2. The Phenomena of Persuasive Emotion

The persuasive use of emotion was common in the world Aristotle was familiar with, and examples might be cited from that era.Footnote ⁷ Likewise, one might note how a number of landmark pieces of persuasive speech through global history, that we think of as clearly legitimate, are clearly emotive in how they bring about their persuasive effect.Footnote ⁸ But I will here use as my central example Sojourner Truth's ‘Ain't I a Woman?’ speech at the Women's Convention in Akron, OH, in 1851, aimed at persuading those present and the wider world to support women's suffrage.

Clearly it persuades, in large measure, through its use of emotion. It invites admiration for the contribution to society of women, indignation at their unjust treatment, and a kind of ridicule of the absurdity of the opposition. Is this problematic? Should people be persuaded to be committed to women's suffrage in a more austere way that dispenses with admiration, indignation, and ridicule? Intuitively: no. If there is any prospect for vindicating as legitimate the persuasive use of emotion, it is in cases like this. Cases like this invite us to endorse both ANTI-AUSTERITY and PROPRIETY.

When Aristotle incorporates the persuasive use of emotion within his account of the sort of rhetorical skill worth valuing, it is perhaps cases of this general kind that he has in view. Insofar as we share this kind of ‘anti-austerity’ intuition, it will be because we too think that Sojourner Truth, like Lysias and Demosthenes before her, is not to be faulted for including the engagement of audience emotions within her persuasive arsenal.

This is reflected in Aristotle's more generalised analytic remarks, and in the advice he offers to orators on the persuasive use of emotion-arousal in speeches.

Aristotle is here specifying the types of ‘proof’ (his word is pistis) that form part of the repertoire of the expert orator. In calling them types of ‘proof’ he is indicating that these are things that constitute proper grounds on which for listeners to be convinced of the orator's case.Footnote ¹⁰ The arousal of their emotions, at least in the ways Aristotle will be recommending, is just such a thing.Footnote ¹¹ Accordingly, he recommends that orators arouse audience emotions in certain ways, in pursuit of their persuasive goals. Of course, it is part of this rationale for the use of emotion that it works! Emotion-arousal is often an effective means of securing persuasive goals.

But not only does Aristotle recognise emotion's persuasive efficacy, he recognises that the precise contours of each emotion that the orator stirs up in their audience make a difference to the nature of its persuasive effects on the listeners.

Feelings incline listeners in particular directions. Hostility towards someone inclines listeners towards judging them guilty, sympathy inclines listeners away from such a verdict: it is this way round and not vice versa. Feelings of eagerness and hopefulness encourage judgements to the effect that some hoped-for prospect is more likely actually to happen, and that it will be good when it happens: not judgements with the opposite content. Those are made more likely by gloomy feelings, or at least the absence of eagerness and hope. Aristotle directs the orator's attention to these very obvious features of how emotion affects judgement: it is the orator's business to develop a systematic understanding of what distinctive kinds of effects particular types of emotions have. And he presents this as part of the understanding that will underpin their skilful deployment of emotional ‘proofs’. This observation about the way the particular contours of a given emotion serve to produce specific persuasive effects is important, in that it highlights that the persuasive role of emotion goes beyond simply making listeners generally more attentive to the speech – what one might call a ‘coffee and cigarettes’ view of the persuasive role of emotions.Footnote ¹² That is to say: emotions are not simply perking listeners up to be more attentive generally, or making them more prone to be decisive and quicker to form a verdict. They incline audiences in particular directions depending on the specific features of the emotion used.

This general position is reflected also in the specific advice Aristotle offers about the use of the various types of emotions, when he treats them one by one in Rhetoric 2.2-11.Footnote ¹³ He reminds the reader at the end of these accounts that these instructions on arousing and dissolving the various passions count as instructions on how to give ‘proofs’ (pisteis, 1388b29-30). In other words, his commitment to the use of emotion-arousal in skilful oratory (i.e. to ANTI-AUSTERITY), is held concurrently with his commitment to the kind of rhetoric that works by providing listeners with proper grounds for conviction (i.e. to PROPRIETY).

When we consider Aristotle's position alongside examples such as the Sojourner Truth speech, we see the attractions of this view. There is intuitively a strong case for vindicating both of the claims, PROPRIETY and ANTI-AUSTERITY, insofar as it seems correct to say that audiences can be justified in their inclination to adopt the speaker's proposed conclusions, on the basis of persuasive devices that include the arousal of their emotions.

