University of Virginia School of Law

Public Law and Legal Working Paper Series 2011- 35

On Open Texture of Law

Fredrick Schauer
University of Virginia School of Law

September 2011

This paper may be downloaded without charge from the

A complete index of University of Virginia School of Law research papers is available at

Electronic copy available at: http://ssrn.com/abstract=1926926

ON THE OPEN TEXTURE OF LAW

Frederick Schauer

University of Virginia

Summary

H.L.A. Hart and others have argued that law is open-textured. It is in the nature of law, they

say, that it necessarily possesses an open texture going beyond the open texture of the

language in which legal rules are written. But when we examine the question of open texture

in light of Harts claim that the open texture of law entails the necessary defeasibility of legal

rules, we discover that Hart and his followers are mistaken. Both the alleged open texture of

law qua law and the defeasibility of legal rules are contingent features of certain legal rules in

certain legal regimes, but neither are necessary components of the nature of law or the nature

of rules.

1. Introduction

In The Concept of Law, H.L.A. Hart famously claimed that the law is open-textured (Hart

1994, 123, 128-36). Hart borrowed the term from his Oxford colleague Friedrich Waismann

Electronic copy available at: http://ssrn.com/abstract=1926926

as well as language is open-textured. But is Harts claim correct? Waismann argued that

language is necessarily open-textured, but is there an open-texture to law distinct from, or at

least going beyond, the open-texture of the language in which legal rules are written? Hart

seemed to believe so, but it is not at all clear that Hart was right in so believing. Nor is it

apparent that Hart and others are correct in maintaining that laws open texture produces the

necessary defeasibility of legal rules. For even if law is open-textured, it does not follow for

law, as it may follow for language, that the uncertainty generated by an instance of open

texture is necessarily an occasion for revising an existing legal rule. Indeed, resistance to

precisely such revision may well be what characterizes rules. If that is so, then Harts insistence

on laws essential defeasibility may well be mistaken. Or so I shall argue here.

2. Waismann on the Open Texture of Language

I begin with a prcis of Waismanns basic idea. Arguing against the logical positivist

credo that meaning was a function of verifiability (Ayer, 1946), Waismann observed that even

the most precise and carefully delimited empirical terms might nevertheless produce

uncertainty in the face of unforeseen and virtually unimaginable instances. Because language

cannot anticipate all possible occurrences in all possible worlds, he argued, there persists the

ineliminable potential that a definition of an empirical concept bounded in all now-foreseeable

dimensions can break down in the face of unforeseen and unforeseeable events. To illustrate

the phenomenon, Waismann observed that if we attempted to verify the statement, There is a

cat next door, no amount of seeing, touching, or patting the cat to verify that it was indeed a

2
cat could foreclose the logical possibility that the creature we thought was a cat according to all

of the existing criteria of catness might then grow to gigantic size, or die and then come back

from the dead. Were such eventualities to occur, Waismann argued, the existing criteria for

application of the term would collapse. An uncertainty that had not previously existed would

arise precisely because of the appearance of the unexpected, and it is central to Waismanns

idea of open texture that the possibility of the unexpected, at least for material object

statements, can never completely be foreclosed.

At about the same time that Waismann was writing about open texture, J.L. Austin

captured the identical idea even more vividly. Austin asked us to imagine a creature that

appeared to be a goldfinch according to the existing definition of what it is to be a goldfinch,

but which then proceeded to explode, or quote Virginia Woolf (Austin 1946, 160). Under such

circumstances, he says, we dont know what to say. It is precisely the lack of knowing what to

say (or do) in the face of the unforeseen and unforeseeable that Waismann originally called, in

German, Die Porositt der Griffe (the porosity of concepts), and then, at the suggestion of

W.B. Kneale (Waismann 1945, 121), relabeled in English as open texture. Thus, as Stewart

Shapiro puts it in summarizing Waismanns idea, P exhibits open-texture if there are possible

objects p such that nothing in the established use of P, or the non-linguistic facts, determines

that P holds of p or that P fails to hold of p. In effect, Pp is left open by the use of language, to

date. (Shapiro, 2011).

As Waismann explicitly stressed, open texture is not to be confused with vagueness

(Waismann 1945, 122). Some terms are indeed vague in that they are indeterminate with

3
respect to known (existing or currently imaginable) applications. It takes no great foresight to

understand that some collections of things may or may not be heaps, that certain kinds of

behavior may or may not be kind, or that a car traveling at 90km/hour on a particular stretch of

highway may or may not be traveling quickly. In such cases, we perceive now that some

applications of the term are indeterminate. But open texture is different. It is the possibility of

vagueness the potential vagueness of even those terms that appear to have no

uncertainties with respect to known or imagined applications. We could list all of the existing

criteria for something being a goldfinch, and thus the term goldfinch would not be vague for

any of the known existing goldfinches and non-goldfinches in the world. But because such a

precise and non-vague definition of goldfinch would not likely contain the criteria of does

not explode or does not quote Virginia Woolf, the occurrence of the previously unimagined

would render the hitherto non-vague term newly vague with respect to that application. And

because such a possibility could never entirely be eliminated for empirical concepts, Waismann

maintained, attempting to define the meaning of such concepts in terms of verifiability was

destined to failure.

