John Culler

Department of Philosophy - Clemson University

Tel. 864-656-2584

Fax 864-656-2858

e-mail address: Culler@clemson.edu

Areas of professional specialization:

Philosophy of science, theories of confirmation and evidence, scientific methodology

Science Studies logo

Synopsis of Presentation:

Meeting of the Science Studies Group on Thursday, February 6th, 1997. Our visitors were two philosophers of science from Clemson University, John Culler and Tom Oberdan. From different vantage-points, both dealt with a very fundamental dilemma of empiricism and modern science: what is the role of mind and reason in an enterprise that is devoted to learning from nature? And how does factual evidence or the testimony of our senses enter into our reasoning processes since, surely, facts and sense impressions are very different from axiomatic principles in systems of pure reason?

John Culler pursues what some would call the problem of the Philosophy of Science: how can we best describe scientific method? He does not want to give in to the skeptical and disturbing questions by Berkeley and Hume. Instead, he advances and probes arguments in defense of the idea that the scientific method gives us privileged access to the world, i.e., the truth. More particularly, he asks under what conditions a scientist is justified to accept a claim upon the evidence.

To give a proper description of these conditions means two things. On the one hand, it should conform to the actual practice of scientists, but it should also serve as a norm for rational conduct: "This is how scientists use evidence to justify their claims. Moreover, this allows us to see that their claims truly are justified by the evidence." Over the years, philosophers have made numerous proposals in this regard. John presented four theories of confirmation. One is the classical hypothetico-deductive account of confirmation, another is the much-discussed Bayesian account, a third is Clark Glymour's famous "bootstrapping" positive-instance account, and then there is John Culler's own proposal (see Madison Culler: "Beyond Bootstrapping: A New Account of Evidential Relevance," Philosophy of Science, vol. 62:4, 1995, pp. 561- 579).

In light of these four competing major strands of confirmation theory, which is properly normative and descriptive of scientific practice? Episodes from the history of science are notoriously malleable and can be framed more or less to confirm all of these accounts. For the purposes of the hypothetico-deductive approach the philosophical historian of science will focus on the logical relationship between background assumptions, a proposed hypothesis, and the evidence, trying to ascertain that the hypothesis is necessary to predict or explain the evidence. The Bayesian confirmation theorist will attempt to identify how the evidence has enhanced or diminished a scientist's degree of belief or confidence in a hypothesis. And so on. In order to discover a proper account of confirmation in science, we have to find the one which best agrees with our intuitions of what constitutes good evidence. Hypothesis-testing in the realm of competing confirmation theories therefore involves the discussion of numerous hypothetical scenarios, trying to ensure that theories do not allow for counter-intuitive results. For example: does the proposed theory of confirmation allow us to treat 'old evidence' (the well-known facts for which the hypothesis was contrived in the first place) to count as confirming evidence that will increase our degree of belief in the hypothesis? How can we immunize the proposed theory of confirmation against this? For another example, John has identified a condition of adequacy which should be satisfied by any and all theories of confirmation: if some evidence gives us reason to believe that a certain generalization is true, that same evidence should also give us reason to believe that any particular instance of that generalization is true. As it turns out, only John's "entailment confirmation relation" satisfies this requirement. Since it also avoids the known counterexamples to the three other theories of confirmation, it comes highly recommended (but rest assured that the debate won't end here).

John developed his ideas most extensively in his dissertation The Course of Confirmation: An Investigation of Evidential Relevance Relations and the Verification of Phenomena. Forthcoming articles include "Hypothetico-Deductivism: New Work and Old Problems" and "The Entailment Confirmation Relation."

Alfred Nordmann