Tuesday, August 16, 2011

Intelligent Music Tastes

Although people do talk about taste as though it can obviously be ordered in terms of good to bad, most would be hard pressed to come up with a good reason justifying those judgments. Indeed, upon reflection they would probably concede that taste is simply a given and that judgments on the relative merits of different tastes boil down to prejudice. After all, one cannot really say that there is something wrong with liking, or not liking, peanut butter. Some say that tastes are just the way we were wired at birth, but that’s not entirely true. We do inherit different amounts of the various kinds of taste buds so the same food may taste different to each of us right from the start. Nevertheless our food preferences – like our religion and many other things – are usually those we were raised with. Preferences are also conditioned by subcultures as well as individual or random associations.

So if tastes are largely determined by genes and culture does that mean they are beyond comparative judgment? Politically correct opinion these days is that it is simply prejudice to say that one culture is better than another. That’s a bit short sighted. When you look at some aspects of culture it becomes obvious fairly quickly that some variations are better than others. For example the Arabian number system that we all use today is clearly an improvement on the Roman number system; writing was a big improvement over having people remember everything; the adoption of systematic error correction - in the scientific method and in legal procedures - has improved our lives immensely; where cultures differ in the degree of senseless cruelty they accept, the more humane cultures are better in that respect, etc, etc.

In principle therefore we can make unprejudiced judgments on phenomena that are at least partially culturally determined. That judgment might legitimately take the form of saying that a taste is to some extent more or less intelligent than the alternatives. Or can it? The cultural aspect the person prefers may well be better than the alternatives but he may simply prefer it because he was raised with it and thinks “my culture is best”. That’s hardly being intelligent. What we need is some evidence that intelligence leans toward certain preferences within a pool of alternatives, within and across cultures and sub-groups.

We know that men like action movies more than women and that women tend to like romance movies more than men, but what we want to know is do intelligent men and women like action and romance movies more or less than the average men and women. Suppose 75% of men and 45% of women like action movies and 30% men and 80% of women like romance movies. If we find that 50% of very intelligent men and 85% of dull men like action movies (and the percentages are 35% and 50% of smart and dull women respectively) then we might conclude that action movies appeal to stupidity. Of course it doesn’t follow that romance movies would therefore appeal to intelligence. We might find a similar shift away from enjoying romance movies too. It might turn out that smart people disproportionately like ‘art movies’.

I was looking at various items in the General Social Survey and came across a number of questions on musical preferences and noticed that almost all of them correlated in some way with the GSS’s mini-IQ scale (a 10 item vocabulary test that correlates 0.7 with a full IQ scale). Many people I shared this with expressed a great deal of skepticism about the notion that musical preferences could be better or worse – let alone smart or stupid. Some suggested the correlations I noticed must be an artifact of say socio-economic status being partly a function of intelligence and that the music tastes simply reflect the SES differences. On the other hand SES may be related to music taste because it reflects intellectual differences. One of the virtues of the GSS is that it allows one to test such hypotheses.

Class and education, race, sex and generation each represent an obvious subculture with respect to music preferences. Who hasn’t heard someone say that classical music is pretentious, that show tunes are for women and gays, that blacks like rap and jazz and whites country, or an old person say the music of today’s youth is just noise? If music tastes are to some degree really due to intelligence then I need to show that intelligence is still related to music preference even after accounting for these sub-cultural influences. Multiple regressions are a statistical technique that allows one to look at the influence of each sub-culture, and of intelligence, independently of each other. So I ran a multiple regression on the degree to which each type of music was liked or disliked, using the vocabulary scale, social class, race (black and white only), sex, age and amount of education as independent variables.

In the table below I show whether the relationship was significant (unlikely to be due to chance) and the direction of the relationship.

- A * indicates significance at the 10% level, all others are at 5% level.

Some well known results pop out - we see that being black increases the chances that you will like Blues, Gospel, Jazz, Latin, Opera, Rap and Reggae, and reduce the chances of liking Bluegrass, Country and Folk. (OK so Opera was a surprise to me.) Older people are less likely to like newer music, like any kind of Rock, Jazz, New Age, Rap or Reggae, and more likely to like older music genres, than the youth. Women like Musicals more, and Heavy Metal less, than men. Classical music does live up to its pretentious reputation but strangely Rap is an upper class taste too.

Also note that the racial, class, gender and generational differences in taste cut across both smart and dull tastes i.e. are not due to intellectual differences between any of these groups. However educational differences in taste do track IQ differences in taste – even though the two factors are independently evaluated in the analysis. One would expect that the role of education would reflect intellectual factors but here it reflects the intelligence of those who designed the curriculum, rather than that of the student.

The interesting result is that IQ is frequently significantly related to how much people like or dislike music. Even after the cultural influences are accounted for 13 of the 18 music genres still show a relationship to intelligence. Furthermore the independent effect of intelligence is usually relatively strong – about equal to all the cultural influences combined. Clearly differences in music tastes are partially due to differences in intelligence, and the connection isn’t explicable in terms of arbitrary connections between IQ and culture.

But what is it about the music that appeals to intelligence or stupidity? One way we can shed light on this is to look at pairs of musical tastes that typically coincide – in the sense that if you like one kind of music you are very likely to like the other. If one of the pair correlates with IQ and the other doesn’t, then we can use this to isolate the intellectually relevant appeal of the music. A liking for Rap tends to coincide with liking Reggae, Jazz and Blues, and the 4 share features that appeal to some people and which others dislike. In the Table you see that the latter 3 are positively related to IQ while enjoyment of Rap is negatively related. What this implies is that it is specifically not those aspects which Rap shares with Reggae, Jazz and the Blues that is relevant. It is what is unique about Reggae, Jazz and the Blues on one hand, and Rap on the other, that taps into the intellectual side of taste. We would need to zero in on what makes people like Reggae, Jazz and the Blues that wouldn’t help them like Rap, and visa versa.

Similarly I found that what separates Folk music from Country & Western or Gospel music, and what separates Oldies Rock from Contemporary Rock and Heavy Metal, also isolates the intellectual side of musical tastes. There is something in Folk music that Country music doesn’t have (or something unintelligent in Country that Folk doesn’t have). There was something in Early Rock music that the Rock music of today, and Heavy Metal, is just missing. What did Jazz have that Rap has lost, or what features has Rap introduced (which appeal particularly to the dull) that Reggae and Jazz never had? The answers aren’t immediately apparent and I’m afraid the GSS data doesn’t have any more to add to the subject.

I would be willing to bet that with the right data I would find that intelligence is relevant to taste differences in the visual arts too i.e. paintings, photos, interior decoration, clothing and hair fashions, architecture, car design, jewelry, etc.

A concept I’d like to introduce here is something I call The Smart Vote. This is based on the notion that IQ is the ability to be correct given a fairly standard experience with the subject matter. High IQ people have shown that they are reliably correct and low IQ people that they are reliably incorrect, across a wide range of subject matter and problem types. So where high IQ people agree on a different answer to that agreed upon by low IQ people, the chances are extremely good that the high IQ preferred answer is correct. The Smart Vote is not what the majority of smart people prefer but rather the direction in which their preferences differ from those of the dull. Where a systematic difference of opinion between high and low IQ groups exists, it usually points to the most correct alternative.

I have shown that The Smart Vote exists for music tastes and that it isn’t accounted for by cultural influences, so it would seem that musical tastes can probably be correct or incorrect. Good and bad taste is not merely prejudice.