Currently, law enforcement in the United States largely uses the Reid method of interrogation, which trains police to lie to both suspects and witnesses whenever they deem it useful to do so. This essay argues that we must rethink and reform this entire approach because it is wrong for police officers to lie to suspects and witnesses during interrogation and investigation, and that outright lying by interrogators should be forbidden. The argument of this essay rests on broadly Kantian considerations about respect for persons and their autonomy: By lying to suspects, police officers attempt to subvert suspects’ capacities for reasoned action and autonomous choice for their own purposes. In doing so, they use them as mere means and thus fail to show them the respect they deserve as rational beings capable of autonomous action, which is wrong even if police officers can truthfully say that they lie to achieve worthy goals like protecting the public from criminals and assuring that the guilty are punished.
I. Why Kantianism?
Why think about the issue of police interrogation practices through the lens of Kantian theory? It will not do simply to assume that Kantianism is the right or best moral theory and then apply it to the concrete issue of police interrogations. Even many philosophers would find this question begging, and to many of those outside academic philosophy it would look like the worst sort of armchair speculation. I suggest that we should think about police interrogation through a Kantian lens because Kantianism provides the best way to make sense of what we might call two fixed points in the ethics of policing, or less metaphorically two claims that almost everyone who thinks seriously and sincerely about these issues whether philosopher, jurist, criminologist, or principled police officer agrees to: (1) it is wrong for police officers to perjure themselves in a criminal court to obtain convictions of criminals; and (2) it is wrong for police officers to engage in physical abuse as a means of interrogation. If I am correct, then Kantian moral reasoning is internal to the ethics of policing and that denying their validity in this area would have the high cost of being unable to give a good account of these principles.
I begin with police officers committing perjury to gain a conviction. Two things are obviously wrong with perjury, and the best way to explain how and why each is wrong is in Kantian terms. First, when police officers do so, they attempt to manipulate jury members or even judges into doing what they want instead of giving them true information and allowing them to reach their own decisions. This sort of manipulation is a near textbook case of using other persons as mere means: Police officers who perjure themselves seek to undermine the rational capacities of jury members or judges so that they will do what they want. And a Kantian can account for why this is wrong even if police officers can truthfully claim that perjury is most certain, or even only, way to get a dangerous criminal off the streets and give him the punishment he deserves. These are good ends, but it is wrong to use other rational beings as mere means even to such good ends.
The second intuitive way that perjury to gain convictions is wrong is that in doing so police officers are acting contrary to the norms of honesty that are necessary for the court system to function. For the court system to function, those who testify under oath must generally tell the truth as they understand it, and judges and juries must know that this is the case. One could try to account for the wrongness of police officers committing perjury to gain convictions in something like the following consequentialist terms: by lying police officers at least threaten to undermine the norms of honesty the court system needs to function, and the consequences of undermining these norms are bad. This explanation is insufficient in two ways. For one thing, it could be empirically false. If lying by police officers goes largely unnoticed such lying would do little to undermine the norms of honesty and expectation of honesty a functioning court system requires. More importantly, even if it were true, it would not explain why it is in principle wrong for police officers to give lying testimony to gain convictions. It might explain why doing so is generally bad. It could justify a policy of trying to root out such perjury by police officers and severely punishing those who are caught committing perjury to set an example. However, it could not explain why it would be wrong for a police officer who was confident both that she could get away with lying and that lying would help to convict a criminal that it would be wrong to do so. As Jerome Skolnick (Reference Skolnick, Elliston and Feldberg1985) notes, police officers themselves are hardly ignorant of the implications of such utilitarian reasoning when consistently applied. Many of them find it bizarre that lying in investigations is encouraged on grounds of expediency while it is absolutely forbidden in court even when expedient. In fact, Skolnick believes that many police officers who choose to lie in court justify it by falling back on what they see as the flat out contradiction between the ethics of lying in investigation and lying in the court room (Skolnick Reference Skolnick, Elliston and Feldberg1985: 79, 83). I believe that these officers are right to see a contradiction here but disagree with them on must how one ought to resolve it.
Kantianism on the other hand offers a natural way to explain why behavior that is contrary to the norms of honesty the legal system requires is in principle wrong. When police officers lie to the court, they act on a principle that they could not will as a universal law. If everyone lied in court whenever they thought that doing so had good consequences, then either we could not have a functioning court system, or we would have one where testimony carried relatively little or practically no weight. In neither scenario could the police accomplish their aims by lying in court.
