Get Access to Print and Digital for $23.99 per year.
Subscribe for Full Access

Lax Americana

As the Biden Administration escalates one conflict with a nuclear power in Europe, and stokes another in Asia, it’s encouraging to read Daniel Bessner’s call for a foreign policy predicated on restraint [“Empire Burlesque,” Essay, July]. Liberal internationalism was thoroughly discredited by the adventures of the George W. Bush and Barack Obama presidencies, only to be rehabilitated by the perceived success of U.S. policy in Ukraine. As a result, we seem once again oblivious to the risks of military escalation.

In reality, the war in Ukraine should be a stark reminder of the limits of armed primacy. Between the global economic fallout, the West’s failure to deter or restrict Putin’s criminal actions through force, Ukraine’s gradual but dramatic loss of territory, and the mission creep that’s bringing two nuclear superpowers closer to direct conflict, the war is far from the triumph many initially cheered. This should be a grim warning as the American right fantasizes about staging a similar confrontation with China.

When all you have is the military, everything looks like Europe in 1938. Even as Congress has rubber-stamped huge military budgets, it has kept the State Department under-resourced, hobbling U.S. diplomatic capability. This indifference toward non-military solutions is backed by the talking point that U.S. military presence underwrites peace. But Donald Trump’s threat to disengage the United States from the Middle East actually fueled regional diplomatic efforts, as the prospect of non-reliance on Washington suddenly made potential warfare far costlier for the region’s squabbling governments.

Such episodes give us a possible preview of the world Bessner envisions, in which states take greater responsibility for their own security. In such a world, states must cooperate with one another, not through military alliances, but by finding areas of mutual interest, of which there is no shortage.

Branko Marcetic


While Bessner does an adequate job of cataloguing the many failures of U.S. foreign policy, he fails to consider the counterfactuals of the United States’ supremacy. It is fair to note that the World Bank and the International Monetary Fund “aren’t exactly beloved in the Global South,” but it is increasingly obvious that “the glimmers of the alternative world” dominated by totalitarian China and its debt-trap economic diplomacy are far more ruthless and predatory.

The U.S.-led world order is coming to an end whether we like it or not, Bessner argues. Yet that assumption, based on the sole fact of catch-up growth in countries such as China, disregards America’s unique advantages, including its ability to build consensual, values-driven partnerships with like-minded countries.

Our decline, in other words, is a choice. If it does occur, it is bound to make the world a worse place. For all the shortcomings and hypocrisies of the U.S.-led order, the United States is the first and only superpower in history to take national self-determination and rules constraining the use of power seriously—hence the decolonization of the Fifties and Sixties, postwar Germany’s comfort with U.S. leadership, and the current embrace of the United States by small, vulnerable democracies from Estonia to South Korea.

It is surprising that an essay ostensibly concerned with empire spends no time engaging with imperialism’s most recent manifestation: Russia’s genocidal war against Ukraine. While a U.S.-led world is far from perfect, it is enough to look at the images of Irpin and Bucha, or to listen to any of our allies from Europe and the Indo-Pacific, to get a sense of the brutality of relevant alternatives.

Dalibor Rohac
Washington, D.C.


Daniel Bessner responds:

Dalibor Rohac’s response to my essay contains the usual, tired defenses of U.S. hegemony traditionally offered by the foreign policy establishment.

First, he warns of a world “dominated by totalitarian China and its debt-trap economic diplomacy,” despite the fact that China has shown little evidence of desiring global hegemony.

Second, Rohac refers to the United States’ “ability to build consensual, values-driven partnerships with like-minded countries,” happily ignoring the partnerships we presently have with Bahrain, Brazil, Egypt, Qatar, and the United Arab Emirates, as well as the relationships we have historically enjoyed with the Shah’s Iran, apartheid South Africa, various military regimes in South Korea, Mobutu’s Zaire, and many more countries that could hardly be described as democratic.

Third, he asserts that “our decline … is a choice,” ignoring the arguments I made about the material and philosophical foundations of America’s decline. It is simply wrong to state that my claims are “based on the sole fact of catch-up growth in countries such as China.”

Fourth, no honest observer of the history of the twentieth and twenty-first centuries could say, as Rohac does, that the United States has ever abided by the “rules constraining the use of power” or expressed genuine concern for “national self-determination.”

Rohac ends his letter with the canard that my essay—which is about U.S. foreign policy, especially toward China—“spends no time engaging with” Russia’s war in Ukraine. He warns about the “brutality of relevant alternatives” to U.S. empire. I would only ask him what people in Fallujah, Kandahar, or Sanaa might say about the benevolence of the “U.S.-led world” he cherishes.



Watch and Learn

Reading Hari Kunzru’s recent column [“Objectivity,” Easy Chair, July], I couldn’t help but quibble with the author’s explanation of the human observer’s role in Bayesian inference. Kunzru asserts that a Bayesian model “starts out with a guess—perhaps an informed one, perhaps not,” and that “without knowing anything, without a theory or a sense of what’s important, the machine begins to understand the world.” What he seems to be claiming is that a Bayesian model makes very few (or no) assumptions about its world before it collects observations with which to make inferences. The idea that these initial guesses matter progressively less as the model observes more about the world is contrasted starkly with the Goethean notion that the aim of a well-trained scientist is to suss out archetypal forms from a series of unreliable observations.

But in many feats of machine learning, for which Bayesian probability theory often serves as the inferential engine, the prior assumptions instilled by a model’s architects play a critical role in its interpretation of its observations. For example, a Bayesian model that “watches” videos of people performing quotidian tasks and identifies their actions makes a number of assumptions about the structure of the world giving rise to these images. In identifying a video of a person making coffee, the model makes assumptions about the probable order of the actions being performed—grinding the beans comes before pouring the water. This structured prior knowledge, created by the model’s human architects, is critical for making sense of the incomplete, disjointed stream of observations that algorithms must contend with in the real world. Without such assumptions, all that’s left is noise.

Ross Otto
Associate Professor of Psychology, McGill University


The Victim Cloud” by Hannah Zeavin [Report, July] inaccurately described Herman Melville’s book The Confidence-Man as a novella. In fact, it is a novel. We regret the error.

| View All Issues |

September 2022

“An unexpectedly excellent magazine that stands out amid a homogenized media landscape.” —the New York Times
Subscribe now