Vous êtes sur la page 1sur 12


Patricia Greenspan
Department of Philosophy
University of Maryland
College Park, MD 20742

Draft: August 31, 2002

Written for oral presentation

The innateness issue comes up persistently in cross-disciplinary literature on ethics

and emotion. My own aim in recent work within philosophy has been to explain how
an emotion-based account of ethics can still understand it as recording moral facts.(2) I
don't think this requires presupposing anything about innateness. But sociobiologists
and evolutionary psychologists thinking about the sources of morality in the social
emotions tend to follow eighteenth-century British moral philosophy, which works
from a conception of innate human nature.(3) The literature specifically on emotions
sometimes features a debate between opposing camps on innateness: theories of
"basic" emotion in psychology versus a position derived from anthropology known as
"social constructivism," which in extreme form holds that all emotion types are
variable products of culture.(4) If ethics were based on emotions on this latter sort of
account, that might indeed seem to rule out moral facts. But there's room for less
extreme views within both this and the nativist position, which holds only that some
original subset of emotions precedes social influence.

In recent philosophy Paul Griffiths has defended the notion of basic emotions as
"affect-programs," clusters of responses to selected classes of stimuli that might be
said to be programmed into us by evolution.(5) These are found panculturally, but
without the sort of fixity or unchangeability by either culture or further development
that the term "innate" (which Griffiths rejects) can be taken to imply.(6) Griffiths
acknowledges responses resembling emotions that are set up entirely by culture, but
he thinks of these as "social pretenses" of emotion that should be dealt with in a
separate category. (7)

Griffiths also raises questions about whether the full range of states that do seem
offhand to be genuine emotions can be explained in terms of the basic set - whether
we can "get there from here," as it were, with "there" taken as covering what he calls
"the higher cognitive emotions." He isn't himself concerned with moral emotions,
though his discussion of some attempts to "get there from here" by appeal to evolution
deals with authors like economist Robert Frank who do assign a central role to moral
(and other social) emotions.(8)

What I want to do in this paper is to explore how we might at least move toward
morality and moral emotions from a fairly minimal innate basis in emotion, along
with whatever conceptual equipment we also possess innately. I'll retain the term
"innate," acknowledging its ambiguities, but essentially just meaning what Griffiths
has in mind when he speaks of evolutionary "programs": unlearned responses or
clusters of response or response tendencies, emerging as a consequence of genetic
endowment, but subject to serious cultural modification, in ways that will be
particularly significant in the moral sphere.

Although I think it's plausible to suppose that there's an innate basis of morality
understood in these terms, this shouldn't be equated with the eighteenth-century
notion of an innate "moral sense." There's more learning in this area than meets the
eye - particularly an inexplicit (and initially undifferentiated) kind of moral learning
that rests on educating the emotions. After discussing various ways of arguing for an
innate basis, emotional or conceptual, I'll turn to some examples of how such a basis
might still be modified in essential ways by our early emotional interaction with
children. I'll also indicate at various stages how it might be altered and expanded by

1. The evidence for basic emotions in psychology concerns the ability to identify
emotion types cross-culturally, initially on the basis of facial expression. Exactly
which emotions are included in the basic set is a matter of dispute, but familiar states
like fear and anger are on everyone's list and also represent emotions we'd be inclined
to attribute to animals.(9) The application to animals also involves appeal to other
forms of behavior and bodily expression, along with confirming evidence from
physiology and neuroscience.

Psychologists' work on basic emotions suggests an evolutionary approach to emotions

but doesn't connect in a clear way to evolutionary accounts of morality. Positive or
altruistic social emotions such as love and gratitude don't appear on the standard lists
of basic emotions, and there's disagreement about whether to include shame and guilt,
which require broadening the criterion for inclusion to features of bodily posture
rather than simply facial expression.(10) But this area is still in flux, and there's also
evidence from other fields for attributing some social emotions, or at any rate
evolutionary predecessors of emotion, to animals.

