- 1 See Vittorio Villa, A Pragmatically Oriented Theory of Legal Interpretation, Revus – European Const (...)
- 2 Pierluigi Chiassoni, Tecnica dell’interpretazione giuridica, Bologna, Il Mulino 2007, 50.
- 3 See Villa 2010 (n. 1), 99.
1 This essay aims to examine the theory of legal interpretation recently proposed by Vittorio Villa:1 a theory that, I think, has a lot of merits and highlights some important and often neglected aspects of legal practice. Villa’s theory concerns the study of the narrow notion of legal interpretation, i.e. the textual legal interpretation, ‘the activity that consists in determining the meaning of a disposition [i.e. a provision] (the basic component of every legal text), deriving from it one or more explicit norms, accredited as legally correct interpretation of them’:2 in particular, Villa’s theory aims at analysing the structural profile of textual legal interpretation, i.e. at asking which type of relation should be established between interpretation and meaning.3
- 4 Actually, as Villa notes, interpretative anti-formalism also stresses the importance of the context (...)
2 Villa expressly presents his position as a pragmatically oriented theory, that is, as a theory which focuses on the pragmatic aspects of legal interpretation and the inter-linguistic dimension of meaning. This approach is clearly original: in fact, almost always within the analytical tradition (although, perhaps, not within the hermeneutic tradition), the problems of legal interpretation are confined within semantic-syntactic boundaries, i.e. they are seen as semantic-syntactic flaws (such as vagueness) or they are viewed as connected to judicial discretion and, in particular, to the power of judges to use different canons of construction that exceed the literal interpretation.4 Instead Villa’s approach also connects legal interpretative problems (their onset and solution) to pragmatic factors, and, doing so, he throws a bridge between legal practice and ordinary communication, where context plays a fundamental and undeniable role.
- 5 Villa 2010 (n. 1), 106–7, emphasis in the text.
3 In particular, Villa proposes a dynamic theory of (legal) interpretation and sees the concept of meaning ‘as a notion that appears as stratified (in the sense that it contains different levels), inclusive (it includes both sense and reference) and [..] having a progressive formation (in the sense that the attribution of meaning is a “process in several phases”)’.5
4 It seems to me that Villa’s theory can be reconstructed as being based on two assumptions (a1; a2) and as articulated in three, interconnected, thesis (t1; t2; t3) – that I will rename freely below.
-
(a1) Semantic contextualism: ‘It is only within a specific context of use that the sentence, expressed by a given speech act, enacts a complete communicative message’.6
-
(a2) Reductionist assumption: ‘a prescriptive meaning strictly speaking does not exist’;7 the problems connected to the semantic content of a sentence (the phrastic, in Hare’s terms) are identical (or not so different) for both assertions and prescriptions.
-
- 8 See Villa 2010 (n.1), 113ss.
(t1) Club sandwich thesis: meaning is (can be constructed as) a stratified concept which includes, at least, meaning in a weak sense (the reference and the sense of single words), meaning in a narrow sense (the topic of the sentence, the phrastic) and meaning in a broad sense (the autonomous quantum of communication that is inferable through interpretation).8
-
- 9 It is worth noting here since, according to Villa, the conventional (or literal) meaning does not c (...)
(t2) Mixed thesis: legal interpretation always consists of the discovery of conventional meaning9 and in the creation of the complete meaning through reference to the various contexts of reception.
-
- 10 Villa 2010 (n. 1), 112.
(t3) Monistic thesis: the interpretative process remains ‘unitary because it starts from a common semantic basis’;10 this common semantic basis constitutes the framework of the complete meaning of each sentence (i.e. meaning in a broad sense) and represents a major constraint for the process of specification and the concretization of the complete meaning.
5 In the next section I will develop some objections to the previous assumptions and theses, and, in particular, to (t1): I will argue that (t1) can be intended in two different ways, but, however we intend it, it is neither coherent with other aspects of Villa’s thesis, nor heuristically valuable. Moreover, I will try to show that Villa’s intuitions about the dynamic and mixed nature of (legal) interpretation would be better developed if he abandoned the club sandwich thesis and adhered to a more radical form of contextualism (in the sense that will be clarified later). Then, in the last section, I will consider some conditions that Villa’s theory should meet to adequately account for the peculiarities of legal interpretation.
