Nudging, Bullshitting, and the Meta-Nudge

Abstract

In “Nudging, Bullshitting, and the Meta-Nudge”, the author responds to William Simkulet’s claim that nudging is bullshitting (according to Harry Frankfurt’s analysis of bullshit and bullshitting), and therefore nudging during the process of informed consent renders consent invalid. The author argues that nudging is not necessarily bullshitting and then explains that although this issue is philosophically interesting, practically speaking, even if nudging is bullshitting, it does not follow that nudging necessarily renders informed consent invalid. This is obviously true in those situations in which nudging during the process of informed consent is unavoidable. The author concludes with a discussion of the meta-nudge and suggests that physicians can use the meta-nudge to eliminate or decrease the power of inappropriate, problematic, or undesirable nudges.

Introduction

In Nudge: Improving Decisions About Health, Wealth and Happiness, Richard Thaler and Cass Sunstein explain that nudging, which involves creating a choice architecture “that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives,” has many applications. ¹ A well-known example of nudging involves a cafeteria worker who can increase the likelihood that students choose healthy foods if the worker moves the healthy foods to the beginning of the line and places them at eye level. This choice architecture has the potential to increase student (and societal) well-being without limiting the students’ choice set or changing the prices of the available foods. ² Increasing societal well-being without limiting choices or altering prices should, according to Thaler and Sunstein, be acceptable (or even desirable) to those on the political right (especially libertarians) and those on the left. ³

Thaler and Sunstein explain that we can use simple nudges in a variety of healthcare situations, including the decision to register as an organ donor. In the United States, unless a person affirmatively chooses to be registered as a donor, the person retains nondonor status, that is, the default status is nondonor. One criticism of this default is that the donor status of many people is inconsistent with their stated desires. In the United States, the percentage of people who state that they support organ donation is far lower than the percentage who are registered as donors. ⁴ Might the percentage of people who are registered as donors be increased if the default was changed from nondonor to donor, if the system was changed from an opt-in to an opt-out system? ⁵ The answer to this question is complicated, but regardless of the answer, changing the default in this manner is consistent with the requirement that nudges neither decrease the choice set nor change economic incentives. Furthermore, and this is crucial, changing the default is considered a nudge because the change is aimed at increasing the likelihood that a person’s registration status will be consistent with the person’s desires or values. Thaler and Sunstein explain that nudges must be aimed at increasing the likelihood that people will make choices “that will make choosers better off, as judged by themselves.” ⁶ Thaler and Sunstein believe that this requirement should make nudging acceptable to both liberals and conservatives (especially libertarians). They state: “For all their differences, we hope that” Democrats and Republicans “might be willing to converge in support of some gentle nudges.” ⁷

The issue of whether it is permissible for a physician to nudge her patient to consent to what the physician believes is the best treatment for her patient has received considerable attention as of late. Consider a situation in which a physician believes that surgery is best for her patient (and that surgery is consistent with her patient’s values). This physician, who is familiar with the research and literature surrounding nudging, believes that wearing surgical scrubs (as opposed to a simple white coat) will increase the likelihood that this patient will consent to surgery. Accordingly, the physician chooses to wear scrubs and thereby nudges her patient to consent to surgery. Or, using a different nudge, the physician informs her patient about the success rate of surgery (say 85%) as opposed to the complimentary failure rate (15%), knowing that empirical research demonstrates that discussing rates of success as opposed to complimentary rates of failure increases the likelihood that patients will consent to surgery. ⁸ Some, like Shlomo Cohen, claim that nudging in this situation is consistent with both the physician’s duty of beneficence (and nonmaleficence) and her duty to respect patient autonomy and is, therefore, ethically permissible (or even obligatory). ⁹ Others argue that nudging during the process of informed consent is unethical. Eric Chwang, for example, claims that nudging may render informed consent invalid in that it does not respect patient autonomy. Chwang claims that the lack of respect for patient autonomy associated with nudging is similar to the lack of respect for patient autonomy associated with obtaining informed consent from a patient who is ignorant. ¹⁰

William Simkulet takes the debate concerning nudging during the process of informed consent in a different direction. Simkulet argues that nudging during the process of informed consent is what Harry Frankfurt calls bullshitting. He furthermore claims that because nudging is bullshitting, nudging renders informed consent invalid. ¹¹

In this essay, I attempt to demonstrate that nudging, like many other behaviors, is complicated and complex and cannot, therefore, neatly fall under the umbrella of bullshiting or not bullshitting. I then explain that although this debate is philosophically interesting, practically speaking, even if nudging is bullshitting, it does not follow that nudging necessarily renders informed consent invalid. This is obviously true in those situations in which nudging during the process of informed consent is unavoidable. I conclude this essay with a discussion of what I call the meta -nudge, and I claim that physicians can use the meta-nudge to undermine the power of inappropriate or problematic nudges.