Recent work on brain circuits by neuroscientist Joseph LeDoux indicates that even
standard entries on the list of basic emotions such as fear fall into different "tracks" of
neural response, some involving the cerebral cortex and some, the more primitive set,
only the amygdala.(11) But a similar line of thought would apply at least to primitive
elements of the social emotions such as the various neurotransmitter responses
involved in love.(12) All LeDoux uses to identify subcortical fear in nonhuman
animals, after all, is the behavioral response of "freezing."

One important feature of this deconstruction of "basic" emotions is that cortical

involvement would allow for more fundamental modification of the primitive set than
we'd get on the model of derivation from "primaries" inherited from seventeenth-
century philosophy. Descartes understood the "passions of the soul" as compounds of
simple emotions on his own "basic" list.(13) But the more advanced forms of emotion
needn't be seen as combinations of states that are already fully constituted as
emotions. Instead, the more primitive behavioral or neurotransmitter responses on the
"two-track" approach might be thought of as proto-emotions that take on the features
of full-fledged emotions when the link is made to the cerebral cortex. On a familiar
philosophical view they thereby come to be about something, or to possess what's
called "intentional content." The element of intentional content (whether itself learned
or innate) may then be modified on the basis of experience and cultural influence.
Sometimes it may be modified in ways that in turn affect the phenomenal "feel" of a
given emotion. Consider some of the variants of love: possessive, devoted, or
dependent; wild or subdued.(14)

Some philosophers would want to exclude the more primitive, pre-intentional states -
"objectless" anxiety, for instance, as a state that persists into adult human life - from
the category of emotions. In my own work I've mainly confined the term "emotion,"
along with common terms for particular emotions, to states that do have a content
capable of recording reasons for action. (15) I've treated the more primitive states as
"deficient" cases of emotion or of particular types of emotion such as fear or love. It
wasn't my aim, though, to establish the uniqueness of human emotional response but
just to assess the force of human emotions as factors in practical reasoning.(16) In fact,
I wouldn't deny a few contentful states to animals.(17) But in any case, which states we
should treat as paradigm cases of emotion, the more primitive or the more cognitively
complex, will vary with our theoretical purposes, normative or scientific.

Is there a theoretically principled reason for slanting things differently for different
types of emotion? Some of us would be inclined to deny love to nonhuman animals
while readily granting them fear. Apart from the moral or religious significance
assigned to love, one possible reason has to do with variable cultural influences on
expression. If expressions of love or other of the more developed social emotions in
humans are more subject to cultural influence, then they won't form the same
relatively stable clusters of response that we get with fear, anger, and similar "basic"
emotions on the standard list.

The work on neural circuits doesn't put any special emphasis on facial expression; nor
does it imply stable clusters of bodily response perceptible to the agent and others.
However, according to evolutionary arguments for emotions as "exaptations,"
emotion types that originally evolved as mechanisms promoting behavioral
readinesses - "fight-or-flight" in the case of fear, say - later came to be selected for
their communicative function.(18) The central role of facial expression in humans
essentially serves to concentrate this secondary function of emotions into an area that
facilitates social transfer, via patterns of shared attention, or "gaze-following," along
with the tendency of the human infant to imitate facial expression.(19) These are social
mechanisms below the level of culture but capable of conveying cultural influence,
even if themselves innate.

