Self-control and mechanisms of behavior: Why self-control is not a natural mental kind

Abstract

In this paper, I argue for two main hypotheses. First, that (philosophical) self-control is not a natural mental kind and, second, that there is no dedicated mechanism of self-control (indeed, the latter claim forms part of my argument for the former). By the first claim, I simply mean that those behaviors we label as “self-controlled” are a somewhat arbitrarily selected hodgepodge that do not have anything in common that distinguishes them from other behaviors. In other words, self-control is a gerrymandered property that does not correspond to a natural mental or psychological kind. By the second claim, I mean that self-controlled behaviors are not produced by a mechanism (or a set of them) that is not utilized in the production of other (non-self-controlled) behaviors. Not only is there no natural mental property of self-control, there is no mechanism (such as willpower) that is dedicated to producing self-controlled behavior. I further evaluate whether this account of self-control has enough explanatory power to account for a range of phenomena related to self-control (systematic self-control failures, etc.). I argue that my account does a better job of explaining these phenomena than accounts which appeal to a dedicated self-control mechanism.

Keywords:

• Behavior
• ego depletion
• mechanisms
• natural kind
• resources
• self-control

Acknowledgments

I would like to thank Stephen Kearns for his extremely helpful feedback on various drafts of this paper. I would also like to thank Al Mele and Randy Clarke for their helpful contributions.

Notes

1. Such a grouping is not entirely arbitrary—there are (non-distinctive) similarities between different self-controlled behaviors. However, as I argue in section 3, such a collection of behaviors is still largely arbitrary because these behaviors do not share a (natural) property unique to them.

2. Though I understand resources as mechanisms, it will be sometimes appropriate to treat resources and other mechanisms separately. In those cases, I talk of mechanisms and resources.

3. It is beyond the scope of my paper to examine any technical notion of self-control used in psychology. It is, however, a worthwhile endeavor to run a project similar to this for such technical notions. I expect that many issues discussed here can also be raised for various psychological notions of self-control. Given that there are many different and (seemingly) successful operationalizations of self-control, one may worry that self-control, as understood in (academic) psychology, is not a natural kind either.

4. Levy (2011) argues that weakness of will is not a psychological kind. I will not draw here any conclusions about weakness of will from my claims about self-control. My arguments here differ greatly from Levy’s.

5. By focusing on Mele’s account, for the reasons mentioned above, I have to omit discussing other prominent accounts of self-control in philosophy. I do, however, in section 6, look at Richard Holton’s account of willpower (2003, 2009) in relation to the question of the explanatory power of those accounts that posit a dedicated resource of self-control. I argue, contra Holton, that there is no distinct faculty of willpower and that accounts that don’t posit such a dedicated self-control mechanism are in fact preferable. (This raises problems for the overall plausibility of Holton’s but not of Mele’s account—which is one reason why my starting point is the philosophical notion of self-control as captured in Mele’s account).

6. Please note that Kennett’s discussion/account of self-control is far from exclusively focused on such issues.

7. In his many publications, Mele offers many details regarding the nature of self-control and provides a more refined characterization than the one given above. However, it is not necessary for my purposes here to go into the details of Mele’s analysis. Any other relevant details of Mele’s account are mentioned in section 3.

8. While Mele here talks of better judgment, I talk of best judgment. Nothing of importance turns on this.

9. See Mele (1990, 1995, p. 60).

10. See also Mele (1987, p. 54).

11. Please note that Mele understands deciding as a “momentary mental action of intention formation” (1992, pp. 8, 12).

12. Some disagree. For example, Henden argues that self-control may only serve “what one takes oneself to have most reason to do” (2008, pp. 73–74). Henden argues for this conclusion on the basis that “self-control involves more than control over action; it also involves control over self” (p. 73) which, according to Henden, does not apply to Mele’s case described above. I do not here have room to go into details and problems with Henden’s argument. However, one thing to note is that Henden seems to consider our best judgments—but not our desires and intentions—as being central parts of our selves. I see no reason to accept this dichotomy. We may deeply identify with our desires and intentions even if they do not stem from our best judgments.

13. I hesitate to attribute this argument directly to Mele, as he does not make this argument explicitly, but certain points he makes suggest it.

14. For more details, see Mele (2012, pp. 95–96). According to Mele’s experimental results, 46% of participants agreed that in a case similar to the one above, the agent displays self-control, and 64% participants agreed that such a case involves a display of willpower. These results also, to a degree, answer Kennett’s worry (2001) that the so-called orthonomous self-control, the exercises of which are limited to bringing “the agent’s actions into line with her view about what is, all things considered, desirable in the circumstances … best answers to the common-sense notion of self-control” (2001, pp. 33–34).

