The Golem
November 1, 2007
Following some heretical comments by me regarding the nature of scientific activity, I was referred by commenter “dsquared” to a book called “The Golem” by two sociologists, Harry Collins and Trevor Pinch. This fairly reasonable book makes the point that, despite pretenses by scientists and non-scientists, science is not a formal system in which crisp models generate crisp predictions which can then be neatly confirmed or refuted by experiments with unambiguous results. This is a good point, but the book somehow still manages to miss, or at least to obscure, the main issue – as dsquared’s summary of the “the basic conclusions of the literature” shows.
By discussing mainly “big science” (relativity, spontaneous generation, gravity waves, solar neutrinos and cold fusion) and paradigm threatening ideas (chemical transfer of memory and cold fusion), the book seems to imply what dsquared thinks it does, namely, that normal scientific activity breaks down in certain, rare, problematic cases. In fact, I believe the authors make a better point, namely, one close to the one I am making (repeatedly) – that science, including normal, everyday science, is a political rather than a rational activity – but their presentation obscures this point, inadvertently or deliberately.
It is therefore the chapter about the small, run-of-the-mill science, the sex life of the whiptail lizard, that is the best one in the book. Even here, the profile of the events is much higher than usual, as the appearance of a Time magazine article about the matter shows. However, despite some attempt on the side of Collins and Pinch to explain the importance of the zoological issue, to the reader (to me, at least) the whole point seems so trivial that it is quite clear that the science of the matter (in its pure, mythical, objective sense) is of marginal impact on the events described, while personal politics is at the forefront.
It seems, as I wrote above, that Collins and Pinch do in fact know that science in general, rather than only revolutionary or controversial science, is political. For example, they do occasionally (but only occasionally) mention the issue of the careers of the scientists involved (p. 133-134, for example, discuss physicist John Bahcall’s maneuvering on the missing neutrinos issue in the light of potential impact on his career). Why, then, do they soft-sell or obscure this point enough so that dsquared can misunderstand it?
The likely answer is, again, political. Even the relatively benign reading of the ideas in the book, leading to the understanding of scientific politics as put forward by dsquared, is surely unpleasant to many (and in particular to many established, powerful scientists) and is thus dangerous to the authors. Going the whole way would be much more dangerous. Not wishing to take such a large risk, but not wanting to deny their own understanding, Collins and Pinch compromise – they write their findings in such a way that it can be interpreted by the reader according to his tastes. I have mine.
Pro Bono Statistics 
December 4, 2007 at 8:32 am
dsquared is Daniel Davies of Crooked Timber.
December 6, 2007 at 7:11 pm
Hmm, it’s been a long time (12 years of so?) since I read the book, but I was underwhelmed. My thought on why they focused on big, revolutionary changes is science was not because they thought that was safer, but because they thought that was more dangerous. Although in my opinion they never really had a “so what?” point, I think they wanted to show that science was a human activity even in major advances, and as such the whole edifice is somehow questionable or compromised.
Whereas talking about some minor issue in narrow, run of the mill science, the fact that personal and political issues are part of the process is probably less surprising to people, and isn’t seen as a big deal.
It makes a big splash to say that a huge experiment on General Relativity was compromised. It doesn’t have much impact to say that personality conflicts enter into evaluation of results on the third order resonant spectroscopy of heavy metal doped II-VI semicondcutors at ultralow temperatures, or some other obscure field that only has 3 people doing research on it.
I had three main complaints about the book (again working from memory.) First, I think much of their argument is somewhat banal–science doesn’t work in a clean, perfect, ideal fashion as it is usually described in a high school science class. But I don’t know that their examples prove that much, unless you had a very naive view of how the scientific process works.
Accepting their arguments, there’s still the question of what it means. While they seemed to want their results to imply large and important, given that science seems to produce theories that work and can be used in technology, apparently all the messiness they worry about doesn’t really matter.
A second complaint was that, in several of their examples, they focused exclusively on experimental evidence, but ignored existing theoretical arguments which properly biased researchers for or against particular positions.
There were reasons aside from lab results why most people were skeptical of cold fusion claims. It wasn’t an issue of power and politics. Similarly with the relativity example, the experiment is treated as a one off, yes/no proof, ignoring all the other outside reasons why most theorists were pretty sure that General Relativity was valid. They ignore these contextual issues which help to explain a lot of what they passed off as arbitrary or political decisions.
A third complaint is that they cherry picked examples to emphasize ones where political or outside issues played a role. But there are many, many examples where the scientific process is clean and close to the ideal. Where experimental tests were performed and gave clear and unequivocal confirmation of a theory or disproved it. What’s missing from their analysis is an appreciation of this or any attempt to quantify how often there are equivocal results that end up influenced by political or other issues.
December 8, 2007 at 7:21 am
Hi Doug,
I disagree with much of what you say. For one thing, you seem to hold the self-contradicting view that the Golem thesis is both “banal” and mostly untrue (based on “cherry picked examples”).
I also think that the authors make their “so what” point (even if you restrict that to mean a point with clear practical implications) quite explicitly. In their concluding remarks they state that in their view the misunderstanding of science makes it a less-useful tool for policy decisions. I recall that they specifically discuss the use of science in the courtroom as one area where a better understanding could have important implications.
As the example of forensic science indicates, your view of science as something that produces technology is much too narrow. Even in that function, the political nature of science influences which technologies are produced, but the effect of science goes much further.
In particular, for scientists science is a source of livelihood and a process by which status and social relations are determined. Non-scientists are also affected by science in many ways beyond technology. For example, science is used to influence public opinion. In all those effects, the fact that science is mostly political rather than mostly rational, is a decisive factor.
