Journal for the Theory of Social Behaviour 37:3

0021–8308

Measuring Self-Respect
Kristján
Original
Measuring
Blackwell
Oxford,
Journal
JTSB
©
0021-8308
XXX Kristjánsson
2007 TheArticles
UK
for Self-Respect
Publishing
the
Author
Theory
Journal
Ltd
of Social
compilation
Behaviour
© The Executive Management Committee/Blackwell Publishing Ltd. 2007

KRISTJÁN KRISTJÁNSSON

I. SELF-RESPECT IN PLACE OF SELF-ESTEEM?

Can “self-respect’
ABSTRACT
main components supplant
of Aristotelian
the now
self-respect
much-maligned
are then
“global
workedself-esteem’
out. The paper
in psychological
concludes byresearch
exploring
and
how,
therapy?
in orderThe
to make
aim ofthose
the present
components
paperobjectively
is to examine
measurable,
this suggestion
certainand
methodological
develop it further.
pitfallsItmust
is argued
be avoided.
that there are two distinct philosophical concepts of self-respect abroad in the literature, Kantian and Aristotelian, between which psychologists need to choose. The

Self-esteem was, not so long ago, the Holy Grail of much psychological and educa-
tional research, hailed both as an essential social vaccine and as a panacea for an
array of personal and educational problems. Every flow has its ebb, however, and
in recent years we have seen the ideal of (indiscriminately) boosted self-esteem
come under sustained attack from various quarters.
Psychologists grumble that the expected correlations between high self-esteem
and salient positively valenced factors, such as above average school achievement
and prosocial behavior, have failed to materialize in empirical research. Instead,
null findings abound and, if anything, there seems to be a link between high
self-esteem and various types of risky and antisocial behavior (Baumeister et al.,
2003; cf. Emler, 2001). Educationists point out that if self-esteem is understood,
in line with the typical psychological instruments, simply as satisfaction with
the global ratio of one’s achievements to one’s aspirations, then the easiest way to
enhance self-esteem is by diminishing the aspirations: dulling educational
standards and dumbing down the curriculum. The catch there, however, is that
although the house of such (deluded) educational self-esteem may be easy to
build, it is draughty to live in and liable to fall once students enter the “real” (read:
meritocratic) outside world (Damon, 1995; Stout, 2000). Finally, philosophers
have long looked upon the social-science notion of global self-esteem with
suspicion. It appears to be an artificial construct with little, if any, grounding in
our rich ordinary-language repertoire of “self ”-conceptions, ranging from arro-
gance and boastfulness to meekness and diffidence (Smith, 2002; Cigman, 2004).
Moreover—many ethicists complain—the concept of self-esteem lacks an objective
moral grounding. Given that what self-esteem instruments measure is simply
subjective self-reported satisfaction, amongst those individuals epitomizing
high self-esteem may be the big-headed bully, the smug drug baron, and the
Machiavellian tyrant (cf. Kristjánsson, 2006a).
If self-esteem’s star has faded of late, it is salutary to ask what could replace it.
It seems counter-intuitive to suppose that people’s self-concept has no bearing
on their achievement in life, for example how well they do at school or in the

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007. Published by Blackwell
Publishing, 9600 Garsington Road, Oxford, OX4 2DQ UK and 350 Main Street, Malden, MA 02148, USA.
226 Kristján Kristjánsson
workplace. If self-concept is simply equated with global self-esteem, as has often
been done in the psychological literature of late, then we are stuck with the
counter-intuitive implication. If, however, we understand self-concept as the total-
ity of an individual’s attitudes towards himself or herself, involving a number of
distinct aspects or dimensions, then various candidates for research other than
global self-esteem may emerge. It could well be that domain-specific self-esteem—
esteem predicated upon a specific domain of one’s life rather than that life as seen
from a global perspective—matters (one’s self-esteem as a philosophy student for
one’s philosophy grades, for instance). Another candidate is self-confidence, global
or specific. Bandura (1997: Ch. 3) has thus argued convincingly that students’
“perceived self-efficacy” (a type of domain-specific self-confidence) substantially
influences their likelihood of accomplishing tasks. Yet one more candidate for
research is self-respect. Indeed, Roland and Foxx argue, in a suggestive and astute
article (2003), that the star of self-esteem in psychological and educational research
could reasonably be eclipsed by that of self-respect. The aim of the present paper
is to gauge the aptness of that suggestion, to ameliorate certain shortcomings to
be found in it, and to develop it further. In order to do so, I need to respond to
two competing pulls: the philosophical one of elucidating the notion of self-respect
and the psychological one of suggesting ways for conducting empirical research
into self-respect. However tricky it is to respond adequately to both pulls in a single
article, it is worth a try, given the interdisciplinary focus of the present journal.
What is at issue in Roland and Foxx’s article is quite a radical suggestion; just
how radical it is can be seen by comparing the respective number of abstracts
yielded by PsycINFO (1985–May 2006) when using the search terms “self-esteem”
and “self-respect”. While the former results in no less than 11,313 abstracts, the
latter turns up only 239. To put it bluntly, self-respect has not exactly, so far at
least, riveted the attention of psychologists; moreover—tellingly—few if any of
those 239 abstracts point to any empirical research into self-respect. In their
article, Roland and Foxx claim to have found no “scientific studies” of self-respect
(2003: 279). To be sure, showing that self-respect provides a lean counterpart to
self-esteem in recent psychological research is not indicative of the futility or
dispensability of research into the former; the current research focus might simply
be wrong. That is, more or less, what Roland and Foxx argue: now that the
“self-esteem fallacy” has been exposed, it is time to explore “the couch on which
the cushion of self-esteem resides.” More specifically, they propose that “a relationship
exists between self-respect and self-esteem such that self-respecting individuals
may experience either high or low levels of self-esteem and individuals with high
levels of self-esteem may or may not possess self-respect.” They hypothesize that
this could explain, inter alia, why people reporting low levels of self-esteem often
fail to exhibit the expected psychological or social dysfunction: namely, that such
people have a secure seat on their “couch” (the identity-conferring moral principles
to which they aspire) although the “cushion” (the satisfaction with their attainment)
happens to be missing (2003: 247, 268, 271).

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 227
Returning to the typical characterization of self-esteem in psychological
research as one’s level of satisfaction with the global ratio of one’s achievements
to aspirations, a shift of research focus from self-esteem to self-respect would
mean that closer attention be paid to the nature and content of the relevant
“aspirations”—or “pretensions” as the father of the notion of global self-esteem,
William James, used to call them (1950, Ch. 10)—but less to the reported
subjective satisfaction. Any remaining concern with the “ratio” part would focus
on domain-specific, rather than global, self-esteem: that is, exclusively on that
type of self-esteem which is predicated upon self-respect. How happy is the person
with his or her attainment of self-respect-grounding aspirations? Philosophers
have termed this particular type of self-esteem “appraisal self-respect.”
So, having forsaken global self-esteem, should social science turn its empirical
compass to self-respect and (possibly) appraisal self-respect? I assume that many
educationists would take well to that suggestion. It seems more intuitively plausible
that there exists a positive correlation—and even a causal connection—between
self-respect and school achievement than between global self-esteem and school
achievement. At any rate it is, I believe, a common view amongst academics
and laypeople alike that self-respect imparts in us the general zest necessary for
pursuing our life plans. Perhaps high self-respect can help keep students focused
and working hard, inducing them not to let their talents lie fallow. It has been
suggested, moreover, that self-respect is easier to achieve than many other educa-
tional competences, improving one’s maths score for instance (Nesbitt, 1993).
Philosophers will hardly protest either at such a change of compass. After all,
self-respect has a more secure grounding in ordinary parlance than does global
self-esteem (at least pre-Oprah Winfrey and the self-esteem industry’s colonization
of the public media) and can, additionally, draw on long-running discursive
traditions in the philosophical literature. Furthermore, self-respect may provide
precisely what self-esteem sadly lacks from a moral point of view: namely, an
objective basis for, or true measure of, moral worth. It is, after all, common in the
philosophical literature to see self-respect referred to as the guardian of the other
moral virtues: the column of true majesty in human beings which preserves com-
mendable character traits and contributes to the continuation of morality. Roland
and Foxx specifically mention that “objectivity is what differentiates self-respect
from self-esteem” and that it is the presence of a “moral code” in self-respect
which provides its objective components (2003: 269, 282).
The people having to carry out the relevant measurements in the end are,
however, empirical psychologists and I suspect that many of them would not be as
enamored of Roland and Foxx’s suggestion as their colleagues in philosophy and
education. Let me just touch here on some of the doubts that psychologists might
entertain. First, self-respect is a moral concept and, as Roland and Foxx note
correctly, many psychologists are made uneasy by issues of morality, to the point
even of being averse to the very word “morality,” viewing it as blatantly value-
laden with connotations of a holier-than-thou attitude. The chasm between social

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
228 Kristján Kristjánsson
scientists who (allegedly) trade only in the descriptive and moral philosophers
who deal with the normative is, after all, long-standing and well-entrenched.
Nevertheless, two simple observations may allay some of doubts of empirical
psychologists in this regard and persuade them to tread on the moral terrain with
less trepidation.
The first observation is that normativity does not necessarily entail relativity.
Judgments such as “It is good to teach children honesty” or “It is morally wrong
to sexually molest children” do not so much evaluate the world of description as
describe the world of evaluation. Although they are value-laden, they are neither
inherently relative (to place or time), nor merely subjectively true. This much at
least will be acknowledged by those who subscribe to some form of moral
naturalism, whether Aristotelian eudaimonism, or a biological evolutionary perspect-
ive on morality such as that to which Roland & Foxx allude (2003: 280). Roland
and Foxx also note helpfully the existence of ethnographic evidence for the uni-
versality of prosocial behavior (280). The second observation is that normativity
does not necessarily entail prescription. Doing research based on a moral concept
and possibly finding correlations between its normative components and posit-
ively or negatively valenced psycho-social variables does not mean that empirical
psychologists have turned themselves into moralists: have started to prescribe
rather than describe. Describing a positive outcome is not the same as prescribing
it. For example, even if the normative components of self-respect turned out to
have positively valenced correlates, such findings would not by themselves
prescribe those components, except to persons who wanted to be moral and to
embrace those components in the first place. I will leave the “we-are-not-
moralists”-qualm out of further scrutiny in what follows.
The second doubt entertained by psychologists, at least those of a more dis-
cerning kind, concerns whether or not the notion of self-respect could be opera-
tionalized in such a way that a non-logical, and hence non-trivial, correlation
could be found between it and salient psychological and social variables such as
school achievement, prosocial behavior, and so forth. Almost the only robust
correlations that have been found between global self-esteem and other relevant
variables have been between global self-esteem, on the one hand, and happiness
and depression, on the other, with high self-esteem being linked to happiness and
low self-esteem to depression (Baumeister et al., 2003). What many psychologists
have failed to notice is that these correlations are most likely logical since the
typical measures of global self-esteem zoom in on the same characteristics, at least
in part, as do measures of (self-reported subjective) happiness and depression.
Measures of self-respect must avoid the same pitfall: that is, the measured com-
ponents of self-respect must not include those very characteristics (say, diligence
in school work) that inform the other variables to be explored. Indeed, empirical
research into self-respect can scarcely get off the ground without a careful
and critical analysis of its components—which will surely mean some notable
tightening of our ordinary language notion of “self-respect.”

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 229
The third possible doubt would relate to the question of whether or not there
is a single concept of self-respect abroad in ordinary language and the philosoph-
ical literature, or whether there are not perhaps two or more overlapping or even
radically distinct concepts. In case there is more than one concept at stake, the
preliminary task will have to be the selection of the one that is to form the basis
of the empirical research and to give good reasons for that choice.
I would not want to be understood to be mired in a positivistic conception
of methodology which refuses to accept any construct as “scientific” that cannot
be operationalized and measured. Obviously, there are established qualitative
research paradigms within social science which would not require prior consensus
on the meaning of the research term and would not aspire to “measure it” in
any way. Consider, for example, possible phenomenological studies into people’s
understandings of self-respect, or case studies of exemplary individuals who have
been widely thought to possess self-respect (cf. Colby & Damon, 1992). However,
I understand Roland and Foxx’s suggestion about replacing self-esteem with
self-respect in psychological research to mean replacing it within the discursive
field where self-esteem has been studied most assiduously. That discursive
field happens to be a quantitative one. In such a field, the best reason for psycho-
logical investigation of a variable is considered to be the expectation that interesting
and illuminating correlations—and better still, causal connection—will be found
between that variable and others, whereas an important scientific reason for psycho-
logists not investigating a variable is that it cannot be adequately operationalized.
Roland and Foxx have done well to spark attention to the salience of self-
respect and to suggest it as a possible topic of exploration after the slaying of the
self-esteem dragon. Pointing to self-respect in that way, however, has merely the
status of a promissory note. To establish whether or not the note can be paid off,
the second and third “doubts” introduced above must be subjected to a more
thorough analysis than the two authors undertake in their article. I will try to do
so—in reverse order—in the following two sections.

II. KANTIAN OR ARISTOTELIAN SELF-RESPECT?

A quick and superficial glance at what contemporary philosophers have written

about self-respect may create the impression that they are all contributions to a
unitary discursive tradition about a single concept. Philosophers thus seem to
agree that self-respect is a complex character state involving a disposition not to
behave in a manner unworthy of oneself: that is, a disposition to shun behavior
that one views as contemptible, degrading, or otherwise immoral. A person with
a sense of self-respect identifies with a project, code, or status that provides a
standard of worthy conduct, a line past which one does not go. The person is
committed to this standard, confident that by and large they are the right
commitments, and tries to live accordingly (see, e.g., Kristjánsson, 2002, S. 3.1,

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
230 Kristján Kristjánsson
for references to the current literature). Philosophers also seem to agree, at least
implicitly, that there is no such thing as “domain-specific self-respect”: a person who
behaves self-respectfully within the family circle but un-self-respectfully outside
of it will simply be considered to lack self-respect. Self-respect is thus understood
holistically, as a global trait of one’s character as a whole (see, e.g., Russell, 2005:
104). Moral situationists, who argue on the basis of dubious psychological
experiments that human behavior is essentially context-dependent and lacks any
evaluative integrity (e.g., Doris, 2002), are at least right in the following: if our
behavior is as fickle and erratic as they claim, then there really is no such thing
as moral character, let alone self-respect.
When one probes more deeply into exactly what stands behind the relatively
converging formulations of self-respect, however, considerable divergence begins
to emerge. Roland and Foxx are aware of the fact that current discussions of self-
respect are grounded in different historical traditions, and they make a reasonable
attempt at grouping those discussions with respect to their originators and their
historical precedents, ranging from Aristotle, Aurelius and Augustine to Hume,
Kant and Nietzsche (2003: 248–258). Unfortunately, their taxonomical ground-
work seems to betray an inadequate grasp of just how radical some of the internal
divisions are. This can be divined from the fact that those divisions do not survive
to the latter and more substantive part of their article where they explore the
differences between self-esteem and self-respect and the use that psychologists
could make of the latter. Indeed, in this latter section the authors quietly ignore
their earlier rhetoric about different philosophical conceptions of self-respect and
clandestinely help themselves to one such conception, namely the Kantian one.
For instance, they claim that “cognition and the law of respect for persons would
be considered the predominant properties of self-respect” and that “self-regulation
and self-control is an integral component of a self-respect system” (269, 273);
moreover, in the very abstract of their article they say that “autonomy is central
to self-respect” (247). This is all well and good on a Kantian conception but it fits
an Aristotelian conception only tangentially, if at all.
After delving through the contemporary philosophical literature on self-respect,
I have become convinced that there are two main concepts of self-respect at work
there which can be characterized, broadly, as Kantian and Aristotelian. Notice
that I say “concepts” rather than “conceptions”; I do not believe that these are
simply contestable conceptions of the same concept, revolving around a common
core, but rather radically divergent notions. I hope that the reasons for this will
become evident during the course of the present section. Notice also that I say
“Kantian” and “Aristotelian,” rather than “Kant’s” and “Aristotle’s.” In recent years
we have witnessed various attempts to undermine or transcend the traditional
distinction between Kant as a moral formalist and Aristotle as a moral naturalist
by naturalizing (or de-formalizing) the former and formalizing (or de-naturalizing)
the latter (see, e.g., Korsgaard, 1996; Mcdowell, 1996). Although I think that
those interpretive maneuvers are ultimately misguided, this is not the place to

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 231
Table 1. Contrasts between the Kantian and the Aristotelian concepts of self-respect

