The ‘trait list’

Over the last 40 years the so-called ‘trait list’ has become the dominant methodology for the identification of modernity from Palaeolithic archaeological traces (e.g. Mellars Reference Mellars and Renfrew1973; Reference Mellars1989a; Reference Mellars, Mellars and Stringer1989b; McBrearty and Brooks Reference McBrearty and Brooks2000; d'Errico Reference D'Errico2003; Klein Reference Klein2008). The trait list is a suite of archaeological features or properties that, when present, are held to record past behaviours or practices possible only if directed by a modern mind. The list has been subject to much revision since Mellars (Reference Mellars and Renfrew1973) produced the prototype. Some traits that figured in early manifestations, such as prismatic blade technology, have fallen decisively out of favour. The utility of others remains the subject of disagreement, as with the question whether a subsistence focus on a single game species (e.g. Mellars Reference Mellars1989a; Reference Mellars2004b) or broad-spectrum subsistence strategies (e.g. Hockett and Haws Reference Hockett and Haws2003) denote modernity.

There is, however, widespread agreement in palaeoanthropology that symbolically mediated thought and behaviour are the hallmarks of the modern human mind. Recent reviews by Nowell (Reference Nowell2010, 447) and Marean (Reference Marean, Mellars, Boyle, Bar-Yosef and Stringer2007, 367) agree that symbolism lies at the core of modernity. As Texier et al. (Reference Texier, Porraz, Parkington, Rigaud, Poggenpoel, Miller, Tribolo, Cartwright, Coudenneau, Klein, Steele and Verna2010, 6180) put it, ‘symbolically mediated behavior has emerged as one of the few universally accepted markers of behavioral modernity’. Chase and Dibble (Reference Dibble1987; Reference Chase and Dibble1990); Mellars (Reference Mellars1989a; Reference Mellars, Mellars and Stringer1989b; Reference Mellars1996; Reference Mellars2004a); Gargett (Reference Gargett1989; Reference Gargett1999); Klein (Reference Klein1995; Reference Klein1999; Reference Klein2000; Reference Klein2008); Knight, Powers and Watts (Reference Knight, Powers and Watts1995); Mithen (Reference Mithen1996); Ambrose (Reference Ambrose1998); Gamble (Reference Gamble1998); McBrearty and Brooks (Reference McBrearty and Brooks2000); Wadley (Reference Wadley2001); Henshilwood (Reference Henshilwood, Mellars, Boyle, Bar-Yosef and Stringer2007); Henshilwood, d'Errico and Watts (Reference Henshilwood, d'Errico and Watts2009); and Henshilwood and Dubreuil (Reference Henshilwood and Dubreuil2011), for example, all cite one or more putative expressions of symbolism – mortuary ritual, art, personal ornamentation, geometric decoration, standardized and regionally specific artefact style or the extension of social relations in time and space – as distinctive expressions of the modern human condition. Henshilwood and Marean go further and recognize only a single trait list element – symbolism, visible archaeologically through the presence of ‘art work, personal ornamentation and the social use of space’ (Henshilwood and Marean Reference Henshilwood and Marean2003, 635). Archaeologists who champion the cognitive and behavioural modernity of Neanderthals do so on the basis that their archaeology also features symbolic objects (d'Errico Reference D'Errico2003; Zilhão Reference Zilhão2007; Zilhão et al. Reference Zilhão, Angelucci, Badal-García, d'Errico, Daniel, Dayet, Douka, Higham, Martinez-Sanchéz, Montes-Bernárdez, Murcia-Mascarós, Pérez-Sirvent, Roldán-García, Vanhaeren, Villaverde, Wood and Zapata2010; Caron et al. Reference Caron, d'Errico, Del Moral, Santos and Zilhão2011). Even the growing number of archaeologists working towards an understanding of modernity as an emergent property of social, demographic and ecological dynamics feel impelled to explain the emergence of symbolism in those terms (e.g. Hopkinson Reference Hopkinson2007, 125–26; Powell, Shennan and Thomas Reference Powell, Shennan and Thomas2009).

Symbols and culture

Despite this consensus, symbolism in modern human origins is seriously undertheorized. Nevertheless, its basic conceptual architecture is clear: symbolism, modern culture and language are indivisible aspects of a unitary cognitive phenomenon, arisen in the course of human evolution, that entails systems of meaning mediated through symbols. One cannot exist, at least as potential, without the others. Consequently, evidence for the presence or absence of symbolism is taken also to be evidence for the presence or absence of modern culture and language. The necessary association of these elements is repeatedly affirmed in the literature (e.g. Dibble Reference Dibble, Mellars and Stringer1989; Mellars Reference Mellars1989a; Reference Mellars2004a; Whallon Reference Whallon, Mellars and Stringer1989; Marshack Reference Marshack1996; Conard Reference Conard2003; Reference Conard, Malina and Münzel2009; d'Errico Reference D'Errico2003; Carbonell and Mosquera Reference Carbonell and Mosquera2006; Bouzouggar et al. Reference Bouzouggar, Barton, Vanhaeren, d'Errico, Collcut, Higham, Hodge, Parfitt, Rhodes, Schwenninger, Stringer, Turner, Ward, Moutmir and Stamboul2007; d'Errico and Henshilwood Reference Henshilwood, Mellars, Boyle, Bar-Yosef and Stringer2007; Marean et al. Reference Marean, Bar-Matthews, Bernatchez, Fisher, Goldberg, Herries, Jacobs, Jerardino, Karkanas, Minichillo, Nilssen, Thompson, Watts and Williams2007; d'Errico et al. Reference D'Errico, Henshilwood, Vanhaeren and van Niekerk2005; d'Errico, Vanhaeren and Wadley Reference D'Errico, Vanhaeren and Wadley2008; Henshilwood, d'Errico and Watts Reference Henshilwood, d'Errico and Watts2009; Texier et al. Reference Texier, Porraz, Parkington, Rigaud, Poggenpoel, Miller, Tribolo, Cartwright, Coudenneau, Klein, Steele and Verna2010; Henshilwood and Dubreuil Reference Henshilwood and Dubreuil2011). It is stated with admirable clarity by Chase and Dibble (Reference Dibble1987, 264):

