Words Can Speak Louder Than Actions more

Good afternoon, everyone! 1 I’d like to begin by saying that we are all at pains to make our work rigorously empirical, whether that means refining our excavation techniques or taking care to examine as full a range as possible of explanations for what we observe. For, it is in the interpretation of our observations that we succeed or fail in our efforts to make accurate knowledge of our past. 2 Tracing language origins is no exception. It is a little like tracking an animal by following its spoor; it involves the uniformitarian assumption, and employs inductive inference and wellwarranted analogies to processes that can be observed in the present. For example, we recognize the distinctive imprint of a paw or a hoof as having been made by a unique species of animal, one that we can observe making its tracks in the present. However, tracing language origins is more like interpreting the kangaroo crossing sign, in that it requires inferences that go beyond the data. 3 For example, interpreting this sign involves first inferring that the area covered in black pigment is a depiction, indeed it s a stylized representation of a kangaroo in motion. We couple it with the knowledge that a piece of yellow-painted metal or wood in the shape of a square tipped on its corner is the conventional signal to exercise caution while driving an automobile on a public road. Such an interpretation is several removes from the physical nature of either the paint or the material on which it has been applied. Thus, our interpretation involves meaning beyond what is signified, and agreedupon conventions that are linguistic in origin. 4 A scrap of wood and some paint has become a meaningful object--but only through our commonly shared system of meaning and interpretation. Tracing language origins using the archaeological record employs much the same reasoning. However, our search among the present-day objects which are the remnants of our past is guided by no knowledge like that of the tracker. Trackers know that a certain animal made a certain track. 5 For example, we have no knowledge of what language looked like when it arose, or the kinds of archaeological traces the earliest language-capable hominins would have left behind. And that is the motivation behind my paper this afternoon--my guiding purpose is to refine our ability to recognize the earliest evidence of language. We all know that the earliest direct evidence of language is writing. But that only takes us back a few thousand years. 6 But language was around for a long time before writing came along. We know this because there is clear and abundant evidence of language before writing. Depiction predates writing by as much as 30 or 40 ky. But how do we know that depiction and other artifacts are evidence of language ability? For this, we need to be explicit in the way we link our observations to language. 7 To link depiction to language, we need to understand the process of interpretation that occurs when an enculturated human being confronts such markings. Without language, and in the absence of enculturation, these lines at Grotte Chauvet are just marks made by charcoal and other pigments. If we were not capable of language we could not understand their purpose, even though, to you and me their message is clear. Language alone enables us to interpret these markings,by convention, as depictions of bison and lions. 8 The same is true of beads and harpoons, clay sculptures and disposal of the dead. 9 And we know that by 40 to 50 kya there were humans in Australia, a feat which almost certainly required language for its achievement. But what about earlier hominine achievements? 10 As we delve deeper in time, and encounter fewer and fewer traces that we recognize, inferring language becomes manifestly more difficult, as we ll see, because it involves an even more complex series of arguments linking the archaeological observations to the inference of language. Is there, in fact, evidence of language prior to the advent of depiction? Many sitting here today would answer this question in the affirmative, and point to several lines of evidence. Well, then, what sorts of traces came before depiction, and how do they relate to language? 11 In the period immediately preceding the earliest evidence of depiction, we see lots of retouched flakes. We see the so-called handaxes, and a process of stone working known as the Levallois technique. There are also, apparently, some sharpened sticks. 12 Also in the Middle Stone Age / Middle Paleolithic, we see anomalous objects claimed to be art, such as this water-rounded piece of mammoth tusk from Tata, a pierced cervid tooth and that of a wolf from castelperronien deposits in France, the so-called Neanderthal mask from Roche Cotard, and the now thoroughly discredited flute from Divje Babe. 13 There have also been claims for the construction of elaborate shelters, such as this one pictured, at Terra Amata. Professor Villa knows why we should be suspicious of claims for this shelter. 14 And what about purposeful burial of the dead. As many of you know, I ve questioned the basis for these claims. 15 As for sharpened sticks--I m not certain what they might imply about language (or about the Paleolithic javelin throw), but I think at best it s equivocal. 