Extended Prolepsis 3: CAPS to the Rescue

‘CAPS’ stands for ‘cognitive-affective personality system’ – a framework developed by Walter Mischel and Yuichi Shoda (1995) to bring together personists and situationists under an interactionist big tent.  The framework is complicated, and seems to be ill-understood even by some of the philosophers who draw on it.

In this context, the most important thing to bear in mind about CAPS is that it is not a theory of personality.  As Shoda and Mischel (2006) emphasize, it is better understood as a framework or meta-theory from which to build a theory of personality, and is intentionally “content free.”  Despite this, Nancy Snow (2009, p. 13) suggests that it is “plausible to understand virtues as traditionally conceived as a subset of CAPS traits.”  Daniel Russell (2009, p. 269) contends that a virtue as traditionally conceived is particular kind of “pattern of cognitive-affective processes” in the CAPS vein.[1]  Christian Miller (2009) initially seemed warm to CAPS, though in subsequent work (forthcoming b), his enthusiasm has cooled.  Jonathan Webber (2013) argues that CAPS “confirms the possibility of virtue.”  All of the above are philosophers arguing in favor of virtue ethics; however, some defenders of virtue epistemology also draw on the CAPS framework to bolster their arguments (Fairweather & Montemayor unpublished manuscript; Axtell current volume).

CAPS cannot be so easily harnessed.  It is a theory about which kinds of entities belong in the ontology of a first-order psychological theory may.  Those entities include (features of) situations, cognitive-affective units, and behaviors.  Situations themselves are subdivided into intra-psychological situations (e.g., moods), inter-psychological situations (e.g., being threatened or teased), and extra-psychological situations (e.g., being in a loud environment).  Cognitive-affective units are subdivided into encodings (e.g., categories for the self, others, events, and situations), expectancies and beliefs, affects and emotions, goals and values, and plans (Mischel & Shoda 1995).  According to CAPS, situations differentially influence cognitive-affective units, which differentially influence one another, and which together conspire to produce behaviors, which in turn influence situations.  That’s the CAPS framework.  A first-order psychological theory counts as CAPS-theoretic if and only if it refers to the entities in the CAPS ontology and specifies the relations between them as the CAPS framework indicates.  Thus, a CAPS-theoretic virtue psychology – a first-order theory that might substantiate the hypothesis that virtues as traditionally conceived are empirically supportable – would be framed in terms of the ontology furnished by the CAPS framework.  Such a theory, were it to be constructed, could then be tested to the extent possible given constraints of time, budget, experimental design, and research ethics.

Screen Shot 2013-08-31 at 7.18.01 PM

No such theory exists.  Moreover, it’s unclear why anyone thinks that it would be possible to construct one because the CAPS ontology is incommensurable with the ontology of traditional virtue theory.  One example: CAPS countenances powerful, unconscious influences of subtle, seemingly irrelevant intra-personal situations on cognitive-affective units.  I defy anyone to find a traditional virtue theory that does this.  Another example: virtue theories tend to countenance reasons in their ontologies (Russell 2009).  Where are the reasons in CAPS?  They’re not behaviors or cognitive-affective units, so presumably they are features of the situation.  But none of the situational subcategories seems to be a good place to bin them.  Reasons are certainly not moods (intra-psychological situations).  Nor are they things like being in a loud environment (extra-psychological situations).  That leaves only inter-psychological situations, but that someone needs help and that the consensus opinion is wrong are not the sorts of inter-psychological situations that Mischel & Shoda (1995) have in mind.  A third example: virtue epistemologists tend to countenance evidence in their ontologies.  But, just as there appears to be no place for reasons in the CAPS ontology, so there appears to be no place for evidence.[2]  This suggests not only that there is no CAPS-theoretic virtue psychology, but also that there could not be such a psychology.

But suppose that I’m wrong about this and that someone will soon come along and construct a CAPS-theoretic virtue psychology.  The question then would be whether there is any empirical support for that theory.  I have my own hunch about what the answer would be, but that’s not the point.  The point is that one would have to actually do the theoretical and empirical dirty work to establish whether the theory was supported.  There is some support for some first-order CAPS theories (along, of course, with evidence against others and – more commonly – a dearth of evidence either way).  For instance, Shoda, Mischel, & Wright (1994) drew on nearly a decade of empirical research to argue for a CAPS-theoretic psychology of children.  This theory had five situations (peer approach, peer tease, adult praise, adult warn, and adult punish), five behaviors (verbal aggression, physical aggression, whining, compliance, and friendliness), and no specified cognitive-affective units.  Their meticulous work on this relatively modest model revealed consistency coefficients of .19 for friendliness, .28 for whining, .41 for compliance, .32 for physical aggression, and .47 for verbal aggression – higher than is often found (Mischel 1968), but still nothing like what one would need to reliably predict and explain particular behaviors (let alone cognitions and affects).

In context, then, we can see that the task of crafting and confirming a CAPS-theoretic virtue psychology would be monumental indeed.  Thus, even in the best-case scenario, CAPS lends no credibility to traditional virtue theory.  To reiterate, the best-case scenario is that a merely possible theory will be confirmed by undreamt-of experimental data.



[1] Russell also mistakenly claims that CAPS is a situationist, rather than an interactionist, psychological framework.

[2] This should be unsurprising to anyone who sympathizes with the view that reasons are evidence – evidence for what to do in the moral case and what to believe in the epistemic case (Kearns & Star 2009).

7 thoughts on “Extended Prolepsis 3: CAPS to the Rescue

  1. Pingback: Epistemic Situationism: An Extended Prolepsis | philosophy and other thoughts

  2. Re your comment ” I defy anyone to find a traditional virtue theory that does this” —

    Hasn’t Kamtekar (Ethics 2004) already met this challenge in her discussion of Plato, Aristotle, and Marcus Aurelius? And to what extent have Sarkissian (Phil Imprint 2010) and Slingerland (Ethics 2011) met it in their discussions of Confucianism?

    • I don’t see how Kamtekar responds to this in the slightest.

      Stoics such as Marcus Aurelius:
      1) “Wipe out your external impression. Stay the movement of the puppetstrings.”
      2) “adoption of a personal role model as one’s ideal”
      3) “imagine yourself in situations that will test your commitment to your values”
      4) “analyze what is happening to you into what is “given” and what is your contribution”
      What does any of that have to do with “unconscious influences of subtle, seemingly irrelevant intra-personal situations on cognitive-affective units”? It’s a call for heroic levels of reflection, deliberation, and disengagement from the immediacy of action that characterizes the vast majority of agency.

      The Plato/Aristotle stuff is even less helpful. According to Kamtekar, they think that your ultimate values straightforwardly influence and predict your behavior. Inconsistency arises because of conflicts in ultimate values. But according to the CAPS model, inconsistencies arise not (just or primarily) because of conflicts between values but because of the unconscious influence of subtle, seemingly irrelevant intra-personal situations on cognitive-affective units. There’s no place for such influence in the ancient character ontologies.

      • I meant her point that for the ancient Greeks the regulation of behaviour by endorsed values is an achievement of the virtuous and, to a lesser extent, the continent. The default position from which we should ascend to that achievement is one of inconsistency due to irrelevant situational influences. These are the puppet strings that manipulate the akratic, the vicious, and the ethically disengaged.

Leave a Reply