> Ok, so there's the difference between us. I am a
> true empirical scientist trying to find the most
> likely explanation for events. You, on the other
> hand, are a pseudoscientist, trying to impose your
> unquantifiable views as facts. If this is
> established in archaeology, then archaeology has a
> big problem. Explains a lot.
Fascinating. So archaeologists categorizing motifs using diagnostic criteria are not doing science. Geologists categorizing rocks using diagnostic criteria are not doing science. Entomologists categorizing beetles using diagnostic criteria are not doing science. Paleontologists categorizing fossils using diagnostic criteria are not doing science. Etc.
No, Martin, you have an astonishingly narrow view of science. There are myriad perfectly legitimate scientific methodologies where statistics are not involved, are complementary, or are even inappropriate. And you clearly have not a clue how archaeology works, but you feel competent to dismiss it. There's Dunning-Kruger for you.
When faced with a new archaeological assemblage, a good first step is to organize a systematic, rigorous typology based on the objects' or images' observable attributes, informed by comparison with other assemblages if applicable. With a large artifact collection to deal with - pottery or lithics, for example - statistical methods have been common for decades; for iconographic corpora, the categories are usually more discrete and obvious. At any rate, with systematic categories defined, the archaeologist can then move on to quantitative questions, with much of the GIGO factor eliminated.
Martin, you blithely skipped over that first step. That is how you were able to end up with, for instance, these clearly dissimilar images ending up in the same category...
...and these clearly similar images ending up in different categories:
Your treatment of the corpus was inconsistent and self-serving; given how confident you are in your results, however, I think you've reached a new level of GIGO: Garbage In, Gospel Out.