According to Wikipedia, generative grammar is distinguished by its putting the object of a sentence inside a VP. How reasonable is this assumption?
-
5Question in your title is completely different from the question in the body. you should change the title because the title question is just an invitation for opinions. Wikipedia is just totally wrong if that's how it characterizes generative grammar.– user6726Commented Sep 29, 2020 at 19:40
-
1That's hardly the defining characteristic of generative grammar. As to how reasonable it is, that depends on what kind of "reason" you're looking for. La grammaire a ses raisons que la raison ne connaît point.– jlawlerCommented Sep 29, 2020 at 20:26
-
2I thought most non-generative frameworks also put object NPs inside VPs.– curiousdannii ♦Commented Sep 29, 2020 at 22:11
-
@curiousdannii What are some examples?– JeffreyCommented Sep 29, 2020 at 23:13
-
1The object is complement of the verb, so where else could it be put? Descriptive grammar says the same.– BillJCommented Sep 30, 2020 at 7:04
2 Answers
Recall that the generative framework has been around for 60+ years, and there are numerous specific generative models. It is not an axiom of GG that there even exists a VP, though it is an assumption made more often than it is denied. In all word orders besides OSV and VSO, the verb and the object are adjacent, which is a minimum condition for "reasonably" concluding that the verb and object are a constituent (conventionally called a VP).
In the vast majority of instances where people talk about GG, they really mean "in contemporary Minimalism". In that framework, the question is, what does Merge do, and what does it do it to? It is a basic premise of the theory that Merge combines 2 things into 1 (not 3 or 4), and that subcategorization plays a role in how a verb and a noun can combine. The subject-predicate distinction is a basic fact of language and logic, recognized since Aristotle, which motivates taking the subject "out" of the VP, leaving the object as the remaining nominal that may combine with the verb.
The Minimalist analysis of basic word order is much more complex that that sound bite, because there is a lot of movement, and many more nodes than just V, VP, N, NP. Koeneman & Zeijlstra Introducing syntax is an easy intro to that analytic framework. The most-viable alternative to V+O=VP is flat structure with no structural distinction between S and V. Perhaps then what you are uncertain about is the merit of the alternative of less or no constituency, which would be a separate question.
It’s reasonable for a few languages such as English but in many languages objects are located at the clause level. In fact in many languages objects in focus are located within the VP while topic objects are “detached”, which might be due to the universal tendency of focussed constituents to huddle together (unless one of them is contrastive). In sum, it’s language-specific and largely determined by pragmatics (information structure)