3. The Cognitive Mechanisms of Emotional Persuasion

We may, however go further than these general observations, and see that the way emotion-arousal achieves its persuasive effects seems to be by contributing to or undermining the inferences that listeners appropriately make in arriving at their judgements or decisions. The role looks cognitive. This is most easily seen in the instructions Aristotle offers to orators on how to use emotion persuasively.

Aristotle's definition of the passions in Rhetoric 2, and his general instructions about how to use an understanding of the passions in rhetoric, are as follows:

The passions, as well as involving pleasure and distress, are things that account for changes in judgement, and perhaps are themselves changes in judgement (various interpretations are possible here). They are aroused by deploying an understanding of what kind of (say) distress and/or desire is involved in each passion, what kinds of objects elicit that passion, and what grounds for the passion need to be apprehended by the listener if that passion is to be produced in them. The most natural interpretation is that the orator is to present the person in question as a suitable object of (say) anger, as having features (or having done things) that are grounds for feeling angry, and in such a way that it is distress (not pleasure) that is aroused at these things, and a desire for revenge. The process of doing so is to mirror the process of constructing explicitly-presented arguments: i.e. a process of methodically selecting premises based on the orator's systematic knowledge of what premises support what conclusions. Indeed, one might plausibly read the final sentence of the passage above as indicating that the orator's method of emotion-arousal will literally involve ‘premises’ for the arousal of emotions, and that what we are to be offered in the ensuing chapters on the individual emotion types constitutes the construction of a systematic bank of knowledge regarding which premises will support the arousal of which emotions.Footnote ¹⁴ The premises arouse the emotion, and the emotion in turn gets the audience to draw a particular ‘verdict’. The emotion is an interim conclusion which then functions as a premise of a further inference by which the eventual ‘verdict’ is derived.

That Aristotle sees the role of emotion-arousal in persuasion in terms of how they contribute or block the availability of premises for inferences is suggested by his remarks on specific types of emotions. Good examples are the arousing and calming of anger, and the blocking of envy by the arousal of pity.

It is certainly clear from these passages that Aristotle envisages the arousal of emotions to occur at the hands of the skilled orator through the presentation of evidence that the person in question is an apt target of that particular emotion – the orator gets the audience to see the world in the ways typical of (say) an angry person, with their opponent as the object of their anger, and what they have done as grounds for their anger. The second and third passages equally indicate how the presentation of such emotion-relevant premises can block the arousal of a particular emotion. The third passage indicates that this can be expected to bear then on what the audience does next, i.e. on what benefits or punishments they confer by upholding one side or the other's case, in their verdict and sentence.

Emotion-arousal serves a persuasive function because of the role that emotions play in inferences, that is in arguments – broadly-construed – presented by the speaker. This role consists in their contribution of premises or their serving to block the use of premises by which a further conclusion is supported.

The kind of way in which Aristotle's view might apply to the Sojourner Truth case could be spelled out as follows. She advances an argument which, of course, could be followed in an emotionless way by the audience. But it can also work emotionally (perhaps simultaneously),

1. 1. Women have muscle, and plow, reap, husk, chop, mow, carry and eat on a par with men.

2. 2. Women contribute impressively to society (from 1.). [ADMIRATION].

3. 3. It is unjust for such admirable contributors (cf. 2) to be treated in such a substantially inferior way (not even getting their ‘pint’ to a man's ‘quart’). [INDIGNATION].

4. 4. Opposition to women's suffrage is based on a fear that women will take more than their ‘little pint’ can hold, on the part of men ‘all in confusion’. [RIDICULE].

5. 5. If (or perhaps, insofar as) it is unjust and absurd to oppose women's suffrage, we should not oppose it.

6. 6. It is unjust (3.) and absurd (4.) to oppose women's suffrage.

7. 7. THEREFORE: we should not oppose women's suffrage.

It is obviously an empirical matter whether this is indeed what goes on within the psychological mechanisms of an audience on any particular occasion, and whether in general this tends to be the mechanism by which seemingly legitimate emotional persuasive methods such as those to which I am calling attention achieve their effects. This cannot be established from the armchair. Nevertheless, we can see that this is a perfectly intelligible way in which emotional persuasion could work; it seems to be how Aristotle saw emotional proofs working in his Rhetoric; and it seems to offer an intelligible explanation of cases such as that of Sojourner Truth. In what follows, I undertake to explore the implications of this understanding of how emotional persuasion can work. In particular, what must be true of the emotions and the justification of emotions, if this view is correct?

