Posted by: Johan Normark | June 17, 2009

Climate and the ”Maya collapse” pt 7: Is the Signor-Lipps effect relevant in archaeology?

The main proponent of the mega-drought hypothesis for the “Maya collapse”, Richardson Gill, uses the so-called Signor-Lipps effect to suggest that the collapse was drastic rather than slow. What is this effect? It is first of all named after two palaeontologists: Phil Signor and Jere Lipps who suggest that there are reasons why, for example, the extinction of dinosaurs appear to have been going on before the Chicxulub/Shiva impacts rather than being blown away by these astronomical objects and the climate changes they brought about.

The Signor-Lipps effect states that two types of sampling effects can cause taxa to appear to decline before they go extinct. This is because the sampled rocks control the apparent frequency of fossils. It may lead palaeontologists to the conclusion that some species disappeared before they really did (the last known fossil of any species is likely not the last individual of this species that actually existed). Neither is the last recorded date at a Maya site the very date people left the site, they left sometime later but before they could erect a new dated monument. Thus, the fossil/archaeological record is discontinuous in time and apparent extinctions/collapses will begin well before a mass extinction/collapse and increase in frequency until the actual event which will give an impression of gradual extinction/collapse. Signor, Lipps, and Gill argue that animals/societies die randomly, their remains are preserved randomly, and palaeontologists/archaeologists dig randomly. This result in an apparent gradual die-off/deterioration but the extinction/collapse may have been abrupt, such as meteor impacts or a series of severe droughts.

There is also a boundary effect caused by the sampling effect. The relative frequency of species/artefacts in a stratum can cause palaeontologists/archaeologists to place the chronological boundary lower than it should be. For archaeological contexts this means that the collapse occurred later than the last recorded dates on monuments in the Maya area. The apparent gradual cessation of monument building is therefore not necessarily evidence of a gradual collapse. Gill focuses on sites with longer sequences of stelae and calendar dates since some small sites only erected one or two dated stelae. Based on this, he defines four 50 year periods in which cities collapsed in various parts of the Maya lowlands: 761, 762-810, 811-860, and 861-910. These sites are also geographically grouped.

I find his use of the Signor-Lipps effect to be somewhat problematic. I am not only referring to his use of an old chronology for the northern lowlands. For example, he claims that Chichen Itza collapsed around 910 but archaeological data suggests that the site was a major power until 1050. He further argues that Mayapan emerged as a major political power around 1200, but this is likely to have occurred at 1050.

What he misses is the differences in paleontological excavations and archaeological excavations at Maya sites. Maya sites are often large and easily detectable on surface, fossils are most often not as easily spotted even if you happen to be located at the right sedimentary rocks. Maya archaeologists do therefore not dig as randomly as palaeontologists have to. Of course, archaeologists cannot cover everything, so in one sense we dig randomly, but the main underlying assumption in Gill’s hypothesis is that he sees culture as an organism.  You may never find some dinosaur species in the Early Tertiary period but you do find hierarchies, pyramids, ballcourts, etc. after the “Maya collapse”. When organisms goes extinct they remain so, what we call “cultures” (a term I seldom use) do not go extinct, they transform into something else. Palaeontological sampling problems related to the Late Cretaceous mass extinction is therefore not a suitable analogy for the archaeological sampling problems related to the Maya collapse.



%d bloggers like this: