This Article offers the first targeted study of the Supreme Court's use of canons and other tools of statutory interpretation in a "dueling" manner--that is, in both the majority and dissenting opinions in the same case, to support opposing outcomes. Taking its inspiration from Karl Llewellyn's celebrated list of canons and countercanons, this Article examines how often and in what ways the members of the Roberts Court counter each other's references to particular interpretive tools when disagreeing about the proper reading of a statute. Many of the Article's findings are unexpected and undermine the assumptions made by some of the most prominent theories of statutory interpretation. Most notably, the data reveal that several of textualism's most-favored interpretive tools are at least as susceptible to dueling use as the purposivist tools that textualists have long denigrated as indeterminate and readily subject to judicial manipulation. For example, the study shows that the Justices dueled extensively over the meaning of statutory text. By contrast, they dueled at far lower rates over legislative history, purpose, and intent. Moreover, the Justices dueled over dictionary references, the whole act rule, and language canons at rates that were virtually identical to the rates at which they dueled over the purposivist-preferred tools. The study also reveals that the canons do not seem capable of constraining the Justices to vote against ideology and that noncanon tools of analysis, including precedent and practical-consequences based reasoning, lead to higher rates of dueling than do most traditional canons or tools of statutory interpretation. After reporting the data, the Article examines doctrinal patterns in how the Justices duel over individual canons and explores the theoretical implications of the Justices' dueling canon use.
Table of Contents Introduction I. Background A. Llewellyn's "Thrusts" and "Parries" B. Prior Empirical Work II. Findings A. Methodology B. Dueling Statistics 1. Frequency of Dueling 2. Statutory Subject Matter 3. Dueling Canons and Ideology III. Doctrinal Findings A. Inside Story: Patterns of Dueling Canon Use 1. Text / Plain Meaning Rule 2. Whole Act Rule 3. Statutory Purpose 4. Practical Consequences B. Other Observations 1. Dictionary Definitions 2. Legislative History IV. Theoretical Implications A. Textualism B. Purposivism C. Pragmatism D. Methodological Stare Decisis Conclusion Appendix: Dueling Canons by Case
Judge A is an appellate-court judge tasked with interpreting a criminal sentencing-enhancement statute. She believes that there is a strong argument based on the expressio unius canon (1) that the enhancement provision does not apply to the defendant in the case. Judge B, who is sitting on the panel with Judge A, believes that the purpose of the sentencing statute covers conduct like the defendant's and that the enhancement therefore should apply. Will Judge B write an opinion that relies on statutory purpose alone, or will he search for an opposing language canon to counter Judge A's expressio unius argument? Will Judge A's opinion confine itself to discussing the expressio unius canon, or will it also seek to demonstrate that her interpretation is consistent with the statute's purpose--perhaps framing that purpose in different terms than Judge B's opinion? How easy will it be for Judge A and Judge B to find countervailing purpose or language canon arguments, if they seek to counteract each other in this manner?
If we extrapolate from the above hypothetical, two larger questions emerge: To what extent do judges tailor the interpretive tools they discuss in statutory opinions to counter the tools referenced by opposing opinions in those same cases? Or, put somewhat differently, to what extent do various canons or tools of statutory interpretation lend themselves to "dueling" use, such that they can be employed to support competing statutory constructions? These are two important, yet rarely asked, questions in statutory interpretation. Both questions implicate some of the key theoretical debates in the field--such as whether certain interpretive tools constrain judges more than others--and both are deeply tied to underlying views about the proper role of judges.
In 1950, Columbia Law Professor Karl Llewellyn offered a memorable answer to the second question, in what has been called "one of the most celebrated law review articles of all time." (2) The article famously challenged the view that the canons of statutory construction provide neutral, predictable legal rules that lead courts to one "correct" reading of a statute. (3) Its most feted feature is a list of twenty-eight pairs of canons and countercanons, which Llewellyn labeled "thrusts" and "parries." The article's demonstration that the canons readily can be used to cancel each other out has been taken to demonstrate the canons' illegitimacy and has served as a catalyst for discussions about the ability of interpretive tools to constrain judges. (4)
But to show that the canons (or other tools of construction) are capable of being used by judges in a discretionary, canceling-out manner is not to say that they will be used in such a manner. Theoretical possibility is not the same as actual judicial practice.
This Article takes Llewellyn's famous juxtaposition of canons and countercanons as its inspiration and examines the extent to which the modern Supreme Court actually duels over the most common statutory interpretation canons and tools in opposing opinions in the same case. That is, the Article identifies the extent to which a majority (or occasionally, concurring) opinion's reference to a particular canon or tool is countered by a dissenting opinion's offsetting reference to the same canon or tool, in the same case. Notably, the Article measures a different kind of judicial dueling over interpretive canons and tools than that suggested by Llewellyn's "thrusts" and "parries": the Article counts as dueling those instances in which a majority and dissenting opinion in a case invoke the same interpretive tool to reach different readings of the statute. It does not measure instances in which a majority opinion invokes one tool (e.g., purpose) and the dissenting opinion counters by relying on another tool (e.g., text or a language canon). The reasons for this methodological choice are elaborated in Part II.A.
Despite the Llewellyn article's tremendous influence on statutory interpretation scholars, only two empirical studies to date have considered, even in passing, the extent to which majority and dissenting opinions rely on the same interpretive canons to counter each other's constructions of the same statute. (5) One of the studies was limited to employment law and only briefly discussed dueling references to language and substantive canons. (6) The other focused exclusively on the Supreme Court's use of legislative history and paid only passing attention to judicial dueling over this tool. (7) No study has examined whether, or to what extent, the Court uses the canons in a dueling manner when construing statutes dealing with subjects other than employment law. And, importantly, no study to date has compared the extent to which the Court uses other traditional statutory interpretation tools such as purpose, dictionaries, the common law, and the like in a dueling fashion, period. This Article offers the first empirical evidence of this kind, going beyond the linguistic and referential canons contained in Llewellyn's list, (8) and also exploring the extent to which majority and dissenting opinions duel over the full array of other statutory interpretation tools. (9)
Thus far, statutory interpretation theory has operated with a blind spot regarding the extent to which particular interpretive tools and canons are used to support competing statutory constructions in the same case. Scholars and jurists have made assumptions about how manipulate they think certain interpretive tools are, but no one has tested those assumptions systematically or examined how judges use interpretive tools to counter each other in practice. This Article begins to address that lacuna.
The Article's findings are surprising, and intriguing, for statutory interpretation theory. Five points stand out: (1) the overall rate of dueling canon or interpretive tool use in the Roberts Court's first five terms was low--at or below 25.0 percent for most tools; (10) (2) the rates of dueling for text/plain meaning and precedent were much higher than for other tools, at 42.7 percent and 63.3 percent, respectively; (11) (3) the rate of judicial dueling over most textualist-preferred interpretive tools was roughly the same (about 25.0 percent) as the rate of judicial dueling over purposivist-preferred interpretive tools; (12) (4) statutes dealing with certain subject areas--criminal law, environmental law, and antidiscrimination law--showed particularly high rates of dueling; (13) and (5) none of the canons or tools seemed capable of constraining the Justices' tendency to vote consistently with their ideological preferences, at least in divided-vote cases. (14)
These empirical findings have important implications for some of the key debates in statutory interpretation. They are particularly relevant to theoretical approaches that emphasize predictability or seek to constrain judicial discretion. For example, the findings call into question some of textualism's (and in particular Justice Scalia's) claims about the unique manipulability of interpretive tools like purpose, intent, and legislative history as compared to text, language canons, dictionary references, and "other statutes." (15) The findings also cast some doubt on purposivism's claims that reliance on legislative history, relative to other interpretive resources, helps courts act as faithful agents of the legislature. (16) The findings even have implications for pragmatism--a theoretical approach that does not claim to promote predictability or constrain judges--because the data confirm that practical consequences play a significant role in the judicial interpretation of statutes, and because they show that judges duel along predictable lines with respect to this interpretive tool. Specifically, the data reveal that most of the Court's dueling over practical consequences involves one opinion that emphasizes administrability-type practical concerns and an opposing opinion that emphasizes policy-constancy concerns. (17)
The Article also engages in doctrinal analysis of how precisely majority and dissenting opinions invoke the same interpretive resource to reach different readings of a statute. This analysis reveals some noteworthy patterns in the Court's dueling over particular canons. For example, in applying the plain meaning rule, the Justices often divide over the prototypical, core meaning versus the legalist meaning of key statutory text. (18) Similarly, when dueling over a statute's purpose, the Justices who favor one reading of the statute sometimes focus on a generalized purpose, while those who favor an opposing reading focus on a narrower, more specific purpose. These patterns provide valuable insights and add texture to our understanding of how the Court applies the tools of statutory construction. The patterns also suggest that it may be possible to eliminate some of the dueling that occurs over certain interpretive tools--through clearer meta-rules about how those tools should be applied. (19)
The Article proceeds in three parts. Part I examines the theoretical and empirical background for this study. Part II reports the study's findings on the Roberts Court's dueling use of interpretive canons and other tools from the middle of the 2005 term, when Justice Alito joined the Court, to the end of the Court's 2010 term. It also provides doctrinal analysis of the Court's dueling over several individual interpretive tools. Part III explores the theoretical implications of the data and doctrinal observations.
A. Llewellyn's "Thrusts" and "Parries"
Ironically, Llewellyn's "fiendishly deconstructive" (20) attack on the canons--the famous list of "Thrust But Parries"--was not the focus of his article. Indeed, Llewellyn included the list only at the end of the article, as an illustrative exercise. (21) His focus, instead, was on the broader legal-realist claim that judges do not decide cases based on neutral legal rules, and that neutral legal rules capable of producing a single correct answer do not in fact exist. Llewellyn had made this point earlier in the common-law context, arguing that case law precedents are highly malleable and subject to multiple, conflicting applications. (22) His article was designed to show that the same is true in the statutory interpretation context, despite the existence of numerous seemingly definitive "canons of construction." (23)
One reason why Llewellyn's list of "thrusts" and "parries" has endured, and proved so influential among academics, is because the list seems to demonstrate concretely that the canons are easily manipulated and, therefore, incapable of constraining judges. Llewellyn did not actually argue that courts would use his canon--countercanon pairs to reach opposing statutory constructions in the same case, but the strong implication was that judges looking to counter a statutory reading based on one canon could readily find an opposing canon to support their preferred outcome. Indeed, Llewellyn called it a "foolish pretense" for courts to justify their statutory constructions based on "a set of mutually contradictory ... rules on How to Construe a Statute." (24)
If Llewellyn was correct that judges can readily find countercanons for every canon and that judges use the canons as after-the-fact justifications for their decisions, then we might expect, in practice, to see judicial opinions pitting canon against canon to neutralize the tools that supposedly support an opposing reading--at least in cases where judges disagree about the best reading of the statute. For example, we might expect to see dictionary definitions in a majority opinion countered by opposing dictionary definitions in the dissenting opinion, references to statutory purpose in a majority opinion (e.g., "the Civil Rights Act was enacted in order to remedy the nation's long history of discrimination against racial minorities") countered by an opposing statutory purpose in the dissenting opinion (e.g., "the Civil Rights Act's primary goal was to create a color-blind society"), and a majority opinion's reliance on the whole act rule met by reliance on a contradictory subset of the whole act rule in the dissenting opinion. (25) Although this Article does not measure the Court's use of Llewellyn's precise form of canon-for-countercanon pairings, it takes his canons-can-point-in-different-directions heuristic as inspiration for examining how often majority and dissenting opinions in the same case use the same canon to support different statutory constructions.
In a less-celebrated passage of his article, Llewellyn also embraced a purposivist approach to statutory interpretation. He argued that "if a statute is to make sense, it must be read in the light of some assumed purpose." (26) Llewellyn did not claim that judicial reliance on statutory purpose would be more determinate than reliance on the canons, but his simultaneous rejection of the canons and endorsement of statutory purpose raises the comparison. This is particularly so given that the New Textualism popularized by Justice Scalia takes precisely the opposite view--hailing the canons as neutral rules that enhance predictability, while arguing that statutory purpose is malleable and easily shaped to justify an individual judge's policy preferences. (27) Indeed, both Llewellyn's list and New Textualism's claims about the relative merits of different interpretive tools invite empirical analysis of the extent to which statutory purpose--as well as other noncanon tools, like legislative history--are susceptible to Llewellyn's "thrust and parry" criticism.
Part II of this Article answers these questions with data from the Roberts Court's first five terms. Before turning to the data, however, the next Section reviews the two empirical studies that have addressed judicial dueling over the canons of construction and legislative history, respectively.
B. Prior Empirical Work
As noted above, only two studies to date have attempted to measure the extent to which judges use the canons of construction against each other to support opposing readings of the same statute in the same case. The first study, conducted by James Brudney and Corey Ditslear, reviewed the Supreme Court's use of interpretive canons in every workplace-law case decided between 1969 and 2003. (28) The study focused on the extent to which the canons of construction operate as neutral legal rules, constraining the Justices' ability to interpret statutes according to their ideological preferences. Brudney and Ditslear treated "dueling canons" in passing, as a way to test claims that the canons enhance consistency and predictability and perform an important "gap-filling" function. (29) They made two findings that seem to support Llewellyn's skeptical view of the canons: (1) dissenting Justices were significantly more likely to rely on language and substantive canons when language canons were part of the majority's reasoning than when they were not; and (2) the canons failed to constrain the Justices' ideological preferences, as conservative Justices tended to use the canons to reach conservative results, while liberal Justices tended to use the canons to reach liberal outcomes. (30) Based on these findings, the authors concluded that, "in divided decisions, the Justices themselves are more prone to view the canons as reasonably amenable to supporting either side." (31)
In a subsequent article expanding on their original study, Brudney and Ditslear found some evidence that judicial reliance on legislative history does have a constraining effect on judicial ideology. Specifically, they reported that liberal Justices were more likely to vote in favor of employer interests when they invoked legislative history, (32) and that, during the Burger Court, conservative Justices were more likely to vote in favor of worker interests when they relied on legislative history. (33) However, Brudney and Ditslear found that after 1986, conservative Justices ruled increasingly in favor of employers (consistent with their ideological preferences), relying on canons and even legislative history to do so. (34) Brudney and Ditslear also reported what they called a "Scalia Effect," finding that liberal Justices opted not to rely on legislative history in a series of proemployer majority opinions that Justice Scalia joined, and that when liberal Justices did rely on legislative history, Justice Scalia was significantly less likely to join their majority opinions, even when he voted for the same result. (35) Conversely, the authors found that Justice Scalia's resistance to legislative history did not extend to majority opinions authored by his conservative colleagues; that is, he was just as likely to join his conservative colleagues' majority opinions when they relied on legislative history as when they did not. (36)
These findings are significant, particularly given the large sample size of workplace-law cases examined by Brudney and Ditslear. But the findings also are limited in their generalizability because, as Brudney and Ditslear acknowledged, their study examined cases in only one subject area--workplace law. (37) It is thus unclear whether the significant level of dueling that Brudney and Ditslear found over language and substantive canons is representative of how the Justices use the canons in all statutory interpretation cases. Moreover, Brudney and Ditslear, like Llewellyn, focused exclusively on language and substantive canons. But there are many other interpretive tools that courts regularly use to construe statutes--including the plain meaning rule, purpose, legislative history, dictionary definitions, and congressional intent. And there are non-statute-specific tools of legal analysis, such as precedent and practical consequences, which recent empirical work has shown to play a significant role in the judicial interpretation of statutes. (38) The Brudney-Ditslear study thus provides an incomplete picture of how the Court duels over statutory interpretation tools other than language and substantive canons, leaving open the questions of how susceptible other tools are to competing inferences and whether such tools provide any constraining effect on judges' ability to vote according to their ideological preferences.
