Paper presented at the British Educational Research Association Annual Conference, University of Manchester, 16-18 September 2004
The ‘c’ of creativity, like that of culture, may be writ both large and small. Education also makes this distinction, with some support from psychology. I want to explore this majuscule/miniscule disjunction from an evolutionary perspective. In the past two decades, studies in the field of human evolution, including findings in palaeontology, microbiology and primate studies, have offered insights into the origins of humanity; and thereby into the way we behave. One finding is that creativity appears unique to our species. For this peculiarly human capacity I have, elsewhere, appropriated the term ‘technicity’ – the attempt of [behaviourally] modern man to dominate the earth by controlling beings that are considered as objects – from the philosophy of Martin Heidegger. Technicity has a direct relationship to little c creativity and innovation for which, I argue, it provides the mechanism: a deconstructive/reconstructive cycle. Big C creativity I comment on in the context of evolved competitive behaviours. There is one matter that we seem to have difficulty bringing to consciousness: the relationship between technicity and language. Indicators for their respective temporal emergence in the evolutionary sequence now make a challenge to language primacy inescapable.
The concept of the selfish gene (Dawkins 1989), like that of conservation of energy or the second principle of thermodynamics, serves to keep our thinking within realistic bounds. With this caveat, we may begin by noting that creativity is poorly defined in education. Craft (2001) proposes that we consider creativity to come in two flavours: Big C and little c, which she implies lie on a continuum running from high creativity to coping. Citing Gardner, she rehearses the difficulties psychologists have had trying to pin down creativity using psychometric methods. My approach the question is from an evolutionary perspective. I am aware that in human evolution I tread on interesting ground. Not the least difficulty I face is the precise definition of our species and its precursors. For instance, the chronology I use differs significantly from that on public display in Manchester Museum. My chronology is based on two desiderata: the first is the Darwinian dictum that nature progresses without sudden jumps; the second the Skinnerian criterion that behaviour must be present to modify. I also take as axiomatic that a totally unexpected phase transition can arise from a steady, gradual, natural process: the boiling of a kettle is more than just the means to a nice hot cup of tea. The key characteristic of phase transitions is that entirely new properties emerge. We have developed the capability to examine, modify, and genetically engineer our own genes, and those of all other species. This capacity may signal such a phase-transition.
Craft discusses creativity. She speaks on the subject. She writes papers. I take as my text one of her writings. She employs two methods: story telling and lexical selection, to connote and denote features of little c creativity (LCC). I want to consider language in terms of its fitness for the purpose to which Craft bends it; and to hint at resonance with the “two cultures” of CP Snow. So, let us enquire in language of language.
I want it to be very clear that I restrict the term ‘language’ to natural language, speech, tongues, and exclude analogous connotations such as the language of genes, computer language, the language of art, and the like, for reasons that will later become apparent. I pose three ‘wh’ questions about the evolution of language: when, where, and why?
The answer to Where? is Africa (Lewin 1998), the continent of origin of genus Homo; for When? the answer appears more equivocal; and Why? rides on a wild hypothesis.
The genetically most closely related species to us are the great apes, with Bonobos and Chimpanzees (98%) the nearest. Our lineages split some five million years ago. The genus Homo is (Diamond 2002 Lewin op cit Tudge 1996) traced sequentially through H habilis, H erectus, and archaic Homo sapiens including the Neanderthals, to our species H sapiens sapiens. Associated stone tools, the systematic use of fire, and increasing brain size are touchstones for the classification Homo. Speech leaves no such direct evidence. Primate studies and palaeontology suggest, however, that the power of speech evolved within the genus Homo, and is not unique to our species. Lewin (op cit) summarises the evidence from brain endocasts, which show Broca’s and Wernike’s (speech) areas in all of Homo. The descended larynx, which provides wide vowel range at a risk of choking, is evidenced by flexion of the basicranium. This, Lewin notes, begins with H erectus and reaches modern configuration in Homo sapiens about 300Kya ago. The finding of a hyoid bone in a Neanderthal skeleton confirms a descended larynx and hence speech in this species. Barrett et al (2002:331) cite MacLarnon and Hewitt (1999) who report marked expansion in the chest region of the spinal chord, associated with fine control of breathing in speech, in modern humans. Absent in living primates, this expansion is seen in archaic H sapiens fossil vertebrae some 500Kya ago. These three evidence streams point to a modern speech system having evolved by some half a million years ago. By ‘modern speech system’ I mean the physiological capacity to make the range of sounds used in modern speech.
