Anaphoric definiteness in the ACA

« previous post | next post »

The following is a guest post by Graham Katz. It makes an interesting point (which I haven't seen elsewhere) about the phrase that's at the center of King v. Burwell: "an Exchange established by the State".


Today the Supreme Court hears argument on the King v. Burwell case challenging the subsidies for health insurance put in place by the Affordable Care Act (or "Obamacare"). At issue in today's argument is the interpretation of a part of the law which specifies the premium subsidies that are crucial to making health insurance affordable to lower income Americans. This part is section 1401(b):

In General- In the case of an applicable taxpayer, there shall be allowed as a credit against the tax imposed by this subtitle for any taxable year an amount equal to the premium assistance credit amount of the taxpayer for the taxable year.

(2) PREMIUM ASSISTANCE AMOUNT- The premium assistance amount determined under this subsection with respect to any coverage month is the amount equal to the lesser of–

(A) the monthly premiums for such month for 1 or more qualified health plans offered in the individual market within a State which cover the taxpayer, the taxpayer’s spouse, or any dependent (as defined in section 152) of the taxpayer and which were enrolled in through an Exchange established by the State under 1311 of the Patient Protection and Affordable Care Act, or

(B) the excess (if any) of–

(i) the adjusted monthly premium for such month for the applicable second lowest cost silver plan with respect to the taxpayer, over

(ii) an amount equal to 1/12 of the product of the applicable percentage and the taxpayer’s household income for the taxable year.

Although states are required by the law to establish an Exchange for residents to use, many did not. The law anticipates this and allows for the federal government to set up and run an exchange in its stead.

The plaintiffs argue that subsidies are restricted by the phrase "an Exchange established by the State" to taxpayers in those states that set up their own exchange (California, New York, …) and are not available to taxpayers who use the HealthCare.gov exchanges run by the federal government. Why else would Congress use this phrase, argue the plaintiffs, if not to distinguish federal from state exchanges?

There are many legal issues at play in this argument, but there are also important linguistic issues. For example, does a state "establish" an exchange by failing to act and thereby triggering the creation of the exchange by the federal government?

While the lexical semantics of this verb can be debated, there is little room for debate over the meaning of the definite article "the." Semanticists agree that definites are used in one of two ways: to specify the (contextually restricted) uniqueness of a referent (there is presumed to be exactly one PPACA, exactly one applicable second lowest cost silver plan, etc.) OR to anaphorically to refer back to a previously introduced discourse element. (Some have argued that these are two aspects of a univocal meaning, but the present point is only that there are two interpretations, whatever their source.)

In the phrase "an Exchange established by the State under 1311 of the Patient Protection and Affordable Care Act" the definite expression "the State" is used anaphorically. Its antecedent is the phrase "a State" in "health plans offered in the individual market within a State." The anaphoric relation enforces a legal requirement that the state that the health plans plans are offered in be the same state as that which set up the enrolling Exchange. These are the kind of anaphoric uses of definites that launched the Discourse Representation revolution of the early '80s, which provided an compelling account of both anaphoric interpretations of defnite expressions and the contextual variability of interpretation of indefinites. In short, indefinites introduce novel discourse referents which may be bound by discourse-level operators, while definites refer to already introduced discourse referents.

The anaphoric relation induced by the use of the definite creates an important requirement. If an indefinite had been used instead (or, as some have suggested, the disjunct "…or the federal government"), the required relationship between the health-plan offering State and constraints on its Exchange would be broken. A plan could be offered in one state and (theoretically) enrolled in by an exchange set up by another. In fact, each of the eight uses of the phrase "an Exchange established by the State…" in the ACA involves such an anphoric use
— this is clearly the idiom favored by Congress to set up this kind of linked requirement. Here are some (simplified) examples:

"In the event that allotments .. are insufficient to provide coverage to all children who are eligible to be targeted low-income children under the State child health plan … a State shall establish procedures to ensure that such children are provided coverage through an Exchange established by the State …"

"A State shall establish procedures for … enrolling, without any further determination by the State ,.. individuals who are identified by an Exchange established by the State … as being eligible for … assistance"

"With respect to each State, the Secretary … shall review the benefits offered for children … offered through an Exchange established by the State …and shall certify those plans .. offer benefits … at least comparable to the benefits offered .. under the State child health plan."

