The videos showing an unruly scene in and around Brookings last Thursday during the visit of Turkish President Recep Tayyip Erdoğan make for distressing viewing. The incongruity of what occurred—think Trump security meets Turkish nationalism—invites introspection about how scholarly institutions manage appearances by controversial leaders.

There are legitimate criticisms to be made of the Turkish government in general and of Thursday’s security detail in particular. But lost in the melee—and in the past year of terrorism, arrests, and media closures—is the message that Erdoğan most needed to convey.

Dramatic changes in the geopolitical neighborhood now present the most serious challenge to Turkish territorial integrity since the founding of the Republic. With the aid of Western intervention, the wars in Iraq and Syria accomplished more for the Kurdish cause than decades of terrorism and negotiation. Since the addition of a second stronghold in Syria to the de facto Kurdish territory in northern Iraq, Turkey is paying a price for conflicts not wholly of its own making.

It is not quite a century since European armies last seized Ottoman territories or supported national Kurdish independence from Istanbul. Whether or not now is the moment an autonomous Kurdish state takes legal form, the model is being proven nearby under Western protection. It does not make things easier that this time it is not Western countries’ intention to hurt Turkey’s national interests. Adding insult to Erdoğan’s injury, in 2015 the Kurdish cause met unprecedented support among urban elites around Turkey—and the United States and Europe—for a political party (HDP) that spoiled Erdoğan’s institutional ambitions by denying him a supermajority in parliament.

The Turkish president is criticized for allowing feelings of personal betrayal to color his strategic relationships—for example with Israel and Syria—yet many in the U.S. foreign policy community also now react to him emotionally. Because their high hopes were dashed after Gezi Park and the Gülen scandals, he can do no good again. This fuels Erdoğan’s outrage: Turkey gets no respect for its current role absorbing waves of refugees or for “taking the fight to terrorists.”

Erdoğan alienated Western allies with a take-no-prisoners approach in domestic politics and bears some responsibility for the government’s disastrous relationship with the country’s two major dissident groups—one ethnic (Kurdish) and one spiritual (Gülenist). But that should not relax similarly robust democratic expectations of these groups’ own political behavior, and the impression of such a double standard is at the root of the Turkish president’s annoyance. The suicide bombs and illegal wiretaps his country endured have failed to capture the American imagination. Instead, he perceives friends who would tie his hands as he defends the rule of law against terror and treason.

The Turkish government should be discouraged from abusing executive power, squelching dissent, or other acts of overzealous majoritarianism and break the cycle of retaliation against political opponents. But it is bizarre to equate the Turkish president with former Venezuelan leader Hugo Chavez or Russian President Vladimir Putin: Turkish democracy is battered but not yet sunk, and its government is still not a strategic opponent—it remains a NATO member in accession talks with the EU.

The Brookings incident is said to have exposed the regime’s true colors and thin skin, and to emblematize how polarized and undemocratic Turkey has become in the last decade. But a lopsided and illiberal democracy also preceded AKP rule: a quarter century of single party rule followed by four military coups in as many decades, with strict limits on free speech and religious exercise. American enthusiasm for democratization in the region must include a commitment to remain constructively engaged when the spring recoils and conservative parties win power—including those who appear to abuse that power—through exactly this kind of visit.

Because last week’s scene unfolded in the same auditorium where a younger Erdoğan appeared as a promising democratic leader years ago, it is fair to ask which was the real one. He who walked down the aisle with files on 57 imprisoned journalists in March 2016? Or the Erdoğan who arrived in government with dossiers on negotiations with PKK leader Abdullah Öcalan, opened talks with the Alevi religious minority, and established a ministry for European Union affairs?

Critics now say that was all just a show and diversionary tactic that have finally given way to his true attitudes towards the proverbial “tram of democracy.” But with friends who are deaf to some of Turkey’s legitimate concerns it is fair to ask what may now be an academic question: Is President Erdoğan an ex-liberal who simply got off the tram, or was he mugged by reality while on board?

Editors’ Note: The mass-casualty terrorist attacks in Paris and now in Brussels underscore an unsettling truth: Jihadis pose a greater threat to France and Belgium than to the rest of Europe. Research by Will McCants and Chris Meserole reveals that French political culture may play a role. This post originally appeared in Foreign Affairs.

