A passage from Walter Mosley’s Devil in a Blue Dress:
Joppy stopped wiping [the bar] for a moment and looked me in the eye.
“Don’t get me wrong, Ease. DeWitt is a tough man, and he runs in bad company. But you still might could get that mortgage payment an’ you might even learn sumpin’ from’im.”
That ‘might could get’ was a serendipitous phrase to encounter. Over the preceding days I’d come across several treatments of what are known as double modals or multiple modals, and had been considering a blog post about them. Hint taken.
Brad Dourif in ‘Deadwood’
First, a technical note on modals. These are a small and grammatically unusual family of verbs. They’re a subset of the auxiliary (helper) verbs and so are sometimes called modal auxiliaries. They qualify other verbs in a verb phrase, influencing the overall meaning: I can go, you may be, she must try. Geoffrey Pullum says there are 8–12 of them in English:
can, may, shall, will, dare, must, need, ought
He and Rodney Huddleston mention could, might, should and would as the preterite forms (past tense marked by inflection) of the first four. Grammarians differ slightly in naming the family members; this depends on the category boundaries, and needn’t concern us here.
Modals are used to indicate modality, or ‘mood’ – not in the sense of atmosphere, but to express possibility, permission, obligation, necessity, deduction, prediction and such things. Heather Marie Kosur writes that modality ‘allows language users to express what is, what would be, what may be, and what should be’.
Modern grammar generally divides modality into two or three branches: epistemic (probability, deduction, necessity) and deontic (duty, obligation, permission), and sometimes also dynamic (factual). See this glossary, or Kosur’s essay for a more detailed treatment.
Unlike lexical verbs, modals have no to-infinitives, no –s forms for subject agreement, and no tenses formed with be or have. So you don’t see oughting, mights or musted, etc. At least, not normally (James Joyce, Finnegans Wake: ‘when cherries next come back to Ealing as come they must, as they musted in their past’).
…
*
And so to double or multiple modals: might could, may would and the like.
Megan Risdal, in a recent post at For the Love of Linguistics, used a map of ‘might could’ usage to gauge its geographic distribution in the U.S. She also studied the reactions double modals inspire, and shared her thoughtful observations.
As I wrote in a comment there, double modals are not in my idiolect, but I find them charming. They’re also interesting grammatically, semantically, and sociolinguistically. They may be used with subtlety by those to whom they come naturally: to modify the degree of likelihood or speculation expressed, for example.
Multiple modals also popped up in an article on the influence of Scotch-Irish [PDF] on East Tennessee grammar, which John Cowan shared in a comment to my recent post on Hiberno-English till. The article’s author, Michael Montgomery, is one of the people behind MultiMo: The Database of Multiple Modals, which launched last week.
MultiMo offers, among other things, a multi-page table of reported examples, including some rare and delightful triple modals:
I might could should write home.
It’s a long way and he might will can’t come, but I’m gonna ask.
Aren’t they amazing? What is grammatical in standard English is often erroneously equated with what is grammatical, period. But grammaticality differs with dialect, and standard English is just one dialect (or a set of them) — privileged socially but not linguistically.
If you’re still with me, and you might would be hungry for more, Language Log has analysed double modals on several occasions; for starters see this post by Ben Zimmer and the pages it links to.
I’ll conclude as I began, with Devil in a Blue Dress:
I always tried to speak proper English in my life, the kind of English they taught in school, but I found over the years that I could only truly express myself in the natural, “uneducated” dialect of my upbringing.
Update: More discussion of double modals at Language Hat, who says:
They are a peripheral part of my dialect thanks to my Ozark ancestors, and while I don’t use them on a daily basis, I delight in tossing them into the mix once in a while; they give me that warm down-home feeling.
Sentence first 
Stan:
One of the popular triple modals going around over here is “shoulda, coulda, woulda,” where you might have done something, but didn’t.
An interesting thing about U.S. double modals is that they aren’t productive: indeed, they tend to fall off as you move from east to west, with just one or two left in Arizona, the extreme westernmost point. Even in the East, might could is far more common than may can, and may will is I think unknown.
Marc: I decided not to include that (or any version of it) because I wouldn’t say it’s a genuine triple modal: it seems more like an abbreviated “[I] should have, could have, and would have…”
John: Right. There does appear to be huge variation in the frequency of different double modals. It’s a pity MultiMo disappeared yesterday; I had time to go through the table only briefly. I expect DARE will have lots of information on them.
EDIT: Joe Salmon posted a table showing double modal combinations in North and South Carolina Corpus.
The ‘musted’ in Finnegans Wake (which is fairly conventional grammatically) is more likely to be a past tense of a verbal form of ‘must’ in the sense of pulp in preparation for alcoholic fermentation (OED must n1, 2), hence the ‘pressing season’.