The authors go on to argue that there is no evidence for symbolism in the Middle Palaeolithic, so Neanderthals cannot have had modern minds, language or culture. Similarly, Byers (Reference Byers1994) draws an all-or-nothing distinction between end-goal-directed material behaviour – the purely functional application of tools to the achievement of a material goal – and end-product-directed material action, in which both the artefact and the end to which it is directed are also governed by symbolically constituted cultural rules. The Middle–Upper Palaeolithic transition records the evolutionary leap from non-symbolic behaviour to symbolically constituted action (Byers Reference Byers1994, 370).

Only recently have some researchers begun to grapple with these issues. Chase now embraces the possibility of symbolic thought in apes and archaic humans (Chase Reference Chase2006, 64). But beyond these revisions Chase's dualism is undiminished: only human beings create cultural meanings (ibid., 2), while only human codes and conventions are arbitrary (ibid., 12, 30). In a similar vein, Shea (Reference Shea2011) has critiqued the equation of modernity with symbolism on the grounds that much past symbolic behaviour might not have left material traces, and rejects the very notion of ‘modern behaviour’ as essentialist. Yet his objections are primarily methodological, and he perpetuates the problem through accepting symbolism as a universal feature of living humans, by presenting symbolism as those aspects of behaviour that contradict cost–benefit (i.e. functional) considerations, and by seeking to replace one essentialist dualism with another, between behavioural flexibility and inflexibility (Porr Reference Porr2011). To all practical intents and purposes, the theoretical understanding of symbolism, culture and modernity developed over 25 years ago is still in place today.

Symbols and things

The crux of the matter is how archaeologists can recognize symbolism when they see it. As we have seen, Henshilwood and Marean (Reference Henshilwood and Marean2003) identify art objects and personal ornaments as unequivocal symbols, while McBrearty and Brooks (Reference McBrearty and Brooks2000) point also to artefact style, pigment use and ritual burial. In his original trait list Mellars recognized that the transition from the Upper to the Middle Palaeolithic in Europe entailed ‘a range of striking cultural innovations – including . . . the adaptation of animal teeth, marine shells etc. as personal ornaments, and the emergence of a sophisticated naturalistic tradition of art’ (Mellars Reference Mellars and Renfrew1973, 255). Claims and counterclaims for ritual (e.g. Solecki Reference Solecki1975; Gargett Reference Gargett1989; Reference Gargett1999; Hovers et al. Reference Hovers, Ilani, Bar‐Yosef and Vandermeersch2003; Hovers, Kimbel and Rak Reference Hovers, Kimbel and Rak2000); body decoration (e.g. d'Errico et al. Reference D'Errico, Henshilwood, Vanhaeren and van Niekerk2005; Bouzouggar et al. Reference Bouzouggar, Barton, Vanhaeren, d'Errico, Collcut, Higham, Hodge, Parfitt, Rhodes, Schwenninger, Stringer, Turner, Ward, Moutmir and Stamboul2007; Zilhão Reference Zilhão2007; Texier et al. Reference Texier, Porraz, Parkington, Rigaud, Poggenpoel, Miller, Tribolo, Cartwright, Coudenneau, Klein, Steele and Verna2010; Zilhão et al. Reference Zilhão, Angelucci, Badal-García, d'Errico, Daniel, Dayet, Douka, Higham, Martinez-Sanchéz, Montes-Bernárdez, Murcia-Mascarós, Pérez-Sirvent, Roldán-García, Vanhaeren, Villaverde, Wood and Zapata2010; Morin and Laroulandie Reference Morin and Laroulandie2012); art (e.g. Conard Reference Conard2003; Reference Conard2009; Marquet and Lorblanchet Reference Marquet and Lorblanchet2003; Pike et al. Reference Pike, Hoffmann, García-Diez, Pettitt, Alcolea, De Balbín, González-Sainz, de las Heras, Lasheras, Montes and Zilhão2012); and, more recently, musical instruments (Kunej and Turk Reference Kunej, Turk, Wallin, Merker and Brown2000; Conard, Malina and Münzel Reference Conard, Malina and Münzel2009) in the Palaeolithic have dominated debate over when, where, and in which hominin species modernity emerged.

But what makes these objects and practices symbolic? It is the absence of physical function, in the narrowly utilitarian sense, in resource extraction and maintenance. This is Byers's distinction between symbolic end-product action, which incorporates rule-bound arbitrary style, and end-goal behaviour, the artefactual instruments of which display form wholly determined by objective functional factors. Symbolism is the residue when all functional dimensions of an archaeological artefact or pattern have been accounted for. By this reckoning artefacts and practices, or aspects thereof, that have no direct utilitarian function must be symbolic. Equally, tools such as scrapers and awls possess a symbolic dimension if they can be shown to display arbitrary, non-functional form or style (Chase and Dibble Reference Chase and Dibble1987, 266). If not, their makers cannot have been modern. Formal types express cultural norms or values and, by extension, linguistically structured symbolic categories.

