 

Personality and Character

. An Ideal
Character is the final, ‘boss level’ of the affective life. Ideally, it is the
ultimate arbiter of the emotions we have, as well as the manner in which
we express and act on them. I say ideally, since it is clear that character is
not something we always manage to show. Most of us will find at least
something lacking about ourselves when reading the descriptions of fully
robust and reliable characters celebrated by the philosophers. While those
who read these descriptions thinking ‘yep, that’s me’ are probably deceiv-
ing themselves. However, I think it is fair to say that over the course of
long, painful years of struggle, most of us attain a bit of character.
To understand character, we must differentiate it from personality.
Personality and character have similar meanings. Both refer to enduring
psychological features that help to distinguish one individual from another.
Both, moreover, are ways in which the individual systematically prioritises
the pursuit of some concerns over others. Ontologically however, the two
notions are distinct. I will argue that personality corresponds to individual
differences in the relative sensitivity and strength of our concern-regulating
systems. This potentially applies to every concern-regulating system dis-
cussed in this book. In contrast, character is a definite level of affective
control. It helps us to manage our concerns with respect to the widest
range of personal contexts and time periods.
My basic claim about character is that it develops out of the struggle to
maintain sentiments. I define sentiments as enduring commitments to the
welfare of specific individuals, institutions or causes. It is in virtue of
maintaining such sentiments that the distinctive normative properties of
character are generated. Meanwhile, the focus of sentiments on specific
individuals requires from us a representational capacity that is able to
pick out individuals uniquely. I claim that this is achieved by means of
narratives.



Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind

. Personality Traits

I will begin by reviewing some of the key ideas in contemporary personal-
ity theory. This summarises a fairly well established body of research, and
my goal, initially, is to incorporate these ideas. In the following section,
I will draw connections with my control theory.
Among psychologists, the main approach to personality is by appeal to
traits. This derives from our ordinary folk-psychological use of terms like
‘irritable’ or ‘phlegmatic’ when trying to explain each other’s behaviour.
These terms refer to apparent dispositions or settled tendencies in the way
that people behave, think or feel. Thus if we say that a person is irritable or
phlegmatic, we mean that it is predictable that the individual will act in an
irritable or phlegmatic manner, given the opportunity. For example, to be
stuck in traffic is a common source of frustration, but the irritable person
will more frequently display frustration in this situation or display it with
greater intensity. Personality psychologists then further operationalise traits
as bivalent scales upon which all people can be placed to some degree or
another. The distribution of the population upon these scales will approxi-
mate a normal bell curve with the majority sitting near the middle, and
relatively few at either extreme (McCrae & Costa : ).
Significant personality traits should not be simple dispositions to react
in stereotyped ways to a specific stimulus, like a phobia, but should be
identifiable in the way a person behaves across a range of circumstances.
However, this aspect of personality theories came in for considerable
criticism in the s and s. Various studies seemed to indicate that
small situational factors have a much bigger impact on behaviour than
would be predicted if people possess generalised traits. One of my favour-
ites is a study that found that people are five times less likely to help an
apparently injured man when a power lawnmower is running nearby
(Mathews & Canon ). Another example is the well-known ‘bystander
effect’ whereby people are far slower to help someone obviously in trouble
if others are standing around doing nothing (see Fischer et al.  for a
review of various permutations).
Actually, we should expect traits to display variations across circum-
stances, because of their strong connections with emotions. Recall that
emotions are specifically sensitive to the social or environmental context.
Thus an individual may display bravery one day (when facing social
pressures) and cowardice the next (when seeing a snake) without under-
mining the genuine possession of a trait. Even if we narrow our traits to
say, social interactions, features like social status rely on different

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
characteristics within different contexts (athletic prowess, wit, age, hier-
archical role, etc.). Sensitivity to these contextual variations will have
resultant effects on one’s manifestation of the social trait.
The situationist critique has to some extent been mollified by gradually
accumulated evidence that trait attributions remain very stable over time
and that they can predict important life outcomes, such as relationship
stability, job performance and even longevity (reviewed in Fleeson et al.
; see also Mehi et al. ). The nature of traits has also been
qualified. It is allowed that for an individual to possess a trait to a certain
degree is for his or her behaviours to cluster around an average, displaying
more or less intensity on different occasions (e.g. Fleeson et al. :
–). This means that an individual’s traits will not always have a
completely consistent effect on his or her behaviours, and that sensitivity
to the circumstances will make a significant difference.
Webber () makes a related qualification that to possess a trait is to
raise the probability that one will act in accordance with it. This means
that sometimes the situation will facilitate the trait sufficiently for it to be
manifested in behaviour, where on other occasions facilitating conditions
will fall below the threshold for manifestation. Webber situates this point
within a general connectionist account of personality. According to this
model, evaluative attitudes towards objects are made up of clusters of
beliefs and affective states, which are connected with varying degrees of
associative strength. The extent to which an attitude is activated in a given
situation (e.g. the pro-attitude towards helping people in need) will be a
function of how easily that attitude can be associatively accessed from the
individual’s current thoughts. Webber suggests that the reasons why, in
situationist experiments, people fail to act according to values that they
would normally profess, is that these experiments temporarily make certain
attitudes less accessible. For instance, in cases of the bystander effect,
when we see that other people are apparently unwilling to act, this raises
the accessibility of the attitude to conform to social norms and inhibits the
accessibility of the attitude to help others in need. This can explain why
people tend to intervene much more slowly.
Webber also notes that there are traits that concern one’s general
cognitive style or needs, such as the degree to which one takes situational
details into account (considerateness, circumspection) or the degree to


