Large Language Models Demonstrate the Potential of Statistical Learning in Language

To what degree can language be acquired from linguistic input alone? This question has vexed scholars for millennia and is still a major focus of debate in the cognitive science of language. The complexity of human language has hampered progress because studies of language–especially those involving computational modeling–have only been able to deal with small fragments of our linguistic skills. We suggest that the most recent generation of Large Language Models (LLMs) might finally provide the computational tools to determine empirically how much of the human language ability can be acquired from linguistic experience. LLMs are sophisticated deep learning architectures trained on vast amounts of natural language data, enabling them to perform an impressive range of linguistic tasks. We argue that, despite their clear semantic and pragmatic limitations, LLMs have already demonstrated that human-like grammatical language can be acquired without the need for a built-in grammar. Thus, while there is still much to learn about how humans acquire and use language, LLMs provide full-fledged computational models for cognitive scientists to empirically evaluate just how far statistical learning might take us in explaining the full complexity of human language.

Statistical learning or phonological universals? Ambient language statistics guide consonant acquisition in four languages

What predicts individual differences in children’s acquisition of consonant production across languages? Considerations of children’s development of early speech production have traditionally emphasized inherent physiological constraints of the vocal apparatus that speakers generally have in common (i.e., articulatory complexity). In contrast, we propose a statistical learning account of phonological development, in which phonological regularities of the ambient language guide children’s learning of those regularities in production. Across four languages (English, Spanish, Japanese, and Korean), we utilized recent meta-analytic dataset of age of consonant acquisition spanning 28 studies. High-density measures of children’s ambient language environment from over 8,000 transcripts of speech directed to over 1,000 children were used to assess how well the frequency of consonants in child-directed speech predict the age of consonant acquisition. Our results suggest that both frequency and articulatory complexity are related to age of acquisition, with similar results found for English, Spanish, Japanese, and Korean. Consonants heard frequently by children tended to be incorporated into their production repertoires earlier and consonants heard less frequently are incorporated into production repertoires later in development. We discuss future directions that incorporate a statistical learning pathway towards learning to produce the sound patterns of the ambient language.

Bibliometric analysis of the phenomenology literature

It is not clear to what extent these divisions are reflected in the organization of work done within the field, as expressed by patterns of communication between authors. To assess this, we analyze the phenomenology literature using bibliometric methods (Osareh, 1996). This makes it possible to supplement an intuitive understanding of its structure with an empirically grounded analysis of citation patterns. In particular, we extract an author-wise citation network (Radicchi et al., 2012) for the published phenomenology literature, a network of nodes and connections, where nodes correspond to authors of articles or books about phenomenology, and connections correspond to citations from one author to another. The resulting graph has 11,980 nodes and 69,324 connections.1 We then study clusters in this network, that is, groups of authors who cite each other more than they cite authors in other clusters and compare this more bottom-up image of communication dynamics withinthe field to the different sub-groups identified in expert historical reconstructions.

Networks in philosophy: Social networks and employment in academic philosophy

In recent years, the “science of science” has combined computational methods with novel data sources to understand the dynamics of research communities. Many of the questions investigated by science of science are also relevant to academic philosophy. To what extent can the discipline be divided into subfields with different methods and topics? How are prestige and credit distributed across the discipline? And how do these factors interact with other factors, such as gender, to shape job market outcomes? Using job market data for anglophone academic philosophy, this paper finds, first, evidence that is consistent with the analytic-continental divide but is also consistent with other, more complex ways of organizing academic philosophy into distinct intellectual traditions; second, a clear prestige hierarchy, dividing Ph.D. programs into two distinct prestige categories; and, third, evidence that gender, prestige, and country have notable effects on academic job market outcomes for recent philosophy Ph.Ds.

We need a comparative approach to language acquisition: A commentary on Kidd and Garcia (2022)

The study by Kidd and Garcia is long overdue. Their analyses of published research on language acquisition highlight the lack of typological diversity in studies of how children acquire their native tongue. We concur with their conclusion that more research on understudied languages is urgently needed. However, we argue that what the field needs is not just wider cross-linguistic coverage but a systematic comparative approach to language acquisition – one in which investigations of well-studied languages still has much to contribute.

Toward a comparative approach to language acquisition

The world’s languages vary in almost every conceivable way, yet children readily learn their native language. Understanding how children can acquire such a diversity of different languages has been a long-standing goal for psychological science, yet current acquisition research is dominated by studies of children learning one particular language: English. In this article, we argue that progress toward this goal will require systematic comparisons between different languages. We propose three levels of comparison: coarse-grained comparisons contrasting unrelated languages to confirm or refute broad theoretical claims, fine-grained comparisons between closely related languages to investigate the impact of specific factors on acquisition outcomes, and within-language comparisons targeting the impact of socio-communicative differences on learning. This three-pronged comparative approach to language acquisition promises to provide new insights into the mechanisms and processes by which children acquire their native tongue under such varied linguistic and socio-communicative conditions.

Quantifying Interdisciplinarity in Cognitive Science and Beyond

Recent publications have lamented the dominance of psychology in cognitive science. However, this relies on a limited definition of collaboration between fields. We call for a renewed conception of interdisciplinarity as a “mixture of expertise.” We describe an information-theoretic measure of interdisciplinarity and apply it to multiauthored published articles. Results suggest that cognitive science journals mix expertise more than topically related journals. We suggest that perceptions of diminishing interdisciplinarity may in part be due to the emergence of different theoretical perspectives and use a semantic model to illustrate this argument. We conclude by describing some benefits of this broader conception.

Models of Language and Multiword Expressions

Traditional accounts of language postulate two basic components: words stored in a lexicon, and rules that govern how they can be combined into meaningful sentences, a grammar. But, although this words-and-rules framework has proven itself to be useful in natural language processing and cognitive science, it has also shown important shortcomings when faced with actual language use. In this article, we review evidence from language acquisition, sentence processing, and computational modeling that shows how multiword expressions such as idioms, collocations, and other meaningful and common units that comprise more than one word play a key role in the organization of our linguistic knowledge. Importantly, multiword expressions straddle the line between lexicon and grammar, calling into question how useful this distinction is as a foundation for our understanding of language. Nonetheless, finding a replacement for the foundational role the words-and-rules approach has played in our theories is not straightforward. Thus, the second part of our article reviews and synthesizes the diverse approaches that have attempted to account for the central role of multiword expressions in language representation, acquisition, and processing.

Dissipation, Integration and Practical Pluralism: The Case of Cognitive Science

In this chapter, we focus on (i), the theoretical usefulness of pluralism. We are interested in how a robust pluralism could influence theoretical debates in different scientific fields. Our area of expertise, cognitive science, will serve as the primary vehicle for our discussion. To organize this discussion, we identify a tension that simmers under much work on pluralism, between “dissipative” and “integrative” pluralism. The former highlights differences among theories, seeing proliferating accounts of even similar phenomena as drifting apart, taking on their own character. A purely dissipative pluralism would lead to distinct accounts for virtually every single observable phenomenon. In contrast, integrative pluralism highlights potential similarities and seeks to link theories by various formal or informal strategies. In an extreme form, an integrative pluralist may invest too much in seeking overly abstract linkages that account for very few specific phenomena or get caught in an attempt as futile as the fundamentalist’s to integrate over all our diverse knowledge, but in ever more abstract ways.

Phonological cues to semantic membership across hundreds of languages

Categorization is a fundamental function of minds, with wide ranging implications for the rest of the cognitive system. In humans, categories are shared and communicated between minds, thus requiring explanations at the population level. In this paper, we discuss the current state of research on the cultural evolution of categorization. We begin by delineating key properties of categories in need of evolutionary explanation. We then review computational modeling and laboratory studies of category evolution, including their major insights and limitations. Finally, we discuss remaining challenges for understanding the cultural evolution of categorization.