This amounts to an argument for modernity as typology. Neanderthals cannot have thought symbolically, according to Dibble (Reference Dibble1984; Reference Dibble1987; Reference Dibble, Mellars and Stringer1989; Rolland and Dibble Reference Rolland and Dibble1990) because archaeological typologies of Middle Palaeolithic artefacts are illusory and do not reflect arbitrary cultural norms in the past. Instead they simply represent points on continua of form through which stone tools passed as they were repeatedly resharpened and reduced before eventual discard. The same argument for modernity as artefact style can be seen in Pfeiffer (Reference Pfeiffer1982), Sackett (Reference Sackett1982), Mellars (Reference Mellars1989a, 365), Klein (Reference Klein2000, 26; Reference Klein2008, Table 1), and McBrearty and Brooks (Reference McBrearty and Brooks2000, Table 3), among others. Those who insist that arbitrary form can indeed be seen in Middle Palaeolithic artefacts, and that Neanderthals were therefore cognitively modern (e.g. d'Errico Reference D'Errico2003; Hopkinson Reference Hopkinson and Conard2004), are relying on identical logic; they are simply quibbling over the details.

There have, of course, been dissenting voices. For Leroi-Gourhan (Reference Leroi-Gourhan1964) and Holloway (Reference Holloway1966; Reference Holloway1981), all artefact fabrication necessarily indicates linguistic capacities and the imposition of arbitrary form on the environment. On the other hand, Wynn (Reference Wynn1985; Reference Wynn1991; Reference Wynn, Gibson and Ingold1993) and Chazan (Reference Chazan1995) have expressed scepticism that artefact form records any information about past linguistic and symbolic capacities. But these objections have not succeeded in challenging the dominant view that symbolism, culture and language are the hallmarks of ‘people like us’. Between these extremes, Botha (Reference Botha2008; Reference Botha, Botha and Knight2009; Reference Botha2010) has championed a narrower concept of symbolism by challenging the symbolic status of the pierced shell beads of the African MSA since we cannot be confident that they were repositories of arbitrarily attributed meaning, and are therefore better understood as indexes. But Botha does not abandon symbolism. Instead he argues that symbolism, and thus cultural and linguistic modernity, appear only in the Upper Palaeolithic after 45 kya. The origins of symbolism and the origins of the modern human condition are still seen as one and the same.

Theoretical issues with symbolism

Palaeoanthropology has been remarkably uninterested in the theorization of the symbol in even closely related disciplines. Neither the structural and symbolic movement in theoretical archaeology (e.g. papers in Hodder Reference Hodder1982) nor structuralist and symbolic anthropologies (e.g. Leach Reference Leach, Hugh-Jones and Laidlaw1968; Reference Leach1976; Lévi-Strauss Reference Lévi-Strauss1978; Needham, Reference Needham1979) have exerted any significant influence. Even C.S. Peirce's seminal 19th-century work, distinguishing between icons (signs whose meaning is directly related to their form or properties), indexes (signs that possess some sensible feature that implies or points to something else) and symbols (signs whose relationship to their referent is entirely arbitrary) has had little impact until very recently (Botha Reference Botha2008; Reference Botha, Botha and Knight2009). Yet this trichotomy of signs, which remains a foundation of semiotics, has the potential to challenge even the categorization of the Aurignacian's most impressive art objects as symbols.

The remarkable figurines from the Aurignacian site cluster of Vogelherd, Hohle Fels, Hohlensteinstadel and Geissenklosterle in southern Germany (Conard Reference Conard2003; Reference Conard, Malina and Münzel2009) are a case in point. The figurines, which include animals, a Venus and two therianthropic ‘lion men’, are all more or less naturalistic, and some, such as the Hohle Fels waterfowl, and the horse and mammoth from Vogelherd, are exquisitely so. But, in the trichotomy of signs, naturalistic representations are iconic rather than arbitrary and symbolic; one can know simply by seeing it that the horse figurine represents a horse. Even if these figurines instead referred to values implied by their form (e.g. speed, the hunt or kinship relations) they would still constitute indexes, not symbols. The presumption that these objects are obviously symbolic reveals that the symbol concept is used without real exploration of its theoretical ramifications.

More seriously, perhaps, the application of symbolism in palaeolithic archaeology has, with few exceptions (e.g. Gamble Reference Gamble1998; Reference Gamble1999; Reference Gamble2012; Hopkinson and White Reference Hopkinson, White, Gamble and Porr2005) been immune to theoretical developments in archaeology since the early 1980s. There is little or no recognition of the ambiguous and contested character of material-culture meaning, or of Palaeolithic hominins as agents (but see papers in Gamble and Porr Reference Gamble and Porr2005). Archaeological theory today is deeply distrustful of the sort of dualism that has been discussed. If meaning is arbitrarily attributed to symbolic artefacts then the artefact and its meaning have no organic connection and occupy separate ontological realms; the artefact is a passive receptacle and symbolic meaning becomes an extension of Descartes's ‘ghost in the machine’, a spirit that animates matter but is not of it. Contemporary theoretical perspectives concerned with praxis (that in making and transforming things in social practice, people also make and transform themselves as social beings, so that people, things, performances and meanings are mutually constitutive) offer an escape from this dualist impasse. Although relational archaeologies of praxis and being have begun to make some headway in Palaeolithic and, more especially, in Mesolithic archaeology (e.g. Conneller Reference Conneller2010), they have had little impact on the symbolism orthodoxy.

It appears that, as a discourse, palaeoanthropology is concerned with symbolism only insofar as it is an observable diagnostic trait of a particular class of organism, the modern human being. As we shall see, this is crucial to understanding symbolism's rise to prominence in this field of enquiry.

