172A. Cognitive Psychology of Music (Introduction) Undergraduate, non-major course Week 4

Summary of lectures

 

i s s u e s   c o v e r e d

Summary of Seashore & introduction to Davies: two different takes on the Psychology of Music.

Davies on the relationship between music and productivity, consumption, & emotion. Musical 'laws'.

Models of knowing. Implicit / Explicit knowledge.

Memory limits. Data reduction; Categories.

 

 

 

Seashore versus Davies
    

 

As a summary, Seashore's model can be classified as Cartesian for the following reasons (among others):

        a)     The instances where his isomorphic mapping between physical and perceptual variables breaks down he calls normal illusions. Such a term reflects the belief that the physical world is true and right while our perception of it is false, wrong or inadequate.
        b)    True to the idea of reductionism/atomism he insists that studying musical units (defined in terms of the 4 acoustical variables that describe sound and therefore music) will tell us all we need to know about music. The whole is the sum of its parts.
        c)    Although his simple model is reliable (thanks to its simplicity and relative internal consistency) it is nothing more than a model (and one that, as we have seen leaves many questions unanswered). Still he uses it as a basis for unjustified over-generalizations. Based always on the physical frame of reference, he argues for a genetic basis of musical talent, raising numerous validity problems.

With reference to the model of musical communication put forth by Kendall & Carterette, 1996. Seashore's model is limited to a small segment of this larger model, the segment dealing with the acoustical signal. The more holistic model of musical communication suggests that in order to adequately represent musical phenomena we need to introduce more frames of reference, other than the physical and physiological ones. (As we will see, however, most empirical studies in music still favor the physical and physiological frames of reference).

 

J.B.Davies, in contrast to Seashore, offers a rather Humean approach that is closer to the Kendall & Carterette model:

        a)     Perception defines reality (one of the most critical aspects of a Humean thought). ".....the ear is the final court of appeal..." (compare this with Seashore's view of reality as objective and perception as subjective and the source of "normal illusions").

        b)     Symbols (notes, graphs, etc) are our inventions; they are attempts to represent music in a manner that communicates its various operational definitions, without necessarily strong or proven connections to the way we experience music as listeners.
"...signal graphs are simply convenient modes of description..."  compare this with  Seashore's exclusive concentration on symbolic (on paper) representations of music and his belief that the sound signal is isomorphic to sound and music. 

        c)     Music is determined by the listener, meaning that in essence people effect music.
Compare this with Seashore's underlying belief that music effects people. It was such a belief that supported the rise of Muzak (see below) around the 1950's.
Music is at every level created by people. And this is as much the case for listeners as it is for composers or performers. As mentioned previously, listeners are not passive observers being effected by music. "Listening music" means "configuring music". The act of listening involves a configuring operation that turns a series of abstract incoming sounds into "meaningful" patterns, relationships, temporal wholes: "...patterns exist in people and not in sequences of tones..."

Davies in his book: "The Psychology of Music" (1978), essentially argues that
a) one does not have to be Cartesian to do a science of music,
b) adopting a more Humean approach does not have to be accompanied by a total dismissal of Cartesian models but by a recognition of their limits, and
c) such an approach "... can serve a very useful purpose in exposing certain beliefs as pure mysticism..[by] making such things as 'good taste', 'sensitivity', 'musical understanding', and so on, into comprehensible entities...rather than magical properties of some sort of musical priesthood."

   

 

Davies on music & a) productivity b) industry &
c) emotions (introduction to the semantic differential method.)
Davies on 'Musical Laws'
     

 

Davies launches an attack to what he refers to as "cherished notions" of a cause-effect relationship between music and:

a) Productivity,         b) Consumption,           c) Emotions

 

a) Productivity: 'introducing music in the workplace has a direct and predictable effect on the workers' output.'

i.e. Study on the effects of music on productivity by H.C.Smith. Smith used mainly 3 independent variables:
i) type of music (popular / semi-classical),
ii) time of presentation (day / night),
iii) presentation format (continuous / periodic).
He concluded that we get a 7% increase in productivity (average) when popular music is introduced in the workplace , for 12% of the time during the day and 50% of the time during the night. (He is referring to assembly lines: repetitive task.)
His results have been largely dismissed because of the great variability of the data (the results were spread over such a large range that the 'average' does not give an adequate representation of the observations.)
As Davies points out, two very important factors are usually ignored in such studies:  
a)  the relationship of the listener to a given piece of music (essentially what you know), and
b)  general contextual issues (general state of the worker/listener, type of task, etc.)
He therefore concludes that: "...Studies like these [introduce models that] have little scientific value, [they do not have enough predictive power to base firm decisions on] but appear in large part to be responsible for the 'music in industry' myth."

 

b) Consumption: 'spending attitudes of consumers can be directly manipulated through music.'

