The Concept of Judicial Discretion

  • Marisa Iglesias Vila
Part of the Law and Philosophy Library book series (LAPS, volume 49)


The debate on judicial discretion and its limits gained considerable attention from legal theorists in the 1930’s. One of the reasons accounting for this increase in interest was the loss of influence of certain maxims of theoretical positivism as a way of describing the structural characteristics of legal orders1. Some currents of thought such as American legal realism challenged completeness and the formalist perception of the role of judges as simply implementing the will of the legislation2. Extreme realism, which presents legal theory as a mere psychological theory of adjudication, did not have a lasting influence. However, it did trigger an interest to analyse judicial decision-making and its connection to normative production. Particularly in the 1970’s, intense debate arose about the problem of adjudication of norms. Controversy focused on two main issues: the technical necessity for judicial discretion on the one hand, and the justification of discretionary decisions on the other.


Legal Order Judicial Decision Legal Positivism Discretionary Decision Semantic Problem 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    See Bobbio (1995, 37–66) for a characterisation of theoretical positivism.Google Scholar
  2. 2.
    See Frank (1930; 1949) and Llewelyn (1951).Google Scholar
  3. 3.
    See Hart (1994, Ch. VI) and Raz (1982, Ch. Ill).Google Scholar
  4. 4 S.
    ee Dworkin (1977, 31–39, 68–71) and Sartorius (1971, 151–160).Google Scholar
  5. 5.
    See Greenawalt (1975, 392–393; 1977, 1004–1007), Dworkin (1977, 84–86), Barak 1989, 192–203) and Sartorius (1971, 158–159).Google Scholar
  6. 6.
    See Aamio (1987, 3–8), Dworkin (1977, 84–86), Barak (1989, 167–171, 203–206) and Burrows (1976, 1518).Google Scholar
  7. 7 S.
    ee Isaacs (1933, 344–352), Galligan (1986, 4–14), Traynor (1978, 1025–1040) and Barak (1989, 235–260).Google Scholar
  8. 8.
    See Wellman (1988, 153–165), Davis (1969, 216–217), Aarnio (1991, 31–35) and Kadish and Kadish (1973, 40–45).Google Scholar
  9. 9.
    For a detailed account of Dworkin’s interpretive theory, see Dworkin (1985, 1986).Google Scholar
  10. 10.
    See Hawkins (1992, 32–35) and MacCormick (1978, 242).Google Scholar
  11. 11.
    See Barak (1989, 3–5).Google Scholar
  12. 12.
    With regard to administrative discretion, see Davis (1969) and Galligan (1986).Google Scholar
  13. 13.
    Isaacs (1933, 339, 341–342).Google Scholar
  14. 14.
    Here ‘responsible’ is meant in a laudatory sense. Lucas states that an individual is responsible in this sense when he or she acts reliably. For further details see Lucas (1993, 11–13).Google Scholar
  15. 15.
    Dworkin (1977, 31–32). Pattenden’s ‘concealed discretion’ refers to this idea. See Pattenden (1990, 2). See also Greenawalt (1975, 365–366), Galligan (1986, 14), Skubik (1990, 26–27) and Sartorius (1971, 152).Google Scholar
  16. 16.
    This is one of the reasons that account for the confusion in the controversy over the relationship between weak and strong discretion, particularly between Dworkin and authors such as Greenawalt and Galligan. There seems to be some implicit conceptual disagreement about the meaning of weak discretion. As we shall later see, most authors agree on the above characterisation of weak discretion, but they emphasise different aspects of this concept: some highlight the linguistic indeterminacy of language, whereas others stress the existence of a right answer that is not obvious to the interpreter.Google Scholar
  17. 17.
    See Barak (1989, 13–14), Isaacs (1933, 339) and Pattenden (1990,3–4). This meaning of discretion also appears in the works of some classics of the analysis of the judicial process. See Cardozo (1921, 100) and Frank (1949, 57–58).Google Scholar
  18. 18.
    Dworkin uses discretion in a further weak sense to speak of the irreversibility and finality of certain judicial decisions. See Dworkin (1977, 32). See also Isaacs (1933, 339, 341–342), Pattenden (1990, 2), Greenawalt (1975, 365 and ff.), Galligan (1986, 14), Shiner (1992b, 32), Guest (1992, 214–215, 220–221) and Rosenberg (1971, 637).Google Scholar
  19. 19.
    See Pattenden (1990, 2). Rosenberg (1971, 637) refers to this meaning of discretion as ‘secondary discretion’.Google Scholar
  20. Le.
    gal realism takes this meaning of judicial discretion as its starting point to reach the conclusion that it is the judiciary who determines the content of law. Both American and Scandinavian realism have been interested in a factual approach to the legal phenomenon, that is to say, in law as it is applied in reality. Hence, they tend to argue that legal norms exist only in so far as judges to settle conflicts use them. Judicial decisions that are not subject to review become the very contents of law. About this point see Hart’s remarks (1994, 141–142) on the difference between the ‘finality’ and the ‘infallibility’ of a judicial decision and his criticism to scepticism.Google Scholar
  21. 20.
