British warplanes struck a large bunker on Aug. 26, 2011, in Moammar Gaddafi’s home town of Sirte, Libya, his largest remaining stronghold, as NATO turned its attention to loyalist forces battling advancing rebels in the area. (AP)

These are difficult days for defenders of the Responsibility to Protect (R2P) doctrine, which holds that the international community must be prepared to act when countries “manifestly fail to protect their populations from genocide, war crimes, ethnic cleansing and crimes against humanity.” All member states of the United Nations endorsed this language in 2005.

In the past year alone, however, mass atrocities against civilian populations in Syria, Central African Republic and South Sudan have unfolded in plain sight while international efforts to halt these crimes have ranged from tentative to nonexistent. When, contrary to this trend, the Obama administration employed military force last summer to rescue members of the Yazidi minority in northwestern Iraq, some observers asked: Why protect the Yazidis and not the multitude of other threatened groups?

The R2P doctrine was supposed to answer this question. It says that civilian populations have a right not to be subject to mass atrocities and that all states have the responsibility to uphold this right, a formulation that has gained a broad following. It has helped elevate the importance of “human protection” in the United Nations and elsewhere, including in the White House. Two years ago, President Obama issued a directive that “the prevention of mass atrocities and genocide is a core national security interest and a core moral responsibility of the United States.”

But for all this attention, R2P did little to resolve the toughest questions of armed humanitarianism: When, for whom and how should coercive force be employed?

For a while, the international air campaign in Libya in 2011 seemed to render such questions moot. After Libya’s then-President Moammar Gaddafi threatened to overrun Benghazi, the U.N. Security Council authorized military force to protect “civilians and civilian-populated areas” in the country. Rapid intervention by NATO-led air forces quickly stopped Gaddafi’s forces – a remarkable demonstration of R2P’s utility, or so it seemed at the time.

In retrospect, however, the Libya intervention was problematic for R2P. After neutralizing the immediate threat to Benghazi, the NATO-led coalition provided de facto air support for Libyan rebels, who counterattacked and ultimately destroyed the Gaddafi regime. This, in turn, provoked an angry response from several countries, including some that had voted for intervention and now accused NATO of using R2P as a cover for regime change. Meanwhile, inaction in the face of mounting atrocities in Syria elicited a very different criticism of R2P: That it was an ineffective, hollow doctrine that offered false hope to threatened populations.

To some extent these criticisms were specific to the circumstances of Libya and Syria, but in a recent International Peacekeeping article, I argue that they have revealed deeper tensions in the strategic logic of humanitarian intervention and R2P:

1) The mixed-motives problem – The legitimizing rationale for a R2P intervention is its altruistic aim. However, decisions to use armed force usually involve a mix of motives, including self-interest. This is not only unavoidable, but to some extent it is also desirable and necessary; unless humanitarian interventions are partly rooted in self-interest, intervening states may lack the political commitment to complete the tasks they undertake. On the other hand, self-interested motivations make R2P interventions inherently prone to having their legitimacy called into question.

2) The counterfactual problem – It is virtually impossible to demonstrate that such missions have accomplished their main objective because the primary evidence of success is a non-event: That is, a mass atrocity that did not occur. Consequently, defenders of the mission must resort to “counterfactual” arguments that involve imagining a reality that might have taken place.

3) The conspicuous harm problem – What is actually visible, instead, is the destructiveness and costs of the intervention itself. This is bound to have a more immediate impact in public debates about the operation, not only because the detrimental effects of military intervention are obvious and material whereas an averted atrocity must be imagined, but also because the stated purpose of the intervention is to prevent harm.

4) The end-state problem – If an operation achieves its immediate goal of protecting a threatened population, it must then devise an “exit strategy” of one kind or another. The problem, however, is that the requirements for terminating such a mission are different – and more expansive – than the initial goal of preventing mass killing. As a result, humanitarian intervention appears to have a built-in propensity toward mandate-expansion.

5) The inconsistency problem – There will be cases in which mass atrocities loom, but outsiders do not intervene to protect civilians for any number of reasons, including because R2P itself counsels against intervening in circumstances in which military action is likely to do more harm than good. The inevitable result is an appearance of inconsistency, which in turn erodes the credibility of R2P.

Each of these problems is daunting in itself, but they coalesce in a way that makes it difficult to establish the usefulness and legitimacy of R2P and humanitarian intervention. The central issue is not a lack of political will or intervention machinery, but that the strategic logic of this kind of mission gives rise to a profound dilemma. On one hand, if there is no intervention in the face of looming mass atrocities, R2P is likely to be criticized as phony or hollow, because of the inconsistency problem. On the other hand, if a preventive operation is launched and achieves its initial goal of averting an atrocity, it is still likely to be judged harshly because of the combined effects of the first four problems.

In other words, the more R2P is employed as a basis for military action, the more likely it is to be discredited, but paradoxically, the same will hold true if R2P’s coercive tools go unused. This helps to explain why the Libya intervention was simultaneously a triumph and setback for R2P. The doctrine is trapped by its own internal logic.

This does not mean that all humanitarian interventions are destined to fail, or that R2P has no future. But it does highlight the inherent problems facing such interventions, while also dampening hopes for the development of R2P into a reliable mechanism for averting atrocities. As in the past, states will analyze each crisis through the filter of their respective interests and values, considerations of what is feasible, and calculations of the costs and benefits of action versus inaction.

Roland Paris is founding director of the Centre for International Policy Studies and associate professor in the graduate school of public and international affairs at the University of Ottawa. He tweets at @rolandparis.