Built on Facts

If you’re reading this the morning it’s published, there’s a good chance that right at this very second I am sitting with pen in hand doing battle with a statistical mechanics final. Topics: fermions at zero and low temperature, virial expansion of the equation of state, critical constants and critical exponents of a Van der Walls like equation, and critical constants and exponents in an Ising lattice. The first two (I hope) aren’t so bad, the third is difficult, the fourth may be well-nigh impossible. We’ll see.

I do have something for y’all to cogitate over while I’m sweating. I came across a Vox Day post that I think warrants some commentary. He’s a semi-frequent bete noire around ScienceBlogs by virtue of being (among other things) a radical libertarian Christian skeptic of global warming, vaccination, and science in general.

He presents a list of definitions of science in various science textbooks. The definitions tend toward the ambiguous, and a few are borderline contradictory. Almost all of them are from biology texts. I own well over two dozen physics textbooks, and on a little thinking and poking, I note that as far as I know not a single one of them tries to define science. I’m pretty sure none of them even try to define physics. Even the freshman intro text only makes passing comment about how physics requires measurement before launching into a discussion of the SI units.

While I’m not certain this was the conscious intent of the authors, I suspect the reason is that physics is so obvious as to not need explanation. It’s like a cookbook defining cooking. Why bother? If you end up with edible food, you’ve cooked. If you end up with a mathematical statement of the way nature is observed to work, you’ve done physics.

So am I being too flippant, or is physics just a lot easier to define than science in general? I know science isn’t all that hard to get a handle on, but of course there’s intrinsically some fuzziness or else that whole “philosophy of science” thing would have been put to bed a long time ago. Not that that’s necessarily a bad thing, mind you. But I wonder if some of the less mathematical science textbooks dither too much in trying to cover every possible base in their definitions. It’s not so hard to do science. Step 1: Get some facts. Step 2: Build on ‘em.


  1. #1 speedwell
    May 13, 2009

    Vox Day is as libertarian as he is Venusian, Matt. Insane /= libertarian.

  2. #2 Russell
    May 13, 2009

    The problem is that homeopaths, to take an example, claim they are doing science exactly as you prescribe, taking facts, and building on them. It takes some effort to point out how they go wrong. This leads philosophers to want to describe the One, True Right way to do science as a well-specified methodology. While the philosophers and the methodologists and especially the statisticians have improved how science is done, I don’t think they ever will succeed in finalizing The Method, because a) there are many ways to go wrong, not all of which we have discovered, and b) those who want to cloak their nonsense in the mantle of science will make sure to describe it as following the latest proposed recipe.

    I’m content with the notion that there always will be some fuzziness about what science is, just because we will no more wrap up methodological issues once and for all than we will wrap up science once and for all. Were I to attempt a definition that allows for that, it would be something like:

    Science is the search for descriptive theories based on empirical evidence, with a collective refinement of methodologies and disciplined effort for intellectual honesty.

    It is the “intellectual honesty” part that disqualifies most pseudo-science, no matter how much it is couched in the latest methodological rhetoric.

  3. #3 Russell
    May 13, 2009

    BTW, there are plenty of mathematical descriptions of how nature works in chemistry and biology and some even in the social sciences. So, it may take a bit more to say what is physics.

  4. #4 Eric Lund
    May 13, 2009

    Step 1: Get some facts. Step 2: Build on ‘em.

    You omitted Step 1.5: Verify that these facts are indeed factual.

    Russell correctly points out that homeopaths and other alt-med types do your Steps 1 and 2. The problem is that they take whatever principles their alt-med is based on as given, without actually verifying that they are factual. The handful of formerly alt-med practices that have graduated into science-based medical practice are the ones for which somebody performed Step 1.5 and found that it passed.

    It’s also closer to how science is supposed to operate. You have a hypothesis which explains some previously observed phenomenon. Based on that hypothesis, you predict the outcome of some experiment. You then verify the prediction, and if it is correct you have an additional fact on which to build.

  5. #5 Steven N. Severinghaus
    May 13, 2009

    While I’m not certain this was the conscious intent of the authors, I suspect the reason is that physics is so obvious as to not need explanation.

    That’s certainly a possibility. More likely it’s that biology (evolution) has been under attack for decades by people desperate to caricature it as unevidenced hand-waving. Authors or publishers of biology texts have had to go on the defensive in a way that is simply not needed in other fields. Part of that defense is to explain the framework of science in generic terms so that the reader can see that biology (evolution) is just as much of a science as physics.

