4453RE: [Synoptic-L] Length of Luke (and Acts)
- Aug 4, 2012I would put it this way: Stylometric tests look for patterns in the usages of various words or phrases when comparing two or more pieces of text. Depending on what words or phrases are used in such a test, we may find patterns that we interpret as being related to genre, style, or authorship. Further, the selection and/or the grouping of the words or phrases used in any such test can affect the patterns seen, as can the particular statistical procedure used. Therefore, I would say that no individual stylometric test can be relied upon, but where several different tests all seem to be showing the same patterns (or lack thereof) of usage of words or phrases then we should be able to, at least, state that the patterns either exist or don’t. Whether this then can be attributed to style, authorship, or something else, is a different question.
If we find that the same patterns do exist, or not (at a significant statistical level), in a number of different tests, and have structured our tests to eliminate (or at least minimize) genre differences, then we have a phenomenon that requires an explanation. The question then becomes whether particular patterns of usage are a reliable indicator of authorship or not, which appears to be the sticking point. However, I think we are on safer ground if we say that the lack of common patterns of usage is a good indicator of different authorship than common patterns of usage are an indicator of common authorship. In other words, I think we can at least eliminate some hypotheses in situations where we don’t find the patterns that the hypotheses predict.
David Inglis, Lafayette, CA, 94549, USA
From: Synoptic@yahoogroups.com [mailto:Synoptic@yahoogroups.com] On Behalf Of David Mealand
Sent: Saturday, August 04, 2012 9:24 AM
Subject: RE: [Synoptic-L] Length of Luke (and Acts)
Bruce claims I said
stylometric indicators tend to be sensitive for style, but less so for authorship, a matter for which they were not designed.
Not so. What I said was that stylometric tests tend to note genre differences as more prominent than source or author differences. Tests have to be structured to allow for this.
I have no idea what people may have said at SBL. My point is that my tests corroborated Walters' findings, giving support I had not expected to a view towards which I had previously been resistant. I have read some of the reviews of the book and noted that most of the points made in resistance to the critique of Lukan authorship of Acts did not deal with the tests reported, but appealed to matters
such as similar themes in both works, a line of argument I find particularly unconvincing but not surprising.
[Non-text portions of this message have been removed]
- << Previous post in topic Next post in topic >>