Monday, 19 December 2011

Science Magazine and the AAAS Take Note: Sibudu Cave's Fantastical Insecticidal Palaeo-Mattresses

In my last post I laid out a blueprint for the subversive archaeologist--examine every argument's premises to see if they are well warranted or not. I really don't mean every claim. Practically speaking, the vast majority of archaeological inferences are (relatively) straightforward and require little (if any) scrutiny--'this is a rock; this is a flake,' 'that is ash; this is charcoal,' and so forth. These are (usually) inferential statements of such high probability (backed up by innumerable instances of the same kind), that it's reasonable to treat them as 'facts.'

Sibudu Cave is in shadow in this 2011 topographically exaggerated, artificially oblique-ified view looking northeast, remembering that in the southern hemisphere the sun is in the northern sky (Thanks, Google earth!)

In this post I want to provide a case study of an 'argument from want of evident alternatives,' about which I wrote in the previous post. I'll walk you through the problem I've identified in the recent report from Sibudu Cave, South Africa--where archaeologists claim to have discovered evidence of the oldest purposefully constructed vegetation mattresses in human history. 
[I promise not to mention that many animals collect plant material to make 'nests,' and that it's just the Sibudu archaeologists' speciocentric bias that compels them to treat a human 'nest' as something special. Oops! I guess I just broke my promise. Mea culpa.]
Archaeologist Lyn Wadley and others were excavating in a rockshelter with a Middle Palaeolithic (MSA) component when they came across 15 layers of decomposed and oxidized vegetation. While horizontally quite extensive these layers are thin--about 10 mm--and comprise a number of species that still grow in the area. These laminae of vegetation also include small lithic debitage and broken and burnt bone. One of the layers, about 1000 mm by 2000 mm in plan, comprises leaves of just one species. Wadley contends that, with evidence of a number of different plant species in the other layers, an accumulation of a single species seems unlikely--full stop. The remainder of their claims about the 'bedding' follow from this, and depend on the premise that such an accumulation is 'unlikely.' Insecticidal properties of the single species points to an intention to use it in preference to others to keep the work/sleep area free of pesky critters. The inhabitants regularly burned their old bedding to rid them of vermin. These subsidiary claims are totally undermined if the original premise is shown to be in error, and rendered virtually toothless if natural processes could have been responsible.
     I suppose it goes without saying that the Sibudu Cave excavators have not attempted to rule out natural processes in the production and preservation of these vegetation layers. Instead, they simply assert improbability and leave it at that. I'm not accusing them of disingenuousness. Yet, I find it amazing that they didn't take a moment to consider the possibility of other, expectable, natural causes. All I need to do is propose an alternative natural process and down goes the house of cards.
     Sibudu Cave [rockshelter, really] is a large, low, semi-circular escarpment formed in Ordovician sandstone about 40 km north of Durban, South Africa. It's the head scarp of a small valley that extends southwestward from the cliff face. As can be seen in the Google Earth overview above, the present-day natural vegetation is dense. By all accounts it was similarly dense in the archaeological past. 

Sibudu in sunshine. In this view you get an idea of the hollow formed by the escarpment and the surrounding terrain--in other words, a sediment trap (Click here for photo source).

In the site plan (below) you can see that the cave's aspect is due west, and that the talus slope formed by the weathering bedrock slopes sharply downward as you move from north to south along the 'rockshelter wall.' In such circumstances you can expect local sedimentary input from the weathering bedrock, as well as from the surrounding terrain. And because of the sloping rockshelter floor you can anticipate a relatively constant downslope movement of surface sediments due to a number of natural processes, among them short-distance fluvial and colluvial transport.

Sibudu Cave site plan (Villa et al. 2005)

Archaeological sites in such landforms develop in unpredictable ways. At times they can be net depositional, at others, they can be net erosional. They are very prone to flushing out by running water, which creates temporal breaks of unpredictable magnitude, and they are easily and predictably susceptible to what's called 'reverse stratigraphy,' where objects that find their way into low-lying areas can be buried by older sediments transported from upslope. 

