A Textualist Approach to Legal Interpretation

Here is a draft of a paper on legal interpretation that is similar to origanilism in emphasizing the meaning of the text as enacted, but differs from standard forms of originalism, especially with respect to normative words and phrases in the text and in disavowing as determinative drafter and legislative or ratifier expectation and intent. The first section by itself might be of interest if you are a little uncertain about the difference between ambiguity and vagueness, or types of ambiguity, or what problems of interpretation there are beyond ambiguity and vagueness.

A Textualist Approach to Legal Interpretation

Lawrence Crocker

Draft 6/17/14

1. The Tasks of Interpretation

Most specific issues in legal interpretation are about vagueness. That is, the dispute is over what is within the coverage of a legal word or phrase and what lies outside the coverage. A stock non-legal example, “tall woman” applies to a woman who is 6’8, and does not apply to a woman who is 4’10”. About a woman who is 5’9”, we will want to know more of the context in which “tall woman” is used before committing ourselves on the in/out/borderline determination¹. 

 

Where borderline cases are frequent or important, we say that the word or phrase is “vague.” Examined closely, all words and phrases are in some respects vague with the exception of the terminology of mathematics and other precise and regimented vocabularies. An inquiry into the scope of “due process” or “freedom of speech” or “right to bear arms” is usually about eliminating part of the vagueness of these phrases.

The second familiar interpretive problem is the elimination of ambiguity. To eliminate vagueness we draw an in/out line across shades of gray. To eliminate ambiguity, we determine in favor of one of two (occasionally of three, rarely of four or more) distinct alternatives.² 

Ambiguities are sometimes divided between semantic ambiguities and syntactic ambiguities. An example of a semantic ambiguity from the illustrious twentieth century philosopher of language, W.V. Quine, attaches to the “bore” in “Our mothers bore us.” The illustrious twentieth century plunderer of language, Groucho Marx, contributed an example of syntactic ambiguity: “One morning I shot an elephant in my pajamas. How he got into my pajamas I'll never know.”

“Age of 35 years” might pose a problem of semantic disambiguation for someone aware of both the western convention that a baby is one year of age after living 365 or 366 days and the east Asian convention that a baby is one at birth and two after making it through that first year. 

 

Prior to the adoption of the 25^th Amendment, there was arguably an ambiguity in the US Constitution as to whether, upon the President’s death, the Vice President became President, or only assumed the powers of the President.

In Case of the Removal of the President from Office, or of his Death, Resignation, or Inability to discharge the Powers and Duties of the said Office, the same shall devolve on the Vice President, and the Congress may by Law provide for the Case of Removal, Death, Resignation or Inability, both of the President and Vice President, declaring what Officer shall then act as President, and such Officer shall act accordingly, until the Disability be removed, or a President shall be elected. (Article II, Section 1)

In fact I think it is clear, upon reading the whole provision, that the Vice President would assume only the powers and duties and not the presidency. So I would not count this as a serious ambiguity, were it not that Presidents Tyler, Fillmore, Arthur, Theodore Roosevelt, Coolidge, Truman, and both Presidents Johnson apparently thought otherwise.

An issue of syntactic ambiguity attends the Second Amendment phrase “A well regulated militia being necessary to the security of a free state.” Hotly debated is which of handful of different ways this phrase affects, if at all, the following right to keep and bear arms.

In addition to disambiguation and the clearing up of vagueness, there is an important, if less easily characterized, component to legal interpretation. It is nailing down the implications of the legal language. Of course all interpretation is a matter of getting the implications of the language right. In removing vagueness or ambiguity we are working out implications. An example of implication in the narrower sense is application of general language to a specific case. “All dams over twenty feet in height must be inspected annually.” “The dam at issue is 25 feet high, therefore it must be inspected annually.” 

 

The sounding out of implication deserves recognition as an important component of interpretation because of those cases in which the implication is less direct and potentially controversial.³ John Marshall found judicial review to be implied by Article III as well as structural features of the United States Constitution. Arguments supporting stronger or weaker states’ rights are, again, often predicated largely on the implications of the structure of the federalism as set out in the Constitution. In Griswold v. Connecticut, Justice Douglas concluded that “specific guarantees in the Bill of Rights have penumbras formed by emanations from those guarantees,”⁴ inter alia, supporting a constitutional right to privacy. This last exercise in finding the implications of the text has been regarded by both friend and foe as aggressive.

       2. An Idealization of Legal Interpretation

A judge, or anyone else, who is interpreting a statute or constitution, may need to disambiguate any one or more of the ambiguities of the text, make precise any one or more areas of vagueness, and work out any one or more implications as they affect the issue for which the interpreter is looking to the legal text. Some points of ambiguity, vagueness, or implication will never need to be worked out, but the nature of the application of law and of disputes about law leave few interpretive issues immune from the possibility of coming into issue. For this reason, the ideal of legal interpretation would be the elimination of every possible issue stemming from ambiguity, vagueness, or the scope of implication. 
 

This would amount to the translation of the statute or constitution into a language that was completely free of ambiguity, vagueness, and unresolvable issues of implication. The idea of a language with these characteristics was worked out in the early mid-twentieth century by the Vienna Circle logical positivist Rudolf Carnap. Ambiguity and vagueness were to be banished by explicit, univocal definitions of all terms. Implication was taken care of by complete axiomatization and formalized rules of inference. 

 

A first thing to notice about the ideal of translating legal texts into a regimented language of this sort, is that it would be far from ideal for most other kinds of interpretation. A good translation of a poem by Goethe into English would not eliminate but preserve the ambiguity and vagueness of the German original. If you read the preamble of a constitution for its rhetorical force, the last thing you want to do is to redo it in a Carnapian language. It is only if you may be required to remove any ambiguity or vagueness or work out any question of implication that this kind of translation is the ideal. 

 

A second thing to notice is that my ideal language is its impossibility. We can construct Carnapian languages, but only if we are willing to restrict ourselves to very well behaved subject matters. A language rich enough to be the medium expression of law in the real world, could not be regimented in this way. Think of a single statute. It should be a finite, if perhaps lengthy, task to find and eliminate every ambiguity. Eliminating vagueness, however, might well be a task without end. Each vague term may only require the drawing of a single in/out boundary, but it may be a boundary through a very complex space of possibilities. Implications will often be infinite, and not necessarily infinite in any tidy way.

Still the idea of translating legal text into a language free of ambiguity, vagueness, and with determinate implications is the right ideal for purposes of law’s application and adjudication. It is no more a real possibility than is an ideal gas something found in the real world, but both provide methodological guidance.

I understand this idealization of legal interpretation to be neutral with respect to most particular theories of the right way to go about the process of translation. ⁵ It is not an idealization that automatically yields any form of textualism. Those who believe that the interpretation of legal texts should look to old traditions or new public opinion or a comprehensive political theory or ideology can incorporate these ideas into their translation manual. 

 

That the legal interpreter is to start with text and go on to produce from that text a translation ideally suited to the resolution of legal disputes, is, however, obviously congenial to a textualist approach.

3. Textualism and Legitimacy

I argue in what follows for one specific method of legal interpretation, which I think is fairly called “textualism,” although it will differ in some respects from other forms of textualist legal interpretation that have been elaborated, defended, and attacked. My commitment to textualism comes primarily from political theory. There is a strong normative argument for textualism as an interpretation of statutes and constitutions passed by a voting citizenry or its representative bodies⁶. What was before the voters or ratifiers was a text. It was a text that became law and behind which lies whatever legitimacy the law making process possessed. 

 

I concede that it is not any form of text-determinative interpretation that comes first to the popular mind when the question is interpretation of law, especially when it comes to construction of the United States Constitution. Instead we think of interpretation based on framer intent and expectation. 

 

Drafters undoubtedly did have some specific intentions in selecting the words and phrases of their drafts and they had some expectations as to the real world effects of the enacted law. Theorists skeptical of drafter intent, however, have observed, rightly, that intents may have been only partially formed, that expectations may have surveyed only a few possible futures, that anticipated futures may not have corresponded to the actual futures, that plural drafters may have had conflicting intentions and expectations, and that the intentions and expectations of drafters may not have been known to the voters or may even have been misapprehended. These observations support strong arguments against drafter intent as the interpretational touchstone. The decisive argument, however, is the simplest. Neither the expectations nor the intentions of the drafters were passed or ratified – only the statutory or constitutional text was.

The understanding and expectations of the citizens or representatives whose votes turned the text into law (or were the penultimate step towards same) are not quite so clearly irrelevant to an interpretive method based on considerations of legitimacy as is drafter intent. Each voter’s understanding and expectation no doubt entered directly into his or her decision whether to vote yay or nay. Yet the voters who deliberated with clarity would have understood that their predictions of the effects of the law might well be made erroneous by future circumstances, including the circumstance that the law would be interpreted by those governed by it, those who administered it, and those who adjudicated it. Different voters might well have a different understanding of the implications of the text. One thing that a voter knew with confidence was common to all his or her fellow voters and would remain common to all future interpreters was the printed text. That would remain exactly as it was when voted upon until amended, repealed, or overturned.

4. Text, Meaning, and Textualist Requirements

The observant reader will have noticed that in the text constancy claim I just made, I added in the word “printed.” Without that addition, I could not have guaranteed text constancy. It is important that the printed text of statute or constitutional provision endures through time, absent copying or printing errors. But does the text necessarily remain unchanged? Marks on paper (or pixels on a screen) are not by themselves text. They become text only when meaning is associated with the marks. It caricatures textualism to leave out meaning. The meaning of the text is a matter of the way the text is to be understood. 

 

You may sense a danger of circularity at this point, inasmuch as the grand interpretive question is how the text is to be understood – how its meaning is to be found. Circularity is avoided because textualism, in the sense I examine here, imposes three conditions upon the task of finding meaning. First, it looks for meaning at the time when the text became law. Among its other justifications, this follows from the concern for legitimacy. It is the text with the meaning that it had at this moment of time that was made into law. Second, it should embrace meaning as understood by the entire linguistic and political community subject to the text when it became law. Third, it should look, as nearly as possible, to the essential meaning, core meaning, basic meaning, or unvarnished meaning of the text.

 

The motivation for this third criterion should be obvious. It is, again, that only what is actually in the text that was enacted that is law. Law is not what anyone wishes were in the text or could be found in the text through an exercise in creative imagination. Not every expectation about the effects of the law are part of the text. Perhaps some northern ratifiers not only expected that the constitutional fugitive slave law would be unenforced in their states, but would not have voted to ratify the Constitution had they not so expected. Even were this expectation key to the ratification, that would not make it part of the meaning of the text.

This third condition, admittedly, poses something of a theoretical problem because there is no such thing, in any very strict sense, as essential, core, basic, or unvarnished meaning. This problem is exacerbated by the second textualist condition in that different segments of the linguistic and political community may speak somewhat different versions of the common language. To see how the textualist program can survive these problems, we need to take a look at how individual speakers connect words up to reality and how the specifics of that connection vary from speaker to speaker.

5. The Meaning/Collateral Beliefs Continuum and the Agreed Meaning/Disputed Meaning Continuum

Suppose that we have a statute on ponds. You are shown a computer image of what appears to be an ordinary farm pond. It could be circumambulated in about four minutes at a moderate pace. In one popular formulation of cognitive semantics, this is a “prototype” of a pond. (“Paradigm” might have been less misleading, but “prototype” has stuck.) Our imagined computer program then grows the size of the pond in relation to its setting. At the extreme the farm appears to be on the shore of a large lake with three sloops racing offshore. At some point well short of this extreme, you would have thought, “I am no longer comfortable calling this a 'pond;' it is more nearly a lake.” (This experiment could also be run in the other direction, shrinking the prototypical pond down until you wonder whether it should be labeled “pond” or “puddle.”)

There are three lessons that I want to draw from this thought experiment touching the vagueness of “pond” and “lake.” The first and most obvious is that the point at which you change from “more like a pond” to “more like a lake” is going to seem pretty arbitrary. That is typically the case when we push into the vagueness of language.

 

The second lesson is subtler. It comes from the observation that your response will probably change from “pond” to “lake” more quickly if the body of water seems natural rather than artificial, e.g. if there is no dam visible. You may also stick with “pond” longer if there is a cow wading in the water, and less long if a small sailboat appears on its surface. These responses might seem puzzling, initially, because everyone knows that there are large artificial lakes, that there are natural ponds, that cows are perfectly capable of wading in lakes, and that small sailboats can be launched onto ponds. Cow friendliness, dammedness, and slooplessness are clearly not necessary components in the meaning of “pond.” There is a respectable question, however, whether they are nonetheless relevant to the meaning of “pond.” If not, then they are merely “collateral beliefs” about many ponds or typical ponds. 

 

What cognitive semantics teaches is that there may well be no either/or answer to this question because, in general, there is no hard and fast line to be drawn between those beliefs so central as to our use of a term as to count as part of the term’s “meaning” and collateral beliefs about what the term refers to. Some beliefs about the nature of a pond are more important than others for the meaning of 'pond' but no non-arbitrary division can be made between “meaning” and “other beliefs.”

The third lesson from our pond example is that in moving from the individual speaker to the language community, the individual’s zone of discomfort in whether to call something a large pond or a small lake will show up as significant variation in the group in making this discrimination. Some people will only use 'ponds' for very small bodies of water, while others will count as ponds bodies many times larger. There will be overwhelming agreement on prototypical ponds and again on prototypical lakes, but there will be a middle region in which there is a great deal of disagreement.

It is because of research backing up the ideas that I here very roughly illustrate that I concede that the words and phrases from natural language that are turned into law do not, in general, have, in a strict sense, “essential meanings,” “core meanings,” “unvarnished meanings,” or “basic meanings.” A hurried conclusion would be that this makes textualism impossible. If we cannot rely on cut and dried meanings for the text as specified by neutral and expert linguists, then mustn’t we turn back to drafter or ratifier intent on the one hand or move forward to a judicial determination of the practically, morally, or politically preferred interpretation on the other? 

 

No. It isn't that we must, and I contend we mustn't. Linguists, including historical linguists for old law, are not going to hand us cut and dried meanings, but we can make use of their expertise about what language meant in a particular community at a particular time. The semantics of 'pond' may lack the sort of neatness for which Aristotle strove, but it has a semantics. That there is no sharp distinction between meaning and collateral beliefs, doesn’t mean there is no distinction. That people would disagree whether certain bodies of water are ponds, is consistent, as just noted, with a very large amount of agreement as to what constitutes a pond. 

 

Let me give examples for the latter two points in reverse order. In 1787-88 there may have been some speakers who understood 'shall not have attained the age of 35 years' as excluding from the presidency only those who had not completed their 34^th year of life. The overwhelming majority, however, would have understood the provision as excluding those who had not yet lived 35 years. That was the dominant semantics of the phrase in the linguistic community, and so the proper textualist interpretation. Where the majority and a minority dialect conflict, the text should be interpreted according to the majority dialect. (An exception for terms of art will be discussed shortly.) 

What if the linguistic community splits evenly on an ambiguity or point of vagueness, within the margin of error of the linguists’ investigation? In that event, the textualist interpreter needs a tie breaker, a matter important enough to have a separate section below.

