Derek Hodgson

Abstract. Lower and Middle Palaeolithic abstract marks have recently become more available in the archaeological record. Theories concerning the significance of such marks have invariably concentrated upon an intended referentiality, thereby ignoring the possibility of an analysis based upon evolutionary criteria predicated on a gradual, cumulative scenario which does not necessarily have to depend on symbolic explanations. This paper seeks to redress the balance by determining how early mark-making can be sufficiently explained through evolutionary mechanisms in conjunction with how the brain and visual cortex process information. In approaching the origins of art from this perspective, parallels will be drawn with how the brain deals with concomitant types of information, e.g. language, which can give important clues as to why and how mark-making might have developed.

KEYWORDS: Palaeolithic art – Evolutionary psychology – Information processing – PerceptionIntroduction

If, as evolutionary theory suggests, all physical and behavioural change is due to an adaptation to shifting circumstances it seems unlikely that ‘art’ (or what can be termed ‘proto-art’ when referring to early mark-making) can be dissociated from those forces by which the very fabric of life has been determined. The evolutionary factors, which have shaped the brain and complex behaviour of humans over millions of years, are, therefore, unlikely to be completely overturned by cultural developments arising out of a relatively short period. In this respect evolutionary psychology, premised on the recent findings of cognitive science and the brain as a means of processing information, has much to offer by recourse to a nativist-oriented Integrated Causal Model as opposed to the predominating relativist and contextually based Standard Social Science Model. Crucially, an Integrated Causal Model would be inclined to regard the acquisition of mechanisms of the brain as both specialised and universal, having gradually evolved in response to evolutionary principles (Tooby and Cosmides 1992: 23, 100). As Dawkins (1991: 90) has indicated, small changes over a long period can explain ‘an enormous range of things that would be otherwise inexplicable’. One of the ideas I hope to clarify in this paper is how art originated and developed as part of a similar, gradual, evolutionary dynamic. It will also be argued that the origin of art was inherently bound to the fundamentals of natural selection (Dennet 1995: 144) and an evolving visual system as Gombrich (1984: 1-4, 7) has implied.
With a greater insight, in recent years, of the functioning and structure of peripheral and central mechanisms of perception has come, likewise, a greater understanding of how the visual system has been forged by the exigencies of evolution (Gregory 1990). It has been suggested that over half the human cortex is involved in vision (Gregory 1990: 7-8) and the cortex, in toto, grew, initially, out of the need to process visual information (Gregory 1970: 12-13) — a case of form (graphic marks) following function (the neurophysical structure of the visual cortex). Consequently it is now possible, by graphic analysis of the Lower Palaeolithic (L. P.) and Middle Palaeolithic (M. P.) mark-making (and by implication later Upper Palaeolithic [U. P.] representational forms), to make a comparative analysis linking such artistic phases to the stages through which the early human visual system, shaped by the demands of survival, had already passed before the onset of mark-making, i.e. from a reactive to a more pro-active engagement with the immediate environment. This has opened the way to a possible investigation of how the earliest manifestation of ‘art’ can be understood from a position free of the usual constraints contained in contextual or aesthetic approaches. Indeed, the only realistic way of hoping to understand the earliest marks produced by hominids, before the supposed 40 000 BP threshold, is through a perceptual and Darwinian analysis and, in this respect, such marks should be regarded as vitally important for the very reason that they are the earliest, and may provide valuable insights, not only about the early humans responsible for their production, but also the structure of the visual system and the origins of art. By rephrasing the question ‘What does early Palaeolithic art tell us about the socio/economic structure of early humans?’ to one of ‘What do the first graphic forms used by early humans tell us about perceptual evolution?’ might a new understanding be possible.
An investigation of this order will, of necessity, be obliged to take a broader sweep of history than is usual while considering, at the same time, how art must be viewed both as a single trend and part of a multifaceted tendency. In taking this position the genesis of art has to be seen from the perspective of evolutionary time, that is, slowly unfolding, with many false starts, retrogressions, important developmental accomplishments and long plateaus of achievement, reinforcement or recapitulation but still continuing to evince a discernible trend. Regarded in this way the L. P., and later M. P., will be considered as pre-representational, or proto-art phases, which led, eventually, to the representation of the U. P.
In order to examine how evolutionary imperatives have moulded art from this standpoint, Part I of this paper considers the general determining evolutionary and environmental parameters, which will provide a ground-ing for Part II, concerning a more detailed analysis of relevant examples of early art with regard to perceptual factors. This will embrace theories of perception, language and art, anthropology, ethology and child development, as well as precise examples of pre-Historic art and other relevant sources, such as infant drawing, shamanism, phosphene theory and neurophysiology. Part III will attempt to bring the findings of the first two parts together to provide a comprehensive model tracing the development of early art in relation to commensurate activities. The conclusion, it is hoped, will provide some insight as to the origins of art within the course of evolutionary history and illustrate how art has figured within the context of other important traditions arising during and subsequent to pre-History. Although the Integrated Causal Model brokered by evolutionary psychology is acknowledged as having provided a useful framework by which these various disciplines can be appropriated in relation to early art, it will be necessary, here, to accommodate the model according to the ensuing analysis.PART I

Defining art: order, repetition and information processing
Art, by its very nature, is not, as a subject, amenable to any straightforward delimitation. For present purposes a working definition will be proposed which seeks to encompass the majority of Western and non-Western art in accordance with a universal component implied by an evolutionary analysis (Tooby and Cosmides 1992: 78-9) but, at the same time, is not wholly incompatible with cultural interpretations. To assess how art can be considered from this perspective it will be necessary to begin with the theoretical context by which graphic ‘primitives’ and simple repetitive lines can be understood, and how particular formative perceptual processes might be related to such elemental configurations.
By definition, art is inherently bound to the processes of visual perception as Gombrich (1977) has suggested. Gombrich, however, regards the perception of novelty in nature, by way of ‘top-down’ hypothesis testing (which determines the artistic depiction of pattern), as a primary predisposing factor (Gombrich 1984: 7-10, 107-8). He bases this view on Gregory’s theory of perception (see Gregory 1973: 74, 79-80 and below). Contrary to Gombrich, the definition to be elaborated in this paper defines the capacity to depict pattern as dependent on the ability to transcribe or represent order by way of continuity (what Gombrich describes as the ‘etc. principle’; 1984: 99) and, therefore, would antecede the aesthetic pleasure of novelty as a critical feature. More simply, pattern has to be seen, analysed and constructed (as determined by a precept or rule) before its more novel attributes can be exploited. Pertinently, Mach suggests, ‘A rule always presupposes a repetition. Repetitions, therefore, will probably be found to play some important part in the production of agreeable effects’ (Mach 1943: 91), and ‘Any figure, no matter how crude or poor, if several times repeated, with repetitions placed in line, will produce a tolerable frieze. Also the pleasant effect of symmetry is due to the repetition of sensations’ (Mach 1943: 92) (my emphasis). We would hardly have evolved in the first instance from the ‘blooming, buzzing confusion’ if at every moment it was necessary to entertain a myriad of hypothesis as to what was transpiring in the world! In reality we tend only to question perceptual phenomena, in the form of various possible hypotheses, in the event of a doubt arising which might threaten the prevailing, tried and tested norms (rule-governed system). Zeki (1993: 355) summarises the point admirably: ‘The brain strives to acquire a knowledge about the permanent, invariant properties of objects and surfaces in our visual world’, which is only possible ‘if it is able to discard the continually changing information reaching it from the visual environment’. Although Gombrich alludes to the concept of order as repetition (1984: 4-5) it is consistently regarded as ‘redundant’ (1984: 8-9) and of passing importance. Repetitive mark-making would, however, in itself, have constituted a novelty in the face of the flux of nature, a preliminary stage in the representation of form.
In short, the reason repeated lines and geometric shapes are so appealing is not so much because of rarity in nature, but more because they are already an integral feature of the brain. Continuity, or repetition, would have signified what was safe, secure and ‘understood’, whereas change (or novelty) represented threat.
Early humans, through a process of trial and error and successive approximation were, in extracting pattern from the world, unwittingly engaged in the appropriation of a technique which would become the basis for encoding thought in a tangible, stored and retrievable form. In the material realisation of this striving, art becomes a means of embodying and preserving information relatively immune to the ravages of time, and as Gombrich (1991: 65) himself admits, a way of stopping, or arresting, the incoming perceptual stream. As a perceptual phenomenon art is, then, an attempt to render permanent and tangible that which was formally intangible and fleeting, the seeking of order in the midst of disorder, the expression of the sense of pattern, harmony and symmetry synthesised from the immediate, ambient confusion. This definition need not exclude cultural interpretations, but would tend to regard such accounts as more relevant to the art of the U. P., as a precocious vocal language ability might well apportion meaning to the represented forms prevalent during this period. Such interpretations, however, only begin to become credible either during, or after the representational stage, because it is only at this point that two-dimensional picture-making becomes a powerful vehicle for the transmission of ideas and the prevailing cultural norms — representation would have been contingent upon the ‘learning’ of a graphic vocabulary before a more advanced, vocal competence was able to attribute any coherent meaning. ‘Learning’ in the sense used here has a particular significance for reasons which will be outlined below.

Previous explanations of Palaeolithic art
Notwithstanding the difficulties concerning definitions of what is to constitute ‘art’, and the ambiguities thrown up by relativism, pre-Historic art brings with it countless problems to do with its remoteness in time and gaps in the archaeological record. Recent discoveries of L. P. art, together with archaeological finds throughout the world have, however, begun to reveal a more complete but still far from adequate description of the formative art produced by early humans.
With more reliable dating techniques the genesis of ‘art’ has been consigned further to the distant past (Bahn 1998: 162; Bednarik 1992, 1995a). In addition, a growing catalogue of diverse marks and various artefacts from the period before 40 000 BP has served only to increase the controversy as to their possible significance (Bahn 1998: 164; Bednarik 1992). Yet, the later cave paintings of the U. P., such as Lascaux, Pech Merle, Cosquer and Altamira continue to be accepted as indicative of a sudden upsurge of artistic activity — part of the cultural explosion after the 40 000 BP M./U. P. boundary (White 1982; White 1989). Theories have, consequently, been fabricated in order to explain this upsurge (e.g. Mithen 1996; White 1992), leading to a spate of hypothesis which have attempted to account, more precisely, for the images of the U. P., e.g. symbolic (Marshack 1976; White 1992), shamanistic (Lewis-Williams and Dowson 1988), as evidence of hunting magic (Breuil 1952) or, as suggested by Leakey (1994: 114) and Bahn and Vertut (1997: 213), according to whatever has happened to be the prevailing hypothesis. Some of these socio/cultural explanations may well be compelling and applicable to rock art from the time of the so-called cultural explosion, i.e. when discernible representational art had become an important feature, but questions have been raised as to the relevance of such theories to mark-making before the advent of representation (Davis 1986).
In focusing on possible explanations for the sophisticated naturalism of U. P. art, commentators, by resorting to a sudden onset theory, have tended to misunderstand or discount earlier geometric marks (Chase and Dibble 1987; Gamble, 1993: 168), usually as a function of natural events or part of the butchering or crafting of an object, as Bahn (1998: 86) and Bednarik (1992, 1995a), amongst others, have noted. Despite this failing some serious attempts have been made to explain the abstract geometric marks produced both before and concurrent with representational art, including decorative function (Conkey 1980, 1981), sexual (Leroi-Gourhan 1958), notational/calendric (Marshack 1972), as a depiction of natural or fossil objects (Feliks 1998a), etc. Ultimately, a definitive explanation regarding abstract motifs of this kind will continue to remain elusive, not least because concrete evidence to substantiate any given theory is unlikely to become available.

The gradualist approach
Davis (1986) is of the opinion that the image-making of early humans is a ‘predictable adaptation which should be coherently situated in the overall trajectory of hominid evolution’, which can be ‘derived logically from simple and archaic perceptual and cognitive processes’.
This view presupposes a gradualist scenario based upon the development of perception as manifesting the appearance of an increasingly diverse and sophisticated use of mark-making similar to that suggested by Bednarik (1992), who regards the idea of a ‘cultural explosion’ at the beginning of the European U. P. as hindering the acceptance of a more protracted process (1995a: 606). Bahn (1998: 186) sees this division as artificial and based upon the lack of evidence in the archaeological record because of the greater time span involved.
The position taken, in this paper, in regard to the evolution of art, is essentially gradualist, in the sense that there was no mysterious, cognitive or aesthetic cultural explosion at the M./U. P. transition, receptive and spoken language, graphic schema and tool-making having incrementally increased over a protracted period at different rates. The perceived proliferation after the 40 000-year boundary is, accordingly, considered as artificially exaggerated due to the more recent rock art having survived the time scales involved (Bahn 1998: 87)
Once the trick of representation had finally been mastered there would probably have been an increased tendency to use, and value, the technique as a means of embodying information. The commonly accepted ‘upsurge’ could well have been but one, albeit significant, event in a string of corresponding small, cumulative events. The assumed upsurge can then be regarded as no more than to do with a synthesising or fusing of particular ways of processing information, i.e. visual depiction and spoken language, as the continuation of an integrative trend which had been taking place throughout hominid history.
If art is defined as a means of storing and transferring information (Bahn 1998: xxiii) within and between generations, its supposed widespread appearance from 40 000 years BP onwards can be explained more readily by an increasing ability of humans to disseminate such information materially with ever more efficiency and economy — a development which had probably taken some hundreds of thousands of years to perfect ‘climaxing’ in the full, representational images of rock art during the U. P. Information is defined here as a continuum from the non-referential to the fully symbolic, where the representational art of the U. P. can be seen as occupying the middle ground. The abstract schema of L./M. P. art can be regarded as information driven more directly by way of evolutionary mechanisms, in contrast to the subsequent, more culturally derived representational art, as part of a progressive, exponential process whereby the co-ordinated integration of abstract schema into representational forms led to less time-specific constraints on communicative expertise — an expertise which was to culminate in symbolic depiction.
The recently discovered cave paintings of Chauvet have somewhat confused the issue in being dated to more than 30 000 BP, while at the same time being similar in sophistication to the much later paintings of Lascaux and Altamira (c. 16 000 BP). Whitney Davis’s thesis as a developmental one is, therefore, seen to fail as his premise is based upon the latter being the consummation of a historically linear process on which Aurignacian images (34 000-28 000 years BP) are deemed the earliest. Halverson’s case (1992) fails for similar reasons, as do other explanations based on the evolution of elements of style limited to U. P. art (Stringer and Mckie 1996).
Attention should rather have been focused on the abstract schema which has come to light before the Aurignacian period, i.e. before the M./U. P. transition and the appearance of fully-fledged representational art. The outline drawings and abstract geometric forms and patterns, still prevalent during the U. P., it will be shown, can be accommodated by taking this longer view and may also help to shed further light on why Davis’s and Halverson’s accounts have failed to account for the facts.

The relationship of art and language
The decisive question arises: if art is not part of a developmental process whereby abstract schema gravitates towards representational form, as White maintains (White 1992: 539), why is it representation never precedes the more ancient, abstract marks or, conversely, why is it that this schema actually predates representation? (Bahn 1998: 223). As Bednarik (1984) has indicated, ‘The more frequently phosphene types occur in rock art, the more archaic it is. In the oldest traditions they may dominate to the point of exclusiveness’, and as Bahn (1998: xv) emphasises, ‘Apparently non-figurative art-motifs which convey nothing to our eyes other than the patterning, has existed from the beginning. Indeed it often dominated the art of the Palaeolithic period and its study is one of the long-neglected challenges of archaeology’.
The willingness to attribute symbolic intent to art before the representation of the U. P. has caused much confusion with a continuum of theories, from those that would regard such marks as already fully symbolic to those, at the opposite extreme, taking a more asymbolic position, usually from a neurophysiological standpoint, with all shades of opinion in between (Byers 1994). This confusion may reside in the fact that symbolic ability was developing differentially across various cognitive domains — speech racing ahead of activities such as mark-making, the latter of which was still striving towards symbolic realisation but nevertheless remained constrained by functional evolutionary processes to do with the visual cortex. Moreover, simply to regard early geometric marks as somehow mysteriously symbolic ignores the fact that before symbolic representation could come about a stage of pictorial representation had, essentially, to be transcended. Hominids before the U. P. did, it would appear, have a capacity for symbolic behaviour — evidence suggests this ability was already present in speech — but that this was not yet present in graphic depiction.
Why should early mark-making be devoid of meaning in contrast to spoken language at this time? The answer may reside in the fact that as the development of speech would have been contingent on a separately functioning cortical mechanism to that of artistic competence (Gazzaniga 1970), it is possible, from the outset, the two abilities developed independently. It is logical for vocalisation to have evolved first as this function would most likely have been primed earlier by social demands, or the need for communication in the making and using of tools.
Gombrich (1977) makes the point that art evolves through a process whereby ‘making comes before matching’, in other words, the previously learned ways of depiction in the plastic arts are generally improved upon by later practitioners — a process he sees as equally applicable to the realisation of representation in Palaeolithic art after a protracted period of image making (Gombrich 1977: 91-3). Through this means an increasingly elaborate array of graphic devices came to construct the illusion of three-dimensions on a two-dimensional surface as a capacity with discrete rules of engagement to that of speech. This segregation becomes more apparent when it is realised that it had taken modern humans some ten thousand years, from the Mesolithic period to the Renaissance, to discover the laws of linear perspective despite a sophisticated language competence in vocal and written form together with an elaborate conceptual ability. How much more formidable would it have, thereby, turned out to be for early humans to achieve iconic representation on a two-dimensional surface? Arguably a greater achievement than the discovery of perspective! If Renaissance man had found it so difficult to master such laws with a sophisticated language facility long intact it, then, follows that for early humans the trick of representation would have been equally, if not more, problematic and possible only through an appeal to the modus operandi of graphic depiction.