Kantian self-respect Aristotelian self-respect

a) Moral basis of Dignity of persons as ends in Virtues as ends

self-respect themselves
b) Psychological processes Cognitive Cognitive and affective
involved
c) Relation to desires Self-control Moralization
d) Extension to others Members of the kingdom of ends Philia
e) Concern for others Their human rights Their eudaimonia
f ) Formation/maintenance Autonomous Heteronomous
of self-respect plus autonomous
g) Moral worth of persons Equal Unequal
h) Basis of morality Rational (formal) Substantive (non-formal)
i) Possible loss of self-respect Internal Internal or external
j) Excessive self-respect Impossible Possible

argue that point or to engage in textual exegesis. Rather, let us rely on the
traditional analyses distilled from Kant’s (1967) and Aristotle’s (1985) main ethical
works, and the views customarily trotted out under the banners of “Kantianism”
and “Aristotelianism,” respectively.
I first unpack in Table 1 what I take to be some of the fundamental differences
between the Kantian and the Aristotelian approach to self-respect, and then
expand on each difference. While this list is by no means exhaustive, the items on
it are not arbitrarily chosen. I have tried to single out those differences that
I consider most germane to the task of contrasting the two concepts. I will
leave it to readers to explore how and to what extent the lines of descent of
various contemporary accounts of self-respect can be traced to either the Kantian
or the Aristotelian approach. Readers will be helped in that pursuit by Roland
and Foxx’s categorizations (2003: 250–258); indeed, I hope that the following
comparisons will throw clearer light on what precisely it is in such modern
accounts as those of Thomas, Hill, Boxill, and Sachs that makes them Kantian,
and in those of Downie, Telfer, and Taylor that makes them Aristotelian
(see references in Roland & Foxx, 2003). For present purposes, the key aim is
to establish that there are actually two distinct concepts of self-respect at
issue, between which we need to choose before we can design psychological
measures of self-respect.

(a) Moral basis of self-respect. Personhood and its inherent dignity form the core of
Kantian moral philosophy. Persons are ends in themselves, in virtue of their
ability to rationalize, think and choose, and must respect themselves and others
as such. Meanwhile, in the Aristotelian system, persons’ eudaimonia (wellbeing)
forms the core of morality. Persons achieve eudaimonia, and come to deserve

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
232 Kristján Kristjánsson
dignity, insofar as they actualize the moral and intellectual virtues, which are at
once conducive to and constitutive of human wellbeing. Persons are entitled to
respect themselves and others not as ends in themselves, but only to the extent
that they have mastered the virtues, which are the real ends. Simply invoking one’s
personhood or humanity does not count. “Respect” (including “self-respect”) also
has a more earthbound, and less detached, understanding in the Aristotelian than
the Kantian model (see (b)).

(b) Psychological processes involved. According to the Kantian concept, conformity to

a pure rational principle (namely, his well-known categorical imperative) is essen-
tial to human agency and to morality. Emotions are, on the other hand, intruders
in the realm of reason. The psychological processes which guide true self-respect
are thus exclusively cognitive. In contrast to this, the assumption that emotional
dispositions may also constitute virtues is essential to Aristotle’s virtue theory.
Emotions can, no less than actions, have an intermediate and best condition
proper to virtue: that is, when they are felt at the right times, about the right
things, towards the right people, for the right end, and in the right way. Persons
can only be fully virtuous if they are disposed to experience emotions in this
medial way on a regular basis. Guiding their self-respect will not only be cognitive
processes and beliefs, but also emotions such as pride and a forward-looking sense
of shame which prevents them from engaging in behavior that may jeopardize
their self-respect. While those emotions have an important cognitive core, they
also include pure feelings: the non-intentional states of pain (lupe) or pleasure
(hedone).

(c) Relation to desires. Virtue is, in the Kantian model, about strength of will.
Reason must therefore constrain desires. It is not enough to try to sublimate and
purify the desires, for it would still remain possible that they might run amok and
counter to reason. The disposition to overcome, through an act of will, conative
obstacles to moral behavior is, for Kant, an ineradicable feature of human nature:
the self-respectful person is essentially the self-controlled person. In the Aristote-
lian system, desires belong to the irrational part of the soul. Policing them through
self-control is, however, only a second-best tack—a semi-virtue at most—for the
self-controlled person still has base desires. Truly virtuous persons have infused
their desires with reason, and hence moralized them. For instance, the “generous”
Kantians who force themselves to give money to the poor in deference to the
moral law (and whose moral virtues are not enhanced but rather compromised
by the presence of co-operating inclinations) only count as “resistant” or “con-
tinent” in the Aristotelian model. The truly generous Aristotelians give money to
the poor because they truly want to: they are emotionally disposed to do so. In
contrast to the bifurcated nature of Kantian agents, Aristotelian self-respectful
persons are essentially at one with themselves, as manifestations of their own
properly felt desires.

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 233
(d) Extension to others. Once Kantian agents have understood their uniqueness as
rational agents and the dignity that such a status necessarily involves, they extend
their respect for themselves to respect for other rational beings, belonging to the
same kingdom of ends, where each being must be treated as an end in itself and
never as a means to an end. This extension is again purely cognitive; there are
no emotions (such as compassion or kindness) involved. For Aristotle, it is an
empirical fact that the virtues are essential to our own good; they help us fulfill
what is central to us. Applying the virtues is therefore necessary to our own
interest. But the virtues require precisely that we pursue the good of others in the
ways required by morality; indeed, the greatest virtues are those most beneficial
to others. This is best brought out in Aristotle’s discussion of true self-love as
involving love of others, with the two developing together inseparably. Self-
respectful persons extend their respect to others through the process of philia: of
friendship and (non-erotic) love, which incorporate various other-regarding
emotions. Indeed, self-respect may turn out to be the very maturity that makes a
character capable of self-love and, hence, of loving others (see further in Russell,
2005: 119).

(e) Concern for others. After extending our self-respect to others, the Kantian con-
cern for others rests primarily on their human rights. As members of the kingdom
of ends, human beings are thus entitled to certain inalienable rights, and respect
for others means respecting those rights. The indebtedness of modern liberalism
to Kant is here obvious: the human good is defined in terms of the right,
which in this case leaves us with a relatively thin notion of the common good.
In Aristotle’s view, concern for others is expressed through our concern for their
eudaimonia. We want them to be able to actualize all their essential human capab-
ilities. Respect for others thus implies a thick conception of the common good, a
direct descendant of which is the moral position nowadays referred to as “the
capabilities view” (often connected to the works of Amartya Sen and Martha
Nussbaum, at least before the latter’s liberal turn a decade ago).