Internal problems with the ‘Symbolism = modernity’ orthodoxy

These objections are of little significance to most palaeoanthropologists since they derive from theoretical paradigms to which they do not subscribe. But even on its own terms, the ‘symbolism = modernity’ paradigm has serious weaknesses. One of the more obvious is the possibility that archaic hominins might have made symbolic objects on perishable materials that have not survived, or that their ritual practices left no traces on durable materials (Speth Reference Speth2004; Shea Reference Shea2011). Another is that landscape features such as rivers, hills and forests can be imbued with meaning (Bradley Reference Bradley2000), though Palaeolithic archaeologists prefer to render them as resources or problems for hominins to exploit or solve through adaptive behaviour. A possibly more serious problem is that the arbitrariness of the relation between the symbol and its referent is simply impossible to reconcile with the criteria used to distinguish symbolic from non-symbolic artefacts. If the relation is truly arbitrary, then any meaning can surely be assigned to anything. This directly contradicts the denial of meaning in artefacts that appear to lack arbitrary style.

If Middle Palaeolithic stone artefacts were the products of exclusively functional considerations, lack standardized arbitrary form and therefore cannot have been produced by symbolically thinking hominins (Dibble Reference Dibble1984; Reference Dibble1987; Reference Dibble, Mellars and Stringer1989; Rolland and Dibble Reference Rolland and Dibble1990; Davidson and Noble Reference Davidson, Noble, Gibson and Ingold1993), then some classes of things-in-the-world, and not others, are capable of carrying meaning, and the relationship between symbol and referent cannot be entirely arbitrary. This is an unsustainable position. Not only is it a theoretical oxymoron, it also contradicts experience. If the informality of Middle Palaeolithic stone artefacts precluded them from bearing meaning, then they could not be meaningful for archaeologists in the present. That is obviously not the case. Through practical engagement with them, these artefacts can come to embody for us our professional identity and form links in chains of association with places, people and events that contribute to situating us in the world. Since these artefacts can carry meaning, the inference of a cause (a non-symbolic Neanderthal mind) from the alleged effect (non-stylistic artefacts) cannot be sustained.

The symbolism orthodoxy also struggles to make sense of the paucity of symbolism in the archaeological record of Pleistocene Homo sapiens in some regions of the world. The Niah Cave ‘Deep Skull’ demonstrates that anatomically modern Homo sapiens was present in Borneo between 35 and 44 kya, at the same time as Homo sapiens was entering Europe equipped with Aurignacian material culture. But the archaeology associated with the Deep Skull contains nothing that can be described as non-functional or stylistic (Barker et al. Reference Barker, Barton, Bird, Daly, Datan, Dykes, Farr, Gilbertson, Harrisson, Hunt, Higham, Kealhofer, Lewis, McLaren, Paz, Pike, Piper, Pyatt, Rabett, Reynolds, Rose, Rushworth, Stephens, Stringer, Thompson and Turney2007). Are we to conclude that Homo sapiens at that time was cognitively modern in Europe but not in Borneo? Of course, nobody draws any such conclusion. Because it is Homo sapiens the cognitive modernity of the brain that was once inside the Deep Skull is presumed, despite associated archaeology that does not fulfil the stipulated criteria for symbolism. This is simply a double standard (Roebroeks and Corbey Reference Roebroeks, Corbey, Corbey and Roebroeks2001).

Neanderthal symbolism

Claiming modernity in Neanderthals on the ground that their archaeology also features symbolic objects does not in itself challenge the ‘symbolism = modernity’ paradigm; it is just quibbling over the details of whose archaeology passes the symbolism test. However, it seems unlikely that the symbolism paradigm could long survive a convincing demonstration of Neanderthal symbolic capacities.

The trait list of which symbolism is the core element was originally devised with the explicit intention of systematizing the behavioural differences between Homo sapiens (understood in terms of the European Upper Palaeolithic) and the Neanderthals, as represented by the Middle Palaeolithic. Neanderthals by definition could not have been modern humans. Any convincing demonstration that the Neanderthals produced and used symbolically charged artefacts therefore presents an existential challenge to the ‘symbolism = modernity’ consensus. That this is so is illustrated by amendments to the modernity trait list, driven primarily by Middle Palaeolithic archaeology's awkward habit of meeting its conditions. Bluntly, the rules of the game have been repeatedly changed to keep the Neanderthals out. Prismatic blade technology, game species specialization, the systematic exploitation of marine resources, fowling and the occupation of marginal environments is just a sample of the traits that have faded from the modernity list since it became clear that, at least sometimes, Neanderthals practised them (Conard Reference Conard1990; Ameloot-van der Heijden Reference Ameloot-van der Heijden1993; Grayson and Delpech Reference Grayson and Delpech2002; Reference Grayson and Delpech2008; Stringer et al. Reference Stringer, Finlayson, Barton, Fernández-Jalvoe, Cáceres, Sabin, Rhodes, Currant, Rodríguez-Vidal, Giles-Pacheco and Riquelme-Cantal2008; Blasco and Fernández Peris Reference Blasco and Fernández Peris2009; Krause et al. Reference Krause, Orlando, Serre, Viola, Prüfer, Richards, Hublin, Hänni, Derevianko and Pääbo2007). This ‘ticking off’ of allegedly modern traits by the archaeology of the Neanderthals has contributed to the increased reliance on symbolism as the sole robust indicator of modernity. Yet recent work has now begun to erode even that last bastion of our species’ specialness, and thrown into doubt the claim that the western Eurasian Middle Palaeolithic was symbolically impoverished.