    See Dworkin (1977, 32–33). For a detailed analysis of Dworkin’s meaning of strong discretion, see Waluchow (1983).Google Scholar
  22. 21.
    The distinction between valid and admissible courses of action will be established in the next section, which is devoted to strong discretion. This definition of strong discretion is generally accepted although sometimes with slight differences in meaning, or even a different terminology. See Hart and Sacks (1958, 126), Barak (1989, 7), Greenawalt (1975, 363–366), Galligan (1986, 6–8, 21–22), Davis (1969, 4), Burrows (1976, 1), Raz (1972, 843), Rosenberg (1971, 637), Bell (1992, 92) and Shiner (1992b, 33).Google Scholar
  23. 22.
    This is why Dworkin defines strong discretion as that action which is not subject to standards set by the legal authority. See Dworkin (1977, 32–33) and Galligan (1986, 6, 8, 22).Google Scholar
  24. 23.
    This idea is very clearly outlined by Raz (1972, 847). See also Hart (1980, 9), Galligan (1986, 7–8) and Barak (1989, 24–27).Google Scholar
  25. Th.
    e demand for general justification of a discretionary decision allows for distinction between strong discretion on the one hand, and a further sense of discretion that links this concept to arbitrariness, on the other. This point will be discussed in detail later on, when we define and analyse arbitrariness and its relationship to discretion. It is worth noting that some authors consider strong discretion to be synonymous to arbitrariness. See, for instance, Hayek (1960, 228, 453) and Dicey (1945, 188).Google Scholar
  26. 24.
    For a detailed distinction between explicit and implicit discretion, see Luzzati (1990a, 130–137). For Kelsen’s distinction between intentional and non-intentional indeterminacy in the application of law, see Kelsen (1970, 349–351).Google Scholar
  27. 25.
    See Hart (1994, 141–147), Aamio (1987, 217). Every judicial decision establishes rights and duties, but only in the sense that it creates an individual norm for the instant case. See also Dworkin (1977, 31–33), Isaacs (1933, 339, 341–342), Greenawalt (1975, 365–367), Galligan (1986, 14) and Sartorius (1971, 152).Google Scholar
  28. 26.
    ’Application of law’ is the use of legal norms to justify judicial decisions. For the time being, I will not use this term to refer to the meaning of ‘application’ proposed by Ross in connection with validity. As Bulygin points out, Ross considers that a norm is being applied when there is a certain normative ideology, a certain intent that induces the judge to follow a rule. External conduct is not sufficient to determine whether the judge is using a normative standard. The judge should feel internally obliged to follow that pattern of conduct. Bulygin claims that Ross uses both meanings of ‘application’ without realising the differences between them. Usually, conceptual proposals for strong discretion understand ‘application of law’ in the first meaning set out above, that is to say, they only take into account the external conduct of judges when they use norms to back up their resolutions. On this point, see Bulygin (1991a, 346–348). For Ross’ different meanings of’application’, see Ross (1958, 37–38).Google Scholar
  29. 27.
    See Greenawalt (1975, 365). See also Galligan (1986, 2–3).Google Scholar
  30. 28.
    Dworkin (1977, 31). For a similar example, see Bell (1992, 93–94).Google Scholar
  31. 29.
    A clear account of the three duties of the judicial role is given in Alchourrón and Bulygin (1971, 151153) and Mendonca (1997, 49–61). They distinguish among the duty to dispose of every dispute, the duty to provide justification to every resolution, and the duty to apply the law.Google Scholar
  32. 30.
    See, for instance, Hart (1994, 127), Barak (1989, 8–9) and Rosenberg (1971, 636–637).Google Scholar
  33. 31.
    The concept of choice has been widely developed in Economic Theory, more specifically within the analysis of rational action. The defining features of choice are: voluntariness, preferences, different real possible courses of action, and mutually exclusive options. See, for instance, Afriat (1987, 14–15), and Hollis and Sudgen (1993, 2–7).Google Scholar
  34. 32.
    About this point, see Galligan (1986, 8, 21) and Bell (1992, 92–94).Google Scholar
  35. 33.
    Dworkin claims that in cases of strong discretion the decision-making agent is not subject to standards set by the legal authority. See Dworkin (1977, 33).Google Scholar
  36. 34.
    For a discussion on the importance that such a judicial duty should exist, see Shapiro (1985, 577–578).Google Scholar
  37. 35.
    An overview of the concept of right answer can be found in Aarnio (1987, 158–161). The absence of right answer is usually encountered when a) there exist several legal solutions for a case and the judge must decide in favour of one of them by means of extra-legal criteria, or b) there exists no legal solution for the case and the judge makes her choice among different extra-legal solutions.Google Scholar
  38. 36.
