Justice Breyer, Professor Austin, and the Meaning of 'Any'

« previous post | next post »

[Tap tap. Is this thing on?

I guess as a freshly minted Language Logger, I should introduce myself: I am a professor of linguistics at MIT. I work on meaning: semantics, pragmatics, philosophy of language, and the intersections thereof. I am also a part-time denizen of the academic Dark Side, as Associate Dean of MIT's School of Humanities, Arts, and Social Sciences. I blog on academic, geeky, abstruse, and personal aspects of my life on my personal blog "semantics etc.". Together with fellow LanguageLogger David Beaver, I co-edit the new kid on the block journal Semantics and Pragmatics, for which we maintain an editors' blog as well. So what's one more gig, right?

In this post, I partially recycle some notes that appeared on my blog in 2005. But there is a new angle.]

In a recent interview, Supreme Court Justice Breyer lists the five books that have influenced his thinking the most. Among them: J.L. Austin's How to Do Things with Words. Breyer says:

JL Austin was an ordinary language philosopher. When I studied in Oxford, I went to one of his classes and I read his books. How to Do Things with Words teaches us a lot about how ordinary language works. It is useful to me as a judge, because it helps me avoid the traps that linguistic imprecision can set. If I had to pick a single thing that I draw from Austin's work it would be that context matters. It enables us to understand, when someone makes a statement, what that statement refers to and what that person meant.

When I see the word "any" in a statute, I immediately know it's unlikely to mean "anything" in the universe. "Any" will have a limitation on it, depending on the context. When my wife says, "there isn't any butter," I understand that she's talking about what is in our refrigerator, not worldwide. We look at context over and over, in life and in law.

Austin suggests that there is good reason to look beyond text to context. Context is very important when you examine a statement or law. A statement made by Congress, under certain formal conditions, becomes a law. Context helps us interpret language, including the language of a statute. Purpose is often an important part of context. So Austin probably encourages me to put more weight on purpose.

It is very interesting that Breyer should choose the word "any" as an example of why context matters. A few years back, there was in fact a Supreme Court decision (Small v. United States) that hinged on the meaning of "any" (pdf of the decision here]). And as it turns out, Justice Breyer wrote the decision for the majority (made up of Breyer, Stevens, O'Connor, Souter, and Ginsburg; ah the good old days).

The background:

Petitioner Small was convicted in a Japanese Court of trying to smuggle firearms and ammunition into that country. He served five years in prison and then returned to the United States, where he bought a gun. Federal authorities subsequently charged Small under 18 U. S. C. §922(g)(1), which forbids "any person … convicted in any court … of a crime punishable by imprisonment for a term exceeding one year … to … possess … any firearm."

Small subsequently argued that any court was not meant to encompass foreign courts, only domestic ones. The Supreme Court agreed.

The arguments in the decision are a good case study of semantics/pragmatics in the real (well, legal) world. Here are some excerpts:

The question before us is whether the statutory reference "convicted in any court" includes a conviction entered in a foreign court. The word "any" considered alone cannot answer this question. In ordinary life, a speaker who says, "I'll see any film," may or may not mean to include films shown in another city.

In law, a legislature that uses the statutory phrase " 'any person' " may or may not mean to include " 'persons' " outside "the jurisdiction of the state." See, e.g., United States v. Palmer, 3 Wheat. 610, 631 (1818) (Marshall, C. J.) ("[G]eneral words," such as the word "'any,' " must "be limited" in their application "to those objects to which the legislature intended to apply them"); Nixon v. Missouri Municipal League, 541 U. S. 125, 132 (2004) (" 'any' " means "different things depending upon the setting"); United States v. Alvarez-Sanchez, 511 U. S. 350, 357 (1994) ("[R]espondent errs in placing dispositive weight on the broad statutory reference to 'any' law enforcement officer or agency without considering the rest of the statute"); Middlesex County Sewerage Authority v. National Sea Clammers Assn., 453 U. S. 1, 15-16 (1981) (it is doubtful that the phrase " 'any statute' " includes the very statute in which the words appear); Flora v. United States, 362 U. S. 145, 149 (1960) ("[A]ny sum," while a "catchall" phase, does not "define what it catches"). Thus, even though the word "any" demands a broad interpretation, see, e.g., United States v. Gonzales, 520 U. S. 1, 5 (1997), we must look beyond that word itself.

