Attributing Creative Agency: Are we doing it right?
Oliver Bown
Design Lab
University of Sydney
NSW, 2006, Australia
oliver.bown@sydney.edu.au
Abstract
When contemplating the creativity of computational
systems, a host of factors have been taken into consideration,
many of which people have attempted to measure
or otherwise operationalise: novelty, value, P-creativity
versus H-creativity, exploration versus transformation,
the subjective evaluation and contextualisation of the
artefact, and so on. Whilst of equal importance, the
systematic and rigorous attribution of creative agency
to different actors in the production of a specific output
has been given less attention. It is common to make the
simplifying assumption that the most direct contributor
to an artefact is that artefact’s sole author, but arguably
this is never the case: all human creativity occurs in the
context of networks of mutual influence, including a cumulative
pool of knowledge.
This paper looks at how we might better formalise creative
authorship such that for any artefact, a set of agents
could be precisely attributed with their relative contributions
to the existence of that entity. It asks only what the
nature of this formalisation should be, and concludes
that a more critical approach is needed to the creative
agency of human actors, and thus the expected creative
agency of machines.
I draw on two critical notions that can inform a methodology
for the ascription of creative origins in computational
creativity: becoming, and the agency of networks
of interaction.
I look at a example from both historical human creativity
and computation creativity, to consider how we can
break down creative agency and ascribe it to different
sources. Practical implications are dicsussed.
Introduction
In contemplating creativity, we are comfortable with taking
at face value statements such as “Ludwig van Beethoven
composed Beethoven’s Fifth Symphony” or “Leonardo da
Vinci painted the Mona Lisa”. At the same time, we are
well aware that such attributions are rough at the edges when
scrutinised. Creativity does not occur in a vacuum. All creators
are subject to influence from their culture or environment,
and other forces at play in the creative process include
chance, the influencing of opinions such as value attribution,
the emergence of outcomes through collective action,
and the need to consider the potentially active role played
by passive objects, as discussed most famously by Latour
(1996) and Clark (2003), but with recently renewed interest
by Malafouris (2007), Miller (2010) and Ingold (2007).
Longstanding theories of creativity have successfully
managed these apparently conflicting perspectives, most notably
the work of Simonton (2003) and Csikszentmihalyi
(1999). In both cases, creativity is properly understood as a
process that operates at a macro level (sometimes described
as a network or systems level). For Csikszentmihalyi the
macro perspective is critical because the process of creativity
involves the interaction between heterogenous groups of
participants, and for Simonton it is because creativity is best
modelled as a stochastic process across a population, which
cannot be properly understood when looking at single instances.
However, it has been difficult to translate such knowledge
into practical methods for evaluation in computational creativity,
which despite its strong acknowledgement of such
theories does not successfully draw on this macro-level perspective
in evaluating individual systems. In this paper I
present this challenge in terms of recurring misconception
that evaluation can be performed on isolated individuals, i.e.,
at a micro-level, which I refer to as the “islands of creativity”
view. Drawing on literature from creativity research, philosophy
and the social sciences, I consider how a macro-level
view of creativity can work in the applied task evaluating
computationally creative systems.
I suggest that a critical step is to recognise how the objects
of evaluation are dynamic, in flux, and have boundaries that
shift at different stages in their history, as they interact with
other people and things.
I propose a “dynamic analysis” of any system, which details
(i) the fluid and temporary boundaries between entities,
and when these aggregations act as agents, (ii) when
and where influence occurs, (iii) what constitutes an output.
Such an analysis, it is proposed, could help us better attribute
creative agency in the evaluation of computationally creative
systems, by clarifying how novelty and value are determined
(by whom) and what influences feed into the creative system
at different times.
Simonton’s macro-creativity model
Simonton, for example (Simonton, 2003), showed through
quantitative analysis of scientific achievements that the
Proceedings of the Sixth International Conference on Computational Creativity June 2015 17
arrival of creative breakthroughs was sufficiently unpredictable
as to be effectively random. This does not mean to
say that a member of the population chosen at random might
make an advance in quantum physics. Naturally, strategies
for creative success involve becoming expert in a field, focusing
on problems, working hard, knowing what to look
for, and so on. Indeed, Simonton showed that success was
proportional to activity: the more active you were in a field
the more likely you were to produce creative outcomes, but
equally the more likely you were to produce uncreative ones.
What remained stable was the rate of success, measured as
the ratio between successful output and total output.
From this perspective, in computational creativity what
we might describe as strictly micro level focus – privileging
the creative agency of individual creators without considering
how these agents interact with each other and with other
elements in the world – is a detrimental but seductive simplification,
which is often assumed to be reasonable where
in fact it is problematic. Simonton’s micro-level view of creativity
tallies with his macro-level view: at the micro-level
an individual iterates through many trial-and-error attempts
at a solution, understood in creativity research through cognitive
processes such as incubation. This trial and error is
the best that can be done in an unknown search space; there
aren’t reliable analytical or inductive approaches available
to the kinds of problems that we would define as creative,
because the problem spaces are unknown – at least, in the
case of Boden’s ‘transformational’ creativity (Boden, 1990).
Thus we may imagine a population of individuals searching
for solutions to the same problem, working at the same
rate. When one individual discovers the solution, in Simonton’s
view, we should not leap to the conclusion that there
is anything fundamentally different about the creative process
used by that individual. Simonton also draws on evidence
from ‘simultaneous scientific discoveries’ to support
this view, arguing that the common occurrence of such discoveries
is due to the fact that it is the discovery context,
and not the creative ability of the discoverer, that is key to
the arrival of the discovery.
Such work is widely acknowledge in computational
creativity research, but this macro-level thinking remains
largely absent in the methods that we apply to the evaluation
of computationally creative systems.
The “islands of creativity” problem
Such approaches have been successfully applied in the context
of studies of traditional human (i.e., not computational)
creativity. But in computational creativity, although we frequently
pay homage to these macro theories, we have yet to
find a way to incorporate them into a working methodology
in the complex area of evaluation. I suggest that a significant
obstacle to computational creativity evaluation lies in
the idea of “islands of creativity”, the idea that creativity is
situated in specific systems (mostly humans, now also computers),
without any fluidity between these systems and the
rest of the world:
Definition: The “islands of creativity” problem in creativity
is the misuse of the simplifying view that individual
human actors (or individual computer actors) are sole
originators of specific creative artefacts. It conflicts with
the more holistic view that stochastic and network macro
processes involving interactions between heterogeneous elements
underlie the big picture of creative production.
Is this view actually a misconception, and what have been
the implications of holding it? Would our approach to evaluation
benefit from avoiding it, and shifting towards thinking
about creation occurring through the relationships between
entities? I will argue that looking at creativity only by reference
to the human cognitive capacity for creativity continues
to be problematic for computational creativity, not least
because the kinds of computational systems that will do creative
things in the near future may not do them in particularly
human-like ways. Rejecting the “islands of creativity”
problem is a necessary part of stepping away from a humancentric
frame.
Specifically, the embrace of an alternative, macro-level
theoretical framework may enable two important contributions
to computational creativity: (i) in the way we understand
what we mean by human cultural activities such as art
and music. There is a tendency to trivialise such questions
in pursuit of simple computable targets, whereas these areas
of activity are some of the most ethnographically rich
that humans exhibit, so as to be far from easily reducible;
and (ii) by providing practical methods to help us attribute
creative agency properly when asking questions of the form
“did system x do something creative?”
Defining creative production in terms of
interaction
In the words of Heraclitus, via Nietzsche, “the whole flows
as a river”, the river’s evident dynamism, by which it is
constantly in a state of re-creation in the movement of water,
is an apt way to understand those less obviously fluid
things in our environment: “being is an empty fiction” (Nietzsche,
1998). We tend to take the consistency of objects
at face value, but for practical, not only philosophical, reasons
it can be preferable to view things not as entities that
have the property of being; instead their existence is in constant
re-creation, captured through the notion of ‘becoming’.