A good example of the application of this "cherished notion" is 'Muzak':
Muzak:    N.Y. company that sells music programming, claiming to have a scientific psychological understanding of the ways music can effect people in the workplace and in commercial environments. The company creates its own musical pieces based on the idea that music you do not know does not attract your attention (allowing you to concentrate on other tasks/activities) but can still act on you subliminally (influencing your performance on those other tasks/activities).
Muzak pieces are always instrumental (no voice), familiar sounding but not really recognizable. They are arranged in such a way as to have gradual increases and sudden drops in stimulus intensity (speed/tempo, density of orchestration, loudness levels, etc.) that repeat in 15min (approx.) intervals. The term Muzak is often used generically to describe background music for the workplace and elsewhere, supposed to have causal/predictable/isomorphic relationship to the resulting behavior of the listeners, at a subconscious/subliminal level. (Good example of Cartesianism.)

 

c) Emotions: 'music is the language of emotions.'

Question: How can we operationalize emotion (so that we can subject it to the process of empirical investigation) in a way that does justice to its variability and multidimensionality?

The Cartesian approach (since the end of the 19th century) operationalizes emotion in terms of physiological responses:
a) heart rate,  b) respiration rate,  c) GSR, (galvanic skin response),  d) ECG (measuring brain signals),  e) electromyography (measuring muscle response),   f) blood pressure, etc.
In 1978 Dainow reanalyzed all available physiological data and showed that they do not reveal any consistent pattern/relationship between changes in specific emotional states and changes in physiological states.
    As one of few exemptions we can mention electromyography studies by W.Sears, in the 1950's. He showed that for major/fast background music, there was a relative increase in muscle tension, with subjects subconsciously assuming an erect posture, while for minor/slow background music there was a relative decrease in muscle tension, with subjects subconsciously assuming a more slumping posture.
    As was the case with 'productivity', emotion must involve some kind of knowing. But how do we get to an explicit definition of something that is supposed to represent an internal state?

Related questions:    a)   are all emotions a single emotion varying in magnitude?    b)   if not, what differentiates emotions?    c)  do emotions, feelings, or moods describe the same thing?    d) does experiencing music relate the same emotions as experiencing other aspects of everyday life?   e) if not can we construct a meaningful link between music and 'mood'?
Such questions will be addressed after the mid-term exam. Emotion in music will then be linked to the idea of musical meaning and will be understood in cognitive terms, with music suggesting rather than eliciting emotions.

 

Davies points out that Science is simply one of the possible ways to look at the world (the way he adopts) that is based on the ideal of replicability/reliability and the resulting possibility for a) reduction of uncertainty and b) prediction.
He draws attention to the fact that science does not examine reality but models of reality that are based on definitions invented by scientists. Such definitions run the danger of introducing abstractions which cannot support the desired inferences/generalizations, reducing the validity of models to absurdity.
Still, with proper precautions, Davies claims that issues of value, beauty, feeling, or emotion -traditionally addressed by Aesthetics- can also become subject to scientific investigation (experimental aesthetics), "...making such things as 'good taste', 'sensitivity', 'musical understanding', and so on, into comprehensible entities common in varying degrees to almost all people, rather than magical properties of some sort of musical priesthood."

Since emotions cannot be operationally defined (and therefore measured) in terms of physiological variables (Cartesian model) an alternative has to be introduced. 'Measurement' is a necessary ingredient of science whether 'Cartesian' or 'Humean'.

One way to find out about peoples emotions (other than connecting them to electrodes) is asking them. Of course such an approach introduces numerous complications: Will the people asked (experiment subjects) tell the experimenter
i) how they actually feel?
ii) how they think they feel?
iii) how they think they are expected to feel?
iv) what feelings they are being reminded of? etc.

As expected, by recognizing that a model describing the relationship between music & emotion has to go beyond a simple physiologically-based mapping, we have increased the validity of the model (it better represents observed relationships) at the expense of its reliability (it becomes increasingly harder to attribute results of an experiment to something specific.)
One way to operationalize 'asking people' is through 'verbal attributes' methods such as the 'semantic differential' method.

Q.: Where does the term semantic differential come from?
A.:"Sema" is a Greek word meaning "sign". In the way it is used here it refers to "words" that act as signs pointing to something non-linguistic. I.e. The word "book" is just a collection of scribbles on paper and means nothing unless we understand it as a sign pointing to the "thing" that we call book. If we agree on this label then we can use this and other such labels without having to actually have present in front of us the things we are talking about.
Any "semantic" operation therefore attaches labels that we assume refer reliably to the "things" that -in terms of communication- often replace.
The "semantic differential" method is a method that tries to put down in terms of words what someone understands (meaning) or feels (emotion). In other words it assumes that we can grasp someone's understanding or emotional state based on the labels he/she chooses to attach to them.
The "differential" part indicates that instead of asking people to give us single labels (adjectives i.e cold) we ask them to provide a label as it is "differentiated" from another label (usually its opposite i.e. cold-hot). Such an approach allows us to "differentiate" between degrees of 'coldness' AND see how they relate to degrees of 'hotness" SO THAT the degrees of coldness/hotness -being on a comparable scale- can be "differentiated" in a comparable way. If that was not the case statistical analysis and meaningful interpretation of the results would be very difficult/'unreliable'.

