The doxastic profile of the compulsive re-checker

ABSTRACT

Incessant checking is undeniably problematic from a practical point of view. But what is epistemically wrong with checking again (and again)? The starting assumption for this paper is that establishing what goes wrong when individuals check their stove ten times in a row requires understanding the nature of the doxastic attitude that compulsive re-checkers are in, as they go back to perform another check. Does the re-checker know that the stove is off, and is thus looking for more of what she already has (Whitcomb, D. 2010. “Curiosity was Framed.” Philosophy and Phenomenological Research 81 (3): 664–687.)? Or is she an inquirer who repeatedly loses her knowledge and finds herself inquiring again and again into the same question (Friedman, J. 2019. “Checking Again.” Philosophical Issues 29 (1): 84–96.)? I present what I see as the three main hypotheses currently available, and propose a refinement to Taylor's ‘what-if questioning’ account (2020).

KEYWORDS:

• Checking
• belief
• compulsion
• suspended judgment
• inquiry
• emotion

1. Introduction

Checking is one of the most common compulsive actions performed by patients with Obsessive-compulsive disorder (OCD) (APA 2013; Abramowitz, McKay, and Taylor 2008). Incessant checking is undeniably problematic from a practical point of view; it is an impairment to the achievement of one’s everyday goals. But what is epistemically wrong with checking again (and again)? While we have the intuition that this is not good inquiring practice, it is hard to pinpoint what exactly is epistemically deficient about it. After all, certain philosophers have argued that if it is not absolutely certain that p, we have an epistemic duty to seek more evidence about p (Hall and Johnson 1998). If checking counts as evidence gathering, what could be wrong with it, from an epistemological standpoint?

The starting assumption for this paper is that establishing what goes wrong when individuals check their stove ten times in a row requires understanding the nature of the doxastic attitude that compulsive re-checkers are in, as they go back to perform another check. Does the re-checker know that the stove is off, and is thus looking for more of what she already has? Or is she an inquirer who repeatedly loses her knowledge and finds herself inquiring again and again into the same question?

There are three main hypotheses on the market regarding the nature of the re-checker’s doxastic stance as she goes back for another check. The first consists in considering the re-checker as an insatiable knowledge-seeker (Whitcomb 2010). According to this account, the compulsive re-checker knows that p (the stove is off), but keeps greedily looking for what she already has (knowledge that p), apparently insatiable in her desire for more of the same knowledge. The second hypothesis is that the compulsive re-checker is a repeatedly suspended inquirer, who keeps (unjustifiably) suspending on a matter that she has just settled through successful inquiry (Friedman 2019).¹ According to this view, it is not that the re-checker knows that p and keeps inquiring into whether p nonetheless, it is rather that after she has performed a check, she goes back to suspending on the question, so that she needs to perform another check to reach a firm judgement again. A third view has recently been put forward, which combines elements of the first two views. According to this third view, the compulsive re-checker in fact knows that p, but is also simultaneously in a ‘what if?’ questioning attitude, thinking to herself: ‘I know that the stove is off, but what if it is not?’ (Taylor 2020).

It is possible that different kinds of re-checkers check for different reasons, and that all three views combined are able to explain most cases of re-checking. Nonetheless, I will argue that the suspension view is a superior alternative to the insatiable knowledge-seeker view, and that the knowledge and ‘what if’ questioning account faces a serious issue. In particular, I will take issue with the claim that the re-checker retains knowledge throughout the ‘what if’ questioning and subsequent checking. I will argue that, if wondering ‘what if’ can be taken to act as a motivating reason for subjects to re-check, then this questioning is likely to raise non idle doubts, and thus be strong enough to defeat one’s knowledge. I argue that this is particularly likely when re-checking occurs in the context of OCD, where the pathological anxiety which systematically accompanies ‘what if’ questionings precisely functions to make salient epistemic fallibility and high stakes, thereby raising serious doubts.

Understanding the epistemic state that Obsessive-compulsive re-checkers are in as they form the intention to perform another check is important for furthering our knowledge of the disorder itself, but also for designing efficient interventions which correctly identify and target the mental states that are problematic in motivating these patients to re-check.

2. Compulsive re-checkers as insatiable knowledge-seekers

OCD is characterized by the recurrence of unwanted, unpleasant thoughts, and is diagnosed when obsessions and compulsions cause marked distress and interfere with the person’s functioning. OCD most often implies two elements: obsessions, which are ‘recurrent and persistent thoughts, urges or images that are experienced, at some time during the disturbance, as intrusive and unwanted’ (American Psychiatric Association 2013, 237); and compulsions, which are ‘behavioral or mental rituals according to specified ‘rules’ or in response to obsessions’ which are often understood as an attempt to neutralize or suppress the obsessions themselves (Abramowitz, McKay, and Taylor 2008, 5). For instance, compulsive checking is generally performed in response to an obsessive thought about possible harm which could occur, and compulsive washing is often associated with obsessive thoughts about contamination.

