My assumption: English modal verbs are non-tensed (i.e. we don't say shoulds or shoulded).

Yet, in X' bar theory, modal verbs appear under the inflection node I', precisely where we find the inflectional morphemes -s and -ed.

Is my assumption wrong, and modal verbs are actually tensed?

If my assumption still holds, why do modal verbs appear under I'?

  • Is this a homework question?
    – Draconis
    Commented May 19, 2018 at 19:50
  • 2
    No. Simple curiosity. I never had a chance to study linguistics in a formal setting (i.e. university), so I just read and try to make sense of things on my own (and I'm often left with questions, such as this one).
    – Puzzled
    Commented May 19, 2018 at 20:19
  • 1
    X' theory is somewhat outdated, if not obsolete. 'I' nodes are generally meant for auxiliaries & Co. Many linguists use I for modal verbs because they behave the same way syntactically. Note that X' bar theory isn't terribly consistent when applied to English - in languages with rich(er) inflection I is reserved for all finite verbs.
    – Atamiri
    Commented May 19, 2018 at 22:20
  • Related question: linguistics.stackexchange.com/questions/17324/… Commented May 19, 2018 at 22:34
  • 2
    Most are: "could, might, would and should" are undoubtedly preterite forms, since they are the forms required in backshift.
    – BillJ
    Commented May 20, 2018 at 11:17

5 Answers 5


I start out with declaring my ignorance. I do not know what X’ theory is. I am a mere historical linguist. For me, the reason we do not say (in English) “*shoulds” is that in the Germanic languages “should” is a pretero-present verb: it is in form past tense, but in meaning present. Like “I will, he will”, not “*wills”. In many other languages (Indo-European or not) modal verbs are conjugated normally in the present tense. But of course English is the unquestionable paradigm for linguistic “universals”.

  • 1
    English isn't even particularly consistent about this: other (non-modal) auxiliary verbs do conjugate fully.
    – Draconis
    Commented May 20, 2018 at 2:26
  • 1
    Sorry, I'm confused. How is will past tense in form (especially considering that would is arguably the past tense of will)? Commented May 21, 2018 at 11:08
  • 1
    @Araucaria. “shall” is a pretero-present: it conjugates like the past tense (“he shall” not “*shalls”), but it is functionally present tense, and it has the past tense “should”. Similarly “can” > “could”.
    – fdb
    Commented May 21, 2018 at 13:16

I wouldn't say that modals are I': rather, I'd say that they're I. In other words, a modal verb is syntactically an inflection, not a full verb. (It's easy to see that modals don't act like Vs: they can't be put in non-finite contexts, like *to should.)

In X-bar theory, the I is an inflection applied to the V.

  • In "Alex walked home", the I is "-ed", and the V is "walk".
  • In "Alex should walk home", the I is "should", and the V is "walk".

The trees for these two sentences are almost identical, except for the changed I. (The fact that "-ed" then goes off and affixes itself to "walk" is attributed to movement, or explained away by "that's morphology instead, not syntax, so it's someone else's problem".)

Notably, this is the same place where "do"-support happens: the auxiliary "do" appears as an I. However, this is different from auxiliary "be" and "have", which are actually Vs.

Hopefully that clears it up!


In current English, the common modals are paired up in a fashion similar to present/past pairs: "will/would, can/could, shall/should", and sometimes "could" has the sense of a past tense "can". Of course, the "-d" occurs regularly as a past affix.

Both "would" and "could" occur as past-shifted forms when direct discourse is shifted to indirect speech:

  He said "I will" --> He said that he would  
  He said "I can help" -->  He said that he could help

Otherwise, I can't think of any reason to regard modals as tensed.

