Non-Adaptive Pleasure Seeking
If music is not in itself an adaptation, then all the effort we put into making music and listening to music is a form of non-adaptive pleasure seeking (NAPS), or to put it more colloquially, a waste of time, effort and money. Which raises the question: why hasn't music evolved itself into non-existence? Why haven't we evolved the ability to ignore it altogether?
Of course no adaptation is perfect, and the continued existence of music may be a function of the difficulty of evolving a perfect defense against its wasteful seductions.
One thing to remember is that music is not one thing forever fixed. Rather it is a large set of individual musical items – a large set that is constantly growing larger, because people keep on composing new music.
If we consider each item of music as being a separate example of non-adaptive pleasure seeking, then the problem of evolving resistance to music is similar to the problem of evolving resistance to germs. If there was just one kind of germ, which never changed, then no doubt we would have evolved near perfect resistance to that germ long ago.
But actually there are many different kinds of germ, and they are constantly evolving into new kinds of germ. And, even worse, the germs can evolve much faster than we can.
The best that our species can do in response to the threat of germs is to evolve an immune system which adapts itself on the fly to whatever germs it encounters.
So it's plausible that our musical "immune system" operates in a similar manner. Because the set of musical items cannot be known in advance, the system must be programmed in advance to expect some form of overly-musical false speech to exist, and when this category of false speech has been encountered and identified, to label it as "music". Once labelled as music, the false signal generated by its musicality (see my previous article for details) can be suppressed.
In effect such an "immune system" is learning to divide all forms of "speech" into two categories: real speech, with a low typical level of musicality, and music, with a much higher level of musicality.
Mother and Child
Motherese is the special way that mothers talk to their babies. There is also a special way that mothers sing to their babies. If music is non-adaptive, why is it so necessary for mothers to sing to their babies at all? If music is a drug, shouldn't mothers be keeping it away from their young vulnerable children?
But if we assume there is a musical immune system whose operation depends on our ability to identify speech and music as distinct categories, then there is a very good reason why mothers should sing to their children: they are giving their children a headstart with the task of learning this distinction.
A very famous metaphor for music as a non-adaptive pleasure is Steven Pinker's catchphrase "auditory cheesecake". The idea is that cheesecake tastes very nice, even though it is not very healthy in the long run to consume large quantities of cheesecake. Of course our appreciation of cheesecake is not completely non-adaptive, since it does contain food value. If you are starving to death (or even just chronically hungry), and you happen to find some cheesecake, then it's probably a good idea to eat it.
Because cheesecake is a very recent invention, we have not yet evolved any specific resistance to the temptation that it poses.
If music is a more anciently invented pleasure which we have evolved some resistance to, then the "cheesecake" metaphor is an understatement.
A more realistic metaphor would be that of auditory super-cheesecake. A cheesecake utterly devoid of any nutritional value, and indeed slightly poisonous, yet so powerfully delicious that it distracts us from any thought of eating normal food. We cannot learn to ignore its deliciousness, because the tastes and smells that make the super-cheesecake delicious are the same tastes and smells that signal to us the nutritional value of real food, and if our preferences evolved to ignore super-cheesecake, then we would also evolve to ignore all kinds of food.
Furthermore, we can't evolve a resistance to the specific taste of super-cheesecake, because there are thousands of deliciously different cheesecakes, and creative cheesecake chefs are constantly inventing new kinds of cheesecake.
The task of evolving resistance to this seductive pleasure seems almost hopeless, but the problem can be solved by categorisation and disposal after the fact. We eat delicious cheesecakes, but when our brain identifies a cheesecake, it sends it via an alternative route directly to the disposal department. And, perhaps because these cheesecakes never reach our stomachs and reduce our hunger, we still seek to eat normal food to satisfy that hunger.
Mothers feed super-cheesecake to their children, not because the cheesecake has any nutritional value, but because the children need to start the process of learning to discriminate between the category of super-delicious cheesecakes and the category of mundane ordinary foods, so that in future their digestive systems will be able to efficiently send each category to its correct destination.