Obsessions and compulsions are easily understandable, because we are all familiar with their more innocuous variants. As Neal et al. rightly remark:

Who among us has never stopped to verify that their travel documents are in the bag on the way to the airport? Or that the alarm has been set properly the night before an important early morning appointment? (Neal, Alcolado, and Radomsky 2017, 363)

However, when confronted with OCD, we realize how much most of us take for granted in our everyday life. We seem to exhibit a basic trust, that we are out of danger, that things will, most of the time, turn out the way we expect them to, and that we have the ability to perform routine actions (such as washing our hands) in a correct, acceptable manner (De Haan, Rietveld, and Denys 2013). OCD sufferers experience recurring doubts about some of these domains, and some patients (the ‘checking subtype of OCD) respond to these doubts by performing repetitive checking behaviors.

Now, the repetitive nature of checking behaviors does not only raise questions concerning the psychological underpinnings of such doubts, but also on the norms that should govern evidence gathering. One question which has attracted the attention of several philosophers, and epistemologists in particular, is: What exactly is epistemically wrong with re-checking? My suggestion and the starting point for this paper is the idea that, in order to understand exactly what goes epistemically wrong with re-checking, we first need to take a step back from the behaviors themselves, and consider the mental state that the re-checker is in as she goes through the moves. What is she looking for, as she goes back for another check? Thus, the question that will be the focus of my present inquiry is: What is her epistemic position as she does so, and how does the checking affect it?

A first intuitive answer to this question is: as she goes to perform another check, the compulsive re-checker is looking for knowledge, or an equivalent epistemic good such as truth or accurate information. Although she has previously settled the question of whether the stove is off, the compulsive re-checker seems to keep trying to answer this question. In other words, it seems that the compulsive re-checker already knows that p (the stove is off), but that she is looking for more knowledge through performing another check. According to this view, compulsive re-checkers may becharacterized as insatiable knowledge-seekers. ²

What compulsive re-checkers suffer from is thus an incapacity to respond to evidence in such a way as to recognize or remain aware that they in fact possess knowledge and are justified to form and maintain a firm judgement. Along these lines, Whitcomb (2010) suggests that the individual who checks their alarm clock five times in a row is like the glutton who keeps eating after he has been sufficiently nourished. As he suggests, in much the same way as nourishment is the satisfier to hunger, checking is like keeping on eating after one’s hunger has been satisfied:

In both cases there is a desire (hunger, curiosity); an attempt to sate that desire (eating, inquiry); a sating of that desire (by nourishment, by knowledge); and a continued attempt to sate the desire after it has already been sated (more eating, more inquiry). (Whitcomb 2010, 674)

As Whitcomb suggests, what is epistemically wrong with checking one’s alarm clock five times in a row is that one keeps inquiring although one’s desire for knowledge has already been satisfied by the second check. For Whitcomb, checking is epistemically problematic because it repeatedly seeks to produce what is already there, namely knowledge.

According to this first account, what makes checking epistemically deficient is the fact that checking is aimed at acquiring knowledge, while subjects who re-check already possess knowledge. In what follows, I argue that this first account of the problematic nature of checking is mistaken for several reasons. First, while the analogy between hunger and inquiry seems appealing, it does not reflect the complexity of our relation to inquiry. Answering a question is not always a process as linear as eating a meal; some inquiries are long-standing endeavors, involving many steps and shifts in our epistemic attitudes. Moreover, while an eater might still derive pleasure from eating more once they are no longer hungry, the inquirer derives no specific pleasure from obtaining a piece of knowledge they already possessed.

Second, while the re-checker might be in a state of knowledge as she goes back to check, her re-checking is not necessarily aimed at ‘more knowledge’. Whitcomb’s account relies on the idea that what the re-checker is looking for is simply what she already has: knowledge. However, it is also possible that what the compulsive re-checker is looking for, through performing her check, is rather something else, perhaps ‘more than knowledge’. Perhaps she is looking for certainty, or higher-order knowledge (knowledge that we know), for instance.

Third, Whitcomb’s account relies on an analogy between hunger and curiosity (the desire for knowledge), or between knowledge and nourishment. However, we must draw a distinction between hunger and the mere desire to eat. While hunger is a ‘rational desire’ that is sensitive to reasons, the mere desire to eat is more akin to an impulsion or a-rational desire (see Alvarez 2010). In the same way, it seems reasonable to distinguish between the desire for knowledge and the more impulsive desire towards absolute certainty. If we accept such a distinction, then Whitcomb’s explanation as to what makes checking epistemically problematic collapses. This is because one may satisfy one’s hunger without satisfying one’s desire to eat, and keep trying to satisfy one’s desire to eat even after one’s hunger has been satisfied. Since the desire to eat is a-rational, and is not actually aimed at nourishment, it cannot be evaluated as irrational, or as a desire for more of what one already possesses. Following the analogy proposed by Whitcomb, perhaps the desire manifest in compulsive re-checking is not a rational desire aimed at acquiring knowledge, but rather an arational desire for absolute certainty. If so, one could keep on trying to satisfy this desire for absolute certainty through re-checking even once the rational desire for knowledge is satisfied

