• 'The Origins of Monsters' Book Club

The Origins of Monsters: A Précis

The origins of The Origins of Monsters lie in an interview I conducted more than a decade ago with Maurice Bloch (Wengrow 2003). We were talking about the work of Dan Sperber and Pascal Boyer, and more specifically the problem of how to explain striking recurrences in the content of human culture, especially when these relate to such apparently arbitrary things as the form taken by mythological or otherwise imaginary beings.

I had in mind some specific examples, which have intrigued art historians and archaeologists for more than a century. They concern the transfer of images depicting composite beings such as griffins, sphinxes and so on. Images like these have “epidemic” features. They spread with amazing bravado from one society to another, often in periods of major social change. The whole phenomenon begins at a suggestive time – about 6000 years ago, when the world’s first states and urban societies were taking shape in North Africa and the Middle East. Another famous and much studied monster-epidemic occurs in the Iron Age, when imaginary beasts from the East invade the image worlds of archaic Greece and Etruria along with a variety of other social and technological innovations, as part of what is sometimes called the “orientalising” of the Mediterranean. There are plenty more examples in between, which I try to document in the book.

Lots of people have studied these cultural transfers from the perspective of their own period and regional specialization. But I think my book is the first to treat them as a group, and possibly as part of a unified phenomenon. Maurice’s comment on the whole affair was along the lines that monsters are interesting, anthropologically and psychologically, precisely because they appear to be acts of pure and free imagination, yet their characteristics recur so widely within and between societies.

I made a mental note to return to the topic but did so only seven years later, when I was invited by Roger Bagnall to give a series of lectures at New York University’s newly established Institute for the Study of the Ancient World. The lecture series is named in honour of the historian Mikhail Rostovtzeff. Speakers are supposed to address topics that explore his cosmopolitan vision of ancient Eurasia as a world shaped by far-flung connections that, at least since Bronze Age times, have bound its societies together by land and by sea, from China to the Mediterranean.

I had no idea at the time that Rostovtzeff, best known for his (1926) Social and Economic History of the Roman Empire, had himself written a series of lectures on the transfer of images, and in particular the fantastic ‘animal style’ of the Central Asian steppe. I have the 2010 Eyjafjallajökull eruption to thank for this discovery. It left me stranded at Princeton University, where the library has a copy of the original lectures. Never mind that most of Rostovtzeff’s ideas were wrong: their existence offered the prospect of academic credibility for what must otherwise have seemed an unlikely project.

Around that time, and for reasons initially unconnected with my research, I was reading an increasing amount of work in experimental psychology, especially about mental modularity, symbolic expression, and theory of mind. A few prehistorians were already exploring the implications of this kind of work for their own material, but in a way which seemed problematic to me (see pp. 4-5 of the book in particular). Still this seemed better than the default position of most culture historians, which is simply to ignore what is going on in fields such as neuroscience and cognitive psychology.

Susanne Küchler (2005) has made similar observations about anthropology, where approaches to human cognition – if made explicit at all – seem torn between incompatible notions of what the mind is and how it works. Is it a highly malleable organ, constantly re-trained through its encounters with a culture-laden world? Or is each of us, by contrast, carrying within ourselves a pristine hunter-gatherer brain that fights its primeval battles through a modern world of its own mysterious making? In their own ways, both alternatives are highly romantic and appealing; but my book joins many others in arguing that the reality is somewhere in between.

I conceived of the book as a test case for what Dan Sperber calls the ‘epidemiology of culture’, and am absolutely thrilled that it has been taken up for discussion by researchers who are serious about exploring the relationship between cognition and culture. But I am also aware that my subject matter raises some immediate problems. I half anticipate that the “epidemiologists” will take the opportunity to explain that I have simply misunderstood what they are trying to do. But I have read their work closely, and taken them at their word when they say that they are interested in understanding how cognition might underpin the spread of culture (culture, in all its manifestations, as opposed merely to language).

The main difficulty with epidemiological approaches to culture, it seems to me, is the mismatch between theory and data. The theory is all about diffusion, popularity, distributions of culture, spreads of representations. But the data used to test this hypothesis seem so far to derive either from experimental (laboratory-type) conditions or from the ethnographic record. This would have been fine a century ago, when mapping out distributions of culture on a large scale was precisely what a lot of anthropologists did. But that kind of work is rarely done nowadays. What we have instead are snippets of culture, isolated examples uprooted from their histories of circulation, and arguably unsuited to an empirical study of how things spread.

Most archaeologists, by contrast, have never lost their love of map-making and the study of distributions. It’s arguably what we’re best at. So it seemed to me that, in this case, the anthropologists and evolutionary psychologists had proposed a great hypothesis that could not really be tested from within their own material, but which at the same time offers a wonderful invitation for archaeologists to get involved. A second problem is that the definition of ‘culture’ in evolutionary psychology often seems heavily weighted towards spoken discourse. It has relatively little to say so far about either images or objects (Alfred Gell’s [1998] work surely addresses some of this, but Gell was rarely explicit about the kind of cognitive models he was using).

An exception is Sperber and Hirschfeld’s (2004) piece on ‘The cognitive foundations of cultural stability and diversity’. They present a suggestive discussion of how artifacts such as outlandish masks, caricatures, and cosmetics may stimulate – yet at the same time violate – specific mental modules for face recognition. Such things are intuitively recognised as both face-like and un-face-like, in a way that makes us pay special attention to them. And according to the theory, this may enhance their chances of transmission. It makes them culturally catchy. Pascal Boyer, if I understand him correctly, argues on similar lines that some such cognitive balancing act is necessary or at least desirable if a given concept of the supernatural is going to catch on, and become embedded in some wider cultural milieu:

In any cultural environment, indefinitely many representations of religious entities are constantly created and communicated. Only some of them, however, have the potential to support both imaginative scenarios and intuitive references. These are the ones that combine a rich intuitive base, with all its inferential potential, and a limited series of violations of intuitive theories, which are attention-demanding. Because of these characteristics, such assumptions are more likely than others to be easily acquired, memorized, and transmitted than other assumptions. It should not be surprising, therefore, that they constitute the most recurrent aspects of religious systems.

(Boyer 1994: 122)

Can this kind of theory be legitimately extended from spoken discourse about supernatural beings to the world of images? I don’t see why not, and two particular sources of inspiration here were the work of Barbara Stafford (especially her 2007, Echo Objects: The Cognitive Work of Images) and La fabrique des images, an exhibition hosted at the Musée du quai Branly in 2010-11. Philippe Descola and Anne-Christine Taylor, its creators and organisers, were kind enough to show me around on its final day. The timing could not have been better for me, and it will already be obvious that – like its subject matter – my book is very much a composite of influences, thrown together more or less by chance.

From La fabrique des images I took two things. The first was a way of approaching modes of figuration as types of visual experimentation, at once anatomical and theoretical. The exposition showed how such experiments offer points of entry to more general principles, by which people try to organize the world around them in some sort of systemic relationship with other worlds “beyond” – the imagined worlds of the supernatural. The second thing I took away was a more technical point about the production of composite figures, which Descola associates with an ontological stance called “analogism”. It’s the same point made by Da Vinci in his notebook entry on How to Make and Imaginary Animal Look Natural: that precisely because of their fictional character, the creation of visually compelling composites requires enhanced empiricism and accuracy in the depiction of individual body parts, each of which should be independently identifiable as belong to a certain kind of species (Arcimboldo’s 16th century faces take the same principle to another kind of extreme). I also found it fascinating that nearly all the examples of analogism derive from large-scale, hierarchical societies such as Han China, medieval Europe, or the historical kingdoms of Benin. I discussed this point with Philippe and Anne-Christine, but I suspect this is roughly where our interests begin to diverge.

In Barbara Stafford’s work I found a scaffold that bridged research on mind and image. It emboldened me to expose data of an archaeological and art historical kind to an “epidemiological” approach. Stafford uses the term “compressive compositions” for images that ostentatiously combine elements from different species, and she suggests that such images draw attention to otherwise unconscious systems of visual processing. For example, experimental studies show that human cognitive processing of animal forms is highly sensitized to part-whole relations (Davidoff and Roberson 2002). A total presence can be inferred from quite limited visual cues (horns, tails, feet and so on). This inferential capacity we owe to an intuitive repertory of biological information that is hard-wired, and is part of our evolutionary make-up (New et al. 2007). Pictures of animals – even when jumbled, distorted, or incomplete – may thus activate neural pathways attuned to the recognition and classification of living kinds.

In violating some limited part of this intuitive biology, composite figures nevertheless affirm many of its underlying structural principles. Even such fantastic beings as a dragon, mermaid, centaur, minotaur, unicorn, sphinx, or whatever, will have legs correctly positioned for walking, eyes for seeing, wings for flying, horns for gorging, fins for propulsion, and so on, allowing us to infer (often extraordinary) properties of movement and vitality for such figures. They are, in short, the visual counterparts to Sperber and Boyer’s “minimally counter-intuitive representations”. And if the latter are right, then such representations might be expected to exhibit special “epidemic” properties of transmission, of a kind that should leave traces in the record of human image-making. Or at least that was the premise from which I began my own experiment in tracking distributions of composite figures.

Of course this was not entirely a “blue skies” experiment. I already knew that, from at least around 4000 BC, “monster epidemics” were a striking and recurrent feature of the visual record. I diligently began to compile evidence for the earliest widespread appearance of composite beings in the art of Egypt, Mesopotamia, Pakistan, Greece, and so on, and to map out the likely relationships between them. In nearly all cases the appearance of monsters on the scene correlates with a number of other historical features, which seem non-random and important in terms of modes of transmission. Not all of them will apply to every case, but for western Eurasia and the Mediterranean world they can be summarized under five broad headings:

  1. Mechanical image-making devices: the use of stamps, moulds, and seals (remote ancestors of the modern printing press) to replicate and promulgate officially sanctioned images in standard formats and on a wide scale. This took place mostly through the administrative offices of temples, palaces, and other urban institutions.
  2. Centralisation: composite figures seem most at home in (what was then) the newly emergent social world of the city, whether the ancient city-state or some other kind of large-scale territorial polity, such as the Bronze Age kingdoms of North-East Africa, or the Neo-Assyrian Empire. Monsters are urbanites, and highly cosmopolitan ones at that.
  3. Social transformation: composite figures typically appear in times of major social and economic restructuring – the phases that archaeologists love to call ‘formative’, ‘archaic’ or ‘proto’ (as in Archaic Greece or Protodynastic Egypt) because they precede and foreshadow the emergence of states with their official canons of cultural representation.
  4. Conservatism: these composite figures are rarely original or local flights of imagination. More commonly – as with the serpent-necked felines on the ceremonial cosmetic palette of King Narmer – they are imports from an exotic and distant source (and this they have in common with those much later “marvels of the East”, famously discussed by Wittkower and later Mitter, whose movements from India to Europe can be traced in images from late antiquity to the age of the modern printing press).
  5. Apotropaism: where composite figures can be associated with rich iconographic repertories and/or ancient written sources, these frequently implicate them in the use of magic and medicine to avert misfortune, including protection from the spread of harmful diseases. The Neo-Assyrian empire furnishes wonderfully detailed examples, such as the corpus of cuneiform texts from the House of the Exorcist at Assur, with their pedantic instruction-manuals for the making and placement of protective images, and injunctions to send sickness-bearing demons “3600 miles” away from their intended human targets.