The second empirical study to date on judicial dueling is David Law and David Zaring's comprehensive study of the Supreme Court's use of legislative history. (39) Law and Zaring studied every Supreme Court statutory interpretation case decided from 1953 to 2006. Their focus was on identifying legal factors that might influence the Court to rely on legislative history, such as the age or length of a statute. (40) Like Brudney and Ditslear, they examined judicial dueling over legislative history only in passing, as a small part of their larger project about legislative history use. (41) And, like Brudney and Ditslear, Law and Zaring found that dissenting Justices were significantly more likely to cite legislative history when a majority or concurring opinion also cited legislative history than when no other opinion cited this interpretive resource. (42) Law and Zaring hypothesized that these data "suggest that Justices are sensitive to the types of arguments made by their colleagues and feel an obligation or desire to respond in kind, especially when they disagree with one another on the merits." (43)
Law and Zaring's findings, like Brudney and Ditslear's, are important, especially given the depth and breadth of their sample size--all cases before the Court that involved statutes over a fifty-year period. Their study was not limited by subject area, so their findings are less likely to be distorted by partisanship, judicial impressions of congressional expertise, age of a statute, obscurity, or other concerns that might be particular to one area of the law. But Law and Zaring examined judicial dueling over only one interpretive tool--legislative history--shedding no light on competing judicial invocations of the numerous other canons and tools of statutory construction. Thus their study, too, only scratches the surface in exploring how Justices authoring an opinion respond to interpretive canons and tools used by an opposing opinion.
The next Part seeks to provide a more complete picture of the Justices' dueling over multiple interpretive tools as well as to explore doctrinally how the Justices engage each other's use of particular interpretive tools.
The findings and conclusions presented below are based on empirical and doctrinal analysis of all decisions in the Roberts Court's 2005 (post-January 31, 2006) (44) through 2010 terms that confronted a question of statutory interpretation. Every case decided during that period was examined through the Supreme Court's online database to determine whether it dealt with a statutory issue. (45) Any case in which the Court's opinion contained a discussion about statutory meaning was included in the study. Cases interpreting the Federal Rules of Civil Procedure were not included, (46) but a handful of constitutional cases in which the Court construed the meaning of a federal statute before deciding the constitutional question were included. (47) This selection methodology yielded 255 statutory cases over five-and-a-half terms, with 255 majority or plurality opinions, 103 concurring opinions, 156 dissenting opinions, 12 part-concurring/part-dissenting opinions, and 2 part-majority/part-concurring opinions, for a total of 528 opinions. (48) Of these, 115 cases were decided unanimously, and 140 were decided by a divided vote.
In coding and analyzing these cases, my primary goal was to determine the frequency with which the Court referenced a range of interpretive sources when giving meaning to federal statutes. The cases in the study were examined for references to the following interpretive tools: (1) statutory text, including appeals to plain meaning; (2) dictionary definitions; (3) grammar rules; (4) the whole act rule; (5) other federal and state statutes; (6) common-law precedent; (7) substantive canons; (8) Supreme Court precedent; (9) statutory purpose; (10) practical consequences; (11) legislative intent; (12) legislative history; and (13) language canons such as noscitur a sociis and expressio unius. (49)
These interpretive sources are consistent with those examined in other empirical studies of the Court's statutory interpretation practices. (50) A few differences in definitions used for the different sources were inevitable and will be pointed out where notable. For example, unlike the Brudney and Ditslear study, which grouped together several interpretive tools under the heading "language canons," I counted separately references to grammar canons, language canons, and the whole act rule. (51) However, in order to allow comparison with Brudney and Ditslear's data, I also created a combined variable that coded for reliance on any one of these interpretive resources (so that if an opinion referenced at least one of these interpretive sources, it was coded as a "yes"). Further, unlike some previous studies, (52) I recorded as a reference to "practical consequences" any reliance on the absurdity of a result, the administrative or other burdens caused by an interpretation, the fairness of an interpretation, an interpretation's coherence or incoherence, the workability of an interpretation for lower courts, or other effects that an interpretation could be expected to produce. I also further disaggregated this interpretive tool, coding for administrability-type practical consequences concerns versus policy-constancy-type practical consequences concerns. (53)
In recording the Court's reliance on the above interpretive tools, I counted only references that reflected substantive reliance on the tool in reaching an interpretation. Opinions that mentioned a particular interpretive tool but rejected the tool as unconvincing were not counted. Similarly, I did not count instances in which the Court merely acknowledged, but did not accept, a litigant's argument that a particular canon or tool dictated a particular result. (54)
Secondary or corroborative references to an interpretive tool were counted. That is, when the Court relied primarily on one interpretive tool but went on to note that x, y, and z tools further supported that interpretation, the references to x, y, and z were coded along with the primarily relied-upon source(s). (55)
In addition, the vote margin in each case was recorded, and each case and opinion was recorded as unanimous, close margin, or wide margin (cases with six or more Justices in the majority). (56) Each Justice's vote in each case also was recorded, as was the author of each opinion. This methodology comports with my previous empirical study. (57)
To measure the Court's use of interpretive canons and tools in a dueling fashion, I sorted the cases in the dataset by docket number and identified those that showed a majority and dissenting opinion (or a concurring and dissenting opinion) referencing the same canon or tool. For these purposes, I defined "same canon or tool" to mean that both the majority and dissent argued that a specific interpretive resource--purpose, dictionary definition, substantive canon--supported their respective readings of the statute. I did not count as "dueling canon" cases those in which the dissenting opinion mentioned, or even criticized, the majority's application of a particular canon or tool, unless the dissenting opinion also argued that the interpretive tool supported its reading of the statute. (58) I made this methodological choice because in my view, judicial rejections or disagreements over whether an interpretive canon applies in a particular case do not constitute disagreements over the meaning dictated by the canon. Rather, such disagreements show merely that the canon has limitations or exceptions and that the opposing opinion author found the argument from the canon powerful enough to require criticism. Further, while battles over applicability do provide some evidence that a canon does not constrain judges--who retain the discretion to refuse to apply the canon in a particular case--they do not show that the canon itself is malleable or indeterminate. That is, disagreements over applicability do not necessarily reflect an underlying looseness as to what construction the canon, if applicable, directs the Court to adopt or the canon's ready susceptibility to judicial massaging to support a judge's chosen construction. Thus, counting such disagreements as instances of "dueling canon" use would be misleading, measuring something other than how often the Justices used particular interpretive tools to support opposing statutory readings. (59)
I also did not count as "dueling canon" cases those in which a majority opinion relied on one interpretive canon, such as the whole act rule, while the dissenting opinion relied on another interpretive canon or tool, such as statutory purpose or a dictionary definition. (60) On the one hand, a high rate of this kind of interpretive resource dueling would provide some evidence that the interpretive canons and tools do not constrain judges by showing that judges retain significant discretion about which canons to apply, and which ones to privilege when different canons point in different directions. But measuring this type of dueling would not reveal very much about the indeterminacy of individual interpretive tools or show how often the Justices counter each other's references to specific interpretive resources. Instead, it would demonstrate only, or at least primarily, that there is no settled methodology dictating which interpretive canons and tools judges should use to construe statutes, and no hierarchy indicating the order in which they should prioritize particular tools when different tools point toward different constructions. The whole act rule is not necessarily indeterminate simply because statutory purpose points toward a different interpretation; and the entire practice of consulting canons and tools of statutory construction is not meaningless simply because judges faced with different tools pointing in different directions disagree about which tools to privilege. (61) Accordingly, in order to provide a better picture of individual canons' malleability, this study counted as dueling only those cases in which a majority opinion's reliance on purpose was countered with a dissenting opinion's reliance on purpose, or a majority opinion's reference to a dictionary definition was countered with a dissenting opinion's reference to a dictionary definition, or a substantive canon reference in a majority opinion was countered with a substantive canon reference in a dissenting opinion, and so on.
B. Dueling Statistics
Before reporting the data, it is important to note some limitations of this study. First, the study covers only five-and-a-half Supreme Court terms and only 255 statutory interpretation cases, decided by some combination of the same eleven Justices. While this dataset is large enough to teach us some things about the Court's use of canons and interpretive tools in a dueling manner, the data reported may reflect trends specific to the Roberts Court. Second, although the number of cases reviewed is large enough to provide some valuable insights, the focus should be on the patterns that emerge rather than on precise differences in the percentages reported. Third, in noting the canons and other interpretive tools referenced in majority and dissenting opinions, I make no claims to have discovered the Justices' underlying or "true" motivations for deciding statutory cases; the data do not reveal whether a particular opinion relied on a tool because the opinion's author was persuaded by that interpretive tool, or merely because the author felt it necessary to counter an opposing opinion's reliance on that tool. The study's empirical and doctrinal claims are confined to describing how the Justices publicly justify their statutory constructions, and to theorizing about discernable patterns in the kinds of public justifications the Justices regularly provide.
1. Frequency of Dueling. At the outset of this study, I expected to find higher rates of judicial dueling over dictionary definitions, legislative history, and statutory purpose and lower, but still high, rates of dueling over language canons and substantive canons. Specifically, because dictionaries contain multiple definitions for each word--and because the Justices have multiple dictionaries to choose from--I expected to find the Justices countering each other's dictionary references in the vast majority of cases--i.e., in the vicinity of 60 percent to 70 percent of the cases. I expected to find similarly high rates of judicial dueling over legislative history because resourceful lawyers can be expected to dig through the (often copious) legislative history of a statute and find some snippet to support their clients' preferred statutory construction. Further, because both proponents and opponents of a proposed law often make competing statements on the House and Senate floor about the likely effects or scope of the bill, there is a wealth of material for the Justices to invoke in support of opposing readings of a statute. With respect to statutory purpose, I expected to find high rates of judicial dueling because statutes often have multiple purposes, again providing the Justices with ample fodder for fashioning an argument that their chosen statutory construction is more consistent with a statute's purpose than is the opposing opinion's.
For substantive canons, I expected to find lower, but still meaningful rates of judicial dueling (perhaps in the range of 30 percent to 40 percent). This expectation stemmed from the fact that substantive canons are judicially created policy norms, so they should prove relatively easy for the Justices to craft anew, or to tweak and invoke in order to lend an aura of inevitability and consistency with larger legal norms to their chosen interpretation. I similarly expected to find noteworthy rates of judicial dueling (perhaps in the 30 percent range) over language canons. This expectation was based on the fact that it is often possible to characterize the common denominator connecting statutory terms in different and competing ways, so as to support competing noscitur a sociis or ejusdem generis arguments about what a term in a list means. I also expected the rates of dueling over text/plain meaning to be lower than the rates of dueling over legislative history, purpose, and dictionary definitions. For the other canons and interpretive tools, I did not have specific expectations about how often the Justices would employ them in a dueling manner.
Table 1 lists the frequency with which the Justices on the Roberts Court employed the various canons and tools of statutory construction in a dueling fashion--that is, in both the majority and dissenting opinions in the same case--in the 255 cases decided and the 528 opinions issued from the time Justice Alito joined the Court in 2006 through the end of the Court's 2010 term. (62) For each interpretive tool, the Table first reports the number of unanimous opinions, majority opinions, and dissenting opinions that referenced the tool. It then reports the percentage of cases in which a majority and dissenting opinion in the same case both referenced the interpretive tool. This figure is calculated using as a denominator the total number of divided-vote cases in which the interpretive tool was invoked by at least one opinion (second-to-last-column) and, separately, using as a denominator the number of all cases--including unanimous cases--in which at least one opinion referenced the tool (last column). (63)
I believe the former is a better measure of the rate at which the Court "dueled" over the canons and other interpretive tools. If Table 1 used the total number of cases in the dataset as the denominator for calculating the rate of dueling, that would count as potential dueling canon cases those cases in which no member of the Court referenced a particular canon or tool. But cases in which no opinion referenced a particular interpretive tool might be cases in which no on-point legislative history existed, or in which no other statutes were analogous to the one at issue, and so on. In any event, it would not provide a good measure of how frequently the Justices found it necessary to counter each others' claims that a particular interpretive resource supported a particular statutory construction. For the sake of thoroughness, I note that if Table 1 were to use the total number of cases in the dataset (255) as the denominator for its calculations, the rates of dueling would be significantly lower for every interpretive canon and tool (e.g., SCP--22.7%, Plain Meaning--17.3%, Dictionary--6.3%, Other Statutes--7.8%, Language/Whole Act--8.2%, Purpose--7.8%, Practical Consequences--11.4%, Intent--5.1%, Legislative History--7.5%).
