Physics Envy on a Cold Day

I read a post by a biologist philosopher about how supposedly physicists are undertaking “stamp collecting.” And I’m fairly annoyed that they missed the entire point. They read this article on attempted classification of string vaccua, and likens this attempt to taxonomy in biology, in their eyes lowering physics to a descriptive science. This whole flame war started with Rutherford claiming “In science there is only physics. All the rest is stamp collecting.” Let me liberally translate to clarify: “In science the only quantitatively predictive field is Physics, and everything else is descriptive.” This certainly was true in Rutherford’s time, but discovery of atomic structure elevated Chemistry to predictive, and the discovery of DNA elevated biology. Now we’re all buddy-buddy. In real science, 99% of the time you first gather data, then classify, then build models, and then you have a quantitative principle to make predictions. This went faster with physics because we deal with much more homogeneous systems. An old friend and ex-ZEUS student I know is now writing Monte Carlo simulations for Neuro Bio. That aint taxonomy. Heck, even economists have real models these days, and their particles have agency.

The big missed point is the following: String Theory isn’t predictive, or descriptive, yet. It has never made better than tenuous qualitative connection with reality, and I wouldn’t bet any money that it will improve soon. Why? Because its history is all bass-ackwards. It may have started from meson spectroscopy and QCD flux tubes in the late 60’s, but then everyone thought they smelled gravity and got the great idea to start from a unprobeable scale and grope their way back down to experimentally plausible energies. In string theory you first assume an underlying principle, then build a phenomenological model, then gather data. This isn’t to say its totally unmotivated, but it does incur a large investment risk.

The authors whole point is that stamp-collecting is a type of science, and shouldn’t be looked down on. This is half true. It is necessary for developing falsifiable theories, but should not be considered a form of science in itself. People can gather data, and arrange it in clever ways, but without predictions and basic principles, There’s very little objectivity, or application.

So biology in Rutherford’s time was like stamp collecting, but all those stamps they collected were issued from a genuine authority. Woit claims the String Vaccua Project “…is stamp-collecting done by people who don’t have any stamps, just some very speculative ideas about what stamps might look like.” He breakdown of the string vaccua project on his blog, if you want to read more. He also comments on the original biologist’s philosopher’s post.

Edit: Dr. John S. Wilkins, the author of said blog, is a postdoctoral research fellow of philosophy, not biology, as originally stated.


7 responses to “Physics Envy on a Cold Day

  1. I’m imagining the physics blogosphere in 1915:

    “In general relativity, Einstein first assumes an underlying principle, then builds a phenomenological model, then waits for others to gather data.”

  2. First off, I’m a philosopher, not a biologist. Second I think that my point was in fact there is no “ass-backwards” or “ass-forwards” in science – all that matters is that the program is fruitful (and I am incompetent to judge if that is true or not about string theory, and frankly don’t care).

    Criticisms of new models have often argued that this or that scientist was unscientific for not following “established rules of method”. It was said of Darwin, of Einstein and I am sure of pretty well every major innovator in science. But what makes it science is, in the end, whether enough scientists use that model or theory to do their work.

  3. Hi Blake, The big difference is that GR was an extension of SR, in which frame invariance was originally developed and experimentally established. You are right that every theorist must step beyond knowledge at a present time to make an advancement, but how far to jump and how long to wait for fruits of labor is a matter of personal taste.

  4. Dr. Wilkins, I have edited the article to properly describe your title. My apologies. While I do agree that the proof of a theory is in its final results, there is historically a companionship between gathering of data, processing it, and describing it with maximum rigor. Science is this trinity, and no single part of it alone should be described as science. Sometimes scientist must press forward without some component, but the goal should always be all three in accord.

  5. I agree on the “personal taste” aspect, an issue which is probably under-addressed. I just think it’s interesting to contemplate:

    0. Einstein originally developed SR without knowing about the Michaelson-Morley experiment. Instead, he was just contemplating the symmetry properties of the Maxwell equations and their physical implications. Furthermore, that experiment was on the bleeding edge of the possible when it was first performed, so if it hadn’t been done by 1905, I’m sure a few strong voices would have deemed it impossible. Sure, we’d look back and file them with Lord Kelvin and his “heavier-than-air flying machines are impossible” judgment, but if science were easy, it’d all be done by now.

    1. The most quantitatively exact test of general relativity available in the 19-teens was the orbital precession of Mercury. This had already been measured, so “predicting” it from GR meant giving a figure for something already known.

    2. The first prediction of a quantity which hadn’t been measured before was the gravitational deflection of light. This got measured pretty quickly, of course — we only had to wait until 1919. However, Eddington and company still had to traipse to the ends of the Earth, camp on a desert island and account for the thousand-and-one factors of error which confounded their photographs of the solar eclipse. (“Twinkle, twinkle, little star: are my readings off that far?”) This was, again, an experiment at the limits of the possible, and its result was only borderline: promising, but not conclusive.

    3. A laboratory test of GR wasn’t possible until the Mössbauer Effect had been discovered and understood. Rudolf Mössbauer didn’t make that breakthrough until 1957. Is forty years and change a long time to hold one’s breath, reassuring oneself that we understand frame invariance so very well in the case of flat spacetime, and that such elegant mathematics involving such wonderful curvature tensors could never lie?

  6. Stellar points.
    Einstein may never have heard of Michaelson-Morley before 1905, but he probably had heard of a few other experiments like those listed above. Whether it came to him as an amusing symmetry of EM or of utility is anecdotal. What is important here is how it came to popularity in the community, which is in the context of the need to describe observations.
    1: You are right, agreement previous observations isn’t proof, but its a sight better than not agreeing, or worse yet not predicting anything.
    2: I’m not surprised this was borderline. I remember the seeing pictures in school. Again, marginal agreement with a non-null experiment isn’t proof, but its better than not agreeing.
    3: I had to look this up. That’s an awesome experiment. I saw a talk by an atomic clock guy a few years ago who claimed his group could see time dilation from a height difference on the order of a few meters. What I wanted to say here was that Forty years is a long time for significant confirmation with a single experiment, but not nearly as long as 40 years without a prediction at all.

  7. I’m not an expert in this by any stretch of the imagination, just a guy who spends too much free time reading about gauge/gravity duality. So, what I don’t know outweighs what I do kn0w by an astonishing factor. Consequently, I ask a whole lot of hypothetical questions which doubtlessly seem completely brain-damaged (brane-damaged?) to people who do know their stuff.

    To that end:

    What if the weather had been cloudy in 1919, and Mössbauer had got too drunk at Oktoberfest 1956 to do anything about gamma rays, and we were left without any testable predictions of GR at all? Oh, sure, we could talk about the Schwarzschild and Kerr solutions to the Einstein field equations, and discuss their abstract mathematical properties — but who can make a black hole in the laboratory? Without some basic experimental confidence in the equations, would we even look for compact massive bodies in the sky, or know what to do with anomalous X-ray sources when we found them? Speculating about event horizons when we haven’t even seen a gravitational redshift might be like, I dunno, saying that at extraordinarily high energies, scattering cross-sections should look “stringy” rather than particulate, regardless of the D-brane/orientifold background on which those strings are propagating.

    As far as I can grok the history of this subject, it seems that twenty-five years ago, there were good reasons to expect that no theory of quantum gravity could be worked out unless you knew all about the interactions all the way up to the Planck scale. In such an atmosphere, following even the ghost of the smell of a falsifiable prediction with all the tenacity of a bloodhound seems rather a sensible thing to do.

    Your mileage will, of course, vary.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s