Colombetti

Damasio's distinction between primary and secondary emotions assumes a traditional tiered model of the brain. I argue that the distinction is problematic and that, at best, it describes two extreme classes of emotional episodes. There are many other cases of emotion that seem to require a more integrated and dynamical account of brain and bodily mechanisms.

1 Primary and secondary emotions

Antonio Damasio (1994) divides the realm of the emotions into two groups, the primary and the secondary ones. Primary emotions are defined as "innately wired", "preorganised responses" to stimulus-features such as size, type of motion and type of sound. The baby chick's reaction (hiding the head) to wide-winged objects flying overhead at a certain speed is a typical primary emotion. All that is needed to elicit a primary emotion is a rough, quick and dirty detection of certain features; consciousness is not required. Primary emotions are generated in the part of the brain often called the limbic system, and from there they project onto the body (nervous and endocrine systems) and other parts of the brain. The limbic system, or "emotional brain", was first mentioned by Paul MacLean as early as the 1940s; he saw it as sandwiched between the oldest "reptilian brain", constituted mainly by the thalamus, and the "neomammalian brain" or neocortex, the seat of thought (MacLean, 1970). Although the concept is quite old and not everyone adopts it (e.g. LeDoux, 1996), many distinguished neurologists still use it as if it were unproblematic (Damasio and also, for example, Freeman, 1999). One of its most studied components is the amygdala, generally seen as importantly involved in the production of somatic responses typical of fear and anger (e.g. release of adrenaline, noradrenaline, freezing, etc.).

Secondary emotions, in Damasio's definition, "occur once we begin experiencing feelings and forming systematic connections between categories of objects and situations, on the one hand, and primary emotions, on the other" (1995, p.134, emphasis in original). For example, if we imagine meeting an old friend, we may undergo bodily changes typical of primary emotions (the skin may flush, the heart race, etc.). But in this case the bodily reaction would be triggered by what Damasio calls a "mental image" - namely, it would be triggered "off-line", not directly by the perception of a stimulus, but rather via (consciously entertained) thoughts. The limbic system is not sufficient for secondary emotions, because it is not able to support "categories of objects and situations". The brain circuitry for secondary emotions must thus be larger, and - Damasio speculates - include mechanisms for primary emotions as well as parts of the cortex responsible for the elaboration of complex stimuli (such as evaluations and expectations); it seems more plausible and economical for nature to fashion new emotions from older mechanisms, rather than adding separate new ones.

2 Problems with the distinction

I think that Damasio's distinction illustrates, at best, two extreme classes of emotional responses; in-between there is room for cases that are not easily classified as either primary or secondary.

Consider primary emotions first. According to Damasio's characterisation, they are triggered by a specific set of stimuli to which the organism is innately disposed to respond in a specific, pre-packaged way. "Innateness" thus refers to both the input side of the emotion (the stimulus) and the output side (the bodily responses) (Griffiths, 1997). This excludes, for example, that fear responses acquired via fear conditioning are primary. Fear conditioning is the association of a frightening stimulus such as a mild shock (the "unconditioned stimulus"), with a neutral one (the "conditioned stimulus"). Exposure to a number of presentations of the two stimuli together can bring about fear responses in the presence of the conditioned stimulus only. In his studies of auditory fear conditioning in rats, Joseph LeDoux (1996) has discovered that damaging the cortex after conditioning has been induced does not prevent fear responses. Rats keep showing typical fear reactions (freezing, increase in heart rate) even when parts of their thalamus are severed; eventually, only the pathway between the auditory thalamus and the amygdala is necessary to support those responses. This "thalamo-amygdala pathway" can thus trigger a fear response bypassing the cortex entirely; it is an automatic, quick-and-dirty response whose brain machinery spans the thalamus and the limbic system. However, Damasio's definition must exclude it as a case of primary emotion because it is acquired.