Let me now turn to meaning and collateral beliefs. Suppose that a late eighteenth century statute governed the issuance of licenses for oceangoing ships. In 1860 an unlicensed ship owner argued that the statute did not apply to his vessel, because it was a steamship and everyone in the political and linguistic community when the statute was passed would have understood all oceangoing vessels to be sailing ships. 

 

Clearly, the argument should fail. Even were the ship owner correct that everyone at the time the statute was passed believed only sailing ships were oceangoing, and even though collateral beliefs begin to shade into meaning at some point, that ships had sails was clearly on the collateral belief side of the line and not a part of the meaning of “oceangoing ship.” Within a few years people of the statute's time would hear of oceangoing steamships, and, although they might have been surprised, they would hardly have regarded the news reporting such ships to a violation of the meaning of “oceangoing ship.”

A licensing case from just a few years ago. A same sex couple applied for a marriage license, and the question was whether it should issue. Suppose that history suggests that the only people being married at the time the marriage license law went into effect were mixed sex couples. It might be said, in fact has been said, on the basis of linguistic evidence not this good, that “marriage” meant “union of one man with one woman.” 

 

In any very strong sense of “meant,” this is clearly wrong. Were it true that “marriage” had meant the union of one man with one woman, then the newspaper headline reporting plural marriages in Utah would have been as unintelligible to the readers of that period as would have been a headline reporting the discovery of a round square. 

 

Still, if gender difference was not a meaning implication of the word “marriage,” gender difference was also not merely a matter of purely collateral belief. Give speakers of the statute’s generation the task of distinguish “marriage” from “not a marriage” in borderline cases (say because of deviant ceremonies or exceptionally easy divorce procedure), and they would surely more readily put a case on the “not a marriage” side of the line if the couple were same sex. So “different sex” was somewhere on the meaning relevance spectrum, below meaning implication but above flatly collateral belief. 

 

Presented with a steamship, fluent speakers of the ship licensing era would have said, “Wow, look an oceangoing ship without sails.” When presented with a same sex couple emerging from a ceremony having all the solemnities of a traditional Episcopalian wedding, it is not nearly so clear what those fluent speakers of an earlier day would have said. 

For this reason, whether a marriage license statute was compatible with same sex marriage (prior to any constitutionally motivated reinterpretation) depended upon details of its wording. A statute requiring issuance to “all persons over the age of consent so long as neither member of the couple has a still living spouse” should be interpreted permissively. The “all persons” language with explicitly stated exceptions is easily strong enough to overcome the meaning relevance of diverse genders. Other formulations of wedding license statutes might well not have been properly interpreted as same sex permissive, as the reader will infer from the next section. They would need repeal and replacement.

 

Because of its importance, I would like to give one additional example of beliefs sufficiently collateral so as not to count as meanings for textualist interpretation. I think it quite probable that most of those who voted to ratify the Fourth Amendment would have firmly believed that what constituted an “unreasonable search” would vary with the social position of the person searched. It would probably have seemed unreasonable in 1791 to search the owner of extensive estates on the basis of the same modest cause that would trigger a reasonable search of a slave or an itinerant peddler. Likely it would also have been thought unreasonable to search the patrician as intrusively as the peddler, if searched on the same cause. 

 

Had these beliefs about what would count as a reasonable search some weight in the decision whether to vote to ratify the Fourth Amendment? They may well have. Should they then control the textualist interpretation of “unreasonable searches and seizures”? No. These beliefs about the connection of reasonable intrusions to social position were less collateral to the meaning of “unreasonable” than were any northern expectations of non-enforcement to the fugitive slave clause, but they are still well into collateral territory. They were not enacted with the text.

6. Breaking the Ties

There are bound to be times when the best of historical linguistic research cannot resolve a key ambiguity or vagueness and when the linguist and logician together cannot resolve a question of implication. How should we should handle such failures? 

The answer follows directly from the root idea that interpretation should follow legitimacy. Only if it can be ascertained that fluent, clear thinking, speakers at the time of adoption would have understood the text along the lines of a more extensive reading should that reading be adopted. Otherwise legitimacy cannot be shown to attach to it. By the “more extensive” reading, I intend the reading that makes more law: prohibits more acts previously permitted, permits more acts previously prohibited, or puts the government in motion on projects for which it was previously at rest. 

 

If we are interpreting a constitution for the purpose of judicial review, then an interpretation on which the statute would be struck down should be rejected in case of a tie in interpreting the constitutional provision. Surely it would be dubious respect for legitimacy that would strike down a statute having strong democratic bona fides on the basis of a constitutional interpretation that is under the shadow of the equal claims of a contrary interpretation.

 

There will be cases, although they seem to be rare, in which neither of two interpretations is “more extensive” than the other. On one interpretation, the dam is to be built on Marsh Creek; on the other interpretation, it would be Swamp Creek. The two projects would be of similar scope. Here the interpreter should simply give a best guess as to which side the historical linguist should have come out. It would, you might contend, be only a small conflict with the spirit of legitimacy to decide in favor of that alternative that would produce the best consequences. You would be right; still that would be to move from interpretation to legislation.

What if the interpreter has no guess at all, not even a gut feeling, as to how the linguistic question should come out? Am I suggesting that the interpreter flip a coin? I suppose in that case, if the interpreter is a judge, I would follow the spirit of Hart and suggest: “Because I cannot conscientiously decide as a matter of interpretation of the statute between Marsh Creek and Swamp Creek, and because I am required to make a decision, I give judgment in favor of the Swamp Creek party because a dam on swamp creek will have better consequences (or will violate fewer or less important rights).” 

 

Hart thought that the need for such judicial legislation arose very frequently – wherever there was a crucial vagueness or ambiguity. In my view it will arise rarely. First, the crucial vagueness, ambiguity, or implication question must be unresolvable through application of linguistics and careful working out of implications. Then, it must be impossible to conclude that one of the candidates tied after that stage is legally more extensive than the other. This is a powerful tie breaker. For example, it immediately settles Hart's toy car hypothetical in favor of non-prohibition. I made up the Swamp Creek example because I could not think of a real world case requiring judicial legislation a la Hart. (The best place to look for real world examples is not simple cases of precisely balanced extensiveness, as in the Swamp Creek and Marsh Creek dams, but in complex cases in which one alternative is more extensive in some respect and the other alternative more extensive in other respects.) 

 

Unlike our ideal interpreter, the real world interpreter has only a limited amount of time to make an interpretation, an amount of time that should vary with the importance of the decision at issue. It not often happens that at the moment at which efficiency requires putting an end to the linguistic investigation, the interpreter will have no suspicion as to how more complete research would come out. It will be even rarer that the same interpreter will next be left with no suspicion as to which of two interpretations yields the more extensive legal change. 

 

7. Words of Art

Consider a statute that permits law enforcement to swab an arrestee’s mouth for a DNA sample. Do we interpret the statute in accordance with the understanding of “DNA” or “deoxyribonucleic acid” in the general population at the time the statute was passed? Clearly, no; to do so would produce hash. An interpreter of the statute should look instead to the definition of DNA that would be given by experts in the field. 

 

We all understand that there is a division of labor when it comes to words or phrases that have their home in a specialty. It is to the specialists that we properly look for the meaning of technical terms or words of art. Because this is the general understanding of the way the language works, it is not a departure from the textualist thesis that we should look for our interpretation to the language as it was understood by fluent speakers at the time the text became law. It is only a refinement that the typical fluent speaker refers us to specialists for the meaning of some words and phrases.

This is obviously not a phenomenon restricted to scientific words. Laws regulating horse racing will make use of some vocabulary from the track, and it is that argot that will normally control. The most important of the special institutions and practices when it comes to legal interpretation are the institutions and practices of the legal profession itself. “Bill of Attainder” is an example. A late eighteenth century livery owner probably had only the vaguest idea what this phrase meant unless he read English parliamentary history or had a loquacious lawyer uncle. 

