Four PPPPerspectives on Computational Creativity

Anna Jordanous 1

Abstract. From what perspective should creativity of a system be in the light of how they are relevant to computational creativity re-
considered? Are we interested in the creativity of the systems out- searchers.
put? The creativity of the system itself? Or of its creative processes?
Creativity as measured by internal features or by external feedback?
Traditionally within computational creativity the focus had been on
1.1 The product/process debate in computational
the creativity of the systems Products or of its Processes, though this
creativity evaluation
focus has widened recently regarding the role of the audience or the As a research community, we have largely focussed on assess-
field surrounding the creative system. In the wider creativity research ment of creativity via assessment of the artefacts produced. [8,
community a broader take is prevalent: the creative Person is consid- p. 1]
ered as well as the environment or Press within which the creative
entity operates in. Here we have the Four Ps of creativity: Person, As illustrated by the ICCC 2014 call for papers, one important
Product, Process and Press. This paper presents the Four Ps, explain- debate in computational creativity is about whether evaluation of a
ing each of the Four Ps in the context of creativity research and how it creative system should focus exclusively on the output produced by
relates to computational creativity. To illustrate how useful the Four the system, or whether the processes built into the system should
Ps can be in taking a fuller perspective on creativity, the concepts also be taken into account. Should both product and process should
of novelty and value explored from each of the Four P perspectives, be included in evaluation [39, 8, 20], or should evaluation concen-
uncovering aspects that may otherwise be overlooked. This paper ar- trate solely on the product of systems [45]? Ritchie [45] stated that
gues that the broader view of creativity afforded by the Four Ps is examining the process is unimportant for creativity, arguing that hu-
vital in guiding us towards more encompassing and comprehensive mans normally judge the creativity of others by what they produce,
computational investigations of creativity. because one cannot easily observe the underlying process of human
creativity. Ritchie therefore advocated a black-box testing approach,
where the inner program workings are treated as unknown and eval-
1 Introduction uation concentrates on the systems results. Later, however, Ritchie
[46] conceded that it can be important to consider a systems mech-
A practical issue arises when considering the evaluation of a com- nisms in the case of more theoretical research[46, p. 147].
putational creativity system: from what perspective should creativity While it is true that we can only use the material we have avail-
of a system be considered? Are we interested in the creativity of the able to form an evaluation, evaluation experiments [36, 19] show that
systems output? The creativity of the system itself? Or of its creative people often make assumptions about process in their judgements
processes? Creativity as measured by internal features or by external on product. As Hofstadter pointedout, covert mechanisms can be
feedback? deeply probed and eventually revealed merely by means of watching
The computational creativity community has traditionally consid- overt behaviour ... [this approach] lies at the very heart of modern
ered creativity from the perspective of the creative output produced science. [15, quoted in p. 10, [39]]. Pearce & Wiggins [36] discussed
by a system, or the processes employed within creative systems (with how our interpretation of how something was produced is important,
notable exceptions, such as Saunders [48]). The call for this ICCC even if the actual method is unknown, and that such an interpreta-
2014 conference invites papers addressing the Process vs. product: tion can be derived if people are repeatedly exposed to the compo-
addressing the issue of evaluating/estimating creativity (or progress sitional systems (human or computational) that they are evaluating.
towards it) in computational systems through study of what they pro- Collins [6] discussed how making reasonable assumptions can assist
duce, what they do and combinations thereof. the reverse-engineering3 of program code from output, in scenarios
This paper argues that to consider process and product is not where white-box testing (evaluation with access to the program code)
enough; computational creativity should be considered and explored is not possible.
from four different perspectives, known as the Four Ps: the creative Colton [8] acknowledged Ritchies arguments but quotes exam-
Person, Product, Process and Press (or environment) [43, 26]. ples from art to demonstrate that process is as important as the end
The Four Ps have long been prevalent in creativity research re- product when evaluating creativity, at least in the artistic domain. As
lating to humans2 and enable a more inclusive and encompassing evidence, Colton cites conceptual art for details on conceptual art in
approach to the study of creativity and accommodating multiple rel- the context of this debate, where the concepts and motivations be-
evant perspectives. Here the Four Ps are presented and considered hind the artistic process are a significant contribution of the artwork.
Sol LeWitt defined Conceptual Art [25] as an art form where the
1 University of Kent, UK, email: [email protected]
2 Variants of these Ps also arise in slightly different guises in non-related 3 Reverse-engineering is the process of identifying and perhaps replicating
areas, such as software project management [16] or education [2]. how a product is made, through analysis of that product.
idea or concept is the most important aspect of the work. ... The idea models [9] which consider various aesthetic features and interactions
becomes a machine that makes the art. Two examples are Tracey between audience and system. Work on computationally creative so-
Emins controversial exhibit My Bed (1999) and Duchamps Foun- cieties has also developed in the last few years [48, is a significant
tain (1917). Jordanous [20] makes similar arguments for creativity example].
in musical improvisation, finding that the process of improvisation is Along a similar broadening of perspectives, the next section
often seen as more relevant for creativity than the end result. brings in work from the wider creativity research community, ex-
If assessing how creative a piece of conceptual art or a musical amining further viewpoints - the creative person operating in a
improvisation is, solely by evaluating the product, then there are two press/environment - and relating these viewpoints to a computational
negative consequences: creativity standpoint.

