This is the fifth post in a series (Part One, Part Two Part Three Part Four Part Five Part Six Part Seven Part Eight Part Nine) that replies to Stephen Griffin's very helpful comments about Semantic Originalism on Balkinization. In this post, I reply to the fifth of Griffin's comments. Quotes from Griffin are in ordinary roman type, indented, and black. My discussion is in blue.
The core of "Semantic Originalism" is a theory of linguistic meaning (or "semantic content") of the Constitution. That theory draws on the work of Paul Grice and argues for two claims about constitutional meaning:
The Fixation Thesis is the claim that the meaning of a constitutional utterance (a clause in the original constituiton or an amendment) is fixed at the time of framing and ratification. For example, the phrase "domestic violence" now means "spouse abuse" but it did not have that meaning in 1789: the subsequent change in linguistic practice does not change the semantic content of the Constitution.
The Clause Meaning Thesis is the claim that the meaning of the Constitution is the conventional semantic meaning of the words and phrases given the linguistic facts about usage and syntax at the time each clause was uttered, but subject to some modficiations including (1) publicly available context of each clause including the whole Constitution, (2) the division of linguistic labor that accounts for terms of art such as "Letters of Marque and Reprisal," (3) constitutional implicature, the idea that the consitution can imply things it does not directly say, and (4) constitutional stipulations, some phrases such as "House of Representativies" have meanings that were created by the Constituton itself.
Griffin's fifth question about "Semantic Originalism" was:
5. Solum identifies Dworkin's theory of constitutional interpretation as a competitor to his own. But I didn't find one of Dworkin's most famous claims clearly addressed. This is the idea that the abstract provisions of the Constitution, such as the equal protection clause, are not vague because they directly invoke important concepts of political morality. I know this idea was developed in greater detail in Chris Eisgruber's 2001 book, perhaps also in Jim Fleming's. I would be interested to see Solum's response to these second-generation Dworkinians, as well as to this characteristic claim of Dworkinians.
This is a deep and important question. I'm sure that Griffin will understand if I don't fully meet the challenge (answer Dworkin, Eisgruber, and Fleming) in one blog post, but I will do my best to address the central challenge, which I take to be twofold. First, how does famous distinction between concepts and conceptions affect the theory of meaning offered in "Semantic Originalism." Second, how is this related to the idea that the Constitution may invoke "moral kinds" such as equality or cruelty--the meaning of which is fixed by moral reality rather than conventional usage.
I want to be very clear that I am not offering a full answer to Dworkin in this post. Dworkin's theory operates at several different conceptual levels. Dworkin addresses "meaning" and "interpretation," but the primary thrust of his theory goes to a different question--the relationship of institutional history (including but not limited to legal texts) to legal content. Dworkin's arguments on that topic are addressed in "Semantic Origianlism" in depth in connection with the contribution thesis--the claim that the semantic content of the Constitution contributes to the legal content of constitutional law. I will not be addressing that topic in this blog post.
Concepts and Conceptions
Let's go back to the concept-conception distinction. Here is how I describe that distinction in "Semantic Originalism":
So far as I know, the concept-conception distinction originates with Essentially Contested Concepts, a paper written by the philosopher William Gallie in 1956. The core of Gallie's argument was the idea that certain moral concepts are "essentially contested." "Good," "right," and "just," for example, are each moral concepts which seem to have a common or shared meaning. That is, when I say, that the alleviation of unnecessary suffering is good, you understand what I mean. But it may be that you and I differ on the criteria for the application of the term "good." You may think that a state of affairs is good to the extent that it produces pleasure or the absence of pain, while I may think that the criteria for "good" make reference to the conception of a flourishing human life as a life of social and rational activity lived in accord with the human excellences or virtues. We share the concept of "good" but we have different conceptions of what constitutes a good life.
Sometimes, when there is this sort of disagreement, we want to say, "Ah, you and I are referring to different concepts." If by "cause," you mean "legal cause," whereas I use "cause" as a synonym for "cause in fact," then we are using the same word to refer to two different concepts. If this were the case, then "cause" would be ambiguous and once the ambiguity had been identified, we would come to realize that we were talking past one another.
But in the case of "good," we seem to be using the same concept. I think that the good really is human flourishing and not pleasure; you have the opposite opinion. So we are contesting the meaning of the concept "good," and each of us has a different conception of that concept. We are not talking past one another; we are disagreeing. Gallie thought that some concepts were essentially contested. That is, Gallie believed that some concepts were such that we would never reach agreement on the criteria for application of the concepts. If a concept is essentially contested, then it is in the nature of the concept that we disagree about the criteria for its application.