Open science and modified funding lotteries can impede the natural selection of bad science

Assessing scientists using exploitable metrics can lead to the degradation of research methods even without any strategic behaviour on the part of individuals, via ‘the natural selection of bad science.’ Institutional incentives to maximize metrics like publication quantity and impact drive this dynamic. Removing these incentives is necessary, but institutional change is slow. However, recent developments suggest possible solutions with more rapid onsets. These include what we call open science improvements, which can reduce publication bias and improve the efficacy of peer review. In addition, there have been increasing calls for funders to move away from prestige- or innovation-based approaches in favour of lotteries. We investigated whether such changes are likely to improve the reproducibility of science even in the presence of persistent incentives for publication quantity through computational modelling. We found that modified lotteries, which allocate funding randomly among proposals that pass a threshold for methodological rigour, effectively reduce the rate of false discoveries, particularly when paired with open science improvements that increase the publication of negative results and improve the quality of peer review. In the absence of funding that targets rigour, open science improvements can still reduce false discoveries in the published literature but are less likely to improve the overall culture of research practices that underlie those publications.

Cultural evolution of categorization

Categorization is a fundamental function of minds, with wide ranging implications for the rest of the cognitive system. In humans, categories are shared and communicated between minds, thus requiring explanations at the population level. In this paper, we discuss the current state of research on the cultural evolution of categorization. We begin by delineating key properties of categories in need of evolutionary explanation. We then review computational modeling and laboratory studies of category evolution, including their major insights and limitations. Finally, we discuss remaining challenges for understanding the cultural evolution of categorization.

Interacting Timescales in Perspective-Taking

Through theoretical discussion, literature review, and a computational model, this paper poses a challenge to the notion that perspective-taking involves a fixed architecture in which particular processes have priority. For example, some research suggests that egocentric perspectives can arise more quickly, with other perspectives (such as of task partners) emerging only secondarily. This theoretical dichotomy–between fast egocentric and slow other-centric processes–is challenged here. We propose a general view of perspective-taking as an emergent phenomenon governed by the interplay among cognitive mechanisms that accumulate information at different timescales. We first describe the pervasive relevance of perspective-taking to cognitive science. A dynamical systems model is then introduced that explicitly formulates the timescale interaction proposed. This model illustrates that, rather than having a rigid time course, perspective-taking can be fast or slow depending on factors such as task context. Implications are discussed, with ideas for future empirical research.

Exploratory mapping of theoretical landscapes through word use in abstracts

We present a case study of how scientometric tools can reveal the structure of scientific theory in a discipline. Specifically, we analyze the patterns of word use in the discipline of cognitive science using latent semantic analysis, a well-known semantic model, in the abstracts of over a thousand academic papers relevant to these theories. Our results show that it is possible to link these theories with specific statistical distributions of words in the abstracts of papers that espouse these theories. We show that theories have different patterns of word use, and that the similarity relationships with each other are intuitive and informative. Moreover, we show that it is possible to predict fairly accurately the theory of a paper by constructing a model of the theories based on their distribution of word use. These results may open new avenues for the application of scientometric tools on theoretical divides.

Palabra y concepto: acercamiento a un eliminativismo conceptual en ciencia cognitiva

En este artículo, me concentro en una pregunta de carácter metateórico respecto de lateoría de conceptos en ciencia cognitiva: ¿es necesaria la postulación de conceptos? Pararesponderla, inicio mi argumentación desde el punto de partida de que los conceptosson entidades teóricas inobservables postuladas con fines explicativos y de coherenciacon una teoría. Me baso en esto para dividir los desiderata de una teoría ideal presentadospor Fodor (1998) y Prinz (2002) en desiderata explicativos y desiderata teóricos.Los desiderata teóricos son sólo compromisos con la estructura de la Teoría Representacionalde la Mente, por lo que no es menester aceptarlos. Así, identificaré los explanandade la postulación de conceptos mediante el análisis de los desiderata explicativos. Unavez definido este punto, presentaré tanto descripciones alternativas de los fenómenoscomo explicaciones plausibles de ellos en esos términos. Con esto, pretendo minar loscimientos de lo que ha sido considerado como el mayor argumento a favor de la existenciade conceptos: su exclusividad como explicación de sus explananda. Concluyo conalgunas observaciones acerca de las consecuencias teóricas y metateóricas de los argumentosdesarrollados.