Based on these five factors it might be reasonably argued that composite figures do, in fact, exhibit (almost literally) epidemiological features in the record of human image making, far outstripping most other kinds of imagery in their scale of distribution (and this includes, of course, other types of image transmitted by mechanical reproduction). In which case we would have discovered a good empirical demonstration of how cognitive dispositions works in practice to create spreads or distributions of culture. The difficulty, of course, is that all my examples derive from the last 6000 or so years of human history – a relatively shallow period of time in evolutionary terms. If the striking distribution of imaginary composites in the visual record is to be explained in terms of innate pre-dispositions, then what about the many preceding millennia of human image making. Where, in short, are all the prehistoric monsters?

Here we enter slightly murky waters. Some of the initial reactions to my argument were along the lines: but what about the famous ‘Sorcerer’ of Les Trois-Freres Cave, which dates back to the Upper Palaeolithic? Or the ‘Bison-Man’ of Chauvet and the ‘Lion-Man’ of Hohlenstein-Stadel, which at 30,000 BC are twice as old still? But even if we accept the identification of these and some tens perhaps of other very early prehistoric figures as composites (and this is not uncontroversial – as pointed out by Dale Guthrie, for instance, Lion-Man may simply be “standing bear”), they remain strikingly isolated within the much wider corpus of images that survive from Palaeolithic and Neolithic times. The majority of those images, as I discuss in chapter 3, are more or less schematic representations of animals and other subjects that could actually be seen in the worlds of Ice Age hunter-gatherers and of the first farmers.

The question is not whether early humans were capable of producing images of fantastic, composite beings: they undoubtedly were, and they undoubtedly did. The question (of my book at least) is whether such images ever caught on, becoming stable and widely transmitted features of prehistoric culture. And while there may be limited exceptions, the general answer, it seems, is an emphatic: ‘no’. As Andre Lerio-Ghouran pointed out half a century ago, ‘Palaeolithic art offers very few examples of what might be construed as flights of the imagination. Its monsters can be counted on the fingers of one hand’ (or today, perhaps, two hands and a foot).

Modern scholars may love to disseminate and talk about images of prehistoric bull-men and so on. They open the door to discussions of early religion and the imaginary worlds of our hunter-gatherer ancestors. But the ancestors themselves just don’t seem to have shared this fascination. So rare, in fact, are these kinds of images that one almost suspects a policy of avoidance or prohibition. Such a policy would in fact be quite familiar to more recently documented groups of hunter-gatherers or small-scale cultivators, who follow the ontological precepts of ‘animism’ and ‘totemism’ (in Descola’s sort of sense). Consult the catalogue of the quai Branly exhibition and you see that the kind of anatomical reshuffling required to produce composites is generally at odds with the plastic and visual arts of such groups.

It is, by contrast, a pervasive feature in their performance arts, where ritual actors take on attributes of animal and other non-human bodies, and vice-versa. These are the image worlds of masks and short-lived effigies, destined to vanish in spectacular rites of expiation. For reasons of prudence, among the San or Inuit for example, such effigies were not traditionally rendered as permanent images. More often, crossing the boundaries between species also meant having to navigate a safe return – as epitomized in the figure of the trickster, shaman, or shape-shifter – or in the multi-layered masks of the Kwakiutl that flicker open and shut in ritual performances, affording glimpses of a human face lodged in an animal body, but never more than a glimpse.

To depict such states of mediation or mid-transformation in durable form may invite danger, by leaving open an extended trace of a relationship (between human and “other”) that should be properly circumscribed by rites of passage. Images of composite beings, rigid and unchanging, thus evoke the principles of metamorphosis and liminality, only to subvert them. They fix transformations in stable media that render them capable of being replicated and disseminated, over and over again, in canonical forms. Composites thus typically belong to the image-worlds of cities and hierarchical states: mechanized, modularized, standardized, and centralized. The basic point was made long ago by Elias Canetti, in those sections of his (2000 [1960]) Crowds and Power that deal with the topic of transformation. Over the decades related points were developed by, among others, Victor Turner, Ruth Benedict, Fredrik Barth, and more recently Harvey Whitehouse (2004) with his distinction between ‘imagistic’ and ‘doctrinal’ modes of religious transmission.

In some other respects too, the ‘cultural ecology’ of cities seems especially well suited to the production of composites. Not just in imagery but also in other spheres of technical production. In the central chapter of the book I try to show how principles of modularity and standardisation – all basic to composite figuration – can in fact be found at work more generally in styles of craft and industry that developed in the early urban centres of western Asia, north Africa, and the Mediterranean. From the fourth millennium BC onwards, new levels of uniformity and precision are evident across a whole range of activities from the ceramic arts to techniques of making furniture and buildings, and also in modes of depiction. Similar points have been made for the bronze-working systems of Shang China and their associated forms of ornamentation. There, from around 1500 BC, images of anatomical composites (like the taotie figures on metal vessels) proliferated within an urban industrial complex that allowed manufacturers ‘to assemble countless combinations from a limited repertoire of motifs and compartments’ (Ledderose 2000: 25).

Of course, generating signs from standard components is also a favourite pastime of bureaucrats and administrators the world over. Some of the earliest evidence for this kind of activity can be traced back to ancient Mesopotamia and Elam – today’s Syria, Iraq, and parts of western Iran. There the first known forms of literate administration were concerned with classifying, archiving, and monitoring the flow of resources – both raw materials and finished goods – as they passed through large urban institutions: so many hundred jars of beer or oil, so many thousand sheep, so much grain or textiles. Such institutions combined temple and factory functions, forming the religious and economic hubs of cities such as Uruk and Susa – this was bookkeeping for the gods. The surviving inscriptions (written in scripts known as proto-cuneiform and proto-Elamite) also show that resident scribes sometimes engaged in what Robert Englund calls “fanciful paradigmatic name-generating exercises”, producing long lists of signs, many of which seem never to have been used for any practical purpose whatsoever.