Simkulet on Nudging

Simkulet, in the first of two essays in which he argues that nudging is bullshitting, explains that the duty to obtain informed consent requires a physician to provide her patient with adequate disclosure of the risks and benefits associated with the different treatment options. He then states that he will “assume, not argue, that adequate disclosure requires truth-telling.” ¹² Simkulet claims that nudging is inconsistent truth-telling, “thus a physician who nudges does not provide adequate disclosure to her patients, and thus cannot reasonably conclude that she receives genuine informed consent from her patients.” ¹³

Put more formally, Simkulet argues:

1. 1. If a physician does something that is inconsistent with truth-telling during the process of informed consent, the informed consent is rendered invalid.

2. 2. Bullshitting during the process of informed consent is inconsistent with truth-telling.

3. 3. Therefore, bullshitting during the process of informed consent renders the informed consent invalid.

4. 4) Nudging is bullshitting.

5. 5) Therefore, nudging during the process of informed consent renders the informed consent invalid.

I hope to demonstrate that premises 1 and 4 are false (or may be false).

Although Simkulet claims that premise 1, truth-telling is necessary for valid informed consent, is uncontroversial, I suggest that it is (or should be) controversial and that it may, in fact, be false. Simkulet does recognize that one might argue that premise 1 is false, but he asserts that claiming “that genuine informed consent does not require truth-telling … seems to undermine the rationale for requiring informed consent in the first place.” ¹⁴

For the utilitarian or consequentialist, lying during the process of informed consent is not necessarily wrong or problematic. If lying maximizes the good or happiness or brings about the best state of affairs, lying is permissible (or obligatory). ¹⁵ But those of us who are committed to a notion of patient rights that grounds rights in something other than utilitarianism or consequentialism would claim that although consequentialism may be related to the duty of beneficence (and nonmaleficence), it does not adequately explain or justify the duty to respect patient autonomy, which purportedly grounds the patient’s right to be told the truth.

As stated above, Simkulet claims that the rationale for informed consent, which he believes is grounded in the obligation to respect patient autonomy, is undermined when the requirement of truth-telling is not strictly upheld. I suggest that this claim is not obviously true and may, in fact, be false. Physicians do have a duty to respect patient autonomy, but this duty is one duty among other duties, and it is not lexically prior to these other duties. There may situations, for example, in which the duty to respect patient autonomy conflicts with the principle of nonmaleficence, primum non cere, above all do no harm, which dates to Hippocrates. Even those who attach a lot of weight to the duty to respect autonomy do not claim that this duty always trumps the duty of nonmaleficence (or the other physician duties). In an essay defending the value of the four principles of bioethics, Raanan Gillon famously stated: “I personally believe that emphasis on respect for autonomy is in many circumstances morally desirable and why I personally am inclined to see respect for autonomy as primus inter pares—first among equals—among the four principles.” ¹⁶ Nevertheless, Gillon does not claim that autonomy always trumps the other three principles. Gillon states:

To give but one example, the nonprovision of a universal health service in the richest country in the world (in contrast to its acceptance of what seems to be a universal gun service) is in my view too, an example of a political infrastructure that gives excessive weight to respect for individual autonomy over concerns to benefit the sick. ¹⁷

Gillon’s point is that sometimes the principle of beneficence (or one of the other principles) may trump the principle of autonomy.

It might be helpful to see lying as being prima-facie wrong. Thus, even if one believes that a physician has a duty to tell her patient the truth, there may be situations in which the prima facie duty to tell the truth can be overridden by a stronger or more pressing duty. I would like to explore several situations in which lying during the process of informed consent is (likely) permissible.