The data reveal several surprising results. First, prior Supreme Court precedent and text/plain meaning show very high rates of dueling during the Roberts Court's first five-and-a-half terms--the Justices dueled over the application of Supreme Court precedent in 63.7 percent of the divided-vote cases in which at least one opinion cited precedent, and they dueled over text/plain meaning in 42.7 percent of the divided-vote cases in which at least one opinion found a clear statutory meaning. (64) (The percentages are significantly lower when unanimous cases are included in the count, but still higher than for other interpretive resources). Practical consequences and other statutes exhibited the next-highest rates of dueling, generating competing judicial references in 30.9 percent and 34.5 percent of the divided-vote cases that referenced either resource, respectively. By contrast, the Justices dueled over purpose, legislative history, and intent in roughly 25.0 percent of the cases in which they invoked these tools. The rates for these interpretive tools are almost identical to the rates at which the Justices dueled over textualist-preferred tools, such as dictionary definitions (28.6 percent), the whole act rule (21.8 percent) and the combined grammar/language canons/whole act rule (24.7 percent), and the common law (22.2 percent). Finally, the data revealed that the Justices rarely dueled over substantive canons, at a rate of just 10.4 percent. (65)
In short, the Court engaged in the highest rates of dueling canon or interpretive tool use when invoking traditional tools of legal analysis--i.e., precedent, text, and to a lesser extent, practical consequences--not when invoking statutory-interpretation-specific tools such as the language or substantive canons, statutory purpose, legislative history, the whole act rule, or dictionary definitions. The only statutory-interpretation-specific tool for which the Justices exhibited rates of dueling comparable to those exhibited for traditional legal tools was other statutes--an interpretive tool which requires traditional legal analysis of prior Supreme Court precedents interpreting analogous statutory provisions, in addition to the statutory-interpretation-specific task of making analogies across statutes. (66)
The data are intriguing on many levels. First, the relatively low rate of dueling over dictionary definitions is noteworthy and unexpected. Dictionary definitions are bountiful; there are numerous dictionaries in print and each dictionary typically contains several definitions for the same word. Moreover, a recent study of the Supreme Court's dictionary usage from 1986-2010 found that the Justices' choices regarding which dictionary to cite to be "largely ad hoc, based on the appeal of particular dictionaries in particular cases." (67) That study also found that in the vast majority of cases, judicial opinions that referenced dictionary definitions used them as "ornament[s]"--meaning that the dictionary played a minimal substantive role in the opinion's reasoning, but nevertheless was used to "lend a patina of objectivity and legitimacy" to the chosen construction. (68) Given the abundance of dictionary definitions and empirical evidence indicating that the Justices use them in an ad hoc manner, one would expect that jurists seeking to use this interpretive tool readily could find some definition to support almost any construction of a statute--and that majority and dissenting opinions in the same case frequently would employ competing dictionary definitions, however marginally helpful, to cancel out any aura of objectivity that an opposing opinion might gain from using such definitions.
Thus, the low rate of dueling dictionary use--28.6 percent--is puzzling. (69) One explanation could be that the Justices tend to prefer a handful of dictionaries, so that despite the abundance of dictionaries and definitions in publication, competing definitions may not be readily available within their preferred lexical universe. (70) But upon closer examination, this explanation seems insufficient for at least two reasons. First, the Court's opinions in recent cases like Taniguchi v. Kan Pacific Saipan, Ltd. (71) demonstrate that despite the Justices' individual preferences for certain dictionaries, they are perfectly willing to reference other dictionaries (as many as fourteen in Taniguchi!). (72) Second, empirical research shows that even those Justices with the most pronounced dictionary preferences deviate from their preferred dictionary in a substantial number of cases. (73)
Another possible explanation for the relatively low rate of judicial dueling over dictionary references is that, despite the availability of numerous dictionaries, there may have been no competing definition available in the forty cases in the dataset in which majority or dissenting opinions declined to counter an opposing opinion's dictionary reference(s). In order to test this possibility, I (or a research assistant) examined the briefs in thirty-eight of the forty cases in which a majority or dissenting opinion referenced a dictionary definition, but the opposing opinion(s) did not counter that reference. (74) In the vast majority of the cases (67.5 percent), both the petitioner and the respondent (or, in some cases, amici) had provided dictionary definitions supporting their respective statutory constructions. Moreover, in another handful of cases, the party whose interpretation was favored by the opinion that did not invoke a dictionary definition did provide a supporting dictionary definition in its brief. (75) Ultimately, in 75.0 percent (thirty out of forty) of the cases in which a majority or dissenting opinion declined to counter a dictionary reference in an opposing opinion, at least one brief supporting the opposing opinion provided a helpful dictionary definition. Thus, the lack of availability of opposing definitions does not seem to explain the low level of dueling--nor does a lack of framing by litigants and attorneys.
A third possibility is that the Justices are motivated to counter an opposing opinion's dictionary references only, or primarily, in those cases in which the opposing opinion relies significantly on dictionary definitions to reach its statutory construction--and tend to leave unanswered those dictionary references that are used merely as "ornaments." I explore this possibility further in the doctrinal analysis conducted in Section III.B. (76)
Second, the Roberts Court's relatively low rate of dueling over legislative history is quite surprising. The legislative history of most federal statutes is extensive, and debate on the House and Senate floor often produces competing statements about a statute's meaning. (77) Accordingly, as with dictionary definitions, one would expect there to be "something for everyone" in the legislative history--and for this to translate into frequent dueling over this interpretive resource. Yet the data from this study reveal judicial dueling over legislative history in only 25.3 percent of the divided-vote cases. Part III will explore the theoretical implications of this finding in detail, but for now it bears noting two possible explanations for the data: On the one hand, the conventional wisdom could simply be wrong, and the legislative history of many statutes may not in fact contain support for competing characterizations of the statute's meaning. Alternatively, the Justices who make legislative history references may not be citing directly on-point "smoking gun" statements, and those authoring opposing opinions accordingly may not consider it crucial to counter such legislative history citations.
Third, the low rate of judicial dueling over substantive canons also is unexpected. Substantive canons are judicially created interpretive presumptions and rules based on background legal norms, policies and conventions. (78) They reflect judicially preferred policy positions, expressed as rules of thumb about how to treat statutory text in light of constitutional priorities, common-law practices, or specific statute-based policies. Because substantive canons are policy-based, there often are two or more that point in opposite directions. Indeed, Llewellyn lists in his "thrust" column the substantive canon that "[s]tatutes in derogation of the common law will not be extended by construction" and, as its "parry," the countercanon that "[s]uch acts will be liberally construed if their nature is remedial." (79) Accordingly, it seems surprising that the members of the Roberts Court dueled over substantive canons at a rate of only 10.9 percent. (80)
Last, the high rates of judicial dueling over precedent and practical reasoning, as compared to the low rates of dueling over purpose, legislative history, the whole act rule, and the like suggest that the canons and tools of statutory construction may not be as readily manipulate as the traditional tools of legal analysis that Llewellyn sought to compare them to in his article. Or, at the least, the Justices on the Roberts Court do not seem as inclined to use the tools of statutory construction to counter each other as they do to use the common-law tools of analysis in a dueling manner.
Before we rush to conclude that the Court rarely duels over statutory-interpretation-specific tools, however, it is worth looking at the rate of dueling by case type rather than by individual interpretive tool. As Table 2a shows, only 88 of 255 statutory cases decided during the period studied (34.5 percent) involved competing references to the same interpretive tool in both a majority and a dissenting opinion, excluding cases in which the Justices dueled only over Supreme Court precedent or practical consequences. (81) This is not a trivial number, but neither does it constitute a terribly high rate of dueling. If we separate out unanimous cases and focus on the 140 cases in which the Justices divided over a statute's construction, the rate climbs significantly, indicating that the Roberts Court engaged in dueling canon use in 62.9 percent of such cases. Put differently, a large majority of the cases in which the Justices disagreed over the construction of the statute involved majority-dissent dueling over the same interpretive canon or tool.
In addition, when the Justices used the interpretive tools in a dueling fashion, they tended to do so on multiple levels, dueling over numerous resources. Indeed, the members of the Roberts Court dueled over two or more interpretive tools in 83.0 percent of the eighty-eight dueling canon cases. Only fifteen of the eighty-eight dueling canon cases (17.0 percent) involved majority and dissenting opinions that dueled over only one interpretive tool. (82) Further, as Table 2b below shows, the data from the Roberts Court's first five terms are consistent with the Brudney-Ditslear and Law-Zaring findings that dissenting opinions were significantly more likely to invoke a canon or legislative history when the majority opinion relied on that interpretive resource than when the majority opinion did not. On the one hand, this finding could be taken to demonstrate that the Justices are strategic in their interpretive tool use--tailoring their references to counter the specific tools relied on by an opposing opinion rather than referencing only those tools they independently find compelling. This is the conclusion that Brudney-Ditslear and Law-Zaring reached based on similar data. (83) On the other hand, the cases in which a majority opinion references a particular interpretive tool may also be cases in which citation to that tool is more appropriate than in the average case. In many cases, for example, the legislative history may not contain on-point discussion of the relevant statutory issue; in those cases where it does, and the majority references that discussion, a competing comment about the statutory provision may also be available. Thus we should be cautious before ascribing too much meaning to this statistic. (84)
Overall, the data regarding the frequency of dueling canons provide a mixed picture. The rates of judicial dueling for most individual tools of construction are low--certainly lower than I expected to find for legislative history, dictionary references, language/whole act, and substantive canons. At the same time, a large percentage of the Court's divided-vote cases involved some level of majority-dissent dueling over the same interpretive tool. Indeed, in 52.1 percent of the cases in which the Justices disagreed (73 of 140), majority and dissenting opinions dueled over multiple canons or tools. Thus, on the one hand, the worst implications of Llewellyn's "thrusts" and "parries" do not appear to be borne out, in that most individual tools of construction are not being pitted against themselves in the vast majority of cases. On the other hand, however, the tools of statutory construction hardly seem to constrain the Justices or point neatly to one correct interpretation, as nearly two-thirds of the Court's divided-vote cases (88 of 140) contained competing applications of the same interpretive tool.
2. Statutory Subject Matter. If we break the cases down by statutory subject matter, the data become still more interesting. As Table 3a illustrates, subject matter seems to have a notable correlation to the rate at which the Roberts Court dueled over one or more interpretive canons when interpreting the statute. Table 3a reports the percentage of all dueling canon cases that involve statutes in a particular subject area. (85) Caution should be used in interpreting these figures because once the data are disaggregated by statutory subject matter, the number of case observations dips significantly. Accordingly, this Article focuses on only the most striking data from Table 3a: criminal and antidiscrimination statutes each made up a substantial proportion of the Roberts Court's dueling canon cases; nearly 40 percent of the cases in which the Court dueled involved a statute that fell within one of these two subject areas. (86) If we add in cases construing environmental statutes, nearly half (46.5 percent) of the Court's dueling canon cases are accounted for. (87)
This subject-matter sensitivity could be one reason why the Brudney-Ditslear study reported greater overall dueling canon use than Table 1 suggests. Because Brudney-Ditslear reviewed only employment cases, including many antidiscrimination cases based on statutes such as Title VII, the Fair Labor Standards Act (FLSA), and the Age Discrimination in Employment Act (AEDA), their study may unwittingly have selected a subset of statutory cases that are particularly susceptible to judicial dueling.
The subject-matter data naturally raise the question: Why? What is it about criminal, antidiscrimination, and environmental statutes that makes the Roberts Court particularly likely to duel over canons and interpretive tools? One possibility is that there is something about statutes in these subject areas that gives rise to greater indeterminacy or to competing applications of interpretive tools than is the case for statutes in other subject areas. For example, Congress often borrows language from existing criminal statutes when it writes new ones, particularly when a new statute deals with the same type of underlying conduct as older statutes. (88) Similarly, Congress often uses Title VII or the Fair Labor Standards Act as the template for other, newer antidiscrimination statutes. (89) As a result, there may be multiple analogous statutes available for the Justices to choose from when applying the other statutes tool to help decipher the meaning of a criminal or antidiscrimination statute. (90) Alternately, statutes in these subject areas may reflect more ideologically charged legislative battles than do other statutes, giving rise to cross cutting statements in the legislative record, including competing claims about statutory purpose. Thus, there might be a large quantity of contradictory legislative history or statutory purposes available to the Justices when construing statutes in these subject areas.
As Table 3b shows, the data contain some evidence that is consistent with these explanations. But the evidence is modest and does not hold for all three subject areas. For example, nearly one-fourth of the dueling canon criminal cases contained dueling references to the other statutes and legislative history tools, as did one-fourth of the dueling canon antidiscrimination cases--but none of the environmental cases dueled over either of these tools. (91) Conversely, one-third of the dueling canon environmental cases contained dueling references to statutory purpose, but the figures were much lower for dueling criminal and antidiscrimination cases. (92)
Another possibility is that statutes in the criminal, antidiscrimination, and environmental law areas reflect clearer ideological dividing lines than do statutes dealing with other subjects. (93) With respect to the environmental law and antidiscrimination law cases, the statutes being interpreted typically are sweeping progressive or social-justice statutes enacted during the 1960s by liberal Congresses--in a political climate very different from the one that has existed during the Roberts Court's tenure. (94) Thus, the clash between majority and dissenting opinions may be political, with the liberal Justices seeking to preserve the original, broad goals of a statute enacted in a more liberal era and the conservative Justices seeking to curtail the modern application of the statute. As a result, when construing statutes in these subject areas, the Justices' ideological preferences may predominate and they may be more inclined to use the canons to justify their preferred statutory constructions than is the case with other, less controversial subject areas. That is, the members of the Roberts Court (1) may be more inclined in criminal, antidiscrimination, and environmental law cases to start with their preferred outcomes and then look for ways to shape the canons and interpretive tools to justify those outcomes; or (2) may regard the stakes as higher for these cases and, accordingly, consider it more necessary to neutralize the interpretive tools used by opposing opinions.