A further function of temple administration was to guarantee the authenticity of finished goods by applying seals to them, or rather to their containers. The miniature impressions, thus applied, offer one of our main sources of evidence on the contemporaneous development of pictorial design, and often show lively arrangements of animals and people engaged in a variety of activities and postures. Such markings also testify to a close relationship between skilled depiction (in intaglio carving) and bureaucracy as linked domains of urban activity. The entire system of marking and recording was predicated on the constant generation of new visual signs that could fulfil their designated purposes, as discrete signifiers within an expanding system of administration. Under such conditions it is easy to see why the “bureaucratic eye” was drawn to the possibilities of composite figuration which – quite apart from being fun and memorable – would have greatly multiplied the range of possible subjects for depiction (i.e. instead of just ‘cow’, one suddenly has the possibility of ‘body of cow + bird head + wings’, or any other number of possible combinations).

This precis is now wandering a bit too far from its initial questions concerning the relationship between culture and cognition, which is supposed to be the main concern of this forum. So let me take a step back and return to where I started, posing again the question: can the origins of monsters (or, more precisely, of composite figuration) be taken as a test case for an “epidemiological” approach to the evolution of culture? And, if so, then what are its main conclusions? On this point I fully expect a lack of consensus from the readers, to whom I am deeply grateful for their attention to my book. My own conclusions are roughly as follows:

As minimally “counter-factual” or “counter-intuitive” representations, images of composite figures do indeed have “epidemic” qualities, which enhance their chances of transmission within and between populations. How else would we account for the cultural survival, in our own modern imaginaries, of such arbitrary Bronze Age creations as the griffin or unicorn, and for the apparently endless capacity of such figures to acquire rich new meanings and associations across the ages? But I would also argue that these epidemic qualities were unleashed only under certain sets of conditions, which are much more limited in range than the universal cognitive capacities that underpin them. Such conditions have been prevalent only in certain parts of the world, and only for what is – on an evolutionary time-scale – a relatively short episode in the history of our species.

Another way of putting this would be to say that the “intuitive” basis for the cognitive reception of monsters has not one, but two foundations – it is a sort of “double intuitive”. On the one hand, it rests upon the mind’s innate and evolved tendency to compensate for gaps in the visible world, conjuring organic-seeming wholes out of ill-fitting parts, and ascribing them properties of living beings. This inferential capacity has no doubt been with us since the origins of our species. But at another level, the distinct epidemiology of monsters is a product of institutional environments existing only since the appearance of the first cities, and with them (as Jim Scott might have it) our first systematic attempts to “see the world like a state”: as a realm of standard and divisible subjects, each comprising a multitude of separable and combinable parts.

The latter and much more recent development is, I argue, what lay foundations for the enhanced orders of “monster-reception” that would be predicted by an epidemiological approach, and that are so evident a feature in the record of human image-making from the Bronze Age onwards. More generally, I suggest, it is in reconstructing such emergent properties of cognition over historical timescales that an archaeological – rather than purely lab-based or ethnographic – approach remains important for understanding cultural transmission and cultural evolution.


References

Boyer, P. 1994. The Naturalness of Religious Ideas: A Cognitive Theory of Religion. Berkeley, CA and London: University of California Press.

Canetti E (2000 [1960]) Crowds and Power. New York: Farrar, Straus and Giroux.

Davidoff, J. and D. Roberson. 2002. Development of animal recognition: a difference between parts and wholes. Journal of Experimental Child Psychology 81: 217-34.

Descola, P. ed. 2010. La fabrique des images: visions du monde et forms de la representation. Paris: Somogy/Éditions D'art.

Englund, R. K. 1998. Texts from the Late Uruk period. In Mesopotamien: Späturuk-Zeit und Frühdynastiche Zeit, Annäherungen 1, ed. P. Attinger and M. Wäffler, pp. 15-233. Fribourg: Universitätsverlag.

Gell, A. 1998. Art and Agency: an Anthropological Theory. Oxford: Oxford University Press.

Guthrie, R. D. 2005. The Nature of Paleolithic Art. Chicago and London: University of Chicago Press.

Küchler, S. 2005. Materiality and cognition: the changing face of things. In Materiality, ed. D. Miller, pp. 206-30. Durham and London: Duke.

Ledderose, L. 2000. Ten Thousand Things: Module and Mas Production in Chinese Art. Princeton: Princeton University Press.

Mitter, P. 1992. Much Maligned Monsters: A History of European Reactions to Indian Art. Chicago: University of Chicago Press.

New, J., L. Cosmides, and J. Tooby. 2007. Category-specific attention for animals reflects ancestral priorities, not expertise. Proceedings of the National Academy of Sciences of the United States of America 104: 16598-16603.

Rostovtzeff, M.I.  1926. The Social and Economic History of the Roman Empire. Oxford: Clarendon Press.

- 1929. The Animal Style in South Russia and China, being the Material of a Course of Lectures delivered in August 1925 at Princeton University under the auspices of the Harvard-Princeton Fine Arts Club. Princeton Monographs in Art and Archaeology 14. Princeton; London: Princeton University Press; H. Milford, Oxford University Press.

Sperber, D. and L. A. Hirschfeld. 2004. The cognitive foundations of cultural stability and diversity. Trends in Cognitive Sciences 8: 40-46.

Stafford, B. M. 2007. Echo Objects. The Cognitive Work of Images. Chicago and London: University of Chicago Press.

Wengrow, D. 2003. Machiavellian moments: an interview with Maurice Bloch. Journal of Social Archaeology 3: 299-311.