The innate emotional basis of ethics on this more complex kind of account wouldn't
be limited to discrete emotion types, or "basic" emotions, but would also include
general mechanisms of emotional learning, or what I think of as mechanisms of social
transfer, that also affect the more culturally malleable forms of emotion. Some authors
appeal particularly to the mechanism of emotional empathy, originating in the
tendency exhibited by newborns to cry at the sound of other infants' cries, as a basis
for guilt - a distinctively human emotion.(20) However, there's no implication on these
accounts of a culturally invariant "moral sense" as a feature of human nature.
The universal element of morality, or of the more advanced forms of morality, lies in
the application of moral norms to everyone, at least potentially, regardless of culture
or some narrower set of personal affiliations, including emotional involvement with
the speaker or simply presence to current perception. This is a question of the content
or scope of morality and does seem to require cognitive advancement beyond animals
- specifically, the development of language - if only to generate a full range of objects
of moral response. It would seem to be facilitated by cross-culturally recognizable
response tendencies just insofar as they enable us to learn moral responses from
anyone, regardless of culture. However, social influence obviously allows for
variations in the circumstances thought to justify a given type of emotion - what
counts as an appropriate object of guilt, say, in a given culture. For that matter, there
seem to be potentially competing emotional mechanisms involved at the more basic
level - empathy, but also retributive tendencies - that might be mixed or balanced
differently in different cultures, or even different personality types. So there's plenty
of room for divergent moralities or moral codes - another point of contrast with the
relevant philosophical traditions.

2. What we seem to have, then, as innate bases of ethics, are first, some primitive
states or elements of emotion, some of which we're happy to think of as full-fledged
emotions in animals, and secondly, a set of mechanisms for emotional learning, or the
social transfer of emotions. What we don't have is either the seventeenth-century
notion of emotional primaries whose combination yields the full range of emotions or
the eighteenth-century idea of an innate "moral sense" grounded in human nature.

Note that what I've been discussing, in any case, is innate emotional bases of ethics,
not (or not necessarily) innate moral emotions, even as influenced by cultural norms
in their application or developed content. Though animals might be said to have
rudimentary social emotions - or even the rudiments of moral emotions, as in
philosopher Allan Gibbard's explanation of guilt in terms of the submissive response
in animals - full-fledged moral emotions are usually understood in a way that puts
them clearly into the "higher cognitive" category.(21) In Wild Minds, for instance Marc
Hauser denies nonhuman animals genuinely moral emotions (and genuine empathy)
on the grounds that they lack self-awareness - or more fundamentally, they lack
awareness of others or themselves as bearers of the intentional states, such as beliefs,
desires, and needs, to which moral behavior in humans responds.(22)

It's still possible, of course, that moral and other cognitively complex reactions might
be innate in humans - in a sense compatible with emerging well after infancy and no
doubt also requiring certain conditions to emerge, on the model of gene-expression or
language-learning. But this can't be established just by appeal to psychologists'
evidence for basic emotions or affect-programs. A distinct line of argument for
innateness that Susan Dwyer has suggested from a Chomskian perspective would
appeal to Elliot Turiel's finding of a distinction made panculturally by preschool
children between moral and merely conventional rules or rule-infractions.(23)

This amounts to a variant of the "human nature" approach, though it substitutes for
universal moral emotions a universal conceptual basis for morality. For the "basic
emotions" variant, however, biology and brain studies supply some confirming
evidence of processes independent of learning. Without this, one might just be
inclined to explain the universality of the distinction between moral and other rules by
the way it fits the obvious facts of social life. For an analogy outside the social sphere,
consider the distinction between the important and the trivial, which I assume is found
cross-culturally in preschool children. We're innately equipped to learn such things,
perhaps, and we may not require instruction from others to learn them, but these
distinctions are "there" independently of us to be learned.(24)

Dwyer suggests an interpretation of Turiel's results in terms of a moral analogue to

Chomskian "poverty of stimulus" arguments in linguistics: there's not enough
teaching of moral rules or judgments of wrong in contrast to conventional rules or
prohibitions to explain the universality of the distinction without an innate conceptual
basis. But moral teaching takes many forms apart from explicit instruction, as my
discussion of the innate emotional bases of ethics should indicate. We educate the
moral emotions in children, for ni stance, by reacting to characters in the stories we tell
them in a way that encourages shared emotional response. Whether we express anger,
say, or amusement at a certain character's exploits can convey the idea that harm or a
slight to another is of serious moral concern in a way that mere flouting of
conventional rules is not. The ability to process this information cognitively may
depend on some innate mental structures, but the point is just that "poverty of
stimulus" arguments don't clearly establish this.(25)