A possible move here would be to say that primary emotions include responses to conditioned stimuli that have been paired with unconditioned stimuli to which subjects react "innately". This move, however, is not unproblematic. Hardcastle (1999) argues, for example, that cortical activity may play a role in the development of fear conditioning. She mentions a study on auditory fear conditioning by Schneidermann et al. (1974), in which rabbits were exposed to two tones. Only one of the tones was associated with an electrical shock and, accordingly, rabbits started to exhibit fear conditioning to that tone. But when the auditory cortex was lesioned after conditioning had been induced, the rabbits exhibited fear conditioning to both tones. For Hardcastle this shows that the amygdala and the cortex cooperate to develop fear conditioning to a specific tone. The brain takes advantage of its own complexity and exploits all its resources: "without cortex, you get one neuronal configuration; with cortex, you get another" (Hardcastle, 1999, p.243).

Damasio's characterisation of secondary emotions is also problematic. In his example he asks us to imagine meeting an old friend, but he does not say anything about the case of actually meeting her. The two cases are importantly different. When you imagine, you consciously entertain a thought (an "image", as Damasio says), and indulge on its details. When you meet the old friend, your response is more similar to the quick-and-dirty one: you see her (or hear her voice) and your heart starts to race, without the need of any detailed conscious thought. The response is not primary (in Damasio's sense), however, because we are not hard-wired to respond automatically to a specific friend's face. So, what is it? Maybe it is another case of emotion acquired through conditioning. The unconditioned stimuli were events that happened in the past in association with that particular friend, so that seeing her now automatically triggers an emotion. Once again, Damasio's characterisation of primary and secondary emotions does not cover this case.

The following example presents a more complicated and intriguing case. Consider Shakespeare's portrait of Othello's jealousy. In one sense, it perfectly fits Damasio's characterisation of secondary emotions (as Damasio himself writes, explaining how primary emotions are elicited "does not do justice to what Othello goes through in his mind before he develops jealousy and anger"; op.cit., p.130). Because of the vivid images conjured up by Iago, Othello undergoes strong feelings and bodily reactions, from headaches to epileptic crises. But see what happens in Act IV; Cassio's mistress, Bianca, storms in and, angry because Cassio has neglected her, hurls Desdemona's handkerchief to him (Bianca had found Desdemona's handkerchief in Cassio's drawer, where it had been put by Iago). Othello witnesses the fact and, convinced to have a proof of Desdemona's unfaithfulness, makes up his mind to kill her.

What kind of emotion is the one triggered by the perception of the infamous handkerchief? What happens to Othello when he sees it? We imagine him being overwhelmed by several "gut feelings", sensations, a sense of unreality maybe, and a storm of thoughts and images. Again, according to Damasio's criteria this is not a primary response, because surely Othello has no innate, hardwired predisposition to feel jealousy when he sees handkerchief-features. Nor has Othello been exposed to the joint association of the handkerchief and the perception of Desdemona and Cassio together; his reaction certainly depends on what he has "gone through in his mind" before. However, it does not involve the conscious entertainment of the images induced by Iago (or some other image): the perception of the handkerchief seems enough to trigger Othello's response. The reaction induced by the sight of the handkerchief is thus different from the one triggered by just imagining Desdemona and Cassio together, and it is also different from seeing them together (the latter is perhaps more similar to actually meeting the old friend). It has the quick-and-dirty qualitative aspect of primary emotions, and it triggers a bodily reaction - as well as further emotions and deliberations - in an automatic or quasi-automatic way.

3 Blending cortical, limbic and somatic processes

It thus seems that cortical, limbic and somatic processes are more integrated that Damasio's framework implies. Quick-and-dirty responses may involve cortical activity, and thought-dependent emotions may not require consciously entertained images or thoughts. Features typical of "primary emotions", in other words, belong also to thought-dependent emotions, and vice-versa; and consciously entertained images are not the hallmark of thought-dependent emotions. This is compatible with more recent neurological frameworks which, unlike MacLean's tiered model, posit a deep integration of older and newer brain parts. For example, for Quartz and Sejnowski

the increasing complexity [of the brain] does not arise by creating new brain areas, but by enhancing the function of those that already exist. This is a different view from that of the triune brain of Paul MacLean … [I]t is a mistake to think that one area is more “primitive” than the other, since every brain system has components from all of these levels, strongly interacting and evolving together … (Quartz and Sejnowski, 2002, p.91)

In particular, they look at how the ventral tegmental area in the basal ganglia (located beneath the cortex), which contains dopamine, influences the development of the cortex. It does not seem possible to isolate MacLean's three brain layers without disrupting their functionality.