Whitehouse, H. 2004. Modes of Religiosity: A Cognitive Theory of Religious Transmission. Walnut Creek, CA: Altamira.

Wittkower, R. 1942. Marvels of the East: A Study in the History of Monsters. Journal of the Warburg and Courtauld Institutes 5: 159-197.

8 Comments

  • Olivier Morin 10 January 2016 (22:40)

    Thank you very much David for this great introduction to your book. Two questions somehow fell from the last draft of my comment post. Here they are, in no particular order. As a non-specialist, I expect to be wrong about many particulars, so take these as curiosity questions rather than objections.

    • The Origins of Monsters (TOM) is keen to focus on iconography, as opposed to live performances (and the masks that often accompany them), or verbally recounted tales. I understand that, as an archaeologist, you want to focus on what fossilises, so to speak. Yet TOM gives the impression of dismissing the evidence for cultural belies and practices involving hybrid animals in tales or performances as not worth considering, your investigation being about images and images alone. This might be problematic. The “MCI monsters” theory that informs your account was developed to account, first of all, for beliefs (and the tales and legends reflecting them)—not for images alone. Besides, taking tales and performances into account might reveal that (contrary to TOM’s apparent claim), representations of hybrid animals are very common in non-state societies.

    • Speaking of which, TOM seems to ignore a line of evidence that might be relevant. First, iconographic material from contemporary (i.e., from the 19th c. onwards) small-scale societies depicts hybrid animals. (I guess you might reply that these societies could have adopted this imagery from surrounding states, but you also claim that state institutions determine a society’s readiness to borrow composite imagery, and not just to invent it.) Relatedly, TOM addresses prehistoric cave paintings in a very illuminating way—but, even if your interpretation of it is right, can we turn absence of evidence into evidence of absence? After all, normal animals are (it seems to me) more numerous than composites in many visual cultures, including some that are truly fond of monsters. Taphonomic biases may well have erased most prehistoric monsters (except for a few items).

  • David Wengrow 11 January 2016 (13:13)

    Thanks for getting things in motion Olivier. I think I deal with the relationship between images and other media a bit more carefully than you seem to give me credit for (for instance, pp. 28-32 of chapter 2, and then also in chapter 3, ‘The hidden shaman’). You will also find there some discussion of taphonomy, and the importance of considering what has not survived.

    As you say most cognitively based studies of cultural transmission have so far taken language and verbal expression as their main source of evidence. I think it’s high time to reconsider the ‘uses of images’, as Ernst Gombrich put it, but with a more up-to-date psychology.

    You ask how my argument relates to the visual products of recent hunter-gatherers and small-scale cultivators. This is something I would certainly like to discuss more. But perhaps we should await the contributions of Susanne Kuechler and ErhardSchüttpelz? I wonder, in the meantime, what you make of the Descola exhibition, and my thoughts on it?

    Your questions raise a couple of other points that I want to try to clarify.

    From the point of view of cognition, I think it would be an analytical dead-end to collapse all cultural representations into a single category or genre. Images, verbal expressions, and ritual enactments – even if based on shared beliefs – are not the same thing. They have different properties of transmission, and the differences are in my view important.

    I think this is the problem with some current applications of experimental psychology in the social and historical sciences – they treat external (public) representations as direct reflections of internal (mental) representations, without attending to how different media and different genres of expression organise thought into patterns that are historically distinctive.

    From the standpoint of cultural transmission, I wonder whether the whole issue of “belief” (which you raise) may be slightly beside the point, or at least not central. I doubt whether many people today actually believe in the existence of “monsters” (or "witches", or whatever), but their cultural epidemiology is surely more rampant than at any time in the past.

    I’m not sure I actually share your view of archaeologists as being attracted to “what fossilises”. Most of the ones I know (including me) would love nothing better than to access living performances and stories from the remote past. But unless we base our interpretations on something empirical, we are engaging in pure speculation. I think it unlikely that the known corpus of Palaeolithic parietal art is statistically trivial as you suggest. There is just too much of it, and I recommend Dale Guthrie’s (2005) 'The Nature of Palaeolithic Art' for a survey.

    What is striking, I think, is the degree of emphasis that has been placed on a tiny handful of surviving figures from the Upper Palaeolithic that do appear to depict composite or imaginary body-forms. On the one hand this says a lot about archaeologists’ determination to access the intangible dimensions of prehistoric experience. But on the other it gives a misleading impression of the frequency of such figures in prehistoric art, which is in fact remarkably low. This to me is the more interesting point.

    No doubt there were costumed performances and such like, mixing body parts from different species into “ritual composites”. We even catch a glimpse of these in prehistoric burial customs (pp. 37-38 of the book). But an important part of my argument, picking up from Canetti and others, concerns the issue of transformation, and the crucial difference between performance and image in this regard. A fixed and static image of a composite figure is not the same thing as an enacted transition between bodily states (and then back again) – the latter represents a genuine transformation, the former a refusal to transform.

  • Olivier Morin 14 January 2016 (11:04)

    Thanks a lot David, that is an illuminating answer. My first two questions were extremely clumsy and masked a key point: Your treatment of the representations, practices and performances surrounding composite animal imageries is admirable, especially in the book’s chapter 3.

    The book is making a very ambitious claim—hybrid animals thrive in state societies with a complex technical and commercial infrastructure, not, or much less, elsewhere—and this requires you to explain away a lot of evidence pertaining to the mental representations (“belief” is, I agree, an awkward term in this context, to be avoided) and practices surrounding hybrid animals in societies that did not leave monsters set in stone (or carved in clay). TOM’s level of engagement with the ethnographic, art-historical, and psychological literature is quite exemplary. What I wonder is this: Does TOM give definite proof that mental representations of composite animals are much more prevalent in state societies with modular technologies and mechanical reproduction of images, than elsewhere? I am not quite sold to this claim yet, and perhaps neither are you.

