Some Remarks on Precedent and Analogy

Some Remarks on Precedent and Analogy Giovanni Battista Ratti* Foreword In his excellent and intriguing book on legal reasoning, Thinking like a Lawyer, Fred Schauer strongly defends the thesis according to which we should not conflate precedent and analogy, since both have different aims and work in different ways. In particular, precedent preempts axiological considerations about which would be the best solution for a legal case by imposing a certain solution to subsequent legal questions, whereas analogy is a persuasive tool in order to reach what appears the best solution (or, in any case, a good solution) to a certain legal case. In this paper, I shall try to show that, though Schauer is no doubt right in holding that precedent and analogy are many times different, there are some conceptual connections and similarities between both that deserve a closer look. 1. Schauer on Precedent and Analogy Schauer holds the view that “analogical reasoning is not the same as constraint by precedent” (2009: 85) and that “a genuinely constraining precedent is different from a previous case that may be used analogically” (2009: 86)1. The difference lies in the fact that “in the latter a * Tarello Institute for Legal Philosophy, Department of Law, University of Genoa, [email protected] This draft is a part of a research that was conceived and partially carried out during my stay as a visiting scholar at the Law School of the University of Virginia (JuneSeptember 2015). I would like to thank Fred Schauer for the invitation, his warm and generous hospitality, and his guidance on the topics on which the paper bears on. I would also like to thank the UVA Law School for welcoming me so kindly and for actively supporting my research. Many thanks also to Andrej Kristan and Jorge Rodríguez, for very helpful comments on a first version of the present draft. 1 See also Schauer (2009: 91): “Whereas in the case of analogy the lawyer or judge is looking for assistance in reaching the best decision (or in persuading someone else of previous decision is selected in order to support an argument now, while in the former a previous decision imposes itself to preclude an otherwise preferred outcome” (2009: 88). This distinction is also backed by considerations about the perceptions that accompany both kinds of argument, according to Schauer. He writes in this regard: “Perhaps the most striking difference between precedential constraint and the classic case of reasoning by analogy is the typical lack of freedom that a follower of precedent perceives in the selection of the precedent. Whereas analogical reasoners are widely understood to have a choice among various candidate source analogs, such freedom is typically absent with respect to the genuine constraints of precedent” (2009: 90). This lack of choice (real or perceived as it might be), eventually leading to what is perceived as a bad legal outcome, is the foundation of the distinction between precedent and analogy, according to Schauer (2009: 91). He writes: “Once we understand that in the case of precedent the choice of source decisions is often not perceived as a choice at all—and is often simply not a choice at all—we can grasp the difference between analogy and precedent. Whereas in the case of analogy the lawyer or judge is looking for assistance in reaching the best decision (or in persuading someone else of the best decision), in the case of precedent the effect is just the opposite. When there is an unavoidable similarity between the source and the target, and when the judge because of the constraints of vertical precedent or stare decisis is required to decide the instant case (the target case) in the same way that the precedent case (the the best decision), in the case of precedent the effect is just the opposite. When there is an unavoidable similarity between the source and the target, and when the judge because of the constraints of vertical precedent or stare decisis is required to decide the instant case (the target case) in the same way that the precedent case (the source case) was decided, a judge will be constrained to reach what she will sometimes believe to be a poor outcome. Whereas in the case of analogy a lawyer or judge is looking for a source case in order to help her make the best argument or the best decision now, with respect to genuine precedent the judge will be compelled to make what she may believe to be the wrong decision. Analogy is hugely important in law, and good lawyers and good judges know how to use analogies effectively. But they also understand that precedent may on occasion constrain, and they understand as well that using an effective analogy and recognizing the constraints of precedent are hardly the same”. 2 source case) was decided, a judge will be constrained to reach what she will sometimes believe to be a poor outcome”. In sum, the main differences between precedent and analogy may be summarized as follows: (1) precedent is constraining, whereas analogy is not; (2) who argues from analogy, unlike who argues from precedent, has a choice among the possible sources of the analogy; (3) analogy is used to reach the best outcome possible (or, at any rate, a satisfactory outcome), whereas precedent constrains the applicator in a way that it might turn out (and often turns out) to bring about a suboptimal legal outcome. 