Never has anyone ever said that the this is a usage of imperfect verbs of speech.David Lim wrote:Sure, the text is segmented, but Mark does not consistently use the imperfect only at these breaks. For example, can you explain Mark 2:23-26 which I quoted?
...
If the places where direct discourse in a narrative text is segmented is only a subset of the places where the author uses the imperfect, then segmentation cannot be considered as a usage of the imperfect verb of speech. Otherwise I can claim that all conjunctions and particles are also used to segment direct discourse. The more reasonable assertion in my opinion is that the context of the discourse itself segments the discourse, not any single word or word form in it, though it is of course more likely for one to use unusual words or word forms at the junctures between sections. To prove it, my test would be to remove the verb conjugation of all instances of "λεγειν" in an extended length of direct discourse that you are not familiar with and you should be able to determine where an imperfect should be used. If you are very accurate, then I would agree with you, because you would clearly demonstrate a correlation between the imperfect verb of speech and the segmentation of direct discourse.
You've still got it backwards. We're not classifying instances of the imperfect. Imperfects used in breaks within direct speech aren't a class of their own. The only reason it appears that way is because Wallace made the claim that imperfects use in breaks in direct speech are aoristic. And for that reason, the nature of the argument used by Runge and Levinsohn needs to start there. But they're making a broader (and different) claim.David Lim wrote:Ah okay. Anyway I still claim that segmentation of a long speech is not quite the author's intention when he uses the imperfect verb of speech, because all the examples fit the broader classification of the imperfect literally signalling an incomplete event. I therefore think there is not much reason to divide them into two groups, one for segmentation, the other not for segmentation. I choose the simpler, even if more vague, explanation.
Let's try a different starting point for this.
When humans speak, we know that there is only so much that can be said before the audience struggles to process it. At a lower level, the proposition/sentence/clause becomes a useful place to break a discourse into segments. Breaking text into segments allows for the audience to more easily process the text. And its done at a higher level, too. And when you're writing down a text in a language whose writing tradition doesn't have a system for marking larger chunks such a paragraphs, there needs to be some way to do it. So Mark breaks a longer piece of direct discourse in to smaller pieces. The discourse isn't over, so only imperfective verbform would be acceptable for starting it up again (as you say, the general category of continuous work perfectly here) and since this is narrative, it also needs to be a past tense verbform (assuming the semantic-pragmatic requirements for a historical present aren't met). And thus ἔλεγον is really the only option.
So, from the perspective of the author, the segmentation of the text comes first and the use of ἔλεγον is a natural result.
However, from the perspective of the person analyzing the text, knowing this information makes ἔλεγον a useful linguistic entity to watch for in making decisions of where the author wanted to segment the text. The segmentation there; ἔλεγον is a functional guidepost for the reader.