In determining the scope of the statutory phrase we find help in the "commonsense notion that Congress generally legislates with domestic concerns in mind." Smith v. United States, 507 U. S. 197, 204, n. 5 (1993). This notion has led the Court to adopt the legal presumption that Congress ordinarily intends its statutes to have domestic, not extraterritorial, application. See Foley Bros., Inc. v. Filardo, 336 U. S. 281, 285 (1949); see also Palmer, supra, at 631 ("The words 'any person or persons,' are broad enough to comprehend every human being" but are "limited to cases within the jurisdiction of the state"); EEOC v. Arabian American Oil Co., 499 U. S. 244, 249-251 (1991). That presumption would apply, for example, were we to consider whether this statute prohibits unlawful gun possession abroad as well as domestically.

The statute's language does not suggest any intent to reach beyond domestic convictions. Neither does it mention foreign convictions nor is its subject matter special, say, immigration or terrorism, where one could argue that foreign convictions would seem especially relevant. To the contrary, if read to include foreign convictions, the statute's language creates anomalies.

For example, the statute specifies that predicate crimes include "a misdemeanor crime of domestic violence." 18 U. S. C. §922(g)(9). Again, the language specifies that these predicate crimes include only crimes that are "misdemeanor[s] under Federal or State law." §921(a)(33)(A). If "convicted in any court" refers only to domestic convictions, this language creates no problem. If the phrase also refers to foreign convictions, the language creates an apparently senseless distinction between (covered) domestic relations misdemeanors committed within the United States and (uncovered) domestic relations misdemeanors committed abroad.

The Supreme Court's doctrine therefore seems to be that "any" like other quantifiers can be contextually restricted, that what the restrictions are depends on the intentions of the speaker (here: Congress), and that one can infer the intentions by seeing what interpretations make sense in the context of other utterances in the same text.

What makes "any" so interesting in this context is that there is a tension between the natural tendency of quantifiers to be contextually restricted and the peculiar properties of "any". In a seminal article on the semantics of "any" (unfortunately there's no open access version, here is the JSTOR version which some of you may have access to), Nirit Kadmon and Fred Landman argue that what "any" contributes is a widening of the meaning a sentence might otherwise have. They suggest that the difference between "we don't have bananas" and "we don't have any bananas" is that in the latter case we claim to not even have questionable bananas. Justice Breyer argues in his decision, quite plausibly, that this widening effect has its limits. "Any court" can mean "any court in the US" without being interpreted as widely as "any court anywhere in the world".

By the way, Justices Thomas, Scalia, and Kennedy dissented, saying that the court's decision "institutes the troubling rule that 'any' does not really mean 'any,' but may mean 'some subset of "any,"' even if nothing in the context so indicates" (hmm, three levels of quotation do stress the English punctuation system).

As a professional semanticist, I concur with Breyer and dissent from the dissenters.

[Thanks to Brian Leiter and Brian Weatherson for linking to the Breyer interview.]<



63 Comments

  1. Viktor said,

    July 6, 2011 @ 4:20 am

    Funny: I am a semanticist – but for programming languages (which language log is teaching me is both very different and very similar to natural languages).

    For programming languages all always means every member of a set. But, perhaps this is not so inconsistent, since sets are often left implicit anyhow when they are obvious (or the author of the paper finds them obvious).

    It is also worth noting that if we agree with the dissenters then that would imply for instance that the rulings of courts in Nazi germany, or other dictatorships, are accepted as having some legal power in the US. Which is unlikely to be the intention of congress.

  2. Matt said,

    July 6, 2011 @ 5:01 am

    Well, three levels of quotation certainly stress the *American* punctuation system. But wouldn't that sentence be noticeably easier to read (and write) with the commas outside the quotation marks, UK-style?

  3. My first post on Language Log | Kai von Fintel said,

    July 6, 2011 @ 5:15 am

    […] As hinted yesterday, I have joined the Language Log juggernaut (thanks to Mark Liberman and Geoff Pullum for recruiting me). Here's my first post: Justice Breyer, Professor Austin, and the Meaning of 'Any'. […]

  4. Eric P Smith said,

    July 6, 2011 @ 5:43 am

    Justice Breyer’s words confuse two uses of any that are very different. In the predicate forbids any person convicted in any court of a crime punishable by imprisonment for a term exceeding one year to possess any firearm, both occurrences of any are the non-affirmative anyn, deriving their meaning partly from the implied negative forbids. In the sentence, I’ll see any film, the word any is the free choice anyf, not deriving its meaning from anyn express or implied negative. The two uses are so different that I don’t think the usage of one sheds much light on the usage of the other.