Speech, art, brain size and information
The suggestion that the early geometric art of the L./M. P. is mediated by innately defined mechanisms can be regarded as analogous to Chomsky’s theory of language whereby a deep, hard-wired structure is deemed to underlie the more overt, surface structure — which ultimately dictates the language a child will speak (Chomsky 1957: 234; similarly, most linguists would accept that an infant’s spontaneous ‘meaningless’ babble constitutes a crucial part of later language development — Weir 1962; Chukovsky 1968; Schwartz 1980; Bee 1992: 298-9). Here, Pinker (1994: 223) has proposed not only that some important aspects of speech vocalisation are contingent on simple mechanisms which antedate more sophisticated, culturally-dependent elements, such as grammar, but, in addition, evolution, having established the basic computational units of language as innate, may have, finally, seen no need to replace every bit of learned information with innate wiring. From the point of view of proto-art this can be translated into innately-conditioned, abstract marks produced during the L.P. as ‘hard-wired’, in the sense that they mirror the functioning of the visual cortex (in the same way as the first utterances of vocalisation are hard-wired), while the representational art of later cultural periods (corres-ponding to specific grammatical rules of language) can be interpreted as the ‘surface structure’.
Although there may appear to be certain parallels common to the functioning of art and speech this does not imply that, initially, there was any synthesis of modal functions, as a slowly evolving account would imply that separate channels merge incrementally — given favourable circumstances and the particular level of development inherent to each modality. Surprisingly, Chomsky (1988) does not subscribe to a gradualist evolutionary model of speech and, in this respect, the proposition of Davidson and Noble (1989) that language came about through the agency of artistic expression and, hence, art had to predate language — although sympathetic to Chomsky’s ‘threshold’ approach — cannot be supported here, as language is also taken to be incremental. The step-by-step growth in brain size over the past two million years (which strengthens the gradualist case) can, according to Leakey (1994: 128), be related to a language facility, and Pinker (1994) has amassed a copious amount of evidence to suggest that spoken language is, indeed, contingent on the micro-circuitry of the brain moulded by natural selection over a considerable time span. Anatomical evidence clearly supports this view (Holloway 1983; Laitman 1984; Deacon 1989; Kay et al. 1998). In addition, Deacon (1997) has shown how human language is different by an order of magnitude to that practised by other species (whales, monkeys etc.), i.e. abstract symbolic referentiality as opposed to innately defined iconic Y indexical referentiality — which reinforces the case that brain growth in humans was both contingent on language and such language was originally built on the foundations of the iconic-indexical axis.
Taking these studies together it seems clear that brain growth would almost certainly have been mediated by the demands of speech (Jerison 1991). Wills (1993) has suggested an evolutionary-mediated, culture/brain, feedback loop to explain the accelerated growth, language playing the key role — although this expansion would, ultimately, have come to an end due to a naturally-imposed limit. If, therefore, the human brain, as a metabolically expensive organ, has not increased in size for the past 500 000 to 350 000 years (Rose 1976: 172; Deacon 1997: 343) (to do with the demand for oxygen and the neonate’s head size leading to birth difficulties etc.), one way of circumventing corporeal restrictions would have been to realise, extrinsic to the brain, those traits which had already conferred selective evolutionary advantage, i.e. spoken language and the ability to perceive structural form. Speech, having emerged first as a capacity of the left brain hemisphere, continued to be a circumscribed, transitory capability tangibly unable to act as a reliable means of storing and retrieving information. In order to by-pass physical constraints a means of representing the existing processes of visual perception, as embodied in the visual cortex, would have been a prerequisite (Davis 1986: 196). Thus, the need for continued brain growth would have been circumvented through the ‘discovery’ of a means of holding ‘information’ in a more reliably stable and relatively permanent form (Gregory 1970: 148) providing a kind of surrogate cortex with unlimited capacity. As Gamble (1993: 168) has indicated, art can act as a sign to chunk enormous amounts of information into a manageable proportion, hence, in a very practical sense, expanding memory.
In much the same way as extra modules of the brain had developed by over-laying the more ancient, deeper, automatic, reflex-oriented levels, so extra modular systems continued to develop as an increasingly sophisticated way of dealing with incoming stimuli (Jackson 1932; see Allman 1999 for an updated and more sophisticated account of this process) and, being modular, infers separate, functional development of brain structure (Fodor 1983; Hirschfeld and Gelman 1994) and related behavioural components (Dean 1982; see Pinker 1994: 314-17, who addresses any reservations, such as Damasio and Damasio’s [1992], concerning this model). This is confirmed by the asymmetrical nature of the cortex, where the left side is broadly devoted to language, the right more to artistic expression and visual perception (Solso 1993; Kolb and Whishaw 1996; Pylshyn 1999; Hollinworth and Henderson 1999). Speech, as a separate, modular, parallel increment, would have constituted the first stage in a wider processing of information one step removed from the brain. The initial scratch marks made by early humans can be interpreted as an extension of this quest, still mediated, but not wholly bound to the brain, i.e. the beginning of a search for order and structure as a reflection of the neurological patterns on which the human cortex had previously depended for survival in the face of a hostile world. As with language (Pinker 1994), an emerging depictive function can, then, be seen as having arisen out of a similar adaptive striving.
To summarise, as evidence from brain morphology seems to suggest that speech developed significantly ahead of, and separate to, depictive modes, it appears highly unlikely that symbolic meaning would have been accorded to abstract, graphic schema until this schema had evolved into representational forms during the U. P. Speech as a first decisive step towards a pro-active concern with information processing would have had inherent practical limitations, to do with its temporal or transitory nature, further restricted by a naturally imposed ceiling on brain expansion. Only ‘art’ as an important alternative mode was able to take up, at this point, the reins in the form of an aide memoire par excellence.

Evolutionary mechanisms in early art: play, curiosity, arousal and a two-construct aesthetic factor
The question remains, however, what was the precise nature of the underlying mechanism which had led Homo to first begin ‘experimenting’ with form and line if meaning or symbol initially played no part? Evolution, as Dawkins (1991: 72) has explained, has no end-purpose or distant goals in mind, rather it works on the basis of anything that would improve the chance of individual survival.
For early humans, time devoted to mark-making is apportioned at the expense of some other important preoccupation, e.g. the gathering of food, provision of shelter, mating rituals etc., which suggests art must have had some decisive value in the game of survival (Conkey 1978). But this begs the question of how art can be translated into a motivational drive determining mark-making in hominids. Only by looking at the principle of how ontogeny recapitulates phylogeny can some clues be gained (for a more recent, updated version of this principle see McNamara 1999).
Play has been proposed as being: biologically important (Groos — see Millar 1968: 19-20), a risk-free learning strategy (Boulton and Smith 1992), exploratory behaviour as a function of increased arousal and energy levels (Loizos 1967: 193-204; Piaget 1963), and independent of any external reward (Harlow et al. 1950; Morris 1967). This activity, in infants (and by implication early humans), would have been elicited in the same way as Berlyne (1970, 1971) has proposed in demonstrating how art elevates pleasure by acting on arousal. Arousal is induced in two ways, moderately through the use of familiar, regular patterns, such as ones high in symmetry, or sharply through the use of patterns high in novelty, surprise, or complexity which cannot be immediately assimilated. Further evidence for an ‘aesthetic’ factor of this kind and at this level comes from Götz et al. (1979) and Iwawaki et al. (1979) where subjects, asked to choose the better Gestalt from a set of two designs, preferred the more regular and symmetrical. The high degree of consistency found in this study proved to be independent of sex, intelligence, personality and culture. Similarly, young children have been found to prefer those elementary geometric figures manifesting greater symmetry and simplicity (Brody 1970; Brighouse 1939). Kagan (1970) has, in addition, discovered that once infants have become familiar with a particular pattern they go on to prefer those which are moderately different rather than the same or radically different, suggesting an incremental process congruent with a gradualist, evolutionary scenario.
Such studies confirm the existence of a two-factor construct of artistic arousal, where simple lines can be regarded as a forerunner to more complex features. The development of graphic, depictive strategies in infants and early humans is, thereby, construed as a corollary of this two-factor perceptual process to which play and an insatiable curiosity are closely linked. For early humans and infants, simple pattern, although elemental, would have represented, at first, a kind of novelty which, in turn, became an accepted schema on which later more complex graphic elements came to be based. For children, the playing with sounds in learning the rudiments of speech or the making of the first marks and scribbles on the way to representational drawing is a recapitulation of how early humans first began to ‘play’ (by recourse to an incorrigible curiosity) with sounds and graphic primitives as a species-specific means of maximising the chances of survival by inventing communicative and perceptually-mediated analytical strategies to contain environmental threat through the control and manipulation of information.
This may well have been instigated in hominids by the biological need to seek out different or alternative kinds of stimulation to the norm as more time became available due to greater tool efficiency (see White 1959, and Berlyne 1971, for an explanation of the effects of inactivity on arousal). Appositely, Getzels and Csikszentmihalyi (1976) concluded that the artist, as a problem-solver, epitomises the stimulus-seeking individual — in this guise an ‘artist’ would have been a valuable asset to any hunter-gatherer community.
The important question of why early humans should have resorted to producing crude scratches and lines cannot be fully answered by simply implicating an incorrigible curiosity, although this may be part of the answer. The question can only be addressed when it is realised that it was a curiosity driven by the fact that the repetition of a simple line would have been evoking that part of the visual cortex (in much the same way as Mach has suggested), which had previously allowed Homo to see the continuity and pattern in the world at large and had led to the greater survival of those individuals capable of this aspect of perception. Perhaps the gene cluster responsible for particular aspects of the structure of the visual cortex, due to subtle, individual modifications and mutations, had led, over time, to the selection of more elaborate ways for the analysis of pattern. Stimulation of this function through mark-making, in addition to being a means of sharpening and honing important perceptual skills through better eye-hand co-ordination and visual memory, may also have been a way of increasing awareness to environmental signals with accordingly greater survival rates for those so disposed.


Gestalt theory, infants’ early drawings and Palaeolithic abstract marks
Gestalt theory comes half-way between Direct Registration (Gibson 1950, 1979) and Constructivist theory (Helmholtz 1867; Gregory 1970: 10-14). Direct Registration and Constructivist approaches tend to emphasise a particular aspect of the perceptual process, the sufficiency of the information already present in the visual world (bottom-up) and the importance of higher-order interpretation of ambiguous perceptual stimuli (top down) respectively. Each theory, in attempting to account for the total range of visual experience has, hence, sought to exclude possible alternative explanations. In reality, all three approaches probably have something to say about the perceptual processes pertaining (Neisser 1976: 53; Treisman 1986), the disagreement simply reflecting the dynamic which exists between bottom-up and top-down stances. While agreeing that optical information is transformed in some way but at the same time denying that raw visual information is adequate to meet all perceptual demands, the Gestalt approach, as midway between the two positions, seeks to explain perceptual phenomena by recourse to such factors as the law of good continuation, simplicity, closure etc. In the present context, Gestalt theory is regarded as more applicable to how the neurones in the visual cortex are organised as an important stage post in the processing of one component of incoming visual stimuli rather than as a sufficient account of all aspects — more directly, as an expression of how patterns emerge in the neuronal system at the pre-attentive level of the striate cortex. Primitive marks then become the starting point for the assignment of form according to Gestalt principles in the same way that the combining of such primitives into basic form may well take place at the level of the striate cortex (see below).
Kellogg (1979: 14-15) states: ‘The fact that children’s art can be categorised for similarities of Gestalt formations means that they are reflections of mental and physical processes common to the species’. This universality, Kellogg goes on to suggest, is also apparent in the spontaneous basic configurations which appear between the ages of two to three years in infants’ drawings (Kellogg 1969: 220, 235) — although this scheme has not gone unchallenged (Golomb 1981; Cox 1993). Moreover, Kellogg regards the development of infant art as paralleling the development of vision (1969: 260), and makes a direct comparison between abstract U. P. forms and similar forms which she had identified in infants’ pre-representational art (1969: 217). Given this correspondence what, then, are the implications for a similar positive correlation involving L./M. P. abstract marks?

Direct comparisons between early mark-making in the Lower and Middle Palaeolithic, early infant art and phosphene primitives
From the age of eighteen months to three years the developmental sequence of infants’ drawings, according to Kellogg, is as follows: the meandering, scribbled line, as the first graphic element to be explored by infants, is, by approximation, placed ever more carefully within the borders of the rectangular paper resulting in the emergence of basic geometric forms, such as a scribbled circle, rectangle or triangle. Initially, it is difficult to distinguish any form from the indiscriminate array but, eventually, shapes based upon, for example, the circle, become increasingly oval or elliptical. By two and half to three years of age children begin to draw with a single outline the latent shapes inherent in the drawing instead of just filling-in each existing form with further scribbling. Six recurring ‘diagrams’ were found to arise as a consequence: circles/ovals, squares and rectangles, triangles, crosses and an irregular odd shape — although Gestaltists, like Arnheim (1974), regard the circle as being one of the first shapes to emerge because this could be due to the ease by which the form can be realised in terms of gross motor control (Piaget and Inhelder 1956). The child next begins to play with the vocabulary of discovered forms, first a square, then a triangle, then a patch filled with colour, then a series of dots, for long periods at a time without any prompting. Afterwards, children begin to incorporate two diagrams, producing ‘combines’ where a cross is conjoined with a circle, rectangle or square, to produce a ‘mandala’ (e.g. Figure 1 [16E]). This is regarded as a crucial stage in the transition from ‘scribble’ to representation. Soon afterwards the mandala becomes a more complex form known as an ‘aggregate’, where the child adds further crosses so the circle appears like a wheel with spokes; other similar configurations are created by a juxtaposition of lines, circles and rectangles to produce six basic shapes. The ‘spokes’ then begin to radiate out from the circle. By limiting the number of radiating lines and adding two ‘eyes’ the child can turn the form into a tadpole figure as the first representation of a head (or head and body combined). The development of other schemas — animals, buildings etc. — are explained in similar terms by highlighting the origins of a prior invented graphic repertoire (Kellogg 1969). Children rarely depict representational forms by the age of three.
The development of drawing in this manner can be viewed (following Arnheim 1974) as congruent with the tendency of mark-making to proceed from simple, global forms to complex, more differentiated ones. This is in keeping with the fact that most psychological development, including perception in infants (Slater 1996), evolves from global to differentiated structures.

Figure 1. Comparison of early human mark-making with phosphene motifs and early mark-making of infants.

