Systematicity between Forms and Meanings across Languages Supports Efficient Communication

TL;DR

This paper investigates how cross-linguistic variation in form-meaning mappings can be understood through systematicity within word forms. It introduces CETL, a learnability-based complexity measure that captures internal structure by encoding forms as character sequences and measuring learning speed under a need-weighted encoder. Operating within an information-theoretic framework that balances accuracy and simplicity, CETL is applied to verb and pronoun paradigms across typologically diverse languages and contrasted with the Information Bottleneck baseline. The results show attested paradigms achieve more favorable complexity-accuracy trade-offs, with natural syncretism patterns driving easier learning, thereby linking efficient communication theory to natural language systematicity. The framework offers a nuanced account of how internal form structure contributes to communicative efficiency and provides a tool to discriminate genuine systematicity from superficial syncretism.

Abstract

Languages vary widely in how meanings map to word forms. These mappings have been found to support efficient communication; however, this theory does not account for systematic relations within word forms. We examine how a restricted set of grammatical meanings (e.g. person, number) are expressed on verbs and pronouns across typologically diverse languages. Consistent with prior work, we find that verb and pronoun forms are shaped by competing communicative pressures for simplicity (minimizing the inventory of grammatical distinctions) and accuracy (enabling recovery of intended meanings). Crucially, our proposed model uses a novel measure of complexity (inverse of simplicity) based on the learnability of meaning-to-form mappings. This innovation captures fine-grained regularities in linguistic form, allowing better discrimination between attested and unattested systems, and establishes a new connection from efficient communication theory to systematicity in natural language.

Figures (23)

• Figure 1: Turkish pronouns show systematic form-meaning mappings: person is consistently marked by prefixes (e.g., s- for second person), number by suffixes. Language evolution research demonstrates that such systematicity supports learnability. Our model connects these findings, proposing that learnable, systematic mappings contribute to communicative efficiency.
• Figure 2: Communication model, adapted from Zaslavsky2018colorzaslavsky2021pronouns. Our model encodes the form $w$ as a sequence, and decodes it as an atomic unit.
• Figure 3: Example input and output used in training.
• Figure 4: Complexity and accuracy in our model. Across domains, real, attested paradigms (colored, one dot per language) are more efficient than nearly all counterfactual structural (gray) and surface (blue) permutations.
• Figure 5: Complexity plotted against unnaturalness for Afro-Asiatic verbs by model, CETL (left) and IB (right). We group results by paradigm type (reflecting different grammatical categories; cf. App. \ref{['app:feat_repr_APP']}). CETL positively correlates with unnaturalness, meaning more natural paradigms have lower complexity, while IB shows no correlation.