Would love comments:

I'm delighted to report that my new article, Choosing Interpretive Methods: A Positive Theory of Judges and Everyone Else, is going to be published in the June 2008 issue of the NYU Law Review. Here's the abstract:

In this Article, I propose a theory of how rational, ideologically motivated judges might choose interpretive methods, and how rational, ideologically motivated laymen—legislators, litigation organizations, lobbyists, scholars, and citizens—might respond. I assume that judges not only have ideological preferences but also (perhaps merely strategically) want to write plausible opinions. As a result, if judges decide to use any particular method of statutory or constitutional interpretation, the plausibility demands of the method they use will make them deviate from their own ideal points in the direction of the "most plausible point" of that method.

When judges can choose their interpretive method, they select the method that (taking these deviations into account) comes as close as possible to their favored outcome. This creates a selection bias, which makes interpretive methods' observed distributions differ systematically from their true distributions. This bias explains how one can favor mandating an interpretive method even though one is politically closer to the current practitioners of a different method.

Judges can also choose whether to use the same method from case to case. I explain why, even though ideologically motivated judges (or litigation groups) might want to make the method they prefer in most cases mandatory for everyone, they do not personally have much effect on whether other judges use that method, and so it is rational for them to deviate from that preferred method in those cases where they prefer a different method.

It's going to be quite a while before this appears in print, so comments are welcome. The paper is available here.

curious about submission processes:
through which channels do you and your co-conspirators normally submit articles for publication?

specifically, do you (1) target a select group of journals and submit via transparent, online submission forms/protocols, (2) leverage faculty contacts at prestigious schools to secure a "private placement," or (3) submit the article via ExpressO (or similar service) to the broader law review market?
9.13.2007 10:38am
Sasha Volokh (mail) (www):
I use ExpressO (or the law reviews' own online submission forms, which are basically equivalent -- some law reviews prefer it that way), submitting to the "top x" (where x can vary). I wish I had faculty contacts at prestigious schools to secure a private placement!
9.13.2007 10:52am
Without having read the whole thing (in fact having skipped to the practical consequences for judges and litigators), I agree that this advances the ball beyond simple legal realism. It may have been revolutionary to say that judges just decide cases according to their policy preferences, but it's at odds with the reasoned decisionmaking we see in most appellate and Supreme Court decisions. I look forward to a full read, and as a VC contributor, thanks for the FN1 kudos!
9.13.2007 12:22pm
Bruce McCullough (mail):

I think there is a typo in footnote 114.

For the first term of x* you have
d / (d + (1-d)c)j

According to the standard rules for operator precedence, in which multiplication occurs before division, "j" is in the denominator when clearly it should be in the numerator. So rewrite as
(d / (d + (1-d)c))j or better
dj / (d + (1-d)c)

Similarly for the second term, where your numerator should be (1-d)cb

Then in the simplification of x* to involve eta (but let me call it "n") I get the first term nj but not the second term (1-n)b

Doublecheck my work and email me to discuss if you want.

Additionally, Figure 5 has its ordinate labeled "Distribution" which is wrong. Figures 6 and 7, 10 and 11, 13 and 14, have their ordinates labeled "Distributions" which is also wrong.


9.13.2007 1:07pm
1. The paper makes plausible arguments, and quite well. (For all that my opinion counts.)

2. There is a transposition of two words somewhere fairly early in the paper. Sorry, I can't remember where.

3. Hire Jim Steranko to redo the graphics.
9.13.2007 1:18pm
Sasha Volokh (mail) (www):
Bruce McCullough: I see what you mean with the numerator and denominator; I used the space to substitute for parentheses and brackets, but I might reconsider. But explain why "distribution" is incorrect as an axis label.
9.13.2007 1:44pm
DJR, could you name a few "legal realists" who believe that "judges just decide cases according to their policy preferences"? I'd sure love to see them. Or am I correct in assuming you've never read anything written by philosophers who are legal realists, but have instead relied on inaccurate characterization of their position?
9.13.2007 1:55pm
An interesting paper that is far, far too long. Do you really need 8 pages to introduce an article that is 58 pages in length?