There's lots of moral instruction, when you think of it, that only much later gets pulled
apart from other forms of instruction in which we use emotional display to train
children. Just saying "no" in an emphatic or warning tone counts as moral instruction
in this not-yet-differentiated sense, which doesn't distinguish wrong from dangerous
action. Even if we do allow, then, that an innate concept of morality has to be
deployed at an advanced stage in order to make all the relevant distinctions, we
shouldn't see this as the basis for morality in the sense of moral motivation and
behavior - as opposed to thought about morality or reflective moral knowledge. In
normal cases morality has already in large part been learned before the concept of
morality comes on the scene.

3. Some of the ways we teach moral motivation via emotion in advance of the concept
of morality - and without reliance on specifically moral emotions - can be seen by
looking at our play rituals with prelinguistic children. The child thrusts one of its toys
at us and we respond by elaborately portraying surprise and gratitude. The point is not
to teach "thank you" or conventions of politeness but rather to convey a kind of joy at
social exchange. We give the toy back, are handed it again with an expectant look,
and repeat the emotional display. Our feelings or pretended feelings are contagious -
not as gratitude per se, at this stage, but rather a combination of simpler positive
responses to giving and taking: excitement, delight, and amusement. The elements of
affectionate interest and attention involved in care-giving and general play begin to
get transferred by these means to other forms of social exchange.
This is a kind of early moral learning, though it hasn't yet been put to an explicitly
moral use. When we include such interactions, though, the stimulus for moral learning
can be seen to be quite rich, even if susceptible to all sorts of further development by
way of its pairing with language. At this stage we're not (or not merely) teaching
language but rather using it as one vehicle of emotional expression in an effort to
shape social emotions in children.

A more specifically moral example involves encouraging sympathy: a child kicks its
mother and the mother hams up her expression of pain and hurt feelings, conveying
the sadness and resentment of the victim that becomes moral guilt. We also
discourage certain emotions, or in the first instance overt expression of them, by
various maneuvers such as not making too prolonged a fuss about a child's own minor
hurts and pains. Clearly all of this is subject to the influence of cultural or other local
norms - "display rules," in Ekman's term - even if based on innate emotions, innate
mechanisms of social transfer of emotions, and innate conceptual categories or
structures of moral thought.(26) Whose pain counts, what you should feel about
causing it, and whether and when you may express or acknowledge those feelings are
questions that can be answered differently at this early stage of social learning.

There also are other concerns of morality besides harm or pain that can be emphasized
and interpreted differently by different cultures. Disgust is one of the entries on the
standard list of basic emotions, as I've noted, and it's put to use by moralities that
depend on a concept of vice or moral sin. In current terms, it's the emotion we express
in response to behavior seen as "beneath" a morally scrupulous agent. It's particularly
interesting in connection with the Chomskian analogy because it seems to exhibit in
emotional form the "principles-vs.-parameters" structure of Chomskian Universal
Grammar: though the emotion itself appears to be pancultural, apparently there's little
or no cultural uniformity in its "natural" objects - in which food-sources evoke
disgust, say.(27) In that sense the parameters of a universal emotional principle (that
some things warrant fear of contamination) get set by culture.

A different sort of example is provided by feelings of respect or deference. They're

essential to the workings of social hierarchy - and arguably, in a different way, to
social equality - but it's not obvious how they could be derived from entries on the
standard list of basic emotions. They might initially have emerged from fear, but in
adult human life they seem to have lost any tight connection to it. Perhaps they might
be understood as reactions to certain basic emotions in others - acquiescence in their
anger, say, as on Gibbard's account of guilt.(28) However, understanding them as
affective responses would seem to involve augmenting the basic set with further
materials provided by social interaction.