Note that Webber’s main aim in this article is to defend the existence of character. However, he does
not draw a principled distinction between character and personality. At best, there will be a
difference of degree in attitude strengths. On this point I disagree with Webber.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
which one prefers to plan or control things. These traits are not generally
addressed in the situationist studies. Moreover, Webber emphasises that
traits are not simply a matter of how one reacts to a situation, but they also
concern the situations that one actively seeks out. Here accessibility is less
of a problem, because the way the individual weights his or her evaluative
attitudes or preferences can better assert itself.
Overall, I want to incorporate the model of traits described by the
personality psychologists. In particular, it is important to recognise that
many personality traits equate to emotion dispositions. Our emotions
display enduring individual differences with respect to their degree of
sensitivity and the sorts of objects to which they are attuned. At the
same time, some of the most important traits concern psychological
features that are more basic and pervasive than single emotion dispos-
itions. Still, in the next section I will argue that even the most funda-
mental traits can be explained in terms of our dispositions to regulate
concerns.

. The Big Five

According to the trait approach, an individual’s personality as a whole will
be the product of the interactions between his or her various traits. In
pursuit of such individual summations, psychologists have been interested
in identifying a small set of highly generalised traits that represent the
most basic dimensions of individual variability. A near-consensus has
apparently emerged since the s on what are called ‘the big five’
personality traits: extraversion, neuroticism, agreeableness, conscientious-
ness and openness to experience (McCrae & Costa ; Block ).
These five traits were arrived at independently by different personality
researchers by means of factor analysis, which is a method whereby
participants rate the similarity of a large number of trait-related linguistic
items, allowing researchers to discern the clustering of items along factors
or dimensions. The five traits have also been found in factor analytic
studies in different languages (though see Ashton & Lee , who review
some inconsistencies).
Given the popularity of the big five model, it is worth spending
some time analysing the specific traits identified. From a philosopher’s
perspective, the terms are rather vaguely defined. Note in particular that
while factor analysis clusters terms together, it is up to researchers to
provide a descriptive summary and label for these clusters. Thus there is
room for interpretation (and potentially differences in application across

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
experiments and translations into other languages). In each case, I will note
the connection with affective states.
Extraverted-Introverted: This is one of the most clearly recognisable
traits of the big five. To be extraverted is to be outgoing, talkative, fun-
loving or boisterous, where to be introverted is to be quiet, reserved,
undemonstrative, or shy. Extraversion is sometimes described as ‘positive
emotionality’ (e.g. Shiner & Caspi : ). It also seems particularly
linked with social dominance and confidence (e.g. McCrae & Costa
: ).
Neurotic-Calm: Neuroticism is the big five trait that is most closely
associated with specific emotion dispositions, towards for instance, anxiety,
depression or anger. However, neurotic people apparently have exagger-
ated tendencies towards all of these emotions, thus neuroticism is sup-
posed to be more general than any single emotion disposition (McCrae &
Costa : ).
Agreeable-Disagreeable: Agreeableness is principally focused on how
we relate to other people. Agreeable people tend to be friendly, trusting
and generous, while disagreeable people are ruthless, stingy, critical and
antagonistic. It seems then that agreeableness is closely associated with the
social emotions, or the general ways in which people regulate their social
affiliations.
Conscientious-Unconscientious: The conscientious person is
described as well-organised, hardworking, punctual and ambitious where
the unconscientious person is disorganised, careless and aimless. A general
link has also been made with conscientiousness and the ability to delay
gratification (Roberts et al. ). This trait thereby links most closely to
the stability of desires.
Open Minded-Close Minded: Open-mindedness is the trait most
closely associated with intellectual virtues. The person who is open-
minded is receptive to new ideas, creative, heavily invested in the imagin-
ation and tends to be socially liberal. Meanwhile, the close-minded person
tends to prefer the familiar and be socially conservative. The affective
connections are less clear here than with the other big five traits. However,
McCrae and Costa particularly link this trait with valuing ideas and
experiences for their own sake (: ). This is a kind of pleasure-
seeking. The contrast with familiarity preferences also suggests that anxiety
regarding the unfamiliar or new may be present to some degree in close-
mindedness.
One of the most noticeable features of each of the big five traits is that it
looks better to be on one side of the scale than the other. Extraverted

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
people are reportedly happier (indeed, happiness is built into the defin-
ition). Agreeable people are more pleasant to be with. Conscientious
people are more hard working and successful. Open-minded people are
smarter, and calm people are less prone to distressing negative emotions or
general instability. This is not to say that one can’t be too agreeable (e.g.
servile) or too extraverted (e.g. brash and annoying). But there is a clear
bias in these terms towards ways of being good.
On reflection, this should not be surprising. The factor analysis method
that was used to derive the traits works with the language that we
ordinarily use to describe people. It makes sense that when describing
people, we are most interested in the ways in which they are good or bad.
What the clustering of the traits into five factors tells us is that there are
five relatively independent ways in which people can be good. Indeed, a
sixth proposed factor of humility or honesty (Ashton & Lee ) also
clearly links to a traditional religious virtue and a sense of justice.
It should, however, be troubling for personality psychologists interested
in providing a descriptive account of human nature to rely so heavily on
normative factors like these. It raises methodological worries about poten-
tial bias in both self-reports and external ratings. Trait descriptions may be
susceptible to cross-cultural or cross-historical disagreements rooted in
different cultural norms. More generally, it is just worth getting a clear
sense of what people are like before we mix it up with how people ought to
be. The big five traits risk obscuring that distinction.
However, I think we can construe the traits in less normatively loaded
ways. Since personality traits such as the big five are indicated by observed
behaviours, they imply not just the ways in which people are good, but
also the ways in which people actively seek out value. Ashton and Lee
() develop a related view by associating the different traits (including
their sixth trait of humility/honesty) with adaptive strategies. The key
reason for treating personality traits in this way is that each trait has its
own rewards and costs. For instance, the heightened sociability of extra-
version tends to deliver social gains, but it is also energetically costly, and
exposes the individual to social risks. Similarly, neuroticism increases one’s


