« Terrorism Trials and the Constitutionality of the Religious Insanity Defense | Main | AALS Happy Hour in NOLA: Thurs Jan 7th, 9pm at the Whiskey Blue Bar, W Hotel »

Thursday, January 07, 2010

An AALS smackdown on philosophy of language?

As I complained in an earlier post, I'm a bit disappointed by what I take to be the dearth of interest in the philosophy of language at law schools. But Larry Solum's work is an exception that proves the rule: Larry has defended a dense and rigorous brand of "semantic originalism" using Gricean semantics. (You can download the Reader's Digest version here). Given that Larry, Mitch Berman, Ian Bartrum, Laura Cisneros, Amy Barrett, and I are all panelists at the AALS conference charged with discussing "The Interpretation – Construction Distinction in Constitutional Law" (on Saturday, January 9th) and given that Mitch and Larry have tussled over Larry's originalist argument in print, this seems like a good opportunity to provoke some pre-conference discussion on a linguistic issue close to my heart -- the principle of charity. Specifically, is the principle of charity a normal part of "interpretation," akin to dictionaries, rules of grammar, etc., as a guide to norms of linguistic usage? Or is it a matter of "construction" to which we should resort only when the text is otherwise ambiguous?

"Charity" is simply the idea that we construe someone else's words to correspond as closely as possible to what we believe to be true: Interpretation, on this account, should maximize normative or factual agreement between interpreter and speaker. In Donald Davidson's words,"it is impossible for an interpreter to understand a speaker and at the same time discover the speaker to be largely wrong about the world." (Subjective, Intersubjective, Objective at 150). When some theory of usage makes speakers say something really weird (to our lights), we discount that theory, favoring other plausible theories of usage that eliminate the oddity as much as possible.

Larry seems to argue that our own normative or empirical notions cannot play a role in fixing the semantic content of constitutional text: In his words, the belief that "semantic content is determined after the fact of communication on the basis of what we would like an utterance to have meant given our practical concerns" is a "fundamentally confused" way to think about interpretation (page 103). Larry includes three pages (pages 146-49) on the principle of charity, Quine, and Davidson in Semantic Originalism, but he treats charity as a way of reducing ambiguity through "construction" rather than as a form of "interpretation."

I think that I disagree with Larry on the role of charity, which I take to be built into interpretation just as deeply as any anthropological examination of dictionaries and rules of grammar. But an illustration from some specific constitutional text -- Article I, Section 2, Clause 3, calling for a decennial "enumeration" or census -- might help clarify why I think that Larry is mistaken.

Article I, section 2, clause 3 requires that an "actual Enumeration shall be made within three years after the first meeting of the Congress of the United States, and within every subsequent term of ten years, in such manner as they shall by law direct." Are statistical sampling techniques that do not involve the actual individual counting of citizens an "actual enumeration" within the meaning of this clause? SCOTUS has not decided this question, preferring to resolve the issue on statutory grounds in Department of Commerce v House of Representatives, 525 U.S. 316 (1999). But Justice Scalia, concurring separately, suggested that "[f]or reasons of text and tradition, fully compatible with a constitutional purpose that is entirely sensible, a strong case can be made that an apportionment census conducted with the use of sampling techniques is not the actual Enumeration that the Constitution requires." Dep't of Commerce at 349. "Text and tradition" were established to Justice Scalia's satisfaction by "[d]ictionaries roughly contemporaneous with the ratification of the Constitution" as well as the early republic's conducting its censuses of 1790 and 1800 through the actual counting of heads.

I want to draw attention to Justice Scalia's reference to "a constitutional purpose that is entirely sensible." Even a good originalist like Scalia felt the need to supplement his dictionaries with some account of the purpose served by limiting the definition of "actual enumeration." Moreover, Scalia felt that he should describe this purpose in a way that rendered it intelligible -- "entirely sensible" -- to us. (According to Justice Scalia, the term excluded enumerations made through complex statistical techniques, because such techniques would give Congress too much discretion to skew the population count for partisan reasons). We could not eliminate the possibility that statistical sampling constituted an acceptable "enumeration" simply by looking at dictionaries or early censuses. Even if these sources indicated that head counts were the universal method for conducting censuses in 1790, we could not exclude the possibility that those folks simply had not considered the possibility of using advanced statistics. If statistical sampling were a much more accurate and cheaper way of counting the population, then it would be akin to taking a spoonerism literally to exclude this practice from the term: Why in the world would we assume that the framers and ratifiers and general public had some fetishistic aversion to good social science? To make sense of the constrained definition excluding sampling, we need some account of 18th century purposes that rings true to us.

