Content uploaded by Eric Werner
Author content
All content in this area was uploaded by Eric Werner on Dec 13, 2018
Content may be subject to copyright.
Music and Distributed Programs*
†
Eric Werner ‡
Oxford Advanced Research Foundation
eric.werner@oarf.org
https://www.ericwerner.com
Abstract
A theory of music is presented that views musical scores as distributed programs.
The motivation is to contradict a view held by some in biology that the genome
is like a keyboard that the cell plays. Instead, it is argued the genome is more
like a musical score that is read and interpreted by multiple players, namely, the
cells. Rather than the cell playing the genome, each cell interprets and executes
its distributed copy of the score that is in the genome.
Key Words: Distributed algorithms, distributed programs, musical score, partitur, developmen-
tal biology, genome semantics
1 Historical conceptual introduction 2018
In the many years while I was teaching and doing research at the University of Oxford
(simultaneously at DPAG and the Computer Science Dept.) and especially in Balliol
College, I had frequent debates on the conceptual foundations of biology with Denis
Noble and with Evelyn Fox Keller while she visited us at Balliol from MIT. At the
time of its writing and delivery in February 2008 to a Philosophy of Music Seminar at
the University of Oxford, the motivation for this short essay (below) was to contradict
*Delivered 27 February 2008 to the Philosophy of Music Seminar, University of Oxford, held by
Christoph Denoth and Denis Noble.
†Please Cite As: Werner, E., Music and Distributed Programs, Preprint, 2018 DOI: *Insert DOI
here*
‡©Werner 2008-2018. All rights reserved.
1
Eric Werner: Music and Distributed Programs 2
the view, shared by Denis Noble, Evelyn Fox Keller and others that the genome can
be viewed as a organ keyboard where its genes are the keys that are "played" by the
cell[4,3][1,2][5,6]. On my view, this metaphor of the genome as just a keyboard is
a throwback to preformationist ideas of the egg as homunculus. It confuses the cell
with the score. It misses the essential point that the same cell with different genomes
will produce different morphologies. It also places too heavy a burden on the cell has
container of the complexity of information required to generate the resulting space-
time event of embryonic development.
On my view, each cell in the developing embryo is more like the conductor together
with the performers who are jointly reading, interpreting and executing the musical
score[6,7]. The score is the text in the genome. A text that is much more than just
the genes. For me they are developmental control networks or Cenes (for control
genes)[8,9,7]. For some who hold the view that there are programs in the genome
this essay may seem obvious. For others my view will seem obviously wrong. Such is
the philosophy of biology when mixed with the philosophy of music. Deeply hidden
are actual empirical hypotheses lurking and not just analytical or logical confusions.
More on this in a future essay. My original 2008 essay follows1:
2 Directives versus declaratives
2.1 Directives
Directives tell an agent what to do, e.g., “Open the door!” Directives are interpreted by
actions. There are two basic types of directives. Basic directives that call for a action
or action strategy by the agent and meta-directives that tell the agent which of a set
of directives to do next. Thus, if the directives are indexed by some number, do the
command with number X next, tells the agent to find the command with number X and
then execute it.
2.2 Declaratives
Declaratives talk about the state of the world, the way it is, its properties and relations,
e.g., “The door is open”. Declaratives are true or false. Directives are not true or false,
but rather are intended to direct the receiver of the directive to do the action described
by the directive.
1The citations and references were not in the original text
Eric Werner: Music and Distributed Programs 3
3 Musical Scores as Encoded Strategies
Each note in a reduced score for one player may be viewed as a directive to the player
at an instant in an ordered time series. The entire reduced score for a particular player
in a full score is then an encoding of a musical strategy of action for that player. Event
coordination of directives in parallel is achieved by notation that places each player’s
score parallel and in synch with the other players. Coordination of directives in se-
quence can be achieved by the use of communication signals (e.g., a score sequence
or visual cues from the conductor). Coordination by signalling cannot be achieved by
the score itself, because the player needs to know where in the full score the play (the
performance of the full score) is at.
4 The Full Score as a Description
Once the event or play or performance has occurred and it has followed the full score
accurately then the full score can be viewed as a declarative history of the event.
Hence, in this sense a musical score can serve both as a description (sequence of
declaratives) and as a complex directive.
5 Musical Scores are like Programs
Like programs, musical scores are passive in that they need an agent that interprets and
executes them for the actions they denote to be achieved. Like programs musical score
can be executed in the imagination, in other words, modeled and simulated without
actually being performed by players. Like programs, musical scores may contain meta-
directives that refer not to an action to be performed on the instrument of the player,
but rather a directive to what to do next. In musical scores as in programs what to do
next is often just the next directive (note or command) in the score sequence. But, there
may be jumps that tell the player to go to another section of the score (or program).
These are meta-directives.
6 Performance versus experience
We must distinguish between the experience of a performance from the directives that
control the execution of the performance. In musical performances, as we saw above
Eric Werner: Music and Distributed Programs 4
in the case of coordination, the players need the experience of the partial history of
the performance to cooperate and execute the next part of the score. So too programs
need to record their previous results (memory) as well as signals to coordinate their in
parallel and or distributed actions.
7 The music of life
We may view the genome as being analogous to a musical score that directs the multi-
tude of cells. Each cell follows its own score, or that part of the score that is currently
active. There is no central conductor but each cell coordinates its activity through com-
munication with other cells. This communication may affect which part of the score or
genome is played when and where. Things are more complicated because players join
the group as the score is read. More and more players join as the players divide giving
themselves and their offspring instructions as to where to start reading their copy of
the score.
References
[1] E. Fox Keller. The Century of the Gene. Harvard University Press, 2000.
[2] E. F. Keller, L. S. o. E. Science, and Political. Is There an Organism in This Text? London School
of Economics, Centre for the Philosophy of the Natural and Social Sciences, 1995.
[3] D. Noble. The Music of Life. Oxford University Press, 2006.
[4] D. Noble. Genes and causation. Philos Transact A Math Phys Eng Sci, 366(1878):3001–15, 2008.
[5] E. Werner. Genome semantics, in silico multicellular systems and the central dogma. FEBS Letters,
579(7):1779–1782, 2005.
[6] E. Werner. How central is the genome? Science, 317(5839):753–754, 2007.
[7] E. Werner. On programs and genomes. arXiv:1110.5265v1 [q-bio.OT],
http://arxiv.org/abs/1110.5265, 2011a.
[8] E. Werner. Cancer networks: A general theoretical and computational framework for understanding
cancer. arXiv:1110.5865v1 [q-bio.MN], http://arxiv.org/abs/1110.5865, 2011b.
[9] E. Werner. Combinatorial limits of transcription factors and gene regulatory networks in develop-
ment and evolution. arXiv:1508.03531 [q-bio.MN], http://arxiv.org/abs/1508.03531, 2015.
