I am about to do something highly uncharacteristic, namely to agree with Donald Trump on something. Recently, Trump made a statement, apparently shocking to many, that “I want surveillance of certain mosques.” If you know anything about terrorism and counter-terrorism, there is only one thing shocking about this remark, namely that anybody is disturbed by it. Only the shock shocks.

Put another way, look at another recent remark from British film-maker Frankie Boyle, who complains about the French and Russian bombing campaign in the Middle East. Wrong solution, he says. What we really need is “an urgent debate about how to make public spaces safer and marginalized groups less vulnerable to radicalization.” Those are excellent and critical suggestions. Lacking extensive surveillance and intelligence operations, though, such strategies are irrelevant and suicidally dangerous.

Here is the problem. In order to investigate terrorism, law enforcement agencies absolutely must use various clandestine methods, including surveillance, infiltration and double-agent tactics. Agencies wishing to suppress terrorism must, of necessity, operate in a complex and dirty clandestine world. In order to succeed, they must often do things that they cannot publicize frankly.

The best and perhaps only means to defeat subversive or terrorist groups is to keep them under detailed and constant surveillance, a strategy that must be coupled with infiltration and penetration. This needs to be stated because of the common official emphasis on enhancing security to protect airports and public buildings against terrorist attacks. At least, that is what agencies say publicly.

Americans are all too familiar with the searches they deal with before flying on commercial airlines. Some of these precautions are useful and necessary, but most have no impact whatever on the likelihood of a terrorist assault. If you fortify aircraft, terrorists attack airports; if you fortify airports, they can bring down aircraft with missiles fired from remote locations; if you defend aircraft, they attack ships; and so on. Does TSA operate perfect security at its checkpoints? Then gun down the long lines waiting to pass through the checkpoints.

It is a near miracle that nobody has yet blown up a ship packed with high explosives in a U.S. port, creating something close to the effects of a nuclear blast. To see what I am talking about, just Google “Halifax 1917”.

Overwhelmingly, security precautions and airport searches are solely designed to raise public consciousness, a feel-good strategy without any real effects or benefits. The fact that we have not to date had “another 9/11” has next to nothing to do with airport security operations. It is the result of intelligence, pure and simple.

No government can defend itself against terrorism solely by means of protection and security. This is all the more true when we consider the international dimension. Even if we assume the impossible, and the U.S. became wholly invulnerable from terrorist attack, this would still leave an almost infinite number of targets worldwide.

Think of Europe, an area somewhat smaller than the continental U.S., though divided into over 30 separate nations, with competing police jurisdictions. Most also operate within the Schengen system, so that there are effectively no internal border controls. This makes it easy for activists to escape detection, to operate freely across several countries, and to drive across several countries in a day. Meanwhile, Europe is host to hundreds of major U.S. targets, including embassies, military bases, business offices, airline offices, and tourist resorts. Add to that tourist sites where Americans gather in large numbers.

And that is just Europe: American targets are just as exposed in the Middle East, in the Asia-Pacific region, or in Latin America. There is simply no way that the U.S. can mobilize forces to defend every single potential target in the world against a global organization like ISIL/Daesh or al-Qaeda, or to place U.S. troops at every port and airport, every embassy and tourist destination. Of itself, talking about “making public spaces safer” is delusional.

Let me say this simply: in a terrorist war, any effort that is solely defensive is bound to fail. Just to protect is to lose.

The only way to defeat attacks is to prevent them being launched, and that means finding out what the extremists are going to do before they do it, and stopping them. This demands the use of surveillance and infiltrators, both what is known as electronic intelligence, and human intelligence.

But surveillance over what? In an ideal world, terrorists would wear T-shirts with the word TERRORIST printed in large letters, so they could be picked out easily by the security forces. Unfortunately, they do not do so, and strenuously resist attempts to isolate them from the people. They operate through above-ground organizations and institutions.

One critical concept here is insulation, which British counter-terrorism expert Frank Kitson defines as “a functional system of associations, clubs and other groupings designed to carry out specific tasks.” It is, in short, a means by which terrorists can hide among the larger population. Through legal and above-ground organizations, terrorists spread propaganda, and recruit. The groups can be used as testing grounds, to observe the efficiency and dedication of young militants who might eventually be drafted into the terrorist organization. In advanced stages of a campaign, they can be used to smuggle and store weaponry, and train militants in their use.

Depending on the context, those above-ground groups might take many forms—labor unions, political parties, social clubs, ethnic and religious pressure groups—but in the Islamist world, that chiefly (but not exclusively) means mosques. I could easily list a hundred European mosques that presently serve this crypto-terrorist function, and many do so quite flagrantly.

This point might be obvious, but let me say it clearly. The vast majority of U.S. mosques presently serves no such role, and their members would utterly reject radicalism, extremism, or violence. Even where there is an extremist presence, that would be absolutely contrary to the wishes of the mainstream in the congregation. The main thing the imams in those places want is to have the police help them kick out the extremists, and not to be too gentle doing so.

But any terrorist Islamist presence in the U.S., present or future, does and will use mosques in this way. If you do not maintain such mosques under surveillance—and particular “certain mosques” already leaning in radical Salafist directions—you might as well abandon any and all pretense of trying to limit or suppress terrorism on U.S. soil.

“Surveillance” in this instance emphatically means human intelligence within the mosque. That means recruiting informants within it, and trying to bring radicals over to your own side, to see what extremists are going to do before they do it. Just how and where is radicalization being undertaken? Who are the key militants? Are there weapons present? What are the overseas connections? And if that means recruiting and controlling imams and religious teachers, all the better.

These tactics are absolutely fundamental to European counter-terrorism approaches, and nobody has the slightest doubt of that fact. That fact is public knowledge, and effectively beyond political criticism. If U.S. agencies claim that they are not doing the same things right now, in American mosques, they are simply deluding the public. They will worry about the freedom of religion lawsuits later.

So, God help me on this, in this instance, Trump is right. The only thing he is doing wrong is talking about it publicly.

Again: In a terrorist war, any effort that is solely defensive is bound to fail. To protect is to lose. We must understand and accept all the necessary consequences of that fact.

Philip Jenkins is the author of The Many Faces of Christ: The Thousand Year Story of the Survival and Influence of the Lost Gospels. He is distinguished professor of history at Baylor University and serves as co-director for the Program on Historical Studies of Religion in the Institute for Studies of Religion.