3. Hart and the Open Texture of the Law

In analyzing the operation of legal rules, Hart purported to adopt Waismanns idea and

adapt it to law. Hart claimed that rules necessarily possess a fringe of vagueness or open

texture (Hart, 1994, 123), and he attributed the idea to Waismann (Hart, 1994, 297). But it is

not obvious that Hart and Waismann were talking about the same thing. One problem is that

Harts language is ambiguous as between the view that vagueness and open texture are

4
distinct, as Waismann argued, and the view that open texture is for Hart simply a synonym

for vagueness. Perhaps he is merely following the common practice of using the x or y

locution to say that y is an alternative and possibly better way of saying x, as when I describe

my dog as my pet, or my cherished companion.

Although Harts language is susceptible to that interpretation, a more plausible reading

is that Hart did see vagueness and open texture as different, but not in the same way as

Waismann. After introducing the term open texture, Hart defines it as the fringe area of a

term with clear and unclear core and fringe (or core and penumbra) applications. Thus, a

man who is two meters in height is clearly tall, and a man who is one and one half meters is

clearly not tall, and so a man who is in the area of uncertainty 1.85 meters, say exists in the

fringe of the term tall. And the same applies to terms less pervasively vague than tall that

is, to terms which, unlike tall, are determinate for most of their known applications. Take

wine, for example. Its extension is quite precise, such that the vast majority of substances,

and even the vast majority of potable liquids, are easily classified as wine or not wine. But

whether non-alcoholic wine is wine at all is not so clear. Consequently, we can say that non-

alcoholic wine is located in the relatively small fringe of a term with a large core of settled

meaning a term containing, overwhelmingly, clear cases of application or non-application.

That open texture was to Hart the label not for potential vagueness but for the actual

vagueness surrounding the determinate applications of partially non-vague terms becomes

even more plain when we examine his introduction of the now-famous example of a rule

prohibiting vehicles in the park (Hart 1994, 129; Schauer 2008). For most of the paradigm,

5
clear cases, the application of the term is straightforward, and Hart offers motor-cars, buses,

and motor-cycles as examples of such clear cases. But with respect to an electrically propelled

toy motor-car, he says, the application of the term vehicle is no longer straightforward.

Accordingly, with respect to the question whether the no-vehicles-in-the-park rule applies to

the electrically propelled toy motor-car, Hart continues, the law can be seen to have exhibited

its open texture. He then goes on to say that in such cases we confront the issues at stake and

[] then settle the question by choosing between the competing interests in the way that best

satisfies us. In doing so we shall have rendered more determinate our initial aim, and shall

incidentally have settled a question as to the meaning, for purposes of this rule, of a general

word. (Hart 1994, 129). The open texture of the law means that there are, indeed, areas of

conduct which must be left to be developed by courts or officials striking a balance, in the light

of circumstances, between competing interests which vary in weight from case to case. (Hart

1994, 135).

4. The Open Texture of Law and the Open Texture of Language

Harts explication of the open texture of legal rules is as interesting as it is important,

but Hart may be less explaining or using Waismanns idea than he is modifying it. Of course

many terms do have cores of settled meaning and fringes of now-debatable applications. What

Hart said about the term vehicle in a legal rule holds also for most terms. And it holds even

when those terms are not components of prescriptive rules. In numerous contexts, there are

core and fringe applications of bus, run, blue, and dead, for example, and so the core

and fringe feature of most of the words in any language will be part of the language of legal

6
rules as well. Waismanns point, however, is different. It is that even terms with no currently

known fringes may still develop such fringes in the face of currently unforeseeable applications,

and that this possibility can never completely be eliminated. The claim of languages open

texture is thus a claim about (empirical) languages inevitable potential uncertainty, and it is

this idea that Hart seeks to transfer to legal decision-making.

In one respect Hart is plainly correct. Law is inevitably potentially uncertain just

because language is inevitably potentially uncertain. Legal rules are ordinarily canonically

written in language, and are typically (or perhaps necessarily) understood in language even

when, as is often the case in common law systems, they may have no single formal canonical

inscription in a book of statutes, regulations, or constitutional provisions. Because legal rules

are thus language-dependent, the properties of language in general are present with respect to

legal rules expressed in language. More particularly, the phenomenon of open texture, as

identified by Waismann, must be a feature of all legal rules.1 To illustrate, imagine a legal rule

pertaining to goldfinches say, a rule declaring the goldfinch to be an endangered species and

thus protected against hunting, capturing, and the like. The rule would specify what could or

could not be done with or to goldfinches, and, because goldfinch is a term with no special

legal meaning, might then implicitly or explicitly incorporate the ordinary meaning of

goldfinch then present in the language. But if someone were charged with violating the law

by, say, capturing an exploding goldfinch, the very uncertainty of whether such a creature was a

goldfinch at all would make the legal consequences of the capture pro tanto uncertain as well.

