Shared publicly  - 
 
Interesting debate by +Gary Marcus and +Geoffrey Miller on the origins of music. Is it instinct, or invented?

I have been asked by many to give my reaction to the issue. As to whether music is instinct or invented, in my view it is...neither.

Most "cultural invention" viewpoints would focus on the extreme diversity of musics, and point to how we can learn loads of things we never evolved to do. I believe, on the contrary, that there are lots of universals. And not just for music, but also for language and writing. And my book, Harnessed, points out lots of them (many new). Generally, I'm of the belief that when we do brilliant things (like reading, language or music), it's because it relies on our instincts running as nature roughly intended them. We do genuinely new things very poorly -- like mathematical logic. So, I'm not a cultural-invention guy. Instead, I'm on the instinct side.

But, in the case of writing we know we didn't evolve by natural selection to read. It's way too recent. And although it is more plausible that there may have been time to evolve spoken language and music by natural selection, it is not particularly plausible, time-wise. And reading has the signatures of instinct: we learn it early with relatively little training, we become ridiculously good at it, it pervades our lives, and we seem to have specialized regions for it. But we know it's an illusion of instinct. ...and it seems like an instinct only because cultural evolution shaped writing to harness another instinct, namely one we have for visually recognizing objects. Written words look like nature (specifically like opaque objects strewn about in 3D space). And, as I argue in Harnessed, speech sounds like solid-object events of nature, and music sounds like humans moving about -- two of our auditory instincts are harnessed for language and music, respectively.

If music (and writing and language) came by virtue of nature-harnessing, then our capabilities will seem innate, yet not be. Not innate, and yet not mere cultural invention either.

Rather, it is a very specific sort of cultural evolution, one that nature-harnesses us.
11
2
Gary Marcus's profile photoMark Changizi's profile photoPatricia F Anderson's profile photoYael Mendez's profile photo
9 comments
 
What do you think the universals for music are? My reading of the literature is that there a very few; even the logarithmic spacing between notes seems less than universal.

Reading and music seem different from (spoken or signed) language in that language is acquired by everybody, with any sort of formal instruction, but not everybody learns to read, especially without instruction, and many people are unable to carry a tune.
 
I would add that all inventions co-opt earlier brain circuits to some degree, with more successful ones fitting more tightly to what's there earlier. Western tonal music fits more naturally to our brain circuits than Schoenberg's tone rows; both are inventions; neither plausibly innate, but one is more readily accepted by the listener.

(I wrote about some of this in more depth in Guitar Zero).
 
Rather than "universal," I should say "strong cross-cultural tendencies." Cultural evolution stumbles toward nature-harnessing solutions, so one doesn't expect strict universals. The tendency for folks to like Western tonal music over Schoenberg is one such example.

I think we're closely aligned on music (because, after all, the nature-harnessing view is committed to music not being an instinct), but one wants one's "cultural origins" hypothesis to have the ability to explain the tendencies within music.

On reading and music requiring formal instruction but not spoken language, I'm skeptical that music comprehension requires formal instruction. And the processing of written words also doesn't seem to require it (e.g., the recent baboon study, and my own son).

In my view, the simplest hypothesis (i.e., greatest prior probability) is that none of these three cases evolved via natural selection, and that cultural evolution shaped each to fit us. ...by structuring the stimuli to sound like three fundamental facets of nature we're already brilliantly competent at.

Guitar Zero in queue, but only poked at thus far... Soon.
 
The verdict is still out about language, but I disagree that the prior hypothesis should be on language being no different from music or reading: the latter are two simply aren't inevitable for all members of the species in the way that language is.

Leaving music out of the picture for the moment, language and reading just aren't on a par, either developmentally, or historically. Virtually very normal four-year-old can talk (or sign), but only an exceptional one can read. Language is also at least an order of magnitude older. Why presume then that language and reading are on a par? (Similarly, four years old have a very weak grasp of harmony, and even the fact that minor chords are sad; there are limits even in the comprehension of music.)

Reading is manifestly a cultural invention that fits decently well with the prior structure of the mind; language could be the same, but its ubiquity and the speed with which it is learned suggests otherwise.

Maybe Chomsky and Pinker are wrong about language being instinct but the mere fact that language fits well with prior adaptations (something I discussed in a 2006 article in Cognition) doesn't mean its not innate, and I think that's your main argument. (Note for example that wings fit well with the tetrapod limb scheme, but that doesn't mean their structure is learned.)

Bottom line: it seems odd to collapse something that is obviously learned and effortful (reading) with something (language) that is apparently learned effortlessly and more robustly across individuals. Both learning and evolution might be considered instances of "harnessing", but they reflect very different mechanisms.

In my view, the most reasonable notion then is to assume that language is different from reading, and to ask whether music patterns more with the former or the latter.
 
Children are getting "talked at" orders of magnitude more than they're getting "written at." And getting talked at happens much much earlier. But even so, most children end up reading very early, even before they're competent on monkey bars (to give some context).

To the extent that language can culturally evolve to have the structure a non-language brain can already process, the less one has to suppose natural selection needed to give us specialized language-processing software. Language will only seem to be an instinct -- the typical "language is not instinct" cultural hypothesis can't explain why it seems to be an instinct. Nature-harnessing can.

My arguments for language in Harnessed get only at speech processing, not semantics or syntax. For semantics, there are signs that the large-scale structure of the lexicon has been culturally shaped to be good for the brain as well: http://www.changizi.com/dictionary.pdf
 
Reading is actually quite hard for many humans. According to http://nces.ed.gov/fastfacts/display.asp?id=69 nearly a quarter of adults never develop beyond "below basic" skills, even after a lifetime's exposure to written language -- but almost all of them can talk fluently.

You seem to be seriously underestimating how nontrivial reading is, and how much instruction it requires for many people. Spoken language just isn't the same in this regard, even when you factor exposure into account.
 
The exposure that matters is in the early years, not in late childhood or adulthood. And we already know that children can learn visual language at roughly the same age as they can learn spoken language, it just happens that we know this for visual sign language. If a child were instead raised with visual written language (e.g., logographs) instead of visual sign language, it is reasonable to expect a similar result.
 
Sign language is much like spoken language, for sure, both are acquired naturally without instruction.

Reading isn't. Plenty of dyslexic kids see logographs (written words) every day, and still struggle.

I just don't see how what you are saying connects with the scientific or educational literature on dyslexia or real-world literacy; I shan't loop further on this point.
 
I doubt anyone has raised a human child purely on logographs or written words (instead of sign language), and I'm sure the dyslexic kids are getting to reading quite late compared to when kids get speech or sign-language thrown at them. Which means there is a perfectly good reason for why reading happens later and with greater difficulty.

And the fact that sign language can be acquired naturally without instruction suggests that a cultural invention -- which is what sign language is -- can have much or all the same signs of instinct as spoken language comprehension. Sign language surely harnesses us, and probably nature-harnesses us in some way, although I haven't tried to rigorously address this.
Add a comment...