Wikipedia explains how the Inflectional Phrase has a VP as its complement and an NP (the subject of the phrase) as its specifier. It is long ago that I studied this, but a quick look at Sprachliches Wissen by Grewendorf confirms it. And the analysis works alright for many languages.

So I am unsure of how we can account for cases like the following Welsh sentence:

Naeth a   dyn brynu car
did   the man buy   car
'The man bought a car'

I can't figure out where the IP even is. Is naeth the complement of the IP? Then why is it to the left of the subject? Or has naeth somehow moved to the complementiser slot? Then why is it getting conjugated?

Either way, I would love to see some syntax trees of entire sentences for Welsh and maybe other VSO languages.

  • 1
    My guess would be that the IP scheme was almost exclusively motiavted by English (which happens to be SVO) and is simply not applicable to languages that behave a bit differently - like many linguistic theories. Commented Aug 31, 2016 at 21:15
  • @lemontree While I agree that the IP scheme was mainly motivated by English, there's a plausible analysis for VSO languages which works well for Welsh (though I'm more familiar with Irish Gaelic). In the above example naeth is an I and brynu car a VP (they form a constituent that can be moved around the sentence) which forms an exocentric category together with the subject. The difference is that in Gaelic all finite verbs are Is whilst nonfinite verbs are Vs. Joan Bresnan wrote a paper on Welsh phrase structure but I can't remember its title for the life of me.
    – Atamiri
    Commented Sep 1, 2016 at 1:41
  • @Atamiri What would this look like in a phrase structure tree? naeth is then I, the head of IP, brynu car is a VP, that much is understandable, that somehow goes together with the subject a dyn, but what is their syntactic category, and woud this then be the complement of I so the whole clause is an IP, does it have any specifier or can you not apply the classical X-bar notions here? Commented Sep 1, 2016 at 9:35
  • @lemontree I see you found the paper which I guess answered your questions. One small remark: I don't think this is an example of a mixed category, in this case I think it's just a plain periphrastic construction.
    – Atamiri
    Commented Sep 1, 2016 at 13:10

1 Answer 1


I googled for Joan Bresnan which @Atamiri mentioned in the comment and think I've found the analysis that he/she was mentioning and should be what you are looking for.

According to Bresnan (1997), just as Atamiri described, the finite auxiliary naeth is in I, the head of IP, while the infinite VP brynu car and the DP a dyn combine to the complement of I, labelled S.
As far as I understood it from quickly skimming the article, this DP + VP constituent constitutes a so-called mixed category, a "construction in which a single word heads a phrase which is a syntactic hybrid of two different category types" (Bresnan (1997: 2)).

Adapted from the example in (27), the phrase structure tree for your sentence should look something like this:

enter image description here

Bresnan (1998: 13) further explains that "S is univerally available as an exocentric category having no fixed categorial head and projecting no higher category. [...] Configurational S consists of a subject constituent and an XP predicate. [...] A configurational 'internal-subject' language (e.g. Welsh [...]) would have S under IP, and VP under S."
Assuming additional X' levels for the XPs, but without specifiers, the tree would look like this, along the lines of Bresnan (1998), (11):

enter image description here

Bresnan, J. (1997). Mixed categories as head sharing constructions. In Proceedings of the LFG97 Conference. CSLI Publications Online.
Bresnan, J. (1998). Optimal Syntax1.

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.