I seem to have a gift for being drawn into trenchant conversations. The latest involves the usefulness of philosophy of science:
I stated:
…normative PoS has little bearing on what we (a geotechnical engineering consulting firm) do. This worries me a bit.
LolaWalser, asked:
Why does this worry you?
I think that the philosophy of science should work to advance the sciences and, in turn, the sciences should take methodological recommendations (in part) from PoS. I see problems in both directions. I worry that philosophers, with some important exceptions, do not perform the requisite work to bring PoS into the implementation level, and scientists ignore the potentially helpful insights and recommendations of philosophers and other methodologists.
For example, many sciences need to select models (curves) from data in order to make predictions and make important decisions. Model selection is a nontrivial philosophical problem with practical consequences. The usual method in many businesses involves using Excel’s built-in curve fitting by selecting a trend line from a set of equations of varying degrees and eyeballing the fit. This is “good enough” in some ways, but it certainly could be better, given the research done in the subject, and given the problem of over fitting (that is, ramping up the number of parameters to perfectly fit the data as well as noise).
A great example of applied PoS is Causation, Prediction, and Search, Second Edition by Peter Spirtes, Clark Glymour, and Richard Scheines. Therein they set forth a formal theory of causal inference involving Bayesian networks. Blurb: “The authors show that the relationship between causality and probability can also help to clarify such diverse topics in statistics as the comparative power of experimentation versus observation, Simpson’s paradox, errors in regression models, retrospective versus prospective sampling, and variable selection.” The authors have (with the help of others) developed TETRAD– a software implimentation of their philosophical/formal ideas.
Your thoughts? Can/should philosophy of science be applied? How? Examples?
LolaWalser’s condescending reply and more below the fold:
I think that the philosophy of science should work to advance the sciences and, in turn, the sciences should take methodological recommendations (in part) from PoS.
Do you also think philosophy of art should work to “advance art'”? What about, um, the philosophy of LOVE and its subject? ๐
Frankly (and I’m fairly sure I speak for a large number of my colleagues) I have no interest in philosophy of science–NOW, after 16 years of benchwork. Once upon a time, a gentler, high-schoolish time, I may have entertained some such notions of philosophy–well, not exactly “advancing” science (science has done remarkably well all on its own, pre- and post-Popper et al.)–but at least, I don’t know, mattering to it. As it turns out (she said, eyeing her Philosophiae Doctor wistfully), it doesn’t.
But perhaps I’m a narrow-minded intellectual hick–and after all, what’s true for my science/field may not be true for another. Still, as far as I can glean from my fairly wide trade reading (I have never in my entire professional career TALKED about PoS with my colleagues–nobody brings it up), it doesn’t appear scientists in general think PoS any more useful or pertinent than I do. What’s the problem?
Could it be that it’s naive, useless and ultimately simply wrong to imagine modern philosophers “advancing” science, because they seem to have no idea of the realities and practicalities of research?
scientists ignore the potentially helpful insights and recommendations of philosophers and other methodologists.
I’d love to see a demonstration of a real life example of a scientific study which would have profited from professional insight from philosophers.
If I need to learn better modelling or statistical analysis, I’ll turn to mathematicians; for help with experimental design, first of all to people with practical experience. I couldn’t begin to trust a person with no practical knowledge of experimental research (students, of course, are there to gain that knowledge). I couldn’t begin to trust someone who has had less math than I did (we might run into some geographical/national differences here as well… European scientific education, at pre-grad school levels especially, tends to be considerably more rigorous than the American one) and above all, did not use those skills in the situations I’m encountering.
Can/should philosophy of science be applied? How? Examples?
My answer: Doesn’t apply/Don’t know any. However! One way to determine whether PoS can contribute to science would be to have mature, finished PoSers (the more advanced their degrees the better) study science–go through the whole process, of course, grad school to postdoc to, ahem, probably another postdoc…
If these candidates perform significantly better than mere non-philosophical scientists (appropriately matched for age etc.–maybe also throw in a control group of, I don’t know, sociologists? art historian-scientists?–don’t want my experimental design called in question here!), your case is made.