For grammar, semantics or pragmatics, there is no fossil evidence. But the argument that archaic H Sapiens (and the Neanderthals) lacked language of modern complexity is a dangerous one in the evolutionary context: complexity, of the eye, was deployed to discount the very concept of evolution. The why question may provide some light.
Speech development in children begins around 18 months and is essentially complete by the age of four. Larynx descent will not be complete until puberty. (Descent in H erectus is considered to be about that of an eight-year-old child (Lewin 1998: 460).) We have so early a speech development that Chomsky (1965:47) proposed an innate “language acquisition device” and Pinker (1994, 2000) wrote of “instinct,” Deacon (1998) of co-evolution with the brain; and yet Dennett (1996:381), a Darwinian to the core, can still asserted that language is “an innovation that has invaded our brains and no others” that makes scientific method, ‘generate-and-test,’ hence progress possible.
Dunbar (2002) first published his hypothesis on the origin of language in 1996, so Dennett did not have his work to consider. Eight years on, the second edition requires no modification. A primatologist, his work is founded not in attempts to teach chimps language, but in the evaluation of primate social relationships. Unencumbered by the baggage of philosophy, he asks in what manner language might confer an adaptive advantage: a straight Darwinian question. A flavour of chimpanzee social structures is to be found in the fieldwork of Goodall (1986), who noted alliance making; hierarchy; cooperation; male-female relationships; warfare; and tool making. Her findings were expanded on by later studies, e.g. Strum & Fedigan (2000), Boesch et al (2002). New findings included troupe specific tool-making passed on by social learning i.e. culture.
Humans are unique in that they work in large groups and cooperate with strangers. In addition, our neonate is some twelve months premature relative to other primates: this makes huge demands on the female and, given the difficulty of human birth, on the social group as a whole. From a genetic perspective, these are a high-risk strategies. How does our species, uniquely, make this reciprocal altruism work? It is not easy. Take a population of blind co-operators who share and share alike. Game theory shows that this form of cooperation is genetically advantageous to the individual because peaks and troughs of resource are evened out. However, if a few freeriders (who do not reciprocate) arise they will quickly drive the co-operators to extinction. The solution is to cease to be a ‘blind’ co-operator. Reciprocal altruism is an adaptive strategy if three prerequisites are fulfilled: a) the ability to recognise other individuals; b) a memory of others, and ones own, behaviour; c) following a tit-for-tat algorithm. This means that individuals can evaluate the reliability of others and react adaptively. The tit-for-tat algorithm, however, requires ‘co-operate on first meeting’ to test out a stranger. Here lies the opportunity for freeriding. An evolutionary stable state ensues, however; co-operators outnumbering freeriders. The social brain hypothesis of human neocortical expansion is rooted in the evolutionary arms race driven by this conflict.
The evolved phoneme space (vocal tract) enables accent to vary within as little as 25 kilometres. Mutually incomprehensible languages arise through systematic syntactic and morphological drift. Groups, be they street gangs or academic schools, tailor language to provide self-definition. Only very young children can learn a second language without betraying the fact. And all languages turn out to be equally complex and expressive: a pidgin will develop into a fully structured creole in a generation. Dunbar (2004a) suggests that speech evolved to serve a function similar to primate grooming, which is used to form and maintain alliances. Specifically, Dunbar notes that the main use of language is for social cohesion, or gossip. Nettle (1999) verified this computationally for co-operator/freerider conflict and group cohesion attributes.
The language phenotype
I cannot overemphasise the fact that speech is a property of the human phenotype. Languages are tongues. They are spoken and heard, said and listened to. Speech is in nowise outwith the phenotype. The language system, used by a competent speaker, is capable of an infinite range of expression; yet every language-using member of genus Homo kept all their words in their heads. In addition, to the phoneme, morpheme and syntax, we must add the song of intonation; powerful enough to modify, negate even, the system-generated utterance. The phenotype is also a co-operator or freerider with a large brain to store memories; to recognise people, places and events; and to engage in complex social interactions. The bounding limitation on the language phenotype is brain capacity. Some archaic H sapiens had brains larger than ours: the Neanderthal brain was about 150ml greater. As a species of recent origin, no more than 150Kya old (Barrett et al 2002:17), it seems that language came to us from our precursors.
