Cognitive Science of Language Lab

Evidence for the representation of non-hierarchical structures in language Nature Human Behaviour - Language is often thought to be represented through hierarchically structured units. Nielsen and Christiansen find that non-hierarchical structures are present across...

Specially, in 4 pre-registered experiments and 2 corpus analyses we find that sequences of Parts-of-Speech elements (such as noun and verb) can be primed in language processing and production

The paper itself can be read freely here: rdcu.be/eZ26u

Discovery challenges assumptions about the structure of language | Cornell Chronicle We can improvise new sentences so readily, language scientists believe, because we have acquired mental representations of the patterns of language that allow us to combine words into sentences.

Exciting new paper from the CSL Lab by @yngwienielsen.bsky.social and @mh-christiansen.bsky.social showing that we mentally represent abstract non-hierarchical structure during language use.
Here's a short write-up about the paper:
1/2
news.cornell.edu/stories/2026...

AI gør op med myten om vores indre Er menneskets unikke kommunikationsevne en evolutionær biologisk gave eller kulturelt opbygget over tid?

CSL Lab's @mh-christiansen.bsky.social is featured in a Danish article, discussing what bird alarm calls and Large Language Models can tell us about the evolution of human language and whether our unique communicative abilities relies on an innate language module
blog.minlaering.dk/blogindlaeg/...

Many congrats to Dr. Serene Wang @sere-yu-wang.bsky.social -- the newest PhD to come out of the @csl-lab.bsky.social. 🥳

She expertly defended her dissertation on "Chunking In the Second Language: Connecting Sentence Processing, Proficiency, and Memory Outcomes" 👏

LLMs highlight the importance of interaction in human language learning Recent years have seen large language models (LLMs) achieve impressive performance with core linguistic abilities. This can be taken as a demonstration that, contrary to long-held assumptions about in...

📣 New publication from the CSL Lab arguing that Large Language Models need interaction and feedback to become human-like in their production of grammatical language -- just like people.

Read the paper for free 👇

www.degruyterbrill.com/document/doi...

I was just in Tokyo a couple of weeks ago. Let’s talk next week!

📣 If you're at #AMLaP2025 in Prague, come see the poster #184 by CSL Lab's Cris Rivera and @mh-christiansen.bsky.social:

👉 "Comparing natural language statistical learning and human intuition for chunking language"

🗓️ Thursday afternoon (Sept 4), 17:20-18:50

📣Come see the two posters from the CSL Lab at #CogSci2025! 👀

👉 @elmlingersteven.bsky.social is presenting his on Friday 10:30AM-12PM

👉 Calen MacDonald is presenting his on Saturday 1-2:15PM

Language Evolution

Want a short and accessible introduction to #LanguageEvolution? 🤔

👀 Take a look as CSL Lab's @mh-christiansen.bsky.social new contribution to the Open Encyclopedia of Cognitive Science. Get up to speed in about 1K words—check out the references for more in-depth info

oecs.mit.edu/pub/18miikqb...

📣 New paper in @cp-trendscognsci.bsky.social by CSL-Lab's @mh-christiansen.bsky.social and @yngwienielsen.bsky.social reappraising what structural priming tells us about the representation of language: it's about context — not grammar.

Free download until Aug 8: authors.elsevier.com/a/1lIFK4sIRv...

The paper is a collaborative effort with Ram Frost, @bogaertslab.bsky.social, Art Samuel, @jimmagnuson.bsky.social, and Lori Holt

New paper from CSL Lab's @mh-christiansen.bsky.social and colleagues, rethinking the goal of #StatisticalLearning as providing the baseline against which novelty can be detected by a cognitive information foraging system
doi.org/10.1037/rev0...

Results are presented from a lab-based cultural evolution, computational modeling, and psycholinguistic experimentation. Together, these studies suggest that cultural evolution, constrained by chunk-based learning and processing mechanisms, has shaped the emergence of linguistic structure 4/4

He argues that, to deal with this Now-or-Never bottleneck, the brain must incrementally chunk and integrate the linguistic input as rapidly as possible before it is gone. This perspective has profound implications for the nature of language processing, acquisition, and evolution. 3/4

During normal linguistic interaction, we are faced with an immense challenge by the combined effects of rapid input, short retention of sensory information, and severely limited sequence memory. How, then, can the brain successfully handle the continual deluge of linguistic input? 2/4

In September 2024, CSL Lab's @mh-christiansen.bsky.social gave a talk in the Language Circle series on how chunking affects language across multiple timescales with implications for the nature of language processing, acquisition, and evolution:
www.youtube.com/watch?v=xOD1...

Abstract in 🧵 1/4

New CSL-Lab paper out by alumni Pablo Contreras Kallens and @mh-christiansen.bsky.social: Distributional Semantics: Meaning Through Culture and Interaction (including discussions of meaning in LLMs)
👉 Read it for free here:
onlinelibrary.wiley.com/share/author...

If you're at #psynom24 go see the poster by CSL-Lab's Serene Wang about how to best measure real-time language skills: "Utterance Recall as a Measure of Second Language Proficiency: Cross-Linguistic Applicability and Fine-Grained Automatic Scoring" # 7028 in the poster session tonight, 7:45-9:15PM.

kermit the frog is standing behind a red curtain . ALT: kermit the frog is standing behind a red curtain .

Very happy for the CSL Lab to have joined bluesky! Look here for posts about the various lab members' papers and presentations. 😀

More to come soon! 👀