The adverb 'only' is known to be able to come in a variety of positions. The following examples demonstrate that it may be generated in positions that aren't so simple to syntactically analyse. The examples below involve control/subj-to-obj raising:

1) Only he wants Max to eat the chicken.
2) He only wants Max to eat the chicken.
3) He wants only Max to eat the chicken.
4) He wants Max only to eat the chicken.
5) He wants Max to only eat the chicken.
6) He wants Max to eat only the chicken.
7) He wants Max to eat the only chicken.
8) He wants Max to eat the chicken only.

Adverbs are usually analysed to modify VPs, so the AdvP is usually generated adjoined to VP. This enables the adverb to occur before or after the VP on the surface. Moreover, adverbs can be preposed to SpecCP. The problem arises when analysing the adverb 'only' when it comes between the verb and (raised) object (3), the subject and HeadIP (4), and between the determiner and noun (7).

In (3), it is conceivable that AdvP is generated as an adjunct of VP with "want" moving from VP to vP. The split VP hypothesis does deal with this problem in fact. 'Only' is unexcitingly generated as an adjunct of VP.

[IP He [vP wants1 [AgrOP Max2 t1 [VP [AdvP only] [VP t2 t1 [CP t2 [IP t2 to eat the chicken]]]]]]]

However, does V always move to v even in simplex, non-raising/control declaratives such as "he only ate chicken"? If it does, is AdvP an adjunct of vP when the adverb precedes the verb?

(By the way, which is a more recent and widely accepted analysis of control - null PRO or subj-to-obj raising?)

An additional problem is when raising verbs such as 'appear' are involved as in the example below. How is 'only' able to appear between HeadIP and SpecIP? The split VP hypothesis doesn't seem to be able to account for this.

[IP Max1 appears [IP t1 only to have eaten the chicken]]

Perhaps in the examples involving intra-DP 'only', 'only' is an adjectival of sorts as it seems to be modifying the DP/NP in examples (1), (3), (6), (7) and (8). This could be supported by 'only' being infelicitous with expletives such as raising-it and weather-it.

*Only it seems that Max is eating the chicken.
*Only it is raining.

So, adjectival 'only' modifies "he" in (1), Max in (3), "the chicken" in (6), and "chicken" in (7). The adjective 'only' is known to modify the NP within the DP, e.g. "the only child", which is different from 'only' coming before D or after N as in "only the child" and "the child only". So, how us this formalised? Could it be a quantifier?

What do you all think?

  • This is a matter of what the focus of only is, and where it's located in the sentence. Only can go immediately before its focus, or immediately before any constituent that contains its focus. For more detail, take a look at Larry Horn's classic paper on only and even, (which work the same syntactically).
    – jlawler
    Sep 26, 2015 at 18:15
  • @jlawler True, but what I want to know is where in the tree 'only' is located. Sep 26, 2015 at 18:44
  • Which tree? In any syntactic tree, with only in any grammatical location, only will appear chomsky-adjoined to the next constituent (the one containing the focus).
    – jlawler
    Sep 26, 2015 at 19:59
  • @jlawler Yes, in any syntactic tree. Is it really possible for an AdvP to adjoin to constituents other than VPs? I ask this because I have never seen it before in the generative literature. Oct 22, 2015 at 7:28
  • 1
    You really ought to read McCawley's account of only in SPHE. Only modifies its focus and can be moved so as to become a modifier of any constituent containing the focus within its scope.
    – Greg Lee
    Oct 23, 2015 at 1:12


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.