It has been found that primates such as chimpanzees (Morris 1967) produce paintings with a striking similarity to drawings made by children at the early stages (this similarity adds weight to the evolutionary paradigm of mark-making). Both create vertical, horizontal and curved lines as well as dots. Chimps follow this sequence up to the diagram stage but produce only the circle and cross compared to the six basic forms of infants. One of the major forms the chimp was unable to master was the square, as well as being incapable of any repre-sentational depiction.
Given the transitional graphic sequence of infant’s drawings, how does this compare with marks made in the L./M. P. by early humans? Figure 1 illustrates some of the scribble and geometric primitives commonly employed by infants from eighteen months onwards, as identified by Kellogg (1969), together with concomitant L./M. P. marks. The similarity of many of the forms is immediately obvious, including phosphene primitives, the significance of which will be discussed more fully in the following section. We can now explore in greater detail the extent and ramifications of any positive comparisons arising.
One of the earliest instances of marks made by early humans comes from the Auditorium Cave, Bhimbetka, India, dating back some hundred thousands of years (Bednarik 1993, 1994a), namely a petroglyph in the shape of a large cup mark, or cupule (Figure 1 [11A]), which will be discussed presently, and a pecked meandering line (1A). Meandering lines, in the form of digitally made marks, can be found in rock art throughout the world and are thought to be Pleistocene and among the most ancient, predating later forms and iconic representation (Bednarik 1984; Bahn and Vertut 1997: 106). A meandering line, it can be seen, has distinct similarities with the first scribbles made by infants of eighteen months (Fig. 1 [1D]). There are also obvious commonalities between the irregular shape of this line and the two phosphenes illustrated (lB and 1C). Other marks dating c. 350 000 years ago from Steinrinne, Bilzingsleben, Germany, on non-utilitarian elephant bones (Bednarik 1995a), seem to be patterned with vertical (2A), horizontal (3A), diagonal (4A), and fan-shaped incisions (5A) (the latter coming from Stránská Skála, Czech Republic, on an elephant vertebra), which can be favourably compared to concomitant lines made by infants and with related phosphenes. Repetition of a particular line direction seems to be a key feature of such marks, fan-shaped lines being a resume of what had previously been achieved. A similar development in infants is highlighted in 2D Y 5D where the explored directionality of line is unified to create a fused, fan-shaped, scribble/combine form (fan-shapes are another common feature in the paintings of chimps; Morris 1967). L./M. P. radiation lines have also been unearthed engraved upon a modified phalanx (finger bone), horse canine and bone splinters from the Micoquian of Prolom (Bednarik 1995a: 609-10). Support for such lines being the beginning of an informational search comes from Arnheim and Booth (Goodnow 1977: 35, 40) who suggest a line provides, as one of the first marks to be discovered, a way of expressing order through repetition. Arnheim (1971) also proposed that an economical use of shape can establish a bit of order in a world of complexity and what appears at first sight to be a limited vocabulary of units may, in the discovery of similarity, represent a strength.
Abri Lartet (Bednarik 1995a: 627) has provided a series of three attempted, acute-angle forms dating from the M. P. (6A) which can be positively correlated with associated phosphene (6C) and infant scribble/emerging combine (6D). The ubiquitousness of chevron-like forms, e.g. at Bacho Kiro, Bulgaria, on a M. P. bone fragment (Marshack 1976), highlights the preoccupation with the attempted depiction of angle (Morris 1967 also reports a zigzag line in chimps, although rare). Motifs 7A, 8A, 9A and 10A are non-utilitarian markings from Bilzingsleben, Germany (c. 350 000 BP), on: a split elephant tusk, elephant bone, quartzite slab and metatarsal bone, respectively, all L. P. (Bednarik 1995a). Motif 7A seems to be half-way between an arc and angle form, the important feature being the nesting of one form within another as also in the accompanying phosphenes and infant emergent combine (7D Y 7E). In Kellogg’s terms these figures would be a demonstration of good visual balance, which would explain the repetition found in the marks of children and early humans. Similar, but more curved, paired lines have been found (13A) (Bahn 1998: 162) in Australia from Panaramitee North, dating back approximately 43 140 BP, which clearly exhibit attempted repetition. Repetition is one step on from simple doodling and the straight line indicative of a definite graphic progression, and anticipates Kellogg’s combine/diagram stages. Motif 8A is a further step on the way to repetitive angles as well as being an attempt to integrate each form into a pattern with longer, continuous, straight lines. The source of this kind of figure may derive from the fortification phosphene illustrated in 8C — note also the chevrons in this illustration. Such patterning demonstrates the groping towards a compounding of angled and straight lines. Motif 9A (engraved on a quartzite slab from Bilzingsleben, Germany; Bednarik 1995a) represents a preliminary attempt at realising a rectangle, corners seeming to have been a significant problem — while 10A (engraved on a metatarsal bone also from Bilzingsleben, Germany; Bednarik 1995a) marks a more successful integration of this form, where the obvious experimentation with angle and corner has resulted in the portrayal of a complete figure. Motif 10A is particularly significant, because not only are right angles successfully depicted for the first time, but the process by which this has been achieved is graphically portrayed in the struggle to create the right-angle at the bottom left and the incorporation into the design of a series of horizontal, vertical, and sloping lines, cross and chevron shapes — important prerequisites before an integrated square can be achieved. It is as if the author has summarised graphically all previous achievements to date (1A to 9A) before creating a fully-fledged rectangle, the lines, with practice, becoming more parallel and perpendicular, and the angles sharper. Motifs 9D Y 9E and 10D Y 10E present a neat synopsis of how the infant attains the equivalent form, where early squares may resemble circles as in 9D (Gardner 1980: 42).
A corresponding evolutionary process of combining previously attained forms can be observed in boxes 11A to 16A using curved lines — in this case, however, the culminating figure, 16A, is a circular mandala. Motif 11A is a large cupule from Auditorium Cave, Bhimbetka, India, of the Acheulian and estimated to be a few hundred thousand years old (Bednarik 1993, 1994a). The cupule (11A) correlates positively with circular scribbles 11D and diagram circular form 11E made by infants, as well as with phosphenes 11B and 11C. Panel 12A, from a Neanderthal M. P. burial site (Bednarik 1995a) and 13A, from Panaramitee, Australia (c. 43 140 BP) (Bahn 1998: 162), show the same inclination towards repetition as with straight lines, and probably repeated for the same reasons already cited (a series of intentionally engraved, connected double arcs have also been found on a bovine rib from Le Pech de l’Azé, Dordogne, c. 300 000 BP; Bahn 1997: 24). Also from South Australia (Wharton Hill, c. 42 000 yrs; Bahn 1998: 162), comes an oval shape (14A), an infant motif which follows that of the circle (Gardner 1980: 42). The markings on the Quneitra artefact (15A) (Marshack 1996: 357), a flat cortex plate from the Levant, incised with four nested semicircles surrounded by vertical lines (about 54 000 BP), is evidence of the importance directed towards exploring the possibilities of the curved line. Of particular interest in this figure is the fact that the curved lines are made up from a series of straight lines knitted together to create the curved form. This simulates perfectly the way in which the striate cortex (see below), according to the Gestalt principle of good continuation, can allow for the construction of geometric forms from simple lines — the perception of straight lines is mapped by striate neurones in revolving order of orientation in exactly the same manner as in the circular motif inscribed on this artefact (for a more detailed discussion of the neurophysiological processes involved see Gilbert 1998; also Eysel 1999, and Siegel 1977: 131 for summaries). The similarity of this artefact with the infant combine is clear (14D Y 14E).
Finally the geometric figure from Tata, Hungary (16A): a silicified semi-transparent nummulite has a crack through the centre visible on both sides and, at right angle, a single engraved line on each face (Bednarik 1995a: 612), constituting the first known example of how a circular and straight line form are arranged together to produce a fully balanced figure. This, perhaps, represents a defining stage in the bringing together of two distinct graphic primitives — an important landmark in the evolution of graphic schema. Bednarik (1995a: 614) sees the realisation of this form as part of a continuity from the L. P. to the M. P., i.e. from the unstructured to the structured, just as Kellogg finds the same form (16E) equally significant for infants reflecting the tendency of perceptual development to evolve from the global to the particular. A similar progression can be seen in some South Australian caves, e.g. Karlie-ngoinpool and Koorine Caves, where more than thirty-five contain nonfigurative markings made by humans embodying three distinct artistic traditions (Aslin and Bednarik 1984a, 1984b; Aslin et al. 1985). The first stage consists of digital flutings, which seem to follow the contours of the wall; the second is dominated by engraved circles which are either simple, concentric or divided by lines; and lastly, the more recent manifestation of engraved lines of various forms.
All the marks from the L. P. and M. P. can be found either in the early marks of infants, up to and including the combine stage, or phosphene types. The gradual evolution of graphic primitives as an explanation has the advantage of being testable, the model allowing us to predict that the present gaps in the record, particularly between the combine/aggregate stage and representational stage, will be made up of further combines as well as subsequent aggregates similar to those produced by infants. A later incarnation of such combines and aggregates (commonly found in much of the later rock art of the U. P., such as grids, crosses, triangular forms and other abstract shapes, e.g. the grids of Altamira and La Mouthe, Dordogne, the cross of Castillo, Santander), parallels the same way as children, well into the representational stage, harken back, from time to time, to earlier stages producing various comparable, geometric designs (Golomb 1981). Following the parallel with infant markings this can provide important clues as to the nature of the aggregates which might exist as part of the missing link between M. P. and U. P. art. In addition, we should expect to see a divergence in similarity between the pre- and early representational forms of children and early humans as higher-order, culturally derived factors begin to impinge on recognisable iconic form. At the same time it should not be forgotten that the preoccupation of U. P. artists with representational depiction, by the use of outline rather than filling-in (Halverson 1992), is also a dominant aspect of children’s art — something which seems to be sufficient to meet the yearning for representation.
If a computer database could be devised documenting a complete existing inventory of early mark-making and representational imagery according to date, and keeping in mind how evolution can evolve haphazardly (but still in one direction), the model allows us to make the further prediction that there would be a gradual drift toward complexity beginning with the earliest, simple marks of the L. P. through to the appearance of diagrams, then combines/aggregates, followed by simple abbreviated and attenuated representational outline forms (as Davis [1986: 184] proposes) to full naturalistic outline representation, finally culminating in representation of the U. P., complete with colour and textural highlights. Figure 1 already gives some clues as to the nature of this ‘drift’ with a steady increase in complexity from the L. P. to M. P., concordant with infant mark-making.
The fact that marks found in the archaeological record are not specific to one site but from many widespread and varying locations and distributed over long periods of time (Bednarik 1995a: 614) is hardly surprising given the existence of a possible universal, predisposing factor. This, together with the close similarity between the early art of infants and L./M. P. art, leads to the visual cortex (Bednarik 1984: 28) and, more specifically, the striate cortex as the one mechanism which can account for this commonality. Before examining more precisely the nature of this mechanism for understanding the emergence of art I will first cite some further evidence alluding to this part of the brain as the possible critical locus for early mark-making.

Phosphene Theory and the visual cortex.
Bednarik (1984, 1990) has proposed the phosphene theory as an adequate explanation of L./M. P. art. Phosphenes have also been implicated as the underlying factor in geometric primitive art by Siegel (1977: 131-2) and the abstract form of Neolithic art by Kellogg (Kellogg et al. 1965). One aspect of the manifestation of phosphenes is the claim that they derive from shamanistic practices involving drug-induced trance states (Lewis-Williams and Dowson 1988; Dronfield 1996) and, it is claimed, representational features arise out of geometric primitives in three stages (Lewis-Williams and Dowson 1988: 203-4; Siegel 1977: 125, 127, 130).
The shamanistic theory has been criticised by Bednarik (1990) who regards such an interpretation as untenable as cultures which are not shamanistic, or do not resort to hallucinatory derived experiences, are also found to have phosphenes in their art (1990: 79). Like Kellogg et al. (1965), Bednarik believes phosphenes are a fundamental universal of early art (1990: 77), all humans experience them (even blind people) but more particularly infants (Bednarik 1990: 78).
Appeal to the central nervous system or visual cortex as a common cause in the manifestation of phosphenes seems to be understood (Siegel 1977:132; Kellogg et al. 1965: 1130; Bednarik 1990: 78; Lewis-Williams and Dowson 1988: 202). More specifically, Knoll and Kugler (1959) and Penfield and Roberts (1959) found phosphenes arising when either the temporal part of the brain or the visual cortex was directly stimulated by electrical impulses.
Further evidence that the visual cortex may be the focal area for phosphene experience comes from migraine sufferers. Siegel et al. (1975), in keeping with Gestalt theory, suggest that geometric forms derived from visual images of those suffering migraine attacks provide information on the lattice arrangements of detector cells in the visual cortex.
More specifically, Dobelle and Mladejovski, in developing an artificial aid to help the blind report that only stimulation of the striate cortex (1974: 559-60) in conscious patients produced phosphenes (either round or short line shapes; Dobelle and Mladejovski 1974: 567), and Brindley and Lewin (1968) found simultaneous stimulation of the same area by several electrodes led to the perception of various predictable patterns. Dobelle et al. (1976) also report a prosthesis, consisting of electrodes inserted into the striate cortex connected to a TV camera, which allowed a blind person to detect horizontal and vertical lines so they were able to recognise simple letters and patterns (Dobelle et al. 1976: 111). Importantly, it has also been established that damage to the striate cortex leads to an inability of those affected to copy the simplest geometric forms; moreover, and inversely, in cases where the striate cortex remains intact, but there has been a lesion to other areas of the visual cortex, the subject is able to draw local elements of form such as angles, simple lines and shapes but is unable to integrate such lines into a complex whole (Zeki 1992: 48-9).
The one unifying thread linking the different accounts as to how phosphenes arise seems to be the striate cortex — whether deriving from the trance states of shamanism, drug-induced hallucinations, migraine attacks or electrophysical intervention etc., each can be viewed as a different route towards the inducement of phosphene primitives as indicative of the structure of this mechanism. Ultimately, phosphenes can be seen as being derived from the underlying neuronal architecture of the striate cortex from which, in turn, phosphene-like marks can be evoked through various means. In this model shamanism can only account for a limited number of cases while phosphene theory is able to embrace a wider spectrum of instances. Significantly, however, as Bahn and Vertut (1997: 182) have indicated, the very act of engaging in mark-making may serve to induce a hyper-reality experience and in this sense it is an activity which can be said to ‘resonate’ with the striate cortex, therefore, providing a further, alternative but inclusive, explanation for the appearance of L./M. P. phosphene-like motifs.
The neurophysiological evidence already presented, together with the correspondence between phosphenes, L./M. P. and infant art, as highlighted in Figure 1, suggests that the striate cortex is the common, predisposing, neurophysiological factor. In support of this hypothesis further neurological studies will now be considered.

The importance of lines and the striate cortex
Latto (1995: 67-8) stresses, as Gombrich does (1973: 201), the way artistic primitive motifs are aesthetically interesting, not because they reflect properties of the world, but, rather, because they simulate properties of the visual system (see also Bednarik 1984; Halverson 1992; Solso 1993: 49, for earlier expressions of this idea). The obvious question arises; how is this realised in the visual system?
The striate cortex (also known as the primary visual cortex, area V1 and Brodmann’s Area 17) in primates plays a critical role in visual information processing, as most visual information reaching the rest of the visual cortex is funnelled through this area (Felleman and Van Essen 1991). This ‘gatekeeper’ role may account for the area being the largest known visual cortical area, if not the largest cortical area (Tootell et al. 1998), thereby underlining its importance. The significance of the striate cortex is further highlighted by Hubel and Wiesal’s discovery that cells in the striate cortex are organised to respond to specific orientation of line and that perception may be fabricated from the accretion of selected features (Hubel and Wiesal 1979). Tootell et al. (1998) also confirmed, by functional analysis of the striate cortex using Magnetic Resonance Imaging (MRI), that this area has lower contrast sensitivity, i.e. it is more sensitive to visual phenomena such as lines, as well as being orientation selective (1998: 815-16).
Hubel and Wiesal go on to describe how the striate cortex may represent an early stage in the brain’s analysis of line orientation and an important aspect of the processing of visual information through recourse to a hierarchy of simple, complex and hypercomplex cells by which the nature of information to do with line becomes more abstract — this may be a continuation of a process that begins in the retina which leads to a successive enhancement of contoured borders (Jung and Baumgartner 1965). Barlow’s (1972) feature detection theory is an extension of Hubel and Wiesal’s analysis, which proposes that cortical cells forming the bottom layer of a hierarchy of cells respond progressively to more and more abstract geometric features. Hence, cells in higher layers could respond to simple, geometrical patterns, such as angles, defined by the activities of particular combinations of complex and hypercomplex cells, leading to the perception of yet more complex features, such as rectangles and circles, and so on right up to representational figures. It should be remembered, however, that this is a simplification of an extremely intricate process which is still not fully understood.
Another aspect of visual processing relevant to the origins of art at this level is the separate, but parallel, processing of different kinds of visual information (Livingstone and Hubel 1987), which becomes more specialised on reaching the striate cortex, dividing it into three channels as opposed to the previous two originating from the eye. It appears that early humans, in making the first marks, may have been extracting one particularly significant aspect of the optical array for possible exploitation, beginning, as in vision, with the analysis of parts of form in terms of line and line orientation (Hubel and Weisal 1979: 36-7). The artists of the L./M. P. were, therefore, probably appealing to this evolutionary earlier, colour-blind system as a visual channel (known as the magno system), that is concerned with, amongst other things, decisions involving which visual elements, such as figure/ground, discontinuities or edges belong to individual objects in the scene as specified by Gestalt theory (Livingstone and Hubel 1995: 61-4). This system operates in tandem with the ‘parvo-interblob’ pathway, which deals with the high resolution of detail found in static form (the parvo-interblob pathway is thought to have originally evolved out of the magno system; Livingstone and Hubel 1995: 64). The perception and analysis of form, which these pathways seem especially suited to deal with, is a critical visual imperative that would, over time, have been specifically selected for by way of evolutionary dictates, e.g. in the detection of concealed, stationary and camouflaged fauna as either threatening, benign or as a possible food source.
Marr (1976) has demonstrated how a visual processing model (leading to what he calls the ‘raw primal sketch’), beginning with the extraction of essential edge features as necessary components of a line drawing, can account for why such drawings are so efficient in depicting representation or carry so much information with such great economy. It is interesting, in this respect, that Marr’s ideas made extensive use of Gestalt principles of perceptual organisation where discrete elements are deemed more likely to belong together which present a similar orientation, or lie next to one another, than those oriented dissimilarly and spaced far apart (Bruce and Green 1990).
Evidence points to the fact that line drawings are universally exploited as a shortcut to representation as exemplified by infants and both ‘primitive’ and modern artists (Kennedy and Silver 1974; Latto 1995). This can explain why early humans began, preferentially, with an investigation of the possibilities inherent in the use of line leading to the artistic equivalent of Marr’s ‘raw primal sketch’, culminating in the animal outline figures of the U. P. (Halverson 1992: 389, 402). Thus, different aspects of a painting are processed by different regions of the visual cortex (Livingstone 1988; Zeki 1993: 355), with the analysis of line in the striate cortex being of primary significance.
Zeki (1993), it should be added, has also identified discrete cells in the visual cortex (V4) beyond the striate cortex responsible for the analysis of colour. The additional colour dimension lent to the image by U. P. artists may, accordingly, also be a simulation of the segregation of visual processing in the visual cortex, which parallels the more recent evolution-formed ‘parvo-blob’/ V4 stream concerning the analysis of the visual array in terms of this further attribute (Livingstone and Hubel 1995: 63).
In brief, this research provides added substance to the striate cortex as the seat of the phosphene/artistic, primitive forms, and suggests, as a graphic component, line, by way of this mechanism, was the first to be exploited during the L./M. P., as a precursor to representational line drawings. A similar drive towards representation is also present in infant mark-making and reflects the perceptual stages through which the neonate passes on the way to making visual sense of the world. Vertical, horizontal, tilted, circular and lattice lines would have been graphic elements indicative of this early visual information processing, the mandala being especially significant, embodying, in one figure, the various features involved. Subsequent fully-fledged coloured paintings appeared in the U. P. as a derivative of a separate, more recently evolved, but parallel, visual channel. Figure 2 summarises all the major factors identified suggesting the striate cortex as common cause, as well as possible interrelationships.

Figure 2. Relationship of factors identifying the striate cortex as common cause.


Incremental trends and common trajectories
Figure 3 sets out the proposed information-processing model illustrating how early art may have played a pivotal role. It delineates the evolution of graphic form as well as how the increasing drive towards the processing of information was incremental and, at significant points, cross-referenced, or cross-fertilised, with other previous, parallel modes. Each cross-fertilisation can be seen to have led to a significant step forward in the creation of more efficient and different ways of dealing with information. Dennett (1995) might describe this cultural development as analogous to a ‘crane-making crane’. In this respect it can be seen that once abstract marks had spawned representational art such marks continued to proliferate along a separate trajectory and, in so doing, became more refined leading, eventually,

Figure 3. A gradualist exponential account of art as information processing.

to the repeated geometric decoration common to nearly all later cultures. More specifically, the first graphic lines fingered into soft surfaces, carved and scratched onto crude tools, bone objects and stones would have provided the foundations for the later representation of fauna on cave walls as well as, through the use of graphic combines and aggregates, the purely decorative abstract forms found widely in much later so-called ‘primitive’ art. A transitional aggregate or combine leading to this fully integrated decorative realisation is apparent in the widespread manifestation of various geometric forms during the U. P. and, as already cited, is a phenomenon which continues to perplex many commentators. It can be argued that such abstract motifs, typical of this later period, are a retracing of familiar, but earlier, geometric forms (see above). Trajectories for other traditions and significant dates can, likewise, be traced according to the model, each cross-fertilisation leading to increasingly powerful and efficient ways of dealing with information.
More speculatively, separate linear traditions involving later amalgamations with other lineages may well be a simulation of the strategy of functional specialisation of the brain as a means of acquiring an understanding of the permanent properties of the world. Although this functionality solved one set of problems, new ones arose, concerning how separately evolved mechanisms were to be synthesised in order to analyse different kinds of perceptual information. The hypothesised association cortex, as a necessary interface, may correspond to how separate traditions of overt information-processing have gradually cross-referenced to produce more efficient, and effective, cross-modal analytical strategies (e. g. Humphreys et al. 1988), rather than as a consequence of some sudden mysterious process taking place in the brain during the U./M. P. transition (e.g. Mithen 1996).
The question concerning why sculptural objects have tended to predate graphic representation can be adequately embraced according to this model by the fact that three-dimensional objects were probably easier to fabricate than the more sophisticated representation of three dimensions in a two-dimensional format, sculpture having arisen directly out of the more ancient tradition of tool making (Halverson 1987: 66; White 1989: 80; Feliks 1998a: 113-6). Indeed, even before the first fully sculptural object was created there would probably have been a long period of passive appreciation of naturally occurring artefacts as proposed by Bahn and Vertut (1997: 23). Hence, sculptural and graphic traditions, it would seem, sculpture having arisen developed separately, but in parallel, the more advanced sculptural tradition reaching representational maturity before graphic representation, as implied by the archaeological record (Bahn and Vertut 1997: 100).