  5. Eric P Smith said,

    July 6, 2011 @ 5:45 am

    Why do my subscripts appear correctly in the preview and incorrectly in the published version?

  6. John Walden said,

    July 6, 2011 @ 6:06 am

    I can see the attraction of "any" having two different usages which don't help each other. But isn't there an underlying usage to them both, either n or f? Something in the way of a challenge, similar to the "not even questionable bananas" mentioned above:

    "There's a shop in Edinburgh where you can buy any brand of whisky"

    I defy you to name a whisky it doesn't sell

    "There's a shop in Edinburgh where you can't buy any brand of whisky"

    I defy you to name a whisky it does sell

  7. richard howland-bolton said,

    July 6, 2011 @ 6:52 am

    I bet a shop in Edinburgh would never sell Yamazaki Malt!

  8. GeorgeW said,

    July 6, 2011 @ 7:53 am

    First, KvF, welcome as a LLer. A good start.

    Second, I would be interested in the interaction of 'any' and 'some.'
    P1 Do you have any money?
    P2 I have some. *I have any.

    Is there a difference between?
    1. A person convicted in some court.
    2. A person convicted in any court.

    Is there a good article that addesses this?

  9. John Walden said,

    July 6, 2011 @ 8:02 am

    It depends on your definition of "never sell". They may never sell any but they list it

    http://www.royalmilewhiskies.com/product.asp?pf_id=0010000027570

  10. Ben Hemmens said,

    July 6, 2011 @ 8:24 am

    GeorgeW, indeed: is "some" always indefinite?

    Do speakers really intend a difference between:

    "So some guy came up and asked me for a light."
    and
    "So this bloke came up and asked me for a light."

    ?

  11. Eric P Smith said,

    July 6, 2011 @ 8:42 am

    Whoever fixed my subscripts: thanks!

  12. Mark Etherton said,

    July 6, 2011 @ 10:01 am

    There's also the classic text from Beyond the Fringe (with Jonathan Miller as Bertrand Russell):

    Russell: One of the advantages of living in Great Court, Trinity I seem to recall, was the fact that one could pop across at any time of the day or night and trap the then young G. E. Moore into a logical falsehood by means of a cunning semantic subterfuge. I recall one occasion with particular vividness. I had popped across and had knocked upon his door. “Come in,” he said. I decided to wait awhile in order to test the validity of his proposition. “Come in,” he said once again. “Very well,” I replied, “if that is in fact truly what you wish.”

    I opened the door accordingly and went in, and there was Moore seated by the fire with a basket upon his knees. “Moore,” I said, “do you have any apples in that basket?” “No,” he replied, and smiled seraphically, as was his wont. I decided to try a different logical tack. “Moore,” I said, “do you then have some apples in that basket?” “No,” he replied, leaving me in a logical cleft stick from which I had but one way out. “Moore,” I said, “do you then have apples in that basket?” “Yes,” he replied. And from that day forth, we remained the very closest of friends.

  13. Emily said,

    July 6, 2011 @ 10:07 am

    @John Walden: I think that "any" is functioning like the universal quantifier. "There isn't a brand of whiskey you can't buy at that store" is logically equivalent to "For all brands of whiskey, you can buy them at that store". Of course there are pragmatic constraints on the extension of "all", though, and I think those utterances do implicate a challenge.

  14. Jonathan said,

    July 6, 2011 @ 10:54 am

    If I might agree with the dissenters — sure, there are contexts which delimit the meaning of "any." But for the Court to pick them out gives way too much scope to the Court to interpret the poor drafting of Congress. The Court surely must come to some decision, and it shouldn't come to an absurd one, but if Congress said "any," and meant "any," then "any" it is. And if Congress didn't mean "any," they shouldn't have said so. Congress is certainly free to redraft and clarify a law, but giving contextual meaning to "any" can create uncertainty about what the law is… and that's bad. Why not simply interpret the law as drafted and then let Congress fix it if they drafted the law poorly?

