Verbs: Modality and the modal system

How possible situations relate to reality

Introduction

The concept of modality relates to the ways in which the possible situations described in a clause can relate to reality.

Note

Modality: Certain kinds of alteration in the relation of clause contents to reality

Grammatical context: The modal system of modal auxiliary verbs can, may, must, will etc.

Reality: Assertion of fact

An assertion of fact about the actual world:

He saw us (fact in the past in the actual world)

  • Male person or animal

  • Visually detected

  • A group of two or more including me

Non modal assertion about the future:

He leaves today (scheduled departure)

Lexical categories to express modality

Several different lexical categories can express modality.

Evidently he saw us (modality expressed as adverb)

Perhaps he saw us (adverb)

It’s possible that he saw us (adjective)

There’s a possibility that he saw us (noun)

Meaning of the modal auxiliaries

  • Epistemic: What must be true or could be true given what we know (from Greek episteme, knowledge)

  • Deontic: What’s required or permitted (from Greek deon, “what is obligatory”)

  • Dynamic: Related to physical capability (from Greek dynamikos, “powerful or effect”)

Epistemic vs. deontic contrast

Epistemic modality modifies clause meanings by expressing notions of what’s necessary or possible given what we know or believe. (“got to be true”, “might be true”)

Deontic modality expresses meanings relating primarily to what is required or permitted. (“got to do”, “allowed to do”)

He must have overslept (epistemic, infer, strong commitment, do not accept any other possibility)

He must apologize right now (deontic, strong obligation)

She may be resting (epistemic, don’t know if she is or isn’t, weaker, a mere possibility)

She may take as many as she needs (deontic, permission)

The storm should be over soon (epistemic, don’t know, but have reason to believe, in between: weaker than must but stronger than may)

They should try again (deontic, milder kind of obligation, attempt to influence what happens rather than to make assertion)

The epistemics: Relation to reality, belief, limitations on the speaker knowledge.

The deontics: Authority and judgment.

Both have varying degrees.

Note

The concepts of necessity and possibility.

These are the key concepts in what philosophers call “modal logic” – apply to both.

The difference is that with epistemic modality, necessity and possibility relate to whether or not something is actually true.

With deontic modality they relate to whether or not something happens or is done.

He must have overslept (epistemic, it’s necessarily true that he overslept)

He must apologize right now (deontic, it is necessary for him to apologize)

Neither example allows for any other possibility.

Ambiguous epistemic/deontic

You must be very tactful. (epistemic or deontic)

  • Epistemic: “I have evidence that leads me to believe that your are very tactful”

  • Deontic: “There is an obligation or need for you to be very tactful” (which might be one way of instructing you to show some tact)

Dynamic interpretations

She can speak (dynamic, ability, (foreign language competence)

I’v asked him to help us but he won’t (dynamic, volition, unwillingness to help)

With can, we find clear cases where both a dynamic and an epistemic or a deontic interpretation are plausible:

You can’t be serious. (could be epistemic or dynamic)

She can drive. (could be deontic or dynamic)

  • Epistemic: denies the possibility you are being serious

  • Deontic: Permission to drive

  • Dynamic: Your personality, a trait, you can’t be serious; attributes ability to drive.

The preterite forms of the modals

Similarities with ordinary preterites

  • Past time meaning: “could”, “would”

I asked him to help me, but he couldn’t/wouldn’t. (past time)

  • In the subordinate part of a remote conditional:“could”, “would”

We’d save a lot of money if you could/would cycle to work. (remote conditional)

Important

Unlike other modal auxiliaries, could and would can occur in the subordinate clause of a remote conditional.

“Could” and “would” express ability and willingness. Because ability and willingness are objective states, they can logically serve as a condition in a subordinate clause:

If I could fix it… (condition: If I had the ability)

If he would listen… (condition: If he were willing)

Other modals express probability. Because the word “if” already establishes possibility, putting a probability modal (like might) inside an if-clause creates a semantic clash.

Note

“Should” is the notable exception.

