School of Human and Environmental Studies
University of New England, Armidale, NSW, AUSTRALIA
Iain.Davidson@une.edu.au

Approaches to the study of the origins of language have three major sources of evidence: 1) From modern humans, including linguistics; 2) From modern non-human primates, including their communicative abilities; 3) From archaeology. This paper is about the third of these – the archaeological evidence of what actually happened. I will review the evidence from stone tools that has provided insight into the emergence of language over the last ten years since the Wenner-Gren conference in Cascais, Portugal. My paper will begin by outlining the argument I advanced in the conference volume Tools, language and cognition in human evolution.

The principal point was that the apparent existence of different tool types is not only a product of deliberate intentions to produce the forms recognised by archaeologists. It is this understanding which gave rise to our definition of a phenomenon we called "the finished artefact fallacy". This phrase has been misunderstood by Gowlett and by Mithen. I will clarify the concept in my presentation. Without plausible argument about the intentions of tool-makers, we cannot approach understanding of what tools tell us about the language-based symbolic representations in the mind.

One of the significant contributions to the discussion of the intentions involved in stone tool making is Mellars’ introduction of the phrase "imposed form". If archaeologists can find a way to identify that hominins or humans imposed a form on an artefact, then it would be possible to comment on the process of conceptualisation among those hominins. For Mellars, imposed form was absent from Mousterian scrapers where the modification of the flake is related to its original form. But it was present in the Upper Palaeolithic of Europe where there were modifications of flakes which are visually distinctive, repetitive and standardised. Such modifications may or may not have been applied to flakes which were themselves of standardised form (often "blades").

At Cascais, I argued that the first appearance of this "imposed form" was with the distinctive geometric microliths of the Middle Stone Age in southern Africa. I would now reinforce this judgement because of the excellent documentation both of the processes of production of these tools and of bone tools in contemporary industries in different parts of Africa. I have been at pains to point out the constraints on artefact form imposed by the mechanics of stone knapping. No such constraints apply to the production of bone tools. In these African artefacts of 90 000 years ago, modification of form was not determined by the requirements of function or the contingencies of use. I believe an appropriate argument can be constructed that these products of the African MSA were made by people who used language. That they are accompanied by quantities of worked ochre is a comfort to some other arguments about language origins.

In Mellars’ terms, the MSA tools were visually distinctive, repetitive, and standardised. The crucial issue is still whether there were earlier artefacts which fulfilled those criteria, and, if so, whether they indicate language use. The most obvious case which still needs discussion is the handaxes called Acheulean which seem to dominate the production of stone tools over 1.5 million years and over most of the world occupied by hominins over that period. One of the novelties of the last ten years has been the increasing recognition that handaxe forms are found in east Asia. I will outline some of Wynn’s arguments about the cognitive significance of handaxes.

At a conference in 1987, Dibble demonstrated that there were two common problems with the interpretation of Acheulean handaxes. First, there have been claims that there were many different types of handaxes. Dibble showed that the divisions were more apparent than real, resulting from the partitioning of continuous variation. There have been other demonstrations in the last decade that archaeologists are particularly inclined to partition continuous variation in stone artefact assemblages, with misleading results. Second, there was some element of the apparent standardisation of form of handaxes which resulted from simple procedures by archaeologists. I will present some more data in relation to this argument, and reinforce the suggestion that a relatively small number of constraints could determine the apparent standardisation of Acheulean handaxes.

In light of this claim, I recently began to look at cores from the oldest layers at the site of Tabun, from the meticulous excavations by Art Jelinek. If I am right in the contention that the form of handaxes is an outcome of the application of a small number of constraints in knapping, then some of the features of handaxes should be present on the non-handaxe cores in the related assemblage. I will show how my analysis of material from Tabun confirms this. In addition, Tabun provides good evidence confirming the claim by Bradley and Sampson that handaxe forms might be a stage in reduction sequence.

Arguments about the uniformity of handaxes over such vast times and regions, bring into question the issue of standardisation which is such an important part of Mellars’s definition. I will discuss the arguments about standardisation by Chase, by Chazan and by Kuhn. The fundamental point here is that there are two major reasons why artefacts are standardised: either there is a technical, procedural or cognitive limit on the possible outcomes, or a single option of many available was chosen.

