Nussbaum

I thus take issue with Aristotle’s definition which was followed, with some adaptations,
by Nussbaum (2001) in her account of compassion. For Aristotle (and Nussbaum, in that
regard), in order to attribute compassion to someone, three crucial criteria have to be
fulfilled. First x, where x is the object of one’s compassion, has suffered some misfortune.
One must also see x’s misfortune as seriously damaging his well-being, i.e. one cannot
see it as merely a minor harm. Second, one must think x did not deserve this misfortune.
One must think it was not his fault or due to actions for which he is to blame: i.e. either x
did not deserve to suffer misfortune, or though he deserved to suffer some, the
misfortune he suffered was out of proportion to what he did deserve to suffer. Third,
one must regard oneself as similarly vulnerable – i.e. one must think the same
misfortune could befall one as befell x. Nussbaum refers, in her account, to these
requirements as the three cognitive requirements for compassion, of size, non-desert
and similar possibilities and she claims that all three are necessary elements of the
complex evaluative judgment that constitutes compassion. In my view, even though I
have not argued for this in full detail here, only the first requirement is actually
necessary. Even though the non-desert and the similar possibilities requirement are
often present, they are not necessary for ascribing compassion to someone. As I argued
above (e.g. Crisp 2008), there are cases in which one may feel compassion for someone
who did a wrong or even if one does not believe that the same misfortune could befall
one as befell the suffering other (I think this last requirement, especially, has a self-
regarding orientation that is not a mark of compassion).1

People’s judgments about their situation can be wrong in a variety of ways.2 Whereas
they may learn a lot from their suffering in some occasions, suffering also often
brutalizes and distorts perception. As Nussbaum notes, suffering often produces
adaptive responses that deny the importance of the suffering which are especially likely

1
To be clear, Nussbaum replaces the requirement of similar possibilities – present in Aristotle – with an
explicitly eudemonistic judgment. She argues that the requirement of similar possibilities is unnecessary
and makes Aristotle’s definition too restrictive, given that there are cases of compassion felt by superior
beings for the misfortunes of inferior ones that the definition fails to reach. She then claims that in its
place, one should introduce a requirement that one see the well-being of the object of one’s compasssion
as important and take it as one’s end. IN her account, this requirement figures as the third necessary
element for compassion. Even though, she claims, judging that one is similarly vulnerable to the
misfortunes that have befallen another is a valuable way in many cases of coming to realize the
importance of another’s well-being and to take it as one’s end, it is this realization and not the judgment of
similar possibilities that is necessary for compassion.
2
This point has been stressed, e.g. by Nussbaum (2001, p. 32).
to happen when this suffering is connected to oppression and hierarchy. Others can
become deeply attached to things that on reflection are trivial or bad for them (and
thereby the loss of these things may produce real suffering for them, even though for
onlookers that may seem exaggerated reaction). In other situations, others may not be
aware of their negative condition (as the example of Adam Smith illustrates, of a person
who has altogether lost the use of reason. Smith argues that ‘of all the calamities to
which the condition of morality exposes mankind…by far the most dreadful’. Such a
person will be an object of compassion, in his view, to anyone who has ‘the least spark of
humanity’. Even though this is the case, the person affected does not judge that his
condition is bad – indeed, that is a large part of what is so terrible about it). A
compassionate agent must act informed by ‘the best judgment [he] can make about what
is really happening to the person being observed – taking the person’s own wishes into
account, but not always taking as the last word the judgment that the person herself is
able to form’ (32). The other’s own perception of his situation matters, but compassion
is also connected to a conception of what is valuable and not valuable for human beings
upon which the compassionate agent relies.

