1. Leiter on Agreement in Legal Judgments

2. Can Dworkin Account for Agreement in Legal Judgments?

Leiter is clearly right that most potential legal disputes are not litigated, most disputes that result in litigation are settled before trial, etc.^{Footnote 22} I also agree that this fact is explained, in part, by the further fact that there is general agreement about the legally correct outcome of most potential legal disputes. There may be other factors at work as well—in particular, the existence of alternative methods of dispute resolution (both formal and informal) which are generally cheaper than litigation. Nevertheless, I think that Leiter is right that the existence of agreement in legal judgments is an important part of the explanation of why most potential legal disputes are not litigated (etc).

Moreover, positivism offers an explanation of this agreement in legal judgments: officials agree about the outcome of most potential legal disputes because they generally use the same tests for determining what the law is.^{Footnote 23} However, this is not the only possible explanation of agreement in legal judgments. Officials might agree on the outcome of most cases even if they fail to converge on criteria of legal validity. Consider two judges, one of whom believes that the law made by a statute is fixed by the plain meaning of the words used, the other of whom believes that it is fixed by the (perhaps counterfactual) intentions of the legislature.^{Footnote 24} Let us call the first judge a “textualist” and the second an “intentionalist”. (Nothing turns on the choice of labels; they are meant only to facilitate exposition.) These judges employ different criteria of legal validity, since they employ different tests for determining what law is made by a statute. Nevertheless, they may agree on the outcome of most cases, provided the intentions of the legislature generally coincide with the plain meaning of the words it uses.

There is reason to think that the intentions of the legislature will generally coincide with the plain meaning of the words it uses, at least in well-functioning legal systems. The words used in the statute are the primary mechanism (or at least the primary official mechanism) that the legislature has for communicating its law-making intentions. Moreover, to the extent that judges consider the plain meaning of the words used in the statute (either because they think that plain meaning is legally significant in its own right, or because they believe that it provides evidence of the legislature’s intentions), the legislature has reason to produce legal texts whose plain meaning matches its intentions.^{Footnote 25}

However, even if agreement in legal judgments can be explained without positing convergence among officials on criteria of legal validity, this does not mean that Dworkin is able to account for such agreement. As we have seen, Leiter argues that law as integrity renders such agreement puzzling, because it holds that the law consists (in part) of certain moral principles, and because members of the relevant community may fail to identify the correct principles.

That members of a particular community may fail to identify the moral principles that constitute part of their law does not necessarily mean that they will disagree among themselves. If they all make the same mistake, then they may agree about which moral principles form part of the law, even though they fail to identify the correct moral principles. This suggests that Leiter means to appeal, not to the possibility of moral error, but to the possibility of moral disagreement.^{Footnote 26} The thought seems to be that, in modern legal systems, there is widespread moral disagreement, and we should expect this disagreement to extend to the moral principles that Dworkin claims form part of the law. Moreover, since law as integrity claims that judgments about legally appropriate outcomes are derived from those moral principles, we should also expect widespread disagreement about the appropriate outcome of particular cases.

This is an important objection to law as integrity, but I do not think that it shows that Dworkin is unable to account for massive and pervasive agreement in legal judgments. Before I explain why, let me briefly address one complication. Positivism, as Leiter characterizes it, claims that, wherever there is law, officials converge on criteria of legal validity. It is natural, then, to think that the truth or falsity of positivism has implications for whether there is massive and pervasive agreement in legal judgments. By contrast, law as integrity claims (inter alia) that certain moral principles feature among the determinants of legal content. One might think that whether this claim is true has no implications for whether there is massive and pervasive agreement in legal judgments. It is whether people believe that it is true that has such implications. On this view, if people believe that law as integrity is true, then they will (presumably) search for the moral principles that it claims form part of the law, and (if Leiter is right) this will lead to widespread disagreement about the content of the law. However, whether law as integrity is true can (by itself) have no such implications.

This line of thought is, I think, misguided. On Dworkin’s view, if law as integrity is correct, this is (in part) because it fits our legal practices. Thus, like positivism, law as integrity makes claims about what officials do, accept or believe. If it is correct, this is (in part) because officials generally treat certain moral principles as part of the law, and seek to identify and apply those principles. If Leiter is right, this means that we should expect widespread disagreement in legal judgments, because officials will disagree about what the moral principles are from which those judgments should be derived.

So why do I think that Leiter is wrong to claim that law as integrity cannot account for massive and pervasive agreement in legal judgments? For one thing, we should be careful not to overstate the amount of moral disagreement there is, even in pluralistic, liberal societies. True, there is much moral disagreement, but at any given time there are also many moral issues on which there is general agreement. This is especially so if we are concerned primarily with the views of legal officials.^{Footnote 27} Many officials may have undergone similar legal training, and officials may also tend to be of a similar socio-economic status—both factors that may tend to produce a higher than normal level of agreement on moral issues relevant to law. Thus, there may in fact be considerable agreement, at least among officials, about the moral principles that form part of the law on Dworkin’s theory.

