Sie sind auf Seite 1von 27

Philosophy of the Social Sciences

Philosophy
10.1177/0048393105284170
Saaristo / NoofEscape
the Social
fromSciences
Philosophy
Volume 36 Number 1
March 2006 40-66
© 2006 Sage Publications

There Is No Escape 10.1177/0048393105284170


http://pos.sagepub.com
hosted at
from Philosophy http://online.sagepub.com

Collective Intentionality and


Empirical Social Science
Antti Saaristo
University of Helsinki and London School of
Economics and Political Science

This article examines two empirical research traditions—experimental eco-


nomics and the social identity approach in social psychology—that may be
seen as attempts to falsify and verify the theory of collective intentionality,
respectively. The article argues that both approaches fail to settle the issue.
However, this is not necessarily due to the alleged immaturity of the social sci-
ences but, possibly, to the philosophical nature of intentionality and intentional
action. The article shows how broadly Davidsonian action theory, including
Hacking’s notion of the looping effect of the human sciences, can be developed
into an argument for the view that there is no theory-independent true nature of
intentional action. If the Davidsonian line of thought is correct, the theory of
collective intentionality is, in a sense, true if we accept the theory.

Keywords: collective intentionality; experimental economics; social identity


theory; Donald Davidson; Ian Hacking; constructivism; action; agency; phi-
losophy of the social sciences

1. The Theory of Collective Intentionality

O ne of the most interesting recent developments in philosophy of social


science is the approach that has become known as the theory of collec-
tive intentionality. The theory is an attempt to move towards a more holistic
view of social action while remaining ontologically naturalistic and individ-
ualistic (Saaristo 2003). In short, standard methodological individualism of,

Received 5 October 2005


Author’s Note: I wish to thank the audiences at the Conference on Collective Intentionality IV
(Siena, October 2004), the Erasmus Institute for Philosophy and Economics (Rotterdam, Febru-
ary 2005), the 7th Annual Philosophy of Social Science Roundtable (New York, March 2005)

40

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 41

for example, mainstream rational choice theory holds that social action
amounts to nothing more than strategic individual action in the sense of inde-
pendent individuals—social atoms—taking the (expected) behaviour of oth-
ers into account when seeking to realise their personal goals (cf. Barnes
1995). The theory of collective intentionality, in contrast, holds that the
notion of strategic action captures at most one aspect of social action. After
all, strategic action is social only in the sense that the agent performing the
action takes the behaviour of others into account when choosing her own
actions. In the jargon of the collective intentionality theory, such action is
essentially individual-mode action, albeit social in its content (taking the
behaviour of others into account) and, perhaps, directedness (whose benefits
the action aims to maximise).
The collective intentionality theory holds that—in contrast to mere strate-
gic individual-mode action—truly social action is essentially we-mode
action based on collective intentionality. The claim is that in truly social
action, agents instinctively switch to we-mode psychology and frame the sit-
uation at hand in irreducibly collectivistic terms so that the instrumental
rationality of social actions is assessed in terms of what is rational for the
group to do. In short, the rationalising question behind strategic individual-
mode action is what I should do in this social situation to realise my ends (be
they egoistic, altruistic, or what have you), whereas the question motivating
we-mode action is what we should do in order to realise our goals. The works
of Margaret Gilbert (1989 in particular), John R. Searle (1995), and Raimo
Tuomela (especially 1995) are usually identified as the main sources of the
theory of collective intentionality in this sense.1
To put it in Searlean terminology, the theory of collective intentionality
holds that sometimes our intentional attitudes such as beliefs, desires, and
intentions are held irreducibly in the we-mode. Individuals who adopt the
we-mode do not assess a situation from their distinct individualistic perspec-
tives, but rather from the perspective of the collective they identify with. In
other words, if I adopt the we-mode, I figure out what is our goal in the situa-
tion at hand and what is our best means for achieving that goal, and then form
1. I should add that in this article, I am concerned with the idea of collective intentionality as
an essentially collective mode of intentional states attributed to an individual, i.e., the view that
individuals are capable of thinking and acting both in the individual mode and in the collective
we-mode. This is the aspect of the theory emphasised by Searle (1995), whereas, e.g., Gilbert
(1989; see also Tollefsen 2004) is mainly interested in collective intentionality in the sense of
attributing intentional states to collectives. Tuomela (1995) analyses both aspects.

and the HPS Workshop on Philosophy of the Social Sciences (Cambridge, April 2005), as well as
John B. Davis, Damien Fennell, Eleonora Montuschi, and Paul A. Roth for very helpful com-
ments and criticisms. Research for this article was supported by the Helsingin Sanomat Cente-
nary Foundation, the Jenny and Antti Wihuri Foundation, and the Emil Aaltonen Foundation.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


42 Philosophy of the Social Sciences

a we-intention of the form “We will do X.” Only when I have reached this
stage do I derive my individual-mode intention from the we-intention. Hence
my individual action intention is subordinate to the collective we-intention in
the sense that I first figure out that we should do X, and only then that in order
for us to do X, I should do Y as my part of X, and hence I set out to do Y.2 In
the standard individualistic picture, sociality is based on the strategies of
individuals. The theory of collective we-mode intentionality holds, in con-
trast, that the behavioural strategies of individuals are based on the more
fundamental sociality of collective intentionality.
This clearly challenges the received individualistic view of intentional
action, according to which all actions (including social actions) are ulti-
mately explainable in terms of individual-mode beliefs and desires. The
received view holds that I may have desires that are directed towards the
well-being of others or the group I belong to, but such desires are my personal
(individual-mode) desires that are social only in their directedness and con-
tent.3 In other words, the received individualistic view holds that humans
may entertain collective goals and the like only if they have a more funda-
mental individual-mode pro-attitude towards sociality. The collective
intentionality view denies this hierarchy and claims that in terms of
fundamentality, individual-mode considerations and collective we-mode

2. This characterisation of collective intentionality satisfies largely what Elizabeth Ander-


son (2001, 28) identifies as Amartya Sen’s requirements for a theory of agency that could see
cooperation in one-shot prisoners’ dilemmas as instrumentally rational. Anderson conceptual-
ises the reasoning corresponding to my characterisation of the collective intentionality approach
in terms of social identifications—a move that agrees very well with my reading of the social
identity approach in social psychology (below). I should also add that my characterisation is
Searlean in the sense that it agrees with Searle (1995), whereas in Searle (1990) the content of
intentions is given by their causal roles, and my collective intention is not an intention to make it
the case that I have an action-intention (cf. Zaibert 2003, 227). Thus, Searle (1990) argues that the
true form of collective intentionality is “I intend that we do X by means of my action Y.” How-
ever, as Zaibert (2003, 228) points out, this is not a collective we-mode intention, but a standard,
individual-mode intention (albeit collective in its content). In this article, I focus on the
Davidsonian view that the contents of intentions are given by their role in rational relations (and
not by their causal role as such), in which case the 1995 characterisation in terms of derivation is
perfectly adequate.
3. This is why I did not mention, for example, Michael Bratman’s (1999) work on shared
intentions as an example of the theory of collective intentionality. In Bratman’s view all
intentionality is in the individual mode and only the content or directedness of our intentional
attitudes can be collective (“I intend that we J” as opposed to “We intend to J”). In Bratman’s sys-
tem, there is no room for collective we-mode intentions, which I take to be the defining mark of
the theory of collective intentionality. As explained in the previous note, for this very same reason
Zaibert (2003) argues that Searle (1990) does not give a theory of collective intentionality either.
See, however, Tollefsen (2004), which counts Bratman as one of the main theorists of collective
intentionality on a par with Gilbert, Searle, and Tuomela.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 43