  6. #6 Uncle Al
    May 13, 2009

    Science is a predictive mathematical mole of observation subject to falsification. In contrast,


    1) That which supports religion supports religion.
    2) That which ignores religion supports religion.
    3) That which contradicts religion supports religion – test of faith!
    4) Anybody who criticizes is thereby proven unqualified to comment – and must be destroyed lest god(s) take offense.

    Step 1: Get some facts. Step 2: Build on ‘em.

    Yea, right – economics, psychology, string theory. A sine function can be aribtrarily perfectly fit by a polynomial. Try extrapolating the fit.

  7. #7 Josh in California
    May 13, 2009

    Vox Day is as libertarian as he is Venusian, Matt. Insane /= libertarian.

    Ah, the old no-true-libertarian fallacy. ;-)

  8. #8 Jason Baur
    May 13, 2009

    As others have already said, science is properly understood as methodology. You describe physics (and cooking) in terms of results. This is problematic. It is possible to produce a mathematical model of the universe without doing science in any meaningful fashion. This is what creationists attempt when they describe the speed of light as changing over time, for example. Additionally, it is possible to obtain edible food without cooking at all. Cooking is a process of altering or combining edible materials. A cookbook does not have to define cooking because we all understand what cooking is, and the Bible doesn’t insist that cooking involves some ineffable act of God.

  9. #9 Peter Morgan
    May 13, 2009

    Step 0.5: decide what/where your starting point is and who your role models are, largely determining what facts you will see/notice, and in what conceptual structure you will place what you notice for future reference. Known in short-hand as the theory-ladenness of observation. Some ideas you will not take on trust when your role models tell you, forcefully, that you must; clarifying or modifying one or some of those ideas may become your unique contribution.

    Anything can be studied meticulously, accurate records kept of experiments, concepts revitalized a few times by new generations of researchers who take extraordinary pains to show the old guard how to change without selling out their principles, and gradually general consent may be given that it has become a science. Being a science is not a big deal in itself, however, what’s important is the extent to which practitioners are intellectually content with how things are done in their field (or creatively content, or spiritually content, …, whatever is important in that field). Interaction with other systems may result in something that the pioneers might not recognize, but perseverance over decades or centuries can set right a lot of problems, if there are enough able people who believe it to be worthwhile to spend their lives.

    A photography blog has an Ezra Pound definition of science by story-telling that may be of interest: http://theonlinephotographer.blogspot.com/2007/03/portraits-of-scientists.html
    The (many!) stories that scientists tell about themselves are more definitional than single sentences that start “Science is …”. Biographies, notes in monographs on what the author took to be guiding principles, etc, … , which typically will be at a level of detail beyond easy summary. Methodology of Science is a very problematic area in Philosophy of Science, particularly since Kuhn, Feyerabend, Lakatos, and Rorty (to name a few), problematized the positivist world-view. Those difficulties have I think been a significant part of the impetus to transform Philosophy of Science into History and Philosophy of Science.

  10. #10 Matt Springer
    May 13, 2009

    The exam is done. My prediction wasn’t too far off. Problems 1 & 3 are mostly right, problem 2 is completely right, problem 4 is fubar.

    W/r/t the libertarian & “no true Scotsman” thing, I should point out that in fact some people who claim to be Scottish aren’t. At any rate libertarianism is not something where there’s a clean definition because it’s intrisically a very individualistic and personalized view of the world. To use a less controversial word, I think Vox fairly cleanly fits the definition of an anarchist.

    For full disclosure, libertarianism also an ideology for which I have some sympathy. I’m not one myself, but generally I side with Jefferson on the whole “An honest man can feel no pleasure in the exercise of power over his fellow citizens” thing.

  11. #11 percy
    May 13, 2009

    From your April 18th post:

    “In the physical sciences and physics especially, science follows an infinite loop procedure. I’ll label them 1 and 2, though it’s really a chicken-and-egg thing. This is the way physics works, simplified:

    1. The experimental results produce data which suggest ways old theories could be refined, replaced, or strengthened.
    2. The new theoretical descriptions and predictions suggest new experiments to test the accuracy of those theories. Go back to step 1.

    And recursively the theories become better and better approximations to reality.”

    I thought that summed things up fairly well.

  12. #12 Nick
    May 13, 2009

    Science is like Pornography: You know it when you see it. Come to think of it, so is Bullsh*t.

    And when I see any reference to Vox Day, I know I’m about to step it the last one.

