The Weak Binding of Music-Generated Feelings
In my previous article I developed the hypothesis that music generates feelings which are only weakly bound, in the sense that those feelings can, subject to various constraints, be transferred to perceptions of other things.
If this is the case, it is a fairly unique aspect of music.
This observation leads to the question: Why are music-generated feelings only weakly bound to the music?
The Binding Problem, in general
The Binding Problem remains a major unsolved problem in neuroscience, and there is more than one theory about what the solution might be.
Given the existing state of knowledge about the binding problem in general, I think the most we can do with regard to music is to identify what properties music has which might, plausibly, account for the weak binding that music-generated feelings exhibit.
Two Properties of Music
There are two properties of music that I have identified which may account for the weak binding of music-generated feelings.
Property 1: Globality
The first such property is that the musical quality of any music item is a global property of the musical item as a whole, and it is not possible to account for the musical quality of a musical item in terms of the properties of the individual components of the music.
If we consider situations where very "tight" perceptual binding occurs, especially if perceptions of sound are involved, there is usually an identification of an event occurring at a specific time and place, and this identification results in the binding of the visual perception of the event and the audio perception. For example, seeing a cup fall on the floor and break, and, at exactly the same moment, hearing the cup hit the floor and the sounds it makes as it breaks – our brains presumably depend on the simultaneous occurrence of visual and audio perceptions, with matching information about location, to determine that they should be bound together into a single integrated perception.
If the feelings generated by an item of music can only be attributed to the musical item as a whole, then this attribution is spread out over a long period of time, ie probably at least 10 seconds, and that spreading out may prevent activation of the perceptual binding mechanisms that depend on identifying events occurring within much shorter time frames, ie much smaller than even 1 second.
Property 2: Invariant Perception
The second relevant property of music is that the perceived musical quality of a music item is invariant under two types of transformation, which are:
- Pitch translation, ie transposing to a different key
- Time scaling, playing faster or slower
Neither of these invariances applies to an unlimited degree, but within a certain range they do apply, ie a musical item is exactly the "same" musical item if it is played in a different key, and it is the "same" item if it is played a bit slower or a bit faster.
The raw input data for music perceptions are not invariant under these transformations, ie notes have absolute pitch values, and time intervals between notes have absolute lengths.
These same invariances apply to speech perception, in particular the "melody" and "rhythm" of speech are invariant under these transformations, and these invariances are biologically functional, because:
- Different people have voices with different pitch ranges.
- Sometimes people talk faster, and sometimes they talk slower.
Here I will introduce a secondary hypothesis of opacity, which is that:
In effect, as far as our conscious awareness is concerned, these computations are a "black box", the inner workings of which are hidden.
A plausible reason for why these intermediate states must be hidden from conscious awareness is that the most important thing is the final invariant perceptions, and, any intermediate states will not be fully invariant, and such non-invariant intermediate states would create perceptual confusion.
Of course we do still have conscious awareness of the raw inputs to these computations, because we are consciously aware that different people speak with higher or lower voices, and that sometimes people speak fast, and sometimes people slowly.
What we are not directly aware of is how our brain enables us to perceive the same speech melody and rhythm in the speech of a low-pitched speaker speaking fast as in the speech of a high-pitched speaker speaking slowly.
How does this relate to the weak binding of music-generated feelings?
The specific feelings generated by any particular musical item are very much a function of the specific musical quality of that musical item, and perceived musical quality is invariant under the same transformations of pitch-translation and time-scaling.
It follows that the computations of those feelings, in as much as they depend on perceived musical quality, are occurring inside the opaque "black box", with the conscious part of our brains having no direct access to the internal details of those calculations.
As a result of this opacity, we are completely unable to deconstruct the globality of the musicality of the musical item, and this enforces the "spread out" nature of the attribution, which is what creates the weak binding.
Extra Note: Opacity and the Difficulty of Composition
The opacity hypothesis can also explain why it is not at all straightforward to compose good quality original music.
If we had direct access to the internal details of how musical quality of musical items was computed in our brains, then it might be quite easy to figure out how to compose new music.
But given that we don't have that direct access, anyone who wants to compose music has to create their own separate models of how musical quality is computed, from empirical observation of musical quality as a purely global property of each musical item. These models will reside somewhere in the composer's brain, completely separate from the "black box" of computation that they are models of.
Also, in as much as they exist, we know that these models are both incomplete and approximate, because there isn't any composer alive who knows how to compose the best possible examples of all possible types of music.