Why this invocation of normative considerations by the arch-originalist? Because this is how ordinary interpretation is always conducted: Charity is essentially similar to conventions of grammar, "textual" canons of construction, dictionary definitions, etc. It is part and parcel of ordinary interpretation. Just as we read a series of terms by deploying ejusdem generis, thereby preventing a broad final term from rendering the other terms redundant, so too, we read apparently wacky assertions in light of our own norms of morality, empiricism, etc. We invoke the canon, because we impose on speakers a norm against redundancy; We invoke charity, because we impose a norm of rationality. But both the canonical and the charitable readings are equally "literal" albeit normative: They eliminate ambiguity only in the trivial sense that all texts are ambiguous until we invoke all of the semantic resources necessary to interpret them. When William Spooner referred to Queen Victoria as "our queer old dean," we do not first discover that the terms "queer" and "dean" are somehow obscure and then deploy some ambiguity-reducing norm because ordinary semantic resources have failed us. Rather, we know from the outset that the terms are perfectly clear when read in light of the appropriate norms of usage: Spooner said that Queen Victoria was a "dear old queen" and did not (even ambiguously) say that she was an eccentric staff member of a university. Moreover, we reverse the letters in Spooner's statement not because we are looking to his subjective intent ("speaker's meaning") or because we are judicial activists who ignore fidelity to the text in favor of our own values. Instead, we read Spooner charitably, because charity is just as much a usage norm as, say, the rule that "and" should be read conjunctively and "or," disjunctively.

Does Larry disagree? I cannot say for sure. But, if he agrees, then why are our norms not built into interpretation from the outset?

Posted by Rick Hills on January 7, 2010 at 05:51 PM in Constitutional thoughts | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference An AALS smackdown on philosophy of language?:


First, I have to thank everyone involved letting me take part in this panel discussion, particularly as my thoughts on interpretation and construction (and the constitutional canon) are somewhat outside (perhaps naive to?) the theoretical parameters that other, more experienced, panelists have established. Thus, I probably should just leave well enough alone, but…

I can’t help myself, and so I want to add that I am very sympathetic to Rick’s point about the principle or “rule” of charity in interpretation; but I think this is but one example of many similar kinds of interpretive “rules” that work to undermine the interpretation/construction distinction as typically posited. I think Grice’s work on implicature is a persuasive account of some, but not all, these kinds of rules. In other words, I don’t think that Grice’s account of maxims and conventions can give us any comprehensive or definitive guidance on what should count as an act of interpretation vs. construction, and to this effect I would point out: (1) that law is still a language-game and we all necessarily “play” it in different ways—including the ways that we “interpret” utterances; and (2) these ways necessarily include *individualized* applications of the charity principle, and many other general, non-quantifiable kinds of rules. The basic point being that I take Grice’s maxims as simply specific illustrations of a much larger and indeterminate body of rules governing our understanding of utterances, and I may select different rules—or apply the same rules slightly differently—than would others in response to the same utterance…and I would still be within the acceptable practice boundaries of the particular game. It is this practice, in a rough way, that allows us to learn something when told that “the morning star is the evening star”. And, ultimately, it is this individualized selection and application of rules that, to me anyway, is the essence of what we call construction.

Now I understand that this is perhaps a more limited kind of construction than what happens when we translate semantic content into rules of law, and perhaps there is something to be gained in differentiating between the less and more expansive kinds of construction. This I take to be Larry’s point about the forest and the trees. To the extent, though, that we want to view the “interpretation” part of the equation as somehow objective or foundational, however, I simply don’t think that is an accurate description of the practice. That is, in all but a small class of uncontroversial kinds of interpretation (say, “2 senators per state” as opposed to “equal protection”), the individual selection and application of interpretive rules is often just as significant as the act of translating semantics into policy. And if we are looking for definitive ways to *justify or explain* the choice of interpretive rules we may make on particular occasions, I think that this, too, is regressive endeavor in which we will ultimately find our spades turned—there is simply no useful accounting for all the individual variables of context, shared knowledge, “trees in front of rabbits”, etc…

Thus, I think our efforts are ultimately better directed to description than to explanation or justification--as unsatisfying as it may be, we simply have to leave those kinds of judgments up to the evolving norms of the practice.

Posted by: Ian Bartrum | Jan 15, 2010 4:40:53 PM

"Charity begins at home." That is, today's home, not the home of hundreds of years ago. So where does that leave originalism, whether the old, new or in between (Goldilocks)? Has the search for the "Holy Grail of Constitutional Interpretation" come to a fork in the road with "construction"? And can constitutional "construction" be other than "living"? Or does originalism also apply to constitutional "construction"? How soon can we expect a determination by SCOTUS of the interpretation-construction distinction in constitutional law and would it be 5-4? (Maybe Chapter III of Francis Lieber's "Hermeneutics" originally published in the late 1830s will be helpful, as it was more contemporaneous with the adoption of the Constitution than the current discussion.)