2. Some Meanings of “Precedent” “Precedent” is a notoriously ambiguous term because it is normally used to refer at least to five different items: (a) A previous judicial decision; (b) A general rule contained in a certain decision used to justify the particular outcome of the legal question at hand; (c) The particular outcome of the question itself; (d) The argument deployed in a previous decision in order to reach a certain outcome; (e) The “argument from precedent”, i.e. a schema of reasoning based on the authority of the person or organ that reached a certain conclusion and recommending the applicator to found his decision on the previous decision mainly (or exclusively) on the basis that it is a previous decision2. Schauer (2009: 50-54) seems to use “precedent” mostly in the sense of a general rule backed by a certain rationale (as any rule in Schauer’s view). In particular, a precedent framed in its “ideal” form will exactly state which is the rule contained in it (or, better put, which is the rule that it consists in)3. When the formulation is less than ideal, the general 2 See Chiassoni (1999). Schauer (2009: 54): “So when, for example, the Supreme Court decided New York Times Co. v. Sullivan, which dramatically revamped American libel law on First Amendment grounds, it described the plaintiff, Commissioner Sullivan, not only as a police commissioner (which itself would have been an abstraction from Sullivan himself and from Sullivan’s particular job), but also as a “public official.” And it described the New York Times not just as the unique New York Times, and not even just as a newspaper, but as “the press.” As a result, New York Times v. Sullivan, from the beginning, stood as a precedent for all libel cases involving public officials suing the press, and that is precisely because, and only because, the Supreme Court said 3 3 rule must be derived from the holding of the case, its supposed justification, and the facts of the particular case. In both cases, the precedent is regarded as a sort of general rule (or something working as a general rule) 4 and, as such, it can be reconstructed in the typical form of an antecedent bearing upon the key operative facts which triggers the application of a rule, and the legal consequences provided by such a rule5. Thus, the reasoning that precedent so understood involves is typically one of subsumption. The first premise of such reasoning is a general rule (i.e. the holding of the precedent), providing a certain normative consequence connected to certain operative facts. The second premise states that such operative facts have obtained in a particular case. And the conclusion states that the normative consequence provided by the rule follows. In its typical form, precedent consequently works the same way as a legislated rule. It is a matter of course that analogy is usually employed when there is no rule providing for a certain case. Consequently, there is an obvious difference between the use of precedent so regarded and the use of analogy, since the former is simply a tool to carry out a deductive reasoning, whereas analogy is a tool used in ampliative reasoning, creating a new rule from a preexisting one. But in the work of Schauer, usages of the first meaning of “precedent” are also found. For instance, when discussing the distinction between holding and dicta, Schauer (2009: 54 ff.) deals with precedent as a previous judicial decision which can, and usually does, contain both elements. It is clear that, regarded as such, precedent is absolutely just that. If the Court had described Sullivan as a police official, and if in a subsequent case it were to have been argued that New York Times v. Sullivan was precedent for a case in which the libeled plaintiff had been a public official having nothing to do with law enforcement, one side would have argued that it would be good to understand Sullivan as being about all public officials, and the other side would have argued for a narrower interpretation, but neither side would have been able to maintain that its preferred interpretation was compelled by the earlier case, as would have been possible under the language of the Supreme Court’s actual opinion”. 4 For criticism of this view, see Lamond (2006). 5 Schauer (2009: 53-54): “When asked to say why it reached the result it reached, a court will describe the facts of the case before it as an example of a type, but the type is necessarily more general than the particular example of the type – philosophers call it a “token” – that happened to arise in the specific case”. 4 compatible with analogy, insofar as analogy may have been used in a previous judicial decision and consequently constitutes a precedent in this sense (2009: 85-86). “Precedent” denotes the argument from precedent, instead, when Schauer (2009: 42) refers to the fact that while precedent is widespread within the legal domain, the same argument is regarded as fallacious virtually in all the other discursive contexts (such as science, politics, etc.). Indeed, analogy and precedent are different schemata of argumentation, based on different premises. 