S e l e c t e d P u b l i c a t i o n s
—”How Much is Left for the Theory-Theory to Explain?” (2026) – In Behavioral and Brain Sciences (commentary) Sept 18.
— “Shared Moral Foundations of Embodied Artificial Intelligence” (2019) – Proceedings of the Association for the Advancement of Artificial Intelligence.
— “Parsimony and the Triple System Model of Concepts” (2011, with Safa Zaki) In Behavioral and Brain Sciences (commentary) 33(2).
— “Is There a Reason for Skepticism?” – (2010) In Knowledge and Skepticism (O’Rourke and Campbell, eds.).
— “The Chimerical Appeal of Epistemic Externalism” (2004, with John Pollock) – In The Externalist Challenge (Richard Schantz, ed.).
— “Epistemology” and “Simulation Theory” (2003) – Entries in the Nature Encyclopedia of Cognitive Science (the latter written with Bob Gordon).
— Contemporary Theories of Knowledge, (Rowman & Littlefield, 1999) co-authored with John Pollock, friend, mentor, and collaborator.
— “Mindreading: Mental State Ascription & Cognitive Architecture” (Mind & Language, 13(3), 1998).
S e l e c t e d T a l k s & P u b l i c L e c t u r e s
—“Illusionism and Artificial Consciousness” University of Glasgow, Glasgow, Scotland, 2026.
Abstract
I’m going to defend two claims that you’re likely to find implausible. Firstly, I’ll argue that illusionism —or, at any rate, a view that is recognizably a kind of illusionism—is true about consciousness. Secondly, I’ll claim that even though familiar current LLMs are very likely not conscious, it is not conceptually difficult to see how they easily might someday, even soon, be. My version of illusionism maintains that our conception of phenomenal consciousness is the byproduct of an inability to represent our fine-grained cognitive, embodied, perception-and-action systems. We misunderstand this self-opacity as the presence of real phenomenal properties when in fact there are none. Instead of asking what sort of special property consciousness is, we should aim to explain why we think that there is such a special property. I am proposing an answer to the latter question. This shows why LLM consciousness is not difficult to conceive. What is required is that the LLMs’ self-representational schemes are inadequate to characterizing their own finest-grain processing, and that they represent the mismatch in such a way as to be ineliminably opaque. An LLM might then understand itself as conscious in just the way that we understand ourselves as conscious. Current LLMs are somewhat ambiguous on this front. They are representationally heterogeneous, but the heterogeneity might be the wrong kind. I’ll try to say something more about what the right kind is, but I admit that LLMs might indeed already be conscious. Handout
—“How Does Having a Body Make a Difference?” University of South Florida, Tampa, FL, 2025.
Abstract
The embodied cognition (EC) research program in cognitive science maintains that understanding mind requires taking the body seriously. On this view, no purely mentalist or internal-state account—and certainly no abstract computational model—can fully capture how biological minds work. In its most radical form, EC holds that the proper unit of explanation for intelligence is the whole organism acting within an environment and in interaction with others. In this talk, I ask what difference the body is supposed to make. Why aren’t representations of bodily states enough? Could a virtual body suffice? And in the case of disembodied generative AI, might a world model derived from the language of embodied beings be enough? I argue that embodiment does make a genuine difference, though saying precisely why turns out to be more difficult than it first appears. Handout.
— “Artificial Intelligence and Ethics” – AI: Philosophy, Ethics and Society, Tehran, Iran, 2019.
Abstract
In this talk, I aim to make explicit how artificial intelligence is a distinct area of moral study separate from wider discussions of technology. The opacity and autonomy of reasons that AIs employ makes them moral agents in their own right, and we need to consider this agency in relation to ours. That consideration applies, of course, to future maximally sophisticated artificial general intelligence, but also to the proto-AI systems that presently have a significant presence in our lives. Overall, I argue that the most pessimistic worries about AI are overstated. In order to achieve intelligence in its full generality, AIs will need to be embodied and will therefore share our moral foundations, namely coordination, sociality, and acknowledgement of shared resources.
— “Shared Moral Foundations of Embodied Artificial Intelligence” – AAAI/ACM Conference on Artificial Intelligence, Ethics, and Society, 2019.
— “Knowing How to Reason” – Epistemic Virtues and Epistemic Skills Conference, 2017.
— “Attention and Reason” – The Intersection of Epistemology and Philosophy of Mind Conference, 2015.
— “Why There Is No Mind/Body Problem” – TEDx Williams, Winter 2014.
— “Irrational Thought” – First Conference in Honor of John Pollock, 2012.
— “Two Worlds” – Models of Mind Gallery Talk, Williams College Museum of Art, 2010.
— “Knowing One’s Mind” – Campus faculty lecture, Spring 2007. Podcast, transcript.
P o p u l a r A u d i e n c e W r i t i n g
— “Knowledge and Reasons” (2013) Philosophy Now. Vol 96, May/June, 19-22.