There is some anecdotal evidence for this explanation: a number of the U.S. Supreme Court's most memorable statutory interpretation cases have involved significant judicial dueling over statutes in the antidiscrimination and environmental law areas. Consider, for example, the Rehnquist Court case, Babbitt v. Sweet Home Chapter of Communities for a Great Oregon f Babbitt involved a section of the Endangered Species Act (ESA) that makes it unlawful for any person to "take" an endangered species within the United States. (96) The ESA defines the term "take" to mean, in part, "to harm." (97) In 1975, the Department of the Interior issued a regulation defining '"[hjarm' in the definition of 'take'" to include any activity that results in "significant habitat modification ... [that] significantly impair[s] essential behavioral patterns, including breeding, feeding or sheltering." (98) Babbitt raised the question whether the Department had the authority, under the ESA, to prevent landowners from harming endangered species by destroying their essential habitats. (99)
The majority and dissenting opinions in Babbitt dueled over a dizzying array of canons and interpretive tools, including dictionary definitions, the plain meaning rule, noscitur a sociis, the whole act rule, and legislative history. (100) Despite their impressive repartee over the canons, however, both opinions give the distinct impression of being about much more than the neutral application of statutory interpretation rules. The strong undercurrent of the conservative dissenting opinion is that environmental statutes should be construed to interfere as little as possible with private-property rights; (101) while the underlying theme of the liberal majority opinion is that species extinction poses grave threats to the environment and that the ESA's broad purpose of species preservation must trump private-property interests. (102) A recent case, Rapanos v. United States, (103) which involved a landowner's right to backfill wetlands under the Clean Water Act, reflects similar majority-dissent dueling on the surface, and a similar underlying ideological battle over private-property rights versus effective environmental preservation. The conservative plurality opinion, for example, began by bemoaning the costly "burden of federal regulation on those who would deposit fill material in locations denominated 'waters of the United States,"' calling the agency responsible for regulating wetlands "an enlightened despot," and complaining that "for backfilling his own wet fields, Mr. Rapanos faced 63 months in prison and hundreds of thousands of dollars in criminal and civil fines." (104) The liberal dissenting opinion, by contrast, emphasized the Clean Water Act's "Herculean goal of ending water pollution" and the importance of wetlands for the preservation of water quality. (105)
In a similar vein is the Burger Court case United Steelworkers v. Weber. (106) Weber involved an early application of Title VII, raising the question whether that statute should be read to bar private employers from adopting voluntary affirmative action plans designed to remedy the present effects of past discrimination against black employees. (107) The majority and dissenting opinions in Weber dueled over multiple interpretive tools, including legislative history, purpose, and statutory language. (108) Most notably, the opinions engaged in an exhaustive back-and-forth about various statements in the legislative history, often refuting each other's claims point-for-point. (109) Underlying this extensive dueling, however, was a fundamental ideological battle over affirmative action. The liberal majority opinion found it unimaginable to read a statute designed to address centuries of discrimination against black Americans in a manner that would prohibit voluntary efforts by private employers to abolish the present-day effects of that historical discrimination. (110) The conservative dissenting opinion, by contrast, found it equally odious to read the statute to allow discrimination against white employees as the means for improving black employees' opportunities. (111) Several of the Roberts Court's antidiscrimination cases reflect similar underlying judicial divides over the appropriate reach of statutes designed to protect minorities and women from employment or voting discrimination, even as the majority and dissenting opinions duel over the application of specific interpretive canons. (112)
What, then, about the Court's high rates of dueling canon use in cases involving criminal statutes? After all, the statutes at issue in the Court's criminal-law cases generally are not the product of sweeping social-justice movements. Despite this difference, criminal statutes nevertheless may reflect clear ideological dividing lines, or disagreement, over how the State should treat those accused of crimes. Again, the Justices' clear policy preferences may color the way they apply the canons and interpretive tools and make them more inclined to counter an opposing opinion's use of such tools--not as a result of longstanding ideological divisions over sweeping statutes enacted in a previous era, but because the Justices have strong views about the moral culpability of certain behavior and about how the government should punish those accused of such behavior.
Many of the Court's criminal cases, for example, involve sentencing statutes, and seem to reflect an underlying disagreement about the harshness of applying a mandatory enhancement to particular conduct. Consider Dean v. United States, (113) in which the Justices employed the whole act rule, other statutes, and common-law rules in a dueling manner in determining whether a ten-year sentencing enhancement applied to a defendant whose gun accidentally discharged while he was robbing a bank. (114) Although the majority and principal dissenting opinions dueled over the structure of the statute, the appropriate inferences to be drawn from the felony murder statute (parallel v. meaningful variation), and common-law rules about mens rea versus unintended consequences of unlawful conduct, (115) a strong policy disagreement about fairness seems to underlie these more technical disagreements. Specifically, the opinions seem to be motivated by a fundamental concern about whether it is fair (majority) or unfair (dissent) to subject a criminal defendant to an extra ten years in prison because his gun accidentally discharged while he was committing a violent crime. (116) Several of the Court's other criminal cases involve the Antiterrorism and Effective Death Penalty Act and reflect ideological divisions about the equity of cutting off death-penalty appeals. In such cases, the majority and dissenting opinions often seem, at bottom, to be clashing over different normative visions about access to the courts for death-penalty appeals. (117)
Overall, then, ideological divisions may explain the high representation of environmental, antidiscrimination, and criminal statutes among the dueling canon cases. Each of these subject areas involves a hot-button issue that tends to divide the major political parties and their supporters. As a result, statutes dealing with these subject areas may be the product of extensive compromise between legislators on opposite sides of the political divide. This in turn may give rise to (1) an abundance of legislative history reflecting competing visions of the statute's reach, (2) statutory text that is intentionally vague or susceptible to competing readings, or (3) a heightened likelihood that the Justices charged with interpreting these statutes will have individual policy preferences regarding how such statutes should be implemented. Consequently, it may be the case that statutes in these subject areas give rise to competing applications of the same interpretive tool more frequently than do statutes in other subject areas. Moreover, when construing statutes in these subject areas, the Justices may be more likely to use the canons and tools of statutory construction to justify a reading arrived at through other means, rather than as a starting point. In contrast, when construing statutes in other, less ideologically charged subject areas, the Justices may be less inclined to use the canons to counteract each other.
Another possibility raised by the subject-matter data is that the Justices tend to duel at higher rates in cases involving a statute not administered by an agency or a statute that is administered by a weak agency. This hypothesis does not perfectly explain the data reported in Table 3a, as environmental statutes are administered by the EPA, a powerful agency. (118) But lack of deference to an agency interpretation could be part of the reason why so many of the dueling canon cases involve criminal and antidiscrimination (119) statutes (and why the next largest category of dueling canon cases involves preemption and bankruptcy statutes). (120) Perhaps the Court engages in more extensive statutory interpretation in cases that do not involve a respected agency interpretation because in such cases it does not begin with a presumption of deference to the agency's construction. And perhaps when the Court engages in more extensive statutory interpretation, without the benefit of agency guidance (or with guidance from an agency to which Congress has given limited power, like the EEOC), the Justices are more inclined to duel over individual interpretive tools.