15. See, for example, Arpaly & Schroeder, 1999; Greenspan, 2016; Hollander-Blumoff, 2012; Holton & Shute, 2007.

16. For more discussion on possible implications, see section 7.

17. For an overview of the natural kinds debate, see Bird and Tobin (2017).

18. See Kripke (1980).

19. This is not to say that philosophical theories of the mental always agree with the folk on the nature of such mental items. Rather, it is to say that philosophers recognize the resilience and usefulness of such everyday categories of the mental. Such categories contrast with more technical notions that one may find in psychology, neuroscience, and the philosophy of mind (such as the notions of memory traces, aliefs, neural circuits, psychopathologies, etc.).

20. Please note that the claim contained in premise 3 is still required in support of premises 1 and 2.

21. There are further reasons to accept premise 1. In particular, those mental terms we use to describe self-controlled behaviors may themselves not denote natural mental kinds. Consider, for instance, the notion of best judgment. If an agent judges it best to A, but also that A-ing is supererogatory, is this a best judgment? If an agent judges it best to A, but believes she ought not to A, is this a best judgment? If an agent judges it objectively best to A, but has no interest in doing what is objectively best, is this a best judgment? The answers to these questions are, at least to me, unclear. However they are answered (whichever judgments count as best judgments and whichever do not), it seems that the boundary we draw will be somewhat arbitrary—an arbitrariness that the kind self-control will thus inherit.

22. Another rejoinder is that both cases (a) and (b) do feature self-control. According to this reply, behaviors in accordance with best judgments (or intentions) count as not self-controlled only if there is no counter-motivation or no effort expended in service of one’s judgment or intention. On this approach, there is a non-arbitrary line between self-control and non-self-control. The main issue with this reply is that it gets the extension of self-control entirely wrong. Neither Mele, nor anyone else, would (or should) accept that perfectly ordinary, unexceptional actions involve exercises of self-control simply because there was a tiny amount of counter-motivation present that took no further effort (than is usually expended) to overcome. Indeed, the vast majority of our actions are performed in the face of at least some motivation to do something else, however small. Unless we countenance the near-ubiquity of self-control, we should accept that successful action in the face of minor counter-motivation does not involve self-control.

23. According to Sheeran (2002), implementation intentions have “medium” effects on moderating ego depletion, a phenomenon discussed in Section 5.

24. Schmeichel and Vohs explain that “People may construe events at different levels of abstraction … A high-level construal … refers to the global, superordinate, abstract features of an event, whereas a low-level construal … refers to the local, subordinate, and concrete features of an event” (2009, p. 775).

25. This thesis is one of main tenets of the strength/resource model of self-control (see Baumeister, Heatherton, & Tice, 1994) which is currently the dominant model of self-control.

26. For a meta-analysis on ego depletion, see Hagger, Wood, Stiff, and Chatzisarantis (2010), also Inzlicht and Schmeichel (2012).

27. At least one exception to this is a set of experiments which test how glucose affects performance in self-control tasks. There are some contradictory results with regard to glucose’s effects (e.g., Gailliot, Peruche, Plant, & Baumeister, 2009; Molden et al., 2012).

28. Baumeister, Schmeichel, and Vohs (2007) explain that “Self-regulation is one the self’s major executive functions … The other major executive function of the self is choice” (p. 517), and “Not only self-regulation, but also acts of effortful choice and volition use the same resource” (p. 526).

29. This proposal is compatible with the general resource model; it is not supposed to compete with/replace the latter.

30. Mental effort has been typically discussed in relation to cognitive goals only. I propose it may be more fruitful to understand it as connected with executive goals too. (These types of goals overlap to some degree.)

31. While it is debatable how best to operationalize mental effort, several plausible candidates are available. Previously, mental effort has been operationalized via different psychophysical variables (e.g., muscle tension, glucose reactivity, etc.) as well as behavioral and subjective measures (for a review, see Fairclough & Houston, 2004). Selecting an appropriate operationalization will depend on the specifics of the mental effort theory and comparisons between competing operationalizations.

32. This claim is compatible with how I describe mental effort—I do not argue that mental effort is used with all cognitive and executive goals.

33. In the above section, I have focused on the strength model of self-control as it is arguably the most prominent psychological account of self-control/self-control mechanism(s). There is, of course, a number of competing accounts, some of which yield other promising candidates for a dedicated mechanism of self-control. One such promising candidate is the process of intertemporal bargaining (Ainslie, 2001) (I thank an anonymous referee for highlighting the promise of this approach). Ainslie explains that:

Maintenance and change of choice will be governed by intertemporal bargaining, the activity in which reward-seeking processes that share some goals (e.g., long-term sobriety) but not others (the pleasure of having some drinks now) maximize their individual expected rewards, discounted hyperbolically to the current moment. This limited warfare relationship is familiar in interpersonal situations … where it often gives rise to “self-enforcing contract” … such as nations’ avoidance of using a nuclear weapon lest nuclear warfare become general. In interpersonal bargaining, stability is achieved in the absence of an overarching government by the parties’ recognition of repeated prisoner’s dilemma incentives. In intertemporal bargaining, personal rules arise through a similar recognition by an individual in successive motivational states, with the difference that in a future state she is not motivated to retaliate, as it were, against herself in the past states where she has defected (2011, p. 68, emphasis added).