The most obvious situation in which the prima facie wrongness of physician lying during the process of informed consent can be overridden is when the duty to respect patient autonomy conflicts with the duty of beneficence (or nonmaleficence). This conflict can be conceived of as a conflict between the physician’s deontological commitments and her consequentialist commitments (or obligations grounded in beneficence or nonmaleficence). This type of conflict requires the physician to engage in a balancing act, and sometimes her consequentialist commitments outweigh her deontological commitments. ¹⁸ Daniel Sokol makes a similar point: “Although deception in medicine is generally wrong, as it tends to undermine patients’ autonomy and erode the trust between doctor and patient, the ethical duty to be honest is not absolute. Some moral goods, such as the avoidance of severe physical or emotional suffering and the preservation of life or long-term autonomy, may over-ride the prima facie duty not to deceive.” ^{19 , 20}

A second situation that may allow for lying during the process of informed consent is pointed out in the quotation above, namely when lying, which normally conflicts with the principle of autonomy, increases patient autonomy. Consider a patient, James, an academic who loves to joke around and rarely pays close attention to his physician, Dr. Benson. In the past, when conversations pertaining to James’s health started to get serious, James said something funny, often self-deprecating, and tried to change the topic by asking Dr. Benson what she thought about a paper he was planning to present at an upcoming conference. Dr. Benson discovers that James has a rare and fatal cancer, and she knows that there are two reasonable treatment options. The first involves chemotherapy and radiation and is accompanied by painful side-effects, but it is expected to provide James 6–8 months of life. The second results in much less pain, but James’s life expectancy will likely be 2–3 months. Dr. Benson begins explaining these two options, but James keeps interrupting and making jokes. Dr. Benson tries to get James to listen to her, but she is concerned that James is not paying attention and does not understand the details pertaining to each of the options, so she tells James: “I had a family member, a cousin, who had this form of cancer, and she died 2 months after being diagnosed. This is serious. Please pay attention.” When Dr. Benson is confident that James is paying attention, she informs James of the different treatment options, the implications of the different options, the costs, and so forth. James consents to chemotherapy. In truth, Dr. Benson did read a case study involving a patient who experienced what she described, but this patient was not a family member.

Simkulet, when discussing what constitutes lying, asserts that he utilizes “the most widely accepted definition of lying.” ²¹ This definition states: (1) A does not believe X; (2) A acts to convey X to an audience; and (3) A intends the audience to come to believe X (or at least, that the audience comes to believe X). Based on Simkulet’s definition, Dr. Benson lied. She did not believe that one of her family members had the same type of cancer that James had; she conveyed to James that one of her family members had the same type of cancer that he had; and she intended James to come to believe that one of her family members had the same type of cancer that he had. I suggest that even though Dr. Benson lied to James during the process of informed consent, the informed consent was not rendered invalid.

Simkulet might respond by asserting that Dr. Benson’s lie did, in fact, render the informed consent invalid, or he could claim that even if Dr. Benson was lying when she claimed her family member had the same type of cancer that James had, the lying took place before Dr. Benson began the process of informed consent. Hence, the informed consent was not rendered invalid. ²²

The first response, claiming that Dr. Benson lied and that the lie renders the informed consent invalid, is not convincing. There is widespread agreement that valid informed consent consists of several components: (1) The patient must have capacity; (2) The physician must provide the pertinent or relevant information, including the different treatment options, the risks and benefits of these options; and (3) the physician must provide information as to the probability associated with the different risks and benefits. Dr. Benson met these requirements; thus, we have reason to believe that her lying during the process of informed consent did not render informed consent invalid. Furthermore, although being lied to usually undermines one’s autonomy and negatively affects one’s ability to make a rational choice, this lie enhanced Dr. Benson’s patient’s ability to make a rational choice.

In order to respond to Simkulet’s second possible response, that Dr. Benson’s lie took place outside of the process of informed consent, I can change slightly the hypothetical. Consider Dr. Manuel, who has a long relationship with a patient, Drew, who was recently diagnosed with a particularly aggressive form of cancer. Dr. Manuel, while explaining what she believes are the two best options, is concerned that Drew does not recognize the seriousness of his situation. She decides that she must be blunt and exaggerate/lie in order to get Drew to focus. “If you choose the chemotherapy, you will probably lose your hair, experience severe, maybe uncontrollable, nausea, and at times you will suffer extreme pain. Not only that, but in all likelihood, the therapy will only allow you to live for 3 months. If you choose to manage the disease without aggressive treatment, you will live for 1 month, maybe a little longer, but you will avoid much of the pain and the nausea.” Dr. Manuel, confident that she has Drew’s attention, seeks to discover which option Drew thinks is best. After learning that Drew is leaning toward chemotherapy, Dr. Manuel says: “Drew, I probably exaggerated a little. In reality, chemotherapy will give you about 6–8 months of life, and although there will be periods of nausea and pain, there will be periods when you will feel normal. If we manage the progression of the cancer, you will likely live 3 or 4 months, and we will be able to manage the pain; you’ll be able to stay home for most of the time.” Drew opts for chemotherapy.

