2001 Annual Meeting
Wednesday, January 3, 2001 - Saturday, January 6, 2001
San Francisco, California
Saturday, January 6, 2001, 1:30-3:15 p.m.
Section on Legislation
Theories of the Political Process and Theories of Interpretation
Interpretation, Politics and the Fallacy of Division
Today Iíll confine myself to a single point about the relationship between theories of interpretation and theories of the political process. The point is that existing accounts of interpretive dynamics draw a false inference, called the fallacy of division. Suppose that we are advising a newly appointed judge about what approach to statutory interpretation the judge ought to adopt. The judge says the following: ďWhich account of the political process should I adopt in choosing my interpretive approach?Ē The problem with this question is that it canít be answered; it rests on a false premise. The question ought to be, ďWhich account of the political process should I adopt in choosing my interpretive approach, given that other judges will disagree with whatever account I do adopt?Ē The fact of inevitable and irreducible disagreement knocks out any interpretive approach that, in order to succeed, would implicitly require that all judges agree to adopt that interpretive approach. Any such approach commits the fallacy of division: the fallacy of supposing that what is true of a group is therefore true of the members of the group taken individually. A family may be large without any of its members being large; and while it may be good for the army to advance into enemy territory, it doesnít at all follow that itís good for Private Smith. Here, the division fallacy is to suppose that an interpretive strategy that would be beneficial if adopted by all judges should therefore be adopted by any particular judge. Whether the latter point is true depends on what the other judges are doing.
My claim is certainly not that positive and normative accounts of the political process are in principle irrelevant to the judgeís problem. In principle, normative and positive accounts of the political process may dictate, constrain or at least inform our judgeís choice of interpretive approaches. A normative account of the political process, for example, would indicate what a well-functioning process would look like and would provide an extrinsic standard for evaluating its output. A positive account would tell the judge how inputs into the legislative process are translated into outputs, as a static matter, and would help predict legislative reactions to judgesí interpretive decisions, as a dynamic matter.
But this sort of account of politics is often irrelevant in practical terms, owing to a critical feature of our judgeís institutional position. The judiciary is a they, not an it; a collective bureaucracy, not a single jurist. Judges vote on cases within a collective judicial bureaucracy that, at appellate levels, always sits on merits cases in multi-judge panels. Whatever theory of politics our judge adopts, other judges will adopt different ones. The practical consequence of the collective character of the judiciary is to produce irreducible disagreement within the judiciary about the truth of competing accounts of politics - disagreement both within panels or courts and, over time, across panels or courts. If you believe, with Dan Farber, that theories of statutory interpretation donít matter very much, perhaps because you believe with Cass Sunstein that judges with different theories often reach consensus on particular outcomes, then disagreement isnít consequential; but the price of that position is that interpretive theories generally, whether or not process-based, arenít consequential, in which case talking about the relationship between interpretive theory and the political process isnít very interesting.
So if weíre to take the relationship between process and interpretation seriously, the persistence of disagreement undermines the normative, empirical and predictive accounts of the political process embodied in the leading interpretive approaches. For a normative example, take the claim that judges should adopt information-forcing or democracy-forcing default rules: interpretive doctrines designed to encourage legislatures to reveal their preferences explicitly. These rules have become a favorite both of textualists, who argue for the primacy of text over legislative history on democracy-forcing grounds, and of process theorists, some of whom argue that canons may be used to elicit useful information from legislatures. There are related suggestions that judges may use interpretive doctrines in an experimental fashion, gauging legislative response and then readjusting doctrines as necessary. But these positions all ignore the heterogeneity of interpretive doctrine across courts and over time. If all (relevant) courts consistently eschewed legislative history, or consistently adhered to the rule of lenity, Congress might encode policy choices in text and speak clearly when imposing criminal punishments. But the federal judiciary disagrees about the merits of textualism and the rule of lenity, both within courts, such as the Supreme Court, and across courts, such as the courts of appeals; these disagreements produce shifting coalitions and mutable doctrines; so there is no consistent, unified set of incentives to which Congress might respond, and no steady course of controlled judicial experiments that might reveal valuable information.
