The Volokh Conspiracy

that the Defense of Marriage Act is unconstitutional. So said Sen. Rick Santorum on the NewsHour (see here, roughly at the 9 minute mark), explaining why the Federal Marriage Amendment is needed:

The Supreme Court case in Lawrence versus Texas last year was very clear. The majority opinion was very clear. It signaled clearly that the Defense of Marriage Act was not going to stand. Every left, right, and middle constitutional lawyer in this country have all said that as a result of the Lawrence decision, the Defense of Marriage Act will not stand.

Yesterday on CNN, Sen. Hatch likewise said "most likely all constitutional authorities say the Defense of Marriage Act will be ruled unconstitutional" (I'm quoting a transcript on this).

Now I don't want to quibble with the "every . . . constitutional lawyer" — Sen. Santorum is entitled to a bit of hyperbole. But even taking that into account, it semes to me the Senator is just wrong. The Court did not clearly say or signal that DOMA was unconstitutional; the Court held that criminalizing sexual conduct violated people's liberty, not that homosexual couples were entitled to equal access to the benefits flowing from marriage. And my sense is that most constitutional scholars (not all, but most) that have considered the issue believe DOMA would be upheld. Indeed, some people, mostly liberals, have argued that DOMA is unconstitutional — but not remotely the broad swath that Santorum is suggesting.

One can plausibly argue that courts might strike down DOMA. I doubt that they will, but if one thinks that would be a really awful result, one could argue that we should preempt it now, rather than waiting for later. Of course, if one's concern is really about courts forcing the pro-gay-marriage position on states that oppose gay marriage, the solution would be an amendment that constitutionalizes DOMA. There'd be no need for an amendment that would force the anti-gay-marriage position on states that support gay marriage (which is what the FMA would do).

But in any event it seems to me incorrect to argue that somehow the courts' striking down DOMA is a foregone conclusion — and especially to argue that all or nearly all constitutional lawyers make such a prediction.

NOTE: It would also be wrong, I think, for those who say DOMA is unconstitutional to use DOMA's existence as an argument for why the FMA is unnecessary. (I suppose they could argue that DOMA is unconstitutional but courts will still uphold it, so the FMA is unnecessary because of that; but that still doesn't seem quite right to me, because they'd essentially be arguing that their fellow citizens should just rely on the government's violating the constitution. At the very least, such an argument ought to be made explicitly.)

If anyone can point me to specific examples of both of these arguments being made, likely at separate times, by a particular person or organization, I'd love to see that. Note, though, that I'm looking for the person or group making specifically those arguments. It's not just that they argue that DOMA is bad policy and that the FMA is unnecessary, or that DOMA is unconstitutional and that the FMA is unwise, or anything else. To deserve condemnation, they need to have argued that DOMA is unconstitutional and have also argued that the FMA is unnecessary because of DOMA.

Interesting little illustration of the way lawsuits against federal officials sometimes work, from the AP -- both as to who pays, and as to the resulting limits on punitive damages. I express no opinion about whether this is the proper result, but it seems interesting:

Former U.S. Rep. Bill Janklow was on duty when he caused a fatal accident last summer, so taxpayers should pay any civil damages in a wrongful death lawsuit, according to a court ruling Tuesday.

U.S. Magistrate Arthur Boylan sided with U.S. Attorney Tom Heffelfinger's conclusion that Janklow, 64, was on official business Aug. 16 when he sped through a stop sign near Trent and collided with motorcyclist Randy Scott, 55, of Hardwick, Minn.

Boylan concluded that the federal government, not Janklow, should be listed as the defendant in the lawsuit filed by Scott's mother, sister, son and daughter.

The family's lawyer, Ronald Meshbesher of Minneapolis, wanted the case moved back to state court in Minnesota so the family could get punitive damages, something not allowed if the case stays in federal court. . . .

Janklow had appointments or appearances over two days in Rapid City, Pierre and Aberdeen and was on his way to his Brandon home when he went through a stop sign at an intersection of two county highways. . . .

The magistrate also concluded that Janklow's driving habits did not remove him from coverage by the Federal Tort Claims Act that legally protects federal employees.

"It is foreseeable that improper driving conduct, including conduct that can be construed as reckless, would occur," Boylan wrote. . . .

Janklow, who was elected to Congress in 2002 after serving a total of 16 years as governor, spent 100 days in jail after being convicted of speeding, running a stop sign, reckless driving and second-degree manslaughter. He resigned from the House in January. . . .

Thanks to reader Dennis Callahan for the pointer.

