H. L. A.

Hart and the "Open Texture" of Language

Author(s): Brian Bix
Source: Law and Philosophy , Feb., 1991, Vol. 10, No. 1 (Feb., 1991), pp. 51-72
Published by: Springer

Stable URL: https://www.jstor.org/stable/3504835

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Springer is collaborating with JSTOR to digitize, preserve and extend access to Law and
Philosophy

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 BRIAN BIX

H. L. A. HART AND THE

"OPEN TEXTURE" OF LANGUAGE*

ABSTRACT. H. L. A. Hart and the "Open Texture" of Lan

the writings of both Hart and Friedrich Waismann on
Waismann's work, "open texture" referred to the potent
under extreme (hypothetical) circumstances. Hart's use o
different, and his work has been misunderstood because t
underestimated. Hart should not be read as basing his a
discretion on the nature of language; primarily, he was put
argument for why rules should be applied in a way whic
discretion.

INTRODUCTION

In this article, I will offer a detailed analys

discussion of "open texture". In The Concept of Law
position on judicial interpretation halfway betw
rule-scepticism.2 H. L. A. Hart's middle position w
least, justified by) a theory of the open textur
concept comes from the work of Friedrich Waism
turn probably based on a constructivist view o
Wittgenstein put forward in the early 1930s.5

* I wish to thank Joseph Raz and Gordon Baker for thei

earlier drafts of this article.
H. L. A. Hart, The Concept of Law (Oxford: Clarendon Pres
2Id., pp. 121-44.
3 Hart, The Concept ofLaw, pp. 120-32.
4 See id., p. 249; Waismann, 'Verifiability', Proceedings of t
Supplementary Volume 19 (1945): 119-50.
See Baker, 'Defeasibility and Meaning', in P. M. S. Hacker
Morality and Society (Oxford: Clarendon, 1977), p. 51 and n. 7

Law and Philosophy 10: 51-72, 1991.

? 1991 Kluwer Academic Publishers. Printed in the Netherlands.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
52 Brian Bix

article will focus in turn on Hart's text and on its origins in the works
of Waismann and Wittgenstein.

In a chapter in The Concept of Law called 'Formalism and Rule Scepti-

cism', H. L. A. Hart argued that legal rules, whether promulgated by a
legislature or derived as the ratio of a prior case, characteristically have
a core of plain meaning. The decision whether a rule applies to a
particular situation often turns on delimiting the range of meaning of
a general term. For example, the application of the rule, "No vehicles
in the park", will usually turn on whether a particular object is a
"vehicle" for the purpose of the rule (or whether a particular area is a
"park" for the purpose of the rule). In plain cases, "the general terms
seem to need no interpretation . . . the recognition of instances seems
unproblematic or 'automatic', ... there is a general agreement in
judgments as to the applicability of the classifying terms."' However, in
cases in the "penumbra" of the term's meaning (for the purpose of the
rule), it no longer seems clear whether the general term should apply
or not. "[T]here are reasons both for and against our use of the general
term, and no firm convention or general agreement dictates its use."7
The tendency of rules to have "a fringe of vagueness",8 to become
indeterminate in their application to borderline cases Hart calls the
"open texture" of rules (and of language in general).9 Hart added that
the "open texture" of legal rules should be considered an advantage
rather than a disadvantage, in that it allows rules to be reasonably
interpreted when they are applied to situations and to types of prob-
lems that their authors did not foresee or could not have foreseen."'

') Hart, The Concept of Law, p. 123.

7 Id., pp. 123-24.
Id., p. 120.
Id., pp. 124-25.
IO Hart, The Concept of Law, pp. 125-26. Compare Anthony Quinton's discussion
of Waismann's idea of the "open texture" of concepts: "[T]he kind of linguistic
indeterminacy it implies is a positive advantage. It allows for the continuous

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 53

In this chapter, Hart was concerned with the problem of social

control through law, not questions of strategy or political theory, of
how social control could best be effected, but the preliminary question
of how social control could be possible. How can a government guide
its population's actions on the basis of legislation and precedent, and to
what extent will those means necessarily need supplementation? Hart
stated: "If it were not possible to communicate general standards of
conduct, which multitudes of individuals could understand, without
further direction, as requiring from them certain conduct when occa-
sion arose, nothing that we now recognize as law could exist"."
Hart considers two forms of guidance, corresponding to two sources
of law: examples, analogous to precedent, and verbal instructions,
analogous to legislation.12 Among those two, examples seem far less
clear and determinate. When someone tells us to do as he does, we
cannot be certain what aspects of his performance must be imitated
and where deviation is condoned because irrelevant. Transforming the
example into a verbal rule seems to avoid these problems. Now the
citizen "only" need "'subsume' particular facts under general classifica-
tory heads and draw a simple syllogistic conclusion".13 However, Hart
showed how general rules actually retain both the character and the
problems of guidance by example. According to Hart, when the rule
(for example, "No vehicles in the park") is enacted, both the legislators
and the public have in mind a particular problem, particular situations
that are to be brought about or avoided. In the "no vehicles" example,
the image is of excluding normal motor-car, bus, and motor-cycle
traffic from the park.'4 Interpretation of the rule is thus seen as similar
to reading a rule off an example, here the example being the problem
the legislation was meant to meet.

development of a language to accommodate new discoveries, as exemplified by

the progressive amplification of the scope of the concept of number from the
positive integers to complex numbers." Quinton, 'Introduction', in F. Waismann,
Philosophical Papers (Dordrecht: D. Reidel, 1977), p. xiii.
"Hart, The Concept ofLaw, p. 121.
12 Ibid.
Id., p. 122.
Id., pp. 125-26.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
54 Brian Bix

Hart used a mixture of a "paradigm" and a "criteria" approach to

meaning. According to Hart, our first move in defining a general term
for the purpose of a rule is to invoke the image, example, or particular
situation at which the rule was aimed. In interpreting the rule, "No
vehicles in the park", we might begin by thinking "If anything is a
vehicle a motor-car is one".'5 In deciding whether, for the purpose of
the rule, "vehicle" applies to roller skates or toy cars, one would "con-
sider ... whether the present case resembles the plain case 'sufficiently'
in 'relevant' respects".16 We begin with the plain case or the paradigm
(the motor-car) and then consider a list of criteria which allow us to
begin to evaluate how similar a purported extension would be. For
example, like a motor-car, roller skates make noise (but not nearly as
much) and they threaten safety and order (though the threat is on a
much lower scale). Further dissimilarities include the facts that roller
skates are far smaller than motorcars and that they do not pollute the
air. There are both similarities and dissimilarities; some criteria are
fulfilled, others are not. In Hart's language, "there are reasons both for
and against our use of a general term".17 This is the "open texture" of
rules, that particular situations arise that we were not thinking of
when proffering the rule and which are different in some ways from
the situation we had in mind (the paradigm) at that time.'8
Sometimes the extension of a general term from the original para-
digm case to a different case is clear, not because there are no differ-
ences between the two cases, but because the problem of extension has
come up many times before, and a consensus has developed as to

15 Id.,p. 123.
16 Id., p. 124.
17 Id., p. 123.
18 Cf. White, 'What Can a Lawyer Learn From Literature?' (book review),
Harvard LawReview 102 (1989): 2014, 2035:

It is ... the genius of law that it is not a set of "commands," but a set of texts
meant to be read across circumstances that are in principle incompletely foresee-
able. This is what it means to pass a piece of legislation, or to decide a case - or
even to draft a contract - at one point in time, with the knowledge that it will
in the future be brought to bear by others (or ourselves) in contexts, and with
meanings, that we cannot wholly imagine.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 55
H. L. A. Hart and the "Open Texture" ofLanguage

whether the term should apply.19 For Hart, the problem of "open
texture" will recur regularly, because there are "fact-situations, con-
tinually thrown up by nature or human invention, which possess only
some of the features of the plain cases but others which they lack".20
The slow building of a consensus about whether to apply a general
term to particular; relatively common, borderline cases will do little to
mitigate the problem of "open texture", for life will soon provide more
uncertain borderline cases to replace those convention has transformed
into "plain cases".

II.

I want to consider the intellectual origins of Hart's concept of "open

texture". I will trace the concept back to the writings of Friedrich
Waismann, and will also consider whether it can be traced one step
further back to the work of Wittgenstein.21 This account of "open
texture", though analogous to Waismann's account that Hart credits,22
differed from it in a number of ways.
To understand Waismann's concept of "open texture", it is useful to
see it within the larger context of his work in general. Waismann's
work was devoted largely to presenting Wittgenstein's ideas in a more
accessible form; however, some of Waismann's concepts were his own

19 See Hart, The Concept of Law, p. 123: "The plain case[s] ... are only the
familiar ones, constantly recurring in similar contexts, where there is general
agreement in judgments as to the applicability of the classifying terms."
2( Id., p. 123.
21 While Hart's idea of "open texture" was derived from Waismann, his talk of
"a core of certainty and a penumbra of doubt", Hart, The Concept of Law, p. 119,
may have come from Bertrand Russell, though no attribution for those ideas was
given. See Russell, 'Vagueness', in Collected Papers of Bertrand Russell, vol. 9
(London: Unwin Hyman, 1988), p. 149 ("The fact is that all words are attribut-
able without doubt over a certain area, but become questionable within a
penumbra, outside which they are again certainly not attributable."). (That article
was first read to the Jowett Society in 1922, and published in the Australasian
Journal of Psychology and Philosophy in 1923.)
22 Hart, The Concept ofLaw, p. 249.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 56 Brian Bix

extension of
belonged t to
the philosoph
with the Rea
many of the
For example,
ment against
ments are e
sense-datum s
"Open textu
verification
concepts, Wa
translated in
cannot be co
actual use m
in that we c
doubt may se
Like Hart, W
tions we hav
however fain
that may be
that means t
stances in wh
where, Wais
be constructe

23 See Quinton
24 Waismann,
25 Waismann, 'V
Volume 19 (19
Meaning and U
1979), suggesti
possible world
26 Waismann,'
27 Id., 123. Wai
121 n. *, could
28 Id., 123.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 57
H. L. A. Hart and the "Open Texture" ofLanguage

unforeseen factor emerging", "the process of defining and refining an

idea" to meet each new factor "will go on without ever reaching a
final stage".29
To try to understand Waismann's argument better, I will consider
similar arguments from his other writings. Though discussions of
verification are often the context for Waismann's analysis of "open
texture", the concept is not concerned primarily with problems in that
area. In The Principles of Linguistic Philosophy,3" Waismann seemed
almost indifferent regarding the question of verification. He wrote:
"We were asking the question whether the assertion that a ball is lying
on the table can be finally verified. The answer to this question is that
can be decided on our part by an arbitrary determination."31 It all
depends on what we mean by "verified", and there is no a priori reason,
according to Waismann, to choose one approach over another. Under
some approaches, the statement would never have final validity.32
Waismann argued that our language, as well as our usual
approaches to verification, is organized to respond to normal back-
ground conditions and to the small-scale problems of everyday life.
Our language and our grammatical rules do not serve us well if we
start to imagine wildly unusual circumstances or deceptions of a
Cartesian magnitude.33 Here Waismann's comments are quite relevant
to "open texture":

29 Id., 125.
"' Waismann's long work, cited as often as an explication of Wittgenstein's ideas
as it is to show Waismann's ideas. F. Waismann, The Principles of Linguistic
Philosophy (London: Macmillan, 1965). There are two long segments in the work
relevant to the concept of "open texture". See id., pp. 68-86, 221-25. It is
difficult to date the material, as Waismann revised the text continually over the
last decades of his life, at some places even incorporating or respondng to ideas
from Wittgenstein's Philosophical Investigations. The text was not published until
nine years after his death. Quinton, 'Introduction', in F. Waismann, Philosophical
Papers, p. ix.
31 Waismann, The Principles o Linguistic Philosophy, p. 74.
3 Id., pp. 74-75.
33 Id., pp. 75-76.
Anthony Quinton, in his reading of Waismann, chose to emphasize changes

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
58 Brian Bix

The laws of any

habits and need
time, and able t
foundation to st
as a rule, notice
events. Similarly
circumstances a
to provide for
deplore the insuf

In a later cha
topic of defin
at all like an
concept in a
hypothetical s
nobody can g
but emitted a
comments on

Try as we may,
any doubt. We
example, "This i
practical purpo
other direction

in "background
unrelated to tho

Waismann's poi
there are borde
not, though he
their application
included among
conceivable case
conditions are no

Quinton, 'Intro
Reidel, 1977), p.
34 Id., p. 76.
3 Id., p. 222.
3' Id., pp. 222-2

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 59

could imagine hundreds of situations which would necessitate new limitations.

Are our concepts therefore incomplete, inexact? But what then would be an exact
concept? One which anticipated all cases of doubt, one which is outlined with
such precision that every nook and cranny is blocked against entry of doubt? But
then we have to own, that no concept satisfies this demand; and we begin to see
that there is something utopian in the demand for absolute precision. A concept
is good if it fulfils the purpose for which it has been devised.37

In the article "Language Strata",38 Waismann argued that different

types of statements - e.g., sense datum statements, material object
statements, aphorisms, and natural laws - must be analyzed in dif-
ferent ways. "Statements may be true in different senses, verifiable in
different senses, meaningful in different senses. Therefore the attempts
at defining 'truth', or at drawing a sharp line between the meaningful
and the meaningless, etc., are doomed to fail."39 It is a mistake to try to
apply the analytical tools of one language stratum to another, or to try
to reduce one stratum to another (as Phenomenalism and behav-
iourism attempt to do40). Here, and throughout the article, Waismann's
discussion of different language strata resembles Wittgenstein's discus-
sions of different "language games".
There are two arguments in "Language Strata" relevant to the
concept of "open texture". First, material object statements cannot be
reduced to a collection of sense-datum statements: "a statement about
a cat is a statement about a cat: and not a truth-function of sense-
datum statements, or an infinite group of sensibilia, or heaven knows
what".41 Second, the description of material objects (as contrasted with,
e.g., geometrical figures) is never complete:

However many features I may assert of a thing, say of this chair, or however

7 Id., p. 223.
38 Waismann, 'Language Strata', in A. Flew, ed., Logic and Language, Second Series
11 (Oxford: Basil Blackwell, 1961). Though not published until 1961, the article
is an unrevised version of the paper that had been read to the Jowett Society in
1946. Ibid.
39 Waismann, 'Language Strata', p. 26.
40 Id., pp. 28-29.
4 Id., p. 29.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 60 Brian Bix

many relations
many statemen
where my descr
increment in k
edge of it is alw

Both of these
"Verifiability"

III.

Though Waismann's concept of "open texture" is said to derive from

a middle period of Wittgenstein's thought, corresponding with the
material eventually published as Philosophical Remarks and Philosophical
Grammar,43 related ideas remain in Wittgenstein's later writings as well.
The following quotations are from Philosophical Investigations:44

I say "There is a chair." What if I go up to it, meaning to fetch it, and it

suddenly disappears from sight? - "So it wasn't a chair, but some kind of
illusion." - But in a few moments we see it again and are able to touch it and so
on. - "So the chair was there after all and its disappearance was some kind of
illusion." - But suppose that after a time it disappears again - or seems to
disappear. What are we to say now?45
It is only in normal cases that the use of a word is clearly prescribed; we
know, are in no doubt, what to say in this or that case. The more abnormal the
case, the more doubtful it becomes what we are to say.4

42 Id., p. 27.
43 L. Wittgenstein, Philosophical Remarks (Chicago: University of Chicago Press,
1975); L. Wittgenstein, Philosophical Grammar (Oxford: Basil Blackwell, 1974).
44 Which was first published eight years after the article which contained
Waismann's primary discussion of"open texture".
45 L. Wittgenstein, Philosophical Investigations (New York: Macmillan, 1958),
section 80; cf. L. Wittgenstein, Philosophical Grammar, p. 220 ff.
46 L. Wittgenstein, Philosophical Investigations, section 142; see generally G. P.
Baker and P. M. S. Hacker, Wittgenstein: Rules, Grammar & Necessity (Oxford, Basil
Blackwell, 1985), pp. 229-32. There is even one place where Wittgenstein
seemed to be discussing an analysis similar to "open texture" in a legal context:
"It is as if our concepts involved a scaffolding of facts."

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" of Language 61

Despite these surface similarities, Gordon Baker and Peter Hacker

claimed that there is actually a large conceptual distance between
Waismann's concept and the ideas of the later Wittgenstein, at one
point stating: "according to the outlook of the Investigations", "Wais-
mann's concept of open texture is doubly incoherent".47 They offer
two criticisms of Waismann's concept. First, that Waismann's "hypoth-
eses" of material object statements inappropriately "transcend all
possible experience".48 The criticism refers to an earlier Wittgenstein
comment:

[H]ow can I even make the hypothesis if it transcends all possi

How could such a hypothesis be backed by meaning? (Is it not like
not backed by gold?)49

Second, Waismann's concept allegedly "presupposes a dis

ception of what it is for a set of rules to be comple
plete").50 I will consider the two criticisms in turn, in th
to determine whether Waismann's analysis is correct, but
whether or to what extent Waismann's analysis actually
that of Wittgenstein (as Baker and Hacker implied).

That would presumably mean: If you imagine certain facts othe

them otherwise, than the way they are, then you can no long
application of certain concepts, because the rules for their appl
analogue in the new circumstances. - So what I am saying come
is given for human beings and a jurisprudent may well be capa
consequences for any case that ordinarily comes his way; thus the
has its use, makes sense. Nevertheless its validity presupposes all s
and if the being that he is to judge is quite deviant from ordinary
then e.g. the decision whether he has done a deed with evil int
not difficult but (simply) impossible.

L. Wittgenstein, Zettel (Berkeley: University of California, 1970), sec

47 G. P. Baker and P. M. S. Hacker, Wittgenstein: Understandin
(Oxford: Basil Blackwell, 1980), 383 n. 12.
48 Id., p. 432.
49 L. Wittgenstein, The Blue and Brown Books (Oxford: Basil Blackwell, 1958), p.
48.

35 Baker and Hacker, Wittgenstein: Understanding and Meaning, p. 432.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 62 Brian Bix

I am not su
experience"
object statem
of sense-da
possibility o
applied in un
concepts tak
now; they d
ences that w
having. Wit
possible expe
text, it seem
different ta
our knowled
The second
corrective t
stein that ou
completely v
this "deficie
in normal ci
here from a
make the dis
sense to spe
delimiting
being "incom
characterizat
see this dispu
Finally, Bak
texture" by
hypothesis (H

51 See L. Wittg
52 See, e.g., W
tic Philosophy,
53 See, e.g., Ba
432-33.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 63

esis can be made only more or less probable by any relevant evidence
.."54 The belief that material object statements are "hypotheses"
made more or less probable - but never completely verified or falsi-
fied - by our experiences apparently intrigued Wittgenstein during his
"middle period"55 and has some resonance with an early Waismann
article, "Hypotheses", that was not published during his lifetime.56
However, Waismann did not elaborate, or even mention, a "hypoth-
esis" position in the later articles (discussed in detail above) where the
concept of"open texture" was put forward.57

54 Id., pp. 383 n.12, 432. In their text, Baker & Hacker connected the discussions
of "hypotheses" with the criticism of "transcend[ing] all experience". Id., p. 432.
However, the connection is difficult to follow, for the passages from Wittgen-
stein cited for the criticism did not refer to the "hypotheses" approach. For
example, Zettel, section 260, seems clearly to refer back to the immediately
preceding sections, which do not deal with the "hypotheses" approach:

"Philosophers who think that one can as it were use thought to make an exten-
sion of experience ...."

"Generality in logic cannot be extended any further than our logical foresight
reaches... .

Zettel, sections 256, 258. The other citation, Blue Book, p. 48, discussed a par-
ticular problem in the philosophy of psychology (whether I should believe that
other persons have the same sort of feelings that I do, and whether this belief
should be characterized as a hypothesis), not a general "hypotheses" approach to
material object statements.
55 Baker and Hacker, Wittgenstein: Understanding and Meaning, p. 432. Wittgen-
stein's ideas about "hypotheses", and their similarities to some of Waismann's
writings, can be seen by looking at F. Waismann, Ludwig Wittgenstein and the
Vienna Circle (Oxford: Basil Blackwell, 1979), pp. 99-101, 158-62, 210-11 (the
book contains Waismann's transcriptions of Wittgenstein's conversations with
members of the Vienna Circle in 1929-1932); and L. Wittgenstein, Philosophical
Remarks, pp. 200-201, 282-97.
5' F. Waismann, Philosophical Papers, pp. 38-59.
57 In recent private correspondence, Gordon Baker argued that even if the
conceptual framework of "hypotheses" was not explicitly present in Waismann's
discussions of "open texture", it is likely that this conceptual framework was
presupposed by the discussions. This conclusion is based on the ability to trace
back some parts of the article 'Verifiability' to earlier writings explicitly about the

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
64 Brian Bix

IV.

Returning to Hart's adaptation of Waismann's analysis, what is dif-

ferent between the two concepts of "open texture" is the type of
unforeseeability - the type of exceptional circumstances - that
is being considered. In this, Waismann was far more extreme; he
wrote of cats growing to gigantic sizes and people disappearing
(asking how those events would affect our labelling of the objects
as "cat" and "a friend" respectively).58 When he referred to the
"unforeseen", he meant "some totally new experience such as at
present, I cannot even imagine" or "some new discovery ... which
would affect our whole interpretation of certain facts".59 Com-
pare this to Hart's legislators, who just happened to be thinking
about motor-cars when they promulgated their rule about access
to the park, but who certainly could have imagined the possibility
of roller skates, skateboards, or golf carts in the park.6?
One could connect the two conceptions by seeing Waismann's
idea of "open texture" as Hart's idea taken to its limit. Imagine an
extremely careful legislative draughtsman who spent many hours

"hypotheses" approach (see Waismann, 'Hypotheses', reprinted in F. Waismann,

Philosophical Papers, pp. 38-59, an article composed in 1936) and to Wittgen-
stein's dictation on "hypotheses" (see F. Waismann, Ludwig Wittgenstein and the
Vienna Circle, pp. 99-101, 158-62, 210-21. Letters from Gordon Baker, 27
March 1990 and 8 May 1990.
However, Dr. Baker also wrote that he now shared my doubts about the
contrast the Baker and Hacker text had drawn between Wittgenstein and
Waismann. He wrote that he now had "qualms about whether the contrast
between hypotheses and [Wittgenstein's later writings on] criteria is as sharp as I
once thought them to be". Letter from Gordon Baker, 27 March 1990.
58 Waismann, 'Verifiability', 121-22.
59 Id., 127.
60 A third analysis, which has a "family resemblance" both to Waismann's idea
of "open texture" and Hart's idea of "open texture" was Waismann's discussion
about whether the limits of certain concepts have been "anywhere determined
accurately". He wrote that for some concepts one could speak of "a nucleus of
meaning surrounded by a haze of indeterminacy". F. Waismann, The Principles of
Linguistic Philosophy, p. 222.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 65

listing dozens of objects that might be in a park and that might be

considered vehicles, and then writing detailed classificatory clauses to
clarify the original, "No vehicles in the park". At that point, the type
of "unforeseen" situations for Hart would begin to resemble those for
Waismann, and the reasons why judicial discretion could not be
avoided in applying even the most meticulously drafted statute would
begin to resemble (without actually reaching) the reasons Waismann
gave for our not being able to define an empirical term completely.

V.

Hart argued from the "irreducibly open-textured" nature of language

to the need for judges in some cases to make "a fresh choice between
open alternatives".61 Even if the conclusion (partial indeterminacy)
follows from the premise (the "open texture" of language), the basis for
that premise is not well-established in the text. "Open texture" is more
asserted than argued for. Gordon Baker claimed that Hart's argument
is circular: Waismann's notion of "open texture" derived from his
argument/assumption that a term's sense is constituted by the rules
governing its application and that no rule can be formulated in a way
such that the rule's application is never in doubt; given that indeter-
minacy of application is built into the idea of "open texture", it is not
surprising to find it as one of the idea's consequences.62 Baker went on
to note: "Although this is not generally recognized, the notion of open
texture makes sense only within a particular form of semantic theory.
... As a result it might well be impossible for Hart to incorporate it
into his philosophy of law."63 Searching for the philosophical presup-
positions and consequences of Hart's concepts becomes even more
complex if one believes, as I do, that his idea of "open texture" is in
fact substantially different from Waismann's idea.
I do not think that Hart's conclusion of partial indeterminacy - in
my view, summarized by more than based upon the idea of "open

61 Hart, The Concept ofLaw, p. 125.

62 Baker, 'Defeasibility and Meaning', in Law, Morality and Society, p. 37.
63 Id., p. 37, n. 46; see id., pp. 50-57.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 66 Brian Bix

texture" - d
how people
about creati
phy of lang
with the w
applied. Hart
must have d
be interpret
law.

In Hart's discussion of "open texture", he often seemed to refer

to words, sentences and rules interchangeably. This may reflect an
inexactness in transcribing an idea, not being sufficiently careful in
describing the idea's domain or scope. It may also reflect a tension
within Hart's concept, which arose because he was adapting an analysis
of descriptive terms (Waismann's "open texture") to an analysis of rules
- and not just an analysis of rules as such, but analysis of the applica-
tion of rules by judges in modern legal systems.
Waismann was writing about language in general; Hart was writing
about language in the context of law - in particular, in the context of
applying and interpreting rules - and the problems to which his ideas
responded derive from that context. If while we are walking through
the park, and my friend talks about "that vehicle" while pointing
towards a toy car or a skateboard, I may find her usage strange or
quaint, but I understand what she said; I can understand the use of the
term "vehicle" to refer tb an object to which that label is not usually
applied. Because the extension (from the usual usage of the term) is
not radical or bizarre, I do not react by correcting my friend, as I
might if she had used the term "vehicle" while trying to refer to a
banana or a book. A certain tolerance or laxness in the application of
terms is beneficial in normal conversation; all is well as long as I think
I understand roughly what my friend was trying to say and she thinks
I understood roughly what she was trying to say.65

64 Cf. Baker, 'Defeasibility and Meaning', p. 37.

65 See e.g., Davidson, 'On the Very Idea of a Conceptual Scheme', reprinted in
D. Davidson, Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984),
p. 196.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 67

The situation with commands, instructions, suggestions, and so on,

is different. With such uses, because the focus is on the guidance of
behaviour and because such sentences are often meant to be applied
("followed") on an indefinite number of occasions, the exact scope of
the rule's application - determined at least in part by the exact scope
of the rule's terms - is important. As Gerald Graff, in a slightly
different context, wrote: ". . . the practical concerns of the law occasion
the imposition of a number of artificial restrictions on interpretive
procedure, restrictions that do not apply outside the legal context.
These restrictions arise from practical, ethical considerations rather
than epistemological ones...."6 When language is used to guide and
coordinate behaviour, the problems of interpretation and meaning will
necessarily be different from those that accompany language qua
method of expressing one's thoughts and method of communication
between persons.
Properly seen, Hart's approach was not based on a theory of lan-
guage, at least not if that is defined as a theory about the meaning of
particular terms. While Hart at times seemed to argue that something
about language makes it inevitable that judges will have discretion, he
seemed at other times to concede that judges could interpret rules in
such a way that they would not have discretion. However, Hart
argued, any such attempt to get rid of judicial discretion would have
negative consequences (for example, an inflexibility, an inability to
recharacterize the rules to meet changing circumstances).67 Hart con-
sidered and rejected a way of clarifying the meaning of terms within
rules based simply on language: attaching necessary and sufficient
conditions which an object or event must satisfy if it is to be sub-
sumed under that term.68 The problems of focusing on particular
terms as a way of understanding daily communication, let alone legal
rule?, was illustrated by Lon Fuller's "improvement" example. Fuller
showed that the meaning of "improvement" in the sentence fragment

"6 Graff, "'Keep off the Grass", "Drop Dead", and Other Indeterminacies: A
Response to Sanford Levinson', Texas Law Review 60 (1982): 405, 411 (emphasis
omitted).
67 Hart, The Concept ofLaw, pp. 126-27.
(8 Ibid.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
68 Brian Bix

"all improvements must be promptly reported to ..." cannot be

understood outside the sentence's context (not only what the require-
ment relates to, but also who gave the order to whom, and what set of
practices surround the situation).9
At this point, it is helpful to introduce a distinction between what
meaning a writer (or speaker) tried to convey by his or her words, and
what the words, considered by themselves, actually mean ("the well-
known distinction between what a speaker means and what his words
mean"70). This distinction appears in various forms throughout the
philosophical literature. For example, the nineteenth century herme-
neutic theorist Friedrich Schleiermacher wrote that an act of speaking
had to be understood in two separate ways: in its relation to the
language (that is, the meaning of the words spoken) and as an expres-
sion of the speaker's thoughts.7' The distinction is also illustrated by
the fact that we can understand malapropisms and code-name refer-
ences.72 The distinction appears as well, explicitly and implicitly, in
many of areas of law.
The distinction can be used to clarify even some of the more

69 Fuller, 'Positivism and Fidelity of Law', Harvard Law Review 71 (1958): 630,
664-67.

70 Dummett, 'A Nice Derangement of Epitaphs: Some Comments on

and Hacking', in E. LePore, ed., Truth and Interpretation (Oxford: Basi
1986), p. 460.
71 Schleiermacher, 'General Hermeneutics', reprinted in K. Mueller-Vollmer, ed.,
The Hermeneutics Reader (Oxford: Basil Blackwell, 1986), p. 75.
72 See Davidson, 'A Nice Derangement of Epitaphs', in E. LePore, ed., Truth and
Interpretation (Oxford: Basil Blackwell, 1986), p. 433. As an example of code
names, in American political rhetoric "States' Rights" has been used as a racist
code word for racial segregation, and "cosmopolitan" in Eastern European
rhetoric was (and still is) an anti-Semitic code-word for Jewish. An experienced
observer hearing a politician from the American South refer to "States' Rights"
would know that the politician meant to refer to racial segregation, and that the
speaker knew that the relevant audience would understand the phrase that way,
even though this was not the phrase's literal meaning.
73 See e.g., J. W. Harris, Law and Legal Science (Oxford: Clarendon Press, 1979),
pp. 132-43 (the "will model" and the "natural meaning model" as alternative
"models of rationality" for justifying legal decisions).

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 69

obscure arguments in legal academic writing. For example, J. M. Balkin

offered a "deconstruction" of the idea that a text has a clearly definable
"core" of meaning independent of context, and only the text's "periph-
eral" meanings are affected by context. Balkin wrote: "If two parties
have adopted a code for contracts involving livestock where 'cow'
means 'horse,' the core meaning of 'cow' will shift radically.... [I]t is
the 'normal' context in which we use the word 'cow' that gives us its
'core' meaning."74 Balkin concluded that "core" meanings, like "periph-
eral" meanings, are context dependent. The problem with this argu-
ment is soon clear. Balkin's contracting partners are not using the
English word "cow" in an "abnormal context"; in a sense, they are not
using the English word "cow" at all. They could be described either as
using the word "cow" incorrectly, or of using a language of their own,
where the word "cow" has different rules for usage and different
applications than it does in English.75 In terms of the discussion above,
we could distinguish the fact that (in English) the word "cow" means
cow from the fact that these parties used the word to mean horse. If
the point of Balkin's "deconstruction" is only that the meaning (the
acontextual meaning, the "core" meaning) of a word can change as we
move from a common language to an idiolect (e.g., from English to
these parties' code) or from one common language to another (e.g.,
from English to French), then his point is correct, but not interesting.

VI.

Hart's discussion in The Concept of Law in the section on "open

texture" seemed to rest halfway between emphasizing speakers'
meaning and emphasizing words' meaning, and halfway between a
theory of meaning and a theory of statutory interpretation. The

74 Balkin, 'Deconstructive Practice and Legal Theory', Yale Law Journal 96

(1987): 743, 780 n. 106.
75 See Baker and Hacker, Wittgenstein: Rules, Grammar and Necessity, p. 332: "If one
follows deviant grammatical rules it does not mean that one is saying something
wrong .... Rather ... one [is] speaking of something else (which one may have
to explain), just as if one follows rules other than those of chess one is playing
another game" (citation omitted).

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
70 Brian Bix

approach focused on what the speaker meant (and resembled a theory

of statutory interpretation) in that Hart seemed to want the judge to
focus on the problem the rulemakers had in mind. However, Hart did
not have the judge try to discover the rulemakers' aim from legislative
(or judicial) records, but rather from the rule's words alone: rule-
makers' aims as equated with the "clear examples" that fall under "the
language used in this context".76 Because the rulemakers formulated
the rule the way they did, Hart implied, they must have had those
cases (the "clear examples") in mind, "and [their] aim in legislating is
so far determined because [they] have made a certain choice".77
For Hart, the rulemakers' aim is embedded in the language, and
where the application of the words to a particular case is no longer
clear, "[the rulemakers'] aim is, in this direction, indeterminate".78 If
Hart had been concerned only with what the rulemakers meant, or
only with implementing their intentions, he would not have us reach
that conclusion so quickly. He would have had us try to discover
whether, the words the rulemaker chose notwithstanding, their aim
might still have been determinable and determinate. He would have
advised us to look through relevant records or to ask counterfactual
questions (e.g., "even if the rulemakers had not considered the question
of skateboards in the park, what would they have said had they
thought about it?") to try to discover the rulemakers' aim. However, he
did not do so.
On the other hand, Hart's approach was also not dependent solely
on what the words meant. In fact, for Hart decisions can sometimes
actually contribute to meaning: "[w]hen the unenvisaged case does
arise" and we make our decision ("by choosing between the competing
interests in the way which best satisfies us"), we will have "incidentally
. . settled a question as to the meaning, for the purpose of this rule,
of a general word".7)

76 Hart, The Concept of Law, p. 125.

77 Ibid.
78 Id., p. 126.
79 Hart, The Concept of Law, p. 126. Cf. L. Wittgenstein, Zettel, section 120: "'This
law was not given with such cases in views.' Does that mean it is senseless?"

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
H. L. A. Hart and the "Open Texture" ofLanguage 71

One matter which may explain some of the more paradoxical and
counterintuitive ideas in Hart's discussion here was his use of "we"
throughout the relevant section: "we ... frame some general rule of
conduct", "our aim... is so far determinate", "[w]e shall have rendered
more determinate our aim", and so on.80 Hart's analysis of rule-appli-
cation appeared to occur in a hypothetical context where the same
person, group, or institution which had created the rule had the
responsibility of applying it and modifying it. Such a situation is, first,
far different from the most common situation in most legal systems,
and second, far less troubling. The problems legal theorists face in the
area of rule-application come largely from the fact that those who
apply rules usually are applying rules written by someone else; in the
American context, this means judges applying statutes and constitu-
tional provisions or judges applying rules set down by other judges.
Within such a context, one's view about how rules should be applied
(interpreted, modified, supplemented) will depend on one's theory
about the proper role of various institutions and the relationships
among them. For example, those who see judges' role as merely imple-
menting the will of the legislature,81 and who believe that any judicial
action that cannot be so characterized is illegitimate, would probably
recommend an approach different from Hart's approach for facing an
"unenvisaged case". (They would probably want judges only to make
decisions that could reasonably be characterized as implementing the
legislature's aims or values.)
Thirty years after he wrote it, Hart's short discussion of "open
texture" and legal determinacy remains a rich and provocative text.
However, to gain the most from Hart's discussion, we must disentangle
the various strands: conclusions based on the nature of language, con-
clusions based on the nature of rules and rule-application, general
recommendations for how rules could best be applied, and recommen-

8( Hart, The Concept of Law, pp. 125-26 (emphasis added).

81 See e.g., R. Posner, Law and Literature (Cambridge: Harvard University Press,
1988), pp. 220-47.

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms
 72 Brian Bix

dations for r
situations or p

Balliol College,
Oxford OX1 3BJ,
England

This content downloaded from

3.6.255.29 on Wed, 19 Feb 2025 03:27:59 UTC
All use subject to https://about.jstor.org/terms