  15. KevinM said,

    July 6, 2011 @ 11:20 am

    @Matt. I take your point about punctuation. But that still leaves us to wonder why Breyer would be so influenced by Austin if he's just an "ordinary language philosopher"? :)

  16. slobone said,

    July 6, 2011 @ 12:25 pm

    Sorry to say I don't quite understand the point about the bananas. "We don't have bananas" is not terribly idiomatic in American English, but if I heard it, I would assume the speaker meant "we don't carry bananas." Whereas "we don't have any bananas" would probably mean "we usually have bananas, but at the moment we're sold out." And what is a questionable banana?

  17. Sivi said,

    July 6, 2011 @ 1:05 pm

    @Slobone

    A questionable banana would be one which could, theoretically, be found in a crate in the back? The theoretical speaker wishes to assure the customer than they really do not have ANY bananas, even in back, so they'd better not expect the shopkeeper to go looking for them.

  18. Dan Hemmens said,

    July 6, 2011 @ 1:08 pm

    Non-linguist question. Is this distinction really about the definition of "any" or is it about the definition of whatever word is *qualified* by "any". As in, is this really about the meaning of the word "any" in "any court" or is it about the meaning of the word "court" in "any court"?

    To work with the example of the Royal Mile Whiskey Shop, while they may stock Yamazaki Malt, but they presumably don't stock – for example – brands of whiskey which nobody has made for a hundred years, or brands which don't exist yet, or purely fictional brands. They don't – for example – stock "Urbon" the trial whiskey brand which Stella English created for the final task on the 2010 series of /The Apprentice/ (which as far as I understand was never actually produced past the prototype stage).

    Of course in this context it's reasonable to argue that fictional brands of whiskey or brands of whiskey that were created for an episode of a TV show don't really count as "brands of whiskey" and therefore aren't included in "any brand of whiskey". At least, that's the way I'd see it, the issue isn't that "any X" – depending on context – can refer to a subset of X but rather that "any X" can take a narrow definition of X (for example, "any brand of whiskey" can define "brand of whiskey" as referring only to brands of whiskey which actually exist).

  19. marc said,

    July 6, 2011 @ 1:21 pm

    If anything, this ruling is an interesting example of how the the liberal and conservative halves of the court could occasionally switch sides. Here we have the liberals supporting gun ownership and condemning the application of extraterritorial statutes (and judgments) to domestic cases. Considering that Breyer, the internationalist par excellence of the court, was part of this opinion, this is quite unusual indeed.

  20. Ellen K. said,

    July 6, 2011 @ 1:41 pm

    Two things,

    First, @Eric P Smith, you refer to "both occurrences of any", yet the passage you quote has three instances of any, not two.

    Second, I like what Dan Hemmens wrote. I think it also fits with what Viktor (first comment) wrote about "all". All is all members of some set. Any is any member of some set. The question is, what set? Or, in this case, what does "court" mean in this context. (Well, this applies to non negated usage, which is the case under discussion, "convicted in any court".)

  21. Nat said,

    July 6, 2011 @ 1:56 pm

    I like the overall argument, but I'm having difficulty understanding the last excerpted paragraph from the decision, the one about "anomalies" produced by the interpretation which includes international courts. The claim is that there is "no problem" if the context is limited to domestic courts, but that there is a "senseless distinction" being made if "any" quantifies over international courts.

    As far as I can tell, this is precisely backwards. The distinction is perfectly intelligible on the more general interpretation. This is because this clause is particularly about domestic violence misdemeanors. While the other convictions included under this statute can be international, when it comes to domestic violence misdemeanors, the statute only cares about "Federal and State law". If the assumption being made at the beginning of the statute is that "any" covers international courts, then if you don't want to include international convictions for domestic violence, this restriction must be explicitly spelled out, as it is here.

    On the other hand, the language is anomalous on Breyer's preferred reading. If it's implicit in the context that "any" is limited to domestic courts, then the restriction to "Federal and State law" is redundant when discussing domestic violence. Perhaps Breyer is thinking that this explicit language is reinforcing the implicit restriction on "any". But then why is it only being made explicit in this particular portion of the statute?

  22. hector said,

    July 6, 2011 @ 2:06 pm

    My gut reaction to the "bananas" example was exactly the same as slobone's. The "any" doesn't widen the sentence, it changes its implied meaning, from "we don't have bananas (ever)" to "we don't have bananas (today)." You could argue it restricts the meaning (temporally), not widens it.

    But on this quibbly a point, usage may differ from one locale to another.