I hope to put out a second edition of my Academic Legal Writing book eventually. If any of you have used it, can you suggest any changes, or let me know what you'd like to see added? Please let me know, at volokh at law.ucla.edu. And if you run a legal Weblog that has some law student readers, and could pose this question to them, I'd be very much obliged. Many thanks in advance.

Some research I was doing recently reminded me of the familiar line, "rape is a crime of violence, not a crime of sex." It's certainly true that rape (setting aside statutory rape and a few other unusual situations) is a crime of violence. I also suspect that most rapists aren't just after ordinary sexual gratification, but also want the feelings of domination (or something like that) that come from the violence of the act. Given the serious penalties for rape, and even taking into account the difficulty of catching and convicting the rapist, committing rape just to get ordinary sexual gratification is a pretty expensive proposition.

But it seems to me pretty likely that there's a false dichotomy here. In fact, rape seems to be both a crime of violence and a crime of sex -- the rapist is motivation by sexual desire as well as by the desire for domination (and the two may well be intertwined). The best evidence that I've seen for this is the breakdown of rape by age of victim (see National Crime Victimization Survey data, table 4):

Rapists seem to select victims in age ranges that are pretty highly correlated to the generally understood peaks of sexual attractiveness. Yes, there are rapes of older women; yes, women outside the highest-risk groups are sexually attractive; but the correlation is still quite striking.

Now of course there may be other explanations: For instance, younger women may spend more time in places where rapes might occur. Indeed, the simple assault and robbery rates peak in the same age ranges, too (see the same table) -- but the rates level off far more gradually, for instance with simple assault per 1000 women in the age group going from 32.8 for 12-15 to 37.1 for 16-19, 23.7 for 20-24, 15.9 for 25-34, 12.1 for 35-49, 7 for 50-64, and 1.4 for 65 and over. (Simple assault is defined as successful or attempted "[a]ttack without a weapon resulting either in no injury, minor injury . . . or in undetermined injury requiring less than 2 days of hospitalization"; the way the NCVS treats this, it seems to exclude the great majority of rapes or attempted rapes.)

Also, younger women tend to date younger men, younger men tend to be more likely to commit crimes, so younger women may be more likely to be victimized by date rape. But though roughly 2/3 of rapes involve nonstrangers (see table 27 here), 1/3 involve strangers, so the drop-off even for stranger rapes seems to be quite striking (see table 29 for data, though note that again there are relatively few cases in the survey for each category).

Again, perhaps there's some other explanation that completely eliminates the connection between rape and the offender's likely sexual interest in the victim. But given the data, I find that pretty implausible. Rape seems generally to be a crime of violent sex, not of violence or sex alone.

Some research I was doing recently reminded me of the familiar line, "rape is a crime of violence, not a crime of sex." It's certainly true that rape (setting aside statutory rape and a few other unusual situations) is a crime of violence. I also suspect that most rapists aren't just after ordinary sexual gratification, but also want the feelings of domination (or something like that) that come from the violence of the act. Given the serious penalties for rape, and even taking into account the difficulty of catching and convicting the rapist, committing rape just to get ordinary sexual gratification is a pretty expensive proposition.

But it seems to me pretty likely that there's a false dichotomy here. In fact, rape seems to be both a crime of violence and a crime of sex -- the rapist is motivation by sexual desire as well as by the desire for domination (and the two may well be intertwined). The best evidence that I've seen for this is the breakdown of rape by age of victim (see National Crime Victimization Survey data, table 4):

Rapists seem to select victims in age ranges that are pretty highly correlated to the generally understood peaks of sexual attractiveness. Yes, there are rapes of older women; yes, women outside the highest-risk groups are sexually attractive; but the correlation is still quite striking.

Now of course there may be other explanations: For instance, younger women may spend more time in places where rapes might occur. Indeed, the simple assault and robbery rates peak in the same age ranges, too (see the same table) -- but the rates level off far more gradually, for instance with simple assault per 1000 women in the age group going from 32.8 for 12-15 to 37.1 for 16-19, 23.7 for 20-24, 15.9 for 25-34, 12.1 for 35-49, 7 for 50-64, and 1.4 for 65 and over. (Simple assault is defined as successful or attempted "[a]ttack without a weapon resulting either in no injury, minor injury . . . or in undetermined injury requiring less than 2 days of hospitalization"; the way the NCVS treats this, it seems to exclude the great majority of rapes or attempted rapes.)