It took around 4500Mya from their accretion in space, in conditions peculiar to earth, for the elemental chemicals forged in the heart of dying stars to evolve the particular combination of DNA that describes our species. It took only about 150Kya for our species to unlock this secret. We are now tracing our own evolution from the joint ancestor with the great apes by DNA, as well as by associating human remains with tools. It is to the transmission of tool making knowledge that we need now turn.
There are two ways an organism can gain tool making capabilities. The first, and most common, is the extended phenotype (Dawkins 1999). Examples might be birds’ and bees’ nests. The extended phenotype is an inbuilt behaviour pattern that is transmitted genetically in the same way that mating behaviour or colouration is. The alternative is learned, culturally transmitted, behaviour. An example of this is chimpanzee termite fishing sticks, or nut cracking behaviour. Not all chimpanzee troupes indulge in either activity. The nuts cracked differ in troupes that crack them; and of the two troupes known to fish for termites, one makes and uses a long stick and the other a short one. Such knowledge may be transmitted between troupes, usually by an incoming female, but cultural spread is very slow. Such learned behaviour is unstable and may be lost.
The stone tool assemblages of H habilis and H erectus remained virtually unchanged for all of their existence: the bifacial pear-shaped ‘handaxe’ the signature of the latter. The more sophisticated assemblage of the Neanderthals was similarly conservative; changing only when, in Europe, they were in a position to learn from our species. This suggests that in these early language-using humans, tools were constructed to a series of inherited templates rather than learned patterns. Such conservatism in the making and materiel so contrasts with our flexibility that a ‘revolution’ is often proposed, cf. the industrial revolution, about 40Kya ago. But McBrearty and Brooks (2000) make a compelling case for a relatively long lead-in, preceding our arrival, to our style of tool making. The idea that I would like to pursue is based on chimpanzee termite fishing. The hyper-co-operation possible with reciprocal altruism makes cultural transmission a less risky and more adaptive strategy than inherited templates. I suggest that in early H sapiens inter-generational learning displaced inherited extended phenotype. There are two small pieces of evidence to support this idea. Firstly, the bigger Neanderthal brains occupied a somewhat differently shaped skull, larger at the back. Secondly, the Tasmanians, isolated for 10Kya from the Australian mainland, posessed a toolkit little greater than that of chimpanzees; had lost the art of making fire; and lacked seaworthy boats (Diamond 2002:252). They had lost learning relative to their mainland cousins.
Let’s look at child development post language acquisition. We see in primary school the genesis of drawing; drawing that is composed from the simple forms of point, line and arc: ‘tadpole’ figures that have more in common with engineering drawing than representational art. For most children realistic representation is not attainable, but the majority can combine the simple forms to note language in writing. Certainly, in this skill of drawing we find the foundation of our technology; but from whence come the elemental forms? Our best clue is to be found in the psychology of visual perception (Gregory 1998) where we learn that our visual nervous system appears to analyse the world into such elements. But we do more than draw. Craft’s example was cooking. Again we may perhaps look to the chimpanzee. It has been observed that when certain chimps are suffering from intestinal parasites, they will seek out a particular herb and chew it. This appears to deal with the problem. We, of course do the same. However, we also cook: we blend flavours, textures, and colours. How do we do this? Again, it seems that, somehow, we are able to access the system that analyses taste, touch, and smell. Consider the vocabulary of the wine buff: all indirect comparisons with fruits. The fiction I use to explain away this is a borrowed Maxwell’s (1877) information processing demon. I imagine that within our brain there is an element (adaptationists would say module) which somehow accesses the deconstructed features within the old pre-human processing system and makes them available for novel recombination.
This is what I mean by technicity. The best I can do to categorise it is to ammend the concept coined by Heidegger (Krell 1993): the capability of behaviourally modern humans to dominate the earth by controlling beings that are considered as objects. Technicity is the capability to access our internal processes, externalise them, and re-input them as novel information. It enables us to propose a different reality. This is where creativity comes from. But there is a catch: prerequisite for creativity is the hyper-co-operation that evolved before our species, which itself turns upon language.
Were we to try and tease out just what it is about our species that has made us more successful, to date, than all our predecessors, we had better look at ourselves today than seek for the answer in the spark that ignited our progress. Clearly, we have extended our phenotype extensively. We have done so by taking the risky path of learning rather than evolving a better internal template. This can be successful only because we previously evolved hyper-co-operation and the language adaptation necessary both to sustain it and mount a defence against freeriders. Processes that precipitate phase transition generate unforeseeable consequences. The capacity to draw has enabled us to store our memory externally. By not doing everything in our heads, as the Neanderthals did, we run the risk of nakedness when technology fails: but it appears to have paid off. Getting things off our minds seems to have become institutionally necessary as agricultural city civilisations arose. Writing it down and drawing it out preserves what was, after memory fades or flood occludes. Writing has served to bring language under a degree of control. Symbol systems do not suffer from phonetic and syntactic drift; they can disambiguate; and they place language in the relation of object to us, whence we may examine it objectively. Consequent upon writing was philosophy, followed by logic. This led, in the nineteenth and twentieth centuries, to the confluence of mathematics and language; from whence computing.