  23. anonymous said,

    July 6, 2011 @ 2:10 pm

    The same kind of argument was made by Breyer with regard to the word "licensing" in the Arizona Immigration case (Chamber of Commerce v. Whiting) earlier this year.

    "Dictionary definitions of the word 'licensing' are, as the majority points out, broad enough to include virtually any permission that the State chooses to call a 'license.' But neither dictionary definitions nor the use of the word 'license' in an unrelated statute can demonstrate what scope Congress intended the word 'licensing' to have as it used that word in this federal statute. Instead, statutory context must ultimately determine the word's coverage. Context tells a driver that he cannot produce a partnership certificate when a policeman stops the car and asks for a license. Context tells all of us that 'licensing' as used in the Act does not include marriage licenses or the licensing of domestic animals. And context, which includes statutory purposes, language, and history, tells us that the federal statute's 'licensing' language does not embrace Arizona's overly broad definition of that term. That is to say, ordinary corporate charters, certificates of partnership, and the like do not fall within the scope of the word 'licensing' as used in this federal exception."

  24. Xmun said,

    July 6, 2011 @ 2:18 pm

    I doubt if location makes much of a difference on this matter. I don't know where slobone and hector live, but I live in New Zealand, and my reaction to the bananas example is just the same as theirs.

  25. BlackHumor said,

    July 6, 2011 @ 3:15 pm

    @Nat: Suppose someone happens to commit the kind of domestic violence demeanor mentioned abroad, and gets a sentence of more than one year for it.

    If "any" means literally any, than the law contradicts itself; he can't buy firearms under the text which says "any court" but he can under the text which says "state and federal courts". So he can under the law both buy firearms and not buy firearms.

    But if any means only any US court there's no contradiction.

  26. Kenneth said,

    July 6, 2011 @ 3:22 pm

    @ Viktor – the point is not whether all means all members of a set. It's what the scope of that set is – programming languages work exactly the same way. If I loop over a restricted set (whether explicitly, or due to the way the compiler works), then "all" means something different – thus, the meaning of "all" (or "any") is very context dependent, which is the point of the post.

    People arguing that this should be straightforward, and that it's up to the legislative body to clarify, are not acknowledging just how much ambiguity there is in natural language – there are competing imperatives here: make laws that ordinary humans can read, versus remove all traces of ambiguity. You really can't do both.

    And yet, we seem to be able to communicate with each other most of the time, as long as we speak the same language. We make educated guesses about intended meaning and scope, and we're usually right. I'm not a lawyer, but I read the majority decision here as essentially admitting that there is interpretation in every communicative act.

  27. Adrian said,

    July 6, 2011 @ 4:33 pm

    Another banana-agreer. In BrEng too. (We haven't got bananas vs We haven't got any bananas.) The first is a bit unidiomatic, and there's no difference in meaning, unless you put emphasis on the any. If you don't want to use emphasis you'd have to say something like "We haven't got a single banana (even manky ones)."

    Reminds me of something that happened to me. I went into a cafe in a certain European country and asked for tea. The lady said "No tea." This clearly wasn't true, since I could see rows of boxes of tea on the shelf behind the counter. It turned out that they were out of lemons.

  28. Eric P Smith said,

    July 6, 2011 @ 4:47 pm

    @Ellen K.: yes, thankyou for correcting my counting. Three occurrences of any: any person, any court, any firearm. All three occurrences are of the non-affirmative any (what I used privately to call the "egative" – negative without the n – until I learned the accepted name). My point is unchanged.

  29. Rubrick said,

    July 6, 2011 @ 5:05 pm

    I hereby proclaim that the technical term for an item which may or may not fall in a certain category is "questionable banana".

  30. Dan Hemmens said,

    July 6, 2011 @ 6:21 pm

    What's the technical term for something that might or might not be categorized as a questionable banana?

  31. Xmun said,

    July 6, 2011 @ 6:31 pm

    A metaquestionable banana.

  32. Craig said,

    July 6, 2011 @ 7:00 pm

    A metaquestionable banana on the street, asked him if he had the time, said he didn't have any clue other than that time flies like an arrow yet fruit flies like a banana (even if or especially if you've got any wonky ones).

  33. Nat said,

    July 6, 2011 @ 8:13 pm

    Thanks BlackHumor, that's a helpful way of looking at it. Still, I don't think it's quite right, since being a misdemeanor and being punishable by more than a year's imprisonment are generally exclusive categories. So there's not a threat of a direct contradiction in that way.