4. Some Conditions for Legitimate Use of Emotions to Support Inferences

Do emotions have the right kinds of features to play the role assigned to them in this view? The kind of role in question seems simply to mirror the uncontroversial role of beliefs in persuasion. Clearly one kind of legitimate persuasion involves presenting someone with reasons to believe something, getting them thereby to form justified beliefs, and then using those justified beliefs (plus other knowledge or justified beliefs of theirs perhaps) to get them to accept some further conclusion on the basis that it is entailed or recommended by those things they have come to believe. This is a familiar process, and involves what we might call ‘belief-arousal’. A variant on it might be where persuasion involves literally ‘showing’ something, such that perception supplies one or more of the premises involved in the inference. In both cases, the adoption of the conclusion inherits the epistemic merits of the way in which the premises are apprehended. Can ‘emotion-arousal’ play the same uncontroversially legitimate role as is played in each of these cases by directing perceptual attention (‘showing’) and by giving reasons to believe something? Or is there something about the nature of emotions, and how they differ from beliefs or perceptual states that rules this out?

I propose that in order for emotion-arousal to contribute to inference-based persuasion in this kind of way, at least the following conditions must be met.

1. 1. The inference must be performed, such that the conclusions are believed because of the truth (or likelihood) of the premises.Footnote ¹⁶

2. 2. The premises must contribute justification to the conclusion.

3. 3. If the premises are not perceived, they must be asserted.

The first of these is straightforward: to the extent that the inference is to contribute to the justification that the subject has for believing its conclusion, they must to that extent believe the conclusion on the basis of the truth (or likelihood) of the premises. In order for this to be the case, they must actually perform the inference and believe the conclusion on the basis of having done so.

The second condition is that insofar as the subject's justification for believing the conclusion is derived from the inference, it must be derived from the epistemic good standing of the premises. Take the case from the Hebrew Bible of the prophet Jonah's (emotional) persuasion of the people of Nineveh to change their ways. The inference involved is this:

1. 1. If there is a terrible destruction coming, then we must do all we can to avoid it.

2. 2. There is a terrible destruction coming.

THEREFORE:

1. 3. We must do all we can to avoid it.

Perhaps premise 1 might be known to Jonah's audience in some very robust sense. Their justification for believing the conclusion on the basis of this inference will most significantly depend on the epistemic justification contributed by their grasp of premise 2. The audience may believe premise 2, but the weaker their justification for holding it, the weaker their justification for adopting the conclusion on the basis of the inference. Or persuasion might proceed by showing someone that a key premise is true – i.e. enabling them literally to perceive the content of the premise, and establish perceptually for themselves that it is true. Perhaps Jonah might, in a variant on the biblical narrative, have shown the people of Nineveh the divinely-appointed armies massing over the brow of nearby hills to destroy them. But either way, their epistemic justification in believing the conclusion (insofar as they believe it on the basis of this inference) is contributed by the premises. And this may be either through the subject's being justified in affirming the premises, or through their perception of the premises being a source of justification in the way that a perceptual experience in favourable conditions is. On the latter suggestion, the idea is not that perceptions are themselves justified, but that they are evidentially valuable, and can confer justification on beliefs based upon them. If subjects are to be justified in believing the conclusion of an inference, that justification must be conferred, in one way or another, by the premises.

The third condition simply highlights that one cannot be justified in believing the conclusion of an inference whose premises one merely entertains in thought, without asserting them. In relation to this condition, we set aside the possibility that the premises might perceived in an evidentially valuable way. That possibility apart, it seems as though the subject can only be justified in believing the conclusion if they actually assert the premises.

In the remaining sections, let us consider how these premises apply to emotional persuasion, and what light they shed on how we would need to understand the emotions for them to play the kind of legitimate role in persuasion that we have been tracing out.

5. ‘Because of the Truth / Likelihood of The Premises’

Let us take the first of these conditions first. Insofar as the subject's justification for believing the conclusion is derived from the premises, the conclusion must be adopted because of the (supposed) truth or likelihood of the premises and because of the merits of the inference. Nevertheless, this needs to be understood in a way that avoids the pitfalls of over-intellectualising the process. It should not be a condition of one's being justified in adopting the conclusion of a valid inference based on justified belief in its premises that one have reflected explicitly upon (still less have a justified explicit belief in) the validity of the inference used.Footnote ¹⁷ One does not need beliefs explicitly about good inferences to be competent in using them. This is important in the emotional case. For it might have seemed significant that dispassionate persuasion has a higher likelihood of involving explicit reflection on the validity of the inferential connection between premises and conclusion than does emotional persuasion. But if it is clear that this is not a condition of the transmission of justification even in the dispassionate case, it would be a mistake to impugn the emotional case because it is less likely to be fulfilled. What is required in both the emotional and the dispassionate cases is the same: that the conclusion be accepted because of the supposed truth or likelihood of the premises, through the competent performance of a good inference.

6. That the Premises Contribute Justification

We recall that if a successful case of persuasion is legitimate, this will be because the audience comes to be justified in believing the speaker's proposed conclusion (‘PROPRIETY’). On the model of emotional persuasion set out in section 3 above, the emotions aroused contribute premises to an inference the speaker intends the audience to perform, and thereby to believe their proposed conclusion. The relevant kind of justification for their believing the conclusion is epistemic. Hence, the kinds of justification of the premises that the emotions used must confer is epistemic justification.Footnote ¹⁸ So, they must either be sources of epistemic justification, or themselves possess epistemic justification.

The idea that emotions can themselves be justified epistemically will be unproblematic on a number of views of emotions. If emotions can be had for reasons (for example, because of evidence that their object is dangerous) and extinguished by reasons (for example, by evidence that their object is not dangerous after all), and if they can be assessed for how well justified they are in the light of the reasons available, then they would seem to be capable of possessing precisely the kind of epistemic justification that in legitimate emotional persuasion gets conferred onto the adoption of the conclusion. This can be readily accommodated on lots of views of emotions, not just judgement-based views such as Solomon's or Nussbaum's, but also on those views, such as Goldie's view of emotions as sui generis states, that accommodate the recognition of emotions as reasons-responsive and reasons-assessable.Footnote ¹⁹

Conversely, it seems equally clear that on so-called ‘thought-theories’ of emotions,Footnote ²⁰ insofar as emotions may be based on unasserted thoughts, or partially constituted by such thoughts, it is very difficult to make sense of how one could be justified in believing the conclusion of an inference in which one or more premises was contributed by such emotions. For we would not be epistemically justified in believing the conclusion of an inference whose premises were merely entertained in thought, unasserted. Such premises could not provide justification for believing the conclusion.

Of course, it is characteristic of such views that while holding that emotions involve cognitions, they are pluralist about what kinds of cognitions can be involved in different cases: thus, some emotions are based on beliefs, others on thoughts, others on vivid imaginings, and so on. A central motivation for such views is the need to accommodate cases of emotion where it seems that the subject lacks beliefs or judgements with the requisite content. These include cases of recalcitrant emotion, such as fearing a dog one knows to be harmless, and cases of emotional response to what one knows to be fictional, such as pitying Anna Karenina. Since emotions, it is claimed, can be based on thoughts or imaginings, there is no difficulty about the possibility of such cases occurring, nor the rationality of having such feelings. Nevertheless, it is important for our purposes to notice that thought theories are not supposed to exclude the possibility of emotions’ being based on beliefs or judgements. Emotions, on this view, can be based on a variety of different kinds of cognition.Footnote ²¹ However, only some kinds of cognition can be sources of epistemic justification. On thought-theories of emotion, the use of emotion in persuasion will be legitimate only where the emotion in question can contribute epistemic justification to the adoption of the speaker's conclusion. Emotions involving unasserted thoughts and imaginings will not fit the bill. It seems as though it will only be emotions that are themselves candidates for epistemic justification, such as those involving beliefs or judgements, that can play this role, and possibly also those based on perceptions if the problems already highlighted for perceptual views of emotions can be overcome. None of this amounts to any objection to thought theories: the claim here is that they will accommodate the legitimate persuasive use of emotion by using the explanatory resources that they share in common with judgementalist and sui-generis views, or with perceptual views.