16 Finally, we have the spatial co-occurrence of artifacts and features, including hearths. I really don t know what to make of the tens of thousands of hearths at Kebara and Hayonim. In what remains of my time today, I want to focus on how the stone artifacts--particularly those made using the Levallois technique--and intrasite spatial analysis are used as evidence of the linguistic ability of the hominins who created them. 17 Leroi-Gourhan launched the idea of the chaine operatoire, or operational sequence, as a way of systematizing the conceptual framework, the decision-making, that underpins certain processes evident in the hominin past. Indeed, he contended that the steps in an operational sequence bely a syntax of behavior, behavior that was intimately linked to structures in the brain that relate to language. And so, where Leroi-Gourhan saw increasingly complex operational sequences, he inferred increasingly complex linguistic ability. Notice here that the Human Parameters include tradition and symbolic constraints. These kind of terms belie the presumption of humanness in the chaine operatoire concept. 18 Ofer Bar-Yosef puts it this way: The operational sequence demonstrates "the different stages of tool production from the acquisition of raw material to the final abandonment of the desired and/or used objects. By reconstructing the operational sequence we reveal the choices made by ... humans." (Bar-Yosef et al. 1992, 511). 19 And Bleed puts it this way: “...the most distinctive feature of the chaîne opératoire approach is its emphasis on ideational aspects of material systems. It has helped archaeologists recognize that the patterned activities that can be reconstructed to link prehistoric materials into technological sequences must have been informed by cognitive structures.” As I will argue in a moment, Leroi-Gourhan s faith in the relationship between operational sequences and language was based in large part on the assumption that a given artifact was a consciously desired end product. 20 But, how do we know when conscious decisions are being made, i.e. when an artifact is a desired end product, and when we re simply seeing the results of highly determined sequences of behavior in an otherwise nonlinguistic animal? I ask this question because there is plenty of behavioral complexity in the non-human world. The beaver, Castor canadensis, transforms a wilderness stream into its own private domain, first by damming the stream with mud and rocks and sticks and logs to create a lake, and then building a substantial dwelling there, complete with a living space, underwater entrances and ventilation. 21 Yet, there is no language here; it s only the natural behavior of complex higher mammals. If I were to view the dam, lake and lodge as desired end products, the operational sequence here would be long, indeed, and incredibly complex. And on Leroi-Gourhan s account, these constructions would surely have required language, although he would be very wrong indeed. So, why not think the beaver conceives of the final form of its dam, lake and lodge? I suggest it s because we can t ask them about it and expect an answer--they are not human. 22 Admittedly, such perishable materials as twigs, sticks, saplings and trees, would not normally form part of the fossil record. Nor would any dams or structures if they had been created by Neandertals out of wood. That is the frequently heard rejoinder to the question, If the Neanderthals are so much like us, why don t we see carvings and decorations like those in the later period? They worked in perishable media. Well, I ll give you an example of complex animal behavior that does preserve nicely in certain circumstances, and with which you ll be more familiar. 23 A complex sequence of events, representing decisions of a sort, characterizes the way that hyenas and wolves use a carcass. 24 Hyenas take apart an animal carcass in a sequence that follows the ease with which meaty and bony portions can be removed. This varies somewhat depending on the size and the species of the carcass, but for each size and species the sequence is fairly regular. Moreover, the order in which the carcass parts are consumed is the result of choices made by individual hyenas at each point in the process. But none of these animals has its eye on the final product, even if the final product is the same each time. 25 1.  Enter through the anus and clean out the innards 2. Including the tail 3. Then the pelvis 4. The rib meat and most of the ribs are removed 5. Then the forelimb 6. Followed by the hindlimb 7. Then the remaining ribs 8. And the accessible portions of axial skeleton 9. And finally off with its head! 26 Unlike wolves who will normally leave the backbone and ribs at the death site because they can neither transport the heavy remnant nor reduce it further, hyenas are adapted to consuming bone, and have enormously powerful jaws and extremely large and strong teeth that enable hyenas to consume the entire carcass of certain size classes of animals. Nevertheless, the sequence of carcass part consumption by wolves and hyenas is highly correlated, as is the order in which all extant and extinct hyena species reduced a carcass. 27 Here is a case where an orderly sequence occurs due in part to choice and in part to the varying architecture of animal carcasses. One could easily compare hyena and wolf carcass consumption sequences to the operational sequence of the Levallois technique. The only difference that I can see is in the presumption that the end product is desired from the outset of the sequence. Imagine if we assumed hyena behavior was purposeful? 28 Leroi-Gouhan looks at the Oldowan and sees choppers as the desired end product. Most of us now see this as just a source of sharp flakes. Nevertheless, these concepts drew L-G to believe that even these first stone artifacts were the products of the relationship between the language areas of the brain and the hands. Thus, the Oldowan was the dawning of language for L-G. 29 L-G also looked at Acheulean bifaces and saw some of them as the desired end product. Most paleoanthropologists and paleolithic archaeologists would still agree, and would identify four or five variants, the cordiform, the lanceloate, triangular (not shown), the cleaver and the pick. No doubt typologists leave out discoid bifaces because they have no apparent sharp end . 