One of the significant changes over the last ten years has been the clarification of the chronology of Middle Pleistocene archaeological sites. In particular, it is now quite clear that there is no simple sequence from Oldowan to Acheulean to a late Mousterian and then the Upper Palaeolithic characterised by stone tools made on parallel-sided blades. I believe that the new chronology requires the abandonment of the notion that there is a continuous 1.5 million year long Acheulean tradition. Instead we should recognise that there are relatively small numbers of possible outcomes from stone-knapping. There is some confirmation of this alternative view from the discovery of bifacially flaked handaxe-like cores in Australia, a region colonised after the end of the Acheulean from a region where handaxes were said never to have been made.

There remains the issue of the Levallois technique, if that had not been abolished by my argument in 1990. The appearance of "Levallois cores" in the early stone industries of Olduvai Gorge would be a product of the limited number of outcomes from knapping. Van Peer’s work provided new evidence in support of my position. This work showed that, when conjoining flakes from single knapping events, non-Levallois flakes were very often missing (presumed used). This was confirmation of the interpretation of use-wear studies by Beyries that unretouched non-Levallois flakes were more often used than unretouched Levallois flakes. The argument that Levallois technique implies a long time-depth of intentionality is looking very shaky.

Nevertheless, the Levallois technique has come back into prominence because of argument by Foley and Lahr that it provides insight into the evolutionary history of hominins. I will outline the difficulty of accepting the position adopted by Foley and Lahr.

There is new evidence and argument which could be presented about the earliest industries which I do regard as showing "imposed form" – those from the southern African Middle Stone Age (MSA) – and about the contemporary industries in Europe and adjacent areas. Although some of these industries may represent the emergence of modern human behaviour, I will confine my remaining comments to the question of the importance of long, narrow flakes known as blades.

By commenting on the various industries in Africa, Europe and the east Mediterranean which appear to have long parallel-sided flakes (much) earlier than the Upper Palaeolithic, I will argue that the appearance of "blades" alone cannot be used as an indicator of technical or cognitive abilities of hominins.

The evidence I present draws attention to two issues: the problems arising from our language-based practice of naming categories; and the need to think more fundamentally about the factors contributing to variation and to lack of variation in stone tool assemblage.

The problem of naming is well illustrated by the various examples of the partitioning of continuous variation in identifying tool "types", but it applies equally to the naming of the assemblages of tools as "industries". Importantly, it is highly likely that issues of continuity or discontinuity in patterns of producing stone artefacts result partly from different approaches to giving names to categories in analysis. This problem is best illustrated by the continuing importance given to the issues surrounding the larger category "Upper Palaeolithic". In particular, there are issues about whether one industry or another belongs to the Middle or the Upper Palaeolithic, and then whether the Upper Palaeolithic necessarily indicates modern human behaviour. Arising from this is a widespread perception that somehow or other the Upper Palaeolithic indicates the emergence of language (despite the overwhelming evidence to the contrary). Finally, the issue of the reality of the names invented by archaeologists and palaeoanthropologists is starkly revealed by the problems of naming fossil hominins. One important issue here is that the naming also involves the partitioning of continuous variation – producing a conflict with an evolutionary approach to changing hominin morphological variation. In addition, the naming practices of archaeologists and palaeoanthropologists raise the question of whether a single stone industry could be made by more than one hominin species, as appears to be the case in the east Mediterranean where Mousterian industries were made by both Neandertals and anatomically modern humans.

Explanations of stone artefact variation need to take into account all of the various factors which create variation between stone artefacts, as well as the many different factors which cause there to be little variation. These factors may be mechanical, mental or cultural. Mechanical constraints include the nature of the raw material, the requirements of knapping, knapping technique, and the requirements and contingencies of use. Mental factors include the absence of conceptualisation as well as the ability to conceive of new forms and how to make them. Constraints due to the culture of the knappers involve the fundamental issue of the context of learning to knap, as well as the cultural significance attached to style and any cultural constraints on producing new forms. It seems highly unlikely that, if a single stone tool industry was made by two different species, we should regard the lack of variation in that industry as the product of culturally determined intentions.

The study of stone artefacts clearly has huge implications for understanding the origins of culture, particularly if we can tie down the appropriate criteria for recognising that the forms of artefacts were determined primarily by the self-conscious choices of the knappers. In this context, we will also need to understand not only how there can appear to be imposed form in non-language using industries (as in the Acheulean), but how there can appear not to be imposed form in industries made by people who certainly used language (as inAustralia). The story is undoubtedly more complicated than the text-books tell us, but we are beginning to get answers despite the difficulties of interpretation. To do this, we must understand the extent to which some of the story was determined by assumptions about the way hominin behaviour changed in the Pleistocene.

 Conference site: http://www.infres.enst.fr/confs/evolang/