In this view it's possible to see the steepness of the north-south slope (If you took this photo or know its ascription, please advise how I should best handle the credit).

Villa et al. (2005) publish a profile of the excavation's north-wall (shown below), which illustrates nearly horizontal layers in the west-to-east direction at the surface. By presenting the site stratification using this profile the authors are [one hopes] inadvertently downplaying the potential for re-entrainment and secondary deposition across the excavation area. 

Excavation profile illustrating the north wall of squares B5 and B6 (Villa et al. 2005). 
To illustrate the potential for downslope transport in the excavated area, I developed a generalized slope diagram from the site plan, seen below, which allows you to see the north-to-south slope perpendicular to the plane of the profile drawing. The slope is fairly steep, and it wouldn't take much surface water or treadage to mobilize small particles on the surface. A downpour would no doubt create a torrent on a slope of this magnitude, capable of transporting sediments up to and probably exceeding small pebbles.

Schematic north-south slope of site sediments based on contours in the site plan above, showing the location of the profile in the excavation profile reproduced below.

 You can also see in the excavation profile that most of the sub-surface layers dip gently toward the bedrock. Regardless of the reason for this west-to-east, downward trend, the resulting low-lying areas would tend to act as sediment traps. This means that, in addition to sediments transported from upslope to the north, those sediments from west of the cave wall would, when mobilized, tend to fill in the lower places, nearer the wall, with older sediments. All in all this is not an archaeological site that lends itself to inferences that depend on fine chronological control, or horizontal spatial integrity for that matter. By contrast, these depositional circumstances mean that it's no surprise that blankets of leaves collected and were preserved at times in this place, because of the more-rapid burial expectable in the lower lying areas near to the wall.
     The excavators claim that the strata of preserved vegetation are the result of deliberate preparation of a comfortable working/sleeping surface. Moreover, they claim that the bed builders regularly burned old vegetation mats, and that in one case they selectively collected the leaves of one species because it conferred an insecticidal property to the area of the leaf mat. In support of these claims the authors employ just a single premise. And it all boils down to this statement: 
'Many woody plants grew near Sibudu during the MSA; thus, single-taxon windborne leaf litter seems improbable.' 
As far as I can tell this is the only reason they concluded that purposeful human/hominid behaviour is the explanation for such a deposit. But, why must we accept that only wind-borne leaves could have contributed to the deposits?  
     I'm mystified as to their reasoning with regard to the source of the vegetation mats. Given the likelihood that, in the MSA, dense vegetation grew everywhere in front of and above the cave (indeed, all but in the drier areas underneath the overhang, much as in the present day photo, above), let's imagine instead a more likely scenario--leaf-fall. A plant grows, loses leaves, and those leaves come to rest beneath or very near to the stalk or trunk. It's the origin of the term leaf litter. Makes sense, doesn't it? We don't need to construct an elaborate scenario whereby the palaeo-humans went about collecting leaves for bedding--they merely had to use what nature would in all likelihood have provided for them. Or, am I missing something?
     And what about the single species mat? No real mystery there. The species in question is Cryptocarya woodii. This is no bush, nor is it a shrub. It's a hulking great tree! Moreover, it's not a mostly leafless, splindly thing like the one in the photo above. It's a tree that can grow to 20 m high, and from the look of the growth habit visible in the photo below, the tree is every bit as wide as it is high, and has an abundance of leaves. Hmmm.

Cryptocarya woodii grows from 5 to 10 m high, but it can grow to 20 m in favourable conditions (Photo credit).