This required anaphoric relation — specifying, for example, that the state establishing procedures for ensuring childhood coverage is the state whose exchange is providing coverage – makes a crucial contribution to the meaning of the statute, and provides a rationale for the inclusion of the phrase "…established by the State…" in these eight passages in the statute. The phrase "an Exchange established by the the State under section 1311 " is not contrasted with "an Exchange established by the Federal Government under section 1321" but in contrast to the bare "an Exchange established under section 1311…", which would have failed to make explicit the link between the Exchange and the State. Leaving out "the State" would have allowed for situations in which, for example, New Mexico could claim to satisfy the child health benefits requirement by pointing to benefits offered on the California exchange.

The participle "established" came along for the ride. But the crucial thing was that a relation should be set up between an Exchange and a (previously mentioned) State.


The above is a guest post by Graham Katz.



44 Comments

  1. D.O. said,

    March 4, 2015 @ 7:36 pm

    I am completely unqualified to comment on legal issues as well as on intricacies of anaphoric determinants, but a hypothesis that "New Mexico could claim to satisfy the child health benefits requirement by pointing to benefits offered on the California exchange" is so pragmatically absurd that it's not possible to imagine the drafters of the ACA guarding against such interpretation.

  2. Raghav said,

    March 4, 2015 @ 7:49 pm

    So why not (as Justice Alito proposed at oral argument today) "established in the State"?

  3. Steve T said,

    March 4, 2015 @ 8:17 pm

    People who draft legal documents (contracts or statutes) are always guarding against interpretations that seem pragmatically absurd at the time.

  4. Roadrunner said,

    March 4, 2015 @ 8:45 pm

    As someone who writes laws for a living, that sort of "absurd" interpretation is exactly what we work to guard against in our drafting. And in fact, since cross-state insurance sales are an active policy controversy (could we reduce the cost of insurance by allowing people in high-regulation states to buy cheap policies in low-regulation states?) clarifying that the policy in question must be sold in the state and on the exchange in question is a critical piece of specificity.

  5. J. W. Brewer said,

    March 4, 2015 @ 9:44 pm

    "The" is not the only way to mark definiteness, and it's not always the strongest. I think you could make the anaphoric-ness (anaphoricity?) discussed here even clearer by making the phrase "an exchange established by that State." Although maybe "established by such State," which has a nice legal-jargon feel to it, would be more idiomatic for statutory language. But would these alternative phrasings driving home the must-be-the-same-State angle make the also-covers-federally-established-Exchanges argument feel stronger or weaker?

  6. D.O. said,

    March 4, 2015 @ 10:58 pm

    Steve T. and Roadrunner, thank you for setting me right.

    On the larger question, can there be a commutativity problem? Let's look at two outlines

    1a. Each state must establish an exchange.
    2a. Fed. gov. subsidizes the exchange established by the state under 1a.
    3a. If a state does not want to establish the exchange, fed. gov. does it for the state.

    or

    1b. Each state must establish an exchange.
    2b. If a state does not want to establish the exchange, fed. gov. does it for the state.
    3b. Fed. gov. subsidizes the exchange established by the state under 1b.

    Is it me or in the first case it is much clearer that subsidies are available no matter who ends up establishing the exchange. In the second case, the reading of "under 1b" (as opposed to 2b) feels much more disjunctive, while "under 1a" reads like usual overabundance of precision.

  7. Raghav said,

    March 4, 2015 @ 11:30 pm

    Incidentally, this point is made in the Government's brief at pages 33-35, and challenged in the petitioner's reply brief at pages 6-7. The petitioners point out that "an Exchange established by the State" is not used anaphorically in 26 U.S.C. § 36B(c)(2)(A)(i), which says in part: "For purposes of this subsection – The term 'coverage month' means, with respect to an applicable taxpayer, any month if – as of the first day of such month the taxpayer, the taxpayer's spouse, or any dependent of the taxpayer is covered by a qualified health plan described in subsection (b)(2)(A) that was enrolled in through an Exchange established by the State under section 1311 of the Patient Protection and Affordable Care Act."

    [(myl) Sort of. The government's brief is here, and says on p. 33: "The phrase 'established by the State under Section 18031' serves to identify the Exchange in a particular State, not to exclude a federally facilitated Exchange."]

