Theory or law?

Andrew Sullivan has been posting comments from readers about why evolutionary biology is comprised of "theories" rather than "laws." I found these via Razib, who naturally has more interesting things to say than Sullivan or his commenters. But no one in the conversation has really given an answer to the question, other than some vague idea of what it takes to "qualify" as a "law." I think the answer is historical, and demands that we consider a number of so-called "laws of evolution" and their fates.

As evolutionary biology got underway in the late nineteenth century, its main proponents called many of its major ideas "laws." I give a number of examples toward the end of the essay, but to begin with, Darwin referred to the "Law of Natural Selection." When Mendel was rediscovered, his two main ideas were called "Mendel's Laws of Heredity": the "Law of Independent Assortment" and the "Law of Segregation." Those two have stuck around.

The "Hardy-Weinberg Law" is hardly ever called that anymore; "law" being replaced by "principle."

It seems to me there's an obvious historical reason why evolutionary biologists reject the term "law" -- too many of the so-called "laws" of evolution ended up being false!

The premier example is Haeckel's "Biogenetic Law": "Ontogeny recapitulates phylogeny." The theory of recapitulation came with an entire suite of concomitant mechanisms to explain its exceptions, In the end the idea was conclusively falsified. The mode of evolution of developmental processes can generate similarities between ontogenetic and phylogenetic sequences, but there is no causal mechanism constraining these other than the normal mechanisms of genetics.

There are other outmoded "laws" of a similar vein. The "Law of Orthogenesis" is now rarely referred to as a "law", but its main proponents certainly called it one. Then there was the "Law of Irreversibility" -- the idea that evolution couldn't "go backward" to bring a population to an ancestral morphological pattern.

The empirical and theoretical failure of these so-called "laws" did not suppress evolutionary biology's taste for overarching statements about patterns and process. But their failures did tend to make people skeptical of the idea that "laws" of evolution would really be found. Also, the discovery of actual "laws" of heredity yielded a theoretical interest (enshrined in population genetics) for reducing the overarching pattern of evolutionary history to the mechanisms of heredity.

The influence of recapitulation over embryology has been well-documented, and of course the main detractors from the early development of population genetics as the mechanism of evolution were morphologists -- embryologists, paleontologists, biometricians. This possibly influenced early population geneticists like Fisher to refer to their mathematical formulations as "theories" or "principles" rather than laws -- although this is just a speculation and I would like to see documentary evidence.

What seems clearer is that after the biogenetic law was rejected, empirical generalizations in biology tended to be called "rules" rather than "laws". Consider "Romer's rule", the "Island rule" (also called "Foster's rule"), "Rensch's rule", and (this is a genetic example) "Hamilton's rule."

Maybe the best example is "Cope's rule", which started out as a "law" but was turned into a rule by people who still found it useful at midcentury. I find this sentence from the Wikipedia article on "Cope's rule" quite relevant to the shift:

Note that semantically the "rule" in this context (unproven assumption with exeptions) refers more to a rule of thumb, trend or a belief than to a truth, law, fact or a norm.

That's certainly the way that biologists today think of these "rules." Interestingly, Cope was also responsible for the "Law of the Unspecialized", which is uncommonly enough invoked that the "law" name has stuck.

Still, in physiology and anatomy (biological fields, to be sure), "law" is widely used. "Kleiber's law" relates body mass and metabolic rate. "Sherrington's law" refers to the simultaneous stimulation and inhibition of opposing muscles. "Wolff's law" relates bone growth to mechanical loading. Most of these have just been held over from the nineteenth century, but not all -- Kleiber's law was formulated in the 1930's just as "law" was going out of vogue.

In short, I think that the reasons why evolutionary biologists don't call ideas "laws" are basically historical. It has nothing to do with whether a "mathematical formulation" can be found -- there's certainly none underlying Wolff's Law, which is still called that. It has entirely do do with the rejection of over-ambitious "Laws of Nature" as applied to the outcomes of the evolutionary process. The worst offender was the biogenetic law, but there were others as well.