3. Precedent and Analogy Reconsidered What makes things more complicated, though, is the fact that “A common view is that the precedent case is a precedent not only for more or less identical cases arising in the future, but also for similar cases – cases involving similar facts” (2009: 46). To determine when two cases are similar for these purposes one must mainly look at the background reasons underlying the general rule justifying the decision: “We need to know not only what the precedent court decided, but why it decided it” (2009: 50). If precedents are applied to similar cases (and not only, as it were, to “identical cases”), in what way are they different from analogies? One obvious answer is that the question is ill-formed, since it refers to different domains: precedents are rules (or rule-like items), whereas analogies are schemata of reasoning. To let the question have a sense, we must understand “precedent” in another of the five senses listed above: i.e. as a kind of reasoning based on previous cases. But if we do so, the argument from precedent and the argument from analogy seem hardly dissimilar, at least from a structural point of view. They both aim at providing an unregulated case with a solution, they both use a preexisting rule as their starting point, and they both use the background reason of such a rule to produce a new rule. So, in what aspects might they differ? The first option consists in pointing to the fact that, even though both arguments are sometimes used to fill-up gaps by arguing from similarities, they do so in different ways. While analogy aims at providing reasons to regulate the similar cases alike, precedent aims at 5 barring reasons not to regulate the similar cases differently. So, even if they look structurally indistinguishable, they differ regarding the grounds on which both arguments are predicated. Schauer (2009: 90) suggests that precedent (again understood here as an argument) is binding in the legal domain, whereas analogy is not. This is of course contingent on the content of each legal system. However, let us assume that this is generally true regarding common law legal systems (and partially true also regarding civil law ones). This means that, as a matter of law, judges are obliged (because of an existing rule to this effect) to use precedents, whereas they are not obliged to use analogies6. However, if precedents are structurally indistinguishable from analogies when previous cases are used to solve similar cases, judges are also obliged to use analogies whenever they are obliged to use precedent cases to solve new unregulated cases. Both kinds of reasoning neatly differ only when precedent cases are used as providing the basis for a subsumption: but here the difference is not specific to the precedent/analogy dichotomy, it pertains to the general diversity between deduction and analogy, and has nothing to do with bindingness. At this juncture a precedent, understood as a general rule contained in a previous decision, is used as the major premise of a deduction, and not as the premise of an ampliative kind of reasoning such as analogy. Of course, the general rule used as the major premise can itself be the product of any kind of legal reasoning, including of course analogy, but this relates to how the rule is created, not to how the rule is applied: which is what interests us in this paper. Let me elaborate on this point. Consider the case in which the Italian Constitutional Court held that same-sex marriage was not allowed under Italian law7. The general rule contained in such a case can be opposed by any public official, abstractly competent to celebrate a marriage, to any same-sex couples asking to be 6 This can be due to two different situations: either because there is no rule imposing analogical reasoning or because there is a rule making it optional. 7 See C. Cost. 138/2010. Someone could think that this is not a real precedent (in the sense of a general rule created by judges), since here what is applied is the constitutional provision (art. 29 of the Italian Constitution). But this would be incorrect, since the constitutional provision at hand provides no definition of marriage nor provides an expressed prohibition regarding same-sex marriage. It is the Italian Constitutional Court that creates the rule (“the precedent”) according to which same-sex marriage is not permitted within the Italian legal system. 