1
For purposes of this article, I assume that Waismann is basically correct. My question is not
whether Waismann is, ultimately, right, but what, if Waismann is right, this says about law.

7
 If Waismanns idea is sound, then open texture is indeed an ineliminable feature of law,

but it is ineliminable in law precisely because it is ineliminable in language. As long as law is

written and understood in language, then the characteristics of language infuse law as well.

Law is thus open-textured, and necessarily so, just because of the ineliminable open texture of

language, but this conclusion does not yet suggest that legal open texture is any more than, or

any less than, the open texture of the terms in which the law is written and understood.

5. Is the Open Texture of Law More than the Open Texture of its Language?

The open texture of language thus produces open texture in law, but this conclusion is

neither startling nor particularly interesting. Insofar as language is open-textured, it follows

that any use of language is pro tanto open-textured. Law is thus open-textured, but so too are

novels, theatrical productions, everyday conversations, restaurant menus, and the instructions I

give to the automobile mechanic. If Waismann is right, then the phenomenon of open-texture

pervades language use, and Austins exploding goldfinch can appear and make the immanent

open texture of language present and problematic in any use of language, including but by no

means limited to legal language.

Hart, however, seems to make a broader claim, as his examples demonstrate. The

important difference for our purposes between Austins exploding goldfinch and Harts

electrically propelled toy motor-car is that we can not only imagine the latter, but can also

describe it with existing linguistic tools. A toy electric motor-car does not exemplify the open

texture of language, and so for it to be an example of the open texture of law, Hart must have

had in mind a feature of law distinct from the open texture of language.

8
 That there is something about law, or more precisely about legal rules, that seems to

Hart analogous to the open texture of language, but which is not simply the open texture of

language, is suggested by the no-vehicles-in-the-park example. The no-vehicles-in-the-park rule

is a straightforward example of a rule. Many rules, including this one, are presently both under-

and over-inclusive with respect to their background justifications (Alexander & Sherwin, 2001;

Schauer 1991). Consider the typical highway speed limit. The speed limit is a rule, designed to

serve the background justification of fostering highway safety. But we know now that a

designated speed limit aimed at ensuring safe driving will excessively limit some experienced

and safe drivers under ideal conditions who are in fact driving safely, albeit above the speed

limit. And by reaching some instances of safe driving, the rule reaches beyond its background

justification, and is accordingly over-inclusive. Conversely, some forms of unsafe driving

driving just below the limit in snow and ice, for example, or an inexperienced drivers driving

right at a high motorway limit are not covered by the speed limit rule. The rule is thus under-

inclusive to the extent that some behaviors covered by the rules background justification of

ensuring safety are not covered by the rule itself. But neither such over- nor under-inclusion

represents open-texture in either Waismanns or Harts senses, because neither involves a

confrontation with what was previously unexpected or undecided. For all of the reasons of

predictability and consistency, among others, that we use rules, the speed limit rule is

knowingly adopted with full-knowledge of its under- and over-inclusion, and with, ideally, full

appreciation of the fact that achieving the virtues of rule-based decision-making involves

enduring the vices of a rules under- and over-inclusiveness.

9
 In other instances, however, the tokens of under- or over-inclusiveness may not be

imagined by the rule-makers at the time a rule is made, and thus not anticipated in selecting

the language of the rule. Such instances are presumably what Hart had in mind when he used

the example of the toy electric motor-car (Bix 1993, 17-35). He considered it an example of

legal open-texture because a particular application of a rule was not settled by the language of

the rule, as it would have been had the rule-maker specified that the prohibited vehicles were

of a certain size, or propelled in a particular way, or capable of carrying a designated number of

passengers. But in Harts example, the extension of the rule is not so limited, extending to

anything that is a vehicle. And because we do not know whether the toy electric motor-car is

a vehicle, we do not know whether it contravenes the no-vehicles-in-the-park rule.

The mere fact that a new occurrence is not covered by a rule, however, need not

produce the same we dont know what to say response that the exploding goldfinch

occasions with respect to the previous criteria for proper use of the word goldfinch. The rule

could have carefully defined vehicle in all then-known directions, and could also have said

that anything not a vehicle according to that definition would not be covered by the rule. And

the rule could accomplish this goal with then-existing linguistic resources. Were the rule to be

formulated in this way, and if, say, the rule defined a vehicle in a way that limited vehicles for

the purpose of the rule as being of a certain size, or having the capacity to hold a certain

number of human passengers, then the toy electric motor-car would simply not have been

covered by the rule, even though this particular application was not imagined at the time the

rule was written and enacted. But if the rule, as in Harts example, is specified simply in terms

of vehicles, then the presence of the toy electric motor-car might well be an example of the

10
limited foresight or imagination of the particular drafters of a particular rule, but little more

than that. If the term vehicle, standing alone, is indeterminate with respect to toy electric

motor-cars, then we have an example of the open texture of language as manifested in a legal

rule, but not an example of a feature of law going beyond the open texture of language. And if

the term vehicle, standing alone, is determinate with respect to toy electric motor-cars

(assuming that toy electric motor-cars are vehicles), then this is like the speed limit an

application lying within the scope of the rule but not serving the rules background justification.