  8. Ben Hemmens said,

    March 5, 2015 @ 10:07 am

    Could it just be that someone accidentally typed the second "State" with an uppercase S? Is "the state" used in US laws to refer in a general way to the whole collective of State and Federal governments and their agencies?

  9. Raghav said,

    March 5, 2015 @ 11:02 am

    Sort of. The government's brief is here, and says on p. 33: "The phrase 'established by the State under Section 18031' serves to identify the Exchange in a particular State, not to exclude a federally facilitated Exchange.

    Right, that's the heading of the relevant section, and the brief goes on to say more specifically: "An Exchange is a state-specific marketplace, and Section 36B(b)(2)(A) uses the phrase 'Exchange established by the State under [Section 18031]' because it is referring to the Exchange in the specific State mentioned earlier in the same sentence: The formula for tax credits depends on the cost of one or more insurance plans 'offered in the individual market within a State * * * which were enrolled in through an Exchange established by the State under [Section 18031].'"

    Also, the original post says that "each of the eight uses of the phrase 'an Exchange established by the State…' in the ACA involves such an anphoric use," but as the petitioners pointed out, that doesn't seem to be true of section 36B(c)(2)(A)(i).

  10. Michael Watts said,

    March 5, 2015 @ 11:43 am

    I think you could make the anaphoric-ness (anaphoricity?) discussed here even clearer by making the phrase "an exchange established by that State."

    You're moving in the wrong direction, trying to derive a noun from the adjective. The word you want is "anaphora".

    This is something I noticed recently in an article talking about ethnic Mongols (in China, not in Mongolia). It always referred to them as Mongolians. But I'm pretty sure that most English-speaking people are familiar with the term "Mongol". I conjectured, on the basis of nothing in particular, that it was difficult for the author to imagine that an area might be named after its people rather than the other way around, although in fact regions named after people are quite common. (France, England, Russia, Germany/Allemagne, all of the -stan countries (as far as I know), Mongolia, Thailand… there's no end to this.)

  11. Andrew Taylor said,

    March 5, 2015 @ 12:09 pm

    "Mongol" was formerly used to describe a sufferer from Down Syndrome, so it's not surprising that people now tend to avoid it in describing people from Mongolia. Also, Mongolia -> Mongolian follows the familiar pattern exemplified by Australia -> Australian, Bolivia -> Bolivian, Croatia -> Croatian and many others.

  12. J.W. Brewer said,

    March 5, 2015 @ 12:53 pm

    FWIW the word I wanted was not "anaphora" in its usual sense but something meaning something like "the quality of being an instance of anaphora," which could, but I suppose wouldn't need to be, a different word. But if I were making this argument in a legal brief I would probably spend time figuring out how to phrase it without using "anaphora" etc. in its technical syntactic sense, to avoid readers used to legalese getting bogged down in a different technical jargon with which they lacked previous familiarity. Or I might use some more transparent phrase like "reference back" with a single parenthetical note toward the beginning of the argument that this phenomenon is known in the relevant scholarly literature as anaphora.

  13. J.W. Brewer said,

    March 5, 2015 @ 1:02 pm

    Ben Hemmens: answer to second question for AmEng is no (or at least, not in statutes as opposed to libertarian rhetoric about how The State is oppressing us). For first question, since the U.S. is comprised of 50 states plus other bits of territory that are not states strictu sensu (the District of Columbia, Puerto Rico, etc.) it is not uncommon for particular federal statutes to want to treat some (but not necessarily all, and not necessarily the same subset from statute to statute) of those non-state subdivisions like states, and one standard way to do that is have a definitional section that says that for purposes of this statute "State" also includes Puerto RIco or whatever. It is often thought good practice (although I can't claim it's universally followed) in AmEng legal writing to capitalize a word like that when you want to make sure the reference is understood as being to a defined term used according to the definition given elsewhere in the same statute/contract/whatever rather than to the default common meaning of the word (which of course might be ambiguous).

  14. Raghav said,

    March 5, 2015 @ 1:08 pm

    J.W. Brewer: And in fact, the ACA does define "State" to mean "each of the 50 States and the District of Columbia." 42 U.S.C. § 18024(d).