  13. #13 Chad Orzel
    May 13, 2009

    “Vox Day” is a fucking lunatic. Of course, saying that on my blog led to one of the highest-traffic days I had last year, so you may well be setting yourself up with beer-and-pizza money for the summer…

  14. #14 Ubiquitous Che
    May 14, 2009

    To get back to the actual topic:

    I think Physics benefits from the fact that its easier than Biology.

    Let me qualify that.

    I’ve always understood that physics gets to describe the basic forces by which things work. When those forces get together to form atoms and molecules, and we become interested in what those atoms and molecules are doing, things get a bit more messy and interconnected. So we move into the chemistry building.

    When we’re in the chemistry building, after a while we come across some bits and pieces of interconnected atoms and molecules that behave in even more complex, interconnected ways, leading to reproduction of chemical structures on a macro-level. When things get too messy for chemsitry, we then move into the biology building.

    At each step, things get more difficult. Not to say that very advanced physics isn’t difficult – of course it is. However, every time we take a step, the level of interconnectivity goes up and precision goes down. Things start to get fuzzy. It’s harder to get a clear answer.

    Now before anyone raises the question of quantum mechanics and fuzziness: Quantum mechanics may be a bit fuzzy, but it’s fuzzy in a precisely calculable way. Not so with biology and the soft sciences.

    That’s my understanding. Any other views?

  15. #15 Tybo
    May 14, 2009

    I definitely second Che here, in that difficulty of prediction increases at higher “levels”.

    There’s more potential confounds at the biology level than the physics level, for example, due to the nature of the systems themselves. Certainly fluid flow and related fields see similar problems in prediction, but these are mostly issues of chaos and divergence of the variables at hand. Higher levels, in the scientific sense, have variables bordering on the inconceivable at even the biological level. Statistics becomes a necessity at this point, and laws become generalizations.

    Is this bad for the institution of science? Not really; the soft sciences can still do randomized-grouping experimentation, and while this gives only statistical certainty about “reality” or what-not, it’s still a fairly reasonable assumption. Beyond that, though, and you really do get into that fuzzy “how much validity are you going to give this?” sort of questioning. Using psychology as an example, is an interrupted time-series design scientific? There’s certainly not a definitive yes-or-no to that one.

    Going back down a level, this is the type of stuff biology runs into. A lot of research is after-the-fact investigational, or qualitative comparisons (e.g. anatomy). One really sees problems here and there as to how many potential confounds there are, and even trials that one thinks are randomized can be confounded. A recent issue in biomed research has come about due to cleaning chemicals used in labs, for example.

    One really doesn’t see those sorts of confounds in certain branches of physics; generally, the experimenter either just screwed up or something’s wrong with a detection instrument.

    So I guess I’m in the Churchland/Feyerabend camp, in that biology books trying to define science are really just pidgeon-holing. Then again, define what constitutes language, philosophy, or the like, and the same issues come about. It’s like the Wittgensteinian “family grouping” is rearing its ugly head: can you tell me what is and is not a game, even?

  16. #16 Tom
    May 14, 2009

    One problem is that Vox Day didn’t present a list of definitions of science. It was a list of descriptions, all from different perspectives and each of them incomplete. You can describe a house when standing out in front of it and when standing in the back yard. The two descriptions will differ, even though it’s the same house.

  17. #17 Tercel
    May 14, 2009

    I’d say that science is the process of formulating an explanation of how the world works, which will hopefully, but not necessarily, be predictive as well as descriptive, based strictly on observations, which are hopefully, but not necessarily, quantitative.

    Besides my horrible sentence structure, I believe that captures the fundamental goal and essence of science. Everything else that usually comes up in this context, such as the scientific method, the technique of peer review, the devotion to intellectual integrity, and the willingness to replace or update theories based on new evidence, is not what defines science, but simply the tools of science.

    By analogy, an auto mechanic is a person who fixes cars. He may do that with wrenches and ratchets, but you wouldn’t define a mechanic as “somebody who uses a 5/8″ wrench to tighten lug nuts and a screwdriver to align headlights.” Those things are simply a means to an end.

  18. #18 Anne
    May 16, 2009

    Before we physical-sciences types get too self-congratulatory, the introductory astronomy course I was the TA for this term did explain the scientific method.

  19. #19 JDHuey
    May 25, 2009

    My two favorite “definitions” of science are:

    1.) Science is what scientists do.


    2.) Organized commonsense.