Sunday, 4 December 2011

Putting Middle Palaeolithic Archaeologists on Notice (except Harold Dibble, that is)

Previously on the Subversive Archaeologist...

I was critical of the recent claims from Oman having to do with a until-now unknown 'extension' into the Arabian Peninsula of the Nubian Type 2 Middle Stone Age lithic technology. However much you may disagree with my ire-filled assessment of these claims, you may not be able to ignore what I have to say about the mind-set that in all likelihood contributed to the claims, and which underlies many assemblage-level discussions among Middle Palaeolithic archaeologists.
     The fundamental problem is typological. That is, how and why a given piece of modified stone is grouped with others similarly defined, and how those are distinguished from other groups identified using the same method. François Bordes viewed pieces of worked stone from a purely formal perspective. For this reason he paid less attention to the steps ancient flint-knappers took during manufacture of any given artifact. 
     Archaeologists of modern humans can easily distinguish a used flake from a retouched flake, a formed tool from a core, or a flake core from a blade core. Bordes was confronted by a welter of shapes and technological pathways to the European Middle Palaeolithic (Mousterian) artifacts he was faced with characterizing. [Have a look here for some examples of the variability.] It led him, ultimately, to name dozens of distinct Mousterian so-called types--a far greater number than he would ever have recognized in any assemblage associated with modern humans. As he would have done with assemblages from more recent times, his only choice was to characterize each according to the relative frequency of each type and category of type--e.g. flake, point, or blade; handaxe or Levallois core. 
     Furthermore, Bordes saw what he thought was structured variability between assemblages he characterized in his unique way--both within and between sites. In the end he saw five different assemblage types that he named Typical Mousterian (with the broadest representation of his 'types'), Charentian Mousterian (conflating the Ferrassie and Quina Mousterian assemblage types, which display steep retouch and fewer of the Bordesian types), Denticulate Mousterian (lots of flakes with retouch, leaving margins resembling teeth), Mousterian of Acheulian Tradition A (high numbers of handaxes) and Mousterian of Acheulian Tradition B (with handaxes, but fewer in number or smaller in size, or both, than in the MAT A).
     In other words, in a stratified MP site he sometimes recovered different examples of the five assemblage types in superposition. Bordes fastened on the notion that the variability was the result of culturally distinct groups of Neanderthals whose territorial boundaries had shifted through time. If you're familiar with the history of archaeology in the twentieth century you'll remember the long-running debate between Bordes and Lew and (mostly) Sally Binford over MP inter-assemblage variability. Sally Binford posited the notion that Bordes' five assemblage types might instead reflect the changing use of a site through time, the so-called 'functional' explanation. [Wow! I just found this article while scrounging for images on Google. If you're interested, this set of Sally Binford's reminiscences should make you rethink Lew Binford's career and his person. Talk about 'Revelations'! This is apocalyptic.] 
     Sadly, neither Bordes nor the Binfords examined the theoretical underpinnings of Bordes' typology. Had they done so, and had they arrived at the same conclusion that Harold Dibble did in the mid-1980s, they would have realized that the entire argument was rather like contemplating the number of angels that can dance on the head of a pin. Dibble, a flint-knapper of no small talent himself, perceived a chink in the facade of Bordes' artifact classification. He reasoned that the artifacts Bordes saw as finished end-products were, as likely as not, just pieces of stone that had gone through a predictable reduction sequence that depended on the starting shape of the flake. Thus, if one recovered an artifact that Bordes would have seen as a tool that needed two convergent, convex, retouched lateral margins, Dibble could show where, at an earlier stage in the artifact's life history, it would have been assigned to a different one of Bordes' types--in other words, if it had been discarded at an earlier stage of reduction instead of retouched further. This meant that a number of Bordes 'types' were essentially in the mind of the typologist. In effect, Bordes' scheme is still useful in a purely technological sense, but it falls down in its ability to reflect the mind of the Mousterian flint-knappers with respect to the intended end-product in the life of a flake.
     In reality, Dibble was recognizing what was later referred to as the Finished Artifact Fallacy. However, as far as I know, Dibble's insight has never been taken to its logical conclusion--that of dismantling the notion of Mousterian artifact types--nor has it led to a general examination of some other typological constructs associated with Middle Palaeolithic archaeology, such as the Levallois technique and the handaxe. Dibble and his colleague Nicolas Rolland did suggest that, as a result of Dibbles' deconstruction of Bordes' artifact types the various Mousterian assemblage types were very likely to have been nothing but a house of cards. 
     A very few (Davidson, Noble and I, in a small way) have attempted to flag the handaxe and the Levallois as classifications that are as susceptible to a take-down similar to Dibbles', but it's really hard to be heard when you're talking into a prodigious head-wind of mostly hot air.
     Which brings me to the issue of the Nubian Complex, 'a regionally distinct Middle Stone Age (MSA) technocomplex first reported from the northern Sudan in the late 1960 s [sic],' and its recent discovery in Oman at the southeastern tip of the Arabian Peninsula. Aside from other problematic matters,* Rose and Marks's discussion of the Nubian of Oman is based on the premise that one can compare MP assemblages based on the proportional representation of artifact types so classified using questionable assumptions of what a type is (or represents). Rose et al. report on excavations of lag deposits from various localities in Oman and, using a classification system that roughly mirrors that of Bordes' original MP typology, purport to describe a movement out of Africa on the part of the inhabitants of what's now the Sudan (previously Nubia). 
     This post has gotten far longer than I intended, and so I'll leave you now with a promise to bring in visual, numerical and analytical support for my critique of the Oman results in a future post. In the meantime, have a think about what I've said, and if I've misspoken, or if you simply disagree, I'd very much like to hear from you, and hear your reasons.

* Such as the characterization of an assemblage from what is clearly a time-averaged accumulation of heavier particles in a depositional context dominated by aeolian deflation--in other words lag deposits, which are notoriously difficult to characterize simply because there is no way of knowing the temporal relationships between and among its different constituents.


  1. One of the great sadnesses of this story, of course, is that despite the fundamental importance of the argument by Rolland and Dibble for looking at the intensity of retouch as the most important factor in Mousterian artefact variation, Dibble succumbed to the needs of working in France and wrote a volume with Debenath which, in essence, restored the priority of typological classification, thereby, IMHO undermining his own fundamental insights.

  2. You're dead right. Sad, really. But, hey, he's now show us that Roc de Marsal was a crock, and he's now working at La Ferrassie with the same intent. Like a fox!

  3. I think we need to be careful not to throw out the baby with the bathwater here. I don't doubt that some of the types in question are problematic, but the finished artifact fallacy / reduction thesis is largely an argument of plausibility. There just aren't that many studies that show quantitatively that retouch has the claimed effect of blurring types. Given how popular the finished artifact fallacy / reduction thesis has become in recent years, this is pretty remarkable, I think. Briggs Buchanan and I discuss this issue in the following paper:

    Buchanan, Briggs, and Mark Collard (2010) A geometric morphometrics-based assessment of blade shape differences among Paleoindian projectile point types from western North America. Journal of Archaeological Science, 37(2):350-359

  4. But not every tub of bathwater has a baby in it.

    Thanks for the link to Sally Binford article. Would never have connected her to Susie Bright, but it makes a lot of sense after reading that. Great stuff.

  5. @Mark,
    Good to hear from you. And by the way, thanks, Pal, for throwing cold water on a tub of lukewarm dirty bathwater, with baby or without. ;-)
    As for the popularity of the FAF, I'd dearly love to see some quantitative support for your statement. After all, in the past month, with Qesem cave and Oman prominent, I'd say you might re-examine your data! :) However, if those fellas are in the minority, allow me to be astonished. I have sensed no such surge of acceptance. But then I've been on the fringes for some time. [With no small thanks to articles like 'The Cultural Capacities of the Neanderthals: A Review and Re-Evaluation' emanating from departments like yours! But I'm not bitter.]
    And merçi for the reference. I can haz pdf, please???
    Lastly (if there can ever be a 'lastly' in a tiny part of a decades-long conversation), I think you'll find that Iain Davidson has examined, quantitatively, the issue of blurring with respect to the handaxe, and I have every expectation that once your hoped-for quantitative examinations of Bordes' types comes along, the whole edifice will crumble. I'll tickle Iain to see if he'd be willing to cough up a reprint for you.

    Good about the presence or absence of babies. Mind you, the absence of evidence isn't evidence of absence, said the fringe archaeologist with a predilection for arguments of plausibility in the absence of access to the material that would make those plausible arguments into well supported publications (but I ain't bitter).
    Sally was long-held suspicions validated. Quite amazing. Top o' the season to ya!

  6. The interesting thing about the Sally Binford piece is the affection with which she writes about the early part of the relationship with Lew. I am not sure about doing anything quantitative in the context of an argument with Mark, though. What are you referring to?

  7. @Iain,
    Sorry. I thought the work on Jelinek's Tabun material was of the type Mark is calling for.

  8. Just a comment on Iain’s first post regarding the typology book I wrote with André Denénath. When André first suggested doing that, my reaction was quite negative, for just the reasons that Iain stated – I had just shown (and yeah, there are lots of quantitative data that have been published in support of it) that many of the various scraper types were basically the result of how much they had been resharpened. But there were two reasons why I went ahead with the book. First, in defense of Bordes’ typology, it seemed to me that having a descriptive system such as his was worthwhile, in spite of how it is interpreted. I wrote once about the analogy with Linnaeus’ classification – he did that basically to show the immutability of species and therefore to provide support for the existence of God. In effect, Darwin used that same classification system to demonstrate evolution. All empirical research relies on units of analysis (like Linnaeus’ classification or Bordes’ typology), but advances in knowledge come through our interpretations of those units. So, I used Bordes’ system to present a different interpretation of the various scraper types, but the actual demonstration of my interpretation utilized those same types. Without a standard descriptive system, all we are left with are interpretations (which is the current problem with the French chaîne opératoire studies).
    The second reason for doing the book was that I felt I could present Bordes’ system with all of the warts showing. If you read it carefully, there are numerous places where I underline some of the silly aspects of it – like the famous difference between convergent scrapers and Mousterian points – it’s a point if you can kill a bear with it. Or the difference between a truncation, backed knife, and abrupt scraper – they all have identical retouch; it’s just a question of where the retouch is located (on a truncation it’s on the end) or what the other side of the tool looks like (if sharp, then it’s a backed knife – if it’s dull, it’s an abrupt scraper). Those, to me at least, are silly distinctions in the absence of any data that demonstrates the “reality” of them.
    At any rate, the book was not written to placate the French. I do have lots of intellectual disagreements with my French colleagues (including the interpretation of the Roc de Marsal “burial”), but at the same time I have a tremendous respect for many of them (and not just the ones who agree with me from time to time).

  9. @Harold,
    Thanks for dropping by the Subversive Archaeologist. And merci for the background on your typology book. I don't believe we've ever met, but lately I've been wanting to tell you how much I've always admired your work on Bordes' typology. And, as you may already know, it was your team's vindication of my 22-year-old prediction as to the depositional circumstances of the Roc de Marsal Neanderthal child that spurred me into starting this blog.
    As you can imagine, thanks to your work at Roc de Marsal, I'm finally allowing myself some satisfaction that I had it right in 1989 and 1999. I’m also a little more hopeful that at some point the discipline might see fit to acknowledge my contribution. I've never been able to figure out why your deconstruction of Bordesian typology and my work on the edifice of Middle Palaeolithic burial elicited such different responses from the academy. [But I'm not bitter!]
    Again, thanks for dropping by. And, don't mind Iain's comment about placating les français: I too, thought that by advocating a continued role for Bordes' reified categories, you were enabling the less intellectually agile among your readers to think that there was nothing wrong with continuing to construct palaeo-reality using those artifacts as evidence of cognition on a par with modern humans. Thus, you can probably understand why Iain might have thought the book undermined your revolutionary insights.
    Now I'll go back to envying your team's new excavations at La Ferrassie. I fully expect that you'll find convincing proxy evidence that the 'ceremonial' mounds were artifacts of cryoturbation, and that the other pits and so forth in which some of the remains were found were either solution features or evidence of fluviatile erosion, and not purposefully excavated graves. Good luck!

  10. Hi Rob,

    Regarding the popularity of the Finished Artifact Fallacy / Reduction thesis, my experience has been that more or less every time I've submitted a paper on Paleoindian point variation, at least one reviewer has brought up the Finished Artifact Fallacy / Reduction thesis. Of course, it may have been the same reviewer each time, but I have also heard the thesis discussed at conferences a number of times. In addition, it's being propounded by some of the most influential archaeologists around (Harold Dibble, Mike Shott, Peter Hiscock, Shannon McPherron, and Iain Davidson) So, I don't think it's a minority view any more. Indeed, sometimes it seems close to being the new orthodoxy...


  11. Hey, Mark.
    I'll gladly take your word for it! However, somebody has to break the news to the French! ;-)
    By the way, something Harold said prompted me to upload at the paper I gave at the 2006 Cradle of Language Conference in Stellenbosch, called Words can speak louder than actions. I heard Phil VanPeer's crest falling as I was presenting it, moments after he'd described how the chaine operatoire enabled him to map the decision-making process of his Middle Palaeolithic Egyptians. Awkward. You might find the paper interesting. There weren't any afficionados of the FAF in the room, as best as I could tell!
    Thanks again for dropping by and commenting.

  12. This is getting interesting. I go to sleep only to find I have upset Harold, for which I am very sorry, Mark thinks I have become orthodox, for which I am very sorry, and Rob does not think I was in the room to hear his paper, for which I am very sorry. I slept well.
    For what it is worth, the FAF began with Harold, for me, though I believe that George Frison had said something similar. And I always remember Harold's kindness when I was just starting out on this path. I am sorry for misattributing the motives for the Debenath book. My feeling is closer to that of Rob, though. It does enable those of us who have not read the text carefully to think that it was a return to orthodoxy, and so there was no need to do the backflips necessary to implement the implications of the original resharpening idea. I guess I would like to see an analysis of Mousterian sites with some index of degrees of resharpening and locate those spatially relative to toolstone source and to season of occupation. I have had my suspicions that some of the activities might be a product of being snowed in for the winter without access to new toolstone.
    Mark: it is nice but frightening to be put among the orthodox. I will have to try harder. Better write that culture paper! You are right, there are some other people around who have bought in to it, and there is a nice paper I have seen about the re-use of patinated handaxes as cores. So it is there. But I think you are dealing with another dilemma. My suspicion is that the idea of handaxes as tools (or indeed scrapers as the intended artefact) derives from a naive expectation that the archaeological record should be full of spear points, arrow heads and axes. Difficulty is that there probably were spear points, arrow heads and axes. So the question becomes "can you define the boundary between an intended artefact and one that has formal properties that were unintentional but nevertheless patterned?" I don't have an answer to that. It does not help that your reviewer(s) may not realise the subtlety of that question. Tell them to use me as a reviewer!
    As for the Chaine gang. I commented on the Van Peer and Bar Yosef paper in CA a couple of years back. It seemed to me that the approach raises some very interesting questions about how we write the narrative of the past (as does the resharpening story) and it throws us back on the question of what sort of story we want to tell, and whether we have the means to tell it with our current methods. I mean, we need to have something to say that we can believe in, or the whole story-telling role for the past will be take from us by the geneticists.
    Wow, I did sleep well.

  13. Hey Iain -- don't worry, you didn't upset me in the slightest, and you're not the first person to wonder why I did the typology book. In fact, when John McNabb reviewed it many years ago, he brought up the irony as well. I just figured that I'd set the record straight in my post here. BTW -- I also have fond memories of when you stayed at our house many years ago, and we had lots of good conversations with Phil Chase, Dorothy Cheney and Robert Seyfarth.
    Also -- as far as I'm concerned, it was George Frison who first demonstrated that artifact types can change through reuse/resharpening. Art Jelinek referred to it as the "Frison Effect". What I did was apply it to the MP, and along with Rolland, try to show that at least some of the classic MP variability can be explained by intensity of utilization of the available raw materials.
    Now, as to the status of the whole reduction school (which includes the FAF), my read is that most people are on board with the notion that reduction and resharpening are processes that affect lithic assemblages. However, to this day, most French won't acknowledge any of it. As I tried to explain in my big 1995 paper on scraper reduction (the last one I wrote on the topic), such a notion is very much counter to the French chaîne opératoire approach, which is based on the assumption that we can recognize desired end products, and then experimentally show how they were made. In other words, where I look at objects found in the archaeological record as unwanted discards, they look at them as tools discarded in their prime. And it's true that from time to time, some of the chaine-op people do actually believe that with such an approach they are getting into the minds of Neandertals.

  14. Hi Harold--I am so glad you have such positive memories of my visit. I better buy you a beer in Memphis and reminisce.
    As for the Chain gang, I will try to attach my comment here. I think there is something to be said for the initial approach, but there needs to be an analytical phase later which tries to draw conclusions. But, as I tried to say in the CA comment, the conclusions you draw will depend on the sort of story you want to tell. One of the things that made Bordes' story so compelling was that he was able to go from his typology of artefacts to a typology of assemblage types based on those types, to an arrangement of those assemblage types in time and space (well, I am not sure he did that very thoroughly). We should be demanding of any alternative scheme of identifying variation and clustering in stone artefact assemblages that it be put into that time and space matrix to see whether the variation (AND clustering) is anything more than random. I do not see that happening, but I do think that your work with Nic Rolland indicated the way ahead. By the way, had you met Nic before the Saint Louis meeting in 1993? I remember a lunch/drink with you, me, Nic, Peter Hiscock and Tom Wynn (at least). I know there were some first meetings there.
    Don't know how to upload the paper, so here is the citation:
    Davidson, I. 2009 Comment on Bar-Yosef and Van Peer. Current Anthropology 50(1):119-120.
    By the way, everyone, have you seen this paper just published by Mark Moore?
    Moore, M.W. 2011 The design space of stone flaking: implications for cognitive evolution. World Archaeology 43(4):702-715.


Thanks for visiting!

Note: only a member of this blog may post a comment.