    The same claim is much easier to defend, however, if we restrict its scope to images set in stone, as opposed to representations, tales, performances, and practices that leave no solid trace. I read TOM as making precisely that move.

    For instance, p. 37: “The significant point, to which I will return, is that if such beings [composite animals] did populate the collectie imagery of early hunter-gatherers, then their presence was made manifest predominantly through ephemeral modes of display such as psychotropic visions and masked performances that have left few durable traces.” Earlier (pp. 33 sq.), TOM makes a similar point about contemporary hunter-gatherer societies.

    This is important because, later in the book, you use the claims above to criticise the view that representations of monsters might be universally appealing. The reason you can use these claims in such a way is because TOM carefully defined away the representations and practices involving monsters in hunter-gatherer societies, on the grounds that they usually leave no solid traces (at least that is my interpretation). Yet narratives and performances concerning monsters possess most of the cognitive properties that are also possessed by composites carved in stone or impressed on clay—if anything, in narratives and performances involving metamorphoses, men morphing into animals and back again are more counter-intuitive than simple composites.

  • Olivier Morin 14 January 2016 (11:18)

    I agree that Guthrie's Nature of Paleolithic Art does much to dispel the view that monsters were common in prehistoric cave art, and the way he explains away many so-called monsters as clumsy attempts by children to make realistic animal images feels quite compelling. On the whole, it gives great support to your thesis.
    I would also read his book as containing some evidence that Paleolithic hunter-gatherers had a complete and vivid understanding of the modular nature of vertebrate anatomy: bodies made of interchangeable parts, following a broadly universal body plan. Cave paintings of headless or siamese animals (cf. p. 393 of Guthrie's book, for instance) seem to show as much (and in a way, this is to be expected from a society where so much time was devoted to butchering and processing carcasses).

  • Martin Fortier 14 January 2016 (16:13)

    Thank you for this very interesting and thought-provoking piece. A series of questions/objections came to my mind through this reading. Hopefully I don’t misconstrue the points you have made, but please correct me where I’m missing what you actually meant.

    (1) I quite like the idea of expanding the epidemiological approach (EA) beyond the restricted scope of discourse. The tricky question, however, is whether such a move is legitimate (i.e., supported by experimental evidence). As you notice, so far the EA has been dealing almost exclusively with verbal representations. For example, when Boyer & Ramble (2001) show that a specific kind of counterintuitivity makes representations particularly catchy they are talking about verbal representations.
    Your proposal is to apply the EA to public *material* representations. The problem is that the strand of evidence traditionally used by the EA literature cannot support the move to material representations. Indeed, while a study such as Boyer & Ramble (2001) tells us something about the catchiness of verbal representations it doesn’t say anything about the catchiness of material (artistic, artefactual) representations.
    In your book you adduce two main experimental studies in favour of your move towards an EA of material representations (Davidoff & Roberson, 2002; New, Cosmides & Tooby, 2007). First, it seems to me that these studies should be taken cautiously. To my knowledge, Davidoff & Roberson’s study hasn’t been tested cross-culturally; now it seems that at least in some cultures (e.g., Diamond & Bishop, 1999; Medin & Bang, 2014) contextual cues do play a crucial role in the recognition of animals; it is hence not clear which of the three components (parts, wholes and context/situation) plays the main role in animal detection and how these components relate to each other. Regarding New et al.’s study, I think the importance of domain-specificity (e.g., distinguishing between animals and inanimate objects) in vision has definitely some grain of truth, but still things are more complicated than what is often suggested by the proponents of strong modular views (VanRullen & Thorpe, 2001). However, for the sake of the argument, let’s assume that these studies can be taken at face value. Well, in my view, it still remains that these studies don’t support the move from verbal EA to material EA. Indeed, in order to support the EA, it is not sufficient to demonstrate that *recognition* is based on parts detection rather than wholes or that *change detection* is faster and more accurate with animals rather than with inanimate objects; what needs to be shown is that some material representations are more *catchy* – are more easily *recalled* – than others. As far as I know, you never adduce experimental evidence demonstrating this very last point.
    A recent development of Dan Sperber’s work sheds an interesting light of this question. Let’s put it this way: lately, Sperber has been trying to import the EA from the realm of verbal representations into the realm of (ritual) action. As previously, one may wonder: is such a move legitimate? I think in this case, it is indeed perfectly legitimate. And the reason for this is that there is clear evidence that patterns found in opaque actions (e.g., Gergely & Csibra, 2006) or learned conventionally – as opposed to instrumentally – (e.g., Legare & Nielsen, 2015) are particularly catchy. Not only is the EA of ritual actions experimentally supported; but it can furthermore be very elegantly applied to ethnographic data – e.g., to cargo cults, in which causal opacity plays a key role (Umbres & Sperber, Forthcoming).
    To conclude on this and to put my point in even more general theoretical terms, let’s say that one can think of at least four versions of the EA: (i) one applied to verbal representations; (ii) one applied to material representations; (iii) one applied to experience; and finally, (iv) one applied to action. One the one hand, it seems to me that there is to a certain extent clear evidence in favour of versions (i) (e.g., narratives about spirits) and (iv) (e.g., cargo cults) of the EA. On the other hand, I remain unconvinced that the EA (that is, the key concepts of counterintuitivity, opacity, quasi-propositional format, etc.) can be legitimately and fruitfully applied to material representations (e.g., an engraving representing extraordinary monsters) or to experiences (e.g., a mystical experience).

