Philip Pettit’s A Theory of Freedom:
From the Psychology to the Politics of Agency is remarkable in a number of
respects. Its principle achievement is to offer a kind of holistic perspective
on freedom not often seen in philosophical work in this area. Traditionally in
analytic philosophy, issues of ‘metaphysical’ and ‘political’ freedom are kept
quite distinct. Whether agents possess the freedom necessary for them to be
responsible for their conduct has long been considered a separable issue from
the political question of what range of options it is desirable that people
have within which to choose. One of Pettit’s principal aims in "A Theory
of Freedom" is to bridge this gap. In particular, he harnesses intuitions
regarding both of these questions to develop an overarching theory of freedom:
a theory both of what conditions are necessary for agents to be held responsible
for their conduct, and of what kind of environment is required for this freedom
to have the significance we normally take it to have.

In
this discussion I will focus on the first half of the book where Pettit
develops his account of agential freedom. This account, however, is consciously
guided by considerations of environmental freedom developed later in the book.
Pettit’s discussion of agential freedom is extremely valuable for the way in
which it distinguishes and draws together various disparate themes prevalent in
contemporary discussion of free will. Pettit distinguishes three different "connotations"
of freedom that guide different approaches to the topic: the connotations of
underdetermination, ownership, and of responsibility. Pettit shows how emphasis
of one or another of these connotations of freedom underlies and motivates
different important contemporary theories on the subject.

‘Underdetermination’
connotes the idea that free agents’ conduct be underdetermined, in a
significant respect, by their environments. Coercion undermines freedom in this
sense because it signifies an undesirable environmental determination of an
agent’s conduct — something (rope, for example, tying someone to a chair) that
prevents the agent from being the determining cause of their conduct. Agents
who are free in this sense are said to possess ‘freedom of action’.

‘Ownership’
connotes the idea that free agents identify with their conduct; that they see
it as theirs, as flowing from their wills. Psychological compulsions such as
addiction undermine freedom in this sense precisely to the extent that such
conditions preclude agents from identifying with their actions. While an addict
may not be externally constrained in their conduct, they are nevertheless prevented
from ‘owning’ their actions. Agents who are free in this sense are said to
possess ‘freedom of the self’.

‘Responsibility’
connotes the extent to which free agents are evaluated, interpersonally, for
what they do. The responsibility connotation is brought out most clearly by focusing
on the attitudes involved in attributions of freedom. Agents are free, in this
sense, when they are the kind of beings appropriately resented for bad conduct.
That cats and the mentally ill are thought to lack freedom in this sense can be
seen in the fact that resenting them seems inappropriate. Agents who are
appropriately held responsible for their conduct are said to possess ‘freedom
of the person’.

Pettit
uses this taxonomy of the connotations of freedom in several useful ways.
Firstly, he identifies a distinct skeptical challenge (or "conundrum")
that arises for each connotation of freedom — a reason why it may be thought
that agents lack freedom in that sense. Secondly, Pettit shows how various
different theoretical approaches to the question of freedom may be seen as focusing
on one or another of the connotations of freedom, and as being formed in
response to the conundrum unique to it. Different theories may be seen as
addressing one connotation centrally, and as being applicable, only by
derivation, to the other connotations. Ultimately, Pettit argues that only by focusing
primarily on the connotation of responsibility may all three connotations
adequately be characterized.

The
primary conundrum that arises for freedom of action is a modal conundrum: how
is it possible for an agent to have the ability to do otherwise if the world is
understood as being governed by mechanistic processes of causes and effects?
Seen in this light, it may appear that all agents (not just those tied to their
chairs) are equally coerced to do exactly what they actually do, by the forces
and laws that constitute them. Pettit introduces a class of theory — freedom
as rational control — designed precisely to address this conundrum. According
to freedom as rational control, free actions may be characterized by their
rational antecedents. Free actions are explained and rationalized by the
psychology of the agent who acts. Free actions may thus be distinguished from
twitches and spasms, on the one hand, and coerced activity, on the other. In
neither of these cases is an action properly explained by the psychology of the
agent. Freedom as rational control thus seems adequately to capture the sense
in which a free action needs to be determined by the agent. While there
is clearly some important insight in construing freedom as rational control,
Pettit convincingly argues that it is incapable of adequately explaining both
freedom of the self and freedom of the person. This is because it seems that an
agent may find herself alienated from an action that satisfies the conditions
of freedom as rational control. Moreover, there seems no reason to assume that
agents are necessarily responsible for actions that are free in this sense.

