorative technique, the presence of a relatively greater number of shared type–varieties, and the emergence of

Like San Claudio, the Pom and Naba ceramics display increasing internal variability and a wider range of new attributes, likely stemming from the experimentation noted throughout the lowlands near the Late Formative– Early Classic period transition. Nevertheless, Early Classic period ceramics at both sites display a considerable degree of continuity with the preceding Late Formative period complexes, with Paso Caballos Waxy ware (27%), Uaxactún Unslipped, and types such as Sierra Red (14%), forming a significant minority of the Pom assemblage. This relatively minor continuity at San Claudio and Piedras Negras stands in stark contrast to the assemblages at Tiradero, Mirador, and Revancha, where the earlier waxy, unslipped tradition carries on well into the Early Classic period and forms the majority of the ceramic materials encountered at those sites during this time. When compared to the northern sites, the Early Classic period ceramics at San Claudio display considerably

It does bear mention that the Early Classic period Acropolis ceramic assemblage at El Mirador is relatively small, and its density and distribution are greatly restricted within the site, recovered from the latest, outermost features in the Tigre complex (Forsyth 1989: 129). 19 Alternately, the relatively greater degree of continuity with Late Formative ceramic traditions at El Mirador may simply be a reflection of a smaller sample size rather than the absence of a wider variety of types. As Forsyth (1989: 131) suggests, the relative scarcity of Acropolis complex pottery may indicate a decline in the population, occupational intensity, or influence of El Mirador during this time. 18

17 Such clinal distribution or a down–the–line lag is also suggested by the ceramics from the Sierra de Lacandón and Middle Usumacinta region directly to the south of the study area (e.g., at sites such as La Pasadita, El Cayo, and Yaxchilán), where diagnostic Tzakol attributes only appear well into the Early Classic period (Fournier 1987; Golden 2003; Golden and Scherer 2006; Golden et al. 1998; González Lauck 1993).

114

Interpreting the Results of the Comparative and Statistical Ceramic Analyses polychrome, orange–slipped ceramics firmly locate the Paixbancito and Acropolis phases within the Tzakol sphere and established classification of the east–central Petén, and suggest a closer affiliation with the Early Classic period ceramics at San Claudio.

of the Salinas and Ayn materials encountered across all operations at the site. That said, significantly more types are shared between Altar and San Claudio—including diagnostic Tzakol orange polychrome type–varieties— than between either of those sites and Tiradero, Mirador, or Revancha. Additionally, the ceramic groups represented in smaller quantities at these sites but not at San Claudio (e.g., Balanza, Actuncan, and Dos Arroyos) form a statistically minor percentage of the Early Classic period materials at Altar (Adams 1971: 130). It is further noteworthy that 90% of Salinas and Ayn phase ceramics are calcite tempered, paralleling the frequency of that temper type at San Claudio.

Tzakol phase materials at Uaxactún are also dominated by glossy polychrome types, orange slips with painted decoration, and morphological attributes such as Z– angles and annular ring bases (Smith 1955: 23). Gloss ware is abundant at 64.8% of the Tzakol 1 assemblage, and orange slipped types analogous to Aguila Orange occur frequently, comprising 24% of the Early Classic period materials at Uaxactún. As at many other lowland sites, the Tzakol phase appears to have emerged out of experimentation in ceramic production techniques near the Late Formative–Early Classic period horizon. Nevertheless, some continuity is noted, with unslipped ceramics accounting for 15.3% of the Tzakol 1 ceramics, primarily noted on grey colored, roughly finished jars. Although such materials are found frequently in the sample, they are most prominent at the northern sites of Tiradero, Mirador, and Revancha, where Late Formative period Chicanel–style ceramic traditions persisted well into the Early Classic period, as they did at other sites throughout the northwest Maya lowlands. Such forms and modes are decidedly less common at San Claudio at this time. Basal flanged vessels are ubiquitous during the Tzakol phase at Uaxactún, whereas such forms and morphology is nearly absent in the sample. The few flanged vessels that do occur are found almost exclusively at San Claudio, suggesting a closer affiliation of the ceramics of that site with the Tzakol sphere, compared to the assemblages of Tiradero, Mirador, and Revancha, which appear divergent.

The Early Classic period Junco complex at Seibal is poorly represented, due in large part to a decline in occupational intensity at this site during this time (Sabloff 1975: 14–15). Thus, in a possible parallel to the situation at El Mirador, the continuity of Formative traditions is particularly pronounced at Seibal, despite the fact that Sabloff affiliates the limited ceramic evidence with the Tzakol sphere. Nevertheless, the Junco materials are puzzling in relation to the sample, and do not appear to conform to the general pattern that emerges through comparison of the sample with the ceramic assemblages of other lowlands sites during the Early Classic period. For example, although Petén Gloss Ware is introduced and glossy surface finishes with polychrome painted decoration are prevalent during Junco times, the type components do not find many parallels throughout the lowlands, except at Altar de Sacrificios, some 40km distant. Nevertheless, the dominant ceramic types, Actuncan Orange, Dos Arroyos Orange, and Balanza Black, are not well represented at Altar or elsewhere. Moreover, unslipped and roughly finished dull pottery such as Triunfo Striated and San Martín Brown are, unusually, relatively common at Seibal during the Junco phase. Although bowls with Z–angles and basal flanges appear among the Junco materials, these forms and morphologies are in the minority, whereas unslipped, striated, short–necked jars with slightly outcurved necks are common. The Early Classic period assemblage at Seibal curiously appears slightly more closely related to the contemporary ceramics at the northern sites of Tiradero, Mirador, and Revancha within the sample, although significant discrepancies between the sample and the Junco materials from Seibal are widespread.

At Altar de Sacrificios during the Early Classic Salinas, Ayn, and Veremos phases a similar pattern is evident. The Salinas phase witnessed the introduction of new ideas and ceramic technologies, such as polychrome ceramics, representing a rupture with the previous Late Formative traditions of the Plancha phase (Adams 1971: 93–4). Several types of orange polychrome appear at this time, including the Tzakol horizon marker Aguila Orange, and together comprise 34.2% of the Salinas complex. Aguila Orange becomes even more widespread in the subsequent Ayn complex, forming 39.5% of that assemblage. More elaborate types and decorations are also introduced during the Ayn phase, including the ring–based Z–angle bowl common in the Early Classic period assemblage at San Claudio (Adams 1971: 97). Usulutan–style forms and decoration, such as mammiform supports, is also significantly more common at Altar than within the sample. Nevertheless, some continuity with preceding traditions is manifest, with unslipped types such as Sapote and Triunfo Striated remaining common throughout the Early Classic sequence at Altar through the Veremos phase. Triunfo Striated, in fact, forms an average of 35%

Comparing the sample against assemblages at sites outside the Maya lowlands, it is clear that links with such areas are few, and decline dramatically in the Early Classic period (Adams 1971: 127). The sequence at Cuello, for instance, indicates increasing experimentation and a re–orientation toward the developing regional ceramic traditions of northern Belize (e.g., Holmul I and Floral Park; Kosakowsky 1987: 28). Although certain resemblances to Tzakol 115

Archaeological Paleography sphere ceramics are evident in the Early Classic period Cuello assemblage, this is likely due less to interaction, contact, or shared ceramic technologies than to independent reception of influences from a common area, the northern Petén (Adams 1971: 130). Usulutan– style attributes and characteristics are also much more pronounced in the Floral Park sphere (Demarest and Sharer 1982; Neff et al. 1999). To the west, at Jiquipilas phase Chiapa de Corzo and in the Early Classic period Juspano assemblage in the Middle Grijalva region, a differential firing tradition emerges that is absent in the Maya lowlands. The Yahama Rough and Yatsipo Sandy types that emerged in Late Formative times continue to dominate the Early Classic period assemblages. Flat bottom bowls and tetrapod cylinders are the most common forms. Although such forms are found in the greater Maya lowland area, they are rare, and occur infrequently and in small quantities in Early Classic period assemblages. Lee (1972: 13–14) concludes that the appearance of differentially fired pottery indicates an emerging affiliation of the ceramic traditions of the Middle Grijalva region and western Chiapas in general with the Los Tuxtlas region of southeastern Veracruz, the western coast of Campeche, and, possibly, the incipient grey–black Early Classic period pottery tradition of Oaxaca. In either case, it is clear that the ceramics of the regions to the west of the greater Maya lowlands become increasingly divergent when compared to Early Classic period Tzakol traditions.

Summary of Comparative Analysis The comparative evaluation generally confirms the results of the statistical analysis. Overall, a striking degree of uniformity, both in terms of common type components, morphology, formal characteristics, and ceramic attributes and their distribution throughout the lowlands, is noted in the Middle and Late Formative periods between the sample and the ceramic assemblages at adjacent sites and regions. This inter– assemblage continuity and similarity is likely due to the widespread distribution of the Mamom and Chicanel spheres, both of which exercised a significant degree of influence on developing lowland ceramic traditions during the Formative period. In the Early Classic period, increasing variability and divergence is pronounced in both the sample and regional sequences. The ceramic traditions of the lowland Maya area also begin to show marked differentiation when compared to other areas of Mesoamerica. At this point, ceramic attributes become discretely distributed as assemblages across the Maya world broke with preceding Formative traditions and experimented with new forms and modes of ceramic production. This rupture is slightly more pronounced within the sample at San Claudio, where the Early Classic period assemblage is firmly affiliated with the Tzakol sphere. Ceramics at Tiradero, Mirador, and Revancha, on the other hand, appear to orient themselves toward a developing northwest lowland regional tradition. Although some attributes and type components are shared between these sites to the north of the study area and adjacent sequences, a much lower degree of distributional uniformity is evident. This fact explicates the partial adscription of the ceramic sample into the Tzakol sphere.

In general, the variability between Early Classic period ceramic materials from San Claudio and the sites of Mirador, Revancha, and Tiradero closely parallels the traits identified by Holley (1987: 188–9) that distinguish the pottery of Piedras Negras and Altar de Sacrificios from the northwestern Maya lowlands as a whole. These include the relative thickness of vessel walls, jar morphology, basin elaboration, and surface treatment. Vessel walls are decidedly thinner at Tiradero, Mirador, and Revancha, and do not demonstrate the typical dichotomy between thin walled serving ware and thick walled utility ware evident at San Claudio and Piedras Negras. Jars in the assemblage at San Claudio have longer necks than those at Mirador, Revancha, and Tiradero, and basins at San Claudio are relatively simple in comparison to the shouldered shapes and elaborate rims evident at the other sites and assemblages throughout the northwestern lowlands. Finally, the polychrome and orange slip traditions evident at San Claudio and characteristic of Early Classic lowland conventions in the Petén are noticeably lacking at Mirador, Revancha, and Tiradero. In sum, the Early Classic ceramic sequence at San Claudio appears more closely related to Tzakol sphere assemblages at Piedras Negras and Altar, whereas the remaining sampled materials seem to be associated with a developing northwestern tradition evident at Palenque, Tierra Blanca, and the greater lower Usumacinta region.

Patterns of Interaction and Innovation Revealed through the Ceramic Analyses Based on the statistical results and comparative evaluation presented above, the general patterns of variability over time and through space that emerge from these analyses are presented, both within the sample and between assemblages at the regional level. This section concludes with a discussion of the implications of those patterns for issues of interaction and material innovation. To recapitulate, measured ceramic attributes were grouped into five distinct classes—classificatory, paste, formal, morphological, and dimension—and into high and low visibility attributes,20 following Carr (1995), Parkinson (2002: 400, table 1), and Voss and Young (1995; see also Parkinson 2006: 37, table 1). In both the sample and regional sequences, the data illustrate increasingly greater degrees and ranges of variability over time. Per the interpretive framework discussed previously, such High visibility attributes include the classificatory and formal classes, particularly decoration, surface treatment, slip, and form, whereas low visibility attributes include morphological, paste, and dimension. 20

116

Interpreting the Results of the Comparative and Statistical Ceramic Analyses diachronic variability suggests a much greater amount of interaction over a wider spatial area in the earlier temporal contexts of the Middle and Late Formative periods, followed by a drastic decrease in interregional interaction after the Late Formative–Early Classic period transition.

not always correspond exactly, insofar as ceramic sequences are, to some degree, artificial constructions. Nevertheless, changes in attribute categories and classes throughout lowland Maya ceramic assemblages become markedly more pronounced toward the middle of the Early Classic period, c. AD 400.

Variability over Time

Thus, both the statistical and comparative analyses indicate a trend toward variability and diversity in ceramic materials over time. Uniformity in frequency of attributes and continuity in the distribution of those attributes characterize the Middle and Late Formative period ceramic materials within the sample and across the lowland Maya world. This uniformity and widespread similarity suggests a that a great deal of interaction occurred across southeastern Mesoamerica during the Middle and Late Formative periods. In contrast, the Early Classic period is marked by increasing inter–assemblage variability and intra– assemblage diversity. Such variability, coupled with the widespread experimentation noted in and around the Late Formative–Early Classic period horizon, suggests a diminishing degree of interregional interaction, as well as a significantly greater degree of localized material innovation in lowland ceramic traditions.

Diachronic variability in ceramic materials within the sample is illustrated by Figures 5.64–5.67, which display clustered boxplots of ANOVA–derived ranges of attribute variability for each attribute class for the sample as a whole over time. These graphs indicate that the ranges of attribute values become increasingly wider over time, suggesting a progressively greater degree of variability within the sample across all attribute categories and classes moving from the Formative through the Early Classic period. These results complement the H Score analysis (Figure 5.59), which again illustrates strikingly little diversity within the sample in earlier temporal contexts followed by increasing intra–assemblage heterogeneity during the Early Classic period. A diachronic plot of attribute means (Figure 5.68) illustrates progressively greater means over time, indicating an increasing degree of intra–assemblage variation, again complementing the H Score results.21

Variability through Space

The diachronic variability across attribute categories and classes within the sample illustrated by these statistical results is paralleled in the outcomes of the comparative assessment at the regional level. Across the Maya lowlands during the Middle and Late Formative periods, ceramic attributes display remarkable uniformity in distribution and frequency of occurrence within the assemblages present at sites and regions adjacent to the study area. A trend toward experimentation in ceramic production techniques is evident at many lowland Maya sites nearing the Late Formative–Early Classic period horizon, and a much greater degree of divergence is noted between the sample and regional assemblages in the Early Classic period itself. At this point, the sample becomes split, with San Claudio firmly ensconced in the Early Classic period Tzakol ceramic sphere characteristic of the central lowlands of the Petén, and the sites of Tiradero, Mirador, and Revancha oriented toward a developing northwestern lowland regional tradition. The shifts in ceramic production styles and techniques evident in the Early Classic period did not occur contemporaneously at all sites, and temporal subdivisions and phases in ceramic sequences do

Considering diachronic ceramic similarity and variability in spatial terms, a comparable pattern becomes apparent. The distribution of characteristic Middle Formative period ceramic attributes is widespread and uniform across the whole of southeastern Mesoamerica (Figure 5.69). Typical Mamom sphere types, forms, and techno–stylistic attributes, such as waxy read and cream slipped ceramics with little decoration on simple forms are found across a wide swathe of the lowlands. The geographic expanse of this tradition encompasses the area previously occupied by the Early Formative period Locona tradition (see Figure 3.2), stretching south and west through the highland regions of Chiapas and Guatemala and past the Isthmus of Tehuantepéc and extending north and east into the central and northern Maya lowlands and across the Yucatán peninsula. Such a uniform distribution of shared ceramic attributes during the Middle Formative period implies extensive interregional interaction throughout southeastern Mesoamerica at this time. The general continuity between Early and Middle Formative period ceramic conventions, as well as the absence of significant attribute or distributional variability within the sample or between regional assemblages, suggests a lack of specifically local innovation in ceramic production or technologies during the Middle Formative period.22

21 Although the mean values of some attributes (e.g., decoration) may appear to decrease over time as reflected in Figure 5.68, this is simply a result of the fact that the absence of certain attributes was assigned an artificially high numerical value. In this case, then, a decrease in the mean attribute value over time indicates the presence of more significant variability across this attribute category in later temporal contexts.

Although some innovations in ceramic production (e.g., bichrome slipping, resist decoration, necked jars) do appear first in the Middle Formative period, these new techniques largely developed 22

117

Archaeological Paleography

Figure 5.64. Clustered boxplot displaying ranges of classificatory attribute variability over time.

Figure 5.65. Clustered boxplot displaying ranges of paste attribute variability over time.

118

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.66. Clustered boxplot displaying ranges of formal attribute variability over time.

Figure 5.67. Clustered boxplot displaying ranges of morphological attribute variability over time.

119

Archaeological Paleography Thus, Figure 5.70 more accurately represents the distribution of Chicanel attributes during the latter portion of the Late Formative period. It is also at this time—sometimes labeled the ‘Protoclassic’—that the marked increase in experimentation with new ceramic forms and technologies is noted throughout lowland Maya ceramic assemblages (Pring 1977). Nevertheless, for the majority of the Late Formative period, uniformity in attributes and distribution of those attributes across the region, coupled with a distinct lack of intra–sample or inter–regional variability, suggests a continuation of the wide–ranging interaction that characterized the Middle Formative period. Even the variations in southern highland ceramic traditions that emerged in the later part of the Late Formative period were uniformly distributed and shared throughout the lowlands, unimpeded by any social or cultural boundaries. Innovation in lowland ceramic traditions is again evident during this time, particularly in terms of decorative techniques and formal attributes, but such innovation was uniform across the Maya world. Additionally, despite the emerging complexity in Maya society during the Late Formative period (Fahsen Ortega 1999; Grube 1999), there is little archaeological evidence that indicates any integration into a pan–lowland regional system that would explain uniformity in ceramic traditions. Although absence of evidence is not evidence of absence, the data currently available suggest that interregional interaction is the most likely explanation for this phenomenon.

It is tempting to view such uniform distribution of attributes and techniques as indicative of the presence of a large regional system into which the sites that exhibit evidence of shared ceramic technologies were integrated. Archaeological evidence, however, suggests that no such hegemonic cultural system existed at this time.23 In the Late Formative period, the distribution of characteristic Chicanel ceramic attributes is similarly uniform across lowland southeastern Mesoamerica (Figure 5.70). The geographic extent of the Chicanel sphere was comparatively more limited in relation to the preceding Mamom tradition, becoming confined almost exclusively to the greater Maya lowlands in the later part of the Late Formative period (i.e., after AD 1). At the same time in the highlands of Chiapas and Guatemala, the Providencia/Miraflores ceramic sphere and the Usulutan style came to prominence (Demarest 1986, 2002; Demarest and Sharer 1982; Neff et al. 1999). Although slightly different than Chicanel ceramics, these highland traditions were equally an outgrowth of preceding Middle Formative and early Late Formative period conventions. In addition, there is evidence of widespread sharing of these new styles, and their distribution, although limited, is relatively uniform throughout the lowlands. In any case, both spheres are characterized by a uniform distribution of attributes and the prevalence of thick waxy red, cream, and black slipped ceramics with groove– incised, fluted, and striated decoration. Similar vessel forms and morphologies also dominate assemblages across southeastern Mesoamerica during the whole of the Late Formative period.24 To the west in the Middle Grijalva basin, continuity with Chicanel sphere ceramics is apparent through the Guañoma phase until roughly AD 1, corresponding to the Guanacaste and Horcones phases at Chiapa de Corzo. After this point, a rupture with southeastern lowland and highland traditions becomes progressively more evident.

In the Early Classic period increasing variability in and between ceramic assemblages suggests a much more limited extent of particular lowland ceramic traditions (Figures 5.71 and 5.72). At this time, the uniformity in attributes across regions that characterized the Middle and Late Formative periods gave way to a more discrete pattern of distribution. Early Classic period ceramic assemblages across the lowlands also display a much greater range of attributes, both within the sample and in adjacent areas. Although a degree of continuity with earlier traditions is evident, the experimentation evident around the Formative–Classic period horizon led to a greater deal of innovation in ceramic production techniques during this time. Red slips were replaced by orange ones, waxy surfaces were discarded in favor of glossy finishes, polychrome painting usurped simpler forms of decoration, and an array of new forms and morphological features, such as the ubiquitous basal– flanged bowls and tripod–supported cylindrical vessels, came to dominate Tzakol sphere ceramics in the central Petén lowlands. The distribution of these attributes is discrete, and they are not often found outside of the central lowlands at this time. The innovation noted in the Early Classic period appears to be localized in nature, specific to this particular area at this point in time. Such discrete attribute distribution indicates a dramatic decline in the amount of interregional interaction that occurred in the Early Classic period.

simultaneously from preceding Early Formative period traditions and are found across assemblages throughout southeastern Mesoamerica; their occurrence is infrequent and the distribution of such attributes is relatively uniform. 23 Some sites within the Maya lowlands (e.g., El Mirador) certainly do appear to have begun to grow into larger, complex urban centers, perhaps approaching state–level societies as early as the later part of the Middle Formative period (Hansen 1991). The same may also be true of sites such as San José Mogote in the Valley of Oaxaca (Flannery and Marcus 2000). Despite this fact, there is no evidence to suggest that either site exercised any pan–regional hegemony into which other lowland sites were integrated. 24 The emergence of a somewhat unique ceramic tradition in the Guatemalan highlands and along the south coast at this time may parallel the development of the Izapan/South Coast scribal tradition at sites such as Kaminaljuyú. The broad relationship between Providencia/ Miraflores and Chicanel ceramics may mirror the close relationship between the Izapan/ South Coast and Early Maya writing systems.

120

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.68. Diachronic attribute means plot, illustrating increasing variability over time.

Figure 5.69. The spatial distribution of common Middle Formative period ceramic attributes (e.g., waxy red slip with little decoration on simple forms). The overlay of attribute distribution generally reflects the range of the Mamom ceramic sphere and typical Mamom types, forms, and techno–stylistic attributes. The study area is outlined in the black cross–hatched box.

121

Archaeological Paleography

Figure 5.70. The spatial distribution of common Late Formative period ceramic attributes (e.g., thick waxy red, cream, and black slips, groove–incised, fluted, and striated decoration, flat–based shallow dishes and deep bowls with thick everted rims and divergent walls, and divergent walled utility ware). The red overlay of attribute distribution generally reflects the range of the Chicanel ceramic sphere and typical Chicanel types, forms, and techno–stylistic attributes. The study area is outlined in the black cross–hatched box.

For example, in the lowlands of Belize to the east, assemblages are dominated by the variant Holmul style and the Floral Park ceramic sphere, in which Early Classic period Tzakol–style ceramics are not well represented. Rather, these ceramic materials present a non–glossy slip treatment that is derived directly from the Late Formative tradition. Additionally, there is evidence of the continued production of monochrome red pottery at this time that approaches the glossiness and hardness of Early Classic ceramics but within the slip color range for Sierra Red (Kosakowsky 1987). The persistence of Late Formative ceramic types well into the Early Classic has been suggested elsewhere in both Belize and the northern Yucatán (Matheny 1970; Robles Castellanos 1990). In fact, the northern lowlands appear to have entered a hiatus during the Early Classic period, with ceramic materials scarce in assemblages during this time. In the southern and southeastern highlands of Guatemala and Chiapas, Usulutan–style ceramics and divergent forms not typically associated with Tzakol sphere materials, such

as mammiform supported tetrapod bowls, are prevalent. The Early Classic period Aurora and Esperanza phases, in particular at Kaminaljuyú display an abrupt change in ceramic types and the beginnings of a tradition with no local antecedents, perhaps the result of an intrusive population or the blocking of trade routes (Braswell 2003b). In any case, the Early Classic period ceramics of the southern highlands are scarce, and characterized by a relative lack of development or elaboration. In addition, Usulutan ceramics became increasingly rare in assemblages at other lowland sites in the Early Classic period, suggesting a restricted amount of interaction. Finally, in the Middle Grijalva basin and at Chiapa de Corzo to the west, the ceramics of the Juspano and Jiquipilas phases continue the pattern of divergence begun in the later part of the Late Formative period. The Yahama Rough and Yatsipo Sandy types prevalent at this time are unrelated to the Tzakol sphere, and more closely affiliated with the ceramic assemblages of southeastern Veracruz, Oaxaca, and the western Campeche coast (Lee 1972: 13–14; Matheny 1970).

122

Interpreting the Results of the Comparative and Statistical Ceramic Analyses

Figure 5.71. The spatial distribution of divergent central (Tzakol; dark grey overlay) and developing northwestern (stippled light grey overlay) Maya lowland ceramic spheres in the Early Classic period. The study area is outlined in the black cross–hatched box. The bold dashed line suggests the presence of a boundary between these ceramic traditions, running through the study area. Cf. Willey et al. 1967: 311, fig. 7.

Within the sample from the northwest Maya lowlands, a similar decline in interaction is suggested by the ceramic evidence. In general, the variability between Early Classic period ceramics from San Claudio and the sites of Tiradero, Mirador, and Revancha parallels the traits identified by Holley (1987: 188–89) that distinguish the pottery of the central Petén Tzakol tradition from the northwestern Maya lowlands as a whole. These include the relative thickness of vessel walls, jar morphology, basin elaboration, and surface treatment. Vessel walls are decidedly thinner at the northern sites and do not demonstrate the typical dichotomy between thin walled serving ware and thick walled utility ware evident at San Claudio and Piedras Negras. Jars in the assemblage at San Claudio have longer necks than those at Mirador, Revancha, and Tiradero, and basins at San Claudio are relatively simple in comparison to the shouldered shapes and elaborate rims evident at the other sites. Finally, the polychrome and orange slip traditions evident at San Claudio and characteristic of Early Classic period lowland conventions in the Petén are noticeably lacking at Tiradero, Mirador, and Revancha. Overall, the Early Classic ceramic sequence at San Claudio

appears more closely related to Tzakol sphere assemblages, particularly those at Piedras Negras and Altar (Figure 5.62), whereas materials from the remaining sampled sites seem to be associated with a developing northwestern tradition evident at Palenque and the lower Usumacinta region (Figure 5.63). As I have suggested elsewhere (Englehardt 2010), these data intimate the formation of a boundary that separated these ceramic traditions at or near the Late Formative–Early Classic period transition, impeding interaction and the exchange of ideas or materials between regions (Figures 5.71 and 5.72). The widespread similarity and uniformity in the distribution of ceramic types and forms throughout the Maya lowlands and beyond during the Middle and Late Formative periods implies a great degree of interaction over a large swathe of southeastern Mesoamerica that does not appear to have been hampered or affected by any recognized sociocultural or political boundary. After the Late Formative– Early Classic period transition, a regional separation of sequences and a greater degree of variability and discrete distribution both within the sample and in 123

Archaeological Paleography

Figure 5.72. Detail of the central Petén and northwestern Maya lowlands, illustrating the geographic extent of the Early Classic period Tzakol sphere (dark grey overlay) and the developing divergent ceramic tradition of the northwestern Maya lowlands (cross–hatched light grey overlay). The bold dashed line suggests the presence of a boundary between these traditions, running through the study area.

relation to ceramic assemblages at adjacent sites and regions indicates much less interaction within and across the region during the Early Classic period. The evident correspondence between ceramic materials from San Claudio and those at neighboring lowland centers such as Piedras Negras and Altar de Sacrificios suggests that interaction within the central lowlands continued at this time. The lack of parallels between the assemblages from Tiradero, Mirador, and Revancha and those of the central Petén Tzakol sphere, coupled with the formal and stylistic similarities of those assemblages with materials encountered to the north and west, implies that the intermediate plains of the mid–lower Usumacinta and lower San Pedro Mártir basins north of the low Sierra del Lacandón foothills was a frontier at which interaction between regions was restricted in the Early Classic period. Sites on or near this boundary, such as San Claudio, continued to interact with the central lowlands, generally imitating core trends. The innovation and incipient localized Early Classic styles noted San Claudio and sites such as Piedras Negras and Altar, themselves heavily influenced by the Tzakol sphere trends of the central Petén, speak at once to both their increased integration in a larger, more complex and centralized social system

and their relative isolation and situation at a peripheral area. Moreover, the seemingly clinal distribution and temporal lag in ‘down–the–line’ emergence of Petén–based traits (e.g. the relatively late appearance of basal–flanged bowls at Altar de Sacrificios, and, subsequently, Piedras Negras, and their near absence in the Early Classic period sample from the study area) indicate an emphasis on stricter control and restricted flow of diffused technologies from the core to the periphery as Maya society became increasingly complex, centralized, and inward–focused. Conclusions In sum, both the statistical and comparative analyses of the sampled ceramic materials indicate marked homogeneity, a near absence of statistically significant variability, and striking uniformity in the distribution of individual attributes throughout Middle Formative period ceramic assemblages across lowland southeastern Mesoamerica. Such homogeneity and distributional uniformity continued into the Late Formative period. Although a certain degree of innovation in ceramic production is suggested by the Late Formative period data, in

124

Interpreting the Results of the Comparative and Statistical Ceramic Analyses technologies and production techniques becomes evident. These Early Classic period developments correspond to what appears to be increased control of the western, southern, and southeastern peripheries, respectively reflected in the complete divergence of Juspano and Jiquipilas phase ceramics in the middle Grijalva basin (Lee 1972: 13–14) and the lack of formal or stylistic similarity between the Tzakol sphere and post–Usulutan ceramics on the south– eastern periphery and the materials evident in the Aurora and Esperanza phases at Kaminaljuyú in the highlands to the south.

terms of greater ranges of attribute types and the introduction of new technologies, forms, aspects of vessel morphology, and modes of decoration, there is a great deal of continuity between Middle and Late Formative period ceramic traditions. Additionally, these comparatively innovative attributes were widely shared and evenly distributed across the lowlands, suggesting little locally specific innovation. The spatial and temporal patterns that emerge from the Middle and Late Formative period ceramic analyses thus suggest the existence of extensive interregional interaction throughout the Formative period, both within the study area and across the lowlands (Figure 5.73). Despite the pronounced lack of variability in Formative ceramic traditions, available archaeological evidence does not suggest that integration into any larger, hegemonic cultural system is responsible for the observed uniformity between assemblages.

A boundary between ceramic traditions also appears to run through the study area, reflected in the significant differences between Early Classic period ceramic materials at San Claudio vis à vis those from the northern sites of Tiradero, Mirador, and Revancha. This frontier separated the northwest Maya lowlands from the developing core area of Classic Maya society centralized in the Petén, with sites in the low sierras south of the intermediate plains such as San Claudio becoming more integrated in Classic Maya traditions, and sites to the north of the plains orienting themselves outward. It appears that the development of this boundary was coeval with trends of increasing complexity, integration, and centralization in the Maya lowlands that occurred roughly at the Formative– Classic period transition. San Claudio is integrated into the Tzakol sphere and its materials find numerous correspondences with ceramic materials from other sites in the central Petén. Tiradero, Mirador, and Revancha, on the other hand, appear more related to a

Starting with the experimentation in ceramic production techniques observed in the later part of the Late Formative period and continuing into the Early Classic period, the data indicate statistically significant increases in both variability and heterogeneity within the sample, coupled with an increasingly discrete distribution of ceramic attributes throughout the lowlands following the Formative–Classic period horizon. The comparatively greater degree of variation within the sample, and between the sample and adjacent regional ceramic sequences, suggests a marked decline in interaction during the Early Classic period (Figure 5.74). Also at this time, a significant amount of locally specific innovation in ceramic

Temporal Context Variable Scale Attribute and visibility Decoration (HV) Surface finish (HV) Slip (HV) Form/shape (HV) Morphology (LV) Paste (LV) Temper (LV)

Middle Formative Period Distribution within unit* uniform uniform uniform uniform uniform uniform uniform

between units* uniform uniform uniform uniform uniform uniform uniform

Late Formative Period

Interpretation of pattern interaction interaction within between units* units* high high high high high high high

high high high high high high high

Distribution within unit* uniform uniform uniform uniform uniform uniform uniform

between units* uniform uniform uniform uniform uniform uniform uniform

Interpretation of pattern interaction interaction within between units* units* high high high high high high high

high high high high high high high

Figure 5.73. Interpretation of patterns of interaction within and between units deduced from ceramic analyses for the Middle and Late Formative periods. *At the regional scale, units refer to distinct ceramic traditions (or cultural groups). At the micro–regional scale, units refer to the four sites as a clustered whole (when compared with site clusters from adjacent regions). At the local scale, units refer to the individual sites. After Parkinson 2002: 400, Figure 1, 2006: 37, Figure 1 (Carr 1995; Voss and Young 1995).

125

Archaeological Paleography

Temporal Context

Early Classic Period

Variable Scale Attribute and visibility Decoration (HV) Surface finish (HV) Slip (HV) Form/shape (HV) Morphology (LV) Paste (LV) Temper (LV)

Distribution within unit* discrete clinal discrete discrete discrete discrete discrete

Interpretation of pattern

between units* discrete discrete/clinal discrete discrete discrete/clinal discrete discrete

interaction within units* low low/moderate low low low low low

interaction between units* low low/moderate low low low low low

Figure 5.74. Interpretation of patterns of interaction within and between units deduced from ceramic analyses for the Early Classic period. *At the regional scale, units refer to distinct ceramic traditions (or cultural groups). At the micro–regional scale, units refer to the four sites as a clustered whole (when compared with site clusters from adjacent regions). At the local scale, units refer to the individual sites. After Parkinson 2002: 400, Figure 1, 2006: 37, Figure 1 (Carr 1995; Voss and Young 1995).

burgeoning local tradition specific to the northwestern Maya lowlands. The unique nature of the ceramic sample from the study area, and the variability between the assemblages represented therein, suggests two primary conclusions. First, that significantly less interaction occurred across the lower San Pedro Mártir basin during the Early Classic than in preceding periods. Instead, interaction was impeded by a less permeable and more strictly defined boundary in the Early Classic period. Second, a considerable amount of localized innovation in ceramic production and technologies emerged at this time in and around this newly defined border region, and throughout the more strictly delineated southern central lowlands of the Petén.

The identification of the extent and role of interaction in the development of cultural innovations such as new ceramic technologies represents one half of the objectives of this investigation. In the case of the ceramics from the lower San Pedro Mártir basin, the data support the contention that shifting patterns of interregional interaction were indeed involved in processes of innovation, particularly in and around the Formative–Classic period transition. Returning to the analogy mentioned in Chapter Two regarding the development of writing, material innovation in this case appears to have occurred after a widely shared “open” ceramic complex became “closed” within a specific sociocultural context, which fostered the emergence of locally particular forms and functions for a previously widely–shared material cultural tradition. This process parallels current models of script development.

The data thus suggest that innovations in ceramic traditions and production techniques occurred only after more than a millennium of sustained, extensive interregional interaction across southeastern Mesoamerica. Areas in which less interaction is evident through time, such as the extreme northern and eastern lowlands of the Yucatán peninsula and the far southeastern periphery of central Honduras and eastern El Salvador, exhibit noticeably less innovation in the Early Classic period, tending instead to continue previous Late Formative traditions. The data suggest a positive correlation between innovation, amplified material variation, and sustained interregional interaction. Additionally, the noted increase in variation and decrease in interaction through time appear connected to diachronic changes in socio– political organization, namely increased centralization and integration into a larger, increasingly complex, and specifically lowland Maya cultural system.

If there is a correlation between developmental processes in this case, then the patterns of material interaction revealed through qualitative and quantitative analyses of archaeological indicators of material interaction should correspond closely, in spatial and temporal contexts, to the patterns of iconographic and linguistic interaction involved in the development of Maya writing. Comparing patterns of variability across datasets within and between sites and regions in differing temporal contexts ultimately allows for the evaluation of the hypothesis that interaction drives both material and ideational innovation. If this hypothesis is correct, one would expect patterns of variability in the data to correspond in both spatial and temporal contexts. Specifically, one would expect to observe less variability in all data in earlier temporal contexts, indicating more interaction,

126

Interpreting the Results of the Comparative and Statistical Ceramic Analyses followed by increasing variation that implies a greater degree of innovation, decreasing interaction, and increasing integration into more localized cultural systems in the later temporal contexts of the crucial Late Formative–Early Classic period transition. Before exploring such potential correlations between patterns

and processes of iconographic, linguistic, and material change, however, the temporal and spatial patterns of interaction and innovation evident in contemporaneous linguistic and iconographic data involved in the emergence of the Mayan script must first be explored. It is to this evidence that we now turn.

127

Chapter 6 Comparative Analysis of Iconographic and Linguistic Evidence The comparative and quantitative analyses detailed in the preceding chapter elucidate patterns of interaction reflected in the ceramic data at different points in time. The quantitative analyses also permit an appreciation of innovation as expressed through changes in ceramic materials over time, particularly in terms of the introduction of new technologies and design elements. In terms of the development of the Mayan script, linguistic and iconographic data serve a similar end. Just as a broad distribution of a specific ceramic type, attribute, or form suggests a high degree of interaction between groups, traditions, or regions, the identification of particular linguistic or iconographic elements that are widely shared and broadly distributed across the landscape also suggests the operation of processes of interaction, and are potentially revelatory of the role of such processes in script development. This chapter presents and discusses patterns of interaction and innovation as revealed through linguistic and iconographic data related to the development of Maya writing.

In Mesoamerican contexts, current models hold that the diversity of regional scripts developed from an ‘ancestral’ Middle Formative iconographic complex or Olmec script, which was widely shared throughout Formative period Mesoamerica and inherited by various groups who subsequently employed its forms and conventions as the basis for other representational and script traditions, including Maya writing (Figure 6.1; cf. Coe 1965:769, fig. 57, 1976: fig. 1; Houston 2004a: 286; Justeson 2012: 838; Lacadena 1995, 2010; Mora–Marín 2001: 444–46, figs. 1.7–1.9). Processes of borrowing, adaptation, abstraction, modification, or imitation of this ancestral system among and between multiple groups fostered the conditions necessary for the development of glottographic writing. As representational systems became increasingly disassociated from the interpretive framework of a shared precursor iconography, scribes and readers could exploit emergent tension between visual and auditory perception, assigning new values based primarily on the phonetic or word (i.e., linguistic) values associated with the sign or sign sequence (Carrasco and Englehardt 2015: 4; Jakobson 1971a; Justeson 2012; Justeson et al 1985; Justeson and Matthews 1990: 106; Robertson 2004). In this respect, the development of Mesoamerican writing appears to follow similar patterns in the development of other world scripts (Dixon 1997; Schmandt–Besserat 1992; cf. Bagley 2004; Baines 2004; Cooper 2004; Lamberg–Karlovsky 1986).

To briefly recapitulate the processes of script development presented in Chapter Two, writing emerges from a ‘crisis of meaning,’ in which iconic elements become ‘divorced’ from artistic contextualizing frameworks. In such a situation, an audience can no longer rely on familiarity with the visual interpretive principles of established iconographic conventions to sustain meaning adequately or to determine the relationship among signs. To resolve this interpretive tension, readers may infuse icons with grammatical or linguistic values, thus facilitating their ‘recontextualization’ in the more discretely organized contexts of scribal conventions, in which new semantic meanings or phonological values potentially adhere to a sign that previously depended on visual or iconographic interpretive frameworks. (Carrasco and Englehardt 2015; Englehardt 2005; Hopkins 1997; Justeson 2012, Justeson et al. 1985: 34; Robinson 2003; Rogers 2005). In effect, a system of visual communication previously open to interpretation or visual exploitation among and between multiple audiences or languages becomes ‘closed’ within a single one, in which ‘the organization of signs conforms to the structure of language as opposed to the representational schemes of which a given sign was originally an iconographic element’ (Carrasco and Englehardt 2015: 4; Houston 2004b). In many cases, although the written sign assumed a new value within the context of a script, it retained a visual connection to its former iconic significance (Miller 1989; Robertson 2004: 30; Robinson 2003; cf. Figures 2.4–2.6).

If one accepts this model for the development of writing in Mesoamerica, then it is eminently logical to assume that interregional interaction was intimately involved in the developmental process. Although not a direct causative agent per se, interaction would have facilitated the spread of a shared precursor system in the Middle Formative period. Subsequent interaction between regional iconographic and scribal traditions would have encouraged innovative recontextualizations of common iconic elements as these motifs were deployed in contexts increasingly removed from their ‘original’ conventional frameworks. Over time, interaction between languages, groups, or conventions encouraged human agents to modify the iconographic system to represent more closely the particular linguistic values of a specific language. In the case of the Mayan script, an increased need for specificity in the decipherment of visual messages promoted the scribal innovation that resulted in the locally specific linguistic codification of the shared system of graphic communication.

128

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.1. The development of Mesoamerican writing systems (after Lacadena 2010).

impact of interregional interaction on the innovation of writing systems. Nonetheless, in order to undertake such a comparative assessment, it is first necessary to identify the patterns revealed through iconographic and linguistic evidence. To that end, this investigation departs from several methodological assumptions regarding the identification of interaction and innovation in iconographic and linguistic data.

Methodological Considerations: Tracing the Role of Iconographic Innovation in Script Development Although it is reasonable to assume that interaction played a critical role in the emergence of Mesoamerican writing systems, a considerable difficulty arises when attempting to trace such interaction and its role in script development in concrete terms, or—as in the context of this research—compare patterns of interaction and innovation potentially evident in iconographic or linguistic evidence with similar trends more readily apparent in other datasets. There exists no established method to quantify transformations in iconic elements as they are recontextualized in scripts. Nor is there an accepted technique to objectively determine degrees of distance between recontextualized signs and their ‘original’ iconic referents. Indeed, visual similarity alone is an invalid basis for comparison between scripts or iconographic systems (Alfonso Lacadena, personal communication, 2004). Understanding the relationship between interaction and innovative reformulations of shared motifs is further complicated by the fact that these processes are intertwined in the complimentary historical trajectories of linguistic diversification and script development (Carrasco and Englehardt 2015: 13).