    But after looking at the text of the statute, I think you're basically on the right track. It turns out that the discussion of misdemeanor domestic violence occurs in two distinct sections. There's the part that's immediately relevant to this statute, according to which it's an additional restriction on who can possess firearms. That is, misdemeanor domestic violence is a separate category from conviction which carries a lengthy prison sentence, and either criterion can be used to prohibit firearms. In this section, the language for both of these criteria is "any court" without mention of Federal or State courts. But, there's also a much earlier section which explicitly says that any mention of misdemeanor domestic violence in the subsequent passages only refers to conviction in a Federal or State (or Tribal) court.

    So, there are three passages at issue her. The phrase "any court" occurs twice, once for domestic violence, and once for crimes with a harsher penalty. A third earlier passage explicitly tells us that domestic violence is limited to U.S. courts. In order to make the two passages on domestic violence logically consistent, "any court" in reference to domestic violence must mean "any U.S. court". Then, if you want to have parallel use of language or at least a unified context for both occurrences of the phrase "any court", when applied to both misdemeanors and felonies, then the context must be U.S. in both cases.

  34. anna said,

    July 6, 2011 @ 8:21 pm

    Welcome to languagelog! I've been reading for several years, but as I've otherwise moved on from the field of linguistics, it was a very pleasant surprise to see the post author's name. I had you as a professor for, indeed, Semantics, at MIT several years ago, and it remains one of the most intriguing and fun courses I took while there. Glad to see you'll be contributing here, I look forward to reading more.

  35. Jonathan D said,

    July 6, 2011 @ 8:57 pm

    I imagined the banana statements in the context of a conversation at home, with an inclusive we and no need to spell out the usual state of affairs, and to me at least it was all perfectly idiomatic. (If that's hard to see, try 'milk' instead of 'bananas'.)

    Other Jonathan, the point here is that 'any' doesn't have any meaning with out context – we always need to "impose" a contextual meaning, and the question is which one is correct. That doesn't mean the court got it right in this case, of course.

  36. bloix said,

    July 6, 2011 @ 9:50 pm

    Q; Why does the statute say "any court"? Why isn't "convicted of a crime punishable by imprisonment for a term of more than one year" enough? Isn't "in any court" implied?
    A: Because this is a federal statute, and some defendant is bound to argue that "convicted of a crime etc." must mean convicted in federal court of a crime." So "any court" is intended to mean "any court, state or federal."

  37. Hermann Burchard said,

    July 7, 2011 @ 1:11 am

    @GeorgeW:

    Is there a difference between? 1. A person convicted in some court [may not own a gun]. 2. A person convicted in any court [may not own a gun].

    Using Emily's remark: . . "any" is functioning like the universal quantifier. about the whiskey example and applying it mutatis mutandis to your case, it is a fun exercise to properly formalize the two forms and show they are logically equivalent. I used de Morgan's law and the distributive law, both for quantifiers, plus "if A then B" is logically equivalent to "not A or B."

  38. Ben Hemmens said,

    July 7, 2011 @ 2:16 am

    I think the "any court" must be restricted to US courts, because one has to presume, as a default position, that the intention of Congress cannot normally be to unilaterally make US law subject to acts of foreign jurisdictions. If they really mean foreign courts then they should say so explicitly, but as has been pointed out that would be absurd: the person could have received in some rogue state a sentence that the US would never want to recognize in any way (e.g. 2 years' hard labor for eating questionable bananas in a country from which he is now a refugee, with recognized asylum). I suspect that this legal point is lurking behind the quasilinguistic consideration of "any". In other countries the definition of types of law and their scope might be more explicit and watertight. On the other hand, I'm pretty sure that in other countries, especially in code civil systems, this wording would be sent back to the legislature to get fixed.

    Just because in this case the foreign sentence was by a civilized country and suggests the man has a possibly questionable devotion to his firearms doesn't make it right.

    But "any" really does do odd things. It's hard to teach native-like use of it. In the bananas case, it seems to substitute for the lack of aspect marking in "have" (& the curious borrowing of what might have been the continuous form to mean "eating")
    Do you have bananas? (implied: ever, in general)
    ? Do you have bananas at the moment? (ooh, not quite right, is it?)
    Do you have any bananas at the moment? (oh. it's fixed)
    Do you have any bananas? (implied: at the moment)
    Do you have some bananas? (makes the "at the moment" even more redundant)

    Apologies to Dan for suspecting him of borrowing my surname. I see he really exists. For the record, he's the first Hemmens from outside my immediate family I've run across anywhere. It happened on LL, folks!