    See Alchourrón and Bulygin (1971, 154–157).Google Scholar
  39. 37.
    On this maxim, see Hoffsmaster (1982, 27–28). A detailed analysis can be found in von Wright (1970, 123–132).Google Scholar
  40. 38.
    On the concept of false presupposition, see Ross (1958, 133–134).Google Scholar
  41. 39.
    This judicial duty is acknowledged in most legal orders. See, for instance, article 120.3 of the Spanish Constitution.Google Scholar
  42. 40.
    Along the lines of the conceptual proposals that address this question in the context of strong discretion, ‘permission’ and ‘power’ are assumed to be synonymous. But their equivalence is much debated. Thus, an individual is empowered to do x when the relevant authority explicitly neither forbids her to perform x nor forbids her not to do x. See von Wright (1970, 100–107).Google Scholar
  43. 41.
    See, for instance, Hart and Sacks (1958, 126), Barak (1989, 7), Bell (1992, 92), Lemper (1992, 185) and Greenawalt (1975, 366).Google Scholar
  44. 42.
    Alchourrón and Bulygin note that in these instances the judge has no specific obligation to condemn the defendant or to dismiss the claims. The general obligation to render a decision is what actually prevails, and the judge complies with this obligation both in the event of a conviction as in that of an acquittal. InGoogle Scholar
  45. 45.
    These situations are referred to as ‘easy cases’ or ‘clear cases’ by legal theorists because, as pointed out by MacCormick (1978, 197), in such cases ’justification of decisions can be achieved by simple deduction from clear established rules’. See also Aamio (1987, 1–3). We shall discuss in detail the distinction between ‘easy cases’ and ’hard cases’ in Chapter 2.Google Scholar
  46. 46.
    See Galligan (1986, 1).Google Scholar
  47. 47.
    For this criticism, see Greenawalt (1975, 365, 366), Galligan (1986, 19–20) and Waluchow (1983, 330, 331; 1994, 190–207).Google Scholar
  48. 48.
    Dworkin’s rejection of semantic problems as instances of no right answer, and hence, as instances of strong discretion, is to be found throughout his work. See, in particular, his references to the problem of the ‘semantic sting’. Dworkin (1986, 45, 46; 1987, 14–17; 1977, 33–38, 68–71).Google Scholar
  49. 49.
    See article 566 of the Spanish Rules of Criminal Procedure.Google Scholar
  50. 50.
    As pointed out above, Luzzati refers to this kind of power that judges may be granted by law as ‘explicit discretion’. It was also mentioned that the present book would be focusing on implicit discretion and its connection with legal knowledge. See Luzzati (1990a, 130–137). It may be of interest to note that, in cases of no right answer, some legal orders explicitly empower the judges to act as a legislator. See article 1 of the Swiss Civil Code. This criterion completes the legal system by means of authoritative decisions that solve any particular conflict. Hence, the existence of a right legal answer can be ascertained for every individual case. Its only distinguishing feature would be that the judge is empowered to determine what it is. For this reason, it seems inappropriate to include this explicit power as an instance of strong discretion.Google Scholar
  51. 51.
    Cases for which there is an absence of right answer are often considered to be instances of ‘hard cases’. This common term in the theory of adjudication has been defined in several ways. Later on I will discuss in detail the relationship between no right answer, hard cases, and strong discretion with a view to determining whether they can be considered coextensive.Google Scholar
  52. 52.
    ’Indeterminacy of law’ is the lack of deontic qualification of a certain conduct, that is, the absence of a right legal answer. Navarro observes that it is important to distinguish between legal indeterminacy in generic cases, and in individual ones. A generic case is a combination of properties, whereas an individual case is a particular instance of a generic case. If a legal system has a closure mechanism, it guarantees a right answer to every particular case. However, this does not ensure avoidance of legal indeterminacy with respect to generic cases. On this distinction see Navarro (1993, 249–252, 262–266).Google Scholar
  53. 53.
    In this context, I am referring to linguistic indeterminacy. Adjudication can involve other forms of legal indeterminacy such as indeterminacy of purpose, or what Kelsen called ‘intentional indefiniteness’. Intentional indefiniteness is met in those instances where the appropriate authority has empowered the judiciary to determine the standards that guide its decisions in a particular subject matter. Luzzati claims that this is the only sense of indeterminacy in Kelsen’s theory of law. See Kelsen (1970, 349–351) and Luzzati (1990a, 125–137). On the different sources of legal indeterminacy, see also Coleman and Leiter (1995, 212–219).Google Scholar
  54. 54.
    In this work, the problem of vagueness will be examined in the context of the theory of meaning as use. This is the most common semantic setting of positivist theories based on Wittgenstein’s theory of meaning. If we accept that the meaning of a term or expression depends on the usage within the community of speakers, vagueness occurs when there is no well-established convention on whether a case is embraced by a certain concept or not. See, for instance, Luzzati (1990b, 12–13, 34–39).Google Scholar
  55. 55.