 

A more complex example is the phrase “Law and Equity” as it occurs in the grant of power to the federal courts in Article III of the US Constitution to “extend to all Cases, in Law and Equity . . .” (Section 2). It is the legal tradition, and particularly the English tradition of a dual court system, one set of law courts and another of chancery, that supplies the meaning for the phrase. This, then, calls for the services of a particular kind of historical linguist – one who does double duty as a legal historian. 

 

To separate these functions, think of the linguist as identifying that “Law and Equity” is a term of legal art, perhaps by noting, first, that, although “law” and to a lesser extent “equity” were commonly used by non-lawyers in non-legal contexts, the common meanings do not combine very well in making sense of the power granted to the courts. The linguist would find, next, that there are special legal meanings of “law” and of “equity” that apply naturally and directly to kinds of courts and within which “Law and Equity” in Article III makes very good sense. Finally, the linguist would hand off to the legal historian to supply the details of the translation into our ideal language (or details needed for the resolution of a specific interpretation issue.) 

 

Determining whether a word or phrase is or is not “of art” or “technical” is not always easy or free of controversy, but it is a sort of project for which we can look to historical linguistics in the reasonable expectation of results.
 

It is often strong evidence in favor of the conclusion that a phrase is a phrase of art that it has a well known history. For example, “cruel and unusual punishments” was taken word for word from the 1689 English Declaration of Rights and the Virginia Declaration of Rights of 1777. That a phrase had well known prior use as a phrase having legal effect is enough to trigger an inquiry as to whether the same phrase as it appears in the Eighth Amendment to the US Constitution is one of legal art. It is not, however, by itself enough to establish that the phrase was a technical one for the ratifying generation. To make that determination, the best evidence would be the discussions surrounding the ratification of the Eighth Amendment. Did those learned in the law make an effort to explain “cruel and unusual” to the general population? If little such explanation surfaces from that period, it would support an inference that the lawyers and the general population, although knowing the phrase had a history, were content to rest with the ordinary meanings of the words.

If we concluded that the phrase was a technical one, our next move would not be a painstaking examination of the meaning of the phrase as used by aristocratic England in 1689. It is not how those in the know at the time of the Glorious Revolution used the phrase. It is how the legal community of the United States a century later understood it. English history might serve as evidence for what the lawyerly understanding in the United States was in 1791, but it would only be evidence. It is not unheard of for American lawyers to misunderstand technical terms of English law, or to appropriate them while making changes to their meaning. 

 

I have not done the research, but I suspect that it would turn out that the phrase “cruel and unusual” was not one of art, despite its historical precedents. There was not enough communication to the general population that this phrase was the special property of the lawyers. 

 

If that is correct, then “unusual” raises the following issue. Does it prohibit the punishments that were unusual in 1791, for example thumbscrews and long terms of imprisonment, or should it be interpreted to directed against any a punishment unusual at the time the interpretation is being done. The lash was not unusual in 1791, but having long since fallen out of favor, it is unusual now. 
 

In its indicative use “unusual” picks up the time of utterance unless there is a reference to some other period. “It is unusual for lawyers to arrive at court by horse, but it was not unusual in the eighteenth century.” When there is any danger of ambiguity, as when different periods are under discussion, we add “now” to ensure reference to the usualness standards of the present. “It is now unusual for lawyers to ride a horse to court.” 

 

In the imperative mood and related uses, “unusual” and words like it tend to pick up the context of application rather than that of utterance. “I instruct my trustee to provide $3000 per year to each of my grandchildren for the purchase of clothing in the hope that they will not be unfashionably attired.” Clearly here “unfashionable” is not to be judged by the standards of fashion when the trust was established. Had the grandparent admonished: “Don't wear anything unusual to school,” the same principle of interpretation would apply.

 

In “cruel and unusual” the “and” raises a more significant question than might be expected. In contexts governed by a negative, “and” sometimes functions disjunctively. If a “shall not be” construction was a context in which “and” might well be read as “or” in late eighteenth century American English, it would justify the practice of some interpreters to read “cruel and unusual” as if it were “cruel or unusual,” and to find cruelty alone sufficient for an Eighth Amendment violation. In any event, if the common language rather than the technical vocabulary of law is the proper touchstone here, then “cruel” will deserve some further attention, as do all other expressly normative words.

8. Normative Words.

Two people can speak exactly the same language, and yet one of them affirm that haggis is delicious, and the other deny that proposition with vehemence. Their difference is not about the meaning of “delicious” but about the delectability of haggis. For a partially similar reason, people who have identical understanding of the meaning of “cruel punishment” can disagree as to whether flogging is cruel for theft of a chicken.

 

No one thinks that a list of what public opinion regarded as cruel punishments in 1791 would give us a proper Interpretation of “cruel punishment.” Sensory deprivation, chemically induced psychosis, randomly varied electric shock, would not be on any such list as they had not yet been invented or, at least, perfected. 

There are those who seem to believe, however, that any potential punishment of which the late eighteenth century citizenry knew should be counted as cruel or not depending on their majority view. This would have the consequence that flogging would not violate the Eighth Amendment even for some relatively minor offenses, and that long term imprisonment might well, as it was certainly unusual before the nineteenth century penitentiary movement, and would have struck many as cruel.

 

Making eighteenth century sensibility the touchstone of constitutional cruelty is understandable for an adherent of one variety of social moral relativism: roughly, moral truth is defined only for a particular society at a particular moment in time and consists in the moral propositions that the majority then and there would endorse. If this were the case, then to look to eighteenth century meanings would arguably be to look to eighteenth century morality – at least where there was eighteenth century moral consensus.

 

There are good philosophical reasons for rejecting this kind of moral relativism. Roughly and quickly: slavery was not morally right even when the great majority confidently affirmed that it was.

 

I conclude, then, that our idealized textualist translation of the Eighth Amendment might well rule out as cruel hanging for horse theft or flogging in the public square for blasphemy even if most of the citizenry of 1791 would not have regarded these as cruel. (Most of the citizenry might have regarded these offenses as unlikely to appear among federal statutes, but let that pass.) 
 

There is a complication that requires us to go a little more deeply into moral relativism, or more generally normative relativism. There are some words that carry relativism on their sleeve. Suppose that a statute prohibited “offensive photographs,” assuming for the moment, that the statute would not be constitutionally infirm. Offensiveness is clearly in the eyes of the beholder. As times and manners change, what is offensive will change as well. So a textualist interpreter can be correct in determining that a given photograph violates the statute, and yet a second textualist, also correctly, find that the same picture is not statutorily offensive thirty years later or 300 miles away. 

 

A rough test of whether a normative word in a legal document has the relativity of “delicious,” or “offensive” would be whether a majority of speakers, and not just college sophomores, would say things like “question of taste,” “eye of the beholder,” “different strokes,” or “matter of opinion.”

What about “excessive” as in “excessive fines”? Let me assume that the meaning of the phrase is something like “fine so large as to injure the fined person more than is just for the offense.” Arguably, the justice of the injury done by the fine will vary with such factors as the absolute and relative wealth of the person fined. This is not yet a matter of any sort of normative relativism, but of the variables to be taken into account in assessing the justice of a deprivation. 
 

A thorough going relativist would say that there is no anchor whatsoever for “excessive fine.” So if a judge confiscated all of a person's property for a parking violation, saying “I do not find that fine excessive and neither do my family, friends, and colleagues,” that would be the end of the story. This, however, is surely incorrect. There are some deprivations that are unjust, however they appear in the eye of a particular beholder. 