1. The primary intentions of the artist/musician are ignored (their

focus is on how the creative work is made rather than the end 2 The Four Ps of creativity
result). One major approach in creativity research is to break down creativity
2. The level of creativity presented will probably be underestimated, into four perspectives, commonly referred to as the Four Ps [43, 51,
especially if the creative process results in producing something 34, 26, 49, 53, 35]:
that might seem commonplace outside the context of that art in-
stallation/musical performance. Person: The individual that is creative.
Process: What the creative individual does to be creative.
Colton [8] also posed a thought experiment that considers two
Product: What is produced as a result of the creative process.
near-identical paintings presented at an exhibition. In the first paint-
Press: The environment in which the creativity is situated.
ing, the dots are placed randomly, whereas in the second, the dots lo-
cations represent the artists friendships with various people. Colton
argued that the second painting would be more appealing to pur-
chase than the first, though the end product is very similar, due to the
process by which it was created. Coltons thought experiment illus-
trates how process can impact on our judgement of creative artefacts,
though one could question if the experiment explores perception of
creativity, or of quality/appeal.
The thought experiment described by Ventura [54] gives further
evidence (perhaps unintentionally) on how knowledge of the creative
process affects how we evaluate creativity. Two creative systems, the
RASTER and iRASTER systems, were designed by Ventura to be
decidedly non-creative. If these systems were implemented and their
generated images were given to people to evaluate without telling the
evaluators how they were produced, the evaluators may well rate the
Figure 1. A simplified view of how the Four Ps fit together in creativity
creativity of the system highly. Supplying the evaluators with details
of how a program works, though, could have a detrimental impact on
the subsequent evaluations [11, 8].
One issue with creativity is analogous to the adage that a magi- Rhodes [43] was perhaps first to identify the four P perspectives.
cian never reveals their secrets. This adage is based on the fact that Rhodes collected 40 definitions of creativity and 16 definitions of
tricks do not appear so impressive once you have found out how the imagination. The Four P dimensions of creativity emerged from
magician performed the trick. Similarly things can appear to be less analysis of these definitions.5 Several people seem to have indepen-
creative when you know how they were produced:4 dently identified four similar themes of creativity [26, 51, 34, 35],
boosting the credibility of the Four Ps.
it is not unknown for critics of AI to refuse to accept programs
Plucker, Beghetto & Dow [41] conducted a literature survey in-
as creative (or intelligent) once the mundane mechanistic nature
vestigating the use (or absence) of creativity definitions in creativity
of the inner workings are revealed [44, p. 4]
research. As part of this review, Plucker et al. used their analysis
Colton [8] intentionally sidestepped this issue by reporting on his to derive their own definition by identifying reoccurring themes and
artistic system in high-level terms only, rather than giving details of forming these into an inclusive definition which happens to account
the program [8, p. 8]. for each of the Four Ps:
Until recently, computational creativity evaluation methodologies Creativity is the interaction among aptitude, process, and envi-
mainly looked solely at a systems products [45, for example] or at ronment by which an individual or group produces a perceptible
a combination of the products and the process [39]. Recently it has product that is both novel and useful as defined within a social
been acknowledged that there is more to creativity than process and context [41, p. 90]
product, with the Creative Tripod [8], whose evaluative framework
is influenced by how an audience perceives the creativity of a sys- In reviewing Four Ps research, Kaufman [23] described adden-
tem, SPECS [20] which requires the researcher to investigate what dums that have been suggested for the Four Ps: persuasion [49]
creativity means in the context of their system, and the FACE/IDEA
5 As Rhodes work appeared in a relatively unknown journal, many later ad-
4 If the inner workings of a program are very impressive, complex or novel, vocates of a Four Ps-style approach to creativity seem unaware of Rhodes
then we may still be impressed by the program, but this is a different per- contribution (e.g. Odena, 2009, personal communications), so fail to cite
spective to whether or not we think the program is creative. him.
and potential [47]. In general, however, the Four Ps have been extent, process modelling, rather than on the modelling of character-
adopted as they were originally conceived by various researchers istics of a creative Person in computational format. This is because
[43, 51, 34, 26]. computational creativity systems tend to be oriented towards a par-
ticular goal or domain, rather than being generally creative, as we can
see by the plethora of domain-specific systems (as opposed to mod-
2.1 The Four Ps: Person elling of creative personal characteristics) in the various proceedings
This perspective addresses human characteristics associated with of ICCC conferences (International Conference on Computational
creative individuals or groups of people. Encouraged by Guilfords Creativity). As argued in [20], different types of creativity require
call in 1950 for studying the creative person, an abundance of dif- domain specific skills to some extent, so domain-specific computa-
ferent personal characteristics have been associated with creativity tional creativity systems tend to be built around the most prominent
[43, 51, 24, 53, 35], ranging from personality traits, attitudes, intel- necessary skills for that domain.