  15. J.W. Brewer said,

    March 5, 2015 @ 1:10 pm

    For example, note the contrast between lower-case "foreign state" and defined-term capitalized "State" in this jurisdictional statute: https://www.law.cornell.edu/uscode/text/28/1332. (The definition is at the very end, thus following rather than preceding the substantive sections where you need to understand what the defined term means. That seems to violate some sort of Gricean principle but statutes are a funny genre.

  16. Michael Watts said,

    March 5, 2015 @ 2:07 pm

    I'm having a difficult time really grasping the distinction proposed between using "anaphora" and using "anaphoricness". I think that's because the sentences work out to mean exactly the same thing.

    I can say "You can make the anaphora in this sentence clearer by doing X". I interpret that as saying that the sentence contains anaphora, the phenomenon, and X is advice on demonstrating that.

    I could say "You can make the anaphoricness of this sentence clearer by doing X". I interpret that as saying that the sentence is anaphoric, and X is advice on demonstrating that.

    But, to me, there is no distinction whatever between "this sentence contains anaphora" and "this sentence is anaphoric"; that's the entire meaning of "anaphoric".

    If you see a difference that I'm missing, I'd like to hear about it.

  17. Michael Watts said,

    March 5, 2015 @ 2:15 pm

    @Andrew Taylor

    Croatia -> Croatian isn't an example of a pattern that leads people astray into losing the word "Mongol", it's an example of the same process that produces "Mongolian". Croatia is named after the Croats.

  18. Michael Watts said,

    March 5, 2015 @ 2:28 pm

    Mongol -> Mongolia; Serb -> Serbia; Croat -> Croatia; Rus -> Russia (this one can fairly be called historic), Bulgar -> Bulgaria… there's clearly an active process that produces place names by suffixing -ia to a stem that characterizes the place. (This is true of Australia too, and I tend to assume it's true for all such -ia names.) But the logical endpoint (actually, midpoint) of this process is ethnonyms like canadianianianian. The only explanation I can think of for why that doesn't happen is that people somehow believe that while it's OK to derive an ethnonym from a place name, place names can't be derived from ethnonyms; they are instead fixed by some sort of primal divine mandate. I can't think of a reason, though, for why people would think that way.

  19. Graham Katz said,

    March 5, 2015 @ 2:30 pm

    @Raghav

    FWIW I take it to be the case that 36B(c )(2)(A)(i) provides an implicit antecedent for definite "the State." Section 36B(c )(2)(A)(i) references "a qualified health plan described in subsection (b)(2)(A)," which, of course, is a health plan "offered in the individual market within a State."

  20. Nathan said,

    March 5, 2015 @ 5:28 pm

    Michael Watt, I've got a hypothesis. The idea of an "ethnic group" just isn't that important to Americans, what with the "melting pot" ideal. Mongols who go to China become Chinese.

  21. J.W. Brewer said,

    March 5, 2015 @ 6:10 pm

    Not having felt that anaphora was quite the right word in a specific context is perhaps an idiosyncratic quirk of mine not really worth trying to justify and causing unnecessary thread-drift. But a moment's googling shows that "anaphoricity" is certainly a word that has been used previously in technical linguistics scholarship, so although while I coined or recoined it on the spot without consulting that literature (although I might have a subconscious memory of having seen it previously in such contexts), others seem to have found the word useful.

  22. Eli Nelson said,

    March 5, 2015 @ 8:05 pm

    @Michael Watt
    I think people see the names of geographic names fixed, not by divine mandate, but by the map. It's comparatively harder to find an unknown ethnonym, and there are few groups for which people are more familiar with the ethnonym than the place name. For me, "Bulgar" is alongside Rus in the category of "historical ethnonyms"; I would say "a Bulgarian" without second thought.

  23. Marc said,

    March 5, 2015 @ 10:20 pm

    Re anaphoricity vs. anaphora –

    Isn't this the same distinction as in "There was redness in her cheeks" vs. "there was red in her cheeks"?

    Red is the color itself. Redness describes the red quality of her cheeks.

    Anaphora is the rhetorical device itself. Anaphoricity is the anaphoric quality of the usage of definite articles in the statute.

    I think it's a useful distinction.

  24. Levantine said,

    March 5, 2015 @ 10:31 pm

    Michael Watts, "Pakistan" is an exception among the -stan countries, though easily enough accounted for.