  • Martin Fortier 14 January 2016 (16:15)

    (2) I find the thesis that there is a blatant scarcity of material representations of chimeric creatures in “stateless” or hunter-gatherer societies quite disputable. Arguably, Amazonian societies (to restrict myself to the region I know best) provide numerous examples averring the fact that chimeric representations are indeed widespread among hunter-gatherer (or, to be more precise: in hunter-horticulturist) societies. This is so much the case that very recently, the confidential world of French americanist anthropology has been stirred up by a heated and vehement debate between Dimitri Karadimas and Carlo Severi. The moot point precisely concerned the status of chimeric representations; interestingly enough, many of the ethnographic data on which the discussion was based were drawn from Amazonian societies.
    If I understood you well, in the book you do acknowledge that some chimeric representations can be found in hunter-gatherer societies, but then, you point out that such representations are strictly restricted to masks and performance. Yet, many Amazonian examples of chimeric representations, and notably those discussed by Karadimas (2003, 2007, 2008, 2015a, 2015b) and by Severi (2011, 2015 [2007]), consist not only of masks or of performance clothing, but they also consist of baskets, pottery, more or less lasting body paintings, clubs, etc. It seems to me that this strand of ethnographic evidence somewhat challenges the thesis you are putting forth.
    Detailing Karadimas’ and Severi’s respective take on the question of chimeric representation would lead me beyond the scope of my comment. But to put it in a succinct and rough way, Karadimas has it that a chimeras are nothing but ethnobiological knowledge about ordinary animals gone extraordinary (i.e., gone chimerical) through the distorting mechanisms of analogical inference and knowledge transmission whereas the author of *The Chimera Principle* considers that chimeric arrangements stem from basic cognitive mechanisms related to memory and imagination.
    I don’t want here to adjudicate between the two authors, but suffice it to say that, arguably, each thesis defended respectively by Karadimas and by Severi proves to be incompatible with (and hence challenging to) the main thesis of *The Origins of Monsters*. If Karadimas is right, then there is no genuine chimeric representation, for every chimera can be boiled down to some cryptic ethnobiological knowledge; the trouble for you is that in his view this is true not only for Amazonian chimeric representations but also for any kind of such representation. Indeed Karadimas has applied his theory not only to Amazonian material but also to material pertaining to Mesoamerica (Karadimas, 2014) and to Europe (Karadimas, 2010) (see next point). His view is that the analogical inferences through which chimeras are created is virtually universal. If, on the other hand, Severi is right, then chimeric representations are genuinely such; but again, this is quite challenging for *The Origins of Monsters* for Severi’s claim is that the basic cognitive mechanism he dubs “chimera principle” is widespread in “stateless” societies. It thus appears that Severi’s thesis is, as it were, the opposite of the one advocated in *The Origins of Monsters*.
    So my general question is: how do you deal with this kind of data coming from Amazonian ethnography? And how would you position yourself with respect to Karadimas’ and Severi’s respective theory of chimeric representations?

    (3) Philippe Descola’s grand theory of ontological schemas is broad-ranging and very stimulating indeed; but probably precisely because of its breadth it is also very doubtful. I have spent hours thoroughly reading and reading again *Beyond nature and culture*, and now I tend to think that numerous claims of this book remain (conceptually, ethnographically and cognitively) ill-supported. It would be way too long to mention all the problems of Descola’s theory. Let me simply mention one which is directly related to the topic of your book. One central claim of *Beyond nature and culture* is that a systematic coherence can be found between modes of identification (i.e., the way internal and physical properties are ascribed to surrounding entities) one the on hand and modes of figuration (e.g., whether one represents chimeras or not) or mode of production (e.g., whether one domesticates animals or not) on the other. According to Descola, being an animist (i.e., recognising similar interiorities and dissimilar physicalities between entities) has very specific implications as to how figuration and production will be performed and implemented. Inter alia, Descola states that within the animistic ontology, it is impossible to produce works of art where composite beings are represented or to indulge in herding. Descola claims that his framework is useful and fruitful to the extent that it can explain, for example, why Amazonian people have always been so reluctant to breed animals or to depict supernatural beings in the way Incas or Aztec would do so. The strength of the four ontological schemas would be that they allow us to make very specific predictions: given a specific ontological schema we can confidently infer a set of practices compatible with such a schema and a set of other practices which are incompatible with it.
    So far so good. The problem, however, is that there are plenty of cases which seem to be at odd with Descola’s predictions. Here is one example. In *The Origins of Monsters* (chap. 2, fn 24 and 25), you refer to Karadimas’ work on composite beings (Karadimas, 2010). In his chapter, Karadimas mainly uses European material. Strikingly enough, his ideas nicely fit with Descola’s overall framework (Descola (ed), 2010): Karadimas shows how widespread composite beings were in the iconography of the Middle Ages and of the Renaissance (i.e., in a time where, in Descola’s terms, the analogical ontology was prominent). And yet, as explained before, Karadimas’ work deals only partly with European iconography; most of his work concentrates on the Miraña people of Columbian Amazonia. Now it is particularly interesting to notice that Miraña people, who, by all Descolian standards possess an animistic mode of identification, are nonetheless prone to represent composite beings through various mediums. We can then wonder: how is it that one can be an animist (i.e., interact with spirits before hunting expeditions, transform herself into an animal, and furthermore be very reluctant to breed any animal) and yet represent chimeras while Descola clearly predicts that such representations are only compatible with analogism?
    A way to solve the problem would be to say: in fact, Miraña people are analogists and not animists. But this would sound way too ad hoc. Moreover, it wouldn’t solve anything for we would then have to explain how is it that one can be an analogist and still be reluctant to breed animals while Descola’s framework predicts that the two are not incompatible whatsoever. To be sure, the framework developed in *Beyond nature and culture* is very elegant; but, it seems to me, it is replete with numerous pitfalls similar to the one just mentioned. If you choose to endorse this wobbly framework, as you seem to do in *The Origins of Monsters*, you then have to explain, among other things, why is it that composite beings can be encountered in the Amazon just as easily as they can be encountered in the iconography of the Middle Ages.