In this case, Dr. Manuel lied while talking about the different treatment options, that is, the lie took place during the process of informed consent. Nevertheless, I suggest that Dr. Manuel obtained valid informed consent. Like the previous case, Drew was adequately informed about the relevant aspects of his diagnosis and treatment options, and the lie increased Drew’s autonomy or his ability to make a rational choice. We can assume that Drew was competent and if Dr. Manuel quizzed Drew, he would be able to explain the two different options and their implications. Finally, Drew was not (unduly) coerced; the lie increased Drew’s ability to make a rational decision pertaining to his treatment. I suggest, therefore, that even though Dr. Manuel was lying during the process of informed consent, lying did not render informed consent invalid, and Simkulet’s assumption that truth-telling is necessary for informed consent (premise 1) is false. ²³

Simkulet might again respond that the lie is not part of the process of informed consent. He might claim that when Dr. Manuel was lying, she was attempting to get Drew to listen, and until Drew was listening, Dr. Manuel was not engaging in the process of informed consent. Put differently, Dr. Manuel began the process of informed consent, but when she lied to Drew, she was taking a break in order to get Drew to listen. When she believed Drew was listening, she returned to the process of informed consent.

Informed consent is a process without a clear beginning and end. The claim that anything one says during the process of informed consent, other than facts pertaining to the diagnosis, prognosis, or treatment options, is not a part of the process of informed consent is not obviously true. Furthermore, if Simkulet is committed to the claim that things not pertaining to the diagnosis, prognosis, or treatment options are not part of the process of informed consent, then if a physician nudges her patient and the nudge does not pertain to the diagnosis, prognosis, or treatment options—the physician wears surgical scrubs as opposed to a simple white coat—the nudge should not be considered part of the process of informed consent and should not, therefore, render informed consent invalid.

The previous hypothetical cases do not involve nudging. My point is that not telling the truth (or lying) during the process of informed consent does not necessarily render informed consent invalid; therefore, Simkulet’s premise 1 is false (or at the very least, not necessarily true).

Bullshit and Bullshitting

In this section, I will examine premise 4—nudging is bullshitting. In his well-known essay, “On Bullshit,” Harry Frankfurt attempts to determine what is bullshit and what is bullshitting. ²⁴ At the beginning of his essay, Frankfurt admits that attempting to arrive at a set of necessary and sufficient conditions for what constitutes bullshit would be arbitrary. Perhaps for this reason, Frankfurt limits the scope of his essay to providing “something helpful, even though it is not likely to be decisive.” ²⁵

For Frankfurt, bullshit is what the bullshitter produces. If the speaker has the right state of mind he is bullshitting and what he says is bullshit. In order to understand what state of mind is associated with bullshitting, Frankfurt attempts to distinguish lying from bullshitting. Lying, according to Frankfurt, involves intentionally misrepresenting both the state of affairs to which the lie refers and the beliefs of the speaker, that is, the liar falsely claims both that X is true and that he believes X is true. In addition, the liar intends to convince another that X is true. Frankfurt states that the bullshitter, unlike the liar, does not necessarily deceive another nor does he necessarily intend to convince another that what he says is true. What distinguishes bullshitting and lying is that for the bullshitter, there is a “lack of connection to a concern with truth—this indifference to how things really are—that I regard as the essence of bullshit.” He furthermore states that what the bullshitter “necessarily attempt[s] to deceive us about is his enterprise. His only indispensably distinctive characteristic is that in a certain way he misrepresents what he is up to.” ²⁶

Frankfurt provides many examples of bullshitting. One of the most oft-cited examples concerns a Fourth of July orator who “goes on bombastically about ‘our great and blessed country, whose Founding-Fathers under divine guidance created a new beginning for mankind.’” ²⁷ Frankfurt says the orator was “not trying to deceive anyone about American history. What he cares about is what people think of him.” ²⁸ He wanted people to believe he was a patriot who had deep thoughts and feelings about our country, that he placed importance on religion, and that he was a good person.

Regardless of whether the orator’s claims pertaining to the founding fathers were true or false, he was bullshitting because he was not concerned about the truth or falsity of what he said about the founding fathers. He made the claims about the founding fathers in an effort to increase or strengthen people’s positive beliefs about himself. Furthermore, even if the orator believed the claims were false, given that he was not attempting to convince the audience that the claims were true—he was trying to increase the likelihood those listening would think he was a good person—he was not lying.