The big five traits are not completely independent (contra McRae & Costa : ). For example,
it does not seem possible to be extremely introverted while also being extremely agreeable. A highly
agreeable person should be disposed towards gregarious interactions with others. Similarly, it is hard
to imagine someone who is extremely open-minded while also extremely disagreeable. One cannot
be truly receptive to new ideas if one is also highly antagonistic and critical of others. Tighter
definitions could eliminate these overlaps, but would simultaneously undermine the inclusivity of
the factors.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
sensitivity to dangers or losses. In risky situations, this can be beneficial,
while in other situations it is liable to generate unwarranted emotional
responses (cf. McCrae & Costa : ).
We can ground the appeal to strategies in the process of concern-
regulation. In general, a distinctive personality trait depends upon a
relatively heightened sensitivity to the relevant concern or cluster of
associated concerns. At the same time, sensitivity to a concern implies a
capacity for regulating that concern. This derives from the basic structure
of valent representation. Moreover, concern sensitivities and capacities are
likely to be mutually reinforcing. Where the cost of utilising the capacity is
compensated by the reward, the continued deployment of the capacity is
encouraged. And the more one rehearses the capacity, the more effective it
should become at delivering the relevant value.
Relatedly, it is likely that valent representational systems adapt over time
to the availability of their targets. Suppose for instance that a regulative
response (like crying) simply is not working at affecting the presence of the
target (e.g. getting the caregiver to come). In the short term, the individual
should be capable of terminating the response or adopting an alternate
response. In the long term, repetitive adjustments should lead to lasting
changes in response strategies. We see this in neglected infants who
develop ways to self-soothe rather than rely on parental engagement (as
discussed in Chapter ). Similarly, repeated experiences concerning the
availability of a resource (or lack thereof ) should lead to adaptations such
that expectations are altered, thereby changing the degree to which emo-
tions that depend on expectations are triggered (e.g. fear or excitement).
Both of these factors could have wide-reaching implications for building
on certain concerns or response strategies in later life.
Note also that concern-regulation strategies need not be identified with
a person’s reflectively conscious preferences. Personality traits develop
from the beginning of life, and thus regulation strategies are liable to be
habitual. An individual may well consciously endorse their habitual strat-
egies, but they need not. Thus, an introvert may wish they were more
extroverted, but changing the habits of a lifetime is a tall order. Moreover,
the introvert may not fully appreciate their emotional aversion to social
risks, which raises a high barrier to the extrovert’s strategy.
Overall, I propose that each of the big five traits corresponds to the
relatively heightened sensitivity to a certain concern (or associated cluster
of concerns), which is encouraged and continually reinforced by a certain
capacity (or cluster of capacities) for securing that concern. We can now
analyse the big five traits in these terms. Thus, the agreeable person has a

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
heightened attraction towards social intimacy. This is most likely under-
pinned by the capacity to interact fluently with others or good empathic
skills. The conscientious individual values achievements highly (they may
also be distinctly averse to error). This seems most closely connected with
the capacity to maintain motivational and attentional focus on a given
goal, whereas the unconscientious individual is distracted easily, or finds it
hard to form clear priorities. In the case of open-mindedness, the relevant
value or concern seems to be with the new, as opposed to the familiar.
With regards to capacity, the trait is linked with imagination and creativ-
ity, and thereby a facility for handling new ideas that may conflict with
existing beliefs. Neuroticism, as mentioned earlier, involves a heightened
sensitivity to dangers and losses. Alternatively, we can look to the opposite
side of the scale and suppose that a calm demeanour entails a generalised
preference and capacity for minimising negative emotional arousal. That
is, calm people avoid catastrophising or being too caught up in worries.
Finally, the introversion-extraversion divide has been linked with the
preference and capacity for regulating arousal. Introverts tend to avoid
high levels of excitement (Eysenck ). In this respect they are the
mirror of neurotic people, preferring calm contentment to risky arousal.
An analysis of the big five personality traits can reveal some interesting
generalities in the ways in which individuals pursue their concerns. I am
not, however, persuaded that the big five represent the highest level of
generalities about individual differences. The fixing on five traits is not
simply a natural fact, and both a greater and fewer number of traits can be
justified depending on the level of exactness one specifies in the analysis
procedure. For instance, a combined analysis of several five-factor correl-
ation matrices led Digman () to propose two higher-order factors, one
relating to emotional stability and the other relating to active engagement
or the search for personal growth. More generally, factor analytic models
rely on a rather top-down method of analysing everyday intuitions about
individuals via our ordinary linguistic behaviour. It is essentially a more
statistically sophisticated version of the kind of ordinary language analysis
in which philosophers engage. Such a method should at least be supple-
mented with other routes to identifying personality features.
Two alternative routes in particular seem available. First, there is the
developmental perspective, in which we examine to what extent infants
display distinct psychological characteristics that carry over into adulthood.
Certainly, differences in infant temperaments can be observed from birth.