A question now is whether one should altogether give up the primary/secondary distinction, or amend it so that it can fit a more dynamical picture of the brain. Interestingly, Damasio's views are often more dynamical than the primary/secondary distinction implies (e.g. see Damasio,1999). I suspect that the latter is, above all, the product of traditional dichotomies characteristic of emotion theories. The debate in the psychology of emotion is in fact typically formulated in dichotomous terms (and confused ones): are emotions visceral or cognitive? Are they automatic or conscious? Are they preceded by cognitive appraisals (or "evaluations") or are the appraisals automatic? (Lazarus, 1984; Zajonc, 1980 Ekman, 1980; the tendency to dichotomise the emotions goes back at least to Descartes' Les Passions de l'Ame). Damasio wants to allow for all these cases: emotions are both visceral and cognitive, automatic and conscious, etc. Some can be both (there is the primary fear for an approaching bear, and the secondary fear for tomorrow's examination), some only secondary (existentialist Angst, for example), and maybe some only primary. Damasio has probably inherited (and succumbed to) traditional dichotomous talks, despite the fact that his view on the workings of the brain is often more complex and dynamic, and calls for a richer conceptualisation.

Still, there might be some utility in those distinctions. Some emotions seem more dependent on thoughts than others (compare Othello's jealousy with the fear for an approaching bear), and some seem more culture-dependent than others. Is there a way to account for those differences without adopting a tiered approach to the brain?

A dynamical view of the brain does not necessarily imply a fully bi-directional connectivity among all its parts, nor a completely heterarchical account. Different emotions may involve different brain (and bodily) dynamics, with different mutual influences scattered across different orders of complexity in the brain (and body). Complex human brain processes probably include mechanisms that in survival-threatening circumstances, for example, act in a fast, quick-and-dirty way, and may even involve informationally encapsulated brain pathways ("modules" in Fodor's terms; see Fodor, 1983). However, in different contexts the same brain mechanisms may interact with other brain parts in different ways. Some emotions may be triggered by consciously entertained images through very complex loops that are sustainable by human brains only. In-between these case there is room for various possibilities of mutual interactions among various brain parts, some of which are illustrated by the examples above. Many emotions are likely to involve cortical activity, but not the conscious entertaining of images; some of them may depend on the perception of a stimulus that so far has been associated, in one's mind, with certain thoughts (Othello and the handkerchief); some may depend on the perception of a stimulus that was, in the past, recurrent in someone's experiences and that has come to have a certain emotional meaning accordingly (meeting an old friend, seeing a person who assaulted you). The latter can, in turn, involve different dynamics according to the context that created the association between emotion and stimulus. Some contexts require less instances of the association than others, which in turn depends on the complex interactivity between genetic and environmental resources. The idea, in short, is that different dynamics are implicated as time unfolds, and the interaction of various brain and bodily mechanisms unfolding along different time scales creates an extremely rich range of possible emotions and behaviours.

This view tries to do justice to the complexity of the brain and bodily mechanisms suppporting emotions by avoiding dichotomous talks. It does not assume that emotion is a natural kind, nor that emotion is strictly separate from other faculties (such as cognition). It is probably because we have the label "emotion" that we describe different dynamics of the organisms as emotional, and others as non-emotional (although the fact that we continuously use this label may justify its appropriateness; I will not discuss this issue here).

This does justice also to the fact that emotions usually feel differently according to the context, even when the emotions in question go under the same label. Damasio's distinction implies that primary and secondary emotions must feel the same. If secondary emotions differ from primary ones only because they are triggered by the cortex, their physiological arousal will be the same as the one involved in primary emotions, and thus feel the same (recall that Damasio believes that feelings are perceptions of bodily changes). For example, the fear for an approaching bear (primary, triggered by a quick and dirty subcortical mechanism) should feel like the fear for tomorrow's examination (secondary, triggered by cortical processes and images) because, by definition, they involve the same limbic mechanisms and accordingly the same bodily responses. Introspection suggests that this is not the case. Different feelings in different instances of fear may depend on different organismic dynamics supporting such different instances.