Larry Solum points out "It is important that we not lose sight of the forest for the trees." I would add that we should not lose sight of the trees for the forest. Either way, we, the audience of the panelists, may still get lost; guys, keep on talking so we can find our way. And God bless SCOTUS!

Posted by: Shag from Brookline | Jan 14, 2010 6:18:57 AM

Thank you to both Rick and Brian for the helpful points. My reply to Rick was given in two minutes at the end of my remarks, and I am sure it could have been made much clearer had time permitted.

I agree with Brian (and there really is no room for disagreement about this) regarding the content of the principle of charity. The principle of charity (as reformulated by Davidson following Quine) does not require that we interpret each sentence that someone utters is true. (That version of the principle of charity would be silly and obviously false.) I made a further point. Sometimes the principle requires we interpret particular sentences as false. I referred to this point very briefly in my remarks in New Orleans, and the relevant passage from "Semantic Originalism" is as follows:Donald Davidson extended Quine’s account from translation to interpretation (where interpretation means understanding within a single natural language) and he collapsed Quine’s two constraints into one principle of charity which is maximize agreement. Dagfinn Föllesdal correctly identified a problem with Davidson’s statement (his precise formulation) of the principle of charity, which he explained as follows:
I am together with a person who speaks a language which I do not know, but would like to learn. He frequently uses the phrase 'Gavagai' and I have formed an hypothesis that it has to do with rabbits. While we are in a forest and I note a rabbit I try out the phrase 'Gavagai'. However, my friend dissents. According to Davidson's thesis of maximizing agreement this would be a reason against my hypothesis that 'Gavagai' should be translated by 'Rabbit'. If I now discover that there is a big tree between my friend and the rabbit, I immediately have an explanation for our disagreement: I take it for granted that my friend, like me, is not able to see through trees and that he therefore does not think that there is a rabbit there. I even take my friend's dissent as confirming my hypothesis, I do not expect him to believe that there is a rabbit there.
Thus, we can articulate a modified principle of charity as follows: “Maximize agreement where you expect to find agreement.”
Thus, in the example discussed in New Orleans of Emma and the eggs in the refrigerator, I offered the following hypothetical:
You (the interpreter) observe Emma as she is preparing to cook a meal. She looks in the refigerator and counts the eggs. She then leaves the room. While she is away, you thoughtfully remove the eggs from the refrigerator and put them in a mixing bowl. Emma then returns with Frank who asks, "Emma, do we have any eggs." Emma cannot see the mixing bowl, which is around the corner, and says, "There are twelve eggs in the refrigerator."
If the principle of charity required that we interpret her statement as true, we would have to interpret her statement as meaining there are no eggs in the refrigerator, but this is not what the principle requires. As Föllesdal's reformualted version makes clear, this is the kind of case, where charity actually requires that we interpret Emma's statement as untrue.

As applied to legal texts, a very similar point holds. Of course, legal texts do not make assertions and hence are usually not characterized as truth apt (a provision of the Constitution cannot be "true" or "false"). But by analogy with the case of truth, the modified princple of charity might require us to interpret the whole system of legal norms as "making sense" or "normatively intelligible" from our point of view. (There is more to say about the question whether this is so, but let us assume it arguendo for the purposes of this discussion.) But in the case of particular constitutional provisions, our knowledge of the differences in perspective between us (contemporary interpreters) and the authors of the constitutional text (roughly the framers and ratifiers) will in some instances require an interpretation of a provision that we view as morally unattractive. In New Orleans, I have the example of the rule that gives each state two Senators regardless of population--which we may now believe is inconsistent with the political equality of persons.

This brings me to Rick's comment in reply to Brian, which he summarizes at the end as follows:
So charity, in my view, has two components: (a) The stipulation that the interpreter and speaker must agree on most things and (b) the presumption that, when in doubt, individual sentences are to be construed to be true according to the interpreter's lights. The second proposition follows from the first, no? If I assume that most members of set X must have property Y, then surely I must believe that, ceteris paribus, it is more likely than not that any particular member of X also has property Y.
From the modified principle of charity (the Föllesdal version), it does not follow that there we must always presume that a particular constitutional provision is normatively correct. In the example of Emma and the eggs, given what I know about Emma's situation, I should presume that her statement will be false. In the case of the Constitution, given what we know about the framers and ratifiers, there are particular constitutional provisions that we should presume are normatively unattractive.

It is important that we not lose sight of the forest for the trees. The point of the interpretation-construction distinction is that "interpretation" (which aims at the recovery of linguistic meaning or semantic content) is a distinct activity from "construction" (which produces constitutional doctrine or legal content). The correctness of an interpretation depends on linguistic facts (about patterns of usage) and contextual facts (about the circumstances of utterance). Rick's invocation of the principle of charity was intended to show that normative concerns enter into interpretation. Both Brian's criticism of Rick and my reply in New Orleans show that the particular argument that Rick made in New Orleans does not go through. As Brian put it, Rick would "need a stronger claim."