It does seem that, in the case of the compulsive re-checking that is part of OCD, patients are not experiencing a rational desire aimed at knowledge, but rather an impulsion or a-rational desire aimed at absolute certainty or total absence of uncertainty. It has been suggested that OCD patients have an excessive need for certainty (Makhlouf-Norris and Norris 1973). Intolerance of uncertainty is a central feature of the disorder, and has been linked to the inability to ‘experience a sense of conviction’ (Shapiro 1965), put closure on experience (Reed 1985), or generate the normal ‘feeling of knowing’ (Szechtman and Woody 2004). Empirical evidence has shown that subjects with OCD require a higher level of certainty to be able to form a judgment and make a decision (Banca et al. 2015). Compulsive behaviors can thus be understood as an attempt to attain a higher degree of certainty, or to reduce uncertainty to a minimum (Rotge et al. 2008; Stern and Taylor 2014).

In sum, it is not clear that what the compulsive re-checker is looking for is indeed just knowledge. If compulsive re-checkers are not looking for knowledge, then what is epistemically problematic about checking is not that subjects are looking for what they already have. An alternative explanation of the problematic nature of re-checking relies is the idea that, in fact, as they go back for another check, re-checkers have lost the relevant knowledge that they had previously acquired. According to this view, re-checkers need to check again because, since the last check, they have shifted back out of belief and into suspension of judgement.

3. Compulsive re-checkers as insatiable knowledge-seekers

As Jane Friedman (2019) has proposed, re-checkers are perhaps not best described as insatiable knowledge-seekers, and the problem with re-checking may not lie in the manifestation of a desire for some epistemic good one already possesses.

Friedman offers a different view on why repeated re-checking cannot be considered good inquiring practice. Friedman is interested in robust cases of re-checking, that is, cases in which one is genuinely investigating into the matter at issue (rather than performing habitual movements, for instance). Now, while in severe or long-standing cases of OCD checking compulsions might have become habitual or ritualized, this is typically not the case at early stages of the disorder, or for milder forms of the disorder (Szalai 2016, 2019). I thus believe that Friedman’s proposal can be applied to early stages or milder forms of the compulsive re-checking that is present in OCD.

The originality of Friedman’s treatment of the phenomenon lies in her proposal to view checking as an activity that takes place within the larger framework of an inquiry. This inquiry process can be broken down into a sequence of actions and doxastic shifts. Contrary to other types of inquiry, Friedman remarks, re-checkers do not start their inquiry from a position of ignorance and neutrality but from a settled epistemic position. As the re-checker goes to triple-check whether her keys are in her pocket, she is looking to answer a question she has already settled in the past. Let us take a closer look at this inquiry process:

A re-checker starts settled with respect to a question Q (e.g. is the stove off?) – they already know Q or at least believe an answer to Q (it’s off); then they open Q again (is it off?) and collect more evidence on the matter (look at the dial); they settle the question again as a result of the check (it’s off); they re-open the question again (is it off?) and do another test (look at the dial again or maybe check the burner this time); they re-settle again (it’s off); and so on. (Friedman 2019, 3)

In other words, Friedman provides a diachronic picture of re-checking, in which re-checkers open a question and close it several times in a row. When re-checkers move towards the next check, and put the question back up for inquiry, they suspend judgement on the question of whether the stove is off once over. Between every check, Friedman suggests, re-checkers move from a settled attitude of believing some answer to the question (the stove is off), to suspending judgement again on whether the answer is true.

Even if they antecedently knew or correctly believed that it is, as they go back for a new check, re-checkers are suspended all over again on whether the stove is off. Friedman hence concludes:

I don’t think we should be thinking of them as simply insatiable believers or knowers, subjects who have settled the relevant questions but still want more. It’s exactly because they don’t (or at least no longer) take matters to be settled or think they have enough information that they go back for more […] An accurate account of the doxastic or epistemic state of the re-checker then should have them shifting their view on whether their answer is the right one. (Friedman 2019, 22)

Contrary to what Whitcomb (2010) suggested, the re-checker is perhaps not a knower who is greedy for more knowledge. Rather, the hypothesis put forward by Friedman is that the re-checker’s doxastic position is an ever vacillating one; it is one of constant shifting out of belief and back into suspension of judgement.

Indeed, one might even argue that, once one has given in to the drive to re-check, the checking itself might threaten justified belief or knowledge, as it might raise new worries and questions. Moreover, the act of checking is itself a method of evidence gathering that is subject to epistemic evaluation. Each check might thus add its share of uncertainty, if it gets evaluated as unreliable or flawed in some manner or other.

If the re-checker is genuinely suspended again on whether p, then what makes it epistemically wrong for her to try and settle the question again through engaging in another check? The answer lies, in Friedman’s view, in the fact that any genuine inquiry into a question implies that one is also necessarily suspended on that question, and that this doxastic attitude is not always justified. Suspension of judgement, Friedman argues, is the most general question-directed attitude, much in the same way as knowledge is the most general factive attitude according to Williamson.