Nevertheless, there is also likely to be considerable disagreement about those principles. However, and this is the crucial point, where there is disagreement about the relevant moral principles, there need not be disagreement about the appropriate outcome of particular cases. We have seen that two people may agree about the outcome of very many cases even though they disagree about the basis on which those outcomes are to be reached. For example, two people may disagree about the correct theory of statutory interpretation (e.g., because one embraces textualism and the other intentionalism), and yet agree about the outcome of most cases involving statutory interpretation. There does not appear to be any reason why the same could not be true with regard to the moral principles that Dworkin treats as part of the law. If, in many cases, the competing moral principles accepted by different officials support the same outcome, then this renders the existence of massive and pervasive agreement in legal judgments much less surprising on Dworkin’s account.^{Footnote 28}

This may seem like an unsatisfying response. Leiter might concede that it is possible that, if law as integrity were correct, there would be considerable agreement about the appropriate outcome of particular legal disputes even where there was disagreement about the relevant moral principles. However, he might ask, is there any reason to think that this is in fact likely?^{Footnote 29}

This is a fair question. However, I think that there are features of law as integrity that make it likely that there will be considerable agreement in legal judgments, even where there is disagreement about the moral principles that form part of the law. Consider how law as integrity is meant to work. Legal interpreters seek to arrive at an interpretation of the law that shows it in its best light, by showing it to reflect a coherent (and independently attractive) conception of justice and fairness. Different interpretations of the law consist of different views about which conception of justice and fairness is reflected in the law. But any interpretation of the law must satisfy a threshold of fit—that is, it must fit sufficiently well with the political history of the community (the past decisions made by legal officials on behalf of the community) to be an eligible interpretation of that history.^{Footnote 30} This constraint ensures that there will be some level of agreement in legal judgments, since every legal interpreter shares a common set of “data points” (the previous decisions of legal officials that constitute the community’s political history) and each eligible interpretation must fit a sufficient number of those data points.^{Footnote 31}

Moreover, even where an interpretation fits enough of the data points to be eligible, fit continues to play a role.^{Footnote 32} Everything else being equal, the higher the degree of fit, the better the interpretation portrays the law as being and so the more attractive the interpretation is. Thus, regardless of where they set the threshold of fit, there is reason for legal officials to adopt interpretations that fit with a very considerable number of past legal decisions, at least in a largely just legal system. (In a less just legal system, a considerable degree of fit may need to be sacrificed to promote Dworkin’s other dimension of interpretation—namely, justification.) Again, this is likely to militate in favour of considerable agreement in legal judgments.

It might be objected that this over-estimates how much agreement about fit there would be if law as integrity were true. As Dworkin acknowledges,^{Footnote 33} legal interpreters will disagree about where to set the threshold of fit—that is, about how well an interpretation must fit the community’s political history in order to be eligible—as well as about how to trade off the dimensions of fit and justification when choosing among eligible interpretations. All of this is likely to lead to disagreement about the appropriate outcome of particular cases, since it is likely to lead to different interpretations having different degrees of fit with the community’s political history.

Nevertheless, it remains the case that there is also likely to be considerable agreement about legally appropriate outcomes among eligible interpretations of the law. Even if legal interpreters disagree about where to set the threshold of fit, they all agree that there is a threshold. Perhaps more importantly, they also agree that, even when this threshold is surpassed, the more fit there is, the better (everything else being equal).

Moreover, I suggested above that we should be careful not to exaggerate the amount of moral disagreement there is among legal officials, especially given that many of them are likely to share a particular form of legal training. For similar reasons, we might plausibly speculate that, in familiar legal systems, there is likely to be a significant degree of convergence in terms of where officials set the threshold of fit and how much weight they attach to fit when balancing it against justification. While there are undoubtedly differences between officials, we might think that judges, in particular, are likely to require a high degree of fit before an interpretation counts as eligible, and to give considerations of fit a relatively high weight when choosing between eligible interpretations. This is, in part, because judges are likely to have internalized (though perhaps to different degrees) the view that they should be very significantly constrained by previous legal decisions (respecting legislative supremacy subject to any constitutional limitations, being hesitant to overturn settled common law doctrine, and so on).^{Footnote 34}

In fact, even if judges disagreed significantly about where to set the threshold of fit, or about how to balance the dimensions of fit and justification when they conflict, it would still be likely that there would be considerable agreement about legally appropriate outcomes. This is because most judges are likely to regard their legal system as legitimate, and so are unlikely to believe that dramatic sacrifices of fit are needed to arrive at an interpretation of the law that scores well along the dimension of justification.^{Footnote 35} (That they believe their legal system to be legitimate may also help explain why judges generally attach great weight to legislative supremacy and are hesitant to overturn settled common law doctrines.) All of this suggests that we should expect considerable agreement in legal judgments, since—while different officials hold different interpretations of the law—most of those interpretations (and certainly the more popular ones) are likely to fit with a considerable proportion of the data points referred to above.