Block (: –) also reviews several related studies indicating two higher-order traits.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
Some infants are more irritable, some are more socially attentive and some
are more physically active. One particularly important personality factor
that developmental psychologists focus on is attachment style (e.g. Bowlby
/; Ainsworth & Bell ). Observational studies have revealed
three major attachment styles: secure, anxious and avoidant (a fourth
disorganised style is also observed in which infants show disturbed and
contradictory behaviours). A key experimental confirmation of these dif-
ferent styles is the ‘strange situation’ procedure, in which the caregiver of a
year-old infant leaves the room while the infant plays with a toy, a stranger
enters and then after a few minutes the caregiver returns. Infants with a
secure attachment style show some distress when the caregiver leaves, but
are able to interact with the stranger and are easily comforted when the
caregiver returns. In contrast, infants with an anxious attachment style
show much greater distress when the caregiver leaves, as well as when the
stranger enters, and are harder to calm when the caregiver returns. Finally,
avoidant infants display little distress when either the caregiver leaves or
the stranger enters, though this belies a raised heart rate, indicating anxiety
(Ainsworth & Bell ).
It is believed that attachment style is established early in childhood,
partly as a result of early caregiver interactions, and has wide-ranging
effects on adult social behaviours such as romantic relationships (e.g.
Hazan & Shaver ). A particularly noteworthy use of the theory is
Magai and Haviland-Jones () in which in-depth personality studies of
three twentieth-century psychiatrists (Carl Rogers, Albert Ellis and Fritz
Perls) are provided. Magai and Haviland-Jones claim that the attachment
styles of these psychiatrists had profound effects not only on their day-to-
day adult behaviour, but even on the theories of therapeutic practice that
they ultimately developed. It would not be an understatement to say that
on Magai and Haviland-Jones’ approach, attachment style is the most
significant trait for determining a person’s life course.
The other major route to discerning personality traits is via a ‘first
principles’ approach of examining the various affective and cognitive
systems that human beings possess and identifying the areas where these
will admit individual differences. For instance, when we examine the
capacities I have thus far examined in this book, we can anticipate
individual variations in the introspective sensitivity to bodily feeling, in
detecting the emotions of others (either by means of expressive perception
or imaginative simulations) or in the generation of emotionally stimulating
mental imagery. At the most foundational level, there should be variations
in the thresholds for triggering regulations of attractants and avoidants, or

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
equivalently, for prioritising the pursuit of one concern over another. This
plausibly allows for individual differences in emotional sensitivity, and
possibly, the degree to which the individual is neurotic or phlegmatic
more generally.
When we combine these three different methods of discerning traits,
there do seem to be some convergences on key individual differences such
as the style of social engagement and the sensitivity to emotional arousal.
However, it is doubtful that there is any psychological capacity, or indeed
bodily capacity, that does not have some kind of impact on the manner in
which we pursue our concerns. Since there are always going to be individ-
ual differences in the effectiveness with which people deploy their capaci-
ties, the range of potential differences between people is vast. It seems that
only for the sake of simplifying research projects is there any impetus for
limiting ourselves to a small set.

. The Nature of Character

Personality generally reflects the way in which people pursue their con-
cerns, and reaches into all the affective capacities that we have explored in
this book. Certain recognised personality traits (such as conscientiousness
or calmness) indicate a level of affective control over one’s emotional
dispositions. However a more distinctive elaboration of the control system
is only really found when we turn to character.
The key feature of character is its normativity. To have character (where
this is equivalent to having a strong character) is something good. Of
course, some people have bad (strong) characters – but Satan is at least
fiercely independent. And sometimes real-life criminals appeal for their
daring or fortitude even while their actions are deplored. However, the
normative nature of character does not straightforwardly translate into a set
of normative character traits, distinct from the set of personality traits. As
we saw with respect to the big five, normative traits find their way into
personality models as well. There is no obvious reason why personality
models cannot include paradigmatic virtues like bravery, loyalty and
honesty. The relative possession or lack of these traits can be universally
applied, and it can effectively serve to distinguish different people.
The main distinctive way in which normativity gets into character is via
a demand for consistency. While personality traits may well display cross-
situational consistencies, there is no demand that they ought to (cf. Goldie
: ; Deonna & Teroni ). If a man with a generally phlegmatic
personality also happens to have a spider phobia, this merely adds some

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
colour to his personality. In contrast, phobias tend to undermine the
possession of character. This is because such irrational tendencies threaten
a person’s capacity to maintain a consistent pattern of behaviour.
Possessing character makes a person truly reliable. Ideally speaking, the
person with character is even more reliable than someone with strong
personality traits. This is because a personality trait by itself gives a person
no reason to maintain consistency in the face of countervailing or under-
mining pressures. Indeed, from one point of view, personality is the mere
receptacle of biological or environmental influences, where those influ-
ences could just as easily start pushing you in a different direction. In
contrast, the person with character deliberately commits to certain patterns
of behaviour, and this commitment demands that they resist conflicting
pressures. Of course, the person with character is not insensitive to
circumstantial factors. The point is that they should not be made to
neglect their commitments under the manipulation of such factors. In
his discussion of the ‘sovereign individual’ Nietzsche expresses the same
basic point (in typically hyperbolic fashion) in terms of the capacity to
make promises:

[This individual] has his own independent, protracted will and the right to
make promises—and in him a proud consciousness, quivering in every
muscle, of what has at length been achieved and become flesh in him, a
consciousness of his own power and freedom, a sensation of mankind come
to completion . . . The ‘free’ man, the possessor of a protracted and
unbreakable will, also possesses his measure of value: looking out upon
others from himself, he honors or he despises; and just as he is bound to
honor his peers, the strong and reliable (those with the right to make
promises) that is, all those who promise like sovereigns, reluctantly, rarely,
slowly, who are chary of trusting, whose trust is a mark of distinction, who
give their word as something that can be relied on because they know
themselves strong enough to maintain it in the face of accidents, even ‘in
the face of fate’—he is bound to reserve a kick for the feeble windbags who
promise without the right to do so, and a rod for the liar who breaks his
word even at the moment he utters it. (/, second essay, section )