Consider the prohibition on physically abusing suspects. Here too the most natural explanation for why it is wrong in principle is straightforwardly Kantian. John Kleinig (Reference Kleinig1996) offers a clear and convincing Kantian explanation for the wrongness of using physical abuse as a means of interrogation: ‘One of the problems with torture or, more generally use of the “third degree” [that is physical abuse of suspects] to extract truth from a suspect is its almost exceptional unseemliness, its violation of human dignity. The intentions of the torturer may be high minded—a desire to obtain truth, to see justice done—but treating someone in a way that undermines his capacity to make a rational response, literally brutalizing him—is an unseemly affront’ (Kleinig Reference Kleinig1996: 101). Physically abusing suspects is wrong because it fails to respect their ability to make rational decisions. In Kantian language, it uses them as a mere means.
Kantian reasoning then offers a natural way to make sense of principles against committing perjury to gain convictions and abusing suspects. However, parallel Kantian reasoning shows that it is wrong for police officers to lie to suspects. For one thing, police officers could not effectively lie to suspects and others to get information or confessions, if these people believed that police officers would lie whenever doing so was useful to them. They must believe that, at the very least, police officers are generally honest. Lying to suspects fails the formula of universal law in almost exactly the same ways that perjury to obtain convictions does. By lying to suspects, police also attempt to manipulate them (and others they lie to) into doing what they want. To adapt Kleinig's language, they try to undermine their capacity to make rational responses. This uses them as a mere means and so is contrary to Kant's formula of humanity. (I also suspect that the same reasoning means that most, or perhaps all, forms of coercion and manipulation in police interrogations is wrong since they also treat suspects in ways that at least attempt to undermine their capacity to make a rational decision, but these are issues for another essay).
If we embrace a Kantian account of why using physical abuse as an interrogation technique and committing perjury to gain convictions are wrong, then it seems we must also concede that lying to suspects is wrong and wrong for essentially the same reasons. One could of course reject the Kantian account of the wrongness of these actions. But when one must either hold that either they are not in fact wrong in principle or come up with an alternative account of their wrongness.
II. The Fundamentals of the Reid Method
The vast majority of interrogators in law enforcement in the United States are trained using the Reid method of interrogation as are many interrogators in Canada (Leo Reference Leo2008 provides an excellent, if highly critical, account of for the role of the Reid method in the US policing. For its role in Canadian policing, see Snook et al. Reference Snook, Eastwood, Stinson, Tedeschini and House2010). The Reid method not only recommends lies and other sorts of deceptive techniques at practically every stage of interrogation, but it is essentially built on lying to suspects. Its method is to get suspects to confess by inducing stress in them until they confess to the crime or give up information to relieve that stress. To induce such stress in the suspect and overcome his resistance to confessing the Reid method advocates a number of lies and other deceptive strategies.
The Reid training manuals use the term interrogation in a very particular sense. The Reid manuals sharply distinguish between interviews and interrogations. The official training manual for the Reid method characterizes interviews as non-accusatory and describes their purpose as gathering information. Interrogations on the other hand are accusatory and their primary purpose is to get a suspect to either confess or give other incriminating evidence against himself and his accomplices. Police are to interview all of those who might have information about the crime including witnesses and victims. However, they are to interrogate only suspects who they ‘are reasonably certain’ are guilty of a crime (Inbau et al. 2011: 3–7). The Reid manual does not forbid lying to interview subjects, and even recommend it in some circumstances. However, the manuals urge caution about lying to interview subjects.
The Reid method urges no caution about lying in interrogations, though. Because of the very way that the Reid method recommends that interrogators begin an interrogation they must lie to the subjects about the purpose of the interrogation. Interrogators are told that it is essential that they begin an interrogation by accusing the suspect of the crime and assuring him that there is absolutely no doubt about his guilt (Inbau et al. 2011: 190–202). But starting the interrogation this way creates a problem since if there really were no doubt of the subject's guilt there would be no need for an interrogation to get a confession or other evidence of the suspect's guilt. So, the interrogator must invent a false pretense for the interrogation (Inbau et al. 2011: 193). Further, the manuals emphasize that the level of certainty that investigators need as to a suspect's guilt to justify subjecting them to an interrogation falls far short of the reasonable doubt standard needed to convict someone of a crime, which itself falls short of a standard of no doubt whatsoever. This means that in most cases the interrogator consciously and deliberately lies simply by saying that there is no doubt as to the subject's guilt. Interrogators are also advised to use lies and other forms of deception liberally as the interrogation proceeds. They are encouraged to lie about the physical evidence that links the suspect to the crime as well as about potential witnesses to the crime and whether suspects’ accomplices or friends have implicated them. (See in particular the discussion of ‘theme development’ in Inbau et al. 2011: 203–55).