It can appear in the subordinate clause of a conditional, where it serves a specific syntactic function: to mark the condition as highly improbable:

If he should call, tell him I am out

  • All four preterites are found in backshift:

I thought I could/would/should/might see her yesterday, but I had to work late at the office. (backshift)

Note

Backshift is obligatory here: The adjunct “yesterday” locates the event in past time.

Non-backshifted forms are only permissible when the subordinate proposition remains *open or true at speech time.

See also

Backshift

Differences from other preterites

Restriction to subordinate

With other verbs, the modal remoteness use of the preterite is restricted to a few subordinate constructions.

Any modal preterite can occur in the subordinate clause, only a modal auxiliary can occur in the main clause.

Note

The ordinary modal preterite can not appear in main clause:

  • Ordinary verbs in main clauses: If ordinary preterites expressed remoteness here, you could not distinguish a past fact from a present hypothetical:

I bought a house (ambiguous, without restrictions it could be a fact or hypothetical)

  • Subordinate environment establishes the hypothetical frame:

I wish I bought a house…

Modals are allowed in main clauses: Could, would, might etc. have historically lost most of their past-time meaning. Their default is already modality, so they can stand in main clauses without causing confusion.

The reader immediately registers a modal as a hypothetical possibility (modal remoteness), not as a factual past-time event:

I could by a house (remoteness, not past time)

You could/might have been killed! (you’ve done something reckless, putting you at risk of being killed– but in fact you weren’t killed)

Could/Would you help me move these boxes? (interrogative clause, used as request, sounds more polite than present tense can/will)

You should apologise. (you owe someone an apology)

Note

In present-day English, this use of preterite should is not generally perceived as semantically related to present tense shall.

The use of shall to say what’s going to happen is vastly less common than the use of should (in fact shall is almost extinct for a lot of people), and “should” does not stand in the same semantic relation to shall as could and would in [iii] do to can and will.

There is no use of shall comparable to the epistemic should of:

The storm should be over very soon.

Remote conditional

A modal auxiliary is obligatory in the main clause of a remote conditional

A remote conditional expresses an entire scenario that is hypothetical/unreal. Both parts of the sentence must convey that unreality.

  • In the if-clause: The word “if” provides the structural trigger for modal remoteness. Because of this trigger, an ordinary verb preterite (If I won…) successfully signals unreality.

  • In the main clause: There is no if. An ordinary verb preterite in a main clause (I bought a house) can only express past fact. It cannot express unreality.

Since the main clause needs to remain hypothetical, and only modal auxiliaries can carry a remote meaning in a main clause, you are forced (obligated) to use one.

I could/would/should/might do it if they offered to pay me. (remote conditional construction)

Without the modal auxiliary, the main clause collapses into a past-tense statement of fact:

I did it if they offered me to pay (open conditional, offered referring to past time)

This means: “I used to do it provided/whenever they offered to pay me”. It can’t be understood with “offered” as remote preterite.

Important

The construction pairs a modal preterite in the subordinate clause to establish the hypothetical event, with a modal auxiliary in the main clause to express the resulting modality.

Futurity and modality

Locating a situation in future time:

Sam will be here soon. (will, most common)

I shall carry on no matter what happens (shall, rare)

Note

“Shall” survives in interrogatives, expressing a suggestion:

Shall I open it for you?

Where, in a sense, the opening will happen in the future if the suggestion is accepted.

Everything is going to be okay (be going + to, an idiom)

Note

Be + going idiom

The progressive auxiliary “be” and the lexical verb “go”, the latter has lost its motion meaning and the idiom serves merely to signal futurity – typically the near future. The idiom has to be followed by an in nitival complement with to.

“gonna” is phonological merging of “going” and “to”.

The train is about to leave (about + to, adjective, reference to the immediate future)

Note

“about to leave” looks similar to a raising constructions but is slightly different.

The train + is + about to leave (“about to leave” is a to-infinitival complement of the adjective about, in the construction be about to + verb).

The meaning is: “The event of the train leaving is imminent.”

The subject relation is:

  • The train is the subject of leave.

  • The construction “be about to” contributes the temporal meaning “imminent”.

The train [is about [the train to leave]] (=”The train’s leaving is about to happen.”)