  39. RP said,

    July 7, 2011 @ 3:43 am

    There is a further distinction between "Do you have any?" and "Do you have some?". If you have only one, the answer to the former question is "yes" or "yes, (but) I have (just) one left", whereas the answer to the latter question might be "no, I only have one I'm afraid".

  40. John F said,

    July 7, 2011 @ 4:22 am

    For anyone who has access to or is familiar with the text of the law in question, is 'any court' defined anywhere, either explicitly or implicitly?

    This is where judges have to draw a fine line and why we call them judges. True, we can communicate pretty well most of the time with the ambiguities of our languages, but when you're going to restrict the liberty of someone or compel them to do something against their will, you need to be very precise to protect the innocent and the victims and make sure the guilty get the appropriate sentence.

    Judges may want legislatures to legislate better (a case regarding Police Bail recently came up in the UK rendering illegal 25 years of police practice), but they also have to judge what's right for the people who might be affected by their rulings and second-guess what precedents they are setting (I don't know if this latter aspect affects non-common law jurisdictions so much).

  41. GeorgeW said,

    July 7, 2011 @ 6:08 am

    Hermann Burchard: "it is a fun exercise to properly formalize the two forms and show they are logically equivalent."

    Hmm, I don't think 'any' and 'some' can be freely interchanged. As an example:

    1. I don't have any money.
    2. *I don't have some money.

  42. GeorgeW said,

    July 7, 2011 @ 6:14 am

    RP: Good point, 'some' must be greater than one.

  43. Ellen K. said,

    July 7, 2011 @ 7:21 am

    @Eric P Smith, I don't agree. Seems to me the first usage is distinctly different from the other two. We can substitute "all persons" for "any person" without changing the meaning. Doesn't work with "any court" or "any firearm".

  44. Dan Hemmens said,

    July 7, 2011 @ 9:00 am

    Apologies to Dan for suspecting him of borrowing my surname. I see he really exists. For the record, he's the first Hemmens from outside my immediate family I've run across anywhere.

    Yeah, I was kind of surprised as well – I don't know any Hemmenses outside of my immediate family either.

  45. Ben Hemmens said,

    July 7, 2011 @ 9:18 am

    Roots in Wiltshire? That's where my great-grandfather came from.

  46. Hermann Burchard said,

    July 7, 2011 @ 12:48 pm

    GeorgeW: "I don't think 'any' and 'some' can be freely interchanged."

    "Sigh.."
    No claim of free interchange made in my above comment, just that two particular sentences are logically equivalent statements:

    1. A person convicted in some court [may not own a gun].
    2. A person convicted in any court [may not own a gun].

    Here now the "fun exercise to properly formalize the two forms and show they are logically equivalent." We need some notation (clumsy due to my lacking LaTeX2HTML skill):
    v(c,p) denotes "person p was convicted in court c"
    mog(p) denotes "person p may own gun"
    ~ denotes not
    forall denotes the universal quantifier
    exists denotes the existential quantifier
    A=>B denotes "if A then B"

    Then statement 2 formalizes as follows, logically equivalent forms following:
    forall c [ v(c,p) => ~mog (p)]
    forall c ~[v(c,p) & mog (p)]
    ~ exists c [v(c,p) & mog (p)]
    ~ [exists c v(c,p)] & mog (p)
    ~ [exists c v(c,p)] & mog (p)
    [exists c v(c,p) ] => ~mog(p)

    The last line is the formalized version of statement 2. [End of proof.]

    We applied de Morgan's law and a tautology from the propositional calculus:
    forall c ~A[c] is logically equivalent to ~ exists c A[c]
    ~[A&B] is logically equivalent to A=>~B.

  47. Hermann Burchard said,

    July 7, 2011 @ 1:01 pm

    Correction: The last line is the formalized version of statement 1.

  48. Hermann Burchard said,

    July 7, 2011 @ 1:26 pm

    Another correction: Instead of
    ~ [exists c v(c,p)] & mog (p)
    put
    ~ [exists c v(c,p) & mog (p)]
    (Sigh!)

