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

Abstract

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.Significance statementWhether language comprehension relies predominantly on structural (syntactic) cues or meaning- related (semantic) cues remains debated. We shed new light on this question by examining the language brain areas’ responses to stimuli where syntactic and semantic cues are pitted against each other, using fMRI. We find that the language areas respond weakly to stimuli that allow for local semantic composition but cannot be parsed syntactically—as confirmed in a novel behavioral paradigm—and they respond strongly to grammatical but semantically implausible sentences, like the famous ‘Colorless green ideas sleep furiously’ sentence. These findings challenge accounts of language processing that suggest that syntactic parsing can be foregone in favor of shallow semantic processing.

A 3.5-minute-long reading-based fMRI localizer for the language network

“…This network supports computations related to goal-directed behaviors and is recruited during a broad array of cognitively demanding tasks (e.g., Duncan, 2010;Duncan et al, 2012;Fedorenko et al, 2013;Shashidhara et al, 2019;Assem et al, 2020b;Duncan et al, 2020). Of most relevance to the current investigation, the MD network appears to be engaged in some cases of effortful comprehension, including processing speech in noisy conditions (Mattys & Wiget, 2011;MacGregor et al, 2022;Liu et al, 2022), processing accented speech (Adank & Janse, 2010;Janse & Adank, 2012;Adank et al, 2012;Banks et al, 2015), processing non-native languages (Malik-Moraleda, Jouravlev et al, 2024;Wolna et al, 2024), and processing linguistic inputs that are not syntactically well-formed (Kuperberg et al, 2003;Nieuwland et al, 2012;Mollica et al, 2020;Tuckute et al, 2024;Kauf et al, 2024). However, the full range of conditions under which the MD network is recruited during language processing is not well-understood, yet is critical for understanding the contributions of this network to comprehension.…”

“…In contrast to the costs associated with linguistic processing specifically (e.g., processing unexpected elements or non-local inter-word dependencies; Shain, Blank et al, 2020;Shain et al, 2022), some cases of effortful comprehension, even without external task demands, appear to engage the MD network. Such cases include listening to speech in noisy conditions (Mattys & Wiget, 2011;MacGregor et al, 2022;Liu et al, 2022), processing accented speech (Adank & Janse, 2010;Janse & Adank, 2012;Adank et al, 2012;Banks et al, 2015), processing sentences in non-native languages (Malik-Moraleda, Jouravlev et al, 2024;Wolna et al, 2024), and processing linguistic inputs that are not syntactically well-formed (Kuperberg et al, 2003;Nieuwland et al, 2012;Mollica et al, 2020;Tuckute et al, 2024;Kauf et al, 2024). A possible generalization about these cases is that they all involve difficulty extracting a syntactically parsable word sequence from perceptual linguistic inputs.…”