Rational Communication Shapes Morphological Composition

The Study
200 years of English word-making 

We tested this idea at scale. Using a time-indexed lexicon constructed from COHA and COCA, we collected 4323 naturally occurring English compounds and derivations spanning 1820–2019.

For each word, we generated a large set of alternative morpheme combinations that were available in the lexicon at the same historical moment, and tested if our model ranks the actual attested form above those alternatives. 

We built a semantic model that infers a concept from morphemes (using modern language model embeddings), and a cost model that minimizes production cost (using historical corpus statistics). We then integrate these components to model pragmatic speaker choice as predicted by the rational communication hypothesis.

Here we show, within the Rational Speech Act (RSA) framework, attested compositions are systematically ranked above unattested alternatives generated from contemporaneously available morphemes. Models integrating semantic informativeness with production cost outperform semantic-only and cost-only baselines on Mean Reciprocal Rank (MRR) and top-k accuracy (Acc@k), with the advantage of the Pragmatic Speaker model (𝑆1) over the semantic-only baseline growing as the candidate set expands, where meaning alone leaves morphological choice underdetermined.

These findings suggest that lexicalization reflects a communicative trade-off between expressiveness and efficiency, extending rational accounts of communication from utterance-level choice to the internal structure of words.