Police are neither forbidden from lying in interviews nor trained categorically to avoid doing so. The training manuals do not advocate lying in interviews to anywhere near the degree that they do in interrogations and urge caution in lying to interview subjects whom they do not believe to be directly involved in any crime. However, they still recommend investigators lie in interviews when doing so might be useful. For instance, because there is supposedly ‘safety in numbers’ they are advised to lie to witnesses who hesitate to give evidence out of a fear for their safety, about other witnesses coming forward (Inbau et al. 2011: 45). They are also encouraged to lie to uncooperative witnesses when they judge doing so to be useful. For instance, they are advised that they can ‘break the bonds of loyalty between the witness and the offender’ by lying about the suspect committing some act of disloyalty (Inbau et al. 2011: 336–37). The manual even advises interrogators to accuse falsely witnesses they know to be innocent of a crime when they think it necessary (Inbau et al. 2011: 337). The only time the training materials evince any sort of scruples about lying, whether to suspects in interrogations or witnesses in interviews, is when doing so is judged to be counterproductive. For instance, police are urged not to make specific promises about the legal benefits of confession—such as that it will lead to a reduced sentence—but only because doing so could render a confession inadmissible in court (See Inbau et al. 2011: 293–303, 343–54).
Perhaps more important than any of this is the role of what the manuals for the Reid method call ‘developing themes’. By this they mean that the interrogator is supposed to present the most apparently justifiable and sympathetic version of the crime to the suspect and to convince him that the interrogator subscribes to this version of events and understands and sympathizes with him, though the training manuals are clear that the interrogator should not believe this version of events nor that the crime was in any way justified. (See, in particular, the discussions ‘Theme Development’ and ‘Presenting an Alternative Question’, Inbau et al. 2011: 202–55, 293–303). Interrogators are also advised to tell suspects explicitly that by getting this sympathetic version of the crime, which they are told to describe as ‘your version’ to the suspect, personal consequences such as public shame and damage to relationships with family members and friends will be minimized.
All of this means that if lying to suspects is morally wrong and should be forbidden, then the interrogation practices of American law enforcement will have to be significantly rethought and rebuilt. To clarify: I do not mean to argue that police officers must always tell the whole truth in interrogations or even that all interrogation techniques that could be deemed deceptive must be banned. I take no position here on whether the police must correct suspects who through no fault of the police have mistaken beliefs about the evidence against them or even the law itself. Nor do I take a position on whether it is wrong to mislead a suspect without lying to them, and I admit that there are vexed philosophical questions about how to define lying and to draw the line between lying and misleading. My claim here is only that clear cases of outright lying to suspects should be banned. Just how much bite this position has will of course depend on how one defines lying, but I avoid debate about the definition of lying as, on any remotely plausible account of lying, my conclusion here has considerable practical importance. The Reid method advocates both what its training manuals and representatives themselves openly describe as lying. Moreover, any remotely plausible account of lying must admit that many of the untruths the Reid manuals recommend are clearly lies. In fact, explicitly stating things that one personally believes or even knows to be false with the intent of deceiving another person is obviously a paradigm case of lying. This is exactly what the Reid manuals tell investigators to do at will to interrogation subjects and to do even to interview subjects when they deem it useful.
III. Lying as a Response to Wrongdoing
Any argument requiring truthfulness on Kantian grounds faces an obvious obstacle. Kant is infamous for the rigid and unrealistic stance on lying that he takes in his 1797 essay, ‘On a Supposed Right to Lie from Philanthropic Motives’ (Kant Reference Kant and Gregor1996: 611–15). Kant's claim that one cannot permissibly lie to a murderer on one's doorstep to save his intended is, to say the least, implausible. Any realistic Kantian theory should allow some space for deception to defend the lives and persons of the innocent against unwarranted aggression and to prevent those with evil intentions from using others as means to their immoral ends with impunity. There is also the related issue of the use of force. Using force against someone limits or even destroys that person's autonomy and does so in an even more direct and obvious way than lying. Any plausible moral theory, Kantian or otherwise, must allow some space for using force in self-defense or to defend others from unwarranted aggression generally, and surely any plausible account of the ethics of policing must admit that the police are sometimes justified in using force. So, to avoid a pair of obvious reductios, I need to show that Kantian moral theory can allow lying to the murderer at the door and using force against others while still forbidding lying to suspects.
To do this, I examine two attempts to show how lying to the murderer at the door is acceptable, if not even morally required: Michael Cholbi's ‘The Murderer at the Door: What Kant Should Have Said’ (2009) and Christine Korsgaard's ‘The Right to Lie: Kant on Dealing with Evil’ (1986) to see whether either could justify lying to suspects for information or to elicit a confession. If they cannot, then this suggests that Kantian moral theory can avoid obvious reductios while maintaining that lying to suspects is morally wrong.