  49. Aaron Davies said,

    July 7, 2011 @ 8:33 pm

    this reminds me of arguments where "all that glitters is not gold" is interpreted as Ɐx(¬Glitter(x) → Gold(x)) or Ɐx(Glitter(x) → ¬Gold(x)), when what's clearly meant is ¬Ɐx(Glitter(x) → Gold(x))

  50. Aaron Davies said,

    July 7, 2011 @ 8:33 pm

    (and btw fiddling around on wikipedia and unicode reference sites is one way to get proper logic symbols into a comment)

  51. Paul Turpin said,

    July 7, 2011 @ 11:08 pm

    "Yes, we have no bananas!"

    I'm surprised no one has mentioned this one yet. (Implications: yes, we carry bananas, but no, we're out of them at the moment…?)

  52. Hermann Burchard said,

    July 7, 2011 @ 11:58 pm

    @Aaron Davies:
    Your examples of formalizations of all that glitters is not gold are false and based on easily detectable errors. Please tell where my examples (above) are erroneous (as corrected above).

    [Apologies for not scrounging around the internet for "to get proper logic symbols into a comment." But ~ used to be quite common for not, and I'm used to LaTeX, where:
    \forall –> ∀,
    \exists –> ∃.]

  53. John Walden said,

    July 8, 2011 @ 2:02 am

    Nobody has mentioned the context of:

    "Have you got Wee Jock McSporran's One-Month-Old Scotch-Style Beverage For The Budget-Minded Drinker? "

    "I'm sorry, Sir, we don't sell ANY whisky" (i.e.it has to be good)

    "Have you got any questionable bananas?"

    "I see Sir has come to the wrong establishment. We don't sell ANY bananas either. Like our whiskies, all our bananas are of the highest quality"

  54. BlackHumor said,

    July 8, 2011 @ 2:08 am

    @Paul:

    I don't think that makes sense except in response to a question very similar to "You don't have any bananas, right?"

  55. John Walden said,

    July 8, 2011 @ 6:00 am

    I got the feeling that I'd been here before. I then found:

    http://languagelog.ldc.upenn.edu/nll/?p=3071

  56. Ellen K. said,

    July 8, 2011 @ 10:16 am

    @John Walden,

    How does that post relate to this one? I don't see the connection.

  57. Ben Hemmens said,

    July 8, 2011 @ 11:01 am

    I suspect a shop I saw recently, called "Thistle Do Nicely", sells any amount of questionable goods.

  58. John Walden said,

    July 8, 2011 @ 11:31 am

    @Ellen K. Look at the comments on that link from

    Pflaumbaum, April 6, 2011 @ 4:48 am

    onwards and you'll see why I got déjà vu.

  59. Ellen K. said,

    July 8, 2011 @ 1:05 pm

    Ah, yes, I see clearly. And I nearly made the same argument here that I did in that comment thread, save for the banana/pigeon distinction.

  60. “we don’t have any bananas” « shigekuni. said,

    July 10, 2011 @ 6:51 pm

    […] Fintel, language log and Linguistics Kai von Fintel's inaugural post at LanguageLog. Click here for the full post. The Supreme Court's doctrine therefore seems to be that "any" like other […]

  61. Mark Beadles said,

    July 16, 2011 @ 10:12 am

    A few years back I was getting ready to move to my first house and, being freshly out of college, was doing it on the cheap. So I went to local grocery stores begging boxes that I could use for packing. To my luck, a large shipment of bananas had arrived. Banana boxes being large and very sturdy, they were perfect for my needs, so with permission my companion and I each grabbed a stack of empty boxes.

    On the way out of the store the security by the door asked us a question so serendipitous it was like a gift:

    "Are those banana boxes empty?"

    To which there was only one possible response. :)

  62. Sabbatical Diary 2011-07-25 | Kai von Fintel said,

    July 25, 2011 @ 11:46 am

    […] wrote two Language Log posts: one on the meaning of “any” according to the Supreme Court and one on the difference between […]

  63. Jeffrey said,

    January 26, 2012 @ 8:14 pm

    'Failure to make any such submission' means 'fail to make such submission at all' or 'failure to make one of such submission' ? Such submission refers to a list of 3 types of notices that are to be submitted within certain time frame.

    Q:Do you agree that the meaning of 'any' in the above phase is ambiguous?

RSS feed for comments on this post