It also seems clear that the emotions involved in legitimate emotional persuasion cannot be (or cannot depend for their representational aspects on) what Tamar Gendler has dubbed ‘aliefs’. It is a defining characteristic of aliefs that they are arational, that is to say that they cannot be assessed for how rational or otherwise it is to have them.Footnote ²² This seems immediately to exclude the possibility of their being epistemically justified. Again, it is not obvious that this constitutes an objection to any of Gendler's claims about aliefs: although is it part of the characterisation of aliefs that they are affective, there does not seem to be any suggestion that all emotions involve (or are constituted in part or whole by) aliefs. For our purposes, the key observation is that, as characterised by Gendler, aliefs cannot account for the representational element of emotions involved in the kind of legitimate emotional persuasion we have been exploring.

A further important category of views of emotions holds that they are perceptions, or are closely analogous to perceptions. Since perceptual states are not had for reasons, and are not reasons-assessable, views of the emotions as perceptual states will not be in a position to recognise the possibility of emotions’ contributing to the justification of the conclusion by being themselves epistemically justified. (This, of course, is already a disadvantage of such views.) Perceptual views of emotion need instead to insist that in any legitimate cases of emotional persuasion, the emotions involved contribute to justifying the adoption of the conclusion by being themselves sources of epistemic justification, just as perceptual states might plausibly be understood as sources of justification for beliefs that things are the way perception represents them as being, i.e. they are evidence that things are thus and so.Footnote ²³

In principle, this might be fine. In practice, there are difficulties. There are widely-canvassed doubts about whether emotions can be perceptual states in this way, in the first place.Footnote ²⁴ But beyond these general doubts, Michael Brady has drawn attention to problems in the way the perceptual model conceives of the epistemic role of emotions.Footnote ²⁵ I shall reprise one of his arguments and add one of my own. Brady points out that the perceptual model implies that emotions are evidentially valuable in cases where this is not plausible. In the perceptual case, suppose a reliable colleague tells me that it's raining and everything's getting soaked outside, and then I step outside and perceive that it's raining and everything's getting soaked. Although, in important respects, my perceptual experience tells me some things that I already knew, it strengthens my evidential position – I have stronger epistemic justification than I had before for my conviction that it is raining. The perceptual theory suggests it should be like this in the emotional case. But it isn't. Brady's example is that the sharp horns and rapid advance of the bull in the same field give me good reason to believe it is dangerous. On the perceptual view of emotions, fear is a perceptual experience of its object's dangerousness, and has evidential value accordingly. So, my fear itself should give me extra evidence, perceptual evidence, above and beyond what I have in the bull's sharp horns and rapid advance, for believing that the bull is dangerous. But this is implausible.Footnote ²⁶ In the Sojourner Truth case, if we ask what reasons we would have for believing that women are being treated unjustly, the perceptual theory says implausibly, that in addition to the facts of women's contributing equally and being treated worse, the listeners’ indignation at this injustice itself counts as extra evidence that women are being treated unjustly. I concur with Brady that consideration of such cases suggests that emotions do not serve as sources of epistemic value, i.e. as evidence, in the way that perceptions do.

A further difficulty relates more specifically to cases of emotional persuasion. The calibre of justification that the audience has, as a result of emotional persuasive arguments, varies from case to case. And one way in which it varies is that the level of epistemic justification conferred on the conclusion by the emotional premises varies. The most natural way to explain this is by the variation in how justified the emotions in question themselves are. But the perceptual model does not have recourse to this explanation, since it says that emotions are perceptions, and hence have epistemic value as sources of justification, not by themselves being epistemically justified. Perceptual theories can, of course, rate some perceptions as evidentially more valuable than others – what I see through the fog, or at night time, or with my glasses off, has less evidential value than what I see with my glasses on, in broad daylight, on a clear day. Some conditions are better for perception than others. Could there be some account of better conditions for emotional perceptions that could explain why some emotional arguments provide better epistemic justification, and hence persuade more legitimately, than others?

