This is from a paper titled "What for?" by Bas Aarts:

(35) [NP It [S′ [COMP for] [S Mary see his relatives]]] [M may] [VP distress John]

Bresnan’s account was very influential in proposing that COMP and S together form an S-bar constituent (indicated as S′ in (35)). However, it is not clear how to is accommodated in the clause labelled ‘S’. This analysis was adopted in later Government and Binding Theory in which for was no longer regarded as being coupled with to, but as a single element, although it was recognised that ‘[t]here are evidently relations between COMP and INFL (that-tense, for-to)’ (Chomsky 1981: 54). In more recent work this ‘single complementiser analysis’ has become the standard analysis in Chomskyan grammar, but also in other generative frameworks, such as HPSG (Head-driven Phrase Structure Grammar). For example, in Sag (1997: 458ff.) for is also treated as a complementiser – see (36). enter image description here Here for is regarded as an ‘object-raising element’. What this means is that the NP complement of for (them) is also the subject of the infinitival CP complement (to go to the UK), as indicated by the shared indices, and can hence be seen as having been raised out of the CP.

(Boldface mine.)

The author says that HPSG adopts what he calls ‘single complementiser analysis’, where for is not regarded as being coupled with to, but as a single element in the 'for NP to VP' construction. But the cited HPSG work by Sag treats both for and to as complementizers (C) in the 'for NP to VP' construction.

Does HPSG treat both for and to as complementizers in the 'for NP to VP' construction, as in the cited paper? If so, what does Aarts mean HPSG adopts what he calls ‘single complementiser analysis’?

1 Answer 1


As I understand it, the for-to complementiser is conceived of as a single complex complementiser. In Sag's (1997) analysis, for and to are separate complementisers. The for takes an NP and also a CP headed by to as a complement, where the NP is seen as having moved out of the lower CP.

Having said all of that, Sag subsequently ditched such analyses altogether. In his later work to is not a complementiser at all, but an auxiliary verb. See, for example, Sag et al (2020) Lessons from the English Auxiliary System.

Sag, Ivan A.; Rui P. Chaves; Anne Abeillé; Bruno Estigarribia; Dan Flickinger; Paul Kay; Laura A. Michaelis; Stefan Müller; Geoffrey K. Pullum; Frank Van Eynde; and Thomas Wasow (2020): Lessons from the English auxiliary system. Journal of Linguistics 56, 87–155. Published online at http://dx.doi.org/10.1017/S002222671800052X.

  • Thanks. So did Aarts misquote Sag (1997)? Also, did Sag change his mind about [NP] and [to VP] not being a constituent but being two separate complements of for?
    – JK2
    Commented Mar 15 at 12:50
  • I don't think so. The key bit there is ''for' was no longer regarded as being coupled with to, but as a single element, I think. He's not overly clear, though, granted. Re Sag and for, I don't know, I'm afraid. Commented Mar 15 at 12:53
  • I'm not too familiar with the terms of GENERATIVE GRAMMAR including HPSG, but Sag 1997's (36) labels to go to the UK as a CP, which means that to is a complementizer (C). Then, is 'the key bit' in line with to also being a complementizer? Also, is 'the key bit' also in line with CamGEL, which analyzes both for and to as subordinators (i.e., complementizers)?
    – JK2
    Commented Mar 15 at 23:10
  • @JK2 The key bit is that for them to go has just for as its C, not for ... to. The to is also a C, but it's not part of the head of the larger CP (it's not part of the C of the larger CP), it's just the C for a smaller CP constituent within the larger CP. Aarts's description is slightly confusing because he calls the analysis of the 'CP', the single complementiser analysis - where he means that the larger CP has only one word as its C. However, there are two Cs altogether because there's a smaller CP inside the bigger one. Main takeaway: there's no for-to complementiser. Commented Mar 15 at 23:32
  • 1
    Ah ha! Good to know!
    – JK2
    Commented Apr 3 at 11:06

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.