Viewing things without this frame of dynamism, as neat
bounded entities, may be a practical way of simplifying and
understanding the world in the everyday, but risks missing
the myriad ways in which entities transform, influence each
other, have porous boundaries and fuse and fissure. Such
thinking has been applied successfully in the social sciences,
and may be helpful in thinking about evaluation in computational
creativity, particularly in how we frame the notion of
a creative agent.
Creative agents
Theorists have embraced the idea of fluidity in the context of
social systems, which are more evidently fluid, using a network
interaction approach, most famously the actor network
approach of Latour (1996) and Law (1992) and the extended
mind theory of Clark (2003). More recently, Malafouris
(2007) makes a terse argument for the abandonment of the
human as a privileged category of agency. For Malafouris,
Proceedings of the Sixth International Conference on Computational Creativity June 2015 18
much as for Clark, if a blind man can be said to ‘see’ with
his stick, then the physical matter of the stick is exactly to
the blind man what the optic nerve is to the sighted. For as
long as the blind man is using the stick, we can designate
a transient entity of the form blind-man+stick which is in
some sense capable of sight. Importantly, the stick is part of
that unit, not apart form it. The man does not see with the
stick; the man+stick sees.
Similarly, he argues, as a potter shapes clay on a wheel,
one cannot successfully draw neat lines of causality that
show the potter’s hands influencing the clay, and not vice
versa. The potter is responsive to the clay, and in her adaptivity,
allows causality to flow back in the opposite direction
from clay to action. The right way to understand the resulting
creation of a pot, Malafouris posits, must not presume
potter as agent and wheel and clay as other, but to conceive
of a unity in interaction between them.
In his words:
“If human agency is then material agency is, there is
no way that human and material agency can be disentangled.
Or else, while agency and intentionality
may not be properties of things, they are not properties
of humans either: they are the properties of
material engagement, that is, of the grey zone where
brain, body and culture conflate.” (original emphasis).
(Malafouris, 2007, p. 22).
The purpose of this thought experiment is to preempt and
thus interrogate the implied objection: “surely we can see
that the potter is the active, intelligent agent in this interaction,
whilst the wheel and clay are passive non-agents,
there to be operated or shaped”. This presents a problem:
although it seems mistaken to start to talk of the agency or
intentionality of clay and mechanical wheels, how else can
we handle the fact that the resulting pot owes its form to clay
and wheels, and not merely to a single human actor?
I understand Malafouris as saying here, as with the blindman
and his stick, that it would be more correct to say that
the temporary interaction of potter+wheel+clay is responsible
for the creation of the pot, than to say that the potter
created the pot using the wheel and the clay. Although apparently
a trivial distinction, the question of agency has been
shifted in a way that significantly transforms discussions of
creative authorship in computational creativity, and equally
resolves the “islands of creativity” problem. This is a more
palatable option than talking about the agency of inanimate
objects, and is particularly apt in the context of machines,
for which the perception of agency might slide easily up
and down a scale. It also takes care of collaborative action
between individuals, whether in a clearly bounded working
unit such as a band, or a fluid genre movement.
Turning to computational creativity, we see that attention
to this detail concerning the existence of bounded agents is
generally overlooked. In major mathematical and logical
formulations such as those of Ritchie (2007) and Wiggins
(2006), understandably, this would be a complex step. Here
the focus is more on artefacts anyway. In other work where
the focus is on the individual and the process of production,
there is still little in terms of acknowledging the fluid boundaries
between components of a creative system.