Given the species in question, how unlikely is it [really] to imagine a 1 m by 2 m 'carpet' of leaves accumulating beneath a single individual of the species that would have dominated a 20-m diameter area? Not very, unless I miss my guess. And what about the mat remaining intact while natural processes buried it? Well, in a closed woodland, at the end of a valley, you might not expect too much in the way of air movement that could blow the leaves around once they'd fallen to the ground. Clearly the right circumstances prevailed to have aided preservation, unless we imagine purposeful burial of disused mattresses! What about the follow-on assertion that the vegetation mat was insecticidal? You tell me! Almost all plants produce substances that are toxic to other organisms. Think about it. Deadly nightshade. Oleander. Wormwood. Hemlock. I could go on. Cryptocarya woodii  happens to produce a substance that mosquitos avoid. Whoop-de-do. Given the improbability that the palaeo-humans made this bed of leaves, how likely is it that they chose the Cryptocarya woodii blanket over other accumulations at Sibudu so there wouldn't be clouds of mosquitos buzzing around? I'll leave that up to you to decide.
     I've hope I've said enough about Sibudu Cave to persuade even an antagonistic reader that there are major problems with this article. So, I'm done with Sibudu. 
However, I wouldn't consider my job complete under the circumstances without a word or two to the scholarly society that published Wadley et al.'s 'Middle Stone Age Bedding Construction and Settlement Patterns at Sibudu, South Africa.' Responsibility in this case lies with no-less-august a community than the American Association for the Advancement of Science, publisher of Science, one of the most read journals in the universe. How in the world did Science find such credulous editors and reviewers? You don't have to be an archaeo-botanist or Sir Isaac Newton to know that leaves fall and come to rest underneath trees! A three-year-old knows that! 
     And it's one thing for the media to hang on every fantastic narrative that gets published, but it's another thing entirely for a peerless, peer-reviewed, journal like Science to accept for publication a set of claims based on such an implausible premise and nothing else. Furthermore, even if it were a plausible premise, which it's not, since when does Science publish research that's based on plausibility arguments? Does Science have some sort of social-science quota that they need to satisfy to placate their readers? Are they so hard-up for good original contributions from archaeology that they would lower themselves to publish it? I doubt it. Hey, maybe they have a hidden agenda aimed at portraying the social sciences in a bad light by publishing poorly reasoned submissions! 
     More than likely the scientists who refereed the paper were too enthralled by the cool micromorphological thin sections, and the scanning electron micrographs, and the other science-y stuff included in the Science article to be critical of the argument. Sad. I'm done.

Next up: Qesem Cave and the extirpation of Homo erectus in the Levant: an article entitled 'Man the Fat Hunter.' Another clunker from PLoS ONE
[I think the authors of 'Man the Fat Hunter' should have used a hyphen in the title, unless they wanted to imply that Homo erectus, as a species, was fat! Something tells me that's not the case.] 
[If you find me remiss in not quibbling with the gender-exclusive language in the title, please understand that those of us old enough to remember Sputnik know that the title is an allusion to a pivotal conference in palaeoanthropology--'Man the Hunter'--which took place in the 1960s, and set the terms for the archaeology of human evolution for a generation thereafter.] This is gonna be so. Much. Fun.


  1. So, what have you got against caves, man?

  2. I enjoyed the post, but must comment that I think you have missed an important part of the argument. Namely, the younger part of the bedding in the sequence ~58 ka consists of sedges (Cladium mariscus according to the paper) characteristic of wetland environments. They would have to have moved up from the river valley to the abri. Hence, your alternative scenario does not hold for these deposits.

  3. @Gerrit Dusseldorp
    Thanks for your perspicacious comment, Gerritt. I was hoping that I'd adequately prepared the reader to infer the explanation for the recovery of water-loving plants within Sibudu. It appears that I've failed in that intention. Now I'm forced to spell it out! The answer is straightforward, but I won't produce it here. Instead I'll be posting an addendum in the next 24 hours that will provide an expectable, natural process to explain the presence of sedges and reeds in the shelter itself.

  4. Very informative post! There is a lot of information here that can help any business get started with a successful social networking campaign!
    memory foam vs latex

  5. Very informative post! There is a lot of information here that can help any business get started with a successful social networking campaign!
    memory foam vs latex

  6. Easily, the article is actually the best topic on this registry related issue. I fit in with your conclusions and will eagerly look forward to your next updates. Just saying thanks will not just be sufficient, for the fantasti c lucidity in your writing. I will instantly grab your rss feed to stay informed of any updates.
    online mattress retailers


Thanks for visiting!

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