Emotions and Natural Kinds

In a recent post and its comments, some questions were raised about Griffiths’s theory of emotions and whether emotions are natural kinds.

Griffiths’s (1997) central point was that it is unlikely that all the psychological states and processes that fall under the vernacular category of emotion are similar to one another to a sufficient degree to allow for the emergence of a unified scientific psychology of emotion.  Notice that this is not to say that emotions have nothing in common, or that what they have in common cannot be discovered. It is simply to say that instances of folk emotion categories do not share the sorts of properties that are the focus of investigation in the disciplines comprised within scientific psychology, a vague but not useless umbrella term (e.g. neuroscience and cognitive science belong to scientific psychology, sociology and critical theory do not).
Griffiths’ (1997) account of natural kinds was inspired by Boyd’s (1991) analysis of homeostatic property clusters kinds. Griffiths used the term natural kind “to denote categories that admit reliable extrapolation from samples of the category to the whole category”, and argued that “[i]deally, a natural kind should allow very reliable predictions in a large domain of properties”. The proper method for studying the emotions, the 1997 book argued, is to focus on subclasses of emotions whose members share a dimension of scientifically relevant similarity.

Now, what are such subclasses? The only type  of emotions for which we currently have strong evidence of naturalness in Boyd’s sense are affect programs (or basic emotions), namely biologically based and pan-cultural suites of short-term, coordinated and automated responses which includes measurable physiological changes, stereotyped facial expressions and action tendencies.

Ekman’s most recent list of such programs comprises surprise, amusement, anger, contempt, joy, disgust, embarrassment, excitement, fear, guilt, pride in achievement, relief, sadness/distress, satisfaction, sensory pleasure, and shame (notice: we currently only have strong evidence for basicness with respect to anger, fear, disgust, sadness, surprise and joy and I doubt we will have evidence of basicness for many of the items in Ekman’s latest list).

Differently from Ekman, Griffiths acknowledges that affect programs are not the only things that qualify as folk emotions, but he points out that many folk emotions are sufficiently different from affect programs to require a distinct scientific psychology. Griffiths (1997) suggested two main research paths along which we may develop theories suitable for finding out other natural kinds of emotions. Some emotions appear to require responding in a more cognitively complex way to stimuli than it is the case for basic emotions. This seems to be the case for emotions such as envy, jealousy, moral indignation, resentment and others. Griffiths (1997) proposed to call such emotions higher cognitive or complex emotions. On the other hand, on some occasions emotions appear to “involve an internalized cultural model of appropriate behavior”. Griffiths suggests that this appears to be the case for emotions such as “going postal”, which seem to follow a script “derived from real or fictional incidents that are culturally salient”. Griffiths proposes to call such emotions “socially sustained pretenses”.

These distinctions have been understood as boiling down to a fragmentation of the folk domain into three natural kinds of emotions: basic, higher cognitive and socially sustained ones. I do not think this fragmentation is what Griffiths intended, and at any rate it does not reflect what he currently thinks (I think). As Paul and I argue in “Emotions in the Wild
, the higher cognitive-basic divide is not a useful one (I add: the notion of a socially sustained pretense is even less useful than that of a higher cognitive emotion if understood as singling out a natural kind of emotion). Here is the relevant quote:

“First, [the basic-higher cognitive divide] suggests that the occurrence of these emotions [e.g. guilt, envy] necessarily involves conceptual thought, a view we have strongly questioned. Second, it seems almost irresistible to align the distinction between ‘basic emotions’ and ‘higher cognitive emotions’ with a distinction between two sets of vernacular emotion categories, anger, disgust, surprise being paradigmatically ‘basic’ and guilt, shame and embarrassment being paradigmatically ‘higher’. We believe, however, that there is as much need for pluralism in the theoretical treatment of subordinate categories of emotion as there is in the treatment of the superordinate category of emotion: some instances of anger, disgust or surprise may be adequately accounted for in the affect program framework, but others may require other theoretical perspectives, and the same holds for episodes of guilt, shame, or embarrassment”.

If all of this is right, the bottom line is that (a) we need to come up with theoretical constructs other than “basic emotions” to help us understand the sorts of emotions that are involved in the phenomena we are most interested in (e.g. art, morality, mental disorder, etc.) and (b) we need to stop striving to theorize about the items that qualify as bona fide folk emotions, because no unified theory is likely to emerge about them all in scientific psychology.


  1. Is the Dynamic System Theory, or Embodiment Approach, that views emotions as self-organized (build from the interaction between infant and caregiver, for example) rather than the pre-adapted (or hardwired) account, the proper solution to your claim of a much needed new theoretical framework beyond the universalist view of the basic emotions?

    If this is so, and i´m not defending one view over the other, why is it that the infant when she experiences joy produces a particular facial expression(face signature) and not other? why blind people smile? Why exist cross-cultural invariants even if we recognize Prinz neo-humean remarks on the development of emotions and the role of experience? Why the underlying systems (neural, endocrine…) subserving emotions they have a clear phylogenetic root with certain mechanisms present in all species, and absolutely, in the primate linage? Why the same neural afferent/efferent pathways are involve in all the emotions?, that is, they produce an skeleto-muscular reaction with the same pattern of activation of the autonomic nervous system although with variances. Even if all of the above research questions are ill-posed, with respect of the clash between emotions and consciousness, why all emotions are felt with some specific and univocal dimensions (valence; arousal…) In relation with the communicative role of emotions, if we cannot accord a unified structure of emotions under the universalist account of the basic emotions theory, how we save the realibility of the message convey by emotions to conspecifics? why are emotions perceived as signalling devices to be intentionally interpreted?

    I think that with many things in life truth lies in the middle, and you and Griffiths have right as Calvert with his discoveries about dissociations of emotions; but categorical or unidimensional proposals are right as well.

  2. Today,i bought a new Replica Watches. The elegant appearance and favourab accurate function of the watch appeals me.It has various collections for you to choose from and most important is the conveniet means of payment,you can recieve your Replica Watches soon after you pay it.If you have any problem,you can call online customer service,there you can get a satisfactory reply.070