Of course, this is speculative. However, it does suggest that Leiter underestimates the resources available to law as integrity to explain agreement in legal judgments. Since competing interpretations of the law are rival attempts to fit and justify the same body of previous legal decisions, it should not be surprising that there are many instances in which the competing interpretations support the same legal outcome, leading to considerable agreement in legal judgments.

Leiter might respond that, while there might be considerable agreement in legal judgments if law as integrity were true, there would not be massive and pervasive agreement. Even if different interpretations of the law often overlap in terms of the particular legal outcomes they support, there will still be many cases where they do not overlap (i.e., where the different moral principles that they regard as part of the law have different implications for the appropriate resolution of particular legal disputes). Thus, Leiter might readily concede that law as integrity can account for a considerable amount of agreement in legal judgments, while maintaining that it cannot account for the full extent of the agreement that exists in well-functioning legal systems.

To assess this response, one would need to be much more precise than either Leiter or I have been, in terms of quantifying both how much agreement in legal judgments there is and how much agreement law as integrity can account for. I shall not seek to undertake this task here. Quantifying how much agreement in legal judgments there is in modern legal systems would require a careful empirical investigation. However, if I have succeeded in showing that Dworkin can account for considerable agreement in legal judgments, in ways that Leiter does not adequately take into account, one might reasonably think that the onus is on Leiter to show that this still falls short of explaining the full extent of the agreement that exists in modern legal systems.

That said, I do want to suggest that the evidence Leiter offers in support of his claim that there is massive and pervasive agreement in legal judgments is more equivocal than it may first appear. Recall his reason for concluding that there is massive and pervasive agreement—namely, that this explains why very few potential legal disputes are litigated, why most litigation is settled before trial, and so on. He claims, I think rightly, that the existence of agreement in legal judgments is one important reason for this relative lack of litigation. However, I suggested earlier that there may be other reasons as well, such as the existence of formal and informal alternatives to litigation—most of which tend to be cheaper, simpler and quicker than litigation.^{Footnote 36} Thus, one cannot assume that, because the overwhelming majority of potential legal disputes are not litigated, there is agreement about the appropriate outcome of the overwhelming majority of such disputes. Rather, to ascertain the amount of agreement in legal judgments there is, one would need to ascertain what contribution other factors make to the relative lack of litigation. Again, I cannot undertake this task here, but it is worth noting that, the greater the contribution made by other factors, the easier it will be to show that law as integrity can account for the full extent of agreement in legal judgments.

3. Does Positivism Offer a More Straightforward Explanation of Agreement in Legal Judgments?

As we have seen, Leiter’s official position is that law as integrity cannot account for massive and pervasive agreement in legal judgments. Such agreement, he claims, would be puzzling if law as integrity were correct. I have argued that he is wrong about this (or, more cautiously, that much more would need to be done to vindicate his claim). However, at times, he appears to offer an importantly different objection—namely, that, while law as integrity can account for agreement in legal judgments, its explanation is less straightforward than positivism’s. In this Section, I consider this alternative objection.

In the course of responding to the ATD, Leiter acknowledges that Dworkin might seek to account for agreement in legal judgments along roughly the lines I have suggested—namely, that different interpretations of the law frequently overlap in terms of the concrete results they support.^{Footnote 37} Leiter offers several responses to this suggestion, one of which is that:

It is not entirely clear how this passage should be understood, but the idea seems to be that law as integrity offers a less straightforward explanation of agreement in legal judgments, because it requires us to hypothesize that there are theoretical disagreements lurking hidden beneath the surface agreement in legal judgments. By contrast, positivists have no need to entertain such a hypothesis, since they explain the massive and pervasive agreement in legal judgments on the basis that there is agreement (or convergence) among officials on the criteria of legal validity.^{Footnote 39}

However, it is misleading to claim, in this context, that “agreement is agreement”, even if we leave to one side the possibility that the convergence that positivism posits among officials may not be best thought of in terms of agreement. Agreement about the appropriate outcome of legal disputes is not the same as agreement (or convergence) on the criteria of legal validity, and it is not clear why we should regard the latter as representing the most natural or straightforward explanation of the former. Just as Dworkin imputes to agreement in legal judgments an underlying disagreement about the correct interpretation of the law, positivists impute to agreement in legal judgments an underlying convergence on the criteria of legal validity. Both sides engage in “imputation”, and we have not yet been given any reason to think that one imputation is more plausible than the other.

To put the point differently, it is misleading to suggest that Dworkin’s explanation of agreement in legal judgments involves positing widespread disagreement about the correct interpretation of the law that is “hidden” by the surface agreement. The existence of agreement in legal judgments (i.e., agreement about the outcomes of particular cases) does not, by itself, suggest that there is agreement about the criteria of legal validity, such that Dworkin has to claim that disagreement about these criteria is widespread despite appearances. Convergence on criteria of legal validity is, on Leiter’s view, something that positivists posit, in order to explain what is apparent (namely, agreement in legal judgments).

