Anaphoric Structure Emerges Between Neural Networks

TL;DR

This paper investigates whether anaphoric structures, like pronouns, naturally emerge between neural networks trained on communication tasks, revealing that such pragmatic features can develop without explicit constraints.

Contribution

The study demonstrates that neural networks can spontaneously develop anaphoric structures during communication, influenced by efficiency pressures, shedding light on language evolution.

Findings

Neural models can learn languages with anaphoric structures.

Anaphoric structures emerge naturally without explicit constraints.

Efficiency pressures increase the prevalence of anaphoric structures.

Abstract

Pragmatics is core to natural language, enabling speakers to communicate efficiently with structures like ellipsis and anaphora that can shorten utterances without loss of meaning. These structures require a listener to interpret an ambiguous form - like a pronoun - and infer the speaker's intended meaning - who that pronoun refers to. Despite potential to introduce ambiguity, anaphora is ubiquitous across human language. In an effort to better understand the origins of anaphoric structure in natural language, we look to see if analogous structures can emerge between artificial neural networks trained to solve a communicative task. We show that: first, despite the potential for increased ambiguity, languages with anaphoric structures are learnable by neural models. Second, anaphoric structures emerge between models 'naturally' without need for additional constraints. Finally, introducing an explicit efficiency pressure on the speaker increases the prevalence of these structures. We conclude that certain pragmatic structures straightforwardly emerge between neural networks, without explicit efficiency pressures, but that the competing needs of speakers and listeners conditions the degree and nature of their emergence.