Project Awarded: $6,615
Psychologists study human cognition indirectly by positing constructs such as “episodic memory” and “executive control.” Collectively, these constructs form a “cognitive ontology” (Price and Friston 2005)—a taxonomy of scientifically legitimate psychological kinds. There is now much scientific interest in using neuroscience—particularly functional neuroimaging techniques—to test our cognitive ontology (Poldrack 2010, Lenartowicz et al. 2010, Anderson 2015). For example, Lenartowicz et al. (2010) applied pattern classifiers to fMRI data to test a hypothetical set of cognitive control constructs—e.g., “response inhibition” versus “task switching.” A key assumption of their work, shared by many others (e.g., Price and Friston 2005, Lindquist et al. 2012, Anderson 2014) is that our best psychological theories should align with observed patterns of brain activation.
The idea of using neuroimaging data to revise our cognitive ontology—viz., to discover new cognitive kinds, eliminate existing ones, etc. (Price and Friston 2005, Anderson 2015, Polger and Shapiro 2016) is fraught with empirical and philosophical challenges. One issue is whether informatics efforts—meta--‐analytic databases such as Neurosynth (Yarkoni et al. 2011), the Cognitive Atlas (Poldrack et al. 2011), etc.—can improve inferences from neural to cognitive states (Sullivan 2017). Another is whether analysis techniques such as dimension reduction analyses (Anderson 2014) and multivariate techniques (e.g., multi--‐voxel pattern analysis--‐MVPA) can be used to discover novel cognitive constructs, or to test our cognitive ontology (McCaffrey and Machery 2016, Kaplan and Craver 2016).
The current proposal aims to foster collaboration between philosophers, neuroscientists, cognitive scientists, and computer scientists interested in cognitive ontology. Philosophers studying the mind--‐brain relationship should be aware of developments in the area of cognitive ontology. And, crucially, empirical work on cognitive ontology should be informed by philosophical concerns about the relationship between mechanisms and kinds (Craver 2009, Polger and Shapiro 2016), on the inferential limits of neuroimaging techniques (Ritchie, Kaplan, and Klein 2017), etc. We plan to run a workshop on cognitive ontology at Washington University in St. Louis with and to publish a volume or special journal issue.