Writing, logic, and mathematics are products of technicity, so it is not surprising that technologies other than pen and paper might represent what that medium does, but more powerfully. This, the stored program digital computer does. I now write using a tool that can check my spelling and can get sums right for me. But its challenge to the intellect is deep. The mathematical Entscheidungsproblem (uncertainty) was resolved contemporaneously by Church and Turing. The former used a symbolic calculus; the latter envisaged a reading and writing machine working to an algorithm (Berlinski 2000). That is, mentation and mechanism are the same. The adaptation that makes our species successful is a capability to externalise data and processes that earlier species were constrained to keep within their heads. Herein we see the rationale for education.
The Big C
The notion of creativity (technicity) developed above is clearly the source of Craft’s little c. It may be argued that Big C is just one end of the normal distribution that we associate with so many human abilities. I suspect matters are more complex. For a start we are talking about a recently evolved capability, the population distribution of which may not be Gaussian and, given the character of our hyper-co-operation, may be represented in genetically carried talents. And there is also the primate underlay: social systems based on alliance, male-female relationships, and hierarchy. An Emin, like Michelangelo, may thrive because of what they offer to a high-status sponsor. I would also suggest that some aspects of Big C function as display cf. a peacock’s tail; and may not be of post-reproductive benefit to their possessor (Nettle 2001). Others: a Stephenson, Einstein, (Turing?) appear more embedded in the communal enterprise.
Gossip, gadgets and LCC
I now take a risk by asserting that all creativity entails technology, i.e. is the product of technicity, and that language is, of itself, uncreative. Let me, therefore, be quite precise about creativity: creativity is that capability which has enabled our species to progress from simple toolkits and lifestyles to our present designer society. Recall that our uncreative precursors had a language physiology that enabled them to make as great a range of sounds and tones as languages employ today. So, there is no reason to deny them the capacity to vary the order of actor, action, and acted upon, as we do. The why of language evolution, freerider detection and group cohesion, would imply as much. Language can, therefore, only be creative in the sense of spin, persuasion, or economy with the truth. Yet, creativity is physical; producing novelty and innovation within the constraints of the physical laws that technicity has enabled us to determine. Energy, work and entropy constrain creativity: ideas are only capable of realisation when the resource is to hand (Diamond 1998). Technology validates science. This is why our technological history has a long tail: each generation builds on its ancestors. Our own rapid technological progress is consequent upon a huge increase in energy conversion, itself build upon a post-Columbian transatlantic transfer of resource.
Language and technicity
This brings us to the difficult relationship between language and technicity. Nettle and Romaine (2000) report a discussion in a Papua New Guinea village whose inhabitants decided to mark their identity within their language community by changing the word they used to say “no.” In a modern technological society the language problem is very different: there is the need to find words through which to discuss novelty: quarks of varied colour and flavour. Neologisms being frowned upon, old words with a helpful connotation are bent to denote new entities and processes. The lexicon remains within bounds whilst the knowledge-base expands. This process is psychologically available because it is a feature of language, where ‘bough’ is a homophone that context alone disambiguates, in speech. The process is aided by an uncertain interaction of syntax and semantics, illustrated by Chomsky (1965:149) in the example: “colourless green ideas sleep furiously” vs. “revolutionary ideas appear infrequently.” Language is an infinitely extensible system capable of self reference: arguments may be formulated wholly within language, the lexicon being extended by internal re-definition: a pitfall for the academic unknown to the mechanic. (Craft is in danger of doing this when she uses lexical items to categorise LCC.) Technology has helped bring this feature under a degree of control: writing, systems for noting certain informative features of speech on a surface using a stylus, has made language available for inspection: we can check it for internal consistency. We deploy other technologies to test our assertions against reality: spectacle lenses, suitably mounted, enabled Galileo to negate the assertion that the sun rotates around the earth. Other technologies, and the information they provide, has enabled me to challenge the assertion by Diamond (2002:125) that: “Without language we could never have conceived and built Chartres Cathedral.”; and propose that language only provided the hyper-co-operation necessary to realise the enterprise. Conception and construction required our peculiar successor adaptation: technicity.
