A trans-human source for music?
===============================

Brian D. Josephson
Cavendish Laboratory, Madingley Road, Cambridge CB3 0HE, U.K.


Abstract. Following a discussion of what symbolisms are and the
factors that give rise to their development, it is argued that the
complex phenomenon of music can best be understood as a category of
symbolism. Human culture and evolution by natural selection do not
give suitable conditions for a symbolism of the subtlety of music to
arise, and it is proposed therefore that some agency of a trans-human
nature, some of whose concerns we are able to share, is the original
source of the phenomenon. These considerations lead among other things
to the proposal that biological signals in general may carry more
subtle information than is commonly assumed. 


1. Introduction

Previous papers [1 - 4] have criticised certain conventional views of
music, such as the idea that music is intrinsic to a musical culture
and therefore has no intrinsic meaning of its own, and argued instead
that music constitutes a general symbol system, analogous in many ways
to DNA.  The previous work relied heavily on the hypothesis of there
being close analogies between DNA and music (such as their both being
informational systems of high specificity), but in view of the
existence of significant differences in detail between the
characteristics of the two kinds of symbol system, especially in that
music (like natural language) is processed by a mind that in some
sense understands it, whereas DNA expresses its potential by a more
passive process, a slightly different mode of attack is appropriate.
In the current work the emphasis is put on music as a symbolism,
rather than as before on it being a system closely analogous to DNA. 

2. The nature and origins of symbolisms

The essence of a symbolism, as we shall envisage it in the following,
is that it involves agents making use of symbolic expressions, to
define processes of greater or lesser complexity, which in turn
generate significant outcomes (significance being defined by the
agents, in accord with some prescription appropriate in the context of
their mode of existence). In formal terms, a symbol system can be
regarded as being defined by its associated collection of permitted
symbolic expressions and the expressive processes that generate
responses to these symbolic expressions.  The expressive processes
give the symbolism its expressive power, whilst the ability to make
use of this expressive power requires the existence also of generative
processes which generate symbolic expressions appropriate to the
circumstances. 

In terms of the above description, the evolution of a symbol system
can be viewed as the gradual incorporation of more and more powerful
expressive processes into the system, in essence through the linking
of symbols and concepts, a symbol being the means of providing
controlled access to a useful concept, a concept in turn being an
aspect of a useful process. An idea that is effective in one situation
can, through the mechanism of representing it by a symbol or symbols,
be tried out in other circumstances simply by calling up from memory
the relevant symbol or symbols. Through the process of trial and
error, the concept becomes shaped to have maximal utility. 

The refinement of a symbol system is seen to be the natural outcome of
a succession of acts of creative change on the part of the agents that
make use of it, in conjunction with a process of the collection of the
most successful new variants. A symbol system may in certain instances
be the possession of a single agent (as exemplified by the situation
of a person who has invented a new idea with its own terminology, who
has not yet shared the idea with other persons), but in the general
case aspects of the symbol system become the properties of a cultural
group, as a result of members of the group being able to share uses of
the symbolic expressions involved and to exchange symbols with one
another. 

3. The case for music being a symbolism

A process such as has been described accounts, in its essentials, for
the evolution of natural language. The current proposal is that
similar processes gave rise to the phenomenon of music.  For such an
idea to begin to be worthy of consideration, it would be necessary for
it to be able in principle to account for features of music that would
be hard to understand according to any other point of view. The most
definitive support for the idea that good music signifies something
(originally proposed by Suzanne Langer [5]) comes from the area of
subjective experience, starting with the fact that listeners to music
appear to be able to experience a clear distinction between a type of
music that seems to consist simply of recognisable patterns, and music
of aesthetic value that appears to be in some way 'significant'. With
the 'significant' kind of music there is the feeling that the surface
form of the notes fits the context, as is not the case with music that
consists merely of recognisable patterns. There is also the feeling
that the music involves 'themes' that are 'developed' as a piece of
music progresses. Again, there is the experience that both composers
and performers have that what they are doing is attempting the
accurate expression of an idea in the medium of music. These
perceptions parallel corresponding experiences with language, which
are often described in almost identical terminology. In the case of
language, these features are reflections of the fact that language is
a symbolism, and not only a symbolism but also a symbolism that its
users know to be a symbolism in the sense that they have some
awareness of what that symbolism signifies. While in the above I have
referred only to subjective experiences paralleling those of language,
analysis of the structure of music indicates the existence of
objective parallels also, discussion of which topic lies outside the
scope of this paper. 

At the very least, we can conclude (i) that there is a phenomenon of
'seeming significant' in music;  (ii) this phenomenon is tied to very
individual structures (individual compositions), implying that
appreciable interference with the details of a composition is liable
to modify quite strongly the impression of significance; and (iii)
seeming significant is not the same as appearing recognisable or
familiar. It would seem that significance or aesthetic appeal cannot
readily be tied down to rules, and neither would it appear explicable
on the basis of the kinds of generalisations that neural nets can
make. Both rule-governed structures, and structures formed on the
basis of generalisation capacities of neural nets, appear to pertain
far more to the question of 'what kinds of sound sound like music'
than they do to aesthetic matters. Correspondingly in the case of
language, rules and generalisation procedures appear to relate to
grammatical categorisations rather than to semantic ones. 

Assuming that any objections to such a proposal can be overcome, the
most reasonable way of accounting for all the parallels between music
and natural language would appear to hypothesise simply that music is
a symbolism in the sense defined at the beginning of this paper, that
is that it has as its basis agents making use of symbolic expressions
to define processes of greater or lesser complexity, which in turn
generate significant outcomes, with the whole self-organising system
having a system of values that leads step by step to the evolution of
the symbolism in the direction of increasing expressive power,
particularly in regard to the possibility of using it to define ideas
that accord with the value systems of the agents concerned. Such
processes can, as exemplified by the phenomenon of language, lead to
the development of similar complexity to music, as well as to the
associated development of a sense of significance or otherwise in
relation to symbolic expressions. 

4. Trans-human agencies as the sources of musical symbolism

While the above hypothesis fits the phenomenon of music in many ways,
various questions arise.  In particular, who or what were the agents
concerned, and in what context did this hypothesised evolution of
musical symbolism take place? What were the value judgements that
required the processes concerned with the existence of the symbolism
to become more and more complex in order to express what was needed in
order to conform to the requirements of that value system? 

To phrase this last question more explicitly, the theory being
proposed is that musical symbolism, like natural language, became more
and more sophisticated because more and more sophistication had
well-defined uses. The question is, what might these uses be? One kind
of use would be cultural uses such as use in ceremonies or in song. I
shall ignore this possibility as the subtle problems arise largely in
connection with more exclusively abstract manifestations of music such
as classical music in a purely instrumental context. The claim I shall
make here is that because such music may involve radically new
'inventions' there are no grounds for claiming that the meanings
involved are ones that have been learnt in the usual way. This leads
one to the radical proposal that musical symbolism is in some sense a
pre-existing system, which human beings discover rather than invent. 

The possibility of explanations for apparent meaningfulness on genetic
grounds should also be mentioned; however, such explanations suffer from
the objection that it is difficult to imagine selection pressures that
would lead to the degree of uniformity that does exist regarding aesthetic
preferences among those that have given sufficient attention to the kinds
of music concerned to be able to appreciate them fully. 

If human beings discover music rather than invent it, then questions
arise such as where does the music come from, and how do we acquire
it? To the first question I shall simply give the answer 'from some
source or other. This source must simply be some entity satisfying
the conditions under which we have presumed complicated symbol systems
will develop (i.e. agents making use of symbolic expressions, etc.
etc.). And in answer to the question of how do we acquire it, we can
simply say: by the same means as those by which human beings acquire
language from each other, or to take a more asymmetrical case, pets
from human beings (this process also involving shared meaning, shared
utterances, and processes for acquiring some comprehension of symbol
systems). We are familiar with the way chemicals in the brain have
specific effects on thought, and so it should be not be too difficult
to envisage the possibility of a similar symbiosis between the
mysterious source of music and our own minds enabling us to acquire
some kind of understanding of the symbolism and the uses that the
source makes of music. 

The above picture requires for its completion attention to the
question of what the concerns may be of the agencies or agency
involved with music, these being the concerns that would have driven
its evolution. While music in general has, as already noted, no
specific denotations in the world, it does, subjectively, appear to
relate to general life themes. One might speculate therefore that it
originated as a means of symbolising life processes in general,
perhaps with the goal of indicating archetypal problems and their
resolutions. 

5. Conclusions

The above has indicated an unorthodox, and yet apparently consistent
concept of what music is. In essence, the familiar concept of the
existence of symbolisms is hypothesised to be applicable in an
unusual, not at all clearly specified domain, and to show its presence
only indirectly via the phenomenon of music. Thus a visible regularity
is being accounted for in terms of an invisible one, the justification
being that the very general mechanism of symbolic evolution leads to
phenomena of exactly the kind that are observed. The situation is
frustrating in that we can see very few of the details. However, this
is a situation that is commonplace in science. The question is what
tools may exist to probe the situation in more detail. 

One avenue of research may be through the study of music, which may
provide information about the symbol system and some aspects of its
semantics. This would require the development of some kind of
understanding of the 'thought processes' implicit in music, and may
involve a more perceptive analysis of archetypal aspects of life, that
is those aspects which are largely independent of the details of a
given organism, than there is in the theories that we have today.
Another avenue might involve analyses of possible subtleties of
symbolisms in the biological realm generally.  Signals in biology are
normally assumed to encode information on the basis of only simple
features such as amplitude and frequency, but such an assumption may
be incorrect. Our difficulties with conceptualising what music is may
be a pointer to the existence of a more general limitation of this
kind in our understanding of the biological realm. 


Acknowledgements

Many of the ideas contained in the above paper are ones that emerged
during discussions with a collaborator, Tethys Carpenter, concerning
the nature of music and the ideas described here could not have been
formulated without the benefit of her numerous contributions. I am
also indebted to Dr. D. G. Blair and Prof. N. A. Baas for discussions
related to linguistic development, and cognitive development in
general, respectively. 

References

[1] B. D. Josephson and T. L. Carpenter, New Scientist 129(1762),
2(1991). 
[2] B. Josephson and T. Carpenter, New Scientist 131(1780), 51Ð2
(1991). 
[3] B. D. Josephson and T. Carpenter, Music and Mind - a
Theory of Aesthetic Dynamics, in 'On Self-Organisation', Springer
series in Synergetics Vol. 61, pp. 280-7, Springer, Heidelberg 1994.
[4] B. D. Josephson and T. Carpenter, What can Music tell us about the
Mind? A Platonic Model, in The Physical Basis of Consciousness, MIT
Press (to be published).
[5] S. K. Langer, Philosophy in a New Key, Harvard University Press,
Cambridge, Mass., (1957). 

END