This is similar to raising in one important respect: about to does not assign a semantic role to the train.

The train gets its role from leave.

However, “be about to” is not normally analysed as a raising verb like seem.

The reason is that about is not a verb taking an infinitival complement; it is part of a fixed adjectival construction (be about to). The construction as a whole behaves like a modal/aspectual marker of imminence rather than an ordinary raising predicate.

See also: Raising

I am seeing him tonight (present progressive, futurate)

The train leaves in ten minutes (simple present tense, futurate)

English does not have a future tense

Traditional grammar takes the will and shall to be future-tense auxiliaries: English is said to have a tense system contrasting past, present, and future.

The view taken here, however, in a major departure of much modern work from traditional grammars, is that there is no future tense in English.

“Will” and “shall” belong to the modal system, not the tense system.

Our argument is presented in three steps, focusing primarily on will.

“Will” doesn’t contrast with past and present

"Will" is not mutually exclusive to past and present

The auxiliary “will” is itself a verb which inflects for present and past tense: “will” is the present tense form and “would” the past tense form.

Note

The traditional account has “will take” contrasting with “takes” and “took” as future vs present vs past.

She has only just entered the Parliament, but in a few years, she will become a minister (present tense, contrast between becoming a minister in the future relative to now)

She had only just entered the Parliament, but in a few years, she would become a minister (past tense, relative to a time in the past)

I have already asked him, but he won’t help (present tense, contrast is between not helping because of his present unwillingness)

I had already asked him, but he wouldn’t help (past tense, because of his past unwillingness)

Note

The preterite would is here used to indicate past time (not modal remoteness, not backshift)

Auxiliary “will” is not mutually exclusive with the present and past tense inflections: it can combine with either of them.

Treating “will” as a future tense would require to analyze tense on two systems: present/non-present, and future/non-future, leading to clashes, e.g. “future on one system, and present on the other system”.

Note

Difference in the temporal viewpoint.

Present + future relative:

She has only just entered Parliament, but in a few years, she will become a minister. (she entered Parliament very recently, relative to now. Her becoming a minister is presented as future relative to now)

Past + future relative at a past point:

She had only just entered Parliament, but in a few years, she would become a minister. (at some past point, she had entered Parliament only shortly beforehand. Her becoming a minister was still future relative to that past point)

  1. Present viewpoint: The temporal centre is now

  • has entered = anterior to the present

  • only just = a very short time before now

  • will become = future relative to now

“She entered Parliament very recently, and I predict that a few years from now she will become a minister.”

  1. Past viewpoint: The temporal centre has shifted into the past:

  • had entered = anterior to some past reference point

  • only just = a very short time before that past point

  • would become = future relative to that past point

“At that time, she was a newcomer to Parliament. A few years later, she became a minister.”

The entire passage viewed from the past. The two finitie declerative clauses were shifted to the past:

She had only just…

she would become a minister…

“Would” is future-in-the-past.

  1. Mixed alternative

No additional context is required:

When I met her last year, she had only just entered Parliament, but in a few years, she will become a minister.

This can be perfectly grammatical and coherent:

“At the past time when I met her, her entry was very recent; from the speaker’s present viewpoint, her becoming a minister is still in the future.”

The other way around does not make sense on it’s own:

When I met her last year, she has only just entered Parliament, but in a few years, she would become a minister.

Now the interpretation is ambiguous. We don’t know if “would” is interpreted as:

  • future-in-the-past

  • remote/conditional

Note: What happens is a discourse-level temporal interpretation. Both clauses are coordinated and hence independent, yet the first establishes a salient past viewpoint, and the second clause naturally continues from it.

We need to add addtional context to support unambigous interpretation.

When we past viewpoint, not from “has…entered” itself, (since present perfect is anchored in the present) would is interpreted straightforward as future-in-the-past:

She has only just entered Parliament, but I predicted last year that in a few years she would become a minister. (would interpreted as future-in-the-past)

  • has entered → present viewpoint

  • predicted last year → establishes a past viewpoint

  • would become → future relative to that past viewpoint

However, a conditional is more natural, since in a sentence with a present-oriented first clause, a remote conditional reading is often the most readily available way to license “would”:

She has only just entered Parliament, but if everything went according to plan, in a few years she would become a minister. (would is interpreted as expressing modal remoteness)

“She has entered Parliament very recently, and under the imagined condition that everything goes according to plan, she becomes a minister a few years from now.”