Of course, there is more that can be said. I look forward to reading Rick's Constitutional Commentary piece. I am sure it will be illuminating.

Posted by: Lawrence Solum | Jan 13, 2010 9:31:56 PM

Thanks, Brian, for the thoughtful comments: As I am supposed to write up my views on charity for Constitutional Commentary as part of that journal's publication of the panel proceedings, I really appreciate the help.

I accept your indisputably true statement that it is not a requirement of the principle of charity that all of a speaker's beliefs and values conform to the interpreter's beliefs and values: Any such a position would make it impossible for an interpreter ever to infer that a speaker disagreed with her, which is exactly the opposite of Davidson's position: Davidson argued, to the contrary, that a background of agreement on most matters was necessary in order to interpret someone else's sentence as a disagreement with one's own position. So, yes, it surely is the case that "[w]e can think the Framers morally wrong about various issues, and still interpret them charitably."

Here's is my fuzzy notion: Charity is an interpretative resource not dissimilar from, say, rules of logic or true facts. While a speaker can, of course, utter an illogical (say, internally inconsistent) sentence, we would normally presume, absence evidence to the contrary, that a speaker's sentences were internally consistent. Likewise, while a speaker can certainly utter a sentence that is factually untrue, we would presume that, absent evidence to the contrary, that a speaker's sentences are, in fact, true. So, contrary to Larry's view -- if I understand him correctly -- I maintain that interpretation and the inference of semantic content involves the presumption of shared belief and value.

My example of "actual enumeration" in Article I, section 3 was intended to illustrate how we naturally presume that the Constitution's ratifiers share our beliefs about what it means to make an accurate count of population. That 18th century census takers did not use statistical sampling is not a very good argument that statistical sampling falls outside their concept of "enumeration," because we naturally and appropriately impute to that 18th century concept a 21st century goal -- the goal of producing an accurate count. How else could we make sense of what they were doing in visiting people's houses, taking down names on pieces of paper, tabulating results in bound volumes, etc? Of course, all of this assumption of shared ends is just a rebuttable presumption, but it is a semantically appropriate presumption, akin to avoiding the gratuitous assumption that an otherwise ambiguous sentence is internally inconsistent. I think that Justice Scalia implicitly recognizes that interpretation always involves a presumption of shared belief when he buttressed his argument that the 18th century concept of "enumeration" excluded sampling by offering a normative end for such a limit on the concept that we find "sensible" -- namely, constraining legislative discretion.

So charity, in my view, has two components: (a) The stipulation that the interpreter and speaker must agree on most things and (b) the presumption that, when in doubt, individual sentences are to be construed to be true according to the interpreter's lights. The second proposition follows from the first, no? If I assume that most members of set X must have property Y, then surely I must believe that, ceteris paribus, it is more likely than not that any particular member of X also has property Y.

Posted by: Rick Hills | Jan 13, 2010 8:17:58 PM

Rick, I enjoyed your remarks at the panel in question, and I commend you for actually adhering to time limits, which is all too rare at AALS events. I heard Larry Solum's reply, but left before the session was over, so don't know if you or anyone in the audience replied on the principle of charity issue.

It seems to me that you were clearly correct that the principle of charity is basic to all interpretation. Larry, if I understood his reply, wanted to deny that the principle of charity always applies based on the assumption that the principle of charity would require us to interpret every statement as true (I took this to be the point of his example of interpreting Emma to say there were 12 eggs in the fridge, which she had reason to believe true, though it was no longer true unbeknowst to her). But the principle of charity certainly does not require taht we interpret any particular utterance of a speaker as true, only that we impose a semantics on his or her statements *as a whole* that renders them intelligible by our lights--that's compatible with lots of his or her statements being false even by our lights. So Larry's rejoinder seemed to me non-responsive.

On the other hand, I'm not sure how much damage the point about charity does to the innterpretation/construction distinction that Barnett et al. are interested in. If the norms of intellibility necessary to make sense of speaker's utterances in 1789 or 1791 demanded that they conform to our moral and political views, then this would be a problem for the originalists in question, but this is also not a demand of the principle of charity as far as I can see. We can think the Framers morally wrong about various issues, and still interpret them charitably, e.g., in assuming that "religion" refers to something we could recognize as religion and not to backgammon, that "necessary" has something to do with required or essential and not arbitrary, that commerce has something to do with business and economic activity and not silent meditation, and so on. So to cause trouble for the new originalists, you need a stronger claim about the principle of charity, namely, that the norms of intelligibilty it imposes on constitutional interpretation render meaningless the idea of recovering an original public meaning, in particular, an original public meaning we find morally odious or objectionable or discredited.

Posted by: Brian | Jan 10, 2010 8:13:19 PM

The comments to this entry are closed.