( f ) Formation/maintenance of self-respect. The Kantian rational will is essentially

autonomous or free in the sense that it is the original author of the law that binds
it (the categorical imperative) and also the constant protector of that same law.
It is not only negatively free (from external influence) but, more importantly,
positively free as its own author, and this freedom to choose its own law has
inherent worth. Autonomy is, thus, nothing less than the lifeblood of Kantian
self-respect—but, pace Roland and Foxx, not of all self-respect. Aristotle does not
possess a corresponding notion of “will,” nor does he have at his disposal a concept
of “autonomy” in a modern sense. The Aristotelian self is both derived from and
essentially sustained through social recognition and admiration, and to that extent
it is “heteronomous” in the Kantian sense. To be sure, fully self-respectful persons
have developed their own phronesis (practical wisdom), and the decisions they take

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
234 Kristján Kristjánsson
are, in that sense, “their own.” This does not make their choices intrinsically valu-
able, however, but only extrinsically so (insofar as they are morally right and conducive
to eudaimonia), and in any case, all choice is embedded: it takes place within a
framework of moral upbringing that is itself unchosen and non-autonomous.

(g) Moral worth of persons. For Kantians, all persons deserve equal respect and may
respect themselves equally, in virtue of their rationality, irrespective of their con-
tingent characteristics. This is a far cry from the Aristotelian model where people
are of unequal moral worth—and have unequally good reasons for respecting
themselves—based on their actual respective attainment of moral and intellectual
virtues. Immoral persons have thus no good reasons for respecting themselves
here and now, although they would have good reasons for improving themselves
in such ways that they would be worthy of self-respect and the respect of others
(provided they could fathom the need for such improvements). In the Kantian
model, the idea of a person’s worthlessness is inconceivable; for Aristotelians,
in contrast, there can be persons who have no moral worth (and even realize this
fact themselves).

(h) Basis of morality. As we have seen, both the Kantian and the Aristotelian con-
cepts of self-respect have their basis in morality. In that sense they concur, but
the problem is that morality is here understood in very different ways. For
Kantians, the basis of morality itself is the “good will,” the only thing good
without qualification, and good-willed actions are good because of the way they
are willed: that is, in virtue of the formal qualities of choice. For Aristotle, in
contrast, good-willed actions are good because they exemplify the virtues, many
of which are irreducibly moral. Kant is thus a moral formalist (rationalist) while
Aristotle is a moral substantivist (naturalist).

(i) Possible loss of self-respect. If Kantian agents lose their self-respect, it will be
solely for internal reasons for failing to exercise their own good will. Total loss of
self-respect is, however, impossible as long as they remain rational agents. On the
Aristotelian view, persons can lose their self-respect for internal reasons (negli-
gence, carelessness, or, simply, choosing to be bad), but also for external reasons.
Lack of what is nowadays known as “moral luck”—various externalities that are
necessary for the good life—can thus deprive us of the necessary conditions for
self-respect. Aristotle would be quick to point out, however, that these would have
to be extreme deprivations and, moreover, that persons can avoid becoming
totally miserable as long as they retain their equanimity in the face of adversity—
but this does not change the fact that total loss of self-respect is possible, in the
Aristotelian model, both for internal and for external reasons.

(j) Excessive self-respect. This notion is perplexing for Kantians—witness Sachs’s

refusal to accept this possibility (1981: 348–349): How can a person’s will be too

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 235
good? How can persons overestimate the worth of their rationality? For
Aristotelians, there is nothing perplexing about this notion. Persons who make
stronger demands on themselves than they can live up to as natural beings, who
set their moral stakes too high, have excessive self-respect. Self-respect, like any
other virtue, admits of a golden mean between excess and deficiency.
Obviously, the Kantian and Aristotelian concepts of self-respect are not diamet-
rical opposites. Both engage with the issue of human dignity and fortitude, and
both include a demand for public availability: the self-respectful person must not
hide under a bushel—not sit pretty on it, but rather do something with it.
Nevertheless, I hope that the above comparisons suffice to show that the Kantian
and Aristotelian concepts of self-respect are radically distinct on a number of
scores, and that it is futile to attempt to wrench from the two some sort of unified
account of self-respect, or to disregard the differences with impunity. This can be
seen, for instance, in the previously noted shift from a concept of self-respect in
Roland and Foxx’s article to what is actually the Kantian concept. These differences
need to be preserved and engaged rather than overlooked. But which of the two
concepts is, then, preferable? Which one should form the basis of psychological
research into self-respect? That is, of course, a tricky question to answer. The
Kantian concept will recommend itself naturally to liberals, while communitarians
and like-minded people will tend to favor the Aristotelian concept. The general
reservations that communitarians entertain about the thin liberalist notion of the
good and the disembeddedness and anomie of the liberal self will certainly carry
over into their assessment of Kantian self-respect. Nevertheless, we should bear in
mind that the rampant disputes that raged between liberals and communitarians
twenty-odd years ago have gradually subsided. My aim here is not to adjudicate
between these two moral standpoints though I may allow myself to suggest certain
practical reasons why I think social scientists should consider measuring Aristotelian
rather than Kantian self-respect.
First, Aristotelian self-respect falls neatly into line with contemporary virtue
ethics which is, whether we like it or not, the currently most fashionable moral
theory in academic circles. Moreover, ample anecdotal evidence seems to suggest
that virtue ethics—with all its down-to-earth references to human virtues and
vices, flourishing and follies—holds strong appeal for the general public, much
more so at least than the Kantian discourse which tends to be pitched at quite a
high level of abstraction. Second, Aristotelian self-respect acknowledges the psycho-
logical and moral salience of the emotions (their affective as well as cognitive
features), which is in full accordance with current research trends in psychology
and philosophy, and also with the increasing emotionalization of public discourse.
Third, the Aristotelian notion of the self, underlying Aristotelian self-respect, as
one created and sustained through the recognition of others, corresponds well
with the (symbolic-interactionist) looking-glass view of the self that continues to
hold sway in social-science circles. Fourth, the Aristotelian notion of excessive self-
respect will appear more intuitively plausible to most people than the Kantian

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
236 Kristján Kristjánsson
denial of this possibility. This, in point of fact, is precisely what laypeople tend
to refer to as perfectionism, which can, according to folk psychology at least, be a
pretty disabling condition. It should be noted however, that “perfectionism” in
this lay sense is not the same as “moral perfectionism” (a respectable philosophi-
cal outlook).

