Raw Meat is a quoteblog by Aaron Swartz.

Edinburgh Revisited

I’m finally getting around to reading Beyond the Hoax and really appreciating the chance to revisit Sokal’s clear-thinking elucidation of such fundamental philosophical issues. But there is one thing that I was persuaded by the first time I read some of these arguments (in the papers posted to his website and in Fashionable Nonsense), but now I’m convinced isn’t quite accurate. p. 159:

Let’s take a concrete example: Why did the European scientific community become persuaded of the truth of Newtonian mechanics somewhere between 1700 and 1750? Undoubtedly a variety of historical, sociological, ideological and political factors must play a role in this explanation — one must explain, for example, why Newtonian mechanics was accepted quickly in England but more slowly in France[footnote omittted] — but certainly some part of the explanation must be that the planets and comets really do move (to a very high degree of approximation, though not exactly) as predicted by Newtonian mechanics.[footnote omitted]

I think the issues with this can be seen more clearly if we take Sokal’s suggestion on the following page and replace the scientists with detectives. A detective is called to the scene of a murder and, after investigation, concludes the victim was killed by Rubin Carter. Why did he conclude this?, you may ask. Surely one can make the case for sociological factors playing a role. Perhaps the detective was under great pressure from his bosses to name a killer and close the case. Perhaps he held racist beliefs and thus was eager to pin the crime on a black male. If you asked the detective, presumably he’d point to the evidence for his conclusion: an eyewitness from the murder scene, a lack of alibi from Carter, etc. But surely it is of no use at all to say that the detective concluded Rubin Carter did it because Rubin Carter actually did it.

Since the detective cannot directly perceive the reality of the murder, it’s impossible for his belief to be caused by the reality without some intermediate step (the evidence). And if the evidence is sufficient to cause the belief, the belief can be held whether it is true or not (since any evidence could be caused some other way). So why should the belief being true be part of the explanation?

I think this is actually a pretty common sense point. If a friend asks me why I think the restaurant is open at this hour, I can’t just reply “because it is!” It seems odd to think sociologists of science should behave any differently.