Also, younger women tend to date younger men, younger men tend to be more likely to commit crimes, so younger women may be more likely to be victimized by date rape. But though roughly 2/3 of rapes involve nonstrangers (see table 27 here), 1/3 involve strangers, so the drop-off even for stranger rapes seems to be quite striking (see table 29 for data, though note that again there are relatively few cases in the survey for each category).

Again, perhaps there's some other explanation that completely eliminates the connection between rape and the offender's likely sexual interest in the victim. But given the data, I find that pretty implausible. Rape seems generally to be a crime of violent sex, not of violence or sex alone.

reminded me of this old chestnut: A man goes to his minister. "Reverend, help me -- everything is going badly for me, I've almost given up hope, nothing is going right, what should I do?"

"Well, my son, do what I always do: Just open up a Bible to a random page, put your finger in a random place, and whatever you're pointing to is sure to be good advice."

A month later the man comes up to the minister again. "Thank you so much, Reverend," he says. "I followed what you said to the letter, and everything is much better now."

"I'm so glad to hear it, my son," the minister says. "But I'm curious: What verse did the Lord guide you to?"

"Well, I opened up the Bible, put my finger on a random place, and there it was, in black on white: Chapter 11!"

with a special emphasis on infant CPR, for Benjamin's benefit. Glad to know how to do it, though I hope I'll never have to.

Ice cream production is closely correllated with the rate of forcible rape. Yes, that's right — Professor Eugene Volokh, of the prestigious UCLA School of Law and the even more prestigious Volokh Conspiracy Web log, has uncovered scientific evidence that ice cream production is closely correlated with the forcible rape rate.

Professor Volokh used 2000 data from the Federal Bureau of Investigation's Uniform Crime Reports, and 2000 data from the international Dairy Foods Association's Dairy Facts publication; the correlation was 0.84, which is very high (1 would be a perfect correlation and 0 would be no correlation at all) and statistically significant. Unfortunately, data on monthly ice cream consumption was unavailable, but ice cream production data seems likely to be a good rough proxy for ice cream consumption. For the spreadsheet containing this data — The Volokh Conspiracy is a stickler for openness in research — see here.

It is time, especially in this election season, that Big Dairy be asked some serious questions about this disturbing phenomenon and about its impact on the women and children of our nation. Professor Volokh (volokh at law.ucla.edu) is available for interviews and commentary on this important research finding.

(Read here before sending me e-mail accusing me of statistical ignorance.)

I am told that when the New York election scheduled for 9/11 was postponed, the votes cast on the original day were discarded, and people were expected to revote on 9/25. Joel Graber, of the New York State AG's office, posted something more about this to an election law list I'm on, and I pass it along with his permission:

In New York City on 9-17-01, nine African-American, Hispanic and Asian-American Democratic candidates for City Council, and some voters, represented by highly-experienced civil rights attorneys, sued in the EDNY to postpone the primary rescheduled from 9-11-01 to 9-25-01, alleging violations of their "right to vote, free speech and political participation, freedom of association, right to due process and equal protection of the law," and sections 2 and 5 of the VRA. The claims boiled down to the arguments that the board of elections, which is located close to the WTC, was not up to the task, that voters were too distracted by the attack, and that Rosh Hashanah fell on Sep. 18 and 19 thus hampering plaintiffs' opportunities for political activity. A [temporary restraining order] and [preliminary injunction] were denied by [district judge] I. Leo Glasser on 9-20 and the case was abandoned. Barbour v. Board of Elections, EDNY 01-cv-6216 (ILG).

Legislation was passed on 9-13 resheduling the primary (and liberalizing absentee voting and other things), on the authority of N.Y. Const. art. III, sec. 25, which provides that the Legislature, "in order to insure the continuity of state and local government operations in period of emergency caused by enemy attack or by disasters (natural or otherwise), shall have the power and the immediate duty . . . to adopt such . . . measures as may be necessary and proper . . . ." The legislation also provided that valid absentee and military ballots cast by 9-11 would be counted and that votes cast at polling places on 9-11 would not be counted. 2001 N.Y. Laws ch. 298, ch. 303.

So at least in New York, there is already state law authorization for the state legislature to take care of election interruptions.

Reader David Smith reports that many people did leave work on September 11, because they wanted to be home, though that didn't seem to keep them away from the polls:

I live in Cambridge, work in Boston, run a small company, and was at my desk on 9/11. I remember the attacks clearly -- we turned on the office television. People were in quiet shock. Around 10:30 people started leaving their offices -- I could see them from my fifth floor window, on the streets, heading for subways and trains. There was a sense that we had no idea what would happen yet, and it seemed better to be home with loved ones and not in large buildings. Just because we did not know what better thing to do.