One might want to say that feelings like respect that don't constitute relatively
stereotyped sets of responses like the items on the standard list of basic emotions don't
really count as distinct emotions. I don't have a fixed answer to classificatory
questions, as I say. When you think of it, though, gratitude is subject to similar
questions, though it's definitely an entry on standard lists of nonbasic emotions, and it
plays a central role in evolutionary accounts of morality as what motivates reciprocal
altruism. In fact, I think we'd be missing much of relevance to moral and practical life
if we confined attention to recognized categories in discussing emotions.
If we don't restrict emotions to relatively stereotyped clusters of response, the role of
language as the ultimate vehicle of content makes the number and variety of emotions
potentially limitless. We just need to attach some affective element - not necessarily a
whole affect-program - to a thought or cognition.(29) This might be a thought about
emotions, as in Gibbard's account of moral emotions, but in the simplest cases,
cognition and affect just need to register a corresponding kind of significance of the
situation for the organism, on the model of threat in the case of fear. Specifically
moral emotions result from having the linguistic resources to express social norms,
plus some way (whether evolutionary or social) of programming in an aversive
reaction to a norm violation.

Guilt itself wasn't a recognized emotion category until relatively recently. The noun as
used in English to name an emotion was apparently the result of linguistic error in the
period of the Reformation, a time when emotions of religious despair and self-
accusation were all the rage.(30) That it can be reconstructed from items on the list of
basic emotions - variants of sadness or anger, conveyed by the mechanism of
emotional empathy - may explain why the word seems so clearly to name something
that was there all along, but it's something that not all societies put to a moral use.
Often the function of social control of behavior is assigned to shame, as a sense of
failure to measure up to social norms.(31)

Gibbard explains the distinction between guilt and shame by pairing them with
different social emotions: guilt involves submitting to others' anger (viewing it as
warranted) whereas shame involves accepting their contempt. This explanation makes
appeal to others' basic emotions, but it doesn't obviously derive guilt and shame from
basic emotions in the individual's own repertoire - from sadness, say, considered as a
particular cluster of expressive and physiological responses, rather than just emotional
distress or discomfort in a broader sense.

For that matter, guilt also includes "aroused" variants such as moral self-loathing or
self-hatred that would be harder to derive from sadness. There's no requirement that
the higher cognitive emotions fall into categories that correlate one-one with sources
in the basic set. But why should they be limited to sources in the basic set, either?
Probably both shame and guilt, along with forms of respect and similar states, also
have some basis in adult behavioral responses to the child - incidents of social
rejection in an overall situation of acceptance and trust - that can be said to evoke
generalized emotional unease. We apply this to a cognitive content when we build in
reference to warrant for others' emotions.

In short, there's no need to "get there from here," in the sense that implies explanation
in terms of a primitive set of emotions in our own repertoire, as on the Cartesian
model. We have lots of initially disorganized materials to work with - from the
standpoint of categorization into distinct emotions - in whatever pre-emotional
feelings and extra-emotional conceptual structures we can harness to the task of
constructing moral reactions. The important thing that basic emotions supply is a link
to facial and other bodily expression that helps make our moral reactions extensible to

Robert Frank's account of moral emotions as "commitment devices" at first sight

might seem to require making them out as response clusters stamped in by evolution.
In a nutshell, Frank argues that the communicative function of emotions depends on
their role as supplying expressive signs of our intentions that others can reliably
determine to be genuine. We can't easily fake blushing out of shame, say, so the
emotion serves as a check on our ability to conceal behavior that violates social rules.
Interpreting basic emotions as affect-programs serves to take emotional expression
out of the realm of voluntary behavior, which is relatively easy to fake. Others have a
corresponding ability to tell which signs of emotion are genuine.

However, all that's strictly needed for Frank's argument to apply is the reliable
connection - and awareness of the connection - between some nonvoluntary
expressive signs of emotion such as blushing out of shame and longer-range
intentions such as those made likely by the urge to conceal something. Even if moral
learning begins with preprogrammed clusters of response, their elements might later
be separated, recombined (or combined with further cognitive elements), and
reconstituted into culturally variable affective complexes - on the basis of both the
way we learn expressive behavior and the possibilities that language offers for
differently specified cognitive contents of emotion. One should not exaggerate our
ability to tell whether the resulting complexes - guilt with its submissive posture, say;
or respect as an impulse to social deference - are genuine. But emotions on this more
complex account will still at least be difficult to fake, providing an external handle on
others' intentions and a way of establishing our own.