30 Here s the Acheulean reality at Olorgesaile. You might be able to see a couple of the formally classifed types here, a lanceolate and a pick, perhaps a cleaver or two. But other than that, we are presented with quite a variable assemblage. I only have to look at this collection of bifacial cores from Olorgesaile to see that there is no such thing as a type in the Acheulean of East Africa. 31 I think it s safe to say that, in defining types such as the handax, we are simply anthropomorphizing certain shapes that fall along a continuum that goes from disc to cordiform to rectilinear to triangular to lanceolate. Iain Davidson calls this the finished artifact fallacy Viewed as by-products and not end-products, L-G s argument is seriously weakened in the case of the Acheulean. 32 But it was truly the Levallois technique that convinced L-G, by virture of the complex chaine operatoire, that these hominins had achieved modern linguistic abilities. It s widely assumed that the Levallois flint knapper wishes to produce flakes of a predetermined shape by reducing a nodule in a way that will allow the removal, at the end of a long sequence, of flakes of certain shape. On the surface of it, this seems non-sensical, given the sheer number of flakes removed in advance of the final flake. 33 So, what are the traces of the Levallois technique? The debitage is not shown. However, the debitage can be used to infer that this technique was employed. Bottom left, a core with a flake removed. The classic Levallois tortoise-shell core In the upper left, a flake, the final flake removal, and called a Levallois flake In the upper middle, a Levallois Point. Upper right is a series of proximal views of Levallois flakes with evidence of platform preparation, the so-called Chapeau de Gendarme. 34 As the standard story goes, a nodule of workable rock is reduced in such a way that a flake of a predetermined shape can be removed. After dozens of flake removals, all seen as preparation for the final flake, the core is then, apparently, discarded, as seen here on the surface of a site in the Negev Desert. The Levallois technique would indeed be evidence of a singular talent and considerable forethought, were it not for some annoying realities, beginning with the true variability of these artifacts. 35 Francois Bordes, who literally wrote the book on Middle Paleolithic typology, decided that there were at least nine core types, each with its special technique of preparing the nodule for removal of a certain shape of flake. 1. Classis levallois core 2. Classic elongated levallois core 3. Classic elongated levallois core for flake/blade production 4. Classic levallois core with pointed flake 5. Long levallois core with parallel preparation 6. Short levallois core with parallel preparation 7. Victoria West core (for some reason we re supposed to think that the flintknapper hit this one from the side) 8. Levallois core producing triangular points prepared distally 9. Levallois core producing triangular points prepared proximally 36 Notice that the first 4, termed Classic levallois cores, differ only in the length to breadth ratio of the flakes. Types 5 and 6 are roughly rectangular, and about twice as long as they are wide, and the flakes are removed parallel to the long axis of the core. Type 7 is simply a wide Classic. Types 8 and 9 differ only because the penultimate flakes either diverge from or converge on the eventual platform. The common denominator is the facetted platform preparation that frequently occurs, the chapeau de gend arme. 37 But, when one looks on the ground, as it were, one sees that the real artifacts are uncooperative, and do not adhere to Bordes idealized types. Here s a quiz. Is the flake outlined in red a Classic? A Classic Elongated? A Blade type? A slightly-more-elongated Classic? A rather-broad blade? Well, I think you get the picture. 38 Here I have reproduced 7 of the original Levallois core types outlined in green. I have used the original plans to create intermediate examples. These are outlined in red. B1 is simply a laterally compressed Type B. C1 is a laterally inflated Type C. D1 is a laterally compressed Type D. D2 is a laterally inflated Type D. And so on. Including near morphological variants in the display illustrates continuous variability. 39 One can also see that the so-called parallel preparation in Types F and G could easily be explained by including variability beyond the 9 Bordesian types. This suggests that the shape of the core may be not the result of choice, but rather of subliminal circumstances. But what about the flake itself? If it isn t desired from the outset, why do we see that last flake removal, and then a discarded core? I think that this question is begged by the assumption that the final flake was the intended product. 40 And the reality is nothing like the ideal types erected by Bordes, as these examples from Doura Cave demonstrate. Seriously! What about these artifacts persuades us that the final flake removals were aimed at from the beginning? It seems to me that we are dealing with reified categories and not evidence of a deliberate sequence of conscious choices, nor of language. 41 Recapping to this point: 1. It is legitimate to suggest that the notion of an operational sequence depends very much on our ability to recognize a desired end-product. 2. Calling the Levallois core prepared may be unwarranted. 3. Thus, using the Levallois technique as evidence of language is at best equivocal. 42 And finally, Space. 