Writing systems, as conservative entities, keep through time certain features related to their own history, including historical episodes of borrowing and/or interaction with other scripts or systems of visual notation (Lacadena 2010). The existence of shared elements and motifs in culturally or geographically distinct scripts of representational systems is therefore most parsimoniously explained by interaction between the groups that employed those systems. Although care must be taken in postulating the nature of the relationship between those groups or systems (i.e., not necessarily ‘genetic;’ see Proskouriakoff 1968, 1971; Quirarte 2007), it is further reasonable to presume that shared elements are a product of the prior dissemination from a common source of the antecedents to the same iconographic and scribal depictions. As discussed in Chapter Two, in Mesoamerica this source would appear to be the Middle Formative symbolic-ceremonial complex (Reilly 1995: 29–30). Middle Formative period ‘Olmec’ (or ‘Olmec– style’) symbols of power and authority—artifactual, iconographic, and linguistic—were widely shared throughout Mesoamerica and subsequently adopted and deployed in differing spatial, temporal, and cultural contexts. In this sense, Mesoamerican scripts and systems

Of course, these difficulties, in part, motivate the primary objective of this investigation. The positive correlation of spatial and temporal patterns of interaction and innovation more evident in alternate datasets with processes of recontextualization involved in script development may offer researchers an alternate methodological route to quantitatively evaluating the

129

Archaeological Paleography of visual notation shared a common iconographic base, in terms of a collective core pictographic and ideographic visual vocabulary. Initially, a common interpretive framework was also shared, since in order to deduce meaning distinct audiences must have been familiar with the underlying narrative conventions. But these narratives changed over time, as they encountered themselves within increasingly diverse spatial and cultural contexts. As the narratives changed, so too did the system(s) of interpretation.

incipient script system—or that are shared between contemporaneous scripts and iconography—could suggest specific functional changes of those elements as ‘ancestral’ icons were excised from pictorial contexts and recontextualized within emergent systems of writing. Such shifts may suggest decreasing interaction, as users of those reformulated signs became increasingly integrated in regionally specific cultural systems or scribal traditions. These transformations may also imply localized innovation, reflected specifically in those variant recontextualized forms of ancestral icons (see, e.g., Englehardt 2005; Lacadena 1995, 2010). Thus, stylistic similarities and differences in visual data suggest trends in the nature, extent, temporality, and directionality of the processes of regional iconographic interaction. Innovation, on the other hand, is detectable through the recontextualization of originally iconic elements—shared widely across Mesoamerica—within new, locally specific, and, in this case, ostensibly Mayan grammatical and linguistic organizational frameworks. From a diachronic perspective, patterns become apparent in the data, which suggest the origins and directions of influences on incipient scribal traditions. These patterns allow for the determination of the degree to which Formative period interregional interaction is associated with the process of recontextualization involved in the development of Maya writing.

In linguistic and iconographic terms, then, interaction is hypothetically traceable through formal features, orthographic conventions, and linguistic aspects shared between Mesoamerica scripts in both synchronic and diachronic contexts. Interaction may thus be reflected in the sharing of visual, stylistic, and syntactical elements between iconographic and script systems. Interaction may be further inferred from the presence of shared general characteristics of the systems, the shared graphic designs of the signs that comprise those systems, ‘intermediate’ visual forms and/or texts that are legible in two distinct languages, ‘frozen’ semantic values that continue to be employed in scripts and/or the presence of fossilized reading values (belonging to the original donor), or identifiable problems of adaptation1 of the donor script (which was created originally to write a different language; Lacadena 2010; Mora–Marín 2001; cf. Justeson 1986, 2012; Justeson et al. 1985). The directionality and temporal contexts of interaction can then be extrapolated from the contextual data of the objects themselves (where available), or through formal analysis of the icons and/or signs in the text.

Because iconic recontextualization may occur over a long span of time—and is itself contextualized within the interrelated processes of linguistic diversification and script development—historical linguistic data and glottochronological reconstructions provide clues regarding the temporal contexts in which pictorial interpretive matrices ceased to function as the sole organizing framework for systems of visual communication. Linguistic data thus have the potential to frame the spatial and temporal contexts of the interaction potentially involved in the innovative recontextualization of iconic elements within the emergent Mayan script. Therefore, I first discuss the linguistic evidence before turning to the iconographic and epigraphic data that more graphically illustrate the patterning of the processes of interaction and innovation at play in the development of the Maya writing system.

In terms of innovation, Justeson et al. (1985) distinguish between shared formal traits that develop from independent invention and those that result from inherited or diffused innovations. This study follows their assumption (see also Justeson 1986; Justeson and Matthews 1990; Mora–Marín 2001) that a greater degree of arbitrariness of a shared feature indicates greater likelihood of common descent from an ancestral iconography or script. Moreover, as Reilly (1991: 151) notes, the identification of elements of an antecedent iconographic system within a later script must be predicated on the testable hypothesis that certain elements of the writing system can be visually identified in iconographic contexts (or some other Mesoamerican or scribal tradition), and that these elements perform similar functions in both the ‘donor’ and ‘recipient’ systems. In those cases, linguistic data (and shifts) latently related to script development are critical to deducing ‘new’ potential semantic or syntactical functions of specific visual elements.

Linguistic Data: Framing Interaction and Innovation Mesoamerican linguistic history is a highly complex and oft–debated topic. Although the linguistic record in Mesoamerica is considerably more complete than the archaeological or iconographic records, debate persists regarding topics such as linguistic affiliations of script traditions and the language represented in the Mayan script (Campbell 1976; Campbell et al. 1986; Hopkins 1997; Houston and Coe 2003; Houston et al. 2000; Houston et al. 2004; Hruby and Child 2004; Josserand 1975; Justeson and Campbell 1997; Justeson and Kaufman 1993, 1997; Justeson

Nonetheless, modifications in the formal or stylistic aspects of the visual elements that comprise an E.g., potential or suggested syntactical or functional values of a particular sign that do not correspond to prior visual readings or the interpretive–organizational frameworks of the ancestral system.

1

130

Comparative Analysis of Iconographic and Linguistic Evidence et al. 1985; Kaufman 1976; Kaufman and Norman 1984; Lacadena 2008; Lacadena and Wichmann 1999; Mora– Marín 2001; Stross 1982; Wichmann 1995; Wichmann et al. 2008). Because the development of Mesoamerican writing systems was a long–term process for which there is a distinct lack of evidence surrounding incipient stages, historical linguistic data and glottochronological reconstructions are often the most reliable sources of contextual data capable of clearly illustrating the spatial and temporal extent, direction, and duration of interaction during the Formative period.2 The linguistic data thus complement and frame the iconographic evidence presented below by suggesting the spatial and temporal contexts in which principles of visual interpretation ceased to function as the sole organizing framework for iconographic systems. Additionally, the syntactical information pertinent to script development and crucial to deducing the function and operation of specific visual elements is often intricately connected to linguistics, and therefore most visible through linguistic data (Hopkins 1997; Justeson 1985; Lacadena 1995, 2010). Finally, following Justeson et al. (1985: 4), in this investigation it is assumed that linguistic interaction is coeval with material cultural and technological exchange.

Linguistic Affiliation of the Ancestral System and the Impact of Mixe–Zoque on GLM Languages Given the developmental dynamics of the Mayan script, in order to appreciate problems of adaptation and the chronology of linguistic interaction, it is first necessary to establish the linguistic affiliation of the ancestral system from which it developed. Alfonso Lacadena (2008, 2010) has recently made compelling arguments regarding the phonological features of the language of the Middle Formative ancestral script of Olmec iconographic system with which developing Maya writing interacted. His conclusions suggest a Mixe–Zoque linguistic affiliation, indicated by several key linguistic aspects: the absence of /b’/ and presence of /m phonemes, the absence of a /ch/ phoneme, the absence of an /l/ phoneme, the absence of an /x/ phoneme, the absence of glottal consonants (C’), and the presence of one back spirant phoneme (Figure 6.2). These findings fit well with traditional considerations of Olmec linguistics, which have consistently suggested a Mixe–Zoque linguistic identity for Middle Formative period Olmec groups (Campbell 1976; Josserand 1975; Justeson et al. 1985; Kaufman 1976; Mora–Marín 2001: 45; Wichmann et al. 2008). The case for an Olmec Mixe–Zoque linguistic affiliation is based primarily on the fact that the geography and chronology of proto– Mixe–Zoque (pMZ) and the Olmec culture correspond closely. Such archaeological–linguistic correspondences suggest Mixe–Zoque languages as the most probable candidate for the linguistic identification of the Olmec (Campbell1976; Justeson 1985; Justeson and Campbell 1997; Kaufman 1976; Wichmann 1995; Wichmann et al. 2008).

This investigation considers a limited selection of evidence and cogent examples that highlight processes of linguistic interaction and the temporal contexts of transformations that frame the development of Maya writing.3 These include the chronology of linguistic divergence in the Greater Lowland Mayan (GLM) language family, the linguistic affiliation of the ancestral Middle Formative iconographic system or scribal tradition, the linguistic affiliation of the early Mayan script, general linguistic diffusion into the GLM language family, the influence of Mixe–Zoque on the GLM language family, and the influence of Cholan–Tzotzilan4 and lexical diffusion within the GLM language family. A selection of morphological shifts is also considered, such as the * k’ > * č shift in the Cholan–Tzotzilan language group and the * m > * b’ shift in the GLM language family (Hruby and Child 2004; Justeson 1985; Justeson and Campbell 1997; Justeson et al. 1985; Kaufman 1976; Kaufman and Norman 1984; Mora–Marín 2001; Wichmann 1995). The spatial and temporal contexts of linguistic interaction and divergence reflected in these lines of evidence illustrate how general patterns of interaction varied over time and through space. Additionally, the examples chosen are particularly well– represented, documented, and relatively easy to trace in the emerging Mayan script (e.g., Justeson 1986; Justeson et al. 1985; Lacadena 2010; Macri 1991; Mora–Marín 2001; Stross 1982, 1990).

Evidence for Mixe–Zoque linguistic influence on the GLM language area is abundant. The GLM language area is defined by linguistic features shared among GLM languages, including common morphological patterns, the presence of ‘verbal nouns’ (words that can be inflected either as nouns or as intransitive verbs, without derivational modification), the sharing of a large number of suffixes and grammatical particles that have no apparent correspondences in other Mayan subgroups,5 the phonemic contrast of b’ and p’, and the exclusive sharing of a large body of vocabulary across GLM languages, particularly Cholan–Tzotzilan and Yucatecan (Justeson et al. 1985: 9–12, tables 3 and 4). These traits define an extensive area, and likely reflect an early era of interaction within the GLM language area. Recognizing a GLM language area centered on Cholan and Yucatecan implies that these languages share phonological, grammatical, and lexical features that distinguish them from other Mayan languages and that are not the direct legacy of common ancestry (Justeson et al. 1985: 9; Mora–Marín 2001: 369–71,

In spite of valid criticism of the method of glottochronological reconstruction (see, e.g., Dixon 1994, 1997). 3 Although the linguistic data considered here is, an exhaustive presentation and discussion of all available evidence is unwieldy and beyond the scope of this investigation. 4 This language group is sometimes referred to as Greater Tzeltalan (e.g., Justeson et al. 1985; Kaufman 1976). 2

E.g., * –tal suffixed to positionals; * –na to affect verbs; * tuhl as a numeral classifier for people and animals; * iwa: l as a particle marking progressive aspect.

5

131

Archaeological Paleography

NO COINCIDENCE

COINCIDENCE ANCESTRAL SCRIPT

MAYAN

OTO–MANGUE

NAHUA

MIXE–ZOQUE

/m/, no /b’/

/b’/ , /m/

/b’/ , /m/

/m/, no /b’/

/m/, no /b’/

no /ch/

/ch/

/ch/

/ch/

no /ch/

no /l/

/l/

/l/

no /#l_/

no /l/

no /x/

/x/

/x/

/x/

no /x/

no glottal consonants

C’

no C’

no C’

no C’

one back spirant

two back spirants

one back spirant

no back spirants?

one back spirant

Figure 6.2. Phonological aspects of early writing or ancestral script adopted by the Maya as compared with four Mesoamerican language families (after Lacadena 2010:36, table 3).

widely throughout Mesoamerica appear to have roots in Mixe–Zoque languages.6 The presence of Mixe–Zoque loan words in a variety of Mesoamerican languages across spatial–temporal contexts is significant, in that such presence suggests a very broad extent of Mixe–Zoque linguistic influence. Reconstructed pMZ vocabulary items of cultural content imply a sophisticated cultural complex for speakers of pMZ around 1500 BC. Archaeological data indicate that the Olmec were the only Mesoamerican culture to have developed such complexity by the Early Formative period. These facts support the Mixe–Zoque–Olmec affiliation, since one would expect other groups to borrow cultigens and a ritual vocabulary from the Olmec—the first highly complex Mesoamerican agriculturalists. That so many Mixe–Zoque loanwords were diffused into such a variety of Mesoamerican languages at an early date attests to the prestigious and powerful position the original speakers of Mixe–Zoque—likely the Olmec—must have had (Flannery 1968a). The evidence thus indicates that Mixe–Zoque languages exercised significant influence on the Greater Lowland Mayan language area, and throughout lowland Mesoamerica, during the Formative period (Figure 6.4; cf. Campbell et al. 1986; Justeson et al. 1985; Lacadena 2010; Wichmann 1995).

373, tables 2.4–2.6, 2.8). In the absence of a close genetic relationship between Cholan and Yucatecan (i.e., no common ancestor since proto–Mayan), these shared features can only have developed through direct contact and linguistic interaction of great intensity and duration after the diversification of GLM languages into Western and Eastern Mayan groups during the Early Formative period (Justeson et al. 1988; Campbell 1984: 11; Justeson et al. 1985: 9; 13–14). The importance of grouping these languages together lies in the fact that morphological and lexical diffusion into this areal group can be shown to have specific and demonstrable subsequent consequences in the linguistic descendents of these languages (Campbell 1976; Campbell et al. 1985; Justeson 1985; Justeson et al. 1985; Kaufman 1976; Kaufman and Norman 1984; Lacadena and Wichmann 1999; Mora–Marín 2001). Prior investigations have established the presence of a strong linguistic influence of Mixe–Zoque on many Mesoamerican entities, including GLM languages and Maya cultural groups, especially in terms of diffused agricultural and ritual vocabulary (Figure 6.3; cf. Justeson et al. 1985: 22–3, tables 9–11). Many detectable Mixe–Zoque loans into Mayan involve calendric and ritual concepts; for example pMZ may is a word that has specific calendric associations in most Maya languages that adopted it (Justeson et al. 1985: 23). Additionally, many day names in the 260–day ritual calendar shared

E.g., #ok tenth day name (see Table 6.2); *(h)iʔ(i)š fourteenth day name ‘jaguar’ < Oaxaca Mixe ʔi:š ‘cacomixtle’ (a wild feline); #čowen/ čuwen eleventh day name ‘monkey’ < pMZ *¢awiʔ ‘monkey.’

6

132

Comparative Analysis of Iconographic and Linguistic Evidence Mixe–Zoquean Source Borrowed Cultigens

* kakawa  ‘cacao’ [pMZ]

kakaw   (pan–Mayan);    kƏ kƏ w (Chol); kokow  (Tzotzil)

* ¢ ima  ‘gourd’  [pMZ] (* ¢ iʔ wa  ‘gourd’  [pMi])

koyaʔ  (Chol); ʔ iš–ko:yaʔ (Mam, Aguacateca); (š)–ko:ya: ʔ (Cakchiquel, Tzutujil)

pátaŋ  ‘guava’  [Sierra Popoluca]

páta(h)  (Tzeltal); póto  (Tzotzil)

Maize Preparation Complex

* way  ‘to grind corn’  [pMZ]

* waye  ‘pozole’ (proto–Mayan)

* pi¢ i  ‘nixtamal’  [pMZ]

po¢ a  ‘tamal’  (Tzeltal); pa¢  ‘tamal’  (Tzotzil)

* hƏ ¢  ‘to grind’  [pMZ]

* xuč’  (proto–Mayan)

Ritual and Calendric Terms

* po: mV  ‘incense’ (copal)  [pMZ]

po: m  (pan–Mayan)

may  ‘to count, to divine’  [pMZ]

may  ‘20, 20 years’  (Kekchi, Pokom); may  ‘20 years of 400 days each’  (Quiché, Cakchiquel); mayex (?) ‘offering’ (Kekchi)

(*ʔ ok  ‘dog’  [pMi])

* pus  ‘axe,’ ‘to cut with knife/axe’    (* pusan  ‘metal’   [pMZ])

¢ ima   (pan–Mayan) ¢ iw   (Huastec), ¢’ iwan  (Lenca); č'iwan  ‘chayote’  (Chortí)

koya  ‘tomato’  [pMZ]

* ʔ ukA ‘dog’ [pMZ]

Linguistic Derivatives

ok ‘dog’  (Yucatec); ʔ oʔq, ʔ oq   ‘coyote’  (Kanjobalan) pos  ‘stone war axe,’ pus ‘witch,’ ax pos  ‘wonder worker’ (Pokom); pos  ‘polished stone’ (Cakchiquel); pos, pus  ‘to sacrifice men by removing their hearts,’ ‘to cut,’ ‘polished stone,’ ‘magic power’   (Cakchiquel, K’iché)

Figure 6.3. Mixe–Zoque loans into Greater Lowland Mayan languages. These loans are widespread in Mayan and other Mesoamerican languages, and probably reflect contact with the Olmec (Justeson et al. 1985: 23). This table provides only those loans into Lowland Mayan languages, though these pMZ words were adopted into a variety of other, distinct Mesoamerican language groups, families, and specific languages (Campbell et al. 1986). (1)

grounds, this is unsurprising. There is little material evidence to suggest that the Maya lowlands, and thus the GLM language area, had begun to coalesce culturally or integrate in such early temporal contexts—an observation complemented by the conclusions drawn from the ceramic data in the previous chapter. Thus, in the Middle Formative period, no special Mixe–Zoque influence on GLM languages is detected specifically (Justeson et al. 1985: 23). Nevertheless, by examining the temporal contexts of the diversification of GLM languages, as well as morphological transparency, reconstructibility, and phonological and grammatical

Temporal Contexts of Interaction: GLM Linguistic Diversification and the Influence of Cholan–Tzotzilan The presence of such a large quantity and variety of Mixe–Zoque loans within the Greater Lowland Mayan language area suggests that linguistic influence was directed eastward from the Mixe–Zoque area straddling the Isthmus of Tehuantepec, although this does not imply a specifically ‘Olmec’ influence. The timing of such linguistic interaction and diffusion from Mixe–Zoque into GLM languages appears to postdate the Middle Formative apogee of Olmec culture. On archaeological 133

Archaeological Paleography Mixe–Zoquean Source

Linguistic Derivatives

Other MZ Loans

* hΛyΛ  ‘man’s brother–in–law’ [pMi]

* xaʔ an  ‘man’s brother–in–law’ (pan–Lowland Mayan)

* ¢ iku  ‘coatimundi’ [pMZ]

* čiʔ k  ‘coatimundi’  (Yucatec, Cholti)



* sah  ‘wing’ [pZo];     šáhpak ‘fathom’ [Sayula Popoluca]

* sahp  ‘fathom’ (Yucatec, Eastern Cholan)

* ma¢  ‘to grasp’ [pMi]

* mač  ‘to take, grasp’ (Yucatec)

ša: k  ‘to guard’  [Totontepec Mixe]

ša: k–t  ‘to examine’  (Mopan)

* soʔ k  ‘grass’ [pZo]

* ¢ um  ‘to tie’ [pMi]

suʔ uk  ‘grass’ (Mopan)

sum  ‘rope’  (Mopan)

tu: h¢  ‘palm’ [Sayula Popoluca]

tu¢  ‘corozo’  (Mopan)

mageʔ č  ‘cockroach’  [Sayula Popoluca]

makeʔ č  ‘escarabajo camaleón’ (Yucatec)

múguy  ‘topote’ [Sayula Popoluca] * pači  ‘lizard’ [pMZ]; * pač ‘lizard’ [pMi]

sok  ‘snail’  (Yucatec)

* ¢ uk  ‘mouse’ [pMZ] * ʔ uma  ‘dumb’ [pMZ]

mok  ‘gata del mar’  (Yucatec) – pač  ‘male iguana’  (Yucatec)

* soki  ‘snail’  [pZo]

¢ uk  ‘mouse’  (Chol, Chontal) * umaʔ  ‘dumb, stutterer’ (Cholan– Tzotzilan)

* kono  ‘short’ [pMZ]

* ko(h)m  ‘short’  (pan–Lowland Mayan)

* ʔ une  ‘child’ [pZo]

* une, * unin  ‘infant, tender’ (Cholan–Tzotzilan)

* wetu  ‘fox’ [pZo]; * wa: š  ‘fox’ [pMi]

* we: t, * waʔ š  ‘fox’ (pan–Lowland Mayan)

Figure 6.3. Mixe–Zoque loans into Greater Lowland Mayan languages. These loans are widespread in Mayan and other Mesoamerican languages, and probably reflect contact with the Olmec (Justeson et al. 1985: 23). This table provides only those loans into Lowland Mayan languages, though these pMZ words were adopted into a variety of other, distinct Mesoamerican language groups, families, and specific languages (Campbell et al. 1986). (2)

Figure 6.4. Extent of Mixe–Zoque language area (L) and probable movement of Olmec ethnic groups/Mixe–Zoque language groups and Olmec artistic styles and ceramic technologies during the Middle and Late Formative periods (R); based on linguistic data and artifact distribution (after data in Campbell 1976, 1988; Kaufman 1976; Justeson et al. 1985; Wichmann 1995).

134

Comparative Analysis of Iconographic and Linguistic Evidence

KEY: CHJ=Chuj; CHL=Chol; HUA=Huastecan; IXL=Ixil; KAN=Kanjobal; KEK=Kekchi; MAM=Mamean; MOCH=Mocho; POQ=Poqomam; QUI=Quichean; TZO=Tzotzil; USP=Uspantec; YUK=Yucatec Figure 6.5. Phylogenetic grouping of Mayan languages detailing glottochronological estimates for divergence (cf. Campbell 1984: 2–3, figs. 1–2; Justeson et al. 1985: 3, fig. 1).

immediately followed by speakers of Cholan–Tzotzilan (CT) between 1000–600 BC (Justeson et al. 1985; Kaufman 1976; Mora–Marín 2001: 43). During the Middle Formative period, when CT speakers began to migrate into the Maya lowlands, they came into contact with speakers of Yucatecan who were already there. Close linguistic interaction between these two languages followed, and is reflected in the temporal contexts of the incipient diversification of Western Mayan beginning c. 1000 BC (Campbell 1976, 1984; Justeson et al. 1985: 57–8; Kaufman 1976: 110–12; Kaufman and Norman 1984; Mora–Marín 2001: 43). At this point, phonological, lexical, and grammatical innovations were diffused from CT into Yucatecan, evident through loan words and subsequently reflected in glyphic representations of month names, numeric head variants, and other ritual terms (Englehardt 2005: 435–7, figs. A2.1–A2.3; Mora–Marín 2001: 44, 372–3, tables 2.7 and 2.8).7 At the same time, CT borrowed verbal

anomalies among GLM languages, it is possible to establish a temporal framework for interaction. Such episodes of linguistic modification and transformation are doubly important insofar as they speak to issues of innovation and the integration of the Maya lowlands within the GLM language area, as well as the role that local factors played in the historical development of individual languages within the Maya lowlands. At the same time, these linguistic data, particularly the evidence regarding the influence that the Cholan–Tzotzilan language family exercised within the GLM language area, further elucidate processes of interaction and clarify the temporal contexts in which such interaction occurred. Figure 6.5 illustrates the phylogenetic grouping of GLM languages, detailing glottochronological estimates for divergence. Although Yucatecan and Eastern and Western Mayan languages diverged from proto–Mayan at an exceptionally early date—at or about the Archaic– Early Formative period transition—archaeological and linguistic correlations suggest an initial settlement of the Maya lowlands c. 1000 BC by speakers of Yucatecan,

CT exerted vastly more lexical and phonetic influence on the Lowland Maya language group. The preponderance of GLM vocabulary was diffused from CT (estimated at nearly ninety percent of loan words and descendents reconstructible back to proto–Maya; Justeson, et al. 1985:

7

135

Archaeological Paleography * m > b’ shift, b’en does not exhibit the * e > i shift, and manik does not exhibit the * k > č shift indicative of CT diversification (Justeson et al. 1985: 21–2, 57). These words were not original to proto–Zapotecan. Given the morphological forms the words take in GLM languages, they must have entered GLM vocabulary after the diversification of proto–Zapotecan. Linguistic reconstructions place this diversification event between 500–400 BC. In addition, on culture–historical grounds, neither Zapotec nor any other Oto–Mangue language was likely to have been in the position to make substantial or significant contributions to GLM vocabulary until at least the Middle Formative Monte Albán I phase, c. 500 BC (Justeson et al. 1985: 57). Given these temporal bookends, and allowing for a time lag between the initial phases of interaction, the adoption of the loan, and specific sounds changes and morphological shifts in CT,9 it appears that the most intense episodes of interregional linguistic interaction in southeastern Mesoamerica occurred between 500 BC–AD 100, in the late Middle to Late Formative periods.

morphology and split ergativity from Yucatecan (Mora– Marín 2001: 44, 378, table 2.12), suggesting an intense episode of sociolinguistic interaction. Concurrent with this extensive interaction within the GLM language area, a great deal of lexical diffusion from Mixe–Zoque into Cholan–Tzotzilan—and subsequently into other Mayan and non–Mayan languages—was taking place. Several critical pieces of linguistic data suggest that the timing of this inter– and intra–regional interaction can be placed squarely in the Late Formative period. The primary evidence that indicates a Late Formative period temporal context for extensive and intense linguistic interaction involves morphological shifts within the CT language family (e.g., (p)CT * k > kʸ > č; * e: > i; and * o: > u; see Campbell 1984: 11; Justeson et al. 1985: 9, 13–4, 18–9, tables 5–7, 58, table 17; Justeson et al. 1988). This shift accompanied the diversification of the Cholan and Tzotzilan branches of CT c. 300 BC–AD 100 (Justeson et al. 1985: 21–4, 57; Kaufman 1976; Mora– Marín 2001: 44). Significantly, no Mixe–Zoque words containing k and loaned into GLM languages underwent the * k > č shift noted in the Cholan–Tzotzilan family (e.g., *kakaw). Mixe–Zoque loan words that appear in Yucatecan do not exhibit this morphological shift, suggesting that they were incorporated into the Yucatec language at roughly the same time as they were diffused into CT.8 This fact implies that those Mixe–Zoque words were borrowed prior to the diversification of Cholan and Tzotzilan in the Late Formative period (Justeson et al. 1985: 23). Additionally, such contemporaneous borrowing into multiple GLM languages suggests that linguistic interaction was widespread and relatively uniform throughout the Maya lowlands.

Archaeological–Linguistic Correlations The suggested pattern of linguistic interaction complements available material evidence. Archaeological data indicate the movement of groups of people into the southern Usumacinta and central Petén lowlands in the Middle Formative period (c. 1000–500 BC), associated with the Xe ceramic complex (Lowe 1977, 1989; Mora– Marin 2001: 46; Ochoa 1983). The Xe complex displays affiliations with ceramic assemblages in highland Chiapas, potentially created by Mixe–Zoque speakers, as well as highland Guatemala, possibly Greater K’iche’an speakers. These affiliations, as well as the fact that the Xe complex was distributed across an area later dominated by Cholan–Tzotzilan languages, it is probable that the Xe complex was developed by CT speakers (Justeson et al. 1985; Kaufman 1976: 112; Mora–Marín 2001: 47). At roughly the same time, the Swasey ceramic complex developed in northern Belize. Based on cultural continuities between the developers of Swasey ceramics with the preceramic inhabitants of the region and the later occupation of the area by speakers of Yucatecan, it is likely that the developers of the Swasey complex were in fact Yucatec speakers (Kaufman 1976; Mora– Marín 2001: 46–7). These archaeological linguistic correlations are supported by glottochronological estimates that suggest a slightly earlier settlement of the northern and eastern lowlands by Yucatec speakers c. 1000 BC (after the split of Yucatecan from late proto– Mayan c. 1800–1400 BC), followed by the arrival of CT speakers in the southern and central lowlands c. 1000– 600 BC, after the diversification of Western Mayan into

Framing the other extreme of the temporal context of extensive interregional linguistic interaction is the adoption of several Oto–Mangue words into GLM languages, specifically lexical diffusion from Zapotecan. For example, three day names of the 260–day ritual calendar, lamat, b’en, and manik, appear to have entered GLM languages from Zapotec. Like Mixe–Zoque loans, these words must have been borrowed into Mayan prior to the diversification of CT. Lamat does not exhibit the 20). Even highland and non–Mayan languages (e.g., Lencan, Xincan) exhibit lexical borrowings from CT, and display sound changes of a simple merger type with CT, presumably by imitating the phonetic characteristics of Cholan languages (Kaufman 1976: 112; Mora–Marín 2001: 375–77, tables 2.9–2.11). 8 Loans that can be traced to an original Mixe–Zoque source in other, later developing Mayan languages (e.g., Kanjobalan, c. AD 150), previously diversified Eastern Mayan languages (e.g., K’iche’an, Mamean, c. 650 BC), and non–Mayan languages of southeastern Mesoamerica (e.g., Lencan, Xincan) likewise do not exhibit evidence of a * k > č shift, implying that the episode of borrowing occurred before the Late Formative period breakup of CT reflected by this morphological shift. In the absence of a close genetic relationship between these languages (i.e., no common ancestor since proto– Mayan), such shared features can only have developed through direct contact and linguistic interaction of great intensity and duration.

Such a time lag may be attributed to the fundamentally conservative nature of linguistic systems; loan words and morphological shifts take time to be fully assimilated into and appear clearly in the linguistic system that is borrowing them.

9

136

Comparative Analysis of Iconographic and Linguistic Evidence Greater Cholan–Tzotzilan and Greater Kanjobalan (see Figure 6.5; Justeson et al. 1985; Kaufman 1976, 1984: 3, fig.2; Mora–Marín 2001: 47). Again allowing for a time lag between the initial episodes of interaction and subsequent linguistic changes, these affiliations suggest a period of intense linguistic interaction and possibly significant CT–Yucatecan bilingualism in the GLM language area beginning in the Middle Formative period (Mora–Marín 2001: 44). Significantly, the Xe and Swasey ceramic complexes were eventually assimilated into the Mamom ceramic sphere c. 600–500 BC, at which point the central lowlands appears to have moved toward increasingly homologous ceramic styles, as suggested in the preceding chapter. This cultural assimilation may reflect the intense linguistic interaction and diffusion between CT and Yucatecan speakers that paralleled the development of the GLM language area in the Middle Formative period (Mora–Marín 2001: 47).

exerted on the GLM language area strongly suggests a great degree of contact, mixture, exchange, and diffusion, both linguistically and otherwise. Such contact may be reflected archaeologically in the widespread similarity in ceramic attributes and distributional uniformity between the Maya lowlands and the Middle Grijalva region through the first half of the Late Formative period (Lee 1972; cf. Josserand 1975; Kaufman 1976; Wichmann et al. 2008). The emerging influence of Cholan–Tzotzilan during the Late Formative period following an extended period of CT–Yucatecan bilingualism within the GLM language area (c. 600 BC–AD 1) and contemporary with an influx of Mixe–Zoque linguistic elements is doubly significant when considered in the context of the development of the Mayan script. Writing develops when a system of iconography is shared between multiple groups—in this case, speakers of Cholan– Tzotzilan, Mixe–Zoque, and Yucatecan. Over time, purely visual interpretive frameworks break down, and locally specific linguistic values are assigned to shared icons in order to more precisely determine the relationships between signs, and to clarify semantic meanings. I suggest that CT linguistic principles were juxtaposed onto the shared iconographic system immediately following the intense, extensive linguistic interaction between and within the GLM language area and Mixe–Zoque for at least the 500 years prior to AD 1. The earliest evidence for Maya writing occurs precisely in the Late Formative period, in areas where sustained linguistic interaction10 was most intense and CT languages were likely exercising the most influence, such as the traditional ‘homeland’ of CT in the southern central lowlands (e.g., the San Bartolo murals, El Mirador Stela 2) or the marginal border areas at the limits of the Maya lowlands (e.g., Kaminaljuyú Stela 10, El Polol Altar 1, Takalik Abaj Stela 5). It is significant that the Mayan script may be read and understood in Yucatec, but that it most likely represents specifically Cholan linguistic values (Campbell 1984: 9–12; Houston et al. 2000; Justeson 1985; Justeson et al. 1985; Justeson and Campbell 1997).11 The time range for the emergence

Following this first episode of extensive linguistic interaction within the lowlands during the Middle Formative period, interaction continued across the lowlands in the Late Formative period in the form of the post–Olmec Mixe–Zoque lexical diffusion into GLM languages (see Figure 6.3). Also at this time, loanword evidence suggests much borrowing of ritual, cultigens, and commerce technology from CT into other GLM languages—almost 90% of GLM vocabulary can be traced to a CT source (Campbell 1976; Josserand and Hopkins 1999; Justeson et al. 1985: 20; Kaufman and Norman 1984; Mora–Marín 2001: 45). This development suggests that Mixe–Zoque loans were adopted first into CT during the Late Formative period prior to CT diversification and subsequently diffused into other contemporaneous and descendent languages within the GLM language group (Justeson et al. 1985). Two significant conclusions may be inferred from the emerging primacy of CT within the GLM area during the Late Formative period. First, it suggests that localized influences were becoming increasingly important in the linguistic development of the GLM area. These localizing trends may be reflected in the Late Formative period diversification of Greater Cholan–Tzotzilan into Cholan, Tzotzilan, Chujean, and Kanjobalan (see Figure 6.5), as well as the morphological shifts within the CT family associated with the processes of linguistic divergence. Second, it suggests that linguistic interaction was most intense in the marginal border regions of the lowlands and the GLM language area (Justeson et al. 1985: 73, note 15). The historical distribution of CT languages extended to the limits of the Maya lowlands in central Chiapas to the west and the Soconusco to the southwest, areas occupied by speakers of Mixe–Zoque (Campbell 1988; Campbell et al. 1986: 538–39; Campbell and Kaufman 1976). The geographic proximity of the Mixe–Zoque and CT language groups, which effectively abut each other directly to the east of the Isthmus of Tehuantepéc, coupled with the substantial lexical and morphological influence that Mixe–Zoque

Either between CT and Yucatecan or CT and Mixe–Zoque. E.g., T301 b’e / b’i, derived from pCh * b’ih ‘road’ and illustrating the pCT * e: > i shift (cf. pM * b’e: h); T528.560v ku–tz’u, kutz’ kú: ¢: kutz ‘turkey’ from pYu ku¢ derived from pTz * ko¢ and pCT * ko: ¢, both of which also demonstrate the results of the Cholan * o: to u shift; T25/T203 ka, a syllabic sign representing fish/fish fins (cf. pM * kar ‘fish’). The last example is especially telling. Fish fins as ka (vs. cha, or ‘kay’ vs. ‘chay’) indicate clearly that the development of the Mayan script had occurred prior to the Cholan * k to č shift, and therefore prior to the diversification of the CT family in the early Late Formative period (c. 300 BC–AD 100). See Campbell (1984: 12–3) for other examples. Campbell (1984: 12) argues that the crucial stage in the development of phoneticism lies in the development of phonetic complements (determinatives). These complements derive from logograms which have the form (C)VC with a ‘weak’ final consonant (i.e., h, w, y, l, r, ʔ ; nasals). Monosyllabic logograms of this formulation come to be used as phonetic determinatives to distinguish between different values of a logogram. For example, consider T614:514v.59 OTOT otot ‘house,’ carrying the T59 phonetic complement ti– from

10 11

137

Archaeological Paleography of the Mayan script is coeval with the breakup and diversification of CT c. 300 BC–AD 100, which suggests that locally specific linguistic factors were becoming increasingly important and that the borders of the lowland Maya world, at least in linguistic terms, were becoming increasingly more defined. In other words, linguistic interaction on the interregional scale was diminishing toward the end of the Late Formative period.

this time, since its first examples appear early in the Late Formative period c. 250 BC. Given the time lag that must have existed between episodes of interaction and their subsequent reflection in linguistic and scribal traditions, it is likely that processes of linguistic and scribal experimentation had begun well before the material innovation exhibited around the Formative– Classic period transition. These processes of incipient divergence would have been coeval with the influx of Mixe–Zoque linguistic influence beginning in the late Middle Formative period c. 500–400 BC. It is significant that the phylogenetic and diffusional ties of the Mayan script are closer to the Greater Izapan rather than the West Isthmian area of Mixe–Zoquean speech (Justeson et al. 1985: 41, table 16, 57, table 2.2). This fact suggests that linguistic interaction and subsequent experimentation in the context of script development were most intense along the southern and western frontiers of the Maya lowlands, an area in which numerous early examples of Maya writing are evident (Fahsen Ortega 1999; Justeson 2012; Saturno et al. 2006).

These Late Formative period linguistic and scribal developments correlate well with archaeological data. For example, there is evidence for population movement of CT speakers from the Petén lowlands to the central Chiapas highlands coinciding with the 300 BC–AD 100 diversification of CT (Justeson et al. 1985). This movement is postulated to have caused the collapse of traditional Zoque capitals along the Grijalva River (e.g., Chiapa de Corzo, La Libertad). These events, in turn, may subsequently have been reflected in the introduction of new ceramic technologies and the re– orientation of the ceramic tradition of the Middle and Upper Grijalva region, marked by the introduction of Yahama Rough and Yatsipo Sandy wares (Lee 1972; Mora–Marín 2001: 48). In addition, it is at this time that experimentation in ceramic production techniques becomes increasingly evident in lowland assemblages. The drastic divergence in lowland ceramics noted in the Early Classic period may be reflected in the numerous episodes of diversification across GLM languages beginning roughly at or immediately following the Formative–Classic period transition (see Figure 6.5). Finally, the somewhat more inward–focused interaction among GLM languages in the Late Formative period may parallel the slightly diminished spatial extent of the Chicanel ceramic sphere at this time (c. 150 BC–AD 300). Thus, just as the Mamom complex could correspond to an extended period of interregional linguistic interaction, the Chicanel sphere potentially reflects a period in which locally specific factors, in terms of ceramic production and linguistics, came to the fore, and interaction was greatest within the Maya lowlands.

To recapitulate the linguistic evidence presented to this point, extensive interregional linguistic interaction is evident throughout southeastern Mesoamerica and between Cholan–Tzotzilan and Yucatecan within the Greater Lowland Mayan language area beginning in the late Middle Formative period, contemporary with the Mamom ceramic sphere. Interaction intensified greatly in the earlier part of the Late Formative period with an influx of Mixe–Zoque influences and the emerging primacy of Cholan–Tzotzilan among GLM languages. This end of this period of intense interaction is marked by the diversification of CT, dated between 100 BC–AD 100 on the basis of various morphological shifts within CT and the breakup of CT into Cholan and Tzotzilan branches. The diversification of CT and the end of extensive linguistic interaction appears to have occurred slightly earlier than the experimentation in ceramic technologies noted in late Chicanel sphere assemblages that accompanied the Formative–Classic period transition.

It appears that the diversification of CT had already begun prior to the ceramic experimentation noted at the end of the Late Formative period. Morphological changes in CT languages are evident prior to AD 1. In any case, the Mayan script had certainly emerged by

Textual Evidence Indicative of Linguistic Diffusion With these general patterns in mind, it is possible to appreciate evidence that speaks to specifically interscribal interaction and exemplifies potential problems of adaptation in the emerging Mayan script. The proposed decipherment of Epi–Olmec script (Justeson and Kaufman 1993, 1997) and evidence for script diffusion between Epi–Olmec and Maya scribes (Mora–Marín 2001; Stross 1990) suggest a great deal of both interlingual and interscribal interaction during the Late Formative period. If, as Lacadena (2010) suggests, the ancestral script represents Mixe–Zoque, and interaction between Epi–Olmec and early Maya scribes was extensive at this time, then there should be

Ch * tah, which, by virtue of its weak final consonant (–h) can be used as phonetic determinative ‘ta,’ specifying that the house logogram ends in –t, and therefore has the Cholan value * otot, and not, for example, Yuk otoch. Mayan phonetic and syllabic signs with CV values, when their sources are clear, can be seen to have developed from this origin. The form of the specific individual glyph indicates that its iconic origin is that of the object named in its gloss, since the appropriate phonetic value is available only in Cholan as a result of specific sound changes (Campbell 1984: 12). The importance of the T59 locative and its antiquity within the script tradition makes a strong argument for CT influences on the developing Mayan script. T671, like T59, presents a rebus that works only for CT— only in this language can the linguistic forms render the subsequently developed phonetic values.

138

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.6. Sign reformulation to reflect the m to b’ linguistic shift and problems of adaptation. The glyphs are identical except for the two circles with infixed cross– hatched patterns (T741) affixed to the protuberance in T19 and T21 (Englehardt 2005: 470, figs. A4.24 and A4.25; Lacadena 2010).

Figure 6.7. MS130, T528, and T548.

identifiable problems of adaptation in one or the other script, as distinct linguistic values were juxtaposed onto a shared tradition of graphic representations. This turns out to be the case. Figure 6.6 illustrates one cogent example of how the graphic representation of the /m/ phoneme that exists in both Mixe–Zoque and Mayan languages had to be reformulated to reflect the /b’/ phoneme, which is absent in Mixe–Zoque but present in Mayan. In early examples of the Mayan script, such as on the Hauberg Stela, /m/ is present where /b’/ should appear. This fact suggests that interethnic and interscribal interaction between Mayan and Mixe–Zoque groups continued after the diversification of CT, given that the Hauberg Stela dates to AD 197.12 This is unsurprising; one would not expect interaction to cease full stop following CT diversification, but to continue, albeit to a relatively lesser degree as local factors come to the fore. In any case, only until well into the Early Classic period (c. AD 400) do we find evidence that Mayan scribes had developed a specific glyphic representation for the /b’/

phoneme, as on Tikal Stela 31—and even in that case /m/ continues to substitute for /b’/ in some contexts (see Figure 6.6). Justeson and Kaufman (1993: 1709–10) briefly outline a complex historical relationship between the Epi– Olmec and Mayan scripts (cf. Houston and Coe 2003; Lacadena 2010; see also Figure 6.2). The final line of linguistic evidence presented here, which leads into the following discussion of iconographic interaction and script development, concerns the linguistic values of certain signs shared between the Epi–Olmec and Mayan scripts. Justeson and Kaufman (1993, 1997) identify three kinds of shared signs or sign compounds in the two scripts. The first are shared logographic signs with Zoquean values in Epi–Olmec writing and Mayan values in the Mayan script. Examples of these are calendrical signs and sign compounds, such as Long Count notations or signs representing day names in the 260–day ritual calendar (Englehardt 2005: 460–63, figs. A4.1–A4.10; Justeson et al. 1985; Méluzin 1992, 1995; Mora–Marín 2001: 257). Second, there are also shared syllabic signs derived acrophonically from Zoquean lexemes in Epi– Olmec and Mayan lexemes in Mayan. Examples of these might be Epi–Olmec MS130 tza and Mayan T528 TUN, both of which depict a shiny stone (Figure 6.7;

12 As Mora–Marín (2001: 252–53) notes, some of this interaction likely predated the Late Formative period, given the antiquity of the Epi–Olmec script. Unfortunately, in most cases it is impossible to distinguish between cases of diffusion of signs among separate scripts and cases of inheritance of signs among artistic traditions that later developed distinct writing systems.