I focus on these specific accounts because their reasoning for why lying to the murderer at the door seems to parallel the two main attempts at justifications the Reid training manuals give for lying. According to the Reid manuals one reason that lying is justifiable is that it is necessary to protect the innocent from criminals and another is that it is justified as a response to the lying of suspects and recalcitrant witnesses. Korsgaard's account explains the justifiability of lying to the murderer at the door largely in terms of the fact the murderer himself is lying, while Cholbi holds that lying is morally required in terms of allowances for the defense of innocent lives in the face of aggression. If neither of these accounts justifies lying to suspects then it offers powerful evidence that the reasons offered for lying by the Reid manuals cannot excuse much less justify lying.
As one would expect Korsgaard's argument is detailed, creative, and subtle, but at its root are two basic moves. First, Korsgaard argues that while Kant's formula of humanity would rule out lying to the murderer at the door the formula of universal law, when properly understood, would not (Korsgaard Reference Korsgaard1986: 328–34). Second, she argues that the formula of universal law and formula of humanity are on some level distinct and are appropriate for different sorts of circumstances. Korsgaard argues that the formula of humanity is an ideal principle that holds in situations where others are at least trying to comply with the principles of morality, while the formula of universal law is appropriate to non-ideal circumstances where others are not even trying to follow the principles of morality (Korsgaard Reference Korsgaard1986: 346–48). This means that lying to the murderer is acceptable. Since the would-be murderer is bent on flouting moral principles and using the person who answers the door as a tool to help him do so, the formula of universal law is applicable here. Because the murderer is himself lying one would have no problem universalizing the principle that everyone will lie to people like him in circumstances like this.
Initially, it seems that Korsgaard's strategy would also justify police interrogators lying to many suspects, and we should note that one of the main rationales that the Reid manuals offer for the practice of lying to suspects is that they routinely lie to investigators. They offer the same reasons for lying to some witnesses. They also emphasize that if suspects did not lie to the police, police officers would not need to lie to them. Have not suspects, as well as witnesses who lie, placed themselves in the same situation as the murderer at the door? Could we not justify lying to them as a response to their own dishonesty?
According to Korsgaard's analysis would not the formula of universal law be applicable here and would not the formula of universal law allow the police to universalize lying in cases where they believe suspects, or even witnesses, are lying to them? It would also seem to give the right result when it comes to police officers lying to judges and juries. Since neither the jury members nor the judge have lied to the police, the more stringent formula of humanity holds and lying is forbidden as it would use the judge and jury members as mere means. It appears that we have an explanation for why lying to suspects is permissible even on Kantian grounds and one that preserves the Kantian explanation for the wrongness of perjury to obtain convictions.
However, Korsgaard's strategy for dealing with the murderer at the door does not justify the sorts of lying police interrogators engage in, and it does not because we could not universalize this policy of lying. Part of the reason that lying to the murderer at the door passes the formula of universal law is that we presume that he is lying to the person who answers the door. However, as Korsgaard emphasizes, another essential reason that lying to the murderer passes the formula of universal law is that it is a rare and odd case. On the other hand, an official police practice of lying to suspects is a common policy that a large and prominent group of public officials follow. Moreover, in the murderer at the door case we do not suppose that the murderer publicly identifies himself as a murderer and a liar, which means that he does not know that we know he is both. Police officers do publicly identify themselves as such. If everyone knew that police interrogators always felt free to lie whenever they even had a hunch that interrogation suspects might not be telling the truth and to witnesses when useful, then these lies would become entirely ineffective.
One could counter that given the facts about the law enforcement system in the United States, it must be possible to universalize principles that allow the police to lie. After all, police officers, or at least those in the United States, have adopted a more or less universal policy of lying to those they interrogate when they think it useful, and nonetheless they still obtain confessions from suspects and other useful information from those they interrogate by doing so.
This last statement might be true, but it brings to light the deep issues in just what it means to will a maxim as a universal law. Police officers and other interrogators trained in the Reid method can often lie effectively, but the fact that they can do so rests on the fact that those they interrogate do not know that interrogators lie so freely and so often. The Reid method simply could not be effective on anyone who knew its workings. No one who knows how likely the police were to lie to them likely would voluntarily undergo interrogation and if forced to do so would insist on having an attorney present and would dismiss any claims that their guilt was beyond doubt.
The fact that the police can use deception effectively rests entirely on the fact that most of the people they try to deceive do not know that they have such policies of deception. This is a point that the training manuals for the Reid method are themselves quite clear on. For instance, they expressly admit that, ‘the average citizen has little appreciation for the pervasive efforts required to convince a guilty suspect to offer admissions against self-interest’ (Inbau et al. Reference Inbau, Reid, Buckley and Jayne2011: 100).