This, to repeat, is a familiar (and often intentional) feature of rules, but does not generate the

quandary or uncertainty that characterizes open texture.

6. On the Possibility of Closure Rules

As noted above, rules may contain closure provisions, excluding (or, in theory, including)

all applications not plainly encompassed by the rule. And as Hart observed (Hart 1994, 272),

Jeremy Bentham imagined and discussed just this kind of rule, one with its own specified

closure. For Bentham, who notoriously abhorred the idea of judicial legislation, a gap in the law

was not the occasion for judicial gap- filling, but for legislative action, albeit legislative action

prompted by the a proposal from the judge. Bentham, with his penchant for neologisms, called

this the suspensive power or sistitive function, whereby a judge would suspend execution

of the law pending legislative revision (Postema 1986, 434-439). For present purposes,

however, the issue is not principally about the strategy to be followed when a legal rule does

not plainly cover some occurrence in the world. Rather, Bentham shows that a set of rules can

in theory provide a closure or default rule specifying what is to happen in all events not plainly

11
covered by the existing rules (Gardner 2001, 212). Consider the implicit closure rule of the

Anglo-American criminal law, according to which all behavior not clearly prohibited by an

enacted legal prohibition is permitted. Or consider a hypothetical rule of private law, pursuant

to which claimants could recover against defendants only if an explicit rule authorized such

recovery, and only if the claimants actual claim clearly fell within such an explicit rule. In all

other cases, the claim would fail. This might well be a bad rule, neglecting to account for the

fact that in private law an uncompensated wrongfully caused injury is as regrettable as

compensation where there is no injury or wrong. But the justice or injustice of the rule is a side

issue. What is important is simply that a closure or default rule is entirely possible.

Any closure rule would, of course, remain subject to the open texture of the language of

the closure rule itself. But if the closure rules language were determinate, then there is no

reason why law would necessarily be compelled to face a situation in which the law simply runs

out, in the way that language runs out when confronted with the exploding goldfinch. Indeed,

Hart appears to concede as much, in his posthumously published Postscript, when he is

describing Benthams approach (Hart 1994, 272). By characterizing Benthams proposal a

form of closure rule -- as an alternative to judicial discretion, Hart implicitly concedes that there

are systemic options for law other than judicial discretion or judicial law-making. In so

conceding, however, Hart sets himself at odds with what he, earlier in The Concept of Law,

appears himself to have maintained, and sets himself at odds with what others have said about

legal defeasibility. And it is to this that we now turn.

7. The Possibility of Non-Defeasibility

12
 With the foregoing analysis in mind, consider Harts claim that a rule with an unless

clause can still be a rule (Hart 1994, 136). This itself seems uncontroversial. A rule that is

specified in If A, then , unless B form is merely a different way of formulating a rule with

multiple conditions for its application, as in If A and not-B, then (Schauer 1991, Stone 1985,

68-73).2 Moreover, using an unless clause is sometimes a method of formulating a genuine

reason-providing but still non-absolute rule. In American constitutional law, for example, a

governmental classification by race is unconstitutional unless the government can justify such a

classification by a compelling interest where it is understood that a compelling interest is a

heightened standard of justification substantially stronger than that minimally necessary a

mere rational basis typically needed to justify ordinary legislative classifications (Loving v.

Virginia 1967). But neither using an unless clause to formulate a rule with multiple conditions

for its application nor using one to express a non-absolute rule genuinely implicate the idea of

open texture. The latter is, to be sure, a way of accommodating for the possibility that the

details of all potential overrides cannot fully be anticipated in advance, but still differs from the

case in which specific law confronts the genuinely unexpected. Making rules defeasible and

using a vague standard for defeat is a way of managing the uncertainty of the future, but is

different from the case in which a non-vague rule confronts the genuinely unexpected.

Although intentional vagueness to manage an uncertain future is more or less

straightforward, Hart makes a stronger and more problematic claim when he maintains not

only that a rule with an unless clause is still a rule, but also that the list of unlesses cannot

2
I put aside burden-of-proof considerations, which may make the two formulations more
different than they would be absent such considerations (Chapman 1998; Finkelstein, 1999).

