My research is organized around two interconnected projects in the philosophy of mind and artificial intelligence. The first concerns the nature of semantic memory. Against the long-standing view that semantic memory is a stored repository of facts, I develop a proceduralist account on which it is a constructive, embodied capacity to construct knowledge. The second asks how artificial systems reshape human epistemic agency, that is, what kinds of epistemic agents are AI systems shaping us to become? 

Proceduralism about Semantic Memory

My dissertation builds on research in philosophy and cognitive science to develop a new account of semantic memory. While episodic memory has recently been the primary focus of philosophical inquiry into memory systems, semantic memory has often been treated as theoretically unproblematic: a storehouse of factual knowledge, language use, and general information about the world. But the conceptual underpinnings of this “orthodox view” are far less stable than often assumed. It presupposes a dedicated storage space of explicit facts, shaped by unconscious mechanisms of abstraction and generalization. Yet this view faces persistent challenges from both philosophical and empirical work: the lack of evidence for stable, dedicated semantic storage; difficulties locating or characterizing the neuroanatomical substrates of semantic memory; and the rise of enactivist approaches that reject stored content altogether.

In response, I propose a proceduralist alternative. On this view, semantic memory is not stored content but the procedural skill to construct semantic knowledge. Rather than treating semantic recall as retrieval from a mental database, we can understand it as the exercise of learned procedural capacities for constructing knowledge, shaped by prior experience, embodied habit, and social practice. This account integrates computational multi-trace approaches to memory, enactivist accounts of cognition, and insights from deep learning, while drawing on Wittgenstein’s ungrounded hinges, and Rowlands’s Rilkean memory to develop semantic memory as a procedural capacity for constructing knowledge. 

Accordingly, I understand semantic memory construction as taking multiple, mutually reinforcing, forms, among them mnemonic effort and mnemonic habit. Mnemonic efforts are the construction of semantic knowledge through the conscious use of prior experience and often take the form of skill based memory strategies. Mnemonic habits are the habitual use of memory attributable to embodied processes that become causally connected to the constructive processes of mnemonic efforts. This constructive approach also allows us to critically examine the ways in which the metaphors of “storage and search” influence our conceptualization of memory and memory traces as storage systems. My constructive approach imagines the (re)constructive operations of memory like the folding and sculpting techniques one uses in origami. On the “origami view” of memory, memory traces are like sheets of origami paper out of which, using the right set of procedural skills, a memory can be (re)constructed. In these cases the procedural skills “unfold” in ways characteristic of prior constructive processes, thus potentiating a memory. 

My long-term goal is to develop these ideas into a book-length treatment of semantic memory, offering both a critical examination of the dominant view and a positive account of memory as a dynamic, constructive procedural capacity. Beyond that project, I plan to extend my procedural account into a broader account of memory and human cognition, investigating how flexible, knowledge-like capacities emerge from learned constructive procedures.

Epistemic Agency and AI Mediation

My second line of research examines how generative AI reshapes human epistemic agency. I take up the question of how habitual use of AI systems can reorganize the epistemic relations through which people exercise judgment and come to stand behind what they take themselves to know. This project explores the question: what kinds of epistemic agents are AI systems shaping us to become?

My current work is organized around two connected ideas: the Cyrano de Bergerac Problem and what I call testimonial outsourcing. The Cyrano de Bergerac Problem, named for the play of the same name, identifies a form of mediated speech in which the apparent speaker relies on an external source for the epistemic work of determining what they are prepared to say and stand behind. Assistant- and companion-style AI systems instantiate this structure in a technologically intensified form. The ease of the exchange and the system’s apparent epistemic authority can make it seem better positioned than the user to determine what should be said, gradually drawing users away from their own judgment and toward an artificial interlocutor as the place where their claims are settled.

Testimonial outsourcing describes the state that results when repeated routing of one's epistemic activity through such systems alters the conditions under which users form and own their judgments, as authority over what they are prepared to say shifts toward the system. This weakens the user's ability to stand behind their own judgments, as they remain answerable for claims whose content and warrant have increasingly been settled through the AI system. Over time, this pattern erodes cognitive and epistemic self-trust: users come to regard their first-order judgment as provisional until it has been confirmed by the system. In this respect, testimonial outsourcing functions as a technologically mediated analogue of gaslighting: the user’s own epistemic position begins to appear deficient in advance, while an external system is elevated as the authority before which their judgment must be settled.

My long-term goal for this project is to extend this research into a broader account of AI-mediated cognitive agency. Testimonial outsourcing provides the initial case, showing how artificial interlocutors can displace the capacities through which users form and answer for their own claims. Future work will examine how increasingly invasive forms of AI mediation affect human cognitive capacities, including memory, attention, and perception, as these systems become more deeply embedded in everyday life and more directly involved in the exercise of cognition.