One can speculate as to some likely reasons why police interrogators could adopt a policy of lying and still lie successfully. For one thing, as the Reid manuals themselves admit, while most of the general public has likely seen enough crime dramas to have a vague idea that the police sometimes use deception in interviews, they do not know how far they are allowed to go with such deceptive practices or how willing they are to use them. As Richard Leo notes many people assume that the police are trustworthy and wrongly assume either that police do not lie or that there are at least some things that they would not lie about such things as the effectiveness and results of polygraph tests (Leo Reference Leo2008: 105). I would also wager that much of the effectiveness of lying by the police can be traced back to the fact that those who are the most likely subjects of police interrogation are of relatively lower socioeconomic status and educational level, which means that they generally know less about police procedure and the workings of the legal system than the average member of the population might. Could a maxim of lying that relies on its targets being uninformed or even misinformed in these ways really be universalized in the sense that Kantians have in mind when they talk of the universalization test?
It could not. For one thing, if it could be then the universalization procedure would okay some absurd results. Take Kant's case of lying to a friend to get money. Suppose the friend in question was someone who I knew to be terrible at spotting lies but who thought she was quite savvy at spotting any and all forms of deception. Some maxim of roughly the form, ‘I will lie to extremely credulous people who do not realize they are credulous to get money whenever I need money and can get it by doing so’ could be universalized in the same way as it seems lying to the murderer at the door could be. Even if my friend knew there was such a universal policy, I could still lie to her since she does not even suspect that she is one of the credulous people to whom this maxim applies.
Now of course, in this case, my friend is not behaving immorally, and so on Korsgaard's suggestion the formula of humanity would still be operative, and the formula of humanity would forbid such deception. However, we can imagine cases where other people lie, and so remove themselves from the formula of humanity's protection and where we can formulate maxims that allow lying but pass the formula of universal law, but where it would still clearly be wrong to lie to them. Suppose someone is desperate for money and decides to sell an antique desk she owns and which she believes to be somewhat rare and valuable. To try to keep herself from being taken advantage of she lies to the potential buyer about her reasons for selling it—say she says, ‘Oh I just want to get rid of it since it doesn't match the décor’. However, the buyer sees through her lie and realizes not only how desperate she is, but also that the seller has a piece that is much more rare and valuable than she knows it to be. Would it be morally acceptable for the buyer to lie to the seller about what type of desk she had and how much it was worth? Even most people who are not committed to Kantian ethics would say that this is clearly wrong and that the seller's own lie does not give the buyer license to lie in this way. But we could tailor a maxim here that allows it. Say something of the form, ‘I will lie to people who are desperate and do not have enough information about antiques to spot my lies about the type of and value of their antiques’. Even if there were a universal principle of lying of this sort, lying in these circumstances could still work because many people may not realize that they know much less about antiques than they think, and of course desperate people often fail to do due diligence even when they have the knowledge and abilities that render them capable of doing so in normal circumstances.
It is an odd form of Kantianism indeed that allows us to frame maxims that pass the formula of universal law because of the ignorance, desperation, or other cognitive weaknesses of some people. Surely it is not enough that an action could be universalized and remain effective because some people might remain ignorant of it, lack the knowledge to realize that the maxim applies to their own situation, or are too desperate or stressed to think clearly and apply the maxim to their situation. If that is so, the fact that lying and other deceptive interrogation techniques might prove effective even if universally used by interrogators due to the ignorance, stress, or other impairments of those they interrogate is not enough to make a policy of lying and other deceptive interrogation techniques acceptable.
Of course, Korsgaard realizes that some ways of thinking about the formula of universal law make it too easy to satisfy as her comments on Kant's publicity criterion make clear (Korsgaard Reference Korsgaard1986: 330). My point is not that Korsgaard herself would think deceptive interrogation practices acceptable but rather that one could not use her reading of Kant to defend such practices without opening up her strategy for Kantianism to allow some lying to allow an absurd amount of it. (I do think the considerations here indicate that Korsgaard's solution to the problem created by the murderer at the door case might come at the cost of making the empty formalism charge against Kant's ethics much worse. I suspect severing the formula of humanity and the formula of universal law in the way Korsgaard does is responsible for many of the difficulties we see here, but all that is a subject for another essay.)
The rationale that Cholbi offers for lying at the murderer to the door—that it is necessary to defend the innocent against unwarranted aggression—is also a rationale that the Reid training manuals offer for lying to suspects. Just as lying to the murderer at the door is justified because we must do so to protect his victim, lying to criminals is justified because police officers must do so to protect their future victims. Cholbi argues that according to Kant's own writings we have a moral duty to defend our own lives and that this duty not only allows us to lie to save our lives but can even require that we do so (Cholbi Reference Cholbi2009: 24–27). He then argues that according to Kant if we have a duty to perform an action to protect our own rational nature, we have a duty to perform similar acts to protect the rational nature of others (Cholbi Reference Cholbi2009: 34–45). Since we have a duty to lie to save our own lives from unwarranted aggression, we have the same duty to lie to save the lives of others from such aggression. If this is so then we are not only allowed to lie to the murderer at the door, but we are morally required to do so.