13
be exhaustively specified in advance. (Hart 1994, 136; MacCormick, 1974, 117). In this he

marks a position now quite common in the legal literature. Richard Posner, for example, insists

that courts always (the word is crucial) retain the power to append ad hoc exceptions at the

moment of application even to existing, precise, and exceptionless rules (Posner 1988, 834-

835). Similarly, Richard Tur argues that courts necessarily have the power to set aside the

indications of rules in the service of equity or justice when applying the rule would produce an

unjust result or a one at odds with the rules background justification (Tur 2001), Ronald

Dworkin has described the list of legal principles as numberless, (Dworkin 1977: 43-44), and

Neil MacCormick has offered a slightly more qualified version (MacCormick 1995) of the same

claim. And although Ronald Dworkins later insistence that there is a real rule existing behind

the formal words of a written rule is mysterious (Bix 1993, 34-35), the heart of his argument is

the view, related to those just set out, that written rules are necessarily defeasible in the

service of the real rule lurking in the background (Dworkin 1986).

Thus a new argument is now before us -- that laws open texture is not merely the open

texture of laws language, and not simply the conventional claim that legal rules, like all rules,

can be treated as defeasible,3 but that legal rules are necessarily defeasible. The claim is not

that treating legal rules as defeasible is desirable, useful, or common, but that it is in the nature

of rules or the nature of law that legal rules are necessarily and thus invariably defeasible.

3
Indeed, there is a vast literature exploring the logic of legal defeasibility (Hage 2005; Prakken
1997; Sartor 1995; Stelmach & Broek 2006), but that literature attempts to model and analyze
defeasibility when chosen by a legal system, rather than addressing the question whether such
defeasibility itself is a legal or logical necessity.

14
 To be distinct from the defeasibility occasioned by the open texture of language, legal

defeasibility must exist within the area of linguistic determinacy. Were it otherwise, we would

simply be rehearsing in legal context the issues of linguistic open texture and consequent

linguistic indeterminacy. Thus the claims of Dworkin, MacCormick, Posner, and Tur, among

others, and suggested by Hart in asserting the impossibility of listing in advance all of the

possible unlesses, are best understood as claims positing defeasibility even in the face of

linguistic determinacy. Consider, to take the obvious example, the case featured by Ronald

Dworkin, the 1889 decision of the New York Court of Appeals in Riggs v, Palmer. In Riggs,

Elmer Palmers grandfather had named Elmer is his will, but was contemplating changing the

will, to Elmers disadvantage. In order to accelerate his inheritance, Elmer killed his grandfather

with poison. He was eventually caught, charged with murder, tried, convicted, and imprisoned.

But Elmer argued that he was still entitled to inherit his grandfathers estate, relying on the

New York Statute of Wills, which provided, to oversimplify, that anyone named as a beneficiary

in a will would inherit upon the death of the testator. Importantly, the Statute provided no

relevant exceptions to that rule, and the judges who thought Elmer should inherit and even

those who thought he should not agreed that the rule as written was clear and did not contain

grounds for an exception.

The judges who thought that Elmer should lose were in the majority, relying on the no

man should profit from his own wrong principle to produce that outcome. But it is important

that there was agreement among all the judges that a literal application of the rule would allow

Elmer to inherit. This was not, therefore, a case of linguistic vagueness, ambiguity, or open

texture in Waismanns sense. The language was clear. It is just that the clear language

15
happened to produce an unfortunate and seemingly unfair result. But is such a result an

indication of the open texture of law qua law? Is it an example of the necessary defeasibility of

legal rules?

Riggs illustrates the distinction between a rule that gives no answer and a rule that gives

a bad answer, and precisely this distinction undercuts the claim of necessary defeasibility. The

rule in Riggs did not contain an unless clause, and thus we might conclude, or Hart might

have concluded, that Riggs presents a situation not envisaged when the rule was drafted. And

we might also conclude that the situation presented by Riggs is one in which a previously

unimagined situation shows how a hitherto unneeded and unanticipated unless clause might

have been a good idea. Yet even if that is so, the rule without the unless clause is still capable

of generating a clear outcome, albeit a morally (in this case) unpalatable one. Consequently,

the existence of hypothetical unless clauses whose addition would have prevented an

unfortunate outcome is insufficient to make the case that legal rules are necessarily defeasible.

If there is such a case to be made, essentially a claim that all of the rules in a system should

have open-ended unless clauses, it must be a moral or policy claim rather than a logical or

linguistic one. A rule that ill serves its background justification in a particular case is still a rule

capable of producing an outcome, and whether such an outcome should be tolerated when it is

unsound as a moral or policy matter is different from whether such an outcome presents an

exploding goldfinch type of genuine uncertainty.