Much has been written about Frankfurt’s essay, and many counterexamples have been proposed. In an interesting essay discussing these counterexamples, Don Fallis attempts to provide an understanding of what Frankfurt meant when he claimed that the essence of bullshit is a “lack of connection to a concern with the truth.” ²⁹ Essentially, Fallis states that conversations or inquiries can get us closer to the truth, further from the truth, or no closer to or further from the truth. The truth-teller intends to get us closer to the truth, and the liar intends to get us further from the truth. But the bullshitter, according the Fallis, does not care whether we get closer to or further from the truth, that is, the bullshitter’s agenda is something other than contributing to an inquiry. I think Fallis is correct, but I will go a little further and say that not only is the bullshitter not interested in contributing to an inquiry, but in most instances, the agenda of the bullshitter is to hide the fact that he does not know what he is talking about and/or to increase the likelihood that the listener has a good impression of him. Numerous examples in the literature back up this interpretation, ³⁰ as does a comment Frankfurt makes toward the end of his essay: “Bullshit is unavoidable whenever circumstances require someone to talk without knowing what he is talking about.” ³¹ Frankfurt explains that having the obligation to talk about something that exceeds one’s knowledge “is common in public life, where people are frequently impelled—whether by their own propensities or by the demands of others—to speak extensively about matters of with they are to some degree ignorant.” ³²

As stated in the introduction to this essay, Simkulet states that nudging is bullshitting, bullshitting is inconsistent with truth-telling, and that nudging, therefore, renders informed consent invalid. Simkulet, interpreting Frankfurt, asserts that bullshitting involves: (1) A conveys X to an audience and (2) A intends the audience to believe some belief (or set of beliefs) Y or perform some action (or set of actions) Z. Simkulet further asserts: “The intent of the bullshitter has nothing to do with the meaning of the words she says; she might say ‘X’, but she does not care whether her audience comes to believe X, but rather she intends her audience to come to believe Y or do Z, where believing Y or doing Z are not predicated on the audience coming to believe X.” ³³

Simkulet, while attempting to demonstrate that nudging is bullshitting, significantly limits what behaviors constitute nudges, thereby excluding from the discussion some of the most commonly discussed nudges. For example, he claims that framing, which includes choosing to wear scrubs as opposed to a simple white coat and discussing the success rate of a treatment as opposed to its complementary failure rate (both of which have the potential to increase the likelihood a patient will opt for surgery), is not necessarily nudging. Essentially, Simkulet, contra Cohen and others, ³⁴ claims that nudging necessarily exploits “non-rational aspects of patient psychology to persuade patients through means other than rational persuasion.” ³⁵ He concludes, therefore, that the physician who informs her patient of the success rate of a treatment as opposed to its complimentary failure rate, which has the effect of increasing the likelihood that a patient will consent to a procedure, is not nudging because this behavior does not necessarily increase the likelihood that the patient’s choice will be the byproduct of irrationality; that is, since knowing the success rate can help the patient make a rational choice, informing the patient of the success rate (as opposed to the complimentary failure rate) is not a nudge. ³⁶

After limiting what should be considered a nudge, Simkulet focuses primarily on one example in the literature, namely Cohen’s example of the physician who crafts disclosure in such a way that her patient will likely stop listening to the different treatment options when the physician mentions a reasonable option. Consider a physician who knows that there are at least five different treatment options, and this physician knows that many patients stop listening after they learn about the first option that is good enough. The physician believes X is the best option, but she is concerned her patient may choose Z. Accordingly, she begins with descriptions of R, S, and T, options that are not likely to be attractive to her patient. She then introduces X, the option she believes is best for this patient, knowing that this choice architecture has the potential to increase the likelihood that her patient will stop listening when she moves onto the final option, Z. Simkulet claims that the physician who utilizes such an approach is nudging and bullshitting. He furthermore claims that the informed consent obtained when a physician nudges in this way is not valid.

Before attempting to determine whether this behavior is in fact bullshit, I want to point out that if we take seriously Simkulet’s requirement that a nudge must increase the likelihood that the patient’s choice will be a byproduct of irrationality, crafting disclosure in the way described in this example might not be a nudge, that is, this choice architecture might increase the likelihood that the patient will make a rational choice. Cohen, when he discusses this example, claims “decisions in life are normally taken under conditions where it is not cost-beneficial to attempt to acquire all relevant information. It is pragmatically rational under such conditions to engage in ‘satisficing,’ that is, choosing the first option that crosses some threshold of acceptability.” ³⁷ Cohen states, therefore, that if this nudge is successful, it increases the likelihood that a patient will not make a choice that is a byproduct of irrationality. If Cohen is correct, Simkulet, given his assertion (that I and others believe is incorrect) that nudging necessarily undermines rational choice, is committed to the conclusion that this behavior, his primary example of nudging, is not nudging.

