Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!wuarchive!ukma!seismo!dimacs.rutgers.edu!aramis.rutgers.edu!athos.rutgers.edu!christian From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) Newsgroups: soc.religion.christian Subject: Re: Analysis of a sample text passage from Richar O'Keefe Message-ID: Date: 27 Mar 91 02:30:03 GMT Sender: hedrick@athos.rutgers.edu Organization: Comp Sci, RMIT, Melbourne, Australia Lines: 78 Approved: christian@aramis.rutgers.edu I have quite a number of comments to make on the earlier postings on this topic, which will have to wait until after Easter. I have had one paper posting from New Zealand which will earn its author US$10 (he got 7 right out of the 14). The biggest comment is to say THANK YOU VERY MUCH. My posting was to some extent a "put your money where your mouth is" exercise, and Mike Siemon, by taking the risk of exposing how he thinks the method would work, has demonstrated beyond reasonable doubt the sincerity of his belief in the DH, and by the clarity and detail of his analysis, he may well have convinced a great many readers that it is reasonable. I must admit that I was biting my fingernails as I read through his presentation, because although I had reduced the required number of right answers from 14 to 10 in order to roughly double my expected loss, I really didn't want to part with $100, and his presentation was so convincing. I turned to my original copy (with the sources labelled) in genuine apprehension; but oh what a relief. Amongst the things Mike Siemon did that I liked: he showed a clear appreciation of uncertainty, limiting his estimates to 10 sentences, and he gave an estimate of how reliable he thought _those_ were. Splendid! > This makes 2,4,6,11,14 A sentences, and 1,3,5,9,12 B sentences; > Of these ten, I expect to have 7 +/- 1 correct. In fact 6 are correct, which is well within chance, but it is also within Mike Siemon's estimate of how well he has done. Excellent! The study doesn't _prove_ anything (it was only a pilot study after all), but Mike Siemon's posting has convinced me that I ought to look into DH methods rather more. (I did start to read Gerhard von Rad, but when I found him talking about "saga" in the Torah I decided not to bother any more.) If I add in Mike Siemon's last estimate for the remaining sentences, his score rises to 7 out of 14. > It is difficult for me to believe that the same author would -- in a short > passage from a single work -- use both the correct "Fleischl" and the hyper- > correct "von Fleischl" to refer by surname to Freud's friend ... Believe it. > That > and the shortness of the test passage, and the tight binding of BOTH sources > to recent, well-documented events (thus reducing the scope of traditional > variation in telling the tale) are my main direct complaints about O'Keefe's > test sample. Plus the omissions from his sources that I think I have detected > suggest to me a *different* redactorial procedure than is posited for the DH > redactor. The shortness of the sample is regrettable. I could very easily have made it twice as long. I had to make it short enough so that there'd be some chance of people being willing to try. Precisely two people did, and then only after a somewhat heavy-handed "dare" following up the original posting. (No, the second posting was *NOT* prompted by any recent postings in soc.religion.christian.) However, the "Freud story" is one which has been subjected to a great deal of retelling. The distance between the events and the texts in question is rather greater than the distance between the Gospels and their events. Freud himself began the mythologising (see Eysenck's "The Decline and Fall of the Freudian Empire" for details), and Jones continued it. In fact, one of the texts is out to challenge the "traditional" story. > But as in my "generic" protest > before I looked at this data, I find it hard to believe that O'Keefe has not > biased the question, consciously or unconsciously, by taking selections which, > though they differ in style and content, do so to a FAR lesser degree than is > claimed for the sources of the Torah. This is fair comment. For a pilot study, it's not such a problem; I've learned a lot about what a realistic test should be like. -- Seen from an MVS perspective, UNIX and MS-DOS are hard to tell apart.