The Upper-Middle Palaeolithic ‘transition’ as a product of information processing
The 40 000 BP boundary can, in Figure 3, be seen in the context of an incremental account and is accommodated by the increasing ability to pass down and disseminate information, the hypothesised threshold simply being one of several continuing, but significant steps involved where representation began to cross-reference with spoken language. Equally significant boundaries can be identified anywhere along the sequence, e.g. the earliest mark-making, the invention of symbol, the termination of brain growth. Representational pictures, in this model, would have served as the first, universal protowriting (Bahn 1998: xxiii-xxiiv), providing a necessary and direct precursor to written language and graphic symbol (pictograph, hieroglyph and cuneiform) in much the same way as Gregory (1970: 137, 151) has suggested. Diringer (1962) postulates that pre-Historic representational art, as a form of embryonic writing, is equivalent to no more than nouns, i.e. despite the obvious stylistic merit it remains static and lacks narrative capacity (1962: 27-30); put another way, there is little relationship between the separately defined representations. This may help to explain the lack of any discernible order in the placing of images, the phenomenon of superposition and the scarcity of ‘scenes’. The ability to be able to represent one-to-one, i.e. the iconic image capturing the essential features of a natural object and the probable capacity to be able to name such objects, accordingly heralded a significant stage in the realisation of a symbolic code where ultimately the symbol would become arbitrary to the referent as had already occurred with speech. The emergence of a graphic/symbolic means of representing language can be traced from the naturalistic realism of the U. P., to the increasingly stylised, cartoon-like figures of the Mesolithic Spanish Levant and the first known pictograph 10 000 BP (Stordeur and Jammous 1995).
Symbolic representation, as a means of preserving, spreading and manipulating information, provided a powerful way of increasing the capacity of the brain, disengaged from the usual evolutionary constraints, helping to catapult humans into a dominance over nature. The Integrated Causal Model insists that the cultural imperatives deriving from this ability are as much subject to the vagaries of evolution as anything else (Tooby and Cosmides 1992: 122; Cosmides et al. 1992: 5) but its proponents fail to realise that culture may embody the way brain potential has been amplified through the propensity to represent reality symbolically, thereby storing representations for future analysis, revision and improvement which the cortex, alone, was unable to achieve. The Integrated Causal Model is correct, however, in drawing attention to the slow-moving evolutionary and biological constraints to which modern humans are still largely subject but, in the modified version outlined in Figure 3, the extrapolation must be that culture is not so much a sophisticated embodiment of evolutionary imperatives, but a realisation of the diversity due to the increased ability to process information in abstract ways. The implication is, then, that there is a disparity between the evolutionary factors to which humans continue to be biologically subject and the swiftly changing imperatives of culture mediated by ever more rapid, efficient and varied ways of processing information. It is not so much that ‘culture’ fuels the growth in brain size through a feedback loop as Wills (1993) suggests, which may well have been the case with early humans (e.g. speech), but that culture now provides the sustenance through which it can, itself, evolve (it is worth noting here, that during the past 35 000 years the brain has actually been shrinking in size [Lewin 1999: 60], which may well be due to the way culture has provided a ‘cushion’ in the face of previous ‘raw’ evolutionary pressures). Symbolic and abstract modes of expression have, thereby, conferred a substantial measure of independence to culture which does not reside in any one individual but is a consensus of what kinds of information are significant and relevant to a society’s improvement or survival. The specific cultural elements involved Dawkins (1991: 158) has described as ‘memes’.


The principles of Darwinian evolution have provided an important means of comprehending the significance of various types of early graphic form and how they might relate to other cultural specifics. Together with an appreciation of information processing in the brain it has helped to elucidate the probable sequential development of early mark making and how this might be related to different but parallel ways of encoding the world, e.g. language. Individual modes of information processing, although initially contingent upon discrete qualities inherent to a particular mode, throughout the course of time gradually began to integrate by cross-referencing with other modes, sometimes leading to the inception of new and more efficient ways of dealing with information. Early geometric marks can be sufficiently explained through perceptual imperatives involving ecological factors and mechanisms concerning the striate cortex, thus avoiding accounts involving symbolic explanations. Phosphenes can also be explained by recourse to the same mechanism and, moreover, the fact that every early mark reported since the phosphene theory was first mooted is itself a phosphene motif adds compelling support to the evolutionary hypothesis. Such marks played an important role in the development of ways of dealing with visual information by recourse to Gestalt principles creating a platform for the emergence of graphic representation whereby symbolic interpretations did, eventually, begin to become relevant.

I would like to thank Dorothy Phipps for her support and help in clarifying some of the issues concerning language and Molly Wood for the many thoughtful comments and encouragement, as well as Dr M. Ludvigsen for his kind assistance with the numerous drafts and advice on the compilation of figures. Finally I thank the RAR referees for their constructive criticisms.

Derek Hodgson
2 Belle Vue Street
North Yorks YO10 5AY
England, U.K.


In science, falsifiability rules

Derek Hodgson’s paper is one of the most worthwhile additions of recent years to the discussion of art origins. Like the recent paper by Feliks (1998a), also in this journal, it enriches a debate that has become bogged down in unfalsifiable, often idiosyncratic hypotheses about the meanings of palaeoart whose principal feature it is that they are assumed to appeal to the public. Serious work such as that of Feliks and Hodgson receives scant attention in this scramble for public favour. Indeed, Hodgson’s paper illustrates the issue well. It is based largely on a model presented almost twenty years ago, the phosphene theory. While that theory is perfectly falsifiable and thus scientific, it has been ignored for all these years, but the competing shamanistic model, which its originator has freely admitted is profoundly unfalsifiable, has been widely preferred since it was proposed in 1988. This prompts a simple question: why do untestable (and, one might add, unlikely) hypotheses receive this favourable treatment in archaeology, while hypotheses that are fully testable, and that have never been falsified, are consistently ignored?
Hodgson reminds us that all the hypotheses trying to explain ‘abstract’ marks before and during the U. P. will not lead to a satisfactory model because they cannot be substantiated. Or to rephrase this prediction as a scientific proposition: none of the various ‘interpretations’ we have for this corpus are falsifiable. The phosphene theory, by contrast, does not ‘explain’ these marks, it merely suggests derivation, but it is falsifiable and thus scientific. It can be significantly weakened by the discovery of a major art body of non-phosphenic marks predating the introduction of two-dimensional iconic art. Until such evidence is presented, the phosphene theory remains unrefuted. In fact I regard Hodgson’s paper as a testing of my phosphene theory. He quite correctly points out that that theory made certain predictions: ‘the model [allows] us to predict that the present gaps in the record … will be made up of further combines as well as subsequent aggregates similar to those produced by infants’. In the present paper he fills some of these gaps. A second way to test the phosphene theory is, as I have proposed many years ago, to look for disconfirming evidence, in the form of very early palaeoart that is clearly not phosphene inspired. The phosphene theory is preferable to the untestable alternatives we have in the literature, and yet Hodgson’s paper is the first discussion of it to appear in print. As an epistemologist who is profoundly interested in the heuristic dynamics of the discipline I welcome the opportunity to comment.
I have bemoaned for a good twenty years the fact that non-iconic palaeoart has been so severely neglected. The preference for ‘nice pictures’, which has distorted Pleistocene palaeoart research for a century, is biased, and it should be self-evident that iconic art is semiotically less sophisticated than non-iconic. In iconic art iconicity provides an obvious referent, whereas any referent in non-iconic art must be culturally negotiated and transferred.
Hodgson’s proposition that palaeoart developed as a means of expanding human cognitive capacities beyond the limits biologically imposed by brain size, functioning as a kind of surrogate cortex, seems to be a novel and original idea. It likens palaeoart to back-up storage of computer data: the ceiling in brain size imposed by natal limitations led to external means of storing data. This is the most exciting twist in Hodgson’s paper, and it adds a new facet to the phosphene theory. Until now I favoured the explanation that the discovery of shared phosphenes, through their externalisation, led to the realisation of a shared consciously experienced reality. Hodgson’s idea of a ‘surrogate cortex’ offers attractive features and could be explored further, and it does not exclude the mechanism I had favoured; the two could both be valid components of an interpretation of art origins tied to the ‘catalyst’ role of the phosphenes.
An interesting point made by Hodgson is that repetition is appealing to us because it is an integral feature of the brain, it signifies ‘safe-ness’. This is also worth pursuing further, because there are ‘elements of repetition’ evident in much if not most early palaeoart: consider, for instance, multiple parallel lines, finger flutings, concentric circles, multiple arcs (Quneitra), multiple cupules, paired markings generally, series of notches, multiple zigzags or meanders, or objects made in large numbers to a single pattern (e.g. disc beads). This concept offers considerable prospects for further development: the establishment of synaptic pathways is contingent upon repetition, as is the production of many early marks. There is a good deal in this paper to prompt some stagnant synapses in my own cortex to begin firing again.
I am not convinced that Hodgson’s separation of (U. P. and later) ‘art’ from (L./M. P.) ‘proto-art’ serves any useful purpose, it reinforces the artificial discrimination between non-representational and representational art (are depictions of phosphenes representational?). Worse still, it emphasises the even more artificial separation of the M. P. and U. P. (Fiedler 1999; Bednarik and Kuckenburg 1999). Also, it begs the question, ‘What is art?’, a question archaeologists and art historians have not dealt with in a scientifically meaningful way (cf. Rosenfeld 1999). I therefore prefer the collective term palaeoart to describe all art-like phenomena of the past, including beads, pendants, engraved marks and ‘non-utilitarian’ objects generally.
Concerning an epistemologically anchored definition of art, Hodgson’s preferred version seems to correspond closely to mine: art is the medium conveying the artist’s awareness of perceived reality to human sensory perception; it externalises human concepts of reality. In an epistemologically more satisfying definition, art is the one phenomenon in human experience of which there can be no crucial common denominators of phenomenon categories (CCDs) that are inaccessible to humans; art thus consists of the only collective phenomena which humans can experience ‘objectively’ (Bednarik 1994b).
Hodgson cites Bahn as stating that the 40 000-year boundary is exaggerated due to more recent rock art having survived better. While this is, in an over-simplistic way, a valid statement, the role of taphonomic logic requires much more careful consideration, and 40 000 years is not a taphonomic threshold in rock art. Besides, it is evident from Hodgson’s mention of an ‘assumed upsurge’ in cultural activity at the M./U. P division (there is quite probably more M. P. rock art in the world than U. P., conversely) that he ignores the concept of the taphonomic threshold. We need to be reminded here that archaeology does not deal with cultural events or developments or anything of the kind, it deals exclusively with material residues of events, and unless one understands how the characteristics of these residues can be translated into scientific propositions about events in the past our speculations are scientifically irrelevant. The formulae for such translation are encoded in taphonomic logic, which decrees that the statistical or other quantitative, and often even qualitative, characteristics of the material record are frequently irrelevant to its interpretation. This is the most fundamental law in all of archaeology (Bednarik 1995b).
Hodgson sees the cave art in Altamira and Lascaux as being about 16 000 years old. I think some or all of the Altamira art might be somewhat younger, perhaps 14 000 years, while some of the Lascaux art is probably not of the Pleistocene at all (Bahn 1994).
The marked bovine rib from Pech de l’Azé should not be cited in the context of discussing secure palaeoart. It has always been controversial, and d’Errico and Villa (1997) have persuasively shown that judgment should be reserved on this specimen. Similarly, Hodgson cites radiocarbon-derived datings of petroglyphs from the Olary region in South Australia. It must be emphasised that these dates are no longer regarded as valid, the analyst who produced them has himself withdrawn them by casting severe doubts on the method he used to calibrate these ‘dates’.
I fail to see the significance of the proposed ‘hyper-reality experience’. If the making of certain markings had induced a trance-like state, that in itself would not constitute an ‘explanation for the appearance of L./M. P. phosphene-like motifs’, as Hodgson suggests. I have drawn many phosphene motifs but have never experienced the slightest ‘altered state’ as a result. Nor do people who draw nothing but phosphenes today (such as infants) enter such a state.
Concerning Lewin’s ‘shrinking’ human brain, I doubt that the rather slight difference in average cranial volume between Neanderthaloid sapiens and later sapienoids is of any great significance. It is one thing to accept that encephalisation was essentially complete by the time archaic Homo sapiens forms appeared, but such minor differences in brain size are negligible, and they are well within the range of modern human cranial sizes. Moreover, I understand that the size of a human organ is a poor indicator of its performance.
Concerning the issue of symboling before the Aurignacian, speech or language are forms of symbol use. Colonisation by seafaring undeniably requires complex communication (Noble and Davidson 1996), forward planning and abstract thought of an order of magnitude not significantly different from that available to present-day humans. Since we have known for some time that seafaring hominids navigated the sea many hundreds of millennia ago, this record demonstrates the use of symbolism as well as technological sophistication (Bednarik 1999).
Concerning phosphenes, particularly their physiological aspects, the principal resource remains the work of Eichmaier and Höfer (1974), which Hodgson has not cited. The published works he cites on this subject are eclipsed by Eichmaier and Höfer’s more comprehensive work.
Concerning the Bilzingsleben markings, I refer readers to the recent laser-microscopic study of the principal specimens from that site by Steguweit (1999), according to which the engraved grooves were made intentionally and are not incidental utilitarian marks. The method used for this study is technically superior to the microscopy used elsewhere by Marshack, d’Errico and myself. For instance it eliminates subjective aspects of judging groove profiles and replaces visual judgments with hard metric data.

Robert G. Bednarik
P.O. Box 216
Caulfield South, Vic. 3162

Though art’s hid causes are not found
(Benjamin Jonson, 1572-1637)

Hodgson proposes a bold and imaginative synthesis of art, perception and information processing, within which he makes certain assumptions, some new, some old, that I would like to question. I do feel he relies too much on the heavily-criticised 19th century view of Ernst Haeckel, that ontogeny recapitulates phylogeny, that an organism’s physiological development reflects the evolutionary trajectory of the (often incomplete) fossil record. At a psychological level, Piaget’s stages of cognitive development (themselves recently severely criticised) were probably mistakenly borrowed by students of the evolution of tool use to draw conclusions concerning the cognitive capacities of various hominid species. We should be wary of invoking such parallelisms too uncritically.
Similarly, Hodgson probably relies too much on the striate cortex in developing his argument. Other regions — inferior temporal, temporo-occipital and infero-mesial association cortex, lingual and fusiform gyri — are crucial to face and object recognition and, certainly, to representational art, as shown by object agnosia and prosopagnosia following localised injury.
The fact that the striate cortex is set up to process perceptual primitives such as points, simple linear (and possibly curvilinear) contours, and their intersections (‘stars’), which many in the archaeological world like to call phosphenes, and the observation that the elements of contour drawing correspond roughly to such basic perceptual primitives, do not imply a necessary or an evolutionary connection between them — other than a third observation that, in drawing, similar motor primitives are coincidentally likely for biomechanical reasons. Thus there is only a small set of such basic physical acts possible — drawing a dot, a line, a curve, a circle, an intersection of two or more linear contours. There is no necessary link between a privileged striate cortex (with its set of perceptual primitives) and the motor primitives of drawing praxis, except those constrained by the basic laws of geometry. Hodgson may be placing too much emphasis on ‘resonances’ between ‘mark making’ and ‘phosphene-like activity in the striate cortex’.
It is attractive, but not necessarily correct, to argue that simple dot, line, curve, circle, intersection … contour drawing had to precede representational art in the archaeological record, just as it may seem to in ontogeny. Note, however, that studies of young artist savants like the autistic child Nadia (who possessed superb, imaginative, pictorial draughtsmanship, with full perspective, in her first few years of life) show that such ‘bottom-up’ progression is not in fact the only one.
It may also be dangerous to conclude that ‘art’ in the archaeological record necessarily involved a pre-linguistic attempt to transmit information to contemporaries, or to record it for posterity. Even complex representational art may be produced as an apparent end in itself, rather than as such a means — whether as a doodle, or even as seems to be the case recently with much Australian Aboriginal rock art. Indeed, art (as symbolism) may have very little to do with the evolution of language, except as manifestations of increasing brain capacity.
In the latter context, Hodgson’s observation that human brains have been getting smaller over the last half million years may have less to do with alternative means of recording or transmitting information, and more to do with the evolution of tool use; we need less muscle power now, and, consequently, an allometrically-smaller brain to drive a less-bulky body.
While tool use, language and art clearly relate to increased brain capacity, it is far from clear that they always drove the evolution of the latter — in certain respects, one or more may have been ‘spin-offs’ from a brain-size increase that had other origins. In his writings, the evolutionary theorist Stephen J. Gould frequently invokes the phenomenon of the spandrel, an architectural device in strengthening arches in medieval cathedrals, but which also provided the opportunity, a niche in both senses of the word, for artistic and decorative elaboration which would not otherwise have been available. Maybe art is itself merely a spandrel in the architecture of the brain. Indeed, is an evolutionary perspective even relevant to art if the latter is essentially non-utilitarian and therefore non-adaptive? Because of our desire to find a Darwinian explanation for all human phenomena, maybe we are too ready to invoke a communicative or informational role for art, to link it to the much-more-clearly adaptive phenomenon of symbolic language.
Some minor points: we may be predominantly visual creatures, with a large ‘visual brain’, but it is certainly incorrect to say ‘the cortex, in toto, grew out of the need to process visual information’. Moreover, it is most unlikely that the structure of the visual system differed much between present-day humans, earlier hominids, or even higher primates generally. Any differences are likely to be in ‘higher’ cognitive processes, an ‘aesthetic sense’ (whatever that may mean), and motor control.
Gombrich’s (1984) concept of order as repetition, and therefore as ‘redundant’, is not pejorative; ‘redundant’ is a technical term in mathematical information theory, meaning predictable, as with (aesthetically pleasing, incidentally) symmetrical patterns.
It is not necessarily true that time devoted to mark-making came at the expense of other important preoccupations (food gathering, shelter-construction, mating rituals … ), and so must have been of adaptive significance. Larger mammals, especially predator species, generally devote decreasing periods of time to basic survival activities, with corresponding increases in ‘non-constructive’ activities, or simply inactivity.
I wonder whether Hodgson is falling into the seductive trap of evolutionary teleology when he discusses changes in art in the archaeological record as suggesting a ‘groping towards …’, a ‘struggle to create …’, ‘preliminary attempts’, ‘a yearning to represent’, ‘successful depiction for the first time’.
I am uneasy that representational art somehow reduced processing load and the need for further, metabolically-costly, brain expansion. More brain mechanisms would probably be needed to ‘drive’ the production end than might be somehow saved by ‘exteriorising’ representations or images.
I agree that there is an important difference between geometric and representational art, with possibly different significance and capacities. These differences have not been adequately explored psychologically or archaeologically. Interesting also is Hodgson’s concept of geometric abstract art as reflections of hard-wired cortical activity analogous to Chomsky’s invariant deep structure in language, while representational art may correspond to the variable surface structure of a particular language. However, just as with Piagettian stages, Chomskian perspectives have come in for recent criticism.

