Hi Peter,Well, Ken, I am glad to see that you have identified the issue (see http://groups.yahoo.com/group/PennerThesis/message/2 section 2) and have clearly stated a methodology intended to minimise the problem (same section, and http://socserv.socsci.mcmaster.ca/westerholm/thesis/Penner%20Proposal%205.pdf section Methodology). You are of course correct to restrict your analysis to those verbs whose temporal reference you can be reasonably sure of, although you need to watch out for some statistical bias introduced by this. I still feel you are being rather optimistic, but I hope not.
The methodological problem of circularity is indeed significant, but not
enough to be called a "flaw". I am not sure how Rolf deals with the issue,
but it is my experience (analyzing Qumran Hebrew verbs) that context is more
helpful than one may have expected. My pilot project (available from the
link in the signature) notes only 6 instances (out of 96 verbs) where the
time reference is ambiguous in 1QSa. Circularity in the statistical
correlation between form and meaning can be minimized by noting those
ambiguous cases and leaving them out of the primary analysis.
So you are technically right, "some" verbs but not "all" verbs have a clear
contextual time reference. However, that "some" is the majority, and enough
to make a strong statistical case.
Ken Penner, McMaster/DSS
http://groups.yahoo.com/group/PennerThesis
Archive powered by MHonArc 2.6.24.