Dividing individuals
Further to this, thinking from philosophy of mind, AI, evolutionary
psychology, anthropology, and other disciplines, has
in different ways converged on a notion that human agents,
equally, should not be viewed as unitary in action, but consist
of networks of interaction themselves. This thinking
can be found in Minsky’s society of mind (Minsky, 1988),
Baars’ global workspace theory (Baars, 2005), Barkow, Cosmides
and Tooby’s (Barkow, Cosmides, and Tooby, 1992)
multi-domain model of the evolved mind, and many psychological
accounts that reveal conflicting drives and processes
and dedicated channels of activity. In anthropological theory
we have the notion of the ‘dividual’ (Marriott, 1976). This
concept was initially specific to an ethnographic analysis of
how South Asians viewed personhood, but it may also describe
Western conceptions if we admit them to have more
variability:
“Single actors are not thought in South Asia to be ‘individual’,
that is, indivisible, bounded units, as they
are in much of Western social and psychological theory,
as well as in common sense. Instead, it appears
that persons are generally thought by South Asians to
be ‘dividual’ or divisible. To exist, dividual persons
absorb heterogeneous material influences. They must
also give out from themselves particles of their own
coded substances, essences, residues, or other active in-
fluences that may then reproduce in others something
of the nature of the persons in whom they have originated
. . .What goes on between actors are the same
connected processes of mixing and separation that go
on within actors.”
(Marriott, 1976, p. 111)
Although framed in terms of a distinction between Indian
and Western perspectives, it is fair to say that in all world
views there is some freedom to flip between different conceptions
of personhood and individuality. It is common to
talk about feeling like you are ‘defined’ by your family or
friends or the objects you possess. We are also familiar with
the idea expressed at the end of the quote, that two people
can ‘think together’, for example through brainstorming,
and that this is in some way isomorphic to the same process
happening within an individual.
In our computationally creative systems, this fluidity is
more evident. A piece of software is itself an assemblage
of subsystems and may communicate beyond its nominal
boundaries to form supersystems, including with humans.
We should expect that in some cases it is clear that agency
is more strongly associated with a specific subsystem than
with others, whereas in other cases, agency takes the form
of interaction between subsystems or the system and its environment.
An evolutionary framework
As others have discussed (Dawkins and Krebs, 1978; Boyd
and Richerson, 1985; Aunger, 2000; Shennan, 2002), Darwinian
evolutionary theory provides a good template, recogProceedings
of the Sixth International Conference on Computational Creativity June 2015 19
nising in natural evolution exactly that agency lies in ‘processes
of interaction’ rather than in specific entities (Nietzsche
was also heavily inspired by Darwin). It is interesting
to contemplate the non-human creativity of evolution in contrast
to what we typically think of when considering human
creativity. Given a specific organism and asking, “what created
that organism?” we see very clearly that such an act of
creation can only be understood as a continuous process of
interaction between organisms and their environment, and
amongst individual organisms. We cannot pin our form on
the creativity of our parents, nor even on our entire ancestral
history. This view naturally takes into account the the many
interesting cases of coevolution, runaway sexual selection,
niche construction and, in humans, gene-culture coevolution
which produce things through diverse forms of interaction.
When we talk of function in such systems we are actually
referring to teleofunctions (Sperber, 2007), specifically,
functions that serve their own existence. This is in contrast
to the functions of things we build, which are imposed upon
them and are external to the existence of the thing. But cultural
traits and artefacts can and often do have teleofunctions
too and can come about in ways that are more or less similar
to evolutionary processes occurring at a cultural level.
Sperber (Sperber, 2007) discusses the interesting case of the
perception of suntanning. Furthermore, machines that learn
or evolve can have teleofunctions by virtue of the fact that
their goals can be adaptive, but mostly, today, are built with
regular functions.
Dynamic analysis of fluidity in creative systems
Our earliest efforts at building machines that create have resulted
in superlatively weak creative agents when held up
against human beings, as would be expected. But the contemporary
language of creativity is geared towards the superlative
creativity of humans. It does not do well at describing
the simple forms of computational creativity we are
developing today. For this reason, an “islands of creativity”
view, that works for humans, needs to be replaced by a more
fluid conception of creativity that will work equally well for
computational systems. By comparison, a view of this process
of production based on networks of interaction between
elements (whether brain, body and culture, as Malafouris
suggests (Malafouris, 2007), or some other active ingredients)
makes less of a conceptual meal of that scenario.
