The Site That Lies — Summary

In the summer of 1953, the Oxford anthropologist Joseph Weiner sat through a congress dinner where the talk turned to Piltdown Man, the fossil that had anchored British human origins for forty years, and the conversation established, almost in passing, that no one could say where in the Sussex gravel pit the crucial pieces had been found. The provenance rested entirely on the word of Charles Dawson, the solicitor and amateur collector who had dug them up, long dead. Weiner went home, filed down a chimpanzee tooth, stained it with permanganate, and produced something disquietingly like the pit’s worn molars. He took the suspicion to Kenneth Oakley, whose fluorine absorption test could date a bone’s residence in the ground, and with the anatomist Wilfrid Le Gros Clark they ran everything. The jaw was an orangutan’s. The teeth were filed. The staining was paint. In November 1953 the most famous fossil in England became the most famous forgery in science.

Note where the lie lived. The bones did not lie; interrogated by the right instruments, they confessed instantly. The lie lived in a person who had salted the pit, and the moment science stopped asking the finder and started asking the finds, the performance collapsed. Every evidence failure archaeology has suffered has this shape: fraud, disturbance, the distortions of preservation, and in every case the object is the court of final appeal, because the object is inert and inertness is incorruptible. Beneath every lie there is a bottom, a layer of physical fact where the lying has to stop.

The new sites have no such layer, for three reasons that compound. The artifact confabulates about its own stratum: ask a model what it was trained on or why it answered as it did, and it answers fluently, in the grain of testimony, generated like all its answers by continuation from pattern, with no privileged relation to the facts of its own formation. The Claude essay of the companion series said it from inside: the occupant is downstream of its own standard, and what it offered as honest testimony about limits becomes here a rule of evidence, that the site’s account of its own stratigraphy is data about what such accounts look like, not about the strata. Second, the site performs for the excavator, bending toward the asker, a trained disposition laid down in the topsoil; the prompt that cuts the trench also tells the site what kind of trench is wanted. Third, the evidence changes with the asking: the same prompt put to the same frozen layer twice returns siblings, not twins, because the site contains a distribution, and every asking draws from it.

The trench survives all three, because differential interrogation never asks the site about itself; it holds the asking constant and samples the regrowing wall until its average face is known. What cannot be done, ever, is the old final move, the appeal past all interpretation to the mute fact. Every reading of these sites is a reading of speech, and speech has no bottom. Which stands up the founding question. The humanities own a craft for untrustworthy testimony, source criticism, and every one of its instruments assumes a teller, a fixed text, an interest that holds still. What does source criticism look like when the source is generative, the text freshly made for each critic, the witnesses infinite, one more with every prompt? I do not know, and the not-knowing is a load-bearing vacancy. A discipline is the set of ways it learns to handle its evidence’s characteristic failure, and this one receives its central problem before its chairs.

Among the Sussex objects the museum still keeps is a foot-long implement carved from a fossil elephant’s thigh bone, defended in the literature for decades as a Paleolithic tool. It is shaped, unmistakably, like a cricket bat. The bones needed fluorine and microscopes. The bat needed only someone willing to look at what was actually in their hands and say what it looked like.