In my opinion, you should eliminate most of the graphs. They're not helpful and will look horrible on westlaw. Your chart on page 20, for example, seems like it was just thrown in their to make a simple point into some abstract math formula. Most readers are just going to gloss over them. Quite frankly, once you start substituting in variables instead of just explaining your concept cleanly makes it impossible to quickly read. If something can't be read quickly, you're going to lose most legal readers. I mean, look at this paragraph on page 24:

Then the distribution of political preferences from Figure 5 becomes a distribution of outcomes under theory T as depicted in Figure 6 below. The right-hand star in Figure 6 is your value of J, which we have been using all along. Theory T skews your result from the right-hand star to the left-hand star; but every other J also experiences a similar skew. The entire distribution of J is squeezed toward MT.

Is that really the clearest way to express your point?

I mean, pages 20 through 24 seem to be trying to make a simple point, but it is masked in charts and graphs. Just say what you mean.

Here's another paragraph on page 30:

The true distributions of T and I—what results would look like if everyone were constrained to use a single method—are shown in Figure 7140 (or the dashed lines in Figure 10). The true distributions are not too distant from each other—the true means are shown as TT and IT in Figure 11—and in fact substantially overlap. But the observed distributions are completely distinct and, in fact, are separated by a gap (the result of two adjacent points above moving in opposite directions because they were on opposite sides of the T--I cutoff). The observed means are shown as TO and IO in Figure 11. Thus, self- selection exaggerates the political bias of interpretive methods: TO [less than] TT and IO [greater than] IT.

What are you trying to say?

Even though I thought your introduction was too long, it was really clear and I followed exactly what you were saying. Your idea sounds interesting and thought provoking. You're obviously a great writer with a talent for explaining things clearly. Why mask that ability behind abstract charts, graphs, and variables?
9.13.2007 2:12pm
Bruce McCullough (mail):

For some reason I can no longer access the paper, so I have to work from memory.

I looked at the graph and I wondered, "Distribution"? Distribution of what? What on earth does this mean? I teach stats and in my ph.d courses give lectures on statistical graphics. And the figure _as labeled_ makes no sense.

The line in the figure represents a distribution, with the abscissa representing "policy" (or some aspect of a policy) on a numerical scale. Let the policy be "x". Then the ordinate should represente f(x), the value of the line when the policy takes on the value x.

Usually, for a density function, x is the value and f(x) is the probability associated with that value. For a distribution function, F(x) is the cumulative distribution function.

IANAL, so I can't really help you figure out what the proper labeling of the ordinate should be, but if the paper comes on line again, I'll re-read that part and try to make some guesses.


9.13.2007 2:27pm
Sasha Volokh (mail) (www):
The ordinate is of course the relevant density function. But lawyers have no idea what a "density function" is. And colloquially, "distribution" means more than just the cumulative distribution function.

If this were a math or econ paper, I would just not label the y-axis at all, but call my functions f(x) or whatever and write that next to the curve itself. But lawyers get the willies when you have names of functions -- even referring to things like M_T etc. is a stretch, as you can tell from one of the comments above -- and everyone wanted me to put some label on the axes. This is the best I could think of, definitely better than "density" or "density function" and I think even better than "number of judges" or "frequency."
9.13.2007 2:45pm
Truth Seeker:
Why does this remind me of the Sokal Affair?
9.13.2007 3:17pm
Dilan Esper (mail) (www):

I think it's wonderful that you chose to write about this topic. One of the things that ticks me off (as a liberal) is seeing conservatives simply pronounce as fact that, for instance, Scalia's methodology is the only proper methodology because of how "neutral" it is, without at least acknowledging that the selection of that methodology was anything but neutral. Indeed, Scalia clearly has strong personal feelings about such issues as abortion and gay rights, and his selection of his methodology is not a coincidence.

That doesn't make Scalia wrong-- it just means that people should not proclaim him or anyone else as the tribune of "neutrality" in interpretation.
9.13.2007 3:28pm
Should not the law mean what it says, and say what it means?

