Periodic Reporting for period 4 - SAPPHIRE (South American population history revisited: multidisciplinary perspectives on the Upper Amazon)
Reporting period: 2024-02-01 to 2025-03-31
The central challenge faced by linguistics is to understand linguistic diversity. Despite influential programmatic studies and sustained interest, many aspects of the evolution of linguistic diversity remain elusive. Yet, the potential value of understanding linguistic diversity is enormous. It could help us uncover specific past events such as migration and historical contact between societies, perhaps even in very deep time and, more generally, it can give us insight into the interaction between ecology and human behavior and human cognition.
This project addresses these challenges by focusing on the Upper Amazon in western South America, an area which forms a unique microcosm of diversity in which several sociohistorical layers of the continental population history are represented. The north and south are extremely diverse and possibly represent very old diversity patterns. Both areas also show clear signs of regional, and possibly deep-time language contact. The central part is dominated by three language families, Panoan, Arawakan, and Quechuan. This microcosm of diversity forms an unparalleled natural laboratory for understanding more about sociological dynamics and the evolution of linguistic diversity.
The project takes an innovative multi-disciplinary approach to unearthing the socio-historical dynamics that have led to these various diversity patterns. In order to do so, geographical, socio-historical, and communicative layers of history are reconstructed and systematically compared, achieving a level of synergy between the disciplines hitherto unseen. Interpreting the (mis)matches between these layers allows for a more informed reconstruction of past events (e.g. migrations, language shift, exogamy, trade), and sheds light on how these social dynamics shaped linguistic diversity, in turn unlocking the potential of linguistic diversity as a unique window on our past.
This project addresses these challenges by focusing on the Upper Amazon in western South America, an area which forms a unique microcosm of diversity in which several sociohistorical layers of the continental population history are represented. The north and south are extremely diverse and possibly represent very old diversity patterns. Both areas also show clear signs of regional, and possibly deep-time language contact. The central part is dominated by three language families, Panoan, Arawakan, and Quechuan. This microcosm of diversity forms an unparalleled natural laboratory for understanding more about sociological dynamics and the evolution of linguistic diversity.
The project takes an innovative multi-disciplinary approach to unearthing the socio-historical dynamics that have led to these various diversity patterns. In order to do so, geographical, socio-historical, and communicative layers of history are reconstructed and systematically compared, achieving a level of synergy between the disciplines hitherto unseen. Interpreting the (mis)matches between these layers allows for a more informed reconstruction of past events (e.g. migrations, language shift, exogamy, trade), and sheds light on how these social dynamics shaped linguistic diversity, in turn unlocking the potential of linguistic diversity as a unique window on our past.
The work within the project has focused on the following three themes:
Infrastructure for quantitative analyses and data collection
We collected saliva samples from 200 individuals, speakers of several indigenous languages in the Ucayali-Urubamba region. These samples have been processed for whole-genome sequencing. We furthermore collected cultural data on ca. 63 societies and grammatical data on ca. 100 languages.
Whereas the databases for genetics and biogeography follow established designs, the database designs for cultural and linguistic data were developed specifically for the project. They are based on low-level, fine-grained variables, which in turn are grouped into higher-level classifications. This design allows us to assess signals in the data at different levels of granularity.
In order to streamline the workflow, the project has developed an R package called glottospace (https://CRAN.R-project.org/package=glottospace(opens in new window)). This package contains several analytical tools and visualization techniques that allow for the spatial analysis of linguistic and cultural data, based on distance measures.
In collaboration with a number of colleagues, we are furthermore developing spatial representations of language locations in a consensus map based on previously published maps, which we have digitized, georeferenced and annotated. This allows for the representation of language locations in the form of polygons, which is a more accurate approach to reality. The polygon data will also be integrated into the glottospace package. An accompanying paper is currently under review.
Approach and method development
No established framework exists for the analysis of multidisciplinary data to reconstruct the past. This part of the project, therefore, has focused on the development of a framework for combining signals from different disciplines to reconstruct population history, in particular historical contact scenarios. This framework relies heavily on both new and established quantitative methods but firmly embeds the signals that result from the quantitative analysis in the existing ethnohistorical literature. An important methodological tool involves several methods of several types of distance measuring (between societies, languages), which allows for direct comparisons between the disciplines, and the establishment of (mis)matches. This part of the approach has also been included in the aforementioned Glottospace package. Another recurring element in our approach is to test hypotheses that have been proposed in the literature, which imply predictions about the patterns we would expect to find in the different disciplines.
Regional studies
Using the infrastructure and methodology mentioned above, we have initiated a number of regional and thematic studies. These studies focus in different ways on uncovering patterns that are suggestive of past contact scenarios in the Upper Amazon and between Upper Amazonian and their Andean and Amazonian neighbors. See more on these projects in the next section.
Infrastructure for quantitative analyses and data collection
We collected saliva samples from 200 individuals, speakers of several indigenous languages in the Ucayali-Urubamba region. These samples have been processed for whole-genome sequencing. We furthermore collected cultural data on ca. 63 societies and grammatical data on ca. 100 languages.
Whereas the databases for genetics and biogeography follow established designs, the database designs for cultural and linguistic data were developed specifically for the project. They are based on low-level, fine-grained variables, which in turn are grouped into higher-level classifications. This design allows us to assess signals in the data at different levels of granularity.
In order to streamline the workflow, the project has developed an R package called glottospace (https://CRAN.R-project.org/package=glottospace(opens in new window)). This package contains several analytical tools and visualization techniques that allow for the spatial analysis of linguistic and cultural data, based on distance measures.
In collaboration with a number of colleagues, we are furthermore developing spatial representations of language locations in a consensus map based on previously published maps, which we have digitized, georeferenced and annotated. This allows for the representation of language locations in the form of polygons, which is a more accurate approach to reality. The polygon data will also be integrated into the glottospace package. An accompanying paper is currently under review.
Approach and method development
No established framework exists for the analysis of multidisciplinary data to reconstruct the past. This part of the project, therefore, has focused on the development of a framework for combining signals from different disciplines to reconstruct population history, in particular historical contact scenarios. This framework relies heavily on both new and established quantitative methods but firmly embeds the signals that result from the quantitative analysis in the existing ethnohistorical literature. An important methodological tool involves several methods of several types of distance measuring (between societies, languages), which allows for direct comparisons between the disciplines, and the establishment of (mis)matches. This part of the approach has also been included in the aforementioned Glottospace package. Another recurring element in our approach is to test hypotheses that have been proposed in the literature, which imply predictions about the patterns we would expect to find in the different disciplines.
Regional studies
Using the infrastructure and methodology mentioned above, we have initiated a number of regional and thematic studies. These studies focus in different ways on uncovering patterns that are suggestive of past contact scenarios in the Upper Amazon and between Upper Amazonian and their Andean and Amazonian neighbors. See more on these projects in the next section.
The literature on the reconstructed past of the Americas has focused mainly on large-scale patterns, and the promotion of a single or at least dominant historical explanation for the unexpected linguistic diversity found in the continent. In this project, we have examined these claims, taking a more regional approach, as well as an approach that takes into account signals from genetics, geography, cultural anthropology, and linguistics. Generally speaking, we found that claims made in the literature did have some validity, but only to a limited extent. The past of the Upper Amazon is far too intricate to be explained by a single explanatory model.
For the northern Upper Amazon, we found that both isolation and contact have played a role in creating and, particularly, maintaining linguistic and cultural diversity in the northern Upper Amazon. Genetic patterns in the northern Upper Amazon suggest different sociohistorical phases, with local groups reacting differentially to expanding groups. The different phases of social history were also highlighted in a study that focused on two societies in the northern Upper Amazon. Preliminary results for the central Upper Amazon suggest that, while there is ample evidence for a historical trade network, different groups in the central Upper Amazon were involved in the trade network to different degrees. Linguistic convergence effects seem to be clearest in the southern Upper Amazon, where some of the groups have been heavily influenced by Arawakan languages. The difference with the northern parts is likely due to the effects of identity preservation in the north. It must be said that, here as well, there are important differences between the extent to which groups were influenced by Arawakan languages.
A more general conclusion is that different aspects of geography, culture, biology, and language are affected differentially by the same histories. For instance, cultural behavior associated with women shows different contact effects than those associated with men. This is also true for e.g. Y-chromosomal vs. Mt DNA, phonology versus grammar, and different aspects of the biophysical environment. All of these differential results are potentially meaningful signals.
All of these conclusions point to the more general conclusion that we must abandon ideas that a single mechanism is responsible for the rise and maintenance of linguistic diversity. Reconstruction of the past is possible through careful multidisciplinary regional research.
For the northern Upper Amazon, we found that both isolation and contact have played a role in creating and, particularly, maintaining linguistic and cultural diversity in the northern Upper Amazon. Genetic patterns in the northern Upper Amazon suggest different sociohistorical phases, with local groups reacting differentially to expanding groups. The different phases of social history were also highlighted in a study that focused on two societies in the northern Upper Amazon. Preliminary results for the central Upper Amazon suggest that, while there is ample evidence for a historical trade network, different groups in the central Upper Amazon were involved in the trade network to different degrees. Linguistic convergence effects seem to be clearest in the southern Upper Amazon, where some of the groups have been heavily influenced by Arawakan languages. The difference with the northern parts is likely due to the effects of identity preservation in the north. It must be said that, here as well, there are important differences between the extent to which groups were influenced by Arawakan languages.
A more general conclusion is that different aspects of geography, culture, biology, and language are affected differentially by the same histories. For instance, cultural behavior associated with women shows different contact effects than those associated with men. This is also true for e.g. Y-chromosomal vs. Mt DNA, phonology versus grammar, and different aspects of the biophysical environment. All of these differential results are potentially meaningful signals.
All of these conclusions point to the more general conclusion that we must abandon ideas that a single mechanism is responsible for the rise and maintenance of linguistic diversity. Reconstruction of the past is possible through careful multidisciplinary regional research.