ligence and temperament to habits and behaviours (for example cu- In terms of evaluating creative systems, Coltons Creative Tripod
riosity, persistence, independence and openness). Some of these are [8] emphasises the need for systems to demonstrate skill, imagina-
closely related; others are contradictory. Rhodes mentioned the rele- tion and appreciation before they can be considered as a candidate
vance to creativity of peoples personality traits, attitudes and habits, creative system, all three of which are alluding to personal character-
physique and intelligence and the identifiable features of creative istics.
people, as well as referring to peoples temperament, habits, self- Features, traits and aspects of the creative system can be studied,
concept, value systems, defence mechanisms, and behaviour [43, p. and it would be fascinating to explore how general creative personal
307]. characteristics could be specifically modelled within creative sys-
Empirical studies up until 1968 were summarised by Stein [52] tems (see the Process section, next). Computational modelling of
and were combined into a list of 18 distinct personality characteris- characteristics that encourage creativity could help us progress our
tics of a creative person, including aspects such as curiosity, persis- systems to be able to be creative in more than one system which
tence, independence and openness. Stein used these characteristics they were originally designed for; this would be significant progress
to identify creative individuals for study. There is a risk of circularity in our pursuit of modelling creativity as a phenomenon which tran-
here, as the selection criteria for creative individuals chooses people scends different types of creative activity.
to be studied, then the study involves examining those characteris- The Person could also entail the individual(s) interacting with a
tics and criteria. Steins work has not stood the test of time, with few creativity support system or co-creative system which interacts with
current citations. people[27, 22]. Another possible interpretation of the Person in
Several researchers subdivide the Person category into finer- computational creativity would be to acknowledge the role of the
grained groups. Three sub categories of the pupil perspective programmer(s), tester(s), researcher(s) and peers involved in shaping
emerged during Odena & Welchs work [35]: personal characteristics the project.
of the pupil, their individual learning style (either adapting to new in-
formation or deriving new information themselves) and the influence
2.2 The Four Ps: Process
of the pupils background. Koestler [24] described three types of cre-
ative person: the Artist, the Sage and the Jester. Through Tardif & The creative process has been broken down into a series of sequential
Sternbergs review of definitions of creativity, three main categories or cyclic stages occurring over time [42, 55] or subtasks [35].
were identified with which to describe creative people: cognitive at- In their work on student creativity in school music lessons, Odena
tributes, personality attributes/motivation and developmental influ- & Welch [35] broke down the creative process into subtasks, identi-
ences. Tardif & Sternberg suggested three resultant modes of study fying various types of process (e.g. different activities, group process,
of human creativity: cognitive psychology; psychometric testing; and the structuredness or otherwise of a process and composition by im-
study of human development. provisation) rather than tracing a linear progression of subprocesses.
These discrepancies and the sheer quantity of attributes together It is often stressed that creativity is not just the first flash of in-
place an obstacle in the way of compiling a definitive list of attributes spiration, but is also the activity that validates, develops, and refines
of a creative person and instead provoke disagreements on exactly that first idea; rather than occurring at one point in time, creativity
which cognitive characteristics should be attributed to creative peo- develops over a period of time [55, 42, 53]. Tardif & Sternberg [53]
ple. Tardif & Sternbergs review showed that as of 1988, different questioned whether creativity is a social or an individual process. The
authors highlighted a variety of characteristics, with no general con- social view of creativity has notably been promoted by Csikszentmi-
sensus and no characteristics common to all reports [53, Table 17.1, halyi [12].
p. 434].
2.2.1 The Process in Computational Creativity
2.1.1 The Person in Computational Creativity
In computational creativity, the creative process might be that em-
In computational creativity, the creative person could be analogous to ployed by a single piece of software, or the interactions between mul-
the computer, or perhaps more accurately, to the computer program, tiple machines or programs, or the interactions between machine and
software, or to a creative agent within a multi-agent system. Here human users. As described above, the computational creativity com-
the machine is the hardware hosting the creative agent, much as we munity has given some attention to the concept of creative processes
might distinguish between physical and functional characteristics of employed within computational creativity, with growing attention
a Person. paid to this aspect in recent years. For example, the FlowR frame-
Interesting work has been done on modelling creative agents, for work [5] is designed to facilitate creative computational workflows
example by Saunders [48], although the emphasis in computational by chaining together processes in a linear pattern, and from personal
creativity software tends to be on product generation and to some communications with members of the project team, there are plans
to consider non-linear chains of processes as well. Additionally, the exclusively focuses on evaluating the products of computational sys-
work by Joanna Misztal on poetry generation [31] specifically fo- tems without considering any of the other three Ps.6
cusses on the processes required to generate poetry, at various levels
of abstraction.