On Cholbi's account, police officers, and others, could use force to protect themselves and others from unwarranted aggression. In fact, officers have a duty to protect themselves and if Cholbi is right they have the same duty when the rational nature of others is threatened. When someone directly threatens the lives or persons of other people or even intentionally creates a situation that does so, then we are morally justified in doing things to him that would otherwise be forbidden as a violation of his autonomy to defend ourselves and others. The rationale that Cholbi offers for lying to the murderer at the door is quite similar for the rationale that most forms of just war theory offer for making exceptions to the requirements of deontological moral theory. Does Cholbi offer a strategy that would justify the sort of lying that the Reid method advocates?
He does not. If this strategy were successful in showing that lying to criminals in interrogations is justifiable it would still not justify lying to witnesses. Even if a Kantian or other deontological moral theory allows us to do things to another person that would otherwise be forbidden when that person threatens the lives or persons of others, those allowances do not extend to people who have done nothing to create the threat in question. In cases of interrogation investigators may seem to have a rationale that better parallels what Cholbi takes to be the reason for lying to the murderer: we lie to protect the public from criminals, and we should do so in the same way that one should lie to protect the innocent person hiding in his house. However, even the analogy between suspects in interrogations and the murderer at the door and others who threaten the lives or persons of others does not stand up to closer scrutiny.
For one thing, we must reckon with the fact one has absolute certainty of the murderer's bad intentions while investigators can claim no such certainty in interrogation. Moreover, there is every reason to believe that there should be a higher standard of justification for lying in an interrogation than there is for lying in a case like the murderer at the door. Suppose that we do have some reason to doubt that the person at the door is a murderer. The cost for telling the truth to him if he is in fact a murderer is the death of his victim (and quite possibly our own as well). Even if we grant the claim that police lie to suspects to prevent them from harming future victims with the same or similar crimes, it is not the case that the police have reason to believe that this person will immediately commit a crime if they do not lie to him. In practically no real cases do they have any reason to believe that failing to lie to a suspect will directly lead to the death of another person. Moreover, the potential harm to the subject of interrogation is greater as well. It is hard to see how lying to the murderer at the door really harms him, but since the aim of an interrogation is to gain a confession or gather information to use against suspects it is quite easy to see how these lies can harm the subject of an interrogation. For these reasons alone we should set the bar for justifying lying in the murderer at the door case lower than we would in cases where the harm we aim to prevent is less imminent and immediate.
This brings me to yet another difference between the rationale police and the Reid manuals offer for lying to suspects and the reasons for thinking that we are justified in lying to the murderer at the door. Most of the people whom the police interrogate who have actually committed crimes have not murdered anyone, nor is there any special reason to think that they will in the future. In fact, many of them have not even committed violent crimes. We may have a duty to lie to protect the person hiding in our house from murder, but that does not mean anyone has a duty to lie to prevent someone from picking another person's pocket or attempting to sell that person drugs. This also shows how the analogy between lying to suspects and police justifiably using force to protect themselves and others breaks down. The police are not justified in shooting or beating a suspect to stop him from committing theft, selling drugs, or similar offenses. They are only justified in doing so when that suspect's actions present a direct threat to the lives and physical persons of others.
It appears strategies like Cholbi's that allow us to lie to the murderer at the door argument could at most justify interrogators lying to suspects only in cases of murder or, perhaps, other violent crime. However, it does not license even that because there is a deep disanalogy between lying to the murderer at the door and lying to a suspect, even one who actually is a murderer.
The disanalogy here at root goes to what someone who lies to the murderer at the door and an interrogator who lies are trying to do. If someone lies to the murderer at the door, she is trying to prevent him from carrying out a specific murder that he intends to commit. In many ways, lying to the murderer at the door is like shooting someone who is taking aim at another person with a loaded gun or charging them with a knife. The murderer at the door is in the process of attacking another person, and by lying to him or shooting him, we interrupt that process. That is accurately described as defending the life of another person.
Someone who supports using deceptive interrogation techniques could argue that by doing so they are protecting the criminal's future victims. However, this rationale would not license any claim of defense of the lives of others. An analogy with using force against someone to prevent him from harming or killing others is also useful here. One could not justifiably shoot another person on the grounds that while he was not currently threatening the life of anyone, he had threatened the lives of others in the past and was likely to do so in the future. To justify using force against another person he must intend to harm someone and pose an imminent threat of doing so. On Cholbi's account, the same conditions hold for lying.
