That there were forces to be considered other than judgements was evident to the Stoics as Posidonius’ appeal to Plato's horses makes clear. Already Zeno had identified distress and pleasure with contractions and expansions. But the main Stoic tendency had been to marginalize these forces as side effects. Chrysippus did so by making contraction and expansion into mere effects of emotion Seneca through making first movements which included physical symptoms into mere effects of appearance.
PART I: Emotions as Judgements versus Irrational Forces
10: What is Missing from the Judgemental Analysis? Brain Research and Limitations on Stoic Cognitive Therapy
The objections of Posidonius and others have shown that Chrysippus’ two judgements are not always necessary nor always sufficient for emotions to occur. What then is missing from the judgemental analysis? The debate has continued unresolved through most of the twentieth century with different values being attached to bodily reaction sensation desire or various forms of affect. It has also become clear what a large role is played by unconscious beliefs. But I believe that fresh clarity can be brought to the subject through research on the brain performed during the 1990s by Joseph LeDoux1 to whose work I shall come shortly.
Two suggestions have recently been pressed by philosophers as alternatives to the analysis in terms of judgement. One is the analysis of emotion in terms of desire which I mentioned briefly but disagreed with in Chapter 9.2 The other suggestion is that emotion involves affectivity or feeling and that this notion is not further reducible to anything else.3 I believe LeDoux's work shows how it can be reduced.
Why the brain's amygdala reactions need to be added to the role of judgement
Brain research we shall see was already producing relevant results in the nineteenth century. Even concerning the part of the brain called the amygdala on which LeDoux concentrated much was already known about its importance for emotion.4 But for our particular purposes the extra element added by LeDoux happens to be important because it gives the amygdala's reactions a new independence from judgements thoughts or ideas. The more obvious cases of emotion and the only cases mentioned in a widely noticed book published two years before LeDoux's are those in which the thoughts come first and the physical reactions of the amygdala and other brain systems follow as effects.5 Writing two years later LeDoux pointed out that the order of precedence can also be the other way round. The physical reactions so far from being mere effects of thoughts can actually precede the emotional thoughts in time. He has as yet confirmed this precedence only for the case of fear induced in rats by sound but he takes it the same is likely to be true for fear induced by sound in humans and that it may also be true for other signals associated with harm or benefit in the past experience of the individual or the race. LeDoux writes—but in this I cannot follow him—as if the more obvious order of precedence is not to be found at all (pp. 19 146).
What LeDoux discovered is that a sound associated with danger signals its message to the brain by two routes which diverge at a fork one route being fast one slow (pp. 157–65). The fast route travels to a part of the brain called the amygdala which immediately activates a large number of bodily systems. At this stage the person knows no more than that there is some danger (p. 163) and even that information may not be conscious (pp. 64 69). Almost any philosopher would have thought that one could not know something was dangerous unless one had at least some idea what it was. This turns out not to be so even in ordinary people. The effect can be clearer still in people with brain damage. In a split-brain patient who had lost much of the interconnection between the two hemispheres of his brain the left hemisphere was able to say of an object presented only to the right hemisphere whether it was good or bad but not what it was (p. 14). This apparently is the situation of all of us at the stage when signals have only reached the amygdala.
After the amygdala has been alerted it takes about twice the time for the message to reach the cortical parts of the brain. When the message arrives the person may then but also may not be able to identify what the supposedly dangerous thing is. If the thing is not dangerous the cortex may be able to send messages to quieten the amygdala. But these mechanisms may fail to work. First there may be no recognition. The sound of a car horn associated with a past car crash or near miss may have alerted the system. But one may not recognize that it is a sound that has done the alerting. After all one is seeing a lot of things as well as hearing so one may fail to recognize what has triggered the system (p. 203). The past trauma may even have caused the kind of physical damage that impairs memory of the past mishap and of the horn (pp. 239–45).