A great example of applied PoS
I’ll look this up and report if I find a way to use it.
Lastly, I hope I didn’t sound unfriendly–and there was no intention to offend. Simply trying to convey a bit of your average scientific Joe’s view.
I tried my best to be civil in my response:
LolaWalser,I had anticipated this sort of reaction from a scientist. Generally, there is a great deal of contempt for philosophy, perhaps due to its construal as being synonymous with fluffy, pomo, intellectual onanism. This is wrong. I’ll try to explain.
Frankly (and I’m fairly sure I speak for a large number of my colleagues) I have no interest in philosophy of science–NOW, after 16 years of benchwork.
This is a strange group for you to sign-up for. ๐
I’d love to see a demonstration of a real life example of a scientific study which would have profited from professional insight from philosophers.
When I have time, I’ll chose a specific example I am most familiar with, but for the time being:
Scientific disciplines founded by or having major contributions from philosophers: logic (Aristotle, Frege, Russell and countless others), economics (Adam Smith, Hume), psychology (William James), computer science (Godel, Church, logician precursors), probability theory, welfare economics (Rawls, Nozick), sociology (Auguste Comte, Marx), physics (Mach, Reichenbach)…
Methodological and conceptual contributions: inductive methods (Bacon, Mill), bayesian inference methods (Carnap), bayesian statistics (Ramsey, Seidenfeld), causal discovery (Spirites, Glymour), belief revision (Levi), object language/metalanguage distinction (Carnap), the role of simplicity in inference (Aristotle, Ockham)…
Problems faced by scientific inquiry and identified by philosophers: the problem of induction (Hume), underdetermination of theory by evidence (Quine), Dutch book problem (Ramsey)…
Specific software for discovery: *edit* MYCIN (created by a student of Hempel’s), TETRAD…
The history of science is to a large degree intermingled with philosophy. See History of Tweentieth Century Philosophy of Science for more details.
Now, you might be tempted to make an ad hoc move to say that when it was helpful, it was science (or math/logic), and when not, it is philosophy, but that is a philosopher’s (re)definition game. ๐
No offence taken, by the way.
LolaWalser couldn’t resist further rudeness:
I had anticipated this sort of reaction from a scientist.
Glad I didn’t disappoint!
Generally, there is a great deal of contempt for philosophy,
There is? You shock me. Yes, I’m well aware of the travails of postmodernism and its spats with science, but, first, we don’t all need to be lectured on what philosophy is (or was) beyond this category or cajoled into “tolerating” it. I love classical philosophy and if I were talking to an ancient Greek I’d very proudly describe myself as a philosopher to hiim. (I still sometimes talk of science as “practical philosophy” but only to people who don’t need my background and idiosyncrasies explained.)
This is a strange group for you to sign-up for. ๐
Actually, I didn’t sign up for it, nor do I intend to. Unless you wish to pay me to jeer from the sidelines… ๐ No, sorry, as I said I personally have no interest in PoS, but I will be checking occasionally to see what sort of crowd you draw.
When I have time, I’ll chose a specific example I am most familiar with, but for the time being:
Really, a specific example of a scientific study which would profit from professional philosopher’s insight is all I need. I certainly DON’T need this appallingly condescending rundown of history of philosophy and science. Are you SERIOUS?! I’ve been to school, sir–probably for much longer than you. I’ve even read a few books on the subjects!
Look, I’ll try to keep my good humour (I started replying to your post as I was reading it, never expecting this ridiculous “History of philosophy for dummies” capsule) and assume you’re simply naive, or have never in your whole life talked to a scientist. How in the world do you expect me to have graduated HIGH SCHOOL without philosophy? History? Logic? Mathematics?
Piece of advice: never underestimate the general education of a stranger. Is it because Americans seem to have lower expectations when it comes to schooling?
Now, you might be tempted to make an ad hoc move to say that when it was helpful, it was science (or math/logic), and when not, it is philosophy, but that is a philosophers (re)definition game. ๐
Is it now? Told you I’m a philosopher!