According to Timothy Williamson (2000), knowledge plays a central role for a particular category of mental states: factive mental states. A type of mental state is factive if it is only directed at true propositions. States such as seeing, remembering, and regretting can be said to be factive, insofar as one can only genuinely remember or regret something which is true, in the sense that it is actually occurring or has actually occurred. Williamson famously argues that knowledge is the most general factive mental state: one is in this mental state whenever one is in any factive mental state. Being in a factive mental state about p implies that one knows that p. Friedman makes a similar claim about the relationship between suspension of judgement and the interrogative attitudes (curiosity, wondering, etc.): whenever one is in an interrogative attitude towards a question, one is suspended on that question.

Since genuine inquiry entails suspension of judgment, there are epistemic constraints pertaining to having a question open for inquiry, because there are circumstances in which suspending judgement is not justified. As Friedman suggests, certain epistemic circumstances are ‘suspension-proof’: suspending is not justified in such circumstances. What is epistemically wrong with re-checking, according to Friedman, is that after a few checks the re-checker quickly finds herself in suspension-proof epistemic circumstances, and she will suspend in spite of the fact that suspending is not justified in those circumstances. After having settled on the basis of good evidence (through performing a first check), the re-checker finds herself in an epistemic position that permits belief, and does not justify suspending judgment. Given this, her decision to re-open the question and shift back into an inquiring mode is epistemically problematic.

Hence, Contrary to what Hall and Johnson (1998) advocated, the fact that one could still run another test is not sufficient to justify further inquiry. Most epistemic positions we hold are positions which can be improved; they can be strengthened by seeking and gathering additional evidence. If this is so, and if seeking additional evidence amounts to inquiring again, then we should be justified in re-opening an inquiry (and hence suspending judgement) on most questions we hold epistemic positions towards. Yet, Friedman argues, further evidence gathering is not always justified, and can in fact be epistemically problematic. And this can be explained by the fact that inquiring implies suspending judgement, and that certain circumstances do not permit this doxastic attitude.

Just to be clear, this does not mean that it is never epistemically appropriate to engage in the revision of one’s beliefs. Re-opening the question on something we already believe or even know is good epistemic practice. Proper epistemic maintenance of a belief may involve double-checking. Certain forms of re-checking, however, are problematic because the conditions in which the subject suspends are suspension-proof.

Friedman’s account faces its own challenges. It of course raises the question of which circumstances qualify as ‘suspension-proof’ circumstances. If the problematic nature of re-checking can be explained in terms of suspension-proof circumstances, one important question this account should address is: what exactly makes for suspension-proof circumstances? According to which criteria should we evaluate that suspension on a question one had previously settled is not part of proper epistemic maintenance? Friedman tentatively suggests that what might be necessary for being in suspension-proof circumstances is that one both knows p and is aware that one knows p. Do compulsive re-checkers really know p and are aware that they know p? It is not so clear. As we will see next, we have reasons to believe that compulsive re-checkers both know p and at the same time question whether p.

It has indeed been argued that what is puzzling about the epistemic state of the compulsive re-checker is precisely that, although they generally judge their doubts to be unfounded, they nonetheless experience them. Thus, while compulsive re-checkers might know that the stove is off, it is not at all clear that they are aware that they know that the stove is off. Moreover, another factor which could be taken into account when assessing whether or not an agent is in ‘suspension-proof’ circumstances is the importance of the practical stakes involved. It is reasonable to consider that an agent is more justified in suspending again on a question when (it seems to the agent that) an epistemic mistake would imply serious negative practical consequences. For those compulsive re-checkers which do expect negative practical consequences to result from their negligence, it is therefore questionable whether their circumstances are in fact ‘suspension-proof’. Taken together, these remarks undermine the argument according to which re-checking is epistemically problematic due to the ‘suspension-proof’ circumstances in which it is performed. In the next section, I introduce a recent attempt to characterize the epistemic state of the compulsive re-checker while accounting for its complexity.

4. Compulsive re-checkers as knowers with ‘what if' questions in mind

While it does seem that obsessive-compulsive re-checkers shift back to an inquiring attitude as they launch a re-check, they also seem to have insight into the unjustified nature of their behavior. In fact, it is established that a large majority of obsessive-compulsive re-checkers have good insight: they are well aware that their doubts are unreasonable (Jacob, Larson, and Storch 2014; APA 2013). This suggests that the doxastic position of compulsive re-checkers towards the proposition of interest is complex: re-checkers both know that the stove is off, and at the same time doubt that it is.

However, as Taylor claims, it is puzzling to think of the OCD re-checker as both knowing that p and at the same time believing that not-p, or as knowing that p and at the same time suspending on whether p. These characterizations are not helpful in understanding the epistemic state of the compulsive re-checker, because we have a hard time conceiving how they could be compatible with each other, that is, simultaneously occur in an individual. In order to resolve this puzzle, Taylor (2020) has recently proposed an alternative way of accounting for the complex epistemic state of the compulsive re-checker. While re-checkers do know that the stove is off, he suggests, they also repeatedly find themselves wondering ‘what if it is not?’. In other words, the OCD re-checker is in a question-directed attitude of the form ‘what if not p?’ all the while knowing that p. This interrogative attitude is what motivates her to perform another check, in spite of her knowledge that the stove is off and that doubting is unwarranted.

