Ohio University Philosophy Forum Public Lecture: Dr. Cameron Buckner

Thirty-First Annual Ohio University Philosophy Forum Generative Deep Neural Networks as models of imagination, creativity, and planning” A Public Lecture by Dr. Cameron Buckner, University of Florida Thursday, November 5th, 2026, 4:00PM - 6:00PM, Scripps Hall 111 LECTURE ABSTRACT: In current debates in Artificial Intelligence, deep learning researchers have noted a historical kinship to the view in philosophy called "empiricism", which holds that abstract knowledge is produced by learning from experience. Critics of deep learning have in turn taken up mantle of philosophical "rationalism", which denies that the most sophisticated forms of human cognition can be derived from learning. The stakes of this philosophical debate are vividly illustrated in a disagreement between the empiricist David Hume and the rationalist Jerry Fodor over the role that the imagination plays in rational cognition, which centers on the question as to whether statistical learning procedures could be bootstrapped to perform distinctively human forms of creativity. Fodor alleged that empiricists are unable to explain how minds synthesize examples for use in reasoning, compose novel concepts from simpler concepts, engage in consequence-sensitive planning, or distinguish between causal and intentional relations amongst thoughts. Fodor applauds Hume for observing that rational cognition involves all these creative operations, but criticizes Hume’s delegation of these operations to the faculty of the imagination, because Hume never explains how learning could enable the imagination could perform all these feats. In this talk, I show how a variety of different "generative" deep learning architectures demonstrate how an empiricist imagination could perform such operations. I conclude with more general morals about the prospects for abstraction and reasoning in empiricist theorizing, about a mutually beneficial interaction between philosophy and computer science in this context, and about how