Readers may find these above reasons variously weighty. I consider them, at
least, when taken together, weighty enough to warrant further scrutiny into possible
measures of Aristotelian self-respect, a task which follows in the final section.

III. MEASURING THE COMPONENTS OF ARISTOTELIAN SELF-RESPECT

Roland and Foxx make no bones about psychological research into self-respect
being in its infancy. If self-respect is to supplant the now much-maligned global
esteem in quantitative psychological research, the concept has to be strengthened,
its main components demarcated and made operationalizable. Roland and Foxx
devote only a single short paragraph to the potentially measurable components
of self-respect (2003: 281). Some of those components, such as “self-control,”
“humility,” and a life plan based on “moral law,” are not parts of the Aristotelian
concept. Others, such as “behaviors that demonstrate respect for others,” seem
too vague, as they stand, to serve as guidelines for instrument-design. It may be
that Roland and Foxx give questions of measurements such short shrift because
they are more interested in the future uses of self-respect for psychological therapy—
helping therapists engage in character-building moral discourse with their clients
(see 2003: 275–282)—than for psychological measurements of the kind which
have proliferated in recent years for global self-esteem, where that construct
has been gauged in order to establish potential correlations to salient social,
educational, and psychological variables. Nevertheless, even if the future uses of
self-respect are to be predominantly therapeutic, one would think that some
standards for measuring clients’ self-respect would also be required in order to
monitor the benefits of the relevant therapeutic interventions.
In the preceding section I claimed that psychologists would first have to choose
between Kantian and Aristotelian self-respect as the one to be measured. Assum-
ing that my reasons for favoring the latter are accepted, our earlier glimpse of it
must now be enhanced. The exploration in the preceding section focused on the
areas of discordance between the Aristotelian concept of self-respect and the
Kantian one; now it is time to say more about its actual components. We need
not be overly deterred there by the fact that Aristotle’s corpus does not contain
any term that is equivalent to “self-respect” in contemporary English: a clear
concept of self-respect can be teased, and has been teased, out of his account of
“great-mindedness” (megalopsychia) and some of its surrounding virtues (see, e.g.,
Kristjánsson, 2002, Chs 3–4; Russell, 2005).

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 237
The first component of Aristotelian self-respect is an impeccable command of all the
moral virtues. The great-minded person both “thinks himself worthy of great things”
(that is, has subjective appraisal self-respect) and really is worthy of them (is objec-
tively self-respectful) in virtue of possessing “greatness in each virtue.” Another
character type, the pusillanimous, also has self-respect, but unfortunately lacks
appraisal self-respect, underestimating his moral worth (Aristotle, 1985: 97–99
[1123b2–31]). Let us leave appraisal self-respect (a kind of “domain-specific
self-esteem,” in the current jargon, as already explained) out of our account at this
time and concentrate instead on the objective self-respect part. On the Aristotelian
understanding, self-respect cannot be divorced from other moral virtues: it is a
higher-order virtue makes those other virtues greater and “does not arise without
them” (1985: 99 [1124a1–3]). Russell explains this unique status of self-respect as,
at once, a separate virtue and a consummation of the other virtues by noting that
self-respect and (other) moral virtues shape each other: “one must start with some
form of self-respect in order to develop a person of virtue, and as a person so
develops, the self that he respects changes into a person that is more and more
worthy of his own respect” (2005: 105). This does not mean that the self-respectful
person never errs morally. As Curzer (2005) has carefully sketched out, Aristotle
describes at least seven ways in which even persons of full virtue may act wrongly
on occasion (succumbing to overwhelming pressures, acting temporarily out of
character, having tiny glitches in their virtues, etc.), while still remaining virtuous
overall. Indeed, the painful feeling when one occasionally loses self-respect may
help secure the characterological foundations of true self-respect, just as painful
feelings of remorse attest to and strengthen one’s sense of duty (cf. Telfer, 1995:
114). I mentioned earlier that in the Aristotelian model, those virtues are
considered greatest which are most beneficial to others. Amongst those are justice,
courage, and generosity. Measures of Aristotelian self-respect thus need to capture
both a person’s overall mastery of moral virtues and, more specifically, a person’s
exhibition of the greatest virtues.
The second component of Aristotelian self-respect is pridefulness. Self-respectful
persons have an acute sense of their own dignity; they are proud of their own
worthiness and expect recognition of it from others. They are also liable to shame
not only if they fail to live up to their own standards but also if they feel that the
external recognition they deserve is not forthcoming (see further in Kristjánsson,
2002, Chs 3–4). Measures of Aristotelian self-respect need to capture the
self-respectful person’s sense of both pride and shame.
The third component of Aristotelian self-respect is strength of character. By that I
mean that self-respectful persons have steadfast convictions and also the courage
of those convictions; they nail their colors to the mast and stand up for them-
selves, always willing to defend their convictions with fortitude—even to the
point of their own physical peril—but unwilling to overlook insults or to com-
promise simply for the sake of compromise (see further in Russell, 2005).
Measures of Aristotelian self-respect need to capture this fortitude and strength

