Me too.Barry Hofstetter wrote: ↑December 18th, 2017, 3:02 pm I'd also be interested in Stephen's take on this if different from the SIL definitions he quoted, and in Mike Aubrey's perspective.
Sure, and every model will say that. I think my point is largely an epistemological one: words like direct objects and accusative are not in the language, they are in the model, which is an attempt to account for the things that happen in the language. More than one model may account for a phenomenon found in the language equally well, or one model may account for it better than another. (And I think you understand this perfectly well.)Barry Hofstetter wrote: ↑December 18th, 2017, 3:02 pm I think we need to start strictly with how the constructions are handled in Greek, and not let English or comparisons with other languages get in the way.
To use Wittgenstein's analogy, there are many maps of our earth, but only one earth. Two maps may be equally good, but different. Or one map may be better than another. Regardless, neither map is the earth. So when we compare models, it's important to remember that they are all just models, and keep turning back to the language they describe. No matter how many maps I have of the Appalachian Trail, none of them describes it perfectly, and none of them is a substitute for hiking the trail.Barry Hofstetter wrote: ↑December 18th, 2017, 3:02 pm I have always taught that direct objects are accusatives strictly within the framework of Greek syntax, and that if something else is used that looks like a direct object in translation, we have to conceptualize it differently. Now having said that... what exactly is a direct object but an essentially adverbial construction which limits the action of the verb? And for that matter, isn't that what dative or genitive complements (such as with verbs of filling) also do?
In other words, there multiple perspectives for viewing these things, and most of them valid within a certain context.
Within a given model, it's important to be consistent. And when we ask questions like ones in this thread, it's helpful to ask ourselves if a given question is a question about the language or a question about the model. I suspect that Mocciaro's model and yours are pretty much equivalent, they just account for things differently.