8. On the Grounds for Defeasibility

16
 When the New York court concluded in Riggs that Elmer could not inherit because of

the no man shall profit from his own wrong principle, it treated the most directly applicable

legal rule as defeasible in the service of justice. The no man may profit from his own wrong

principle is narrower than the full domain of justice, but few dimensions of justice probably

no dimensions of justice -- are not instantiated by some common law principle. Thus, the court

avoided the unjust outcome by implicitly concluding that all legal rules are defeasible in the

service of justice. And a similar situation arises when the rule-generated outcome result is

unreasonable or ridiculous, even if not unjust. In United States v. Kirby, for example, a

Kentucky law enforcement officer had been convicted under a federal law making it a crime to

interfere with the delivery of the mail. And that is exactly what Kirby had done. He had

unquestionably interfered with the delivery of the mail, but did so in the process of boarding a

steamboat to arrest a mail carrier named Farris who had been validly charged with murder by

Kentucky court. The case reached the Supreme Court, where the Court treated the statute as

defeasible, rejecting its literally dictated outcome where that outcome was inconsistent with

the purpose of the statute, inconsistent with commonsense, and inconsistent with justice.

If we depart the realm of the real and enter that of the hypothetical, we see the same

phenomenon captured in Lon Fullers famous response to Harts no-vehicles-in-the-park

example (Fuller 1958, responding to Hart 1958). Fuller imagined a group of patriots who

installed a fully functional military truck in the park as a war memorial. This would clearly be a

vehicle, Fuller argued, but it would just as clearly be absurd to exclude it from the park on the

authority of the no vehicles in the park rule.

17
 Had this been an empirical debate, Fuller may well have won. If Hart is understood to

claim that the plain meaning of the terms of a rule provides a conclusive answer in most real

cases in real legal systems, and if Fuller is understood as responding that the plain meaning

answer was typically defeasible and rarely conclusive in well-functioning legal systems, then

Fuller is closer than Hart to the reality in many modern legal systems (Schauer 2008).

But Fuller made a broader claim. For him (and Dworkin, MacCormick, Posner, and Tur),

it is not merely a contingent empirical fact that legal rules are typically (and, to Fuller,

preferably) defeasible in common law legal systems, but also that the defeasibility of legal rules

is an essential feature of legality itself, on a par with those other desiderata of legality that for

Fuller came close to defining law itself (Fuller 1969). For Fuller, failing to treat a rule like the

no-vehicles-in-the-park rule as defeasible was simply to abandon reason, and for him it was of

the essence of law that it be reasonable. A system that did not allow purpose-based or

equitable override of the plain indication of a legal rule when necessary to achieve a reasonable

outcome was for that reason just so much less of a legal system, and perhaps not even a legal

system at all. Indeed, we can understand Fullers claim in the best light, and without saddling

him with the view that nondefeasible law is not law at all, by interpreting him as maintaining

that non-defeasible law is necessarily defective as law, even if the defective law is still law. Just

as any boat that leaks is defective as a boat even as it remains a boat, so Fuller is best

understood as insisting that any legal system is necessarily defective as a legal system, and as

law, insofar as it treats its rules as non-defeasible.

18
 The course of action that Fuller rejects, however, is conceptually possible, often

adopted, and often far from unjustifiable. The New York Court of Appeals could have said that

Palmer would inherit despite the wrong he committed, just as the United States Supreme Court

could have concluded that Kirby violated federal law even though he did it for good reason, and

even though punishing Kirby was inconsistent with the purpose of the law he literally violated.

And some hypothetical judge could conclude that a war memorial made from a functioning

military truck was nevertheless a vehicle, and thus to be excluded from the park by virtue of the

no-vehicles-in-the-park rule. Such outcomes might be condemned as ridiculous or absurd, or,

even more pejoratively in some legal circles, formalistic, but they are neither conceptually nor

linguistically impossible. As long as words have literal meanings with a context-independent

core, and as long the meaning of a word is not entirely a function of the particular context in

which it is used on a particular occasion (Bix, 1993, 21-22), then rules which are written in

words can indeed generate poor outcomes, and judges can embody those poor outcomes in

their rulings.

Indeed, not only can this happen, but it does frequently. The result in Riggs is more

exceptional than normal, even in the anti-formal American judicial system. There are other

cases in which the outcomes do resemble those in Riggs, but in many others beneficiaries who

were culpably responsible for the death of the testator were allowed to inherit. (Schauer 2004).

Indeed, suboptimal or unjust applications of statutes are often allowed to stand, as in

Tennessee Valley Authority v. Hill, in which a literal application of the Endangered Species Act

mandated the preservation of the habitat of a small, unattractive, and ecologically unimportant

fish called the snail darter, even at great cost to the aggregate public welfare by virtue of

19
blocking of an important public works project. And in United States v. Locke, a statute setting a

filing deadline of prior to December 31 was upheld even when its literal application resulted

in the arguably unfair exclusion of a claim filed on December 31 by someone who assumed, not

unreasonably, that the statute really meant to say on or prior to December 31.

