Linguistic inputs must be syntactically parsable to fully engage the language network.

Human language comprehension is remarkably robust to ill-formed inputs (e.g., word transpositions). This robustness has led some to argue that syntactic parsing is largely an illusion, and that incremental comprehension is more heuristic, shallow, and semantics-based than is often assumed. However, the available data are also consistent with the possibility that humans always perform rule-like symbolic parsing and simply deploy error correction mechanisms to reconstruct ill-formed inputs when needed. We put these hypotheses to a new stringent test by examining brain responses to a) stimuli that should pose a challenge for syntactic reconstruction but allow for complex meanings to be built within local contexts through associative/shallow processing (sentences presented in a backward word order), and b) grammatically well-formed but semantically implausible sentences that should impede semantics-based heuristic processing. Using a novel behavioral syntactic reconstruction paradigm, we demonstrate that backward-presented sentences indeed impede the recovery of grammatical structure during incremental comprehension. Critically, these backward-presented stimuli elicit a relatively low response in the language areas, as measured with fMRI. In contrast, semantically implausible but grammatically well-formed sentences elicit a response in the language areas similar in magnitude to naturalistic (plausible) sentences. In other words, the ability to build syntactic structures during incremental language processing is both necessary and sufficient to fully engage the language network. Taken together, these results provide strongest to date support for a generalized reliance of human language comprehension on syntactic parsing.

Lexical-Semantic Content, Not Syntactic Structure, Is the Main Contributor to ANN-Brain Similarity of fMRI Responses in the Language Network.

Representations from artificial neural network (ANN) language models have been shown to predict human brain activity in the language network. To understand what aspects of linguistic stimuli contribute to ANN-to-brain similarity, we used an fMRI data set of responses to n = 627 naturalistic English sentences (Pereira et al., 2018) and systematically manipulated the stimuli for which ANN representations were extracted. In particular, we (i) perturbed sentences' word order, (ii) removed different subsets of words, or (iii) replaced sentences with other sentences of varying semantic similarity. We found that the lexical-semantic content of the sentence (largely carried by content words) rather than the sentence's syntactic form (conveyed via word order or function words) is primarily responsible for the ANN-to-brain similarity. In follow-up analyses, we found that perturbation manipulations that adversely affect brain predictivity also lead to more divergent representations in the ANN's embedding space and decrease the ANN's ability to predict upcoming tokens in those stimuli. Further, results are robust as to whether the mapping model is trained on intact or perturbed stimuli and whether the ANN sentence representations are conditioned on the same linguistic context that humans saw. The critical result-that lexical-semantic content is the main contributor to the similarity between ANN representations and neural ones-aligns with the idea that the goal of the human language system is to extract meaning from linguistic strings. Finally, this work highlights the strength of systematic experimental manipulations for evaluating how close we are to accurate and generalizable models of the human language network.

Event Knowledge in Large Language Models: The Gap Between the Impossible and the Unlikely.

Word co-occurrence patterns in language corpora contain a surprising amount of conceptual knowledge. Large language models (LLMs), trained to predict words in context, leverage these patterns to achieve impressive performance on diverse semantic tasks requiring world knowledge. An important but understudied question about LLMs' semantic abilities is whether they acquire generalized knowledge of common events. Here, we test whether five pretrained LLMs (from 2018's BERT to 2023's MPT) assign a higher likelihood to plausible descriptions of agent-patient interactions than to minimally different implausible versions of the same event. Using three curated sets of minimal sentence pairs (total n = 1215), we found that pretrained LLMs possess substantial event knowledge, outperforming other distributional language models. In particular, they almost always assign a higher likelihood to possible versus impossible events (The teacher bought the laptop vs. The laptop bought the teacher). However, LLMs show less consistent preferences for likely versus unlikely events (The nanny tutored the boy vs. The boy tutored the nanny). In follow-up analyses, we show that (i) LLM scores are driven by both plausibility and surface-level sentence features, (ii) LLM scores generalize well across syntactic variants (active vs. passive constructions) but less well across semantic variants (synonymous sentences), (iii) some LLM errors mirror human judgment ambiguity, and (iv) sentence plausibility serves as an organizing dimension in internal LLM representations. Overall, our results show that important aspects of event knowledge naturally emerge from distributional linguistic patterns, but also highlight a gap between representations of possible/impossible and likely/unlikely events.

Lexical semantic content, not syntactic structure, is the main contributor to ANN-brain similarity of fMRI responses in the language network.

Representations from artificial neural network (ANN) language models have been shown to predict human brain activity in the language network. To understand what aspects of linguistic stimuli contribute to ANN-to-brain similarity, we used an fMRI dataset of responses to n=627 naturalistic English sentences (Pereira et al., 2018) and systematically manipulated the stimuli for which ANN representations were extracted. In particular, we i) perturbed sentences' word order, ii) removed different subsets of words, or iii) replaced sentences with other sentences of varying semantic similarity. We found that the lexical semantic content of the sentence (largely carried by content words) rather than the sentence's syntactic form (conveyed via word order or function words) is primarily responsible for the ANN-to-brain similarity. In follow-up analyses, we found that perturbation manipulations that adversely affect brain predictivity also lead to more divergent representations in the ANN's embedding space and decrease the ANN's ability to predict upcoming tokens in those stimuli. Further, results are robust to whether the mapping model is trained on intact or perturbed stimuli, and whether the ANN sentence representations are conditioned on the same linguistic context that humans saw. The critical result-that lexical-semantic content is the main contributor to the similarity between ANN representations and neural ones-aligns with the idea that the goal of the human language system is to extract meaning from linguistic strings. Finally, this work highlights the strength of systematic experimental manipulations for evaluating how close we are to accurate and generalizable models of the human language network.

The neural architecture of language: Integrative modeling converges on predictive processing.

The neuroscience of perception has recently been revolutionized with an integrative modeling approach in which computation, brain function, and behavior are linked across many datasets and many computational models. By revealing trends across models, this approach yields novel insights into cognitive and neural mechanisms in the target domain. We here present a systematic study taking this approach to higher-level cognition: human language processing, our species' signature cognitive skill. We find that the most powerful "transformer" models predict nearly 100% of explainable variance in neural responses to sentences and generalize across different datasets and imaging modalities (functional MRI and electrocorticography). Models' neural fits ("brain score") and fits to behavioral responses are both strongly correlated with model accuracy on the next-word prediction task (but not other language tasks). Model architecture appears to substantially contribute to neural fit. These results provide computationally explicit evidence that predictive processing fundamentally shapes the language comprehension mechanisms in the human brain.