Marshack (Reference Marshack1976; Reference Marshack and Otte1988; Reference Marshack and Mellars1990; Reference Marshack1996; Reference Marshack1997) and Bednarik (Reference Bednarik1992; Reference Bednarik1995) have sought, largely in vain, to draw a sceptical discipline's attention to a significant and long-known corpus of apparently non-utilitarian Middle and even Lower Palaeolithic finds, including ochre pigment blocks and modified animal parts such as the Tata carved mammoth molar, bone and tooth pendants from La Quina, and incised bone from La Ferrassie and Bilzingsleben. Until recently, the most intensely discussed possible evidence for Neanderthal symbolism was the early Upper Palaeolithic Châtelperronian industry of central and south-west France and northern Spain, notable for non-utilitarian artefacts including pierced and grooved animal teeth and well-made bone tools often bearing decorative incisions. Neanderthal remains recovered from Châtelperronian contexts at Saint Césaire (Lévêque and Vandermeersch Reference Lévêque and Vandermeersch1980) and the Grotte du Renne, Arcy-sur-Cure (Hublin et al. Reference Hublin, Spoor, Braun, Zonneveld and Condemi1996) imply that the symbolic artefacts were made by late Neanderthals. The issue has unfortunately become bogged down in disputes over whether the Châtelperronian was regionally contemporaneous with the Aurignacian and so might represent Neanderthal ‘acculturation’ (e.g. Harrold Reference Harrold, Mellars and Stringer1989; Mellars Reference Mellars1989a, 353; Reference Mellars2005; d'Errico et al. Reference D'Errico, Zilhão, Baffier, Julien and Pelegrin1998; Klein Reference Klein1999; Zilhão and d'Errico Reference Zilhão and d'Errico1999; d'Errico Reference D'Errico2003; Zilhão Reference Zilhão2006), and the stratigraphic integrity of the deposits at Grotte du Renne (Higham et al. Reference Higham, Jacobi, Julien, David, Basell, Wood, Davies and Bronk Ramsey2010; Caron et al. Reference Caron, d'Errico, Del Moral, Santos and Zilhão2011). The comparable Uluzzian of Italy is similarly disputed (Riel-Salvatore Reference Riel-Salvatore2010; Benazzi et al. Reference Benazzi, Douka, Fornai, Bauer, Kullmer, Svoboda, Pap, Mallegni, Bayle, Coquerelle, Condemi, Ronchitelli, Harvati and Weber2011).

Recently, however, plausible symbolic objects have been reported from several unequivocally Middle Palaeolithic contexts. Three perforated marine shells have been found alongside pigments in the Middle Palaeolithic levels at Cueva de los Aviones, southern Spain, along with a Spondylus shell bearing residues of a red colourant produced by mixing three pigments. A broken perforated Pecten shell, partly pigmented with a mixture of goethite and haematite, is reported from the Cueva Antón, also in southern Spain (Zilhão et al. Reference Zilhão, Angelucci, Badal-García, d'Errico, Daniel, Dayet, Douka, Higham, Martinez-Sanchéz, Montes-Bernárdez, Murcia-Mascarós, Pérez-Sirvent, Roldán-García, Vanhaeren, Villaverde, Wood and Zapata2010). All these finds are dated to some 50 kya. Most remarkably, it now seems that Neanderthals across a wide swathe of Europe were systematically removing wing feathers and/or claws from raptors and corvids for display purposes (Peresani et al. Reference Peresani, Fiore, Gala, Romandini and Tagliacozzo2011; Finlayson et al. Reference Finlayson, Brown, Blasco, Rosell, Negro, Bortolotti, Finlayson, Marco, Pacheco, Vidal, Carrión, Fa and Llanes2012; Morin and Laroulandie Reference Morin and Laroulandie2012).

If this emerging picture of a symbolically competent Neanderthal mind is confirmed by further finds and analyses, there will be no remaining characteristic that can be called upon to distinguish Neanderthals cognitively from Pleistocene Homo sapiens. Neanderthals might then be admitted into membership of modern humanity, but the ultimate consequence is likely to be the collapse of the ‘symbolism = modernity’ edifice. A theoretical perspective and derived methodology which were designed to identify and describe distinctively modern humans and their behaviour, but which can no longer pinpoint any qualitative behavioural transformations in the 300,000-year spans of the Eurasian Middle–Upper Palaeolithic and the African Middle–Late Stone Age, must suffer a disabling loss of credibility.

Summary

It is becoming increasingly difficult to sustain the ‘symbolism = modernity’ paradigm as it struggles to deliver on its historical mission – the delineation of the singular difference between Homo sapiens and all other hominin species. Since at least some of its problems are inherent in its theoretical and methodological structure, it is pertinent to ask why palaeoanthropology came to put all its eggs in the symbolism basket. A consideration of when and from where palaeoanthropology came to embrace symbolism as modernity's essential characteristic suggests a route to a possible answer to that question.

Symbolism in early 20th-century American social science

Discussions of symbolism and modernity in palaeoanthropology rarely cite any sources from which the symbolism concept has been derived. I would argue that the ultimate source lies in early 20th-century American cultural anthropology, which from its inception in the work of Franz Boas accorded symbols a central role in human culture. Two of Boas's early ethnographic studies explored the role of material-culture symbolism in indigenous cultures of the American North-west, looking first at the symbolic role of masks in the social organization in the Kwakiutl (Boas Reference Boas1897) and then at the significance of formal variation in Alaskan needlecases (Boas Reference Boas1908). He went on, particularly in The mind of primitive man (Boas Reference Boas1911) to argue that all human beings share the same intellectual capacities, and that all cultures are equally founded on biology, language and material and symbolic culture. Boas trained most of the leading figures in American cultural anthropology in the first part of the 20th century, including Alfred Kroeber, Margaret Mead and Leslie White. Indeed, Kroeber's doctoral dissertation dealt with decorative symbolism in the Arapaho, and he subsequently defined culture as ‘patterns, explicit and implicit, of and for behavior acquired and transmitted by symbols, constituting the distinctive achievement of human groups, including their embodiments in artefacts’ (Kroeber and Kluckhohn Reference Kroeber and Kluckhohn1952). A conviction that culture and symbolism are inseparable can therefore be seen as a foundational concept in American cultural anthropology. As the discipline diversified theoretically, symbolism retained its centrality in such disparate movements as interpretive symbolic anthropology (e.g. Sahlins Reference Sahlins1976; Geertz Reference Geertz1973) and cultural materialism (Harris Reference Harris1979). This stands in contrast to much of British and French social anthropology. The former certainly produced important works of symbolic anthropology (e.g. Douglas Reference Douglas1966; Turner Reference Turner1967) but, unlike the structural-functionalist mainstream, they exerted negligible influence on British palaeolithic archaeology. French anthropology's focus on the significance of exchange and technique, and on structures of cultural meaning, had little impact on palaeolithic studies beyond the work of Leroi-Gourhan (Reference Leroi-Gourhan1964; Reference Leroi-Gourhan1965). In the early and mid-20th century it was in America that archaeologists, trained in ‘four-field’ departments of anthropology, were systematically exposed to the idea that symbolism is the key to culture.