Vincent: Yes, that’s probably the primary reading — but the modal sense of the word seems to be intended too, for the fun of it.
Agreed; as long as we bear in mind that the past tense here is more boozy than modal and that FW, although almost completely derived from conventional texts, is a minefield for philologists.
Though, interestingly, Joyce’s own note, on which this is based, reads “as it was musted / as it must at / the moment, as / hereinafter / it will must ^must will^”. No mention of cherries here. This was probably based on some book or article he was reading in late 1925, which has not so far been identified. The cherries are based on a completely separate note, taken in 1923. Intentionalism may be on your side!
Does the phrase “must needs” fall into this category?
Vincent: Thanks for looking into this, and for reporting back: it is must kind of you! As for FW being a “minefield for philologists”, yes indeed. I believe Joyce had a copy of Skeat’s etymological dictionary by his side as he wrote, but I don’t know if that was confirmed or just inferred.
Charles: I don’t think so. This note suggests it does, but an article at Random House says the needs in the phrase is not a verb (or noun) but an adverb. A Google Books search shows the phrase falling in frequency since the early/mid-19th century, and COHA shows the same, but I come across it fairly regularly, usually in oldish books.
The eponymous Stephen Hero reads Skeat “by the hour”. Burchfield conjectured that Joyce may have taken unusual words from the OED. But by the time he was working on Ulysses Joyce had moved from reference books to books in general (fiction and non-fiction) and newspapers, reading them in much the same way as lexicographers do. So he did his own fieldwork. The OED was of course only available in fascicles and was not completed until six years after the publication of Ulysses. Joyce did not have copies of any of the fascicles, nor did he ever own the OED.
Double/multiple modals were something of a legend in my linguistics classes in college (I went to school in Boston, Massachusetts), but I had never heard anyone use them in real life. Then I moved to Maryland after graduation–and I hear them all the time! The first time I heard it I did a mental double-take and smiled quietly to myself, thinking “Yes! I’ve finally heard it for real!” But now it’s become a common occurrence and I’ve even heard them slip out of my own mouth once or twice.
Another one I had never heard before moving here (but was also often discussed in linguistics classes) was the deletion of the verb “to be” in phrases like “needs to be washed”–thus, it becomes “needs washed,” as in, “The car needs washed.” I’d heard “The car needs washing,” but never “needs washed.” And then the other day at work, a coworker sent an email saying that some of the things in the document she was sending “needed changed” and that she had taken care of it. Score another one for new linguistic experiences! Yay!
That’s a charming report, Bridget: thank you. It’s striking how an idiom so rare and all but mythical in one part of the world can be ubiquitous and second nature in another. Needs washed isn’t part of my normal speech either, but I’m tempted to try it out! You might enjoy this post on the phrase at Language Log.
In my experience, the more rural, and eastern a locality is, the more likely you are to encounter an American using redundant modals. It’s still pretty strong in rural southeast localities.
@sundowniest: The modals are not “redundant”. They don’t duplicate each other, but are simply a different way of expressing more than one modality. Unless you think “perhaps you should” is redundant, of course, that being what “might should” means. “Might could” is simply “might be able to”. We who use them – and I grew up with them (in East Tennessee) – find them useful for shading modalities, though we learn not to use them in formal speech or writing.
The Ridger-
I called them redundant because they could be omitted and we would still take the meaning. But,if used to shade modalities, it’s not at all clear to someone outside of the dialect that that’s what’s going on.
That’s probably true. But it seems clear to those of us who do that the meanings of the modals aren’t the same (might =/= could, for instance, or will).
@Bridget: Need + past participle is a pattern in Scots that was transplanted to parts of the US. It’s stereotypically Pennsylvanian, though many in Maryland, Ohio, Kentucky, and elsewhere also use it – it’s considered part of the North Midland dialect. (http://www.jstor.org/pss/455549 links to Murray, Frazer, & SImon 1996 discussing the construction, but it’s at JSTOR and probably not fully available)
[…] featured on this blog before: in a post about double or multiple modals, I quoted the narrator of Devil in a Blue Dress (Easy Rawlins again) saying he “tried to speak […]
[…] The last one also serves as a nice examples of a double modal.] […]
I know this is an old thread, but might all the ”might will” examples be mishearings of “might well” examples?
I don’t think so. (Comments on old threads are welcome, by the way.)
I’m a double-modal speaker. I say both of those, but “might will” is what I’m saying when I say it. It doesn’t mean the same thing; it means “maybe I will”.
[…] might could know’), Fear Itself by Walter Mosley – who also featured in my earlier post on multiple modals – but I never hear them spoken in […]