- 11 See Villa 2010 (n. 1), 107.
- 12 For a theoretical reconstruction which is fully in line with Villa’s methodological approach to leg (...)
- 13 Against this commonly accepted view see Gordon P. Baker & Peter M.S. Hacker M.S., Language, Sense a (...)
- 14 Therefore, the distinctions drawn by (t1) are heterogeneous with respect to the distinction between (...)
- 15 In fact, the general presumption according to which they express norms , although they are formulat (...)
6 First of all, we have to specify what is the status of Villa’s theory: is it a descriptive theory or a normative model? It seems to me that Villa does not intend to prescribe a model that the interpreter should follow, therefore his theory, not being normative, can be considered as descriptive. What does this theory describe? Villa is very clear on this point: he underlines the fact that his theory is not a psychological description of what really happens in the interpreter’s mind,11 rather Villa proposes a theoretical reconstruction,12 which must be evaluated on the basis of its theoretical adequacy, i.e. on the basis of its internal coherence and congruence, and of its external heuristic value, its power of explanation. It seems to me that the club sandwich thesis (t1) cannot meet these conditions. Although the stratified thesis of meaning is commonplace in the philosophy of language,13 I think that it is misleading in so far as it suggests that the distinct dimensions of meaning do not interact with each other in the actual dynamics of any communication process. In other words, this analysis would be appropriate only if Villa thought that the three levels of meaning that he distinguishes do not interact with each other or, perhaps better, that between these levels there is only a one-way interaction (i.e. the lower level determines the upper, but not vice versa). However, I do not think that Villa shares this view, as it emerges, among other things, by the conjunction of the thesis (t1) and (t2). According to (t2), we can distinguish between conventional meaning and the complete (contextual) meaning of a sentence; according to (t1), within a sentence, we can distinguish the meaning of single words (meaning in a weak sense), the meaning of topics (meaning in a narrow sense) and the complete meaning (meaning in a broad sense). However, in Villa’s theory, the meaning of single words and the meaning of topics does not coincide with conventional meaning, but they are also determined by pragmatic factors.14 Thus, for example, the meaning in a weak sense includes the attribution of sense, and the attribution of sense includes both a semantic dimension (that is, the recognition of a stable conceptual basis) and a contextual dimension (that is, the reconstruction of a complete notion). Now, this latter phase not only determines (and, in fact, belongs to) the third level of meaning (meaning in a broad sense) but, and above all, it is hard to think that it is not guided by the previous reconstruction of the other two upper levels. Consider the example of the word ‘vehicle’, proposed by Villa. How is it possible to reconstruct a complete notion of a‘vehicle’ without taking into account that this word is related to entering in a park (topic), it is employed in a prescriptive way (background assumption),15 it the object of a prohibition, it is (perhaps) defined by other legal provisions or it is employed in other legal provisions (co-text), it is part of a norm to which, in the legal culture in question, certain purposes can be attributed (e.g. to prevent the risk of accidents) and not others (e.g. to encourage physical activity), etc.? In other terms, while (t1) seems to suggest that among the three levels of meaning (meaning in a weak sense, meaning in a narrow sense, meaning in a broad sense) there is only a one-way interaction, instead it is more in keeping with Villa’s own theory to recognize that each level of meaning determines and, simultaneously, is determined, by the others.
- 16 Villa 2004 (n. 1), 124: ‘si può benissimo comprendere l’argomento (il significato in senso stretto) (...)
- 17 Villa 2010 (n. 1), 111.
