Ergative and Nominative Languages.

junk Marc writes:

Beekes give some compelling arguments that PIE may have descended from an older ergative language, eg, the old genitive became used as the new nominative (the agent in a passive construction is often a genetive, eg, in German ("von") & Dutch dialects ("van").

I was unaware that Beekes embraced the 'ergative theory of PIE'. I had thought this was a minority view, one of those interesting spins that cannot really be proven.
The main problem for me is you have to have an awful lot of pure grammar, including lots of stuff that does not occur in English to really get into the topic. And examples via Basque or Georgian grammar are beyond me. The only advice I really understood was that you have to analyze a sentence from the point of view of the verb -- and NOT from that of the subject. If the verb is transitive, the agent is in the ergative case and the patient in the absolutive case; otherwise, what English would call the subject is in the absolutive case. But ergative languages, so far as I understand, can have subjectless sentences, where the whole point is the action of the verb, as with the pale English imitation "The door opens."
Having said this, the idea is fascinating. Switching from being an ergative language to being a nominative language is a huge innovation, one that suggests a very small group of people in isolation, which seems to be the only scenario where such a fundamental linguistic re-analysis could occur. From what I've read, the suggestion is that such an ergative period would be deeply ancient PIE, even pre-PIE, in that such a change would represent a clean break from the previous, ancestral language family.
So. Is there any evidence in Uralic for an equally ancient ergative system?
Mark.