The mass-casualty terrorist attacks in Paris and now in Brussels underscore an unsettling truth: Jihadists pose a greater threat to France and Belgium than to the rest of Europe. The body counts are larger and the disrupted plots are more numerous. The trend might be explained by the nature of the Islamic State (ISIS) networks in Europe or as failures of policing in France and Belgium. Both explanations have merit. However, our research reveals that another factor may be at play: French political culture.

Last fall, we began a project to test empirically the many proposed explanations for Sunni militancy around the globe. The goal was to take common measures of the violence—namely, the number of Sunni foreign fighters from any given country as well as the number of Sunni terror attacks carried out within it—and then crunch the numbers to see which explanations best predicted a country’s rate of Sunni radicalization and violence. (The raw foreign fighter data came from The International Centre for the Study of Radicalisation and Political Violence; the original attack data came from the University of Maryland’s START project.)

What we found surprised us, particularly when it came to foreign fighter radicalization. It turns out that the best predictor of foreign fighter radicalization was not a country’s wealth. Nor was it how well-educated its citizens were, how healthy they were, or even how much Internet access they enjoyed. Instead, the top predictor was whether a country was Francophone; that is, whether it currently lists (or previously listed) French as a national language. As strange as it may seem, four of the five countries with the highest rates of radicalization in the world are Francophone, including the top two in Europe (France and Belgium).

Knowledgeable readers will immediately object that the raw numbers tell a different story. The English-speaking United Kingdom, for example, has produced far more foreign fighters than French-speaking Belgium. And fighters from Saudi Arabia number in the several thousands. But the raw numbers are misleading. If you view the foreign fighters as a percentage of the overall Muslim population, you see a different picture. Per Muslim resident, Belgium produces far more foreign fighters than either the United Kingdom or Saudi Arabia. 

[W]hat could the language of love possibly have to do with Islamist violence? We suspect that it is really a proxy for something else: French political culture.

So what could the language of love possibly have to do with Islamist violence? We suspect that it is really a proxy for something else: French political culture. The French approach to secularism is more aggressive than, say, the British approach. France and Belgium, for example, are the only two countries in Europe to ban the full veil in their public schools. They’re also the only two countries in Western Europe not to gain the highest rating for democracy in the well-known Polity score data, which does not include explanations for the markdowns.

Adding support to this story are the top interactions we found between different variables. When you look at which combination of variables is most predictive, it turns out that the “Francophone effect” is actually strongest in the countries that are most developed: French-speaking countries with the highest literacy, best infrastructure, and best health system. This is not a story about French colonial plunder. If anything it’s a story about what happens when French economic and political development has most deeply taken root.

An important subplot within this story concerns the distribution of wealth. In particular, the rate of youth unemployment and urbanization appear to matter a great deal too. Globally, we found that when between 10 and 30 percent of a country’s youth are unemployed, there is a strong relationship between a rise in youth unemployment and a rise in Sunni militancy. Rates outside that range don’t have an effect. Likewise, when urbanization is between 60 and 80 percent, there is a strong relationship.

These findings seem to matter most in Francophone countries. Among the over 1,000 interactions our model looked at, those between Francophone and youth unemployment and Francophone and urbanization both ranked among the 15 most predictive. There’s broad anecdotal support for this idea: consider the rampant radicalization in Molenbeek, in the Parisbanlieus, in Ben Gardane. Each of these contexts have produced a massively disproportionate share of foreign fighters, and each are also urban pockets with high youth unemployment.

As with the Francophone finding overall, we’re left with guesswork as to why exactly the relationships between French politics, urbanization, youth unemployment, and Sunni militancy exist. We suspect that when there are large numbers of unemployed youth, some of them are bound to get up to mischief. When they live in large cities, they have more opportunities to connect with people espousing radical causes. And when those cities are in Francophone countries that adopt the strident French approach to secularism, Sunni radicalism is more appealing.

For now, the relationship needs to be studied and tested by comparing several cases in countries and between countries. We also found other interesting relationships—such as between Sunni violence and prior civil conflict—but they are neither as strong nor as compelling.

Regardless, the latest attacks in Belgium are reason enough to share the initial findings. They may be way off, but at least they are based on the best available data. If the data is wrong or our interpretations skewed, we hope the effort will lead to more rigorous explanations of what is driving jihadist terrorism in Europe. Our initial findings should in no way imply that Francophone countries are responsible for the recent horrible attacks—no country deserves to have its civilians killed, regardless of the perpetrator’s motives. But the magnitude of the violence and the fear it engenders demand that we investigate those motives beyond just the standard boilerplate explanations.