There are other cases too in which recognition fails to occur. People subject to panic attacks may have their panic triggered by some bodily sensation associated with past danger without being able to recognize any of these things as operating now (p. 258). In another case a brain-damaged patient with no ability to create new memories of people was unable to recognize her assailant. But the assailant's proffered handshake associated with the earlier assault alerted the amygdala system so that the handshake was refused (pp. 180 222).
But suppose recognition does occur. The shell-shocked person may realize that it is only a door slamming not gunfire that he hears. But his past experience at the battle front may have been so traumatic as to damage the systems used for quietening the amygdala (p. 256). In phobias the amygdala's response to snakes or heights may have been innately wired in for evolutionary reasons and so be all the harder to overcome (p. 255). The Stoics had already tried to explain without ascribing rationality to chickens how newborn ones can recognize the bull as safe but the weasel or hawk as dangerous.6 Avicenna had distinguished a special estimative faculty to explain the sheep's knowledge that the wolf is dangerous.7 In a way the hard-wiring of the amygdala provides his faculty.
The reactions of the amygdala system provide the modern counterpart of Posidonius’ horses. We no longer believe like him in spatial movements of a physical soul but Posidonius’ instinct was right. Important as judgements or appearances are to many instances of emotion there is another factor to be considered: the physical reactions of the amygdala or other comparable brain systems and our awareness of those reactions.
We can now see why there were so many examples available to Posidonius in which judgement is either not necessary or not sufficient for emotion. Disowning the judgement that a height is dangerous does not automatically calm the amygdala. When music arouses an emotional response which we cannot attach to a particular object what may be at work is either the amygdala or some part of the brain that plays a comparable role in emotions other than fear. Some animals non-mammals having little or no cortex may react standardly in this sort of way rather than through cortical pathways. Conversely our judgements when we foresee a danger by intellectual means may fail to arouse the amygdala if we are exhausted or for other reasons inattentive or unimaginative. No wonder Posidonius was able to find holes in the purely judgemental analysis.
What then is missing from the judgemental analysis of emotion? LeDoux lists a huge range of reactions set up by the amygdala and another range set up by the cortical response (pp. 109–10 149 158–9 207 284–5 287–8 290–1). Moreover there is sometimes in addition a consciousness of some of these reactions. In fear the amygdala may cause the animal to freeze motionless may alter heart rate and blood pressure reduce sensitivity to pain release stress hormones activate reflexes such as blinking or widening of the eyes. The animal may defecate urinate or have its hair stand on end—humans get goose pimples. Other reactions include sweating fleeing fighting and changing facial expressions. The amygdala may also help activate a number of arousal systems which keep various organs alert for anything important sometimes for a matter of days without supplying any information as to whether the importance takes the form of danger. Something which has all these reactions according to LeDoux will be in a state of fear and this is surely plausible for some cases. To feel afraid as well as being afraid LeDoux suggests is to be conscious of the activity of the brain's emotion system (p. 268).8
This last remark on consciousness may help to explain some of the phenomena I have called ‘feeling-as-if’ the situation were bad without judging it bad. This might in some cases of fear be no more than consciousness of the reactions set up by the amygdala reactions which would be appropriate to imminent harm. The idea of imminent harm need not have entered into consciousness in the case of a lower animal or of a human moved by sad music. On the other hand a shell-shocked person who knows he has only heard a door slam may well have entertained the idea of harm before dismissing it. Only in the last case will the idea of harm be doing much work.
If we put LeDoux's findings together with the earlier discussion of judgement we can see that in many cases fear in the human will involve both systems both certain judgements and the automatic reactions of the amygdala system. But in many cases the judgements come a fraction later and in some cases there is no judgement that is actually conscious not even the rudimentary judgement of some danger.
We can also understand more about the phenomenon of first movements. Tears trembling and sinking feelings in the chest may well be produced by an automatic quick-route system before anyone has had time to register what exactly the situation is. LeDoux seems prepared to concede that the person may be described as already thinking there is danger but that thought he says may as yet be entirely unconscious. I would prefer to express the situation by saying that talk of thinking does not really add to the description of the physical fear reactions since the idea of danger is not yet doing the work. The first movements then can arise even earlier than Seneca supposed so that there is all the more case for his claim that they do not yet constitute emotion.
