I read up a bit more on theories of music’s function – quotes below.

For any signal a key question is why observers should believe it. For example, many animal sounds come with clear simple reasons to believe them. A loud roar can show the physical strength of its source, and its willingness to expend energy. A soft coo, in contrast, can show that the source is relaxed and comfortable.

With language we humans can say far more things than we can directly show via the way we say it. Even so, speakers and listeners usually have a common interest in agreeing at least on what was meant by what was said. So a listener can often at least credibly believe that the speaker meant to claim a certain thing, even if that listener doesn’t have further reasons to believe that what was said is true.

Music seems to tell emotions, and sometimes it is enough to tell what emotion you claim to have, even if listeners have no further reason to believe that claim. But we seem to have far more types of music than we have kinds of emotions to tell. And the main music puzzle seems to be on the listener end – why are we built so that hearing emotional music evokes similar emotions in ourselves, at least when we are in a receptive frame of mind. Why does music evoke emotions more reliably than does hearing a simple verbal description of the same emotions?

A similar thing happens with stories, at least when we are in a receptive frame of mind. Story-tellers can make us like some people, things, and events, and dislike others, using simple tricks that we all know are evidentially unfair. Why are we as listeners so often eager to enter story and music receptive frames of mind, allowing others to more directly control our emotions, attitudes, and opinions about people, acts, and events?

Yes going through a process like this together that can bond people, as they end up knowing that they have acquired similar emotions and opinions. But why does such a capacity even exist? Why are our minds so vulnerable to raw back door control over what we think and feel?

Yesterday I suggested that this capacity exists exactly because it lets people see that they acquire similar emotions, attitudes, and opinions, a similarity that goes beyond any similar evidence they may each hold supporting such things. As long as we are confident that our associates’ emotions and attitudes are so manipulable, we can be reassured that their attitudes are similar to ours.

But if so, why haven’t some of us evolved a capacity to fake such vulnerability, appearing to enjoy music and stories, but not allowing them to change their attitudes and opinions toward people, acts, and events? A similar issue arises with wanting our associates to internalize key social norms, such as against murder. If our associates could act horrified by murder, but not actually be reluctant to murder, we’d have to be a lot more wary of them.

While some of us do seem especially good at faking feelings, most of us are lousy liars. That combined with our strong censure of apparent fakers produces an equilibrium where most of us stay pretty close to acting vulnerable to key norms, stories, and music.

Those promised quotes:

Steven Pinker … argues that music is merely “auditory cheesecake” – it was evolutionarily adaptive to have a preference for fat and sugar but cheesecake did not play a role in that selection process. Adaptation, on the other hand, is highlighted in hypotheses such as the one … which posits that human music evolved from animal territorial signals, eventually becoming a method of signaling a group’s social cohesion to other groups for the purposes of making beneficial multi-group alliances.

Another proposed adaptive function is creating intra-group bonding. In this aspect it has been seen as complementary to language by creating strong positive emotions while not having a specific message people may disagree on. … A different explanation is that signaling fitness and creativity by the producer or performer in order to attract mates. …

Steven Brown … argues that “music and language are seen as reciprocal specializations of a dual-natured referential emotive communicative precursor, whereby music emphasizes sound as emotive meaning and language emphasizes sound as referential meaning.”
…
Joseph Jordania … suggested that rhythmic laud singing and drumming, together with the threatening rhythmic body movements and body painting, was the core element of the ancient “Audio-Visual Intimidating Display” … a key factor in putting the hominid group into a specific altered state of consciousness which he calls “battle trance” … for many social animals, silence can be a sign of danger, and that’s why gentle humming and musical sounds relax humans. (more)

Critics of this [signaling fitness theory] note that in most species where singing is used for the purposes of sexual selection (through the female choice), only males sing (as it is male, who is mostly trying to impress females with different audio and visual displays), and besides, males as a rule sing alone. Among humans both males and females are ardent singers, and making music is mostly a communal activity. (more)