Large language models are not about language
This challenges the relevance of LLMs for understanding human language, which is incremental as it critiques existing approaches without proposing a new method.
The paper argues that large language models are not useful for linguistics because they are probabilistic models requiring vast data to analyze word strings, whereas human language is based on an internal computational system that generates hierarchical thoughts with minimal input and can distinguish real from impossible languages.
Large Language Models are useless for linguistics, as they are probabilistic models that require a vast amount of data to analyse externalized strings of words. In contrast, human language is underpinned by a mind-internal computational system that recursively generates hierarchical thought structures. The language system grows with minimal external input and can readily distinguish between real language and impossible languages.