A Hypothesis About Two Perceptual Systems

21 January, 2020
The Dual Perception System Hypothesis: The perception of music depends on two distinct perceptual systems – an immediate but imprecise perceptual system and a slow but precise perceptual system.
Music generates feelings when the precise system perceives a change, and the imprecise system does not "understand" why the change occurred.

The Two Perceptual Systems

Many features of music and music perception can be explained if we assume that the perception of music depends on two different perceptual systems that are largely separate from each other.

The fundamental properties of these two distinct systems are precision and immediacy.

The first system is immediate, but imprecise. It is the perceptual system that gives the listener direct and immediate awareness that a musical note has occurred.

The second system is precise, but it is not at all immediate. This system is sensitive to the accuracy in timing and pitch of a musical performance, and it provides information about the harmonic relationships between all the notes that occur in a musical item, as well as information about the relationships between beat frequencies and periods that occur in the rhythm of the music.

In the rest of this article I will label these two systems as the imprecise perceptual system and the precise perceptual system (or, for brevity, the imprecise system and the precise system).

We can describe the two systems in terms of what they know, when they know it, what internal state they maintain, and how much that internal state is exposed to conscious awareness.

The Imprecise System

The imprecise system responds immediately to the occurrence of notes and other sounds in music. It knows that some sounds have high pitch, and other sounds have low pitch. It knows if the melody is going up or down. It recogises exact repetition of notes and it recognises exact repetition of simple phrases.

The imprecise system does not necessarily know that a note occurring at one point in a tune is the same note that occurs at a different point in the same tune.

The imprecise system is not very good at distinguishing or recognising particular musical intervals.

The imprecise system maintains perceptual state relating to what has just happened, but this perceptual state decays very rapidly, which limits how much it can know about relationships between musical events separated by moderate amounts of time, even when separated by as little as one or two bars of the music.

The imprecise system tends to simplify its perception of scales by regarding each step in the scale as a single step of indeterminate size.

For the sake of discussion in this article, I will consider a musical example constructed from the justly-intoned diatonic scale (which consists of steps or 1 or 2 semitones each), and harmonic intervals of 3, 4, 5, 7, 8, 9 or 12 semitones.

To give a specific example, consider a musical phrase consisting of the notes C D E, followed by a second musical phrase D E F:

The Precise System

The precise system is very sensitive to the precise values of pitch and timing.

However, the information that it exposes to conscious awareness is a function only of relative pitch and relative timing.

(To put it more mathematically, the output of the precise system is invariant under pitch translation and invariant under time scaling.)

Also, even though the precise system performs calculations that depend on the sizes of intervals between individual pairs of notes, it does not directly "tell" consciousness what the identity is of any particular interval between any particular pair of notes.

The precise system retains considerable persistent state, and this state retains precise information about pitch and timing. However, consciousness does not have any direct access to the internal details of this persistent state.

The time scale of this persistence is of the same order of magnitude as the duration of a simple item of music, ie one or several minutes.

Additionally, the persistent state of the precise system has two major sub-components, one of which is resettable, and the other which is persisted indefinitely, ie for the full duration of a musical item.

The resettable state tends to be reset, if and when:

In modern popular music, the resetting of the resettable state usually happens when a chord changes.

The non-resettable state contains information about a musical item that is constant for all of the item, such as the scale that the melody notes belong to, and the regular beat frequencies underlying the rhythm of the music.

Interaction between the Two Perceptual Systems

The differences between the two perceptual systems have consequences in terms of what they can "tell" about what they know to our conscious awareness:

The Hypothesis of Inexplicable Change

The hypothesis of Inexplicable Change is that music has a strong effect when perceptual changes are reported by the precise perceptual system, and the imprecise perceptual system does not understand why these changes have occurred.

The strongest type of change will occur when the resettable state of the precise perceptual system is reset, ie when a chord change occurs.

Such a change will be most inexplicable when it arises from a change that the imprecise system is not capable of perceiving.

For example, in the very simple example given above, of C D E followed by D E F, a straightforward chord accompaniment would be C major for the first phrase and D minor for the second phrase.

The imprecise system "knows" that the note D is higher than the note C, but it does not "know" that the interval D-F (3 semitones) is a different interval to C-E (4 semitones).

So when the precise system reports a partial change of state as a function of the chord change from C major to D minor, the imprecise system does not fully understand all the things about the D minor chord that are different from the C major chord.

Thus the inexplicability.

The "World" Hypothesis, and "World Change"

In my previous article, The "World" Hypothesis, I described the concept of "world change", and I suggested that each musical item is its own "world".

One problem with this identification is that just "being" a world is not the same as the world "changing". The concept of resettable state can fill in this gap in the explanation.

For example, as follows:

Expectational Theories of Music, and Failure to Learn

Expectational theories of music are theories that attempt to explain music by supposing that:

The basic problem with this type of theory is that it can't explain how we manage to enjoy listening to the same individual music item on many different occasions.

There exist other forms of entertainment that do depend on interactions with the audience's expectations. These include humour, very specifically, and also story-telling in general.

However, these forms of entertainment are of the "once-is-enough" variety. This is because their entertainment value depends on the occurrence of things which are different from what the audience expects, but, after the first occasion, the audience learns to change their expectations, so on a second or third occasion, the joke or the story no longer mis-matches the audience's expectation, and it is no longer entertaining.

But in contrast to the situation with jokes and stories, when a person likes an item of music, they will almost always listen to it more than once, and usually much more than three times.

If music has something to do with expectations, then we have to suppose the existence of some kind of "failure to learn", in order to explain how a person can repeatedly enjoy listening to the same item of music.

Failure to Learn with the Imprecise and Precise Perceptual Systems

The Inexplicable Change Hypothesis is a type of expectational theory. As I have explained, it proposes that musical effect occurs when a "change of world" is reported by the precise perceptual system, if the change is unexpected based on information provided by the imprecise perception system.

We can suppose that those parts of the brain responsible for learning expectations will have a strong tendency to make their predictions based on information from more immediate forms of perception, because, usually, time is of the essence.

So in the case of music, expectational learning systems will preferentially base their learning processes on information provided by the imprecise perceptual system. But, given that music is based on things that the imprecise system does not know about, that learning will fail to happen.

(Presumably the expectational learning systems do eventually take account of the information provided by the precise perceptual system, and this explains why music eventually gets boring when you have listened to it too many times.)

Learning to Sing

Expectational learning is not the only learning process related to music that fails.

Most people have very precise perception of music, and most people are quite sensitive to the difference between a mediocre performance and a high quality professional performance.

But when people try to learn to sing, the result is often quite bad. And often people do not realise how bad their own singing is, unless they go to the trouble of recording their own singing and then listening to the playback.

We can explain a lot of the difficulty of learning to sing if we suppose that the learning processes preferentially only use information from the imprecise perceptual system. We would then expect to see things like the following:

One consequence of the dual perceptual system hypothesis is that there cannot exist any musician whose "imprecise" perceptual system is as precise as their precise perceptual system, because such a person would not enjoy music in the first place.

This further implies that anyone who wants to learn to sing well has to find ways to overcome the limitations of the imprecise perceptual system. This might include: