The musical mind
19 Jan 91
How do we perceive music? Do Mahler and Madonna have
anything in common? A forthcoming television programme
reveals some of the answers
The challenge of making a one-hour documentary
programme about the activity of listening to music
at first appeared to hold about as much potential
for prime-time television viewing as watching wood
warp. But Michael Howes, who wrote and produced
the programme, had previously produced both
music and science documentaries, and was determined that this mix of
arts and science could work as entertainment. He also wanted the
programme to follow the trend towards 'live science' on television: doing
the experiments in front of the camera, so that any shortcomings in
theory or method would be there for all to see.
We also had to explore ideas for turning an aural experience into a
visual one, reasoning that because psychologists have repeatedly
stressed the dynamic, constructive and above all the active nature of
listening to music, there ought to be something entertaining as well as
instructive that the cameras could be pointed at. We thought that any
results worth their salt ought to be sufficiently robust to translate out
of the sound laboratory into the real world.
The study of how people perceive and understand music has
mushroomed over the past 10 years. Psychologists have revealed the
importance of higher-level mental processes, with implications for
cognitive models and internal representation that go way beyond the
subject of music. Because of this broadening basis, it is no longer a
simple matter to decide what to include in a single programme about
listening to music. But the first point to make is that listening is a
musical activity, rather than a passive process. A high level of 'listening
performance' requires complex mental skills, involving perception and
cognition. To obtain satisfaction from this exercise, listeners also rely
as
much on what they know, or 'bring with them' in terms of musical
culture and history, as on the musical material itself. This must be true
for non-musicians as well as musicians, and apply equally to music by
Mahler, Miles Davis and Madonna. If there are differences in people's
response to music, they are for the most part differences of degree
rather than type.
Yet it remains difficult to define what music 'means', because a string
of
notes seldome refers to anything specific in the world in the way that
strings of letters forming words do. Attempts to make direct linguistic
analogies have for the most part proved unsuccessful and arbitrary; for
example, the fact that a group of musicians might agree that a
particular string of notes means 'an ascending feeling of joyous release'
or whatever, reveals more about the personal bias of the subjects and
the experimenter than about the way we understand music.
So the answer to the question 'what does music mean?' is to a great
extent philosophically imponderable. But it is not difficult to show that
people respond to music, and that different people respond in the same
ways, and often at the same point in a piece of music. A simple
experiment in this programme demonstrates this convincingly by taking a
two-channel polygraph into a jazz club to observe the changing
characteristics of pulse rate and skin conductance. Such changes are
linked to altered mood states, and back up the listeners' statements
that music 'does things to them' - a simple demonstration that our
observations are not just an illusion or a self-deception. And even our
brief test reveals a degree of agreement between different people,
suggesting that people can form a consensus about a musical
experience.
Further support for this idea comes from an experiment to identify
'musical hot-spots', in which 18 adults listened to an extract from
Beethoven's Pathetique sonata. They were asked to keep their eyes
closed and raise one hand when they felt partially aroused by the
music, and both hands when they felt a more intense arousal. The
results show that while a few individuals experience arousal when others
do not, there is a marked consensus at particular points in the music.
Furthermore, this consensus is not simply due to prior knowledge. Even
with unfamiliar pieces, particular forms and structures create 'hot-spots'
on first hearing.
While it is probably absurd to suggest that a particular note or harmonic
grouping has any precise meaning, there are nonetheless more general
mood states that particular music evokes. We demonstrate this by
asking 18 test subjects to choose which one of six pictures they would
match to the music being played at the time. Three pieces were used:
Oceans by Anthony Phillips, Wieniawsky's Scherzo Tarantelle, and a
piece of traditional funeral gamelan music from Java.
Two interesting observations emerged from this study. The most popular
choice for Oceans was a picture of sailing boats, in spite of the fact
that none of the subjects had heard the music before and none knew
its name. Six subjects, however, favoured a 'heavy industry' scene. For
the Wieniawsky, most people selected our specially created drawing of
an imaginary 'spiky object'. Both these results were fairly predictable,
and illustrate an appreciation of conventional thematic forms of
reference on the part of our listeners.
We also predicted the confusion created by the gamelan music, which
employs different tonal conventions. Only one person associated this
music, which traditionally accompanies Javanese funerals, with our
pictue of a funeral in Pontefract. This points to another important issue,
namely the role of culture in determining what is seen as musically
appropriate to particular events and circumstances.
A reasonable, but unprovable, explanation for musical behaviour of all
kinds is that it hs its roots in the natural world. Particular sounds
common in nature, such as the cries of animals or birds, the keening of
the wind, the beat of the sea, are seen as the bedrock from which a
musical aesthetic experience springs. Such an idea is supported by the
fact that sensory components of our auditory system respond to alarm
calls - sudden sounds, the wailing of a baby, and so forth.
Consequently, music is seen as the formalisation of these in-built or
preprogrammed mechanisms, into what is loosely referred to as a
system of musical communication. From this basically evolutionary
theory comes the idea that certain types of musical response are
inevitable or 'instinctive'.
Instinct, however, has never been a satisfactory explanation for
anything, and the reality is in fact rather more complicated.
Nonetheless, it is a short step from the notion of 'instinct' to the
proposition that listening to Mozart is basically an act of communication
between the listener and the composer, with the listener abstracting
from the music 'things' that were put there because of the inevitability
of the response they would evoke. Some music therapists, for example,
have attributed the calming and other beneficial effects of music on
brain-damaged children to the message put into the music by the
composer, rather than to simpler effects deriving from soft and gentle
sounds, and a caring social interaction with the therapist. Similarly,
music teachers occasionally attribute failure to 'appreciate' Mozart to
deliberate perversity or stupidity, rather than to a rational learning
process leading to the logical decision 'I don't like this stuff'.
In fact, the basic culture-free aspects of music experience are quite
simkple, and have far more to do with the physics of sound than the
intentions of the composer. Thus loud sounds represent a higher
energy, or more aroused state than quiet sounds; high-frequency tones
represent a higher energy state than lower tones; and faster music
represents a faster rate of information-flow than slower music. Naturally
these characteristics interact in ways that are less easy to predict. But
musical appreciation depends more on learning and cultural exposure
than on basic sensory reactions to the properties of sound.
In The Listening we attempted to illustrate these points by using
different musical contexts and a variety of musical types. One of the
most intriguing theories of musical preference (why people like some
tunes and not others) is derived from work on experimental aesthetics,
mainly in the visual arts, but developed into a theory of music
preference by David Hargreaves at the University of Leicester. The
theory has its basis in information theory, but the important insight
comes from the distinction between this conception of 'information' and
its psychological counterpart. Fundamentally, the coding of physical
information contained in a music composition, as in information theory,
predicts very little of interest, but coding the information in 'subjective'
terms predicts quite a lot. Whether a person likes a particular piece or
not depends on the information they are able to take out of it, rather
than information that is already 'in there'.
People tend to prefer music which provides them with sufficient
information to be interesting, but no so much that they become
overloaded. So if someone hears a variety of pieces, ranging from the
very simple to the very complex, then their liking of the pieces will rise
as complexity increases up to a point where their own personal
preferred rate of information-flow is achieved. After this, increasing
complexity will cause a decline in preference. In theory, this produces
an inverted U-shaped curve.
Unfortunately, due to the wide differences in musical experience
between individuals, experimental studies have not always revealed the
perfect inverted U predicted by this psychological theory. So it was
with some trepidation that we attempted to reproduce this effect in
front of the cameras.
Our experiment involved a group of 10 long-suffering choirboys from St
Paul's choir school. We started with the assumption that more complex
music is harder to remember and to sing back, so the chorus master's
scores for accuracy were taken as some measure of how complex each
boy found the music. Eight weeks before, the choristers had judged
their preferences for the same tunes. In a further test the choristers
selected the tune they would most like to have carried out further
studies on, and as expected, we found that the boys chose the more
complex pieces, even though the pieces were not necessarily those
they liked.
Play it again
But there is another way of manipulating subjective complexity, and we
tried that too. Playing the same piece of music over and over again
tends to lead the listener to find it progressively less complex, as it
becomes more faimilar and predictable. So some judgments of aesthetic
preference depend on how complex the material was to start with.
Using an audience of 36 members of a light operetta society we played
a three-minute extract of an opera that none of them had heard before,
eight times during the course of an evening. About one-eighth of their
number were questioned after each hearing, and when their average
scores out of 10 'liking' were plotted against the number of times they
have heard the extract, the graph resembled something like the
predicted curve. There was something odd about the results the second
and third times that the music was played which we cannot explain, but
otherwise these results resembled the predicted inverted U curve.
We carried out a similar test in front of the cameras for a very different
kind of music: a disco audience of 58 people aged between 10 and 30,
who were asked to dance eight times during the course of a day to a
piece of heavy metal music, again that none of them had heard before.
Their average 'liking' scores, when plotted against the number of times
they had heard the track, revealed a clear inverted-U shape, although
nothing like the height of a chart-topping single.
