About Rationally Speaking

Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.

Tuesday, May 27, 2008

Soft vs. Hard science, part I

[This post is an excerpt from the draft of a chapter of my forthcoming book, Nonsense on Stilts: How to Tell the Difference Between Science and Bunk," to be published next year by the University of Chicago Press]

“Scientists these days tend to keep up a polite fiction that all science is equal. Except for the work of the misguided opponent whose arguments we happen to be refuting at the time, we speak as though every scientist's field and methods of study are as good as every other scientist's, and perhaps a little better. This keeps us all cordial when it comes to recommending each other for government grants.” Fighting words about the nature of the scientific enterprise as seen from the inside by a participating scientist. And what makes these sentences even more remarkable is that they were not uttered behind close doors in a room full of smoke, but printed in one of the premiere scientific magazines in the world, Science. It was 1964, the year I was born, and the author was John R. Platt, a biophysicist at the University of Chicago. The debate between scientists on what constitutes “hard” (i.e., good, sound) and “soft” (i.e., bad, sloppy) science has not subsided since.

Platt was frustrated by the fact that some fields of science seem to make clear and rapid progress, while others keep mucking around without seemingly being able to accomplish much of relevance. As Platt put it, in the same article: “We speak piously of ... making small studies that will add another brick to the temple of science. Most such bricks just lie around the brickyard.” Physics, chemistry and molecular biology are considered by Platt (and many others) as hard sciences, the quintessential model of what science ought to be. Ecology, evolutionary biology, and even more fields like psychology and sociology, are soft sciences, and the maximal aspiration of people working in these fields ought to be to find a way to make them as hard as physics. Platt’s article is a classic that should be read by anyone interested in the nature of science, and he was right in pointing out the problem; he was not quite as right in diagnosing its roots however, and even less so at suggesting a possible cure. Nonetheless, Platt’s critique of soft science provides us with an excellent starting point to explore the idea that, in fact, there may be more than one kind of science, that “science” is a heterogeneous category, a notion that would surprise many in the general public, and that will likely be resisted by most scientists.

Platt’s attack on soft science begins by stressing the fact that some disciplines seem to make fast and impressive progress, while others have a tendency to go around in circles, or at best move slowly and uncertainly. Before we examine why this is and what could possibly be done about it, a more fundamental question is whether Platt is correct at all in identifying the existence of a problem. It seems clear from even a cursory examination of the history of science that Platt is at least partially correct: some sciences do progress significantly more than others. However, the pattern seems more complex than a simple line dividing “hard” from “soft” disciplines: it is true, say, that particle physics and molecular biology have made spectacular advances during the 20th century; but it is also true that physics itself went through long periods of stasis, for instance the long interval between Newton and Einstein, during which little or nothing was known about the intimate structure of matter. And such periods of slow progress may occur again in the future, even for “the queen” of sciences: for all the talk about a “unified theory of everything,” physicists have been trying to reconcile the known discrepancies between their two most successful theories, general relativity and quantum mechanics, for now close to a century; as yet, they have not gotten very far.

Organismal biology (ecology and evolutionary biology) is often considered a quasi-soft science, and yet it has seen periods of great activity and progress -- most obviously with Darwin during the second half of the 19th century, and more recently during the 1930s and 40s. Moreover, there is currently quite a bit of excited activity in both empirical and theoretical evolutionary biology, which may be leading to another major leap forward in our understanding of how organisms evolve and adapt to their environments. Molecular biology, on the other hand, hailed by Platt as a very successful hard science on the model of chemistry and physics, may be running into the limits of what it can achieve without falling back on “softer” and more messy approaches to its subject matter: it is true that the discovery of the structure of DNA in 1952 is one of the all-time landmarks of science; but it is equally clear that the much-touted sequencing of the whole human genome has provided very few hard answers for biologists, instead leading to a large number of “bricks laying around the brickyard,” as Platt would have put it. We know a lot more about the human (and other) genomes, but much of what we know is a complex mess of details that is difficult to extricate and to make into a clear picture of how genomes work and evolve.

All in all, it seems that one can indeed make an argument that different scientific disciplines proceed at dramatically different paces, but it is also true that the very same science may undergo fits and starts, sometimes enjoying periods of steady and fast progress, at other times being bogged down into a spell of going nowhere, either empirically (lack of new discoveries) or theoretically (lack of new insights).

If we agree that the nature of science is along the lines I have just described, and that Platt therefore has a point, next we need to ask why it is so.


  1. [P]hysics itself went through long periods of stasis, for instance the long interval between Newton and Einstein, during which little or nothing was known about the intimate structure of matter.

    You are not being at all fair here. You're picking one specific topic as the measure of progress; whereas physics made enormous progress in other areas: optics, electricity, mechanics, thermodynamics and other topics between Newton and Einstein.

    I think that when we're talking about "soft" sciences, we're talking about sciences such as psychology, anthropology, evolutionary psychology and (at the extreme) psychoanalysis [see Frederick Crews] and parapsychology.

    It seems fairly obvious that the soft sciences show a disconnect between theoretics and actual experiment: its practitioners do not say, "We have no clue what's really going on; we're fumbling about in the dark trying to find something to latch on to."

    There's nothing at all disreputable about stumbling about in the dark, so long as one is honest about it. Indeed, the "stumbling about" phase is probably the most interesting and exciting part of science.

