[alife] Evolutionary Music

[Simon Kirby]

------


ANNOUNCEMENT:

The issue Vol. 22 No. 3 of Contemporary Music Review (Routledge) is
dedicated to the emerging topic of evolutionary music:

"Evolutionary Music: At the Crossroads of Evolutionary Computing and
Musicology." (Edited by Eduardo R Miranda)

It features papers by Phil Husbands, Simon Kirby, Peter Todd, Rodney
Berry, Ian Cross, and other pioneers in the field. List of contents &
abstracts, plus ordering information can be found here:

http://taylorandfrancis.metapress.com/link.asp?id=0afmrvf0kuj2

Excerpt of the preface:
-----------------------
Composers, perhaps more than any other class of artists, have always
been acutely aware of the scientific developments of their time. From
the discovery almost three thousand years ago of the direct relationship
between the pitch of a note and the length of a string or pipe, to the
latest computer models of human musical cognition and intelligence,
composers have always looked to science to provide new and challenging
ways to study and compose music.  Today, with the great scientific
advances being made in the field of evolutionary biology, from new
insights into the origins of humans and other species to complete
mappings of the genes that control our growth and development, it is
time to build a new approach to the study of music: evolutionary music.

For instance, the evolutionary approach to music composition is built
upon the three basic components of biological evolution: reproduction,
variation, and selection. Reproduction with variation has been a
hallmark of music production probably since its beginnings: people would
learn the songs of others, and perhaps make slight changes (accidentally
or intentionally) that could be incorporated into further performances.
In fact, technology has helped to take some of the variation out of this
reproductive process: printed scores allowed at least the underlying
musical intention to remain constant for some period of time, even if
the performances varied (like the distinction between an individual's
genetic instructions and the actual person who is built), and
reproducing pianos in the early 1900's enabled the exact replication of
the performance itself from the information stored in punched paper
rolls.  The separation of musical information (akin to genes) from
musical performance (akin to phenotypes, or bodies) and the
specification of the process to convert the former into the latter (akin
to
growth) also allowed tinkering with the information (and even the
conversion
processes) to produce new performances, reintroducing new possibilities
for variation into musical reproduction.

With the advent of faster information-processing technologies, faster
forms of musical reproduction and conversion from information into
performance have allowed these processes of reproduction and variation
to be applied to many musical ideas in parallel.  That is, computers
have enabled the processing of whole ³populations² of musical pieces,
rather than just the slow repetitive change of a single idea over time.
And once populations are possible, the third component of evolution can
be introduced: selection of some individuals in the population to
reproduce more than others.  So for instance, a computer system for the
evolution of musical themes could contain several themes at once, and
those that were somehow better  (as judged by a listener or an automatic
evaluation program, for example) could be used to form the basis of the
population of new themes making up the next generation, again through
the process of reproduction with variation. Exactly this idea has been
introduced in the visual realm by Richard Dawkins's Blind Watchmaker
program (in his book of the same name), and now musicians are beginning
to use it in the audio domain as well.

Evolutionary musical systems are an exciting new development for
composers and musicologists alike.  For composers, they provide an
innovative and natural means for generating musical ideas from a
specifiable set of primitive components and processes-reflecting the
compositional process of generating a variety of ideas by brainstorming
followed by selecting the most promising ones for further iterated
refinement.  For musicologists, evolutionary systems can model the
cultural transmission and change of a population's body of musical ideas
over time-for example, the development and maintenance of musical styles
within particular cultural contexts and their reorganization and
adaptation in response to cultural exchange.  In both cases, the musical
evolution can be influenced by a variety of constraints and tendencies
built into the system, such as realistic psychological factors that
influence the way that music is experienced, learned, stored, modified,
and passed on between individuals.

This issue of Contemporary Music Review features articles introducing
not only the applications of evolutionary music, but also the tools
needed to create and study evolutionary music systems. These tools are
drawn in part from research into the origins and evolution of biological
organisms, ecology, and cultural systems on the one hand, and in part
from computer simulation methodologies on the other. Contemporary
scientists seem to be increasingly shifting their attention from the
study of the composition of isolated matter to the study of the
functional characteristics of nature's tangled systems, including the
interactions between components of a system, the interconnection of
different systems, and the emergence of global patterns from local
behaviour. Consequently, new scientific methodologies and tools are
being devised for these new studies, most of which would not be possible
without computer technology: genetic algorithms to model evolutionary
processes applied to populations, cellular automata to study emerging
high-level patterns from simple local interactions, neural networks to
explore complex information processing via large numbers of simple
neurones, and adaptive games to understand the strategic interactions
between agents with possibly different goals. These tools can also be
combined, often under the rubric of artificial life, to create surrogate
artificial worlds populated by interacting simulated organisms in which
complex experiments can be performed that would otherwise be impossible.
These systems, as we show in this special issue, can also be used to
create and to study musical compositions and cultures in ways that have
never before been achieved.

--
Simon Kirby 
Language Evolution and Computation Research Unit
School of Philosophy, Psychology & Language Sciences 
University or Edinburgh
www.ling.ed.ac.uk/~simon