In such a world, it does not follow that any particular judge ought to vote in the way that she would if all other judges were to vote the same way. If other courts, or other judges at different times, donít use some information-forcing canon, then using it locally or episodically may produce bad case-specific consequences without the beneficial systemic and dynamic effects that are the canonís rationale. If the justification for the rule of lenity is that it forces Congress to speak clearly to criminal punishments, but other courts arenít using it, an isolated instance of its use by one court will have no democracy-forcing effect; it will just give the particular defendant a windfall victory even though the usual interpretive sources, minus the rule of lenity, suggest that the statute covers the defendantís conduct. That might be a good thing, but it canít be justified on political process grounds.
The division fallacy also infects positive models of the dynamics of legislation. Consider the ingenious models recently advanced by sophisticated intentionalists to rebut textualist claims about legislative history. In these models, intentionalist statutory interpretation produces a self-fulfilling equilibrium: courts consider legislative history as accurate, and rational legislators anticipating the practice monitor legislative history to ensure that it does indeed accurately reflect the deal struck by the enacting coalition. Textualism might also produce a self-fulfilling equilibrium, albeit with the players coordinated on text rather than history. But if disagreement causes different judges or courts to do different things with legislative history at different times, thereís no particular reason to expect that either equilibrium will arise. At the very least weíd need a much more complicated model, in which legislators discount the probabilities of different interpretive strategies, and this sort of model would put even more strain than usual on the strong assumptions about actorsí rationality that are needed to make the equilibrium stick.
In both the positive and normative cases, the problem of disagreement canít be sidestepped by describing interpretation as a pure coordination game. Itís true that many interpretive doctrines and canons have a coordination component. Maybe weíd be better off if all judges adhered either to a rule of lenity or to a rule of severity, rather than mucking about inconsistently. But choosing canons, or choosing between textualism or legislative history, isnít really like choosing whether to drive on the left side of the road or the right; itís more like a multi-person battle of the sexes, in which it matters that all judges do the same thing, but different judges would each prefer that all judges coordinate on different rules. The sophisticated intentionalists I mentioned argue that the self-fulfilling equilibrium centered on legislative history is superior to the one centered on text, on the ground that the former better suits the organizational constraints under which legislatures act. Likewise, the rule of lenity will produce fewer convictions than the rule of severity will, and we might have extrinsic reason for preferring one to the other. These distributive consequences are precisely what create the persistent disagreement that in turn prevents even the coordination benefit of interpretive doctrines from being realized.
If all this is sensible, then the question remains whether and how political process accounts are relevant to a particular judge deciding what interpretive approach to adopt, given that other judges will disagree with her. I think the answer is that those accounts arenít very relevant, and that normative theorizing has to proceed along different lines altogether. In a collective judiciary riven by persistent disagreement over interpretation, any particular judge or even Justice just canít do very much, for good or for ill, to affect the dynamic interaction between Congress and the judiciary as a whole. The judge is like a participant in a competitive market who can choose what quantity to sell or buy, but who can do very little to affect price. Congress has always faced, and probably will always face, and interpretive regime of a mixed character, one that sends inconsistent signals over time.
So disagreement and discord within the judiciary make the dynamics of legislative and judicial interaction sufficiently opaque that political process accounts arenít very helpful to interpretive theory. On what lines, then, should normative theorizing proceed? Whatís needed are interpretive theories that take account of the collective character of the judiciary, of dissensus and disagreement. Such theories have to tell particular judges what marginal or divisible contributions they can make to the aims of interpretation - that is, how they may act in a world where coordinated judicial action is unlikely. Elsewhere Iíve outlined a theory that I think meets this condition, and if it doesnít it should. My aim today isnít to defend an interpretive theory, but to emphasize an irony of legislation as a mature field: weíve spent so much time exploring the consequences of the legislatureís collective character that weíve devoted inadequate attention to the same problem on the judicial side.