The
primary conundrum arising for freedom of the self is that whenever the
psychology of an agent is described in third-person terms, it seems that no
room is left to understand how the agent may relate to their conduct in
first-person terms, as theirs, and not simply as some alienable activity inside
of them. Pettit introduces a class of theory designed to address this conundrum
— freedom as volitional control. This kind of theory is exemplified by the
work of Harry Frankfurt on freedom of the will. According to Frankfurt, an agent identifies with a desire only when it is
supported by a ‘second-order volition’, that is, a desire that their desire be
effective in action. Frankfurt thus explains the phenomenon of
alienation in cases such as addiction by saying that while the addict may
desire heroin, she fails to desire that her desire for heroin lead her to
action. It is in this sense, according to Frankfurt, that the heroin addict fails to
identify with her desire as her own, as reflective of her self. An agent who
possesses volitional control may be seen as responsible by the lights of her
own volitions, (that is, to the extent that what she does is ratified by her ‘deep
self’). Pettit’s worry, however, is that volitional control nevertheless seems
incapable of explaining an agent’s responsibility to standards beyond her own
self-conception. Whether an agent identifies with cruelty or altruism, as a
standard of behavior, may be beyond her control, so even though she possess
volitional control, it’s not obvious that she may be held responsible for her
conduct.

The
primary conundrum challenging freedom of the person is a recursive worry. If an
agent is responsible for an action, it seems that she must be responsible for
the psychological causes of that action. And if she is to be responsible for
these psychological causes, she must be responsible for their causes, and so
on. Clearly a vicious regress looms that seems to threaten all attributions of
responsibility. It is in light of this connotation of freedom that Pettit’s own
preferred theory comes to the fore. On this theory — freedom as discursive
control — an agent is free to the extent that she possesses the capacities
required to be a legitimate co-deliberator in ‘discursive’ contexts — familiar
contexts where people reason together about matters theoretical as well as
practical. Plausibly, Pettit takes the ‘authorization’ of participants in such
discursive activities as being constitutive of responsible agency. For it is
precisely within such contexts that the interpersonal features of
responsibility — praise, blame, resentment punishment, etc. — find their
natural source. Pettit argues that the notion of discursive control provides
all the resources necessary to block the recursive regress that challenges
freedom of the person. It needn’t be the case that an agent be responsible for
the causes of her behavior ad infinitum. As long as an agent’s behavior
is explained, ultimately, by her capacities to respond in the ways required for
co-deliberation, she will be a legitimate object of responsibility attribution.
Pettit then goes on to argue that freedom, construed as discursive control,
provides adequate resources to explain both freedom of the self, and freedom of
action. An agent who possesses the capacities sufficient for co-deliberation
will necessarily be in a position to see the exercise of those capacities as ‘hers’.
Moreover, actions that emerge from the exercise of such capacities are clearly
free in such a way that distinguishes them from twitches and coerced activity.

Before
concluding, I want to highlight what I take to be the most significant problem
that faces Pettit’s theory. Pettit raises the worry himself on pp. 16-17 where
he distinguishes between "developmental" and "non-developmental"
ways of holding people responsible. Holding someone responsible in a
developmental way is essentially strategic — your goal in praising, blaming or
resenting them concerns the effects of doing so; you want to encourage certain behavior,
or discourage other behavior. Pettit points out that this conception of
responsibility doesn’t really match up with the intuitive justification for
responsibility attribution. Intuitively, responsibility attributions are
non-developmental in their justification. We think of ourselves as responding
to people in a ‘backwards-looking’ way; responding to them for having
acted in certain ways, not just to the prospective benefits or drawbacks of
their having acted so.

The
worry is whether Pettit’s conception of freedom as discursive-control is ultimately
able to capture this non-developmental aspect of responsibility. Pettit thinks
it can, arguing that it is a tacit condition of engaging in interpersonal
discourse that one presents oneself as fully responsive to the norms of the
discourse, and thus fully liable for any failures to conform with such norms
(pp. 100-103). Pettit argues that agents thus ‘permit’ others to react to them
in the normal way for violations of the norms of discourse. Does this account,
however, give us enough reason to see responsibility as non-developmental in
its justification? I don’t think so. All Pettit ultimately offers is a story
according to which we naturally respond to each other as if we were
fully capable of responding to the norms of discourse, even when we don’t successfully
respond. However, it seems quite possible that this psychology is founded on a
systematic illusion. The mere fact that we are disposed to think of each other
in certain ways in contexts of co-deliberation doesn’t prove that we actually
are that way.

While
I have reservations about Pettit’s conclusion, I nevertheless found A Theory
of Freedom to be an exciting and original exploration of the notion of
freedom. In a systematic yet eminently readable fashion, Pettit’s book provides
an invaluable overview of contemporary theory both on the metaphysics and the
politics of freedom.

©
2004 Daniel Cohen

Daniel Cohen is a graduate
student in Philosophy at the Australian National University, but he is currently visiting The
University of North Carolina.

Categories: Philosophical, Psychology