Leslie White

The cultural anthropologist whose ideas most closely prefigured the current palaeoanthropological conception of symbolism was, however, Leslie White. Though a student of Boas, White soon rejected his teacher's culture-historical particularism in favour of a neo-evolutionary orientation. His first major book, The science of culture (1949), assembled a number of essays published over the preceding 11 years. White's central concerns at that time were twofold: first, the ontological status of culture, and second, the symbol. His position on the former is set out in the book's preface: culture is not inseparable from its human carriers, but is a discrete real-world phenomenon in its own right, moving according to its own laws and processes, ‘a self-contained, self-determined process; one that can be explained only in terms of itself’ (White Reference White1949, xviii). White developed this further in The concept of culture (Reference White1959a). He argued that cultural phenomena – ideas, acts, artefacts – can be investigated either in ‘somatic’ context, their relations with embodied ‘human organisms’, as components of behaviour; or in their ‘extrasomatic’ context, i.e. their relations between themselves, independently of embodied human organisms, as components of culture (ibid., 230–31). Since cultural phenomena and the relations between them are observable, so culture itself has phenomenal reality (ibid., 239).

But how to distinguish cultural phenomena from other classes of observable phenomena? On this, White is clear: ‘Culture . . . is a class of things and events, dependent upon symboling, considered in an extrasomatic context’ (1959a, 234; emphasis in original). Symbolism makes culture. In the somatic realm symbolically dependent things and events (‘symbolates’) influence behaviour, but relations between them in the extrasomatic realm constitute culture (ibid., 233). All things and acts have both objective (functional) and subjective (symbolic) components (ibid., 236). White draws direct equivalences between behaviour and speech, and between culture and language; the latter both entail relations between symbolates (e.g. syntax and grammar) that are independent of embodied behaving individuals, and govern behaviour, including speech utterances (ibid., 234).

In ‘The symbol’, the second essay in The science of culture, White declares that Darwin was wrong; the difference between people and animals is qualitative, not quantitative. It is the symbol that sets humanity apart. As White puts it, ‘An organism has the ability to symbol or it does not; there are no intermediate stages’. On the same page he cites approvingly Descartes's ascription of animals and humans to separate realms (White Reference White1949, 24). A symbol's meaning is arbitrarily assigned to it by human beings (ibid., 29), who can only apprehend its meaning in symbolic terms (ibid., 26). White briefly acknowledges the evolutionary origin of the symbol, but with no attempt at any account of how that might have happened. However, his conviction that there can be no semi-symbolic organisms implies that it must have been an evolutionary leap – a point he makes more explicitly in ‘The primate revolution’ (White Reference White1959b). Having come into being, it was ‘the symbol . . . which transformed our anthropoid ancestors into men and made them human’ (White Reference White1949, 22).

It seems clear from this brief review that White first formulated the framework that now dominates thought on the matter of modern human origins. His ideas reverberated through cultural anthropology, inspired his students (e.g. Sahlins Reference Sahlins and Spuhler1959; Spuhler Reference Spuhler and Spuhler1959) and are alive and well in palaeoanthropology today. All of the elements discussed above are to be found in contemporary palaeoanthropological discourse on modern human origins, sometimes with structurally minor adjustments, such as replacing White's human–animal dualism with our modern–archaic counterpart. Indeed, both Kuhn and Stiner (Reference Kuhn and Stiner2006) and Shea (Reference Shea2011) speak explicitly of symbol use as ‘extrasomatic’.

The rise of American palaeoanthropology

How, then, to explain this four-decade delay between White's ideas and their acceptance in palaeolithic archaeology? One key factor is the growth of American palaeoanthropology in the years following the Second World War and an influx of American archaeologists, schooled in the four-field tradition of American anthropology, into Old World palaeolithic archaeology from the mid-1960s.

Before the Second World War anthropological and archaeological studies of human evolution were dominated by European figures in whose homelands and their imperial dominions palaeolithic research was carried out. The only pre-war American evolutionary physical anthropologist of genuine international stature was Ales Hrdlicka, while American prehistoric archaeology before 1945 was primarily concerned with the material record of its own indigenous peoples. Only during and after the Second World War did American researchers become internationally significant in palaeoanthropology in figures such as Hallam Movius and Sherwood L. Washburn. In subsequent decades, as American cultural power and reach grew, American palaeoanthropology expanded to the point where it now dominates the discipline.

Leslie White himself was doctoral supervisor to two figures of significance to this question: Lewis Binford and Arthur Jelinek. Although Binford was never a palaeolithic archaeologist per se, his influence on the discipline has been profound. His early programmatic works (e.g Binford Reference Binford1962) were heavily influenced by White's ideas, particularly his systematic view of culture, its objective reality and its separateness from psychology. From 1981, with the publication of Bones. Ancient men and modern myths, Binford was primarily responsible for introducing into anglophone palaeoanthropology a rigid archaic–modern dualism, in which non-modern hominins were depicted as wholly lacking in culture and thus not at all human. This position, reiterated in several subsequent publications (e.g. Binford Reference Binford and Trinkaus1989), clearly perpetuates White, though Binford rarely showed much interest in symbolism, focusing instead on the organization and planning depth of subsistence, mobility and technology.