Even if these various perspectives may be technically
true, is it any use to try to use them to rethink evaluation
in computational creativity? It would be counterproductive
to take clearly delineated elements and blur them
into a loosely defined muddle of interaction purely for
the sake of being more accurate. A danger with adopting
this perspective is that useable categories disappear to
dust. Evoking a Beethoven-piano-stave-pen-church-kingorchestra-etc.-etc.
network complex to explain the creation
of the Fifth Symphony may not have any practical value and
if so, should not be pursued. But as part of a wider investigation
into how qualitative, situated human science methods
can contribute to the understanding of evaluation in computational
creativity (Bown, 2014, 2012; McCormack et al.,
2014), I believe that it will be necessary to take on the “islands
of creativity” problem by introducing such thinking to
form a method of “dynamic analysis” of creative systems.
As a first step in a dynamic analysis approach, we would
need to look at where we have pre-emptively identified creative
agents. Mostly, these will be either individual people,
or the computational systems we have built. For each presumed
agent, we should investigate what assumptions we
hold about their boundedness, their autonomy (any cases in
which we say the system did something “on its own”) and
the origins or their actions. We can also investigate where
different systems might be seen to unite in co-action or break
down into interacting components, and we can look at how
each system is influenced to change its state or structure over
time. In each case, this will be a temporal process where
different system boundaries are recognised over time. In
the case of many computationally creative systems, the full
analysis of such a process would include the role of the system
developer, observing outcomes and iterating their design
in order to improve it (what Colton, Pease, and Ritchie
(2001) refer to as “fine tuning”). We may also find that the
process is so widely distributed across elements that such
descriptions take on a more statistical nature, as we have
seen in both Simonton’s theories (Simonton, 2003), and in
Darwinian evolutionary thinking. In this case, it should be
fine to attribute some degree of creativity to a macro-level
stochastic process itself.
Through the examples below it is proposed that a simple
but effective way to dynamically analyse creative events is
through simple dot-point timelines that discuss sequences
of events, the influence of systems on each other, and the
potential coupling of systems. This is relatively crude, but
may have the potential to feed ultimately into more formal
frameworks such as that of Wiggins (2006).
Application
Without adopting a strong cultural Darwinism – which is
contrary to what I would argue for, and what Sperber’s article
(Sperber, 2007) emphatically argues against – it follows
from all of the above that every creative act should be framed
in terms of processes of interaction. The issue still remains
of showing that this is practically useful. I consider the following
instances and how such an approach serves to clarify
the creative agency.
The Violin
In a recent article (Nia et al., 2015) evidence was given
to support the theory that the shape of sound holes in violins
emerged through an essentially evolutionary process
whereby apprentices copied their masters’ designs with random
variation, and those designs with louder sounds, due
to the shape of the holes on the body of the violin, were
over time more successful. The winning design, the familiar
f-shape that we know today, maximises the ratio between
the perimeter of the hole and its size, providing greater amplification
of the sound, whilst providing a pleasing visual
appearance. Who designed the violin as we know it today?
If the above account is correct we could answer as we would
Proceedings of the Sixth International Conference on Computational Creativity June 2015 20
with the design of organisms in the biological world, that
there is no one designer, and there are not really any designers
in the sense of psychological creative discovery. The
design came about through a macro-level process. Indeed,
we could go so far as to say that the design of the optimised
sound holes was not due in any way to a human creative
capacity, although, difficulty arises when we ask whether
a given luthier’s new design was actually a conscious improvement,
or a random variation that turned out to be successful.
As with Simonton (2003), we may be mistaken in
attributing creativity to the individual mind instead of to the
broader cultural process.
The creative process, as described by Nia et al. (2015),
might look something like the following if represented as a
dynamic analysis timeline:
1. An existing design is copied and modified in ways that do
not explicitly attempt to optimise sound amplification
2. Given time, the louder designs make more money, and
these workshops grow and reproduce whilst the workshops
responsible for the quieter designs diminish.
Paul Hession / Arne Eigenfeldt Live at Cafe Oto
At a recent concert of live algorithms1, drummer Paul Hession
and flautist/saxophonist Finn Peters performed with
a number of live algorithms. I consider the performance
between Paul Hession and Arne Eigenfeldt’s (Eigenfeldt,
2014) system2 (a discussion of the factors underlying such
concerts can be found in Bown et al. (2013)). Clearly, as
an improvised duet, the interaction between the two participants
is critical to understanding the creative output. Musical
improvisation is possibly the most unambiguous case
of a process of interaction underlying a creative result. But
over a longer timescale we can consider Eigenfeldt’s development
of the system, and his interaction with Hession during
rehearsal as part of the creative process. It has been proposed
in various ways (e.g., McLean and Wiggins, 2010),
that creative software development involves a cycle of interaction
between developer and software, and we can see
this as directly analogous to the case of the potter described
by Malafouris, with the same arguments applying. Such notions
have also been discussed in the case of Cohen’s work
with AARON (McCorduck, 1990).
A full picture of the development of the outcome might
look something like the following. Through discussions
with many live algorithm developers, this seems typical, and
really it is just a specific case of what any musicians do in
preparing for a collaborative performance:
1. Designer takes on project, listens to recordings of Musician
in order to approach design of System;
2. Designer iteratively develops System;
3. Designer, System and Musician rehearse;
4. System and Musician perform.
1
Cafe Oto, London, June 29th 2014, as part of the New Interfaces
for Musical Expression 2014 Conference. 2
https://www.youtube.com/watch?v=vL6Jty5hOFc
In this we can look at the moments where there is influence.
Of interest, in Stage 1, the musician has influence on
the System. In Stage 2, the system has influence on the designer,
and in Stage 3 the System has influence on the musician,
influencing how they might choose to perform. Under
Malafouris’ framework, these interactions, no matter how
consciously or authoritatively the subject of the influence is
receiving this input, imply that boundaries between these entities
are fluid, or porous. We should be aware that that design
of the system contains iterative, hence albeit minutely
autopoietic, development, and the final form of both system
and musician are the result of a longer-term interaction.
Still, does this matter? It is not burningly evident that it
does. But it provides a more complete analysis than if we say
that a system, all of a sudden, stands alone as an autonomous
agent and ‘produces’ things. A rich qualitative description
takes account of the actual pathways that lead to something
being produced.
Conclusion
In this paper I consider what is still, despite its long standing
in social sciences, quite a radical approach to thinking about
attributing creative agency. This view removes the privilege
of the human actor, making place for the idea of humans and
other actors forming temporary networks of interaction that
produce things. It does not unfortunately offer us a powerful
analytical framework that makes agency attribution easy or
formulaic, but asks us to avoid making mistaken and simple
agency attributions, whether to humans or to creative machines.
<references_biblio/>
References
Aunger, R. 2000. Darwinizing Culture. New York: Oxford
University Press.
Baars, B. J. 2005. Global workspace theory of consciousness:
toward a cognitive neuroscience of human experience.
Progress in brain research 150:45–53.
Barkow, J. H.; Cosmides, L.; and Tooby, J. 1992. The
Adapted Mind: Evolutionary Psychology and the Generation
of Culture. New York: OUP.
Boden, M. 1990. The Creative Mind. George Weidenfeld
and Nicholson Ltd.
Bown, O.; Eigenfeldt, A.; Martin, A.; Carey, B.; and
Pasquier, P. 2013. The musical metacreation weekend:
challenges arising from the live presentation of musically
metacreative systems. In Proceedings of the musical
metacreation workshop, AIIDE conference, Boston.
