He must can parse

« previous post | next post »

From an interview with a high-school pitching prospect at a Milwaukee Brewers' fan site, "BCB Interview: 26th-round LHP Lex Rutledge", 6/12/2009":

BCB: So, speaking of football and late-round baseball picks, did you hear about this Florida State defensive tackle recruit the Brewers drafted? Jacobbi McDaniel. He's a 285-pound third baseman.
LR: [laughs] No, I didn’t. Dang. Does he even play baseball anymore?
BCB: He said he wants 1.5 million to sign, and now the Noles fans are freaking out because there's a report the Brewers offered him 800k.
LR: Wow. I wish they would offer me that. He must can hit.
BCB: Yeah, no kidding. Gain some weight and become a five-star DT recruit and you can make the big bucks.
LR: [laughs] I just don’t see that happening. Oh well, maybe I can throw 103 and get the big bucks like Strasburg.

There are two links in this passage. The first one is to a note on an FSU fan site, about whether Jacobbi McDaniel plans to play baseball or football. The second one is to a Language Log post by Geoff Pullum, "Do double modals really exist?", 11/20/2007.

In that post, Geoff observed that of 64 logically possible two-modal combinations of the eight most basic modals, very few are actually attested, and most of these are consistent with the hypothesis that might and may are sometimes re-analyzed as adverbs. The quote "He must can hit" adds another combination to the list of attested double modals that Geoff (and others) came up with, and thus bears on an interesting question of syntactic analysis.

The pseudonymous author of the Lex Rutledge interview, "battlekow", is clearly a Language Log reader, who demonstrates that despite the widespread signs of ignorance and confusion about linguistic description, you can find examples of well-informed linguistic commentary in some unlikely-seeming places. In this case, the commentary is nothing but a hyperlink — but that's all that was needed.

[I noticed the link in battlekow's interview, by the way, because our referrer log shows that some 60-odd readers have clicked on it. And I should add that I've always thought that double modals exist, so I was glad to see this one, and feel that the others ought can be found somewhere…]


  1. Sili said,

    June 13, 2009 @ 7:30 am

    That is certainly a pretty clear example. It seems obvious to me that in this case "can" = "be able to", but the latter is of course not a modal.

    I guess this is off topic – let me know is self-immlation is called for – but how does this relate to "needs must" which seems a set phrase, but I'm never quite sure how to analyse it.

  2. scratchdaddy said,

    June 13, 2009 @ 7:37 am

    "He must can hit." I like that. Makes me smile. Works so much better than, "He must be able to hit."

    I'm Southern, and "might could" is definitely part of my usage. It occurs as a conditional I hope or intend to fulfill, and as with this "must can," I usually intend it to be jocular.

  3. marie-lucie said,

    June 13, 2009 @ 8:32 am

    It seems to me that "must" is being used as an adverb here, as in GKP's interpretation of such constructions.

  4. Trond Engen said,

    June 13, 2009 @ 8:50 am

    It seems to me that 'can' is a back-formed infinitive, as Sili's parallel to 'be able to' suggests. But then, we are both Scandinavian.

  5. Mark F. said,

    June 13, 2009 @ 9:00 am

    I'm pretty sure double modals really are double modals, at least in my idiolect. I have

    You might not can do that.

    but not

    * You definitely not can do that.

    As Geoff noted, it is far from being a fully productive construction, but I have also heard "may can", from my brother.

    Can "used to" be treated as a modal (via some kind of reanalyzing)? I want to analyze "used to could" as a double modal, and "might used to could" as a triple modal.

  6. Sili said,

    June 13, 2009 @ 9:12 am

    I hadn't considered that Danish "kan" might well could interfere with my intuitions (smiley face goes here).

    Rereading Pullum's post I see that he calls "maybe" an adverb of probabilty. If "might" is considered likewise to mean "possibly", then "must" fits nicely in the other end of the spectrum as "(almost) certainly".

  7. Brett said,

    June 13, 2009 @ 9:43 am

    I've seen that particular double modal before, and, as a native English speaker, I have always analyzed it the way Sili initially suggested.

  8. peter said,

    June 13, 2009 @ 9:54 am

    I'm reminded of Thurber's proposed advertising slogan for a brewery faced with falling market share: "We still brew good, like we used to could."

  9. CIngram said,

    June 13, 2009 @ 10:49 am


    In the expression "needs must", I have always assumed that "needs" was a noun, and therefore the subject of the verb structure, which then has no main verb, of course. It is not easy to see how to flesh out the ellipsis though, so I may have this wrong.