It is because of the demand for consistency that observations about the

susceptibility of individuals to trivial situational influences are much more
damaging for claims about character than they are about claims for
personality. Consider, for example, the notorious experiment conducted
by Darley and Batson () in which seminarians were observed to be
less likely to stop and help someone who had apparently collapsed if they
were in a hurry to give a presentation (ironically, on the parable of the

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
good Samaritan). We can suppose that, given their chosen profession, the
seminarians would express strong commitments to the value of helping
others in need. But most of them did not act in accordance with this value
when under a relatively mild situational pressure. So we can say that those
subjects who did not stop to help lacked character. If they truly valued
helping others (above for instance, their own reputation) then they should
have striven to ensure that they consistently behaved in this way, despite
the situational pressure.
Indeed, John Doris (: ) argues that we cannot confidently
identify character unless it is manifested in circumstances that are not
conducive to its appearance. That is, anyone could be expected to save a
drowning child given the encouragement of others, average swimming
ability and the lack of competing goals. We would condemn them if they
did not try to do so. Someone with character however, acts in accordance
with his or her commitments where there is the smallest opportunity, so
long as it is not blocked by another, stronger commitment.
I believe the sceptic about character is right to demand a reasonable
degree of robustness. To truly possess character, one should show greater
resistance to relatively unimportant situational factors than is frequently
shown in the situationist experiments. However the average person seems
only rarely to resist the pressures put upon them, such as regulating fear
when the prospect of defending a concern bears a significant risk. This is
why, as Doris claims, we can best detect character when it is manifested in
nonconducive situations. To resist such influences is the clearest sign of a
powerful internal drive. Joel Kupperman defines strong character in similar
terms:
X has a strong character if and only if X’s normal pattern of thought and
action, especially in relation to matters affecting the happiness of others or
of X (and most especially in relation to moral choices), is strongly resistant
to pressures, temptation, difficulties, and to the insistent expectations of
others. (: ).

In line with this approach, we talk of character building as a result of enduring

hardships, cold showers and public school. But I do not think that suffering
straightforwardly builds character. What it does is reveal character. If one
comes through suffering having retained certain values, then one has shown
that those values are less vulnerable to the vicissitudes of circumstance.
Alternatively, if one comes through suffering and acquires a totally new set
of values, i.e. to do whatever it takes to avoid such suffering in the future, then
one’s character has been broken or was never there to start with.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
Overall, the demand for consistency in character explains why the
paradigmatic character traits all involve resistance or resilience of some
kind. To be temperate is to resist temptation, to be brave is to resist fear, to
be self-reliant is to resist the influence of other people, and to be respon-
sible is to keep your promises. Such traits are not simple emotional
dispositions. Rather, they reflect how one’s concerns stand in relation to
each other. To maintain one’s commitment to a concern in the face of
pressures requires that one’s pursuit of that concern will not be over-
powered by the call of other concerns. This entails that the priority of
that concern is firmly established within one’s mental economy. That is,
character demands that one have a definite hierarchy of concerns.

. Sentiments
The foregoing analysis suggested that a hierarchy of concerns lies at the
heart of character. Moreover, the establishment and maintenance of the
hierarchy is deliberate. One displays consistent patterns of feeling and
behaviour because one chooses to do so. Character is the apotheosis of
personal autonomy.
Aristotle already made a relevant distinction between personality and
character in this respect. He recognised that we may have inborn or natural
inclinations towards certain virtues. Just as dogs show a natural disposition
towards loyalty, so different humans are naturally predisposed towards
different attitudes. Some of us will incline towards caution where others
are more reckless. This is personality. However, as we acquire wisdom, we
calibrate these natural tendencies, such that we display bravery when it is
appropriate to do so and not out of the blind manifestation of a disposi-
tional impulse. This is character (Aristotle  BCE/: b; cf.
Wolf ).
The question now is what makes one choose to be brave? How does one
detect the appropriate circumstances for bravery or the other virtues? Articu-
lating a response to these questions is my goal for the rest of this chapter.
My strategy is once again to build upon the general model of affective
states that I have developed in this book. In particular, I want to draw on
our emotional attitudes of love and hatred. These attitudes seem powerful
enough to deliver the regulative force of character. I shall group together
varieties of love and hatred under the term ‘sentiment’ (cf. Deonna &
Teroni ). The key distinctive feature of sentiments, as I understand
them, is that they are directed at specific individuals enduring through
time, such as a person, an institution or a social cause.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
In English, our concepts of love and hatred are associated with certain
specific feelings. In the case of love, there is a distinctive feeling of devotion
towards one’s beloved – a sense of wanting to indefinitely sustain or
deepen one’s connection with them. Meanwhile, in the case of hatred,
there is the yearning to destroy the hated object. However, love and hate
are much broader attitudes than any single episodic emotion or even a
standing disposition for an episodic emotion. In the first place, love and
hatred are revealed in other emotions serving one’s attachment. For
instance, in the case of love, we feel jealousy where our attachment is
threatened and grief where our beloved is irretrievably lost. In addition to
such powerful feelings, to treat the object of a sentiment as an individual in
their own right is to be receptive to a variety of emotions felt on their
behalf. That is, we treat them as having concerns of their own. In the case
of love, anything boosting the status of the object is treated as an attract-
ant, and anything undermining its status is treated as an avoidant. Thus,
the individual tends to feel happiness when the status of the object is
promoted, sadness when it is undermined, fear when it is threatened, and
relief when the threat is avoided and so on. Meanwhile, hatred will involve
a reverse pattern of emotions.
The key reason to appeal to sentiments in the construction of character
is that sentiments can make sense of the normative force of character. This
is because the objects of sentiments are the sorts of things that can demand
consistent patterns of feeling and behaviour from us. In the case of love for
another person, the demand for consistency may be manifested in the
literal demand from the other to respect the attachment. But even where
one loves an institution or a political cause, these objects have a way of
making demands upon us, even if it is just a matter of the rules imposed by
the collective membership. Thus, to maintain a sentiment is to respond to
demands that one ought to regulate one’s emotions or behaviours in ways
that preserve the sentiment. So you ought to be loyal and expose yourself
to potential risks, because you love your partner. You ought to be brave
because you hate Nazis.
In an ideal case, of course, love is not just one-way. Where love is
reciprocated, both parties are disposed to synchronise their emotional states.
One feels happy when the other is happy, sad when the other is sad and so
on (this reiterates the points about intimacy I made in Section .). Thus, as


