As a speculation it’s really pretty good. Many years ago there was a Scientific American article about why people like music. It was long and complicated but the tl;dr would go something like:

Well-liked music of any genre tends to contain fractal patterns. Doesn’t matter if it’s jazz, classical, rock or whatever. If you probe our peripheral nervous system you get a lot of white noise, but the closer you get to the central nervous system the more fractal the signal becomes, as if our nervous system is filtering out the noise and letting the fractal part of our perceptions get through to our brains. This makes it very likely that our thoughts and memories are fractal patterns, which means that on a purely mathematical level there could be similarities between patterns that encode ideas that aren’t related by context - for example, when a piece of music makes you think of the ocean, or flying birds, or the big city, it’s probably because the music’s pattern and the idea’s pattern in your head are mathematically similar.