These essays critically reassess the widely accepted view that statistical methods of analysis can, and do, yield causal understanding of social phenomena. They emphasize the historical, philosophical and conceptual perspectives that underlie and inform current methodological controversies…. More >>
Causality in Crisis: Statistical Methods and the Search for Causal Knowledge in the Social Sciences
This book offers interesting reading, particularly the Carnegie Mellon contributions. But I am disappointed in the book’s atavistic philosophy of science – notably about causality.
In his “Introduction” McKim distinguishes experimental and nonexperimental testing for identifying causality. I see this difference as merely a technical matter about testing that affects the degree of confidence in a theory, not a philosophical issue about causality.
When McKim writes of “underlying structure of causal relationships” I believe he has misconceived causality. In his book titled Patterns of Discovery: An Inquiry Into the Conceptual Foundations of Science philosopher Russell Hanson wrote that cause words are “theory-loaded” (p.60). This is a variation on Willard Quine’s ontological relativity. Causality is a reality. Thus causal claims are ontological claims, and all questions of ontology are subordinated to the empirical adequacy of a tested theory. And that includes models. I am surprised that McKim, once a student of Hanson’s at Yale, has overloked these contemporary pragmatist views.
Causality is claimed by a nonfalsified theory, whether the theory’s testing was experimental or nonexperimantal, and a theory’s causal claim is acceptable until falsified empirically.
I also believe the authors have misconceived language, when they are dismayed that the same explained data admit to multiple models that are equally adequate empirically. As Quine notes, all language is empirically underdetermined, and that includes models.
These contributors seek a Holy Grail called “causality”, and so they are in crisis. But there is no need to unravel the Gordian knots in these papers to find causality. Tested theories make causal claims, all theories are subject to further testing, and all causal claims are subject to revision. And they are often revised; there are no guarantees in science.
Contemporary pragmatism thus cuts through their Gordian knots, and consigns their mythical Holy Grail to the dustbin of history of philosophy. For more extensive comments Google to my site, History of Twentieth-Century Philosophy of Science, for free downloads.
Thomas J. Hickey, Econometrician
Rating: 3 / 5