Perhaps there is a better way of understanding the suggestion that positivism provides a more straightforward explanation of agreement in legal judgments. It might be argued that positivism’s explanation follows directly from a central tenet of that theory (namely, that there is convergence among officials on the criteria of legal validity), whereas Dworkin needs to add an additional hypothesis (namely, that the different interpretations of the law overlap in many cases) to the central claims of his theory. For this reason, it might be thought, positivism offers a simpler explanation of agreement in legal judgments than does law as integrity, and so (everything else being equal) its explanation is preferable.^{Footnote 40}

This is an interesting suggestion, but I doubt whether it shows that positivism’s explanation of agreement in legal judgments is preferable to law as integrity’s. For one thing, it is not clear that Dworkin does need to add an additional hypothesis to the central claims of his theory in order to explain agreement in legal judgments. As we have seen, it is central to law as integrity that any eligible interpretation of the law must fit sufficiently well with the political history of the community, and that—everything else being equal—the better an eligible interpretation fits with that history, the more attractive the interpretation is. To the extent that Dworkin’s explanation of agreement in legal judgments follows from these claims, that explanation does not require him to formulate any additional hypothesis.

Nevertheless, it might be argued that these claims do not suffice to enable Dworkin to account for massive and pervasive agreement in legal judgments. On this view, it is also necessary that most legal officials: 1) operate with a very high threshold of fit (such that, to be eligible, an interpretation of the law must fit with the overwhelming majority of the data points referred to in Section 2), 2) attach a great deal of weight to considerations of fit when weighing this dimension against the dimension of justification (so that fit with the community’s political history is rarely sacrificed in order to promote other considerations of political morality), and/or 3) believe that the dimensions of fit and justification rarely push in different directions (so that there is not normally any pressure to sacrifice fit in order to promote justification). If this is right, then Dworkin does need to add one or more additional hypotheses to the central claims of his theory in order to account for massive and pervasive agreement in legal judgments.

Even if this is so, it is still not clear that positivism enjoys an advantage over law as integrity in this respect. Consider again the explanation of agreement in legal judgments that Leiter ascribes to positivists. The claim is that legal officials employ the same criteria of legal validity, and this explains why they agree about the legally appropriate outcome in the overwhelming majority of cases. However, the mere fact that two people employ the same tests for determining what the law is does not necessarily mean that they will arrive at the same outcomes. Whether they do so depends also on whether they apply those tests in the same way. It seems, therefore, that positivists, too, need to add an additional hypothesis to the central claims of their theory (i.e., that, not only do officials employ the same tests of legal validity, but they apply those tests in the same way in the overwhelming majority of cases), in order to account for massive and pervasive agreement in legal judgments.

Of course, it is open to Leiter to argue that the additional hypothesis required by positivism is more plausible than the one required by law as integrity. However, this is a different argument from the argument that positivism provides a simpler explanation of agreement in legal judgments. Moreover, to the extent that it proceeds by trying to show that Dworkin’s hypothesis is implausible, it is likely to lead us back to many of the considerations canvassed in Section 2.

For these reasons, it is doubtful whether positivism provides a more straightforward explanation of agreement in legal judgments than does law as integrity. To be fair, Leiter does not rest content with the suggestion that it does. He offers further grounds for rejecting the claim that underlying the massive and pervasive agreement in legal judgments is widespread disagreement about the criteria of legal validity. I consider some of those grounds in the next two Sections; my aim in the present Section has simply been to address one of the bases on which Leiter might argue that positivism offers a better explanation of agreement in legal judgments than does law as integrity.

4. Explaining Theoretical Disagreement

I have argued that law as integrity can account for agreement in legal judgments, and that its explanation is not less straightforward than positivism’s. This has significant implications for Leiter’s response to the ATD, which involves contending that positivism can not only account for Dworkin’s examples of theoretical disagreement, but can do so better than law as integrity.

Leiter begins his response to the ATD by suggesting that positivists have two alternative accounts of theoretical disagreement.^{Footnote 41} On either account, while theoretical disagreements appear to be disputes about what the law is, or more precisely about what the criteria of legal validity are, and hence appear inconsistent with positivism’s claim that there is convergence on the criteria of legal validity, they are in fact disputes about what the law should be, in circumstances where the existing law does not dictate a uniquely correct answer.

The first account, which Leiter labels the “Disingenuity” account, explains theoretical disagreements on the basis that the judges who participate in those disagreements are disingenuous. They realize (either consciously or subconsciously) that there is no settled law in the case before them, but present their disagreement as being about what the law already is. The second account, which Leiter labels the “Error Theory” account, explains theoretical disagreements on the basis that the judges who participate in those disagreements are simply mistaken. They honestly believe that there is a fact of the matter about what the law is in the case before them, but the absence of convergence on criteria of legal validity that would govern that case means that they are wrong about this.