As Taylor remarks, this resolves the epistemic puzzle, since there is no doxastic incompatibility between knowing that the stove is off and asking ‘what if it is on?’. However, as I will argue, while Taylor’s model rests on this compatibility between ‘what if’ attitudes and knowledge, he does not provide solid support for it. He instead merely appeals to our intuition that such statements as ‘He knows that his hands are clean, but he keeps thinking what if they're contaminated?’ seem acceptable. In fact, as I will show in the next section, we have reasons to believe that this compatibility claim rests on an equivocation.

Taking stock, I have so far exposed three different accounts of the doxastic attitude that re-checkers are in as they go back to check. While I found several shortcomings to the view of compulsive re-checkers as insatiable knowledge seekers, I found the repeatedly suspended inquirers view to be a compelling alternative. Nonetheless, it fails to account for the apparent complexity of re-checkers’ doxastic state. The final account I presented suggests that we view compulsive re-checkers as wondering what-if knowers.

Now, the main difference between the repeatedly suspended inquirer view and the wondering what if knower view which I have just presented is that, while in the latter re-checkers are in an interrogative ‘what if’ attitude while they know p, in the former re-checkers have shifted out of belief and into suspension of judgement as they go back for another check. In other words, the wondering what if knowers view holds that re-checkers know that the stove is off as they go back to check.

In the next and final section of the paper, I argue that while the wondering what if knowers view is intuitively attractive, its plausibility is diminished when we take a closer look at the psychological compatibility between entertaining a ‘what if not-p?’ question in the way compulsive re-checkers do, and retaining the knowledge that p. I conclude that we should in fact not view compulsive re-checkers as knowers.

5. Losing knowledge with ‘what if' questioning

The idea that what motivates the re-checker to go back is a questioning attitude, rather than a belief, is intuitively plausible. Moreover, and this is Taylor’s main point, having a questioning attitude is epistemically compatible with both the world-directed and the self-reflective insight that compulsive re-checkers have. Persons with OCD both know that the content of the obsessive thought (i.e. my house will burn down) is false (this is world-directed insight), and they also know that it is irrational (or epistemically unjustified) for them to entertain this obsessive thought (this is self-reflective insight). Here I am focusing on the re-checkers’ doxastic attitude towards the relevant world-directed propositions, and not on their higher-order evaluations of their attitude. In a nutshell, while I agree with Taylor’s proposal that the (world-directed) ‘insight’ possessed by OCD sufferers amounts to knowledge, I do not think it plausible that this knowledge is preserved through the checking process. As we will see, upon further examination it is clear that if this questioning is strong enough to motivate another check, it is doubtful that the compulsive re-checker’s knowledge remains intact through this questioning. As I will show, we instead have good reasons to think that compulsive re-checkers lose their previously held knowledge through the questioning and checking process.

Let us first take a closer look at the reasons why Taylor’s view is intuitively appealing. As Taylor argues, viewing obsessive thoughts as mere ‘what if?’ questionings resolves the puzzle of the doxastic state of compulsive re-checkers because ‘what if?’ questionings are, so to speak, innocuous: they are no threat to knowledge. The issue is that, in fact, I believe Taylor’s compatibility claim – the claim that wondering ‘what if not p?’ is doxastically compatible with knowing that p – rests on an equivocation.

In Taylor’s view, ‘what if?’ questionings are to be assimilated with mere wonderings or with the psychological act of entertaining a possibility that not-p while one knows that p. For instance, while I know that sanitary measures such as mandatory mask-wearing in public spaces have been lifted in my country, I might nonetheless wonder ‘what if those measures were still in place?’, entertaining this possibility and engaging in counterfactual reasoning about the impact that such measures would have on everyday life. In this case, we can easily see how wondering ‘what if not p?’ is compatible with knowing that p. It is worth noting, nonetheless, that according to certain infallibilist accounts of knowledge, knowing that p is incompatible with the possibility of not p. Within such a framework, insofar as genuinely wondering ‘what if not-p?’ can be taken as a mark of the possibility that not p, it is not compatible with knowing that p.

In any case, and whatever view of knowledge one adopts, if this is the kind of ‘what if?’ questioning that Taylor has in mind, then I believe it is irrelevant to the phenomenon of compulsive re-checking, and is unable to illuminate it. Indeed, ‘what if?’ questioning understood in this weak sense is not strong enough to motivate one to check again. The ‘what if?’ questioning that is able to motivate subjects to re-check is distinct from the ‘what if?’ questioning we engage in for the sheer sake of exploring counterfactual scenarios which bear no relation to reality. If obsessive thoughts were mere wonderings of scenarios one knows not to be actual, they should be the kinds of thoughts which can easily be dismissed. Instead, the ‘what if?’ questionings that OCD sufferers engage in tend to overshadow all other ongoing thought processes.