Perhaps the most difficult aspect we tackled was the abstraction of
tonal centres by means of 'internal templates'. Several studies have
convincingly revealed that the identity of the tones in a piece of music
and an appreciation of the size of the intervals between notes becomes
possible only after the listener has abstracted the 'tonal centre' of the
music. Once this is achieved, the listened knows where 'doh' is, and
hence the identity of other tones in the scale. At this point, the music
starts to sound 'musical' and is perceived in a way which is
psychologically different from the perception of sets of random tones.
To achieve this, however, the listener must possess some sort of
internal template based on the modes prevalent within a particular
musical culture.
In Western music the major scale forms the basis for such a
psychological template, with its characteristic spacing of the notes of
this scale system. In previous studies, psychologists have inferred the
properties of such a template from painstaking experiments involving
recognition and reaction time. Our problem was that while this is a
cornerstone of modern theories of music cognition, films about people
waiting for things to happen and then pressing a button do not add up
to compelling viewing. So we attempted something quite new.
We called once again on the services of our St Paul's choristers, who
were asked to listen to a tune being played and sing the signature note
or the whole scale that it had been played in, and then sing the tune
back from memory. The tests did show that these templates are
specific to a culture: the choristers could easily identify the signature
note of the scales of tunes played in the Western scale system and sing
back the same tunes from memory. But when a Javanese gamelan
player was substituted for their piano-playing chorus master, their
ability to sing the intervals of one of the Javanese pentatonic scales
was matched by their inability to sing the simple tunes being played to
them. And when the gamelan player was replaced by a clarinetist
playing the 12-tone music of Luciano Berio, the experiment had to be
abandoned amid chaotic laughter.
The importance of a learned cultural template that is derived from the
modes available within a culture is thus nicely illustrated. The reason
why non-European music may seem untuneful and unsatisfactory (or
'less well-developed') to Western ears thus derives from the fact that
the listeners have not brought with them the learnt templates that
make musical listening possible within that culture, rather than because
of some inadequacy in the music.
The programme also explores the ways in which the expressive aspects
of music are heightened by visual cues and social situtions, rather than
being a sole function of auditory input. For the most part these
illustrations work well enough, as well as adding colour and a sense of
context to the overall programme, with one exception. A
well-documented effect shows that people think more highly of a piece
of music if they believe it to be by a respected composer, and even
musicians have been shown to play with fewer errors when they are
confronted with a new piece by a respected composer rather than by
someone new or unkown.
This effect was one of the aspects of the programme that we thought
we could rely on. We chose a pice of music and attributed it alternately
to Wagner and to the little-known contemporary composer Gersdorf
(acutally a small town in Germany), and produced some wonderfully
tongue-in-cheek programme notes to support these claims. Yet the
effect simply failed to materialise during the tests involving our light
operetta society audience, for reasons that remain unclear. But in a
series of questions that required the listener to award marks for both
'liking' and 'quality' for the opera extract they had heard, there was
a
small trend towards awarding lower marks for 'quality' than for 'liking'
among those who thought it was written by Gersdorf, and a significant
trend towards awarding higher marks for 'quality' than for 'liking' among
those who thought it was by Wagner. The extract was actually from
The Queen of Sheba by the Austro-Hungarian composer Goldmark.
Despite these aberrations, however, the main intention of the
programme - to suggest that musical experience has more to do with
the knowledge and culture of the listener than the raw materials of the
notes, sharps, flats, modulations and bar lines from which the music is
constructed - remains more or less intact. There are no musical
absolutes, even though a limited number of physiognomic aspects of
sound probably cross some cultural boundaries. The stuff of music lies
primarily in the people who listen to it, rather than in inherent properties
of the noises made by orchestras or bands. Consequently, if we believe
that we can communicate the highest levels of human creativity to the
rest of the Universe by beaming Beethoven into the Galaxy and beyond,
we are probably labouring in vain - and our gamelan player would have
no greater success either. After all, the greatness of Mahler generally
fails to survive a trip to Brixton, and the greatness of Jimi Hendrix fails
equally to survive a trip in the opposite direction.
John Davies is professor of psychology at the Centre for Occupational
and Health Psychology at the University of Strathclyde. He was the
consultant to The Listening, which will soon be broadcast on Channel 4
television.
---------------------------------------------------------------------
Oceans Scherzo Traditional Picture by Anthony Tarantelle funderal
Phillips by Wieniawsky gamelan
---------------------------------------------------------------------
Sailing 8 6 0 Funeral 1 0 1 Industry 6 0 5 Beach 1 3 6 Guardsmen 0 0 0
Spiky object 2 9 4
---------------------------------------------------------------------
JOHN DAVIES
From New Scientist magazine, vol 129 issue 1752, 19/01/1991, page
© Copyright New Scientist, RBI Limited 2001