Leiter then asks whether these explanations of theoretical disagreement are preferable to Dworkin’s. To ascertain which explanation is preferable, he suggests, we should appeal to the same theoretical virtues (such as simplicity, consilience and conservatism) that we use to judge rival scientific theories.^{Footnote 42} Leiter pays particular attention to the virtue of consilience, which he characterizes in the following terms:

As we have seen, Leiter regards agreement in legal judgments as a particularly important kind of datum, and he believes that only positivism can account for it. He argues that the fact that positivism’s explanations of theoretical disagreement are part of a theory of law that can account for agreement in legal judgments “may be [the] primary virtue [of those explanations].”^{Footnote 44} By contrast, he contends, Dworkin’s explanation of theoretical disagreement is part of a theory of law that cannot account for agreement about the law.

Thus, if I am right that Dworkin does have a plausible explanation of agreement in legal judgments, then this deprives Leiter of an important strand of his response to the ATD. If both positivism and law as integrity can account for agreement in legal judgments, then Leiter cannot point to positivism’s ability to explain such agreement to support his claim that positivism’s explanation of theoretical disagreement is more consilient than Dworkin’s.

Another important part of Leiter’s strategy for responding to the ATD is to downplay the frequency with which theoretical disagreements occur. He contends that “theoretical disagreements about law represent only a miniscule fraction of all judgments rendered about law, since most judgments about law involve agreement, not disagreement.”^{Footnote 45} He draws two important conclusions from this contention. First, he concludes that Dworkin places too much emphasis upon the ability of a theory of law to account for theoretical disagreement, given that the amount of such disagreement is insignificant when compared to the massive and pervasive agreement in legal judgments.^{Footnote 46} Second, he concludes that positivism’s Error Theory account of theoretical disagreement does not encounter the problem faced by global error theories—namely, the problem of explaining “why a particular discourse persists when all its judgments are false.”^{Footnote 47} This is because positivism ascribes error to legal officials only in those relatively infrequent cases where they engage in theoretical disagreement, and so the error arises only at “the margins” of legal practice.^{Footnote 48}

Both of these conclusions depend, however, on Leiter’s claim that theoretical disagreement is relatively rare, “since most judgments about law involve agreement, not disagreement.”^{Footnote 49} By now, it should be clear why this claim is problematic. It follows from the fact that there is massive and pervasive agreement in legal judgments that disagreement about the appropriate outcome of particular cases is relatively uncommon, but it does not necessarily follow that theoretical disagreement is relatively uncommon. Whether it follows depends on whether the best explanation of the existence of massive and pervasive agreement in legal judgments is that there is convergence on criteria of legal validity—in other words, it depends on whether positivism’s explanation of agreement in legal judgments is correct.

This would not be a problem if positivism was the only theory of law that could account for agreement in legal judgments, but—as we have seen—this is not the case. Dworkin also has an explanation of this phenomenon. Moreover, if it is an open question whether positivism or law as integrity better accounts for agreement in legal judgments, then it is also an open question as to how common theoretical disagreement is. If Dworkin’s explanation is correct, massive and pervasive agreement in legal judgments is compatible with widespread theoretical disagreement, because different interpretations of the law often support the same outcome. This calls into question Leiter’s contention that Dworkin overstates the importance of theoretical disagreement as a phenomenon for theories of law to explain, at least to the extent that this contention is based on Leiter’s assumption that theoretical disagreement is relatively uncommon.^{Footnote 50} It also calls into question Leiter’s claim that positivism’s Error Theory account of theoretical disagreement does not face the problem that global error theories face, since it can no longer be assumed that legal officials make the error in question only in relatively rare cases.

In this Section, I have considered only two strands of Leiter’s multi-layered and challenging response to the ATD. (I consider a third strand in Section 5.) Thus, the brief remarks in this Section cannot be regarded as a refutation of his response, which is itself only one of several ways in which positivists might respond to the ATD. Nevertheless, these remarks are sufficient to show that, if Dworkin can account for agreement in legal judgments, this not only deprives Leiter of one of his main objections to law as integrity, but also casts some doubt on his response to Dworkin’s objection that positivism cannot account for theoretical disagreement.

5. Choosing between the Competing Explanations of Agreement in Legal Judgments

Of course, the fact that law as integrity has an explanation of agreement in legal judgments is not the end of the story. It may be that the positivist’s explanation of this phenomenon is preferable to Dworkin’s. To ascertain whether this is so, we need to consider how we should choose between the competing explanations.

As we saw in Section 4, Leiter confronts an analogous question with regard to competing explanations of theoretical disagreement, since he believes that both positivism and law as integrity can account for that phenomenon. His answer is that we should judge the competing explanations in the same way that we judge rival scientific theories—that is, by reference to theoretical virtues such as simplicity, consilience and conservatism.^{Footnote 51}

Leiter would deny that we need to appeal to these virtues to adjudicate between competing explanations of agreement in legal judgments, because he denies that law as integrity can account for this phenomenon.^{Footnote 52} However, if (for the reasons already given) we regard Leiter as mistaken on the latter point, then it might be thought to make sense to appeal to the theoretical virtues to adjudicate between the competing explanations of agreement in legal judgments, too. In the first part of this Section, I explore how this might work, before considering (in the second part of the Section) whether there are alternative ways of adjudicating between the competing explanations.

