Last Monday, the Cairo daily al-Ahram published (in Arabic) an interview with David Welch, U.S. ambassador to Egypt. Toward the end of the interview, Welch was asked whether the United States would seek to remake the Middle East by redrawing its map. “Absolutely not,” replied Welch. “We do not want to draw new maps in the Middle East; we are not an imperial power. We have enough problems with the maps that were drawn in the past.”

There was nothing inappropriate about the answer, but it does get one thinking. The assumption of U.S. policy is that the region can be stabilized, “reformed,” even democratized, without creating any new states, eliminating any existing ones, or redrawing any of the region’s borders. It’s the regimes that have to be changed in some places, the nature of regimes in others.

Is that assumption a valid one? For example, is it a coincidence that the two most egregiously failed states in the region have been Iraq and Afghanistan? These are heterogenous polities that are difficult to rule at all, except through a centralizing dictatorship. Saddam Husayn is often cast as the man who destroyed Iraqi civil society. But the country has been plagued by strong-man interventions since its founding, precisely because none of its constituent elements—Sunni, Shiite, Kurdish—has ever felt secure. In Afghanistan, the lack of trust among its components—Tajik, Pashtun, Hazara—has given rise to warlordism and perpetuated it.

There are two possible solutions for these states. One is federalism, which has no record of success in the region, and many instances of failure (one thinks in particular of various “Arab federations” that came and went). Another is a kind of power-sharing that worked in Lebanon for about thirty years. It was the late Edward Shils who underlined the weakness of this kind of solution, well before it collapsed. Such a polity “revolves around an empty center.” It therefore has to “be kept completely still politically” in order to prevent distrust from exploding. Keeping Iraq and Afghanistan “completely still” will be an expensive business. A new study by the U.S. Army’s Center of Military History says its would require a long-term commitment of 300,000 troops in Afghanistan, and 100,000 in Iraq.

Might it be less costly to the United States and more conducive to democracy, to allow the kinds of modifications that occurred in the map of Europe after the fall of communism? After all, the states of Iraq and Afghanistan were founded at about the same time as Czechoslovakia. And if the present configurations of Iraq and Afghanistan inherently encourage dictatorship and warlordism, what is the U.S. interest in perpetuating them? More important, is the United States prepared to sacrifice treasure and lives to maintain them?

I ask these questions; I haven’t answered them. I am familiar with the objections, and the quarters from which they might come. And the eve of an Iraq war is not the time for planners to discuss the issue, at least in public. But it is certainly not too early to begin thinking about it—and where can that be done, without commitment, if not on a weblog like this?

One thing I will say: only someone with no sense of the history of the Middle East believes that the region’s map will never change again. The question is not whether it will be redrawn, but when and where, by whom and in whose favor.