It matters that LeDoux maintains that what the amygdala system produces without any help from the cognitive systems of the cortex is already to be counted as emotion. His is a rather unusual position. He appeals to a principle of economy Occam's razor to avoid postulating cognitive states unnecessarily and cites Pavlov as a model (p. 143 p. 146 with n. 15). But Pavlov not only denied thoughts to the dogs who salivated on hearing the dinner bell he also refused to describe their response as an emotion. It was only salivation whereas LeDoux does ascribe emotion to his subjects. I think he is right to do so in many cases though not for the reasons he gives.
LeDoux sees both the amygdala system and the cognitive systems of the cortex as principally constituting a survival mechanism (p. 128). The great advantage of the amygdala system is that it is fast so that the animal freezes even before it can tell what the danger is although this may often be inappropriate. The advantage of the cortical system is that it enables us to act rather than merely to react and to respond more appropriately. But it carries the disadvantage of enabling us to foresee failure and so creates anxiety (p. 175). In LeDoux's view the core of emotion is not provided by conscious appraisal of the situation nor by consciousness of our reaction to it (pp. 297–9). These are dispensable. It might be protested that he concedes that the amygdala system involves at least an unconscious judgement that there is danger. But I have expressed doubt whether that judgement does any work in his explanation rather than being a mere redescription of the automatically operating mechanism. LeDoux thus reverses Seneca's verdict on whether the physical reactions are more important or the judgements. The only point of agreement is that neither attaches importance to awareness of the physical reactions (pp. 18 125 128).
I think LeDoux is right about some cases. We do sometimes think of the non-cognitive reactions as constituting and of the cognitive reactions as not constituting emotion. But we need to go case by case. Like LeDoux I would find it hard to insist that a merely intellectual foresight of danger constitutes fear. And it is hard to deny that a person with shell-shock or vertigo who disowns the idea of danger may be afraid and that a person moved by music for unidentified reasons may be feeling emotion even though they may not have an identifiable object to feel emotion about.9 But LeDoux's account fits less well I think with people who have and feel the physical reactions characteristic of anxiety but who have become accustomed to the idea that there is nothing to be anxious about and behave perfectly calmly. We may well think of them as having only feelings of anxiety not as being anxious.
LeDoux claims to throw light also on the tradition which makes emotion depend on bodily feedback (pp. 45–50 291–6) most famously exemplified by William James's view that we do not cry because we are sad but are sad because we cry.10 The biggest problem for James is how to explain our tears or trembling occurring in the first place unless we are already sad or afraid. LeDoux offers an answer although it seems to concern only feeling afraid as opposed to being afraid (pp. 268 295). We will not feel afraid until there has been enough feedback from the unconscious amygdala circuits for us to be aware of our bodily reactions. But this defence of James is limited because on LeDoux's view we will by this stage already be afraid.
Other brain research reducing the role of cognition
LeDoux's is not the first brain research to reduce the role of cognition in emotion. The nineteenth-century neurophysiologist David Ferrier studying animals with their cerebral hemispheres including the amygdala removed commented:
They start at sounds flinch at light… direct their movements in relation with visual impressions; respond with movements expressive of disgust or discomfort… The centres of emotional expression are therefore situated below the centres of conscious activity and ideation.11
Another set of findings concerns sudden emotional experience lasting say 20 to 40 seconds at the onset of epileptic fits.12 The feelings are typically recognized by patients as baseless and when they take the form of fear the fear is said not to be about the impending attack. So the feelings do not consist in the kind of cognition Chrysippus postulates. There are often accompanying sensations which in line with Galen's observations about bites are felt in the epigastrium or pit of the stomach. And it is interesting that the emotional experiences have been described as ‘ictal’ from the word Seneca used for first movements Latin ictus ‘shock or blow’.13 The very same emotional experiences were also aroused so the patients said during surgery to remove malfunctioning portions of the brain. The range of emotional feeling reported is enormous and the patients describe it in terms of cognition but of cognition they recognize as baseless. There are feelings of depression sorrow sadness intense fear loneliness a feeling of being alone a homesick feeling shame or guilt anxiety wanting someone near or wanting to be alone paranoid feelings especially of there being someone behind one a sense of futility of unreality a depersonalization a sense of being far off outside the body and watching oneself a terrified feeling as if one were in another world a sense of disgust a moroseness. There are also triumphal feelings described as feelings of completeness contentment elation exhilaration fascination gladness satisfaction security eternal harmony immense joy intense happiness paradisiacal happiness well-being of all the senses like whisky taking effect a relaxed feeling a wonderful feeling. And there are feelings as of the revelation of truth of clarity or of conviction.