6 married. If the rule works as a precedent in the sense that Schauer suggests, it also imposes on other courts to stick to what it provides. Therefore, any case of same-sex couples wanting to get married will (probably) find a judicial halt whenever the precedent is applied. Here the precedent is the outcome of a combination of constitutional and legislative provisions, supplemented with some judiciary doctrines. No analogy seems to be involved in the main line of argument. However, the rule, which constitutes the major premise of the argument from precedent, can obviously be the outcome of analogical reasoning. When the Italian Supreme Court held that the tenant (and not the owner) of quotes of a limited company had the right to vote in assemblies, it did so by analogy to what was provided by the law regarding the shares of corporations 8 . But once the general rule was stated, it was applicable (and indeed applied) as a precedent, though it was derived by means of analogy. So we must be careful in distinguishing the way in which the general rule to be applied is created from the way in which the general rule is used. As regards the creation of a general rule, analogy and precedent can perfectly coexist, insofar as analogy can be a means to constitute a precedent. Here four possible scenarios exhaust the relations between analogy and precedent: there can be (a) precedent with analogy, (b) precedent without analogy, (c) analogy without precedent, and (d) neither precedent nor analogy9. By contrast – as already observed – both strategies, qua ways of usage of a general rule, seem to be quite different (at least in the case of “ideal precedents”). In fact, we may say that we have a precedent whenever a certain rule, created in a previous case, is used as a rule, in the sense that the normative consequences it provides are enforced 8 Corte di Cassazione, n. 7614/96. Such relations are graphically represented in the following table: PRECEDENT ANALOGY + + + – – + – – 9 7 whenever the operative facts of the rule obtain10. By contrast, we have an analogy whenever a general rule is used to create another general rule regarding a legally similar case. We have also seen, however, that when precedent is applied to similar cases, differences between precedent and analogy seem to fade away substantially. 4. Constraints on Legal Applicators We have seen that, according to Schauer, who argues from analogy, unlike who argues from precedent, has a choice among the possible sources of the analogy; and that analogy is used to reach a satisfactory legal outcome, whereas precedent constrains the applicator in a way that it might turn out that a suboptimal legal outcome is reached. Here I want to discuss briefly both points. It is true that who argues from analogy normally has different sources at his disposal. Cases like Adams v. New Jersey Steamboat Co11 are in this sense paradigmatic: to decide whether a steamboat cabin is more similar to an open sleeping compartment in a railroad car or to a hotel room gives the applicator a wide discretion in the kind of liability one wants to impose on the owner of the steamboat. However, it is also true that precedents are (possibly felt as) constraining insofar as they are uniform. If we have two conflicting precedents on the same subjects, and they both are considered as binding, the applicator is in the same situation as in the analogy case. Recall the Italian case of the tenant of quotes of a limited company mentioned above. Regarding that question, previous judicial decisions were not uniform. There were judicial decisions dwelling on the regulations bearing upon corporations’ shares as a source of solution of the unregulated legal question and other judicial decisions dwelling on the “nature” of property and tenancy as the main source of regulation. In the first line of cases, the tenant had the vote in assemblies, whereas the owner of the quotes had the vote according to the second line of precedents. The second feature of arguing from precedent to be discussed here relates to the fact that legal applicators might feel compelled to reach a legal 10 11 Guastini (2011: 254). 45 N.E. 369 (N.Y. 1896). 8 outcome that they deem wrong12. It seems to me that this feature is quite debatable as the basis for the distinction between analogy and precedent. If I do not get Schauer wrong, the formulations of feeling of judicial compulsion towards a previous decision are taken as a “symptom” of its bindingness. However, it can be also taken as a symptom for the judge’s conservatism, or lack of fantasy, or sometimes even disingenuity. What I mean is that the formulation of feeling of compulsion towards previous decisions is an underdetermined proof, which can be accommodated into different theories and explanations of precedent. If this is so, it cannot constitute the basis for neatly distinguishing between analogy and precedent. However, we should stress the fact that the argument from precedent naturally revolves around the bindingness of a previous conclusion or argumentation: this is the inferential kernel of the argument. A conclusion ought to be used just because it is previous and/or it comes from an authority. However, it is so by definition: it is a conceptual fiat, rather than an empirical one. It is not the actual formulation of the judges about the bindingness of previous cases that shows that they are constraining. It is rather the other way round. If one wants to use precedents, one has to take them as genuinely constraining. Otherwise, precedent would not do as the basis for an inference13. 