The Problem of “Theory”

When I was writing about the seemingly contradictory meanings of “adiabatic” the other day, I almost gave “theory” as an example of a word with nearly opposite meanings. After all, as anyone who has even glanced at the evolution-creation “debate” has heard, a “Theory” in science is something more exalted than a mere guess– it’s a guess that has been confirmed by observations and experiments, and can thus be regarded as true with a high degree of confidence (and assigned a capital letter in this post, to set it apart). (Also, in physical science at least, it makes quantitative predictions.).

I didn’t use it for two reasons. For one thing, it seemed to be inviting an argument that wasn’t remotely relevant to the subject of the post. More importantly, though, it’s not actually consistent with normal scientific usage. To put it more bluntly, I do not recall ever hearing a professional scientist make the theory/Theory distinction in a scientific context (as opposed to a public debate– lots of people cite it when arguing against crazy people).

In physics, at least, the operational definition of “theory” is “that which is not experiment.” The word gets used to apply to anything that isn’t a direct observation of reality. We even tend to use it for things that are demonstrably untrue– “local hidden variable theory” is a common phrase in quantum circles, and the predictions of local hidden variable models have been shown to be wrong by something in excess of 100 standard deviations at this point. There’s no way that phrase could make any sense if scientists really hold to the more exalted meaning of “Theory.”

In fact, if you think about it, if the distinction between the common use of “theory” as a sort of guess and the supposed scientific use of capital-T “Theory” as a well-confirmed and tested model of the universe was really generally held, you would never see a data plot with a line marked “theory” that failed to hit all the data points. And yet, for an experimentalist at least, there are few pleasures greater than showing a data plot where the theory line doesn’t come anywhere near the data– that’s where science gets interesting.

Am I saying there’s a valid point to the “evolution is just a theory” rants? Absolutely not. The scientific use of “theory” is not restricted to the capital-T version, but it unquestionably encompasses it– the response “Yes, evolution is just a theory. So is gravity.” remains the correct one. But given that the day-to-day use of “theory” in science is so muddled, it’s no surprise that this is an area that can cause confusion for non-scientists.

(And let’s not even talk about the humanities, where “theory” (as in “critical theory”) is often used in something close to the pejorative sense meant by denialists– a pre-existing ideological framework which is used to interpret any text you happen to run across, whether it makes obvious sense or not…)

9 thoughts on “The Problem of “Theory”

  1. In a broader scientific sense, I think “Theory” can refer to a framework that is wrong or incomplete; it makes predictions, but the predictions fail to hold. That’s still much stronger than the WAG misuse of the term.

  2. I am wondering: when physicists speak of “theory”, don’t they rather mean “model” (in the simplest case a set of equations)? As opposed to, say, “set theory”, “graph theory” etc?

  3. Seems to me that “theory”, both in scientific usage and in colloquial one, is a attempt at a comprehensive explanation of a set of facts. There can be a well-formed, predictive and successful theories, and there can be theories failing those standard on various counts. I don’t think the word “theory” by itself is an endorsement of the model it refers to, but equally it is not a dismissal of the model. The reply to “it is only a theory” should make the distinction between good and bad theories, but I don’t think it is efficient to endow the word “theory” with meaning it doesn’t have in common usage, and then pedantically tell everyone their use of the word is incorrect.

  4. I have always disliked the crass theory/Theory distinction used in the arguments over evolution, so thanks for being brave enough to say something about it. I just have one thing to add to the debate:

    String theory is called a theory.

    That is all. Now, let the games commence.

  5. I just have one thing to add to the debate:

    String theory is called a theory.

    I thought of that, but I try to only provoke one religious war at a time.

  6. If you have static particles surrounded by an arbitrary border
    (Gaussian law of statics) you can make it dynamic by adding a time varying field while maintaining equilibrium FACT
    Maxwell’s equations of radiation is now applicable FACT
    Therefore static “particles” and not “waves” are the carriers of communication FACT
    So why does the uncertaincy principle even exist?
    Because physicists cannot handle change. The math is ignored
    and the count of “crackpots” increase. More than a hundred years have passed while holding back the advances of science by physicists.

  7. “the predictions of local hidden variable models have been shown to be wrong by something in excess of 100 standard deviations at this point.” You missed out the “that are straw men” that you meant to put after “local hidden variable models”. But it’s true that local hidden variable models that are not straw men –that is, random fields, in large part a half-way house between classical continuous fields and quantum fields– do not meet the empirical standards required of a Physical theory, and are in any case similar enough to quantum field models to make it hard or impossible to tell them apart.

    Quantum field models for experiments are themselves hidden variable models for low-dimensional quantum mechanical models, although one could have a nice time characterizing the difference between microcausality and locality.

    ‘In physics, at least, the operational definition of “theory” is “that which is not experiment.”‘ Alternatively, the operational definition of theory is that which suggests what experiment to do. What might give an interesting result? “there are few pleasures greater” than an interesting result, but I’m sure that rubbing salt in the theorist’s wounds is no part of the pleasure. Deciding what experiment to do is part of the experiment, right? You need us for something, maybe even as much as we need you.

    What is a “successful theory” is so conditioned by how Whiggish you’re prepared to allow your history to be that it’s a matter of definition whether Newtonian Mechanics is a successful theory or not. But I digress, four times.

  8. Ah! My favorite subject. I think the best approach (not really different from that of earlier posters here) is to think of theory as a description of observations in a causal context–a replacement of the misleading and obsolete concept of “scientific law.”

    We make observations (“data” — previously “facts”) and construct a mental understanding of the observations–especially how they relate to each other–with theory (almost always in the sense of things causing other things). So, the conservation of energy is a theory with which we think we understand how it is that energy persists through its alterations.

  9. But “theory” is *exactly* a word with opposite meanings, and I think it would be good to recognise this. The scientific meaning came first. The popular meaning is a language change that came from misunderstood sarcasm, Bugs Bunny style. “Oh yeah, so what’s your theory, wise guy?” implied that the idiot in fact had only a wild guess and not actually a theory at all.

    Very much like “What a Nimrod!” was once sarcasm – as everybody knew, Elmer Fudd was certainly no fabled “Nimrod the Mighty Hunter”! It became a direct insult when people stopped recognising who Nimrod was.

Comments are closed.