In Section 3, I examined whether positivism provides a simpler explanation of agreement in legal judgments than does law as integrity. I begin the present Section by considering the virtue of consilience, Leiter’s characterization of which was presented in Section 4. This requires us to consider whether positivism’s and law as integrity’s explanations of agreement in legal judgments also make sense of other legal phenomena.

Dworkin might argue that one advantage of his explanation of agreement in legal judgments is that it is part of a theory of law that makes sense of the phenomenon of theoretical disagreement. As a version of interpretivism, law as integrity can allow that officials disagree about the appropriate justification of their legal practices, and about the best way to adjust their pre-interpretive understanding of those practices in light of the relevant justification. Even where two officials both accept law as integrity, they may disagree about which moral principles form part of the law, and this disagreement may extend to fundamental principles such as those that constitute the doctrines of legislative supremacy and precedent. All of this leaves much scope for theoretical disagreement. By contrast, positivism struggles to account for theoretical disagreement because it explains agreement in legal judgments by positing convergence among officials on criteria of legal validity.

One of the interesting features of Leiter’s response to the ATD is that he seeks to turn this dialectic on its head. Consider Riggs v. Palmer,^{Footnote 53} which is one of Dworkin’s examples of theoretical disagreement. The judges in Riggs disagreed about whether a beneficiary was precluded from inheriting under a will, given that he murdered the testator (his grandfather) in order to inherit. The Statute of Wills did not explicitly address this situation, and the grandfather’s will satisfied all of the explicit requirements set down in the Statute. Leiter concedes that the judges in Riggs appeared to disagree about which test they should use to determine if the grandson could inherit. More precisely, he says, they appeared to disagree about whether they should look to the plain meaning of the Statute or to the counterfactual intentions of the legislators had the legislators considered the application of the Statute to the present situation.^{Footnote 54} Judge Earl, who wrote the majority judgment, held that the grandson could not inherit, in part because the legislators would not have intended that he inherit had they contemplated this possibility. By contrast, Judge Gray, who wrote the dissent, looked to the plain meaning of the Statute, which (he held) required that the grandson be allowed to inherit despite the fact that he killed his grandfather in order to do so.

However, Leiter maintains that the disagreement in Riggs should not be taken at face value. Rather than there being a fundamental disagreement between the judges about the correct approach to interpreting statutes, the judges in Riggs were opportunistic in their choice of a theory of statutory interpretation. When we consider his opinions in other cases, there is reason to believe that Judge Earl embraced a version of intentionalism in Riggs in order to support his preferred result in that case, rather than that choice “reflecting his deep theoretical commitments about interpretation[.]”^{Footnote 55} Similarly, Leiter points out that Judge Gray adopted a very different approach to statutory interpretation in other cases (including a case decided around the same time as Riggs), trading his textualism for a version of intentionalism.^{Footnote 56} Based on his decisions in other cases, Judge Gray’s “opportunistic literalism in Riggs in all likelihood has more to do with his ideological opposition to state interference with property rights than with a considered view of statutory interpretation.”^{Footnote 57}

Once we realize the true nature of the disagreements in cases like Riggs, Leiter claims, we can see that positivists have a perfectly satisfactory explanation of those disagreements. While judges generally converge on the same criteria of legal validity, that convergence breaks down in certain exceptional cases (e.g., where it would lead to results that some judges regard as absurd or repugnant). Precisely because these cases are not governed by criteria of legal validity on which officials converge, judges are required to make new law in order to resolve them (and, because judges may not realize that this is what they are doing,^{Footnote 58} or because they may be disingenuous,^{Footnote 59} they may appear to engage in theoretical disagreements when they disagree about the appropriate resolution of those cases). Moreover, in these cases, but only in these cases, judges tend to act opportunistically, using whichever purported criterion (e.g., textualism or intentionalism) supports their preferred outcome in the case before them. In this way, Leiter suggests, positivists can explain theoretical disagreements in a way that is consistent with their explanation of agreement in legal judgments. They can also explain why judges who decide cases involving theoretical disagreement often appear to adopt a contrary position on other occasions (such as Judge Gray’s embrace of textualism in Riggs and a version of intentionalism in other cases).

If Leiter is right that the judges in Riggs were acting opportunistically, then it is Dworkin who will struggle to explain the disagreement in this case. It now seems implausible to claim, as Dworkin does, that the disagreement in Riggs was “pivotal”—that is, that it was a disagreement, “not only about whether [the grandson] should have his inheritance, but about why any legislative act, even traffic codes and rates of taxation, impose the rights and obligations everyone agrees they do[.]”^{Footnote 60} If so, Dworkin cannot infer from the existence of theoretical disagreement in Riggs that there is latent theoretical disagreement in other cases as well. And this weakens his case that judges frequently act on conflicting interpretations of the law, though those interpretations produce the same outcomes with regard to many potential legal disputes. To avoid these unpalatable conclusions, it seems that Dworkin would have to ignore the way that the judges in Riggs decided other relevant cases—that is, he would have to offer an explanation of the disagreement in Riggs that is less consilient than positivism’s. Leiter concludes from this that, not only does positivism have an explanation of theoretical disagreement, but its explanation is preferable to Dworkin’s.^{Footnote 61}

This is an ingenious and important response to the ATD. It differs in important respects from a response anticipated by Dworkin—namely, that judges simply pretend to engage in theoretical disagreement.^{Footnote 62} Nevertheless, it might be thought to be subject to a similar objection. If, in certain disputed cases, judges are opportunistic in the way Leiter claims, would we not expect this opportunism to be exposed for what it is? The losing party in such cases would, it seems, have an incentive to do so. And, while the losing party may not themselves have the necessary expertise—it may require careful analysis of the decisions made by a particular judge, of a sort not easily undertaken by laypeople, to expose her opportunism—we might expect the losing party sometimes to have access to people (such as lawyers) who do have the necessary expertise.^{Footnote 63} There may also be interested observers (such as legal academics) who possess that expertise. It might seem strange, then, that legal practice appears to continue as if theoretical disagreements were genuine disputes about the criteria of legal validity.

There are possible responses open to the positivist. Perhaps lawyers and legal academics have nothing to gain from exposing judicial opportunism, though it seems doubtful that this would always be true. More plausibly, perhaps some lawyers and legal academics have sought to expose judicial opportunism (think of the Legal Realists), but have been largely ignored by the judges themselves. However, some explanation is needed of how judges have gotten away with ignoring those members of the profession who have spoken out.^{Footnote 64}

For these reasons, one might harbor doubts about whether positivism can make sense of theoretical disagreement along the lines Leiter suggests. However, Leiter is right to point out that, even if there are problems with positivism’s explanation of theoretical disagreement, that explanation might still be preferable to Dworkin’s.^{Footnote 65} If Dworkin rejects Leiter’s claim that the judges in Riggs were opportunistic, what alternative explanation can he provide of the apparent mismatch between the approach they took in Riggs and the approach they took in other cases?

Here is one alternative explanation. In presenting his theory of law, Dworkin makes reference to Hercules, “an imaginary judge of superhuman intellectual power and patience who accepts law as integrity.”^{Footnote 66} Hercules is able to arrive at an interpretation of the law that reveals it to be fully coherent in principle. However, real-life judges—including those in Riggs—are not Hercules. According to Dworkin, they strive to render the law fully coherent in principle. However, they have only mixed success in doing so. Sometimes, they struggle to find a way to reconcile their various substantive and theoretical commitments, potentially causing them to toy with different theories of statutory interpretation in different cases. Like anyone else, they may also fail to live up to their own aspirations. They may sometimes depart from what integrity requires because they feel strongly about the substantive merits of a case, while acknowledging that the inconsistencies in the law that this produces are undesirable.^{Footnote 67} All of this gives rise to a pattern of decision-making that may sometimes appear to be opportunistic, but in fact represents a flawed but genuine attempt to arrive at the best overall interpretation of the law.

This explanation of the appearance of judicial opportunism in cases like Riggs requires Dworkin to concede that judges sometimes depart from the requirements of integrity, and that they sometimes struggle to ascertain what integrity requires of them. However, these are concessions that he has already made. He acknowledges that real-life judges are not Hercules, and that they are likely to operate with much more incomplete and partial interpretations of the law.^{Footnote 68} He also acknowledges that our “legal structure” often violates integrity in various ways, though he insists that legal actors regard such violations as wrongful.^{Footnote 69} Thus, the need to explain away the appearance of judicial opportunism in cases like Riggs does not require Dworkin to make any new concession.

Obviously, much more would need to be said before one was in a position to conclude that Dworkin provides the better explanation of the disagreements in cases like Riggs, after all. However, I shall leave the discussion of theoretical disagreement here. When judging the consilience of their competing explanations of agreement in legal judgments, we should not focus only on the ability of positivism and law as integrity to account for theoretical disagreement. There are many other legal phenomena that a theory of law should also make sense of. Leiter mentions several that he thinks law as integrity cannot account for, while positivism can. For example, he contends that positivism offers a general theory of law, whereas law as integrity “can only make sense … of legal systems whose institutional history falls above the threshold required for moral justification of that legal system to be possible[.]”^{Footnote 70} This is a version of the familiar, and important, objection that Dworkin cannot account for evil legal systems, given the role that moral considerations play in constituting the law on his account.^{Footnote 71}

On the other hand, theoretical disagreement may not be the only legal phenomenon that positivism struggles to explain. For example, there is a well-known problem (or set of problems) for positivists in accounting for law’s normativity, given their claim that the existence and content of the law depends (ultimately) on social facts alone.^{Footnote 72} There is considerable disagreement about how this “normativity problem” should be understood, but—at a minimum—there appears to be a puzzle in explaining how the law can purport to create obligations (whether moral, legal or both) if the existence and content of the law ultimately depends only on social facts. Is the law (or, if you prefer, are legal officials) committed to claiming that it (they) can derive an “ought” from an “is”? If so, this would be a puzzling feature of law, to say the least.^{Footnote 73}

This is obviously far from an exhaustive list of the challenges, in terms of consilience, facing both positivism and law as integrity. If we take the approach to theory choice that Leiter utilizes when responding to the ATD, and apply it when assessing the competing explanations of agreement in legal judgments offered by positivism and law as integrity, we would need to assess each of these challenges before we could reach a final conclusion as to which explanation is preferable. (We would also need to consider other theoretical virtues, such as conservatism, which I have not discussed.)