I also think it important to preserve a vivid sense of compassion as a not-fully-cognitive

phenomenon. The Aristotelian account of compassion (which is the account, with a few
modifications, followed by Nussbaum, that has attracted most attention in recent years),
requires the presence of three cognitive requirements – that the evil in question must be
seen as serious rather than trivial, that the other does not deserve it, and that one might
expect it to happen to oneself or someone close to one –, is too rationalistic. In many
cases, not many beliefs are necessary for having compassion. Even though such
requirements are present in many cases, a full-fledged account of compassion should
allow for the possibility of ‘a natural capacity to be directly moved by, and perhaps
immediate distress at, the misery of or suffering of some evil by another being’ (238).3 I
thereby agree with Crisp (2008, 241) that we should be careful not to allow the
acknowledgment of this fact lead us to view compassion ‘in an excessively narrow and
cognitivist way’, as if compassion was merely a matter of having certain beliefs or
making certain judgements about the situation of others. By making compassion above
all a certain sort of thought about the well-being of others, Nussbaum’s account carries

3
Crips claims that contemporary neuroscience has produced strong evidence for believing that this
capacity is present ‘in human beings almost from birth as well as in other animals’ (2008, 238).
the risks of leaving the important non-cognitive aspects of compassion in the
background, not sufficiently stressed. Because of this, a developmental account of
compassion on the lines that authors such as Crisp (2008) suggest may be more
appropriate, according to which, in its most primitive forms, compassion involves a
direct, instinctive distress caused by the suffering of others and it is only when it further
develops that it acquires more sophisticated forms of expression and may involve ‘either
as cause or consequence, various other cognitive and non-cognitive states, such as the
belief that someone who has lost their partner is grieving, the desire to alleviate pain, or
the wish that the evil in question were not occurring’ (241).4 In such an account, what all
cases of compassion have in common is only the non-cognitive element of pain or
distress at the suffering of others.5

I also think it important to preserve a vivid sense that, taken as a virtue or valuable
character trait, compassion is a not-fully-cognitive phenomenon. The Aristotelian
account of compassion (which is the account, with a few modifications, followed by
Nussbaum, that has attracted most attention in recent years)6, requires the presence of
three cognitive requirements – that the evil in question must be seen as serious rather
than trivial, that the other does not deserve it, and that one might expect it to happen to
oneself or someone close to one –, is too rationalistic. In many cases, not many beliefs
are necessary for having compassion. Even though such requirements are present in
many cases, a full-fledged account of compassion should allow for the possibility of ‘a
natural capacity to be directly moved by, and perhaps immediate distress at, the misery
of or suffering of some evil by another being’ (238).7 I thereby agree with Crisp (2008,
241) that we should be careful not to view compassion ‘in an excessively narrow and
cognitivist way’, as if compassion was merely a matter of having certain beliefs or
making certain judgements about the situation of others. By making compassion above
all a certain sort of thought about the well-being of others, Nussbaum’s account carries

4
Such account seems to also to be better in line with available empirical evidence on compassion (e.g.
Goetz et al. 2010).
5
Crisp claims that, in its most primitive form, as in the case of young babies who appear to be distressed
by the sound of other babies’ crying – an episode sometimes called of ‘emotional contagion’ in
psychological literature; e.g. Hatfield et al. 1994 –, compassion ‘is little more than instinctive distress
caused by, rather than at, the suffering of others’ (240, and is not connected to the making of any
judgments in relation to those others (e.g. that their suffering is undeserved, that is serious etc.).
6
Even though this is the case, my account of the virtue of argumentative compassion, as I explain below is
still broadly Aristotelian.
7
Crips claims that contemporary neuroscience has produced strong evidence for believing that this
capacity is present ‘in human beings almost from birth as well as in other animals’ (2008, 238).
the risk of leaving the important non-cognitive aspects of compassion in the background,
or not sufficiently stressed. Because of this, a developmental account of compassion on
the lines that Crisp (2008) suggests may be more appropriate. According to this account,
in its most primitive forms, compassion involves a direct, instinctive distress caused by
the suffering of others and it is only when it further develops that it acquires more
sophisticated forms of expression and may involve ‘either as cause or consequence,
various other cognitive and non-cognitive states, such as the belief that someone who
has lost their partner is grieving, the desire to alleviate pain, or the wish that the evil in
question were not occurring’ (241).8 In such an account, what all cases of compassion
have in common is only the non-cognitive element of pain or distress at the suffering of
others.

Second objection: Is argumentative compassion really distinctive?