139

Archaeological Paleography Englehardt 2005: 308–20, 465–66, figs. A4.14–A4.17; Lacadena 1995, 2010). The former was based on proto– Mixe–Zoquean *tza:ʔ ‘stone’ or proto–Zoquean *tzaʔ, while the latter was based on proto–Mayan *tooŋ ‘stone’ or Cholan *tuun ‘stone; year.’13 This example warrants further examination, as it offers insights into the temporal contexts of both Epi–Olmec–Mayan scribal interaction and the diversification of Cholan–Tzotzilan.

sign in calendrical contexts within the Epi–Olmec script implies that even more time must have passed between the completion of the Cholan sound change that allowed for the homophonous relationship in Cholan between ‘stone, year’ and ‘slit drum.’ Additionally, this fact necessitates a greater temporal lag between the morphological shift and the invention by Cholan scribes of the T548 sign and its reading as HAAB’ ‘year,’ as well as for the borrowing of that sign by Epi–Olmec scribes (Justeson et al. 1985: 42; Mora–Marín 2001: 254). Consequently, 100 BC is confirmed as a conservative estimate for the completion of the *o > *u shift, and therefore, for the completion of the breakup of Cholan–Tzotzilan into separate Cholan and Tzotzilan branches.14 This example thus suggests an earlier time frame for CT diversification. At the same time, the evidence also points to an earlier temporal context for linguistic and interscribal interaction between speakers of Mixe–Zoque and CT, as well as the existence of distinct scripts by at least the early Late Formative period.

T528 is homophonic with T548, both of which carry the phonetic value tuun, meaning ‘stone.’ T548 is also polyvalent, meaning that it carries two distinct linguistic values: the aforementioned tuun ‘stone,’ as well as haab’ ( *u shift marking the diversification of Cholan– Tzotzilan had already begun to take place as early as c. 100 BC. The earliest use of T548 with the probable value HAAB’ ‘year, anniversary’ occurs on Takalik Abaj Stela 2. This monument displays a damaged and incomplete Long Count date but can nonetheless be dated in absolute time between 236–19 BC. As Mora–Marín (2001: 253) concludes, given that the use of a drum sign for the word haʔb’ ‘year’ can only be explained through the near or full homophony between Cholan *tuun ‘stone, year’ (from proto–Mayan *tooŋ) and a likely Cholan term *tuun ‘slit drum’ (from Late proto–Mayan *tuun), the use ofT548 in a year count on Takalik Abaj Stela 2 suggests that the Cholan–Tzotzilan *o > Cholan *u shift had taken place already by the time of its carving. Even if Stela 2 was carved as late as 19 BC, a conservative estimate allowing for time lags would place the completion of the morphological shift within CT at c. 100–50 BC.

Returning to the third kind of sign shared between the Epi– Olmec and early Mayan scripts, there are shared syllabic signs with shared values in both writing systems. These signs, and this last kind of sharing, are the most useful in determining the direction of influence or diffusion. One particularly cogent example of this kind of relationship is that which exists between Epi–Olmec MS44 na and Mayan T23 na (Lacadena 2010; Mora–Marín 2001: 254–57). Stross (1990) has argued that Epi–Olmec MS44 (Figure 6.8a) has a value na based on its visual identity with the down–turning ground motif common in Izapan art that can be traced back in time to Olmec iconography (Figure 6.6b), and on the Mixe–Zoque term *naas ‘earth, land,’ but not to a Mayan reflex of proto–Mayan *kab’ ~ *kaab’ ‘earth.’ The main context Stross (1990: 49, figs. 8b and 8c) draws attention to is the very close similarity between the Maya ‘sun–at–horizon’ glyph collocation and its Epi–Olmec equivalent (Figure 6.8c; Englehardt 2005: 92, figs. 3.9 and 3.10). He also suggests that Epi– Olmec MS44 may be graphically related to Mayan T23 na and therefore that the phonetic value ‘na’ for Mayan T23 may be explained as a case of diffusion from a Mixe–Zoquean–speaking Epi–Olmec scribe. The same conclusion was arrived at independently by Justeson and Kaufman (1993, 1997) through a controlled grammatical analysis of La Mojarra Stela 1 and the Tuxtla Statuette.

This conclusion is supported by another early example of the drum sign in an anniversary celebration context: in 32 BC on Tres Zapotes Stela C. Indeed, even though Stela C has an Epi–Olmec text written in pre– proto–Zoquean (Mora–Marín 2001: 254), the use of a DRUM sign to refer to ‘year’ or ‘anniversary’ in the Long Count context on Stela C can only be explained through the homophonous relationship between Cholan *tuun ‘stone, year’ and *tuun ‘slit drum,’ and not through pre–proto–Zoquean *ʔameʔ ‘year’ and *kowa ‘drum,’ which do not provide a basis for rebus–usage of the drum sign as both ‘year’ and ‘drum’ (Justeson and Kaufman 1993). Ultimately, the usage of the drum

Indeed, there are examples of Mayan T23 na in Early Classic monuments (Figure 6.8d) that display a visual affiliation with the down–turning ground motif of Izapan and Olmec art and with Epi–Olmec MS44 na. There are also early examples from Late Formative and Early Confirmation that the DRUM sign had two readings— as both ‘year’ and ‘drum’— is not found until AD 157, in the text on La Mojarra Stela 1(Justeson and Kaufman 1993, 1997). In any case, the evidence regarding T548 strongly suggests that the CT *o > *u shift occurred as early as c. 100 BC, but in any case by c. AD 100 (Mora–Marín 2001: 254). 14

13 Tú:n: tun ‘stone,’ ‘year (ending)’ from Cholan tun ‘stone’ (cf. pM * to:ŋ; pTz, Tz, and Tzo * ton; pCT * to:n; Mot to:ŋ; Toj ton, all with semantic value ‘stone’); illustrates the Cholan * o: to u shift, which did not affect Yucatecan.

140

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.8. Epi–Olmec sign MS44 and Maya signs T23, T526, and T529. a: MS44 in different contexts; b: down–turning ground motif in Izapan art and Olmec iconography; c: Epi–Olmec ‘sun–at–horizon’ glyph collocation and its Maya equivalent; d: early examples of T23 na; e: T529 WITZ; f: T526 KAB’.

Figure 6.9. Frozen uses and continuing visual associations of T23. a: Palenque Tablet of the 96 Glyphs; b: Palenque Palace Tablet; c: unidentified text from Tikal; d: Ahuelicán greenstone tablet; e: Dos Pilas Stela 8 (after Mora–Marín 2001: 680, Fig. 7.13).

Classic period Maya art and writing that suggest that the WITZ ‘mountain, hill’ glyph likewise has visual origins in the down–turning ground motif (Figure 6.8e). T23 na soon loses its original iconic motivation in Mayan writing and a derivative form becomes the standard throughout the Classic period (see Mora–Marín 2001: 754, fig. A2.8).15 Nevertheless, even though Mayan scribes likely borrowed T23 na from Epi–Olmec scribes during the Late Formative period, and even though they did not use it later as a logograph with the meaning EARTH/ LAND based on its iconic origin (T526 KAB’ ‘earth’ was used instead; Figure 6.6f), Mora–Marín (2001: 254–57) details several pieces of evidence suggesting that the Mayan scribal tradition retained knowledge of the original iconic motivation of T23.

The first comes from some frozen uses of T23 as an ideographic EARTH/CAVE in later Classic period Maya texts. For example, a glyph block from the Tablet of the 96 Glyphs at Palenque reads 7AJ–5–PYRAMID– NAH, ‘he of the 5 [pyramid] house’ (Figure 6.9a). The same glyph also appears on the Palenque Palace Tablet as 7AJ–5–PYRAMID–na–NAH (Figure 6.9b). In the latter example, however, T23 follows the PYRAMID glyph. The reason for the optionality of T23 na in these instances in unclear. There are two possible explanations: T23 na either serves as phonetic complement to the PYRAMID glyph, or it is an iconographic component of the PYRAMID glyph without orthographic value in this context. If the logographic value of the PYRAMID glyph is K’UH/CH’UH–NAH ‘temple’ or MUL–NAH ‘mound,’ the first option makes sense orthographically as a phonetic complement. But there is no secure reading of

15 Original forms of T23 were preserved iconically instead as the top part of T529 WITZ (Figure 6.8e).

141

Archaeological Paleography this glyph. In any case, the second option is more likely, and would illustrate shared semantic content across scripts. Comparing these examples with a related glyphic compound evident in later Classic period texts at Copán, Tikal, and Caracol, and with a third glyphic motif found on a Middle Formative period Olmec–style greenstone tablet reportedly from Ahuelicán, Guerrero (see Freidel and Reilly 2010: 652, fig. 10), the likely orthographic optionality of T23 is clarified. The example from Tikal (Figure 6.9c) reads 7u–PYRAMID–na–3.STONE/ HEARTHSTONES. This is a cosmological place name, indicated by the sign for 3.STONE/ HEARTHSTONES (Michael Carrasco, personal communication, 2010). When one compares this glyph compound with the motif on the Middle Formative period Ahuelicán Tablet (Figure 6.9d), it appears that the two examples are illustrating the same semantic content: PYRAMID–EARTH/CAVE–3. STONE/ HEARTHSTONES. The three dots under the down–turning motif in the Ahuelicán Tablet correspond to the three occurrences of T528 TUN ‘stone’ (see above) to the far right of the glyph block from the Tikal text. Found between the PYRAMID and the 3.STONE/ HEARTHSTONES signs in the greenstone tablet is the down–turning ground motif, which corresponds iconically to Epi–Olmec MS44 na and Mayan T23 na. Thus, T23 na may not be functioning phonetically in any of the glyphs in Figure 6.9, but instead as an iconographic complement clarifying the meaning of EARTH/CAVE.

presenting the limited examples discussed above is to elucidate the extent and primary temporal contexts of interlingual and interscribal interaction between speakers of Mixe–Zoque and Mayan languages in the context of emerging script development, rather than solely or exhaustively illustrating the considerable impact that the Epi–Olmec script had on incipient Maya writing. In any case, I will return to the development of T23 and a detailed exploration of its visual relationship with signs in Olmec iconography and other Mesoamerican scripts shortly. The particular examples of Mixe–Zoque and Mayan interlingual and interscribal interaction employed here illustrate how linguistic interaction can affect script development. Justeson and Matthews (1990: 127) suggest that linguistic representational principles derive largely derived from the grammatical structure of words in the languages they represented, with respect both to the distinctions they fail to represent and to how they represent what they do represent. Representational conventions appear to have been spread by processes of analogical change. These analogies are carried mainly via correlations with particular types of linguistic information. At this point visual arguments related to iconographic sharing and reformulation as a function of interaction begin to acquire increasing significance. Summary of Linguistic Evidence To summarize the linguistic data presented above, on the basis of diffusion an loanwords between Mixe– Zoque and GLM languages, and between Cholan– Tzotzilan and other Mayan languages within the GLM area, interregional linguistic interaction in southeastern Mesoamerica was most intense in the late Middle Formative through the Late Formative period. The phylogenetic classification of GLM languages, coupled with glottochronological estimates for their divergence (see Figure 6.5), suggests that such interaction was relatively uniform across the lowlands. These temporal contexts are framed by linguistic transfers between Oto– Mangue and GLM languages in the late Middle Formative period and morphological shifts such as the * k > č shift associated with the diversification of the Cholan– Tzotzilan language family in the Late Formative period (Figure 6.4). Allowing for time lags between interaction, diffusion, adoption, and subsequent linguistic reflection, a conservative estimate for the period of greatest interregional interaction suggested by the linguistic data would be c. 450 BC–AD 1. This temporal frame finds numerous correlations in archaeological evidence and corresponds generally to the extent of the Xe and Mamom ceramic spheres.

Taken together, these examples suggest that T23 na did in fact originate in the Middle Formative period Olmec and later Izapan down–turning ground motif, and also that the linguistic source behind the T23 representation was Mixe–Zoque *naas ‘earth, land.’ The examples in later Classic period Maya texts at Palenque and Tikal further suggest that Classic Mayan scribes and artists clearly understood the original motivation of T23, even several centuries after its borrowing (Mora–Marín 2001: 256). The final piece of evidence that confirms both the origin of T23 and the persistence of its original visual association and iconographic motivation in Late Classic texts is an example of T23 na from Dos Pilas Stela 8, dated to AD 726. This example (Figure 6.9e) clearly reflects the Late Formative iconic origin of this sign, insofar as it shows the original split mountain outline that survived as the top part of T529 (Figure 6.6e). Continuing interscribal interaction and contact between Epi–Olmec and Maya scribes in the Early Classic period may have contributed to the maintenance (or reintroduction) of the association between MS44 and T23 established during the Late Formative period (Mora–Marín 2001: 257). There exist numerous other examples that illustrate grapheme diffusion and the glyphic reflections of linguistic interaction between Mixe–Zoque and GLM in the Epi–Olmec and Mayan scripts.16 The point in 16

In the Late Formative period, roughly contemporary with the increased experimentation noted in later and T528, and MS82 and T45 (see Englehardt 2005; Justeson and Kaufman 1993, 1997; Justeson et al. 1985; Mora–Marín 2001: 258).

E.g., MS49 and T178, MS125 and T360, MS98 and T126, MS124

142

Comparative Analysis of Iconographic and Linguistic Evidence Chicanel sphere ceramics, linguistic interaction appears to become more pronounced along the southwestern border of the Maya lowlands, and is most evident between Mixe–Zoque and Cholan–Tzotzilan. Also at this time, CT began to exercise increasing influence on other GLM language groups, as indicated by reflexive CT linguistic values found across Mayan languages, a situation that also suggests extensive interaction— on a more limited scale—within the Maya lowlands. The episode of interaction appears to have diminished following the breakup of CT, followed by the marked increase in linguistic diversification across the GLM language area in the later Early Classic period (see Figure 6.5). Allowing for temporal lags, the increase in diversification during the late early Classic period suggests that linguistic interaction in the GLM area decreased roughly at or about the Formative–Classic period transition. At and immediately prior to this point, one observes the effects of linguistic interaction between Mixe–Zoque and CT languages in the Epi–Olmec and emerging Mayan scripts (Figures 6.5 and 6.6). This last point relates back to the identification of Mixe–Zoque as the language represented by an ancestral Mesoamerican script (Lacadena 2010), and allows for the observation of graphic reflections of problems of linguistic adaptation between scripts (Campbell 1984; Lacadena 1995). At the same time, one can also begin to appreciate the extent of visual relationships between distinct iconographic and scribal traditions (Figures 6.9 and 6.10). Common semantic meanings and iconographic associations of signs shared in disparate spatial and temporal contexts in the context of emerging scribal traditions suggest interaction between users of these graphic systems, whereas discrepancies and modifications to shared signs in distinct scribal contexts imply a decrease in interaction between traditions of visual representation. Below, evidence of iconographic sharing and transformations, as well as inferences regarding interregional interaction that may be drawn from such examples is presented and discussed.

of developing writing systems is necessarily limited, especially in Mesoamerica (Houston 2004a; Justeson 1986, 2012). Although an ancestral Olmec iconographic system was distributed widely throughout Formative period Mesoamerica (Coe 1965, 1975; Graham 1989; Joyce et al. 1991; Lee 1989; Lowe 1989; Parsons 1979, 1986; Quirarte 2007; Reilly 1991; Sharer 1989a), not all of the iconographic elements in this system were reformulated for use in distinct Mesoamerican scribal traditions. Additionally, it is impossible to measure precisely or determine degrees of distance between ancestral and reconfigured iconographic forms, or to establish definitive genetic associations between contemporary or overlapping artistic or scribal traditions (Proskouriakoff 1968, 1971). Relationships are evident precisely in those shared or borrowed elements, but these methodological limitations do not adversely affect my analysis, insofar as shared elements are considered as indicative of interaction rather than as representative of ‘hereditary’ relationships.

The Visual Dataset: Iconographic Transformations and Mayan Script Development

This section considers documented observable changes in the form, function, syntax, distribution, and textual recurrence of certain widely–shared Mesoamerican iconographic motifs that existed in both the ancestral Olmec system and in subsequent iconographic and script systems. These include the Lazy–S motif (Figure 2.6; Reilly 1996: 414, fig. 3), hand motifs (Englehardt 2005: 366, figs. 5.40, A4.29, A4.30; Lacadena 1995, fig. 6.134, 2010), thrones and basal bands, ritual bundles and Olmec ‘torches,’ and calendrical notations (Justeson 1986; Marcus 1992a). All of these initially iconographic motifs were later converted into written signs and assigned distinct linguistic values or readings within the Mayan script. Some of these elements (e.g., Olmec ‘torches’ and ritual bundles) are particularly indicative of long–term regional iconographic interaction, evident in the wide distribution of these motifs through time. Others are better suited to the evaluation and discussion of scribal innovation. Patterns of stylistic, syntactical, grammatical, and functional variability in these data demonstrate the quality and extent of regional iconographic and linguistic interaction—and subsequent recontextualization.

In order to correlate the processes of iconographic and linguistic transformation and recontextualization involved in the development of Maya writing with varying degrees of interaction and integration, it was necessary to construct a database of iconographic evidence that illustrates the dynamic operation of these processes over space and through time. Because this undertaking alone is sufficient to stand in its own right as the topic of scholarly inquiry (as it indeed has before; e.g., Justeson 1986; Justeson and Matthews 1990; Justeson et al. 1985; Lacadena 1995; Mora–Marín 2001), This study instead sought to create a representative dataset of previously reported evidence. A representative dataset is also desirable insofar as evidence for the incipient stages

Calendrical elements are also considered, such as day and month signs across systems (Justeson 1986; Justeson et al. 1985; Marcus 1992), and particularly early examples of the Initial Series Introductory Glyph (ISIG; Englehardt 2005: 158, 191–192, figs. A4.38, A4.39; Mora–Marín 2001: fig. 1.24), as evidence of long–term, wide–ranging interaction between cultural groups or users of an ancestral iconographic system. Calendrics, especially day signs, are particularly important and illustrate linguistic as well as iconographic interaction. For example, as the only element of content shared by all Mesoamerican scripts, day names were probably inherited from the ancestral representational system (Justeson 1986: 445). At the same time, many day names 143

Archaeological Paleography in the Maya tzolk’in 260–day count illustrate borrowing between Oto–Manguean and Greater Lowland Mayan language groups (e.g., lamat, manik, b’en; Justeson et al. 1985: 47–8, 50). Finally, a selection of linguistic changes is also considered, such as the * k’ > * č shift in the Cholan–Tzeltalan language group, and the * m > * b’ shift in Greater Lowland Mayan (Hruby and Child 2004; Justeson 1985; Justeson and Campbell 1997; Justeson et al. 1985; Kaufman 1976; Kaufman and Norman 1984; Mora–Marín 2001; Wichmann 1995). The temporal contexts of linguistic divergence reflected in these shifts illustrate how patterns of linguistic interaction vary over time. Moreover, the linguistic shifts examined here are particularly well–represented, documented, and relatively easy to trace in the emerging Mayan script (e.g., Lacadena 2010; Macri 1991; Stross 1982). The basic corpus of textual evidence that I employ is based on the work of Mora–Marín (2001: 359–60, table 1.5). Ideally, this study would employ materials and textual evidence from within the study area. This was impossible, due to the scarcity of Formative texts in general, and their near total absence in southeastern Tabasco. The only texts of which I am aware are a stone bowl, currently at the Museo Municipal de Tenosique in Tenosique, Tabasco and a waxy ware ceramic bowl, currently housed in the Museo Municipal de Emiliano Zapata, Tabasco. These objects are unprovenanced, and their respective dates are unclear. INAH investigators informed me that these objects were part of a private collection donated to the Museos de Tabasco, and that they likely originally are from Tabasco, although little can be definitively said about their original contexts of production, consumption, or deposition (José Luis Romero Rivera, personal communication, 2009, 2010). The evidence from southeastern Tabasco, although limited in nature, is significant in that it illustrates a contemporaneous Middle and Late Formative period link between the Olmec iconographic antecedents of Maya writing with origins in the Gulf coast lowlands to the west (Ochoa 1983; Pohl et al. 2002; Pohl et al. 2008) and the burgeoning scribal innovation occurring in the Petén lowlands directly to the east (Saturno et al. 2006).

Hruby and Child 2004; Josserand 1975; Josserand and Hopkins 1999; Joyce 1991; Kaufman 1976; Kaufman and Norman 1984; Macri 1991; Martin 2002; Méluzin 1992, 1995; Mora–Marín 1997, 2001; Ochoa 2008; Parsons 1979, 1986, 1988; Pohl et al. 2002; Pohl et al. 2008; Quirarte 1977, 2007; Reilly 1991, 1996; Rice 2007; Riese 1988; Schele 1991, 1999; Stross 1982, 1990; Urcid 2001, 2005; Wichmann 1995).17 The linguistic data presented above frame the discussion of iconographic transformations in the context of script development by illustrating the temporal contexts and outlining the spatial extent of linguistic interaction. Linguistic evidence suggests that interregional interaction across southeastern Mesoamerica was most intense in the Late Formative period, but that it had begun at an earlier date, at least in the Middle Formative period. Similarly, iconographic evidence illustrates shifting patterns of interaction in Formative and Early Classic period Mesoamerica. Since the imposition of unique linguistic values on a shared base iconographic system in order to clarify meaning is at the heart of script development, the temporal and spatial distribution of shared iconographic motifs allows for the detection of diachronic patterns of interregional interaction through which such elements were integrated into diverse representational systems. Detectable transformations or modifications to shared motifs, on the other hand, speak to localized innovation. Similarly, the appearance of analogous or related iconographic motifs or signs in disparate textual frameworks is suggestive of processes of innovation within the context of developing scribal traditions. Documented observable changes in the form, function, contextual syntax, distribution, and textual recurrence of certain widely–shared Mesoamerican iconographic motifs that existed in both the ancestral system and in subsequent iconographic and script systems are examined below. The incipient process of linguistic codification in emerging writing would be reflected in formal modifications and reformulations of the symbols already present in a shared iconographic or textual registry. Some of these elements are particularly indicative of long–term interregional iconographic interaction, evident in the wide distribution of these motifs through time. Researchers generally agree that an Olmec–style artistic canon established the use of stylized, nonrepresentational symbols of widespread intelligibility and meaning (Marcus 1992a). This process involved the segmentation or ‘divorce’ of independent symbols from ostensibly pictorial contexts, establishing juxtaposition as a viable organizing principle for referent units; in effect, pre–adapting them for interpretation via sequences of linguistic units to which they correspond

The database employed here is also based on a prior investigation of the development of the Classic Mayan script (Englehardt 2005) as well as the previous work of Alfonso Lacadena (1995, 2010; Lacadena and Wichmann 1999) and John Justeson (1985, 1986, 1989; Justeson and Campbell 1997; Justeson and Kaufman 1993, 1997; Justeson and Matthews 1990; Justeson et al. 1985).The previously documented work of a variety of other leading scholars on Mesoamerican iconography, linguistics, and the development of Maya writing is also incorporated (e.g., Campbell 1976, 1984, 1988; Campbell et al. 1986; Coe 1965, 1976, 1977; Dahlin et al. 1987; Fields 1991; Fox and Justeson 1980; Grube 1994; Hansen 1991; Hopkins 1997; Houston et al. 2000; Houston et al. 2004;

For the complete database, see Appendix B in Englehardt (2011). Of course, the iconographic database employed in this investigation is not intended to be exhaustive. 17

144

Comparative Analysis of Iconographic and Linguistic Evidence de facto (Justeson and Matthews 1990: 126). The consideration of elements of the Olmec iconographic system within the later Mayan script is predicated on the assumption that certain elements of the Maya system can be visually identified in Olmec iconography (or some other Mesoamerican iconographic systems or scribal tradition) and that these elements perform similar functions in both systems (Reilly 1991: 151). For example, elsewhere (Carrasco and Englehardt 2015) I suggest that many signs on the Cascajal Block appear to have descendants that maintain a relatively constant semantic value and serve similar syntactical and grammatical functions in later scripts (see also Mora–Marín 2011). Such analogous relationships among signs, when considered in conjunction with an Olmec–Mixe–Zoque linguistic affiliation and evidence of extensive diffusion between Mixe–Zoque and GLM languages, suggest that the representational basis of the developing Mayan script was a derivative of a Middle Formative Olmec iconographic complex which gave rise to an ideographic system homologous to a primitive logographic writing system (Justeson and Matthews 1990: 126). In any case, the existence of shared signs in distinct scribal traditions would indicate the presence of sustained, extensive interaction between the groups that used those scripts. Nonetheless, as Lesure (2000:74) observes, ‘...Olmec iconography was widely but unevenly distributed across Mesoamerica. In some periods and places it seems very pure; in others, it is mixed with more localized themes and styles.’ Thus, although interregional interaction is inferred, as Rosenswig (2010:49) notes, ‘the uses and meanings of Olmec imagery may have been employed in locally specific ways.’

between scribal traditions. Rather than invent an entirely new symbol, the adopting script may simply reformulate a preexisting sign to reflect a new linguistic value.18 Such processes of sign reuse and reformulation in contextually parallel situations are evident in the evolutionary development of other distinct writing systems in the world (Englehardt 2005: 469, figs. A4.21 and A4.22; Lacadena 1995). Thus, localized innovation is suggested by the particular reformulation of a commonly shared sign, often to reflect a distinct linguistic value assigned to a glyphic notation in a particular spatial, temporal, or cultural context. Additionally, subsequent variation may be expected in the evolution of the specific glyph within the script system that borrowed, reformulated, or linguistically encoded its iconographic precursor (Lacadena 1995). Such variations often formally exhibit a specific, readily apparent pattern over time (Englehardt 2005: 290, fig. 5.17, 469, figs. A4.21 and A4.22, 473, figs. A4.29 and A4.30). Formal variation and sign reformulation can also illustrate specific linguistic shifts and morphological adjustments between script systems, as suggested by linguistic data, as well as similar changes within one specific script system. Often, such scribal variation is significantly easier to identify in developed script systems, where clear predecessors for reformulated signs have been previously established and frequently employed prior to reformulation. Figure 6.10 offers three such examples of this sort of intra–script variability. In the case of the Maya writing system, these internal or script–specific reformulations occurred following extensive interaction between speakers of Mixe–Zoque and GLM languages. Linguistic evidence, as well as the developmental history of the Mayan script, thus indicates that such recontextualizations of individual signs began in the Late Formative period, reaching a peak in Early Classic times. Additionally, these reformulations suggest a decrease in interregional interaction, as the Maya writing system ceased to borrow signs from contemporary scripts (e.g., Epi–Olmec) or previously shared iconographic traditions, and instead focused on localized and divergent reformulation and linguistic codification of a widely disseminated base script or iconographic tradition.

Other motifs examined here are better suited to the evaluation and discussion of scribal innovation. Certain signs appear to be reformulations—on formal, contextual, or semantic levels—of specific glyphic or iconographic notations in antecedent spatial–temporal contexts. Systems of graphic representation and communication, even more so than spoken languages, are naturally conservative organisms, inherently resistant to sudden, arbitrary changes in their formal appearance and internal structure. Thus, when variations and modifications are observed, one can infer that there existed a valid and necessary functional motivation to predicate the alteration itself, and effect the observed transformation. Often, such modifications in visual signs presuppose a change in the specific linguistic value of the symbol. Slight modifications, combinations, or formal alterations of signs are effected to reflect new linguistic values, or, in some cases, to impose a specific linguistic value upon a preexisting iconographic notation that did not previously exist in overtly textual contexts. Further, the reformulation or reuse of specific signs may point to their renewed, revived, or sustained use in distinct linguistic contexts. It is in these cases that one may observe problems of adaptation in signs borrowed or adapted

Similarly, certain signs may be discarded if the language of the adopting script had no need of them (i.e., their phonetic value did not exist in the language behind the adopting script), if the adopting language did not have the need to differentiate certain phonetic values so finely, rendering the excess signs redundant, or if such phonetic values could be formed as diphthongs using a combination of characters already present in the shared symbolic registry (e.g., the classic Greek symbols for psi [Ψ], ‘long O’ omega [ Ώ], xi [Ξ], and the bilabial fricative phoneme phi [Φ], which were never adopted into the Etruscan or Latin alphabets, since the linguistic values represented by these signs could be formed using preexisting symbols within the Etruscan and Latin scripts). 18

145

Archaeological Paleography

Figure 6.10. Examples within the Mayan script of the reformulation of established signs to reflect new or alternate linguistic values. These examples show the versatility of single signs, and the range in which they can be reused or reformulated to reflect distinct values or linguistic variation.

If clear evidence for textual reformulation within established scripts suggests a decrease in interaction, the difficulty lies in identifying the nature and extent of interaction evident in earlier contexts. Using linguistic reformulation and sign recontextualization as a bookend, one may examine specific individual glyphic notations or iconographic motifs to determine whether these follow the same pattern of transformation, adjustment, and reformulation as their linguistic counterparts. In other words, tracing the developmental process backward in time illustrates certain signs that exhibit evidence of having resulted from the injection of localized, divergent linguistic factors into a diffused, ancestral iconographic system. In these cases, patterns of stylistic, syntactical, grammatical, and functional similarity and variability in the data demonstrate the quality and extent of interregional iconographic interaction in both spatial and temporal contexts. If the central hypothesis of this study is correct, one would expect to find a great degree of stylistic and formal uniformity in widely distributed iconographic motifs in distinct cultural contexts in the Middle and Late Formative periods, which would suggest a relatively higher degree of interaction between iconographic or scribal traditions. Such similarity should then be followed by increasing variability across systems of graphic communication, as interaction decreased and localized innovations came to the fore.

Contextualizing the Evidence Prior to presenting and discussing the iconographic and epigraphic data, several caveats are in order. First, it is necessary to point out that iconographic motifs, like ceramic traditions and, to some extent, linguistic data, can only be dated generally. This evidence is intended solely to illustrate broad patterns and general long–term processes of interaction and innovation. Second, motifs and signs are presented individually and independently. In other words, in the figures that follow, signs are divorced from their overall iconographic or textual contexts (cf. Taube 2000a). Third, as with the linguistic data, the iconographic examples presented here are not intended to be exhaustive. There exists a great deal of previous scholarship (Englehardt 2005; Fields 1991; Joyce et al. 1991; Martin 2002; Reilly 1991, 1996) on a wide variety of iconographic motifs found in the later Mayan script that are not examined here, including sprout motifs (Schele 1999) and the Olmec ‘U–element’ (Mora–Marín 2001: 260, 346; Pohl et al. 2008). A detailed examination of the Zapotec script would likely yield far more examples of shared iconography than are considered in this study(cf. Urcid 2001, 2005). Where appropriate, however, selected examples from the Zapotec repertoire are presented, which are indicative of particular shared elements and, therefore, general interaction on a wider scale. 146

Comparative Analysis of Iconographic and Linguistic Evidence This study does not engage in an in–depth examination of shared calendric elements. It is appropriate to note that calendrics are excellent indicators of interaction (Justeson 1986; Justeson et al. 1985; Marcus 1992). Widespread similarity across Mesoamerican cultural traditions and through time in the names and representations of the 260– day ritual calendar suggests long–term, wide–ranging linguistic as well as iconographic interaction between cultural groups or users of an ancestral iconographic system. In fact, as the only element of content shared by all Mesoamerican scripts, day names were probably inherited from the ancestral representational system (Justeson 1986: 445). Calendric examples have been extensively studied in previous investigations (Justeson 1986; Justeson et al. 1985; Marcus 1992a; Rice 2007; Stross 1990). As the intent is not to be exhaustive, and since such evidence has been discussed at great length elsewhere, only one calendrical element is examined as an indicator of interregional interaction: early examples of the Initial Series Introductory Glyph (ISIG; Englehardt 2005: 158, 191–92, figs. A4.38 and A4.39; Mora–Marín 2001: 461, fig. 1.24).

than in identifying interaction among scribal and iconographic traditions. Thus, even if formal parallels among signs suggest nothing about genetic relationships or affinities between script systems, such similarities certainly suggest some degree of interaction between the groups that employed those signs. This last observation ties in to the next point, which is that the presence of iconographic similarities between scribal traditions suggests the existence of a common interpretive framework in addition to a degree of iconographic interaction. That is, in order to make sense of a visual message, viewers must be familiar with the underlying narrative. Sharing a common interpretive narrative certainly implies that interaction existed between groups that employed formally similar signs in analogous contextual or syntactical circumstances. Underlying interpretive narratives themselves, however, will change over time, as will the specific system of representation and interpretation. A telling example of such a shared narrative reflected in iconographic parallels is the formal similarity between the content of Kaminaljuyú Stela 10 and the San Bartolo West wall mural, both of which display a central figure to the left of a kneeling woman with a disembodied, floating figure hovering above the head of the central figure (Michael Carrasco, personal communication, 2011; cf. Taube et al. 2010). One may conclude that both monuments are referencing a common cultural narrative, despite the fact that they were produced in disparate spatial and temporal contexts by arguably different cultures. These conclusions are debatable and not presented here as conclusive. Nonetheless, the production of a visually similar narrative in distinct spatial and temporal contexts implies the existence of interaction between the groups that created these images. Even if both narratives were produced by Maya groups the similarity in the narratives would still suggest a significant degree of highland–to– lowland interaction within the Greater Maya lowlands during the Late Formative period.

Fourth, when discussing iconographic motivations, proposing continuity with a later stage of the same script, or positing a relationship between two scripts, whether contemporaneous or not, it is necessary to consider three different sources of similarity: shared descent, contact diffusion, and independent innovation. If two signs from different scripts exhibit the same iconic motivation, one must consider the possibility that such similarity is due to the fact that the real–world source that serves as the pictorial model for the sign may be available to scribes in different social contexts. Simple signs— those containing a very general outline and few or no internal elements—can resemble one another as a result of stylistic simplification, even when they are derived from different iconographic sources. When comparing formally similar signs in different scripts, or even different historical stages of a single script, it is tempting to conclude that a genetic relationship exists between the systems. In other words, it may seem a natural and logical conclusion that one script borrowed the sign from another, or that one script assigned a distinct value to a sign shared between the two. Nonetheless, formal similarity is not always a valid basis for comparison and does not in itself suggest any sort of genetic relationship or affinity between two script systems (Proskouriakoff 1968, 1971; Quirarte 2007). In contrast, a complex sign present in two distinct scripts offers a better test case for relationship. In addition, those cases in which formally similar signs in glyphic or iconographic contexts can be shown to reflect distinct linguistic qualities in differing scripts are excellent indicators of historical relationships, suggesting that a modification of the base symbol occurred due to a shift in its linguistic value between the related scripts. This investigation is less concerned with demonstrating direct relationships between scripts

Finally, little iconographic evidence from the study area is presented. This is due to the fact that few systematic archaeological investigations have been conducted in the study area. Subsequently, available data are scarce. Where possible, iconographic motifs identified on objects found in southeastern Tabasco state, the lower San Pedro Mártir basin, and the northwest Maya lowlands are included (see Figure 6.11). Nevertheless, the presence of such objects, even in limited quantities, certainly does suggest that interaction occurred across the study area. These artifacts indicate the existence of some form of interaction between the inhabitants of the lower San Pedro Mártir basin and Olmec populations— or populations in contact with or influenced by Olmec groups—beginning in the Middle Formative period at the latest. Further evidence of iconographic interaction between the study area and other regions is the San 147

Archaeological Paleography

Figure 6.11. Objects with Olmec–style iconography found in the study area. a: Unprovenanced Olmec low relief, currently in the Museo Municipal de Tenosique; b: Olmec–style incised lápida from Balancán, Tabasco; c: Olmec–style incised lápida from Emiliano Zapata, Tabasco; d: Olmec stela from El Mirador (Ochoa 1983: 165, Fig. II.4.4); e: Incised celt from Simojovel, Tabasco/Chiapas; f: Incised celt from near Emiliano Zapata, Tabasco (Ochoa 1983: 169, Fig. II.4.8).

Andrés roller stamp (Pohl et al. 2002: 1985, fig. 2). Recent analyses of this artifact indicate that the clay used to fashion the object can be traced to the lower Usumacinta and San Pedro Mártir basins (Mary Pohl, personal communication, 2010). Given the age of the San Andrés roller stamp—dated to c. 650 BC—it again appears that interaction was occurring across and through the study area and its environs in the Middle Formative period, likely continuing well into the Late Formative period (Lowe 1977, 1989; Ochoa 1983). These observations are supported by available linguistic evidence. Despite the paucity of evidence that proveniences directly from the study area, the little data that does exists clearly suggests that processes of interaction were at play in and across the study area from a very early date, beginning at least in the Middle Formative period. The evidence offered here is intended to illustrate broad patterns of long–term interaction linked to the development of Maya writing, not to suggest specific times and places that it may have first emerged.

T23 was widely distributed throughout Mesoamerica and is evident in the iconographic registers of several cultures and scribal traditions (Figure 6.12; Justeson and Kaufman 1993, 1997; Mora–Marín 2001; Stross 1990). Temporally, the motif is evident in a number of Middle Formative period Olmec–style objects, most often in iconographic contexts (Figure 6.13). That is, the motif appears as an integral part of the iconographic composition of these objects, not divorced from the visual syntax of the narrative. One example of the motif, found in Group A of the Oxtotitlán Cave paintings, a composition known as El Diablo, (see Figure 6.13; see Grove 1970: 25, fig. 25; Joralemon 1971: 66, fig. 190), does appear to suggest the beginnings of a separation of the motif from strictly visual contexts. Nonetheless, this example is far from conclusive, and it is safe to say that the preponderance of the examples occur in strictly pictorial contexts. The occurrence of formally similar motifs in syntactically parallel situations, particularly on objects in the Izapan or South Coast tradition, suggests that other iconographic systems were familiar with the semantic and contextual meaning of the sign and employed it to signify the same concept of ‘base’ or ‘earth.’ Since the sign itself is a relatively abstract representation, it is unlikely that distinct traditions would have independently invented or chosen this particular motif to signify this particular concept. Thus, one may

Earth Bands and T23 na With these caveats in mind, we may return to the discussion of T23 na begun above. Strictly in terms of iconography, the down–turning basal band motif that many scholars have posited to be at the root of

148

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.12. Spatial distribution of Mesoamerican down–turning ground or ‘basal band’ motif related to Mayan sign T23, Formative–Early Classic period, detailing iconographic or scribal affiliation.

Figure 6.13. Rough temporal distribution of the down–turning ground or ‘basal band’ motif related to Mayan sign T23 in distinct Mesoamerican iconographic and scribal systems.

149

Archaeological Paleography conclude that the motif was spread among discrete Mesoamerican traditions via extensive interregional interaction in the Middle Formative period.

scribal traditions, suggesting that extensive interscribal interaction had waned immediately prior to and during the Early Classic period, when these modifications began to occur.

In the Late Formative period, particularly within the Epi–Olmec or Isthmian script, there are clear signs that the motif had split from a purely visual frame of reference. Rather than appearing at the base of larger pictorial compositions, it now appears to be able to stand alone as a sign in purely scribal contexts. Nonetheless, the motif continued to appear in iconographic contexts in the Izapan corpus, and in Early Maya iconography such as the Leiden Plaque. In some early Zapotec contexts at Cerro de la Campana and Monte Albán, the motif occurred somewhere in between purely visual and purely scribal contexts—sometimes standing alone and sometimes as part of a larger visual narrative. Evidence for its use in the Late Formative period thus suggests the beginnings of experimentation with and linguistic codification of the sign. The fact that the motif continued to be employed in strictly pictorial contexts implies that the interaction that facilitated the adoption of the symbol and its underlying visual narrative continued into the Late Formative period. Nonetheless, the innovation and incipient linguistic codification noted in the Isthmian and early Maya traditions also suggests a diminishing degree of interaction, or at the least, more limited or concentrated interaction between the Epi–Olmec and Mayan scripts, or speakers of Mixe–Zoque and Mayan to the east of the Isthmus of Tehuantepéc.