    (4) My final question concerns the specific scope of your thesis that “composites typically belong to the image-worlds of cities and hierarchical states”. What do you exactly have in mind when you speak of “cities and hierarchical states”?
    Agriculture, urban centres, hierarchy and State-structure are often lumped together, as if they were all tightly co-dependent, but ethnographic and ethnohistorical data suggest otherwise: there are cases in which these four components are not co-present. For example, the Calusa were famous for being a hierarchical and yet a non-agricultural society (Goggin & Sturtevant, 1964; Widmer, 1998).
    So, which component(s), among the four above-mentioned, is (are?) definitely required for representations of chimeric creatures to appear and to spread around? The reason why I would like to know which of the four components you deem to be crucial is that it could help us clarify what the situation is in the Amazon. At first sight, no agriculture, no urban centres no hierarchy and not state is to be found in Amazonia. But here are some subtleties and nuances. First, pre-Columbian and pre-historic Amazonia looked very different from what it is now. The forest was traversed by plenty of roads. Populous towns had been settled in many places and especially in the regions contiguous to the rivers. The mode of production and the economy as well were very different from what they are now. Even if we consider present-day Amazonia, things are a bit more complicated than what appears at first glance. Some (relatively weak) forms of social stratification can be encountered in groups of the northwest amazon while it is extremely difficult to find similar patterns among the Pano groups. All these points have been abundantly documented and discussed in the last years (Balée & Erickson, 2006; Denevan, 1991; Hill, 1996; Hornborg, 2005; Neves, 1999; Santos-Granero 1992, 2002). As I suggested above, the existence of composite beings in Amazonian iconography is problematic; given the very diversity which can be encountered within the Amazon, things turn out to be even more complicated (e.g., Arawak socities are different from Pano societies). And, on the top of that, the various heterogeneous historical layers which define contemporary Amazonia make things even more difficult.

    To conclude, I want to say that I have mainly addressed ideas of your book I found disputable and I haven’t say a word about those which I found utterly convincing, but I want to make clear that, by and large, I did find your ideas utterly convincing!

    • David Wengrow 26 January 2016 (01:19)

      Martin I am working towards a fuller response with regard to Carlo Severi’s book and how it relates to mine. The ‘Origins of Monsters’ and the ‘Chimera Principle’ tackle much the same problem from different angles, and this makes the comparison exciting. Thank you for pushing the discussion in that direction.

      For the moment I just want your help in clarifying a few things. I am no expert on Amazonian societies, but I have read Severi carefully, admittedly in the new English translation. It seems to me that what you describe as ‘material representations of chimeric creatures’ actually have a very specific character and place in his argument, and that he himself is keen to differentiate them from the kind of composite figures discussed in my book.

      This distinction seems crucial for Severi, and he goes into it in some detail: see for instance the careful contrast he draws between the Greek chimera and the Hopi snake-lightning-bird figure (on p. 67 ff. of ‘The Chimera Principle’), in terms of their distinct principles of integration. The difference, as I understand him, resides in the relationship between image and memory.

      Old World composites, of the kind I’m concerned with in my book, are constructed in a way that leaves very few gaps on the visual plane – they are full and often compelling visual representations of animals that don’t exist, and as such they could serve among other things as illustrations to sequential narratives.

      The 'chimerical representations' of the New World that Severi writes about cannot possibly be represented in such a simple and unambiguous way on the visual plane, at least not as he describes them. They are not illustrations to narratives, but part of a far more complex memory apparatus that is activated in shamanic journeys between the visible and the invisible worlds.

      If I understand Severi right, the ‘chimera principle’ in Amazonian art – and other American examples he gives – works through incompleteness, using visible images as indirect cues for mental images that have no tangible existence. Take for instance his discussion of the celestial jaguar, whose oscillation between bird and jaguar is in fact conveyed mainly on the auditory plane, and only very obliquely through visible images.

      This seems completely consistent with the idea that stable and tangible images of mid-transformation are generally to be avoided in such contexts, because the shaman’s capacity as a heroic interlocutor with the invisible is based on remaining in a state of perpetual transformation. Whether the shaman happens to be a hunter-gatherer, horticulturalist, or city-dweller is a secondary point.

      As Severi himself puts it, the Kuna shaman is trapped in a series of unremitting metamorphoses, while his celestial foe – the bird-jaguar – cannot be represented in images other than those that ‘appear either through dazzling sunlight or else as dream-images that can only be seen by closed eyes’.

      That is not to say there are no images of hybrid figures in indigenous American art, but rather that they have an indexical role in relation to language, ritual, and memory quite distinct from that of my Old World composites. As Severi points out this is also reflected in their distinct morphology, which has more in common with the prehistoric hybrid figures discussed in my Chapter 3 (‘The hidden shaman’; e.g. Fig 3.7). These are not composites. Their ambiguous forms result instead from a ‘thoughtful play on continuities and resemblances in the appearance of various species’, just like the Siberian toggle pin discussed early on in Severi’s book (his Fig. 13), which can be wolf or seal depending on the perspective we bring to it, and also like so much of the indigenous American art he discusses.

      The comparison between Severi’s book and mine could be extended in other ways that I hope to get around to. But on these particular points I would see them as mutually reinforcing, rather than contrary arguments. Or have I misunderstood what his chimera principle is really about?

  • David Wengrow 20 January 2016 (11:07)

    Thanks Martin very much for your detailed and stimulating remarks - some of my initial reactions are interspersed through the subsequent responses to Jeremy, Alberto, and Pascal, but I realise they are not adequate. I hope to return to these issues in the coming days. Regards, David