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
238 Kristján Kristjánsson
of character (which must, as already noted, not be confounded with Kantian
self-control).
The fourth component of Aristotelian self-respect is stability of character.
Self-respect is a continuing, global trait of one character as a whole (see Russell,
2005: 104). Self-respectful persons can be trusted to act and react consistently
under similar circumstances. While not exempt from the occasional error,
they do not have regular “off days” or exhibit systematic moral bias in dealing
with different people. In that way, they display maturity of character. Generally
speaking, they do not distinguish clearly between their personal and their moral
selves (see Colby & Damon, 1992). Measures of Aristotelian self-respect need
to capture this consistency and maturity of character. I consider those four
components to be necessary conditions of (Aristotelian) self-respect, in terms of
which the concept can be analyzed. Of course each of them may be studied
in psychological research as a separate variable, just as there is already an
abundance of investigations of particular moral attitudes. However, for a psy-
chological study to be one of self-respect, the four components would have to
be studied conjunctively.
Some allowances have to be made if the subjects of measured self-respect are
children: children, for example, cannot be expected to show the same stability
and maturity of character as adults. Some writers interpret Aristotle’s account in
such a way that children cannot possess virtue at all (see, e.g., Welchman, 2005:
150) and, hence, lack self-respect. That is too drastic a conclusion. Aristotle goes
out of his way to introduce two moral virtues specific to young people: emulousness
and shamefulness. He also describes some morally praiseworthy characteristics
that virtuous adults should ideally possess but that come more easily to young
people for reasons of developmental psychology: open-mindedness, optimism,
trust, courage, guilelessness, and friendship. And he unflinchingly refers to these
characteristics as “virtues” (1991: 165–166 [1389a16–b3]; see further in Kristjánsson,
2006b). I see no particular problem in measuring these virtues, as well as
children’s pridefulness and (relative) strength of character, should the aim be,
say, to explore the link between self-respect and school performance.
Let us return finally to Roland and Foxx’s contention that it is “objectivity”
which differentiates self-respect from self-esteem (2003: 282). Measures of
self-esteem rely on self-reported subjective feelings that are likely to contain
substantial biases (see Baumeister et al., 2003: 7–8). That consideration, coupled
with the fact that most measures of global self-esteem have failed to yield any
significant correlations, makes self-respect an attractive alternative research
option. Measures of the components of Aristotelian self-respect, delineated above,
should utilize the potential advantage of objectivity. Self-reports—particularly
of the “How much self-respect do you feel that you possess”-kind—would then
be avoided; instead, the focus would be on instruments and/or experiments to
determine whether or not people actually possess the components in question
(and those components could, subsequently, be examined with regard to possible

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 239
correlates). I am not an empirical psychologist and it hardly behooves me to
advise experts on the backroads and byways of instrumental/experimental design.
I hope I may be forgiven, however, for sounding certain warnings about potential
pitfalls, in particular, faults that philosophers commonly encounter when examining
empirical research into morality.
The first common pitfall arises from trying to measure people’s moral maturity
via responses to far-fetched moral dilemmas: scenarios that they are unlikely to
ever come across in their daily lives. This is precisely what was wrong with
Kohlberg’s research instruments and what gave rise to his bleak stage theory of
moral development (1981). The theory might better be termed one of moral
underdevelopment, especially with regard to children who tend to score particu-
larly badly when the scenarios to which they are asked to respond lie outside their
immediate world of experience (see further in Kristjánsson, 2002: 184–186). Even
if self-respect involves—at least in adults—the stable and committed exhibition of
moral virtue, the stability of virtue will be partly dependent upon the stability of
daily experiences, and when people are prompted to go beyond those experiences,
their self-respect may flounder for a while, taking time to adjust. So if prospective
instruments for measuring self-respect are to rely on responses to morally charged
scenarios, it is advisable to make those scenarios as this-worldly and reasonably
realistic as possible. For the same reason, longitudinal studies of moral behavior
will be preferable to single-case studies.
Furthermore, if self-respect is to be measured through psychological laboratory
experiments, it is better to avoid experiments in which other people—especially
persons of authority—are enlisted to persuade the subjects to do the wrong thing,
through peer pressure or direct exhortations, in artificially created circumstances.
Milgram-type experiments (1974) have already taught us that in such circum-
stances, subjects often choose a demoralizing course of action. Experiments of
that type hark back all the way to Hartshorne and May’s famous studies of
honesty (1928), which were taken to indicate that children do not really act
consistently over time or possess dispositional honesty or dishonesty: studies which
prompted Kohlberg’s infamous “bag-of-virtues” dismissal of virtue-based ethics.
The conclusions that moral situationists want to draw from such studies—namely,
that the whole idea of stable character traits, including self-respect, is a chimera,
and that human behavior is radically situation-dependent (see, e.g., Doris, 2002)—
seems, however, much too strong. What can be learned from those experiments
is that when people are confronted—without allies, in alien circumstances—with
a view of things very different from their own, and even pressured by those whom
they consider figures of authority to act in certain ways, the majority cannot be
trusted to follow the dictates of their self-respect. Even in such perilous situations,
however, where people’s sense of morality encounters its severest trials, the
correlation in the character domain from situation to situation tends to range
between .2 and .3, which would be consistent with 20% of the subjects exhibiting
stable moral character traits—perhaps not such a pessimistic conclusion after all