My analysis of creativity from an evolutionary perspective has brought into focus the divide that Snow (1993) called “The Two Cultures.” This divide, which we see reflected in the academic/vocational schism in education and in verbal/non-verbal psychometrics, finds support in evidence from the emerging field of evolutionary psychology. We know that language develops before children go to school and that the period from reception to KS2 sees the emergence and blossoming of children’s graphic skills (Anning and Ring 2004). This is the stage in which the co-option of speech by technicity begins. Language-development, upon which educationalists place such emphasis, is in actuality the process of turning the adaptation of gossip into a means of concept communication. The status we afford the means of objectification of the concept carrying aspects of speech – literacy – would tend to support this view.
Technicity is constructive in character rather than imitative. This places it at odds with the meme, cf. gene, proposed by Dawkins (1989:192) and elaborated by Blackmore (2000), an imitative agent of cultural transmission. Although traditional psychology, founded in what Cosmides and Tooby (Barkow et al 1992) call the ‘standard social science model,’ has much to say about perception, language, psychiatric, and social behaviour; the texts are silent on the ‘how’ of technology. The ‘externalisation of the internal’ proposed for technicity does resonate, somewhat, with constructivist ideas founded in the works of Vigotsky (1964) and Piaget (1971) and the constructionism of Papert (1994). The notion of technicity as an adaptation, however, goes beyond the social brain of evolutionary psychology Dunbar (2004b). Map and compass have we none; but there are a few pointers: the nervous system evolved in an environment the phenotype expects, and hence anticipates; a Turing machine can reprogram itself; a heat engine performs work and thereby alters entropy distribution in an open system. I.e. both creativity and progress are immanent in our brains. If we allow access to the features that our brains analyse sense data into and permit their output, we are enabled both store them externally and to re-combine them in novel ways. This would make creative progress a system feature. And they suggest areas for research on creativity.
Implications for education arise. The stage from birth to year one is largely one of extra-uterine maturation. Once the brain has stopped growing and started wiring itself up there are distinct maturational and learning phases. Using a very coarse scale, 1-7 years covers the phase during which access to our internal systems and the output of features develops, alongside the process of co-opting language to talk about them. Years 7-14 see the development of skills in combining this learning into individual and social capability sets. The bridge to adulthood, 14-21, is the age of differentiation. The critical years for building the foundations of creativity are in the 1-7 range, where children need a) to explore and express their perceptions and to learn how to analyse them, whilst b) developing language skills beyond gossip to communicate about them. This could suggest that the academic emphasis on social language in this phase may not be well directed: a focus on poly-technique and associated specialised vocabulary might be preferable. This would imply a shift in focus from the narrative of the story to sensitivity to the multiple meanings of words and their relationship to sensation: the model being language use by wine buffs, perfumers, pigment manufacturers, and the descriptors of tools of trades. But maybe analysis would show that we do this already.
There is also a need, long identified by Papert (1980), to re-evaluate classroom ICT. If my analysis is correct, the computer is the next phase of our external memory system, our extended phenotype. It seems perverse to use ICT to support learning based on the precursor external memory medium, text, and to teach the skills associated therewith. Surely it is better to make use of the new tools to avoid what Pascal referred to as the tedium of computation, and the tribulations of hand-writing and spelling. No longer is literacy prerequisite for learning. Computers can turn speech into text and vice-versa.
Much of what I have proposed is speculative. The fossil evidence is not unequivocal. Nevertheless, I hold that the proposal I float offers far more promising foundation for further research than previous proposals. My prerequisite is the evolution of language and hyper-co-operation. Thereafter, and specific to our species, comes the technicity adaptation. This is the adaptation that enables us to modify the world. My hypothesis directs us to look objectively at what children do as technicity develops in childhood. It points to questions about the manner in which technicity interacts with language. I believe that further research will demonstrate that language is intrinsically uncreative and that all creativity flows from technicity, which co-opts language to communicate.
Anning A & Ring K 2004 Making Sense of Children’s Drawings Maidenhead: Open University Press
Barkow JH Cosmides L Tooby J eds The Adapted Mind Oxford: Oxford University Press
Barrett L Dunbar R Lycett J 2002 Human Evolutionary Psychology Basingstoke: Palgrave
Berlinski D 2000 The Advent of the Algorithm New York: Harcourt