 

Yet it is possible that parochial practices and values will and should sometimes affect what fines count as excessive. Suppose that there are two communities that are very similar vehicularly: similar number of autos, delivery trucks, traffic flow issues, and available parking. In one community parking fines have never in the past exceeded $15. In the other, municipal ordinances have included parking fines up to $100, and $75 dollar fines have not been infrequent. Both communities, in unrelated action, pass ordinances giving the town judge the discretion to impose such parking fines “as shall not be excessive.” Faced with the task of removing the vagueness of “excessive” with respect to a proposed $80 parking fine, textualist town judges in the two communities might properly reach opposite conclusions. 

 

It has sometimes confounded common sense and bedeviled theorists, but normative terms can have both a relative and an absolute component. Some fines are excessive, some punishments cruel, and some procedures violate due process, end of story. There will be borderline cases, however, where excessiveness, cruelty, and dueness are properly tied to community standards and values, variable over time. The extent of the borderlands is a question for research and analysis.

Sensitivity to the relative dimension of normative language has sometimes dulled the awareness of commentators on legal interpretation to the fact that most normative terms do have an absolute component, even interpreters who know as a matter of abstract theory and in their heart of hearts that moral relativism is not unreservedly true. As suggested above, absolutizing the values of the enactment moment is an overreaction to the possibility of relativism and a wholly unnecessary retreat from textualism. 

It is the task of the historical linguist to sound out the rough contours of the scope of relativism in a normative term. Is it more like “tastes good,” or “offensive,” or more like “unjust” or “morally wrong”? The textualist interpreter then places the particular case and applies either generally accepted norms of the interpreters' own time, historical norms, best normative judgment, or some combination thereof depending upon the specifics of the legal text. A plausible answer for “excessive fines” is that the fine may neither be absolutely excessive nor excessive by social consensus at the time at or for which the interpretation is made. 

 

Application of the historical norms of the enactment period would seem to me the least likely exercise in normative clarification. It would clearly be called for if the language were “fines regarded as excessive on the date of enactment of this statute,” but that language never appears. It is conceivable that a legal term of art might, in effect, freeze the values of a particular era of legal history, or, say of the bar of the courts of equity of some period, although I do not know of a good example.

9.  Meaning and the Whole Text

Lawyers and judges in interpreting one particular provision frequently make use of other parts of the statute or constitution. The use of the same term in another provision may shed light on how it is to be understood in the provision at issue. This is typically a move from one or several specific uses of a word or phrase to another specific use. At other times the interpreter moves from the general to the specific. The manifest purpose of the entire statute or the underlying structure of the constitution is sometimes decisive for the resolution of a particular ambiguous or vague word or phrase.

 

We often find the language of drafter intention when an interpreter looks to the wider text. “The drafters must have meant the same thing here as in their earlier employment of the same word . . .” or “the legislative intent of the statute as a whole requires us to interpret . . .”. A closer inspection will show, however, that the intentionalist language is adventitious, and that the process of interpreting in terms of the document as a whole can be understood in a rigorously textualist way. 

 

The key concepts are interpretational consistency and harmony. Admittedly, consistency is not always a virtue in other sorts of interpretation. Even within a single document a translator may sometimes render “Geist” as “spirit” and sometimes as “soul.” Then too, for aesthetic reasons we do not always want to eliminate contradictions in interpreting a poem. When it comes to enforced law, however, we must not be told that we are legally obligated to do something and also legally prohibited from doing it. There are lesser ways in which laws may potentially conflict, and it is with few exceptions the job of the bench and the bar to eliminate conflicts, at least when they become of practical importance.

Suppose we have a statute enacted in 1890 with an ambiguous term and a statute enacted 1990, also with an ambiguous term. The older ambiguity could be resolved either as a or a'; the new statute's ambiguity either as b or b'. Suppose further that, taking each statute in isolation, a and b would be slightly better interpretations. However, a and b conflict, and so we cannot now jointly adopt the a, b interpretation, without saying that the later statute repealed the older provision entirely. Inasmuch as a' was always an eligible, if second best, disambiguation, the better course, other thing being equal, would be to give precedence to the later statute by adopting the otherwise favored interpretation b, while preserving the earlier statute in a way consistent with the later enactment by adopting the a' reading. In reducing conflicts, later statutes get priority over earlier ones because democracies can properly change their minds. Notice that drafter intent need not enter in any way into this consideration.

If we have this sort of conflict within a single enactment, we no longer have temporal difference to help us harmonize potential conflicts. Still, the conflicts need to be harmonized. If the a/a' and the b/b' ambiguities are within the same act, and if a and b are in contradiction, we should determine whether the a-b' or the a'-b combination makes the better sense in the context of the rest of the enactment. If the statute as a whole seems to serve a particular policy, which of the disambiguations best furthers that policy? 

 

Instead of talking about the policy or the purpose that appears on the face of the statute, interpreters in cases like these often talk about intent with an almost imperceptible slide from the intent of the statute to the intent of those who enacted it and sometimes on to the intent of the drafters. Statutes, after all, not being animate, cannot have intents. It is then concluded that any talk about the policy or purpose of a statute is only an indirect way of talking about the intent of the legislators or drafters.

This, however, has things backwards. Just as a sentence may contain a pun without its author's so intending, a text can show a purpose to perspicacious readers that is other than the purpose the author was intending to exhibit. 
 

An extreme case may prove illustrative. In the flurry and confusion of last minute amendments, environmentalists succeed in attaching to an omnibus farm bill a detailed set of restrictive rules for a wide range of operations affecting wetlands. At nearly the same time, legislators more sensitive to developers' interests succeed in passing their own amendment to the agriculture bill, eliminating almost all funds for enforcement of wetlands restrictions. The intent of one successful group was wetlands regulation, that of the other successful faction, no wetlands regulation. If the policy of the law as passed were the intentions of the drafters, the law itself would be in contradiction. 

 

The two provisions of the law are certainly in tension, but they are not in contradiction. A judge need not, indeed ought not, interpret away either provision. The policy of the law, for better or worse, is a rigorous set of wetlands restrictions to be lightly enforced. This was not anyone's favored policy, but it is the policy of the law as enacted. 

 

If one insists on speaking of intentions of written law, they are best understood as the intentions of an idealized (i.e. imaginary) author who intends exactly what a sensitive reader would take to be the purpose of the law. Again, the interpreter can look to the content of the whole law in resolving an issue of interpretation without considering the mental states of drafters or voters.

To go back to the problem of the a/a' and b/b' ambiguities, what if there remains a tie because each disambiguation pair furthers the policy of the law just as well as the other so far as can be determined? Surely in this case, it will be insisted, we should look behind the face of the statute to drafter or voter intentions. 

 

In fact, I think this circumstance will almost never arise. A real world judge will always find that one way of disambiguating furthers the policy of the law more effectively or represents a refinement more nearly consonant with the purpose that appears on the face of the law, at least to a reasonable probability. If a judge concludes that there is an absolute tie as to which of two disambiguations (or eliminations of vagueness or inferences to implications) is the better after looking to the policies reflected in the document as a whole, then the judge should, once again, adopt the less extensive reading to avoid going beyond legitimacy. 

 

In the theoretically possible case in which the two readings are also tied in extensiveness, the judge should settle on what he or she takes to be the better law, that is, legislate judicially. Here, however, we really have entered upon a theoretical possibility that will almost never arise in the real world. It usually transpires that the features of the reading that make it seem the better law to the judge will also make it the better expression of the policy of the text as a whole, as the judge understands it. In those rare cases in which this is not so, the less extensive law principle will almost always yield up a resolution, though not, of course, necessarily a resolution the judge would have personally preferred. 

 

Interpretation in light of the whole text raises the question whether we are to look for the purpose of the text as a whole with the eyes of a fluent speaker of the enactment date or with the eyes of a fluent speaker having the historical and political knowledge that would be typical for someone of that period. When addressing the first stage of the interpretation of a word or phrase, textualism counsels considering only the meanings that would be common to the great bulk of those who lived within the political unit at that time, and that the meanings be more nearly “core meanings” than “collateral beliefs.” When, at that level, we uncover an ambiguity or a relevant vagueness that cannot be resolved following the methods of historical linguistics, and so must look for guidance to the document as a whole, a broader interpretational data base is appropriate. The purpose of a text as a whole may well not be apparent to an interpreter who is entirely innocent of the way of life, problems, technology, and political movements of the period. 

 

Still, there should be restrictions on the information the interpreter relies upon – a line drawn at some point before we get to the specific intentions of the drafters or voters. Without this restriction we risk intentionalism by the back door, even if it be a second or third stage intrusion of intentions.

Suppose that there is an ambiguous clause in the legislative delineation of the line between assembly districts that will put four hundred voters into one district or the other. Linguistics fails to resolve the issue. It appears on the face of the legislation or by dint of modest historical knowledge that one purpose of the statute is to produce as nearly as possible districts with equal populations. Those who voted for the new redistricting map also expected it to give their party an advantage in elections in the next decade. They would not otherwise have voted for it. The equal district purpose is appropriate for disambiguating the provision at issue. The partisan purpose of the legislators is not. That would be to move from statute purpose to the legislators' unenacted intent. 

 

Consider now the establishment clause of the First Amendment of the US Constitution. I am going to assume for present purposes that “establishment of religion” was a technical legal phrase heavily dependent upon history especially that of the Church of England, the legal definition being, roughly, “the relation of church and state sufficiently similar to the relations exemplified by the Church of England, Church of Scotland, the existing laws of Connecticut, Massachusetts, and New Hampshire, and those of South Carolina prior to recent amendment.” If this is correct, then the establishment clause, had it been interpreted only by textualists, would have done less than it has done in historical fact. Just how much less, and in what ways less, would depend how “sufficiently similar” is to be filled in. 

 

It clearly would not be required for the President to be its head for a church to count as established.The governor of Connecticut was not the leader of the Congregational churches of that state. A general tax for the benefit of a particular religion was almost certainly enough, and a tax benefiting the local majority religion seems to be as well. (Massachusetts) Lesser and less specific promotions of religion raise more difficult questions. State endorsement of specific religious doctrine was a key feature of establishment in England, which raises the question whether the establishment clause interpretation has done too little in the “under God” and “in God we trust” cases. 

 

A broader view of “sufficiently similar” is arguably supported by a whole text move looking to the body of the Constitution. Of course, the original text of the Constitution is not the same document as the First Amendment in a narrow sense. Amendments, by their very nature, however, become part of the larger, amended Constitution. We cannot always infer from the purposes of the pre-amended document to resolve interpretational issues in an amendment, as the amendment may well have a facial purpose to annul or restrict the effects of the purposes of the prior law. Still it is always appropriate to look to the document being amended when interpretational problems arise in an amendment.

For the establishment clause, an argument can be drawn for what was for the period the exceptionally non-religious character of the original Constitution. “God” does not appear, even in the preamble. Not only is there no religious test or qualification for public office, as was then common in the state constitutions, but there was an explicit clause prohibiting any religious test. The constitutional oath provisions expressly permitted the substitution of a (non-religious) affirmation. In a document of limited powers, no power was given to the federal government to promote or aid religion, unless it be via the general welfare clause. 

 

Taking into account such history as would be widely known, including the religion clauses of state constitutions, an interpreter of the establishment clause might well conclude that the body of the original Constitution supports a somewhat broader understanding of what is sufficiently similar to the then recognized established churches to fall within the prohibition of that clause.

10.  Statutory Interpretive Directives.

Legislatures sometimes pass statutes mandating interpretive methodologies and standards. For example, New Jersey:

In the construction of the laws and statutes of this state . . . words and phrases shall be read and constructed with their context, and shall, unless inconsistent with the manifest intent of the legislature or unless another or different meaning is expressly indicated, be given their generally accepted meaning, according to their approved usage of the language. Technical words and phrases, and words and phrases having a special or accepted meaning in the law, shall be construed in accordance with such technical or special and accepted meaning. (N,J.S.A. 1:1-1)

As an initial matter, the textualist will try to interpret this statute in the manner I have sketched above, as it is a statute like any other. Yet because it is binding on New Jersey judges and officials, and by operation of federal law, also on United States judges, this statute controls their construction of New Jersey statutes and that of any interpreter, even an academic, who intends to interpret from the prospective of a judge or official. This will, then, include pretty much everyone who is interpreting a New Jersey statute for any real world purpose, excluding only the theoretical project of setting out how a New Jersey statute would be interpreted absent the construction provision. 

 

This construction statute is reasonably congenial with the interpretive method I have sketched. In fact, there is no conflict at all – if two ambiguities in the statute are resolved in one particular, not implausible way. The first of these is whether “generally accepted meaning, according to their approved usage of the language” refers to generally accepted meaning at the time the statute was passed or generally accepted meaning now. 

 

“Generally accepted meaning” has a present flavor as does “approved usage of the language.” It could have been “generally accepted meaning at the date of enactment” and “usage as then approved.” Similarly, it is “having a special or accepted meaning” not “that had a special or accepted meaning.” These observations would favor present meanings rather than enactment date meanings. Yet, we must read all statutes “with their context,” and the context here tells us not to depart from the “manifest intent of the legislature.” Whatever a manifest intent is, a question to be taken up next, it is clearly something manifested at the time of enactment. If enactment date intent is sometimes to control, it must be enactment date meanings that control otherwise. So I conclude that the New Jersey construction statute, like the textualist method defended here, takes the meaning of language as it existed on the enactment date. (It may be worth emphasizing, to return to one of our themes, that the statute speaks of “meanings” not, what is sometimes confused with meanings – the expectations of the language community at the enactment date.)

The more difficult question of interpretation of this interpretation statute is how we are to understand “manifest intent of the legislature.” One possibility would be that it embraces the intent of legislators however evidenced, whether by statutory structure, preamble, current events, floor debates, committee reports, legislators’ private diaries, or the memories of legislative aides thirty years after enactment. I hope I will not seem too partisan in saying that at least the latter two kinds of evidence are beyond the generally accepted meaning of “manifest intent of the legislature” being secret or at least private and probably referable only to a legislator or group of legislators, not the legislature.

If we settle at the other extreme, taking into account only the intent as manifested in the statute, together with background facts about the state of the world that were generally known at the time, then we resolve “manifest intent” in a fashion wholly congenial to what I have above proposed. “Manifest intent” will be the equivalent of my “purpose” shown on the face of the statute. This reading is supported by the natural argument that statutes do manifest the intent of the legislature by their terms taken as a whole and in light of historical fact. 

 

That the best reading of the New Jersey construction provision on its face leaves it consonant with thoroughgoing textualism, does not, however, settle the issue. Another possible reading, even if second best at this stage, would have it that the intent of the legislature is not manifested by the statute alone but by the statute together with legislative history of a publicly available, hence manifested sort. 
 

It seems fair to give this possibility the advantage of its own bootstraps by looking to the published legislative materials. If there were a committee report that gave committee reports as an example of how the intent of the legislature could be manifested, I would tentatively accept this as determinative. I would accept it only tentatively because it would be subject to reconsideration if there were evidence, say, that the committee report was not available until after the statute was passed or simply that legislators did not read the committee reports. Still, if all of the publicly available evidence made it sufficiently clear that it was the common understanding of those voting yay on the statute was that they were voting to elevate committee reports in this fashion, then that is the way the statute’s “manifest intent” ambiguity should be resolved. The case for legislative history in New Jersey would be easier, in fact nearly unavoidable, if the word “manifest” were absent from the statute.

This use of committee reports in interpretation would be unfortunate for reasons already appearing. On the reading requiring their consideration, the New Jersey statute would not only resolve issues of vagueness, ambiguity, and implication, in terms of legislative history, but could require a reading in flat conflict with the generally accepted meaning of the statutory language. Democratically enacted statutes can direct ill-conceived methods of statutory interpretation, and a theory of interpretation grounded in concerns of legitimacy must accept this, even if its advice to legislators would be very different.

11.  Textualism, Originalism, Legislative History, Judicial Activism, Legislative Deference, Strict Construction and the Living Constitution.
 

How does textualism in the form here outlined line up in the debates about statutory and constitutional interpretation, debates that have been especially boisterous in the United States, underlain by sharp ideological differences and often issuing in acrimonious accusations of judicial impropriety? 

 

As noted earlier, textualism, although it looks to original meaning, is in conflict with the most popular forms of constitutional originalism in the United States – drafter intentionalism and voter or ratifier intentionalism. Again, the reason for parting ways with these forms of originalism is that, even if (contrary to fact) there were unanimity of intentions, no intentions were enacted. The drafters were free to put as much of their expectations and intentions into the text as they wanted, subject only to the risk of losing the subsequent vote or votes. There is no reason sounding in legitimacy, at least of legitimacy in the broadly democratic family, that what drafters declined or neglected to put into the text ought to be binding upon the rest of us.

One recently proffered justification for drafter intentionalism in construction of the US Constitution is that Americans do, and should, revere the founders, and that this reverence properly extends to the founders' intentions, whether or not they chose to make them explicit in the text. This is reminiscent of Hegel's theory of world historical individuals through which Spirit works the progress of history. Essentially authoritarian, it is not a theory that is congenial to democratic ideals. 

 

Moving from the theoretical to the empirical, although I concede that the founders do sometimes seem elevated to a pantheon in popular civic culture, there is reason to think the reverence is in some significant part misplaced. The promoters of a the civic faith facilely rationalize the constitutional accommodations to slavery, including even the repellent constitutional fugitive slave clause, as compromises essential to the grand goal of forging a nation. Is nationhood really a cause to which the most fundamental of liberties are to be sacrificed? 

 

With respect to a lesser enormity, it is doubtless anachronistic to fault the founders for failing to establish a federal franchise that went beyond propertied males, yet it is a perfectly good reason for dissenting from their apotheosis – at least an apotheosis that would turn into law their uncodified intentions.

Drafter intentionalism based upon the special revered status of those who produced the constitution is even less appealing when extrapolated to other legal texts. The drafters of Jim Crow laws, and their intentions, never deserved the least respect. Even restricting the question to constitutional interpretation, the special eminence theory as support for drafter intentionalism, seems a good deal less plausible for the drafters of, for example, Amendments 18 ( prohibition) and 20 (repeal of prohibition). I am not taking issue with the content either amendment, though one of them must have been mistaken. I am only contending that the drafters of neither are owed special deference.

Textualism's disinterest in drafter intentions does not quite go so far as to make all legislative history irrelevant to the interpretive process. After all, the drafters were writing and speaking what was presumably reasonably standard English for their period. Their recorded statements are not, in general, better evidence for the way the language was used at the time of enactment than is what we have of the writings or sayings of the general population, but they are just as good evidence, and drafter writings are typically more widely available than those of the general public. 

 

Moreover, there is one kind of drafter language that really does have special relevance for textualist interpretation. As discussed in Section 6, it is of crucial importance whether a word or phrase in a text is language “of art,” relying on a particular sub-community of speakers for its meaning. Often it is obvious on the face of the text alone that a word or phrase has this technical character, and that we are to look, not to ordinary speakers, but to geologists or lawyers of the period for its meaning. In other cases, whether something is a word of art is more difficult. If the drafters in explaining or promoting the legislation said “as defined by the geologists” this would go a long way towards settling the question. It would not go quite all the way because we would still, ideally, want to check to make sure that this drafter language was publicized to the voters or ratifiers or that the latter understood in some other fashion that they were voting to make into law a text that incorporated a specialized vocabulary. 

 

Assuming reasonably good evidence of these sorts, the interpreter might again look at the writings of the drafter if what was understood to be technical language was (as is common) the technical language of law and if (as is also common) some of the drafters were lawyers. Here, however, the textualist interpreter would be perusing legislative history, not for author intention, but as evidence of technical legal meaning as reflected in the speech or writing of a specialist. Lawyerly writing neutral or hostile to the law, or even ignorant as to the proposed law, would be just as useful, because just as good in displaying the legal argot of the time.

Originalism of the voter or ratifier intention sort has closer connections to textualism, although the two are still distinct. The key difference is in the ways the ratifier intentionalist and the textualist understand what is “understood.” The intentionalist looks to the potential voter's understanding of what the law will do, e.g. what specific punishments will be ruled out as cruel, and what punishments will be permitted. The job of the intentionalist interpreter is to try to get as nearly as possible the ratifiers' list. The textualist interpreter, by contrast, stops with the meaning of the language for the ratifier community, not going beyond that to the ancillary opinions of members of that community as to the application of that language. 

 

For example, “cruel punishment” might be found to have meant “a punishment that causes excessive suffering.” Two ratifiers could well agree that this is what “cruel punishment” means yet produce different lists by dint of a disagreement as the degree of suffering needed to rise to the level of the excessive in a particular case. Some of the Quakers of Pennsylvania no doubt thought cruel certain punishments that some descendants of the Puritans in Massachusetts thought not cruel at all. Their disagreement was not about the meaning of “cruel” but about what is cruel. The intentionalist interpreter must decide whether the Quaker influenced or Puritan influenced list, or some third list, is to be taken as constitutionally canonical. The textualist interpreter takes the constitution to prohibit punishment that causes excessive suffering.

This is a good point to take up the important dispute that is often expressed in terms of support or opposition to the idea of a “living constitution” or in terms of the extent to which a judge is free to bring her own values to bear on interpreting the constitution. 

 

It is useful to set the outer bounds of this dispute. No one thinks that the Constitution should exhibit signs of life when it comes to interpreting the age thirty five qualification for the presidency. The observation that the US society now regards such arbitrary age restrictions as an ugly holdover of “ageism” might make it unto Saturday Night Live, but could not be put forward as a serious interpretational argument. At the other extreme, no one thinks that Fifth Amendment “just compensation” as it bears on the taking of a stretch of Virginia beach for a Coast Guard facility should turn on a list of what the drafters or ratifiers of the amendment would have seen as just compensation for given parcels on the Virginia coast. Instead of what the ratifiers thought “just,” judges have, with at least rough concurrence of all schools of interpretation, tried to determine what really was just at the time the taking took place, generally settling on the then fair market value.

So there is universal agreement that the application of some provisions of the US Constitution is not to vary at all with time and that applications of other provisions are to be sensitive to changing circumstances and changing judgments. The dispute is which provisions are on which side of the line, and how those on the “living” side of the line are properly interpreted.
 

Many of the “living constitution” disputes are over provisions containing explicitly normative words or phrases: “general welfare,” “free exercise,” “freedom of speech, “right of the people to keep and bear Arms,” “unreasonable searches and seizures,” “due process of law,” “just compensation,” “excessive bail,” “excessive fines,” “equal protection,” 

Other provisions subject to debate, have a normative character that is arguably just below the surface, “necessary and proper,” “Privileges and Immunities,” “Full Faith and Credit,” “Republican Form of Government,” “high Crimes and Misdemeanors.” Finally, there are disputed provisions that are either non-normative or whose normative character, if any, is revealed only by substantial analysis: “natural born Citizen,” “Commerce with foreign Nations, and among the several States, and with the Indian Tribes,” “establishment of religion.”

