Culbertson, Jennifer

Rabagliati, Hugh

Fletcher, Lauren

UKRI CDT in Natural Language Processing

2026-03-25T18:11:33Z

2026-03-25

Linguistic systems are shaped by those who use them; both cognitive biases and social interaction influence how language evolves. However, a common assumption in research on language change and evolution is that the `majority wins', such that the linguistic features that prevail reflect the cognitive biases and social functioning of the most common (or `typical') individuals. This thesis challenges this assumption, focusing on the contributions of neurodivergent, particularly autistic, language users to linguistic change. I motivate this shift through focusing on the crucial questions of how language change happens, and who does it, and argue that in order to gain a full understanding of human language, we must consider variation beyond the socially imposed norms of cognition. To explore the relationship between neurodiversity and language change, this thesis presents a series of artificial language learning experiments, each investigating different aspects of autistic language learning, use, and change. In Chapter Three, I investigate the relationship between autistic traits, communicative efficiency, and social biases. These experiments first demonstrate that individuals with high autistic traits do adhere to the communicative efficiency trade-off between production and comprehension. Second, they show that in the presence of a social bias, people with higher autistic traits adapt their linguistic behaviour to a greater extent, retaining redundancy to meet a social goal. In Chapter Four, I present a follow-up study that sought to probe why people with higher autistic traits adhere more closely to social biases, and results suggest that the key mechanism driving this behaviour is autistic camouflaging. Finally, Chapter Five examines linguistic accommodation -- a mechanism of language change -- in mixed-neurotype interactions, extending the scope beyond autism to include ADHD. The results suggest that neurotype mixing strongly impacts linguistic behaviour in autistic people, with autistic participants accommodating most to other autistic people, least to allistic people, and intermediately to ADHD individuals. Together, these results demonstrate that neurodivergent people can introduce changes to language not just at the lexical level, but also at the grammatical level, and that the changes introduced by neurodivergent people may differ than those introduced by neurotypical people. This highlights the need for a broader and more inclusive perspective on not just language evolution, but accounts of cognitive functioning in general that have traditionally focused on the neurotypical norm.

https://era.ed.ac.uk/handle/1842/44527

https://doi.org/10.7488/era/7044

en

The University of Edinburgh

Fletcher, L., Rabagliati, H., & Culbertson, J. (2024). Autistic Traits, Communicative Efficiency, and Social Biases Shape Language Learning in Autistic and Allistic Learners. Cognitive Science, 48(11), e70007. doi: 10.1111/cogs.70007

Neurodiversity

Language Change

Autistic Traits

Social Biases

Linguistic Accommodation

Neurodiversity and language evolution: an experimental evolution of autistic individuals' language learning, use, and change

Thesis

Doctoral

PhD Doctor of Philosophy