  • You can also say "He says he will" / "He said he would" (though also "he said he will" and "he says he would", so I'm not sure how relevant that is)
    – anon
    Commented May 20, 2018 at 0:41
  • 1
    @NicHartley, Tense shift in indirect speech occurs when the indirect form is a fair report of the direct speech counterpart. Person shifting also occurs. So it is not just a matter of what can be said, but what it means.
    – Greg Lee
    Commented May 20, 2018 at 1:10
  • Also when embedded under items like past tense think: I thought you can play the guitar. Also in remote conditionals which show backshifting of tenses, and after verbs such as wish. Can you come tomorrow? - *I wish I can/I wish I could Commented May 21, 2018 at 11:12

Leaving aside X' theory¹, an easy thing to do, since it makes far too many invalid assumptions and doesn't work very well to start with, the question of whether English modals are tensed or not depends on what one means by "tense".

Traditionally, tense is a verbal inflectional category, like the Latin, Greek, or Sanskrit tenses. English has only two such, past and present, aside from defective verbs like beware and born. English modal auxiliary verbs, as noted, are not inflected for tense or anything else, except for the special cases that Greg Lee mentioned in his answer, where the old preterite/present modal stems can still occur in that sense. But these are very rare and don't count as inflection, really, since it's simply an opposition.

Many theories, however, require a tense marker on every verb (one of the invalid assumptions mentioned above), and so they put one there. Of course the marker is invisible (as well as inaudible), just a local variable that's needed to process the input and eventually disappears. Why go to all the trouble?

¹Pronounced "X-Bar Theory" by the cognoscenti, which gives you some idea about its degree of abstractness and contact with linguistic reality.

  • I thought that main clauses are always finite/tensed, but in "I should go now," modal should is tenseless (along the lines of what's been said here) just as the infinitive go is tenseless. Is my thinking wrong, and main clauses can be untensed too? Or am I missing something? Thanks. By the way, I'm not sure if this should be a new question.
    – Puzzled
    Commented May 20, 2018 at 3:43
  • @Puzzled The modals are primary forms, i.e. they can head tensed clauses.
    – BillJ
    Commented May 20, 2018 at 13:18
  • 1
    @BillJ, there doesn't have to be a tense-marked verb in the clause for it to count as finite. Main clauses are sposta be finite (though there are all kinds of utterances that aren't, e.g What to do about this? Going to the dance? Never been there). Any subordinate clause that has an infinitive or gerund verb (especially with a gerund or infinitive complementizer) is non-finite. Any other subordinate clause (especially a that-clause or a Wh-clause) is finite. Tense markers don't guarantee finite clauses, and vice versa.
    – jlawler
    Commented May 20, 2018 at 17:02
  • 1
    "Primary forms" is an arbitrary term, meaning forms that can head tensed clauses. That's not explanatory. It's true that the prototypes are tensed. But prototypes are tendencies, not definitions.
    – jlawler
    Commented May 20, 2018 at 17:28
  • 2
    "All verbs are either primary forms or secondary forms". This sounds like the Law and the Prophets. Who says and what's their authority? Let alone evidence.
    – jlawler
    Commented May 20, 2018 at 17:41

First of all, what is tense? Take a look here.

I copied and pasted the following from my 2008 paper:

In the classical Government and Binding Theory (Haegeman 1994, Haegeman & Guéron 1999) it was generally assumed that modal verbs were base-generated in the I-position:

[CP [IP He [might [VP call me tonight]]]].

In the first instantiation of the Minimalist Program, modal verbs were believed to be inserted directly in T via Merge (following Chomsky 1999). However, this position was later revised and now there is a general consensus that modal verbs project into a separate functional category (Adger n.d., Abraham 2002, Butler 2003, among many others), depending on the type of modality it expresses, root or epistemic, according to the Hierarchy of Projections (Adger 2003):

T> Neg> Modepi> Asp> Modroot> v> V

I proposed the following feature checking mechanism for sentences containing modal verbs:

  • Mod intervenes between T and v, thus the Pronouncing Tense Rule (PTR) will fail to apply in English;
  • Since v is no longer the head of T’s sister, the PRT prohibits a valued [uInfl:present] on v being pronounced;
  • do-support does not apply because the tense feature is valued on Mod.

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.