The distinction just war theory draws between preemptive and preventive war is useful here. Preemptive war means going to war when one has good evidence that another country is planning an attack and has made preparations for an imminent attack. Preventive war on the other hand means attacking simply because one thinks that future conflict is likely even though the other country has made no preparations or plans for an imminent attack. Preemptively using force against another person and preemptive lying can be justified in the same way and for the same basic reasons that preemptive wars may be justified. However, both preventively using force and preventively lying are not acceptable. Lying to the murderer at the door is a case of preemptive lying while lying to suspects is, at best, a case of preventive lying, and so could not be justifiable under Kantian theory. The only possible time that lying to a suspect could be justifiable under Kantian theory is when doing so is necessary to stop some process the suspect himself has set in motion that threatens others. So then in so-called ticking time-bomb cases the police may be justified in lying to a suspect if we accept an account like Cholbi's. However, such cases are vanishingly rare. The mere possibility of justifiable lying in incredibly rare cases like this hardly justifies the current policy of allowing and encouraging police interrogators to lie to all suspects.
IV. Lying as a Professional Obligation?
So, then the rationale that the Reid manuals and many police officers offer for lying to suspects fail. However, there is one final strategy for defending the practice of lying by police interrogators that we should consider. One could argue that police officers have some sort of special role-based obligation that allows or even requires them to lie, and which trumps ordinary moral obligations. This idea is hardly an unprecedented one. After all, most legal ethicists think that this is true for attorneys in at least some situations. For instance, if offering a theory of the crime that she did not personally believe were the best strategy for defense attorney to defend her client, then most legal ethicists would think her justified in doing so. Many would even hold that she would be morally obligated to do so. Could something similar not be true when it comes to police officers and lying?
Now we should note that drawing an analogy between lying by police officers and this sort of lying by attorneys is potentially very misleading as they differ in important ways. For one thing, the reason that many of us are willing to grant that defense attorneys are required to engage in this sort of dishonesty is that doing so is part of their role in serving as a check on the power of the state. Granting that sort of power to a representative of the state is a different thing entirely. Perhaps more importantly, no one has to believe that the defense attorney is telling what she personally believes to be the truth for the sort of dishonesty they justifiably engage in to be effective. Indeed, a juror who expected that defense attorneys would say only what they personally believe to be true would be naïve to the point of incompetence. Moreover, a juror need not in any strict sense believe what the defense attorney says for it to be effective. Suppose that a juror in an arson case heard a defense attorney's theory of crime of how the fire could have been accidentally set and assigned it only a low probability (say, 25 percent). This would not qualify as belief on any plausible definition of the term and yet this may well be enough to qualify as reasonable doubt (it likely would for me if I were a juror). When conjoined with other sources of doubt it could certainly serve to tip the scales to reasonable doubt even for someone who interprets this bar as standing much lower than I would. All of this means that even a straightforwardly Kantian moral theory need not condemn such dishonesty. We could likely universalize the practice of at least some dishonesty by defense attorneys since its effectiveness does not depend on a belief that defense attorneys tell the truth as they understand it, and such dishonesty does not treat other people as mere means in any obvious way. As we have seen the same cannot be said of lying by police interrogators.
We can leave all of this aside though because even if we accepted that considerations like those that justify dishonesty by defense attorneys could justify some dishonesty by police, it would still not justify the sorts of lying the Reid method is based on, or indeed any outright lying by police officers at all. This is because such a claim to a special role-obligation that trumps the demands of ordinary morality depends on the otherwise immoral behavior being in some strong sense necessary to fulfill the wider obligations of the role in question. Consider the framework Kleinig (Reference Kleinig1996) develops for evaluating whether claims by police that their special role obligations trump ordinary morality are true. Kleinig draws heavily on the ‘fourfold root of sufficient reasoning’ that David Luban develops to evaluate claims by attorneys to special role-based obligation (for Luban's own account, see Luban Reference Luban1988: 128–47). On Kleinig's slightly modified account role obligations justify departures from ordinary morality only if ‘1. The [otherwise immoral] conduct is essential to the fulfillment of the role obligation. 2. The role obligation is essential to the role. 3. The role is essential to the institution. 4. The institution is justified’ (Kleinig Reference Kleinig1996: 49–50).
By essential both Luban and Kleinig seem to mean something like necessary. So, the question becomes is lying to suspects necessary. Necessary and its cognates are always very hard words to parse in philosophical discussions, but at the very least necessary means something quite a bit more stringent than inconvenient, expensive, or inefficient. Even if one could show that a system that did not allow lying by police cost more, used more resources, allowed more criminals to get away with their crimes, or even all of these things that would not be the same as showing that lying is necessary for effective policing. A claim of necessity is a much stronger thing than claiming simply that lying is justified purely on consequentialist grounds.