For the sake of discussion, let us assume that this choice architecture is a nudge. Consider Dr. Benson, a physician who crafts the choice architecture in this manner in order to increase the likelihood that her patient will choose what she, the physician, believes is the best treatment option and is the treatment that is consistent with her patient’s values. Clearly, Dr. Benson is not lying. She is not uttering a false claim nor is she attempting to get her patient to believe that a false claim is true. In fact, everything that she says is true. But is it bullshit? That depends on the mindset of the physician. Is she trying to deceive her patient about her enterprise? Is she trying to hide or misrepresent her enterprise or what she is up to? The answer to these questions depends on the situation. For example, a physician may believe that she has an obligation to provide extensive descriptions of the treatment options, while at the same time she knows that if she provides extensive descriptions, some people will stop listening when they learn about a treatment that meets certain basic requirements (satisficing). Consequently, believing that Treatment A is best for her patient, she begins the process of informed consent by discussing Treatment A. In this case, the primary motivation of the physician is to adequately inform her patient, and her secondary goal is to nudge her patient to opt for Treatment A. It is not at all clear that this behavior should be labeled bullshitting, that is, it is not clear that the physician is hiding her enterprise or misrepresenting what she is up to. This last is especially true if, as I and others have argued elsewhere, patients should know that physicians want their patients to opt for treatments that will have good outcomes or that physicians want to guide their patients to what they (physicians) believe are the best treatments. ³⁸

If we examine this example in light of Simkulet’s definition of bullshit, Dr. Benson might not be bullshitting. Recall that Simkulet asserts that bullshitting involves: (1) A conveys X to an audience and (2) A intends the audience to believe some belief (or set of beliefs) Y or perform some action (or set of actions) Z. Simkulet further states that the intent of the bullshitter “has nothing to do with the actual meaning of the words she says.” ³⁹ The bullshitter intends her audience to believe Y or to do Z, even though believing Y or doing Z are not predicated on the audience coming to believe X. Recall that Dr. Benson is describing different treatment options and ordering them in such a way that her patient will, at some point, stop listening. Dr. Benson is not conveying X to her patient in order to get her patient to believe Y or do Z, where believing Y or doing Z is unrelated to believing X. She wants her patient to believe what she says about the different treatment options (especially X) in order to increase the likelihood that her patient will opt for a specific option. She believes that her patient will choose this option if her patient believes what she says about the different treatments. Thus, I suggest that according to Simkulet’s own definition, it appears to be the case that Dr. Benson is not bullshitting.

Even if nudging in this case is bullshitting, I suggest that it is not wrong and should not render the process of informed consent invalid. Consider a physician who knows that certain patients have a short attention span and will stop listening when they hear about a treatment that is good enough. This physician believes Treatment A is best for her patient and furthermore believes that her patient would opt for A if he listened to all the treatment options and made a rational choice. I suggest that if this physician begins the process of informed consent by discussing treatments that she believes are not as good as A, believing that her patient might opt for one of these inferior treatments, her behavior would be criticizable. After all, she knows that her choice architecture, the order in which she discusses different treatment options, will likely affect her patient’s treatment choice. Surely, if a physician knows that the order in which she describes different treatment options may affect the patient’s choice, and she orders the options in a way that will increase the likelihood that her patient will not choose the best option (based on both the physician’s and patient’s values), she does not do the right thing. Thus, we see that in this sort of case, the physician ought to bullshit, assuming the behavior described in the example is, in fact, bullshitting. ⁴⁰

If we use Fallis’s understanding of what constitutes bullshitting, we must determine whether the physician is attempting to steer the inquiry closer to the truth or further from the truth or does not care whether the inquiry gets closer to or further from the truth.

Elsewhere, in an essay discussing nudging, I (and several colleagues) argue that when a patient enters a relationship with a physician, the patient understands, or should understand, that the physician has knowledge or expertise that the patient does not have. ⁴¹ Furthermore, the patient recognizes, or should recognize, that the physician’s goal is to help the patient choose what is best for the patient. Of course, there are those rare cases in which there are no established parameters of care, and in those instances, it might be appropriate for the physician to refrain from attempting to steer her patient toward a specific intervention. But in the majority of cases there are established parameters of care, and in those instances the patient should recognize the physician will steer her patient toward what she believes is best for the patient. In fact, we suggest that a patient would not welcome a physician saying: “‘I have no suggestions, what do you think we should do?’” ⁴²

If a physician is attempting to inform her patient as to the different treatment options and attempting to help her patient discover (and then choose) the intervention that the patient believes is best for himself, I suggest that her behavior (even if it involves nudging) can be described at aiming an inquiry toward the truth (in Fallis’s words) and that this behavior is not, therefore, bullshitting.

