Perspectives on Science

Now, by the hidden and admirable Pow'r of the Loadstones, the Steel-Plates were put into motion, and consequently the Gates were slowly drawn; however, not always, but when the said Loadstone on the outside was removed, after which the Steel was freed from its Pow'r, the two Bunches of Scordium being at the same time put at some distance, because it deadens the Magnes and robs it of its attractive Virtue.

-Rabelais, Gargantua and Pantagruel, v.37

Once constituted, scientific facts have a way of roaming about on their own in the world, much divorced from the circumstances of their original constitution. An important part of Latour and Woolgar's discussion in Laboratory Life was to draw attention to how facts are used once they are at the final stage of their constitution. What I propose to do here is to go one step further, and to follow a single fact around in the wild-to tag it, as it were, much as a biologist might tag an animal with a radio collar-and then look to see where it turns up. The fact I have chosen is especially taggable, simply because it happens to be fantastic: I refer to the fact that magnets will lose their power of attraction if they are rubbed with garlic. This fact is also useful because it shows up in authors spanning fifteen centuries, from Plutarch through Rabelais and beyond, and over this time it shows some interesting behaviors. Of course, in the end, the garlic-magnet antipathy was disproved, and so changed its epistemological status, moving from one extreme to the other: from the obviously true to the obviously ridiculous. What struck Rabelais [End Page 326] as self-evident strikes us as being just as self-evidently at the opposite end of the truth-continuum. The reasons for the obviousness on both sides are not so very different, as it turns out, and in particular they share one central common feature: in each, obviousness has more to do with the classifications of facts than it does with the experiences of those facts. But the epistemologies on both sides try to tie that obviousness of kind to obviousness of experience, by surreptitiously including classification under the rubric of experience. We have been aware since Hanson that facts are messy things, and Goodman showed how that messiness of facts could in large part be accounted for by the messiness of kinds. I am adding that empirical epistemologies attempt to purify that messiness by trying to subsume kinds under experience, and that this happens symmetrically, such that under different world views, kinds get subsumed under experience even when the "experiences" are contradictory or impossible. Kinds fabricate facts, but they do so in such a way that those facts behave as though their justification is really experiential. And this so thoroughly that it may be futile to try and systematically distinguish all kind-fabricated facts from empirical facts.