Implications of brain research for different therapies
All this brain research has implications for therapy. LeDoux suggests that the role of bodily feedback may help to explain the kind of therapy I shall associate in Chapter 17 with the figure of Socrates and which has recently been studied by Paul Ekman in which putting on a smile makes you more relaxed (p. 295).
This bodily feedback therapy is of course only one. We can see from LeDoux's research how there could be many therapies. An obvious hope is that drugs might be developed for controlling the amygdala's reactions. But we can see certain other therapies including Stoic cognitive therapy as helping the cortex to gain control of She amygdala or of other automatic systems (pp. 248–50 261). The Stoic cognitive method is for people to change their judgements by saying things to themselves. We can see both why cognitive therapies can often work and why they sometimes cannot. It is no use the shell-shocked person telling himself that he has only heard a door slamming because the amygdala inhibitors have been damaged by earlier trauma.
Different therapies may try to gain control of the amygdala by different means. In extinction therapy someone who fears crowds may be brought through conditioning to see crowds as safe first by imagining going out of doors and then by actually going but at first only as far as the garden gate. He does not like the Stoic have to say things to himself but only to experience the safety of these exercises in reassuring surroundings. LeDoux suggests that a distinct part of the cortex in the medial prefrontal region may be being used here to control the amygdala (pp. 248–50 263–5).
Limitations on Stoic cognitive therapy
The workings of the brain and the impossibility in some cases of controlling the amygdala system by cognitive methods provide only one set of limitations on Stoic cognitive therapy. It has other limitations of its own. I mentioned in Chapter 1 that it is not intended to cure what we should call pathological emotions but it can be very effective with the ordinary ups and downs of life. I should now draw attention to three further limitations.
First Chrysippus is not well placed unlike Posidonius and Galen for dealing with moods that are not directed to a particular situation. Moods like depression for example persist whether directed to a situation or not so cannot easily be treated by Chrysippus whose therapy like his analysis of emotion focuses entirely on the judgements people make about their situations. Looking at a situation differently will not cure depression if depression is independent of any particular situations. By contrast moods are treated not only by Posidonius and Galen but also we shall see in Chapter 23 by the Christian Evagrius who paid particular attention to a form of depression called akēdia.
Secondly Chrysippus has comparatively little to say about children and how to mould their emotions. In this he may be like Socrates: both depend on rational interrogations of a kind more suited to adults. In Chrysippus’ view children do not acquire reason until the age of 14. It is the complaint of Posidonius that Chrysippus has nothing to say about how to train children because he thinks they are naturally virtuous unless corrupted by bad company. This too is due to neglect of their non-rational side.14 Plato Posidonius and Galen by contrast have a great deal to say on the training of children and so do others in Plato's tradition Antiochus and Plutarch. This is true of others too of Aristotle and the Epicureans.
