Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Computational Typology

Gerhard Jäger

TL;DR

The paper tackles how to robustly test language universals and cross-linguistic correlations using computational typology, explicitly addressing non-independence due to genealogy and areal contact. It demonstrates, through hierarchical and phylogenetic Bayesian models, that correlations between affix positioning and adposition type are a genuine diachronic tendency, whereas the apparent link between population size and phoneme inventory size largely reflects shared ancestry. By employing bivariate models for mixed data types and comparing vanilla, hierarchical, and phylogenetic formulations, the work shows substantial gains in model fit and interpretability when genealogical structure is accounted for. The findings underscore the need to integrate phylogenetic information in typological research to distinguish true co-evolution from historical contingency, with practical implications for large-scale linguistic databases and future computational typology studies.

Abstract

Typology is a subfield of linguistics that focuses on the study and classification of languages based on their structural features. Unlike genealogical classification, which examines the historical relationships between languages, typology seeks to understand the diversity of human languages by identifying common properties and patterns, known as universals. In recent years, computational methods have played an increasingly important role in typological research, enabling the analysis of large-scale linguistic data and the testing of hypotheses about language structure and evolution. This article provides an illustration of the benefits of computational statistical modeling in typology.

Computational Typology

TL;DR

The paper tackles how to robustly test language universals and cross-linguistic correlations using computational typology, explicitly addressing non-independence due to genealogy and areal contact. It demonstrates, through hierarchical and phylogenetic Bayesian models, that correlations between affix positioning and adposition type are a genuine diachronic tendency, whereas the apparent link between population size and phoneme inventory size largely reflects shared ancestry. By employing bivariate models for mixed data types and comparing vanilla, hierarchical, and phylogenetic formulations, the work shows substantial gains in model fit and interpretability when genealogical structure is accounted for. The findings underscore the need to integrate phylogenetic information in typological research to distinguish true co-evolution from historical contingency, with practical implications for large-scale linguistic databases and future computational typology studies.

Abstract

Typology is a subfield of linguistics that focuses on the study and classification of languages based on their structural features. Unlike genealogical classification, which examines the historical relationships between languages, typology seeks to understand the diversity of human languages by identifying common properties and patterns, known as universals. In recent years, computational methods have played an increasingly important role in typological research, enabling the analysis of large-scale linguistic data and the testing of hypotheses about language structure and evolution. This article provides an illustration of the benefits of computational statistical modeling in typology.

Paper Structure

This paper contains 12 sections, 9 equations, 5 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Historical Context
  3. Language universals
  4. Correlation between linguistic and non-linguistic traits
  5. Electronic Resources
  6. Statistical non-independence
  7. Two case studies
  8. Statistical evaluation under the assumption of independence
  9. Taking phylogenetic non-independence into account I: Hierarchical models
  10. Taking genetic non-independence into account II: Phylogenetic models
  11. Summary
  12. Conclusion

Figures (5)

  • Figure 1: Distribution of affix types across postpositions and prepositions.
  • Figure 2: Scatterplot of population size and phoneme inventory size.
  • Figure 3: Map of population size and phoneme inventory size.
  • Figure 4: Top: Phylogenetic tree of the world's languages. The tree was pruned to the 589 languages for which WALS data were available. Bottom: Subtree for the Indo-European languages.
  • Figure 5: Ancestral state reconstructions for the Indo-European sub-tree. The left panel shows the reconstructed values for the affix latent variable, while the right panel shows the reconstructed values for the population size and adposition latent variable. Markers at the leaves indicate observed features.