Does the conflict in Iraq amount to a civil war? In many ways, the public debate over this question is largely political. Calling Iraq a "civil war" implies yet another failure for the Bush administration and adds force to the question of whether U.S. troops still have a constructive role to play.

Politics aside, however, the definition of civil war is not arbitrary. For some -- and perhaps especially Americans -- the term brings to mind all-out historical conflicts along the lines of the U.S. or Spanish civil wars. According to this notion, there will not be civil war in Iraq until we see mass mobilization of sectarian communities behind more or less conventional armies.

But a more standard definition is common today:

1) Civil war refers to a violent conflict between organized groups within a country that are fighting over control of the government, one side's separatist goals, or some divisive government policy.

By this measure, the war in Iraq has been a civil war not simply since the escalation of internecine killings following the bombing of a Shiite shrine in Samarra in February, but at least since the United States handed over formal control to an interim Iraqi government in June 2004.

Here's why: Although the insurgents target the U.S. military, they are also fighting the Shiite-dominated Iraqi government and killing large numbers of Iraqis. There is little reason to believe that if the United States were suddenly to withdraw its forces, they would not continue their battle to control or shape the government.

Political scientists who study civil war have proposed various refinements to this rough definition to deal with borderline cases. One issue concerns how much killing has to occur -- and at what rate.

2) For a conflict to qualify as a civil war, most academics use the threshold of 1,000 dead, which leads to the inclusion of a good number of low-intensity rural insurgencies.

Current estimates suggest that more than 25,000 Iraqis have been killed in fighting since the U.S.-led invasion in March 2003 -- a level and rate of killing that is comparable to numerous other conflicts that are commonly described as civil wars, such as those in Lebanon (1975-1990) and Sri Lanka (beginning in 1983).

The organization -- or rather, disorganization -- of the warring communities in Iraq means that a large-scale conventional conflict along the lines of the U.S. Civil War is unlikely to develop. More probable is a gradual escalation of the current "dirty war" between neighborhood militias that have loose ties to national political factions and are fighting almost as much within sectarian lines as across them.

This is roughly what happened in Lebanon and at a lower level in Turkish cities in the late 1970s. Ethnic cleansing will occur not as a systematic, centrally directed campaign (as in Bosnia), but as a result of people moving to escape danger.

And there's another twist to the terminology:

3) If the conflict in Iraq becomes purely a matter of violence between Sunni and Shiite communities driven by revenge and hatred rather than by political goals, many political scientists would say that it is something other than civil war.

Almost no one, for example, calls the Hindu-Muslim violence in India a civil war.

A civil war has to involve attempts to grab power at the center of government or in a given region, or to use violence to change some major government policy.

In Iraq's case, however, the vacuum of power at the center means that communal violence will inevitably be tied to struggles for political power and control.

A final complication concerns the nature of international involvement. Some argue, for example, that the war in Bosnia should be seen as an interstate war rather than a civil war, since the Bosnian Serb forces were armed and directed largely by Belgrade. Post-Mobutu violence in Congo is often termed a civil war, even though fighters have been closely tied to armies from neighboring states.

4) A conflict may be both a civil and an interstate war at the same time.

The Vietnam War, for instance, clearly comprised both a civil war in the South and an interstate war involving the North, the South and the United States.

Iraq may be moving in this direction. The United States and Britain are already openly involved, and such neighboring countries as Iran and Syria are more covertly involved. Not that it matters to the people dying there, but the next debate here may turn on whether what is already a civil war in Iraq should be viewed as an interstate war as well.

The fierce debate on immigration ignores a crucial reality -- what is happening to the United States is only one piece, although a big one, of a much larger global picture.

That hit me a couple of weeks ago when I was in Singapore. The Southeast Asian island nation has long been hailed as an economic model, the business capital for the entire region.

But it is an economy facing demographic peril. Its small population of 4 million is shrinking, thanks to a very low fertility rate. Prosperous Singaporean couples work hard, have fewer children and worry about how to take care of their aging parents. By 2050, Singapore will have a median age of over 52, one of the oldest in the world.

Singapore's answer is to import labor. A third of its workforce are migrants, from construction workers to maids. One out of seven households employs a domestic worker -- low-paid women mostly from neighboring Philippines and Indonesia.

Singapore tries to lure "talents'' -- highly skilled and affluent migrants -- to stay permanently. But the men hauling bricks and the maids washing laundry are in a separate class of temporary guest workers, with no chance to join Singaporean society. If a maid becomes pregnant, she is shipped out within seven days. Employers have to post bonds that must be paid should their servants break the rules and try to stay, putting them in the role of migrant police.

Problems of abuse of domestic workers, including physical and sexual violence and confinement, are serious enough to have prompted a report last December by Human Rights Watch.

Singapore's dependence on migrant labor and its guest-worker policy may be at the extreme end but it's very much on the global spectrum. Labor, like capital and goods before it, is part of a global market. The movement of people across borders in search of wages and work, most of it from developing countries to developed, is growing at a phenomenal pace.

The numbers are staggering. From 1980 to 2000, the number of migrants living in the developed world more than doubled from 48 million to 110 million. Migrants make up an average 12 percent of the workforce in high-income countries. About 4 million migrants cross borders illegally every year.

The demand for labor is driven in part by a demographic disaster -- the falling birth rates of developed countries. Almost all of those countries now have fertility rates that are well below 2.1, the level at which a population replaces itself. At the very low end are Hong Kong (0.94), Korea (1.22) and Singapore in Asia (1.24), along with much of Eastern Europe.

Low fertility means shrinking workforces and aging populations. Without migration, according to a recent study, Europe's population would have declined by 4.4 million from 1995 to 2000. Immigration accounted for 75 percent of U.S. population growth during the same period.

This movement of people cannot be stopped, certainly not by hundreds of miles of fences or even by tens of thousands of border guards. It is an issue that cries out for global cooperation, for common policies that cut across national boundaries. Already, we can benefit from looking at what has worked -- and not worked -- elsewhere.

A Global Commission on International Migration, formed in 2003 by the United Nations secretary-general, has taken an initial stab. Their report, issued last winter, supports the growth of guest-worker programs.

The Senate immigration bill now up for debate includes a provision for a guest-worker program. The bill is clearly preferable to the punitive and ineffective approach of the House version. But the Singapore experience -- and previous guest-worker programs like the German import of Turks -- should prompt second thoughts about going down this road.

One problem is that the guests don't leave. The United States has its own experience with this in the bracero program to import farmworkers, and more recently with the supposedly temporary H1-B visas used so extensively by the high-tech industry here in Silicon Valley.

Most troubling to me, these programs create an underclass of migrants who are never assimilated, as happened in Germany. It sets us on the Singapore road, encouraging inhumane policing mechanisms. And it is a gilded invitation to employers to depress the wages and incomes of American workers, and not just in the dirty jobs that are supposedly so hard to fill.

The United States has been rightfully proud of a tradition that treats all immigrants as citizens in the making. Rather than guest workers, isn't it more American to set realistic immigration quotas and enforce them fairly?