Similarly Damasio's distinction and the tiered model of the brain it assumes cannot easily account for the role of the body in emotion elicitation. Being in a certain physiological state can influence the way in which we evaluate situations. There is evidence, in the psychological literature, that our state of physical arousal influences the way in which we appraise the situations we are in. The famous experiments conducted by Schachter and Singer (1962), for example, showed that subjects injected with adrenaline are more easily manipulated into an emotional state than subjects injected with a placebo (this aspect of their studies is often forgotten by those who believe that emotions depend uniquely on cognitive-evaluative processes). In addition, most emotion theorists agree that some emotions are determined (at least in part) by the state of the body; e.g. fear is associated with high rates of adrenaline, anger with high rates of noradrenaline (LeDoux, 1996). This means that if your organism happens to be in a state characterised by a high rate of noradrenaline, you will be more likely to be angry than scared, even in a fear-eliciting context. The body thus provides a limit to the range of possible emotions one can feel in a certain context. It is not clear that Damasio's (1994) characterisation of secondary emotions does enough justice to these studies. He only mentions the role of conscious images in triggering a bodily response, but he does not mention the influence that the latter may have on the kind of thoughts one entertains. In the tiered model he assumes, this would have to be explained with a feedback mechanism from the body to the limbic system and/or to the cortex. Even in this case, however, my hunch is that the "sequentiality" implied by the tiered model would not do justice to the rapidity and complexity of the ways in which the body affects our evaluations. My impression is that a dynamical approach extended over the whole organism (i.e. the brain and the body strongly coupled together) would more naturally account for such a role. Bodily states could be seen as part of the organismic dynamics that leads to certain evaluations and emotions; Othello's evaluation of the handkerchief, for example, could be seen as crucially influenced by his organismic state, upset by previous strong physiological reactions.

[add conclusion]

References

Damasio, A.R. (1994). Descartes’ error: emotion, reason and the human brain. New York: Putnam.

Descartes, R. (1644 [1964]). Les passions de l’'âme. Paris: Vrin.

Ekman, P. (1980). Biological and cultural contributions to body and facial movement in the expression of emotions. In A.O. Rorty (ed.). Explaining emotions.

Fodor, J.A. (1983). The modularity of mind. Cambridge MA: MIT Press.

Freeman, W.J. (1999). Consciousness, intentionality and causality. Journal of Consciousness Studies 6:143-172.

Griffiths, P. (1997). What emotions really are. Chicago: University of Chicago Press.

Hardcastle, V.G. (1999b). It's OK to be complicated: the case of emotion. Journal of Consciousness Studies 6:237-249.

Keijzer, F.A. (1998). Doing without representations which specify what to do. Philosophical Psychology 11, pp.269-302.

LeDoux, J.E. (1996). The emotional brain. New York: Simon and Schuster.

Lazarus, R.S., Coyne, J.C. and Folkman, S. (1984). Cognition, emotion and motivation: the doctoring of Humpty-Dumpty. In K.R. Scherer and P. Ekman (eds.). Approaches to emotion. Hillsdale, NJ: Erlbaum.

MacLean, P.D. (1970). The triune brain, emotion and scientific basis. In F.O. Schitt (ed.). The neurosciences: second study program. New York: Rockefeller University Press.

Quartz, S.R. and Sejnowski, T.J. (2002). Liars, lovers and heroes. What the new brain science reveals about how we become who we are. New York: Harper & Row.

Schachter, S. and Singer, J.E. (1962). Cognitive, social, and physiological determinants of emotional state. Psychological Review 69:379-399.

Schneidermann, N., Francis, J. Sampson, L.D. and Schwaber, J.S. (1974). CNS integration of learned cardiovascular behavior. In V. DiCara (ed.). Limbic and Autonomic Nervous System Research. New York: Plenum.

Zajonc, R.B. (1980). Feeling and thinking: preferences need no inferences. American Psychologist 35:151-175.