An important feature of an account that allows for reshuffling affective and cognitive
emotion components is that it yields an emotional basis of ethics that isn't necessarily
"sentimentalist," in the sense of making out the content of moral judgments in
emotional terms. Moral judgments themselves may be supplied by fairly sophisticated
reasoning embedded in cultural practices (but subject to individual rethinking) about
ways of effectively coordinating group life, though emotions figure centrally in our
ability to teach morality at an earlier stage, as we have to do in order to implant it in
individual motivation.(32)

This amounts to a further departure from the philosophic tradition. Philosophers who
base morality on emotion typically follow eighteenth-century figures such as
Hutcheson or Hume in making out moral emotions as perceptions (possibly illusory
perceptions) of moral facts.(33) But there's another way of doing the job without either
innate moral emotions or a denial that moral truths involve a kind of fact. The content
of ethics may be supplied by cognitive judgments whose motivational force depends
on the possibility of recruiting emotions for initial learning in the way I've sketched.
In that sense, the innate bases of ethics includes emotions, most importantly in the
form of the mechanisms of social transfer of emotion, along with whatever general
conceptual structures underlie language learning and rational thought. Our innate
emotional endowment constrains the content of ethics by ruling out some conceivable
moral codes as unteachable and hence unsustainable. But it leaves room for variation,
both social and personal - at least some of it a matter of moral error.

1. Let me thank Stephen Leighton for comments on an earlier draft of this paper. Most
of the footnotes to the paper will be omitted from oral presentation.

2. See esp. Practical Guilt: Moral Dilemmas, Emotions, and Social Norms (New
York: Oxford University Press, 1995); cf. also "Moral Responses and Moral Theory:
Socially Based Externalist Ethics," Journal of Ethics, 2 (1998), 103-22.

3. One well-known example of a sociobiological account is James Q. Wilson, The

Moral Sense (New York: The Free Press, 1993). For the Darwinian account see
Charles Darwin, The Descent of Man, and Selection in Relation to Sex (Princeton,
N.J.: Princeton University Press, 1981), esp. ch. 3, and The Expression of the
Emotions in Man and Animals (Chicago: University of Chicago Press, 1965). In the
eighteenth-century literature cf. esp. David Hume, A Treatise of Human Nature
(Oxford: Clarendon Press, 1978), Bks. II-III, and Adam Smith, A Theory of the Moral
Sentiments (Indianapolis: Liberty Fund, 1982).

4. On basic emotions see, e.g., Paul Ekman, "Universals and Cultural Differences in
Facial Expressions of Emotion," in J. K. Cole (ed.), Nebraska Symposium on
Motivation, vol. 19 (Lincoln: University of Nebraska Press, 1971); cf. also Carroll E.
Izard, Human Emotions (New York: Plenum, 1977). On social constructivism see,
e.g., James R. Averill, "A Social Constructivist View of Emotion," in R. Plutchik and
H. Kellerman (eds.), Emotion: Theory Research and Experience, vol. 1 (New York:
Academic Press, 1980), and Catherine Lutz, "The Domains of Emotion Words on
Ifaluk," in R. Harré (ed.), The Social Construction of Emotions (London: Oxford
University Press, 1986).

5. See Paul E. Griffiths, What Emotions Really Are: The Problem of Psychological
Categories (Chicago: University of Chicago Press, 1997.

6. Cf. Griffiths, "What is Innateness?," The Monist, (2002), 70-85. Griffiths' point is
that the concept of innateness blurs together three distinguishable notions, roughly
(ignoring some subtleties of his argument): (1) developmental fixity, (2) species
universality, and (3) evolutionary origin.