43 Activity areas. Drop and Toss Zones. Sleeping areas. Mens areas. Hearths. Archaeologists are very used to thinking in terms of spatial patterning. So, it s no surprise that archaeologists would want to look for spatial patterning in the physical traces of hominins of any time. 44 However. There are a couple of hitches. The first is epistemological; the second is empirical. First, the way modern humans organize themselves in space is the only source for analogies with which to interpret spatial patterning in archaeology. This has proven to be a fruitful line of investigation in a large number of late archaeological contexts, even when tangible structures are absent, as is the case with a great many foraging societies. 45 In Early and Middle Paleolithic sites archaeologists are constrained to interpret spatial patterning as evidence of language, be it social divisions, such as the sexual division of labor, or special activities, such as butchering or sewing. Their reasoning seems to follow this path: 1) modern humans organize themselves in space, 2) modern humans leave behind patterned refuse, and 3) modern humans have language. Thus the presence of spatial patterning in Middle Paleolithic sites logically implies that those whose activities created the patterns must have possessed language. 46 There are good, empirical reasons to be skeptical of this kind of evidence. First, it begins from the implicit assumption that only language could have created the patterns observed. Yet the warrant for that premise has rarely been seriously examined. Second, most archaeologists are aware that natural processes are capable of producing spatial patterning in past traces. However, once the geological processes are ruled out, the rigor usually ends and the enthymematic reasoning takes over. 47 In concluding that spatial patterning is evidence for language, archaeologists have overlooked, or side-stepped, the potential for two or more processes to create the same archaeological signature. Indeed, what has been missing from spatial analyses of Middle Paleolithic and earlier sites is any acknowledgement that the implicit equation of language and spatial patterning might be unwarranted. Yet, many animals create patterns in the objects that end up in their dens. The cave bear, for example. 48 At Pod hradem Cave, in the Moravian Karst of the Czech Republic, generations of cave bears hibernated. Some died there during the long winter. I plotted the locations of over 2000 cave bear skeletal specimens recovered by Karel Valoch of the Moravske Muzeum in Brno. The results of a straightforward statistical technique clearly demonstrate that some non-hominine activities were creating spatial patterning in the cave s skeletal remains. 49 In the statistical analysis of the overall distribution of cave bear remains, we see improbably high quantities (white squares) always near the walls of the cave. In contrast, improbably low numbers (blue squares) of cave bear remains occur mainly away from the walls, in the open areas. The hatched squares denote numbers that could easily be explained by chance. From this distribution, we might want to conclude that the cave bears liked to hibernate near the walls, were it not for the finding that wolves and hyenas were scavenging the carcasses of recently dead bears. 50 I think it s safe to say that purely mechanical processes have contributed to the overall distribution of cave bear remains. I am speaking of the phenomenon that others have now documented, too, that obtrusive objects in spatially constrained contexts, such as caves, tend to come to rest near the walls, and are then less likely to be moved back into the open areas. 51 Shown here, the distribution of just the fully grown cave bear remains is slightly different from the total distribution This is more than likely due to their average size. That s because the more obtrusive an object is, the more likely it is to be moved, and thus end up nearest the wall. 52 Here is clearer evidence that the size of an object contributes to its location in the cave. The very longest specimens occur in improbably high numbers nearest the walls, while the improbably small items occur in the middle of the open area. 53 And finally, evidence of the hibernation behavior of pregnant females, a behaviorally meaningfully inference, drawn straight from the spatial patterning. Baby cave bears are about the size of an average rabbit, and such small specimens are not easily moved by trampling. The largest number of perinatal remains occurs near the wall and in the deepest part of the cave. In all likelihood this is where the sows hibernated, and where their offspring died. Thus, cave bears give lie to the idea that spatial patterning can be equated with language. 54 Many here today would say that language ability characterized Neanderthals and their contemporaries. Spatial analysis is a similar story. I think if one looks at the assumptions underpinning both the chaine operatoire and spatial analysis, the same problems arise. 55 The chaine operatoire assumes at the outset that the Acheulean handax or the Levallois flake were the desired end product. Likewise, spatial analysis implicitly equates modern human behavior with spatial patterning. I think it s safe to say that the jury is still out on whether or not the Oldowan chopper, the Acheulean handax, and the Levallois flake were in the minds of their makers. And that spatial patterning alone is insufficient as a means of inferring linguistic ability in our recent hominine relations. 56 And this leaves us where? I guess it all depends on where you stand with regard to the various lines of evidence. I m fairly certain you can guess my stance. I see little in the archaeological record of the Middle Paleolithic/MSA that speaks to me of language. No depiction. No burial. No sophisticated lithics. No spatial patterns. Just sharp rocks, retouched sharp rocks, some platform preparation, sharp sticks, and maybe, fire. Prometheus, anyone? 57