Compositionality may not guarantee that linguistic meaning is computable

Linguistic meaning may not be computable, even when the meaning of a sentence is built systematically from the meanings of its parts, according to a perspective article published in Frontiers in Language Sciences by Prakash Mondal of the Indian Institute of Technology Hyderabad.

The article, titled “What Compositionality Reveals about Meaning, Cognition, and Machines,” examines the computational basis of linguistic meaning from what the author calls a compositional computability perspective. Compositionality, as described in the article, requires that the meaning of a whole expression depend on the meanings of its parts.

According to the abstract, the article argues that linguistic meaning is not computable under certain conditions of compositionality. In other words, the author contends that compositionality does not guarantee the computability of linguistic meanings.

A second strand of the argument concerns syntax. The abstract states that the non-computability result is partly governed by the syntactic underdetermination of linguistic meanings, as conceived within syntactic theories that assume computability of structural operations for compositional mapping. On this view, syntax understood in a purely computational sense does not guarantee computability of semantic composition when that composition is understood as a cognitive mechanism of understanding linguistic meaning.

The article then indicates the limits of computability of linguistic meaning. The author approaches these limits from the vantage point of unifying the cognitive and formal aspects of linguistic meaning in brain dynamics, seeking to connect formal semantic theory with the dynamics of the brain that underlie language understanding.

The analysis is presented as having consequences for Large Language Models and generative artificial intelligence systems. The abstract notes that such systems appear to stumble over linguistically deep or pragmatically rich aspects of meaning. The article projects insights into the limits of the computational basis of linguistic meaning in both humans and machines.

It is important to note the nature and scope of this contribution. The piece is explicitly labeled a perspective article, and it presents a theoretical and philosophical argument rather than new experimental findings. Its claims concern the conceptual foundations of compositional semantics and computability, not measured performance results from any particular model or experiment. The claims expressed are those of the author and do not necessarily represent those of the publisher, editors, or reviewers, as the article’s own disclaimer states.

The article was received on 26 August 2026 and accepted on 13 October 2026, and it is published open access under a Creative Commons Attribution license. It appears in the journal's Bilingualism section and lists keywords including brain dynamics, cognition, computability, linguistic compositionality, and large language models.

Subject of Research: The computability of compositional linguistic meaning and its limits in cognition and machines

Article Title: What compositionality reveals about meaning, cognition, and machines

Article References: Mondal, P. (2026). What compositionality reveals about meaning, cognition, and machines. Frontiers in Language Sciences, 5, Article 1979638. https://doi.org/10.3389/flang.2026.1979638

Image Credits: AI Generated

DOI: 10.3389/flang.2026.1979638

Keywords: linguistic meaning, compositionality, computability, semantics, syntax, cognition, brain dynamics, large language models, generative AI, language sciences, reveals, meaning