NOTE: Skip this page if you want to 'cut to the chase'! This section explains sequencing methodology, and SUMMARIZES the previous section on Bible 'dating'.
The period of the New Testament was somewhat fortuitous. Alexander the Great had effectively spread greek over a very wide area (people forget the greeks were 'in-charge' about 300 years). So by the time the romans arrived, greek was wide-spread but had deteriorated from 'classical' into a 'common' or 'koine' greek. English has similarly shifted from formality into today's polyglot usage (each generation decrying the sloppiness of its youngsters). Spanish from the 1400s has had a similar pattern in Mexico.
The benefit of this is that it's feasible to sort the koine-greek writings in their likely sequence, using the change in the syntactical mix.
Sedona is what is called 'the high desert'. But come winter, the snows sort of slip down from the northern mountains and blanket the town. But noon, it's all gone, so you have to take your pictures quickly!This whole process is described in considerably more detail on our 'Syntax Dating' page.
But for quick understanding of the process, here's a summary:
Step 1. For a given NT book (eg Galatians), move across the greek sequentially and for every 25 words, classify the syntax for each word, summarize the counts, and then feed this into a file.
Step 2. Iterate this file through a neural network hundreds of times until the network can reasonable reproduce the file (learn its internal characteristics). This creates what might be called the book's syntactical signature.
Step 3. Do this for all the various writings you're interested in. You can include not just the literal writings, but specific groups of writings (e.g. Matthew's 'Q', John-Signs, a composite of Paul's early writings etc).
Step 4. After each writing has a 'signature', apply that 'signature' to all the other writings, to see how close each can come. So, for example, a 'Paul' signature can quite easily reproduce the file for 1st Thesselonians or Ephesians ... not too surprising. Conversely 'Paul' generally has trouble reproducing the file for Titus, generating a considerable amount of error.
Step 5. As a last step, you can select a specific 'author' such as 'Paul' or Matthew's 'Q', and then for each signature-applied writing, sort them according to how well the 'author' does (low error to high error).
Step 6. If you do this for all 'authors' and for all 'writings', you can easily find who is on the 'extemes' (early or late), and after that sort the middle group. What you're effectively doing is sorting the syntax signatures by their respective error levels. Pretty straight-forward.
Now, most people new to this bring up several concerns, and I'll list them here for you.
How do you know you're just measuring each author's own style of writing and nothing to do with 'time'? Well, first, the resulting writing 'sort' closely matches what the scholars have already concluded independently. Could that happen accidentally? Here's a specific example. Mark and 1st Corinthians have syntax signatures that are very similar. Why? Mark and Luke's non-markian writing (both gospels) are not close at all. Why? Altogether, there's about 200 comparisons like this, and they generally fit a 'time' scale, not a type of writing, or specific author grouping.
Yes, but the NT writers often used 'secretaries'. Maybe you're just accidentally 'crossing the paths' of the various secretaries?. Actually, there's quite a good chance for this. Mark in the previous example, closely matches 1st Corinthians, suggesting the possibility that Mark possibly wrote for Paul. To analyze this, I've literally mapped out Acts on a time/location basis and then overlayed the writing signatures to make sure the secretarial issue is not the source. By the way, the NT writers having secretaries is actually a 'good thing' since their greek would be likely closer to a shifting time scale (just like any professional's english today).
I'm still not convinced. For example, most scholars believe Matthew used Mark's account in creating his own. So therefore the so-called 'Matthew signature' would be invalid. You are indeed correct. Believe it or not, 'Matthew' can write much of Mark, better than 'Mark' can. The reason is that Matthew has likely preserved 'Mark' better than 'Mark'. And since Mark is close to 'Paul' in time, quite a bit of 'Paul' text is similar to 'Matthew' (because of markian passages). This illustrates how the signatures can be used to evaluate the internal passages to estimate their likely source and 'when written'.
Well, fine. But does anyone else 'buy' this method? Most Bibical analytics are the 'protect-my-reputation' and 'dont-get-me-in-hot-water' type. They use primarily traditional statistics in various ways. In comparison, neural networks don't give you much 'protection'. Instead, they are primarily designed to seek out complex patterns ... similar to how marketers analyze your purchases at a store. I've taken the marketing expertise for these analytics, and applied them with little change to 'greek'. Almost the same exact problem. And interestingly enough, there's far more money banking on the success of neural networks, than are banking on New Testament greek (unfortunately).
To be truthful, you as a reader should be careful about anything you read. Compare it to what you already know. Sift it. Be critical. But also be open to new ideas. And the 'new idea' here is 'Jesus' ... maybe he's worth a second look. The gospels might have some merit!
Next page please ...