How much argument can there be over the definition of two words?

A lot, it turns out, if those two words are "inherently governmental" and if they appear in federal documents defining what functions U.S. agencies cannot contract out to the private sector.

What are inherently governmental functions? Once upon a time they definitely used to include those activities done by the military, intelligence, law enforcement and criminal justice agencies.

The use of private contractors over the years in all those sectors has stimulated concern and debate over what contractors should and should not be allowed to do.

The view of the private sector has always been clear and simple. They are willing to do anything that the government issues a contract for. In a capitalist economy that is perfectly legitimate, and nobody should be surprised by it.

The flip side of that, however, is that just because function or activity is not inherently governmental does not mean the private sector should do it or even compete to do it.

Office of Management and Budget Circular A-76, first issued in 1966 and revised as recently as 2003, and generally regarded as one of the primary governmental drivers for privatization and outsourcing, requires agencies to outsource all functions that are not "inherently governmental."

In 1998 Congress gave added force to Circular A-76 by mandating that all agencies must identify government positions that are not "inherently governmental" and contract with the private sector to fill those positions if doing so would be more efficient than filling them with governmental employees. Inherently governmental functions may not be outsourced, however, and such functions include combat positions.

But what does the government see as inherently governmental? This is not just a rhetorical question. For example, as I noted in early May, some in Congress have proposed prohibiting contract employees from performing "inherently governmental" security operations, including situations involving combat or extremely hazardous duties.

Others think the term may need to be redefined. Laura A. Dickinson, a professor at the University of Connecticut Law School, testified to the Senate Homeland Security and Governmental Affairs Committee in February that Congress might consider designating such functions as intelligence interrogation or private security as "core" rather than inherently governmental, which would permit outsourcing but at the same time impose limits on the percentage of positions that may be turned over to contractors, while mandating higher standards of oversight regarding these positions.

She said, "The State Department should not find itself in the position, as Patrick Kennedy's report on the Sept. 16 Blackwater incident concluded, that it does not have enough Diplomatic Security Agents to even monitor the actions of contractor security guards, let alone protect government officials themselves."

Ironically, the U.S. government itself is not particularly clear on what is inherently governmental. If one searches the U.S. Code, one sees there is far more clarity on what is inherently religious, as opposed to what is inherently governmental.

According to material handed out earlier this year at a conference on the future of private security contracting, sponsored by the New America Foundation and the Center for a New American Security, the phrase "inherently governmental functions" is used multiple times in the U.S. Code, OMB guidance and Federal Acquisition Regulations, and is not used consistently.

The Defense Department, which seems to have the most extensive guidance, used 120-plus pages to describe inherently governmental and commercial activity exceptions, and the term is used for a multitude of different reasons.

The Federal Acquisition Regulations uses the phrase 17 times -- many of them accompanied by parenthetical exceptions. The defense subset of the regulations, known as DFAR, has a section 7.503 that states that requirements documentation must include determination, "that none of the functions to be performed by contract are inherently governmental."

But the Defense Department's implementing instruction, numbered 1100.22 and running to 60 pages, goes into rhetorical gymnastics when it comes to the point of defining what would exclude direct participation in hostilities by contractor employees.

One section states in part that:

"Security in hostile environments �� shall be designated for military performance �� if, in the judgment of the commander, the security operations would entail defense against military forces of another sovereign government or non-state actors (e.g., terrorists or insurgents) whose capabilities are so unpredictable or sophisticated or such a high risk that a military capability is needed to provide the discretionary authority, flexibility, and options necessary to achieve the sovereign interests of the United States."

"If these security operations were performed by a private security contractor," the instruction concludes, "it most likely would constitute an inappropriate relinquishment of the U.S. government's sovereign authority."

And if that is not enough, consider the pending National Defense Authorization Act. It has three sections addressing "inherently governmental" and all are different. Yet more amendments may be added before it is finalized.

If trying to find a common definition seems impossible, what is the alternative? One possibility is to use the Law of Armed Conflict as a guide. LOAC is a part of public international law and regulates the conduct of armed hostilities.

In the United States the Pentagon has a LOAC Program that requires each military department to design a program that ensures LOAC observance, prevents LOAC violations, ensures prompt reporting of alleged LOAC violations, and appropriately trains all forces in LOAC.

Under LOAC one can check to see whether a private contractor is directly participating in hostilities. If they are, do they fit into the privileged combatant status? If they don't, then their role should be changed. And if LOAC is found not to apply, then there is no reason a party involved in hostilities should not use PSCs for security. The same criterion can be applied to other situations, thus avoiding the endless tautological argument about what is inherently governmental.

(U.S. Navy veteran David Isenberg is a military affairs analyst. He is an adjunct scholar with the Cato Institute, a correspondent for Asia Times and the author of a forthcoming book, "Shadow Force: Private Security Contractors in Iraq." His "Dogs of War" column, analyzing developments in the private security and military sector, appears every Friday.)