Patterns of distribution and variability in T23 and its iconographic precursor thus complement linguistic evidence associated with this sign in suggesting extensive interregional interaction in earlier temporal contexts of the Middle Formative period. In the Late Formative period, experimentation with and variability in the representation of the down–turning motif increase, and interaction appears more constrained spatially to the Epi–Olmec and Maya traditions in the southeast lowlands. The marked rise in formal and syntactical variability in the Late Formative and Early Classic periods and the dramatic recontextualization of the sign solely in the Mayan script suggest a decrease in interregional interaction at this time. Such increasing variability in representations of T23 and the down– turning motif within the Mayan script parallels processes of centralization and integration into a strictly lowland Maya cultural sphere as the Maya lowlands turned inward and its frontiers became increasingly demarcated and noticeably less porous (cf. Englehardt 2010).19 Vegetal Bundles or ‘Torches’ The vegetal bundle or ‘torch’ motif is another example that illustrates the extent of iconographic interaction across Middle Formative period Mesoamerica. The motif was widely distributed across the whole of highland and lowland Mesoamerica, suggesting that multiple cultures and representational traditions came into contact with the design, were familiar with its meaning, and employed it in semantically and syntactically analogous compositional frameworks (Figure 6.14). In almost all instances, the motif is held in the hand or the crook of the elbow of a central figure. Like the example of the down–turning motif from Group A of the Oxtotitlán Cave murals, the bundle motif also appears in the Olmec tradition within contexts that are not purely iconographic, as signs CB12 and CB29 on the Cascajal Block text (Figure 6.15a; cf. Rodríguez Martínez et al. 2006: 1612, fig. 4), suggesting that experimentation with the recontextualization of this sign began at an exceptionally early date (Carrasco and Englehardt 2015; Mora–Marín 2011). The motif appears in Late Formative iconographic traditions in semantically parallel situations. For example, the central figure on a painted stucco vessel from Teotihuacán holds a torchlike bundle in his outstretched hand (Figure 6.15h), and two individuals in the West wall murals at San Bartolo hold aloft an iconographically similar element, identifiable

In the Early Classic period, variability in visual representations of the base motif increased dramatically, most visibly in the Mayan script. Maya scribes continued to modify the base down–turning motif, to the point that T23 eventually lost its semantic association with its iconographic predecessor. In addition, new signs were created to clarify distinctions between various meanings associated with the original visual motif and to relate new linguistic or phonetic values (see Figure 6.13; cf. Figures 6.6 and 6.10). Such experimentation with and modification of the base sign suggests a breakdown in the system of visual interpretation associated with the sign and its formulation. This suggests that the underlying visual narrative that once imparted meaning to the motif had been fundamentally altered. That the sign no longer functioned within a strictly visual framework during the Early Classic period in the Mayan script suggests that the users of the sign had lost touch with the system of interpretation that previously established the semantic meaning of the sign, necessitating the injection of linguistic elements to clarify the value of the visual element. A decrease in the level of interregional interaction may have resulted in scribes losing touch with prior interpretive principles and visual organizing frameworks. This conclusion is supported by the fact that processes of reformulation and recontextualization do not appear to have occurred contemporaneously in other

I am inclined to agree with Mora–Marín’s (2001: 252) suggestion that interaction between Epi–Olmec and Maya scribes likely continued well into the Early Classic period. Nevertheless, the paucity of the Epi– Olmec corpus and the absence of Early Classic period Isthmian texts renders such conclusions mere speculation. 19

150

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.14. Spatial distribution of Olmec vegetal bundle or ‘torch’ motif, Middle Formative period.

Figure 6.15. Subsequent iterations of the Olmec iconographic bound vegetal bundle or ‘torch’ motif in distinct Mesoamerican iconographic and scribal systems. a: CS 29 bundle element (L) and CS 12 torch element (R) in glyphic contexts on the Cascajal Block; b: Justeson and Kaufman’s (1993, 1997) Epi–Olmec bundle sign, La Mojarra Stela 1 and Tuxtla Statuette; c: MS36 bundle element, La Mojarra Stela 1; d: San Bartolo West wall mural (note the vertical bindings that identify the element as a bundle); e: ch’ok bundle compounds from Copán Altar Q; f: T122:150 TAJ, taj, ‘torch;’ g: Torch element on Piedra Labrada Stela 1; h: Teotihuacán style torch motif from stucco vessel at Teotihuacán (Taube 2000a: 46, fig. 35; i: Teotihuacán torch motif in glyphic contexts, Plaza de los Glifos, La Ventilla.

through the binding bands that frequently occur in association with this motif (Figure 6.15d, e).

(Justeson and Kaufman 1993, 1997; cf. Houston and Coe 2003). Nonetheless, both signs retain a connection to the original visual motif, with MS36 displaying the characteristic binding bands and the bundle sign on the Tuxtla Statuette possibly being held by a human hand (note the thumb projecting from the top left in Figure 6.15b). The torch on the Late Formative Piedra Labrada Stela 1 (Figure 6.15g) also retains a strong visual connection with the base motif. Similarly, the Classic

Also in the Late Formative period one may again observe processes of increasing separation of the motif from its visual frame of reference. Iterations of the motif in the Epi–Olmec script (Figure 6.15b, c) show the motif divorced from a purely iconographic framework, possibly infused with a Mixe–Zoquean linguistic value 151

Archaeological Paleography Maya sign for torch, T122:150 (Figure 6.15f) clearly reflects the bound bundle element.

or simply chose to ignore—the prior organizational and interpretive principles associated with the motif.

Within the Mayan script during the Early Classic period, the motif appears to have acquired a distinct meaning, as part of the T600 CH’OK compound frequently encountered at Copán and associated with the ‘lineage founder’ compound (Figure 6.15d). Although the sign is clearly analogous to the base iconographic motif—note the binding bands in Figure 6.15d—it has been assigned a new semantic value, one apparently unconnected to its former meaning or connotation. One may again observe the imposition of a new semantic or linguistic value on a widely distributed iconographic motif. The fact that the motif displays such little formal and semantic variability in earlier temporal contexts suggests widespread interaction during the Middle Formative period. Subsequent iterations of the sign display increasing formal variability across iconographic and scribal traditions in the Late Formative period, although the motif generally occurred in semantically analogous contexts. This fact suggests that interregional interaction continued, as the underlying visual interpretive framework had not yet been discarded completely. In the Mayan script during Early Classic period, the motif was completely recontextualized and assigned a distinct value, despite retaining a formal visual connection with its iconographic precursor. Again, the divorce of the sign from a purely visual frame of reference suggests a decrease in interaction, as Maya scribes lost touch with—

Hand Motifs The distribution and variability evident in hand motifs across Mesoamerican representational traditions illustrates processes of interaction and recontextualization (cf. Boot 2003). Hand motifs were common throughout Mesoamerican iconographic and scribal systems, and were distributed broadly across traditions (Figure 6.16). Hand motifs are not, however, complex or abstract visual representations. It is therefore entirely possible that distinct traditions could have developed such similar depictions of hands independently. Nevertheless, when one considers the semantic values associated with distinct forms of hand motifs, as well as their semantic connotations, it does appear that processes of interaction were behind the initial distribution of the hand motif in the Middle Formative period. Four distinct types of hand motif evident in various Mesoamerican iconographic and scribal traditions are considered: a flat, open hand with an upward projecting thumb (the ‘thumbs up’ motif), a closed, ‘grasping’ hand, an outstretched, flat hand, and an open hand that appears to be casting seeds (Figures 6.17–6.20). Not all motifs are found in all Mesoamerican iconographic or script systems, but each motif illustrates similar processes of interaction on a wide spatial scale during the Middle Formative period and increasing semantic, if not formal, variability through time.

Figure 6.16. Spatial distribution of Mesoamerican disembodied hand motifs, Formative–Early Classic period, detailing iconographic or scribal affiliation. Only securely provenienced examples are shown.

152

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.17. Rough temporal distribution of the outstretched, ‘thumbs up’ hand motif in distinct Mesoamerican iconographic and scribal systems.

Figure 6.18. Rough temporal distribution of the flat, outstretched hand motif in distinct Mesoamerican iconographic and scribal systems.

153

Archaeological Paleography

Figure 6.19. Rough temporal distribution of the ‘grasping’ hand motif in distinct Mesoamerican iconographic and scribal systems.

Figure 6.20. Rough temporal distribution of the ‘casting’ hand motif in distinct Mesoamerican iconographic and scribal systems.

154

Comparative Analysis of Iconographic and Linguistic Evidence Despite a large number of objects and texts that contain hand motifs in various Mesoamerican traditions, frustratingly few have a secure provenience. Nevertheless, the ample distribution of hand motifs in nearly identical semantic contexts across Mesoamerican traditions certainly suggests that all iconographic and scribal systems that employed the motif shared a common interpretive framework to make sense of the visual narrative. Creating meaning is significantly less complicated when dealing with a sign that is not formally complex, is easily understood, and retains a great degree of visual similarity with its object of reference. In the cases of the ‘thumbs up’ and outstretched flat hand motifs (Figures 6.17 and 6.18), it is more difficult to determine meaning in the earlier, Middle Formative period contexts of Olmec iconography. Moving forward in time, very little formal variability in representations of these motifs appears across script traditions. Nonetheless, one may observe increasing segmentation of the sign and the gradual abandonment of the pars–pro–toto representational convention in the Isthmian, Zapotec, Teotihuacán, and Maya traditions. In these later contexts, the sign stands alone and functions independently. In the Mayan script, the T670 and T713 compounds associated with the motif acquired linguistic values by the Late Formative period, just as appears to have occurred in the Epi–Olmec and Zapotec scripts (Justeson and Kaufman 1993, 1997; Urcid 2001, 2005). Within the Maya writing system, the specific values associated with T670 and T713, CH’AM and K’AL, respectively, refer to ‘grasping’ or ‘receiving’ and ‘binding.’ These terms are certainly associated with activities performed by or with the hand. Nonetheless, Maya scribes appear to have reworked the underlying interpretive framework. Although still retaining an iconic resemblance to its visual referent, the motif at the heart of both signs depending on syntactical and linguistic principles for its interpretation in the Mayan script. In contrast, the motif depends on its inclusion and relation to a larger pictorial narrative in Olmec iconography and, to some extent, the visual tradition of Teotihuacán.20 In addition, the T217 iteration of the flat, outstretched hand motif functioned as a phonetic sign or a negative marker and by the Early Classic period had become completely divorced from any specific visual interpretation. The process of recontextualization of the shared motif in the Mayan script appears to be directed at conferring an increasing degree of semantic specificity on the sign, although in this case little formal variability ensued.

wide distribution and nearly indistinguishable semantic contexts in which these particular motifs are encountered suggest a common interpretive narrative across early Mesoamerican iconographic systems. This fact implies that the groups or cultures that employed the motifs were in close contact, and that some form of interaction—if not direct borrowing—between them existed. As with certain other motifs discussed above, previous investigators have suggested that a divorce from iconographic contexts occurred at an early time, particularly in the case of the ‘hand casting seeds’ motif evident on the Humboldt Celt (Figure 6.23; Justeson 1985, 1986; Justeson et al. 1985: 35, fig. 3; Mora–Marín 2001). It is difficult to detect any significant formal variability in subsequent renderings of this motif, and in all cases the semantic content remained constant. Although such continuity suggests substantial interaction through time, the lack of formal or semantic variability in later contexts does not suggest any specific conclusions regarding the effects of such interaction on formal innovation of signs in the Mayan script, since both formal and semantic content remained relatively constant over time and across traditions. Like the flat, outstretched hand motif, the linguistic codification of the initial iconographic motif as T733c.710v CHOK (‘to scatter,’ or ‘to throw’) does suggest that Maya scribes sought to assign greater semantic specificity to the sign in the Early Classic period. The base T710 sign at the root of this compound also functioned within the Mayan script as a phonetic sign with the value ye, as well as a third person possessive pronoun and as an adjective meaning ‘revealed.’ Although formal modifications to the sign were minimal, some semantic variability is evident in later temporal contexts within the particular contexts of Maya writing. The grasping hand motif (Figure 6.19), in contrast, displays marked formal and semantic variability through time and across iconographic and script systems. The motif appears in the Isthmian and South Coast traditions early in the Late Formative period, suggesting widespread interaction. Representations of the motif display marked formal variability in the later part of the Late Formative period, and are also evident in the glyphic repertoires of the Zapotec and Teotihuacán traditions at this time. In general, the initial Middle Formative period motif appears to have formed the basis of three subsequent Late Formative period sign formulations: a general down–turning gasping hand, an upturned hand in a grasping position, and an open hand in an apparent grasping pose. These new iterations suggest that distinct traditions sought to establish new values for the motif, or at the least to clarify existing semantic values of the sign. In turn, one may conclude that the need for such semantic clarification stemmed from a breakdown in a common system of narrative interpretation, suggesting decreasing levels of interregional interaction. This conclusion is also implied by the fact that novel formulations and semantic values of the sign only appear in two

The grasping and casting hand motifs (Figures 6.19 and 6.20) are relatively easy to decipher semantically, and display little formal variability through time. As with the ‘thumbs up’ and outsretched hand motifs, the 20 Within Middle Formative iconography, hand motifs do sometimes appear to stand alone, disconnected from wider pictorial narratives (e.g., the Tlatilco ceramic vessel and Tlapacpoya and Tabasco ceramic sherds; Joralemon 1971: 26, fig. 40, 41, fig. 114). In those cases, however, the semantic value of the sign cannot be reliably determined.

155

Archaeological Paleography Mesoamerican script traditions, Epi–Olmec and Mayan. This fact suggests that extensive interaction continued between these systems in southeastern Mesoamerica, a conclusion that parallels available linguistic evidence. The fact that new formal and semantic formulations of the motif do not appear in other traditions implies a lesser degree of interaction, as script systems imposed increasingly localized values to clarify meaning in the absence of a widespread or common program of visual narrative interpretation.

Early Classic period temporal contexts within the Mayan script. Such specificity was necessitated by a breakdown of the common program of visual interpretation that had previously sustained the general semantic meaning of the motif.21 This development, in turn, suggests a decrease in the amount of interregional interaction, or at the least that interaction became more focused in a narrower spatial context between the Epi–Olmec and Maya systems. This process is particularly evident in the development and reformulations of the grasping hand motif.

In the Mayan script during the Early Classic period, the motif was extensively reformulated, modified, and combined with other motifs, resulting in a large number of new signs that not only reflected increasingly specific linguistic and semantic values, but also varied in formal aspects. The T714 sign displays a great deal of formal similarity to the motif observed on San Lorenzo Monument 42 (see Figure 6.19). Its meaning, ‘to grasp’ also appears to reflect a similar semantic value as the original iconographic precursor. This continuity implies the existence of a shared referential basis for visual interpretation of the sign, suggesting that interaction between systems was involved in its development. Other iterations of the grasping hand motif display localized innovation and the establishment of specific linguistic values on formal modifications of a widely distributed iconographic base. The Tnn k’o phonetic sign is of particular significance insofar as it reflects a glottal consonant value. If, as Lacadena (2010; see Figure 6.2) suggests, glottal phonemes did not exist previously, the Maya writing system would have had to invent a way to express them. As shown in other language systems, existing signs were often simply manipulated to reflect such changed phonetic values (Englehardt 2005: 469, figs. A4.21 and A4.22; Lacadena 1995).

The Lazy–S The Lazy–S motif illustrates a similar process of increasing semantic and linguistic specificity while retaining a significant degree of formal similarity through time and across traditions (Figures 6.21 and 6.22). Reilly (1996) has previously made a strong case for both the formal and semantic continuity of the Lazy–S motif as ‘cloud’ in Olmec iconography and the Mayan script. More recently, Urcid (2001, 2005) has also identified the Lazy–S as ‘cloud’ motif in the Zapotec script. In this case, one may observe an abstract sign whose formal properties and appearance changed very little over the course of its use in Olmec iconography and subsequent script systems. The semantic value of the motif and its general syntactical context also remained relatively constant from its Middle Formative period origins through the Postclassic period. The distribution and formal continuity of the Lazy–S motif strongly suggests that interaction was involved in its sharing and transfer between traditions. As T632 MUYAL ‘cloud,’ the motif survived well into the Postclassic period within the Mayan script, although in these contexts the linguistic codification of the symbol suggests the desire to attach an increasing degree of specificity to the interpretation of the motif. Early examples of T632 at Naj Tunich and Río Azul indicate that the process of linguistic codification began in the Late Formative period. The only potential recontextualization of the sign occurs in the Mayan script, first apparent near the Late Formative–Early Classic period transition. Although its meaning is debatable (possibly ‘cloud,’ possibly ‘lifted’ or ‘raised up’), T367 WAK appears to derive formally from the Lazy–S motif, and certainly carries a new linguistic value. Semantically, the interpretive leap between ‘cloud’ and something ‘lifted’ or ‘raised up’ is not great. It is thus not difficult to imagine a slight recontextualization of the sign to reflect an ideographically close semantic meaning. In the Lazy–S example, little variability through time and across traditions is evident. The fact that the widely shared motif was assigned a linguistic value in the Mayan script, however, does intimate a decrease in the overall scope and extent of interregional interaction

On the whole, the widespread formal, contextual, and semantic similarity in the distinct hand motifs evident across Mesoamerican iconographic and script traditions throughout the Middle and Late Formative periods suggests the existence of a great deal of interregional iconographic interaction at this time. In some cases, such as the ‘thumbs up,’ casting, and flat, outstretched hand motifs, it is likely that formal similarity stemmed from a natural commonality in the representation of a base iconographic referent. In other words, there are only so many ways that one can depict a human hand, and similarity in such representations is therefore unsurprising. The establishment of specific linguistic values attached to such depictions nevertheless does suggest a breakdown in interaction, as a common narrative program of visual interpretation gave way to increasingly specific semantic and linguistic values in certain scribal traditions. The instructive point here is that although formal variability is minimal in these cases, increasing semantic variation and a greater degree of linguistic specificity is certainly evident in the later,

Specificity would also have been necessitated by the presence of particular values in the languages of certain script users, such as glottalized consonants, that were not shared among all of the languages of groups that initially utilized the common base motif and the system of visual narration that facilitated its semantic interpretation and meaning (see Lacadena 2010). 21

156

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.21. Spatial distribution of Mesoamerican Lazy–S motif, Formative–Postclassic period, detailing iconographic or scribal affiliation.

Figure 6.22. Rough temporal distribution of the Lazy–S motif in distinct Mesoamerican iconographic and scribal systems.

in the Late Formative period, as Maya scribes sought to introduce linguistic principles of graphic interpretation for a sign whose meaning had previously depended on

a common visual narrative framework. In this instance, formal variability is insignificant; rather, the introduction of linguistic values is most telling. 157

Archaeological Paleography concluded, in concert with the findings of previous studies, that the widespread formal and semantic continuity in the representations and meanings of the individual day names in the 260–day ritual calendar does suggest a great deal of interaction in the distribution and development of that particular time–keeping system. Discrepancies and disjunction in representations of the 365–day solar calendar across traditions, intimate localized innovation and a distinct lack of interaction. I will not recapitulate these arguments here. I mention them only to illustrate that calendrical evidence, although abundant and a frequent subject of inquiry (see Marcus 1992a or Rice 2007 for excellent overviews), often suggests that processes of interaction and divergence operated simultaneously.

Human Foot Motif The foot motif (Figures 6.23 and 6.24) is exceedingly common in Mesoamerican iconographic and script traditions. Like the various hand motifs, the formal continuity of foot signs in subsequent iconographic and scribal traditions is most likely attributable to the relative simplicity and self–evident nature of the base iconographic referent. Within Middle Formative period Olmec iconography, particularly on La Venta Monument 13, it is possible that the motif had already begun to undergo processes of segmentation and divorce from purely visual contexts, as observed with certain examples of hand and basal band motifs. It is not until the Late Formative period that one encounters incontrovertible evidence of the motif functioning in contexts completely divorced from visual principles of interpretation in both the Mayan and Zapotec scripts (Justeson et al. 1988). The motif continues essentially unaltered through the Postclassic period, always with semantic meanings related to motion that derive naturally from its ideographic concept (i.e., foot motif as ‘road,’ ‘to go,’ ‘to ascend,’ etc.; Figure 6.24). It is significant that in these later contexts the particular semantic meaning of the motif appears to be conveyed with an increasing degree of specificity,22 suggesting innovative values associated with the base motif. Given the relative simplicity of the motif, it is entirely possible that specific signs were created independently in particular Mesoamerican script traditions. Consequently, the temporal and spatial distribution of the motif reveals relatively little about the nature and extent of interregional interaction, or the role of interaction in its development. What is noteworthy is that alternate meanings associated with the motif appear at and immediately after the Late Formative–Early Classic period transition. This fact again suggests that it was in this temporal context that new linguistic and closely related alternative semantic values began to be associated with this widely shared iconographic motif. As with the Lazy–S motif, one may observe very slight reformulations or recontextualizations of the sign to reflect an ideographically similar semantic meaning in purely scribal contexts.

One piece of calendrical data is appropriate for consideration here: the Initial Series Introduction Glyph (ISIG; Figures 6.25 and 6.26). The ISIG element was not shared among all Mesoamerican script traditions. Additionally, whereas other timekeeping systems have a much longer history in Mesoamerica, the ISIG does not appear until the Late Formative period, and thus cannot illustrate as long a process of interaction as other motifs examined here. Despite these facts, the ISIG offers an important advantage, in that the sign can be dated securely, offering firm dates for innovations related to this motif. Elsewhere (Englehardt 2005: 478, figs. A4.38 and A4.39; see also Pohl et al. 2002: 1985, fig. 3), I have argued that the ISIG developed from certain calendrical motifs illustrated on the San Andrés roller stamp. It is possible that the motif on the roller stamp is unrelated to calendrical elements (as explored below). In that case, the tripartite ISIG element that first appears on Tres Zapotes Stela C c. 32 BC would seem to be an ex novo development specific to the southeastern subgroup of Mesoamerican scripts. This innovation can be securely dated to the Late Formative period, where it appears on other texts in the Epi–Olmec tradition. At the same time, the motif is evident in the Izapan tradition of the South Coast, most notably at Takalik Abaj. Justeson (1986; Justeson et al. 1985) suggests that a similar motif is evident in certain Zapotec texts, although subsequent investigations of Zapotec writing (Urcid 2001, 2005) do not suggest a shared semantic or linguistic value, although the occurrence of the motif with numerical and other apparently calendrical elements is intriguing. By the beginning of the Early Classic period, the sign appears as T124 in the Mayan corpus, illustrating only slight changes (e.g., the occasional inversion of the central element; see Figure 6.26) throughout its use in the Classic Mayan script.

Calendrical Notations Similarities in calendric systems and the graphic representation of calendric elements have often been cited as evidence for extensive interregional interaction in the Middle and Late Formative periods (Justeson 1986; Justeson et al. 1985; Marcus 1992a; Rice 2007; Stross 1990). Previously (Englehardt 2005: 191–2, figs. 4.17–4.20) I have explored this possibility, and

When considered in conjunction with the development of the TUN/haab’ sign (Figure 6.7), and within the contexts of Mesoamerican calendrical systems, the ISIG is an excellent indicator of the innovation that occurred in time–keeping systems during the Late Formative

E.g., with semantic determinatives or phonetic markers in the Aztec script (Alfonso Lacadena, personal communication, 2004, 2009), or through the visual abstraction of T843 compounds such as T32.843v T’AB–[yi] t’ab’ay, ‘to ascend,’ ‘to rise up’ (cf. T32:843v[17], T46.843[17], T45.843[17]). 22

158

Comparative Analysis of Iconographic and Linguistic Evidence

Figure 6.23. Spatial distribution of Mesoamerican disembodied foot motif, Formative–Postclassic period, detailing iconographic or scribal affiliation.

Figure 6.24. Rough temporal distribution of the disembodied foot motif in distinct Mesoamerican iconographic and scribal systems.

period. Coe (1976) and Justeson et al. (1985) have discussed at length the evidence for the invention and diffusion of the Long Count notation system, which is shared only by the Epi–Olmec and Mayan scripts. The

fact that the ISIG was shared among the Epi–Olmec and Maya traditions suggests the continuation of the extensive interaction that resulted in such widespread formal and semantic similarity between representations 159

Archaeological Paleography

Figure 6.25. Spatial distribution of possible Mesoamerican calendric count sign related to Maya Initial Series Introductory Glyph (ISIG), Formative–Early Classic period, detailing iconographic or scribal affiliation.

Figure 6.26. Rough temporal distribution of possible Mesoamerican calendric count sign related to Initial Series Introductory Glyph (ISIG) in distinct Mesoamerican iconographic and scribal systems (cf. Englehardt 2005: 478, figs. A4.38 and A4.39).

160

Comparative Analysis of Iconographic and Linguistic Evidence of the 260–day calendar. Nonetheless, the new time– keeping system implied by the development of the ISIG, the Long Count, appears to have been unique to the southeastern branch of Mesoamerican scripts. Previous studies (Justeson 1986; Marcus 1992a; Rice 2007) have suggested that the Long Count developed to meet the need for increasing specificity in time–keeping, in much the same way as distinct versions of the 365–day solar calendar were developed by different Mesoamerican cultures. The very fact that the Long Count was not shared outside of the southeastern branch could suggest that interaction had declined at or about the moment of its inception. Alternately, it could indicate that problems of specificity in time–keeping were solved differently in distinct traditions. In either case, an overall decrease in the amount of interregional interaction appears evident. This is doubly true when one considers the possible Zapotec example from the south platform of Monte Albán (see Figure 6.26). The presence of this sign in the Zapotec corpus does suggest that some sort of interaction had occurred after the invention of the Long Count and the introduction of the ISIG. Its scarcity in Zapotec texts, however, suggests an interruption in the interaction that had previously occurred between the southwestern and Oaxacan traditions in or immediately following the Late Formative period. The increasing exactitude in calendrical recordings implied by the development of the ISIG parallels the augmented linguistic specificity suggested by graphic reformulations of other shared iconographic motifs.

or ‘u’ element is evident in several Late Formative period traditions (Pohl et al. 2002: 1985, fig. 3; Pohl et al. 2008). Such examples exhibit little internal variability and do not differ significantly in formal or semantic contexts between script traditions at this time. Once again, the link between the Epi–Olmec and Early Maya traditions appears especially pronounced. Nonetheless, the motif appears in the Zapotec corpus and has been documented on portable luxury objects from as far afield as Costa Rica (Mora–Marín 1997, 2001). These facts suggest a shared system of interpretation facilitated by a great deal of interscribal interaction across regions during the Late Formative period. In the Early Classic period within the Mayan script, variability in the formal characteristics of the motif increases significantly. The motif is combined with other signs, such as the hand motif (T670), to create new meanings. Further, substitution patterns become apparent; in some cases, T584 B’EN substitutes for T533/534 AJAW.23 These developments suggest that Maya scribes were experimenting with both the form and meaning of the motif when encoding the sign with a linguistic value. The value of glyph compounds that employ the motif is much clearer in the Mayan script—vis à vis other systems—implying that the Maya tradition sought to confer a greater degree of specificity on the semantic, contextual, and linguistic value of the sign. There is little to no evidence that a similar process occurred in other Mesoamerican scripts. The absence of such evidence suggests that Mayan scribes had some specific need to codify interpretation of the motif in a more coherent and discrete fashion in order to more precisely determine its meaning. This fact suggests a breakdown in the common scheme of visual interpretation that had facilitated the wide distribution of the motif in earlier temporal contexts. The experimentation and assignation of new and unique linguistic values to the motif that occurred in the Mayan script in the Early Classic period implies a concurrent decrease in the interregional interaction that had served initially to spread both the motif and its iconographic interpretive framework.

Further, the fact that the ISIG invariably occurs with a TUN/haab’ sign—or one of its precursors—suggests that the latter motif had already undergone the semantic transition from ‘stone’ to ‘year,’ and now functioned purely in a linguistic role within the Long Count calendrical system during the Late Formative period. It is possible that the value of T124, tzi, is related to its proposed linguistic value in MZ languages ‘tza’ (Justeson and Kaufman 1992, 1997). It also appears that T124 at some point in the Classic period acquired yet another phonetic value, ‘tzo,’ possibly related to its original ‘tzi’ value. In either case, it is again apparent that Maya scribes were experimenting with both the form and meaning of the sign itself, and fixing new semantic and linguistic values to the motif. The fact that such values are not evident in other script traditions suggests that a decline in interaction had occurred.

In sum, whether this motif was initially a calendrical element or a purely pictographic sign within the Middle Formative period Olmec registry is relatively unimportant. In either case, it is clear that the motif was widely shared throughout Mesoamerica in the Late Formative period and contextualized within a common system of visual interpretation, suggested by formal and

It is possible that the ISIG had no precedent in Middle Formative period Olmec iconography. The element on the San Andrés roller stamp that some have suggested as an early precursor of the ISIG (Pohl et al. 2002) may in fact be more closely related to the ‘foliated ajaw’ motif common in Maya iconography and found in other Mesoamerican iconographic traditions (Figure 6.27). In the case of this motif, Olmec examples are again scarce. Nevertheless, the sprouting element atop an ajaw glyph

It is possible that the T584 sign and its iconographic precursors were involved in the development of the contextual value of T533/ T534 as AJAW, considering the apparent relation suggested by T364 (T774.584v; Mora–Marín 2001: 752–53, figs. A2.6e and A2.7a). The substitution of T584 for T533/534 in certain contexts is therefore logical. Although the linguistic value of such substitution compounds is distinct (i.e., T126.584:670 as ya–CH’AM vs. T126.533:670 as ya–AL), the substitution compounds carry an identical semantic– contextual value as a relationship glyph indicating ‘child of [mother].’ 23

161

Archaeological Paleography

Figure 6.27. Rough temporal distribution of a possible ‘foliated ajaw’ motif in distinct Mesoamerican iconographic and scribal systems.

registry, which itself is an important visual marker of status in Maya elite portraiture (Fields 1991; Freidel and Suhler 1999; Reilly 1991), and certain glyphic elements, such as the double merlon motif and the combination of the crossed bands and ‘u’ element often associated with bloodletting or in iconographic depictions of implements associated with the act of auto–sacrifice (cf. Carrasco and Englehardt 2015: 2, fig. 1, 17, fig. 12; Englehardt 2011; Joralemon 1971; Joyce et al. 1991; Mora–Marín 2010; Pohl et al. 2008; Reilly 1991; Rodríguez Martínez et al. 2006: 1612, fig. 4). A complete exploration of these and numerous other visual elements that are suggestive of processes of interaction is beyond the scope of this investigation. In general, though, there is little formal variation between the various representations of these motifs across iconographic and script traditions, and it is likely that semantic meaning was analogous among traditions. Such continuity suggests that the interpretation of these motifs depended on a shared system of interpretation, implying interaction of some sort between the cultural traditions that employed these elements.

contextual similarities in depictions of the motif across traditions during this period. Interaction and borrowing is especially pronounced in the southeastern subgroups of scripts at this time given the particularly strong formal and semantic parallels in representations of the motif between the Epi–Olmec and Maya systems. In the Early Classic period, Maya scribes experimented extensively with the motif and assigned particular semantic, contextual, and linguistic values to the motif, often combing it with other elements to create new meanings and increasingly specific linguistic connotations within a new framework that no longer relied on visual principles of interpretation. These innovations did not occur extensively in other scribal contexts within Mesoamerica. It appears that the development of the motif as a specific written sign within the Mayan script was contemporaneous with a period of decreasing interaction between script traditions in Mesoamerica during the Early Classic period. Summary of Iconographic Evidence There exist numerous other iconographic motifs and glyphic notations that illustrate similar processes of interaction in script development. These include the knotted vegetal headdress (cf. Carrasco and Englehardt n.d.; Englehardt and Carrasco 2015; Justeson 1986; Justeson and Mathews 1990; Mora–Marín 2001: 491, fig. 3.5; Stuart 2015), a motif that appears to have given rise to the Jester God headdress in the Maya visual

Although it is possible that these and other glyphic elements were adopted directly from the ancestral Middle Formative sign system into subsequent script traditions, processes of interaction, reformulation, and recontextualization are still evident. Formal and semantic continuity through time suggests interaction through 162

Comparative Analysis of Iconographic and Linguistic Evidence the Middle and Late Formative periods. In the Maya script of the Early Classic period, however, the motifs appear to have been recontextualized. Nonetheless, the recontextualizations themselves depended on previously shared iconographic meanings or connotations—hence the great degree of formal similarity evident in these examples. Taken as a whole, the examples discussed above suggest a pattern of interaction and innovation similar to that indicated by the ceramic evidence. That is, the formal and contextual continuity across traditions in the Middle and Late Formative periods is indicative of interaction insofar as such continuity was predicated on a shared system of iconographic interpretation via a common underlying visual narrative. In the later contexts of the Early Classic period, Maya scribes reformulated and recontextualized these elements. Such innovation suggests a breakdown in shared interpretive principles, implying a period of decreasing interregional interaction.

adaptation when new linguistic values are juxtaposed on signs that had previously depended on a shared interpretive framework. As in other instances of developing writing systems, existing signs in the Maya repertoire were simply manipulated, reformulated, and recontextualized to reflect novel and innovative phonetic values (Bagley 2004; Baines 2004; Englehardt 2005; Houston 2004a; Josserand and Hopkins 1999; Justeson 1986; Lacadena 1995, 2010; Robinson 2003; Rogers 2005; Senner 1989; Stross 1990). These examples also reveal the effects of Late Formative period linguistic interaction on subsequent innovative glyphic formulations. As a whole, the linguistic evidence points to widespread interregional interaction in the Middle Formative period, followed by decreased interaction concentrated at the southeastern periphery of Mesoamerica between MZ and GLM languages and within the GLM language area in the earlier half of the Late Formative period. In the later Late Formative and through the Early Classic period, morphological shifts and episodes of linguistic diversification suggests increasingly diminished levels of interaction and the introduction of novel linguistic values on the incipient Mayan script.

Conclusions This chapter has presented and discussed a selection of linguistic and iconographic data suggestive of processes of interaction and innovation. Linguistic evidence suggests a great deal of interregional interaction in the later Middle Formative period. This interaction is reflected in the widespread distribution of Mixe–Zoque loan words across Mesoamerican language families and in the southeast by the lack of diversification of Mayan languages at this time. By the Late Formative period, linguistic interaction appears to have diminished, or at least more concentrated or restricted in spatial contexts to southeastern Mesoamerica, between MZ and GLM languages. Morphological shifts and linguistic diversification within the GLM language family suggest a decrease in interaction between regions during the Late Formative period, although interaction within the GLM language area continued. Such morphological shifts were reflected in incipient glyphic notations. In addition, the primacy of CT languages and the representation in glyphic form of CT linguistic values indicates continuing intra–regional interaction within the Maya lowlands. The diversification of CT around AD 1 and the subsequent linguistic diversification evident in most Mayan languages at the end of the Late Formative period and through the Early Classic period suggest two conclusions. First, that Maya writing had been invented prior to this diversification sometime in the early Late Formative period (c. 250 BC–AD 1), given the presence of reflexive CT linguistic values reflected in glyphic form throughout the lowlands (Campbell 1984: 9–12; Englehardt 2005: 435–38). Second, that linguistic interaction within the GLM language area decreased at or immediately after the Late Formative–Early Classic period transition c. AD 250.

The iconographic evidence, limited as it is, suggests similar patterns. Despite the fact that little iconographic data is available from the study area, the few objects that are known from southeastern Tabasco attest to a long history of interaction across the region, dating to the Middle Formative period and illustrating ties with Olmec visual culture. Formal, semantic, and contextual similarity and variability in other iconographic motifs examined above intimate a period of extensive interaction and exchange across regions during the Middle and Late Formative periods. Such evidence can only be dated generally and is capable of illustrating broad long–term processes of interaction. On the whole, however, the data illustrate a high degree of formal and semantic continuity in selected iconographic motifs employed across cultural traditions in the Middle and Late Formative periods. Such a lack of contextual variability suggests that interregional interaction continued into the Late Formative period, coeval with the periods of linguistic and material exchange discussed previously. Interaction is suggested insofar as semantic continuity between script systems and iconographic complexes depends on the familiarity of the intended audience with the underlying interpretive framework that gives meaning to the iconographic motif. Although continuity is strongly suggested, this conclusion is somewhat problematic given the paucity of evidence for early scribal systems in Mesoamerica. Beginning in the latter half of the Late Formative period—and earlier in some cases—formal variability in previously shared motifs increases, particularly in the Maya tradition. Such variability increases substantially in the following Early Classic period. Maya

Several examples of the interplay between linguistic values and glyphic representation illustrate problems of 163

Archaeological Paleography iconographers and scribes began to experiment with the formal qualities and semantic meaning of iconographic elements. In general, shared motifs were recontextualized and infused with particular linguistic values specific to (CT) Mayan languages. It does not appear that this process was uniform across the Maya lowlands; certain signs appear to have been recontextualized at different times and to different ends. Additionally, it is unclear whether this process occurred in other traditions at this time,24 again due to the paucity of evidence, particularly for the Epi–Olmec script (Mora–Marín 2001). Finally, the presence of certain glyphic elements in the Olmec registry complicates interpretations by suggesting the possibility that scripts themselves were adopted or transferred between systems. Despite these difficulties, the pattern that emerges from the data suggests a shift in the interpretive framework through which meaning was gleaned from visual markers. Through time, motifs appear decreasingly linked to iconographic contexts and a previous widely shared narrative framework of visual interpretation. Instead, such elements became divorced from iconographic contexts and now depended on linguistic principles of interpretation to adequately deduce a more specific and exact meaning, unique to the Mayan script. In other words, the interpretation of visual elements in a given (textual) composition shifted from purely iconographic to overtly scribal contexts over time. Concurrent with this process—or perhaps because

of it—formal qualities of such recontextualized elements illustrate a far greater degree of variability between scripts, as they are reformulated to express increasingly specific semantic or linguistic values. The absence of shared values or a common framework of interpretation among scripts in the Early Classic period is suggestive of decreasing interaction at this time. Ultimately, it is unclear if the recontextualization of iconographic elements evident in the later Late Formative and Early Classic periods was a cause or an effect of changes in patterns of interaction. Discrepancies in the timing of recontextualization evident in the iconographic data coupled with the general lack of evidence and the inability to effectively ‘quantify’ changes in iconographic data preclude a definitive conclusion to the question of causality. Also obscure are the mechanisms through which such interaction and changes took place. Since I am concerned with a more general correlation in long– term patterns of variability across datasets, however, these lacunae may be overlooked for the time being. Nonetheless, the discussion will return to these issues in the concluding chapter. The following chapter treats the holistic interpretation of the patterns of interaction and innovation that emerged from the material, linguistic, and iconographic analyses in an attempt to discover if parallel patterns of spatial and temporal variability are evident between the datasets.

24 Although it is very likely that such reformulations and recontextualizations occurred in the Zapotec script (Urcid 2001, 2005).

164

Chapter 7 Interpretation and Discussion: The Relationship Between Material Interaction, Innovation, and Script Development This chapter synthesizes the results of the quantitative and comparative analyses to explore potential correlations between interregional interaction, material changes, and script development. Departing from a multi–scalar correlational approach, the discussion attempts to elucidate whether patterns of innovation and interaction evident in the ceramic, linguistic, and iconographic datasets correspond in spatial and temporal contexts. Patterns of interaction and material innovation suggested by these results are explored to determine whether a relationship exists between the data. The implications of the patterns of variability through time in these distinct lines of evidence are considered before turning to a final evaluation of the hypothesis that interregional interaction drove the development of both lowland Maya ceramic traditions and the Classic Maya writing system.

Variability in material, iconographic, and linguistic data should exhibit strong spatial and temporal correspondence if interaction drives both developmental processes. Linguistic and iconographic evidence suggest that Maya writing emerged in the Late Formative period. If a relationship exists between interaction, script development, and material innovation, one would expect to encounter a qualitative and quantitative pattern within the micro–regional analytic unit that indicates a greater degree of interregional interaction and less relative variability between the sample and regional sequences in earlier Middle and Late Formative temporal contexts, followed by a decrease in interaction and increase in material variation in the Early Classic as localized imposition of cultural meaning on icons and material artifacts intensified in the Maya lowlands. If such a pattern exists, it would confirm the centrality of interaction to cultural innovation in this case, and would also suggest a correlation between the dynamics of the developmental processes. The absence of an associative pattern would refute the hypothesis.

If the central hypothesis of this investigation is correct, the results of the comparative and statistical analyses of ceramic materials from the San Pedro Mártir basin should indicate the existence of positive correlations between the appearance of the Maya script and regional interaction during the Formative period in southeastern Mesoamerica. In other words, archaeological indicators of material interaction revealed through qualitative and quantitative analyses of ceramic data should parallel the iconographic and linguistic interaction involved in the development of Maya writing. Quantitatively, the depositional distribution of the material evidence suggests the extent, temporality, and direction of interregional interaction through time. Comparing such continuity and disjunction in the formal characteristics of material artifacts with the distribution and recontextualization of iconographic elements allows for the identification of associative patterns. If correct, I would expect to find temporal and spatial parallels between the patterns of stylistic and distributional variability across datasets. At the boundaries of regional systems—such as the San Pedro Mártir basin—users of shared iconographic complexes challenged symbols’ respective meanings through recontextualization. This process should be reflected through variability in material forms and stylistic characteristics over time. In other words, archaeological evidence should indicate a differential distribution and formal variability in stylistically and functionally similar artifacts that parallels divergence in the localized use of widely–shared, recontextualized iconographic elements. In that case, the emergence of writing and the differential use of material culture would express the same processes of interregional interaction.

Interpreting Variability: A Multi–Scalar Correlational Approach The final phase of the analysis involved correlating diachronic variation in the material data with variability in iconographic and linguistic evidence over time. Following the comparative and quantitative evaluations of the ceramic sample, patterns of diachronic variation indicative of interaction and innovation in the material data were compared against transformations in the form, function, syntax, distribution, and textual recurrence of certain widely–shared Mesoamerican iconographic motifs and linguistic elements. This holistic contextual and distributional interpretation allowed me to correlate and clarify the relationships between the data sets across analytical units and in disparate spatial–temporal contexts (Hodder 1987). An advantage of this conjunctive perspective (cf. Fash and Sharer 1991) is that it enables the identification of those integrative units that assume focal or primary roles, as well as the varying degrees of intensity at which patterns of interaction operate between and within those units over time. Since interaction occurs at a variety of levels in ancient societies within a geographic region, this approach allows for consideration of the dynamic quality of processes of both interaction and integration. Further, this interpretive approach allows for the reconciliation

165

Archaeological Paleography of the changing meanings of symbols with the analytical need to proceed from known symbols and meanings back to unknown ones, and to compare these to natural possibilities along the way (Clark 2004: 215–16). Thus, this interpretive synthesis considers patterning in the data at three different social and geographic scales: regional, micro–regional, and site. Below, the methodologies and lines of evidence employed here are briefly recapitulated, relating these to the central objective of this investigation at these distinct analytic scales.

innovation, which may be deduced from discrepancies in shape classes, techno–stylistic attributes, or site–specific distribution. Likewise, interaction is suggested by synchronic and diachronic homologies observed in the majority of defined variables—homologies reflected in formal and functional characteristics shared within both the sample and regional sequences. Conversely, fewer shared attributes would suggest less interaction and a greater degree of localized innovation across scales. The temporal contexts in which formal and functional changes occurred across analytic units again speak to the direction, intensity, and temporality of interaction on both regional and micro–regional scales. If interaction is correlated with processes of material innovation, intermediate forms in the sample would be expected, that is, artifacts that represent a blend of the stylistic characteristics observed in adjacent areal typological sequences.