The questioning at the center of OCD doubt does not seem to stem from voluntarily engaging in supposition or counterfactual reasoning. As noted earlier, OCD patients are unable to dismiss those thoughts and the behaviors they prompt. For this reason, it has been hypothesized that OCD patient’s recurring doubt is the product of dysfunctional sub-personal signals of error. According to an influential view, such ability relies on a mechanism of comparison between our prediction of the outcomes of our actions, and the actual effects we observe. According to the comparator model (Wolpert, Ghahramani, and Jordan 1995) the consequences of our actions are monitored and compared to our motor intentions, in order to establish whether the outcomes match our predictions or deviate from them. This model has been used to study the sense of agency, or the sense that one controls an external event through one’s action (Blakemore and Decety 2001). Accordingly, it has been suggested that a dysfunction in this comparator mechanism leads to hyperactive error signaling in OCD patients. As proposed by Belayachi and Van der Linden (2010), compulsive behaviors in OCD may be primarily motivated by inconsistent feelings that an action has not been satisfactorily completed, that the intention has not been achieved. Hyperactive feelings of incompleteness lead to inferences of failure of action, which trigger overactive performance monitoring for adjustment, and repeated attempts at meeting the intended goal (Belayachi and Van der Linden 2010).

Interestingly, a similar hypothesis has been put forward, not only regarding the sense of agency and action completion, but regarding the general tendency of OCD patients to doubt, raise questions, and pursue endless inquiries. Cochrane and Heaton (2017) have posited the existence of ‘a dedicated cognitive mechanism for signalling uncertainty that is overactive in the OCD sufferer’ (2017, 196). The hypothesis is that a sub-personal signal with the function of monitoring uncertainty goes hyperactive in OCD and creates an intrusive sense of uncertainty, which in turn leads to an active search on the part of the subject for ways in which harms could manifest in her environment. Their proposal is based on the observation that a sense of uncertainty is experienced among all OCD subtypes – it is a unifying element of the disorder – and that it is experienced as intrusive. The compulsion to check, repeat, and replay scenes mentally is then understood as driven by a need to appease and disconfirm an underlying, constantly resurging sense of uncertainty. This sense of uncertainty can be associated by the individual with whichever mental content or action is currently salient for her:

the individual can experience uncertainty about their intentional behaviour across a variety of different domains. This includes mental actions such as deliberately recalling a memory, or deliberately paying attention to something. It can also include the compliance with norms (such as being a moral person, or keeping clean). (Cochrane and Heaton 2017, 196)

In support of this hypothesis, Cochrane and Heaton point to empirical evidence of an increased neural activity linked to error detection in OCD (Riesel 2019). If this idea is valid, then the resurgent sense of uncertainty and consequent need in OCD sufferers to (re-)settle the question at hand has little to do with the voluntary entertaining of counterfactual scenarios which one knows to be unfounded.

What is more, several authors have suggested that checking itself may be a method of inquiry that is particularly prone to threatening one’s state of knowledge. This would suggest that, not only are the questionings experienced by OCD sufferers different from mere suppositions, but the subsequent checking they cause further reduces individuals’ confidence in their antecedent belief. Can certain forms of inquiry threaten our knowledge? This question was famously raised by David Lewis in his article ‘Elusive Knowledge’ (1996), where he argued that: ‘when we look hard at our knowledge, it goes away. But only when we look at it harder than the sane ever do in daily life; only when we let our paranoid fantasies rip’ (1996, 550). If Lewis is on the right track, and scrutinizing one’s belief makes us aware of possibilities of error in a way that threatens our knowledge, then checking, as a way to inquire into the validity of one’s belief, may actually have a counter-productive effect on the strength of one’s epistemic position.

Indeed, several authors in the clinical literature have argued that checking may not reduce, but paradoxically increase one’s sense of uncertainty (Rachman 2002; Tolin et al. 2001; Van den Hout and & Kindt 2003, 2004). As Van den Hout and Kindt claim: ‘checking is a highly counter-productive safety strategy: checking enhances doubt’ (2003, 315). Rachman similarly claimed that: ‘the more checking you do, the less confidence you have in your memory of the checking’ (2003, 630). According to these authors, repeated checking can worsen one’s epistemic position towards whether p, by fostering doubt and weakening confidence. Van den Hout and Kindt (2004) posit a psychological phenomenon responsible for this effect: checking increases familiarity with the checked matters, which means that the processing of high-level semantic aspects of the experience gets prioritized and that the processing of lower level, perceptual elements get inhibited (Roediger 1990). In discussing OCD patients’ memories of performed checks, Neisser (1981) coined the term ‘repisodic memory’ to refer to memories that are neither episodic nor semantic, but which contain events that have occurred repeatedly. Such memories are often experienced as episodic, but in fact they are not a recollection of a single episode, rather a recollection of common features of a series of events. As a result, memory for the experience becomes less vivid and detailed. Lack of detail and vividness may not affect actual memory accuracy, but it will make the recollection appear less trustworthy to the subject. This cognitive cascade results in reduced confidence in memory.