Another objection that could be charged against my account is about whether

argumentative compassion is really distinctive, or rather just a moral virtue that can be
applied in argumentative contexts. As I argued above, argumentative compassion
displays all the elements that are present in other forms of compassion, its distinctive
feature being the means or way in which compassion is expressed, namely, via the giving
of reasons or the construction of arguments. Albeit this objection is less important than
the first one, my account of the virtue of argumentative compassion is not predicated on
it being distinctive. Even though this is the case, there are two aspects that, in my view,
authorizes the claim that argumentative compassion deserves the status of a distinctive
argumentative virtue. First, as I explained before, it can be distinguished by the specific
means that is adopted to express compassion for others (namely, argumentation).
Second, the virtue can be exercised for the sake of others in their specific capacity as
arguers, e.g. to deal with their argumentative distress, doubts in the process of building
their own arguments etc.; though, as Socrates’ case illustrates, it can also be addressed
to others in a more general way, in the attempt to help them release or cope with their
emotional distress, about a certain situation or state of affairs in the world. Even if in the
end the virtue of argumentative compassion is only a sub-type of the moral virtue of
compassion, this bears no consequence to my claim that the former deserves the status

8
Such account seems to also to be better in line with available empirical evidence on compassion (e.g.
Goetz et al. 2010).
of a distinctive argumentative virtue, therefore a virtue that deserves being
acknowledged in a full or complete list of argumentative virtues.

Take Socrates’ case, my working example, to articulate this claim further. Socrates could
have tried to console Crito – perhaps even more effectively – by doing various other
things, such as hugging Crito or singing him a song, or even using a device of purely
causal manipulation that had little connection with reasons, such as drugs. Arguably,
some of those devices might have made Crito immediately release his distress, and thus
be more effective means to achieve this goal. Moreover, as explained before, it is
possible – indeed likely – that no reasons that he could possibly have given to Crito
could have helped to change that situation. Socrates was aware that there were things
he knew – and that it would be good if Crito knew – but also that there was no fully
satisfactory way for him to transmit them to Crito, in a way that Crito could actually
understand. Though a valued friend, Crito was not – and would probably never be – in a
position to understand the reasons that really moved Socrates. Despite all this, Socrates
decides to give Crito reasons. Those reasons were neither good justificatory reasons nor
Socrates believed them to be good reasons in a justificatory sense. But they were
reasons. In the dialogue, Socrates does not simply assert to Crito that he is not going to
escape, nor does he ask Crito to simply acquiesce to it (‘Crito, I appreciate your concern,
but I am convinced to stay’). Indeed, Socrates’ behaviour here is often viewed as
puzzling partly because, in the Platonic dialogues, Socrates is usually not in the business
of giving reasons to his interlocutors. Characteristically, he operates by asking questions,
in the attempt to derive conclusions from the statements that his interlocutors
themselves make or that he makes them to make.9 This is, however, not the case in the
Crito (especially in the second half of the dialogue).

One possible explanation – and, in my view, a particularly good one – for why, despite all
the circumstances, Socrates still gives Crito reasons is that he attributed a very high
importance to Crito’s own capacity for active critical thinking and for Crito’s
participation in the exchange; i.e. he thought that promoting Crito’s active engagement
with reasons was something worth pursuing for its own sake (and thus independently
from any further consequence that this active engagement could or could not bring to

9
E.g. Vlastos.
Crito, such as releasing his distress).10 It is possible – indeed, likely, if my proposed
reading is true – that he also thought that promoting Crito’s active engagement with
reasons could be an effective means to release Crito’s distress as much as possible. In my
view, this is all reason for believing that there is something distinctive in expressing
compassion for another by giving the other reasons, and the distinctiveness of the virtue
of argumentative compassion.

10
Vlastos G, Socrates, Ironist and Moral Philosopher (1991) p. 44, holds that Socrates had a deep
commitment with his interlocutors’ own autonomy and rational capacities, and that though he cared
about the truth and cared that they arrived at the truth, he cared more about them arriving at the truth by
themselves (in the hard way, even if that meant that they would never arrive at the truth). Vlastos suggests
that for Socrates, this was also the only way of promoting a real, long-term and stable effect in their self-
understanding (p. 42).