    [(myl) The OED says:

    App. originally the impersonal use of must (see MUST v.1 3c) with anaphoric ellipsis of the main verb; by the 19th cent. used in isolation (prob. originally as a shortened form of needs must when the devil drives: see sense 5). Now freq. taken to be a plural noun and verb.

    The cited subentry for must compares French "il me faut" and ME "me oughte", and gives examples like

    1604 E. GRIMESTON tr. True Hist. Siege Ostend 195 We beleeue them no more then needs must.

    But some uses seem hard to reconcile with this analysis. Thus in Shakespeare we find

    the rights of time thou needs must haue
    I an accessary needs must be
    Your presence needs must puzzle Anthony
    They needs must shew themselues
    Thou needs must be englutted
    the plague That needs must light on this Ingratitude
    His passion is so ripe, it needs must breake
    This needs must be a practise

    The lack of agreement ("thou needs must have", "I … needs must be") is consistent with an impersonal analysis, but I'm not clear what the syntactic relations are really supposed to be. Perhaps some reader who knows more about the history can explain. ]

  10. Artifex Amando said,

    June 13, 2009 @ 1:06 pm

    To continue on the Scandinavian line of thought, in Swedish, one can also say "must can" – "måste kunna" – to convey "almost certainly be able to":

    "Han måste kunna slå", would be my translation of the sentence.

  11. Trond Engen said,

    June 13, 2009 @ 1:15 pm

    Thou needs must be englutted

    Isn't 'needs' here the noun 'need' (in some case/noun, preferably the genetive), used as an adverb of necessity?

    Thou necessarily must be engluttted

  12. Trond Engen said,

    June 13, 2009 @ 1:40 pm

    For the sake of obfusity: The Scandinavian infinitive ('kunne'/'kunna') isn't back-formed (or, well, not recently and like that, anyway). The cognate English infinitive might have ought to be 'to cun'.

  13. battlekow said,

    June 13, 2009 @ 8:42 pm


    Thanks for the link. When transcribing the interview, I wasn't sure whether I should preserve the "must can hit" line; I thought the non-standard usage might be distacting. Without my memory of reading about double modals on Language Log, I might have elided that entire portion of the conversation, but I ultimately decided that it added color to the interview. I'm glad you found it and enjoyed it.

    I must say, though, that you missed the most appropriate comment of mine here.


  14. Mark F. said,

    June 14, 2009 @ 12:30 am

    I don't think it works just to say that "can" is an infinitive in this construction, because it's modifying another verb, and infinitives can't generally do that. "Come" and "go" are exceptions, but they're a small enough group not to form a basis on which to reanalyze other words.

    Also, you can't generally put "to" in front of it. You can't say

    * He would like to can hit.

    (But you can say

    He used to could hit

    This is part of why I think the phrase "used to" is reanalyzed as a modal by people who use double modals.)

    My guess is that, in double-modalese, the rule that modals take only the infinitive form of the verb is broadened to permit modals to take other modals too in certain cases. Those cases are relatively few, but provide a template that can be generalized, with some accompanying feeling of stretching things. "Must can" definitely feels like a stretch to me.

    Scratchdaddy says he usually uses double modals in jocular contexts. That's not so for me; they're an entirely natural part of my informal speech. They look wrong to me in writing, though.

  15. Peter-Arno Coppen said,

    June 14, 2009 @ 5:18 am

    In Dutch, these kind of constructions are pretty common. In fact, the typically Dutch elliptic expression 'moet kunnen' (lit. 'must can', meaning "should be possible") is surely a combination of two epistemic modals. It is impossible to interpret 'kunnen' ('can') as "to be able to." But more elaborate combinations are also possible, like 'zou moeten kunnen' (lit. 'should must can,' meaning "should actually be possible," or something like that, it's hard to express the intricacies of the Dutch mind in English).

    I think this is also possible in German. Isn't this also a possibility of the American "Pennsylvania Dutch"?

  16. Theo Vosse said,

    June 14, 2009 @ 8:02 am

    Couldn't this be a simple speech error? Perhaps "LR" was just mixing up phrases like "He must be able to hit" and "He surely can hit". Performance errors are not uncommon enough to warrant deduction of a new structure on one or two examples.

