Undoubtedly the United States needs some kind of amnesty (or earned legalization, or call it what you will) for many of the millions of noncitizens now living unlawfully within its borders. Humaneness and pragmatism favor such a solution. But Joseph Carens offers more than just a policy analysis that addresses our current dilemma. He argues that amnesties are obligatory in principle for liberal democracies, and that they should take a particular form.

Although Carens is right to explore whether and when irregular migrants have moral claims to amnesty, the method by which he seeks to establish his claims seems flawed. The flaws of the method may be further confirmed by the conclusion that he derives from it: the recommendation of a permanent rolling amnesty after five or seven years.

The first problem with Carens’s method is that it proceeds from carefully chosen anecdotes and then abstracts by removing the details that give them resonance and reducing them to generalities about time. Time sinks deep roots. Time accumulates experiences. Time builds social connections. Time therefore erodes the moral right of states to deport irregular migrants.

One might ask, which roots and where? Which experiences? Which social connections? But Carens forbids these questions, both to the officials carrying out the amnesty, and to the policy-makers designing it. It is “presumptuous” even to imagine that we can make “nuanced judgments about how deeply another belongs to the society in which she lives.” That move leaves time as the only legitimate qualification for legal status (except, perhaps, for the absence of serious criminality, which does not include routine immigration offenses).

Social rootedness is assumed without asking whether a migrant’s residence within the host state is fixed or mobile; whether the migrant’s personal relationships are stable or fluid; whether the migrant’s social connections are to one heterogeneous community, an enclave of fellow migrants, or a series of unrelated job sites. Nor does the analysis take into account the migrants’ continued ties to their countries of origin—the relatives and friends they left behind, the property they own there, the remittances they send, the visits they may make there.

We live in an age of transnational migrants, enabled by advanced technologies of transportation, communication, finance, and trade to operate simultaneously as partial members of more than one national community. This phenomenon offers different opportunities to different individuals, and those individuals exercise their options in varying ways. Children brought to a new country at an early age may later have no friends, no relatives whom they know, no linguistic ability or cultural competence in a distant country of nationality. Adult labor migrants often remain more strongly connected.

The dual membership of irregular migrants does not provide reasons for keeping them indefinitely in unlawful status, vulnerable to exploitation. But it can contribute to an argument for denying some of them amnesty and returning them to their countries of origin, even after five or seven years.

Second, one might also ask, which states are obliged to grant amnesty? Carens’s reasoning speaks to liberal democratic states in general. Each has its particular history and circumstances, but they are bound by common moral constraints. Still, might the entailments of those moral constraints depend upon the circumstances? The United States is a multicultural giant in which the world language of English dominates. Smaller states, states with more fragile languages, and states that have had tense relations with their neighbors, may be differently affected by the continuous arrival of irregular migrants. If these concerns are irrelevant to the United States, it is not for reasons inherent in liberal democracy.

Finally, Carens dismisses worries about the effects of rolling amnesty on the incentives of migrants. He argues that potential migrants will not be influenced by the opportunity for legal status five years in the future because it is too distant and because there is insufficient guarantee that the legislation will not be changed. Neither branch of this defense is persuasive. Lifelong legal access to the U.S. labor market is a major asset, and many people invest comparable periods of work or training in the hope of achieving it. More importantly, it is problematic to claim that a liberal state has a principled obligation to adopt a permanent policy that is feasible only so long as outsiders do not trust the state to maintain the policy.

Permanent rolling amnesty provides a predictable road map to legal status: enter without authorization, keep a low profile for the pre-announced period of time, and document your compliance with the pre-announced conditions. There is no risk of discretionary denial in Carens’s plan, only a risk of detection. If that risk were very high, the incentive would be much lower, but there is no reason to believe that the risk of detection is high in generic liberal democracies, and we know that it is not currently high in the United States.

A liberal democracy cannot have an obligation to provide such an orderly avenue to future membership for the migrants who clandestinely enter or remain. Amnesty is not a bonus promised to all who persevere. Mass amnesty should be a matter for regret, an extraordinary response to a particular situation. Our current situation calls for some kind of amnesty in the near future for some large subcategory of our millions of unauthorized residents. Practical reasons may necessitate the use of simple qualifications in administering a mass legalization, but that is not because the moral context is simple.