Semantic Differential Method
Subjects are asked to rate a stimulus on a scale outlined by two polar opposites (i.e. good - bad). The responses are recorded as numbers with a 'fineness' determined by the experimenter.
Such a method, along with the above mentioned complications, introduces the additional one: Are the words used being interpreted by the subjects as the experimenter has intended?
Some experimenters avoid using verbal attributes all together to escape such 'semantic' complications. However, even if the facial expressions or body movements of subjects are somehow turned into data, the issues of interpretation will not be resolved. Facial expressions, gestures, and body movements are as much culture dependent and codified as words in language. And just as there is no 'universal' language (one that everyone uses, understands and communicates with in the same way) neither are any other forms of communication, vocal or otherwise.

 

Musical Laws

In the Davies' readings it is argued that the so-called laws of music (i.e. of harmony, melody, consonance etc.) are essentially inventions that have been defined on the notational frame of reference. In general, music theory presumes that there are connections to other frames of reference without testing them.
Questions that can be raised:
        a)     Do symbols (on paper) map on perception as presumed?
        b)    Are there any evidence that such 'laws' have to be what they are (so that they can be called laws)?
        c)    How do we explain that musical examples provide us, in general, with more exemptions to the rules than adhere to them?

    " There are very few musical universals, and the story of the development and evolution of music is better expressed in terms of the progressive and systematic violation of rules than in terms of adherence to them. 'It seems that the voluminous edifice of music theory rests on sand', wrote Moles (1958)."
    " To say that musical theory is based on a set of largely invented rules rather than on a set of natural laws is not to diminish it, however. The rules define the area within which the artistic endeavor takes place. They make possible such things as aesthetic deviation, confirmation or disconfirmation of expectancy, and the deliberate violation of rules for aesthetic effect, all of which are critical events as far as musical cognition is concerned."

 

Models of knowing; Implicit and Explicit knowledge.

The Seashore model can be schematized in terms of an isomorphic mapping machine (below), with an input (i.e. frequency) entering a translator (i.e. the ear), where the "...wondrous transformation of matter to mind..." occurs, giving us a single and predictable output (i.e. pitch).

    All information we receive (input) is processed in memory, and behavior (output) is closely related to that process. We can distinguish two interrelated levels in memory:
    a) Short-term memory dealing with data as they come in ('first perceptions'). It holds data for ~30secs or less.
    b) Long-term memory dealing with existing data (more than ~30secs. 'old').
The implicit knowledge rules effect the translation from short-term to long-term memory; in other words they make understanding possible.

    The human mind is a limited capacity system (it has limited channel capacity, in Information Theory terms.) In other words, there is a limit to the amount of information that can be processed before we have an "overflow". This limit to the rate of information that short-term memory can deal with is expressed by:
    Miller's 7±2 rule: A maximum of only 7±2 unrelated / unpatterned / random events out of a uni-dimensional continuum can be stored in short-term memory. This means that the complexity of the outside world has to be somehow reduced to become perceptually meaningful. Music would be impossible without some sort of data reduction. Implicit rules have evolved to help us deal with this 'data overload' by abstracting criterial attributes (essential aspects) from some input relative to existing schemata (knowledge structures.)

If we apply Miller's rule to music perception we can easily see that, since the accuracy with which perception can resolve sound (in terms of JNDs) is high enough to allow for practically infinite possibilities, the musical systems we invent must involve parameters -beyond JNDs- that reflect this need for data reduction, and processes that allow for maximum variety (maximum possible combinations) within the resulting set of reduced data. 

    Data reduction: A process that takes the infinite variability of the world and reduces it so that it can be workable by the short-term memory. The term actually refers not so much to actual elimination of individual pieces of information but to a reduction in complexity / randomness / uncertainty, and an increase in predictability based on finding redundancies. It refers to the organization / patterning of random information into larger units (categories).
    Categories: As it has been argued by Jerome Brunner, categories are psychological constructs, connected to the idea of data reduction. Large number of data can be enveloped in a single category which is then understood in terms of the criterial attributes that link these data into a unit.
(Criterial attribute: An attribute whose presence is necessary to the existence of a category).
Categories can be multidimensional (meaning not only that they may incorporate many criterial attributes but that those attributes may also be interrelated.) This is why categories can vary in more than one dimensions (attributes) simultaneously without loosing their identity.

Ethnomusicology Department - UCLA©