Perhaps people voted on their way home. Perhaps they voted in the morning before they knew the news of 9/11. (Voting is heavily clustered before 9 and after 5.) Perhaps they voted on a minor local issue because it was all such a shock that we could not interpret what to do next, and the minds were made up.

He also suggests that given the 9/11 shock, people won't be as shocked next time, and there'd be still less disruption of the election (at least in those places that aren't directly attacked):

All those facts would be different if, God forbid, a terrorist attack hit a day or two before the election. We are smarter now, wiser, we have seen things we wished not to see. We have had the Spanish 3/11 and its consequences. There would, in short, be not the same element of emotional shock. And so the Boston (or other) 9/11 voting experience is, in my view, not relevant to whether an election should be postponed in the event the unthinkable occurs.

May well be so.

A reader passes along this letter from the General Counsel to North Carolina's State Board of Elections summarizing the emergency powers given to the Executive Director of the State Board:

Dear Mr. Stephens:

Pardon for the delay getting back to you. Our office has just concluded a major statewide educational conference. Mr. Bartlett asked that I respond to your concerns.

The backbone of election disruption control is G.S. 163-27.1 set out below:

§ 163-27.1. Emergency powers.

The Executive Director, as chief State elections official, may exercise emergency powers to conduct an election in a district where the normal schedule for the election is disrupted by any of the following:

(1) A natural disaster.

(2) Extremely inclement weather.

(3) An armed conflict involving United States armed forces, or mobilization of those forces, including State National Guard and reserve components.

In exercising those emergency powers, the Executive Director shall avoid unnecessary conflict with the provisions of this Chapter. The Executive Director shall adopt rules describing the emergency powers and the situations in which the emergency powers will be exercised. (1999-455, s. 23; 2001-319, s. 11.)

At this time, I would interpret a terrorist attack that adversely affects an ongoing election as coming within the meaning of "armed conflict" as set out in (3) above.

Your concerns as to the effect of terrorism prior to Election Day on voter turnout have a less clear solution. However, the sanctity of both the legal and historical general election date of the first Tuesday after the first Monday in November would weigh heavily against a delay of the general election based upon calamities that make it more difficult for voters to vote, either physically or emotionally. The certainty of election dates sets the Untied States apart from other countries.

Let's look at some of the elements that may effect this situation.

Approximately 10% of the voters vote in a general election during one stop-absentee voting that starts 19 days prior to the election and ends the Saturday before Election Day. Thus we must view the "slow calamity" disruption issue in the context that voters in North Carolina, and many other states, can vote on more than one day. And absentee by mail voting starts 50 days prior to Election Day.

A statewide primary election will cost about $4 million dollars, and the cost of a statewide general election will be around $5.5 million dollars.

A suspended election would involve extraordinary efforts to secure and protect ballots and voting systems. Issues arise as to the terms and conditions of resuming the suspended election.

A cancelled election means that voting systems would have to reprogrammed, and ballots reprinted. These activities are time and labor intensive as well as costly.

A suspended election or new election could prohibit winning candidates from taking office in a timely matter, allowing previous officeholders to hold over until the winning candidates are certified and sworn in.

In addition, North Carolina has some Section 5 counties under the Voting Rights Act, and all actions as to suspended or new elections would have to be precleared by the U.S. Justice Department, a procedure that can take up to 60 days.

As to your three suggestions, let me offer some positive news.

First, the counties are very aware of the situations that can cause Election Day disruption. We have had precincts closed due to bomb threats, gas leaks, ice storms, tornados, hurricanes, and flooding. The county board makes an analysis of the situation and requests a decision under GS 163-23.1.

Secondly, GS 163-23.1 is designed to have a quick response to a situation. Of course, the State Board is consulted over such issues. This agency starting in 2000 encouraged counties to have disaster and recovery plans on file.

Thirdly, Congress is currently dealing with the issues you note. You may want to review H.R. 2844 that deals with replacing members of the House of Representatives in cases of extraordinary circumstances. The entire issue of continuity of government is being addressed by the Continuity of Government Commission (a joint project of the American Enterprise Institute and the Brookings Institution). Go to their website at www.continuityofgovernment.org . They issued their first report in May, 2003.

This agency is currently considering suggesting to the General Assembly a clarifying amendment to GS 163-23.1 to clearly mention terrorism as a basis for emergency powers by the Executive Director.