I shall not seek to undertake this task here. Instead, I want to suggest that any claim that this represents an appropriate way of choosing between the competing explanations of agreement in legal judgments itself requires defence. When he suggests appealing to the theoretical virtues to choose between the competing explanations of theoretical disagreement, Leiter provides little in the way of argument to support the suggestion. It may seem that little in the way of argument is needed—appeal to considerations of consilience, simplicity and conservatism may appear to provide an obvious and helpful way of arbitrating between competing theories of law.^{Footnote 74} Yet this approach would be rejected by some of the key protagonists in the debate between positivism and law as integrity.

Let me explain. We have seen that Leiter motivates his appeal to the theoretical virtues on the basis that we should judge competing explanations of legal phenomena in the same way that we judge rival scientific theories. However, this approach would be rejected by influential positivists such as Joseph Raz, who distinguishes between classifications or concepts “introduced by academics for the purpose of facilitating their research or the presentation of its results” and concepts “entrenched in our society’s self-understanding”.^{Footnote 75} Appeal to the theoretical virtues is, Raz contends, appropriate when assessing the former type of classification or concept, but not the latter, and law belongs in the latter category.

Dworkin would also reject Leiter’s methodology, though for a different reason. In Law’s Empire, he allows that scientific practice, like legal practice, could be said to be interpretive.^{Footnote 76} However, if we do say this, we must recognize that science is a different type of interpretive endeavour from constructive interpretation (of which law is an instance). They differ, says Dworkin, in terms of their standards of success. In science, what counts as a successful explanation is judged by reference to standards that include the theoretical virtues to which Leiter refers. By contrast, constructive interpretation employs standards of fit and justification.^{Footnote 77} On this view, too, it would be inappropriate, when choosing between competing theories of law, or competing explanations of legal phenomena, to appeal to the theoretical virtues that we use to judge rival scientific theories.^{Footnote 78}

The methodological divide between Dworkin and Leiter affects not only how we choose between competing explanations of legal phenomena, but also what it is to explain legal phenomena in the first place. In Section 2, I sought to show that Dworkin can play Leiter’s game, so to speak, in the sense that he can offer an explanation of agreement in legal judgments that makes sense of that phenomenon. Here, “making sense of” the phenomenon involves telling a plausible story about why we might expect there to be massive and pervasive agreement in legal judgments, were law as integrity correct. However, while Dworkin could join Leiter in saying that we seek explanations of legal phenomena that make sense of those phenomena, he would offer a very different account of what this involves. For Dworkin, it involves offering an interpretation that both fits and justifies the relevant aspects of legal practice (or, perhaps, offering an interpretation that meets the threshold of fit, and which can then be compared with other eligible interpretations along the dimensions of fit and justification). It is, therefore, at least in part, an exercise in political morality.^{Footnote 79} On this approach, if we were to ask whether positivism and law as integrity can make sense of agreement in legal judgments, we would be engaging in a different type of inquiry from the one Leiter recommends.

Of course, the fact that Leiter’s methodology for choosing between theories of law is controversial does not necessarily mean that it is mistaken. It does, however, mean that it requires defence. If one were to employ that methodology to adjudicate between the competing explanations of agreement in legal judgments offered by positivism and law as integrity, one would need to explain why that methodology is the appropriate one to use for this purpose. Moreover, simply pointing out that Leiter’s methodology has the virtue of bringing our account of theory choice in law into line with our account in the sciences would not be sufficient, given that we have seen that both Dworkin and Raz deny that this is a virtue.^{Footnote 80}

This is important because different methodological options may well favour different theories of law.^{Footnote 81} While I have suggested that Dworkin’s explanation of agreement in legal judgments fares surprisingly well under Leiter’s methodology, it may fare considerably better under Dworkin’s own methodology. For example, part of the attraction of law as integrity, according to Dworkin, is that it does a better job of portraying our legal practices as morally legitimate, because it shows how those practices give rise to political obligations on the part of the law’s subjects.^{Footnote 82} This sort of argument makes sense if we should choose between theories of law, in part, by reference to the dimension of justification. However, it does not make sense if we should choose between theories of law by reference to the theoretical virtues. While, on Leiter’s methodology, we might (perhaps) ask which theory of law makes better sense of people’s beliefs about the moral legitimacy of their legal practices, this is a different question from the one Dworkin asks.