In order to demonstrate the existence of a correlation between the developmental processes of both the Mayan script and Maya ceramics in the Early Classic period, it is first necessary to establish that interaction was involved in the dispersal and wide distribution of shared scribal and ceramic traditions in the earlier temporal contexts of the Formative period, followed by decreasing interaction, a greater degree of innovation, and increasing integration into specifically lowland Maya cultural systems in the Early Classic period. Variability in material, iconographic, and linguistic data should exhibit strong spatial and temporal correspondence if interaction drives both developmental processes.

Diversity measures of internal variation within the sample (considering measured analytic variables as a whole) permit a further evaluation of ceramic variability from which interaction and degrees of integration within cultural systems may be inferred. The H score diversity measure is interpreted in essentially the same way as the hierarchical tree plots derived through cluster analysis; lower H scores indicate less variability between sites, which suggests a relatively higher degree of interaction across the micro–region. Comparing H scores in disparate temporal contexts gives a rough measure of how the degree or intensity of interaction varied over time at the regional scale. H scores also permit a crude evaluation of innovation, insofar as greater heterogeneity within a sampled population would suggest the existence of a greater range of material technologies, reflected in increasing variability in types, forms, or stylistic attributes. It is expected that the results of the H score diversity analysis correspond to the patterns of the cluster analysis, that is, relatively lower heterogeneity (i.e., more interaction, less integration) in earlier temporal contexts, followed by increasing H scores in later temporal contexts (indicating a decrease in interaction and increasing integration within the lowland Maya material cultural system). Comparing clusters and H scores over time with the reported frequency or occurrence of particular types, techno–stylistic attributes, or forms in neighboring sequences adds resolution to the patterns of interaction and innovation deduced from the comparative analysis by graphically and quantitatively illustrating the operation of such processes within the sample at the site and micro–regional scales.

The comparative assessment of ceramic materials elucidated patterns of stylistic similarity and variability in the material data over time. On a general level, comparing materials in the sample with diagnostic typological collections of ceramics in adjacent areas illustrates linkages between ceramic assemblages at the regional scale and allows for a preliminary evaluation of distribution and degree of heterogeneity of different ceramic attributes within and between distinct clusters within the study area. Statistical cluster analysis also permits an appreciation of the relative kinds and degree of interaction that occurred at the micro–regional and site levels (i.e., within and between sites located in different clusters throughout the study area). At all scales, interaction can be deduced from patterns of similarity and diversity derived through both statistical and comparative analyses. Evaluating differential cluster patterns of recorded formal variation in the analytic variables over time and at different scales elucidates changing levels of variability between sites and the relationships between the sampled assemblages and materials from adjacent sites, regions, and areal sequences. These patterns suggest conclusions regarding the temporality, extent, intensity, and directionality of interregional interaction reflected in the ceramic materials within the study area. Greater similarity (i.e., less variability, fewer clusters) suggests a relatively greater degree of interaction between the sites or areas that exhibit such correspondences, and less strict integration in localized networks, cultural systems, or material traditions.

In essence, the comparative analysis is most useful at the broad, regional scale, whereas the quantitative measures permit finer appreciation of the patterns as reflected in increasingly smaller scales. Comparative and quantitative similarity in ceramic types, forms, and stylistic attributes indicate interaction between regions, whereas differential

Comparative and cluster analyses also permit an evaluation of questions of functional variability and 166

Interpretation and Discussion sources, functions, or distributions of formal and design attributes suggests innovation and the imposition of locally assigned meanings (cf. Clark and Cheetham 2002; Hodder 1978; Neff et al. 1999; Pring 1977; Stark 1998). As a whole, and due to the nature of the sample, the comparative and quantitative analyses primarily illustrate broad processes of interaction, integration, and innovation in the material data over a long period, and without a great deal of temporal resolution.1

the centrality of regional interaction in the development of Maya writing. Variation in iconographic decoration is expected to correspond to formal and functional variation in ceramics between Late Formative and Early Classic period regional ceramic phases (e.g., Mamom, Chicanel, and Tzakol). Also expected are patterns of variability indicative of a systematic transition towards greater technical complexity, formal innovation, and functional variability over time in the material data, correlating with a greater use and complexity of the emerging Mayan script. Increased variability in and localized adaption of material forms should correspond to formal and functional iconographic and linguistic variations as icons transformed into written symbols and acquired new meaning. Insofar as new cultural meanings are reflected in new material forms, changes in the material data thus parallel iconographic reconfiguration as icons become written signs. Variability suggests innovation in terms of localized adaptations of shared cultural technologies or site–specific singularities. In this sense, stylistic and distributional similarity in the material data may reflects interaction driving the developmental process, whereas discrepancies in form, function, or distribution would suggest localized innovation, paralleling the dynamics of ‘diffusion’ and ‘divergence’ evident in the development of Maya writing.

The comparative and quantitative analyses thus allow for basic inferences regarding the role of regional interaction in the developmental trajectory of lowland ceramic traditions. To determine how and whether this interaction corresponds to the development of Maya writing, observed patterns in the ceramic data were compared with documented iconographic and linguistic evidence. Similar patterns of diachronic variability in linguistic and iconographic data are outlined in terms of changes in the form, syntax, distribution, and textual recurrence of widely–shared iconographic motifs and identifiable signs within discrete iconographic and script systems (cf. Fields 1991; Josserand and Hopkins 1999; Justeson et al. 1985; Lacadena 1995, 2010; Reilly 1996; Schele 1999). As with the ceramic materials, greater degrees of formal, stylistic, and functional similarity in linguistic and iconographic elements suggest a relatively greater degree of interaction. Increasing variability over time and the emergence of innovative formal and functional characteristics of script systems would imply decreasing interaction and greater integration in increasingly localized cultural systems.

In sum, available evidence suggests that Maya writing emerged in the Middle to Late Formative period. If a relationship exists between interaction, script development, and material innovation in this case, patterns of variability in the data should correspond in both spatial and temporal contexts. In terms of ceramic data, variability observed within the micro– regional analytic unit should indicate a greater degree of interregional interaction and less relative variation between the sample and regional sequences in earlier Middle and Late Formative temporal contexts, followed by a decrease in interaction and increase in material variation in the Early Classic period as localized imposition of cultural meaning on icons and material artifacts intensified in the Maya lowlands. Likewise, less variability in linguistic and iconographic data is expected in earlier temporal contexts, suggesting more interaction, followed by increasing variation that implies a greater degree of innovation, decreasing interaction, and increasing integration into more localized cultural systems in the later temporal contexts of the crucial Late Formative–Early Classic period transition, during which the Maya writing system emerged. If such a pattern exists and can be demonstrated in the distinct datasets, it would suggest the centrality of interaction to cultural innovation in these cases, and would also suggest a correlation between the dynamics of the developmental processes. The comparative association of diachronic patterns of variability across datasets and within and between both sites and regions in differing temporal contexts ultimately permits an evaluation of the

Because iconographic and linguistic changes, like ceramic data, can only be generally assigned to specific contexts, these data illustrate broad, long–term patterns of spatial and temporal variability, from which processes of interaction, integration, and innovation may be deduced. Variability in the spatial and temporal patterns derived through the quantitative and comparative analyses of all datasets yields clues regarding the extent and temporal depth of interregional interaction, boundaries of interacting units, the degree of regional integration and group coherence, and localized stylistic and functional innovation. The stylistically defined zones of social interaction derived through the ceramic analyses are compared with overlaps in archaeologically identified regional exchange networks, both of which offer a useful comparative baseline and add a degree of spatial and temporal resolution to the inferences presented here (cf. Josserand and Hopkins 1992; Lacadena and Wichmann 1999). Strong correspondence between patterns in the material, iconographic, textual, and linguistic data would suggest This is also true of iconographic and linguistic evidence, for which spatial and temporal contexts and degrees of similarity or distance cannot be easily established with fine detail.

1

167

Archaeological Paleography central hypothesis of this investigation, that interregional interaction is positively correlated with both material and ideational innovation.

These implications are borne out by the linguistic and iconographic data. The presence of Mixe–Zoque loan words in a variety of indigenous Mesoamerican languages suggests that numerous cultures within the region were interacting extensively with speakers of Mixe–Zoque. Tracing the origins of such loan words linguistically—through glottochronological reconstruction, periods of use, and reflexive morphologies—suggests that MZ linguistic elements began to be absorbed into other Mesoamerican languages in the Middle Formative period. Such linguistic interaction becomes most evident toward the end of this period, but given the conservative nature of linguistic systems and the amount of time that it takes for the effects of linguistic interaction to become apparent through altered syntaxes, morphological shifts, or changes in the grammatical structures of the adopting languages, it is exceedingly likely that such interaction had begun by at least the Early–Middle Formative period horizon, and possibly in even earlier temporal contexts. The linguistic evidence also suggests extensive interaction during the Middle Formative period, the effects of which extended into the Late Formative period.

Synthesis of Data The Middle Formative Period The lines of evidence examined here support the contention that interregional interaction was central to innovation in developing material and scribal traditions in the temporal and spatial contexts of this study. The data suggest that a significant amount of interaction occurred within and across the study area in the Middle and early Late Formative periods. During the Middle Formative period, evidence for rank societies with complex political institutions in the lowland Maya region is limited, and primarily confined to the El Mirador basin of the extreme north Petén (González Moreno 2006: 48–51; Hansen 1991). Prior to this, during the Early Formative period, it is generally held that the majority of the lowlands was a relatively homogeneous society of small villages that lacked status differences or centralized political leadership (cf. Fournier 1987; González Moreno 2006: 49). At some point during the Middle Formative period, interaction between the lowlands Mesoamerican regions began to increase. The clearest indicator of this increase in interregional communication and exchange is the Middle Formative tendency toward homogeneity in lowland ceramic industries that had, at the end of the Early Formative period, begun to demonstrate regional variability.

The spread of MZ linguistic elements throughout Mesoamerica during the Middle Formative period was coeval not only with the trend toward increasing homogeneity in ceramic materials at this time, but also with the increasingly widespread distribution of Middle Formative ‘Olmec’ iconographic and artistic styles across the whole of the region. Despite the relative paucity of evidence, issues of provenience, and problems associated with the definition of ‘Olmec iconography,’ on the whole the iconographic data indicates the widespread distribution of similar iconographic forms and styles that have been traditionally classified as Olmec and associated with that culture. Examples of such elements, and objects that illustrate common widely shared motifs, are even evident within the study area of the San Pedro Mártir basin in southeastern Tabasco (Figure 6.11). These widely distributed iconographic motifs are very similar in formal terms and occur in parallel syntactic contexts, suggesting that they attempt to convey similar semantic messages. The presence of formally similar visual elements in analogous semantic contexts across Middle Formative period iconographic traditions implies that the underlying narrative or interpretive framework that gave meaning to visual signs was also shared between cultures. The totality of the evidence examined therefore indicates a high degree of similarity in ceramics, linguistics, and iconographic elements across a broad swath of Mesoamerica during the Middle Formative period. The absence of significant observed variability, coupled with contextual continuity and strong similarities in the formal aspects of the data, suggests the existence of intense and extended interregional interaction throughout Mesoamerica during the Middle Formative period.

This general trend of marked interregional interaction is reflected in the ceramic data by the near complete absence of stylistic or distributional variability in the Middle Formative assemblages from the San Pedro Mártir region examined here. The sampled materials match assemblages across the lowlands and beyond and are dominated by simple forms of plain red ceramics of the Joventud type. ANOVA and H score analyses revealed little statistically significant variability across any measured ceramic attribute within the sample. A comparison of the sample to sequences at neighboring sites and adjacent regions reveals the absence of meaningful variation in ceramic forms, functions, or distributional contexts across assemblages. The near absence of variation in the formal attributes of Middle Formative Mamom sphere ceramics suggests that modes of ceramic production, decorative techniques, morphology, and treatment were shared widely throughout southeastern Mesoamerica during the Middle Formative period and likely exchanged through wide ranging networks of interaction. Innovation in terms of new techniques, decorative styles, or methods of surface treatment is almost totally lacking. One may thus infer that the observed widespread similarity in Middle Formative ceramics resulted from sustained and intense interregional interaction, as distinct societies produced strikingly homogeneous ceramic materials and were engaged in extensive network exchange. 168

Interpretation and Discussion words, the comparative analysis of ceramic data suggests that interregional interaction decreased in both scale and intensity as the lowland Maya world approached the Formative–Classic period transition c. AD 250.

The Late Formative Period The patterns that emerge from the Late Formative period data are more complicated and present a more complex view of interaction at this time. Statistical analyses of the ceramic materials again revealed a low degree of significant variability within the study area, suggesting that the widespread continuity in ceramic forms and attributes continued into the Late Formative period. Data from other Late Formative period regional sequences support this interpretation, insofar as relatively simple forms of plain red ceramics, now primarily of the Sierra type and forming part of the Chicanel ceramic sphere, continue to predominate in regional assemblages within the Maya lowlands and beyond. These facts suggest that the interaction suggested by the Middle Formative period continued into the Late Formative period. In the latter half of the Late Formative period, after c. AD 1, comparative analysis revealed slightly more variability between regional sequences. Innovative forms and decorative techniques emerged and variation between lowland assemblages increased, particularly in peripheral areas to the east, southeast, and west of the central Petén. These developments suggest increasing regionally specific innovation in ceramic technologies toward the end of the Late Formative period as well as a decline in interregional interaction at this time, although interaction within the Maya lowlands continued. From a network perspective, it appears that spheres of exchange were becoming increasingly more defined, and interaction networks concentrated in relatively more focused spatial scales within the Maya lowlands (cf. Figures 5.69 and 5.70).

Linguistic evidence also appears to reflect these general trends. Interaction between Mixe–Zoque and the majority of other Mesoamerican languages tapers off in the Late Formative period. The morphological shifts in Cholan–Tzotzilan (CT), the increase in episodes of linguistic diversification within the GLM language area, and the increasing influence of the CT language family on other GLM language point to a marked decrease in interregional linguistic interaction during this time, as well as a concurrent spike in linguistic interaction within an increasingly demarcated lowland Maya language area. Reflexive morphological shifts derived from the CT language family, as well as CT loan words in other GLM languages, imply a shift from inter–to intra– regional interaction during the Late Formative period. Although many of these linguistic features are not evident until the latter half of the Late Formative period, the conservative nature of linguistic systems again suggests that these processes began in earlier temporal contexts. The linguistic data complement the material data in suggesting a decrease in interregional interaction and a greater degree of centralization and integration within the greater Maya lowlands. These trends are particularly noticeable at the western periphery of the lowlands in the central Chiapas depression—the traditional ‘homeland’ of CT languages—again pointing to the integral role of peripheral areas in interregional interaction and as potential sources of cultural transformation. The Late Formative period iconographic evidence suggests the operation of similar processes and reflects the same trends. The wide distribution of formally similar iconographic elements in structurally similar contexts continued, suggesting continuity in the processes of interaction that had begun previously. Nonetheless, iconographic evidence also indicates certain innovations in the Late Formative period, specifically in terms of iconographic motifs being employed in new contexts and to convey ostensibly distinct meanings across systems. Also at this time a marked intensification between Mayan, Isthmian, and Izapan/South Coast iconographic traditions becomes apparent, suggesting a re–focusing of interaction concentrated along the western and southeastern borders of the greater Maya lowlands.

Given that these new ceramic forms emerged only after an extended period of interaction and at the boundaries of the lowland Maya world, where interaction between regions was likely most intense, it is reasonable to infer that processes of interaction may be positively correlated with the innovations observed through the comparative analysis of ceramic forms in the latter half of the Late Formative period. The data suggest a relationship between increasing material variability and the development of specialized systems of production, which becomes especially clear approaching the Late Formative–Early Classic period horizon. Such innovations may also reflect emergent integration into a more centralized and defined regional system within the Maya lowlands that sought to exert a greater degree of control over its peripheries (Englehardt 2010; cf. Parkinson 2006). Available archaeological evidence supports the contention of a transition toward a more centralized and integrated regional system within the lowlands at this time. In either case, it appears evident that interaction decreased significantly toward the end of the Late Formative period, or at the very least became more focused and concentrated on intra–regional exchange within the lowland Maya region. In other

The simultaneous continuation of formally similar iconographic elements in parallel semantic and syntactical contexts across traditions suggests that the fundamental interpretive framework that allowed for the transmission of meaning through these motifs continued through the Late Formative period. At the same time, the recontextualization of shared iconographic elements that began at this time also indicates incipient 169

Archaeological Paleography transformations in the underlying visual narrative and interpretive frameworks that substantiated the meaning of iconographic motifs. Innovative new elements, particularly related to regionally specific calendrical systems such as the Long Count or Zapotec year–bearer system, also emerged during the Late Formative period. These developments intimate the establishment of regionally specific interpretive frameworks that were no longer dependent on the shared principles of visual interpretation reflected in the widespread distribution of formally, semantically, and contextually similar motifs. This emergent breakdown in interpretive frameworks— and the establishment of new interpretive principles for visual elements—would appear to suggest a decrease in interregional interaction at this time. This conclusion, however, is at odds with the readily apparent persistence of other formally similar elements in syntactically analogous contexts across traditions that carry the same semantic meanings. What is clear is that the reformulation and recontextualization of shared visual motifs could not have occurred without the prior existence of a common interpretive framework. Although the iconographic data may be interpreted as suggestive of two distinct and contrary patterns of interaction in the Late Formative period, the interregional interaction evident in the Middle Formative period iconographic data appears intimately connected to subsequent innovations in iconographic and scribal traditions across Mesoamerica.

Linguistic data, on the other hand suggest a decrease in extensive interregional interaction sometime around the beginning of the Late Formative period c. 250 BC, although the effects of this decrease did not become evident until closer to the turn of the millennium. Linguistic evidence also points to a re–situation of the loci of intense interaction along the western border of the GLM language area, focused on Cholan–Tzotzilan languages. The influence of the CT language family on other GLM languages in terms of reflexive morphologies and the presence of CT loan words throughout GLM languages also speaks to increasing intra–regional interaction within the Maya lowlands during the Late Formative period, which would potentially suggest a decrease in the degree of interregional interaction at this time. A decline in interaction of all sorts is also suggested by the episodes of linguistic diversification within GLM languages that occurred near the end of the Late Formative period—a trend which continued into the Early Classic period. Nonetheless, linguistic data do indicate that MZ linguistic features continued to affect CT languages throughout the Late Formative period, suggesting an ongoing process of interaction that shifted at different points in time along the western periphery of the Maya lowlands. Finally, available iconographic evidence from this period simultaneously suggests the continuation and decline of long–term patterns of interregional interaction through the Late Formative period. The persistence of formally similar iconographic elements in parallel contexts and with apparently analogous syntactical and semantic meanings across traditions suggests that the interaction that allowed for the sharing of common frameworks of visual interpretation endured at this time. Nevertheless, the emergence of new underlying narratives for the semantic interpretation of iconographic elements, reflected in slight formal variability, new and regionally specific motifs, and the innovative recontextualization of some shared symbols in distinct traditions, suggests that interregional interaction was waning during the Late Formative period. Taken as a whole, the discrete Late Formative period datasets examined in this investigation suggest both continuity and disjunction in patterns of interregional interaction.

As a whole, the Late Formative period datasets suggest multiple interpretations. Consequently, the patterns of interaction and innovation that may be deduced from the different lines of evidence are decidedly more complex. Continuity in the formal attributes and stylistic characteristics of Late Formative period ceramic evidence within and between regions suggests the continuation of the intense interregional interaction of the Middle Formative period through at least the latter half of the Late Formative period, at which point certain innovations become increasingly more evident. The absence of significant within–sample statistical variability in the ceramic materials, coupled with comparatively relative homogeneity across ceramic assemblages in the lowlands and in adjacent regions, suggests that interregional interaction remained intense during the majority of the Late Formative period. The widespread similarity in ceramic types and forms throughout the Maya lowlands and beyond during the Late Formative period implies a great degree of interaction over a large swathe of southeastern Mesoamerica. On the other hand, the divergence in the ceramic sequences of the Middle Grijalva region (Lee 1972) occurred earlier in the Late Formative period, suggesting that extensive interregional interaction may have begun to decrease around the turn of the millennium c. AD 1. This fact could indicate that changes in these broad, long–term patterns occurred in stages and were thus reflected at different points in time within discrete spatial contexts.

Thus, during the later part of the Late Formative period, and certainly in the ‘Protoclassic’ at the onset of the Early Classic period c. AD 150–250, interregional interaction decreased significantly. Archaeological evidence suggests that this development is contemporaneous with a marked emergence of more pronounced status differentiation and increasingly complex political organization throughout the lowlands, more defined settlement hierarchies, and a tighter control over sociocultural and political boundaries. Certain distinctive characteristics of Classic Maya society also appear at this time, such as polychrome ceramics, 170

Interpretation and Discussion carved stelae and altars, roads and infrastructure, and a relatively greater amount of monumental architecture. At this point, lowland Maya ceramic production may have become a specialized activity, in such a way that an incipient lowland Maya style developed that was distinctive from contemporary traditions elsewhere in Mesoamerica. That is, the homogeneity in ceramic design and production that characterized the Middle and early Late Formative periods was replaced by regional variability and increasing heterogeneity in ceramic materials, both across regions and within the lowlands. A positive correlation between material variability and emergent systems of specialized production is thus again suggested.

Early Classic period frontier of the Maya lowlands. Sites on or near this boundary, such as San Claudio and Piedras Negras, generally imitated core trends, and the incipient localized Early Classic styles noted at these sites, themselves heavily influenced by central Petén trends, speak at once to both their increased integration in a larger, more centralized sociocultural system and their relative isolation and situation at a peripheral area. Moreover, the clinal distribution and temporal lag in ‘down–the–line’ emergence of Petén–based traits (e.g., the relatively late appearance of basal–flanged bowls at Altar de Sacrificios, and, subsequently, Piedras Negras) indicate an emphasis on stricter control and restricted flow of diffused technologies from the core to the periphery as Maya society became increasingly complex, centralized and inward–focused.2 The data presented here suggest that the development of sociopolitical and cultural boundaries in southeastern lowland Mesoamerica is a multifaceted process. As a whole, the evidence intimates that the formation and maintenance of boundaries between the systems of networked interaction that traditionally characterize the Classic period are far more complex than many current understandings acknowledge.

The Early Classic Period In general, the observed and statistically significant formal and stylistic variability between the Early Classic period ceramic materials from the San Pedro Mártir basin closely parallels the traits identified by Holley (1987: 188–89) that distinguish the pottery of Piedras Negras and Altar de Sacrificios from the northwestern Maya lowlands as a whole. These include the relative thickness of vessel walls, jar morphology, basin elaboration, and surface treatment. The Early Classic ceramic sequence at San Claudio appears more closely related to Tzakol sphere assemblages at Piedras Negras and Altar, whereas materials from the remaining sampled sites seem to be associated with a developing northwestern tradition evident at Palenque and the lower Usumacinta region. This rupture of continuity in stylistic attributes within the sample intimates the formation of a socio–cultural boundary across the San Pedro Mártir basin on or near the Late Formative–Early Classic period transition (Englehardt 2010). Innovative ceramic forms, morphological attributes, and decorative techniques emerged at this time as well.

The decrease in interregional interaction suggested by increasing specialization and variability in techniques of Early Classic period ceramic production is mirrored in the linguistic data. Shortly after the Late Formative–Early Classic period transi tion, multiple instances of linguistic diversification within the GLM language area occurred (Figure 6.5). Although in most cases the observed morphological shifts and lexical transformations were predicated on previous influence from MZ and CT languages, these episodes of diversification suggest a decline in interaction, both on the interregional scale as well as within the Maya lowlands. Lexical and grammatical diffusion between GLM languages became greatly restricted at this time, essentially coming to a standstill after the Early Classic period (Justeson et al. 1985: 67). Particularly telling

In contrast to the broad homogeneity in ceramics noted in the Middle and Late Formative periods, after the Late Formative–Early Classic period transi tion a regional separation of sequences and a greater degree of variability both within the sample and in relation to ceramic assemblages at adjacent sites and regions indicates much less interaction within and across the region, as well as increasing levels of innovation in ceramic production styles and techniques. The corre spondence between ceramic materials from San Claudio and those at neighboring lowland regional centers at Piedras Negras and Altar de Sacrificios, coupled with the lack of parallels between the assemblages from Mirador, Revancha, and Tiradero and the central Petén, as well as the formal and stylistic similarities of those assemblages with materials encountered to the north and west, suggests that the intermediate plains of the mid– lower Usumacinta and lower San Pedro Mártir basins north of the low Sierra del Lacandón foothills was the

These Early Classic period developments correspond to what also ap pears to be increased control of the western, southern, and southeastern peripheries (Englehardt 2010). Increasing boundary demarcation is reflected in the complete divergence of Ipsan and Juspano phase ceramics in the middle Grijalva basin (Lee 1972: 13–14), as well as the lack of formal or stylistic similarity between the Tzakol sphere and post–Usulutan ceramics on the southeastern periphery, as reflected in the materials evident in the post–Arenal, Santa Clara and Aurora phases at Kaminaljuyú in the highlands to the south. In the northwest, the boundary did not remain static. Tzakol 1 types eventually found their way to Palenque, although not until the later Early Classic Motiepa and Cascada phases and in limited quantities. At this point, a cursory examination of the Late Classic ceramic types evident in the San Pedro Mártir basin reveals that the common type–varieties once again became as uniform as they had been in the Middle and Late Formative, with assemblages dominated by Tinaja Red, Cambio Unslipped, Encanto Striated and Fine Orange ware, the last itself notably absent at the new periphery of Palenque (González Moreno 2006; Hernández Ayala 1981; Holley 1987; Ochoa and Casasola 1991; Rands 1972, 1987, 2007a).

2

171

Archaeological Paleography is the split between Tzeltal and Tzotzil within the CT family, as well as the diversification of Chujean, Kanjobalan, and Mam–Tectiteco. These languages were, and continue to be, spoken in the traditionally peripheral zones between highland and lowland Chiapas and the highland and lowland areas of Guatemala. These episodes of linguistic diversification again suggest a new emphasis on more clearly demarcating the boundaries of the lowland Maya world. In addition, increased linguistic GLM diversification intimates the establishment of more centralized sociopolitical systems within the lowlands during the Classic period, marked by less porous boundaries and a significantly diminished amount of interregional interaction (Josserand and Hopkins 1999; Lacadena and Wichmann 1999). Further episodes of diversification are noted after AD 500—well into the Classic period—but again, given the temporal lag between the reception of a particular stimulus and its reflection in linguistic changes, it is likely that such instances reflect a decrease in interaction and consequent increase in linguistic specificity that had begun in earlier temporal contexts. The spike in Classic period diversification likely resulted from the decrease in interaction marked temporally by the morphological shifts that occurred in CT languages c. AD 50 (see Mora–Marín 2001). In any case, linguistic data from the Early Classic period also suggests a dramatic decline in the amount of both inter– and intra–regional interaction within the GLM language area and between the Maya lowlands and other regions of Mesoamerica.

Classic period and beyond. Nonetheless, even in these cases it is clear that the iconographic motif was assigned a new linguistic value specific to a Mayan language. The previously shared framework of interpretation and narrative that was formerly employed to deduce meaning from visual symbols therefore had been replaced by this time. Like the Early Classic period transformations suggested by the linguistic data, the reformulation and recontextualization of iconographic elements at this time appears to be a continuation of similar processes first noted in the Late Formative period. The iconographic data are more difficult to interpret, precisely because certain elements and motifs—such as the Lazy–S— maintained similar formal characteristics, syntactical contexts, semantic meanings, and continued to rely on primarily visual interpretive frameworks through the whole of the Classic period and in some cases surviving into the Postclassic period. Ultimately, it is unclear exactly when those elements that did become divorced from iconographic contexts underwent this transition.3 That is, the data may be interpreted to suggest that such processes occurred in either the Late Formative or Early Classic period, or that the process of recontextualization was spread over many centuries spanning the Formative–Classic period transition. In either case, the iconographic evidence does suggest a diachronic decline in interregional interaction, and a concurrent increase in innovation as related to visual motifs. In addition, during the Early Classic period there are many more examples of novel visual forms, semantic contexts, and reformulations of signs to reflect distinct and specific linguistic values than are evident in the Late Formative period. This conclusion presents difficulties when attempting to correlate interregional interaction as reflected in ceramic materials with the development of the Mayan script.

The Early Classic period iconographic data suggest a decrease in interaction and concurrent increase in innovative reformulations and recontextualizations of shared iconographic elements and motifs. Formal similarities between visual elements in the Maya tradition and those present in other iconographic and scribal systems decline. Likewise, particularly in the Mayan script, previously shared elements that were once formally similar began to occur in new syntactic contexts, suggesting a rupture with the underlying system of visual narrative that had once sustained the meaning of shared elements—even as these motifs began to appear in increasingly textual contexts during the Late Formative period Mora–Marín 1997, 2000, 2001, 2005). The novel reformulations and recontextualizations of shared iconographic motifs evident in the Maya lowlands do not appear to have been shared with or intelligible to other visual or scribal traditions. At the same time, traditions in other parts of Mesoamerica (e.g., the Zapotec script, Teotihuacano iconography and writing) had developed similarly novel reformulations of shared motifs and conventions. The fact that such reformulations were not shared between traditions suggests that interregional interaction had declined by the dawn of the Early Classic period. Nevertheless, some iconographic elements (e.g., the Lazy–S motif) evidence formal and semantic continuity well into the

Summary As a whole, the data suggest a gradual decline in levels of interregional interaction from the Middle Formative through the Early Classic period. Concurrent with this decline is an increase in innovation, whether in terms of ceramic production technologies or systems of visual communication. As inter– and intra–regional variability increased across all datasets, formal similarities decreased while at the same time formal complexity increased. That is, not only are the ceramics or iconographic elements of the Early Classic period stylistically and functionally distinct from those in the Middle Formative period, they are also considerably more elaborate, indicative of elevated levels of technological sophistication and the introduction and development of new and increasingly This interpretive difficulty is compounded by the scarcity of early Maya texts, as well as the general paucity of early texts in Mesoamerican from any scribal tradition (cf. Carrasco and Englehardt 2015).

3

172

Interpretation and Discussion advanced production techniques. This fact is most evident in the ceramic data, although the increasing syntactic and semantic complexity of the iconographic data through time also speaks to this point.

survived, one could conclude that interaction must have continued. One could just as easily conclude that, as with linguistic evidence, the iconographic data reflect a temporal lag between a decline in interaction and the subsequent appearance of the effects of that decline. In the Early Classic period, the substantial increase in formal variability between visual elements in distinct iconographic and scribal systems suggests a decline in levels of interregional interaction. Yet some elements continue, formally unmodified, to function in similar formal, semantic, and contextual interpretive frameworks. It is unclear whether this continuity suggests the persistence of interaction or simply the survival of ancestral paradigms of visual narrative into the Classic period and beyond. The marked rise in formal and contextual variability within and between iconographic and script systems during the Early Classic period, in contrast, suggests an elevated level of innovation in systems of visual representation at this time.

The data do not present a unified picture or suggest any clear conclusions regarding the precise temporal contexts in which processes of interaction and innovation waxed and waned. Indeed, although the evidence does broadly suggest a general decline in interregional interaction and concomitant increase in innovation over time, in many ways the data present three distinct interpretations. Further, certain datasets are better suited to illustrate different processes (i.e., interaction vs. innovation). The statistical analyses of the ceramic data suggest clear cut divisions, with variability remaining relatively constant through the Middle and Late Formative periods before jumping sharply and abruptly in the Early Classic period. The comparative analysis of ceramic materials belies this interpretation in suggesting a more gradual trend toward experimentation and variability in the latter half of the Late Formative period, particularly at the peripheries of the Maya world.

Since this investigation is focused on broad, long–term trends, certain variable interpretations of the data is to be expected. Although the patterns of interaction and innovation illustrated by the data remain muddled, in conjunction the distinct lines of evidence present a more complete image together than any one of them could individually. All three datasets suggest a great deal of interaction in the Middle Formative period, reflected in formal and contextual homogeneity and a marked lack of variability in material and iconographic evidence. All three suggest a decline in interaction during the Early Classic period, during which variability within and between regional ceramic, linguistic, and iconographic traditions increases substantially. Thus all three lines of evidence suggest a diachronic decrease in levels of interregional interaction in Mesoamerican from the Middle Formative through the Classic period. Concurrent with this trend of diminishing interaction, the ceramic and iconographic evidence suggest a diachronic rise in innovation, observable through increasing formal, functional, and contextual complexity over time. Finally, all three datasets point to the 500 years of the Late Formative period (250 BC–AD 250) as a time of significant change for lowland southeastern Mesoamerica, during which interregional interaction began to decline and localized innovations in cultural technologies came to the fore.

In contrast, the linguistic evidence suggests that interaction, especially between the MZ and CT language families, increased closer to the beginning of the Late Formative period and had certainly diminished significantly by the time of the diversification of CT that occurred c. AD 100 at the very latest. On the other hand, linguistic data reveal little regarding questions of innovation—particularly material innovation—as the morphological, grammatical, and lexical transformations observed in GLM languages through the Late Formative and Early Classic periods stem logically from interaction that occurred in earlier temporal contexts, first with MZ languages and then with the CT language family. Interpretation of the linguistic data is further complicated by the temporal lag between changes in patterns of interaction and the resulting reflection of those changes through morphological shifts, lexical diffusion, or grammatical transformations. The iconographic evidence, in turn, supports multiple interpretations. The Middle Formative period data imply sustained and intense interregional interaction. In the Late Formative period, iconographic evidence suggests decreasing interaction, in the form of recontextualizations of shared elements in new and overtly scribal contexts, while at the same time continued formal similarity in shared elements suggests that interaction continued. Further, many shared iconographic motifs, illustrating little formal variability, continued to function firmly within the context of preexisting canons of visual interpretation, suggesting the survival of the common organizing framework that contextualized the semantic meaning of those elements. Since these underlying narrative principles of visual interpretation

Previous investigations have reached similar conclusions (e.g., Fahsen Ortega 1999; Grube 1999). The data at hand do not permit a clarification of the precise temporal contexts—or a specific point in time—at which these processual patterns shifted. The three datasets suggest that patterns of specifically interregional interaction shifted at slightly different times within the Late Formative period: the linguistic evidence pointing to an earlier decline, the ceramic materials suggesting a shift towards the end of the period, and the iconographic 173

Archaeological Paleography data sketching a general long–term shift throughout the Late Formative period. Nonetheless, these data are indicative of broad, long–term trends. It is possible that there is no one temporal context or ‘point of singularity’ at which patterns suddenly changed and are subsequently immediately reflected in the material evidence. Instead, the evidence points to a longer span of gradual change and a steady decline in interaction over time rather than a punctuated equilibrium characterized by short bursts of innovation and relatively rapid shifts in the patterns of cultural processes during the Formative period. This suggestion, however, is at odds with current understandings of the processes of linguistic diversification (Dixon, 1997). If one considers writing as a tool of emerging hierarchies, a possible conclusion that may be drawn is that a decrease in interaction and the development of boundaries between traditions of elite material culture occurred earlier in time, concurrent with the trend toward increasing sociopolitical complexity and emergent hierarchical organization noted in the lowlands in the Middle and early Late Formative period at large regional centers such as El Mirador. Further study is required to clarify this potential relationship.

southeastern Tabasco suggests that the emergence of sociopolitical boundaries between lowland Maya networks in the Early Classic period was a complicated and variable process. Correlations with the Developing Mayan Script The earliest known exemplar of the Mayan script is found on the murals at San Bartolo in the northern Petén, dated to c. 350–250 BCE (Saturno et al. 2005; Saturno et al. 2006; Taube et al. 2010). Many other early monuments contain inscriptions from the Late Formative period—in the central Petén as well as in peripheral areas—written in an ostensibly Mayan language (see Mora–Marín 2000, 2001:359–60, table 1.5), suggesting that the Mayan script emerged early in the Late Formative period. Indeed, as Justeson (2012: 833) notes, ‘all of the earliest definitively Mayan texts—conforming throughout to Mayan grammar, phonology, and vocabulary—date to the Late Formative period.’ Further, the readings of T25 and T203 as ka indicate clearly that the development of the Mayan script had occurred prior to the Cholan * k to č shift, and therefore prior to the diversification of the CT family in the Late Formative period (see Figure 6.5; cf. Campbell 1984: 12–3; Justeson et al. 1985; Justeson 2012: 834, fig. 63.3a). The scarcity of early Maya texts complicates interpretation and precludes definitive conclusions, but the available evidence suggests that by the time the San Bartolo text was created, Mayan scribes had already begun to juxtapose linguistic information on visual icons. A reformulation of the underlying framework that structured and allowed for the interpretation of the semantic meaning of formally similar visual elements had to have begun by this time, although recontextualized elements continued to be linked to their ‘original’ iconic meaning. Further, the processes by which linguistic codification occurred (cf. Carrasco and Englehardt 2015; Justeson 1985, 2012) depended on exploiting the potentially multiple meanings of a given sign within iconographic contexts, which is likely why so many signs in the Mayan script exhibit enduring links to the iconicity of visual perception and interpretation. As noted above, in many scripts—within Mesoamerica and elsewhere—icons may continue to function in semantically identical contexts within a writing system.

All three lines of evidence suggest the centrality of interregional interaction to cultural innovation. The changes observed in linguistics, material culture, and iconography in the Late Formative and Early Classic periods in southeastern Mesoamerica occurred only after a long period of interaction with other regions. Innovative transformations and shifts in the data also appear to stem from widely shared traditions that were themselves fostered by sustained and intense interregional interaction in previous temporal contexts. Thus, the three distinct datasets, taken separately, indicate the importance of interregional interaction to cultural innovation and change in these contexts. They all suggest a similar long–term trend of gradually declining interaction and an apparently correlated increase in cultural innovation beginning generally during the Late Formative period. The various lines of evidence reflect these processes at different points in time. As a result, the relationship among the data—and the results they suggest—is ultimately unclear. Nevertheless, what the evidence does suggest is equally intriguing. As a whole the data indicate that large–scale changes in lowland Maya culture were occurring in temporal contexts earlier within the Middle and early Late Formative periods than has previously been considered. Although all data suggest that transformations in material culture, linguistics, and iconography were linked to interregional interaction, the evidence indicates that changes in long–term processes of interaction occurred in stages and were evident at different points in time in discrete spatial contexts. The data also demonstrate a striking relationship between material variability and emergent systems of specialized production in the lowlands—another potentially fruitful topic of future investigation. Finally, the evidence from

The very processes that facilitated the linguistic codification of iconographic elements were predicated on the widespread sharing and mutual intelligibility of visual narratives and interpretive frameworks. In order for icons to have existed in the multiple contexts necessary to encourage their recontextualization, they first must have been widely distributed among distinct cultural and linguistic groups. Thus, interaction must have occurred prior to the onset of the Late Formative period to account for the widespread distribution of the shared iconographic elements that were subsequently 174

Interpretation and Discussion recontextualized. The ceramic and iconographic evidence examined herein support this conclusion, in suggesting extensive interregional interaction throughout the Middle and Late Formative periods. In addition, many of the earliest examples of Maya writing found on portable objects were inscribed on reused Middle Formative preciosities (see, e.g., Mora– Marín 2001). Such examples appear to link early forms of the Mayan script with objects that were themselves involved in processes of interregional exchange (Mora– Marín 1997). The extensive interaction noted in the Middle Formative period at once accounts for the broad distribution of material goods and iconographic motifs across multiple cultural groups and suggests that such objects and motifs were ripe for reinterpretation when they were encountered in discrete cultural contexts.

not have been enclosed within particular interpretive frameworks at this time (e.g., they were not ‘fully Mayan’), it is likely that a great degree of interaction was still occurring, and that developmental experimentation was continuing. That these scripts both emerged after a period of extensive linguistic and iconographic interaction suggests that speakers of differing languages were experimenting with the juxtaposition of linguistic elements on a shared iconographic repertoire by at least the middle of the Late Formative period. As Lacadena (2010) has demonstrated, Maya writing appears to have been written originally in a script whose syllabograms were developed in or modeled after a Mixe-Zoquean syllabary. The semantic confusion that potentially stemmed from the adoption (or transliteration) of an MZ word in a CT language to describe a visual element common to both iconographies would have created the preconditions that allowed for the communication of increasingly specific linguistic and grammatical information in graphic form to more precisely clarify intended semantic or phonetic meaning.