    I agree with J. W. Brewer on the issue of anaphora/anaphoricity.

  25. Levantine said,

    March 5, 2015 @ 10:44 pm

    Also, the longer and shorter forms of a particular ethnonym can mean or connote different things, as already discussed in relation to Rus/Russian and Bulgar/Bulgarian. Examples that spring to mind are Azeri/Azerbaijani, Bengali/Bangladeshi, and Arab/Arabian.

  26. Michael Watts said,

    March 6, 2015 @ 3:37 am

    Oh, I agree that redness (n.), the quality of being red, and red (n.), the reification of the adjective, are distinct, and usefully distinct, concepts. I was only saying that, as applied to this particular case, I have trouble seeing the distinction between the proposed "anaphora" and the proposed "anaphoricity". That's also true for the example of "redness in her cheeks" and "red in her cheeks", where the semantics are identical. (And, perhaps, also because, as J. W. Brewer says, "'the quality of being an instance of anaphora,' [] could, but I suppose wouldn't need to be, a different word". As applied to a single word, I have a much easier time distinguishing "anaphora", which doesn't really make sense in that context, from "anaphoricity", which does; as applied to a sentence (well, this sentence), I get hazier. But I don't object to the viewpoint that that's a flaw in my processing.

  27. Michael Watts said,

    March 6, 2015 @ 3:45 am

    The idea of an "ethnic group" just isn't that important to Americans, what with the "melting pot" ideal. Mongols who go to China become Chinese.

    This is kind of a strange argument to make concerning an article about ethnic Mongols in China.

    Speaking a little more seriously about the seeming fixation of place names and fluidity of ethnonyms, the basic problem I have is that it's difficult to make the argument "the name Mongoliania is blocked by the prior existence of Mongolia referring to the same place" without simultaneously making the argument "the term Mongolian is blocked by the prior existence of the term Mongol referring to the same group". Yet it definitely appears that the first is true but the second is false.

    And I think Mongols are an especially interesting case of this because they are a rare ethnic group where it's quite plausible for English speakers to be more familiar with the ethnonym than with the country. The modern country doesn't really amount to much on the world stage, but Genghis Khan made a splash and is widely remembered around the world today.

  28. Levantine said,

    March 6, 2015 @ 12:23 pm

    Michael Watts, "red in her cheeks" to me connotes that her cheeks are in any case red, whereas "redness in her cheeks" suggests that something has (temporarily) reddened them. I'm sure many others would agree with you that the two phrases are semantically identical, but I for one am pleased that English offers its speakers so many (near-)synonyms with which to express shades of meaning.

  29. Jerry Friedman said,

    March 6, 2015 @ 5:28 pm

    Michael Watts: I think there's a rule about piling up ethnonymic or toponymic suffixes. That's what prevents *Mongoliania. Someone will probably find an exception, maybe where it's not obvious that the root word contains a suffix, but in general it works pretty well. We do have both Scot and Scotsman.

    I'm only barely aware of Mongol for a person with Down's syndrome. In my (American) youth they were called "Mongoloids". I suspect "Mongol horde" may play more of a role.

    *walks away whistling Finlandia*

  30. Marc said,

    March 6, 2015 @ 6:39 pm

    So does this explain "Britisher"?

  31. Marc Naimark said,

    March 7, 2015 @ 5:03 am

    Re Mongolians. For me, "Mongols" are the ethnc group, and in particular the group active in world history, as opposed to current Mongols. "Mongolia" is a country with a mainly Mongol population. "Mongolians" are citizens or residents of that country, whatever their ethnicity. Likewise, there are "Mongols" who are citizens and residents of countries other than Mongolia.

  32. Established by a state | The Incidental Economist said,

    March 7, 2015 @ 10:00 am

    […] The statute also reads, in parts, "established by the State". The difference is important. […]

  33. @BobbyGvegas said,

    March 7, 2015 @ 10:26 am

    While we're obsessing over the allegedly contentious import of ACA Section 1311(d) (which starts out "IN GENERAL" btw), it helps to also note that Section 1321 supplants and trumps 1311, by setting for the clear, plain English HIX authority of the federal government in the wake of state HIX inaction.

    http://regionalextensioncenter.blogspot.com/2015/03/obamacare-back-at-scotus-orals.html

    The specious 1311 argument is Motivated Reasoning run amuck.