Professor John L. Bradshaw
Neuropsychology Research Unit
Department of Psychology
Monash University
Clayton, Vic. 3800, Australia

Art as biocultural artefact

It is refreshing to read a rigorous and challenging consideration of Palaeolithic art that is informed by an evolutionary perspective. The biological imperative of symbolic modeling has not been adequately dealt with in the literature on pre-Historic art, so Hodgson’s analysis presents a welcomed perspective. While Hodgson’s emphasis is on the evolution of perception, his analysis is broadly relevant not only to whole organism biology, but also to the evolution of culture.
As humans, our relationship with the world has — at least since the Lower Palaeolithic — been deeply affected by the images we use to understand and express our place in nature. While biological representation of art is readily seen in the naturalistic depictions of the Upper Palaeolithic, the biological imperative of the ‘abstract’ marks of the Lower and Middle Palaeolithic are much more elusive. But even within the realm of representational images, this imperative is largely indiscernible.
For example, Upper Palaeolithic depictions are principally of animals, which often misleads the modem viewer into thinking that the paintings are only about animals. But, mindful of cultural evolution, we can posit that the images — being (arguably) arranged in purposeful compositions according to biology — are also symbolic of the connections between humans and the natural world. The animals referenced in the paintings live in an orderly world, a visible ecosystem. Transposed onto the cave surfaces, they bring an aura of coherence and meaning to that which they might symbolise; the human community. Hence, I postulate that the artists were cognisant of a system in which the ecology of animals parallels the society of humans. In looking at Palaeolithic art, we witness, perhaps, the graphic manifestation of an evolving totemism, which finds its contemporary expression as the symbolic association between a social group and a kind of animal, plant or natural phenomenon.
Hodgson’s approach to the evolution of art links it to the fundamentals of natural selection. Here I cannot fully support his thesis. From my perspective, the application of his theory does not so much support an argument for the survival of the fittest as it supports an argument for the survival of the fitness — how things fit together. In an ecological sense this art, as Hodgson acknowledges, demonstrates a way of ordering information about the world around us; about understanding how this world is woven together.
Hodgson touches on this human propensity to find connections in his definition of art as (partly) ‘… the seeking of order in the midst of disorder …’. As the affirmation of order, Palaeolithic art helped our ancestors find a cognitive niche in nature. It was not so much that ‘continuity, or repetition, would have signified what was safe, secure, and understood’ (how, then, is innovation in art explained?), but rather that order was crafted out of the ecological relationships observed in nature. Culture, human biology and art — from this perspective — evolved hand-in-hand. Symbolic intent and graphic markings would have developed in tandem, and Hodgson’s gradualist approach to uncovering the mechanisms whereby this happened is sensible.
Hodgson speculates (correctly, I believe) that symbolic ability was developing differentially across various cognitive domains, which naturally adds complexity to human evolution. In support of this he asserts that the ‘laws’ of linear perspective did not readily develop until the Renaissance, despite some ten thousand years of sophisticated language competence. But perspective in art is, I suspect, culturally acquired rather than biologically mandated. Expressive culture is linked, however, to biology in intriguing and obscure ways. One of the most powerful extensions of Hodgson’s evolutionary theory is that the propensity toward symbolic expression might be encoded in our DNA, and that symbolic modeling of the environment may bestow evolutionary advantage.
While I find Hodgson’s essay to be insightful and important, I question a number of his assertions. One is that early scratch marks reflect a search for order and structure ‘in the face of a hostile world’. This nature-red-in-tooth-and-claw perspective represents nothing other than our perception of the physical environment of the Lower Palaeolithic, and it cannot be substantiated that the people themselves shared our perception.
In discussing graphic primitives, Hodgson asserts that the non-figurative digital flutings from South Australian caves ‘seem to follow the contours of the wall’. After inspecting a half-dozen of the major caves, it is my sense that topo-congruence is the exception rather than the rule. While these caves do contain marvellous examples of topographical influences on human marks, the great majority of finger flutings do not show a discernible relationship with geomorphology. I mention this only as a point of clarification; it does not in any way erode Hodgson’s thesis.
Following Diringer (1962), Hodgson asserts that pre-Historic art ‘remains static and lacks narrative capacity’ and that there is ‘little relationship between the separately defined representations’. I again disagree. While scenes are hard to identify, there are ample examples of apparent associations. As Bahn and Vertut (1997: 195) have noted, there is a grammar at work, but we do not know what it is.
Another area where I differ with Hodgson is his assertion that speech is a first decisive step toward a pro-active concern with information processing. Neuroscience has not shown decisively what other animals can or cannot intellect, and there is sufficient evidence to suggest that many animals do exhibit pro-active concern with processing environmental information. We know that many animals invent strategies for problem solving. Ravens, for example, have the ability to solve difficult puzzles. Elephants exhibit another consciousness trademark, an awareness of death and the emotion of grief. (They even linger over the bones of long-dead relatives, seeming to ponder the past and perhaps their own future.) Dolphins exhibit a keen awareness of self-identity, recognising, for example, themselves in mirrors. Perhaps Hodgson is underestimating the nuances of animal intelligence, emotions, and communication, of which we currently know exceedingly little. My point is that while we certainly are different from other animals in remarkable ways, we are not as different as we have imagined ourselves to be.
I am resistant, also, to tired comparisons between primitive art and children’s drawings. The statement that art elevates pleasure ‘in infants (and by implication early humans)’ suggests that Hodgson views early humans largely as just a more cognitively infantile form of modern Homo sapiens. This is an unsatisfactory analogy, and actually runs counter to an eco-evolutionary perspective. To adopt this perspective is to misinterpret the evolutionary biologists’ mantra ‘ontogeny recapitulates phylogeny’.
Finally, I reject the notion that we (humans) have been catapulted into a ‘dominance over nature’. This is a fallacy we are so eager to believe, but which just keeps getting rebuffed by nature itself. It represents, I think, a modern insecurity with ourselves and our relationship with the natural world. In the complex calculus of the human psyche, the relationship between humans and nature is obfuscated yet essential.
I want to stress that I find real value in Hodgson’s essay, and that my concerns address assertions that are mostly tangential to his thesis. Being my best academic self, though, compels me to comment on perceived weaknesses. Despite my dissent with some of the specifics of this essay, I believe that Hodgson presents an important meditation on a type of marking which has been insufficiently addressed. His paper brings us closer to the understanding that art is a biocultural artefact necessary to the ordering of human life, and that culture and biology are not mutually exclusive. While our regard for art may be predicated on aesthetics, he reminds us our need for art is biological.

Professor Paul Faulstich
Environmental Studies
Pitzer College
Claremont, CA 91711

Where’s the news?

The list of 99 authors of references Derek Hodgson cites in his meticulously academic essay does not include the following:

Gustaf Britsch, 1879-1923, whose work asserted and demonstrated that the mind in its struggle for orderly conception of reality proceeds in a lawful and logical way from the simplest patterns to increasing complexity.
Egon Kornmann, Britsch’s student and Director of the Britsch Institute, who in 1926 published Britsch’s book Theorie der Bildenden Kunst.
Henry Schaefer-Simmern, who until his death in 1978 researched and taught the intuitive problem-solving process of visual thinking, and who in 1948 wrote The unfolding of artistic activity.
Viktor Lowenfeld, who from 1940 to 1960 influenced world thought in art education with his Creative and mental growth and other books.
Moholy-Nagy and his rationale for art.
Herbert Read and his humanistic rationale for art.
John Dewey, author of Art as experience and other works on American thinking in aesthetics, who introduced Henry Schaefer-Simmern’s The unfolding of artistic activity.
Seymour Saracen, Yale University, Challenge of art to psychology.
Kurt Koffka, who in 1924 wrote The growth of the mind.
Conrad Fiedler, whose On judging works of art was written in 1876 and translated into English by Henry Schaefer-Simmern in 1949, and welcomed by Rudolf Arnheim, one of Hodgson’s references, and by Herbert Read who used it as basic theory in his 1953 lectures at Harvard.
Ellen Disanayake, whose What is art for? is composed of clear essays on the place of art in human evolution, and whose ‘Homo aestheticus’ is acclaimed by Nature as ‘probably one of the most intellectually enriching interdisciplinary studies that has ever been written’.

Hodgson must be aware of these antecedents to his discussion. If their absence from his essay dismisses the importance of this vast bank of basic discovery, we can expect that he will supplant it with some new thesis worthy of attention. But nothing new is identifiable and he instead constructs a compendium of achingly subtle assertions from 99 expert scholars. But he fails to make them clear, or to arrange them to instruct us and advance our knowledge. His difficult logical quibblings and minute dissections of a multitude of selected assertions are beautifully, almost poetically, dressed in scholastic process.
The last sentence in Hodgson’s ‘Conclusion’ reads: ‘Such [early] marks played an important role in the development of ways of dealing with visual information by recourse to Gestalt principles creating a platform for the emergence of graphic representation whereby symbolic interpretations did, eventually, begin to become relevant.’ His huge and serious effort is difficult reading and the reward for reading it is meagre — we already know that marks ‘played an important role’, and many of them. Readers are better rewarded by the references Hodgson has rejected, where they will find equal pioneering devotion to the subject and lucid stimulating discussion.
Sylvia Fein
190 Rolling Ridge Way
Martinez, CA 94553

Iconic interface between the worlds

According to Hodgson’s proposed evolutionary model, a long period of familiarity with forms in the physical world was precursory to sculptural representation. He suggests that this preliminary stage involved ‘passive appreciation of natural artefacts’ (Fig. 3) or ‘naturally occurring artefacts’ (p. 17). Oxymorons aside, I would certainly agree, seeing that I developed the case by way of the ‘natural representations theory’ published in this very journal (Feliks 1998a, 1998b) and widely distributed prior. [In brief, the ‘natural representations theory’ suggests that early humans made the cognitive correlation between living forms and iconic, ‘natural representations’ of living forms (fossils), priming the capacity for ‘artificial representation.’]
Hodgson proposes a separate lineage for graphic representation. Although this lineage follows the same preliminary stage as the sculptural lineage, Hodgson develops it as though it has no natural precedent. However, readily identified, two-dimensional iconic representations of three-dimensional forms, as well as ‘geometric’ forms, have long been established in the natural world (Feliks 1998a: 110-11, 116-20). My concern with Hodgson’s graphic marking evolutionary model is that its foundational assumptions are presented in such a way as to make them appear erroneously fact-like. Before we build too tall a structure here, I believe it is important to question the elevated nature of these assumptions and the criteria for what Hodgson calls ‘facts.’ We must also evaluate his dependence on over-generalisation and analogy as they relate to the certainty with which he expresses his conclusions. These matters are all intertwined, but for the sake of clarity, I will discuss them in a numbered sequence.

Assumption 1: Depictions of phosphenes in early art are not depictions of phosphenes
Hodgson comfortably refers to early markings as ‘phosphene/artistic, primitive forms’, ‘phosphene motifs’, ‘manifestation of phosphenes’ and so on. He even goes so far as to state that a particular Middle Palaeolithic motif reproduced in Fig. 1 can be ‘positively correlated with associated phosphene’ (p. 11). But if we accept these humanly-crafted marks as ‘immediately obvious’ images of phosphenes, can we be simultaneously certain that they are not representational? We are expected to accept at face value that Hodgson’s own drawings in Fig. 1, columns B and C, are deliberate depictions of phosphenes. Then, justified by etic presupposition, we are, conversely, asked to accept that the drawings in column A are not deliberate depictions of phosphenes. This double standard seems unavoidable, in that Hodgson requires the early graphic marks be accepted as ‘pre’-representational in order for his evolutionary scheme to make sense.
If phosphenes were entirely subconscious phenomena, Hodgson’s pre-representational interpretation of phosphene-like marks would be on firmer ground. However, the clinical motifs he reproduces were observed and depicted consciously. Since production of the palaeoart motifs likely involved many levels of conscious awareness (Bahn and Vertut 1997: 26), the possibility of intentional depiction cannot simply be ignored, even if it conflicts with a current popular view or with Hodgson’s own predispositions concerning early human creativity. It does not seem reasonable to presume that all such schema represent but a few low-end levels of conscious involvement.

Assumption 2: Phosphene motifs in early art are pure phosphene motifs
Due to the ever-present possibility of alternative explanations, archaeologists are usually careful not to place isolated instances of palaeoart into unequivocal categories. But Hodgson sidesteps this precaution; he groups together a wide range of sparse enigmatic motifs, and treats them as though their collective identification is somewhat certain. Over-generalising the evidence, Hodgson interprets nearly all ‘abstract schema’ in early art as indicative of phosphenes. To suggest that a single interpretation accounts for everything seems untenable to me, for not only are these motifs separated by hundreds of thousands of years, but perhaps more importantly, they were made by several different — including extinct — human species.
By focusing exclusively on the fact that every early mark (reported post-phosphene theory) ‘is’ a phosphene motif, Hodgson forgets that one fact does not cancel out another. Known from physical evidence, these very ‘motifs’ had long prior been concretely visible in the natural world via fossils (Feliks 1998a: 119-120). Taking this dual existence into account, I have earlier suggested (1995: 35-6, 1997a: 33-4, 1997b: 29) that if Palaeolithic people saw coherently stable images in the outer world which resembled the intangible and fleeting images of the inner world, the outer images would have substantiated the inner, and supplied a partial motivation for imitative, iconic expression of such in the physical world. Relatedly, hard-wired internal motifs may have been intermingled with live observation of natural lines, patterns, or forms — or with conscious (Lewis-Williams and Dowson 1988: 203-4), subconscious or cryptomnesic memories (specific or generic) of the physical world (Feliks 1995, 1997a, 1997b; 1998a: 116, 119-20; see also Bradshaw 1998: 126). Hodgson conveniently bypasses all of these possibilities, perhaps because they make his evolutionary model appear less conclusive.

Assumption 3: Graphic marking ‘evolved’ from non-symbolic Lower Palaeolithic forms into Upper Palaeolithic representational forms
I agree that archaeological evidence supports a pre-Upper Palaeolithic scenario for the origins of art, but I question Hodgson’s tenet that non-symbolic expression necessarily preceded or evolved into symbolic expression. Since a non-symbolic interpretation of early graphic schema is not established (and certainly not if cognitive archaeology insists upon scientific criteria and physical evidence), it stands that its progeny, an abstract-to-representational evolutionary sequence is not established either. Yet, in requiring that geometric marks play a precursory non-symbolic evolutionary role, Hodgson treats it as ‘fact’ that pictorial representation came about before symbolic representation (p. 5). Objectively, are we really talking about ‘facts’ here?
As much as Hodgson tries to make a distinction, I have difficulty reconciling his admitted ‘capacity for symbolic behaviour’ in early hominids with his but-not-in-graphic-depiction model (p. 6). First of all, it is common knowledge that symbolism is not dependent upon one’s ability to physically create symbols but simply one’s ability to project symbolic meaning onto something — anything — whether by resemblance, convention or mere association. (Ready-made natural symbols in the form of metonymical and synecdochical substitutions were likely employed long before any attempt to create symbols artificially.) Hence, symbolic meaning could easily have been projected onto the simplest graphic marks. Secondly, let us keep in mind who we are talking about here. These were people capable of creating fire; co-operatively building extremely large free-standing structures such as those at Terra Amata; astounding ocean voyages (Bednarik 1997; Morwood et al. 1998); three-dimensional sculpture; and, as Hodgson apparently accepts, language! Add to all of the above the likely recognition of natural iconic representations (Feliks 1998a; Bednarik 1998: 124-5, ‘invention of a referent’), and you have a pre-Upper Palaeolithic symbolic world of immense diversity.

Assumption 4: The artistic development of modern infants reflects human cognitive evolution
Hodgson suggests (based on cited studies) that modern-day infants naturally progress from abstract to representational drawing, and that this is an analogy for how humanity as a whole (encompassing several species) progressed artistically. However, these children do not go through artistic developmental stages autonomously. Every step of the way, they are effectively guided by researchers who provide them with all the materials and circumstances needed for their mark-making ‘evolution’ (perfectly flat, smooth paper; modern, easy-flow marking implements; comfortable, artistically supportive surroundings; focused, sequential sessions; and plenty of time). Consider the ‘evolution’ of chimpanzee painting skills similarly. Putting it another way, both children and chimpanzees are subtly encouraged to do what their supposedly non-influential guides already know how to do!
Though analogies are useful, we must be careful what we extract from them. We cannot, as Hodgson attempts, isolate a single ingredient from culturally-informed compound studies of modern infants or chimpanzees, treat it as though it were arrived at naturally, and then confidently project the results half a million years into the past, and onto different, not to mention, extinct, species. As much as we would prefer otherwise, definitive real-time observation and prediction testing are simply not possible in cognitive archaeology, and any statements of confidence in analogical testing or retrospective prediction should reflect this fact (Feliks 1998b: 129).

