Greek Syntax: Unordered Group, Part II

In last week Friday’s post, I blogged about something that J.H. Moulton calls the “Pindaric Construction”. In a comment to that post, David Pereira noted:

The other cases I would question are those in which the “singular things” are joined by “or” or “nor” rather than by the word “and” such as in Matt 12:25. Though these might technically fit the description you gave earlier (i.e. where “a group of singular things in the subject have a singular verb in the predicate”), I don’t think they represent any diversion from standard grammar. Take this for example: “Neither a CAT nor a DOG IS allowed inside.” Though there are multiple subjects, the conjunction serves to relate each singular noun to the singular verb individually. So, I don’t think this is anything more than standard grammar for subject/predicate agreement.

Following up with David, I wrote:

On the search generally — I was surely thinking but apparently didn’t write that syntax searching like this is a way to evaluate assertions made in grammars like Moulton’s. Yes, the hits “techincally” match the description; they must be further evaluated to see if they all really do function as proposed. I think, in this sort of application, syntax searching is a way to narrow initial hits (the same search using only morphology and proximity would be complex if at all possible), not always acheive 100% grammatical accuracy (particularly when context can play a role in analysis).

I don’t know how special the structure is. There are instances like Mk 4.41 (joined by και) where the two singular things are combined with a singular verb, and it might be interesting to note them. But there are also the sorts of things you mention. Perhaps the better search would be to skip the ‘anything’ on the second word group, add και as connector, and see how the hit list changes. I smell a follow-up post …

This is that follow-up post. Here’s the modified query I mention above:

There are a few changes to note in this modified form of the previous query.

  1. I removed the anything operator between the two word groups in the Subject component.
  2. I added a Connector to the second word group, the word και in an effort to search for conjunctive relationships between the groups (or, an ‘and’ style relationship) instead of disjunctive relationships (‘or’ relationships) or negative relationships (‘nor’/’not’ relationship).
  3. I added the requirement that the first word group in the query also be the first word group in the Subject. This means that even if there are more than two word groups, the query will only find the structure once — instead of one hit for each combination of two word groups in the structure (as happened in Col 3.11 with the previous query).
  4. I changed some highlighting so the whole subject would be highlighted instead of individual word groups within the subject.

The result? Well, the hit list shrinks, from 275 hits to 81 hits. Many of the sorts of hits that David mentions in his comment are weeded out. Additionally, we only have one hit for verses like Col 3.11 (instead of the many hits of the previous query). That’s all good.
But some other hits are weeded out too. Re 9.12, one of Moulton’s original five examples, is no longer present. Further evaluation leads me to think that Moulton really meant Re 9.2 (which is located by this query) instead of 9.12, which just doesn’t make sense.
What does it all mean? I really don’t know. Chances are this just once again proves that the nice-and-tidy syntactic structures mentioned in passing in grammars (along with examples) aren’t necessarily as nice-and-tidy as they’re made out to be. I’ve said it before, I’ll say it again: Language is messy.
But what is possible now with these syntactic annotations is to begin to evaluate these sorts of statements about grammatical structures. We can now, with the assistance of syntactic annotations, build searches that take these larger-level clause and phrase structures into account, along with morphology, and then examine the supposed structure in greater detail to see if there really is something there.
And that was what I was angling toward in the first blog post (along with showing the new Unordered Group object), though I didn’t really say it: Here’s a structure mentioned in a grammar, what do we find if we actually search the whole corpus for something like it? Well, that is just one of the things we can do now. In the long run, this sort of work will end up making grammars sharper in their discussion and presentation of data.

On Facebook? Join the Discussion

Sorry, Comments are Closed.

You'll have to take it up with the author...