The generate-and-test [30, 38] or engagement-reflection approach
2.4 The Four Ps: Press/Environment
[40] specifically models the creative process as a cycle of generating The Press perspective encompasses a bidirectional perspective be-
artefacts then improving the generation process via evaluating the tween the environment which influences the creator and receives the
generation phase. This is an approach which deserves broader adop- creative work, and the creator who publicises their work and is given
tion within computational creativity; evaluation is a critical part of feedback on what they produce. Tardif & Sternberg [53] considered
the creative process [42, 12]. In terms of post-implementation evalu- both creative domains themselves and the social environments in
ation, the FACE model for evaluation of creative systems [9] places which creative people are influenced as they employ creative pro-
importance on computational systems being able to report on the cre- cess, advertise their creative products and receive feedback. Rhodes
ative process (this report is referred to in the FACE framework as a [43] concentrated on the role that the environment plays on a person
Frame). during the creative process, rather than how the creative produce is
There are multiple theories about how human creativity processes judged by the external world after being created. Rhodes reflected
are structured (see for example [42, 12, 23, 14]). Computational cre- on how everyone is different, so everyone perceives the world in a
ativity research can provide a test-bed for these psychological the- unique way and processes ideas according to their own contexts.
ories and allow us to explore if implementing the theories result in Of the Four Ps, this is the perspective that is often neglected when
creative behaviour. Conferences such as the Creativity and Cognition one takes an individualistic view of creativity. In general creativity
series showcase work that links between theory and practice to some theorists do however acknowledge the influence of the environment
extent, but further activity along these lines would emphasise the va- in which creativity is situated [49, 13]. If one concentrates on an indi-
lidity of computational creativity research, allowing computational viduals creativity, however, the Press perspective is often neglected,
work to contribute to human creativity research and vice-versa. even if unintentionally. For example, although stating that [t]o be
appreciated as creative. a work of art or a scientific theory has to be
understood in a specific relation to what preceded it [3, p. 74], Bo-
2.3 The Four Ps: Product dens treatment of creativity mainly focused on different cognitive
processes of creativity, rather than a detailed examination of social
Many authors advocate that proof of creativity is necessary to be con-
or environmental influences.
sidered creative [21, 53, 41, 44]. The product-centric view adopted
by computational creativity researchers such as Ritchie [45], that cre-
ative products are both necessary and sufficient for creativity, was 2.4.1 The Press in Computational Creativity
present in earlier human creativity research [21]. But, inspired by
Some computational creativity researchers are starting to highlight
Guilfords seminal 1950 address on creativity research, emphasis in
the importance of the environment in which a creative system is situ-
human creativity research shifted from identifying creative individ-
ated [50, 17, 37, 48], with some of this work influenced by the DIFI
uals post-production of creative work, to predicting future potential
(Domain-Individual-Field-Interaction) framework [12]. Social inter-
for creativity in individuals. This change in emphasis is illustrated
action between creative agents and their audience is an area which
in the proliferation of psychometric tests [23, 19] within creativity
has been neglected by all but a few groups of researchers: for exam-
research.
ple nearly 75% of papers in the 2014 International Conference on
Tardif & Sternberg [53] considered the creative product more
Computational Creativity failed to make any reference to social or
briefly than the other three Ps in their review, deciding that while
interactive aspects of creativity. But creativity cannot exist in a vac-
a creative product is essential for creativity, it is not enough merely
uum. A recent increase in development of the interactivity of creative
to generate a product; the product should also be considered in a
systems (especially where this affects the way these systems works)
domain-specific context.
is pleasing to see and deserves further attention [10].
Computational creativity research has long acknowledged the im-
There is a separate point to acknowledge regarding Press in com-
portance of the output or artefacts generated by creative systems, as
putational creativity. As computational creativity researchers, we
described above. To borrow a metaphor from human creativity re-
should stay aware of any potential biases that may be introduced,
search, it has been common (until recently) for computational cre-
should an audience be aware that the creative agent of interest is
ativity to follow the product-centric approach to creativity as advo-
computational rather than human [32, 19].7
cated by Kagan: Creativity refers to a product, and if made by a man,
we give him the honor of the adjective [21, p. viii].
2.5 Interaction between the Four Ps
2.3.1 The Product in Computational Creativity Simonton [49] saw discrepancies between combining the Four Ps in
theory and in practice:
Generating creative products has been an area of significant success
for computational creativity. To see examples, one just needs to con- Now, in an ideal state of affairs, it should not matter which
sult any years proceedings of the International Conference on Com- one of the four ps our investigations target, for they all will
putational Creativity where there are multiple examples to be found converge on the same underlying phenomenon. ... But reality
of systems which are reported in terms of the products they gener- is not so simple, needless to say. The creative process need not
ate. The success of systems is often reported in terms of what kind 6 Recently proposed evaluation methods such as [8, 9, 19] place more em-
of artefacts they generate, as noted in [18]. Some systems have been phasis on the other three Ps.