The conditional does not make “only just” irrelevant, it merely makes the future outcome less categorically asserted.

Intrinsic connection between futurity and modality

Future is always a prediction, not a fact

The key difference between the future on one hand and the past and present on the other concerns our knowledge of them: In general, we can’t make strictly factual statements about the future, but we generally can about the past and present.

She won in under and hour (fact)

She will win in under an hour (prediction)

He likes you (fact)

He will like you (prediction)

Important

You can calibrate a prediction, or have a confident prediction - but it always remains a prediction. Not a fact. There are languages that have genuine future tense - but English indicate future time in a radically different way than past and present

“Factual” statements about the future:

Fred will be 21 on Friday

Fred is 21 on Friday

But present tense for future is limited:

Australia meets Sweden in the Davis Cup final in December. (future time, finalists already been determined, assurance)

Australia will meet Sweden in the Davis Cup final in December (future time, could be used to make prediction, “will” adds modality)

“Will” is similar to can , may, and must

Close resemblence between will and the undisputed modals

  • Will has no secondary inflectional form:

I expect to will work late (no plain form)

I am willing work late tonight (no gerund-participle)

Note

I am working late tonight.

I am staying late tonight.

I am leaving late tonight.

I am studying late tonight.

All prove the “will” is defective.

Note that in:

I am willing to work late tonight.

“willing” is an adjective, that takes to-infinitival complement

I had would work hard that night. (no past-participle)

Note

I had helped work hard that night.

With an intervening object, many more verbs work:

I had made him work hard that night.

I had let him work hard that night.

I had seen him work hard that night.

I had had him work hard that night.

  • “Will” has no distinct 3rd person singular present tense form: the form “will” occurs with all type of subjects:

He wills go

  • “Will” occurs with a bare infinitival complement:

I will/shell work late tonight

  • “Will” can occur with a main clause of a remote conditional construction:

If you had told me earlier, I could/might/would have helped them. (“will” belongs with “can” and “may” in being able to satisfy the restrictive requirement on the form of the main clause in a remote conditional)

If you had told me earlier, I had helped them (does not satisfy and is hence ungrammatical, “had helped” fails to license the ordinary remote-conditional interpretation)

Usage of “will” in contemporary English

Futurity and modality

Grammatically, ”Will” belongs to the same grammatical subclass as “can”, “may”, “must” , etc.: the modal auxiliaries

That said, most frequent use of will is to indicate futurity.

Note

It applies also to “shall”, but it is becoming very rare in contemporary English.

But “will” can be used about the past and the present with a very clearly modal meaning;

She left Paris yesterday (factual)

She will have left Paris yesterday (confident assumption)

That is a plumber (factual)

That will be a plumber (confident assumption)

Note

Very similar to:

She must have left Paris

That must be the plumber

“Will” provides a different modality interpertation.

Note

“Be” is the plain-form copular verb licensed by the modal

“Will” often conveys volition or intention - a kind of dynamic modality:

I will do everything I can to help you (not just stating what will happen but making a commitment)

I’v already asked him, but he won’t (refusal)

I had already asked him, but he wouldn’t (refusal)

The dog will keep digging in the flower beds (meaning the dog *intentionally digs, this is a comment about the dog’s habitual volitional activity, not making a prediction)

Note

Usage of “shall”.

“Shall” was formerly common for expressing future time reference with 1st person subjects:

I shall be there in a minute

But has been falling out of use for many decades.

However, it is still found in a few cases where the meaning is clearly modal:

One more coat of paint should be enough. (here “should” is a modal use of the preterite of shall, expressing epistemic probability, the expectation that if things are as they ought to be, one more coat will suffice)

Shall we give it another try? (interrogative “shall”, commonly used to ask the addressee to decide what we or I should do)

Shall is also found in very formal writing, in contexts such as legal documents to decree what is required:

The chair shall call a general meeting of the membership once a year.

