The Concept of Representation in ML: Beyond Plato and Aristotle

Toward a Radical Gradualism About Representation

It is a central task for theories of representation to give a principled account of the distinction between representational and non-representational systems. In this paper, I begin by arguing that this plausibly involves giving an account that accurately reflects differences in the explanatory practice of the sciences. I then show that one of the most sophisticated and empirically well-informed attempts to give such an account faces a fundamental problem, the problem of adjacent cases, and I suggest that we have reason to suspect that this problem generalizes to other traditional accounts. Finally, I sketch a radically gradualist alternative that solves the problem in a simple and elegant way and provides a plausible view of the transition from mere sensitivity to genuine representation.

The Cybernetic Order-word: Tensors, Tensions and LLM Vector Spaces

Why what is really a matter of data analytics and statistical prediction is so readily assumed to be a display of real intelligence and even emergent cognition is explored by genealogically tracing the relationship between machines, organisms and language.

What the mental lexicon might be

The nature of the mental lexicon remains one of the central controversies in cognitive science, bearing on fundamental questions concerning the representation and processing of linguistic knowledge. In this article, we examine Gary Libben’s contributions to the study of lexical representation and processing, with particular focus on the theoretical and empirical studies leading to the “flexicon” framework. We address four interrelated questions that have guided both Libben’s research program and broader debates in psycholinguistics: (a) What are the basic units of lexical representation? (b) Do morphologically complex words undergo decomposition during recognition? (c) To what extent are perceptual analyses constrained by morphological principles? And (d) does semantic information penetrate the earliest stages of lexical processing? Drawing on Libben’s theoretical writings and a selection of empirical studies spanning four decades of research, we situate his proposals within competing accounts of the mental lexicon, including symbolic, connectionist, Bayesian, and hybrid approaches. We argue that the flexicon framework represents an attempt to reconcile aspects of these traditions by combining morphemes and full-word forms, dynamic lexical representations, and usage-based influences within a system aimed at maximizing opportunities for lexical use. We review evidence from studies involving aphasia, deep dyslexia, compound processing, and morphological ambiguity in trimorphemic words, assessing the extent to which the findings support the framework’s central claims. While the flexicon provides an innovative perspective on lexical flexibility and morphological productivity, we argue that several questions remain regarding the role of constituent structure, morphological computation, and semantic information in lexical processing. We conclude by suggesting that the flexicon is best viewed as a research program that raises important questions about the relationship between lexical representations, linguistic structure, and language use.

Epistemic Subordination: Generative AI and the Infrastructure of Knowledge

Generative AI encodes the majority's way of knowing as the default infrastructure of knowledge itself as the default infrastructure of knowledge itself, and law must learn to govern at that level of model training.

: Linguistics as Part of the Cognitive Sciences

Cognitive Entanglement: Toward a Developmental Framework of the Human-AI Coevolutionary Leap

If AI is to support human cognitive growth, design must move beyond answer provision and efficiency maximization toward the organization of productive human-AI relations: relations that challenge users’ initial assumptions while providing support appropriate to the task and the user's level of expertise.