7
evaluated using Graeme Ritchies empirical criteria [44, 45], which Many thanks to the anonymous reviewer who noted this point.
arrive at a creative product, nor must all creative products ensue characterises computational creativity research as scoffing at mere
from the same process or personality type; and others may ig- generation for more than a decade. Generation of creative products
nore the process, discredit the product, or reject the personality is only a quarter of the full picture of creativity, only one of the Four
when making attributions about creativity. [49, p. 387] Ps. Granted, we have achieved much success in product generation,
as exemplified by exhibitions, concerts and other demonstrations of
From this, one conclusion which seems to follow naturally is that creative products reported in various papers on computational cre-
an accurate and comprehensive definition of creativity must account ativity systems [18]. However, the more mature work and exciting
for the (potential) presence of all four aspects, in order to be com- potential comes from the incorporation of the other three Ps, at least
plete. Simonton, however, concluded that [i]f we cannot assume that to some extent, such as in [40, 48, 31].
all four aspects cohesively hang together, then it may be best to se-
lect one single definition and subordinate the others to that orienta-
tion [49, p. 387], with his natural research inclination leading him 3 Applying the Four Ps: examples of novelty and
to focus his work on persuasion, his term for the Press/Environment value
aspect. Novelty (originality, newness) and value (usefulness, appropriate-
The mysterious impression often associated with creativity [56, 3, ness) form key parts of creativity [28, 3, 45, 20], often being identi-
23] can be explained to some extent when one or more of the Four fied as the two main aspects of computational creativity [39, 45, 4,
Ps are not accounted for: for example].8 Work in computational creativity illustrates both nov-
elty and utility from each of the Four P perspectives, although some
Each strand [of the Four Ps] has unique identity academically,
perspectives are represented more plentifully within computational
but only in unity do the four strands operate functionally. It is
creativity than others. To illustrate the discussions above, we can dis-
this very fact of synthesis that causes fog in talk about creativity
cuss novelty and value in computational creativity from each of the
and this may be the basis for the semblance of a cult. [43, p.
Four P perspectives. Considering novelty from each of the Four Ps:
307]
Product Novelty is well associated with system outputs and prod-
Rhodes argued that creativity research should follow a specific
ucts: how novel are the generated artefact(s)? The novelty of arte-
path: from product to person and thence to process and to press.
facts generated by computational creativity systems is a key con-
[43, p. 309]
sideration in Ritchies empirical criteria for evaluating creative
Objective investigation into the nature of the creative process systems [45].
can proceed in only one direction, i.e. from product to person Process A creative process can take a novel approach or be im-
and thence to process and to press. [43, p. 309] plemented in a novel way, perhaps employing new algorithms or
techniques or different approaches. Efforts at trying new processes
Such a statement makes Rhodess contribution less useful. For ex- and combinations thereof are being encouraged by systems such
ample, the Press (environment) in which one is creative has some as the FlowR framework [5], which focuses specifically on en-
influence on the creative Process, so one may prefer to study how abling us to chain different processes together for creative pur-
Press and Person interact before looking at Process issues. Simon- poses.
ton viewed creativity as how a persons ideas emerge as influential Person Creativity can be performed by a new creative entity,
when that person, by chance, has new ideas and promotes them to which demonstrates or uses novel characteristics relevant to that
influence others. Creative people would not be equivalent to lucky creativity. As is often encountered in computational creativity
people, by this interpretation, but chance would intervene in their work, implementing or running a creative system on new hard-
success. Simonton refers to this as the chance-configuration theory ware or in different software may also impact upon the systems
that outlines the general conditions that favor creativity [49, p. 422]. performance and may have unexpected results. The number of
Tardif & Sternberg [53] treated each of the Four Ps individually, new systems presented each year at the International Conference
as these really are separate levels of analysis, and it is from com- on Computational Creativity exemplifies how novel creative enti-
parisons within levels that coherent statements about our knowledge ties continually arise in computational creativity research.9 (Also,
of creativity can be made [53, p. 429]. Tardif & Sternbergs sum- the novelty of unexpected results is often unintentionally exempli-
mary is weakened somewhat by this as it does not make comparisons fied when live demos of these systems are attempted in unfamiliar
across the Four Ps, despite highlighting Simontons emphasis on the computing setups.)