There are numerous other examples, in the United States and elsewhere, and it would

be mistaken to describe the defeasibility of legal rules as a universal or even overwhelmingly

common feature of legal decision-making. Although legal decision-makers sometimes treat

rules as defeasible, frequently they do not. Often they treat the literal language of a rule

formulation as conclusive, and refrain from adding exceptions at the moment of application,

overriding the indications of rules in the service of justice or equity or efficiency, or modifying

rules at the moment of application. Such non-defeasibility may not be wise, but it is both

possible (Baker 1977, 38) and widespread. Rule-formulations have meanings distinct from the

background justifications lying behind the rules and distinct from what the best (or even a

good) rule-free outcome in some particular instance would be. Although treating rules as

defeasible is widespread in many legal systems, examples like those just offered show that rules

are often applied as written treated as non-defeasible even when what seem to be valid

defeating conditions are present. As a descriptive matter, it is hard to defend the position that

a legal system without widespread and legitimate defeasibility is for that reason no legal system

at all.

Moreover, non-defeasibility is hardly without justification. The arguments for rules are

familiar, and most of the arguments for rules in general are arguments for treating rules as non-

20
defeasible. When the designers of some decision-making environment are distrustful of judges

and other legal decision-makers who might be biased, corrupt, incompetent, ill-equipped for

the job, or just very rushed, then the designers might want to constrain those decision-makers

by rules rather than granting wide discretion. Indeed, when distrust is at its acme, decision-

makers may not be granted the authority to decide when some application of a rule is

ridiculous or absurd, let alone unfair, inequitable, unjust, or inefficient. It is easy to see that it

would be absurd to exclude the truck used as a war memorial from the park, but the real

question is whether and when some class of officials should be empowered to decide which

applications are absurd and which are not. Moreover, rules also serve to allocate decision-

making responsibility. To treat a rule as non-defeasible is often simply to decide that some but

not other officials will have the power to cancel, override, amend, or modify an existing rule.

And insofar as rules also bring the advantages of certainty, predictability, settlement, and

stability for stabilitys sake, treating the rules as defeasible comes at the sacrifice of each of

these values, even though of course it brings the potential advantages of fairness, equity, and,

in theory, reaching the correct result in every instance.

9. The Contingency of Defeasibility

It now becomes apparent that the traditional defenses of the necessary defeasibility of

legal rules whether Harts, or Posners, or MacCormicks, or Turs mostly rest on a certain

view about the powers and abilities of judges. Few people would maintain that police officers

or ordinary bureaucrats, for example, should have the power to revise the rules that constrain

them when those rules appear to indicate a poor outcome in a particular case. And if that is so,

21
then the view that defeasibility in the hands of judges is required by the Rule of Law while

defeasibility in hands of others is not turns out to be a view about the capacities of judges

within particular legal systems. So long as we can imagine something properly called a legal

system in which the power of rule revision and rule override is not entrusted to judges, then we

can imagine something properly called a legal system in which defeasibility is somewhat or

even largely absent.

Defeasibility is thus not a property of rules at all, but rather a characteristic of how some

decision-making system will choose to treat its rules. The Wittgensteinian maxim that rules do

not determine their own applications reminds us that how a rule will be treated is not

something inherent in the rule itself. If the plain or literal (but not necessarily the ordinary

language) meaning of a rule can indicate an outcome, it is a function not of the rule but of how

the rule will be treated whether that indicated outcome is to be taken as conclusive,

presumptive, or even, at the extreme, as having no weight in itself, being but a totally

transparent (to its background justification, or to the all-things-considered best outcome)

heuristic or rule of thumb. The question of defeasibility is not a question about what is in a

rule, but is rather a question about how what is in a rule, or about how what a rule says, is to be

treated, and this is not, and can never be, something that can be determined by the rule itself.

The question of defeasibility is this exposed as descriptive and prescriptive, but neither

logical nor conceptual. It is logically and conceptually possible for rules to be interpreted,

understood, applied, and enforced according to the literal meaning of the component language

of their formulations. Whether in this or that legal system they are in fact so treated is a

22
descriptive question, and it turns out, as the examples above illustrate, that as a descriptive

matter defeasibility is less universal in actual legal systems than we might have thought, even in

the legal systems in which we might have most expected it to exist.

With respect to the prescriptive question, whether the literal meaning of a rule-

formulation will be treated as what the rule indicates, and whether what the rule indicates will

be treated as conclusive, are questions that cannot be answered by reference to the moral

goals of particularized justice. Those goals exist, to be sure, as Plato, Aristotle, and countless

successors have argued. But so too do the Rule of Law goals that might be thought of as the

goals of generalized justice, or aggregate justice, or systemic justice. And as long as those non-

particularistic goals have a place in our moral and prescriptive reasoning, then we cannot

conclude, the actual practices of some parts of some legal common law legal systems

notwithstanding, that the defeasibility of legal rules is a necessary part of all legal systems, or

that the defeasibility of rules in general is a necessary part of all decision-making environments.