Please be assured that this most difficult subject will continue to be a major active concern of our agency.

Don Wright

General Counsel

I wrote earlier today that

I agree with Jack Balkin, though, that:

The fact that a terrorist attack might influence voters one way or the other is not a reason to cancel an election. Lots of things happen before elections that can influence voters. Rather, the reason to postpone an election is that it is simply not possible to conduct the election in a particular jurisdiction, because, for example, there are dead bodies lying everywhere or buildings have been blown up and local services have to be diverted to matters of life and death.

I would probably set the threshold somewhere below "not possible" and "dead bodies lying everywhere." On September 11, 2001, one shouldn't have conducted elections even in Boston, where elections were physically possible and no dead bodies were present, because I suspect that lots of people would have been scared away from the polls (at least for several hours) either by the shock of the event, or by the reasonable fear that there might be still more attacks that very day. Naturally, that's a judgment call, but a categorical delay of several days following a very serious terrorist attack — on the order of hundreds or thousands of deaths, especially coordinated in multiple places — would probably be wise. But in any event, the focus has to be on the risk that the election would be disrupted by lower turnout or difficulty physically conducting the process, not by people's views being influenced.

Well, reader Steve Jens points out that there was an election in Boston on September 11, 2001, a Congressional primary for a vacant seat. The turnout was apparently expected to be about 20%-25% (not unlikely for that sort of election). The election went on as scheduled, despite the attacks, and the turnout was 32%. I can't confidently say the attacks increased the turnout, but there's little reason to think they decreased the turnout. So perhaps Jack is right: If there's a September 11-scale terrorist attack in the area where a state or local election is held, that election should be postponed (as was done for an election scheduled for that day in New York), but if the attack is in a different area, the election should go on as planned. (One could argue that the vote will be distorted by the passions caused by the attack, but as I mentioned in my earlier post I tend to agree with Jack that this isn't reason enough to postpone the election.)

That still leaves the question of what to do with a major terrorist attack during a national election. I tend to agree with Rick Hasen that one should probably postpone the entire election, rather than hold a makeup election in just the affected area.

But in any event I'm pleasantly surprised, happy to have my hypothesis proven wrong, and thankful to Steven Jens for the correction.

The U.S. Court of Appeals for the Second Circuit, sitting en banc, just certified questions related to the constitutionality of the Sentencing Guidelines to the U.S. Supreme Court. So instead of a party's asking the Supreme Court to consider a question (via a petition for certiorari or, in some cases, an appeal), it's the Court of Appeals that's asking.

I didn't know this was possible. I knew federal Courts of Appeals could certify questions to those state supreme courts that allow this procedure, but I didn't know that they could certify questions to the Supreme Court. Some research reminds me that this was a normal procedure in the 1800s for deciding legal issues in criminal cases; since there was no provision for appeal in such cases, which were often heard by two-judge federal trial courts, the case would sometimes go up to the Supreme Court when the judges certified the question to the Justices (especially when the trial court judges were evenly divided). Learn something new every day . . . .

In any case, 28 U.S.C. sec. 1254(2) provides that the Supreme Court may consider a matter "By certification at any time by a court of appeals of any question of law in any civil or criminal case as to which instructions are desired, and upon such certification the Supreme Court may give binding instructions or require the entire record to be sent up for decision of the entire matter in controversy." The Court may refuse to answer the questions (see, e.g., Supreme Court Rule 19.3; In re Slagle, 504 U.S. 95 (1992); Foley v. Carter, 449 U.S. 1073 (1981); Atkins v. U.S., 426 U.S. 944 (1976) (with three Justices dissenting from dismissal of certificate)), and leave it for the Court of Appeals to do the work.

A quick search suggests that the practice was not unheard of (though producing only a tiny fraction of the docket) until the mid-1950s, but in Wisniewski v. U.S., 353 U.S. 901 (1957), the Court seemed to try to discourage the practice. The most recent case I could find in which the Court decided a matter certified to it by a Court of Appeals is Iran National Airlines Corp. v. Marschalk Co., 453 U.S. 919 (1981). Here's the heart of Court of Appeals' argument for certification (besides its argument that it's hard to tell what the recent Blakely v. Washington decision means for the Guidelines — if ceritification were proper whenever a Supreme Court decision led to confusion, then the Court would have to answer certified questions nonstop):

[W]e believe this is one of those "rare instances" when "the proper administration and expedition of judicial business" warrants certification of a question to the Supreme Court. Blakely not only casts a pall of uncertainty on more than 220,000 federal sentences imposed since Apprendi was decided, but it also raises the prospect that many thousands of future sentences may be invalidated or, alternatively, that district courts simply will halt sentencing altogether pending a definitive ruling by the Supreme Court. We are convinced that a prompt and authoritative answer to our inquiry is needed to avoid a major disruption in the administration of criminal justice in the federal courts — disruption that would be unfair to defendants, to crime victims, to the public, and to the judges who must follow applicable constitutional tequirements.