  34. James Wimberley said,

    March 8, 2015 @ 12:33 pm

    The South Slavic people who gave their name to Serbia were Serbs long before there was such a country. "Serbian" strikes me as ignorant. That does not hold for Pakistan, and "Paki" is offensive.

  35. DaveK1955 said,

    March 8, 2015 @ 1:55 pm

    @James Wimberley: There is a distinction between "Serb" and "Serbian". "Serbian" usually refers to a citizen or national of Serbia regardless of their ethnicity, and "Serb" used for people of Serbian background regardless of where they live.
    On the larger question of why Americans tend to ignore ethnonyms like Serb or Mongol, it probably has something to do with the phenomenon of countries named after the dominant ethnicity being restricted to the Eastern Hemisphere.

  36. Marc said,

    March 8, 2015 @ 2:53 pm

    One instance of a distinction like Mongolian/Mongol or Serbian/Serb which we can't make in English is россиянин (Rossianin) v. русский (Russky) – both "Russian" in English, the former being any citizen of Russia (Rossiya), the latter being ethnic Russians (i.e., descendants of the inhabitants of Rus').

  37. BZ said,

    March 9, 2015 @ 8:57 am

    @Marc,
    It's funny you mention Russian because it does not have the distinction for most countries. Even the term "россиянин" for citizen of Russia first appeared in that sense in 1929, and did not gain wide use until after the breakup of the Soviet Union (there was never a single Russian word for "citizen of the Soviet Union". For the individual republics the ethnonym and demonym are pretty much indistinguishable).

  38. KevinM said,

    March 9, 2015 @ 10:03 am

    Billy Carter, in one of many loose-tongued moments, opined that his brother's foreign policy should tilt toward the Arab countries, because there were "a lot more Arabians."

  39. Rodger C said,

    March 9, 2015 @ 11:26 am

    Surely the worst thing you can say to a Serb, or Serbian, is to call his/her country by the old name, "Servia."

  40. GeorgeW said,

    March 9, 2015 @ 11:53 am

    Re 'mongol' aversion: I think this could also be affected by the word 'mongrel' which has similar sounds and a negative connotation. In fact, the two might be confused by some speakers.

    Also, I sense that adjectives are more positive than nouns when referring to nationalities or ethnic groups: Jew/Jewish, Arab/Arabic, Briton/British, etc. As an example, I can't picture an anti-Semite referring to someone as "Jewish."

  41. Levantine said,

    March 9, 2015 @ 1:45 pm

    GeorgeW, although many use it to describe people, food, etc., "Arabic" refers only to the language. I agree, however, that it functions popularly/non-standardly as a politically correct alternative to "Arab".

  42. Alex Bollinger said,

    March 10, 2015 @ 6:14 am

    @BobbyG, I think most people here are on the same page. Not only is the plaintiff's interpretation of that sentence absurd in the context of the law's full text, it's also absurd considering the intent of that law and the legislative history of that law, plus the courts have a long established principle of deferring to the IRS in cases of actual ambiguity. Like you said, if you don't already hate the ACA, then this argument will seem ridiculous.

    On the other hand, let's analyze how that sentence got written! It's probably a typo, but no typo is a totally random accident and someone was probably trying to say something very specific and just never thought that the subsidies would be denied on the federal exchange because no one was talking about that and the idea is totally ridiculous. And here we are.

  43. Oona Houlihan said,

    March 11, 2015 @ 1:44 pm

    The whole thing starts with the federal government wanting to "ram" through a legal "accomplishment" that constitutionally is the sole jurisdiction of its member states. To achieve this it hat to make so many compromises, plug so many holes and at the same time open the legislation to so many interpretative pitfalls that it could but create a legalistic nightmare. Clearly, if at the federal level not even a compromise can be had while at state level there could sometimes have been clear majorities to implement "best of breed" solutions that might have had convincing appeal. Just think for a second: if you were in private business and someone approached you with a business proposal as complex as this – would you contract?

  44. Links and Tweets... (Brad DeLong's Grasping Reality...) said,

    March 13, 2015 @ 2:50 am

    […] Graham Katz: Anaphoric definiteness in the ACA http://languagelog.ldc.upenn.edu/nll/?p=18039 […]

RSS feed for comments on this post