Final thoughts
The origins of art issue cannot be reduced to a simple question of neuropsychological evolution. Alongside hand/eye/brain considerations are variables of conscious, subconscious, and cryptomnesic creative activity and innovation within myriad external environments. Hodgson’s exclusive approach focuses on part of the issue as though it is the entire issue; concerning alternative, even complementary explanations, Hodgson is content to simply ignore them. In my view, the advancement of cognitive archaeology owes a great deal to an abundance of influential, interrelating ideas. Hence, I still believe that a genuine understanding of our ancient predecessors will involve all reasonable theories. With a little tempering, Hodgson’s model can be a valuable part of this comprehensive whole.

John Feliks
32619 Dover Ave
Garden City, MI 48135-1669

Reason, emotion and graphic representation

Hodgson has certainly put forward a challenging series of ideas about the implications of early rock art for our understanding of human evolution. The purpose of this note is not so much to disagree with the thrust of his work, but rather to suggest ways in which his model may well be further extended and made more comprehensive.
I can only agree completely with his rejection of the cultural explosion scenario proposed by Mithen (1996) and many others and his insistence upon a gradualist model of evolution. After all, this fits with the evidence from studying other members of the animal kingdom.. However, I will return below to a re-consideration of Mithen’s concept of cognitive fluidity.
By concentrating upon the evolving capacity of the mind to produce order out of disorder and arguing that the processing of perception as information must have preceded aesthetic pleasure, Hodgson seems to be locked into the long-standing Cartesian separation of reason and emotion, mind and body. There is indeed an almost equally long-standing and wide-ranging rejection of this principle, perhaps first enunciated by Emmanuel Kant (1781) in his Critique of pure reason, and most recently by the neurologist Antonio Damasio (1994, 1999).
My personal assumption would be that the conjunction of cognition and reason on one hand and affect and emotion on the other must have always been, that one cannot normally exist in isolation from the other, and in fact, that the road to survival depends upon the balance between the two. Since considering the very important ideas which Hodgson has advanced, I now look back to a more basic and broader consideration within the evolutionary process and think about the evolution of consciousness rather than the evolution of information processing as that term is normally construed.
Mithen conceptualises the mind as having evolved from specialised and modularised functioning (the ‘Swiss Army knife’ brain) to cognitive fluidity and integrated generalised functioning. This notion is not, in itself, dependent upon the cultural explosion scenario, but might equally well be accommodated within a gradualist model. Further, Damasio (1999), drawing upon both systematic research and clinical experience in neurology, distinguishes between a continuing ‘core consciousness’ which serves to locate us within our environmental context (the here and now) and ‘extended consciousness’ within which we understand our own lived past, our cognitive recognition of the world about us and our anticipation of a future. This too implies an evolutionary development from one to the other, yet maintaining the core consciousness as a vital base upon which evolution has built. My earlier comments (Hamilton-Smith 1986) about the role of stimulus in the production of images and the potential place of stimulus-arousal theory (Berlyne 1971), also referred to by Hodgson, are probably very relevant here, and even more pertinent within the context of extended consciousness.
Hodgson’s ‘causal integrated’ model of functioning seems to have a great deal in common with both concepts of cognitive fluidity and the extended consciousness. The extent of potential agreement underlying these three sets of ideas, each with their own semantic expression, might well be seen as strong support for the basic principles in interpreting the evolution of art as one expression of the evolution of the human mind. At the same time, art may well be much more than information processing, but rather a record of evolving consciousness and the evolving individuality of self which followed.

Professor Elery Hamilton-Smith
P.O. Box 36
Carlton South, Vic. 3053

Art, perception and (archaeological) information-processing

My views of Palaeolithic ‘art’, like those of many archaeologists, have undergone a series of perceptual flips since I first became acquainted with the problem as an undergraduate — in a neat exemplification of the properties of the perceptual process in general. From an initial expectation that art, like toolmaking, should be a story of gradual progress, I came to recognise that the earlier phases (associated with pre-modern species of the genus Homo) had to be approached as something very different from anything about which ethnography could inform us, and I consequently tended to set up a rigid dichotomy between art-making and behaviours of a kind more widespread in the animal world. More recently, I have come to recognise to what extent this ‘revolutionary’ interpretation was based upon a possibly atypical evolutionary episode, namely the replacement of Neanderthal by modern humans in the (then rather marginal) area of western Eurasia; and as a result I have been more alert to the possibility that what is classified as ‘art’ is not necessarily only of Upper Palaeolithic date, and that the whole issue of visual communication must be situated in a much more deeply-rooted process (Sherratt 1999). This personal intellectual history would be of merely anecdotal significance were it not both typical, I suspect, of many practising archaeologists, and moreover forms part of an editing process by which archaeological finds reached a wider community of specialists: for such perceptions impose a filter on the way in which primary observations are packaged for a non-archaeological readership. The importance of this paper is to show how we can retain some of the insights of earlier views, and yet escape from the perceptual constraints which each of them imposed. In short, it explores how to construct a bridge from Darwinian to cultural ways of thinking about the problem.
Like many important problems, it is one to which many of our usual linguistic habits are inappropriate. (‘Art’ is bad enough when applied to medieval and ancient times, if it carries over the common associations of Renaissance attitudes to earlier contexts: the term ‘artist’ should arouse immediate suspicions!) It is obvious that a period of evolutionary acceleration is going to place great strains on any terminology adapted to particular phases of the process. The great merit of Hodgson’s model is to divorce structured marking from representation, and to situate the discussion in motor habits and neural organisation. This helps to overcome the natural tendency to categorise early behaviours simply in terms of familiar contemporary practices: Neanderthal ‘burials’ were certainly intact bodies in earth-covered hollows, but not necessarily the products of deliberate ‘funerary’ practices — they could be insulated sleeping-hollows in which individuals might from time to time (quite often, at the height of a glaciation?) be expected to die, and perhaps be deliberately covered by survivors. Many archaeological phenomena are of this kind, for which everyday ‘common-sense’ descriptions are inherently problematic; and it is a problem shared with any study of animal behaviour (to which Watsonian behaviourism was an extreme solution), in finding an appropriate set of descriptive categories. The answer, in both cases, is to specify the evolutionary conditions in which the phenomenon might have arisen, and the possibly different conditions in which it might be retained or promoted.
The manufacture of artefacts (typically of wood or other organics, even though only the lithic components of the whole process are likely to survive in the archaeological record) involves repetitive actions, whose structured sequences are not only a means of altering the natural world but also have an existence in their own right and are thus transferable to other contexts — some of which may be more connected with communication than immediate material transformation. Selective pressures are as likely to operate in the former context as in the latter; but at different times may act on both. Hodgson’s account tackles the more fundamental level, at which ‘commonsense’ categories (like ‘toolmaking’ or ‘dance’) may be irrelevant distinctions. It is precisely the elaboration of such schemes in practice which allows the differentiation to take place; but both took place within a structured pattern of rule-making (and rule-breaking) which gave coherence to performance. The partly independent processes of (proto-)linguistic performance, and the (increasingly overlapping) cognitive capacities that they imply, can have had a complex history of sometimes separate and sometimes linked development — each arising from the potential social complexity of primate groups, and more specifically from the unique features of extended immaturity and novel adaptive stance (meat-getting) of ancestral human lineages. Infant art rightly takes its place in these discussions, since prolonged infancy is itself a distinctive human characteristic, and provides a degree of insulation from the kinds of selective pressures which might otherwise inhibit modes of cognition and behaviour not immediately relevant to the task in hand. (This is not just of methodological interest, comparable to the ontogeny/phylogeny relationship, but important in providing a context for the development of play-like behaviours which are fundamental to human cognition and sociality.)
All this is to some extent word-play (on the part of those of us whose expertise lies in analysing the archaeological record, rather than understanding the workings of the brains which produced it); but such play is important, since it affects the way in which we process the observations we make in the course of ‘real’ work. In particular, it helps to rescue the gradualist description of the co-evolution of brain and behaviour (including material-transforming behaviour) from the naive assumptions of the traditional, technology-oriented account; but it also avoids the rigid stadial categorisation which has often accompanied more recent accounts of the emergence of ‘artistic’ activity. Both models convey important aspect of the unique evolutionary process which gave rise to modern humanity; but both are too simple to be satisfactory ways of thinking about the process. It is the continuity from pre-symbolic to symbolic material behaviour which must be at the heart of a contemporary model of the pre-History of ‘art’. I enjoyed this presentation of the problem, and welcome the broadening of debate which it invites.

Dr Andrew Sherratt
Ashmolean Museum
The University of Oxford
Oxford OX1 2PH
England, U.K.


Independent and overlapping trajectories relating the simple to the complex

I thank reviewers for their discerning and mostly favourable Comments. My Reply will be to each contributor in turn. Where similar points have been raised by separate scholars I direct readers to my first response unless otherwise stated.

Bednarik’s positive appraisal is encouraging as one of the aims of the essay was to show how the architecture of the brain, and children’s early mark-making, can be related to phosphene theory in terms of the inception of an externally generated, informational-processing system as embodied in the first ‘art’ (further thoughts on phosphene theory are outlined below).
I accept the point concerning the artificiality of the division between U. P. and L/M. P., which can lead to an exaggeration of the separation between representational and non-representational art, when the reality would have been one of successive modification according to a more incremental agenda. Unfortunately, in conforming to accepted terminology such ambiguities are often unavoidable. I affirm, however, the proposition that a distinction can be made between proto-art and representational art, as the former can be regarded as a signifier of a different order to that of the latter; that is, significant according to arousal factors involving bottom-up rather than top-down determinants. From this standpoint Palaeolithic ‘art’ is conceived as the invention of a means of organising and encapsulating visual information commensurate to how this is achieved in the brain and, therefore, has universal correlates. This distinction does not presuppose any disassociation between the M. P. and U. P. but rather a gradual accumulation of depictive strategies during these periods which eventually led to iconic representation.
The reservations concerning the validity of some of the cited artefacts does not in any way compromise the main thrust of the argument, but I thank Bednarik for bringing this to my attention.
As far as the experience of a ‘hyper-reality’ state is concerned, this was merely to stress the fact that the making of geometric marks stimulates certain visual centres of the cortex. Prolonged absorption in this activity through intense visual concentration (which is common in artists), might, therefore, give rise to a heightened sense of reality, thereby imbuing such forms with a capacity beyond the initial cause. In addition, fatigue to specific neural mechanisms, as a result of the prolonged viewing of a single kind of geometric form, can lead to the greater likelihood of perceiving other kinds of geometric forms (called negative after-effects) through differential thresholds of neurones in the visual cortex being either raised or lowered respectively (Favreau and Corballis 1976).
In defence of the resonance proposition, Smets (1973) found that psychological arousal was greatest when the redundancy-repetitiveness of design elements was about 20 per cent — this is equivalent to the amount of order found in a simple maze, in two complete turns of a logarithmic spiral, or a cross with asymmetrical arms. This phenomenon seems to be innate as drawings with this degree of order tend to be viewed longer by new-born infants.
Concerning the decrease in brain size due to cultural influences, Martin (1990) has made the point that the brain of the early modern human averaged 1450 grams whereas the average for contemporary humans is only about 1300 grams. Allman (1999: 207) has postulated that this may be due, from 35 000 years onwards, to the increasing effects of culture as a buffer against adaptive pressures — which is analogous to how the brains of dogs have decreased in size by twenty per cent, compared to that of ancestral wolves, as a function of domestication, other variables being held constant. The question of the relationship between brain size and culturally derived information processing is, admittedly, not a simple one, yet Donald (1991: 115) and McNamara (1999: 4) have also drawn attention to how the two components interact in ways which may influence brain growth.

In relation to Bradshaw’s criticism of the concept of ontogeny recapitulating phylogeny: to reject this would be to reject an idea which has, implicitly or explicitly, been used as a justification for many theories concerning cognition and cognitive development, e.g. Parker (1985, and a debate on this topic, p. 634); Parker and McKinney (1999); Karmiloff-Smith (1992); Mithen (1998). In support of this idea, a newborn baby has a vocal tract like a non-human mammal and it is not until three months of age that its larynx has descended enough for the first adult-like vowel sounds to be produced (Pinker 1994: 265). Allman (1999: 197) has also shown how the chimp and human skull are very similar at the foetal stage. Furthermore, as indicated in the text (McNamara 1999), this concept, in a modified form, has recently received considerable confirmation. As direct archaeological evidence regarding the significance of palaeoart is unlikely to become available we are obliged to fall back on any clues which can be gleaned from paradigms that can provide suggestive hypotheses.
Concerning particular regions of the cortex that may also be involved in object recognition, Bradshaw’s comments are particularly astute. As the paper concentrated more on early mark-making it was inevitable that the striate cortex should figure so prominently. However, the significance of the association cortex was cited in the text, but a consideration of this and other areas would have been beyond the scope of the essay. The regions specified by Bradshaw are pertinent, but more to the later processing cascade involved in object recognition than earlier primitives — which seems rather to reinforce my main thesis.
Regarding the doubts concerning ‘resonance’, I would refer readers to Zeki (1999), who has convincingly demonstrated how this idea can be applied to contemporary art (both abstract and representational) — and how artists, as unwitting scientists, have been engaged in the deconstructing of the image as assembled by the brain, albeit in this case backwards from representational to abstract art. What is compelling in this account is that substantial, scientific evidence is presented to back up the claim, especially in the form of brain scans which have pin-pointed some of the areas to which Bradshaw alludes as a later, higher-order functional integration of object features. Interestingly, Zeki has charted the pathways through the visual system for colour (from abstract to object recognition) in a similar way to how this seems to operate for form.
As to autistic children such as Nadia, Gardner (1980: 187-9) has indicated that we simply do not have enough information to answer the question as to whether she passed through the normal stages of drawing development — though some of her drawings from the ages of three and four suggest that this was, indeed, the case.
The point is well made concerning the possibility that art may constitute an end in itself — this, however, can be related to the initial private experience and immediate reflexivity involved, of which the product inevitably becomes part of the public domain. Thus, the earliest marks, although probably deriving initially from personal experience, were realised in a material form that would eventually become significant for the group. Most art, therefore, fits into the category of shared artistic experience, even though this might sometimes appear obscure.
Bradshaw asserts that, far from being adaptive, art may actually be a spin-off (‘spandrel’) as a function of increased brain size. Dissanyake (1992: 230) has made the point that spandrels may not, after all, be spin-offs, rather they were one of the reasons cathedrals were built! Contrary to Bradshaw, who asserts that we attempt to find a Darwinian explanation for all things human, Dissanayake draws attention to the fact that Gould (Gould and Lewontin 1979), curiously, shares the view of some more conventionally minded adaptationists that art, in fact, has no survival value. This position reflects those who, in still hoping to save art from a functionalist or materialist analysis, fabricate scenarios whereby art is seen as if to emerge from nowhere. It is because there has been a lack of plausible attempts to interpret art within the larger framework of evolutionary history (as Dissanyake has indicated; 1992: 9) that such proposals continue to be taken seriously and, in addition, it is a view that suggests a form of dualism, which serves only to obscure rather than elucidate. As there is now a growing corpus of artefacts which attest to the gradual development of art over a considerable period in line with brain expansion, this alone seems to disqualify the notion of a spandrel.
I regret the fact that Bradshaw feels that I might be falling into the trap of evolutionary teleology, but perhaps I did not make my credentials absolutely clear in quoting Dawkins (1991). The extracts Bradshaw cites from the text are largely used figuratively to convey the idea of successive approximation according to an autocatalytic process rather than any overt ‘purpose’.
Concerning the extent of the visual brain, this point was made merely to emphasise the general importance of visual information for the brain. I agree that the structure of the visual system at lower levels may be similar across species, which supports my main point concerning evolutionary constraints in relation to early mark-making (e.g. chimps); the differences between humans and other species are probably due to higher-order faculties, which is precisely the point I was making in to regard representation in the U. P. apropos an ‘aesthetic sense’.
Regarding the importance of survival activities in relation to protoart, hominids, as inquisitive creatures par excellence, would have become unsettled at the prospect of inactivity thrown up by the benefits accruing from improving technologies. Rather than spend this time on further social rituals (for which the need would have been reduced by the efficiency of communication that came with the rise of speech), it seems that mark-making provided a suitable and convenient alternative outlet. Inactivity as identified in other species may have an evolutionary significance which has not yet been identified.
Mathematicians may use the word ‘redundant’ to describe predictability, Gombrich, however, was using the term to denote boredom and monotony in the perception of repetitive lines and in this sense the word may be regarded as ‘pejorative’. This was included simply to focus attention on the difference between Gombrich’s hypothesis testing approach and one based more on innate criteria.

Faulstich’s comparison between my ‘seeking order in the midst of disorder’, and the hypothesised order of U. P. depictions I regard as tenuous. I suspect a structuralist agenda is being adopted here; which is, in this case, at variance with an evolutionary-biological position. Most commentators tend to agree that the largest perceptual unit in parietal art is a single animal, e.g. Hudson (1998). Bahn and Vertut’s statement regarding a possible grammar, cited by Faulstich, is self-contradictory!
I refer Faulstich to Berlyne (1970, 1971) and Kagan (1970) as cited in the text, in response to how innovation in art can be generated.
Concerning the development of linear perspective, I agree that this most certainly was an acquired cultural convention, but that it was a convention derived from a previous array of artistic innovations, involving a set of self-contained rules relatively independent of language. Interestingly, visual rules of this kind can still be found in many adult ‘learning to draw’ books where simple geometric shapes and lines akin to phosphene forms constitute a necessary practice stage as a precursor to representational depiction involving perspective. Furthermore, I confirm that expressive culture is linked to biology in intriguing ways — specific graphic primitives are embedded in representational art without the observer being fully aware of the fact (this is akin to the word superiority effect).
Although I concur with the notion that there is often a tendency to exaggerate human traits, Faulstich’s view that speech may not be such a special human attribute can be accommodated within the present model by regarding complex human speech as a species-specific adaptation no different to various attributes inherent to other species (Pinker 1994: 342). Complex speech as a human propensity is emphasised by Deacon (1997) in posing the question: why can’t animals talk like humans? This, of course, is not to say that intelligence is dependent on the ability to speak and the pitfalls of trying to assess intelligence (which is a construct) in humans is difficult enough, let alone animals. These reservations aside, the case can still be made that the first utterances, as words, may have represented an initial, pro-active concern with information processing at a level that was intrinsic to humans as a species — but with the admonition that this was premised on a continuity with modes of communication found in other primates.
With respect to the idea that nature represents a hostile force which has come to be dominated by humankind: I accord with the view that this can appear as Panglossian — a Darwinian perspective disallows the concept of progress in these terms. At the same time this does not invalidate the notion of descent with a tendency for modification and increasing complexity (Wilson 1998; Munro 1963), a concept from which the human ‘illusion’ of ascendancy seems to have been erroneously derived (Midgley 1985). Culture, in conforming to this ‘rule’ through co-evolution, seems to imply progress but it is a progress to which the evolutionary dictates of the natural world remains essentially indifferent. This, however, does not discount the fact that culture originally arose out of evolutionary imperatives as a novel way of solving problems to do with day to day survival.

Fein will be pleased to see that I have, above, alluded to one of her suggested references. In fact, some of these references I quoted in an earlier draft (Read’s [1965] thoughts on these issues are particularly interesting) but, unfortunately, space did not allow for their inclusion. I would like to point out that most of the authors listed by Fein have assessed art either from a more general anthropological stance, without considering the recent data concerning mark-making before the U. P., or do not take into account, in any great detail, the neuroscientific implications for art. Nevertheless, I thank her for bringing these authors to the attention of readers.

Paradoxically, Feliks might be interested to know that there is such a thing as a ‘naturally occurring artefact’ — i.e. the Makapansgat jasperite cobble; a found object which, although not in a strict sense an artefact, became one by virtue of perceptual projection which transformed the object into a semblance of a face! Incidentally, the representational significance of this form may not be as obvious as has sometimes been reported; Coss (1985: 256), for example, has established that the signal of two facing eyes, relating to the identification of predators, is embedded in the overall, innate, neural substrate of humans (thus supporting Bednarik’s criticism of Feliks [1998: 125]).
I agree that natural forms would have provided a potent stimulus for passive appreciation before the first sculptural artefacts were produced with a particular status for fossils (the West Tofts handaxe seems to confirm this). It is apparent that repeated lines and symmetry are characteristic features of many fossils and, therefore, would be more likely to invoke striate cortex structures — but I am averse to the proposal that this led directly to iconic representation through what Feliks describes as a ‘priming’ effect. Rather, a non-symbolic model (viz, self-sufficient marks of Davis [1986]) would propose that early mark-making can be adequately explained by how neurones in the striate cortex are excited in response to such activities, which leads to a feedback mechanism involving elevated arousal levels by way of brain structures such as the reticular activating and limbic systems — thereby increasing attention and focal vision and providing the motivation for further, similar engagements. One of the purposes of this essay was to provide documentary cognitive and neurophysiological evidence to support this hypothesis (which is one of the few ways through which the archaeological record can be appropriately assessed).
The lineage for abstract mark-making (Fig. 3) is defined as linked to tool-making, rather than the passive appreciation of natural objects, to emphasise the fact that the invention of mark-making may have been instigated by accidental cuts or striations made in the making and using of tools which stimulated an ‘aesthetic’ response. The representation of natural objects, in a two-dimensional format, according to this account, could, consequently, only have taken place during the U. P. — but this does not exclude the capacity for a passive appreciation of natural objects before this period.
Although phosphenes experience may provide a focus for representation through externalisation, and therefore a compelling explanation for their manifestation in some forms of ‘art’ by way of intentionally, this notion can readily be included in the model concerning the striate cortex, where the rendering of such features need not be, in the first instance, a prerequisite. Consequently, this phenomenon can be explained in more parsimonious terms by proposing that the striate cortex, through resonance, provided the original impetus for the first marks from which phosphenes might later have become a subject for depictive activities.
I concede the point concerning some of my conclusions being drawn from a disjointed archaeological record, but as evidence from the periods under discussion is limited in the extreme one is bound to fall back on over-generalisation. In defence, I draw attention to the caveat that a broader overview consistent with an evolutionary analysis was to be expected.
In response to Feliks’s inability to conceive of a difference between the capacity for symbolic behaviour in speech as opposed to the ‘meaning’ attributed to L. P. marks, I refer to the fact that speech and imagery are asymmetrically located in the brain. As representation is more likely to be constrained by the latter faculty rather than the former, the visual brain would, therefore, have provided the primary instrument by which full representational depiction was achieved — at which point speech began to attribute meaning. It has been established that representation, as processed by the brain, can have a ‘logic’ independent to that of analytical-symbolic thinking (Corballis 1989; Solso 1993; Damasio and Damasio 1992: 63; Davis 1986: 210 for a discussion on this issue). It is the ‘logic’ of this system which is being evoked in protoart. This supports Davis’s (1993: 343-4) assertion that spoken language need not be a prerequisite for other kinds of communication. He also notes that the marks on the bones from Bilzingsleben might constitute a necessary factor providing the raw material for what will later become pictorial image making — which does not presuppose any particular symbolic expertise. Consequently, Feliks’s ‘projection’ would be one of significance being apportioned through a non-specific increase in arousal levels. Moreover, Halverson (reply to Bednarik 1995a: 621) has suggested that, for early hominids, prototypical marks need not have had any obvious meaning beyond a ‘pongid status’ — although I would still hold to the view that what, in essence, chimps, infants and hominids respond to, in the production of geometric lines, is the perception and construction of order and symmetry according to evolutionary factors as outlined. Pertinently, representation, as realised perceptually in the brain, would have preceded symbolic systems as a necessary strategy for survival through the visual mapping of the environment — Corballis refers to this as a ‘default’ system (1989: 501). (The above analysis is also relevant to Bradshaw’s point concerning pre-linguistic informational transmission.)
Concerning cultural influences in children’s drawing, the universality by which children progress through the various stages towards representation argues in favour of a predisposing mechanism which is independent of cultural factors. The similarity with the art of chimps confirms the point.
The theory as presented is one that puts forward a hypothesis which is eminently open to falsification in the light of further archaeological discoveries. I do not, I think, as Feliks suggests, attempt to provide an exclusive explanation for the origin of art — as indicated this is an inclusive theory which leaves open the question of cultural determinants.

The issue of consciousness raised by Hamilton-Smith is an interesting one. As a corollary to the model proposed here, consciousness would be defined as having evolved as an increasingly efficient mode for integrating diverse forms of information — as a kind of shorthand-way for systematising the complexity of incoming stimuli. This does not imply that a hierarchical, but parallel, cascade of re-entrant processing ends up at a central processor where consciousness resides, or that this consciousness arises as an epiphenomenon of the neural complexity in higher centres of the brain; rather this originates in the cumulative interlocking of more complex, modal, sub-systems at macro as opposed to micro levels in ways which we do not, as yet, fully understand. This approach is not wholly at odds with Damasio’s view (1989), although I would not go as far as to say that consciousness operates right up to the boundary where the inner meets the outer world. But here we get into arguments as to whether what is processed automatically is open to consciousness. In this respect connectionist models would be relevant, which can help to delineate how the striate cortex, as a lower-order function, might analyse visual information. Certainly, I concur with the view that the separation between mind and body is artificial; as a theory which places the beginnings of art firmly in the neurostructures of the brain the inference is that consciousness will also, eventually, be similarly explained. These points obviate any accusation of a Cartesian separation as a continuity rather than a discontinuity model is presupposed.
The emotional centres of the brain would, of course, have played a role in information processing which, as already mentioned, can be linked to arousal as Hamilton-Smith has, himself, noted (1986) in referring to ‘intrinsic hedonic value’ — this provided the drive mechanism through which the first marks were fashioned whereby the striate cortex would have played an important determining role. The two positions are, therefore, mutually sustaining and conveniently integrate Davis’s self-sufficient marks and Bednarik’s phosphene theory.
Where I do part company with Hamilton-Smith is with reference to Mithen’s ‘cognitive fluidity.’ Mithen seems to have the principles of evolutionary psychology back to front; it is not cognition which suddenly became fluid but the flow of information as it was differentially, but exponentially, realised through culture (see Fig. 3) as a derivative of cognitive structures which had probably been in place by at least 150 000 BP. In this respect Mithen (1996: 154) misinterprets Sperber’s theory (1994: 55, 62), the latter of which is more in line with evolutionary psychology in seeing culture as constrained by long established cognitive components (Cosmides et al. 1992: 5) — well before Mithen’s supposed cognitive threshold. I note that Mithen’s support for a cultural explosion at the M.P./U.P. boundary has cooled somewhat recently in admitting the possibility of a longer time frame (1998: 186).

I commend Sherratt for highlighting how an approach to the understanding of palaeoart from the perspective of neural function and Darwinian imperatives can reap rich rewards for archaeology. This not only helps to establish a bridge linking evolutionary studies to cultural theories but also how early mark-making can relate to iconic and symbolic processes.
Sherratt’s shrewd observation that the making of artefacts involves repetitive actions, which would have been transferable to other contexts including those connected with communication, gives substance to the idea that the significance of self-sufficient marks could have derived from the accidentally produced scratches or cut marks created in the defleshing of bone or in the making and using of tools. He also cogently summarises how the putative overlapping processes can have different levels of integrative complexity according to particular contingencies.
I endorse Sherratt’s observation which highlights the importance of prolonged infancy for cognitive development in humans; although this may provide for a measure of insulation from immediate selective pressures, the length of the extended period itself would still have been mediated by these pressures. Interestingly, Allman (1999) relates this latency to information processing. He views prolonged vulnerability as connected to the need for offspring to acquire the essential knowledge for survival in the face of environmental uncertainty. Speech, therefore, developed as a means of communication to promote and sustain the extended family, the latter of which acted as an agency to provide a secure background in the light of both this need and the associated larger brain size.

Résumé. Dans les années récentes, les sites du Paléolithique inférieur et moyen ont livré de plus en plus d’exemples de marques non figuratives. Les théories sur l’interprétation de telles marques cherchent invariablement à les expliquer par l’intention de représenter un référent. Elles écartent par là la possibilité d’une analyse basée sur des critères évolutifs découlant d’un scenario graduel, cumulatif, qui n’a pas nécessairement à dépendre d’explications symboliques. L’article vise à corriger ce biais interprétatif: on cherche à déterminer l’époque initiale où la confection de marques peut être suffisamment expliquée par des mécanismes évolutifs conjugués avec la manière dont le cerveau et le cortex oculaire traitent l’information reçue. Lorsqu’on aborde le problème de l’origine de l’art à partir de ces perspectives, on perçoit des similitudes avec la manière dont le cerveau traite des types d’information concomitants, par exemple le langage. Ceci peut nous fournir des indications intéressantes sur la façon dont la confection de marques a pu se développer, et dans quel but.

Zusammenfassung. Abstrakte Zeichen des Alt- und Mittelpaläolithikums sind in der letzten Zeit in archäologischen Niederlagen mehr verfügbar geworden. Theorien betreffend der Bedeutung solcher Zeichen haben sich stets auf beabsichtigte Bezogenheit konzentriert und ignorieren dadurch die Möglich-keit einer Analyse, welche auf evolutionären Kriterien basiert ist, die sich auf eine allmähliche, kumulative Entwicklung beziehen, die nicht notwendigerweise von symbolischen Erklä-rungen abhängen muß. Dieser Aufsatz sucht ein Gleichgewicht herzustellen, indem er bestimmt, wie frühe Zeichen-Herstellung ausreichend erklärt werden kann, durch evolutionäre Mecha-nismen in Verbindung damit, wie das Gehirn und die visuelle Hirnrinde Information verarbeiten. In der Untersuchung der Ursprünge von Kunst aus dieser Perspektive werden Parallelen damit gezogen, wie das Gehirn mit begleitenden Arten von Information fertig wird, z.B. Sprache, was wichtige Anhalts-punkte geben kann, warum und wie sich die Zeichen-Herstellung entwickelt haben könnte.

Marcas abstractas del Paleolítico Inferior y Medio recientemente se han vuelto más frecuentes en los registros arqueológicos. Las teorías respecto al significado de tales marcas invariablemente han estado relacionadas con una referencia determinada, y por consiguiente, ignorando la posibilidad de un análisis basado en criterios evolucionarios con un argumento gradual y cumulativo que no necesariamente tiene que depender de explicaciones simbólicas. Este artículo pretende rectificar el equilibrio al determinar cuán temprano la hechura de marcas puede ser suficientemente explicada a través de mecanismos evolucionarios junto con la manera en que el cerebro y el sistema visual procesan la información. Al tratar los orígenes del arte desde esta perspectiva, se sacarán semejanzas respecto a cómo el cerebro procesa tipos concomitantes de información, p. ej. el lenguaje, que puede proporcionar importantes indicios en cuanto a por qué y cómo la hechura de marcas puede haber desarrollado.