Irrealis “were”

"Were" as "be" preterite on the 1st, 3rd person singular

Some meaning distinctions that once had special in ected forms are now conveyed in part by tense:

He loved Ty (preterite)

He loves Ty (present tense)

If he loved Ty, he’d go (preterite tense, modal preterite, more remote possibility, use of the preterite where the meaning has to do with modality, not time)

If he loves Ty, he’ll go (present tense, modality: open, more likely)

But something special can happen when the verb concerned is "be" and the subject is 1st or 3rd person singular:

He was in love with Ty (preterite tense)

He is in love with Ty (present tense)

If he was in love with Ty, he’d go (preterite tense)

If he’s in love with Ty, he’ll go

Now there is another possibility: The form were can be used.
It expresses the same modality, remoteness from the facts, but with a different inflectional form.

If he were in love with Ty, he’d go (irrealis use)

If he is in love with Ty, he’ll go (present tense)

Irrealis use: When the verb lexeme is “be”, the modal meaning can be expressed by a different form from the temporal meaning, instead of being different meanings of the form “was”.

Important

“Be” is highly exceptional in this respect: no other verb in the language has a special form distinct from the preterite to express the remoteness modality.

And even “be” only has the special irrealis form in the 1st and 3rd person singular, because in the 2nd person, the form is always were regardless of the sense.

This is a tiny relic of an earlier system, and it may be slowly fading away.

Note

“Be” is by far the most irregular verb in the entire English vocabulary

  • Having three different agreement forms in the present tense, instead of the usual two

  • Two different preterite forms, instead of one

“Be” primary and secondary forms

        flowchart TD
    PRIMARY["PRIMARY FORMS"]

    PRIMARY --> NEUTRAL["NEUTRAL"]
    PRIMARY --> NEGATIVE["NEGATIVE"]

    NEUTRAL --> NPRES["Present"]
    NEUTRAL --> NPRET["Preterite"]
    NEUTRAL --> NIRR["Irrealis"]

    NPRES --> NAM["1st sg: am"]
    NPRES --> NIS["3rd sg: is"]
    NPRES --> NARE["Other: are"]

    NPRET --> NWAS["1st & 3rd sg: was"]
    NPRET --> NWERE["Other: were"]

    NIRR --> NIRRWERE["were"]

    NEGATIVE --> GPRES["Present"]
    NEGATIVE --> GPRET["Preterite"]
    NEGATIVE --> GIRR["Irrealis"]

    GPRES --> GDASH["1st sg: —"]
    GPRES --> GISNT["3rd sg: isn't"]
    GPRES --> GARENT["Other: aren't"]

    GPRET --> GWASNT["1st & 3rd sg: wasn't"]
    GPRET --> GWERENT["Other: weren't"]

    GIRR --> GIRRWERENT["weren't"]

    
        flowchart TD
    SECONDARY["SECONDARY FORMS"]

    SECONDARY --> PLAIN["Plain form: be"]
    SECONDARY --> PASTPART["Past participle: been"]
    SECONDARY --> GERPART["Gerund-participle: being"]

  • The form we’d expect for the 1st person singular negative, amn’t, occurs only in certain regional BrE dialects (in Scotland and Northern Ireland); most speakers use the non-3rd-person singular form aren’t in clauses with subject-auxiliary inversion:

Aren’t I clever?

  • The form !ain’t has survived from earlier regional dialects but is notoriously a non-standard negative form.

  • Irrealis forms are included with the primary forms: they are the ones that have a neutral vs negative contrast:

If I were involved

If I weren’t involved

Note

Irrealis “were” does not make reference to past time, and doesn’t occur as a backshifted counter part of “be”.

It’s vital that he be told. (subjunctive construction)

It was vital that he were told. (not a counterpart, hence the irrealis interpretation)

  • The non-negative forms are called neutral rather than “positive”, because they occur in both positive and negative clauses:

That is true

That is not true

See also

No future (grammaticalisation)

Distance (grammaticalisation)