An improved factor based approach to precedential constraint

There is a controversy in legal philosophy regarding the possibility and nature of precedential constraint, see (Alexander and Sherwin 2008; Schauer 2008). In that realm I have argued that precedential constraint is possible and its nature is best understood through the prioritized defaults of (Horty and Bench-Capon 2012) in my (2014b).

More specifically, the notions of ratio and dicta favor an approach that extracts a rule or rules from an individual case, as the definitions in the text make clear.

In earlier formulations, such as Horty (2004), the antecedent of the rule was the set of all reasons for the prevailing party. The more recent and improved version allows the antecedent of the rule to be a proper subset of the reasons for the prevailing party, allowing the improved theory to go beyond a fortiori reasoning. The determination of which reasons compose the antecedent is part of the process of extracting a rule from a case. That process is independent of the theory.

I ignore the subtleties of contract law remedies in these examples for ease of exposition.

A number of logics have been developed that make use of such defaults. See (Antonelli 1999; Brewka 1991; Horty and Bench-Capon 2012; Horty 2012).

Notice that she can distinguish on the basis that R ₂ ^d combined with R ₁ ^d tilts the scale toward the defendant, which yields Rule₂. She could also distinguish on the basis that R ₂ ^d is so potent that it alone outweighs R ₁ ^p _. This would yield a different rule, namely, {R ₂ ^d } → D☺. The weighing introduced by the second option is {R ₂ ^d } > {R ₁ ^p }. The first approach is thus a more cautious method of distinguishing.

Some have argued against this theory on the grounds that there are always novel sets of reasons for both parties in each case. See (Rigoni 2014a) for a response. There are also questions about aggregating relative weights. For example the > relation is not transitive, so it does not follow from R ₁ ^d  > R ₁ ^p , R ₂ ^p  > R ₁ ^d , and R ₁ ^p  > R ₂ ^d , that R ₂ ^p  > R ₂ ^d . These issues are discussed in (Horty and Bench-Capon 2012, p. 199; Horty 2011, p. 17; Rigoni 2014b, Chapter 3).

R ₃ ^d is highly simplified, as equitable holdings typically involve weighing a number of facts. For simplification I compress all the pro-defendant reasons with respect to equitable indemnity into one reason.

For example, Kritzer and Richards (2003) argue that Lemon marked the beginning of an empirically measurable, distinct epoch in how the Supreme Court decided establishment-clause cases.

One might object that this and the following examples use too coarse-grained a characterization of the relevant reasons. Instead, the objection goes, we should adopt an approach that individuates facts that lead to the conclusion that there is a secular legislative purpose. I address this issue later in this section.

For explanation see (School Dist.of Abington Twp., Pa. v. Schempp 1963, p. 223–225).

It also raises the question of why not take the Walz citation seriously as well and then just make Lemon an entrenchment of that rule. But I am ignoring that question here. See supra at Sect. 4.2.

For example, see (Choper 1987; Cox 1984; Giannella 1971).

Once this is adopted, R _n ^p and R _n ^d will no longer represent reasons that obtain in the case. Instead, they will represent potential reasons. This is unproblematic since the theory assumes that the reasoner has already determined the sets of obtaining pro-plaintiff and pro-defendant reasons that obtain independently of her determination of the rules in the case. I refer to these potential reasons simply as “reasons” in what follows.

Rule₃ imposes the familiar weighing where R ₃ ^p outweighs all the pro-defendant reasons present in Lemon.

This is basically the approach taken by the IBP + HB-C hybrid approach discussed in Sect. 5.3.

See infra, Sect. 5.3 for a discussion of exceptions to the Lemon rules.

See, for example, Marshall's dissent in Stencel Aero Eng’g Corp v. U.S. (1977).

One could go further and stipulate that the three prongs always have equal weight. I do not explore this possibility.

For discussion on this point, see (Horty and Bench-Capon 2012, p. 210; Lindahl and Odelstad 2006; Ross 1957). There are representations of the first step, but they do not neatly map on to the concept of precedential rules (Horty and Bench-Capon, 2012, p. 203–210).

There is another issue lurking here, which I ignore in this chapter, namely, what if the past cases involve businesses that are between 20 and 30 feet from the respective homes? How do would we get from that to a conclusion about businesses 15 feet from homes? See (Rigoni 2014a) for further discussion.

I use this strategy because, as will be seen, it prevents S-rules from chaining. It might be best to allow S-Rules to chain, but I adopt the most conservative possible strategy here. I discuss loosening the reins in my (2014b, Chapter 3).