Especially in the first couple of years. This is sometimes called ‘limerence’ (Tennov ).

And neither is hatred! I think there is much more to be said about the role of enemies in the
development of character, but I will regretfully leave this issue aside here.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
Robert Solomon claims, love is ‘an interpersonal emotional dance’
(: ). Such synchronisation is the most powerful means towards
sustaining intimacy, and relationships falter when synchronisation is lost
(cf. Kolodny ).
My claim, then, is that character is constituted by the maintenance of
one’s sentiments. In particular, I want to emphasise that developing
character is not about committing to generic virtues. One is not brave
merely for the sake of being brave. That would be pointless. Rather, one
is brave for sake of defending one’s family, some particular cause, or a
cherished institution. Thus, we have a psychologically plausible way to
explain where the distinctive normative features of character come from,
building on the emotional faculties that I have so far described in this
book.
To fully recognise the normative demand of sentiments, is it also
important to recognise how they are sustained and develop over time,
often over many years. Sentiments generate emotions that display continu-
ity across their episodic instances. Examples include a sustained resentment
towards one’s boss or gratitude towards one’s parents. In such cases, there
is a prevailing background attitude towards the object that is renewed by
episodic emotions when the object comes to mind. Each time this occurs,
one may find new supporting considerations, implications or qualifying
nuances. This updates the prevailing attitudes and influences the way in
which future episodic emotions will be triggered.
Peter Goldie is one of the philosophers who has shown most interest in
emotions developing over the long term. This fits his general narrative
account of emotions. For instance, in a discussion of grief (), Goldie
notes that this emotion presents particular problems for accounts that try
to identify one essential component to emotional states such as bodily
feelings or cognitive appraisals. The problem is that the same emotion of
grief can include all sorts of different feelings and thoughts, and none of
these particular aspects is essential at any particular time (this coheres with
the popular idea that grief has distinctive stages such as denial, anger and
bargaining). Given that the grief will, on occasion, be interrupted by


This is not to deny that we might sometimes engage in the self-conscious pursuit of certain virtues.
We can develop a sense of honour. Yet this fits my claim about loves and hates since the driving force
here is self-love. That is, we become committed to our selves as objects of sentiment. What I want to
emphasise, however, is that while self-love can be a source of character, it is not necessary for
character.

Starkey () makes the related point that emotions combat ‘axiological entropy’; renewing our
sense of the importance of certain values.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
completely different matters, how are we to say that a single emotion
endures across these occasions? Goldie’s response is that we should not
privilege any particular mental state in our characterisation of grief. Rather,
we should conceive grief as an emergent process, where different stages are
tied together by a developmental arc or narrative.
I doubt Goldie’s claim that there is no single element underlying grief.
On the contrary, I think we can say that the process of grief is sustained by
the underlying concern for the dead individual. Thus, we can still make
sense of such cases within the control theory of affective states. Grief is the
attempt to regulate a certain kind of loss. It is so powerfully sustained
because the object of grief is both highly valued and cannot be regained.
The emergent process of grief that Goldie describes is actually the product
of the kind of long-term adaptation that I described earlier with regards to
the shaping of personality. That is, the different stages of grief correspond
to shifts in regulation strategy. Eventually, the grieving individual comes to
reconcile himself with the permanent absence of the beloved – at least to
the point where his responses are not constantly seeking to re-establish the
prior levels of availability of his beloved. But even then, the demands of
sentiment can persist. Grief can often be renewed in its later stages by the
guilt that one no longer feels so sad. One feels like one ought to feel as sad
as possible about the loss of the loved one, because it otherwise seems to
betray one’s attachment.
The really important feature of grief, the feature that makes it so hard to
regulate, is that its object is not replaceable. This is the case for all
sentiments. You can’t just find another person to love, and then equilib-
rium will once more be restored. De Sousa (: –) similarly
identifies this phenomenon when discussing the novel, Solaris, by Stanis-
law Lem. In the book, a mysterious planet replicates a man’s dead wife,
with all her memories apparently in place, but clearly fake. The protagonist
reacts with horror, and we can sympathise with his attitude. The replicated
wife just is not the woman he loved.
But why it is reasonable not to be satisfied with the duplicate? One
explanation is that we empathise with the individual we love to the extent
that we realise we would not ourselves want to be replaced with a
qualitatively identical copy. But this is a relatively superficial reply. We
already know that we treat the object of love as valuable in its own right,
just as we treat ourselves as valuable in our own right. A deeper explanation
is that sentiments latch onto specific individuals by grasping that individ-
ual’s history. That is, to represent an object’s history is to treat it as a
concrete particular, persisting through time and change. This is what

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
allows us to explain our dissatisfaction with the copy. When we value
individuals, we value them for their specific histories. This strikes us in a
cumulative sense. All those happy occasions, all those sad occasions, all the
times you endured together are felt to remain in the present. The copy has
none of that. We have no history with it (cf. Kolodny ).
Overall, sentiments are enduring attitudes that we bear towards specific
individuals that coordinate a broad pattern of emotions that are sustained
and developed over long periods of time. It is the enduring demands of
sentiments that can explain the normative demand for consistency in
character. To possess character is to maintain successfully one or more
sentiments.

