In Defense of Hart’s Supposedly Refuted Theory of Rules

• Published date: 01 December 2021
• Author: Jeffrey Kaplan
• Date: 01 December 2021
• DOI: http://doi.org/10.1111/raju.12331

© 2022 University of Bologna and John Wiley & Sons Ltd.

Ratio Juris. Vol. 34 No. 4 December 2021 (331–355)

In Defense of Hart’s Supposedly

Refuted Theory of Rules

JEFFREY KAPLAN

Abstract. H. L. A. Hart’s “practice theory” of rules is widely thought to face two problems: (1) It

fails to account for the normativity of law and (2) it is susceptible to a decisive counterexample

dating back to Warnock (1971). In this paper, I offer solutions to both problems. In response to

(2), I appeal to an old, but underappreciated, distinction made by Rawls (1955) and argue that

the counterexample is no counterexample at all. In response to (1), I apply a newly popular

distinction regarding the nature of law’s normativity and argue that Hart’s practice theory has

no problem accounting for the normativity of law.

1. Introduction

Where do laws come from? They come from legislative bodies— like congress, or

parliament, or the monarch— with lawmaking authority.1 Where, then, do these leg-

islative bodies get their authority? They get their authority from other laws— like

those contained in a constitution— which specify which bodies can legislate and

under what conditions. Since these second- order laws are themselves created by leg-

islative bodies, the threat of regress is not difficult to see.2

On H. L. A. Hart’s theory of law, the regress halts at one particular law: the rule

of recognition, a second- order rule specifying the conditions under which other rules

and rule- creating bodies have legal status (Hart 1961). But where does the rule of rec-

ognition come from? How does it get its legal status? The answer, on pain of regress,

must not refer to some other valid law or some other already authoritative body.

Instead, Hart suggests that the rule of recognition arises from nonlegal— and indeed

non- normative— states of affairs.

Hart’s explanation for how the rule of recognition arises from nonlegal states of

affairs later came to be called the practice theory of rules.3 This theory attempts to

1 Laws can be understood broadly so as to include regulations, executive orders, and even judi-

cial precedent. Correspondingly, legislators can be understood to include members of the exec-

utive and judiciary in their regulation- and precedent- creating modes.

2 See Green 1999 and Shapiro 2011 for good discussions of jurisprudence framed around this

regress. A prominent alternative response to this regress is the coordination convention theory

of law, which was developed in the early 1980s by Coleman (1982) and Postema (1982) before

somewhat falling out of favor.

3 For the original use of “the practice theory of rules” see Raz 1984.

Jeffrey Kaplan

332

Ratio Juris, Vol. 34, No. 4© 2022 University of Bologna and John Wiley & Sons Ltd.

explain how a rule of recognition comes to exist, and it does so without reference to

any prior legal entities. But the practice theory is even more ambitious than that. It

purports to explain the existence not only of legal rules but all sorts of social— i.e.,

human- created— rules, such as the rules of etiquette, games, clubs, or fashion. That

is, the practice theory is not a theory of the legal rule of recognition in particular. It

does not present conditions that are sufficient for there to be a rule of recognition.

Rather, it is a theory of just that feature of the rule of recognition that allows it to halt

the regress— its being a social rule.

Unfortunately, the practice theory is widely— though not universally— taken to

be a failure (Kramer 1999, 251– 3). This is for two reasons. First, the things to be ex-

plained, which in the legal case are the validity of laws and the authority of legisla-

tors, are normative phenomena. Hart almost never uses the word “normative” in The

Concept of Law. But in the sixty years that have passed since that book was first pub-

lished, talk of “normativity” has exploded within analytic jurisprudence to the point

where “the normativity of law” is considered a central explanandum of legal philos-

ophy (Coleman 2001; Enoch 2011; Green 1999; Marmor and Sarch 2019; Postema

1982). The problem is that the practice theory attempts to explain the foundation of

legal systems merely by appeal to two types of descriptive facts: behavioral facts about

what people do and psychological facts about what attitudes people take. The worry

is that these descriptive facts could never explain the fact that one legally ought to do

something.4 (Though this problem will feel familiar to many philosophers of law,

some will immediately worry that mention of the normativity of law— and mention of

the supposed fact that one “legally ought” do something— is obscure. This worry

puts us on the right track, as the exact sense in which law is normative will determine

whether this is a genuine problem for the practice theory.)5

The second problem is more straightforward. The practice theory provides two

conditions that are putatively sufficient for the existence of a social rule. The problem

is that there is a counterexample, introduced by G. J. Warnock and reformulated and

repeated dozens of times since, in which these two conditions are met but where no

rule exists (see Warnock 1971, 45– 6, 61– 5; cf. Marmor 2001, 3; 2009, 14– 5; Perry 2015;

Shapiro 2011, 103– 4). The vast majority of philosophers who discuss this problem in

print accept it as a decisive refutation of the practice theory.6

Hart’s theory of law is based on a theory of rules that is thought to be, at mini-

mum, troubled, if not altogether doomed. In this paper, I defend the practice theory

by offering a solution to the normativity problem and by arguing that the putative

counterexample is no counterexample at all. Both of these points draw on existing

literature. The original contributions of this paper consist in (a) the application of

existing insights to an old— and some think settled— discussion, but also, and more

interestingly, (b) several original arguments applying existing insights to the sup-

posed problems for the practice theory.

4 See Bix 2006; Coleman and Leiter 1996, 241; Dworkin 1977, 19; Enoch 2011; Green 1999, 35;

Holton 1998; MacCormick 1978 and 1981; Marmor and Sarch 2019; Perry 2006, 1173; Raz 1975;

Shapiro 2011, 46– 9; Smith 1994.

5 This is the topic of Sections 6 through 14.

6 The two exceptions of which I am aware are discussed in Section 4. They are Green 1999 and

Kramer 1999.