Building social cognitive models of language change

Daniel J. Hruschka, Morten H. Christiansen, Richard A. Blythe, William Croft, Paul Heggarty, Salikoko S. Mufwene, Janet B. Pierrehumbert, Shana PoplackTrends 2009 · DOI

TL;DR

This paper argues for a unified framework for studying language change, viewing it as an evolutionary process driven by competing linguistic variants. It emphasizes the power of probabilistic, data-driven models to understand the interplay of social and cognitive factors in shaping how languages evolve over time.

What Problem Does This Paper Address?

Language change, a fundamental aspect of human culture, has historically been studied in isolation by various linguistic subfields. While each subfield offers valuable insights, there's a lack of a cohesive framework that integrates these diverse perspectives. This paper addresses the challenge of building a comprehensive understanding of language change by proposing a unified, evolutionary approach. It aims to bridge the gap between historical linguistic observations and cognitive science by modeling language change as a dynamic, population-based process influenced by both social and cognitive constraints. The authors highlight that while cognitive scientists have focused on language's origins and structure, the dynamic process of change has been relatively neglected, despite its potential to illuminate broader questions about human cognition and cultural evolution.

Methods

The paper advocates for a new emerging framework that models language change as an evolutionary process. The core methodological contribution is the emphasis on probabilistic, data-driven models. These models are crucial for two main purposes:

  • Inferring change: By analyzing linguistic data, these models can identify patterns and estimate the likelihood of specific changes occurring.
  • Comparing competing claims: They provide a rigorous way to assess the relative contributions of various social and cognitive factors to language change, allowing researchers to discriminate between different hypotheses.

The authors discuss several approaches that contribute to this framework:

  • Dynamic population-based perspective: Viewing language as a system of interacting speakers and listeners, where variation is the raw material for change.
  • Variationist approaches: Analyzing variation within and between speakers at all linguistic levels (phonology, syntax, lexicon) as the engine of change.
  • Social-cognitive approaches: Investigating how social factors (community structure, interaction patterns) and cognitive biases (learnability, ease of use) influence the spread and adoption of linguistic variants.
  • Data-based modeling: Utilizing computational models, including agent-based models and iterated learning, to simulate and test hypotheses about language change mechanisms.

Datasets

Not discussed in the paper. The paper focuses on the theoretical framework and methodological approaches rather than specific empirical studies with defined datasets.

Key Insights and Ideas

  • Language change as an evolutionary process: The paper frames language change not as a random drift but as a structured evolutionary process, akin to biological evolution, driven by variation and selection.
  • The unifying power of probabilistic models: The authors argue that probabilistic, data-driven models are essential for moving beyond descriptive accounts to explanatory ones, allowing for rigorous testing of hypotheses about the drivers of language change.
  • Interplay of social and cognitive factors: A central theme is that language change is not solely driven by internal linguistic mechanisms but is deeply intertwined with human cognition and social interaction.
  • Variation as the engine of change: Just as genetic variation is crucial for biological evolution, linguistic variation within a speech community provides the necessary raw material for language to change.
  • The importance of population structure: The way individuals interact within a community significantly influences how innovations spread and how quickly language changes.
  • Bridging disciplines: The proposed framework offers a common ground for linguists, cognitive scientists, and evolutionary biologists to collaborate and gain a more holistic understanding of language and culture.

Key Results

Not discussed in the paper. The paper is a theoretical overview and proposal for a framework, rather than a report of specific empirical findings or quantitative results.

Limitations and Future Work

The authors acknowledge several challenges and suggest future directions:

  • Interdisciplinary collaboration: A significant challenge is the need for deeper collaboration between linguists and cognitive scientists.
  • Artifact-based inference: Like other historical sciences, linguists primarily rely on indirect evidence (artifacts of speech) to infer change, which can be limited by timescale. Comparing results across different timescales is crucial.
  • Modeling dynamic structures: Developing models that can account for the variability in language use and are sufficiently dynamic to enable learning and change over time remains an ongoing challenge.
  • Quantitative fitting of models: While agent-based models are powerful for exploring complex arguments, a key next step is to develop models that are complex enough to capture essential details but simple enough to be fitted to quantitative data.
  • Understanding what is changing: Further research is needed to clarify precisely what linguistic elements are undergoing change (forms, functions, form-function mappings, rules, exemplars).
  • Generalizability of selection theories: Investigating how language change can contribute to a generalized theory of selection applicable to other evolutionary phenomena.
  • Origin of novel languages: Understanding how new languages emerge from existing communicative structures, particularly in cases like sign languages and creoles.
  • Factors influencing speed of change: Identifying the specific factors that accelerate or decelerate language change at different linguistic levels and across different population structures.

Summary updated

BibTeX
@inproceedings{hruschka2009building,
  author = {Daniel J. Hruschka and Morten H. Christiansen and Richard A. Blythe and William Croft and Paul Heggarty and Salikoko S. Mufwene and Janet B. Pierrehumbert and Shana Poplack},
  title = {{Building social cognitive models of language change}},
  booktitle = {Trends},
  year = {2009},
  doi = {10.1016/j.tics.2009.08.008},
}

Loading related content...

Figure