    Note that vagueness arises only when a concept is used to refer to some aspect of reality. Thus, concepts as such are not vague. They become vague when the conceptual system is measured against reality. The use of the term ‘reasonable’ raises no difficulty in a wide variety of cases where it functions in a precise way. The meaning of this term is vague only in those cases in which the conditions for its use do not include in or exclude from the concept the case under consideration. Every concept has a zone of certainty and a zone of penumbra. The zone of certainty covers all those cases to which the term clearly applies, as well as all those cases to which it clearly does not apply. In cases in the penumbral zone it is no longer clear whether the term applies or not. These are borderline cases. See Hart (1994, 122–127; 1958, 607–615).Google Scholar
  56. 56.
    With regard to these cases of vagueness, see, for instance, Carrió (1965, 31–35).Google Scholar
  57. 57.
    See Hart (1994, 123, 127–129). For Waissman’s characterisation of the open texture of language, see Waismann (1951, 119–120). On open texture, see also MacCormick (1991, 72–74), Bix (1991, 51–72) and Carrió (1965, 35–36).Google Scholar
  58. 58.
    The well-known distinction between the ‘intension’ and the ‘extension’ of a term was first proposed by Carnap. Briefly, when considering a predicate, its extension is the class of objects to which it applies, whereas its intension is the property it expresses. See Carnap (1947, 18–23).Google Scholar
  59. 59.
    A useful account of homonimy and how it differs from polysemy can be found in Lyons (1981). On the ambiguity of the legal context, and the difference between semantic and syntactic ambiguity, see Aarnio (1987, 111–112), Ross (1958, 115–125) and Carrió (1965, 28–31).Google Scholar
  60. 60.
    Raz (1972, 846) also points out that appealing to principles does not eliminate the need for strong discretion. Principles are expressed by means of language, and their propositional contents may, too, be vague.Google Scholar
  61. 61.
    Even though both rules and principles are often considered to be correlations between generic cases and normative solutions, on some occasions a structural difference between the two is said to exist. It has been argued that rules are applicable in an all-or-nothing fashion whereas principles have a dimension of weight: they are reasons guiding conduct, but they do not establish the conditions for their own applicability. This is the distinction Dworkin (1977, 22–28) and Alexy (1993, 86) appear to support. Some authors, such as Atienza and Ruiz Manero (1991, 108–109) have taken this logical distinction further by saying that rules shape a case in a closed manner, whereas principles do it in an open manner. In this work, no special attention will be paid to the structural aspect of this distinction. It is not clear to me how a rule that is expressed linguistically, can conform a case in a closed manner, i.e. how a complete list of conditions of applicability can be drawn. Perhaps, as suggested by Laporta (1985, 5) and Aamio (1987, 96–97), rather than setting forth a structural distinction between these two types of norms, it is more appropriate to establish a gradual one that respects the higher degree of generality and abstraction usually displayed by principles. It could also be said that principles are norms of a very fundamental nature and, as such, they reflect the basic guidelines and the core decisions that govern a normative system and justify its existence. See Prieto Sanchis (1992, 58–59). It seems to be the case that no criterion allows drawing a clear distinction between rules and principles. We probably do not even need such a criterion. It suffices to show that certain normative standards are more fundamental and abstract than others; that the more fundamental ones serve as criteria to justify the existence of other norms; and that they are used to interpret them. Whether these remarks reproduce the distinction between principles and norms may be viewed as a mere verbal issue.Google Scholar
  62. 62.
    Hart (1994, 126) claims that rules of interpretation may reduce the need for judicial discretion, but they cannot eliminate it since they, themselves, may be stated vaguely and require interpretation.Google Scholar
  63. 63.
    If we adopt intentionalist semantics, accepting ambiguity and the existence of alternative actions implies admitting that the speaker has no specific intent when uttering a statement. If, however, we accept that an intent underlies every utterance, then ambiguity is an epistemic difficulty arising from incomplete information about the mental state of the speaker, and can therefore not be considered a source of strong discretion. As it is being described, discretion is dependent on the absence of a single right answer. If, however, intentionalist theory of meaning is replaced by conventionalist semantics, irresolvable ambiguity always generates semantic indeterminacy.Google Scholar
  64. 64.
    A legal proposition is the meaning of a normative statement, i.e. a statement describing the existence of a legal norm. On different approaches to the relationship between semantic problems and legal indeterminacy see Dworkin (1985, 119–145).Google Scholar
  65. 65.
    See Raz (1979, 70–74).Google Scholar
  66. 66.
    This idea can be easily grasped from the following passage by Hart (1994, 127): `Here something in the nature of a crisis in communication is precipitated: there are reasons both for and against our use of a general term, and no firm convention or general agreement dictates its use, or, on the other hand, its rejection by the person concerned to classify. If in such cases doubts are to be resolved, something in the nature of a choice between open alternatives must made by whoever is to resolve them’.Google Scholar
  67. 67.
    On this particular example and the different versions of the thesis on the absence of a right answer, see Dworkin (1985, 119–137).Google Scholar
  68. 68.
    From this point of view, vagueness fits Raz’s definition of legal gap. Legal gaps result from those situations in which some legal questions subject to jurisdiction have no complete answer. See Raz (1979, 70–71).Google Scholar
  69. 69.
    See Coleman and Leiter (1995, 216). Note that if it is assumed that semantic problems cause an absence of right answer, discretion may not be defined as the possibility for choice among different equally valid alternatives. ‘Valid’ here means ‘belonging to the legal system’. We are thus urged to determine whether strong discretion can be characterised in terms of validity, which would limit its scope to cases of normative contradiction, or whether it is more interesting to use looser standards such as acceptability.Google Scholar
  70. 70.
    Dworkin (1977, 32–33, 70).Google Scholar
  71. 71.
    Dworkin (1977, 33).Google Scholar
  72. 72.
    Dworkin (1977, 32) argues that: “if the sergeant is told to take the five most experienced men, he does not has discretion in this strong sense because that order purports to govern his decision. The boxing referee who must decide which fighter has been the more aggressive does not have discretion, in the strong sense, for the same reason’.Google Scholar
  73. 73.
    On this point, see Dworkin (1977, 68–71).Google Scholar
  74. 74.
    In order to exemplify the kind of judicial activity required in the context of weak discretion, Dworkin makes up the figure of Hercules. Hercules is an ideal judge with outstanding abilities, no time constraints, and a sound knowledge of law and of the justifying role of legal principles. Hercules has an extremely well-developed conception of law that allows him to provide the right answer even in those cases in which there are important epistemic difficulties to grasp the solution prescribed by the law. Dworkin maintains, in accordance with his own epistemic approach to law, that Hercules’ answer is interpretive, i.e. it is theory-dependent. Therefore, the fact that Hercules is able to find the right answer, Dworkin claims, does not mean he has the ability to contrast legal propositions with reality independently of his own conceptual schemes. It is particularly important to keep the latter consideration in mind to avoid some of the most common misunderstandings of Dworkin’s theory. On the figure of Hercules, see Dworkin (1977, 105–130; 1986, 239–275).Google Scholar
  75. 75.
    This point will be relevant to later chapters. It is worth noting that Dworkin does not claim that disagreement about the meaning of words implies an absence of right answer. Neither does disagreement on the sources of law imply such an absence. In fact, interpretive disagreements lie at the very heart of his view of law as a dynamical and interpretive concept.Google Scholar
  76. 76.
    See Waluchow (1983, 321–331).Google Scholar
  77. 77.
    Waluchow claims that making strong discretion dependent on the existence of standards that pretend to control a decision is equivalent to expecting the interpreter to act as if there exist a single right answer. Moreover, it would entail obliging her to accept that there is, indeed, such an answer, and urging her to find this answer, in spite of her conviction that such an answer does not exist. See Waluchow (1983, 326). It should be borne in mind that this criticism of Dworkin’s characterisation is confined to a particular interpretation of Dworkin’s theory and of his thesis on the right answer. In chapters to come, the context of Dworkin’s discussion of the problem of strong discretion and that of the right answer will be specified.Google Scholar
  78. 78.
    See Waluchow (1983, 325–328). On this point, see also Galligan (1986, 20).Google Scholar
  79. 79.
    On this point, see Galligan (1986, 17). The claim that legal propositions are either true or false has recently been replied. Dworkin argues that linguistic indeterminacy can be avoided by making use of the legislative principle that states that all legal propositions that are not true, should be considered false. His critics claim that if this principle is interpreted as the statement that all legal propositions that are not true are false, then this interpretation implies bivalence only insofar as it is assumed. On the other hand, this principle as a standard addressed to judges has a contingent character, i.e. it may or may not be contained in the legal order. On these and other criticisms to Dworkin’s thesis on bivalence, see Endicott (1997) and Moreso (1998, 63–72).Google Scholar
  80. 80.
    Greenawalt argues that it is legitimate to talk about strong discretion in those cases in which, although it is accepted that a right answer might exist, it is epistemically impossible to identify it. See Greenawalt (1975, 368, 374–375). For a discussion on Dworkin, see Galligan (1986, 17–18). Dworkin has responded to this criticism in two ways: firstly, he claims that a right answer may exist despite the impossibility to empirically prove so. This argument is complex and will be discussed later on. For the time being, let me just point out that Dworkin holds a view on the truth conditions of legal propositions that significantly differs from Galligan’s and Greenawalt’s. Dworkin claims that a proposition is true when it coheres with the set of propositions that offer the best interpretation of a given phenomenon. Thus, Dworkin does not agree that in order for a proposition to be true it has to accord with certain hard facts, i.e. material facts and other facts concerning human conduct. See Dworkin (1985,137–142). About this point, see also Bix (1993, 107). Secondly, Dworkin argues that semantic disagreement does not necessarily imply an absence of right answer for interpreters may continue to think that such an answer exists. They may put forward arguments in favour of a particular interpretation but they may disagree on what the right answer is. This argument has also been misinterpreted. It does not simply blur the distinction between the existence of a right answer and the belief that such an answer exists. It should be analysed within a coherentist theory of knowledge that takes the distinction between what is subjective and what is objective to be less sharp. See Dworkin (1986, 6–15), Waluchow (1983, 332–333, 335–336) and McDowell (1981, 151–152).Google Scholar
  81. 81.
    See Marmor (1992, 54–55), Finnis (1987, 371) and Putnam (1995a, 6–7).Google Scholar
  82. 82.
    On this wide sense of legal gap see, for instance, Raz (1979, 70–74) and Barak (1989, 83).Google Scholar
  83. 83.
    Dworkin (1989, 127–128).Google Scholar
  84. 84.
    Although any situation is liable to be legally regulated, we are compelled to accept that law does not govern every aspect of social life. It thus seems reasonable to claim that certain domains of human life fall outside the domain of law.Google Scholar
  85. 85.
    Raz (1979, 70). Raz defines ‘complete answer’ as the answer provided by the law. Raz (1979, 70–72) distinguishes between these gaps and what he calls jurisdictional gaps’ which arise when courts lack jurisdiction about a particular legal matter.Google Scholar
  86. 86.
    A more technical definition can be found in Alchourrón and Bulygin (1971, 62,). A generic case is that combination of properties whose occurrence or non-occurrence has an implication on the normative solution and which is identified on the grounds of the thesis of relevance assumed by the legislative body. See Alchourrón and Bulygin (1971, 62, 146) and Navarro (1993, 250–252).Google Scholar
  87. 87.
    Alchourrón and Bulygin (1971, 106–107).Google Scholar
  88. 88.
    Alchourrón and Bulygin (1971, 106–111).Google Scholar
  89. 89.
    For Alchourrón and Bulygin, linguistic indeterminacy creates recognition gaps and does not imply an absence of legal answer. Recognition gaps generate semantic-epistemic difficulties, but not legal incompleteness. In these borderline cases, legal propositions may continue to be true or false. See Alchourrón and Bulygin (1971, 33).Google Scholar
  90. 90.
    Recall that Raz divides legal gaps into two types: a) those cases in which it is neither true nor false that law conclusively requires an action, or permits its omission; b) instances for which it is true that law neither requires conclusively a certain action, nor conclusively permits its omission. Linguistic indeterminacy and irresolvable normative conflicts are instances of the former type and are, according to Raz, inevitable. Raz denies the latter type of gap because the principle that everything that is not legally prohibited is legally permitted applies to these instances. This is an analytical truth if we accept the social sources thesis, i.e. that the existence and contents of law shall be determined with reference to social facts. For an account of this argument, see Raz (1979, 72–77).Google Scholar
  91. 91.
    What Raz has called ‘irresolvable normative conflicts’ will be excluded from our account. Their inclusion as cases of absence of right answer requires further consideration. See Raz (1979, 75). I shall moreover point out that our characterisation of legal gap does not presuppose any particular position with respect to the thesis of the existence of legal gaps. So far, I am only interested in a conceptual analysis of strong discretion and in those elements conceptually related to this kind of discretion, regardless of whether the existence of both strong discretion and of legal gaps is accepted at all.Google Scholar
  92. 92.
    Hart calls the creation of norms by the judiciary in instances of legal gaps ‘interstitial legislation’. See Hart (1983, 138).Google Scholar
  93. 93.
    When speaking of the possibility that judges may create norms when they have discretion, we do not mean that judges may create individual norms when applying the law - they always do so. Although sentences have a declarative dimension in cases of discretion, they also have a constitutive dimension for the particular case in hand (except for declarative sentences). The problem of judicial creation of norms by means of discretion turns on the question of whether judges are entitled to determine or create the general normative premise rather than on whether they are entitled to put forward an individual norm that derives from a general norm. On this point, see Kelsen (1970, 236–242). For a discussion of Kelsen’s views, see Bulygin (1991b, 362–366).Google Scholar
  94. 94.
    Ross (1958, 128).Google Scholar
  95. 95.
    See Raz (1979, 77).Google Scholar
  96. 96.
    ’Irresolvable normative contradiction’ refers to that situation in which different norms impute different incompatible legal solutions to the same facts, and in which legal mechanisms of interpretation and principles solving antinomies (lex superior, lex posterior, lex specialis) are of no assistance to select one among the different possible courses of action. On the idea of contradiction as a situation of absence of right answer, see Raz (1979, 74–77), Peczenik (1989, 25), and Moreso (1998, 56–57).Google Scholar
  97. 97.
    Coleman and Leiter (1995, 218).Google Scholar
  98. 98.
    See Bix (1993, 25–26).Google Scholar
  99. 99.
    A similar argument applies to conflicting principles. Leaving aside for the time being the question of whether there is any structural difference between principles and rules (and the potential consequences of such a difference), the solution of an irreducible conflict between legal principles requires strong discretion for the same reasons that the solution of an irresolvable conflict between rules does.Google Scholar
  100. 100.
    A course of action is a valid course of action if there exists a legal norm that justifies it or, equivalently, when it is the result of applying a valid legal norm that has been created by the competent authority in accordance with the prescribed procedure. For a discussion on the different meanings of ‘validity’ see Kelsen (1970, 10–15), and Alchourrón and Bulygin (1979, 41–43).Google Scholar
  101. 101’.
    See Hart and Sacks (1958, 126), Greenawalt (1975, 366), MacCormick (1982, 282), and Bix (1993, 2627). Davis stands as an important exception to this line of thought. He defines discretion as room for freedom to choose among the different courses of action that are within the effective limits of an official’s power. Davis seeks to thus show that neither validity’ nor ‘admissibility’ are adequate concepts to describe strong discretion and discretionary activity as it occurs in reality. In Davis’s opinion, remarkable proportions of discretionary decisions are in fact legally questionable, if not unlawful. See Davis (1969, 4, 12).Google Scholar
  102. 102.
    Barak (1989, 115–118).Google Scholar
  103. 103.
    If there is normative contradiction, the different alternatives in the ‘zone of reasonableness’ are legal answers. This, however, does not imply that the choice among these options is free of constraints.Google Scholar
  104. 104.
    See, for instance, MacCormick (1978, 250–255).Google Scholar
  105. 105.
    See Isaacs (1933, 347–348), Traynor (1978, 1032–1033), Rosenberg (1971, 643) and Galligan (1986, 321–322). Galligan draws a distinction between reasonableness with respect to the aims of law, and reasonableness as ‘moral acceptability’.Google Scholar
  106. 106.
    The concept of reasonableness as moral acceptability has been developed in the theory of legal argumentation and practical discourse. Sometimes reasonableness is linked to the factual consensus on the moral correctness of a specific course of action. On other occasions, it is linked to a hypothetical consensus within an ideal discursive context. On this point, see Aamio (1990, 33–38) and Alexy (1989, 119–124, 160–164, 287–292).Google Scholar
  107. 107.
    MacCormick claims that judicial discretion is restricted by the following three requirements: a) the judicial decision should use legal principles and construct analogies; b) standards such as consistency and coherence should be taken into account; c) consideration should be given to the consequences of the final decision. See MacCormick (1978, 249–251). Similarly, Ruiz Manero (1990, 185–189). It is worth noting that most of these criteria are clearly conservative methods of normative integration for they all aim to draw the reasons justifying the decision as close as possible to settled law.Google Scholar
  108. 108.
    Barak (1989, 117).Google Scholar
  109. 109.
    See also Soper (1984, 3–7).Google Scholar
  110. 110.
    Galligan asserts that understanding discretion as a mere choice between courses of action is too simplistic a view. Decisions are to be made on the basis of reasons because no course of action can be separated from the reasons that justify it. Thus, it would be irrational to follow the course of action X when there are compelling reasons to believe that the course of action Y is better, Galligan (1986, 7–8). Isaacs points out that proceeding without justification would entail an abuse of discretion. In his opinion, this idea is closely related to another basic idea in the conceptual analysis of discretion, namely that asserting the possibility of discretionary proceeding does not mean submitting the case to the arbitrary will of the judicial body. See Isaacs (1933, 349–350). Similarly, Barak (1989, 118–121) points out that judges are not entitled to decide by flipping a coin. The judge must: a) evaluate factors relating to the structure and development of the legal system; b) take into consideration institutional factors; c) take into account the inter-relations among the different institutions. In a similar line of thought, see Wasserstrom (1961, 23–25), Lord Hodson (1961, 1–19), Bell (1992, 96–97), Dworkin (1977, 33) and Pattenden (1990, 395).Google Scholar
  111. 111.
    Raz (1972, 847–848). See also Hart (1980, 9; 1994, 127)Google Scholar
  112. 112.
    Alexy’s definition of the rule of universality rests on the assumption that a speaker can only make those normative judgements that he or she would make in every situation that is identical in all relevant aspects to the one under consideration. See Alexy (1989, 190, 225–228, 275–276), MacCormick (1978, 83–86, 97–99), and Peczenik (1989, 172). Universality can be considered a minimal requirement in formal equality in instances in which formal equality cannot be warranted by means of the principle of legality.Google Scholar
  113. 113.
    On this point, see Raz (1972, 848) and Galligan (1986, 7–8). This demand in fact amounts to the rule of general practical discourse proposed by Alexy, which reads that an agent may only claim what he or she believes. This standard guarantees honesty in practical discourse. See Alexy (1989, 188–189).Google Scholar
  114. 114.