Taken together, this empirical data suggests that motivating OCD patients’s compulsions is an intense sense of uncertainty which repeatedly gives way to serious doubt. Being in this mental state of uncertainty and recurring doubt seems incompatible with being in the mental state of knowing or having "insight". My contention is thus that, once we grasp the difference between these two kinds of ‘what if?’ questioning, it is clear that (1) the ‘what if?’ questioning that is relevant to understanding obsessive thoughts is of the stronger kind and (2) this stronger ‘what if’ questioning is much more likely to threaten one’s state of knowledge. The questioning that motivates patients with Obsessive-compulsive disorder to keep checking the stove or door knob must be understood in a stronger sense: it is not a mere exploration of a scenario one knows to be non-actual, it is the consideration of a scenario that one takes to be both genuinely possible and actually threatening. The practical thinking that OCD patients engage in about whether they should go back and check the stove again involves threatening possible scenarios, such as one’s house burning down as a result of one’s negligence. It is not mere counterfactual reasoning, but practical thinking factoring in catastrophic practical consequences. Moreover, as we have seen, it is plausible that a hyperactive signal of error or uncertainty leads OCD subjects to more easily question their beliefs.

According to the picture I have just presented, obsessive thoughts (‘the stove might be on’) are thus generated by dysfunctional error signals, and sustained by the representation of negative practical consequences that an error could lead to. My contention is that this stronger kind of ‘what if’ questioning is fit to cast doubt on their antecedent belief that the stove is off. If this is so, the reason why compulsive re-checkers feel the urge to go back home a tenth time to manipulate the lock is not that they merely entertain a counterfactual scenario they view as non-actual. They feel the urge to go back and check because their ‘what if?’ questioning has thrown them back into a state of serious doubt. The ‘what if the door is open?’ questioning which assails OCD sufferers is one which overthrows the individual’s current epistemic base and drives them back to a state of uncertainty and doubt (Cochrane and Heaton 2017). It is precisely the doxastic shift that these ‘what if?’ obsessive thoughts trigger which grounds the motivation to launch a new checking. In other words, it is because they are now experiencing serious doubts about whether the door is in fact locked, as a result of engaging in this stronger kind of ‘what if?’ questioning, that they feel the urge to check.

One way for Taylor to respond to this objection would be to appeal to the fact that certain of the doubts that re-checkers have are in fact idle. Doubt is a response to evidence against a given proposition. This evidence constitutes one’s reasons to doubt, that is, one’s reasons not to believe p which, coupled with one’s reasons to believe p, constitute one’s grounds for doubting (Lee 2018; Howard-Snyder 2013). Doubt is a state of epistemic unsettledness that comes in degrees, depending on the strength of the reasons one has not to believe p. Now, some low degrees of doubt (‘slight’ doubt’) may be compatible with a state of knowledge. If the worries that pop up in the re-checkers’ mind were sufficiently idle, they might not undermine their knowledge of the relevant fact.

As has been highlighted by fallibilists, knowing that p requires a probability that is high enough to make the possibility that not-p idle (Fantl and McGrath 2009). Idle doubts are thus compatible with knowledge. However, as soon as the chance of not-p being the case is not idle, that is, as soon as we have serious doubts, knowledge is threatened. Moreover, when doubts regard high-stakes matters where one has a lot to lose if one forms an incorrect belief, these pragmatic factors might encroach on one’s knowledge, and make it more demanding for one to achieve or maintain knowledge in this context (Fantl and McGrath2009). High degrees of doubt (what I call ‘serious doubt’) amount to a ‘cognitive opposition to a proposition’s truth’ (Lee 2018, 155) and are incompatible with believing that p. In cases where people re-check on the basis of idle doubts, we might maintain that their knowledge is intact as they initiate a re-check. However, I believe that this does not reflect the standard cases of OCD re-checking. Standardly, the doubts that are strong enough to motivate the intention to re-check are of the non idle sort. The reason why individuals re-check is precisely that they are experiencing serious doubts.

This is supported by clinical descriptions of OCD. In individuals with OCD, thoughts expressing possible threats become obsessive because they are taken very seriously by the subject, and are typically accompanied by acute anxiety (Abramowitz, McKay, and Taylor 2008). Taylor (2020) himself bases his hypothesis on psychological research into anxiety-related disorders – particularly Generalized Anxiety Disorder – showing the prominence of ‘what if’ questions in the inner speech of patients, thereby highlighting the tight link between ‘what if’ questioning and anxiety. However, Taylor does not elaborate on the role and effects of anxiety in ‘what-if’ attitudes. I suggest that explicating the role of anxiety in these questionings supports the idea that the doubts they generate are non idle doubts.

Anxiety has been defined as an emotion whose role is to signal that we are facing a ‘problematic uncertainty’ (Kurth 2018, 2015). It is an emotion through which we appraise a ‘possible and uncertain threat’ (Miceli and Castelfranchi 2005). The two central components of the evaluative signal carried by anxiety are: a lack of information on our part (an uncertainty), and a possible threat or thwarting of one’s goals, which renders the lack of information ‘problematic’. Therefore, when anxiety accompanies a thought, the content of the thought is appraised as both uncertain and as threatening or problematic in some way.