It is quite clear philosophers are feeling a bit rudderless these days, toppling from all-encompassing “love of wisdom” to… indeed, I’m not sure what.
Trying to maintain the peace, I ignored LolaWalser’s tone and tried to address her concerns:
Wow. I didn’t expect such a relatively rarified subject to recieve so much immediate attention– I am pleasantly suprised. I should clarify my modest claims.
PoS, in order to be a more fruitful discipline, should pay attention to implementation-level concerns in science and try to meet what I call Hausman’s challenge (“Philosophers of science have learned a great deal about science, but the knowledge falls short of any usable algorithm for scientific practice or theory choice,รขโฌย Daniel M. Hausman, The Philosophy of Economics: An Anthology, Introduction): PoS should aim for providing such usable methods, though they have thusfar not (largely) met this goal. PoSt is well situated to do this because of its theoretical, interdisciplinary nature and emphasis on logical and mathematical tools. Also, philosophy has always been the incubator of sciences
I do not claim that PoS is, or should be, the sole authority in things methodological, nor that it always meets this goal. Some work by philosophers of science is junk (maybe even a majority of it), but this does not mean that philosophers have nothing to contribute.
My infodump in #5 was meant to illustrate the many contributions of philosophers to science, many of which can be used to refute the claim that philosophers do not, or cannot, advance science, as I understood LolaWalser to be making.
LolaWalser:
[Quoting a response I made to another discussant] Biochems have more rigorous standards than some disciplines, particularly those out of the research field, and in the applied sciences. My point is that methods should be backed by sound theory.
More rigorous standards of what?
Next, is your contention that biochemistry and applied sciences have laboured all this time employing methods without the backing of “sound theory”? If it is, can you show it is true?
Now, assuming that it is true (to speed things up–I personally think it’s a ridiculous statement–Not Even Wrong), what does it tell you, if the achievements of science so far have been produced “without” the backing of sound theory–such as you envisage?
And thank you for this quote, it provides a fitting conclusion:
“Philosophers of science have learned a great deal about science, but the knowledge falls short of any usable algorithm for scientific practice or theory choice,”
Yeesh. Again, trying to clarify:
Biochems have more rigorous standards than some disciplines, particularly those out of the research field, and in the applied sciences. My point is that methods should be backed by sound theory.
More rigorous standards of what?
More rigorous standards of experimentation, modeling and methodology in general.
Next, is your contention that biochemistry and applied sciences have labored all this time employing methods without the backing of “sound theory”? If it is, can you show it is true?
Now, assuming that it is true (to speed things up–I personally think it’s a ridiculous statement–Not Even Wrong), what does it tell you, if the achievements of science so far have been produced “without” the backing of sound theory–such as you envisage?
To quote a scientist, “There is nothing more practical than a good theory.” We are able to do quite a lot without theory in many areas of human endeavor. However, theory may allows us to systematize, optimize, predict and control where we could not do so before. I could imagine breeders making the same argument before Mendelian genetics.
What makes you think that science as currently practiced is optimal? How would you suggest our methods be improved, if not by reasoning about and a theory of its methods? This is just PoS. Can you give me a specific example of PoS and how it is as useless as you claim?
…
Philosophers of science are concerned with the understanding the nature, methods, theoretical products and success of the sciences. PoS uses the traditional methods of philosophy (conceptual analysis, explication) to this end, but also mathematical, logical and naturalistic methods are used as well. If you wish to call people who pursue such questions methodologists or gavagai, I don’t care– I’m just interested in the questions. Basically, science involves the use of hypothesis formation, modeling, experimentation, and test in an effort to discover, explain, predict and control natural phenomena. PoS and science are primarily distinguished by the level of abstraction and can benefit from each other (see Machine Learning and the Philosophy of Science (PDF) by Jon Williamson for an excellent overview of a mutually beneficial interaction between the disciplines).
I sometimes wonder why I bother with these exchanges.
Labels: philosophy of science