And here's the circuit's plea for unusual speed:

We recognize that the current term of the Supreme Court has ended, but we respectfully request that the Court not only entertain this certification, but do so at its earliest convenience, with an expedited briefing and hearing schedule, cf. Dames & Moore v. Regan, 453 U.S. 654 (1981) (setting case for oral argument on June 24, 1981, after regularly scheduled arguments concluded, and deciding case on July 2, 1981); Iran Nat'l Airlines, 453 U.S. at 919 (answering certified questions seventeen days after in order to minimize, to the extent possible, what we see as an impending crisis in the administration of criminal justice in the federal courts.

And apparently the Court's initial decision whether to consider accepting the certificate will be made solely based on the Court of Appeals' arguments, without briefing by the parties; here's the relevant part of Supreme Court Rule 19 (emphasis added):

3. When a question is certified, the Clerk will notify the parties and docket the case. Counsel shall then enter their appearances. After docketing, the Clerk will submit the certificate to the Court for a preliminary examination to determine whether the case should be briefed, set for argument, or dismissed. No brief may be filed until the preliminary examination of the certificate is completed.

4. If the Court orders the case briefed or set for argument, the parties will be notified and permitted to file briefs. . . .

I would guess, though, that given the unanimous request of the Second Circuit's judges, the Court would at least ask for briefing.

Thanks to Douglas Berman (Sentencing Law & Policy) for the pointer to the Second Circuit decision.

UPDATE: Marty Lederman (SCOTUSBlog) has more.

OpinionJournal's Best of the Web repeats an old math gag:

Let a=1

Let b=1

Therefore a=b

Multiplying both sides by a gives a^2=ab

Subtract 1 from the left and b (which equals 1) from the right: a^2-1=ab-b

If you remember your quadratic equations, this factors to: (a+1)(a-1)=b(a-1)

Dividing both sides by a-1, we have a+1=b, or 1+1=1

Therefore 2=1

Persuades me!

(Click here for the explanation.)

Rick Hasen has more:

In this discussion, I think it is valuable to disaggregate a few different issues:

(1) Who should decide on whether an election should be postponed in the event of a terrorist strike? Jack Balkin's post makes it clear that, so long as we are talking about presidential elections, it is for Congress, not the executive, to make such decisions. This is both constitutionally mandated and politically wise. One caveat: Norm Ornstein and others have been focused on questions related to a catastrophic attack on Congress, and the need for a plan to select replacement members of Congress if necessary. If for some reason the terrorist attack would be directed at Congress, some backup plan might be necessary.

(2) Which criteria should be used to determine when a presidential election should be postponed? As blog readers know, I am a big believer in setting forth clear rules with as little discretion as possible for resolving election disputes before they arise. Some kind of mechanical rule seems desirable here. (Eugene Volokh suggests a few possible rules in his post.)

(3) If a terrorist attack occurs in one part of the United States, should the entire presidential election be postponed, or only the election in directly affected areas? Putting aside the problem of clearly defining which areas are "directly affected," the more fundamental question is whether Congress should use its powers in the event of an emergency to delay the election nationwide. As I mentioned earlier, fairness should dictate the entire election be postponed. Imagine a terrorist attack in a battleground state, where it would be impossible to hold an election on election day. The entire country besides this state votes on Election Day and either election results are announced by the other states or the media report on their exit polling. Such information could then sway votes and/or turnout in the battleground state on the date of the makeup election, and could thereby change the result the would have occurred had the entire country voted on the date of the makeup election. (Think of the Nader voters in Palm Beach County, voting on a hypothetical revote after the butterfly ballot fiasco.)

(4) Should a presidential election ever be postponed because of a terrorist attack that does not interfere with the ability of voters to cast votes? I think the early blogospheric consensus is that simple shock or grief would not be a legitmate reason for the postponement of an election.

Obviously these are difficult issues and ones with which some people would rather not deal. But prudence dictates thinking about these kinds of issues in advance whenever possible.

Again, I agree, especially as to item 2.