3. Dueling Canons and Ideology. While evaluating the data, it became clear that it would be useful to examine the relationship between ideology and judicial dueling, both for purposes of comparison to the Brudney-Ditslear study and because the subject-matter data suggested the possibility that judicial dueling might be more prevalent in cases with strong ideological dividing lines. Before reporting the ideology data, however, a caveat is in order: It is difficult to code case outcomes for ideology because it is not always clear whether an outcome favoring a particular litigant is liberal or conservative and coders necessarily must make judgment calls. In order to minimize errors and to make this study as replicable as possible, I coded for ideology by importing the ideological direction coding from the Spaeth Supreme Court database for the cases in my dataset. (121)
Three results stand out. First, the "dueling canon" cases (i.e., cases in which the Court invoked the same interpretive tools to support competing interpretations) skewed significantly conservative compared to all statutory cases, and somewhat more conservative than all divided-vote cases. (122) Further, the more closely divided cases were significantly more likely to inspire judicial dueling than were cases decided by a wide margin. Second, when the Justices in majority and dissenting opinions invoked the interpretive canons in a competing or dueling fashion, they tended to do so consistently with their ideological preferences. Third, and by contrast, in the unanimously decided cases the Justices tended to employ the interpretive canons to support both liberal and conservative outcomes at much more even rates. (That is, the conservative-liberal differential for most canons was much closer).
Table 4 reports the ideological direction of the majority opinion in all statutory cases, unanimous cases, divided-vote cases, and dueling canon cases in the dataset. As the Table shows, the conservative-liberal split for all statutory cases decided during the time frame of the study was almost even, while the split for unanimous cases skewed liberal by several percentage points (thirteen percentage points). By contrast, the split for cases decided by a divided vote skewed conservative by a decisive margin (over sixteen percentage points). The largest conservative-liberal differential occurred in the dueling canon cases, which skewed conservative by a whopping thirty-two percentage points.
Table 5 reports the Roberts Court's rate of dueling canon use in divided-vote cases by vote margin. As the Table shows, the overwhelming majority of the dueling canon cases were decided with six or fewer Justices joining the majority opinion (72.7 percent), and more than half of those cases were decided by extremely close vote margins, usually 5-4 or 5-3. Cases in which seven or more Justices agreed on a statute's construction represented a much smaller percentage of the dueling canon cases (27.3 percent). Moreover, a large majority of all cases decided by close vote margins contained judicial dueling over the canons and other interpretive tools. In other words, the Justices were most likely to engage the interpretive canons and tools in a dueling manner when their levels of consensus were low.
It is difficult to draw any definitive conclusions from these numbers, as they show only correlation, not causation. But a number of possible explanations are worth exploring. On the one hand, close vote margins could reflect statutory ambiguity. That is, the Justices could be dividing closely because the statutes at issue lend themselves to multiple plausible readings. In cases involving such ambiguous statutes, individual interpretive canons and tools may be particularly likely to point in multiple directions. Thus, the Justices could be dividing closely because individual canons and tools are pointing in competing directions, and they could be dueling over the canons' application because the canons point in multiple directions. Alternately, the close margin cases may involve issues that are particularly ideologically charged. The Justices could be deciding these cases based on their ideological preferences--choosing their preferred readings first, and then employing the canons in a dueling fashion in order to justify their constructions and neutralize opposing arguments.
Statistical observations cannot reveal the Justices' internal motivations for voting a particular way in a particular case--and in this context cannot tell us whether the Justices duel at higher rates in close-vote-margin cases because of statutory ambiguity, ideological preferences, or some other factor. Nevertheless, it is worth trying to understand as much as we can about how the Justices duel over individual interpretive canons when they choose to do so. In an effort to gain a more granular understanding of the Justices' dueling practices, Tables 6a-6g report the ideological direction of each individual Justice's vote in cases in which the justice authored or joined an opinion that employed a particular interpretive tool in a dueling manner. (123)
The results are striking. They reveal that when the members of the Roberts Court duel over interpretive canons or tools, they do so in a manner that strongly correlates with their ideological preferences. Many of the conservative Justices used particular canons or tools to reach conservative outcomes in over 60 percent, and even over 70 percent of the cases, while the liberal Justices used those same canons or tools to reach liberal outcomes at similarly high rates in those same cases. Strikingly, in the nineteen cases in which the Court dueled over legislative history, Justice Scalia authored or joined an opinion that used legislative history to reach conservative results 68.4 percent of the time, and liberal results 10.5 percent of the time. (124) The figures for Justice Thomas were similar, as were those for Justices Alito and Roberts. In contrast, the liberal Justices referenced legislative history to reach liberal results in the overwhelming majority (57.9 percent to 68.4 percent) of these same cases, and to reach conservative results in only 15.8-31.6 percent of the cases. (125)
The Justices' dueling over other canons and interpretive tools shows similarly stark ideological-slanting. (126)
These findings are consistent with Brudney and Ditslear's findings that conservative Justices tend to invoke the canons to reach conservative outcomes, while liberal Justices tend to invoke the canons to support liberal outcomes. (127) But these findings also deepen, or expand, on the Brudney-Ditslear study by demonstrating that the ideological slanting applies to interpretive tools other than the language and substantive canons, and by revealing that the ideological slanting is particularly forceful in the dueling canon cases. Further, the data in Tables 6a-6g suggests that the canons do not constrain the Justices on the Roberts Court to vote against their policy preferences, providing some support for Llewellyn's legal realist view that judges do not decide cases based on neutral legal rules.
The data in Table 6a also seem to contradict Brudney and Ditslear's findings on the ideologically constraining effects of legislative history. Brudney and Ditslear found that liberal Justices were more likely to vote in favor of employer interests (that is, against their ideological preferences) when they invoked legislative history, (128) and that during the Burger Court, conservative Justices were more likely to vote in favor of employee interests when they relied on legislative history. (129) After 1986, the Brudney-Ditslear study reported that conservative Justices ruled increasingly in favor of employers, even when relying on legislative history. (130) Table 6a shows, by contrast, that when the Justices dueled over legislative history, conservative Justices were highly likely to vote in favor of conservative outcomes and liberal Justices were highly likely to vote in favor of liberal outcomes. Even if we set aside the data regarding the conservative Justices' voting patterns on the theory that they are consistent with Brudney and Ditslear's findings for post-1986 cases, the data regarding the ideological direction of the liberal Justices' dueling canon use run exactly opposite to Brudney and Ditslear's findings.
But there is an interesting counterpoint to these data. Table 7 reports the correlation between canon and interpretive tool use and ideology in the Roberts Court's unanimous statutory cases. The Table does not contain entries for the individual Justices' ideological slanting because, in the unanimous cases, all of the Justices voted the same way. Thus, the ideological breakdown for each interpretive tool tells us the rate at which individual Justices invoked that tool to reach conservative versus liberal outcomes. Interestingly, the Table shows that for this subset of cases, the Justices tended to use most of the canons and interpretive tools to support both liberal and conservative outcomes at generally comparable rates. There were some exceptions--for example, dictionary definitions and legislative history tended to be used to support liberal outcomes at much higher rates than conservative outcomes, and the common law tended to be used to reach conservative outcomes much more often than liberal ones. But overall, most canons and tools were used to support conservative outcomes in roughly 40 percent to 50 percent of the cases and liberal outcomes in roughly 40 percent to 50 percent of the cases.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Supreme Court justices' different uses of legal canons in statutory interpretation; Abstract into II. Findings B. Dueling Statistics 3. Dueling Canons and Ideology, p. 909-956|
|Author:||Krishnakumar, Anita S.|
|Publication:||Duke Law Journal|
|Date:||Feb 1, 2016|
|Previous Article:||Class action myopia.|
|Next Article:||Dueling canons.|