I do agree, however, that there is an element of phoniness associated with some nudges. Consider a patient who knows that his physician wants to help him choose the best treatment. If this physician chooses to wear scrubs in an effort to increase the likelihood that her patient will choose a specific intervention, most likely the patient does not know that the physician chose her attire in an effort to affect her patient’s judgment. Thus, there is a sense in which her behavior is (unduly) manipulative or perhaps bullshitting. As I have already explained, it is not always the case that bullshitting in this manner is wrong. (The physician whose patient wants brain surgery when he has appendicitis should attempt through discussion to steer the patient away from this surgery. If this fails, the physician has the option of refusing to perform the surgery or to try to nudge the patient to consent to an appendectomy. ⁴³)

The Meta-Nudge

Elsewhere, I have argued that nudging during the process of informed consent is permissible under certain circumstances. ⁴⁴ This is most clearly the case in those instances in which nudging is unavoidable. Consider the physician who wakes up in the morning and must choose her attire. She must decide whether to wear scrubs or something else (or wear nothing at all). If she wears scrubs, she knows that she will nudge her patients toward surgery. If she wears something else (or nothing at all), she nudges her patients away from surgery. Similarly, in many instances the physician must disclose the success rates of different procedures or their complimentary failure rates (or both). Depending on which she chooses to focus, she will nudge her patient toward or away from the procedures. In these situations, the physician cannot avoid nudging. Assuming the truth of “ought implies can,” nudging is not impermissible when it cannot be avoided.

I suggest that there is a danger associated with these unavoidable nudges (and other nudges discussed in the philosophical literature): They have the potential to increase the likelihood that a patient will choose a treatment or intervention that is not consistent with the patient’s own desires, values, or conception of the good. (Remember, when Sunstein and Thaler introduce the nudge, they limit its permissibility to those instances in which one is attempting to increase the likelihood that the one being nudged will make a choice that is consistent with one’s own desires, values, or conception of the good, that is, they advocate soft paternalism.) Consider a surgeon who makes it a practice to wear scrubs during the process of informed consent (believing that surgery is best for most of her patients and that using this nudge may increase the likelihood that patients will consent to surgery). Although nudging most patients toward opting for surgery may be best for a majority of her patients, the surgeon runs the risk of nudging one patient (or several of her patients) toward consenting to surgery even though this one patient might not consent to surgery if he reflected on the surgery and was not nudged. The same can be said for the surgeon who informs her patient about the success rate of a given intervention as opposed to its complimentary failure rate. There may be patients (perhaps a small minority) who would be swayed by the nudge even though surgery is not what they would choose upon reflection (and if they were not nudged). This is especially problematic because surgeons (and other healthcare providers) frequently do not have the opportunity to get to know their patients or learn about their patient’s values and conceptions of the good before obtaining informed consent. Thus, physicians frequently do not know whether nudging patients toward one treatment or another is consistent with their patients’ values.

Perhaps nudging is appropriate (or at the very least less problematic) in those instances in which the physician is aware of her patients’ values and conceptions of the good. ⁴⁵ In those instances in which one cannot determine whether one’s patients’ values support one intervention or another, however, nudging may be inappropriate (or at the very least problematic). One problem with this last is, as explained above, that nudging may be unavoidable. Recall the physician who is choosing her attire. Her choice has the potential to push her patients toward or away from surgery. Cohen argues that nudging is permissible in these sorts of cases if the physician believes that surgery is the best option. In fact, he claims nudging is obligatory because not nudging would be a violation of the principle of nonmaleficence (and the duty of beneficence). ⁴⁶

I do not deny that in some situations nudging may be unavoidable, that a physician must choose her attire (or choose to wear no attire at all), and that there are instances in which a physician cannot avoid discussing the probability that an intervention will either succeed or fail. I furthermore agree that one’s attire may nudge a patient in one direction or another, and the same is true with respect to discussing the probability that intervention will succeed or fail. That said, there may be a way to undermine the power of some unavoidable nudges and thereby render them impotent; that is, physicians may be able to utilize what I call the “meta-nudge” in an effort to weaken or even eliminate the power of unavoidable nudges.