7 I think that Villa would agree with these considerations. But why then use an apparatus that does not describe the actual usage but provides a misleading theoretical reconstruction? The point is even clearer in respect to the meaning of topics (meaning in a narrow sense). Elsewhere, Villa claims that ‘we can understand the topic (the meaning in a narrow sense) expressed by a sentence without knowing (or at least putting in brackets) the function which it carries, and therefore without knowing what is the complete communication message (the meaning in a broad sense) conveyed by the sentence’.16 Actually, it is doubtful as to what degree of understanding we can have of the topic ‘Entering into the park by vehicles’. Arguably, a low degree. I mean, we understand what a ‘vehicle’ is generically or ‘Entering in a park’ (and we understand that, usually, it is something different from ‘Taking pictures in a museum’), but we cannot understand what this topic communicates. More precisely, if we only consider the topic, the problem is not that we cannot adequately solve the questions related to the specification of the meaning: the problem is that we cannot even ask these questions because, as Villa writes, ‘without reference contexts and speech acts the words that we pronounce and write remain completely inert’.17 The meaning in the broad sense, such as the phrastic (Hare) and/or the locutionary act (Austin), has no autonomy, neither from the temporal point of view (we do not understand it before), nor from a logical point of view (we cannot understand it in isolation). But, instead, (t1) seems to suggest just such an autonomy for each level of meaning: if we deny that each level of meaning is autonomous, if we admit that among the three levels there is a bi-directional interaction, then there seems no reason to support (t1).
- 18 See, e.g., Jean Piaget, Le jugement moral chez l’enfant, Paris, Alcan, 1932; Kay Bussey, Lying and (...)
- 19 Actually, this thesis has made it possible to define the effectiveness of a norm as the truth of (...)
8 Actually, a reason to adopt (t1) seems to endorse (a2): the reductionist view of prescriptive language is maintained precisely on the grounds that (i) the topics (the semantic contents) of both norms and assertions are identical, and (ii) the prescriptive function is parasitic to the assertive one. This point is very complex, and its exhaustive discussion is beyond the scope of this brief essay. Here I shall confine myself to observing that, as a matter of fact, we learn to prescribe and to issue norms much earlier than to make assertions: psychological studies have shown that the basic rule of assertion (the rule ‘Say only what you believe is true’) is learned rather late, at 4-5 years, when of course we already know how to speak and are already able to make requests and commands.18 It is difficult to see reasons to maintain a thesis denied by practice, especially since, until now, it has produced no appreciable result.19
- 20 Actually Villa always connects (t1), i.e. the stratified concept of meaning, to the distinction bet (...)
9 However, one can object that, according to Villa theory, the distinction between levels of meaning must be drawn in another way: it must be drawn as a distinction between conventional (semantic) meaning and contextual meaning. In fact, this view seems to derive from (a1), (t2) and (t3).20 This way of reading (t1), raises at least the following issues: is a totally a-contextual semantic meaning really identifiable? If it is, is this semantic meaning binding? Can it not be overcome by contextual considerations? It seems to me that this way of constructing the club sandwich thesis is internally adequate only if Villa is disposed to answer ‘Yes’ to all of the previous questions. Moreover, it is externally adequate only if these answers are true.
- 21 See Charles Travis, The True and the False: the Domain of Pragmatics, Amsterdam, Benjamins, 1981; C (...)
- 22 The logical priority of the conventional meaning is expressly stated by Villa 2006 (n. 1).
10 As far as t internal adequacy is concerned, I cannot see how the existence of a totally acontextual meaning can be coherent with Villa’s thesis about the distal context (or background). The background argument aims to prove that the syntactic-semantic meaning is always underdetermined, that there is a generalized and pervasive contextual dependence. Against the relevance of this background it is often claimed that, although it is clearly a pragmatic element, it is so deep, shared and hackneyed that it can be considered as part of the conventional meaning. However, this conclusion is controversial. The point is that, if we consider only the syntactic-semantic meaning, we cannot predict which aspects of the background will be significant: it is the context of the utterance (the proximal context) which determines which aspects of the background are relevant to understanding the meaning of a utterance.21 The background argument seems to also prove that a sentence has no meaning without a context: or, better, that what we call literal meaning is only the meaning a sentence expresses in the context which is, for us, typical, statistically more frequent. But the actual context cannot be a typical one, and a context which is typical for us may not be typical for other speakers. If one accepts this argument, it is difficult to argue that there is a totally acontextual meaning which is logically prior and binding with respect to the contextual meaning.22
- 23 Villa 2010 (n. 1), 98.