Jelinek, who became the leading figure in palaeolithic archaeology at Arizona, represents another line of intellectual descent from Leslie White. He maintained an interest in the relation between lithic artefact form, function and style, with style conceived as the arbitrary residue left behind when utilitarian functional and material factors have been accounted for (Jelinek Reference Jelinek and Cleland1976). It is precisely this position which was further developed by Jelinek's Arizona students Chase and Dibble in their seminal 1987 paper. Together, Binford and Jelinek introduced to palaeoanthropology White's ideas about culture and symbolism and set the intellectual agenda for much of American and anglophone thinking on modern human cognitive and behavioural evolution from the 1980s onwards.

But, as was noted earlier, the identification of symbolism as the defining property of the modern human mind and culture was a general feature of the cultural anthropology in which the growing cohort of American palaeolithic archaeologists was schooled. Movius, for example, was fundamentally a Boasian particularist and a culture-historical thinker whose central interest was space–time cultural systematics. Yet, despite the differences between his and White's theoretical positions, Movius's ideas drew him towards problems of stylistic variation and cultural dynamism (Movius Reference Movius1949) and attribute analysis (Movius et al. Reference Movius, David, Bricker and Clay1969) that fed directly into the work of his students at Harvard such as Sackett, Bricker and Brooks. Sackett cannot be described as a neo-evolutionist, but his work on style (Sackett Reference Sackett and Renfrew1973; Reference Sackett1986; Reference Sackett, Conkey and Hastorf1990, 36) shares with Chase and Dibble (Reference Chase and Dibble1987) the conviction that arbitrary style is symbolically active within culture. Brooks, on the other hand, has embraced a non-Boasian evolutionism and champions what is now the symbolism orthodoxy (McBrearty and Brooks Reference McBrearty and Brooks2000). Finally, the Chicago school of palaeoanthropology, led by Washburn and his student F. Clark Howell, also produced important figures in the field, including Sally Binford, whose 1968 paper on symbolism has already been noted, and Richard Klein.

The symbolism orthodoxy should therefore be understood as a historically arisen paradigm introduced to human-origins studies from American cultural anthropology for historical reasons. This is not, however, a sufficient explanation for the orthodoxy's development and entrenchment. Another, quite separate driver of this was a profound change in the understanding of pattern and process in human evolution from 1980.

Phyletic gradualism in human evolution

During the course of the 1960s, under the influence of the neo-Darwinist New Synthesis, the notion of human evolution as a branching process driven by mutation, and of species as invariant essential types, gave way to phyletic gradualism, which understands evolution as a continuous process of adaptation in time and space within a single variable but unbranching population lineage (Cartmill Reference Cartmill, Corbey and Roebroeks2001). This provoked a pruning of hominin taxa. Sinanthropus pekinensis and Pithecanthropus erectus were subsumed into the genus Homo, which contained just three palaeospecies: Homo habilis, Homo erectus and Homo sapiens. These species were cast as grades arbitrarily subdividing an evolutionary continuum. With the exception of the early African Homo habilis, each species was a regionally variable but pan-Old World population. Homo sapiens was defined broadly to include Neanderthals, who were distinguished from anatomically modern humans only at the sub-species level – Homo sapiens neanderthalensis and Homo sapiens sapiens respectively. Other less well understood large-brained specimens were lumped together in a global ‘archaic Homo sapiens’ sub-grade. Homo sapiens sapiens in each region was understood to be the direct descendant of earlier Homo sapiens in the same region, making modern Europeans the descendants of the Neanderthals. This phyletic gradualist scheme subsequently hardened into the multiregional hypothesis for modern human origins (Wolpoff Reference Wolpoff and Trinkaus1989; Frayer et al. Reference Finlayson, Brown, Blasco, Rosell, Negro, Bortolotti, Finlayson, Marco, Pacheco, Vidal, Carrión, Fa and Llanes1993; Wolpoff, Hawks and Caspari Reference Wolpoff, Hawks and Caspari2000; Wolpoff et al. Reference Wolpoff, Hawks, Frayer and Hunley2001; Reference Wolpoff, Mannheim, Mann, Hawks, Caspari, Rosenberg, Frayer, Gill and Clark2004).

The impact of radiometric dating and DNA sequencing

New evidence that emerged from 1980 dealt a severe blow to the phyletic gradualist narrative of human phylogeny. First, the dating of the Saint-Césaire Neanderthal to 33–35 kya (Lévêque and Vandermeersch Reference Lévêque and Vandermeersch1980) meant that it was simply too close in time to the oldest known anatomically modern human specimens in Europe to have been their ancestor (Stringer, Hublin and Vandermeersch Reference Stringer, Hublin, Vandermeersch, Smith and Spencer1984). Similarly, the dating of the robust but anatomically modern specimens from Qafzeh, Israel, to c.125–74 kya (Valladas et al. Reference Valladas, Reyss, Joron, Valladas, Bar-Yosef and Vandermeersch1988), and of the Neanderthal specimen from Kebara, Israel, to 55–60 kya (Valladas et al. Reference Valladas, Joron, Valladas, Arensburg, Bar-Yosef, Belfer, Goldberg, Laville, Meignen, Rak, Tchernov, Tillier and Vandermeersch1987) meant that anatomically modern humans had been in the Levant at least 15,000 years before Neanderthals and so could not possibly have been their descendants.