Empirical studies demonstrate that the effects of priming can be mitigated or reversed. Hans-Peter Erb and colleagues began a study by demonstrating that people’s risk attitudes can be affected by priming. He then discovered that the effects of priming “can be overridden and (over) corrected when attention is drawn to the priming influence.” ⁴⁷

Perhaps a similar mechanism works with nudging in those cases in which nudging unavoidable. Although I am not aware of any empirical studies focused on this issue, it is not far-fetched to suggest that the power of a nudge can be (significantly) diminished if the person being nudged is told how the nudge might affect this person’s behavior. In the case of surgical scrubs, the physician could tell the patient that a physician’s attire, whether she wears surgical scrubs or something else, has the potential to affect this patient’s decision-making and that the fact that she is wearing scrubs has the tendency to increase the likelihood that her patient will opt for surgery. Similarly, the physician could tell her patient that thinking about success rates, as opposed to rates of failure, increases the likelihood one might consent to an intervention. The physician might also discuss both the success and failure rates. ^{48 , 49} I call this a meta-nudge, because telling a patient about the power of an initial nudge is a kind of nudge. This practice alters the choice architecture in a way that does not minimize options or change economic incentives. In addition, this practice has the potential to affect one’s decisions in a predictable way. (That said, the meta-nudge may not be a nudge if, as Simkulet claims, nudges circumvent rationality. In fact, the meta-nudge may reverse the circumvention of rationality associated with some nudges. It may enhance a patient’s autonomy.)

Assuming the meta-nudge works in the way described above, we might ask whether one who uses the meta-nudge has the obligation to tell her patient that she is using the meta-nudge. I do not think this is the case. As explained above, the meta-nudge is, I think, different from normal nudges in that the goal of one who uses the meta-nudge is different from that of normal nudges. Nudging, by definition, aims at pushing the patient into a different decisional space, that is, nudging influences the decisions of patients and moves them toward decisions that their physicians believe are best (or physicians believe that their patients believe are best). The goal of the meta-nudge, whether it works through rational or nonrational means, is to increase the likelihood that patients make decisions that are consistent with their prenudged decisional state. For example, if a risk averse patient is nudged toward a surgical intervention to which she would not have consented if she were not nudged, the meta-nudge would aim to return her to this risk averse decisional state.

As mentioned previously, Chwang argues that nudging is akin to obtaining informed consent when a patient is ignorant, and that nudging is (almost) always, therefore, impermissible during the informed consent process. Cohan, on the other hand, believes that nudging, when a physician believes her patient will choose the less good option, is usually permissible (and that not nudging may be impermissible). I suggest that the truth is somewhere in the middle. Perhaps the best way to balance respect for autonomy and the duty of beneficence (and nonmaleficence) is to attempt to discover how much value patients place on autonomy. Or, in the case of nudging, to attempt to determine whether one’s patients want to be nudged. In some instances, when the patient and physician have an established relationship, the physician may know, or at least have a good idea, whether her patient would welcome being nudged (or not), and she can act accordingly. In other situations, perhaps the physician should ask the patient, early in their relationship, how much she values autonomy or explain how nudging works and ask her patient whether she opposes or supports being nudged. ⁵⁰ Possessing information pertaining to how much value a patient attaches to autonomy can help the physician decide whether to utilize nudges or the meta-nudge. If a patient values his physician’s judgment and wants his physician to lead him to the best treatment option, nudging may be appropriate. However, in those instances in which a patient would not welcome being nudged, the physician can attempt to avoid nudging. If, however, she cannot avoid nudging the patient who values autonomy, she can use the meta-nudge to weaken the power of unavoidable or otherwise problematic nudges.

Conclusion

In this essay, contra Simkulet, I claim that nudging is, most likely, not a practice that falls under Frankfurt’s conception of bullshitting. I furthermore explain that even if Simkulet is correct, even if nudging during the process of informed consent is bullshitting, it does not necessarily follow that nudging during the process of informed consent renders the informed consent invalid.

I conclude by suggesting that whether nudging is (or should be) permissible is, to a large extent, a function of the patient’s and his physician’s values. Physicians must negotiate the tension between the duty of beneficence and the duty to respect autonomy. The issue of whether to nudge during the process of informed consent is one example of this tension. Nudging during the informed consent process may be permissible when a patient does not place high value on autonomy (or, even more obviously, in those cases in which a patient values a paternalistic physician–patient relationship). However, if a patient places a lot of value on autonomy, the permissibility of nudging becomes less obvious. I recognize that nudging is sometimes unavoidable, a physician must choose her attire (or choose to wear no attire) and she must discuss the success or failure rate of different procedures. I suggest that there may be a way to undermine the power of these nudges. Specifically, physicians can use the meta-nudge to undermine or weaken the power of unavoidable or problematic nudges. Whether the meta-nudge will work in the way I suggest should be studied further. Hence, more research and empirical studies must be performed on nudging and the meta-nudge.