DNA: A Metaphor?

To the Editors:

H. Allen Orr in his review of Matthew Cobb’s Life’s Greatest Secret [NYR, June 9] gives us a beautifully clear exposition of modern genetics, perhaps the clearest I have ever read. He notes that developments in mathematical information theory and cybernetics soon after World War II had a strong influence on the way biologists began to talk about life and heredity—specifically the idea that DNA contains a “genetic code” replete with “information.” Much the same point can be made about psychology and neuroscience: they too were influenced by these developments in the mathematical theory of information, introducing this notion into the heart of theories of the mind and the brain. But Orr goes on to remark:

In the end, the information sciences provided biologists with loose but useful metaphors and analogies, a language that allowed scientists to think and speak in new ways. But the high-powered mathematics of these fields proved mostly impotent in biology [I would add in psychology and neuroscience too]. No one, for instance, used Shannon’s equations to say anything especially interesting about organisms.

This raises a troubling question: If these recent ways of talking in biology, psychology, and neuroscience are really just loose but useful metaphors, now deeply ingrained in these sciences, what is the literally true way of speaking for which they substitute? How can we reformulate these sciences in such a way that the information metaphors are replaced by sober statement of fact? And do scientists now agree that the borrowed way of talking really is just loose metaphor, or have they come to take it for literal truth? This question seems to me not sufficiently addressed, though very important.

Colin McGinn
Miami, Florida

H. Allen Orr replies:

I thank Colin McGinn for his letter. As usual, he has recognized a key conceptual problem and one that often goes undiscussed—as, for instance, in my review of Life’s Greatest Secret (though not in the book itself). It’s important to note, however, that Matthew Cobb’s claim that mathematical information theory contributed little to the rise of molecular genetics does not itself settle the question of whether DNA harbors information. At least two possibilities present themselves.

The first is that information theory contributed little because the idea of information in DNA is fundamentally mistaken. This is (roughly) the view of some philosophers of biology. One kind of molecule (DNA) is associated with another (protein) but talk of information is mostly metaphorical.

The second possibility is that mathematical information theory contributed little because Claude Shannon’s theory considered the wrong kind of information, not because there’s no information in DNA. Cobb, for instance, notes in Life’s Greatest Secret that Shannon’s measure of information features no clear sense of meaning. For example, two alternative DNA sequences at a gene might show the same amount of Shannon information although one sequence “means” that the resulting protein, say, detoxifies a compound and therefore allows an organism to live while the second sequence “means” that the resulting protein does not detoxify the compound and so lets the organism die. A mathematical theory that is blind to such distinctions is unlikely to be of much use to biologists. But something that is loosely analogous to Shannon’s information and that takes this sort of biological “meaning” into account could still reside in DNA.

McGinn asks how biologists view the idea of information in biology now. He will not be surprised to hear that few consider the question explicitly. If pushed, some biologists may incline to the first view but I suspect that most would favor the second. Whether they are fundamentally right to do so is subject to some debate among philosophers.