5. Precedent, Analogy, and the Continuum of Normative Relevance As noticed before, precedent and analogy can also be understood as different arguments. Here I shall confine to briefly analyzing them from a structural point of view, with an emphasis on the question of normative relevance. 12 Schauer (2009: 89 ff.) offers several examples in which judges state expressly to be bound by what they perceive to be a wrong precedent. 13 Constraints on courts are sometimes “measured” by a counterfactual hypothesis on how a case would have been decided should the precedent not have been taken into account. See Indiana Law Journal, Note, 1963, at 273: “Although MacPherson v. Buick was cited with approval by the court, the actual decision is questionable authority for the conclusion that the rule is now applicable in this state. The difficulty in viewing the decision in this light results from the realization that the case would have been decided the same way regardless of MacPherson”. 9 In its simplified form, the argument from precedent can be understood as the (external) justification for using a certain general rule as the major premise of a certain deductive reasoning on the basis that such a general rule was created or used in previous cases14. By contrast, in its simplified form, the argument from analogy can be understood as the (external) justification for the creation of a new norm to be used as the premise of a certain deductive reasoning on the basis that such a general rule was created or used in previous cases. As Schauer (2009: 52, 86) correctly observes, both arguments rely on the identification of the relevant properties made in view of the application of the relevant rules. Here again we have to distinguish between “ideal” precedents (those consisting in the formulation of precise general rules) and less than ideal precedents (where rules are obtained from the facts of the case and the principles inspiring the decision). In the first class of precedents, relevant properties are derived from (a certain interpretation) of the rule-formulation laid down by the judges15. Regarding the second class of precedents, relevant properties are determined by generalization from the facts of the cases, understood through the lenses of some normative criteria. For instance, the facts of the case in Donoghue v. Stevenson are the following 16 : Mrs. Donoghue, a patron at the Wellmeadow Café in Paisley, Scotland, was with a companion who ordered a glass of ginger beer for her. Mrs. Donoghue drank about half a glass of the ginger beer, and the proprietor then refilled her glass from the opaque bottle, at which time the remnants of a dead snail tumbled into Mrs. Donoghue’s glass. The sight and smell of the decomposed snail caused Mrs. Donoghue gastric distress and mental shock, and she subsequently sued 14 As previously noticed, and as Schauer (2009: 42) correctly observes, the argument from precedent is a logical fallacy in ordinary discourse, since the fact that a certain conclusion was held true by someone (even an expert in his field) is not a reason to conclude that such a proposition is indeed true. Within the legal domain, it happens sometimes that invalid inferences are considered as totally appropriate, if not obligatory. It does so happen with regard to precedent. A previous case – or, better put, the general rule justifying the legal outcome contained in a previous decision – is regarded as binding only because it is prior in time. 15 On the normative relevance of properties, the seminal contribution is of course Alchourrón & Bulygin (1971). 16 I take the description from Schauer (2009: 46). 10 the manufacturer (which was also the bottler) of the ginger beer. From the description of these facts, we can identify several sets of abstractly relevant properties: the opacity of the bottle, the presence of a decomposed snail, the fact that it was the companion that ordered the beer, etc. Of course, the actual relevance of these facts for the judicial decision is “filtered” by the normative principles regarding the field in which the judge operates. For instance, the circumstance that the bottle was opaque did not allow Mrs. Donoghue to look inside the bottle and so detecting the remnants of the snail. Consequently, the possibility of detecting the content of the bottle seems relevant, because in tort law there exists a principle according to which no fault can be ascribed to someone who had no chance to avoid ignorance of certain factual elements (like the presence of the remnants of snail in the opaque bottle of ginger beer). Precedents share this feature with analogy. One cannot build-up an analogical inference without having previously identified the relevant properties of both the source and the target. This identification, in turn, cannot be carried out if one has not previously detected the reason underlying the rule from which the analogical reasoning starts. Schauer (2009: 92) is of course well aware of this common feature of these two arguments. In fact, he interestingly argues that there is a criterion that allows one to differentiate the analogical argument from an argument from precedent in the conventional and strict sense, whenever there are sufficient