Journal of Archaeological Pseudo-science: Middle Pleistocene Mythopoeism from Kathu Pan 1

J. Wilkins, M. Chazan / Journal of Archaeological Science xxx (2012) 1-18

I don't know whether to scream or pull my hair out. So I guess I'll do both. 

Jayne Wilkins and Michael Chazan, Blade production ~500 thousand years ago at Kathu Pan 1, South Africa: support for a multiple origins hypothesis for early Middle Pleistocene blade technologies, Journal of Archaeological Science  39, 1883-1900, 2012.

The authors seem to think that they have indisputable evidence of a blade industry at give-or-take 500 kya (i.e. the product of Homo beforeneanderthalensis, no less). But when I went looking for their empirical basis I came up, well, virtually empty-handed. I won't go into the details of what they see as the implications of their findings, but I'm gonna reveal their findings for what they are--air and angels. In brief, my finding is that their findings are going to founder on the rocks of the Sea of Reality. Here's why.

     Wilkins and Chazan looked at 3,786 pieces of rock (out of the more-than 2 metric tonnes of lithics culled from the site!). Of the 3,786 bits, 955 were categorized as complete flakes and 972 were called blades. A 50:50 ratio of 'flakes' to 'blades' will be important later in this discussion. So, keep 1/2 and 1/2 in mind.

     The average length-to-width ratio for the blades is 2.5 (s.d. = 0.4). Therefore, approximately half of the blades are in the upper part of the range (i.e. 2.5:1 to 2.9:1) of length to width. The other 50% of the blades would just make it into the 'blade' category (i.e. between 2.0:1 and 2.49:1). Given that a) the minimum ratio to qualify as a blade is 2.0:1, b) the mean is 2.5:1 and c) the standard deviation is 0.4:1, it's impossible to say on the basis of the date presented whether the distribution of 'blade' length to width is uniform, platykurtic, normal, or leptokurtic. That's a crucial datum, as you'll see in a moment. Nevertheless, it is apparent that the distribution is somewhat skewed toward the upper end, given that the actual range of 'blade' length-to-width ratio goes from 2.0:1 to 4.5:1, strongly suggesting that it's more platykurtic than not. 
     You see, a mystery as to the shape of distribution for 'blade' length to width casts a deep, dark cloud over Wilkins and Chazan's whole argument, because it's quite possible that the length-to-width ratio for the entire assemblage (i.e. flakes + blades) is uniform or unimodal, with a mean closer to 2.0:1 than 2.5:1. That would call into question the wisdom of creating the flake/blade cutoff in the first place. Actually, it would make it seem silly.

Once more, an example of how well-meaning archaeologists cherry-pick (never disingenuously) their assemblages for specimens that put their claims in the best light. And, correct me if I'm wrong, but isn't N technically a Levallois point and not a blade? I really, really like M. That must have been one fine flint-knapper to have prepared a core to take off that one at the end of the sequence!

     What of the complete flakes? Alas, Wilkins and Chazan provide no metrics for those flakes that had a length-to-width ratio of less than 2:1. However, I'll take my usual place on a limb. I think it's safe to say that the vast majority of complete flakes in any assemblage are at least as long as they are wide (i.e. >1:1). That would mean that most of the non-blade flakes in the Kathu Pan 1 sample are between 1:1 and 2:1 in a comparison of length to width. Therefore, when compared with the entire collection of 'flakes' and 'blades' the authors examined for this study, nearly 1/2 are just on the shy side of being blades. 

    On that basis alone, I'm willing to bet the farm that, were it not for the arbitrary 2:1 cutoff between flakes and blades, the complete flakes and blades would more than likely form a continuous and uniform distribution of length to width from near 1:1 up to 2.9:1 (with the rest of the range scattered with outliers). If that be the case (and neither you nor I have any reason to think otherwise in the absence of sufficient descriptive metrics in Wilkins and Chazan's article) it would be very difficult for the authors to sustain their inference that the blades they find at 500 kya, and in such abundance, are anything other than artifacts of the arbitrary line drawn between a 'flake' and a 'blade.' 
     If the authors had so much as a shred of evidence that there was preferential blade manufacturing going on at Kathu Pan 1, I'd very much like to see it. In fact, this observer wonders at some of the data presentation decisions the authors have made, such as presenting a histogram of the widths of their 'blades' rather than their lengths, and leaving out the metrics for the non-blade flakes altogether. [It was unfair of me to say this, but I didn't want to revise history by expunging the record. So, I'll leave it in as a sort of persistent mortification of the flesh The lacunae are mysterious, to say the least.]
     If my criticism has any merit, all of the additional analyses contained in their paper (i.e. core analyses and comparisons with other MSA assemblages) would add up to nothing. Chalk up another big miss for the Journal of Archaeological Pseudo-Science and its referees. Is it possible that they didn't ask to see the overall distribution of length to width for the entire sample to make certain that this article wasn't all about how many angels can dance on the head of a pin? Way. To. Go!
     At the outset I promised I wasn't going to say anything about the implications of a paper like this. What the Hell! I can't forbear saying something. Because, once a story like this gets into a refereed journal there are many impressionable minds that'll just consume it without looking closely, and the myth will just continue to grow. There are also those very serious archaeologists out there who're predisposed to expecting stories like this one. Any chance they're gonna be critical? Hardly likely. The referees who gave this paper a thumbs-up ought, truly, to be ashamed of themselves. The Editors of the Journal of Archaeological Science oughta be ashamed of themselves, too. One of them is Robin Torrence, who should know better. Elsevier, the publisher, ought to be asking why this paper was accepted when, presumably, the vast majority of submissions are rejected for better reasons. And then there's me. I'm certainly not ashamed to call a three dressed up as a nine just that. It's so hard to keep an open mind [as my well-meaning colleagues often counsel] when there's so much seriously specious argument about, so much patently silly empirical observation spread around. Whatever happened to critical thinking? Whatever happened to common sense? Oh, yeah, well, it's never been all than common.