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
240 Kristján Kristjánsson
(see further in Sabini & Silver, 2005). Nevertheless, experiments designed to elicit
information concerning people’s self-respect should avoid circumstances of this
kind, which again—fortunately—are not part and parcel of people’s daily moral
experiences.
I have already mentioned self-report instruments as a potential pitfall. One way
to determine subjects’ levels of self-respect without invoking self-reports would
involve asking those closest to them (friends, parents, siblings, etc.) to evaluate the
extent to which the subjects possess the components of self-respect. Obviously, this
method brings with it problems of its own. However, even if one could expect
those significant others to consistently overrate the subjects’ self-respect through
familiarity bias, the relative rankings of subjects (even if not necessarily their total
scores) might yield important results.
The final pitfall that I would want to mention is the behavioristic fallacy of
simply judging persons’ self-respect from the way they behave; witness, for
instance, Roland and Foxx’s contention that in order to measure self-respect, we
need to define and observe “behaviors that are respectful towards the self and
others” (2003: 272). Self-respect is a character trait, not merely a behavior trait.
For example, the merely self-controlled altruist is, by Aristotle’s lights, lacking in
self-respect in comparison with the willing and emotionally engaged altruist, even
if both perform the same actions. Psychologists studying (Aristotelian) self-respect
thus need to try to gauge the considerations that experimental subjects have in
mind when making moral decisions: the inner mediating events. There is some-
thing quintessentially Aristotelian (given Aristotle’s keen interest in physiology)
with the suggestion that in the future, brain scans and saliva tests may perhaps be
used to detect people’s self-respect levels.
Whatever methods are chosen in the end, I think we could reasonably expect
empirical psychologists to come up with findings of people’s self-respect that
would be more valid than those for, say, global self-esteem, given the objectivity
of the components of the former. Once such findings have been established, the
next step would be to look for potential correlates of low and high self-respect.
Variables which have been studied in relation to global self-esteem—including
school performance, risky sexual behavior, and drug/alcohol abuse—immediately
come to mind. Given my earlier account of the components of Aristotelian self-
respect, we could expect such studies to provide non-trivial conclusions. The same
would not apply to general “prosocial behavior,”however, because such behavior
is already included in the first component of self-respect.
Now that research into global self-esteem in psychology and education has
reached its limits and lost its luster, I hope that the considerations explored in this
article may help pave the way for future psychological research into self-respect
and its correlates. As I have emphasized, such research must not avoid taking
a stand on what kind of self-respect is being studied, Aristotelian or Kantian,
nor shirk detailed analysis of the objective components of self-respect and their
manifestations in human character.

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
 Measuring Self-Respect 241
Kristján Kristjánsson, PhD
Professor of Philosophy
University of Akureyri
P.O. Box 224
602 Akureyri
Iceland
E-mail: kk@unak.is

REFERENCES

A. 1985. Nicomachean Ethics, trans. T. Irwin. Indianapolis: Hackett Publishing.

A. 1991. On Rhetoric, trans. G.A. Kennedy. Oxford: Oxford University Press.
B, A. 1997. Self-efficacy: The Exercise of Control. New York: W. H. Freeman & Co.
B, R.F., C, J.D., K, J.I. & V, K.D. 2003. Does high
self-esteem cause better performance, interpersonal success, happiness, or healthier
lifestyles? Psychological Science in the Public Interest, 4(1): 1– 44.
C, R. 2004. Situated self-esteem. Journal of Philosophy of Education, 38(1): 91–105.
C, A. & D, W. 1992. Some Do Care: Contemporary Lives of Moral Commitment. New
York: Free Press.
C, H.J. 2005. How good people do bad things: Aristotle on the misdeeds of the
virtuous. Oxford Studies in Ancient Philosophy, 28(1): 233–256.
D, W. 1995. Greater Expectations: Overcoming the Culture of Indulgence in Our Homes and
Schools. New York: Free Press.
D, J. 2002. Lack of Character. Cambridge: Cambridge University Press.
E, N. 2001. Self-Esteem: The Costs and Causes of Low Self-Worth. York: Joseph Rowntree
Foundation.
H, H. & M, M.A. 1928. Studies in the Nature of Character, vol. 1: Studies in Deceit.
New York: Macmillan.
J, W. 1950. The Principles of Psychology, vol. 1. New York: Henry Holt & Co.
K, I. 1967. Groundwork of the metaphysic of morals. In H.J. Paton (Trans.) The
Moral Law. New York: Barnes & Noble.
K, L. 1981. Essays on Moral Development, vol. 1–3. New York: Harper Row.
K, C.M. 1996. From duty and for the sake of the noble: Kant and Aristotle on
morally good action. In S. Engstrom & J. Whiting (eds.) Aristotle, Kant, and the Stoics:
Rethinking Happiness and Duty ( pp. 203–236). Cambridge: Cambridge University Press.
K, K. 2002. Justifying Emotions: Pride and Jealousy. New York: Routledge.
K, K. 2006a. Emotional intelligence in the classroom: an Aristotelian critique.
Educational Theory, 56(1): 39 – 56.
K, K. 2006b. Parents and children as friends. Journal of Social Philosophy, 37(2):
250 – 265.
MD, J. 1996. Deliberation and moral development in Aristotle’s ethics. In
S. Engstrom & J. Whiting (eds.) Aristotle, Kant, and the Stoics: Rethinking Happiness and Duty
( pp. 19 –35). Cambridge: Cambridge University Press.
M, S. 1974. Obedience to Authority. New York: Harper.
N, W. 1993. Self-esteem and moral virtue. Journal of Moral Education, 22(1): 51–53.
R, C.E. & F, R.M. 2003. Self-respect: a neglected concept. Philosophical Psychology,
16(2): 247–288.
R, D. 2005. Aristotle on the moral relevance of self-respect. In S.M. Gardiner (ed.)
Virtue Ethics, Old and New ( pp. 101–121). Ithaca: Cornell University Press.

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007
242 Kristján Kristjánsson
S, J. & S, M. 2005. Lack of character? Situationism critiqued. Ethics, 115(2):
535 – 562.
S, D. 1981. How to distinguish self-respect from self-esteem. Philosophy & Public Affairs,
10(4): 346–360.
S, R. 2002. Self-esteem: the kindly apocalypse. Journal of Philosophy of Education, 36(1):
87–100.
S, M. 2000. The Feel-Good Curriculum: The Dumbing Down of America’s Kids in the Name of
Self-Esteem. Cambridge, MA: Da Capo Press.
T, E. 1995. Self-respect. In R.S. Dillon (ed.) Dignity, Character, and Self-Respect
(pp. 107–116). London/New York: Routledge.
W, J. 2005. Virtue ethics and human development: a pragmatic approach.
In S.M. Gardiner (ed.) Virtue Ethics, Old and New ( pp. 142–155). Ithaca: Cornell
University Press.

© 2007 The Author

Journal compilation © The Executive Management Committee/Blackwell Publishing Ltd. 2007