Bown, O. 2012. Generative and adaptive creativity. In McCormack,
J., and d’Inverno, M., eds., Computers and Creativity.
Berlin: Springer. 361–381.
Bown, O. 2014. Empirically grounding the evaluation
of creative systems: incorporating interaction design.
In Proceedings of the Fifth International Conference on
Computational Creativity.
Proceedings of the Sixth International Conference on Computational Creativity June 2015 21
Boyd, R., and Richerson, P. J. 1985. Culture and the Evolutionary
Process. Chicago, IL, US: University of Chicago
Press.
Clark, A. 2003. Natural-Born Cyborgs: Minds, Technologies,
and the Future of Human Intelligence. Oxford University
Press.
Colton, S.; Pease, A.; and Ritchie, G. 2001. The effect
of input knowledge on creativity. Technical Reports of
the Navy Center for Applied Research in Artificial Intelligence.
Csikszentmihalyi, M. 1999. Implications of a systems perspective
for the study of creativity. In Sternberg, R. J.,
ed., The Handbook of Creativity. New York: Cambridge
University Press. 313–335.
Dawkins, R., and Krebs, J. R. 1978. Animal signals: Information
or manipulation? In Krebs, J. R., and Davies,
N. B., eds., Behavioural Ecology: An Evolutionary Approach.
Sinauer Associates. 282–309.
Eigenfeldt, A. 2014. Generating structure–towards largescale
formal generation. In Tenth Artificial Intelligence
and Interactive Digital Entertainment Conference.
Ingold, T. 2007. Materials against materiality. Archaeological
dialogues 14(01):1–16.
Latour, B. 1996. On actor-network theory: a few clarifications.
Soziale welt 369–381.
Law, J. 1992. Notes on the theory of the actor network:
Ordering, strategy and heterogeneity. Published online
at www.lancs.ac.uk/fass/sociology/papers/law-noteson-ant.pdf.
Malafouris, L. 2007. At the potter’s wheel: An argument
for material agency.
Marriott, M. 1976. Hindu transactions: Diversity without
dualism. University of Chicago, Committee on Southern
Asian Studies.
McCorduck, P. 1990. AARON’s Code: Meta-Art, Artificial
Intelligence, and the Work of Harold Cohen. W. H. Freeman
and Co.
McCormack, J.; Bown, O.; Dorin, A.; McCabe, J.; Monro,
G.; and Whitelaw, M. 2014. Ten questions concerning
generative computer art. Leonardo 47(2):135–141.
McLean, A., and Wiggins, G. A. 2010. Bricolage programming
in the creative arts. In 22nd Annual Psychology of
Programming Interest Group.
Miller, D. 2010. Stuff. Polity.
Minsky, M. 1988. Society of mind. Simon and Schuster.
Nia, H. T.; Jain, A. D.; Liu, Y.; Alam, M.-R.; Barnas, R.; and
Makris, N. C. 2015. The evolution of air resonance power
efficiency in the violin and its ancestors. Proceedings of
the Royal Society of London A: Mathematical, Physical
and Engineering Sciences 471(2175).
Nietzsche, F. 1998. Twilight of the Idols. Oxford University
Press.
Ritchie, G. 2007. Some empirical criteria for attributing
creativity to a computer program. Minds and Machines
17(1):67–99.
Shennan, S. 2002. Genes, Memes and Human History.
Thames and Hudson, London.
Simonton, D. K. 2003. Scientific creativity as constrained
stochastic behavior: the integration of product,
person, and process perspectives. Psychological bulletin
129(4):475.
Sperber, D. 2007. Seedless grapes: Nature and culture.
In Margolis, E., and Laurence, S., eds., Creations of the
Mind: Theories of Artefacts and Their Representation.
Oxford University Press. chapter 7.
Wiggins, G. A. 2006. A preliminary framework for description,
analysis and comparison of creative systems.
Knowledge-Based Systems 19(7):449–458.
Proceedings of the Sixth International Conference on Computational Creativity June 2015 22