. Narratives
In the previous section, I offered an account of character as constituted by
the maintenance of sentiments. However, I have yet to identify the
psychological capacities underpinning the maintenance of sentiments.
The central issue concerns how we manage to represent specific individuals
as unique beings persisting through time. Only once we can do this can we
develop attachments to such individuals.
My claim is that to track the unique history of an object or person
requires that we construct a narrative about it. To develop this view I want
to adopt an account of narrative provided by Peyton McElroy ().
McElroy argues that the essential feature of narratives is that they track the
development of an aim or purpose across a set of events. This contrasts
with other definitions of narrative that appeal to causal connections. Such
accounts seem vulnerable to attacks on both their necessity and sufficiency.
First causal connections are insufficient because scientifically described
chains of events do not often count as narratives. Meanwhile, events that
bear no causal connection can sometimes form a narrative. For example,
Velleman () relates an example of Aristotle’s, in which the murderer
of King Mitys was years later killed by a falling statue of that very same
king. There is no causal connection between these events, yet we recognise
an ironic resolution in them. McElroy argues that the case counts as a
narrative because we track how the murderer’s project of getting away with
murder was foiled.


They are, however, ways of constructing narratives about natural processes, such as the growth of a
flower or the water cycle. McElroy argues that in such cases we project onto the object a teleology or
purpose.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
Narrative construction is the key new representational capacity that
allows us to develop sentiments towards specific individuals. In accord-
ance with McElroy’s account, narratives about the objects of sentiments
will focus on the development of aims or purposes. Thus, the narrative of
one’s beloved will trace the continuity of their projects and aims. An
institution or cause likewise has its defining projects or aims. Meanwhile,
we see that tracing the narrative of the object of a sentiment indirectly
generates a narrative about oneself. That is, an institution or cause has an
aim that I endorse or oppose. Similarly, an individual to whom I am
attached often has a reciprocal attachment to me. In this way, by tracing
the history of the things that I love or hate, I come, indirectly, to a sense of
who I am, or what I am about. That is, I am the kind of person that loves
these things or these people. The sum of my narratives about these things
is my self-narrative. This is what Alasdair MacIntyre is getting at when he
claims that we are never more than co-authors in our own lives (: ).
We recognise ourselves in the attachments we bear to others.
To elaborate on the capacity for narrative construction, we can supply
two general principles: a principle of selectivity and a principle of conflict.
The principle of selectivity is that we make a series of events intelligible by
selecting only those events or aspects of events that are relevant to the
development of the aim or purpose. Thus, many details about a person’s
life, such as what they had for breakfast or what colour socks they were
wearing will be quite irrelevant. Irrelevant details recede into the back-
ground. Meanwhile, we bring to the fore those events that seem to make
an important lasting difference. Thus if someone were to construct a


De Sousa suggests that singular reference is a necessary condition for fully fledged intentionality
(: ). Certainly, it is the representational capacity that caps our explorations of the
affective life.

Again, in reference to the point made previously in footnote seven, we may come to explicitly
recognise ourselves as objects of sentiment that we track via a narrative. Narratives have long been
implicated in constructionist theories of the self and personal identity. I will, however, limit my
claims to the role of narrative in developing character. It seems to me that most of the claims that
have been made by philosophers about the narrative self (e.g. MacIntyre ; Kristjánsson ;
Schechtman ) apply equally well to the notion of character, without making metaphysically
dubious commitments. That is, I could survive the destruction or radical transformation of my
character. However, it may be fair to say that my character is what really matters to me (in the sense
developed by Parfit ).

These principles are shared with fictional narratives. However, it is worth emphasising that
character-defining self-narratives display some important differences with fictional narratives.
I review some of these in Cochrane (). Bernard Williams also makes the important point
that while fictional narratives are (in virtually all cases) complete and closed, personal narratives are
incomplete and remain open. ‘It is essential to fictional lives that their wholeness is always already
there, and essential to ours that it is not’ (: ).

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
narrative about my commitment to philosophy, they might prioritise those
occasions in my childhood in which I showed some philosophical promise
or the way I ended up rejecting religious belief as a teenager. In many cases it
may be fair to say that the significance of events is magnified beyond the
significance they showed at the time. My childhood thoughts were not
particularly profound. But they were the first tentative steps on a longer
journey, and that lends them greater significance.
The principle of conflict, meanwhile, is basically a statement of the same
point that was made earlier about the revelation of character via struggle.
As Dewey notes, one could hardly become aware of a distinct agent
without a sense of the resistance of the environment to that agent’s goals
(/: –). Similarly, any narrative about an agent will tend to
highlight the barriers that they faced and whether they managed to
overcome them. Interpersonal conflicts in particular are helpful for
revealing how an agent’s commitments contrast with the commitments
of others. Thus a narrative may emphasise how the demands of their
parents were rejected or how they proved themselves, despite the low
expectations of a teacher.