Also in that period, the sequencing of mitochondrial DNA from living people (Cann, Brown and Wilson Reference Cann, Brown and Wilson1984; Cann, Stoneking and Wilson Reference Cann, Stoneking and Wilson1987; Stoneking and Cann Reference Stoneking, Cann, Mellars and Stringer1989) began to point to all living humans sharing an African female common ancestor who had lived between 120 and 240 kya. Subsequent studies (e.g. Vigilant et al. Reference Vigilant, Stoneking, Harpending, Hawkes and Wilson1991) reached a similar conclusion. Though proponents of the phyletic gradualist multiregional model continued (and continue) to argue for an inclusive Homo sapiens in which there had been no evolutionary discontinuities (Clark and Lindly Reference Chazan1989; Wolpoff Reference Wolpoff and Trinkaus1989; Frayer et al. Reference Finlayson, Brown, Blasco, Rosell, Negro, Bortolotti, Finlayson, Marco, Pacheco, Vidal, Carrión, Fa and Llanes1993), opinion shifted decisively towards a replacement model, in which modern humans had originated recently and only in Africa, and subsequently dispersed through the rest of the world, replacing archaic hominins such as the Neanderthals and Homo erectus – the ‘out-of-Africa’ hypothesis for modern human origins (see Stringer Reference Stringer2002 for a clear exposition).

Cladistic taxonomy and symbolism as autapomorphy

Taxonomy aims to resolve phylogenetic relationships between taxa by classifying them according to their observable similarities and differences. The more similar two organisms are, the more recent their last common ancestor must be. Taxonomic methods in use in the mid-20th century, however, were weak at resolving polyphyletic groups, i.e. classes of organisms that shared characteristics through independent convergent evolution, not through common ancestry. In the 1980s cladistics, a taxonomic methodology devised by Hennig in the early 1950s to resolve this problem (Hennig Reference Hennig1950; Reference Hennig, Davis and Zangerl1966) became established as the dominant phylogenetic method. Cladistics rests on a number of critical propositions (Cartmill Reference Cartmill, Corbey and Roebroeks2001):

1. 1. Evolutionary patterns are generated by ‘branching’ speciation events.

2. 2. Taxa must be monophyletic evolutionary branches or clades, and include a founder, the distinctive traits of which define the taxon, and all and only its direct descendants.

3. 3. Species are to be understood as logical individuals, not as variable populations.

4. 4. Organisms’ traits should be weighted as to their phylogenetic significance on the basis of whether they are ‘primitive’ plesiomorphies shared between several taxa and inherited from their more or less distant common ancestor; ‘derived’ synapomorphies, recently arisen traits that serve to identify close evolutionary relationships between sister taxa; or unique derived autapomorphies that serve to define the taxon as a discrete clade.

By the mid-1980s phyletic gradualism in palaeoanthropology was therefore under attack on two fronts. On one front, new evidence had cast very serious doubt on a Neanderthal ancestry for modern Europeans, and promoted a phylogeny in which living people were descended from a speciation event in the African late Middle Pleistocene. On another front, it was under threat from cladistics, which rejects both species as variable populations and evolution through cumulative change in an unbranching lineage. Palaeoanthropology's response was to abandon phyletic gradualism. By the late 1980s Klein (Reference Klein1989, 414–16) was explicitly adopting a cladistic position on hominin classification. Membership of Homo sapiens was restricted to living people and their recent Pleistocene African ancestors; the Neanderthals were expelled and returned to their original status as a separate species, Homo neanderthalensis. The poorly defined ‘archaic’ Homo sapiens was abandoned as a term, at least for non-African specimens, and its former members distributed somewhat contentiously among species including Homo heidelbergensis and Homo rhodesiensis. As large-brained but anatomically archaic specimens formerly attributed to Homo sapiens were hived off into separate species, out with them went the possibility that archaic and modern humans were not essentially different from one another.

But what unique derived autapomorphy defines the new, exclusive Homo sapiens? Neanderthals are readily defined by a suite of unique skeletal traits, but the same cannot be said of ourselves. Many apparently distinctive traits, such as reduced and bipartite brow ridges, are variably expressed, and in any case reflect a secondary loss of skeletal robusticity in recent Homo sapiens. Early specimens are hardly less robust than Neanderthals. Traits such as the canine fossa and the cranium's pentagonal profile in rear view distinguish Homo sapiens from Neanderthals, but not from the early African species Homo ergaster (Klein Reference Klein1999, figure 5.17). We are left with the unique possession of a chin or mental trigone as our species’ most reliable defining anatomical feature, and even that is absent in three of the early Homo sapiens specimens from Skhul (Schwartz and Tattersall Reference Schwartz and Tattersall2000).

Summary

The rise to dominance of the symbolism paradigm should be seen in this specific historical context. Biological anthropologists, of course, must deal with anatomical (and, increasingly, palaeogenetic) evidence when delineating Homo sapiens. But for palaeolithic archaeology the issue at stake has been the evolution of modern human behaviour. In the course of the 1980s the new dispensation – the collapse of phyletic gradualism, the reconfiguration of Homo sapiens as a clade founded in a recent African singularity, the cladistic requirement for species to be essential kinds distinguished by unique autapomorphies, a persistent Eurocentrism that regarded the European Upper Palaeolithic as a universal template for modernity, and a confident phalanx of American practitioners trained in or exposed to cultural anthropology – led palaeolithic archaeology towards the trait list, and towards its supposedly most robust element, symbolism, as the defining property of the modern human being. In terms expressed by Kuhn and Stiner (Reference Kuhn and Stiner2006, 968), symbolism has come to be seen as ‘a uniquely derived and culturally universal’ feature of Homo sapiens behaviour. The symbolism = modernity orthodoxy tells us that we can define ourselves in terms of our unique mind and its special products, symbols. Homo sapiens has become ‘man the symboller’.