Exploring an extinct society through the lens of Habitus-Field theory and the Tocharian text corpus

This research bridges the gap between sociology, linguistics, and archaeology by applying sociological theories and methods to the study of extinct societies. The researchers aim to test the applicability of Habitus-Field Theory (HFT) and MCA to an extinct society and language, specifically the Tocharian civilization and language of the Tarim Basin and Taklamakan desert. Tocharian, an extinct Indo-European language, offers a unique opportunity to explore social structures through its surviving textual fragments. The study acknowledges the inherent bias of the corpus, which primarily consists of texts produced by and for elites, yet the sheer volume and diversity of the texts allows researchers to reconstruct broad aspects of religious views, cultural practices, and socio-economic spaces. The Tocharian corpus, still undergoing edition and translation, presents unique challenges, requiring careful philological and linguistic methods alongside digital tools for analysis. By combining close reading, inductive content analysis, and MCA, coupled with archaeological and linguistic metadata, the researchers aim to develop a blueprint for analyzing extinct societies using sociological, linguistic, archaeological, and natural language processing (NLP) tools.

The study draws upon existing literature in sociology, particularly Habitus-Field Theory (HFT) developed by Pierre Bourdieu. HFT emphasizes the interplay between power relations, ideas, social categorizations, and resource distribution in shaping social order. The researchers highlight Bourdieu's concepts of capital (economic, social, cultural, symbolic), fields, illusio, doxa, and nomos, which help to understand the competition for power and resources among actors within different societal arenas. The literature review also incorporates previous archaeological and historical studies of the Tarim Basin and the Tocharian civilization, demonstrating the importance of understanding the cultural context and historical background of the texts. Existing research on Buddhist patronage and the role of elites in shaping religious institutions provides a comparative framework for interpreting the Tocharian data. The authors specifically draw upon studies that have utilized HFT in historical and archaeological research, demonstrating its applicability to the analysis of past societies.

The researchers employ a sequential mixed-methods design combining close reading, inductive content analysis, and specific MCA. Close reading of the Tocharian text fragments, performed independently by both authors with varying levels of expertise, aimed to identify topics and uncover connections between content and mode of production. Inductive content analysis guided the development of a coding scheme used to categorize text fragments based on identified topics, using regular expressions (regex) in Python to locate specific terms. The coding scheme incorporates topics like religious themes, economic activities, descriptions of elites, geographical locations and material properties of the texts. Metadata from the CETOM database, including linguistic information, material type, and find region, was then linked to the coded text fragments. Specific MCA, a statistical method, was used to analyze the resulting data matrix and reduce the high-dimensional space to interpretable dimensions. The selection of dimensions was based on the elbow criterion applied to the scree plot. The variables included in the MCA were: linguistic data (language, genre, verse, prose), media data (material), content categories (topics), and passive variables like regional data. The analysis was performed using the FactoMineR package in RStudio. The researchers manually interpreted the results from the MCA, visualized as biplots, to identify patterns and relationships between variables.

The MCA revealed a three-dimensional structure within the Tocharian corpus. Dimension 1 distinguishes between spiritual and secular elites. Texts related to religious themes, descriptions of heaven and hell, and mentions of Buddhist figures are grouped together with texts concerning worldly elites like royalty, ministers, and householders. These texts were predominantly found in major societal centers. Dimension 2 focuses on economic activity, particularly related to the caravan trade. This dimension highlights texts containing mentions of signums (caravan passes), economic terms, and non-literary texts. These texts were often written on wood or walls and found outside the major intellectual and social centers, representing a functional differentiation between intellectual and secular elites. Dimension 3 contrasts secular and spiritual poles. Householders and terms related to material wealth are associated with the secular pole, while spiritual themes, Buddhist metaphors and locations associated with spiritual journeys are grouped with the spiritual pole. This dimension suggests a power dynamic where spiritual and political elites have symbolic dominance over economic elites. The geographic distribution of the texts further supports the functional differentiation. The findings indicate a close alliance between political and spiritual elites, exerting symbolic violence over the economic elites by devaluing their practices and perspectives in the written record. The act of writing itself becomes a tool to assert and maintain power. The study observes that while economic elites were essential for the functioning of society, their perspectives are largely marginalized in the surviving texts.

The findings address the research question by demonstrating the applicability of HFT and MCA to the analysis of an extinct society. The three-dimensional structure identified in the Tocharian corpus highlights a functional differentiation within the society, separating political-spiritual elites from economically active elites. This aligns with the known historical context and supports the notion of symbolic power wielded by political and spiritual elites over economic ones. The spatial distribution of texts and their materiality further reinforces this interpretation. The study’s findings resonate with historical parallels, such as the relationships between religious and secular authorities in medieval Europe, but with a more prominent role for economic elites in the Tocharian case. The study also highlights the crucial role of patronage networks in the spread of Buddhism and the conscious strategy employed by secular elites to gain cultural capital by sponsoring religious institutions.

This study successfully demonstrates the use of HFT, MCA, and NLP methods in reconstructing the social order of an extinct society from fragmentary linguistic corpora. It reveals a complex interplay between political-spiritual and economic elites in Tocharian society, with a clear power imbalance favoring the former. This method offers valuable insights into the past, extending the temporal reach of social science analysis. Future research should expand on the current study by incorporating additional data, addressing the limitations and utilizing temporal changes in the language to trace symbolic revolutions and shifts in power structures.

The study acknowledges several limitations. The corpus primarily consists of texts produced by elites, leading to a biased representation of society. The limited amount of variance explained by the three dimensions suggests the presence of further topic dimensions requiring future exploration. The lack of dating for many sources prevents a dynamic analysis of social changes over time. The study also focuses only on the textual evidence without a detailed consideration of habitus for the broader population.