. Character in Charge

The final issue I wish to address in this chapter concerns the sense in which
character plays a definite hierarchical role in our affective lives. First of all,
it is important to recognise that one does not simply decide one day to
form the character-constituting sentiments that one has. One is often
forced to recognise an enduring sentiment because of the insistence of
lower-level emotional responses. For example, one finds oneself constantly
distressed by news reports about refugees until one comes to realise just
how important this issue is to you. Or, one discovers how gripped one is
by philosophical questions and excited by new ideas. In this way, a
sentiment comes to be gradually revealed as one realises narrative connec-
tions between events.
Similarly, while sentiments function in a rather different way to person-
ality traits, the object of a sentiment is likely to be suited to one’s
personality. That is, one is the sort of person that is disposed to be satisfied
by the qualities of the beloved. We see this particularly where one develops
a sentiment towards a political cause. Such a cause may well represent a
world view that satisfies all kinds of lower-level concerns.
The formation of sentiments is receptive to our concerns and capacities
and thus displays a bottom-up influence. However, I think it is fair to say that

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
all control functions are sensitive to lower-level activity. The point is that
higher-control functions impose a new level of organisation on those lower-
level activities and then directs these activities on a larger scale. For example,
while personality influences sentiment, so sentiments can also influence our
personality traits in return. This is because, over the long term, a sentiment
can solidify a whole set of emotional dispositions by pushing us to regulate
our behaviours in a way that is compatible with our commitments, ultim-
ately leading to the development of habitual strategies of concern-regulation.
We should also consider the relationship between character and the
capacities for conscious thought that were considered in the previous chap-
ter. Recall that I identified three principle areas in which conscious thought
elaborated emotional control: () in the reappraisal of one’s situation and the
discernment of new emotional triggers by means of rational inference, () in
the formation of plans to deal with one’s concerns, and () in inferring the
conflict between one’s emotion and some normative standard, thereby
inviting emotional regulation. With respect to each of those three areas,
we can discern ways in which character further refines control.
First, with regards to reappraisal, one of the points that is frequently
made about character is that it is manifested in the way one pays attention
to situations (e.g. Goldie : ; Deonna & Teroni ; Webber
). The kind person is more apt to notice when the opportunity for
kindness arises; to see the right thing to do, as Goldie puts it. The role of
character here seems to be mediated by narrative construction. Discerning
a narrative that reveals one’s commitments will tend to make one more
consciously aware of situations that reflect that commitment.
Second, the formation of plans bears a clear connection with the
construction of character narratives. Our basic capacity for forming narra-
tives may indeed build upon the skill of forming plans, requiring us to
project our lives into the future, and later, monitoring how well our plans
are proceeding. In return, we evaluate the compatibility between our plans
and our longer-term life narratives. Most significantly, having discerned
the connection between a series of events as the development of an aim or
purpose, one may then form a self-conscious plan to live up to it (Velle-
man : chapter ; cf. Solomon ). Thus one chooses a career, or
marries a particular person or joins a community of like-minded fellows.
Of course, this activity will have reciprocal effects on one’s later experi-
ences, which one must incorporate into the developing narrative. In this
way, generating a narrative is a way of exploring one’s individual procliv-
ities. One forms a hypothesis about the kind of person he or she is, and
tests this by trying to live up to it.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
 Personality and Character 
Finally, with regards to emotional regulation, character clearly sets up
normative standards for us to adhere to. One does not simply accept all the
normative standards of one’s community. One evaluates to what extent an
emotional state fits with the pursuit of one’s deeply held commitments.
This may well entail the qualification or even outright rejection of stand-
ards derived from one’s community.
All of these considerations relate to the idea that character increases self-
consciousness. Goldie similarly recommends circumspection – caution and
attentiveness with regards to when one’s values may be implicated – as an
‘executive virtue’ (: chapter ). Goldie is responding to the situationist
evidence that we can be manipulated by situational factors. Indeed, given
what we now know about these sneaky effects, it is up to the person with
character commitments to ensure that their commitments are not under-
mined by manipulative influences.
However, one of the classic Aristotelian claims about character is that
when it is fully realised, one no longer needs to make a special effort to
conform to one’s commitments. A truly kind person should not have to
struggle to notice opportunities for kindness just as they should not be
tempted by cruel desires. Thus, self-reflection may seem like too much
effort for the person who has firmly established character.
It seems to me that this demand upon character is too strong. Certainly
it makes sense that the repeated activity involved in committing to a value
means that one will develop habits of attention and thought. However, it
will only be in people with a very steep hierarchy of drives, who have some
all-consuming master passion that overwhelms everything else, that the call
of other values will not need to be monitored and managed. It is entirely
consistent with the possession of strong character that one have a flatter or
more pluralistic set of commitments. It is not even inconsistent with
character that one’s commitments are in certain respects in tension with
each other. Thus it seems to me that attentive self-reflection will be
stimulated wherever one feels the pull of different values, and that it will
take one’s most refined understanding of the particularities of the situation
to find the best balance.

. Conclusion
In this chapter, I have analysed personality and character as primarily ways
in which we come to prioritise the pursuit of some concerns over others.
Personality traits are analysed as mutually reinforcing combinations of
capacities and values that exist at all levels of the individual’s affective life.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008
  The Emotional Mind
Character, in contrast, is conceived as a distinct level of regulative control,
going beyond the rational regulation of one’s actions or emotions.
Character is distinctly driven by the things that we love and hate – the
sentiments we form towards specific individuals, projects and causes.
Forming such sentiments, in turn, relies on the construction of narratives,
which allow us to pick out individuals, projects and causes uniquely by
means of their histories. By means of such narratives, we can link up
episodes stretching across decades of an individual’s life span. Thus the
drive to maintain a sentimental commitment may well be the deepest
explanatory principle for an individual’s behaviour over time.

Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 05 Jan 2019 at 22:02:45, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108579056.008