A ‘what if?’ question is a specific interrogative form which focuses on the potential defeaters to one’s beliefs (what could make it true that the stove is on) and on the potential negative outcomes at stake (what would happen if the stove was on). ‘What if’ questions focus our attention (1) on the fallibility of one’s epistemic position, and (2) on the potential threats implied if our epistemic attitude towards p is inaccurate. When this process is triggered by or accompanied with anxiety, epistemic fallibility and stakes are interpreted as a ‘problematic uncertainty’. Thus, when a ‘what if?’ question is accompanied by anxiety, it is likely to motivate serious doubt. As Nagel (2010) has argued, anxiety is a mechanism which functions to generate a need for thorough information processing and motivate greater evidence gathering by making high-stakes salient. When signals or uncertainty are responded to with anxiety, the coupling of these two affective states inclines us towards doubt (Vazard 2019) . In other words, when detrimental ‘problematic’ outcomes are attached to a possible epistemic error, one will more easily be motivated to take the possibility that not p seriously, and gather more evidence on whether p. In individuals with OCD, the pathological anxiety which systematically accompanies ‘what if?’ questionings (triggered by the hyperactive error or uncertainty signals) launches a state of doubt and motivates the pursuit of additional evidence. While they might have no good reasons to doubt, the anxiety which tags a thought content as ‘possible and threatening’, as well as the dysfunctional error signalling (Cochrane and Heaton 2017), makes OCD sufferers feel that they do.

As I have argued, the kind of questioning which prompts the need to check in compulsive re-checkers is likely to be the kind of questioning which raises serious doubts, explaining why they form the intention to re-check. Now, the serious doubts experienced by compulsive re-checkers might (perhaps even most of the time) not be justified. As I have explained, in the case of OCD re-checkers, we have reasons to believe that these doubts are systematically generated by a dysfunctional disposition towards affective states of uncertainty, including anxiety. These processes repeatedly and unreliably signal a threatening lack of certainty, thereby prompting serious doubt, and the need to settle the question again through performing another check. Nonetheless, this is a problem for Taylor’s attempt to defend the compatibility of knowledge with the kind of ‘what if?’ questionings that motivate checking. If my proposal is valid, we should revise our view of re-checkers as subjects who both know that the stove is off and at the same time doubt that it is. Instead, as I have shown, it is more plausible that their antecedent judgement that the stove is off is overthrown by recurrent ‘what if?’ questioning accompanied by affective states which signal a threatening possibility of error and raise serious doubts. Thus, while I hold that Taylor is right in thinking that something like ‘what if not-p?’ attitudes repeatedly assail compulsive re-checkers, I have argued that they do not typically preserve the knowledge that p throughout the re-checking process.

This is not to say that my account is identical to Friedman’s account. While I believe that Friedman is right to point out that re-checkers do not know anymore as they initiate another check, I do not think it is accurate to think of re-checkers as being in a state of suspension of judgement on whether the stove is off. I instead believe that re-checkers are in a state of doubt: not a state of absolute epistemic neutrality, but a state of unsettledness which comes in degrees and is sensitive to the strength of the evidence one has against p. Compulsive re-checkers, I contend, are prompted to perform another check as a result of experiencing signals of problematic uncertainty which raise serious doubts and defeat their antecedent state of knowledge.

I take it that the discussion presented here contributes to research on compulsive re-checking by clarifying the plausible epistemic state that re-checkers are in as they launch another re-checking action. Clarifying the doxastic attitudes re-checkers are in is relevant for our understanding of the nature of the disorder itself, but can also be crucial for designing effective interventions. For instance, if we consider patients to be in a settled attitude (they believe or know that p) as they intend to re-check, then bringing their attention to the good evidence that grounds their knowledge might be the way to go. However, if instead we consider that they are assailed by doubts which undermine their state of knowledge, then a focus on examining the weakness of the evidence grounding these doubts might be more efficient.

Acknowledgements

I am particularly grateful to Jane Friedman for many enlightening discussions while hosting me at NYU in 2019-2020. I also sincerely thank Fabrice Teroni, Jacques Vollet, and Arturs Logins for their comments and suggestions on previous drafts of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Juliette Vazard

Juliette Vazard is a philosopher of mind, specializing in emotion–cognition interactions and philosophy of psychopathology. She is currently a Postdoctoral researcher at City University of New York, Graduate Center. In 2021 she completed a joint PhD at the University of Geneva (Swiss Center for Affective Sciences) and the Institut Jean Nicod in Paris. In 2024 she will be a postdoctoral fellow at the University of Toronto, Philosophy department.

Notes

1 Although Whitcomb and Friedman did not intend for their claims to apply to those checking behaviors that are performed by patients with Obsessive-compulsive disorder, I believe their claims are relevant for such cases and may help elucidate such pathological behaviors, at least for mild cases of OCD, and at the onset of the disorder.

2 This is assuming, of course, that no one has turned the stove on in the meantime. This raises the question of whether re-checkers are in fact trying to establish whether p remains true or is still true. However, this does not justify the frequency and chronological closeness of the checks.