Neuroscience Can’t Explain Wagner (or B.B. King) writes Christopher Shea on the Ideas Market blog at the Wall Street Journal, arguing against the claims that are made in my post from last week about day-old chickens preferring consonant music. I find two problems with his argument: the first concerning methodology, and the second concerning what "innate" really means.

Concerning methodology, he writes:

The difference between consonance and dissonance is demonstrated through two short piano passages: One follows all the simplest rules of harmony, the other violates them—and, indeed, the second example sounds abrasive.

We learn that a recent clever experiment revealed that day-old chickens do, indeed, prefer “consonant” music, and this is said to lend weight to the hypothesis that such a preference is universal and innate, including in humans.

The trouble is, the dichotomous division of music into “consonant” and “dissonant” really makes no sense, except as a cartoon, when it comes to describing the music that humans produce and enjoy.

And, he's right, of course. Music is certainly more complicated than simple strings of notes played on a piano, but in order to study complex mental phenomena you have to start from simplicity. Scientists have done this with great success, for example, when it comes to our understanding of vision. Much of that research derives from very simple experiments in which humans and animals must watch flashing checkerboard patterns or find an X from within a field of Ts, and so on. As a testament to this, in the field of artificial intelligence, scientists are perhaps closest to approximating human mental processes when it comes to visual perception. In science, we need a high level of control - this is why it is common to use simplified experimental stimuli - and still, they give us great insight into mental processes.

To that end, the claims made in my post, and in the paper on which it was based, concern those simplified musical intervals, not anything more complex. However, he writes:

It is certainly interesting and important that day-old chicks display a preference for one kind of musical interval over another. But the leap from that finding to any statement about human musical taste is unimaginably vast.

Again, Shea is correct: a leap from chick preferences for music to a statement about human musical taste is indeed vast (perhaps more imaginably than he might be comfortable with, though). But the story is actually the other way around: the human experiments came first. And the data for humans is actually fairly clear: we show a preference for consonant music by two months of age. Understood within the context of an experiment which showed that day-old human infants can distinguish consonance from dissonance, it is actually quite reasonable to conduct a controlled rearing study with infant chicks to begin to investigate whether that distinction may actually reflect a preference, at ages younger than two months. This particular study is, of course, just one piece of a larger puzzle yet-to-be-completed. And neither the study in question, nor my post about it, make any claims about complex musical tastes beyond the simplified musical intervals in question. Nor do they make claims about human adults, with years and years of musical experience; they make claims about human infants. This leads me to the second problem.

Shea also writes:

The dissonance of 20th-century concert music is well known. Indeed, this is the point when people begin to argue that composers departed from the “natural” laws of consonance, and ought to revert to them. (Their defenders argue that they were pushing ahead with the experiments of Wagner and the Romantics.) But sidestep that debate. Here’s a more persuasive point, for most Americans: The minor third played against a major third chord, the essence of the blues (and much rock), creates a minor second interval, the very interval that Scientific American describes as the epitome of a sound that humans hate. (Do you hate B.B. King?) Jazz, with its flatted ninth chords, contains dissonance in abundance. Metal, with its love of the flatted fifth, the “devil’s interval,” fetishizes dissonance.

And here's where I see the second problem with his argument, and it's the more complicated one: human preferences for music (and by "music" in this case, I'm referring to Wagner, BB King, and so on, rather than simple strings of notes played on a piano) are shaped by years of experience. Humans also find pleasure in plenty of other things that are initially aversive - like the taste of alcohol or caffeine or tobacco (okay, those do have some addictive qualities to them) or capsaicin, the compound that makes spicy foods spicy, which is non-addictive. I wrote about this last year:

Most young children, even from cultures known for their spicy recipes, are averse to capsaicin. So maybe, then, instead of actually liking the pain, we're merely desensitizing ourselves: what used to be really painful is now just sort of painful.


While most scientists still do not quite have a handle on the human preference for spicy foods, the best explanation comes from a mechanism called "hedonic reversal", or "benign masochism". Something happens, in millions of humans each year, which changes a negative evaluation into a positive evaluation, like flipping a light switch.

Not that dissonant music is painful, at least not in the same way that capsaicin produces feelings of pain, but there is perhaps something to be learned from this analogy. As I wrote on Wednesday, the innateness of the preference for consonant music is still an open question. But even if it weren't - even if we knew, for certain, without a doubt, one hundred percent, that there was an innate preference for consonant music in humans - an adult preference for dissonant music would have no bearing on that conclusion. For adult preferences are shaped over time by experiences.

Indeed, anything that is considered innate is subject to modification by experience. An Edge working group (on the science of morality) put it best:

The word "innate"...does not mean immutable, operational at birth, or visible in every known culture. It means "organized in advance of experience," although experience can revise that organization to produce variation within and across cultures.

The question is not whether human babies (or infant chickens) display a preference for consonant music at birth - taken together, the evidence gathered thus far suggests that this is likely the case - the question is whether this preference is "organized in advance of experience," or not. The new paper that I wrote about last week provides some evidence that this may be so.

Shea closes his piece with the same consonant and dissonant music clips that I included in my piece, and writes:

Rest assured (although you knew this): Day-old chickens do not prefer the same music that you do. And in the face of Shostakovich and Charlie Parker, not to mention B.B. King, neuroscience stands mute. There's more to music than this.

This leads me to wonder if Shea's real problem with my piece is not the simplification of experimental stimuli, but with science's purported attempt to "unweave the rainbow." Let's be clear: my original piece was decidedly not about human musical tastes, per se, but whether the apparent early preference for consonance is innate and is shared with other animals, and if so, whether that preference might help young organisms (human or otherwise) learn to distinguish animate from inanimate objects.

But, putting aside the actual content of my original piece for a moment, what is scary about an attempt by science to understand art, or beauty? Does an understanding of the refraction of light through water droplets make rainbows less beautiful to gaze upon? Does the understanding of human biology make the birth of children somehow less meaningful to their parents? Does knowing a recipe for the perfect cheese souffle make it less delicious? Not to me.

Human perceptual and cognitive experiences may ultimately be reducible, but that does not make our conscious experience of them less awesome. Rest assured (although you knew this): there is more to music than science.

Chiandetti C, & Vallortigara G (2011). Chicks like consonant music. Psychological science, 22 (10), 1270-3 PMID: 21934134

Update: Shea responds.