The linguistic evidence, in terms of extensive Mixe– Zoque lexical diffusion into other Mesoamerican languages observed in the Formative period, also supports the contention of sustained and intense interregional interaction immediately prior to the emergence of the Mayan script. The fact that the Maya writing system represents the Cholan language—despite the fact that the Mayan script may be read in Yukatek— suggests that the script developed during a time in which CT languages were exerting particular influence over the GLM language area. The linguistic evidence indicates that the Late Formative period was just such a time. Linguistic interaction became more confined to the western periphery of the southeastern lowlands between MZ and CT language groups, evident through morphological shifts and reformulations of MZ words that were subsequently interpreted in CT languages and reflected in glyphic form (e.g., MZ *ʔameʔ > CT/ GLM haab’ č, * m > b’) may not have occurred until the latter half of the Late Formative period, and thus reformulations of existing glyphic notations to reflect these shifts should not be expected until the Early Classic period. It has been suggested that this is particularly true in the case the glyphic depiction of glottalized consonants (Justeson et al. 1985; Justeson and Campbell 1997; Justeson and Matthews 1990). If, as Lacadena (2010) suggests, glottal phonemes did not exist in other languages that employed a common, ancestral iconographic ‘script,’ the Maya writing system would have had to invent a way to express them (Figures 6.2, 6.6). As evident in other writing systems, existing signs are often simply

Thus, although the iconographic data simultaneously indicate that some motifs were extensively reworked while others retained their essential iconicity during and after the Late Formative period. This observation carries several implications for generalized models of the development of writing systems in Mesoamerica and beyond. The data suggest that the reformulation of iconographic elements did not manifest until the Early Classic. It is in this period that extensive modifications to previously shared iconographic motifs are most evident. This fact suggests that iconographic interaction, or at least the underlying interpretive framework that sustained the semantic meaning of visual narratives, survived through the Late Formative period. It is possible that there existed some time lag between the end of interaction between distinct cultural groups employing a common iconographic system and the subsequent appearance of formal variation in regional iconographies. Nonetheless, the iconographic data support no clear conclusions regarding the specific timing of linguistic codification and the emergence of writing in southeastern Mesoamerica, although there does exist evidence for 176

Interpretation and Discussion manipulated to reflect such changed phonetic values. The glyphic reformulation of certain signs (e.g., mu > b’u) to reflect these values does not become evident until the latter half of the Late Formative period. Further, early Mayan texts such as the Hauberg Stela appear to confuse phonetic values, or at the least fail to explicitly specify in glyphic form the ‘correct’ expected value.5 It is possible that early Maya scribes simply accepted prior, possibly MZ, linguistic values that were previously attached to certain motifs, knowing that the semantic content would remain clear since the complete divorce of scripts from visual interpretive frameworks had yet to occur. In the case of the Mayan script, this process of dissociation was clearly well underway early in the Late Formative period, and must have occurred prior to the Formative–Classic period transition, considering that the Hauberg Stela is organized along a more purely textual framework, suggesting that visual interpretive frameworks had already been subsumed by the time of its creation c. AD 197.

The above points allow for a return to a previously alluded possibility, namely that the Maya and other cultural groups in Mesoamerica ‘inherited’ an ancestral script or previously established system of visual communication that was already segmented and independent of pictorial interpretive frameworks. Many scholars (see, e.g., Justeson 2012: 838; Lacadena 2010) have suggested that writing itself was invented only once in Mesoamerica, by the Olmec, and that it subsequently spread to other cultures through interaction with the Olmec in the Early and Middle Formative periods. The continuity of ‘Olmec’ iconography through time and across space in Mesoamerica—and the sharing of certain conventions and even signs—supports this contention, especially those visual elements found in an textual format that illustrate the incipient transcendence of pictographic narratives (e.g., as on the Cascajal Block or the Tlatenco and Humboldt Celts; Carrasco and Englehardt 2015; Mora–Marín 2010). It is also certain that Maya scribes were not the only group in Mesoamerica to experiment with the reformulation and recontextualization of shared iconographic elements to produce a writing system. Indeed, it is even likely that interaction between the Zapotec and other cultural groups engaged either in writing or in the representational practices that led to writing was intimately involved in the emergence of later Zapotec writing (Flannery 1968a; Justeson 2012: 835; Justeson et al. 1985; Urcid 2001, 2005). Even if the Maya, like other cultural groups, inherited a previously developed script that had already begun to recontextualize visual elements, processes of interaction, reformulation, and innovation would still be evident. Moreover, such processes would have transpired in the same temporal contexts outlined above. Interaction would necessarily have occurred in order for the script to be transferred between cultures. Observed reformulations and innovations would also have been necessary as the ancestral script was adapted to reflect the linguistic values present in the languages of those groups that adopted it. Whether the Maya independently developed their script or inherited it from some other group is a secondary concern. The material, linguistic, and iconographic data examined herein would nonetheless illustrate the same patterns related to its development in the same temporal contexts.

It is just such confusion between expected and represented value that would have necessitated the graphic reformulation of iconographic elements in order to more precisely convey intended meaning within a linguistic framework. It is possible that early instances of linguistic codification in the Mayan script took place as specific phonetic values were juxtaposed onto a preexisting system that had already begun to transcend pictographic frameworks. Nonetheless, extensive material and iconographic interaction must have existed previously in order for the dispersal of the base script or iconographic complex to be shared—and mutually intelligible—between disparate cultural–linguistic groups. At the same time, linguistic interaction must have continued and intensified as well, potentially leading to confusion between speakers of different languages regarding the actual or intended value of a visual element. A crisis of meaning would have occurred regardless of whether elements were encountered in a textual or iconographic format, and subsequently encouraged the formal changes that reflected increasing specificity in the Early Classic period iconographic evidence. Alternately, intensive linguistic interaction would have occurred in prior temporal contexts if early Maya scribes simply accepted previously established values and failed to distinguish discrete phonetic aspects in graphic form. Such distinctions, and the glyphic reformulations that reflect them, become evident in the later contexts of the Early Classic period. Thus, despite the absence of formal variability in iconographic and glyphic representations, linguistic variability and the lack of high degrees of linguistic specificity suggest that interaction remained intense during the Late Formative period.

When considered in relation to the development of the Mayan script, the lines of evidence explored in this investigation suggest that the development of Maya writing involved a gradual shift from an iconic to a textual framework. Linguistic and iconographic variations suggest that this shift occurred over the whole of the Late Formative period as grammatical principles replaced pictorial conventions as the primary organizing principles of visual compositions. At the same time, a comparative analysis of the material data suggests that levels of interregional interaction were slowly decreasing during these temporal contexts as regional

Although it is likely that readers would still have been able to discern the intended meaning, despite the phonetic confusion, from contextual information or familiarity with semantic conventions.

5

177

Archaeological Paleography ceramic traditions emerged, particularly in the latter half of the Late Formative period.6 Thus, when considered in relation to the processes involved in the emergence of Maya writing, the data as a whole tend to suggest that shifting patterns of interaction are related to script development in a broad sense.

rather than formal aspects, were the defining innovation involved in the emergence of the Mayan script. The iconographic data suggest that such contextual transformations, in the form of a shift away from visual frameworks as the primary textual organizing principles, occurred at an exceptionally early date, prior to the onset of the Late Formative period itself.

If the processes are in fact correlated, patterns of variability and interaction evident in the ceramic data should correspond spatially and temporally to documented iconographic and linguistic variation. Despite strongly suggesting a positive correlation between patterns of shifting interaction and innovation, ultimately the relationships between the datasets remain obscure. Each individual line of evidence presents a slightly different pattern of interaction and suggests that changes in these patterns occurred in distinct temporal contexts within the Late Formative period. Ceramic data suggest a decline in extensive interregional interaction toward the end of the Late Formative period, although this decline is only evident through a general comparative analysis with regional sequences, as the statistical analyses reveal no significant variation within the sample through the entirety of the period. Further, variability in regional ceramic materials is noted even earlier, c. AD 1, given the divergence of the Middle Grijalva sequence at this time. Linguistic evidence in the context of script development indicates intensification in a more focused interaction between MZ and CT languages along the western border of the Maya lowlands at the outset of the Late Formative period. At the same time, general morphological shifts and lexical variation noticed within GLM languages at and following the diversification of the CT family in the later part of the Late Formative period suggest that extensive interaction had waned at this time—and likely in earlier contexts as well. The iconographic data, in turn, offer no clear pattern. The lack of formal variability across scripts and iconographic systems during the Late Formative period, as well as the observed continuity of certain visual motifs within purely iconographic contexts, suggests sustained interaction. On the other hand, the obvious recontextualization of other visual elements within new textual frameworks—a process that occurred differently in distinct scribal or iconographic traditions—suggests that interaction was declining as the transition from iconography to writing was underway in the lowlands at this time. Most significantly, such recontextualizations of iconographic elements during the Late Formative period suggest that changes in context,

Evaluation and Implications In this investigation, it is suggested that interregional interaction is central to cultural innovation in Mesoamerica. If this is the case, then the emergence of cultural technologies such as writing should be correlated to and traceable through the development of other material culture, such as ceramics. If interregional interaction is indeed at the heart of cultural innovation, then correlations between the distinct datasets should be evident. That is, the data should reflect the same processes through correlated patterns in similar spatial and temporal contexts. The central hypothesis of this study can be broken down into three separate yet related propositions: 1. That interaction plays a critical role in episodes of cultural innovation; 2. That the development of the Mayan script can be traced through material remains; 3. That specific correlations between the distinct datasets employed herein will be evident. If these hypotheses are correct, patterns of exchange in the material evidence should spatially and temporally correspond to iconographic and linguistic data indicative of broad, long–term patterns of interaction and innovation. The data, both individually and as a whole, do support the first proposition. The ceramic, linguistic, and iconographic transformations observed throughout the Maya lowlands in the Late Formative and Early Classic periods occurred only after a long period of sustained and extensive interregional interaction. The almost complete lack of variability in all datasets during the Middle Formative period suggests a high degree of interaction and relatively little innovation across traditions at this time. I expected to encounter a qualitative and quantitative pattern within the micro–regional analytic unit that indicates a greater degree of interregional interaction and less relative variability between the sample and regional sequences in earlier Middle and Late Formative temporal contexts, followed by a decrease in interaction and increase in material variation in the Early Classic as localized imposition of cultural meaning on icons and material artifacts intensified in the Maya lowlands. This is essentially the pattern reflected in the data, although the nature of the evidence does not permit a fine–grained resolution regarding the precise temporal contexts in

There is some evidence of regionally specific innovation in the earlier Late Formative period, particularly along the southeastern periphery of the lowlands (e.g., at Kaminaljuyú). In highland Guatemala, the divergence of the Providencia and Miraflores ceramic spheres from the Chicanel sphere may indicate a developing lacuna in highland–lowland interaction at this time. The emergence of the Izapan or South Coast scribal tradition during the Late Formative period may thus illustrate the same processes of slowly decreasing interregional interaction and a concurrent rise in cultural innovation in slightly earlier temporal contexts.

6

178

Interpretation and Discussion which patterns shifted. Rather, the data illustrate a general decline in interaction and an increase in variability throughout the Late Formative period, followed by a significant increase in variability across the datasets in the Early Classic period. Nonetheless, such a pattern does confirm the centrality of interaction to cultural innovation in this case. In addition, the innovations noted in the Late Formative and Early Classic period took the form of experimentation with and modifications to existing ceramic and iconographic traditions. In the contexts of this investigation the evidence supports the contention that interregional interaction is central to cultural innovation.

indicate recontextualization of shared iconographic elements at an exceptionally early date—near the Middle–Late Formative period transition—is significant. Not only does this fact suggest that considerable cultural transformations occurred in temporal contexts earlier than has previously been posited, but also that changes in the contextual frameworks of use and exchange were paramount in understanding the effects of shifting patterns of interaction on both material and symbolic culture in the southeastern lowlands. The graphic representation of linguistic elements does appear to indicate a degree of intermediate variability in the Late Formative period. The multiple depictions of, for example, conjugational verb endings in early texts support the idea of such a period of intermediacy and scribal experimentation in which interaction was generally declining but still evident. So too do the discrepancies in the glyphic representation of certain syllabic formations (e.g., mu vs. b’u) or glottalized consonants. This process is contextualized by the standardization in glyphic representation of grammatical elements in the later temporal contexts of the Early Classic period. Nevertheless, due to the flexibility of the Maya writing system, a degree of variability in the graphic depiction of linguistic principles remained, complicating this interpretation. In addition, there is still considerable debate regarding the reading of early texts in the lowlands. The conclusions that may be drawn from this data are therefore tentative.

Intermediate material and iconographic forms that date to the Late Formative period, the time frame in which interaction was decreasing, were also expected to be observed. The sampled ceramic materials revealed no such intermediate forms, although the relatively unique vessel shapes noted at San Claudio—and Piedras Negras—may indicate an emergent localized style that represents just such intermediate forms. Alternately, such variation could also simply indicate the relative isolation of these sites from ceramic trends in the central Petén. Similarly, the down–the–line lag in the appearance of common Early Classic forms such as basal flanged bowls at Piedras Negras and Altar de Sacrificios speaks to the peripheral location of the sites under investigation. Nonetheless, these ceramic phenomena occur in the slightly later contexts of the Protoclassic or very early within the Early Classic period, and the relative scarcity of such unique vessel forms within the sample precludes definitive conclusions. The data are inconclusive, and even a tenuous interpretation would suggest that intermediate forms appear in the Early Classic period— later than I would have expected.

In a general sense the data support the proposition that interregional interaction was central to cultural innovation. The evidence suggests that innovative transformations in the ceramic, linguistic, and iconographic datasets examined here occurred after a long period of sustained and intense interaction between groups in distinct regions of Mesoamerica. Patterns of shifting variability in the data suggest that such changes were linked to interaction, insofar as the reformulation and recontextualization of ceramic materials, linguistic elements, and iconographic motifs stemmed from their previously widespread distribution in multiple contexts across cultural groups. Given the timing of these developments, it appears that interregional interaction played a critical role in both material and scribal innovation in the spatial contexts of the lowland Maya region as a whole, as well as the specific study area examined herein.

In terms of intermediate iconographic forms, the evidence illustrates a similar pattern. Formally, there is little variability in Late Formative period iconography, although, as noted above, Justeson (2012) has identified some signs in early, ostensibly Maya texts that may be considered ‘intermediate.’ On the other hand, icons begin to appear in new contexts at this time, that is, within a purely textual framework. If one considers the recontextualization of a formally similar iconographic element as representative of an ‘intermediate’ form, then it could be said that these transitional forms do exist in the Late Formative period iconography. Although the process of recontextualization was certainly significant in the emergence of writing systems in the Late Formative period, this process is distinct from the reformulation of iconographic or glyphic elements. Reformulation, and concomitant formal variability in glyphic signs, is far more evident in the Early Classic period, within the already established Mayan script— again in temporal contexts slightly later than I would have expected. Nevertheless, the fact that the data

The data also suggest that such changes took place in slightly different temporal contexts within the Late Formative period. Further, there exist inconsistencies and discrepancies in the data, as well as methodological difficultiues with and lacunae in the iconographic evidence, that complicates interpretation. It may be concluded that the data reflect a long–term process of declining interaction throughout the Late Formative 179

Archaeological Paleography period, coupled with slowly increasing variability and localized innovation at this time. This broad associative pattern supports the central hypothesis in a general sense, despite the fact that the data at hand do not offer a finer degree of temporal resolution to this question. The processes involved in the development of writing—and indeed, sweeping cultural transformations of any kind— are long–term dynamics. Considering the inherent limitations in the data, as well as the methodological difficulties associated with their interpretation, perhaps this is the most that one can expect of the evidence, especially given the broad nature of the questions that have been asked in this investigation.

correlations are tenuous, and appear in slightly disparate temporal contexts. Nonetheless, considering the broad processes at play, in evaluating this component of the hypothesis, it appears probable that long–term, ideational phenomena such as script development can in fact be traced materially, particularly in broad strokes. That said, in the case of this study, the ceramic data from the San Pedro Mártir basin were incapable of providing a fine–grained resolution to this specific research question. These ceramic materials appear much more appropriate for investigations of the processes involved in other types of socio–cultural innovation, such as boundary formation (Englehardt 2010) or shifting settlement patterns. Of course, these results may be attributable to the limited nature of the ceramic sample, the fact that so little evidence for the early history of the Mayan script is available, the scarcity of iconographic evidence from the crucial Late Formative period (or early texts and intermediate iconographic forms) recovered within the study area, as well as the lack of an accepted method to quantitatively evaluate the linguistic and iconographic transformations involved in script development. These limiting factors make direct comparisons difficult. Further investigation may reveal new data, or suggest alternate lines of evidence that may be more profitably applied to an attempted correlation of developing scripts with material goods in Mesoamerican—or other—contexts.

Alternately, it may be that the data are indicating something different altogether. One possibility is that different types of interaction are reflected in distinct kinds of evidence. Interregional interaction is a complex process. There are many types of interaction: the reciprocal exchange of luxury or prestige items between elite trading partners, symbolic trade in ritual objects by participants in shared ceremonies or journeys to pilgrimage centers, or the village to village exchange of utilitarian goods in a market context, to name but a few. Such processes occur across ethnic, political, linguistic, and natural boundaries that are themselves not always clearly defined and may become more or less porous or defined at different points in time. The ceramic materials examined in this study may reflect smaller scale utilitarian exchange in limited spatial contexts, especially given the secondary and tertiary nature of the sites under investigation and the relative absence of elaborate or obviously ritual or prestige ceramic forms in the sample. The ritual context of the majority of MZ loanwords that diffused into other Mesoamerican languages suggests either ritual exchange with or emulation of a relatively more established or prestigious cultural complex. Considering writing or advanced iconography as a cultural tool of emerging elites, the development of writing in Mesoamerica may reflect some sort of interaction between elites and their constituents, in which emerging elites attempted to take control of visual messages by recontextualizing, resituating, or rescaling narrative frameworks (see, e.g., Guernsey 2012). These alternative interpretations raise several intriguing possibilities that should be addressed in future research and are treated in the following chapter.

The third proposition that specific correlations between the distinct datasets employed in this investigation will be evident is not the case. Although there appears to be a spatial and temporal correlation between patterns in each dataset throughout the Late Formative period along the peripheries of southeastern Mesoamerica, the data do not allow for a definitive interpretation. Additionally, the individual lines of evidence suggest slightly different temporal contexts for the crucial shifts in patterns of interaction at this time. Therefore, the data do not suggest that there existed a direct correlation between the developmental processes of scripts and ceramic traditions in the contexts of this study (although the limiting factors mentioned above also affect the degree of resolution with which this query may be addressed). Thus, although the evidence suggests that interaction was crucial to the development of cultural innovations such as new ceramic technologies and writing, the iconographic and linguistic data indicate that the transformations involved in the development of the Mayan script occurred in earlier temporal contexts than the inventive and locally specific changes in regional material and ceramic traditions. Nonetheless, the fact that the data suggest an earlier temporal context for these changes in patterns of interaction is in itself significant.

As regards the second proposition related to the central hypothesis, that the development of the Mayan script can be traced through material remains, the conclusions suggested by the data are similarly of a general nature. If a relationship exists between interaction, script development, and material innovation, I expected to encounter similar patterns of variability and innovation across the datasets. The data suggest a broad correlation between the developmental processes of the ceramic sequence that I have examined and the emergence of Maya writing in the Late Formative period. Direct

Alternately, the data may be interpreted as indicating stages in a gradually changing process of interaction. This possibility is intriguing, especially given the 180

Interpretation and Discussion different temporal and spatial contexts in which changes in ceramic, linguistic, and iconographic data are noted. For example, the down–the–line lag in the emergence of certain ceramic characteristics in the Classic period Maya lowlands may be mirrored by the turn of the millennium divergence of the Chiapa de Corzo and Middle Grijalva sequences, which in turn was followed by the post–Arenal and post– Usulutan transformations at Kaminaljuyú along the extreme southeastern periphery of Mesoamerica in the first century AD. These latter shifts occur in roughly contemporaneous contexts with the florescence of the Izapan script, which is closely related to the Classic Maya writing system, suggesting a potential correlation that may be explored through further investigation. It may prove particularly productive to consider such changes in conjunction with linguistic data, since the Middle Grijalva sequence appears to have occurred at the same time as a linguistic re–orientation of the region toward Oaxacan Mixe. Perhaps a similar linguistic change may be noted along the southeastern periphery at or shortly after this time. This possibility should be explored through future research.

Conclusions In sum, when considered as a whole and in conjunction with the developmental processes inherent in the emergence of the Mayan script, the three distinct lines of evidence employed in this investigation—ceramic materials, linguistic data, and iconographic evidence— indicate broadly similar patterns of variable interaction and innovation in generally analogous spatial and temporal contexts. The data indicate a high level of interregional interaction during the Middle Formative period, reflected in the lack of formal, contextual, and functional variability of ceramic materials and iconographic motifs across disparate traditions in Mesoamerica at this time. The Late Formative period evidence suggests a general decline in interaction as innovations and ceramic variability between regions increase slightly. At this time, iconographic elements in the incipient Mayan script—and elsewhere—were recontextualized, despite maintaining strong formal similarities, as they were juxtaposed from a purely visual interpretive framework to more overtly scribal contexts, divorced from the principles of pictorial interpretation. Although minor discrepancies between the datasets suggest that shifts in these broad patterns occurred at in differing temporal contexts within the Late Formative period, the data indicate that interregional interaction was waning as innovative recontextualizations of material and iconographic forms emerged at this time. In a broad sense, this pattern correlates generally with the spatial and temporal contexts in which the Mayan script emerged.

In conclusion, the results of this investigation at once confirm and disprove facets of the central hypothesis. Given that the processes involved in the development of both ceramic traditions and scripts are broad, regional, and long–term in nature, the timing of such innovations, at least in the case of ceramic materials from the San Pedro Mártir basin and the Mayan script, appears variable. Both separately and as a whole, all datasets suggest the central role of interaction to cultural innovation, in both general terms and in this specific instance, supporting the first underlying proposition of the hypothesis. Ultimately, the relationships among the datasets are muddled by several limiting factors, and the correlations evident between them are of a broad nature, or, alternately, suggest multiple interpretations in variable spatial and temporal contexts. Thus, the second and third propositions of the hypothesis cannot be confirmed with any degree of certainty with the data employed in this investigation. Nonetheless, the data do carry several significant implications that should be tested through future studies. The data suggest relationships between variability in material culture and emerging specialization, developing hierarchies, and the construction of sociopolitical boundaries in the lowlands. The results suggest multiple potentially productive avenues of investigation that may augment current understanding of the complex, dynamic, and large–scale processes involved in interaction and socio–cultural change. Further research and new data will allow future investigations to revisit the questions posed here, more adequately explore and test the methodology proposed—perhaps with different types of material data—and generate more definitive conclusions.

In the Early Classic period, innovation in material and iconographic forms increased significantly. Formal variability across traditions in distinct regions suggests that interaction had decreased at this time. The transformations in ceramic, iconographic, and linguistic traditions evident in the Late Formative and Early Classic period only occurred after a long period of intense interregional interaction. Additionally, evident reformulations of ceramic and iconographic elements were predicated upon and stemmed directly from their prior forms, which were themselves widely distributed and shared between multiple cultural groups in Mesoamerica. The existence of similar forms in multiple interpretive frameworks, whether material or ideational, was a necessary precondition for their subsequent reinterpretation. Such transformative and innovative recontextualizations in ceramics and iconography took place within the context of a gradual diachronic decline in the very interregional interaction that had once allowed for a clear rendition of their meaning and subsequently necessitated their own reformulation. The illustration of the centrality of contextual frameworks in understanding material and symbolic changes is perhaps the most significant revelation of this research. 181

Archaeological Paleography The results of this investigation at once confirm and refute facets of the central hypothesis. Separately all datasets definitely suggest the central role of interaction for cultural innovation in Mesoamerica, but the relationship among the distinct lines of evidence ultimately remains unclear. Although the quantitative and comparative analyses strongly suggest that interregional interaction was intimately involved with both material and scribal innovation in the spatial contexts of the study area, a direct correlation between these processes across temporal contexts is not clearly reflected in the data employed in this study. Ceramic, linguistic, and iconographic data indicate that a great degree of interaction occurred within and across the study area in the Formative period and that interaction decreased significantly in the Early Classic period. The iconographic and linguistic data also suggest that the transformative innovations involved in the development of the Mayan script occurred in earlier temporal contexts than the inventive and locally specific changes in regional material and ceramic traditions. These results are complicated by the fact that so little evidence for the early history of Mesoamerican writing systems, including the Mayan script, is available. The data do not clearly reveal a direct correlation between the developmental processes of scripts and ceramic traditions in the contexts of this study. Thus, although the evidence suggests that interaction was crucial to the development of cultural innovations such as new ceramic technologies and writing, the iconographic and linguistic data indicate that the transformations involved in the development of the Mayan script occurred in earlier temporal contexts than the inventive and locally

specific changes in regional material and ceramic traditions. Nonetheless, broad, regional, and long–term processes of interregional interaction do indeed appear to be central to cultural innovation, although the timing of such innovation varies in this case. A significant revelation of this research is that the sociocultural and contextual frameworks in which material and symbolic goods were used and exchanged in past societies is equally as important as the formal qualities of the artifacts themselves in achieving a more complex understanding of their developmental histories. The results of this research also carry several intriguing implications that should be tested through future research. Among these is the suggestion that large–scale cultural changes occurred in various stages throughout the Middle and Late Formative periods, in temporal contexts earlier than has previously been argued. Additionally, the results could suggest that different kinds of interaction are reflected in variable spatial and temporal contexts by distinct types of evidence. Further investigation may reveal new data, or suggest alternate lines of evidence that may be more profitably applied to an attempted correlation of developing scripts with material goods in Mesoamerican contexts and elsewhere. In any case, the interpretations presented here demonstrate that the examination of stylistic and distributional variability in ceramic materials may profitably inform archaeological investigation, generating especially satisfying results when the data are considered on several different temporal, geographic, and social scales.

182

Chapter 8 Conclusions This investigation has sought to trace the dynamics of the emerging Maya writing system in Formative and Early Classic period Mesoamerica. Building on current models of the development of Mesoamerican writing systems (Houston 2004a; Justeson 1986; Justeson 2012; Justeson et al. 1985; Justeson and Matthews 1990), models of interregional interaction and cultural development in Mesoamerica (Blanton et al. 1996; Marcus 1992b, 1998; Willey 1991), and systemic complexity theory (O’Sullivan et al. 2006; Walby 2007), this study tested the hypothesis that the development of Maya writing in late Middle Formative through Early Classic period Mesoamerica (700 BC–450 AD) is positively correlated with interregional sociopolitical and economic interaction. To test this hypothesis, the investigation explored the relationship between archaeological indicators of interaction and the emergence of Maya writing, attempting to correlate Mayan script development with material remains. It was posited that if the emergence of Maya writing is a correlate of regional interaction, then its developmental process should be traceable archaeologically through artifactual evidence. In this sense, this research sought to demonstrate not only that interregional interaction is central to cultural innovation, both generally and in the specific case of Mesoamerica, but also that patterns of variability in materials remains would spatially and temporally correspond to linguistic and iconographic transformations involved in the emergence of the Mayan script.

diffused linguistic elements and iconographic symbols in identical contexts. The emergence of writing and the differential use of material culture would thus express the same processes of interregional interaction and innovation in analogous contexts. Following the re–classification and sequencing of ceramic materials from the study area, the sample was analyzed comparatively against ceramic sequences evident in assemblages at adjacent sites and in neighboring regions in each of the three temporal periods under investigation. The distribution of the material evidence quantitatively and qualitatively demonstrated the extent, temporality, and direction of interregional interaction. A longitudinal consideration of variability revealed diachronic patterning in shifting levels of interregional interaction. At the same time, continuity and disjunction in the formal stylistic characteristics of material artifacts was explored by performing the statistical analyses of ANOVA and the H score heterogeneity measure, using a selection of defined stylistic attributes and characteristic ceramic features on the materials within the sample. Comparing statistically analyzed levels of variability over time again illustrated diachronic changes in patterns of attribute variation within the sample, which were then compared to regional ceramic sequences. Finally, diachronic patterning in documented linguistic and iconographic elements involved in the emergence of the Mayan script period was examined, in an attempt to discern how and when patterns of linguistic interaction changed between the Middle Formative and Early Classic periods. To that end, an exploration of formal, contextual, and semantic variability was undertaken in a selection of widely shared iconographic motifs across Mesoamerican traditions to determine the spatial and temporal contexts in which grammatical and/or linguistic elements replaced pictorial–iconographic frameworks as the primary organizational principle for visual narratives—the point at which writing emerges.

Ceramic materials were selected for analysis, since they are an excellent variable by which to measure interaction and its relation to linguistic and iconographic evidence involved in the developmental dynamics of Maya writing. A database of ceramic materials was constructed, employing materials from the collections at four sites in the San Pedro Mártir basin of southeastern Tabasco, Mexico: San Claudio, Revancha, Mirador, and Tiradero. This region forms part of the northwest Maya lowlands, a boundary area between Mesoamerican regional interaction spheres, including interior Chiapas, the Gulf Coast, and the Southern Maya lowlands. This area was chosen because prior research has suggested that innovative recontextualizations of material and iconographic elements often occur at the boundaries of interacting regional systems. At such boundaries, users of competing iconographic systems challenge their symbols’ respective meanings through recontextualization, a process reflected in diachronic variability in material forms and functions within and across regions. The archaeological evidence was expected to indicate different functional and formal attributes of stylistically similar artifacts in discrete spatial and temporal contexts that would parallel divergence in the localized use of

If the central hypothesis is correct, it was expected that expected archaeological indicators of material interaction revealed through comparative and quantitative analyses of ceramic data from southeastern Tabasco would parallel the iconographic and linguistic interaction involved in the development of Maya writing within the same contextual frameworks. That is, patterns of interaction evident through three distinct lines of evidence should correspond in spatial and temporal contexts. In simpler terms, patterns of exchange in the material evidence should correspond spatially and temporally to iconographic and linguistic data indicative of interaction, as revealed by the distribution and recontextualization of iconographic elements across 183

Archaeological Paleography the same region. In this case, similarity between the sample and regional sequences would suggest a greater degree of interaction, whereas discrepancies hint at the emergence of innovations. Temporal and spatial parallels between these two patterns of distribution were expected. These parallels should take the form of differential use of homologous material culture in distinct, locally specific contexts. Thus, the study assumed that if a relationship exists between interaction, script development, and material innovation, patterns in the data would indicate a greater degree of interaction and less relative variability between the sample and regional ceramic sequences in earlier Middle and Late Formative contexts, followed by a decrease in interaction and increase in material variation in the Early Classic as localized imposition of cultural meaning on icons and artifacts intensified in the Maya lowlands. Such patterns would suggest the centrality of interaction to cultural innovation in this case and a correlation between the dynamics of the developmental processes of both material traditions and incipient scripts.

Despite the datasets as a whole demonstrating a steady decrease in interregional interaction throughout the Late Formative period that corresponded to increased innovation and greater variability between traditions, the individual lines of evidence suggest that changes in patterns of interaction occurred in different temporal contexts within the broad framework of the Late Formative period. The three datasets suggest that patterns of specifically interregional interaction shifted at slightly different times within the Late Formative period: the linguistic evidence pointing to an earlier decline, the ceramic materials suggesting a shift towards the end of the period, and the iconographic data sketching a general long–term shift throughout the Late Formative period. In this sense, the data do generally suggest that interregional interaction is central to cultural innovation. Nonetheless, the timing of such innovation, at least in the contexts of the ceramic traditions and developing scripts of Late Formative period southeastern Mesoamerica, appears variable. Thus, although these results are complicated by certain limitations in the data, all datasets suggest that interregional interaction was intimately linked to material, linguistic, and iconographic innovation during the Late Formative period. As a whole, however, and when considered in relation to the development of the Mayan script, the evidence suggests that patterns of interaction shifted at disparate points within this broad temporal context. Subsequent innovations in the data were thus reflected at different times.

Results and Implications of this Investigation The results of this investigation were mixed, at once supporting and refuting facets of the central hypothesis and its underlying propositions. The data do appear to indicate that interregional interaction was intimately involved in both changes in lowland ceramic traditions near the Formative– Classic period transition and the emergence of the Mayan script in the early Late Formative period. Nevertheless, a clear positive correlation between the developmental processes could not be definitively established. Although the quantitative and comparative analyses strongly suggest that interregional interaction was connected to material and scribal innovation in the spatial contexts of the study area, the data do not conclusively indicate that there existed a direct spatial or temporal correlation between the developmental processes operating on the three distinct lines of evidence. Ceramic, linguistic, and iconographic data exhibit little variability within the sample or between traditions in earlier temporal contexts, suggesting that a great degree of interaction occurred within and across the study area in the Formative period. Marked increases in formal and contextual variability among the evidence in the Early Classic period indicates that interaction decreased significantly at this time. The iconographic and linguistic data also suggest that the transformative innovations involved in the development of the Mayan script occurred in earlier temporal contexts— at the outset of the Late Formative period c. 250 BC— than the inventive and locally specific changes in regional material and ceramic traditions, which are most evident nearer the Late Formative–Early Classic period horizon c. AD 250.

Of course, further investigation may bring new data to light, or suggest alternative bodies of evidence that may be more expediently applied to an attempted correlation of developing scripts with material goods in the contexts of Formative period Mesoamerica. Nonetheless, the results of this research highlight the complexity inherent in the Late Formative period and suggest many intriguing possibilities. This study thus carries several potentially illuminating practical and interpretive implications. The two most significant conclusions suggested by this research are the implication that large–scale changes in cultural processes may have occurred earlier in Mesoamerican history than has previously been thought, closer to the Middle–Late Formative period transition, and the implication that subtle transformations in contextual frameworks may prove equally as integral to understanding processes of long–term cultural change as diachronic variation in the formal characteristics of material data. First, inconsistencies in the temporal contexts of shifting patterns of interaction suggested by the various datasets may indicate that large–scale changes in cultural processes were occurring in temporal contexts that are earlier than has previously been thought, closer to the Middle–Late Formative period horizon. The data suggest that the processes traditionally associated with the Formative– Classic period transition in the Maya lowlands—state

It was hoped that the outcomes of this investigation would establish specific correlations between the appearance of the Mayan script and interregional interaction during the Formative period in southeastern Mesoamerica. 184

Conclusions formation, the emergence of writing, the development of elaborate material cultural forms—appear to occur earlier than the present lowland chronology suggests. The linguistic and iconographic data explored here suggest that significant transformations in lowland societies were occurring closer to the outset of the Late Formative period. Even the ceramic evidence from a peripheral area such as the San Pedro Mártir basin intimates that changes in material traditions began in advance of the Formative–Classic period transition—a conclusion long recognized by those ceramicists that posit the existence of a ‘Protoclassic’ period (see, e.g., Pring 1977). More recent data, such as the San Bartolo murals and the monumental architecture at El Mirador that dates to the Middle Formative period, also suggest that fundamental sociocultural transformations occurred well before the dawn of the Classic period. The dynamic processes that possibilitize such grand cultural changes are necessarily long–term in nature and have roots in earlier temporal contexts and distinct spatial milieus. Nonetheless, the possibility that cultural changes traditionally associated with the transition between the Formative and Classic periods in the Maya lowlands were occurring in earlier temporal contexts is a significant implication of this study, and one that may be further explored through future investigation.

Moreover, this study underscores the complex relationships between processes of changing sociopolitical organization and transformative variability in traditions of material culture. If one considers writing as a cultural tool, the development of the Mayan script may reflect interaction between elites and their constituents, in which emerging elites attempted to take control of visual messages by recontextualizing narrative frameworks. This potential relationship should be explored through a fuller examination of the associations between the emergence of writing and the development of complex sociopolitical hierarchies, in both lowland Mesoamerica and in other cultural contexts. In particular, testing the model proposed here may produce especially illuminating results by employing data from the Bronze Age Aegean world. In those contexts, many of the same concurrent processes are evident: the development of complex sociopolitical formations, the emergence of writing, and increasing degrees of variability between traditions of material culture. Developments in the Aegean during the Bronze Age, including the emergence of Linear A and Linear B and the development of Minoan and Mycenaean states, appear to provide an excellent parallel test case that may elucidate the conclusions reached through this research or the suggestions that remain obscure from the perspective of the data employed in this study.

The interaction evident in the Middle Formative data appears intimately connected to subsequent innovations in ceramic, linguistic, iconographic, and scribal traditions across Mesoamerica. Nonetheless, as noted above, timing was variable. The three datasets suggest that patterns of specifically interregional interaction shifted at slightly different times within the Late Formative period. Thus, the data may suggest that changes in broad patterns of interaction—or the shifting spatial extent or boundaries of individual networks of exchange— occurred in stages throughout the Late Formative period and consequently were reflected in the evidence at different points in time. The ceramic evidence in particular support this contention, with episodes of divergence occurring in stages within discrete spatial contexts: first in the Middle Grijalva region c. AD 1, then along the southeastern periphery later in the first century AD, followed by increasing experimentation throughout the Maya lowlands in the latter half of the second century AD nearing the Formative–Classic period transition. Alternately, this investigation may suggest that distinct types of data were linked to specific types of interaction— ritual, commercial, elite—whose patterns shifted in disparate spatial and temporal contexts. This possibility may be clarified through an examination of potential links between material variability and the emergence of specialized systems of production for distinct types or classes of artifacts. Such a study would require greater detail regarding the distinct archaeological contexts (e.g., ritual, utilitarian) of the evidence than the data employed here were able to provide.

A further revelation of this study is the simultaneous continuation of formally similar iconographic elements in parallel semantic and syntactical contexts across traditions, suggesting that the base interpretive framework that allowed for the transmission of meaning through visual motifs continued through the Late Formative period. The recontextualization of shared iconographic elements that began at this time indicates incipient transformations in the underlying narrative frameworks that substantiated the meaning of iconic motifs. Innovative new elements also emerge during the Late Formative period. These elements intimate the establishment of regionally specific interpretive frameworks that were no longer dependent on the shared principles of visual interpretation reflected in the widespread distribution of formally and contextually similar motifs. This emergent breakdown in interpretive frameworks—and the establishment of new interpretive principles for visual elements—would appear to suggest a decrease in interregional interaction at this time. What is clear is that the reformulation and recontextualization of shared visual motifs could not have occurred without the prior existence of a common interpretive framework. This research therefore illustrates the importance of context—of both use and exchange—in understanding long–term trajectories of change in both material and symbolic culture. This investigation has also served to illuminate certain dynamics of interaction and innovation inherent in the developmental processes involved in the emergence of 185

Archaeological Paleography the Mayan script in Late Formative period southeastern Mesoamerica. On a general level, the results suggest that a functional relationship and an inverse correlation exist between levels of interregional interaction and material innovation. In addition, this study has intimated a broad associative pattern between a gradually decreasing degree of interregional interaction and greater material variability in lowland Maya ceramics, including those of the San Pedro Mártir basin. At the same time, the results suggest that levels of interregional interaction were on the wane in Late Formative temporal contexts, a period during which iconographic elements shared across traditions within a pictorial narrative framework began to be recontextualized and infused with linguistic values, leading to the emergence a diversity of regional scripts. These results suggest the outlines of relationships and intimate further avenues of inquiry that may be pursued to clarify the relationship between the complex and dynamic processes that operated in the Late Formative period and the significant sociocultural transformations that took place at this time.

suggests that the development of this boundary was coeval with trends of increasing complexity, integration, and centralization in the Maya lowlands that occurred at—or slightly before—the Late Formative–Early Classic period transition. This conclusion highlights the complexity of both processes of boundary formation and the transformative developments that occurred in the Late Formative period on a general level. The suggestion also emphasizes the need for further study to clarify the potential relationships between material variability, boundary formation, emerging hierarchies, and the development of craft specialization. Additionally, the results of this investigation raise further questions regarding how and why the interregional interaction that appears so central to cultural innovation and change occurred in the first place. In the end, and despite the appearance of a generalized associative pattern, the root causes of the interaction reflected in the data remain obscure, as do the mechanisms by which interaction occurred and the contexts in which processes of interaction functioned. In this study, the specific processual mechanisms that facilitated interaction, such as potential demic migrations or population movements, warfare and colonization, or the establishment of simple commercial exchange through specific and defined trade networks, were not considered. Numerous prior investigations have examined the latter issues, although not necessarily from the standpoint of the relationship between interaction and cultural change (Lee 1978; Lowe 1977, 1989; Navarrete 1978; Ochoa 1983; Ochoa and Vargas 1983; Price 1978; Rathje et al. 1978). A reexamination of the data employed in this investigation against that generated through these prior studies in conjunction with evidence of settlement patterns and data from regional surveys—with the goal of exploring the relationship between interaction and cultural change—may clarify the mechanisms of both material and ideational exchange, as well as the results of such interaction.

Third, this investigation raises fundamental questions regarding the degree to which the transformations observed in the datasets are linked to other processes involved in sociocultural change and state formation, such as integration, centralization, and boundary formation (Parkinson 2002, 2006; Parkinson and Galaty 2007, 2010). Many scholars have traced the emergence of complex archaeological phenomena to interregional interaction, and in the contexts of this investigation the data suggest that interregional interaction goes hand in hand with these processes, as it does with innovation in cultural technologies, but ultimately the relationship remains unclear. It may be that a similar correlative pattern exists; not necessarily a causal relationship, per se, but something more like the general associative correlations suggested herein. Elsewhere (Englehardt 2010), I have employed the same ceramic dataset to argue that diachronic changes in sociopolitical organization, namely increased centralization, integration, and hierarchical control, are positively correlated with both amplified material variation and an increased emphasis on the delineation and control of peripheral boundaries. In the contexts of this investigation, it appears that a more pronounced and far less porous social boundary developed along the intermediate plains of the mid–lower Usumacinta and lower San Pedro Mártir basins in the Early Classic period. This frontier separated the northwest Maya lowlands from the developing core area of Classic Maya society centralized in the Petén, with sites in the low sierras south of the intermediate plains such as San Claudio becoming more integrated in Classic Maya traditions, and sites to the north of the plains such as Mirador, Revancha, and Tiradero orienting themselves outward. Material variability in my ceramic data

Finally, it is unclear why some areas appear more affected by interregional interaction than others. The evidence explored here supports the contention that boundary areas are particularly fertile grounds in which to observe the effects of interaction, particularly as it relates to script development and material change, but this does not always seem to be the case. For example, scripts do not appear to have developed on the western periphery of Mesoamerica, despite the importance of that area as a node of exchange between regional systems throughout Mesoamerican history, although new studies may refine this understanding (John Pohl, personal communication, 2011). Likewise, the effects of the well–documented Formative period interaction between the Gulf Coast and the Valley of Oaxaca appears not along the boundary between these cultural systems (i.e., the highlands of northern Oaxaca), but at central sites within those 186

Conclusions regions, such as San Lorenzo or San José Mogote. On the other hand, both of these cultures appear to have developed script systems after an extended period of intense interaction. Exploring the development of Olmec and Zapotec writing in conjunction with changing material traditions in these cultures (e.g., Cheetham 2007; Urcid 2001, 2005) may prove a significant step in refining the model proposed here.

temporal contexts. Finally, the evidence intimates that diverse kinds of interaction—elite trading, village to village, or ritual exchange—may be reflected in different types of data and are thus reflected in discrete spatial contexts. Although the evidence does not conform exactly to expected patterns, the data do suggest several potentially significant interpretations. Further research is necessary to refine the findings and implications of this study and to pursue some of the possibilities suggested by the data. Such investigation will permit a more complete understanding of the complex processes and relationships involved in interregional interaction and diachronic change in both material and symbolic culture.

