What makes a scientific category more than a convenient label, and when do our classifications really track the causal structure of the world?
My links: https://linktr.ee/frictionphilosophy.
1. Guest
Muhammad Ali Khalidi is Presidential Professor of Philosophy at City University of New York (CUNY) Graduate Center. His work focuses on the philosophy of science, particularly cognitive science and social science, as well as some work on classical Arabic-Islamic philosophy.
Check out his book in Cambridge Elements on "Natural Kinds"!
https://www.cambridge.org/core/elements/natural-kinds/8CA215EA3A1878FC4856B84E28F4C447
https://www.amazon.com/dp/1009005065
2. Book Summary
Khalidi argues that science can’t even get off the ground without classifying its subject matter, and that a central aim of scientific taxonomy is to carve reality “at the joints” by identifying categories that correspond to natural kinds. Adopting a broadly naturalist stance, the book treats scientific practice as a (defeasible) guide to what kinds there are, while still allowing that scientists can misclassify, over-lump, or over-split. The overall structure reflects that agenda: it first lays out metaphysical background (what “natural” amounts to, realism vs. anti-realism, and pluralism), then compares leading theories of kinds, then distinguishes different kinds of kinds, and finally tests the resulting framework on concrete scientific case studies.
In evaluating theories, Khalidi highlights a widely accepted constraint on bona fide scientific categories: projectibility—roughly, whether membership in the category supports nontrivial induction (e.g., from “x is beryllium” to further expectations about melting point, etc.). He then argues that traditional essentialism and even more recent “homeostatic property cluster” approaches can be too restrictive, especially when mechanisms are absent, multiple, or not what really explains a category’s stability. In place of these, he defends a “simple causal theory” (SCT): natural kinds correspond to clusters of properties, but what makes them non-arbitrary is that the properties are bound together by causal relations (in many possible structures), so kinds are the categories that figure in generalizations that “correctly describe the causal structure of the world.”
A major payoff of the SCT is its flexibility about pluralism and crosscutting: the world may support multiple overlapping classifications, so scientists (and philosophers) needn’t choose once and for all between lumping and splitting—one can legitimately “lump for some purposes and split for others,” depending on explanatory and predictive aims. Khalidi also broadens what can count as a real kind: besides “intrinsic” kinds, there are functional kinds grounded in stable causal-functional profiles even when intrinsic make-up varies. He likewise defends etiological (historical) kinds where common origins and trajectories explain present similarities and differences. And he argues that even interactive (mind-dependent) kinds can be real when they participate in robust causal regularities, despite being partly shaped by human classification and response. The closing case studies illustrate the method: “planet” comes out as a real kind (both functional and etiological), “pandemic” (as currently used) does not, and “autism” looks like a promising candidate though it may be an umbrella that lumps multiple kinds together.
3. Interview Chapters
00:00 - Introduction
01:18 - Introducing natural kinds
04:47 - Realism and conventionalism
12:55 - Mixed views
19:27 - Folk discourse
25:05 - Conceptual engineering
27:49 - Science
30:52 - Degrees of naturalness
34:32 - How many kinds?
38:45 - Uninstantiated kinds
44:19 - Other sorts of kinds
48:14 - Ontological commitment
50:42 - Arbitrariness
54:35 - Essentialism
59:42 - Vagueness
1:03:10 - His approach
1:06:31 - Causation
1:09:24 - Another approach
1:15:42 - Non-natural kinds
1:17:32 - Value of philosophy
1:20:17 - Conclusion