So it is possible that the Constitution uses words or phrases that refer to essentially contested concepts. If this were the case, then it might be argued that the constitution does not fix the meaning of these concepts--they are "up for grabs." Things get very subtle at this point. I have two distinct points to make, and I will do my very best to make them clear:
First, the question whether the Constitution does in fact deploy a contested concept is a question of lingustic fact. Gallie's theory of essentially contested concepts was offered as explanation for patterns of usage: it is not a master template that tells us how we must use particular words and phrases. The inquiry whether a particular word or phrase in the constitution is an essentially contested concept involves two stages:
Stage One: Does the constitution use a word or phrase that might refer to an essentially contest concept? For example, is the word "cruelty" used to refer to the "concept of cruelty" and is that concept "essentially contested"? If the answer to this question is "no," then there is no concept-conception distinction problem. If the answer is yes, then we go on to Stage Two.
Stage Two: Given the linguistic facts (the patterns of conventional usage and the context of constitutional utterance) was the public meaning (or conventional semantic meaning) of the word or phrase the essentially contested concept or does linguistic practice and the context of constitutoinal utterance suggest that the meaning of the word or phrase was a particular conception of that concept?
Stage Two is critically important if our inquiry is into the linguistic meaning (or semantic content) of the Constitution. The words or phrases that are used to refer to contested concepts are ambiguous in the following sense: they can be used to refer to the concept or they can be used to refer to a particular conception of the concept. For example, if a hedonsitc utilitarian were to use the word "good" in a context such that it was clear that the usage of the word was internal to the theory, then it would be clear (as a matter of linguistic fact) that the word "good" referred to something like "presence of pleasure and absence of pain". It would simply be a linguistic error to treat this usage of good as referring to the concept. On the other hand if the same writer said: "What is good? Is good pleasure and the absence of pain, or is it something else?"--it would then be clear that the general concept was invoked.
This point leads to another crucial fact about Stage Two. The question whether a word or phrase refers to an essentially contested concept depends is time bound. The very same word could refer to an uncontested concept at T1 and then (because of changes in linguistic practice) come to refer to a contested concept at T2. For example, it is possible that the phrase "equal protection of the laws" did not refer to the the contested concept of equality at the time the 14th Amendment was adopted, but that the same phrase has come to refer to that contested concept in specialized academic discourse today. Whether it did or didn't refer to the contested concept at the time the 14th Amendment was proposed and ratified is a matter of linguistic fact, and such facts are time bound because usages change.
Second, if the Constitution does employ contested concepts, then the theory of semantic originalism tells us that the meaning of the relevant clause is, in fact, the contested concept and not a particular conception of that concept. That is, semantic originalism can and should absorb the concept-conception distinction, to the extent that the distinction actually applies after application of the two stage procedure identified above. This naturally leads to the following question:
If the conventional semantic meaning of the Constituiton does employ a contested concept, how can "Semantic Originalism" determine the semantic content of the contested concept?
But of course, the answer to this question is built into the theory of semantic originalism. Translating a contested concept into legal content requires construction and not interpretation. Recall that constitutional construction enters the scene when constitutional interpretation makes its exit. If the semantic content of a word or phrase used in the Constitution is a contested concept, then interpretation tells us that much but no more. We then need a theory of constitutional construction to do the additional work needed to craft rules of constitutional law and to apply those rules to particular cases.
At this point, I suspect that many readers will be formulating an objection that might go something like this:
But that isn't what Dworkin meant by using the concept-conception distinction. Dworkin believed that there was a best or true conception of the concept, and his theory requires that we interpret the constitution to refer to what "equality" or "cruelty" really is.
We are now in very deep waters. In "Semantic Originalism," I demonstrate that this, in fact, is not Dworkin's position. Dworkin actually does not have a theory about the actual linguistic meaning of the words and phrases in the Constitution. Instead, Dworkin believes that those meanings are only the raw material from which Hercules constructs the theory that best fits and justifies the law as a whole. The content of Hercules's theory determines leal content. Once Hercules has taken the semantic meaning of the Constitution into account in the development of his grand normative theory of all of American law, he is then free to leave the semantic content behind. In other words, Hercules can decide to construct a rule of constitutional law that assigns to the Equal Protection Clause the legal content created by a theory of equality even if the semantic content of the Equal Protection Clause did not refer to the contested concept of equality.