Consider again the indignation premise (3) in Sojourner Truth's argument. What would be favourable conditions for a more accurate emotional ‘perception’ of the injustice done to women? In the perceptual case, favourable conditions are conditions that favour the accuracy of the perceptual representation. In the emotional case, it is not literally the case that (for example) Sojourner Truth's listeners are in perceptual contact with the injustice itself. The ‘accuracy’ of their feelings of indignation at this injustice, surely depends on the quality of the evidence and reasons on which it rests – i.e. on the evidential support for women's contributing equally and being treated worse. But evidence and reasons of this kind are not merely circumstantial factors that create favourable conditions for some distinct perceptual event to occur, i.e. for one's emotional ‘eye’ to gaze directly on the injustice to which Sojourner Truth refers, such that the resulting emotional experience then has high evidential value. No: the indignation she arouses has epistemic merit to the extent that it is formed on the basis of that evidence and those reasons. But if that is what the level of justification contributed by emotions depends on, then the epistemic contribution of the audience's indignation to the persuasive argument stems not from their having the evidential character of perceptual experiences, as sources of epistemic justification, but from those emotions’ being themselves epistemically justified by other reasons and evidence, of the kind that Sojourner Truth partly supplies and partly can take for granted. In other words, accounting for the variation in how much justification is contributed by emotions drives us away from the perceptual model and back towards views on which emotions are themselves candidates for justification. I cannot see a plausible alternative to this available within the perceptual model.Footnote ²⁷

This is already a significant result. There must be at least some emotions that can be epistemically justified in something like the way in which beliefs can be. And it can be important in at least some contexts – such as where emotions are being used persuasively – to assess emotions for their epistemic justification.Footnote ²⁸ Of course, this does not show that emotions must literally involve beliefs or judgements, but it does seem to underscore the credentials of those theories of emotion that recognise belief-like elements in at least some emotions.

We are seeking to understand how emotional persuasion such as that used by Sojourner Truth could be legitimate, where this is understood in the kind of way suggested by ‘PROPRIETY’ above. Following Aristotle's plausible proposal, we are thinking of emotional persuasion as using inferences, in which the audience's emotions play the same role as their beliefs might have done, i.e. they are the way in which one or more premises of the inference are supplied. We highlighted three key conditions for this to result in the audience's being justified in believing the conclusion. The first was that the audience perform the inference – this presents no special obstacle for emotional persuasion. The second was that the premises contribute justification to the conclusion: and it turns out that this is difficult to accommodate on some theories of emotion, and requires (in our central cases, at least) that emotions exhibit the rather belief-like feature of being candidates for epistemic justification. The third condition takes us in a similar direction.

7. That the premises be asserted

The third condition is simple, and applies in cases where the premises are not supplied perceptually. It is this: one cannot be epistemically justified in believing a conclusion on the basis of an inference that depends on premises that one merely entertains in thought without (mentally) asserting them, i.e. without being committed to their being true, or at least likely. This is simple, but very significant for emotional persuasion. For if emotions are the mental state in the listener providing premises for an inference that they perform, and whose conclusion they are being supposed to be justified in believing (on the basis of the inference), those emotions had better involve the subject's commitment to the truth (or likelihood) of the content that constitutes the premise. In Sojourner Truth's argument, if the audience is supposed to be justified in believing on the basis of this argument that it would be unjust to oppose women's suffrage (6), and that they should not oppose it (7), they had better not leave crucial premises (such as 3) unasserted and merely entertained in thought.

However, on some views of what the emotions are, they might involve no commitment to the truth of the way they represent things as being, i.e. they involve nothing like assertion. This is true of views in which emotions involve or are based on what Tamar Gendler has called ‘aliefs’. On ‘thought theories’ of emotion, which usually are pluralist about what kind of cognitive attitude is involved in emotion, this condition will require that any emotions that contribute to persuasion in the way described must involve a cognitive attitude of commitment to the truth (or likelihood) of its representational contents. The emotions must be, to that extent, belief-like.

It doesn't follow from this that emotions literally involve beliefs. Indeed, it has often been supposed that their representational aspect cannot literally consist in beliefs, because of the kinds of ways in which beliefs and emotions can conflict. This is maintained partly on the basis that emotions seem to involve the exercise of a psychological capacity that has some independence from the subject's beliefs and belief-forming processes. Emotions and beliefs can conflict; ordinary processes for belief formation and belief revision can be impotent to form or revise emotions; and emotions seem to be able to be aroused and dispelled in ways (e.g. by music and images) that leave beliefs untouched. This suggests that they are distinct from those beliefs. A second basis for rejecting the involvement of beliefs in emotions has to do with the degree of similarity between the cognitive attitudes involved in emotions and those involved in believing. For example, Sabine Döring points out that when one feels afraid of something one knows not to merit fear, whatever irrationality may attach to the subject, it does not reach the levels involved in, say, Moore's paradox.Footnote ²⁹ A thorough discussion of this issue cannot be pursued here. Nevertheless one of the goals of this paper is to draw attention to the fact that the legitimate persuasive use of emotion has important bearing on this issue of how exactly we should characterise the cognitions involved in the emotions. If the legitimate persuasive use of emotion is to be vindicated, as I suggest it should be, and if it is to be understood as operating through the emotions’ supplying premises to arguments that confer epistemic justification on the adoption of their conclusion, then this seems to suggest that emotions have two belief-like features. These are that they are themselves candidates for epistemic justification, and that they involve commitment to the truth (or likelihood) their represented content. This does not amount to the claim that emotions literally are, in whole or part, beliefs; but it is no less significant for that.

These features, arguably, can be accommodated not only by judgementalist views of emotion, but by sui generis views such as Goldie's, by pluralist views of the cognitive attitudes involved in emotions, such as are held by most ‘thought theorists’, and by others besides. But they cannot be accommodated by all views of the emotions, and if the above account of emotional persuasion is correct, this counts against those views.

8. Brady and Indirect Views of Emotional Persuasion

Before resting content with such a conclusion, we should consider its compatibility with a recent sophisticated account of the epistemic value of emotions, put forward by Michael Brady's views in his recent monograph, Emotional Insight. Footnote ³⁰ The main focus of Brady's position as discussed there is his (plausible) claim that emotions motivate the search for, and evaluation of, reasons that bear on the accuracy of those emotions themselves. As such, Brady is arguing that emotions facilitate the development of understanding of (among other things) our evaluative situation. Let us distinguish two parts of this picture. The first is the motivation to investigate further – i.e. the way emotions get us fixated on their objects with a view (Brady argues) to improving our understanding of them. The second is the appraisal involved in the emotion itself. This receives less attention in Brady's discussion, but it is clearly part of his view – emotions involve an appraisal and additionally (and importantly for his agenda) motivate the subject to understand the objects of that appraisal better. Let us treat these in turn.

The appraisal: since Brady rejects the perceptual model, he may be sympathetic to a model of emotional appraisals that includes the possibility that in at least some cases, these involve commitment to the truth (or likelihood) of the appraisal. This fits with the idea that emotions seem to motivate and rationalise actions. And if so, then it seems as though on this basis, Brady can readily accept that in the kinds of cases of emotion featuring in legitimate persuasion that we are discussing, the emotional appraisal involves commitment of the required kind.

By contrast, it seems as though the motivational aspect of emotions cannot be what enables them to play the kind of legitimate role in persuasion that we are considering. For having a motivation to understand is not itself a way of being committed to the truth of something. Brady supposes that in emotional experience, there are at least the following 2 components: (1) we appraise object x as F, and (2) are motivated to understand reasons why x is F (or why it might not be F). It is clearly the first and not the second of these that is required if emotions are directly to supply premises to orators’ arguments in the manner we have been considering.

Nevertheless, have we too swiftly dismissed the possibility that emotions might contribute to the legitimate functioning of a persuasive argument without themselves providing the acceptance of one or more of the premises? Two options are here considered. One is the possibility that emotions’ legitimate persuasive contribution somehow stems from the feature of emotional response to which Michael Brady has recently called attention – the way in which emotions focus our attention on their objects, in pursuit (as he argues) of an accurate understanding of our evaluative landscape. The other is an option for views of emotion according to which the representational aspect of those involves perceptions, aliefs or unasserted thoughts.

On the latter of these two views, emotions can be formed as a result of the same kind of evidence / basis / reasons that generate accompanying beliefs. So, the person ends up with an emotion and a belief with similar content. The belief contributes to rhetorical persuasion. The emotion is a by-product.

Could this account for the phenomena we have highlighted? On the face of things, it seems unlikely. This ancillary role doesn't seem to do justice to our sense that the emotion itself is doing some important persuasive work in these cases. On this view, it is the audience's beliefs, which are distinct states from their emotions, on this view, that are doing the persuasive work.

But perhaps this presumes too narrow a view of how emotions might make their contribution to persuasion. Perhaps although they don't directly contribute to persuasion, the emotions keep the subject's focus on their objects, and so make the belief in question difficult to dislodge.Footnote ³¹

Perhaps the suggestion will be that the emotions serve to keep the belief in place, by making it emotionally charged and difficult to dislodge. This does not seem to get us what we wanted in order to vindicate the anti-austerity claim we are seeking to vindicate, i.e. that emotional persuasion is legitimate. For, on this view, the original question simply resurfaces – on what basis is this ancillary role for emotions legitimate? What would make it legitimate for emotions to keep beliefs in place, so as to dispose audiences to accept conclusions of arguments in which they play a part? Surely we will wish to answer that by reference to epistemic justification, i.e. that the emotion keeps the belief in place not merely causally, but by justifying its retention. But this will then be subject to the same requirements for the transmission of justification (i.e. that it be itself justified, and involve commitment to the truth of its contents) that we were considering.

What, then, of the first option above, that builds on the resources highlighted recently by Michael Brady? Could it be that emotions contribute to legitimate persuasion not directly, by contributing premises to inferences from which the subject then is justified in adopting the conclusion, but indirectly, by motivating a search for reasons that (in legitimate cases) justify belief in such a premise, and hence contribute to the justification of believing the conclusion? At first sight this seems a plausible suggestion.

The objections to this suggestion may not be wholly decisive, but they seem to me substantial, and are twofold. Firstly, it is not clear that this view gives us the right answer about which cases of emotional persuasion are legitimate and which are not. In the sections above, we have been considering the suggestion that legitimate use of emotions in persuasion takes place when the emotions that are aroused are epistemically justified, because it is then that the conclusions suitably inferred from them can be justified. That is, the audience will be justified in believing the conclusion when they are justified in believing the premises of the persuasive argument. But this justified belief in (one or more of) the premises is, for the Brady-inspired view we are now considering, the indirect result of the emotion aroused by the orator: that emotion gives rise to a search for reasons, and in the favourable case, those reasons justify their belief in the relevant premise(s). The justification or otherwise of that belief in the premises seems on this view to depend not on whether the emotion aroused is a justified emotion, but rather on what the audience themselves do when that emotion motivates them to seek reasons to support a belief in the emotion's content. If that audience, motivated in that way, discovers good reasons, the resulting belief will be justified, if not, not. This seems wrong. For one thing, it seems as though in the cases above, the legitimacy of the persuasive use of emotion ought rather to covary with the justification of the emotion aroused by the speaker; and for another, it seems to put the legitimacy of the persuasive use of emotion outside the control of the orator to too great a degree. Whether the audience is legitimately persuaded seems, on this proposal, to depend primarily on what they do in response to the motivation to search for reasons supporting the evaluative construal involved in the emotion. And that is something outside the orator's control. Perhaps it is something that the orator could predict, and we judge the legitimacy of their persuasive use of emotion, based on what they could reasonably expect of that audience. But this seems at odds with what it is most natural to say about the cases we are considering – namely that the legitimacy of their persuasive use of emotion depends on the grounds they provide for that emotion, i.e. the things they present to the audience as their basis for arousing that emotion.

The second objection is that this proposal does not seem to give a plausible psychological story about what happens in the kinds of cases we are considering. On this proposal, legitimate persuasion occurs because the orator arouses an emotion, which motivates a search for good reasons, which then justify the formation of a belief (with similar content to the emotion), whose content can then feature in the orator's argument and contribute to justifying the audience's adoption of the conclusion. There are a lot of steps in the sequence: orator's speech, resulting emotion, search for reasons, justified belief in the premise(s), and inference, followed finally by the adoption of the conclusion. Some of these will be in common with any view of emotional persuasion of this kind. But this proposal has added significant complexity by introducing a search for reasons and the formation of a belief, both distinct from the emotion itself. Perhaps all this can happen seamlessly and quickly while the orator is speaking. But it is not obvious that that is so. On this proposal, we are being offered what seems to me an implausibly complex and extended account of the psychological chain of events that occurs when Demosthenes, Jonah, Sojourner Truth or Ashley Judd persuades emotionally.

On this basis, I am inclined to think we should prefer the simpler direct account of emotional persuasion that emerged from consideration of our examples and from the explanation of these offered in Aristotle's Rhetoric. On this simpler account, the emotions involved in legitimate persuasion will need to be themselves epistemically justified, and must involve the subject's being committed to the truth or likelihood of their representational contents.