7. Cf. esp. Griffiths' discussion on pp. 140ff. of cases resting on the pretense that the
state in question, as a putative emotion, is difficult or impossible to control such as the
"wild-pig syndrome" in New Guinea described in James R. Averill, "Emotion and
Anxiety: Sociocultural, Biological, and Psychological Determinants," in A. O. Rorty
(ed.), Explaining Emotions (Berkeley: University of California Press, 1980). Cf. my
"Practical Reasoning and Emotion," in A. Mele and P. Rawlings, Handbook of
Rationality (Oxford: Oxford University Press, forthcoming,

8. See Robert H. Frank, Passions Within Reason: The Strategic Role of the Emotions
(New York: Norton, 1988).

9. But for a contrasting treatment of disgust as distinctively human cf. William Ian
Miller, The Anatomy of Disgust (Cambridge, Mass.: Harvard University Press, 1997),
p. 12. Agreed items on the list are six: anger, disgust, fear, happiness, sadness, and

10. Cf. esp. Izard, Human Emotions, esp. pp. 83-92. Note that contempt gets added to
Ekman's initial list in Paul Ekman and Wallace V. Friesen, "A New Pan-Cultural
Facial Expression of Emotion," Motivation and Emotion (1986), 159-68.

11. See Joseph LeDoux, The Emotional Brain: The Mysterious Underpinnings of
Emotional Life (New York: Simon & Schuster, 1996).

12. See esp. Jaak Panksepp, Affective Neuroscience: The Foundations of Human and
Animal Emotions (New York: Oxford University Press, 1998), ch. 13.

13. See Réné Descartes, "The Passions of the Soul," The Philosophical Works of
Descartes, Vol. I, trans. by E. S. Haldane and G. R. T. Ross (Cambridge: Cambridge
University Press, 1970), pp. 331-99.

14. Besides treatments of romantic love in Western culture, cf. the classic study of
amae as a Japanese variant of love based on dependency in Takeo Doi, The Anatomy
of Dependence (Tokyo: Kodansha International, 1973).

15. See Emotions and Reasons: An Inquiry into Emotional Justification (New York:
Routledge, Chapman and Hall, 1988); more recently, cf. "Practical Reasoning and
Emotion," and "Emotions, Rationality, and Mind/Body," Proceedings of the Royal
Institute of Philosophy (forthcoming).

16. For the uniqueness argument cf. Ronald de Sousa, The Rationality of Emotions
(Cambrdige, Mass.: M.I.T. Press, 1987), ch. 4.

17. Cf. Emotions and Reasons, p. 50. Note that what's in question here is not
necessarily "conceptual" content (implying inferential structure) but just "aboutness."
I would say, for instance, that my cat is upset about not getting enough attention when
I work on this paper at night, rather than simply becoming agitated under those
circumstances. Intentional states at this level needn't involve separable thoughts (let
alone beliefs), as far as I'm concerned, but just contents of emotions. This is of course
a complex subject, but in answer to questions about how such contents can refer to
one thing rather than another without the addition of logical structure, specifically
negation, I would bring in attention (and the limits on attention). A better analogy
than linguistic content might be pointing, whereby a simple physical gesture manages
to be about something.

18. Cf. the argument in Robert H. Frank, Passions Within Reason: The Strategic Role
of the Emotions (New York: Norton, 1988).

19. Cf., e.g., the discussion in Paul Bloom, How Children Learn the Meanings of
Words (Cambridge, Mass.: M.I.T. Press, 2000), pp. 62-4. For relevant work more
specifically on emotions cf. David C. Witherington, Joseph J. Campos, and Matthew
J. Hertenstein, "Principles of Emtoion and its Development in Infancy," in G.
Bremner and A. Fogel (eds.), Handbook of Infant Development (Oxford: Blackwell,
20. Cf. Martin L. Hoffman, "Development of Prosocial Motivation: Empathy and
Guilt," in N. Eisenberg-Berg (ed.), Development of Prosocial Behavior (New York:
Academic Press, 1982).

