Source-linked AI summary

Cultural route to the emergence of linguistic categories

Andrea Puglisi, Andrea Baronchelli, Vittorio Loreto

arXiv:physics/0703164v1physics.soc-phcond-mat.dis-nncs.MA

TL;DR

The paper asks how linguistic categories become globally accepted without central coordination. It models individuals co-evolving symbols and meanings through language games, finding a hierarchical system in which shared linguistic categories support communication while perceptual categories continue evolving.

  • Problem

    The paper asks how linguistic categories become accepted at a global level without central coordination.

  • Method

    The paper models individuals co-evolving symbols and meanings through elementary language games, including a discrimination stage in which speakers and hearers may refine perceptions.

  • Results

    The model produces a two-level hierarchy: evolving perceptual categories support fine discrimination, while a common language groups perceptions for communicative success.

  • Takeaways & Limitations

    Linguistic categories emerge through coordination among individuals, forming shared categories with a finite, small vocabulary and communication success above 80%.

  • Takeaways & Limitations

    The model assumes individuals are never replaced and that perceptive resolution limits their ability to distinguish stimuli.

Abstract

from arXiv · show

Categories provide a coarse grained description of the world. A fundamental question is whether categories simply mirror an underlying structure of nature, or instead come from the complex interactions of human beings among themselves and with the environment. Here we address this question by modelling a population of individuals who co-evolve their own system of symbols and meanings by playing elementary language games. The central result is the emergence of a hierarchical category structure made of two distinct levels: a basic layer, responsible for fine discrimination of the environment, and a shared linguistic layer that groups together perceptions to guarantee communicative success. Remarkably, the number of linguistic categories turns out to be finite and small, as observed in natural languages.

Abstract

The model shows that interacting individuals can self-organise a shared linguistic layer above finer, individual perceptual categories. This cultural negotiation produces effective communication using a finite, small set of linguistic categories.

  • Model and motivation: Individuals begin with perceptual categories and co-evolving form–meaning associations, then develop a shared communication system through elementary language games.The model uses analogical stimuli, category partitions, and dynamically updated word inventories without external supervision.
  • Emergence of communication: The population passes from an initial phase of poor understanding to a later phase of full communication through a common language.Synonymy first grows as new words are created, then declines as successful words eliminate competing alternatives.
  • Emergence of communication: 80% is the final success-rate floor across numerical experiments, while most experiments exceed 90%.Success increases as synonymy disappears and remains stable after reaching its maximum.
  • Finite linguistic inventory: The number of linguistic categories saturates over time and becomes increasingly independent of perceptual resolution as d_min decreases.A finite number persists even in the continuum limit d_min → 0, with only weak logarithmic dependence on population size N.
  • Environment and culture: Cultural interaction aligns linguistic categories, while environmental stimulus frequencies influence where those categories become more refined.The model therefore reproduces a small linguistic inventory despite high perceptual resolution.

Appendix

The paper measures alignment between players’ category inventories using an overlap function based on category intersections and widths.

  • The overlap measure quantifies how closely two players’ category inventories align.Its value reaches one when the inventories exactly coincide.
  • The measure is computed over categories belonging to individual players and their intersection sets.The notation identifies iic as one category of the i-th player and ijc as a generic intersection category.
  • Category width is represented explicitly in the alignment calculation.The paper defines cllc as the width of category c.
  • The calculation compares category intersections across players’ category boundaries.Intersection categories are formed from the boundaries of both players’ inventories.
Loading physics/0703164v1…