The following excerpt is taken from “The Preventive State: The Challenge of Preventing Serious Harms While Preserving Essential Liberties,” by Alan Dershowitz (Encounter Books).

* * *

The Secret Service should have five basic responsibilities: (1) short- and medium-term intelligence operations designed to predict and prevent in advance any threats; (2) securing of all venues at which protectees may appear; (3) identifying and preventing immediate threats by suspicious persons at or near the venues; (4) halting the event and the appearance of the protectee if there is any perceived threat; and (5) immediately responding to attacks by protecting the target from further attacks and neutralizing the attacker or attackers.

The first of these responsibilities — predicting and preventing future attacks — is the most difficult (and the most relevant to my book). It will necessarily require many false positives to prevent even one false negative, because attempted assassinations are such rare events over time. Many individuals and groups threaten — either explicitly or implicitly — presidents and other protectees, but very few actually act on these threats. All such threats should obviously be investigated, but not all suspects should be detained. It is a crime to threaten the president — an inchoate crime designed to reduce threats of future harms. But few potential assassins openly threaten and thus violate the statute. Some say and do things that increase the likelihood that they will actually try to attack. But even among those, only a small fraction will cross the line into action. It is the difficult — perhaps nearly impossible — job of the Secret Service to distinguish between true and false positives, while reducing false negatives to zero. It is unlikely that Crooks or Routh could have been identified before the days in question based on their words and deeds — without also identifying many individuals with similar predictive profiles who would never pose a real threat.

When I began to research the issues associated with predicting and preventing violence, including assassinations, I came across a category of detainees known as “White House Cases.” These cases involved individuals who sought entry to the government buildings, especially the White House, in order to talk to officials. If they appeared to be mentally ill or “disturbed,” they were referred to the Secret Service, which then sought their commitment to locked wards of mental hospitals. I studied one case that seemed to typify this category.

Bong Yol Yang, an American of Korean origin, appeared at the White House gate asking to see the president about people who were following him and “revealing his subconscious thoughts.” He also wondered whether his talents as an artist could be put to some use by the government. The gate officer called the Secret Service, which then had him committed to a mental hospital. Yang demanded a jury trial, at which a psychiatrist testified that he was mentally ill — a paranoid schizophrenic — and that although there was no “evidence of his ever attacking anyone so far,” there was always a possibility that “if his frustrations . . . became great enough, he may potentially attack someone.” On the basis of this diagnosis and prediction, the judge permitted a jury to commit Yang to a mental hospital until he was no longer mentally ill and likely to cause injury.

This was almost certainly an example of a false positive, who was confined in an effort to avoid even the slightest possibility that if released he would become a false negative — that is, an actual assassin whose deadly act was falsely not predicted and thus not prevented. The calculus for “White House Cases” was that it is better that one hundred or more non-assassins be erroneously detained than even one potential assassin be erroneously released.

The issue of false positives and false negatives also arose in the moments preceding the wounding of Donald Trump. According to reports, a Secret Service sniper aimed at Crooks and could have killed him before he fired at Trump, rather than immediately after. But what if Crooks had turned out to be an innocent twenty-year-old who was in the wrong place at the wrong time. His death would have been a terrible false positive. So the sniper erred on the side of not shooting, with the result that it was a near-lethal false negative.

All of this could easily have been prevented if the Secret Service had secured the building, even with its slanted roof, before the shooter got near it. A false positive in unnecessarily securing a building that would not have been used to shoot is no big deal. A false negative in not securing a building that was then used to shoot is an enormous deal. So better one hundred buildings that would not be used be unnecessarily secured than even one building that would be used be falsely not secured.

Another example of how the desire to avoid a trivial false positive resulted in a disastrous false negative was the decision — or nondecision — by the Secret Service not to delay Trump’s appearance on stage until all doubts concerning the person deemed “suspicious” were resolved. There were several people who were deemed suspicious before the Trump shooting, as there apparently are at many events. Each such person should be investigated as soon as practical. Had Trump’s appearance been delayed or even canceled and the suspicious person had turned out to be harmless, that would have been a false positive that caused inconvenience. But failure to delay caused a near-fatal false negative.

Trump has said that “nobody mentioned” the presence of a suspicious person. “Nobody said there was a problem.” He went on: “And I could have waited fifteen—they could have said, ‘let’s wait for fifteen minutes, two minutes, five minutes, something.’”

This raises a related and complex issue. Under current law and practice, the protectee has the power to overrule the Secret Service and make the ultimate decision whether to delay or cancel an event deemed dangerous by Secret Service agents. And history shows that protectees have ignored the expert opinions of the Secret Service in several instances, placing themselves in danger. It must be recognized that when a protectee’s life is endangered by a possible assassin, more is at stake than the life of the protectee. The entire nation — indeed the entire world — is affected by a successful assassination, especially of a president or presidential candidate.

Perhaps the law ought to be changed to reflect this reality. The Secret Service should have more power to delay or even cancel events that in its view pose too great a danger to the nation. This will not be easy to implement in practice, but it could have an impact on the acceptance by the protectee of the more expert opinions of the Secret Service.

To summarize, the preventive intelligence aspect of the Secret Service’s role should favor many false positives — investigative leads that turn up empty — over even one false negative as in not pursuing a lead that could have prevented an assassination. The same is true of securing venues, questioning subjects, delaying the appearance of protectees, and surrounding a protectee. Detaining a suspect for more than a brief period should require some degree of probable or at least plausible cause. Shooting a suspected assassin requires a far different calculation, because it pits life against life.

These are the kinds of judgments that must be made by all preventive law enforcement agencies. The Secret Service has a mixed history in regard to this critical mission. To be sure, we do not know how many potential assassinations were prevented by the Secret Service. Neither do they. It is certainly possible that some of those arrested or detained might have gone on to try to kill a protectee. We know about attempts that were thwarted, as was the attempted shooting of President Ford by Squeaky Fromme, or the attempt on the life of President Truman by a Puerto Rican liberation group. But potential assassins who were deterred, or even those who were prevented by early interventions, are difficult to identify and quantify. The utter failure of the Secret Service to prevent or thwart the shooting of Donald Trump should be evident to all. The resignation of the then director was an important first step in the implementation of significant changes in approach to predicting, preventing, and responding to threats to protectees. These changes should reflect the costs of different types of errors — false positives versus false negatives in differing, but always crucial, contexts.

* * *

This excerpt taken from “The Preventive State” is published by permission from Encounter Books.

Alan Dershowitz is the Felix Frankfurter Professor of Law, Emeritus, at Harvard University. Mr. Dershowitz has been litigating, teaching, and writing about law and policy for more than 60 years. He has written 55 books and more than 1,000 articles. Many of today’s world leaders are among the 10,000 students he has taught. He has represented and advised presidents, prime ministers, and business leaders. Called “the world’s best-known lawyer” and its most prominent defender of civil liberties, he has litigated and won hundreds of cases in multiple countries. He taught at Harvard for 50 years, where he offered courses on issues ranging from criminal, constitutional, family, and Jewish laws to psychiatry, neurobiology, mathematics, literature, philosophy, and even baseball. His primary academic interest has been on prediction and prevention of harmful conduct, which he developed into a course and taught during most of his career.

The views expressed in this piece are those of the author and do not necessarily represent those of The Daily Wire.