Assuming always that the provision in question is not governed by language of art, the more explicitly normative the provision is and the less its normativity is like that of “offensive,” the more clearly textualism requires the interpreter to make a normative judgment. The interpreter must make a judgment what process is due, what searches are reasonable, what fine is excessive. This is not because interpreters, even judges, are wiser than drafters, ratifiers, or the contemporary general public. It is because that is what the text requires them to do.

It is easy to write a statute without explicitly or perhaps even implicitly, normative language. In that case a judge is not called upon to make normative judgments. It is less easy, and certainly less common, to write a constitution without normative language.  Where the language is normative to insist that the interpreter make no moral judgment is to ask the interpreter to subvert the text.

That interpreters are sometimes to make normative judgments will be welcomed by the growing constitution school and regretted by those who fear the loss of fundamental protections or who are concerned with the stability and predictability of law. Interpreters, including judges, with the best will and the highest integrity are inevitably going to make different judgments as to what justice really is due, what fine truly excessive, or what punishment objectively cruel. Even if we assume that there is a morally best answer to each of these questions, getting it right is not so easy as to insure judicial uniformity. 

 

In saying that the necessity of making normative judgments in interpreting normative text would be welcomed by growing constitution advocates, I do not want to suggest that the political values that underlie that advocacy are aligned at a deep level with constitutional change through interpretation. Constitutional change does not necessarily mean a growing constitution, either in the sense of an improving constitution or that of an enlarging constitution. We can hope that later interpreters will have a sounder moral sense, but the trend might equally be in the other direction. The human race does not have a record of monotonically improving morality. Later interpreters may curtail rights guaranteed to the individual or diminish powers of the state to improve the general welfare. 

 

I have to this point addressed only expressly normative provisions of legal text. What about provisions that are only implicitly normative, including those crypto-normative phrases whose normative character comes to light only after careful analysis? If the text is normative, whether it is on the surface or at depth, the idealized interpreter in removing all ambiguity, vagueness, and working out all implications will have to make normative judgments. I want here to emphasize, however, that the question whether a provision of legal text is normative is not itself a question of what the law should be. It is a non-normative question of linguistics (for old texts, historical linguistics).

For a provision that is truly non-normative, difficult interpretive judgments may still have to be made, but they are, again, questions of linguistics, or, in application of the purpose of the wider text, of history. Well informed interpreters of widely different values should, in principle, come to the same conclusions. The “less extensive law” principle, as a tie breaker, cuts off the area of indeterminacy for which reduction to a specific and unique right answer would arguably require a value judgment. 

 

In the debate about “judicial activism,” in judicial review, the less extensive law principle tends towards “legislative deference” and against judicial innovation. When it is a matter of interpreting statutes, however, the same principle is not so deferential to the legislature. It requires for the making of binding law that the legislature be clearer than it sometimes is, and sometimes clearer even than legislators would like the law to be. The legislature should not expect their work to be done for them by textualist judges.

One final remark on the ideologically laden arguments about judicial interpretation of constitutions and statutes: I note that in their critiques of judicial decisions, critics from the political right, left, and center almost always find the constitutional or statutory provision at issue to be exactly what it needs to be for the case to come out on the side favored by the critic's own politics. I suppose that there is something comforting about looking, for example, at the US Constitution, and finding that its controverted provisions, properly interpreted, are exactly as you would like them to be. A consistent textualist is very unlikely to find any legal document so uniformly congenial.

Indeed, in my view, a good rough measure of judicial integrity is the frequency with which a judge decides against his or her own political preferences: resolving a statutory ambiguity to give a result the judge finds unfortunate, upholding the constitutionality of a statute the judge regards with abhorrence, or striking down a statute the judge would strongly support, but for its unconstitutionality.

12. Stare Decisis

What I have discussed here is methodology for interpreting legal texts. It is not intended to be prescription to judges for deciding cases, even cases for which the interpretation of a constitution or statute or both is the heart of the dispute between the parties. Where stare decisis is a legitimate part of the legal tradition, precedent may well change everything. A wrong interpretation of a statue, by textualist lights, may well have become a strong precedent that has long service in resolving disputes, supporting expectations, protecting reliance, and grounding lawyerly predictions used to guide clients. There are powerful reasons from the legality family for honoring precedent. 

 

Textualism has some benefits sounding in legality. It tends towards stability in the interpretation of legal documents as compared to some other popular interpretive approaches. Textualism's primary justification is not in legality, however, but in legitimacy, with particularly strong claims in a broad and healthy democracy. Textualism can conflict with stare decisis because a textualist must sometimes conclude that an earlier interpretation is not true to the statutory or constitutional text. Even if the earlier interpreter was trying to be textualist, a later textualist may well conclude that the precedent was wrong. This is especially likely if the provision at issue is normative or if one interpreter found a term of art where another saw, for example, only ordinary eighteenth century English. As should be clear from the earlier discussion, either of these, or both together, can lead to significantly different final interpretations in their application to particular cases. 

 

It is perhaps possible to make a grand pronouncement that legitimacy, and especially democratic legitimacy, is more important than legality in that a band of brigands that seizes power with no pretense to legitimacy could operate its tyranny through a system of proclamations that was public, transparent, coherent, slow to change, possible of compliance, and consistently administered and adjudicated. King Rex could be a tyrannical usurper without committing any of Lon Fuller's eight sins against legality.

This theoretical precedence of democratic legitimacy over legality does not take us very far, however, in deciding how wrong a precedent can be, how untrue to the text, and yet be permitted to stand. Obviously relevant are such factors as the breadth and depth of the reliance on the precedent, the importance of the interests involved and how wrong and in what ways wrong was the precedent decision. 

 

One promising principle might be that if the textualist judge now has to go to the extensiveness tie breaker in settling on an interpretation, then a precedent that went the other way should stand. Perhaps the same should be said for a deliverance of historical linguistics that relies on thin evidence or that is otherwise supported at a lower degree of confidence, even if it did not require a tie breaker. It is widely accepted that wrong precedents interpreting an entrenched constitution, difficult of amendment, should, in that respect, be treated less gently than wrong statutory interpretation that can easily be corrected by the legislature. 

 

To go much beyond these elementary observations would require a thorough examination of stare decisis as well as involve us in the difficult task of balancing two important sets of values against each other. 

 

So textualism is not the last word in interpreting statutes and constitutions in stare decisis jurisdictions. It should, however, be the first word. 

 

1 If said while talking about gymnasts, we would interpret "tall woman" differently than we would in a conversation about volleyball players.

2

If there are many alternatives and especially if there could be a good faith dispute as to how many, ambiguity begins to resemble vagueness. At the extreme, the distinction between the vague and the muliply ambiguous may disappear. Remember, however, that most disputes about disambiguation are binary, and so the distinction between ambiguity and vagueness is for most practical purposes sharp.

3 It may be that most legal disputes about implications come down at bottom to issues of vagueness or ambiguity. We know from mathematics, however, that there is not always a method for finding implications where there is no vagueness or ambiguity involved. It is possible that the very difficulty of working through the logic of a legal text may give rise to disagreement unrelated to any vagueness or ambiguity.

4 Griswold v. Connecticut, 381 US 479, 484 (1965).

5 One point at which there will be disagreement is as to whether the translation could or should be made (in the idealization) as nearly as possible “all at once” or over a period of time as the need arises and the relevant data present themselves.

6 If good arguments exist for the legitimacy of law in some thoroughly non-democractic, non-representative regime, then there will often by a way to construct an argument for textualism in interpreting the laws of that regime. If the proclamations of a king are legitimate because they are the proclamations of one anointed by God, it is still the text of what is proclaimed that carries the legitimacy. It is true, however, that in a divine right of kings theory, there will be an argument that legitimacy is carried primarily by what the king intended rather than what he proclaimed. Not every theory of political legitimacy supports textualism.