What is Music A Superstimulus For, If Anything?
If music is a superstimulus, it has to be a superstimulus for something.
There is one thing other than music which is like music, and that thing is speech.
There are important similarities between speech and music:
- Speech has rhythm.
- Speech has "melody".
- A lot of music actually includes a form of speech, ie song lyrics.
So if music is a superstimulus for something, it's probably a superstimulus for speech.
However, music cannot be a superstimulus for speech per se.
Rather it has to be a superstimulus for some aspect of the perception of speech.
In other words, music is speech that feels like it is very [X], where X is the aspect of perception in question.
So, what is X?
Positive vs Negative
There are many things that X might be.
Whatever X is, we expect the perception of it to have some effect on how the listener responds to speech, or on how the listener processes speech information.
This effect might be positive, or it might be negative.
To give a concrete example, I will consider the possibility that the perception of X influences how strongly the listener pays attention to speech:
- If X is a positive aspect, then a listener will pay more attention to speech with a high level of X.
- If X is a negative aspect, then a listener will pay less attention to speech with a high level of X.
Music is something that people like to pay attention to, so this would seem to argue strongly that music is a positive superstimulus.
I've used attention here as an example of an aspect of processing that the perception of X acts on.
But, if you continue to read, you will learn that the real target of the perception of X, ie the component of information processing that it affects most strongly, is not attention. It's something else.
Positive implies unconditionally "better"
If music is a positive superstimulus for some aspect of speech perception, then this implies that music is perceived as being a "better speech than speech".
If this is the case, then speakers would be expected to exploit this perception by making their speech more musical in any situation where they are speaking and they want their speech to have more effect on the listener.
But this doesn't happen.
In particular, people never musicalize conversational speech.
(You can try, but I guarantee you will ruin the vibe of the conversation.)
If not positive, then maybe negative ...
If the evidence suggests that music is not unconditionally "better" than speech, and we still think that music might be a superstimulus, then we need to consider the possibility that music is a superstimulus for a negative aspect of speech perception, and that music is a "worse speech than speech".
If music is a negative superstimulus, then we need to identify:
- The aspect X of speech perception that music is a superstimulus for
- The component C of speech processing that is negatively affected by the perception of X
A Model of Conversational Speech Processing
I have developed a model of conversational speech processing which is consistent with what I think I know about the effects of music on mental state, and which is consistent with the observation that people do not "musicalize" their conversational speech, not ever.
This model explains how it is that music completely suppresses one of the steps in the processing of speech information by a listener.
In this model:
- X is the perceived non-spontaneity of a speaker's speech.
- C is the determination of the listener's belief about the truth of what the speaker said.
Firstly we assume that a speaker has said something to the listener, as part of a conversation.
The basic steps of processing, in the normal (non-musical) case, are:
- The listener determines the meaning of what the speaker said.
- The listener determines the emotional significance of the meaning of what the speaker said.
- The listener determines their belief about the truth of what the speaker said.
- The listener responds to the speaker, where their response includes an indication of their belief about the truth of what the speaker said.
There is an additional consequence of step 3 on the result of step 2, which is:
- The intensity of emotional response to the determined meaning is reduced according to the determination of belief about the truth of what was said.
In other words, if it's probably not true, then there is no reason to feel any strong emotion about what was said.
A model of conversational value
The basic model is a "happy case" model, where the listener assumes that the speaker's speech is genuine conversational speech.
What do I mean by "genuine"?
Genuine conversational speech is speech that directly reflects what the speaker has decided to say "in the moment", based on the current context of the conversation.
The context of any actual conversation is constantly changing, as a direct result of what the participants are saying.
Any attempt to say something that is rehearsed, or pre-planned, will not be an honest reflection of the speaker's current thought processes based on the current context of the conversation.
This suggests a possible extra component in the speech processing model, which is:
- The listener has to decide whether it is worth the effort to evaluate the truth value of what the speaker said, and whether it is worth the effort to respond to the speaker with an indication of that evaluated truth value.
If a listener responds with a genuine "in the moment" evaluation of what the speaker said, but the speaker's original speech was not "in the moment", then there has been an uneven exchange of information.
In order to enforce a fair exchange, the listener has to determine, as best they can, whether or not the speaker's speech is truly "in the moment".
In other words:
- Is the speaker's speech perceived to be spontaneous?
Including conversational "value" in the model
A revised model, that includes this evaluation, is as follows:
- The listener determines the meaning of what the speaker said.
- The listener determines the emotional significance of the meaning of what the speaker said.
- The listener determines whether of not the speaker's speech is spontaneous.
- The listener determines their belief about the truth of what the speaker said, but, the amount of effort put into this determination is proportional to the perceived level of spontaneity in the speaker's speech.
- The listener responds to the speaker, where their response includes an indication of their belief about the truth of what the speaker said.
Applying the model to perception of music
We can now account for what happens in the musical case, where the "speech", if any, consists of the lyrics of the song:
- The listener determines the meaning of what the speaker said (ie the song lyrics).
- The listener determines the emotional significance of the meaning of what the speaker said.
- The listener determines that the speaker's speech is extremely non-spontaneous.
- The listener's determination of their belief about the truth of what the speaker said does not happen at all.
In the normal non-musical case, the evaluation of truth value results in a reduction of the emotional intensity.
But in the musical case, the evaluation of truth value does not happen, so the reduction of emotional intensity does not happen.
The result is that the negative effect that music has on the listener's processing of speech causes what appears to be a positive increase in the emotional response to the speaker's speech.
But the positive increase isn't really an increase – it's just a non-occurrence of a decrease.
The final result of all this is that the listener's brain remains in an intermediate processing state, where:
- The meaning of what a speaker has said is represented.
- The emotional significance of that meaning is represented, at full intensity.
- No attempt has been made to evaluate the truth value of what the speaker said – the meaning remains in a completely hypothetical state.
Link to my original publication of this theory: Music Is A Superstimulus For The Perception Of Non-Spontaneous Non-Conversational Speech.
That article also includes discussion of other aspects of the theory, including:
- The emotional effect of music on meanings that don't necessarily come from song lyrics.
- Evidence that music is not just a "side effect" of the conversational processing model, and it must serve a biological function.
- Speculation about what the biological function of music might be.