In sum, although this investigation has not precisely clarified the relationships between the emergence of writing systems, interaction, and material cultural complexity, it does represent a significant first step in that endeavor. This research offers several contributions to larger anthropological debates, particularly in terms of constructing a broader, more nuanced understanding of the development of writing, in both general and specific terms. The results likewise suggest that examinations of the development of writing may serve as a proxy for the study of material changes—and vice–versa— thus informing debate regarding the emergence of sociocultural innovations in antiquity. In this sense, the results do suggest that script development may be traced archaeologically using material remains. The outcomes of this investigation suggest new directions for evaluating the relationship between the development of writing and complexes of material culture in various contexts. Further, the interpretations presented here demonstrate that the examination of stylistic and distributional variability in ceramic materials may profitably inform archaeological investigation. The integration of epigraphic and archaeological data confirms that epigraphy remains an integral tool in archaeological research, especially in Mesoamerican contexts. Lastly, this study has shown that a consideration of stylistic variability across multiple, distinct lines of evidence is capable of generating particularly satisfying results—in the form of cogent research questions—when the data are considered on several different temporal, geographic, and social scales.

The attempt to model the development of writing on stylistically defined artifactual zones of interaction permits more pertinent questions about the relationship between material and ideological aspects of past societies. In this sense, the model created in this investigation moves beyond analytic or interpretive schemes that oppose style and function and allows for a more holistic understanding of variability in material culture, in both general terms, and in the specific case of the Formative– Classic period transition in the Maya lowlands. As such, the significance of the model is that it may be applied to archaeological evidence from other areas of the world to elucidate the connections between the data employed in this case and to refine the conclusions suggested by this research. Future investigation may reveal new data or suggest further linkages that will allow for a more holistic understanding of the associative relationship between material artifacts and symbolic concepts in antiquity (Clark 2004; Renfrew 2001). In this sense, and irrespective of the results of hypothesis testing in this investigation, the cross–cultural applicability of the model proposed here may prove the most enduring contribution of this research. Exploring similar questions in discrete spatial, temporal, or cultural contexts not only will refine the results suggested through this research, but also will enhance the archaeological understanding of the complex processes involved in boundary formation, the emergence of hierarchies, the development of specialized systems of material production, and the functional uses of ostensibly elite material culture. The model has strongly suggested an associative correlation between the development of writing and the emergence of increasingly complex material culture and sociopolitical configurations. As others have suggested (see, e.g., Fahsen 1999; Grube 1999) writing, among other phenomena, has often been used to define the Classic period in southeastern Mesoamerica. Nonetheless, the results achieved here, in tandem with new evidence (e.g., the San Bartolo murals), suggest certain changes in the traditional chronology, with large–scale changes occurring closer to the Middle–Late Formative period transition. If applied in other areas of the world, analyses based on the model

Assessment of the Proposed Model and Future Directions Despite mixed results, as a theoretical model this study offers several significant conclusions and suggests multiple avenues of future inquiry. First, the evidence and interpretations presented here illuminate the centrality of context—of both use and exchange—to forming a more complete understanding of diachronic formal change in material and symbolic culture. Second, the data suggest that large–scale changes in cultural processes may have occurred earlier in Mesoamerican history than has previously been thought, closer to the Middle–Late Formative period transition, as opposed to the Late Formative–Early Classic period horizon. Further, such transformations may have occurred in stages and are thus reflected in the data in discrete 187

Archaeological Paleography proposed in this study may suggest that a similar re– assessment of sociocultural chronologies in disparate contexts is in order.

Maya ceramic traditions. An intriguing possibility for future research would be to source the materials used in ceramic production within the San Pedro Mártir basin and to compare those results with a current investigation that seeks to source the Formative period assemblage from San Andrés, near La Venta, Tabasco (Mary Pohl, personal communication, 2011; Cheetham 2007).

In Mesoamerican contexts, the model succeeds in demonstrating how the study of writing can serve as a proxy for the study of sociocultural innovation by highlighting the effects of shifting networks of interregional interaction on lowland Maya material culture, linguistics, and scribal traditions. These results add to the growing body of evidence that underscores the inherent complexity of the sociocultural processes at play during the Late Formative period in southeastern Mesoamerica. Moreover, the outcomes of this investigation suggest new directions for evaluating the relationship between the development of writing and complexes of material culture in various contexts. The model proposed here is thus an important first step in reevaluating the developmental dynamics of emerging writing systems in both Mesoamerica and beyond. Other such analyses in divergent contexts may bring a fuller, long–term comprehension of the processes involved— and the role of interregional interaction—in cultural innovation and change. In this sense, the proposed model is not predictive, and the results of this study do not suggest that writing and material innovation are always positively correlated, nor that material, iconographic, or linguistic interaction always leads to the emergence of scripts. Rather, the model, and the results derived from it, suggest that interaction fosters a particular set of conditions necessary for the emergence of writing, not unlike various archaeological treatments of the development of agriculture. Thus, the proposed model may assist in identifying potential material correlates of scrpt development in other contexts as well.

Once a fuller comprehension of the ceramic data from the San Pedro Mártir region is achieved, these data can be integrated with and compared against other evidence—such as settlement pattern data—that speaks to diachronic changes in sociopolitical organization in order to explicate the framework within which processes of interaction and innovation in material culture occurred. Alternately, it is possible that an attempt to test the model with materials from a site or region with a more established ceramic sequence or a fuller, more holistic set of regional data, such as the central Petén lowlands, would provide evidence that would allow for a more complete evaluation of the proposed model. Comparing the results of this investigation with a similar study of such a site or region would offer the possibility of clarifying the general and somewhat inconsistent correlations between interaction, material variability, and emerging scripts that have resulted from this study. A more refined operationalization of the model and methods employed in this research is complicated by the paucity of evidence for incipient script technologies in Mesoamerica. This simple fact consistently troubles researchers interested in the development of writing in these contexts, and continues to frustrate our limited understanding of the development of this complex cultural technology. An additional complication is due to the fact that there is no established method for quantifying iconographic differences between visual traditions, in either shared elements or discrete reformulations of iconic motifs. Further, there is no method to determine a degree of purely iconographic distance or dissimilarity between a reformulated icon and its original pictographic referent, either within a single tradition or in the case of an element shared between multiple iconographic systems. The development of such a method for calculating degrees of difference between visual elements and motifs would allow iconographic datasets to be more readily and fruitfully compared in quantitative terms against other material and archaeological data.

This project has presented an opportunity to document and publish data from southeastern Tabasco, a potentially significant boundary area, but one traditionally underrepresented in Mesoamerican archaeology (Ochoa 1983). Future studies in this region may yield additional evidence that can be incorporated into the model and clarify the discrepancies and apparent contradictions in the slightly variable patterns of interaction suggested by the data. In addition, ceramic data from other sites within southeastern Tabasco may prove capable of not only refining the ceramic chronology of the San Pedro Mártir basin, but also teasing out and illuminating the subtle processes of experimentation with ceramic forms and production techniques in the Late Formative period. Current work at sites in the northwest Petén (e.g., La Florida) and the re–excavation of Tierra Blanca may also prove useful in this regard. A fuller exploration of the ceramic materials from the San Pedro Mártir region, particularly a more detailed consideration of variability in ceramic temper and paste (Culbert and Rands 2007), would likely augment and nuance current understanding of emerging variability in Late Formative period lowland

Although differences in the changing formal attributes of iconographic motifs were stylistically compared in this investigation, the data revealed that it was contextual rather than formal variation that was most telling in terms of iconographic transformations during the crucial Late Formative period. At that time, formal variability in iconographic motifs across traditions remained low, but innovative recontextualizations of certain shared elements appeared in the scripts of the southeastern 188

Conclusions branch. Although it was hoped that contextual and functional variability in the ceramic evidence could be simultaneously examined, contextual limitations in much of the material data precluded this possibility. In the future, exploring these contextual changes in iconography against a ceramic sample that provides finer detail in terms of defined archaeological contexts would prove capable of providing a fuller understanding of these subtle transformations in both sets of data. In the absence of a method to quantitatively evaluate degrees of iconographic distance, a fuller understanding of contextual changes in iconographic motifs is an important first step, one that I am currently pursuing in ongoing research.

development of sociopolitical complexity, cultural innovation, and long–term processes of linguistic and socio–cultural change, furthering anthropological debate in each sub–discipline. The primary merit of this work is that it adds to the increasingly nuanced understanding of emergent complexity in the Late Formative period of southeastern Mesoamerica. This investigation has quantitatively and qualitatively tested a model that sought to explain the emergence of the Maya script by correlating a theoretical archetype of script development with archaeological evidence. This model may be applied to data from other spatial, temporal, or cultural contexts not only to elucidate the significance of the results suggested by this research, but also to enhance archaeological understanding of the multifaceted processes involved in the emergence of increasingly complex material and symbolic culture in ancient societies. In Mesoamerican contexts, researchers do not fully understand the emergence of writing, since available evidence is scarce, and what little exists does not fit neatly into current models of the development of writing. This investigation offered a rare chance to correlate material and linguistic data and to evaluate the temporal depth of interregional interaction exhibited in distributions of material remains. Further research and new data may allow the model to be refined and clarified in the future.

Final Thoughts Standing at the nexus of archaeology, iconography, and linguistic anthropology, this research contributes to anthropological debate regarding the effects of interaction on material and symbolic culture. This study underscores the effects of shifting networks of interregional interaction on lowland Maya material culture, linguistics, and scribal traditions. By examining the relationship between such transformations and material variability against a backdrop of changing sociopolitical organization, the model proposed in this investigation elucidates more complex understandings of larger archaeological questions related to boundary formation, emergent hierarchies, the development of specialized systems of material production, and the functional uses of ostensibly elite material culture. Regardless of the specific outcomes of the investigation, this study represents a new way of thinking about and exploring the relationships between material and symbolic culture in past societies, especially during critical episodes of sociopolitical and cultural transition such as the Late Formative period in southeastern Mesoamerica. In sum, this research has shown how combining epigraphic, linguistic, and archaeological data can illuminate wider questions related to the

In conclusion, the future directions for research along the lines discussed here should eventually allow for more satisfying conclusions to the fundamental questions that posed in this investigation. Understanding how and why interregional interaction occurs, how and why material goods and ideas are exchanged between groups, how and why scripts emerge from iconography, or how and why these long–term processes are correlated is an integral component in constructing a more complete record of the dynamic and complex histories of past human societies. That goal will continue to be at the heart of archaeological investigation in both Mesoamerica and beyond.

189

Bibliography Acuña Smith, M. J. 2005. La Cerámica del Clásico Temprano de Piedras Negras (350–550 d.C.). Tesis de Licenciatura, Universidad de San Carlos de Guatemala. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Adams, R. E. W. 1971. The Ceramics of Altar de Sacrificios, Guatemala. Papers of the Peabody Museum of Archaeology and Ethnography, Vol. 63, No. 1. Cambridge, Harvard University. Altschul, J. H. 1978. The Development of the Chacoan Interaction Sphere. Journal of Anthropological Research 34(1): 109–146. Anaya Hernández, A. 2002. The Pomoná Kingdom and its Hinterland. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Anaya Hernández, A., Guenter, S. P., and Zender, M. U. 2003. Sak Tz’i’, a Classic Maya Center: A Locational Model Based on GIS and Epigraphy. Latin American Antiquity 14(2): 179–191. Anderson, L. B. 1993. The Writing System of La Mojarra and Associated Monuments. Second edition. Washington, D. C., Ecological Linguistics. Anderson, P. 1999. Complexity Theory and Organization Science. Organization Science 10(3): 216–232. Bagley, R. W. 2004. Anyang Writing and the Origin of the Chinese Writing System. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 190–249. Cambridge, Cambridge University Press. Baines, J. 2004. The Earliest Egyptian Writing: Development, Context, Purpose. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 150–189. Cambridge, Cambridge University Press. Baines, J., and Yoffee, N. 1998. Order, Legitimacy, and Wealth in Ancient Egypt and Mesopotamia. In G. M. Feinman and J. Marcus (eds.), Archaic States: pp. 199–260. Santa Fe, School for Advanced Research. Barabási, A.–L., and Albert, R. 1999. Emergence of Scaling in Random Networks. Science 286: 509–512. Barth, F. 1969. Ethnic Groups and Boundaries. Boston, Little Brown and Co. Bentley, R. A. 2003. An Introduction to Complex Systems. In R. A. Bentley and H. D. G. Maschner (eds.), Complex Systems and Archaeology: 9–23. Salt Lake City, University of Utah Press. Bentley, R. A., and Maschner, H. D. G. 2001. Stylistic Change as a Self–Organized Critical Phenomenon: An Archaeological Study in Complexity. Journal of Archaeological Method and Theory 8(1): 35–66. Bentley, R. A., and Maschner, H. D. G. 2003. Avalanches of Ideas. n R. A. Bentley and H. D. G. Maschner (eds.), Complex Systems and Archaeology: 61–73. Salt Lake City, University of Utah Press.

Berlin, H. 1955. Selected Pottery from Tabasco. Notes in Middle American Archaeology and Ethnography, No. 126. Washington, D. C., Carnegie Institution. Berlin, H. 1956. Late Pottery Horizons of Tabasco, México. Contributions to American Anthropology and History, Vol. XII, N° 59: 95–153, Publication 606. Washington, D. C., Carnegie Institution. Bertalanffy, L. 1972. The History and Status of General Systems Theory. The Academy of Management Journal 15(4): 407–426. Binford, L. R. 1965. Archaeological Systematics and the Study of Culture Process. American Antiquity 31: 203–210. Bird, R. B., and Smith, E. A. 2005. Signaling Theory, Strategic Interaction, and Symbolic Capital. Current Anthropology 46(2): 221–248. Blanton, R. E., Feinman, G. M., Kowalewski, S. A., and Peregrine, P. N. 1996. A Dual–Processual Theory for the Evolution of Mesoamerican Civilization. Current Anthropology 37(1): 1–14. Boone, E. H. 2004. Beyond Writing. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 313–348. Cambridge, Cambridge University Press. Boone, J. L. III. 1987. Defining and Measuring Midden Catchment. American Antiquity 52(2): 336–345. Boot, E. 2003. The Human Hand in Classic Maya Hieroglyphic Writing. Electronic document, http:// www.mesoweb.com/features/boot/human_hand.pdf, accessed May 11, 2015. Bourdieu, P. 1977. Outline of a Theory of Practice. Cambridge, Cambridge University Press. Bourdieu, P. 1990. The Logic of Practice. Stanford, Stanford University Press. Bove, F. J. 1991. The Teotihuacán–Kaminaljuyú– Tikal Connection: A View from the South Coast of Guatemala. In M. G. Robertson and V. M. Fields (eds.), Sixth Palenque Roundtable, 1986: 135–142. Norman, University of Oklahoma Press. Brainerd, G. W. 1951. The Place of Chronological Ordering in Archaeological Analysis. American Antiquity 16(3): 301–313. Braswell, G. E. 2003a. Reinterpreting Early Classic Interaction. In G. E. Braswell (ed.), The Maya and Teotihuacán: Reinterpreting Early Classic Interaction: 1–44. Austin, University of Texas Press. Braswell, G. E. 2003b. Understanding Early Classic Interaction between Kaminaljuyú and Central Mexico. In G. E. Braswell (ed.), The Maya and Teotihuacán: Reinterpreting Early Classic Interaction: 105–142. Austin, University of Texas Press. Brumfiel, E. M. 1995. Heterarchy and the Analysis of Complex Societies: Comments. Archeological 190

Bibliography D.Charnay, etc. (sic) (eds.), América Pintoresca: Descripción y Viajes al Nuevo Continente, por los más Modernos Exploradores. Barcelona, Montaner y Simon. Cheetham, D. 2007. Cantón Corralito: Objects from a Possible Gulf Olmec Colony. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Cherry, J. F. 1984. Emergence of the State in the Prehistoric Aegean. Proceedings of the Cambridge Philological Society 210: 18–48. Cherry, J. F. 1986. Polities and Palaces. In C. Renfrew and J. F. Cherry (eds.), Peer Polity Interaction and Socio Political Change: 19–45. Cambridge, Cambridge University Press. Chinchilla Mazariegos, O. 1999. Desarrollo de la Escritura en Mesoamérica durante el Pre–Clásico. In M. Popenoe de Hatch (ed.), Historia General de Guatemala, Tomo 1: 557–562. Guatemala City, Asociación de Amigos del País, Fundación para la Cultura y Desarrollo. Christin, A.–M. (ed.) 2002. A History of Writing: From Hieroglyph to Multimedia. Paris, Flammarion. Clark, J. E. 2004. The Birth of Mesoamerican Metaphysics: Sedentism, Engagement, and Moral Superiority. In E. DeMarrais, C. Gosden, and C. Renfrew (eds.), Rethinking Materiality: The Engagement of Mind with the Material World: 205– 224. McDonald Institute Monographs, McDonald Institute for Archaeological Research. Cambridge, Cambridge University Press. Clark, J. E., and Blake, M. 1994. The Power of Prestige: Competitive Generosity and the Emergence of Rank Societies in Lowland Mesoamerica. In E. M. Brumfiel and J. W. Fox (eds.), Factional Competition and Political Development in the New World: 17–30. Cambridge, Cambridge University Press. Clark, J. E., and Cheetham, D. 2002. Mesoamerica’s Tribal Foundations. In W. A. Parkinson (ed.), The Archaeology of Tribal Societies: 278–339. Ann Arbor, International Monographs in Prehistory. Clark, J. E., and Pye, M. E. 2000. The Pacific Coast and the Olmec Question. In J. E. Clark and M. E. Pye (eds.), Olmec Art and Archaeology in Mesoamerica: 217–251. Studies in the History of Art, No. 58. Center for Advanced Study in the Visual Arts Symposium Papers XXXV. Washington, D. C./ New Haven, National Gallery of Art/Yale University Press. Coe, M. D. 1965. The Olmec Style and its Distribution. In G. R. Willey (ed.), Archaeology of Southern Mesoamerica: 739–775. Handbook of Middle American Indians, Vol. 3, Pt. 2, R. Wauchope, general editor, Austin, University of Texas Press. Coe, M. D. 1976. Early Steps in the Evolution of Mayan Writing. In H. B. Nicholson (ed.), Origins of Religious Art and Iconography in Preclassic Mesoamerica: 107–122. Latin American Studies Series, No. 31. Los Angeles, University of California at Los Angeles.

Papers of the American Anthropological Association 6(1): 125–131. Brumfiel, E. M. 2003. Factional Competition and Political Development in the New World: An Introduction. In E. M. Brumfiel and J. W. Fox (eds.), Factional Competition and Political Development in the New World: 3–14. Cambridge, Cambridge University Press. Brumfiel, E. M., and Earle, T. K. 1987. Specialization, Exchange, and Complex Societies: An Introduction. In E. M. Brumfiel and T. K. Earle (eds.), Specialization, Exchange, and Complex Societies: 1–9. Cambridge, Cambridge University Press. Bryant, D. D., Clark, J. E. and Cheetham, D. 2005. Ceramic Sequence of the Upper Grijalva Region, Chiapas, Mexico. Papers of the New World Archaeological Foundation, No. 67. Provo, Utah, Brigham Young University. Buchanan, M. 2002. Nexus: Small Worlds and the Groundbreaking Theory of Networks. New York, W. W. Norton and Company. Caldwell, J. R. 1964. Interaction Spheres in Prehistory. Illinois State Museum Scientific Papers 12(6): 133– 193. Campbell, L. R. 1976. The Linguistic Prehistory of the Southern Mesoamerican Periphery. In Las Fronteras de Mesoamérica, Vol. 1: 157–183. Mexico City, Sociedad Mexicana de Antropología. Campbell, L. R. 1984. The Implications of Mayan Historical Linguistics for Glyphic Research. In J. S. Justeson and L. R. Campbell (eds.), Phoneticism and Mayan Hieroglyphic Writing: 1–16. Institute for Mesoamerican Studies, Publication 9. Albany, State University of New York. Campbell, L. R. 1988. The Linguistics of Southeast Chiapas, México. Papers of the New World Archaeological Foundation, No. 50. Provo, Utah, Brigham Young University. Campbell, L. R., Kaufman, T. S., and Smith–Stark, T. C. 1986. Meso–America as a Linguistic Area. Language 62(3): 530–553. Carrasco, M. D., and Englehardt, J. D. 2015. Diphrastic Kennings on the Cascajal Block and the Emergence of Mesoamerican Writing. Cambridge Archaeological Journal 25(3): 1–22. Carrasco, M. D., and Englehardt, J. D. n.d. F l o w e r s , Thrones, and Headdresses: Conventions and Linguistic Tropes in Olmec Art and Writing. Unpublished manuscript in possession of the author. Carr, C. 1995. A Unified Middle–Range Theory of Artifact Design. In C. Carr and J. Neitzel (eds.), Style, Society, and Person: 171–258. New York, Plenum. Carr, C., and Neitzel, J. 1995. Integrating Approaches to Material Style in Theory and Philosophy. In C. Carr and J. Neitzel (eds.), Style, Society, and Person: 3–26. New York, Plenum. Charnay, D. 1884. Mis descubrimientos en México y en la América Central. In C. Wiener, D. Crevaux, 191

Archaeological Paleography Innovation, and the Development of the Mayan Script through Material Remains. PhD dissertation. Florida State University. Ann Arbor, University Microfilms. Englehardt, J. D., and Carrasco, M. D. 2015. Formative Period Interregional Interaction and the Emergence of Mesoamerican Scripts. Paper presented at the 80th Annual Meeting of the Society for American Archaeology. Eppich, K. 2004. La Secuencia de la Tradición Cerámica de El Perú-Waka’, Petén. In J. P. Laporte, B. Arroyo, and H. E. Mejía (eds.), XVIII Simposio de Investigaciones Arqueológicas en Guatemala, 2004. Guatemala City, Museo Nacional de Arqueología y Etnología. Fahsen Ortega, F. 1999. La Transición Preclásico Tardío–Clásico Temprano: El Desarrollo de los Estados Mayas y la Escritura. In N. Grube (ed.), The Emergence of Lowland Maya Civilization: The Transition from the Preclassic to Early Classic, A Conference at Hildesheim, Germany, November 1992: 151–162. Acta Mesoamericana, No. 8, Verlag Anton Saurwein, Markt Schwaber. Fash, W. L., and Sharer, R. J. 1991. Sociopolitical Developments and Methodological Issues at Copán, Honduras: A Conjunctive Perspective. Latin American Antiquity 2(2): 166–187. Feinman, G. M. 1998. Scale and Social Organization: Perspectives on the Archaic State. In G. M. Feinman and J. Marcus (eds.), Archaic States: 95–134. Santa Fe, School for Advanced Research. Fernández Tejedo, I. 1988. Zonas Arqueológicas Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia and Gobierno del Estado de Tabasco. Fields, V. M. 1991. The Iconographic Heritage of the Maya Jester God. In M. G. Robertson and V. M. Fields (eds.), Sixth Palenque Roundtable, 1986: 167– 174. Norman, University of Oklahoma Press. Fischer, H. G. 1989. The Origin of Egyptian Hieroglyphs. In W. M. Senner (ed.), The Origins of Writing: 59–76. Lincoln, University of Nebraska Press. Flannery, K. V. 1968a. The Olmec and the Valley of Oaxaca: A Model for Interregional Interaction in Formative Times. In E. P. Benson (ed.), Dumbarton Oaks Conference on the Olmec: 79–117. Washington, D. C., Dumbarton Oaks. Flannery, K. V. 1968b. Archaeological Systems Theory and Early Mesoamerica. In B. J. Meggers (ed.), Anthropological Archaeology in the Americas: 67–87. The Anthropological Society of Washington, D. C. Flannery, K. V., and Marcus, J. 2000. Formative Mexican Chiefdoms and the Myth of the “Mother Culture.” Journal of Anthropological Archaeology 19(1): 1–37. Foias, A. E., and Bishop, R. L. 1997. Changing Ceramics Production and Exchange in the Petexbatún Region, Guatemala: Reconsidering the Classic Maya Collapse. Ancient Mesoamerica 8: 275–291. Forsyth, D. W. 1989. The Ceramics of El Mirador, Petén, Guatemala. El Mirador Series, Part 4. Papers of the New World Archaeological Foundation, No. 63. Provo, Utah, Brigham Young University.

Coe, M. D. 1977. Olmec and Maya: A Study in Relationships. In R. E. W. Adams (ed.), The Origins of Maya Civilization: 183–195. Albuquerque, University of New Mexico Press. Cook, K. S., and Whitmeyer, J. M. 1992. Two Approaches to Social Structure: Exchange Theory and Network Analysis. Annual Review of Sociology 18: 109–127. Cooper, J. S. 2004. Babylonian Beginnings: The Origin of the Cuneiform Writing System in Comparative Perspective. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 71–99. Cambridge, Cambridge University Press. Cuevas Reyes, F. 1996. Proyecto de Salvamento Arqueológico “Carretera Tenosique–Guatemala.” Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. Culbert, T. Patrick, and Rands, R. L. 2007. Multiple Classifications: An Alternative Approach to the Investigation of Maya Ceramics. Latin American Antiquity 18(2): 181–190. Dahlin, B. H., Quizar, R., and Dahlin, A. 1987. Linguistic Divergence and the Collapse of Preclassic Civilization in Southern Mesoamerica. American Antiquity 52(2): 367–382. Demarest, A. A. 1986. The Archaeology of Santa Leticia and the Rise of Maya Civilization. Middle American Research Institute, Publication 52. New Orleans, Tulane University. Demarest, A. A. 1989. The Olmec and the Rise of Civilization in Eastern Mesoamerica. In R. J. Sharer and D. C. Grove (eds.), Regional Perspectives on the Olmec: 303–344. Washington, D. C., National Gallery of Art. Demarest, A. A. 2002. Theoretical Speculations on the Rise of Complex Society on the South Coast of Guatemala. In M. Love, M. Popenoe de Hatch, and H. L. Escobedo (eds.), Incidents of Archaeology on Central America and Yucatán: Essays in Honor of Edwin M. Shook: 11–34. New York, University Press of America. Demarest, A. A., and Sharer, R. J. 1982. The Origins and Evolution of Usulutan Ceramics. American Antiquity 47(4): 810–822. Dixon, R. M. W. 1994. Ergativity. Cambridge, Cambridge University Press. Dixon, R. M. W. 1997. The Rise and Fall of Languages. Cambridge, Cambridge University Press. Englehardt, J. D. 2005. The Emergence of the Classic Mayan Writing System in the Late Preclassic and Early Classic. Unpublished MA thesis, Universidad Autónoma de Yucatán. Englehardt, J. D. 2010. Crossing the Usumacinta: Stylistic Variability and Dynamic Boundaries in the Preclassic and Classic Period Northwest Maya Lowlands. Archaeological Review from Cambridge 25(2): 57–76. Englehardt, J. D. 2011. Archaeological Epigraphy and Epigraphic Archaeology: Tracing Interaction, 192

Bibliography Golden, C. W. 2003. Sierra del Lacandón Regional Archaeology Project: First Field Season 2003. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Golden, C. W., and Scherer, A. 2006. Border Problems: Recent Archaeological Research along the Usumacinta River. The PARI Journal 7(2): 1–16. Golden, C. W., Barrientos, T. J., Hruby, Z. X., and Muñoz, A. R. 1998. La Pasadita: Nuevas Investigaciones en un Sitio Secundario en la Región del Usumacinta. In J. P. Laporte and H. L. Escobedo (eds.), XII Simposio de Investigaciones Arqueológicas en Guatemala, 1998: 435–454. Guatemala City, Museo Nacional de Arqueología y Etnología. Golden, C. W., Román, E., Muñoz, A. R., Scherer, A., and Romero, L. A. 2004. Reconocimiento y Patrones de Asentamiento en la Sierra del Lacandón, Petén. In J. P. Laporte, B. Arroyo, and H. E. Mejía (eds.), XVIII Simposio de Investigaciones Arqueológicas en Guatemala, 2004. Guatemala City, Museo Nacional de Arqueología y Etnología Goldstein, P. S. 2000. Exotic Goods and Everyday Chiefs: Long–Distance Exchange and Indigenous Sociopolitical Development in the South Central Andes. Latin American Antiquity 11(4): 335–361. González Lauck, R. B. 1993. Informe de Actividades del Proyecto Usumacinta Medio. Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. González Moreno, A. 2006. San Claudio, Un Asentamiento de las Tierras Bajas Noroccidentales del Área Maya, Tabasco, México: Sus Redes de Interacción Vistas a través de la Cerámica Arqueológica. Unpublished MA thesis, Escuela Nacional de Antropología e Historia. Graham, J. A. 1989. Olmec Diffusion: A Sculptural View from Pacific Guatemala. In R. J. Sharer and D. C. Grove (eds.), Regional Perspectives on the Olmec: 227–246. Washington, D. C., National Gallery of Art. Graham, J. A. 1990. Monumental Sculpture and Hieroglyphic Inscriptions. In G. R. Willey (ed.), Excavations at Seibal. Memoirs of the Peabody Museum of Archaeology and Ethnology, Harvard University, Vol. 17, No. 1. Cambridge. Green, M. W. 1989. Early Cuneiform. In W. M. Senner (ed.), The Origins of Writing: 43–58. Lincoln, University of Nebraska Press. Green, S. W., and Perlman, S. M. 1985. Frontiers, Boundaries, and Open Social Systems. In S. W. Green and S. M. Perlman (eds.), The Archaeology of Frontiers and Boundaries: 3–14. Orlando, Academic Press. Grove, D. C. 1970. The Olmec Paintings of Oxtotitlan Cave, Guerrero, Mexico. Studies in Pre–Colombian Art and Archaeology, No. 6. Washington, D.C., Dumbarton Oaks. Grove, D. C. 1989. Olmec: What’s in a Name? In R. J. Sharer and D. C. Grove (eds.), Regional Perspectives

Fournier, P. 1983. Proyecto de Periodificación de Asentamientos en la Región del Usumacinta Medio. Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. Fournier, P. 1987. La Cerámica de Yaxchilán: Informe Técnico Final. Manuscrito en el archivo del Proyecto Yaxchilán. Mexico City, Dirección de Monumentos Prehispánicos del Instituto Nacional de Antropología e Historia. Fox, J. A., and Justeson. J. S. 1980. Maya Hieroglyphs as Linguistic Evidence. In M. G. Robertson (ed.), Third Palenque Round Table, 1978: Part 2: 204–216. Austin, University of Texas Press. Freidel, D. A. 1979. Culture Areas and Interaction Spheres: Contrasting Approaches to the Emergence of Civilization in the Maya Lowlands. American Antiquity 44(1): 36–54. Freidel, D. A., and Reilly, F. K. III. 2010. The Flesh of God: Cosmology, Food, and the Origins of Political Power in Ancient Southeastern Mesoamerica. In J. E. Staller and M. D. Carrasco (eds.), Pre–Colombian Foodways: Interdisciplinary Approaches to Food, Culture, and Markets in Ancient Mesoamerica: 635– 680. New York, Springer. Freidel, D. A., and Suhler, C. 1999. Crown of Creation: The Development of the Maya Royal Diadems in the Late Preclassic and Early Classic Periods. In N. Grube (ed.), The Emergence of Lowland Maya Civilization: The Transition from the Preclassic to Early Classic, A Conference at Hildesheim, Germany, November 1992: 137–150. Acta Mesoamericana, No. 8, Verlag Anton Saurwein, Markt Schwaber. Galaty, M. L., Parkinson, W.A., Cherry, J. F., Cline, E. H., Kardulias, P. N., Schon, R., Sherratt, S., Tomas, H., and Wengrow, D. 2010. Interaction amidst Diversity: An Introduction to the Eastern Mediterranean Bronze Age. In W. A. Parkinson and M. L. Galaty (eds.), Archaic State Interaction: The Eastern Mediterranean in the Bronze Age: 29–51. Santa Fe, School for Advanced Research. García Moll, R. 2005. Pomoná: Un Sitio del Clásico Maya en las Colinas Tabasqueñas. Mexico City, Instituto Nacional de Antropología e Historia. Garraty, C. P. 2009. Evaluating the Distributional Approach to Inferring Marketplace Exchange: A Test Case from the Mexican Gulf Lowlands. Latin American Antiquity 20(1): 157–174. Gelb, I. J. 1963. A Study of Writing. Second Edition. Chicago, University of Chicago Press. Giddens, A. 1979. Central Problems in Social Theory: Action, Structure, and Contradiction in Social Analysis. Berkeley, University of California Press. Giddens, A. 1984. The Constitution of Society. Cambridge, Polity Press. Gifford, J. C. 1960. The Type–Variety Method of Ceramic Classification as an Indicator of Cultural Phenomena. American Antiquity 25(3): 341–347. 193

Archaeological Paleography on the Olmec: 8–14. Washington, D. C., National Gallery of Art. Grove, D. C. 2000. Faces of the Earth at Chalcatzingo, Mexico: Serpents, Caves, and Mountains in Middle Formative Period Iconography. In J. E. Clark and M. E. Pye (eds.), lmec Art and Archaeology in Mesoamerica: 277–295. Studies in the History of Art, No. 58. Center for Advanced Study in the Visual Arts Symposium Papers XXXV. Washington, D. C./ New Haven, National Gallery of Art/Yale University Press. Grube, N. 1994. Observations on the History of Mayan Hieroglyphs. In M. G. Robertson and V. M. Fields (eds.), Seventh Palenque Round Table, 1989: 177– 193. San Francisco, Pre–Colombian Art Research Institute. Grube, N. 1999. Transformations of Maya Society and the End of the Preclassic: Processes of Change between the Predynastic and Dynastic Periods. In N. Grube (ed.), The Emergence of Lowland Maya Civilization: The Transition from the Preclassic to Early Classic, A Conference at Hildesheim, Germany, November 1992: 1–5. Acta Mesoamericana, No. 8, Verlag Anton Saurwein, Markt Schwaber. Guernsey, J. 2006. Ritual and Power in Stone: The Performance of Rulership in Mesoamerican Izapan Style Art. Austin, University of Texas Press. Guernsey, J. 2012. Sculpture and Social Dynamics in Preclassic Mesoamerica. Cambridge, Cambridge University Press. Hall, T. D., and Chase–Dunn, C. 1996. Comparing World Systems: Concepts and Hypothesis. In P. N. Peregrine and G. M. Feinman (eds.), Pre–Columbian World Systems: 11–25. Monographs in World Archaeology, No. 26. Madison, Prehistory Press. Hanks, W. F. 1989a. Word and Image in a Semiotic Perspective. In W. F. Hanks and D. S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation: 8–21. Salt Lake City, University of Utah Press. Hanks, W. F. 1989b. Elements of Maya Style. In W. F. Hanks and D. S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation: 92–111. Salt Lake City, University of Utah Press. Hansen, R. D. 1991. An Early Maya Text from El Mirador, Guatemala. Research Reports on Ancient Maya Writing, No. 37. Washington, D. C., Center for Maya Research. Harris, R. 1986. The Origin of Writing. La Salle, Illinois, Open Court. Hegmon, M. 1989. The Styles of Integration: Ceramic Style and Pueblo I Integrative Architecture in Southwestern Colorado. In W. D. Lipe and M. Hegmon (eds.), The Architecture of Social Integration in Prehistoric Pueblos: 125–141. Occasional Papers of the Crow Canyon Archaeological Center, No.1. Cortez, Colorado, Crow Canyon Archaeological Center.

Hegmon, M. 1992. Archaeological Research on Style. Annual Review of Anthropology 21: 517–536. Hegmon, M. 1998. Technology, Style, and Social Practices: Archaeological Approaches. In M. T. Stark (ed.), The Archaeology of Social Boundaries: 264– 280. Washington, D. C., Smithsonian Institution. Henderson, J. S., and Joyce, R. A. 2006. Brewing Distinction: The Development of Cacao Beverages in Formative Mesoamerica. In C. L. McNeil (ed.), Chocolate in Mesoamerica: A Cultural History of Cacao: 140–153. Gainesville, University Press of Florida. Hernández Ayala, M. I. 1981. Cronología y Periodificación de la Región del Río San Pedro Mártir, Tabasco. Unpublished MA thesis, Escuela Nacional de Antropología e Historia. Hernández Pons, E. C. 1984. Investigaciones Arqueológicas en el Valle del Rio Tulija, Tabasco– Chiapas. Proyecto Tierras Bajas Noroccidentales, Vol. III. Mexico City, Universidad Nacional Autónoma de México. Hirth, K. G. 1998. The Distributional Approach: A New Way to Identify Marketplace Exchange in the Archaeological Record. Current Anthropology 39: 451–476. Hodder, I. 1978. Boundary Maintenance and Material Culture in the Baringo District, Kenya. Hikuin 5. Copenhagen, Aarhus. Hodder, I. 1979. Economic and Social Stress and Material Culture Patterning. American Antiquity 44: 446–454. Hodder, I. 1982. Symbols in Action: Ethnoarchaeological Studies of Material Culture. Cambridge, Cambridge University Press. Hodder, I. 1985. Boundaries as Strategies: An Ethnoarchaeological Study. In S. W. Green and S. M. Perlman (eds.), The Archaeology of Frontiers and Boundaries: 141–162. Orlando, Academic Press. Hodder, I. (ed.) 1987. The Archaeology of Contextual Meanings. Cambridge, Cambridge University Press. Hodder, I. 1992. Theory and Practice in Archaeology. New York, Routledge. Holley, G. 1983. Ceramic Change at Piedras Negras, Guatemala. Unpublished PhD dissertation, Southern Illinois University. Holley, G. 1987. Living on the Edge: The Ceramic Sequence at Piedras Negras, Guatemala. In P. M. Rice and R. J. Sharer (eds.), Maya Ceramics: Papers from the 1985 Maya Ceramic Conference: 183–202. British Archaeological Reports International Series 345(i). Oxford, Archaeopress. Hopkins, N. A. 1997. Decipherment and the Relationship between Mayan Languages and Maya Writing. In M. J. Macri and A. Ford (eds.), The Language of Mayan Hieroglyphs: 77–88. San Francisco, Pre–Colombian Art Research Institute. Houston, S. D. 2004a. Writing in Early Mesoamerica. In S. D. Houston (ed.), The First Writing: Script 194

Bibliography Invention as History and Process: 274–312. Cambridge, Cambridge University Press. Houston, S. D. 2004b. Overture to The First Writing. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 3–15. Cambridge, Cambridge University Press. Houston, S. D. 2004c. The Archaeology of Communication Technologies. Annual Review of Anthropology 33(1): 223–250. Houston, S. D., and Coe, M. D. 2003. Has Isthmian Writing Been Deciphered? Mexicon XXV(6): 151– 161. Houston, S. D., Robertson, J., and Stuart, D. 2000. The Language of Classic Maya Inscriptions. Current Anthropology 41: 321–356. Houston S. D., Stuart, D., and Robertson, J. 2004. Disharmony in Maya Hieroglyphic Writing: Linguistic Change and Continuity in Classic Society. In S. Wichmann (ed.), The Linguistics of Maya Writing: 83–99. Salt Lake City, University of Utah Press. Hruby, Z. X., and Child, M. B. 2004. Chontal Linguistic Influence in Ancient Maya Writing: Intransitive Positional Verbal Affixation. I In S. Wichmann (ed.), The Linguistics of Maya Writing: 13–26. Salt Lake City, University of Utah Press. Hurst, W. J. 2006. The Determination of Cacao in Samples of Archaeological Interest. In C. L. McNeil (ed.), Chocolate in Mesoamerica: A Cultural History of Cacao: 105–113. Gainesville, University Press of Florida. Hurt, T. D., and Rakita, G. F. M. 2001. Style and Function; Conceptual Issues in Evolutionary Archaeology. Westport, Connecticut, Bergin and Garvey. Jakobson, R. 1971a. On the Relation between Visual and Auditory Sign. In R. Jakobson, Selected Writings II: 338–344. The Hague, Mouton. Jakobson, R. 1971b. Visual and Auditory Sign. In R. Jakobson, Selected Writings II: 334–337. The Hague, Mouton. Joralemon, P. D. 1971. A Study of Olmec Iconography. Studies in Pre–Colombian Art and Archaeology, No. 7. Washington, D.C., Dumbarton Oaks. Josserand, J. K. 1975. Archaeological and Linguistic Correlations for Mayan Prehistory. In Actas del XL Congreso Internacional de Americanistas: 501–510. Mexico City, Instituto Nacional de Antropología e Historia. Josserand, J. K., and Hopkins, N. A. 1999. Classic Maya Social Interaction and Linguistic Practice: Evidence from Hieroglyphic Inscriptions and Mayan Languages. Paper presented at the 1999 Palenque Round Table. Joyce, R. A., Edging, R., Lorenz, K., and Gillespie, S. D. 1991. Olmec Bloodletting: An Iconographic Study. In M. G. Robertson and V. M. Fields (eds.), Sixth Palenque Roundtable, 1986: 143–150. Norman, University of Oklahoma Press.

Juárez Cossío, D. 1992. Informe de la primera temporada de campo del proyecto Reforma, Tabasco. Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. Juárez Cossío, D. 1994. Informe de la segunda temporada de campo en Morales–Reforma, Tabasco. Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. Juárez Cossío, D. 1999. Informe preliminar sobre las exploraciones en el sitio de Morales–Reforma, municipio de Balancán, Tabasco. Antropológicas, No. 15. Mexico City, Universidad Nacional Autónoma de Mexico. Juárez Cossío, D. 2003. Moral–Reforma: En la Senda de Xibalbá. Arqueología Mexicana 61: 38–43. Justeson, J. S. 1976. Universals of Language and Universals of Writing. In A. Julliand, A. M. Devine, and L. D. Stephens (eds.), Linguistic Studies Offered to Joseph Greenberg on the Occasion of his Sixtieth Birthday, Vol. 1: 57–94. Saratoga, California, Anma Libri. Justeson, J. S. 1985. Hieroglyphic Evidence for Lowland Maya Linguistic History. International Journal of American Linguistics 51(4): 429–471. Justeson, J. S. 1986. The Origin of Writing Systems: Preclassic Mesoamerica. World Archaeology 17: 437–458. Justeson, J. S. 1989. The Representational Conventions of Mayan Hieroglyphic Writing. In W. F. Hanks and D. S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation: 25–38. Salt Lake City, University of Utah Press. Justeson, J. S. 2012. Early Mesoamerican Writing Systems. In D. L. Nichols (ed.), The Oxford Handbook of Mesoamerican Archaeology: 830–844. Oxford, Oxford University Press. Justeson, J. S., and Campbell, L. R. 1997. The Linguistic Background of Maya Hieroglyphic Writing: Arguments against a “Highland Maya” Role. In M. J. Macri and A. Ford (eds.), The Language of Mayan Hieroglyphs: 41–68. San Francisco, Pre–Colombian Art Research Institute. Justeson, J. S., and Kaufman, T. S. 1993. A Decipherment of Epi–Olmec Hieroglyphic Writing. Science 129: 1703–1711. Justeson, J. S., and Kaufman, T. S. 1997. A Newly Discovered Column in the Hieroglyphic Text on La Mojarra Stela 1: A Test of the Epi–Olmec Decipherment. Science 277: 207–210. Justeson, J. S., and Matthews, P. L. 1990. Evolutionary Trends in Mesoamerican Hieroglyphic Writing. Visible Language XXIV(1): 88–132. Justeson, J. S., Norman, W. M., Campbell, L. R., and Kaufman, T. S. 1985. The Foreign Impact on Lowland Mayan Language and Script. Middle American Research Institute Publication 53. New Orleans, Tulane University. 195

Archaeological Paleography Justeson, J. S., Norman, W. M., and Hammond, N. 1988. The Pomona Flare: A Preclassic Maya Hieroglyphic Text. In E. P. Benson and G. G. Griffin (eds.), Maya Iconography: 95–151. Princeton, Princeton University Press. Justeson, J. S., and Stephens, L. D. 1993. The Evolution of Syllabaries from Alphabets: Transmission, Language, Contrast, and Script Typology. Die Sprach 35: 1–46. Kaufman, T. S. 1976. Archaeological and Linguistic Correlations in Mayaland and Associated Areas of Mesoamerica. World Archaeology 8(1): 101–118. Kaufman, T. S., and Justeson, J. S. 2001. Epi–Olmec Hieroglyphic Writing and Texts. Electronic document, http://www.albany.edu/ pdlma/EOTEXTS.pdf, accessed August 30, 2008. Kaufman, T. S., and Norman, W. M. 1984. An Outline of Proto–Cholan Phonology, Morphology, and Vocabulary. In J. S. Justeson and L. R. Campbell (eds.), Phoneticism in Mayan Hieroglyphic Writing: 77–166. Institute for Mesoamerican Studies, Publication 9. Albany, State University of New York. Keightly, D. N. 1989. The Origins of Writing in China: Scripts and Cultural Contexts. In W. M. Senner (ed.), The Origins of Writing: 171–202. Lincoln, University of Nebraska Press. Kintigh, K. W. 1984. Measuring Archaeological Diversity by Comparison with Simulated Assemblages. American Antiquity 49(1): 44–54. Kintigh, K. W. 1989. Sample Size, Significance, and Measures of Diversity. In R. D. Leonard and G. T. Jones (eds.), Quantifying Diversity in Archaeology: 25–36. Cambridge, Cambridge University Press. Kintigh, K. W. 2002. Tools for Quantitative Archaeology: Programs for Quantitative Analysis in Archaeology. Programs manual published by the author. Electronic document, http:// tfqa.com, accessed June 29, 2010. Knappett, C. A. 2008. The Neglected Networks of Material Agency: Artefacts, Pictures, and Texts. In C. A. Knappett and L. Malafouris (eds.), Material Agency: Towards a Non–Anthropocentric Approach: 139–156. Berlin, Springer. Kosakowsky, L. J. 1987. The Formative Ceramic Sequence of Cuello, Belize. I In P. M. Rice and R. J. Sharer (eds.), Maya Ceramics: Papers from the 1985 Maya Ceramic Conference: 15–36. British Archaeological Reports International Series 345(i). Oxford, Archaeopress. Kowalewski, S. A., Blanton, R. E., Feinman, G. M., and Finsten, L. 1983. Boundaries, Scale, and Internal Organization. Journal of Anthropological Archaeology 2: 32–56. Lacadena, A. 1995. Evolución Formal de las Grafías Escriturarias Mayas: Implicaciones Históricas y Culturales. Unpublished PhD dissertation, Universidad Complutense de Madrid. Lacadena, A. 2008. La Escritura Olmeca y la Hipótesis del Mixe–Zoque: Implicaciones Lingüísticas de un

Análisis Estructural del Monumento 13 de La Venta. In M. T. Uriarte and R. González Lauck (eds.), Olmeca. Balance y Perspectivas: 607–626. Mexico City, Instituto Nacional de Antropología e Historia. Lacadena, A. 2010. Historical Implications of the Presence of Non–Mayan Linguistic Features in Maya Script. In L. van Broekhoven, R. Valencia Rivera, B. Vis, and F. Sachse (eds.), The Maya and Their Neighbors: Internal and External Contacts Through Time. Proceedings of the 10th European Maya Conference, Leiden, December 2005: 29–39. Acta Mesoamericana, No. 23, Verlag Anton Saurwein, Markt Schwaber. Lacadena, A., and Wichmann, S. 1999. The Distribution of Lowland Maya Languages in the Classic Period. Paper presented at the 1999 Palenque Round Table. Lamberg–Karlovsky, C. C. 1986. The Emergence of Writing: Mesopotamia, Egypt, and the Indus Civilizations. In E. Wyllys Andrews V (ed.), Research and Reflections in Archaeology and History: Essays in Honor of Doris Stone: 149–158. Middle American Research Institute, Publication 57. New Orleans, Tulane University. Lansing, J. S. 2003. Complex Adaptive Systems. Annual Review of Anthropology 32: 183–204. Lee, T. A., Jr. 1969. The Artifacts of Chiapa de Corzo, Chiapas, Mexico. Papers of the New World Archaeological Foundation, No. 26. Provo, Utah, Brigham Young University. Lee, T. A., Jr. 1972. The Middle Grijalva Regional Chronology and Ceramic Relations: A Preliminary Report. In N. Hammond (ed.), Mesoamerican Archaeology: New Approaches: 1–20. Austin, University of Texas Press. Lee, T. A., Jr. 1978. The Historical Routes of Tabasco and Northern Chiapas and their Relationship to Early Cultural Developments in Central Chiapas. In T. A. Lee, Jr. and C. Navarrete (eds.), Mesoamerican Communication Routes and Cultural Contacts: 49–66. Papers of the New World Archaeological Foundation, No. 40. Provo, Utah, Brigham Young University. Lee, T. A., Jr. 1989. Chiapas and the Olmec. In R. J. Sharer and D. C. Grove (eds.), Regional Perspectives on the Olmec: 198–226. Washington, D.C., National Gallery of Art. Lee, T. A., Jr., and Cheetham, D. 2008. Lengua y Escritura Olmeca. In M. T. Uriarte and R. González Lauck (eds.), Olmeca. Balance y Perspectivas: 695– 713. Mexico City, Instituto Nacional de Antropología e Historia. Leonard, R. D., and G. T. Jones (eds). 1989. Quantifying Diversity in Archaeology. Cambridge, Cambridge University Press. Lesure, R. G. 2004. Shared Art Styles and Long-Distance Contact in Early Mesoamerica. In J. A. Hendon and R. A. Joyce (eds.), Mesoamerican Archaeology: Theory and Practice: 73–96. Malden, Blackwell. 196

Bibliography Prehispánicos, Coloniales, y Modernos: Memoria de la Tercera Mesa Redonda de Palenque, Volume I: 50–78. Mexico City/ Mérida , Instituto Nacional de Antropología e Historia/Universidad Autónoma de Yucatán. Martin, S. 2003. Moral–Reforma y la Contienda por el Oriente de Tabasco. Arqueología Mexicana 61: 44–47. Matheny, R. T. 1970. The Ceramics of Aguacatal, Campeche, Mexico. Papers of the New World Archaeological Foundation, No. 27. Provo, Utah, Brigham Young University. Méluzin, S. 1992. Tuxtla Script: Steps toward Decipherment Based on La Mojarra Stela 1. Latin American Antiquity 3(4): 283–297. Méluzin, S. 1995. Further Investigations of the Tuxtla Script: An Inscribed Mask and La Mojarra Stela 1. Papers of the New World Archaeological Foundation, No. 65. Provo, Utah, Brigham Young University. Miller, A. G. 1973. The Mural Painting of Teotihuacan. Washington, D. C., Dumbarton Oaks. Miller, A. G. 1989. Comparing Maya Image and Text. In W. F. Hanks and D. S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation: 176–188. Salt Lake City, University of Utah Press. Mora–Marín, D. 1997. The Origins of Maya Writing: The Case for Portable Objects. In T. Jones and C. Jones (eds.), U Mut Maya VII: 133–164. Arcata, Humboldt State University. Mora–Marín, D. 2000. Late Preclassic Inscription Documentation Project. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Mora–Marín, D. 2001. The Grammar, Orthography, Content, and Social Context of Late Preclassic Portable Texts. PhD dissertation. State University of New York. Ann Arbor, University Microfilms. Mora–Marín, D. 2005. Kaminaljuyú Stela 10: Script Classification and Linguistic Affiliation. Ancient Mesoamerica 16(1): 63–87. Mora–Marín, D. 2011. Further Analysis of Olmec Writing on the Cascajal Block: Sign Inventory, Paleography, Script Affiliations. Unpublished manuscript in possession of the author. Muñoz, A. R. 2002. Ceramics at Piedras Negras, Guatemala. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Muñoz, A. R. 2004. The Ceramic Sequence of Piedras Negras, Guatemala: Type and Varieties. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Navarrete, C. 1978. The Pre–Hispanic System of Communications between Chiapas and Tabasco. In T. A. Lee, Jr. and C. Navarrete (eds.), Mesoamerican Communication Routes and Cultural Contacts: 75–106. Papers of the New World Archaeological Foundation, No. 40. Provo, Utah, Brigham Young University.

Levin, S. A. 1998. Ecosystems and the Biosphere as Complex Adaptive Systems. Ecosystems 1: 431–436. Levin, S. A. 2002. Complex Adaptive Systems: Exploring the Known, the Unknown, and the Unknowable. Bulletin of the American Mathematical Society 40(1): 3–19. Liendo Stuardo, R. 2005. An Archaeological Study of Settlement Distribution in the Palenque Area, Chiapas, Mexico. Anthropological Notebooks 11: 31–44. Liendo Stuardo, R. 2007. Proyecto Arqueológico Chinikihá, Temporada 2006, Informe de Actividades. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. López Varela, S. 1991. La Cerámica de Pomoná, Tabasco: Un Ejemplar del Clásico Maya en las Tierras Bajas Noroccidentales. Puebla, Universidad de las Américas. Lowe, G. W. 1977. The Mixe–Zoque as Competing Neighbors of the Early Lowland Maya. In R. E. W. Adams (ed.), The Origins of Maya Civilization: 197– 248. Albuquerque, University of New Mexico Press. Lowe, G. W. 1989. Algunas Aclaraciones sobre la Presencia Olmeca y Maya en el Preclásico de Chiapas. In M. Carmona (ed.), El Preclásico o Formativo, Avances y Perspectivas: 363–385. Mexico City, Museo Nacional de Antropología–Instituto Nacional de Antropología e Historia. McGlade, J. 2003. The Map is not the Territory: Complexity, Complication, and Representation. In R. A. Bentley and H. D. G. Maschner (eds.), Complex Systems and Archaeology: 111–119. Salt Lake City, University of Utah Press. Macri, M. J. 1991. The Script on La Mojarra Stela 1 and Classic Maya Writing. In D. L. Schmidt and J. S. Smith (eds.), Literacies: Writing Systems and Literate Practices: 11–23. Davis Working Papers in Linguistics, No. 4. Davis, University of California. Manson, S. M., and O’Sullivan, D. 2006. Complexity Theory in the Study of Space and Place. Environment and Planning 38(4): 677–692. Marcus, J. 1976. The Origins of Mesoamerican Writing. Annual Review of Anthropology 5: 35–67. Marcus, J. 1992a. Mesoamerican Writing Systems: Propaganda, Myth, and History in Four Ancient Civilizations. Princeton, Princeton University Press. Marcus, J. 1992b. Dynamic Cycles of Mesoamerican States: Political Fluctuations in Mesoamerica. National Geographic Research and Exploration 8: 392–411. Marcus, J. 1998. The Peaks and Valleys of Ancient States: An Extension of the Dynamic Model. In G. M. Feinman and J. Marcus (eds.), Archaic States: 59–94. Santa Fe, School for Advanced Research. Martin, S. 2002. The Baby Jaguar: An Exploration of its Identity and Origins in Maya Art and Writing. In V. Tiesler Blos, R. Cobos, and M. G. Robertson (eds.), La Organización Social entre los Mayas 197

Archaeological Paleography (ed.), The Archaeology of Tribal Societies: 391–438. Ann Arbor, International Monographs in Prehistory. Parkinson, W. A. 2006. Tribal Boundaries: Stylistic Variability and Social Boundary Maintenance during the Transition to the Copper Age on the Great Hungarian Plain. Journal of Anthropological Archaeology 25: 33–58. Parkinson, W. A., and Galaty, M. L. 2007. Secondary States in Perspective: An Integrative Model of State Formation in the Prehistoric Aegean. American Anthropologist 109(1): 113–129. Parkinson, W. A., and Galaty, M. L. 2010. Introduction: Interaction and Ancient Societies. In W. A. Parkinson and M. L. Galaty (eds.), Archaic State Interaction: The Eastern Mediterranean in the Bronze Age: 3–28. Santa Fe, School for Advanced Research. Parsons, L. A. 1979. Post–Olmec Stone Sculpture: The Olmec–Izapan Transition on the Southern Pacific Coast and Highlands. In E. P. Benson (ed.), The Olmec and Their Neighbors: Essays in Memory of Matthew W. Stirling: 257–288. Washington, D. C., Dumbarton Oaks. Parsons, L. A. 1986. The Origins of Maya Art: Monumental Stone Sculpture of Kaminaljuyú, Guatemala and the Southern Pacific Coast. Studies in Pre–Colombian Art and Archaeology, No. 28. Washington, D. C., Dumbarton Oaks. Parsons, L. A. 1988. Proto–Maya Aspects of Miraflores– Arenal Monumental Stone Sculptures from Kaminaljuyu and the Southern Pacific Coast. In E. P. Benson and G. G. Griffin (eds.), Maya Iconography: 6–43. Princeton, Princeton University Press. Peirce, C. S. 1931–1966. Collected Papers of Charles Sanders Peirce, C. Hartshorne, P. Weiss, and A.W. Burks (eds.). 8 vols. Cambridge, Harvard University Press. Perales Vela, R. 2002. Una Aproximación al Patrón de Asentamiento en la Cuenca de San Pedro Mártir, Tabasco. Unpublished BA thesis, Universidad Veracruzana. Perales Vela, R., and Mugarte, J. 1995. Arqueología de Superficie en Santa Elena, Tabasco. In E. Vargas Pacheco (ed.), Seis Ensayos sobre Antiguos Patrones de Asentamiento en el Área Maya: 27–52. Mexico City, Instituto de Investigaciones Antropológicas, Universidad Nacional Autónoma de México. Pérez Robles, G. 2006. La Secuencia Cerámica del Período Preclásico en Piedras Negras, Petén. Tesis de Licenciatura (BA thesis), Universidad de San Carlos de Guatemala. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Peterson, C., and Drennan, R. 2003. Communities, Settlements, Sites, and Surveys: Regional Scale Analysis of Prehistoric Human Interaction. American Antiquity 70(1): 5–30. Peterson, F. A. 1963. Some Ceramics from Mirador, Chiapas, Mexico. Papers of the New World

Neff, H., Cogswell, J. W., Kosakowsky, L. J., Estrada Belli, F., and Bove, F. J. 1999. A New Perspective on the Relationships among Cream Paste Ceramic Traditions of Southeastern Mesoamerica. Latin American Antiquity 10(3): 281–299. Neitzel, J. 2000. What is a Regional System? Issues of Scale and Interaction in the Prehistoric Southwest. In M. Hegmon (ed.), The Archaeology of Regional Interaction: Religion, Warfare, and Exchange Across the American Southwest and Beyond: 25–40. Boulder, University Press of Colorado. O’Shea, J. M., and Milner, C. H. 2002. Material Indicators of Territory, Identity, and Interaction in a Prehistoric Tribal System. In W. A. Parkinson (ed.), The Archaeology of Tribal Societies: 200–226. Ann Arbor, International Monographs in Prehistory. O’Sullivan, D., Manson, S. M., Messina, J. P., and Crawford, T. W. 2006. Space, Place, and Complexity Science. Environment and Planning 38(4): 611–617. Ochoa, L. 1977. Notas Preliminares sobre el Proyecto Arqueología de las Tierras Bajas Noroccidentales del Área Maya. Estudios de Cultura Maya X: 41–52. Ochoa, L. 1978. Estudios Preliminares Sobre los Mayas de las Tierras Bajas Noroccidentales del Área Maya. Mexico City, Instituto de Investigaciones Filológicas, Centro de Estudios Mayas, Universidad Nacional Autónoma de México. Ochoa, L. 1983. El Medio Usumacinta: Un Eslabón en los Antecedentes Olmecas de los Mayas. In L. Ochoa and T. A. Lee, Jr. (eds.), Antropología e Historia de los Mixe–Zoques y Mayas: Homenaje a Frans Blom: 147–174. Mexico City, Instituto de Investigaciones Filológicas, Universidad Nacional Autónoma de México/ Brigham Young University. Ochoa, L. 2008. Una Aproximación a la Historia del Origen Lingüístico de los Huaxtecos o Teenek. Manuscript on file at the Instituto de Investigaciones Antropológicas, Universidad Nacional Autónoma de México. Ochoa, L., and Casasola, L. 1983. Excavaciones Arqueológicas en Tierra Blanca, Tabasco. Mexico City, Archivo del Centro de Estudios Mayas, Universidad Nacional Autónoma de México. Ochoa, L., and Casasola, L. 1991. Tierra Blanca y el Medio Usumacinta: Notas de su Cerámica Arqueológica. Tierra y Agua: La Antropología en Tabasco 2: 7–28. Ochoa, L., and Hernández, M. I. 1975. Informe Preliminar de los Trabajos Arqueológicos en las Tierras Bajas del Área Maya. Mexico City, Instituto de Investigaciones Filológicas, Centro de Estudios Mayas, Universidad Nacional Autónoma de México. Orton, C., Tyers, P., and Vince, A. 1993. Pottery in Archaeology. Cambridge, Cambridge University Press. Parkinson, W. A. 2002. Integration, Interaction, and Tribal “Cycling”: The Transition to the Copper Age on the Great Hungarian Plain. In W. A. Parkinson 198

Bibliography Archaeological Foundation, No. 15. Provo, Utah, Brigham Young University. Pickel, A. 2007. Rethinking Systems Theory: A Programmatic Introduction. Philosophy of the Social Sciences 37: 391–407. Piña Chan, R., and Navarrete, C. 1967. Archaeological Research in the Lower Grijalva River Region, Tabasco and Chiapas. Papers of the New World Archaeological Foundation, No. 22. Provo, Utah, Brigham Young University. Plog, S. 1995. Approaches to Style: Complements and Contrasts. In C. Carr and J. Neitzel (eds.), Style, Society, and Person: 369–392. New York, Plenum. Pohl, M. E. D., Pope, K. O., and von Nagy, C. 2002. Olmec Origins of Mesoamerican Writing. Science 298: 1984–1987. Pohl, Mary E. D., Josserand, J. K., Pope, K. O., and von Nagy, C. 2008. La U Olmeca y el Desarrollo de la Escritura en Mesoamérica. In M. T. Uriarte and R. González Lauck (eds.), Olmeca. Balance y Perspectivas: 685–694. Mexico City, Instituto Nacional de Antropología e Historia. Portugali, J. 2006. Complexity Theory as a Link between Space and Place. Environment and Planning 38(4): 647–664. Possehl, G. L. 2007. The Middle Asian Interaction Sphere: Trade and Contact in the 3rd Millennium BC. Expedition 49(1): 40–42. Price, B. J. 1978. Commerce and Cultural Process in Mesoamerica. In T. A. Lee, Jr. and C. Navarrete (eds.), Mesoamerican Communication Routes and Cultural Contacts: 231–245. Papers of the New World Archaeological Foundation, No. 40. Provo, Utah, Brigham Young University. Pring, D. C. 1977. Influence or Intrusion? The “Protoclassic” in the Maya Lowlands. In N. Hammond (ed.), Social Process in Maya Prehistory: Studies in Honor of Sir Eric Thompson: 135–165. London, Academic Press. Prouskouriakoff, T. 1968. Olmec and Maya Art: Problems of their Stylistic Relation. In E. P. Benson (ed.), Dumbarton Oaks Conference on the Olmec: 119–134. Washington, D. C., Dumbarton Oaks. Prouskouriakoff, T. 1971. El Arte Maya y el Modelo Genético de Cultura. In E. Z. Vogt and A. Ruz Lhuillier (eds.), Desarrollo Cultural de los Mayas: 187–202. Mexico City, Centro de Estudios Mayas, Universidad Nacional Autónoma de México. Quirarte, J. 1977. Early Art Styles of Mesoamerica and Early Classic Maya Art. In R. E. W. Adams (ed.), The Origins of Maya Civilization: 249–283. Albuquerque, University of New Mexico Press. Quirarte, J. 2007. Revisiting the Relationship between Izapa, Olmec, and Maya Art. In L. S. Lowe and M. E. Pye (eds.), Archaeology, Art, and Ethnogenesis in Mesoamerican Prehistory: Papers in Honor of Gareth W. Lowe: 47–276. Papers of the New World

Archaeological Foundation, No. 68. Provo, Utah, Brigham Young University. Railsback, S. F. 2001. Concepts from Complex Adaptive Systems as a Framework for Individual–Based Modeling. Ecological Modeling 139: 47–62. Rands, R. L. 1972. The Ceramic Sequence at Palenque, Chiapas. In N. Hammond (ed.), Mesoamerican Archaeology: New Approaches: 51–75. Austin, University of Texas Press. Rands, R. L. 1977. The Rise of Classic Maya Civilization in the Northwestern Zone: Isolation and Integration. In R. E. W. Adams (ed.), The Origins of Maya Civilization: 159–180. Albuquerque, University of New Mexico Press. Rands, R. L. 1987. Ceramic Patterns and Traditions in the Palenque Area. In P. M. Rice and R. J. Sharer (eds.), Maya Ceramics: Papers from the 1985 Maya Ceramic Conference: 203–238. British Archaeological Reports International Series 345(i). Oxford, Archaeopress. Rands, R. L. 2002. Palenque and Selected Survey Sites in Chiapas and Tabasco: The Preclassic. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Rands, R. L. 2007a. Chronological Chart and Overview of Ceramic Developments at Palenque. In D. B. Marken (ed.), Palenque: Recent Investigations at the Classic Maya Center: 17–24. Lanham, Maryland, Altamira Press. Rands, R. L. 2007b. Palenque and Selected Survey Sites in Chiapas and Tabasco: The Preclassic. In D. B. Marken (ed.), Palenque: Recent Investigations at the Classic Maya Center: 25–56. Lanham, Maryland, Altamira Press. Rands, R. L., and Rands, B. C. 1957. The Ceramic Position of Palenque, Chiapas. American Antiquity 23(2): 140–150. Rathje, W. L., Gregory, D. A., and Wiseman, F. M. 1978. Trade Models and Archaeological Problems: Classic Maya Examples. In T. A. Lee, Jr. and C. Navarrete (eds.), Mesoamerican Communication Routes and Cultural Contacts: 147–175. Papers of the New World Archaeological Foundation, No. 40. Provo, Utah, Brigham Young University. Reilly, F. Kent III. 1991. Olmec Iconographic Influences on the Symbols of Maya Rulership: An Examination of Possible Sources. In M. G. Robertson and V. M. Fields (eds.), Sixth Palenque Roundtable, 1986: 151– 166. Norman, University of Oklahoma Press. Reilly, F. Kent III. 1995. Art, Ritual, and Rulership in the Olmec World. In E. P. Benson and B. de la Fuente (eds.), The Olmec World: Ritual and Rulership: 27–47. Princeton, Princeton University Art Museum. Reilly, F. Kent III. 1996. The Lazy–S: A Formative Period Iconographic Loan to Maya Hieroglyphic Writing. In M. G. Robertson, M. J. Macri, and J. McHargue (eds.), 8th Palenque Round Table, 1993, 199

Archaeological Paleography Vol. X: 413–424. San Francisco, Pre–Colombian Art Research Institute. Renfrew, C. 1972. The Emergence of Civilisation; the Cyclades and the Aegean in the Third Millennium B. C. London, Methuen. Renfrew, C. 1987. Archaeology and Language: The Puzzle of Indo–European Origins. Cambridge, Cambridge University Press. Renfrew, C. 2001. Symbol before Concept: Material Engagement and the Early Development of Society. In I. Hodder (ed.), Archaeological Theory Today: 122–140. Oxford, Polity Press. Rice, P. M. 1987. Pottery Analysis: A Sourcebook. Chicago, University of Chicago Press. Rice, P. M. 1996a. Recent Ceramic Analysis: 1. Function, Style, and Origins. Journal of Archaeological Research 4(2): 133–163. Rice, P. M. 1996b. Recent Ceramic Analysis: 2. Composition, Production, and Theory. Journal of Archaeological Research 4(3): 165–202. Rice, P. M. 2007. Maya Calendar Origins: Monuments, Mythistory, and the Materialization of Time. Austin, University of Texas Press. Riese, B. 1988. Epigraphy of the Southeast Zone in Relation to Other Parts of Mesoamerica. In E. H. Boone and G. R. Willey (eds.), The Southeast Classic Maya Zone: 67–94. Washington, D. C., Dumbarton Oaks. Robertson, J. S. 2004. The Possibility and Actuality of Writing. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 16–38. Cambridge, Cambridge University Press. Robinson, A. 2003. The Origins of Writing. In D. Crowley and P. Heyer (eds.), Communication in History: Technology, Culture, Society. New York, Allyn and Bacon. Robinson, G. 1951. A Method of Chronologically Ordering Archaeological Deposits. American Antiquity 16(3): 293–301. Robles Castellanos, F. 1990. La Secuencia Cerámica de la Región de Coba, Quintana Roo. Colección Científica No. 184. Mexico City, Instituto Nacional de Antropología e Historia. Rodríguez Martínez, M. del C., Ortíz Ceballos, P., Coe, M. D., Diehl, R. A., Houston, S. D., Taube, K. A., and Delgado Calderón, A. 2006. Oldest Writing in the New World. Science 313: 1610–1614. Rogers, H. 2005. Writing Systems: A Linguistic Approach. Oxford, Blackwell. Román, E., Arroyave, A. L., Meléndez, J. C., Pérez, G., Quiroa, F., and Eppich, K. 2004. La Secuencia de la Tradición Cerámica de El Perú–Waka’, Petén. In J. P. Laporte, B. Arroyo, and H. E. Mejía (eds.), XVIII Simposio de Investigaciones Arqueológicas en Guatemala, 2004. Guatemala City, Museo Nacional de Arqueología y Etnología. Romero Rivera, J. L. 1997. Proyecto Arqueológico San Claudio. Archivo del Centro INAH Tabasco.

Villahermosa, Instituto Nacional de Antropología e Historia. Romero Rivera, J. L. 1998. Proyecto arqueológico de San Claudio: Informe de actividades realizadas durante la primera temporada de campo 1997. Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. Romero Rivera, J. L. 2000. Proyecto arqueológico de San Claudio: Informe de actividades realizadas durante la segunda temporada de campo. Archivo del Centro INAH Tabasco. Villahermosa, Instituto Nacional de Antropología e Historia. Rosenswig, R. M. 2010. The Beginnings of Mesoamerican Civilization: Inter-Regional Interaction and the Olmec. Cambridge, Cambridge University Press. Rye, O. S. 1981. Pottery Technology: Principles and Reconstruction. Washington, D.C., Taraxacum. Sabloff, J. A. 1970. Type Descriptions of the Fine Paste Ceramics of the Bayal Boca Complex, Seibal, Petén Guatemala. In W. R. Bullard (ed.), Monographs and Papers in Maya Archaeology, Vol. 61, Part IV, No. 2: 357–404. Cambridge, Peabody Museum, Harvard University. Sabloff, J. A. 1975. Excavations at Seibal: Ceramics. Memoirs of the Peabody Museum of Archaeology and Ethnology, Vol. 13, No. 2. Cambridge, Harvard University. Sabloff, J. A. 1987. Interaction among Classic Maya Polities: A Preliminary Examination. In C. Renfrew and J. F. Cherry (eds.), Peer Polity Interaction and Socio Political Change: 109–116. Cambridge, Cambridge University Press. Sabloff, J. A., and Smith, R. E. 1969. The Importance of both Analytic and Taxonomic Classification in the Type–Variety System. American Antiquity 34(3): 278–285. Sabloff, J. A., and Smith, R. E. 1972. Ceramic Wares in the Maya Area: A Clarification of an Aspect of the Type–Variety System and Presentation of a Formal Model for Comparative Use. Estudios de Cultura Maya VIII: 97–115. Salomon, F. 2001. How an Andean “Writing Without Words” Works. Current Anthropology 42(1): 1–27. Sampson, G. 1985. Writing Systems: A Linguistic Introduction. Stanford, Stanford University Press. San Román Martín, M. E. 2009. Palenque’s Ceramics: Searching for a Methodology for their Study and Classification. Crystal River, Florida,, Foundation for the Advancement of Mesoamerican Studies, Inc. Sánchez Caero, O. F. 1979. Excavaciones Arqueológicas en la Zona de Jonuta, Tabasco. Unpublished BA thesis, Escuela Nacional de Antropología e Historia. Saturno, W. A., Taube, K. A., and Stuart, D. 2005. The Murals of San Bartolo, Guatemala, Part I: The North Wall. Ancient America No. 7. Barnardsville, North Carolina, Center for Ancient American Studies. 200

Bibliography Guatemala: Interaction and Development of Maya Civilization. Philadelphia, University of Pennsylvania Museum of Archaeology and Anthropology. Shennan, S. 1988. Quantifying Archaeology. Edinburgh, Edinburgh University Press. Shepard, A. O. 1965. Ceramics for the Archaeologist. Washington, D.C., Carnegie Institution. Sinopoli, C. M. 1991. Approaches to Archaeological Ceramics. New York, Plenum. Skibo, J. M., Schiffer, M. B., and Kowalski, N. 1989. Ceramic Style Analysis in Archaeology and Ethnoarchaeology: Bridging the Analytical Gap. Journal of Anthropological Archaeology 8: 388–409. Smith, R. E. 1955. Ceramic Sequence at Uaxactun, Guatemala. Middle American Research Institute, Publication 20. New Orleans, Tulane University. Smith, R. E., and Gifford, J. C. 1966. Maya Ceramic Varieties, Types, and Wares at Uaxactun: Supplement to “Ceramic Sequence at Uaxactun, Guatemala.” Middle American Research Institute, Publication 28. New Orleans, Tulane University. Smith, R. E., Willey, G. R., and Gifford, J. C. 1960. Type–Variety Analysis of Maya Pottery. American Antiquity 23: 330–340. Smith, V. G. 1984. Izapa Relief Carving: Form, Content, Rules for Design, and Role in Mesoamerican Art History and Archaeology. Studies in Pre–Colombian Art and Archaeology, No. 27. Washington, D.C., Dumbarton Oaks. Stark, M. T. 1998. Technical Choices and Social Boundaries in Material Culture Patterning: An Introduction. In M. T. Stark (ed.), The Archaeology of Social Boundaries: 1–11. Washington D. C., Smithsonian Institution. Stark, M. T., Bishop, R. L., and Miksa, E. 2000. Ceramic Technology and Social Boundaries: Cultural Practices in Kalinga Selection and Use. Journal of Archaeological Method and Theory 7: 295–331. Stein, G. J. 2002. From Passive Periphery to Active Agents: Emerging Perspectives in the Archaeology of Interregional Interaction. American Anthropologist 104(3): 903–916. Stross, B. 1982. Maya Hieroglyphic Writing and Mixe– Zoquean. Anthropological Linguistics 24: 73–134. Stross, B. 1990. Mesoamerican Writing at the Crossroads: The Late Formative. Visible Language XXIV(1): 38–67. Struever, S. 1972. The Hopewell Interaction Sphere in Riverine–Western Great Lakes Culture History. In M. P. Leone (ed.), Contemporary Archaeology: 303– 315. Carbondale, Southern Illinois University Press. Stuart, D. 2015. The Royal Headband: A PanMesoamerican Hieroglyph. Electronic document, https://decipherment.wordpress.com/2015/01/26/ the-royal-headband-a-pan-mesoamericanhieroglyph-for-ruler/, accessed January 31, 2015. Taube, K. A. 2000a. The Writing System of Ancient Teotihuacan. Ancient America No. 1. Barnardsville, North Carolina, Center for Ancient American Studies.

Saturno, W. A., Stuart, D., and Beltrán, B. 2006. Early Maya Writing at San Bartolo, Guatemala. Science 311: 1281–1283. Schele, L. 1991. An Epigraphic History of the Western Maya Region. In T. P. Culbert (ed.), Classic Maya Political History: Hieroglyphic and Archaeological Evidence: 72–101. Cambridge, Cambridge University Press. Schele, L. 1991. The Olmec Mountain And Tree Creation In Mesoamerican Cosmology. In E. P. Benson and B. de la Fuente (eds.), The Olmec World: Ritual and Rulership: 105–117. Princeton, Princeton University Art Museum. Schele, L. 1999. Sprouts and the Early Symbolism of Rulers in Mesoamerica. In N. Grube (ed.), The Emergence of Lowland Maya Civilization: The Transition from the Preclassic to Early Classic, A Conference at Hildesheim, Germany, November 1992: 117–135. Acta Mesoamericana, No. 8, Verlag Anton Saurwein, Markt Schwaber. Schiffer, M. 1987. Formation Processes of the Archaeological Record. Albuquerque, University of New Mexico Press. Schmandt–Besserat, D. 1977. An Archaic Recording System and the Origin of Writing. SyroMesopotamian Studies 1: 31–70. Schmandt–Besserat, D. 1992. Before Writing (2 vols.). Austin, University of Texas Press. Schoep, I. 1999. The Origins of Writing and Administration on Crete. Oxford Journal of Archaeology 18: 265–276. Schoep, I. 2006. Looking Beyond the First Palaces: Elites and the Agency of Power in EM III–MM II Crete. American Journal of Archaeology 110(1): 37–64. Senner, W. M. 1989. Theories and Myths on the Origins of Writing: A Historical Overview. In W. M. Senner (ed.), The Origins of Writing: 1–26. Lincoln, University of Nebraska Press. Sharer, R. J. 1989a. The Olmec and the Southeastern Periphery of Mesoamerica. In R. J. Sharer and D. C. Grove (eds.), Regional Perspectives on the Olmec: 247–274. Washington, D. C., National Gallery of Art. Sharer, R. J. 1989b. The Preclassic Origins of Maya Writing: A Highland Perspective. In W. F. Hanks and D. S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation: 165–175. Salt Lake City, University of Utah Press. Sharer, R. J., and Sedat, D. W. 1973. Monument 1, El Porton, Guatemala and the Development of Maya Calendrical and Writing Systems. In J. A. Graham (ed.), Studies in Ancient Mesoamerica: 177–194. University of California Archaeological Research Facility, Contribution 18. Berkeley, University of California. Sharer, R. J., and Sedat, D. W. 1987. Archaeological Investigations of the Northern Maya Highlands, 201

Archaeological Paleography Taube, K. A. 2000b. Lightning Celts and Corn Fetishes: The Formative Olmec and the Development of Maize Symbolism in Mesoamerica and the American Southwest. In J. E. Clark and M. E. Pye (eds.), Olmec Art and Archaeology in Mesoamerica: 297– 331. Studies in the History of Art, No. 58. Center for Advanced Study in the Visual Arts Symposium Papers XXXV. Washington, D. C./ New Haven, National Gallery of Art/Yale University Press. Taube, K. A. 2004. Olmec Art at Dumbarton Oaks. Studies in Pre–Columbian Art and Archaeology, No. 2. Washington, D.C., Dumbarton Oaks. Taube, K. A., Saturno, W. A., Stuart, D., and Hurst, H. 2010. The Murals of San Bartolo, El Petén, Guatemala, Part 2: The West Wall. Ancient America No. 10. Barnardsville, North Carolina, Center for Ancient American Studies. Trigger, B. G. 1989. A History of Archaeological Thought. Cambridge, Cambridge University Press. Trigger, B. G. 2004. Writing Systems: A Case study in Cultural Evolution. In S. D. Houston (ed.), The First Writing: Script Invention as History and Process: 39–68. Cambridge, Cambridge University Press. Urcid, J. 2001.Zapotec Hieroglyphic Writing. Studies in Pre–Columbian Art and Archaeology, No. 34. Washington, D. C., Dumbarton Oaks. Urcid, J. 2005. Zapotec Writing: Knowledge, Power, and Memory in Ancient Oaxaca. Crystal River, Florida, Foundation for the Advancement of Mesoamerican Studies, Inc. Vargas, E. 1985. Proyecto Arqueológico El Arenal– Tabasco. Mexico City, Archivo del Centro de Estudios Mayas, Universidad Nacional Autónoma de México. Vargas, E., and Hernández, M. I. 1979. Notas sobre el Sitio Arqueológico de Reforma, Balancán, Tabasco. Revista Mexicana de Estudios Antropológicos XXV: 49–75. Vargas, E., and Ochoa, L. 1982. Navegantes y Viajeros y Mercaderes: Notas para el Estudio de las Rutas Fluviales y Terrestres entre la Costa de Tabasco– Campeche y Tierra Adentro. Estudios de Cultura Maya XlV: 59–118. Vázquez, R., Scherer, A., and Golden, C. 2005. Proyecto Arqueológico Sierra del Lacandón, Informe Preliminar 3. Guatemala City, Dirección del Patrimonio Cultural y Natural de Guatemala. Velásquez García, E. 2008. El Desciframiento de la Escritura Istmeña: Reevaluación del Estado de la Cuestión. In M. T. Uriarte and R. González Lauck (eds.), Olmeca. Balance y Perspectivas: 639–665. Mexico City, Instituto Nacional de Antropología e Historia.

Voss, J., and Young, R. 1995. Style and the Self. In C. Carr and J. Neitzel (eds.), Style, Society, and Person: 77–100. New York, Plenum. Walby, S. 2007. Complexity Theory, Systems Theory, and Multiple Intersecting Social Inequalities. Philosophy of the Social Sciences 37: 449–470. Wallerstein, I. M. 1974. The Modern World–System I. Capitalist Agriculture and the Origins of the European World–Economy in the Sixteenth Century. San Diego, Academic Press. Wang, X. F., and Chen, G. 2003. Complex Networks: Small–World, Scale–Free, and Beyond. IEEE Circuits and Systems: 6–20. West, R. C., Psuty, N., and Thom, B. 1985. Las Tierras Bajas de Tabasco en el Sureste de México. Biblioteca Básica Tabasqueña, No. 8. Villahermosa, Instituto de Cultura del Estado de Tabasco. Wichmann, S. 1995. The Relationship Among the Mixe– Zoquean Languages of Mexico. Salt Lake City, University of Utah Press. Wichmann, S., Beliaev, D., and Davletshin, A. 2008. Posibles Correlaciones Lingüísticas y Arqueológicas Vinculadas con los Olmecas. In M. T. Uriarte and R. González Lauck (eds.), Olmeca. Balance y Perspectivas: 667–683. Mexico City, Instituto Nacional de Antropología e Historia. Willey, G. R. 1970. Type Description of the Ceramics of the Real Xe Complex, Seibal, Petén, Guatemala. In W. R. Bullard (ed.), Monographs and Papers in Maya Archaeology, Vol. 61, Part IV, No. 2: 313–355. Cambridge, Peabody Museum, Harvard University. Willey, G. R. 1991. Horizontal Integration and Regional Diversity: An Alternating Process in the Rise of Civilizations. American Antiquity 56(2): 197–215. Willey, G. R., Culbert, T. P., and Adams, R. E. W. 1967. Maya Lowland Ceramics: A Report from the 1965 Guatemala City Conference. American Antiquity 32(3): 289–315. Willey, G. R., and Matthews, P. L. (eds.) 1985. A Consideration of the Early Classic Period in the Maya Lowlands. Institute for Mesoamerican Studies Publication 10. Albany, State University of New York. Wobst, M. 1977. Stylistic Behavior and Information Exchange. In C. E. Cleland (ed.), For the Director: Research Essays in Honor of James B. Griffin: 317– 342. Ann Arbor, University of Michigan Museum of Anthropology. Yoffee, N. 2005. Myths of the Archaic State: Evolution of the Earliest Cities, States, and Civilizations. Cambridge, Cambridge University Press. Zafirovsky, M. 2005. Social Exchange Theory under Scrutiny: A Positive Critique of its Economic– Behaviorist Formulations. Electronic Journal of Sociology: 1–40.

202