interactions and relations between these four views [49]. In contrast Press The creativity demonstrated by a system can be noted as
Mooney [34] argued that the four approaches should be integrated in being novel in a particular environment, even though it may be
a model of creativity, proposing a model that puts together the four commonplace in other environments. The system may also ex-
approaches by showing them to be aspects of one unifying idea [34, ploit the surrounding press in previously unexplored ways. This
p. 333]. While Mooneys claims become rather grandiose at points, was demonstrated neatly by the combination of two systems in
Mooneys more specific contributions on creativity match neatly with [33], where a textual annotation system interacted with a system
the four Ps approach identified elsewhere at that date [43, 51] that generates emotion-driven music. The combination resulted in
novel interpretations of fairy tales; such results would not have
2.5.1 Interaction between Four Ps in Computational arisen were the systems operating in isolation.
Creativity
Considering value from each of the Four Ps:
This paper argues that we can make significant progress in compu- 8 It should be clarified that for this author, creativity consists of consider-
tational creativity by considering all four Ps in our computational ably more than novelty and value, though these are two key components of
creativity work. Tony Veales tagline for the ICCC2012 conference creativity. See [20].
9 See http://www.computationalcreativity.net/conferences.
sums up current aspirations of computational creativity well; Veale
Product Value is also well associated with system outputs and REFERENCES
products: how valuable or good are the generated artefact(s)? This
is a highly current area of concern within computational creativity, [1] Daniel Allington, Byron Dueck, and Anna Jordanous. Networks of
with much evaluation concentrating on the quality of output [18]. value in Electronic Dance Music: SoundCloud, London, and the impor-
Process The creative processes being incorporated within creativ- tance of place, submitted.
ity can be useful in themselves for learning or studying how cer- [2] J. B. Biggs, Constructing learning by aligning teaching: constructive
alignment, in Teaching for quality learning at university: what the
tain approaches and techniques work or for cross-application to student does, 1133, SRHE & Open University Press, Buckingham,
new areas. Systems with an emphasis on modelling process, such (2003).
as Misztal and Indurkhyas poetry generator [31] bring added util- [3] Margaret A. Boden, The creative mind: Myths and mechanisms, Rout-
ity by what they reveal about the processes being modelled. ledge, London, UK, 2nd edn., 2004.
[4] David Brown, Computational artistic creativity and its evaluation,
Person Some creators become more valuable than others as a con-
in Computational Creativity: An Interdisciplinary Approach, number
tributor in their field, based on their personal characteristics, expe- 09291 in Dagstuhl Seminar Proceedings, Dagstuhl, Germany, (2009).
rience and influence.10 The same can be noted for creative systems [5] John Charnley, Simon Colton, and Maria Teresa Llano, The flowr
to some extent; some are cited more often than others, for exam- framework: Automated flowchart construction, optimisation and alter-
ple Simon Coltons HR mathematical discovery system [7] (which ation for creative systems, in Proceedings of the 5th International Con-
ference on Computational Creativity, (2014).
provides a useful example of creativity in a non-artistic domain). [6] Nick Collins, The analysis of generative music programs, Organised
Press If creative activities benefit the external world in Sound, 13(3), 237248, (2008).
some way, then they have value to the press. As ex- [7] Simon Colton, Automated theory formation in pure mathematics, Dis-
ample, Harold Cohens AARON colouring system has re- tinguished dissertations, Springer, London, UK, 2002.
[8] Simon Colton, Creativity versus the perception of creativity in com-
ceived much external attention, from media discussions [29]
putational systems, in Proceedings of AAAI Symposium on Creative
through to inspiring a screensaver for personal computers via Systems, pp. 1420, (2008).
http://www.kurzweilcyberart.com. [9] Simon Colton, John Charnley, and Alison Pease, Computational Cre-
ativity Theory: The FACE and IDEA descriptive models, in Proceed-
These above lists are not intended to be a full and conclusive por- ings of the 2nd International Conference on Computational Creativity,
pp. 9095, Mexico City, Mexico, (2011).
trait of novelty and value within computational creativity. What these [10] Simon Colton and Geraint A. Wiggins, Computational creativity: The
lists illustrate is the different viewpoints that can be uncovered using final frontier?, in Proceedings of 20th European Conference on Artifi-
the Four Ps as signposts with which to guide our thinking around cial Intelligence (ECAI), pp. 2126, Montpellier, France, (2012).
computational creativity. The breadth of issues mentioned above [11] David Cope, Computer Models of Musical Creativity, MIT Press, Cam-
shows aspects of novelty and value within computational creativity bridge, MA, 2005.
[12] Mihaly Csikszentmihalyi, Society, culture, and person: a systems view
which may not always be accounted for if taking a product/process- of creativity, in The Nature of Creativity, ed., Robert J. Sternberg,
oriented viewpoint; however it is argued here that those perhaps- chapter 13, 325339, Cambridge University Press, Cambridge, UK,
overlooked aspects give us a closer rendition of creativity, guiding (1988).
us away from incomplete viewpoints of creativity in the context of [13] Beth A. Hennessey and Teresa M. Amabile, The conditions of creativ-
ity, in The Nature of Creativity, ed., Robert J. Sternberg, chapter 1,
our computational work. 1138, Cambridge University Press, Cambridge, UK, (1988).
