Acknowledging that large language models have learned to use language can open doors to breakthrough language science. Achieving these breakthroughs may require abandoning some long-held ideas about how language knowledge is evaluated and reckoning with the difficult fact that we have entered a post-Turing test era.

Food habits vary across ethnic groups and geographical regions. However, validated dietary assessment tools accounting for such diversity remain limited. A semi-quantitative food frequency questionnaire (FFQ) was developed and validated to assess the habitual food intake of adolescents and adults across Malaysia. The 147-item FFQ was constructed using commonly consumed foods from five main ethnicities (Malay, Chinese, Indian, and Sabah and Sarawak indigenous groups) identified from national surveys. A cross-sectional validation study was conducted among purposively sampled healthy individuals aged 10–59 years from 16 administrative regions. Trained community nutritionists administered the FFQ to assess monthly intake, alongside a three-day dietary record and recall (3DRR) covering two weekdays and one weekend. Spearman’s correlation, Bland–Altman plots, and quartile cross-classification evaluated the agreement between the FFQ and 3DRR for energy, macronutrients, and selected micronutrients (Vitamin C, thiamine, calcium, and iron). Respondents (n = 361; 50.3% adults, 49.7% adolescents) were 50.4% female and represented five main ethnicities (range: 15.8–25.2%), with 60.4% from Peninsular Malaysia. Energy intake estimated by the FFQ (median: 2285 kcal) was significantly higher than by the 3DRR (median: 1785 kcal; Wilcoxon p < 0.001). Spearman’s correlation coefficients observed for energy (crude r = 0.31), and selected nutrients (energy-adjusted r range: 0.19–0.38), along with <10% of extreme quartile misclassification indicated acceptable ranking ability and agreement for most nutrients. Bland–Altman plots indicated no proportional bias for energy and macronutrients. In conclusion, the FFQ is a valid tool for assessing dietary intake within the multi-ethnic Malaysian population nationwide.

Futrell & Mahowald argue that language models (LMs) discover linguistic structure as “real patterns.” I contend this framing underplays what mechanistic interpretability uncovers: LMs implement specific algorithms to solve linguistic tasks. Under the framework of causal abstraction, we can rigorously test whether LMs converge on algorithms posited by linguistic theory, which further supports the authors’ conciliatory proposal.

Futrell and Mahowald argue that the success of large language models should move the field away from the formal structures of generative linguistic theory. The limited success of these models falls short of formal linguistic theory in explaining both the character of human languages and understanding the trajectory of child language acquisition.

The diversity of variation captured in data can strongly affect the generalization of a learning system – even when that variation occurs along axes orthogonal to the generalization in question. Thus, I argue that data richness both distinguishes current language models from prior linguistic models and may still underlie their remaining linguistic data inefficiency.

This article examines how artificial intelligence (AI) became framed as a critical node of U.S.–China strategic competition between 2015 and 2023, arguing that “hybrid epistemic experts” – figures who straddle technical expertise, corporate leadership and policy influence – played a decisive role in shaping elite understanding of AI. Through a critical review of policy documents, news media, public statements and institutional developments, this article examines how the “U.S.–China AI Race” narrative did not emerge along the usual pathways of state-driven, top-down bureaucratic processes or traditional lobbying but was actively constructed and amplified by figures like former Google executive Eric Schmidt. Schmidt’s role as a hybrid actor allowed him to translate AI from a narrow technological domain into an existential competition requiring massive policy investment. This overarching capability was driven by AI’s speculative, technically complex and general-purpose nature, which has concentrated knowledge production in private hands, enabling hybrid actors to achieve disproportionate influence over AI policy discourse. This phenomenon raises concerns about democratic governance, the collapse of independent expertise and the self-reinforcing dynamics between private power and public policymaking in emerging technologies.

Large Language Models are useless for linguistics, as they are probabilistic models that require a vast amount of data to analyze 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.

Large language models (LLMs) challenge Chomsky’s long-standing mysterian view of the creative aspect of language use (CALU). By exhibiting fluent, situation-appropriate linguistic behavior and offering concrete mechanistic hypotheses, they provide the first viable scientific models of CALU. We endorse Futrell and Mahowald’s call to integrate LLMs into linguistic inquiry and suggest a bolder aim: elucidating the mechanisms underlying linguistic creativity.

This essay considers sign-processes in their open-endedness. The concern is specifically the non-purposiveness of semiosis for what it might intimate about the provisional nature of forms and the possibility of a radical openness to self-transformation. Sikh philosophy (gurmat), semiotics, and deconstruction here offer a heterogeneous problem-space for considering the anteriority of non-purposiveness in cosmic play (līlā) and equipoise (sahaj). The study first finds a recurrent commitment to the non-purposiveness of play across diverse approaches in metaphysics, modern aesthetic theory, and the more recent so-called ludic turn, developing from this thread a constructive treatment of cosmic play consisting of latitude, excess, indeterminacy, generativity, and expenditure, or endless textural differentiation. This essay then elaborates an account of equipoise in practical involvement whose pursuit of action is alive to that which is imperceptible amidst this endless textural differentiation, developing an account of action that realizes its own ludic workings in a non-coercive creativity as intimated by Peircean considerations of musement and cosmic love. Findings offer the study of signs an orientation to the un-fore-seen (im-pro-visus) constitutive of semiosis, in doing so heightening for social semiotic analysis a sensitivity to the indexical excess that conditions the formation of any text. In each instance, Sikh philosophy is found to offer key sources for considering semiosis as open-ended, non-purposive, and ludic-equipoisal.

We argue that language models (LMs) have strong potential as investigative tools for probing the distinction between possible and impossible natural languages and thus uncovering the inductive biases that support human language learning. We outline a phased research program in which LM architectures are iteratively refined to better discriminate between possible and impossible languages, supporting linking hypotheses to human cognition.

According to F&M, both infants and language models (LMs) find attested languages easier to learn than “impossible languages” that have unnatural structures. We review the literature and show that LMs often learn attested and many impossible languages equally well. Difficult to learn impossible languages are simply more complex (or random). LMs are missing human inductive biases that support language acquisition.

Futrell and Mahowald frame the success of neural language models (LMs) as supporting gradient, usage-based linguistic theories. I argue that LMs can also instantiate theories based on formal structures – the types of theories seen in the generative tradition. This argument expands the space of theories that can be tested with LMs, potentially enabling reconciliations between usage-based and generative accounts.

Futrell and Mahowald argue that neural networks have learned non-trivial aspects of language. We argue that these systems have not in fact demonstrated “mastery” of syntax, marshalling recent evidence, and that they further obscure explanatory insights with respect to topics in the cognitive neuroscience of language.

Futrell and Mahowald present a useful framework bridging technology-oriented deep learning systems and explanation-oriented linguistic theories. Unfortunately, the target article’s focus on generative text-based Large Language Models (LLMs) fundamentally limits fruitful interactions with linguistics, as many interesting questions on human language fall outside what is captured by written text. We argue that audio-based deep learning models can and should play a crucial role.

In this commentary, we challenge the idea that Language Models (LMs) provide explanatorily adequate models of human language. Findings from language evolution show that both humans and LMs fail to produce human-like language via inductive biases alone; the communicative function of language is crucial. More generally, both experimental and computational work underscore the fact that language is more than just incremental prediction.

Futrell and Mahowald argue language model (LM) success suggests humans may learn language entirely through domain-general statistical mechanisms. However, children differ crucially from LMs in their ability to surpass their input, their language learning trajectory, and the presence of a critical period. Until LMs account for these phenomena, it remains possible that human language acquisition is supported by innate, language-specific learning mechanisms.