    It's the disconnect between experiment and theoretics that is disreputable. The disreputable theoretics are "Just-So stories". As you note in your earlier post, they are arguments only that the hypothesis entails the data, without the crucial addition of showing why the data justifies the hypothesis over its alternatives.

  2. Well, first of all, this is only part I of a two-part series. Second, this series certainly doesn't exhaust my thinking about soft vs hard science (you'll have to buy the book :) Also, I wouldn't put psychology into the same boat as parapsychology

    Moreover, I was certainly not claiming that there is one and only one measure of progress. As it will be clear from the second part of the post, my point is simply that one needs to be more careful about uncritically accepting the widespread idea that physics is *the* way to do science.

  3. [T]his is only part I of a two-part series. ... [O]ne needs to be more careful about uncritically accepting the widespread idea that physics is *the* way to do science.

    Agreed and understood. When the next book-buying binge rolls around, your book will be on the list.

  4. Are 'hard' sciences those that are 'explained' via mathematical relations between initial conditions and measurements (e.g. physics) and validated via repeated experimentation?

    What about natural history and cases where we believe that we understand the mechanisms similarly to bench science but can't validate that understanding by experiment. Still 'hard' science?

    What about situations like often in medicine where we have no idea what the mechanism is but we can establish strong correlations between treatments and outcomes? Hard science?

    Or situations like in the social sciences where theory often rests on intuition and there is only a loose connection between conditions and outcomes, but enough to inform policy? Soft science.

    For me, science is explanation resting on evidence. The form of the explanation varies from mathematical equations to storytelling to 'who needs an explanation anyway - look at the data'. The strength of the evidence varies from repeated, predicted, deterministic outcomes to consistency between the story and multiple evidence types (e.g. evolution) to statistical correlation suggesting only a 'degree' of truth.

    But the bottom line for it to be science is: Is it the best (or contending) explanation so far for the evidence right now.

  5. I have always thought that the difference between hard and soft acience was whether something was quantifiable.
    Physics is thus clearly hard, because it boils down to equations.
    Chemistry is likewise hard, but for where it touches biology.
    Yet biology is not necessarily soft, because it is not necessarily unamenable to quantification, only that quantifying its basic notions would require hideously complex algorithms, which are currently beyond the ability of even the greatest supercomputer. Contrast this with sociology, where the "rules" are so flexible, quantification is impossible, because not two sociologists can even agree on the terminology. (Except for the idea that all white people are racist....)
    Which is why I refuse to call sociology a science, hard or soft...

  6. Kimpatsu,
    Wherever you draw the line, given that the empty set's on neither side, psychology is a soft science. Psychologists quantify things. Just saying.
    As a former oceanographer, though, discussions of what makes a hard science makes me a little nervous. The ocean's really complex! So what if we were wrong about Ekman flow?
    I'm excited to learn how to tell science from bunk. It seems like your options are doing science and knowing whom to trust. Since science is difficult and time consuming, and people are better at getting other people's trust than knowing whom to trust, neither of those seem like very practical answers. However, it is a quesion to which we'll need an answer, if we start letting science drive policy.

  7. Kimpatsu wrote

    Yet biology is not necessarily soft, because it is not necessarily unamenable to quantification, only that quantifying its basic notions would require hideously complex algorithms, which are currently beyond the ability of even the greatest supercomputer.

    But one can model (parts of) biology in terms specific enough to write them in formalisms (code) and test assertions about the models as they relate to the 'real' world. Evolutionary simulations, for example, are finally getting to the point where one can test real hypotheses in them and evaluate those models against biological phenomena -- see Avida, for example.

  8. RBH, that would indeed make biology "hard".

  9. "...but it is also true that physics itself went through long periods of stasis, for instance the long interval between Newton and Einstein.."

    This is so wrong ... in between Newton and Einstein you had the discovery of electricity and magnetism, and the experimental and theoretical genius of Faraday & Maxwell. Maxwell is up there with Newton and Einstein ... Richard Feynmann said people would on day remember the 1860s not for the American Civil War, but for Maxwell's Equations...

    Instead we remember the Origin of Species in 1859, but just look at any electrical or electronic device near you and consider that Maxwell's influence has not been less than Darwin's.

    Physics went through one of its most exciting periods in the mid-19th century. Einstein cam along just as physicists thought they had solved all the problems posed by the earlier discoveries (e.g. electromagnetic radiation, radioactivity, radio communication, spectroscopy, thermodynamics ...)

    I tend to agree with the Barefoot Bum on this one.

  10. Toby,

    obviously I did not express myself clearly. My statement was not a denial of *any* progress in physics between Newton and Einstein. As you say, that would be so wrong to be ridiculous.

    Rather, the point was that there had been no conceptual rethinking equivalent to the Newtonian or relativistic theories between those two.

    This is relevant to the discussion on soft science because psychology, for instance, cannot be faulted for not having made empirical advances during the last century. But what it can be faulted for is the lack of any new conceptual advancement that may ground it in the way Freud tried and failed to do.

  11. Hellos...

    A little typo there, Massimo: "discovery of the structure of DNA in 1952".

    It was 1953, as far as I remember from reading both Crick's and Watson's accounts of the whole thing. They came up with the Eureka moment in late February, 1953, and rushed to publication, getting the paper out in the April 25, 1953 (a date every genomicist like me should celebrate :-) issue of Nature.


Note: Only a member of this blog may post a comment.