[14] Beth A. Hennessey and Teresa M. Amabile, Creativity, Annual Re-
view of Psychology, 61, 569598, (2010).
4 Summary [15] Douglas Hofstadter, Creativity, brain mechanisms and the Turing test,
in Fluid concepts and creative analogies, ed., Douglas Hofstadter, 467
The difficulty of understanding what creativity is should not discour- 491, Harper Collins, New York, (1994).
age us from such an attempt [43, 41, 8]. In creativity research, the [16] Ivar Jacobson, Grady Booch, James Rumbaugh, James Rumbaugh, and
Four Ps construct ensures we pay attention to four key aspects of Grady Booch, The unified software development process, volume 1,
creativity: the creative Person, the generated Products, the creative Addison-Wesley Reading, 1999.
[17] K. E. Jennings, Developing creativity: Artificial barriers in artificial
Process and the Press/Environment hosting and influencing the cre-
intelligence, Minds and Machines, 20(4), 489501, (2010).
ativity. This framework helps us to consider creativity more broadly. [18] Anna Jordanous, Evaluating evaluation: Assessing progress in com-
For example, if viewing novelty and value from the perspectives putational creativity research, in Proceedings of the Second Interna-
of product, process, person and press, we uncover various interpre- tional Conference on Computational Creativity (ICCC-11), Mexico
tations of these two key concepts within computational creativity City, Mexico, (2011).
[19] Anna Jordanous, Evaluating Computational Creativity: A Standardised
which may otherwise have been overlooked. The Four Ps framework Procedure for Evaluating Creative Systems and its Application, Ph.D.
helps to highlight different perspectives on creativity, to portray cre- dissertation, University of Sussex, Brighton, UK, 2012.
ativity in a fuller context. [20] Anna Jordanous, A Standardised Procedure for Evaluating Creative
Systems: Computational creativity evaluation based on what it is to be
creative, Cognitive Computation, 4(3), 246279, (2012).
ACKNOWLEDGEMENTS [21] Creativity and learning, ed., Jerome Kagan, Beacon Press Press,
Boston, MA, 1967.
Many thanks to Carly Lassig for originally making me aware of the [22] Anna Kantosalo, Jukka Toivanen, Hannu Toivonen, and Ping Xiao,
Four Ps and pointing me towards Rhodes [43]. Thanks also to various From isolation to involvement: Adapting machine creativity software
to support human-computer co-creation, in Proceedings of 5th Inter-
computational creativity researchers who provided feedback on ear- national Conference on Computational Creativity, ICCC, Ljubljana,
lier versions of these thoughts while in development, and the anony- Slovenia, (2014).
mous reviewers of this paper. [23] James C. Kaufman, Creativity 101, The Psych 101 series, Springer,
New York, 2009.
10 This has been found, for example, in the recent Valuing Electronic Music [24] Arthur Koestler, The act of creation, Danube Books, New York, 1964.
project http://valuingelectronicmusic.org [1], where some [25] Sol LeWitt, Paragraphs on conceptual art, Artforum International
peoples endorsements can have a greater influence on the perceived value Magazine, June 1967, (1967).
of an electronic musician and their work. [26] Douglas W. MacKinnon, Creativity: a multi-faceted phenomenon, in
Creativity: A Discussion at the Nobel Conference, ed., J. D. Roslansky, [49] Dean Keith Simonton, Creativity, leadership, and chance, in The Na-
1732, North-Holland Publishing Company, Amsterdam, The Nether- ture of Creativity, ed., Robert J. Sternberg, chapter 16, 386426, Cam-
lands, (1970). bridge University Press, Cambridge, UK, (1988).
[27] Mary Lou Maher, Computational and collective creativity: Whos be- [50] Ricardo Sosa, John Gero, and Kyle Jennings, Growing and destroy-
ing creative?, in Proceedings of the 3rd International Conference on ing the worth of ideas, in Proceedings of the 7th ACM Creativity and
Computer Creativity, Dublin, Ireland, (2012). Cognition conference, pp. 295304, Berkeley, California, (2009).
[28] Richard E. Mayer, Fifty years of creativity research, in Handbook of [51] Morris I. Stein, A transactional approach to creativity, in Scientific
Creativity, ed., Robert J. Sternberg, chapter 22, 449460, Cambridge Creativity: Its Recognition and Development, eds., Calvin W. Taylor
University Press, Cambridge, UK, (1999). and Frank Barron, chapter 18, 217227, John Wiley & Sons, New York,
[29] Pamela McCorduck, Aarons code: Meta-art, artificial intelligence, and (1963).
the work of Harold Cohen, WH Freeman, New York, NY, 1991. [52] Morris I. Stein, Creativity, in Handbook of personality theory and
[30] G. McGraw and Douglas Hofstadter, Perception and creation of di- research, eds., Edgar F. Borgotta and William W. Lambert, 900942,
verse alphabetic styles, AISB Quarterly, 85, 4249, (1993). Rand McNally, Chicago, IL, (1968).