    [(myl) There's a good-sized literature on the history and current distribution of multiple-modal constructions in English. Geoff Pullum's 2007 post frankly asserts that he has read only a little of this — his point in that post was just that the commonly-presented examples (like "might could") are limited to a small fraction of the many possible combinations, and thus perhaps should analyzed in a different way. I know the relevant literature even less well than he does. But a bit of web search easily turns up some additional evidence that "must can" is one of the permitted double modals in some varieties of English. Thus Gradol, Leith and Swann, English: history, diversity, and change, p. 252:

    … there is a rule of Standard English that only one modal verb can appear in a single verb phrase. Thus 'He must do it' is grammatical whilst '*He must can do it' is not. Indeed Standard English has developed a whole battery of 'quasi-modal' verbs to 'stand in' for modals where the meaning requires them but the above rule forbids them. […]

    In Tyneside English, the rule inhibiting double modals does not apply so long as the second modal is can or could. Thus the asterisked sentence would conform to the rules of Tyneside English. These double modals are also found in Scots and some American dialects, but more combinations of modals are allowed in these dialects than in Tyneside. Furthermore, more combinations are allowed in the dialects of rural Northumberland than in those of urban Tyneside. For instance, the combination of would and could appears in the urban area — but only in a negative form — whilst in rural Northumberland the positive form may be found. Examples from McDonald (1981, pp. 186-187) are:

    (3) He wouldn't could've worked, even if you had asked him. (Tyneside)
    (4) A good machine clipper would could do it in half a day. (Northumberland)

    (The cited reference is Christine McDonald (1981), 'Variation in the use of modal verbs with special reference to Tyneside English', University of Newcastle: unpublished PhD thesis.)

    Kortmann and Schneider, A handbook of varieties of English, p. 127, gives this additional double-modal example from McDonald:

    (73) The girls usually make me some [toasted sandwiches) but they mustn't could have have made any today.

    and this one from the NECTE corpus:

    (76) You'll probably not can remember, but during the war there wasn't wool.

    Kormann and Schneider observe that such examples are "rare and probably recessive" (at least in Tyneside), perhaps "due to the fact that the need to use them only arises in certain circumstances", like speculating about baseball draft rumors. (Well, they don't mention that last part explicitly.)

    We've seen a similar geographic distribution before, in the discussion of the traditional distribution of uptalk, where I quoted Daniel Hirst's theory that it's due to the influence of "intense raiding and settlement by Norwegian Vikings in the early 9th century".

    So anyhow, it seems likely that Mr. Rutledge's phrase is a legitimate regional variant — he's from Tupelo, Mississippi — rather than a speech error. ]

  17. Rick said,

    June 14, 2009 @ 6:52 pm

    My understanding from a Middle English class long ago is that "needs" is a fossilized genitive. It is thus equivalent in meaning to the phrase "of necessity."

  18. Yr Funny Valentine – Linkages said,

    June 17, 2009 @ 1:33 pm

    […] must can do something. […]

  19. Christian R. Conrad said,

    June 22, 2009 @ 8:45 am

    Trond Engen said "The cognate English infinitive might have ought to be 'to cun'."

    Or, at least in the North (as in, say, Scotland), "to ken".

  20. Christian R. Conrad said,

    June 22, 2009 @ 9:00 am

    Mark F. said: "I don't think it works just to say that "can" is an infinitive in this construction, because it's modifying another verb, and infinitives can't generally do that.

    But you're reading it backwards — "can" is *being modified by* "must", not modifying it! And yes, it's usually infinitives that are modified by "must", isn't it? (Like in "He must swim", "She must eat", etc; though not "He must swum", "She must ate".)

    It's because "can" in English *doesn't (usually) have* an infinitive that the correct way of saying that is (usually regarded as) "He must _be able to_ hit", where "be" is definitely infinitive, and the whole phrase "be able to" is standing in for "can".

    Mark F: "Also, you can't generally put "to" in front of it. You can't say
    * He would like to can hit."

    That's precisely the thing — if "can" (or "cun", "ken", or whatever form of it) were an infinitive, then you could.

    Mark F: "(But you can say
    He used to could hit"

    In some socio- and dialects, but not all. Definitely not in the English English I learned in Sweden in the mid- to late 1970s (before it went to H in a HB, from watching too many American films, and later interacting with too many Americans on these here Intertube thingies.) Anything other than "He used to be able to hit" would have gotten short shrift from our teacher.

  21. Mar Rojo said,

    February 17, 2012 @ 10:59 am

    Late in the day, but why not read:

    Volitional modality in the double-modal construction in Southern US English

    Cynthia Kilpatrick and Chris Barker, UCSD

  22. Mar Rojo said,

    February 17, 2012 @ 11:05 am

    "the higher member of a modal pair is either might or may, while the lower can be one of several choices, as shown in (22):
    Higher modal – Lower modal
    may *may
    might *might
    *must *must
    *shall *shall
    *will ?will
    *can ?can
    *could could
    *would would
    *should should

    This is, for the most part, a case of complementary distribution. The modals may and might, which are the only options for the higher modal position, are never possible as the lower modal."

    Cynthia Kilpatrick and Chris Barker.

RSS feed for comments on this post