differences between two cases, so that no judge would be faulted on grounds of failing to follow precedent. Contrary to what Schauer seems sometimes to suggest, though, this criterion is not a neatly structural one: i.e., there is no recursive method to determine, in a finite number of passages, whether we are in presence of an analogy or of a precedent only on the basis that there is (or there is not) a certain number of differences (considered that any two cases are necessarily different in some aspects, as Schauer repeatedly emphasizes). What we have, rather, is a continuum that has at one extreme the “ideal precedent” (the one consisting in the formulation of a general rule) and at the other extreme what we, continental jurisprudents, call “analogia juris” (i.e. the derivation of a normative solution to an unregulated case from a principle, and not from a rule). In this continuum, ideal precedent is followed by a less than ideal precedent (where there are differences which are not present in the ideal 11 precedent). After the less than ideal precedent there is analogy starting from a rule, and after that we find analogy starting from a principle. The continuum is thus characterized by a slight constant move from a few differences (if any) to many differences. It is clear that, more the differences between cases, smaller the fault for not having followed a previous case that can be blamed on the legal applicator. For instance, the precedent in Holy Trinity is a pretty clear example of a rule providing (in a simplified form) that “It is not allowed to pay the journey ticket for a foreigner worker, unless he is a religious minister”. We are at the ideal-precedent end of the continuum. However, take the case MacPherson v. Buick Motor Company17, that “held that the Buick Motor Company, a manufacturer of passenger automobiles, would be liable to a purchaser of a Buick for damages produced by Buick’s incorporation into its automobile of a defective wheel manufactured by someone else, despite the lack of privity of contract between the purchaser and the Buick Motor Company” (Schauer 2009: 45). In cases like MacPherson there will be normally more differences from the leading case than in cases subsequent to Holy Trinity. This is clearly suggested by Schauer (id., 45-46) when he affirms that “Yet even though MacPherson would pretty obviously be a precedent for a claim about a defective Oldsmobile or Toyota or for a case involving defective automobile parts other than wheels, most subsequent cases will not be so similar. If the injury in a subsequent case were caused by a foreign substance in a product normally less dangerous than a car, for example, would MacPherson still be considered controlling precedent?” It’s a very thin line the one which distinguishes the uses of cases like MacPherson as a precedent (even if a less than ideal one) from the uses of such cases as a basis for the analogical argument. By contrast, where analogia juris is used, the differences between the involved cases are overwhelming. A case like this would be one where the principle stated in MacPherson (according to which producers are strictly liable towards the purchasers of their goods, even when there is no privity of contract between them) is used in administrative law, say, to assert the responsibility of a state agency for having passed a rule 17 110 N.E. 1050 (N.Y. 1916) (Cardozo, J.). 12 containing a legal defective element (not recognizable to the non-expert’s eye) which produced distress on citizens involved in the enforcement of such a faulty rule. 6. Conclusion Schauer is surely right in neatly distinguishing precedent from analogy. However, as I have tried to show, such a neat distinction only holds regarding precedent understood as a precise rule (what I’ve called “ideal precedent”), whereas it tends to vanish regarding other notions of precedent. In particular, when precedents are regarded as constraining on similar cases (and not only concerning “identical” ones) the differences between both grow dim. Likewise, other strategies focusing on the perceptions of the legal applicators or their freedom of choice are questionable. Finally, if we understand precedent and analogy as arguments, they are not severable in an all or nothing fashion, but rather they occupy a space of a spectrum that has ideal precedents and analogy based on principles at its ends, but a lot of in-between situations occupying the inner spaces of such a spectrum. REFERENCES Alchourrón, C.E, Bulygin, E. (1971). Normative Systems, Wien – New York, Springer. Chiassoni, P. (1999). La giurisprudenza civile. Metodi d’interpretazione e tecniche interpretative, Milano, Giuffrè. Guastini, R. (2011). Interpretare e argomentare, Milano, Giuffrè. Lamond, G. (2006). Precedent and Analogy in Legal Reasoning, in “Stanford Encyclopedia of Philosophy”, http://plato.stanford.edu/entries/legalreas-prec/ Schauer, F. (2009). Thinking like a Lawyer. A New Introduction to Legal Reasoning, Cambridge (Mass.), Harvard University Press. Indiana Law Journal, Note 3 (1963). Has the Rule of MacPherson v. Buick Been Adopted in Indiana?, in “Indiana Law Journal”, 38/2. 13