21. See Allan Gibbard, Wise Choices, Apt Feelings: A Theory of Normative
Judgment (Cambridge, Mass.: Harvard University Press, 1990), ch. 7.

22. See Marc D. Hauser, Wild Minds: What Animals Really Think (New York: Henry
Holt, 2000), esp. p. 224; cf. pp. 250ff. (Note that Hauser's point here is still quite
compatible with the attribution of intentional content to animals.) Hauser also takes as
essential to morality the capacity to inhibit one emotional response by another; cf. p.

23. See Susan Dwyer, "Moral Competence," in K. Murasugi and R. Stainton (eds.),
Philosophy and Linguistics (Boulder, Colorado: Westview, 1999), pp. 169-90; cf.
Elliot Turiel, The Development of Social Knowledge: Morality and Convention
(Cambridge: Cambridge University Press, 1983).

24. Of course, on certain metaphysical views, "there" could only mean "in the mind,"
or a world constituted or constructed by minds. But it's mental contents - or behavior
tendencies or other properties of a particular mind - that are properly said to be
"innate" in the sense that contrasts with being learned. A structural constraint on what
minds can come up with isn't innately in the mind, an innate component of a mind, if
its recognition depends on learning.

25. It's also unclear whether the moral/conventional distinction would be enough to
support a moral analogue to Chomskian Universal Grammar, as Dwyer claims. I take
it that the latter entails, not just the general capacity to learn some natural language,
but more specifically, various innate mental structures common to all language-
learners. At most, the moral/conventional distinction by itself would just assign
morality a place in a Universal Grammar of thought. We need more to give it any
internal content.

26. Cf., e.g., Ekman, "Universals and Cultural Differences," for the well-known
finding is that Japanese tend to display less negative affect in the presence of an
authority figure. If these are "parameters" of the concept of morality, in the
Chomskian term, my point above is that many of them get set before the concept is

27. For some of the original work on this subject see Paul Rozin and April E. Fallon,
"A Perspective on Disgust," Psychological Review (1987), 23-41.

28. Cf. Gibbard, ch. 7; cf. Simon Blackburn's account of "emotional ascent" in Ruling
Passions (Oxford: Clarendon Press, 1998), ch. 1, sec. 3.

29. By "we" I of course mean human development, not just the theorist; and I do
agree that evolutionary responses have a kind of primacy. The point is just that they're
not set in stone. Cf. Justin D'Arms and Daniel Jacobson, "The Significance of
Recalcitrant Emotion (Or: Anti-Quasijudgmentalism)," in Proceedings of the Royal
Institute of Philosophy (forthcoming).
As one small example of a nonstandard category that we might want to recognize for
purposes of psychological explanation, I've noticed a variable factor in adult behavior
in group settings that might be called "conflict-aversion": the tendency to be made
uncomfortable by social disharmony, possibly to the point of needing to cover up
disagreement. I assume that this is culturally variable (as well as varying within
cultures - on the basis of occupational or gender roles, say) and that it results in the
first instance from how others respond to a child's displays of anger and other socially
disruptive emotions.

30. The OED cites Archbishop of Canterbury John Tillotson for an early use. Cf. my
discussion of these topics in Practical Guilt, ch. 4.

31. Cf. Ruth Benedict, The Chrysanthemum and the Sword: Patterns of Japanese
Culture (Boston: Houghton Mifflin, 1946) and Takeo Doi, The Anatomy of
Dependence (Tokyo: Kodansha International, 1973). Bernard Williams, in Shame and
Necessity (Berkeley: University of California Press, 1993), argues that shame
essentially did the work of guilt for the Homeric Greeks.

32. See my Practical Guilt, esp. chs. 3 and 6.

33. For examples of these two main approaches in contemporary philosophy see esp.
Ronald de Sousa, The Rationality of Emotion (Cambridge, Mass.: M.I.T., 1987), and
J. L. Mackie, Persons and Values: Selected Papers, Vol. II (Oxford: Clarendon Press,
1985), ch. IX.