11 It is worth noting that I’m not claiming that (t1) is in contrast with (a1), i.e. with semantic contextualism. Semantic contextualism argues that ‘It is only within a specific context of use that the sentence […] enacts a complete communicative message’,23 and this is perfectly compatible with the existence of an incomplete acontextual meaning, which acts as a constraint on the full contextual specification of the communicative message. Instead, I claim that the existence of a totally acontextual meaning, which is both logically prior and binding with respect to contextual meaning, is in contrast with the background argument: in fact, the background argument proves that the so-called ‘acontextual meaning’ (e.g. the meaning of the sentence ‘The cat is on the mat’) is not at all ‘acontextual’; instead it is determined by background assumptions that, in turn, are determined by the context of utterance (in the sense, that, as we have seen, the context of utterance determines which background assumptions are relevant).
12 Consider now a variation on the example offered by Villa: Does the rule prohibiting vehicles from the park apply to an ambulance called to help someone inside? The ambulance seems to me a paradigmatic case of a ‘vehicle’ and, anyway, it clearly falls into the conceptual basis of the notion of vehicle (‘a means of transport for people or things, especially mechanical and driven by man’). It is worth noting that this is an easy case, in the sense that the decision seems simple: in our present legal culture at least, it is obvious to think that the prohibition does not apply to the ambulance. However, it not clear whether Villa would see this decision as a case of the judicial creation of law since it exceeds the boundaries of the semantic meaning.
13 In fact, if the semantic meaning is binding (i.e. binding as a matter of meaning, if it is an insurmountable barrier in the interpretative process), then the contextual specification can concern only ‘hard cases’: so, for example, within the attribution of reference, the contextual dimension of the construction of a possible field of extension can concern only the denotation of objects that are not clear cases of ‘vehicle’. Since an ambulance is a clear case of vehicle, since it falls within the core (in the conceptual basis) of the notion of vehicle, then there is no room for a contextual specification. But this contradicts Villa’s thesis according to which legal interpretation is always a matter of discovery and creation.
- 24 Villa 2010 (n. 1), 115, emphasis added.
14 For the previous considerations, the internal adequacy of the theory under discussion seems rather doubtful to me; instead, as far as the external adequacy is concerned, it should be noted that legal interpretation sometimes does not start from the semantic meaning of a term, but from definitions expressed by other legal provisions – provisions that represent a contextual element (more exactly, a co-text). In these cases at least, there is not, strictly speaking, a semantic basis, or rather the only semantic basis is offered by the co-text (which is a contextual, pragmatic, factor). Villa is aware of this fact: he claims that his approach ‘in the limits within which it recognizes the – partial – autonomy of the dimension of sense […] makes it possible to reconstruct in the semantic key the complex interpretative operations that jurists carry out when they work on the terms used by the legislator, exploring the systematic connections with other terms in the legislative language’.24 However I cannot see how this could fit with the picture painted by (t1) – which, as we have seen, requires the identification of an acontextual semantic meaning.
15 Summing up, (t1) does not truthfully describe the actual communicative processes and the actual legal interpretative practices, and, moreover, it does not seem appropriate to Villa’s own theory. With respect to the latter point, if (t1) is related to the distinction between meaning in a weak sense, meaning in a narrow sense and meaning in a broad sense, it suggests erroneously that there is only a one-way interaction between these meanings. Instead, if we read (t1) in the light of (t2) and (t3), i.e. if we refer (t1) to the distinction between conventional (semantic) meaning and complete (contextual) meaning, it seems to suggest that there is always a totally acontextual (semantic) meaning, but this is in contrast with both the background argument and (t2). Of course, this latter criticism holds also for (t3) alone: in so far as (t3) implies that there is always a totally acontextual (semantic) meaning which acts as a constraint on the interpretative process, it excludes the possibility that there may be a contextual specification in easy cases (cases determined by the semantic content) and this in turn contradicts (t2).
- 25 The difference between the various forms of contextualism (radical contextualism and moderate conte (...)
16 In general, it seems to me that the club sandwich thesis ties Villa to a still too traditional vision of meaning, while his interesting insights on the specification of contextual meaning would require a stronger form of contextualism. By this I mean a form of contextualism which admits (and which is able to account for the fact) that semantic and pragmatic factors influence each other so much that it is difficult to distinguish between them and, in any way, pragmatic factors can overwhelm the semantic basis25. Metaphorically speaking, meaning is not a stratification of semantic and pragmatic layers, but rather it is like batter – and the radical versions of contextualism are better equipped to account for it.