Defeasibility may well be a desirable component of some parts of some legal systems at some

times, but it is far from being an essential property of law. And if law is not necessarily

defeasible, then law is not open-textured in Waismanns sense. Insofar as language is open-

textured, then so too, necessarily is law. But the open texture if laws language exhausts the

open texture of law, and thus when laws language does not exhibit languages open texture,

there is no residual open texture of law at all.

23
Bibliography:

Alexander, Larry, & Emily Sherwin 2001: The Rule of Rules: Morality, Rules, and the Dilemmas of

Law. Durham, North Carolina: Duke University Press).

Ayer, Alfred Jules 1946: Language, Truth and Logic. London: Victor Gollancz, Ltd., 2d. ed.

Austin, J.L. 1946: Other Minds. Proceedings of the Aristotelian Society (Supp.) 20, 148-187.

Baker, G.P. 1977: Defeasibility and Meaning. In: P.M.S. Hacker & J. Raz (eds.), Law, Morality,

and Society: Essays in Honour of H.L.A. Hart. Oxford: Clarendon Press, 26-57.

Bix, Brian 1993: Law, Language, and Legal Determinacy. Oxford: Clarendon Press.

Chapman, Bruce 1998: Law Games: Defeasible Rules and Revisable Rationality. Law and

Philosophy, 17, 443-480.

Dworkin, Ronald M. 1977: Taking Rights Seriously. London: Duckworth.

Dworkin, Ronald M. 1986: Laws Empire. Cambridge, Mass.; Harvard University Press.

Finkelstein, Claire Oakes 1999: When the Rule Swallows the Exception. In Linda Meyer (ed.),

Rules and Reasoning: Essays in Honour of Fred Schauer. Oxford: Hart Publishing, 147-175.

Fuller, Lon L. 1958: Positivism and Fidelity to Law A Reply to Professor Hart. Harvard Law

Review, 71, 630-672.

Fuller, Lon L. 1969: The Morality of Law (revised edition). New Haven, CT: Yale University Press.

24
Gardner, John 2001: Legal Positivism: 5 Myths. American Journal of Jurisprudence, 46, 199-

227.

Hage, Jaap 2005: Studies in Legal Logic. Dordrecht, Netherlands: Springer.

Hart, H.L.A. 1958: Positivism and the Separation of Law and Morals. Harvard Law Review, 71,

593-629.

Hart, H.L.A. 1994: The Concept of Law, 2d ed., Penelope A. Bulloch & Joseph Raz (eds.). Oxford:

Clarendon Press.

Loving v. Virginia 1967: 388 U.S. 1.

MacCormick, Neil 1974: Law as Institutional Fact. Law Quarterly Review 90, 102-129.

MacCormick, Neil 1995: Defeasibility in Law and Logic. In Zenon Bankowski, Ian White, &

Ulrike Hahn, eds., Informatics and the Foundations of Legal Reasoning. Dordrecht,

Netherlands: Kluwer, 99-117.

Posner, Richard A. 1988: The Jurisprudence of Skepticism. Michigan Law Review 86, 827-

858.

Postema, Gerald J. 1986: Bentham and the Common Law Tradition. Oxford: Clarendon Press.

Prakken, Henry 1997: Logical Tools for Modelling Legal Argument: A Study in Defeasible

Reasoning in Law. Dordrecht, Netherlands: Kluwer.

Riggs v. Palmer 1889: 22 N.E. 188 (N.Y.).

25
Sartor, Giovanni 1995: Defeasibility in Legal Reasoning. In Zenon Bankowski, Ian White, &

Ulrike Hahn, Informatics and the Foundations of Legal Reasoning. Dordrecht, Netherlands:

Kluwer, 119-158.

Schauer, Frederick 1991: Playing By the Rules: A Philosophical Examination of Rule-Based

Decision-Making in Law and in Life. Oxford: Clarendon Press.

Schauer, Frederick 1991: Exceptions. University of Chicago Law Review, 58, 871-904.

Schauer, Frederick 2004: The Limited Domain of the Law. Virginia Law Review, 90, 1909-

1956.

Schauer, Frederick 2008: A Critical Guide to Vehicles in the Park. New York University Law

Review 83, 1109-1134.

Shapiro, Stewart 2011: The Open-Texture of Computability (unpublished).

Stelmach, Jerzy, and Bartosz Broek 2006: Methods of Legal Reasoning. Dordrecht,

Netherlands: Springer.

Stone, Julius 1985: Precedent and Law. Sydney: Butterworths.

Tennessee Valley Authority v. Hill 1978: 437 U.S. 153.

Tur, Richard 2001: Defeasibilism. Oxford Journal of Legal Studies, 21, 355-368.

United States v. Kirby 1868: 74 U.S. (7 Wall.) 482.

United States v. Locke 1985: 471 U.S. 84.

26
Waismann, Friedrich 1945: Verifiability. Proceedings of the Aristotelian Society (Supp.) 19,

119-47, and in A.G.N. Flew (ed.), Logic and Language (First Series). Oxford: Basil Blackwell

(1951), 117-44.

27