I am here thinking of instances where a past case causes one to notice a previously unnoticed reason, not anything involving the application of the rules.

There is a different concern that the reasons involved in the excessive entanglement prong and the primary purpose prong are the same, and that the reasons that establish no primary purpose of advancing or inhibiting religion are exactly those that establish excessive entanglement. This is what the Court would later call a “Catch 22” argument (Bowen v. Kendrick 1988, pp. 615–616). I ignore this flaw with the reasoning in Lemon and operate as if the test is coherent.

See Arizona v. Johnson (2009) for a thorough (and unanimously approved) discussion of the cases fleshing out the Terry Rule.

This ignores S-Rules, but the parallel statement for when they are binding is obvious.

One might ask why judges include reasons favoring the prevailing party that are not in the antecedent of the rule used to decide the case. I think there are a number of mutually consistent explanations. First, these reasons may matter to other decisions made in the opinion, such as the setting of damages or a decision to award attorney’s fees. Second, these reasons may be needed to provide a coherent narrative for the facts of the case. The importance of such narratives is discussed in (Bex et al. 2011). For example, that the plaintiff and defendant in a contract case are related may not matter for the resolution of a contract dispute, but it may explain why these parties contracted with each other, which fills out the narrative of negotiation, contract, and breach. Third, giving a relatively complete account of the facts is a customary part of opinion writing in common law jurisdictions. The exact influence of dicta of this sort is beyond the scope of this project. However, an interested reader can use the following exercise to see how pervasive such dicta are: pick an illustration from any Restatement of the law with the opinion upon which it is based. The illustrations are single paragraph narratives involving “A” and “B” that demonstrate the rule of the case. The actual opinion is, of course, much longer and more detailed. One will find a number of reasons for each party that are omitted from the illustration.

In Roth’s case this difference may be due to differences in the legal systems on which we are focused. I am focused on the U.S. legal system, which gives trial courts fact finding discretion, meaning such findings are unlikely to be overturned on appeal. Moreover, the U.S. system permits juries to make a number of determinations that seem to fall within Roth’s view of precedent. Juries are not subject to precedential constraint and are not supposed to even be influenced by past cases. Roth is looking at the Dutch legal system, where there are no juries and appellate courts review factual findings de novo, meaning an appeal is essentially a fresh trial (Nijboer 2007, p. 399, 409). Such a system seems friendlier to precedents that govern determinations on the factual end of the spectrum.

This is perhaps the flip side of Branting’s (1993) idea that the theory behind a case must be incorporated into the precedent it establishes.

In other work by Atkinson and Bench-Capon acknowledge that inferences from evidence to facts are typically performed by the jury and the “style of argument is rather different” (Al-Abdulkarim et al. 2014, p. 63). However they did not attempt to explain the effect this would have on the earlier ASPIC+ proposal in Atkinson et al. (2013).

A similar argument, in a Wittgensteinian vein, can be made regarding relevance. Atkinson et al. do not start by considering cases as mere or bare facts, but relevant facts. The dimensions determine which facts are relevant to the case as well as the polarity of these facts (which makes them seem more like low level factors than facts, but that is beside the point). Yet parties can also argue about the dimensions that are relevant to the case. Post might have argued that his advanced age entitled him deference, for he was 2 years older than Pierce (Ernst 2009), thereby attempting to make the dimension of age applicable. Pierson would of course resist. To capture their argument we must descend yet another level down the hierarchy and use facts to decide which dimensions are relevant. To make this tractable, we must represent only the facts relevant to this question, but one could dispute the relevance of these facts, so down the rabbit hole we go. However, as Wittgenstein would point out, we do not fall into the rabbit hole. Although it is logically possible to continually dispute the relevance of some facts based on further facts the relevance of which is then disputed, we do not do this. Just why we do not is a difficult question, but not one that undermines theorizing legal reasoning.

This issue, or a strikingly similar one, arises in philosophical disputes regarding the nature of law. In the inclusive positivist tradition, Kelsen writes, “just as everything King Midas touched turned into gold, everything to which law refers becomes law” (Kelsen 1967, p. 161). Exclusive positivists reject this notion (Raz 1979; Shapiro 2013), arguing for a distinct demarcation between legal and extra-legal norms.

This discussion is indebted to the suggestions of an insightful anonymous reviewer.

As one might expect given the rather subtle and seemingly incongruent opinions in these cases, there is a large literature criticizing, applauding, and interpreting the case law in this area, see (Cox 1984; Esbeck 1989; Kahle 2005; Lewis and Vild 1989; Meyers 1999; Paulsen 1992).