Dworkin's Hercules is unbound in the following sense: so long as his theory fits and justifies instituitonal history as a whole, Hercules is free to disregard the linguistic meaning of particular legal texts. The labors of Hercules are truly monumental, he must muck out the stable constituted by the entire history of American law, but (lucky for him) he unencumbered by chains of fixed semantic content.
My description of Dworkin in the prior paragraph may surprise some readers. That surprise has many sources. One possible source is that Dworkin's views may have changed over time. Early Dworkin, e.g., the Dworkin of Hard Cases, may have believed that the move to contested concept was sanctioned by linguistic meaning. Late Dworkin, e.g., the Dworkin of Law's Empire, may have move to the more radical view that legal content is identical to the content of the theory that best fits and justifies the law as a whole. This possible shift in position is compounded by Dworkin's notorious obscurity of expression. Dworkin formulates his ideas in prose that reads easy but thinks hard. By "reads easy" I mean that Dworkin formulates his ideas using words and phrases that are deceptively familiar. By "thinks hard" I mean that the content of Dworkin's system of thought is determined by the architectonic of his whole argument, such that the meaning of individual passages formulated in familar words and phrases takes on a technical meaning when interpreted in light of the theory as a whole.
Natural Kinds
In fact, many sophisticated readers of Dworkin assume that his discussion of the concept-conception distinction points us in the direction of another, quite different but equally important challenge to semantic originalism--the move to natural kinds. Here is how I introduce the discussion of that topic in "Semantic Originalism":
The account of constitutional meaning that we have pursued so far ties clause meaning to conventional semantic meaning, which is fixed by usage. But is this account of meaning correct? Or if correct for some words and phrases, are the others for which a different account of meaning is required. A fundamental challenge to the notion that meaning is conventional is posed by the work of Saul Kripke, Hilary Putnam, and others on the notion of a natural kind. Following Morris, let us define a natural kind as follows:
Natural Kind: A natural kind is a kind whose identity as a kind is fixed by reality, and not by human interests or concerns.
Take the word "gold" one sense of which refers to the shiny metal. Gold is a natural kind the identity of which is fixed by its atomic structure. What counts as gold in this sense depends on our best scientific theory of the elements. Whether a given object really is composed of gold does not depend on usage. It would be quite possible for everyone to be mistaken about the question whether something that seems very much like gold really is gold. So this seems like a case in which conventional semantic meaning does not fix reference. Once we had dubbed gold (the element) with the English word "gold" as its name, then reference was fixed by nature and not conventions.
How does the notion of a natural kind relate to Dworkin's concept-conception distinction? Some readers may think that the two ideas are almost equivalent, but that would be a huge mistake. "Essentially contested concepts" are the polar opposite of "concepts with fixed essences." If a concept is essentially contested, its meaning is not fixed by reality. If a concept refers to a natural kind, then it cannot be the case that it is essentially contested. Many legal theorists run concept-conception and natural kind together, but that move results in horrendous confusion.
So now that we know what a "natural kind" is, how does the notion of a natural kind apply to the Constitution? Does the constitution employ natural kind terms? There may be some natural kind terms in the text of the Constitution, for example "person" and "speech" may be natural kinds. If the constitution does contain natural kind terms, then the core theory would need to be modified. This modification would not challenge the essential thrust of either the account of clause meaning or the distinction between interpretation and construction. The meaning (semantic content) of natural kind terms is determinate. Metaphysically, there is always a âfact of the matterâ regarding the correct application of a natural kind term. Epistemologically, there may, of course, be uncertainty. The essential thrust of the view argued for this Article requires only that the Constitution have semantic content that is not indeterminate. If the semantic content of the Constitution were largely determined by natural kind terms, that requirement would be satisfied.
This modification of the theory is not as radical as it might seem on first blush. The idea that some words refer to natural kinds is not inconsistent with the notion that the relationship between the word and the natural kind is fixed by convention or usage. Indeed, that conclusion seems inescapable. Because words change their meaning it is possible for a given word to refer to one natural kind at one time and then change its meaning as a result of accumulating errors in usage.