Is lying then necessary for police to perform their professional obligations? That is it necessary for them maintain order and effectively enforce the law? There is little to no evidence that it is, and quite powerful evidence that it is not necessary for effective law enforcement. In fact, there is reason to believe that even on purely consequentialist grounds the practice is not justified. For one thing, many countries ban lying and other deceptive interrogation practices outright and yet do a perfectly good job of maintaining safety and order. German legal practice may allow some deceptive interrogation practices, though the issue is controversial, but it is quite clear in flatly prohibiting police from lying to suspects in interrogations (Ross Reference Ross2008). However, by any measure, the crime rate in Germany is much lower than it is in the United States. German police also clear a somewhat higher proportion of cases than do their American counterparts. (See Jehle Reference Jehle2015: 11 for German crime and police clearance rates; see FBI 2020 for the American figures.)
One could of course point out here that the German legal system is different than the American one in many respects so that what works in Germany might not work in the United States. However, recent developments in the United Kingdom, which has an adversarial legal system quite similar to that of the United States, provide even stronger evidence that lying to suspects is not necessary for effective law enforcement. In recent years, England and Wales have reformed police interrogation so that the features most hostile to the autonomy of interrogation subjects have been removed. The United Kingdom's Police and Criminal Evidence Act forbids police from lying about evidence and required them to record their interrogations, which led to a fundamental transformation of interrogation in Britain with interrogations becoming much less confrontational and coercive, but the rate of confessions obtained by police has not gone down (see Leo Reference Leo2008: 326; Sear and Williamson Reference Sear, Williamson, Canter and Alison1999). Many law enforcement agencies in Canada—most notably the Royal Canadian Mounted Police—have also moved away from the Reid method and there is no evidence that these agencies have become less effective as a result (see Rose Reference Rose2019; Gazette 2017).
Conclusion
In the last section I presented evidence that lying by police is not necessary for them to perform their duties. I believe that the truth of the matter is that the current practice of police interrogation in the United States is wrong even when judged on purely utilitarian grounds, and that even a purely utilitarian treatment of police interrogation would heavily restrict if not outright ban lying as an interrogation and interview tactic. To my knowledge none of the Reid method's defenders have offered much in the way of evidence that it is an especially effective method of interrogation; they usually just seem to assume that lying to suspects must be more effective than telling the truth. The experience of the United Kingdom and the Royal Canadian Mounted Police suggests that it is not the case, and that non-deceptive and less coercive interrogation techniques can be just as effective as lying to suspects.
Moreover, methods of interrogation that rely on lying have hidden, but high, costs, a point which Jerome Skolnick and Richard Leo make powerfully in ‘The Ethics of Deceptive Interrogation’ (1992). The biggest of these is that lying by the police, especially any widespread practice of lying, creates distrust between the police and the communities they are supposed to serve. This is especially so in cases of marginalized communities. This hampers effective law enforcement in two main ways. For one thing, it makes witnesses and others with information much more hesitant to cooperate with the police. It also makes jurors discount evidence and testimony presented by the police, which makes it much harder to obtain convictions. In one particularly striking case presented by Skolnick and Leo jurors acquitted a defendant largely because they refused to believe police officers on how long it took to drive from one side of Oakland to the other (Skolnick and Leo Reference Skolnick and Leo1992: 9).
The utilitarian case against lying to suspects and my own Kantian case are not rivals. In many ways the utilitarian case supplements the Kantian case since the stronger the pure utilitarian case against the Reid method in particular and the whole practice of lying to suspects generally are, the less ground there is for any plea of necessity which might excuse them on Kantian grounds. Further, I do not think that any plausible Kantian treatment of this subject can ignore consequences entirely. If the consequences of doing away with the Reid method and forbidding lying were catastrophic that would be a reason to maintain both practices no matter what Kantian theory might say.
However, the Kantian case against lying to suspects is a stronger and clearer one than the purely utilitarian in several ways. For one, on a purely utilitarian treatment, neither the proponents of lying to suspects nor its opponents clearly have the burden of proof. The Kantian case makes it clear that it is those who claim lying to suspects is justified who have the burden of proof. Moreover, on a utilitarian treatment if the consequences of lying to suspects were marginally better, then the practice would be justified. On a Kantian treatment lying to suspects requires much more justification than this. I doubt that the proponents of lying to suspects can find the justification they need even if we judge the matter in purely utilitarian terms, but I am near certain they cannot justify the practice if we approach the issue using Kantian moral theory.
Finally, the Kantian case is clear cut in a way the utilitarian one is not. There are compelling reasons to think that the Reid method and the overall practice of lying to suspects are not optimal in utilitarian terms. However, it is difficult, if not impossible, to prove this conclusively. It is not as though we can run an experiment or even simulation comparing systems that allow lying by the police to those that do not 1,000 times and compare the results. From a Kantian perspective though it is clear that lying to suspects violates their autonomy and fails the universalization test. It is then clearly wrong, and we do not need to know the exact consequences of allowing or banning lying to say this. The practice of lying to suspects is very likely a bad idea in that its costs outweigh its benefits, but even if it were to turn out to be expedient, it is still wrong.