Thanks for making your arguments and focusing attention on those matters. I hope to elaborate on them in future posts.
December 11, 2007 at 6:29 pm
Thanks for the response. I was working on a fuzzy memory at this point, so I must have forgotten the argument they made about the implications.
As far as contradicting myself, I think it’s banal to observe that politics can play a role in exactly how science advances or how it is carried out. It is a human activity, and so interpersonal and political factors are going to be involved.
But I think they cherry-picked examples in which these political factors supposedly played a decisive role in undercutting a truly logical result (although as I argued, I was not convinced in several of their case studies that the process really was undercut by political factors, or whether there were outside logical factors beyond the ones which they considered, which explained their observations.) And then using these specific examples tried to argue that all of science is similarly affected.
I think even in some of their examples science doesn’t work like they claim, and that you could come up with many cases in which the progress of theory and understanding was fairly close to the ideal.
I guess it comes back to the so what question. I think the presence of politics is not in general that important in its impact, but they cherry picked examples to try and make it look really important and decisive. Maybe the distinction is that I’d completely disagree with your addition of the word “mostly,” when claiming that science is mostly political rather than mostly rational. I’d say it’s partly political (and not in a decisive way, generally) but mostly rational.
December 12, 2007 at 6:10 am
Look – this little argument between us is prototypical. It demonstrates exactly the process that I am talking about: We are talking about phenomena that are supposedly observable facts. How is it then that we both look at the same world and draw different conclusions?
At least one of us is irrational (i.e., draws the wrong conclusions from the observable facts). Yet, there is no clear cut way to demonstrate which of us is wrong. I have my examples you no doubt have yours. If the scientific process was as you think it is, you should have been able to demonstrate that I am wrong in a way that any person with a functioning brain and no clear over-riding interest would accept.
(I also invite you to follow the “repeatedly” link in the post above and tell me how you view that matter.)
December 16, 2007 at 4:29 pm
But this is an argument about social science, which is a different animal, and not one that most folks would consider to be in the same class as the hard sciences. The same kingdom, certainly, and maybe the same phylum, but still not that close of a relative. Even worse, we’re arguing about the degree of an effect in sociology, which gets us even more into the fuzz.
If you really wanted to try and prove one side or the other, what would be required would be a large scale survey of a variety of scientific research, something I certainly don’t have time to do (and also something the authors of the Golem didn’t do.) Short of that, it’s dueling anecdotes and generalities. And even if you did such a thing, you’re building a castle on sand because you’re at bottom trying to quantify a non-measurable entity, the influence of political factors vice experimental and logical ones. So I disagree that our disagreement says anything at all about the scientific process.
I have no doubt there are some cases where poltical considerations play a decisive role, at least in the short term. On the other hand, back when I was working in science myself I didn’t see much evidence for that at the working level. So I started out skeptical of the claims in the Golem.
And I found the book unconvincing because in the two sections about which I knew the most (the physics examples), I thought their argument was flawed and didn’t hold up. They oversimplified the case and ignored good prior reasons for skepticism (in the case of cold fusion) or belief (in the case of general relativity), treating both as single experimental tests performed on a blank slate.
As far as the argument in your link, I mostly agree. Peer review is not proof of value, nor is lack of peer review proof of invalidity. Although I think the fact that a theory or scientific claim has never passed peer review does give you some information about it, and makes it less likely to be true than one which has.
As far as the general issue of testing hypoteses is concerned, at the bottom scientific proof or disproof is statistical rather than formal. All measurements have error so it’s possible that any theory at all might hold, just that all the measurements made have been wildly unlikely outliers. But I think in most cases you can get a pretty huge statistical sample and very tiny liklihoods of error, which from a practical standpoint is as good as an absolute proof.
December 16, 2007 at 5:27 pm
A few points:
1. I don’t find the distinction between “hard” sciences and social sciences to be convincing. For one thing, it is usually made by people who are in the “hard” sciences business, and so it is self-serving.
It is true that social sciences deal with more complex phenomena and have more difficulty with experimentation than hard sciences, but the real problem of social sciences is that they have political implications. Whenever “hard” science has some political implications, we meet much of fuzziness phenomenon that you associate with social sciences. See for example the dynamics in the scientific establishment which occur when dealing with the matters of global warming, nuclear waste disposal and many medical issues.
2. My personal experience (in a “hard” science field), and that of others that I have talked to, is that personal politics is ubiquitous in science. All decisions about which papers get published and who gets credit for certain “discoveries” or “developments” are political. Considering that careers are at stake, it would also be very surprising if it were any different.
3. I was talking about the other link from my post:
p-rep-and-the-myth-of-rational-science. This talks about a specific case of scientific politics. The case is extreme in the sense that the science involved is neither social nor “hard” – the issue is one of pure mathematics: one in which the correctness or incorrectness of the argument should be verifiable in a completely crisp, rigorous way. It turns out that even in such cases, politics trump rationality.
November 2, 2008 at 11:23 am
Hello everybody in here. I saw all comments of yours and I decided to ask your help about “The Golem” Collins and Pinch. Actually I was asked to do a presentation on chapter7 “Solar Neutrinos”, but I can’t get the real point of authors. If you do not mind could you help me to better understand this chapter in order to make a good presentation. This is my email address: ayubxon1988@gmail.com
I would be delighted to receive your response. I need your help and suggestion. Thank you in advance.
Sincerely,
Alex
November 2, 2008 at 3:19 pm
Alex,
Welcome. I would be glad to help, but you have to show some willingness to put in some effort – no learning can occur without effort. Why don’t you lay out your thoughts on the matter? It will get the conversation going.