ALLMAN, J. M. 1999. Evolving brains. Scientific American Library, New York.
ARNHEIM, R. 1971. Visual thinking. University of California Press, Berkeley.
ARNHEIM, R. 1974. Art and visual perception. University of California Press, Berkeley.
ASLIN, G. D. and R. G. BEDNARIK 1984a. Karlie-ngoinpool Cave: a preliminary report. Rock Art Research 1: 36-45.
ASLIN, G. D. and R. G. BEDNARIK 1984b. Koorine Cave, South Australia. Rock Art Research 1: 142-4.
ASLIN, G. D., E. K. BEDNARIK and R. G. BEDNARIK 1985. The Parietal Markings Project — a progress report. Rock Art Research 2: 71-4.
BAHN, P. G. 1994. Lascaux: composition or accumulation? Zephyrvs 47: 3-13. [RBG]
BAHN, P. G. 1998. Prehistoric art. Cambridge University Press, Cambridge.
BAHN, P. G. and J. VERTUT 1997. Journey through the Ice Age. Weidenfeld and Nicolson, London.
BARLOW, H. B. 1972. Single units and sensation: a neurone doctrine for perceptual psychology? Perception 1: 371-94.
BEDNARIK, R. G. 1984. On the nature of psychograms. The Artefact 8: 27-33.
BEDNARIK R. G. 1990. On neuropsychology and shamanism in rock art. Current Anthropology 31: 77-80.
BEDNARIK, R. G. 1992. Palaeoart and archaeological myths. Cambridge Archaeological Journal 2: 27-43.
BEDNARIK, R. G. 1993. Palaeolithic art in India. Man and Environment 18: 33-40.
BEDNARIK R. G. 1994a. The Pleistocene art of Asia. Journal of World Prehistory 8(4): 351-74.
BEDNARIK, R. G. 1994b. On the scientific study of palaeoart. Semiotica 100(2/4): 141-168. [RGB]
BEDNARIK, R. G. 1995a. Concept-mediated marking in the Lower Palaeolithic. Current Anthropology 36: 605-34.
BEDNARIK, R. G. 1995b. Metamorphology: in lieu of uniformitarianism. Oxford Journal of Archaeology 14(2): 117-22. [RGB]
BEDNARIK, R. G. 1997. The origins of navigation and language. The Artefact 20: 16-56. [JF]
BEDNARIK, R. G. 1998. Debris of the Biblical Flood. Comment on J. Feliks, ‘The impact of fossils on the development of visual representation’. Rock Art Research 15: 124-5. [JF]
BEDNARIK, R. G. 1999. Maritime navigation in the Lower and Middle Palaeolithic. Comptes Rendus de l’Académie des Sciences Paris 328: 559-63. [RGB]
BEDNARIK, R. G. and M. KUCKENBURG 1999. Nale Tasih: Eine Floßfahrt in die Steinzeit. Thorbecke, Stuttgart. [RGB]
BEE, H. 1992. The developing child. Harper Collins, New York.
BERLYNE, D. E. 1970. Novelty, complexity and hedonic value. Perception and Psychophysics 8: 279-86.
BERLYNE, D. E. 1971. Aesthetics and psychobiology. Appleton – Century-Crofts, New York.
BOULTON, M. J. and P. K. SMITH 1992. The social nature of play fighting and play chasing: mechanisms and strategies underlying cooperation and compromise. In J. H. Barkow, L. Cosmides and J. Tooby (eds), The adapted mind: evolutionary psychology and the generation of culture, pp. 430-444. Oxford University Press, Oxford and New York.
BREUIL, H. 1952. Four hundred centuries of cave art. Centre d’Études et de Documentation Préhistoriques, Montignac.
BRADSHAW, J. L. 1998. Sermons in stones: fossils and the evolution of representational art. Comment on J. Feliks, ‘The impact of fossils on the development of visual representation’. Rock Art Research 15: 125-7. [JF]
BRIGHOUSE, G. 1939. Variability in preferences for simple forms. Psychological Monographs 51(5): 68-74.
BRINDLEY, C. S. and W. S. LEWIN 1968. The sensations produced by electrical stimulation of the visual cortex. Journal of Physiology 196: 479-93.
BRODY, C. 1970. The development of visual aesthetic preferences in young children. Sciences de l’Art: Scientific Aesthetics 7(12): 27-31.
BRUCE, V. and P. R. CREEN 1990. Visual perception: physiology, psychology and ecology. Lawrence Erlbaum Associates, Hove and London.
BYERS, M. A. 1994. Symbolising and the Middle-Upper Palaeolithic transition. Current Anthropology 35: 369-99.
CAVANAGH, P. 1995. Vision is getting easier every day. Guest editorial, Perception 24: 1227-31.
CHASE, P. C. and H. L. DIBBLE 1987. Middle Palaeolithic symbolism: a review of current evidence and interpretations. Journal of Anthropological Archaeology 6: 263-96.
CHOMSKY, A. N. 1957. Syntactic structures. Mouton, The Hague.
CHOMSKY, A. N. 1988. Language and problems of knowledge: the Managua Lectures. MIT Press, Cambridge, Mass.
CHUKOVSKY, K. 1968. From two to five. University of California Press, Berkeley.
CONKEY, M. W. 1978. Style and information in cultural evolution: toward a predictive model for the Palaeolithic. In C. Redman et al. (eds), Social archaeology. Beyond subsistence and dating, pp. 61-85. Academic Press, New York.
CONKEY, M. W. 1980. The identification of prehistoric hunter-gatherer aggregation sites: the case of Altamira. Current Anthropology 21: 609-30.
CONKEY, M. W. 1981. A century of Palaeolithic art. Archaeology 34: 20-8.
CORBALLIS. M. C. 1989. Laterality and human evolution. Psychological Review 96(3): 492-505. [DH]
COSMIDES, L., J. TOOBY and J. H. BARKOW 1992. Introduction: evolutionary psychology and conceptual integration. In J. H. Barkow, L. Cosmides and J. Tooby (eds), The adapted mind: evolutionary psychology and the generation of culture, pp. 3-15. Oxford University Press, Oxford and New York.
COSS, R. G. 1985. Evolutionary restraints on learning: phylogenetic and synaptic interpretations. In N. M. Weinberger, J. L. McGauch and G. Lynch (eds), Memory systems of the brain: animal and human cognitive processes, pp. 253-273. The Guilford Press, New York. [DH]
COX, M. 1993. Children’s drawings of the human figure. Lawrence Erlbaum Associates, Hove and London.
DAMASIO, A. R. 1989. Time-locked multiregional retroactivation: a systems-level proposal for the neural substrates of recall and recognition. Cognition 33: 25-62. [DH]
DAMASIO, A. 1994. Descartes’ error: emotion, reason and the human brain. Putnam, New York. [EHS]
DAMASIO, A. 1999. The feeling of what happens: body and emotion in the making of consciousness. Harcout Brace, New York. [EHS]
DAMASIO, A. R. and H. DAMASIO 1992. Brain and language. Scientific American 267(Sept.): 62-71.
DAVIDSON, J. and W. NOBLE 1989. The archaeology of depiction and language. Current Anthropology 30: 125-56.
DAVIS, W. 1986. The origins of image making. Current Anthropology 27: 193-202.
DAVIS, W. 1993. Beginning the history of art. The Journal of Aesthetics and Art Criticism 51(3): 327-50. [DH]
DAWKINS, R. 1991. The blind watchmaker. Penguin, London.
DEACON, T. 1989. The neural circuitry underlying primate cells and human language. Human Evolution 4: 367-401.
DEACON, T. 1997. The symbolic species. Penguin, London.
DEAN, P. 1982. Visual behaviour in monkeys with inferotemporal lesions. In D. J. Ingle et al. (eds), Analysis of visual behaviour, pp. 587-628. MIT Press, Cambridge, Mass.
DENNETT, D. C. 1995. Darwin’s dangerous idea. Penguin, London.
D’ERRICO, F. and P. VILLA 1997. Holes and grooves: the contribution of microscopy and taphonomy to the problem of art origins. Journal of Human Evolution 33: 1-31. [RGB]
DIRINGER, D. 1962. Writing. Thames and Hudson, London.
DISSANAYAKE, E. 1992. Homo aestheticus: where art comes from and why. The Free Press, New York. [DH]
DOBELLE, W. H. and M. C. MLADEJOWSKI 1974. Phosphenes produced by electrical stimulation of human occipital cortex, and their application to the development of a prosthesis for the blind. Journal of Physiology 243: 553-76.
DOBELLE, W. H., M. C. MLADEJOWSKI, J. R. EVANS, T. S. ROBERTS and J. P. GIRVIN 1976. Braille reading by a blind volunteer by visual cortex stimulation. Nature 259: 111-12.
DONALD, M. 1991. Origins of the modern mind. Harvard University Press, Cambridge, Mass. [DH]
DRONFIELD, J. 1996. The vision thing: diagnosis of endogenous derivation in abstract arts. Current Anthropology 37: 373-89.
EICHMAIER, J. and O. HÖFER 1974. Endogene Bildmuster. Urban und Schwarzenberg, München, Berlin and Wien. [RGB]
EYSEL, U. 1999. Turning a corner in vision research. Nature 399: 641-4.
FAVREAU, O. A. and M. C. CORBALLIS 1976. Negative aftereffects in visual perception. In R. L. Atkinson and R. C. Atkinson (eds), Mind and behaviour, pp. 88-94. W. H. Freeman, San Francisco. [DH]
FELIKS, J. 1995, 1997a, 1997b. The impact of fossils. Preliminary drafts submitted for publication elsewhere and widely distributed. [JF]
FELIKS, J. 1998a. The impact of fossils on the development of visual representation. Rock Art Research 15: 109-24.
FELIKS, J. 1998b. The value of interpretive approaches in archaeology. Rock Art Research 15: 128-30. [JF]
FELLEMAN, D. J. and D. C. VAN ESSEN 1991. Distributed hierarchical processing in the primate visual cortex. Cerebral Cortex 1: 1-47.
FIEDLER, L. 1999. Repertoires und Gene. Der Wandel kultureller und biologischer Ausstattung des Menschen. Germania 77: 1-37. [RGB]
FODOR, J. A. 1983. The modularity of mind. MIT Press, Cambridge, Mass.
GAMBLE, C. 1993. Timewalkers — The prehistory of global colonisation. Alan Sutton, Stroud, Clos.
GARDNER, H. 1980. Artful scribbles. The significance of children’s drawings. Basic Books, New York.
GAZZANIGA, M. 1970. The bisected brain. Appleton-Century Crofts, New York.
GETZELS, J. and M. CSIKSZENTMIHALYI 1976. The creative vision: a longitudinal study of problem finding in art. Wiley, New York.
GIBSON, J. J. 1950. The perception of the visual world. Houghton Mifflin, Boston.
GIBSON, J. J. 1979. The ecological approach to visual perception. Houghton Mifflin, Boston.
GILBERT, C. D. 1998. Adult cortical dynamics. Physiological Review 78: 467-85.
GOLOMB, C. 1981. Representation and reality in children’s drawings. Review of Research in Visual Education 14: 36-48.
GOMBRICH, E. H. 1973. Illusion and art. In R. L. Gregory and E. H. Gombrich (eds), Illusion in nature and art. Charles Scribner’s Sons, New York.
GOMBRICH, E. H. 1977. Art and illusion. Phaidon, London.
GOMBRICH, E. H. 1984. The sense of order. Phaidon, London.
GOMBRICH, E. H. 1991. Topics of our time. Approaches to the history of art: three points for discussion. Phaidon, London.
GOODNOW, J. 1977. Children’s drawing. Fontana Books, London.
GÖTZ, K., A. BORISY, R. LYNN and H. EYSENCK 1979. A new visual aesthetic sensitivity test: 1. Construction and psychometric properties. Perceptual and Motor Skills 49: 795-802.
GOULD, S. J. and R. C. LEWONTIN 1979. The spandrels of San Marco and the Panglossian program: a critique of the adaptionist programme. Proceedings of the Royal Society of London 205: 281-8. [DH]
GREGORY, R. L. 1970. The intelligent eye. Weidenfeld and Nicolson, London.
GREGORY, R. L. 1973. The confounded eye. In R. L. Gregory and E. H. Gombrich (eds), Illusion in nature and art, pp. 49-95. Duckworth, London.
GREGORY, R. L. 1990. Eye and brain (4th edn). Oxford University Press, Oxford.
HALVERSON, J. 1987. Art for art’s sake in the Palaeolithic. Current Anthropology 28: 63-89.
HALVERSON, J. 1992. The first pictures: perceptual foundations of Palaeolithic art. Perception 21: 389-404.
HAMILTON-SMITH, E. 1986. Comment on R. G. Bednarik, ‘Parietal finger markings in Europe and Australia’. Rock Art Research 3: 159-60. [EHS] [DH]
HARLOW, H. F., M. K. HARLOW and D. R. MEYER 1950. Learning motivated by a manipulation drive. Journal of Experimental Psychology 40: 228-34.
HELMHOLTZ, H. von. 1867. Handbook of physiological optics. J. Southall (ed. and transl.). Optical Society of North America, New York.
HIRSCHFELD, L. A. and S. A. CELMAN 1994. Toward a topography of mind: an introduction to domain specificity. In L. A. Hirschfeld and S. A. Gelman (eds), Mapping the mind, pp. 3-35. Cambridge University Press, Cambridge.
HOLLINGWORTH, A. and J. M. HENDERSON 1999. Vision and cognition: drawing the line. Behavioural and Brain Sciences 22: 380-1.
HOLLOWAY, R. 1983. Human paleontological evidence relevant to language behaviour. Human Neurobiology 2: 105-14.
HUBEL, D. H. and T. N. WIESAL 1979. Brain mechanisms of vision. In Readings from Scientific American. Mind and behaviour, pp. 32-44. Freeman, San Francisco.
HUDSON, S. C. 1998. The hunter’s eye: visual perception and Palaeolithic art. Archaeological Review from Cambridge. 15(1): 95-109. [DH]
HUMPHREYS, C. W., M. J. RIDDOCH and P. T. QUINLAN 1988. Cascade processes in picture identification. Cognitive Neuropsychology 5(1): 67-103.
IWAWAKI, S., H. EYSENCK and K. GÖTZ 1979. A new visual aesthetic sensitivity test (VAST): II Cross-cultural comparison between England and Japan. Perceptual and Motor Skills 49: 859-62.
JACKSON, J. H. 1932. Selected writings. J. Taylor (ed.) Vol. 2. Hodder and Stoughton, London.
JERISON, H. 1991. Brain size and the evolution of mind. The fifty-ninth James Arthur Lecture on the Human Brain. American Museum of Natural History, New York.
JUNG, R. and C. BAUMGARTNER 1965, cited in: Art and visual abstraction. In R. L. Gregory (ed.), The Oxford companion to mind, pp. 40-41. Oxford University Press, Oxford and New York.
KAGAN, J. 1970. The determinants of attention in the infant. American Scientist 58(3): 298-306.
KARMILOFF- SMITH, A. 1992. A developmental perspective on cognitive science. MIT Press, Mass.
KAY, R. F., M. CARTMILL and M. BALOW 1998. The hypoglossal canal and the origin of human vocal behaviour. Proceeding of the National Academy of Sciences of the U.S.A. 95: 5417-19.
KELLOGG, R., M. KNOLL and J. KUGLER 1965. Form similarity between phosphenes of adults and pre-school children’s scribblings. Nature 208: 1129-30.
KELLOGG, R. 1969. Analyzing children’s art. Mayfield, Palo Alto, California.
KELLOGG, R. 1979. Children’s drawings/children’s minds. Avon, New York.
KENNEDY, J. M. and J. SILVER 1974. The surrogate functions of lines in visual perception. Evidence from antipodal rock and cave artwork sources. Perception 3: 313-22.
KNOLL, M. and J. KUGLER 1959. Subjective light pattern spectroscopy in the encephalographic frequency range. Nature 184: 1823-4.
KOLB, B. and I. Q.WHISHAW 1996. Fundamentals of human neuropsychology. Freeman, New York
LAITMAN, J. T. 1984. The anatomy of human speech. Natural History (August): 20-27.
LATTO, R. 1995. The brain of the beholder. In R. L. Gregory et al. (eds), The artful eye, pp. 66-94. Oxford University Press, Oxford.
LEAKEY, R. 1994. The origin of humankind. Weidenfeld and Nicolson, London.
LEROI-GOURHAN, A. 1958. La symbolisme des grands signes dans l’art parietal paléolithique. Bulletin de la Société Prehistorique 55: 384-98.
LEWIN, R. 1999. Human evolution. Blackwell Science Publ., Maiden, Mass.
LEWIS-WILLIAMS, J. D. and T. A. DOWSON 1988. The signs of all times. Entopic phenomena in Upper Palaeolithic art. Current Anthropology 29: 201-45.
LIVINGSTONE, M. S. 1988. Art, illusion and the visual system. Scientific American 256: 78-85.
LIVINGSTONE, M. S. and D. H. HUBEL 1987. Psychophysical evidence for separate channels for the perception of form, colour, movement and depth. Journal of Neuroscience 7: 3416-68.
LIVINGSTONE, M. S. and D. HUBEL 1995. Through the eyes of monkeys and men. In R. L. Gregory. et al. (eds), The artful eye, pp. 52-65. Oxford University Press, Oxford.
LOIZOS, C. 1967. Play behaviour in higher primates: a review. In D. Morris (ed.), Primate ethology, pp. 176-218. Weidenfeld and Nicolson, London.
MACH, E. 1943. Popular scientific lectures (transl. T. J. McCormack, fifth edn.). The Open Court Publishing Company, Illinois.
McNAMARA, K. 1999. Embryos and evolution. New Scientist l6 October (Inside Science series 124: 1-4).
MARTIN, R. 1990. Primate origins and evolution: a phylogenetic reconstruction. Princeton University Press, Princeton. [DH]
MARR, D. 1976. Early processing of visual information. Philosophical Transactions of the Royal Society of London. Series B 275: 483-524.
MARSHACK, A. 1972. The roots of civilization. Weidenfeld and Nicolson, London.
MARSHACK, A. 1976. Some implications of the Palaeolithic symbolic evidence for the origin of language. Current Anthropology 17: 274-81.
MARSHACK, A. 1996. A Middle Palaeolithic symbolic composition from the Golan Heights: the earliest known depictive image. Current Anthropology 37: 357-64.
MIDGLEY, M. 1985. Evolution as a religion: strange hopes and stranger fears. Methuen, London. [DH]
MILLAR, S. 1968. The psychology of play. Penguin, London.
MITHEN, S. 1996. The prehistory of the mind. Thames and Hudson, London.
MITHEN, S. 1998. A creative explosion? Theory of mind, language and the disembodied mind of the Upper Palaeolithic. In S. Mithen (ed.), Creativity in human evolution and prehistory, pp. 165-191. Routledge, London. [DH]
MORRIS, D. 1967. The biology of art. Aldine Atherton, Chicago.
MORWOOD, M. J., P. B. O’SULLIVAN, F. AZIZ and A. RAZA 1998. Fission-track ages of stone tools and fossils on the east Indonesian island of Flores. Nature 392: 173-6. [JF]
MUNRO, T. 1963. Evolution in the arts. The Cleveland Museum of Art, Ohio. [DH]
NEISSER, U. 1976. Cognition and realty. W. H. Freeman and Company, San Francisco.
NOBLE, W. and I. DAVIDSON 1996. Human evolution, language and mind. A psychological and archaeological inquiry. Cambridge University Press, Cambridge. [RGB]
PARKER, S. T. 1985. A social-technological model for the evolution of language. Current Anthropology 26: 617-39. [DH]
PARKER, S. T. and M. I. MCKINNEY 1999. Origins of intelligence. The origin of cognitive intelligence in monkeys, apes and humans. John Hopkins University, Baltimore. [DH]
PENFIELD, W. and L. ROBERTS 1959. Speech and brain mechanisms. Princeton University Press, Princeton, N.J.
PIAGET, J. 1963. The origins of intelligence in children. Norton, New York.
PIAGET, J. and B. INHEWER 1956. The child’s conception of space. Routledge and Keegan Paul, London.
PINKER, S. 1994. The language instinct. Penguin, London.
PYLSHYN, Z. W. 1999. Is vision continuous with cognition? Behavioural and brain Sciences 22: 341-423.
READ, H. 1965. The origins of form in art. Thames and Hudson, London. [DH]
ROSE, S. 1976. The conscious brain. Penguin, London.
ROSENFELD, A. 1999. Rock art and rock markings. Australian Archaeology 49: 28-33. [RGB]
SCHWARTZ, J. 1980. Metalinguistic awareness: a study of verbal play in young children. Unpubl. MS, Queens College, City University of New York.
SHERRATT, A. 1999. The Thak Hypothesis: a prestige-good model of early hominine behaviour, Cambridge Archaeological Journal 9: 277-80. [AS]
SIEGEL, R. K. 1977. Hallucinations. In Readings from Scientific American. Mind and behaviour, pp. 125-132. Freeman, San Francisco.
SIEGEL, R. K., J. E. MURRAY and E. JARVIS 1975. Drug induced hallucinations in animals and man. In R. K. Siegel and L. J. West (eds), Hallucinations — behaviour experience and theory, pp. 81-161. John Wiley and Sons, New York.
SLATER. A. 1996. Organization of visual perception in early infancy. In S. Vital-Durand, J. Atkinson and O. J. Braddick (eds), Infant vision, pp. 307-325. Oxford University Press, Oxford.
SMETS, G. 1973. Aesthetic judgement and arousal: an experimental contribution to psychoaesthetics. Leuven University Press, Leuven. [DH]
SOLSO, R. L. 1993. Cognition and the visual arts. MIT Press, Cambridge, Mass.
SPERBER, D. 1994. The modularity of thought and the epidemiology of representations. In L. A. Hirschfeld and S. A. Gelman (eds), Mapping the mind, pp. 39-67. Cambridge University Press, New York. [DH]
STEGUWEIT, J. 1999. Intentionelle Schnittmarken auf Tierknochen von Bilzingsleben — Neue lasermikroskopische Untersuchungen. Praehistoria Thuringica 3: 64-79. [RGB]
STORDEUR, D. and B. JAMMOUS 1995. Pierre à rainure à décor animal trouvée dans l’horizon PPNA de Jerf el Ahmar (Syrie). Paléorient 21(1): 129-30.
STRINGER, C. and R. MCKIE 1996. The origins of modern humanity. Jonathon Cape, London.
TOOBY, J. and L. COSMIDES 1992. The psychological foundations of culture. In J. H. Barkow, L. Cosmides and J. Tooby (eds), The adapted mind: evolutionary psychology and the generation of culture, pp.19-136. Oxford University Press, Oxford and New York.
TOOTELL, R. B. H., K. HADJIKHANI, W. VANDUFFEL, A. K. LIU, J. D. MENDOLA, M. I. SERENO and A. M. DALE 1998. Functional analysis of primary visual cortex (V1) in humans. Proceedings of the National Academy of Sciences of the U.S.A. 95: 811-17.
TREISMAN, A. 1986. Features and objects in visual processing. Scientific American 255: 106-15.
WEIR, R. 1962. Language in the crib. Mouton, The Hague.
WHITE, Randall 1982. Rethinking the Middle/Upper Palaeolithic transition. Current Anthropology 23: 169-89.
WHITE, Randall 1989. Visual thinking in the Ice Age. Scientific American 23: 74-81.
WHITE, Randall 1992. Beyond art: toward an understanding of the origins of material representation in Europe. Annual Review of Anthropology 21: 537-64.
WHITE, Robert 1959. Motivation reconsidered: the concept of competence. Psychological Review 66: 297-331.
WILLS, C. 1993. The runaway brain. Basic Books, New York.
WILSON, E. O. 1998. Consilience: the unity of knowledge. Little Brown, London. [DH]
ZEKI, S. 1992. The visual image in mind and brain. Scientific American 267: 43-50
ZEKI, S. 1993. A vision of the brain. Blackwell Scientific Publications, Oxford.
ZEKI, S. 1999. Inner vision: an exploration of art and the brain. Oxford University Press, Oxford. [DH]


This paper was first published in Rock Art Research, in May 2000, Volume 17, Number 1, pp. 3-34.