[31] Joanna Misztal and Bipin Indurkhya, Poetry generation system with an [53] Twila Z. Tardif and Robert J. Sternberg, What do we know about cre-
emotional personality, in Proceedings of 5th International Conference ativity?, in The Nature of Creativity, ed., Robert J. Sternberg, chap-
on Computational Creativity, ICCC, (2014). ter 17, 429440, Cambridge University Press, Cambridge, UK, (1988).
[32] David C. Moffat and Martin Kelly, An investigation into peoples bias [54] D. Ventura, A reductio ad absurdum experiment in sufficiency for eval-
against computational creativity in music composition, in Proceedings uating (computational) creative systems, in Proceedings of the 5th In-
of the 3rd International Joint Workshop on Computational Creativity ternational Joint Workshop on Computational Creativity, pp. 1119,
(ECAI06 Workshop), Riva del Garda, Italy, (2006). Madrid, Spain, (2008).
[33] Kristine Monteith, Virginia Francisco, Tony Martinez, Pablo Gervas, [55] Graham Wallas, The Art of Thought, C. A. Watts & Co, London, UK,
and Dan Ventura, Automatic generation of emotionally-targeted abridged edn., 1945.
soundtracks, in Proceedings of the 2nd International Conference on [56] Raymond Williams, Keywords: a vocabulary of culture and society,
Computational Creativity, pp. 6062, Mexico City, Mexico, (2011). Fontana/Croom Helm, Glasgow, UK, 1976.
[34] Ross L. Mooney, A conceptual model for integrating four approaches
to the identification of creative talent, in Scientific Creativity: Its
Recognition and Development, eds., Calvin W. Taylor and Frank Bar-
ron, chapter 27, 331340, John Wiley & Sons, New York, (1963).
[35] Oscar Odena and Graham Welch, A generative model of teachers
thinking on musical creativity, Psychology of Music, 37(4), 416442,
(2009).
[36] Marcus Pearce and Geraint Wiggins, Towards a framework for the
evaluation of machine compositions, in Proceedings of the AISB Sym-
posium on AI and Creativity in Arts and Science, York, UK, (2001).
[37] Alison Pease and Simon Colton, On impact and evaluation in com-
putational creativity: A discussion of the Turing Test and an alterna-
tive proposal, in Proceedings of the AISB11 Convention, York, UK,
(2011). AISB.
[38] Alison Pease, Markus Guhe, and Alan Smaill, Some aspects of analog-
ical reasoning in mathematical creativity, in Proceedings of the Inter-
national Conference on Computational Creativity, pp. 6064, Lisbon,
Portugal, (2010).
[39] Alison Pease, Daniel Winterstein, and Simon Colton, Evaluating ma-
chine creativity, in Proceedings of Workshop Program of ICCBR-
Creative Systems: Approaches to Creativity in AI and Cognitive Sci-
ence, pp. 129137, (2001).
[40] R. Perez y Perez, A. Aguilar, and S. Negrete, The ERI-Designer: A
computer model for the arrangement of furniture, Minds and Ma-
chines, 20(4), 533564, (2010).
[41] Jonathan A. Plucker, Ronald A. Beghetto, and Gayle T. Dow, Why
isnt creativity more important to educational psychologists? Potentials,
pitfalls, and future directions in creativity research, Educational Psy-
chologist, 39(2), 8396, (2004).
[42] Henri Poincare, Mathematical creation, in The Foundations of Sci-
ence: Science and Hypothesis, The Value of Science, Science and
Method., volume Science and Method [Original French version pub-
lished 1908, Authorized translation by George Bruce Halsted], chapter
III of Book I. Science and the Scientist, 383394, The Science Press,
New York, (1929).
[43] Mel Rhodes, An analysis of creativity, Phi Delta Kappan, 42(7), 305
310, (1961).
[44] Graeme Ritchie, Assessing creativity, in Proceedings of the AISB
Symposium on AI and Creativity in Arts and Science, pp. 311, York,
UK, (2001).
[45] Graeme Ritchie, Some empirical criteria for attributing creativity to a
computer program, Minds and Machines, 17, 6799, (2007).
[46] Graeme Ritchie, Uninformed resource creation for humour simula-
tion, in Proceedings of the 5th International Joint Workshop on Com-
putational Creativity, pp. 147150, Madrid, Spain, (2008).
[47] M.A. Runco, Creativity, cognition, and their educational implications,
in The educational psychology of creativity, ed., John Houtz, 2556,
Hampton Press, New York, NY, (2003).
[48] R. Saunders, Towards autonomous creative systems: A computational
approach, Cognitive Computation, 4(3), 216225, (2012).