Granted, there are laws that are written to be intentionally somewhat vague, as in many circumstances, we must rely upon the "reasonable person" to deduce a particular meaning from the text, and I do not believe this necessarily invalidates a given law's legitimacy, but I think the Framers would have blanched at the idea that a law may be written in such a way as to be indecipherable even to subjects of even above-average reasoning capacity.

I would even argue that those areas of interpretation and construction that are most contentious have only become so because the Framers believed these areas to be commonly understood and not in need of further elaboration, while ideologues of all stripes in every era have seized upon any area of apparent uncertainty to advocate in favor of a particular political agenda.

I find the idea that it is acceptable for a particular adjudicant to "switch hit", as it were, to be morally and intellectually dishonest. The notion that because different methods appeal to different jurists validates a "mix-and-match" approach seems to me to be fallacious. Differing opinions held by different people are one thing, but a single person changing principles for different cases, simply because it might be more effecacious in a particular circumstance, is entirely another.

That said, it is certainly possible to adhere to a particular ideology and still recognize that there are areas of law under which it is improper to impose a judicial morality, if you will, upon the populace, though certainly many have used the indeterminacy, perceived or actual, of a given law without substantial reproach.

It is an adherence to original meaning, to the utmost possible extent, coupled with a truly "reasonable" approach to construction where substanial indeterminacy exists (and by "reasonable", I mean that one must acknowledge one's ideological opponent when he has grounded a point in solid logic rather than mere ideology and concede the point in the absence of a similar opposing viewpoint), that provides the most sure guard against such manipulation. Where such an adherence may conflict with our political ideals, it should only serve as a motivating factor for legislative change where excessive misunderstanding exists, and for increasing pressure on legislators to write laws which reflect to the maximum extent possible their actual intent for the statute.

In cases where the law is indeterminate enough to admit of more than one reasonable construction, I think we can then begin to use other interpretive methodologies to assist in informing us where a judicial opinion must be conveyed, but these circumstances should ideally be extremely uncommon.

A very interesting paper that may be helpful to those frustrated with the judicial system and the political climate, and it may even be effective in causing some people to reconsider superficial opinions of particular political viewpoints, but in the end, I would have been more satisfied had the paper been less positive and more normative. The fact that interpretive method selection does occur is already obvious. Judges and Justices are, after all, imperfect human beings as much as the rest of us. I find the why and how less important that the mere fact. What I think we need are more normative arguments for a new rationality in this country, and I think this paper runs the risk of being interpreted as an apologia for more partisan bickering.
9.13.2007 4:53pm
Sasha Volokh (mail) (www):
amper: I can sympathize with your wish that this paper would be more normative. However, if that's what you want, this paper is definitely not for you! I deliberately avoided any mention of anything normative until two pages of the Conclusion, pp. 54-55. And you won't be happy with the normative message there either. :)

If it makes you feel better, most of the literature on statutory interpretation is explicitly normative; I cite it in the footnotes. I recommend Scalia's short and highly entertaining book, A Matter of Interpretation.
9.13.2007 5:16pm
GV, you are not correct. Is your problem that "policy preferences" is an inadequate shorthand for the realists' position on indeterminacy of legal outcomes? If so, why not say that rather than resort to ad hominem?

I happen to think "policy preferences" is an adequate shorthand for, as Brian Leiter describes it, "what would be 'fair' on the facts of the case," (the "'core claim' of Realism") (link).
9.13.2007 5:25pm
DJR, because a core claim of legal realism is that a subset of legal issues (typically those that make it up the appellate ladder) are legally indeterminate, is a far cry from saying that judges decide cases based on their policy preferences. If you could find legal realists who claim that the law is so indeterminate that judge's policy preferences decide the cases before them, I'd love to see those links.

(If your claim is that judges sometimes resort to policy preferences to decide cases, then we agree.)
9.13.2007 6:04pm
See the Leiter article I linked to above.
9.13.2007 6:47pm
I've read the Leiter article previously. Nowhere does he discuss legal realists who believe policy preferences always decide cases. You're obviously not interested in having a debate. *shrugs*
9.14.2007 2:30am