If an agent successfully engages in intertemporal bargaining, this may be sufficient for an agent’s behavior to count as an exercise of self-control (depending on how exactly the intertemporal bargaining is executed, it may count both as an orthodox and as an unorthodox exercise of self-control on Mele’s account). If intertemporal bargaining is sufficient for self-control, it can be then said to exclusively produce self-controlled behavior, making it, prima facie, a viable candidate for a dedicated self-control mechanism. Recall, however, that a dedicated self-control resource must produce all self-controlled behavior. But intertemporal bargaining is not necessary on Mele’s account of self-control. In other words, one can successfully exercise self-control without employing inter-temporal bargaining to overcome counter-motivation. Suppose, for instance, that one is attempting to avoid eating a very tempting chocolate cake. One method of exercising self-control would be to imagine (in line with one’s best judgment about one’s diet) that the cake is filled with squiggly worms as a means of reducing how appetizing the cake appears. This method, when executed successfully (there is no reason to think that such method could not work at least sometimes), need not involve employing the process of intertemporal bargaining. Imagining the cake in this way as a means of combating counter-motivation does not require that one acts on any personal rule (such as to “never eat chocolate cake at all”), even a tacit one. (It also should be noted that Ainslie himself does not claim that intertemporal bargaining is necessary for self-control: Ainslie identifies other relevant mechanisms which involve the manipulation of attention, the preparation of emotion and extrapsychic commitments.) Similarly, then, to the resource measured in the ego-depletion experiments, intertemporal bargaining is not a dedicated self-control mechanism.

34. Also see Holton (2009). Holton argues that appealing to a distinct faculty of willpower provides a better explanation than those which explain agent’s actions by appealing to the strength of the different conative inputs (2003, 2009).

35. Holton himself does not spell out in detail to what a faculty of willpower amounts.

36. Also, I can extend my initial thesis so that it applies to faculties as well—self-control does not employ a dedicated faculty.

37. Cf. Holton (2003, p. 42).

38. Holton makes this point in terms of willpower being stronger (2003, p. 43).

39. Another reason may be that one’s self-control abilities are systematically affected other than through the lack of appropriate learning or development. I discuss systematic effects on self-control in section 6.4.

40. Cf. Holton (2003, p. 41).

41. One might argue that self-control as psychologists understand it is not informed by the philosophical notion to any great degree so that this problem does not arise. Perhaps, that is, psychologists really do pick out a natural kind with their use of the term “self-control” even if philosophers do not with theirs. Even if this is so, my point that (philosophical) self-control is of little help to psychologists stands. Further, I suspect that, given the multitude of ways psychologists understand and operationalize self-control, they are in no better a position to claim that they pick out a natural mental kind. Consider, for instance, the following paragraph from Muraven & Baumeister, 2000 (with their references omitted):

Self-control is the exertion of control over the self by the self. That is, self-control occurs when a person (or other organism) attempts to change the way he or she would otherwise think, feel, or behave. Self-control behaviors are designed to maximize the long-term best interests of the individual … People exert self-control when they follow rules or inhibit immediate desires to delay gratification … Without self-control, the person would carry out the normal, typical, or desired behavior … Self-control involves overriding or inhibiting competing urges, behaviors, or desires. (p. 247).

Self-control is characterized in multiple ways, each of which can come apart from the others. What’s more, some of the ways in which self-control is characterized are similar to the notion we have been discussing and thus may be open to similar problems. In essence, though psychologists may be dealing with genuine natural kinds in their research on self-control (e.g., ego-depletion, desires), it is far from obvious that there is a natural mental kind that can be properly labeled “self-control.” Further work concerning psychological notions of self-control is needed to substantiate or refute this hypothesis.

42. See Putnam (1975).

43. Please note that my argument here is not that self-control is analogous to jade in the sense that they both fail to be natural kinds in a similar way, but rather that the term “self-control” and the term “jade” are analogous in that they both manage to refer despite both of them being failed natural kind terms. Jade and the term “jade” are thus brought up here solely for the purpose of showing that some failed natural kind terms do refer. As it happens, I do believe that jade and self-control are analogous in how they fail to be natural kinds—they are both irreducibly disjunctive kinds. Also, just as jadeite and nephrite have some superficial resemblance (which doesn’t extend to other substances) that lead to them being clustered under “jade”, so does overcoming counter-motivation while acting in line with one’s best judgment and overcoming counter-motivation while acting in line with one’s intentions. However, it is not my aim here to argue for any kind strong analogy between self-control and jade.