17 Does Villa’s theory fit the peculiarities of legal interpretation? In order to answer this question, it is clearly necessary to wait for a complete articulation and exposition of his theory. For the time being, I shall highlight two points that his theory should account for.
- 26 Mario Jori, Definizioni giuridiche e pragmatica, Analisi e diritto 1995, 109–144.
18 As stressed by Jori,26 two features characterize legal practice: the pervasive presence of disputes and authorities. In fact, according to Jori, a general pragmatic feature of legal language is that it does not work spontaneously: the relative harmony in its use is not guaranteed by the conformist interest of the majority of its users to understand and be understood (as in everyday conversation), nor by the unanimous interest in using the more efficient linguistic tools (as in technical languages). Roughly speaking, in legal practice, the interest in winning the case always outweighs the interest in understanding and Jori argues that, precisely for this reason, modern law is a language which must be administered and which is meant to be administered.
- 27 Actually, as Mario Jori pointed out to me, even in legal practice there are situations in which the (...)
- 28 Villa 2010 (n. 1), 109.
- 29 I explored this issue in Francesca Poggi, Law and Conversational Implicature, International Journal (...)
19 Up until now, it is not clear how Villa’s theory can account for these two features, and especially for the pervasive presence of disputes. In my view, this fact is related to the difference between legal interpretation and ordinary understanding: while in ordinary understanding the speaker’s intention plays a central role, in legal interpretation it does not.27 In the legal practice, conflicting attitudes can explode because the context is not fixed but must be can be reconstructed in different ways: the conflicts focus precisely on the identification of the relevant context. But Villa does not think so: he claims that ‘[m]uch of what contextualism maintains on the subject of the “meaning of speaker” can very well be applied to the “meaning of the receiver” or to the meaning of the “receiver of message”’.28 This view seems to me problematic: according to the contextualism, the complete meaning is the speaker’s meaning, but how is it possible a speaker’s meaning without a speaker? More exactly, in the ordinary conversation the speaker relies on contextual elements for the communicative outcome that he assumes as mutually known. During personal discussions, Villa has often pointed out that, even in ordinary speech, the relevance of some elements of the context of utterance may be obscure or opaque. He is certainly right. If the speaker is wrong, if some elements of the context of utterance are not mutually known or he is not able to make them salient, then communication fails. But, the point is that, in ordinary conversation, the speaker relies on (some elements of) the context of utterance in order to be understood, whilst in legal interpretation the context of utterance has almost no relevance (and the speakers, i.e. the legislators, know that it is so).29 In legal interpretation, the message is, so to speak, suspended in a vacuum: its specification is assigned to a context of reception ex ante unpredictable and ever changing. As we have seen (§ 2), Villa partially avoids the ineffability of the legislative message anchoring it to its semantic meaning, but at the cost of limiting the relevance of the context of reception (the possibility of a contextual specification) to the hard cases: this amounts to saying that in the easy cases, i.e. in the cases determined by semantic meaning, the context of reception plays no role, and therefore there is no room for the interpreter’s discretion in reconstructing that context. But then how do we explain the pervasive presence of disputes also in the easy cases (as in the previous mentioned case of the ambulance)? Obviously, we can explain the pervasive presence of disputes as due to the bad faith of either party, but this is not a convincing explanation, especially because the party in bad faith – i.e. the part that supports an interpretation which does not rest on semantic meaning – might be, or be held, right.
20 Moreover, it is worth noting that, while the context of reception can change considerably, the judicial specification of the message cannot: in our legal culture at least, the consistency with previous decisions is considered a value and a constraint. This limit is precisely one of the ways in which the authority manifests itself, the other being the presence of statutory definitions, and it is but another pragmatic element that must be taken into consideration, and that, in my view, is stronger than any semantic constraint.
21In conclusion, it seems to me that also with respect to the peculiarities of law, the theory developed by Villa would benefit from admitting that semantic meaning can be overwhelmed by pragmatic factors. Obviously, this would make it more difficult to distinguish between creation and interpretation of the law, but such a distinction is in fact problematic – and it is so despite the reassuring image suggested by (t1) and (t3).