Chrysippus’ comparative silence is only partly offset by later Stoics like Seneca who talk of the value of precepts (praecepta Greek paraineseis) as well as doctrines and theories (decreta Greek dogmata). Seneca reports a Stoic debate on the utility of precepts and believes that they are useful especially for those who are beginning to take control of themselves in youth a point on which he says Brutus agreed.15 Seneca's Stoic contemporary Musonius Rufus gave detailed advice on behaviour partly for the young for example on whether to obey your parents or get your hair cut.16 Such instructions were sometimes called in Greek hupothēkai. However an earlier Stoic Aristo contemptuously asserts that hupothēkai are suitable only for schoolteachers and nurses not for philosophers to give.17
A third limitation is that the Stoics concentrate their attention on understanding one's own emotions not the emotional effect of one person on another nor the emotions of others18 except in so far as these bear directly on how to control one's own. An example of this exception is that you may control your own anger by reflecting: ‘I too have before now committed the offence I am complaining of.’19 The point is not to understand that person but to see that for you to react would be inappropriate.
Modern cognitive therapy
A parallel question has arisen what limitations there may be on modern cognitive therapy. I should first say something of its relation to Stoic cognitive therapy. In one way it is very like as shown by the fact that a leading practitioner of cognitive therapy has described emotions as being judgements.20 Unwanted emotions are treated by changing judgements. In resorting sometimes to behaviour to change judgements modern cognitive therapy is no different from ancient. Seneca we shall see changed his judgements in favour of retaliation by looking in a mirror to see how ugly anger made him. Similarly in modern cognitive therapy behaviour was used to cure a patient of irrational fear of a heart attack which doctors had shown to be delusive. When he believed an attack was coming on he was asked to engage in the strenuous activity of star jumping. When this did not lead to a heart attack the belief in imminent danger abated.
But the differences should be noticed as well as the similarities. Modern cognitive therapy of course only offers to treat unwanted emotions. It does not follow what we shall see to be the Stoic ideal of eradicating all emotion. Again the judgements selected for treatment tend to be factual not evaluative (‘I am going to have a heart attack’ ‘Nobody is interested in what I say’ ‘I am fat’). But even Stoics like Seneca treat factual questions too (assume your health will last assume your son is enjoying life after death). And in each case the factual question is relevant because of its relation to an evaluative one (death being ignored fatness ill health and extinction are all bad). Sometimes too modern cognitive therapy does the same as Chrysippus in addressing the evaluative questions directly (‘Would it be insupportable if this did happen?’). The modern therapy makes one concession to Posidonius’ point that imagination as well as judgement is important. Patients may not be brought to recognize by evidence alone that a heart attack is not in the offing because they have such a vivid picture of themselves lying dead on the floor. It may be necessary to substitute a picture of themselves sitting comfortably in an armchair and merely imagining themselves collapsed on the floor. Early Stoics had sufficiently allowed a role for imagination in therapy as shown for example by the relabelling exercises to be discussed in Chapter 15 but not (this is Posidonius’ point) in the emotion itself.
Chrysippus’ concern with evaluative judgements is due to a further difference that he has a whole ethical theory according to which most evaluative judgements are mistaken because most things are indifferent. But this difference though huge does not impinge very much on therapeutic practice because as will be seen in Chapter 12 [Chrysippus is willing to exclude this theory from therapy.
The question of limitations to modern cognitive therapy has arisen in connection with the fact that its major successes have been with some conditions rather than others. It has been very successful indeed for example with phobias. But among eating disorders it has had far more success with bulimia in which for emotional reasons people eat too much than with anorexia in which they eat too little and in which mistaken judgements about their being too fat persist. An obvious conclusion would be that there are other factors at work here besides judgements and that other non-cognitive therapies will be needed. But one cognitive therapist has in oral discussion offered the following alternative explanation.21 The judgements associated with anorexia are far more numerous than those associated with phobias and it takes time to discover which judgements are the driving ones doing the damage in any given case. Further in some cases it may be hard for a person to identify the judgements at all: they may be unconscious. In time however all emotions will turn out to consist of judgements and all will be amenable to cognitive therapy.
Chapter 17 will be devoted to a non-Chrysippan tradition in ancient philosophy which gave a central role to the body in emotion and recognized the need for non-cognitive therapies. But before that I want to consider how Chrysippan cognitive therapy would work in detail.
From the book: