| Analysis Patterns (methodology) |
| Analysis Patterns (methodology) |
|
GENERAL AND DATING PATTERNS As you can imagine, there's quite a lot of original data plus new data created. First, you have the actual english text, the greek or hebrew text, the strongs numbers, and the morph-tags. Then you have comparable values for all possible authors, for both vocabulary and syntax for each word in each verse. I've run this for various greek versions, plus hebrew. I've also run it for various morph-tag classing schemas, in order to see their effects (each of which is selectable in my dbBible software). Below are some of the principles that I have found to-date. [1] VARIATION Syntax (morph-tag patterns) appears to match 'time'. In other words, books written around the time of Paul (including Mark) have a different pattern than those written later. Indeed the level of differences is closely associated with the time scale from 50ce to about 100ce, Paul early and Peter late. Using this pattern, I 'date' the books (ballpark only) using the Pauline anchor from Corinth and surmising a 100ce end date. Vocabulary (strongs assignments) tends to match the author, and does not vary across a time scale. As a result, the vocabulary patterns are useful for evaluating conformity within a claimed writer (eg a chapter), and also cross-evaluating the syntax patterns above. [2] VERSIONS DIFFERENCES For the New Testament, I did my initial work on the Byzantine, and then compared using the Textus Recepticus, Westcott-Hort, and NA27. Here's some of the differences (recognizing all are eclectic forms): Byzantine: Syntax patterns tend to strongly group around Matthew and 1st Corinthians. 'I' call this the 'Antioch source'. Pattern connectivity tends to 'string out', suggesting slow expansion. There's little inter-connectedness to help confirm 'syntax dating'. Westcott/Hort: Syntax patterns group around John (gospel) and Ephesians. I call this the 'Ephesus source'. Patterns are highly inter-connected, easing 'syntax dating'. Textus Recepticus: Syntax patterns group around Luke and Romans. I call this the 'Rome source'. Patterns are also inter-connected, generally matching the WH 'syntax dating'. NA27: As an update to WH, syntax patterns group around the pauline writers, to the exclusion of 'jewish' writers. Initially, I thought Marcion was resurrected! By the way, if you're highly familiar with each of these forms, you're likely having a heart-attack. That's because you have so little confidence in the eclectic manuscript process of each. Think about it! [3] LANGUAGE DIFFERENCES I did most of my initial work on the greek, but wondered how the methodology would fare with the hebrew. The Old Testament is a little unusual, since the actual manuscripts are more recent (approx 1000ce) and the time frame for the writings is substantial (maybe a thousand years or so?). I well knew there wasn't any telling what would happen with the same approach. As it turns out (and this may generalize to the New Testament as well), the hebrew analyses had the same syntax (time pattern) and vocabulary (author pattern) as did the greek. The primary difference was that the Old Testament 'groups' into three relationships: Samuel up to the Assyrean captivity, the Judean period, and the post-exile period. Books in the first group can be sequenced among each other, but not relative to books in the other groups. 2nd Samuel doesn't relate to Nehemiah for example, but 2nd Samuel does relate to Ruth, which is in the early group (along with Jonah!). Indeed the groups have almost no 'crossover' where a book relates to two of the groups. This suggests to me that the hebrew language (and/or editing) evolved in major stages, and each group has to be evaluated separately. In a later section, I'll go over the groups themselves. Regarding the generalization to the New Testament, I am wondering if the alexandrian problem above might be similar ... a common greek editor near the end of the period? [3] EMBEDDED QUOTES The gospels use quite a number of quotes, as an example. Just generalizing, the quotes more often than not create a 'non-conforming' syntax/vocabulary issue (as they should, since it's another writer being quoted). However, it doesn't happen in all cases, since some of the quotes are paraphrasing, and some are LXX-greek-originated (versus maybe a proto-massortic hebrew or aramaic). [4] MAJOR STYLE BLOCKS The Old Testament has an obvious distinction between prose and poetry. Also, the 'law' sections and the geneologies have their own major style types. The neural network analyses do indeed recognize the differences, and can adjust somewhat, but not completely. Thus, extra interpretive care has to be exercised with the unusual style types. The New Testament is not as sensitive, even with embedded 'psalms' and so forth. [5] ASSIGNED DATING The neural networks evaluate all possible authors against each word. If the writer datings are very close, especially in time, then a randomly assigned authorship occurs between the very-close writers. In my dbBible software, I thus have a option to group the authorship in general time periods for better usability. I'll discuss these in the next section. Next page please ... >> |
| Copyright ©, 2007, dmbarnhart |
| Prev Back Next |