Significance in the 20th century

Sat 28 Jan 2006 10:23 AM

Working on the d-cog paper and teaching Understanding Science again have got me ruminating on scientific significance.

In The Advancement of Science, Philip Kitcher first advocated the view that science aims not at truth but at significant truth. At the time, he treated significance as an objective feature of some truths. To set up what I say below, here is an excerpt from a paper I wrote in Spring '97:

* * * begin flashback * * *

Kitcher sees science as aiming to adopt significant truths. Traditional attempts to understand scientific significance in terms of systems of universal generalizations have led to problematic schemes to measure truth content. So, he writes:

My approach circumvents these difficulties by offering a quite different view of scientific significance. A significant statement is a potential answer to a significant question. What we strive for, when we can get them, are true answers to significant questions. [p.118]
The explanatory schemata of the concensus practice suggest questions of intrinsic significance, and many other questions derive significance as a step toward answering these primary questions. Significance, then, is a shared sense that points the scientific community to investigate different things. When a significant truth is discovered, it is taken up into concensus practice. ... Although Kitcher assumes that significance is uniform across the community, it's not clear why that should be so. A field biologist might find considerable significance in the migratory habits of birds, but little or none in molecular genetics. Conversely, another biologist might find great significance in genetics but none in migrations. This is not simply a matter of caprice-- suppose the first biologist's work relies on her using the best available information regarding migration, but nothing relies on what she think's about the bird's genome. Even if researchers differed over the significance of certain truths, we would like poeple who apply scientific discoveries to be working with the best that science has to offer. Although it may not be critical for the discovery of further significant truths that a clinician use the best medical knowledge we have, it may be necessary for the survival of his patients. So, significance is tied to practical concerns of two kinds: (i) the discovery of further significant truths and (ii) the achievement of certain technological goals.

It follows immediately that the sectors of the scientific community which ought to hold the community's best candidates for truth are those for whom that truth would be a significant truth. Consider, for example, information concerning cancer. Doctors treating cancer patients should clearly employ those beliefs most likely to heal their patients. Given Kitcher's realism, this is just to say they should employ those beliefs most likely to be true. Cancer researchers, in order to discover new truths, ought to begin by employing background knowledge that is likely to be true. Other members of the scientific community (other doctors, marine biologists, theoretical physicists) needn't believe the community's best candidates for the truth about cancer at all, unless doing so is required to have oncologists believe it. So, the aim of science need not be conceived as adopting significant truth into consensus practice or even as spreading particular truths as widely as possible. Instead, all that matters is that truths are held by the people who actually need them; that is, people for whom they are significant. If the class of people who found a question significant were small, then the aim of science would be consistent with the majority of scientists believing anything whatsoever about it. Many of them may well cling to a falsehood, but why should this matter if it's not a significant falsehood for them? ... The notion of significance developed above is, in a sense, recursive. ... Clause (i), by referring to significant truths, forces us back into the definition. Only facts used to achieve technical objectives are significant in themselves. Clause (ii) is just about technological proweress. Does truth then just reduce to the ability to perform technical feats? The concern is that proper deference to the significance of significant truths makes them all significance and no truth. How might such bald pragmatism be softened?

* * * end flashback * * *

In Science, Truth, and Democracy, Philip changed his view and argued that significance depends on what we care about. He also offered an answer to the objection: Significance is not necessarily practical. There are some questions we are motivated to ask just on the basis of natural human curiosity.

The appeal to natural human curiosity always strikes me as thin and philosophically unsatisfying. Nevertheless, I feel the visceral appeal of it. It is simply cool to learn about dinosaurs, for instance, and that coolness lends some significance to paleontological research. It is not simply that constructing big skeletons in museums is cool, because we could more easily construct fictional but impressive dragon skeletons. The coolness of dinosaurs is due, in part, to the fact that they really did exist and that our account of them-- although not true in all its details-- is based on the best evidence available.

Call this the dinosaur argument against bald pragmatism.