    It was pointed out earlier that Davis’s view on the relationship between discretion and illegality is an exception to this opinion. See Davis (1969, 4–30). See also Dicey (1945, 188) and Hayek (1960, 228, 453). It should be noted, however, that Davis, Dicey and Hayek assume different notions of discretion and arbitrariness. It turns out that these concepts are mutually exclusive only when strong discretion is understood in a restrictive sense, and arbitrariness as a lack of rationality. Thus, an arbitrary decision is that which does not result from a reasoning process based on a general normative premise. On this point, and other meanings of ‘arbitrariness’ as inadequacy of means to ends, or as a lack of predictability of a decision, see Galligan (1986, 143–152) and Aamio (1987, 3–5).Google Scholar
  115. 115.
    Alexy (1989, 179) writes: `Discourses are sets of interconnected actions devoted to testing the truth or correctness of the things we say. Discourses concerned with the correctness of normative statements are practical discourses’.Google Scholar
  116. 116.
    Shiner considers that one of the main differences between simple and sophisticated positivism is their respective conceptualisations of strong discretion. Simple positivism accepts that the extent to which a discretionary act is rational can be evaluated in terms of morality and efficiency. However, it does not understand evaluation against these standards to be of a legal nature because they do not derive from the competent authority through the prescribed procedure. Sophisticated positivism, on the other hand, approves of the separation between law and other normative contexts. Nevertheless, it accepts that a discretionary decision is legally constrained by the obligation to provide reasons, and denies arbitrary proceeding. For an account of these different perspectives see Shiner (1992b, 30–35, 80–87).Google Scholar
  117. 117.
    See Waluchow (1983, 335–336). See also Greenawalt (1975, 384–385). Barak observes that strong discretion is not a mental or emotional state, but a legal situation in which the judge is free to decide between alternatives because law does not determine the course of action to be undertaken. See Barak (1989, 8).Google Scholar
  118. 118.
    Waluchow (1983, 334). The same idea is presented in Shiner (1992b, 32).Google Scholar
  119. 119.
    Waluchow (1983, 333).Google Scholar
  120. 120.
    Waluchow (1983, 333–337). See also Greenawalt (1975, 375) and Galligan (1986, 12–14, 20–22).Google Scholar
  121. 121.
    See Galligan (1986, 12–14, 20–22).Google Scholar
  122. 122.
    For the time being, the question whether the truth of a proposition may transcend the epistemic abilities of an individual agent without transcending the epistemic abilities of a community of speakers will not be discussed. We will postpone this discussion until later chapters.Google Scholar
  123. 123.
    On this meaning of judicial discretion, see Dworkin (1977, 31–32, 69–70) and Isaacs (1933, 339, 342343).Google Scholar
  124. 124.
    In this sense, Sartorius points out that cases of weak discretion are controversial cases in which the judge is liable to make mistakes since there exists a right legal answer which she may fail to identify. See Sartorius (1971, 152–153).Google Scholar
  125. 125.
    See Shiner (1992b, 32).Google Scholar
  126. 126.
    It is assumed at this point that epistemic problems are not due to the interpreter, but rather to the language in which standards are expressed. Thus, epistemic difficulties will emerge when a normative text can be interpreted in several ways, depending on the use of language in the community, rather than on the individual considerations of the interpreter.Google Scholar
  127. 127.
    Aarnio (1987, 1). See also Peczenik (1989, 19).Google Scholar
  128. 128.
    See Barak (1989, 36–39).Google Scholar
  129. 129.
    See Barak (1989, 39–40).Google Scholar
  130. 130.
    See Barak (1989, 40–41).Google Scholar
  131. 131.
    Maximal epistemic difficulties may cause hardship in adjudication because they may prevent the interpreter from making a legal decision. This perspective of weak discretion thus faces a serious problem.Google Scholar
  132. 132.
    See Greenawalt (1975, 365–375) and Galligan (1986, 15–20).Google Scholar
  133. 133.
    Greenawalt (1975, 375). On this point, see also Waluchow (1983, 332–338).Google Scholar
  134. 134.
    For Dworkin’s views on the truth conditions of legal propositions, and of interpretive propositions in general, see Dworkin (1986, 13–15, 49–51, 225; 1985, 131–139, 146–149).Google Scholar
  135. 135.
    On this point see Dworkin (1986, 86–90; 1996, 118).Google Scholar
  136. 136.
    Dworkin considers that these judgements are strictly speaking neither descriptive nor evaluative. They are interpretive statements because they seek to show the object, which they refer to, in its best light. See Dworkin (1985, 146–151; 1986, 65–68).Google Scholar
  137. 137.
    Dworkin (1986, 353).Google Scholar
  138. 138.
    Dworkin (1977, 282).Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2001

Authors and Affiliations

  • Marisa Iglesias Vila
    • 1
  1. 1.Department of Legal, Moral and Political PhilosophyPompeu Fabra UniversityBarcelonaSpain

Personalised recommendations