The constitution includes few words or phrases that are plausibly understood as natural kinds identified by science. But it does include several words and phrases that could be understood as having moral content. For example, "freedom of speech," "cruel and unusual punishment," and "equal protection of the laws." Some philosophers have argued that the meaning of moral language is fixed by moral kinds. This is not the occasion to interrogate the correctness or plausibility of this position, but we do need to ask whether and how this view, if true, would affect the theory of constitutional content offered here.
If the constitution employs moral kind terms, then moral reality and not semantic conventions fixes their meaning given conventions that attach a word or phrase to the kind. Conventions do the attaching of the word to the moral kind, but moral reality and not the convention fixes the nature of the kind. In this case, the constitution has semantic content but that semantic content includes (but is not exhausted by) content that is independent of the original public meaning of the constitution.
This possibility creates a set of options for semantic originalism as a theory of constitution meaning. The first option is to absorb "moral kinds" into the theory of clause meaning. In that case the four modifications of clause meaning identified above, would be joined by a fifth modification, and the theory would proceed as modified.
Finally, there is at least the logical possibility that there exist "legal kinds," metaphysically real entities that fix the reference of constitutional terms like "Republican form of government," "privileges and immunities," or "full faith and credit." The view that comes closest to this is expressed by Michael Moore in the context of the term "malice" in the criminal law:
[A] second level realist is committed to something named "malice" that is neither a natural nor a moral kind. Rather than calling this a "legal kind," however, I prefer the label "functional kind." The legal realist view of "malice" is no different than the general realist understanding of terms like "lawyer", "knife", "vehicle", or "paper weight": all these terms refer to kinds of things whose essence is not given by their structure but by their function.
For the purposes of "Semantic Originalism," the existence of "legal kinds" or "functional kinds" would not require modification of the main thesis--that the constitution has semantic content, that its content has the force of law, and that fidelity to that content is warranted absent overriding reasons of morality. The fixation thesis remains intact, and the clause meaning thesis would be reconstructed to incorporate the role of linguistic convention in creating the necessary relationships between words and phrases, on the one hand, and functional or legal kinds, on the other.
One more thing, and it is very very important. My stance towards natural, moral, and legal kinds in this blog post is neutral. All of these ideas are controversial, but the notion of a natural kind draws on a powerful set of views in metaphysics--associated with Saul Kripke, Hilary Putnam, David Kaplan and others. Moral and leal kinds are much more controversial. If you are tempted to incorporate those ideas in your theory of law, you have a lot of homework to do. The best place to start is the work of Michael Moore:
Michael S. Moore, Educating Oneself in Public: Crititical Essays in Jurisprudence
Michael S. Moore, The Semantics of Judging, 54 S. CAL. L. REV. 151 (1981).
Ideally, all of the issues raised by the Kripke and Moore would be addressed in this blog post or at least in "Semantic Originalism." Practically, this requires full consideration of the most fundamental questions in contemporary philosophy of language. Because the existence of natural, moral, and functional kinds can be incorporated within the structure of semantic originalism without fundamental alteration in the structure of the theory, I will not attempt the impossible--settling the deepest questions of metaethics and jurisprudence in a blog post.
Conclusion
I am about to point out ways in which this post does not fully address all of the issues on the table, but I don't want to allow modesty to obscure the real bottom line. Neither the concept-conception distinction nor the notion of a natural kind poses a fundamental challenge to the fixation thesis. These ideas might require some modification of the clause meaning thesis, but they do not affect the fundamental point--that the meaning of a term is fixed by linguistic facts.
Griffin set me a very difficult task in his fifth question, and I have left some important issues unaddressed. Much more about these matters is said in the full text of "Semantic Originalism," but Griffin rightly points out that my current draft does not yet account for important work done by Jim Fleming and Chris Eisgruber. Their work is important, sophisticated, and must be addressed by me (or anyone who wants to move the ball forward in constituitonal theory). Let me recomend three books, one by Fleming, one by Fleming and Sot Barber, and a third by Eisgruber:
Constitutional Interpretation: The Basic Questions by Sotirios A. Barber and James E. Fleming
Securing Constitutional Democracy: The Case of Autonomy by James E. Fleming
The aim of this post was to illuminate the basic nature of the issues and clearly distinguish the concept-conception distinction from the notion of a natural kind. Accomplishing those tasks clears the brush, enabling a clearer picture of actual claims made by "Semantic Originalism."
Griffin's sixth question raises the issue of "interpretive pluralism," and I will address that topic in my next post.