Baboons (Papio papio) Process a Context-Free but Not a Context-Sensitive Grammar

Raphaëlle Malassis,Stanislas Dehaene,Joël Fagot

doi:10.1038/s41598-020-64244-5

Abstract

Language processing involves the ability to master supra-regular grammars, that go beyond the level of complexity of regular grammars. This ability has been hypothesized to be a uniquely human capacity. Our study probed baboons’ capacity to learn two supra-regular grammars of different levels of complexity: a context-free grammar generating sequences following a mirror structure (e.g., AB | BA, ABC | CBA) and a context-sensitive grammar generating sequences following a repeat structure (e.g., AB | AB, ABC | ABC), the latter requiring greater computational power to be processed. Fourteen baboons were tested in a prediction task, requiring them to track a moving target on a touchscreen. In distinct experiments, sequences of target locations followed one of the above two grammars, with rare violations. Baboons showed slower response times when violations occurred in mirror sequences, but did not react to violations in repeat sequences, suggesting that they learned the context-free (mirror) but not the context-sensitive (repeat) grammar. By contrast, humans tested with the same task learned both grammars. These data suggest a difference in sensitivity in baboons between a context-free and a context-sensitive grammar.

Highlights

Assessing the syntactic abilities of various animal species is essential in order to better understand the evolution of the cognitive operations involved in human language processing1–3
Most previous comparative experiments focused on a single artificial grammar, usually denoted “AnBn”
Counting is of little relevance regarding the cognitive operations involved in natural language processing9. This grammar lends itself to the exploitation of alternative, low-level heuristics in non-human11,15 as well as human participants6,20–22

Summary

Introduction

Assessing the syntactic abilities of various animal species is essential in order to better understand the evolution of the cognitive operations involved in human language processing. To assess whether a sequence is grammatically correct, it is sufficient to count and compare the number of elements from each category and reject the sequence if those numbers are unequal This counting strategy requires a supra-regular computational device, because keeping track of an arbitrary (unbounded) number of A elements in order to match them with the number of B elements is something that a finite-state automaton cannot do. The mirror grammar generates sequences in which the second half of each sequence mirrors the first half (i.e., the same items appear in reverse order), such as AB | BA, ABC | CBA Mastering this grammar requires keeping track of an arbitrary number of centre-embedded long-distance dependencies (the first element must be matched with the last element, etc). Processing the repeat grammar requires keeping track of an arbitrary number of crossed dependencies, another type of long-distance relationship which requires a linear tape from which items can be retrieved in a “first-in, first-out” (FIFO) manner, starting with the first stored item

Methods

Results

Conclusion