considerations are on a par. Both are irreducible modes of human psychol-


ogy. Which mode of psychology is activated and whether agents will act
upon individual or collective-mode considerations, regardless of their direct-
edness and content, depends on the particular features of the situation, per-
sonal histories of the agents involved, etc.4
The theory of collective intentionality has been a characteristically philo-
sophical affair, motivated by conceptual problems faced by attempts to pro-
vide atomistic analyses of collective action and social facts. The claim is that
individual-mode intentions, even when combined with mutual beliefs or
common knowledge concerning the intentions of others, will not add up to
the strong sense of intending and acting together that appears to be required
for paradigmatic social action. Hence, irreducible collective intentionality is
needed.
In my view, perhaps the strongest formulation of this line of thought can
be given in terms of collective action dilemmas. The claim is that individual-
mode accounts of cooperation in collective action dilemmas (such as the
prisoners’ dilemma) where cooperation is irrational from the perspective of a
disparate individual but rational from the point of view of the collective in
question, cannot be fully satisfactory (e.g., Tuomela 2000, 2002). Collective
intentionality in the present sense, namely, as the tendency to put aside con-
siderations tied to the individualistic perspective and to adopt a collectivistic
viewpoint to the situation at hand, seems tailor-made to deal with such short-
comings. Moreover, since it is generally accepted that we are in fact able to
cooperate in collective action dilemmas (e.g., Taylor 1987), we seem to have
here a strong argument in favour of the theory of collective intentionality.
I find arguments like this convincing. However, arguably such theoretical
and philosophical arguments in favour of the theory of collective
intentionality can only take us to the point where we can see that the theory is
plausible, that it makes sense, and perhaps that we have good reasons to think
that collective intentionality most probably is an irreducible part of human
intentionality. But who says that the real world is such that what makes sense
is indeed real? In other words, the theoretical and philosophical arguments in
favour of the collective intentionality theory ought to be strengthened by

4. Thus, in my reading the theory of collective intentionality makes conceptual room for
thinking that sometimes the desires that move me are not essentially my (individual-mode)
desires but our (we-mode) desires. Of course, a we-mode desire that moves me is my we-mode
desire—I may be the only person holding it—but I hold it irreducibly in the we-mode. It seems to
me that Hans Bernhard Schmid (2004; see also Anderson 2001) is developing a view similar to
this when he suggests that a theory of collective intentionality could be the way to make sense of
Amartya Sen’s perplexing claim that truly social action requires that a person’s rational choices
are not based on the pursuit of her own goals.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


44 Philosophy of the Social Sciences

empirical studies examining whether humans in fact do show evidence of


collective intentionality.
Since the mainstream individualistic view and the theory of collective
intentionality paint pictures of social action very different from each other,
one would expect the empirical social sciences to lend their support to one of
them in a rather unambiguous manner. However, we will see that this is not
necessarily the case. In what follows, I first review two empirical approaches
to the problem and argue that they cannot settle the issue conclusively (sec-
tions 2-4). From section 5 onward, I offer a possible philosophical diagnosis
of this letdown.

2. The Discovered Preference Hypothesis


and Experimental Economics
Economists studying human action sometimes subscribe to the so-called
discovered preference hypothesis (I take the label from Charles R. Plott
1996). The view can be seen as a reaction to the earlier doctrine of a revealed
preference hypothesis (cf. Hausman 2000). However, part of my argument
will be that, somewhat surprisingly, vis-à-vis the present problem the discov-
ered preference view actually fails to overcome the limitations of the re-
vealed preference view. Therefore it is useful to start with the revealed pref-
erence view and only then move on to the discovered preference hypothesis.
The animating idea behind the revealed preference view is that a scientific
perspective to human action must not engage in folk psychological specula-
tions on what is going on in the minds of the agents. Rather, scientists must
study actual choice behaviours as they become manifest in observable exter-
nal behaviour. The preferences behind the behaviour are reconstructed theo-
retically on the basis of observations. In this sense, the behaviour of an agent
is thought to reveal the (individual-mode) preferences of the individual agent
in question. An agent does what she prefers to do in that situation, and, con-
sequently, observations of behaviour are thought to provide direct empirical
evidence concerning the preferences behind the behaviour.
However, in this picture the individual mode of agency and action is based
on a plain postulation that whatever an individual chooses to do, the choice is
based on the agent’s individual-mode considerations. As it is usually set up,
the framework of the revealed preference hypothesis simply defines a priori
all action and intentionality to be in the individual mode. Thus, the word
“revealed” is rather misleading here. The economists subscribing to this
approach simply construct a theoretical model for explaining and predicting
behaviour, and the psychological-sounding concepts such as preferences,

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 45

desires, and beliefs play purely instrumental roles in the model (for the idea is
not to speculate about mental states). Obviously, however, this kind of
instrumentalism cannot be used as an argument in the debate concerning the
nature of real psychological mechanisms. In short, the revealed preference
hypothesis does not form a relevant argument for the present discussion,
which is about the psychology of real agents, not about the models of
dogmatic economists.
However, some economists tend to drift away from the instrumentalist
and rather simplistic standpoint of the revealed preference hypothesis. They
argue that results of economic experiments5 do provide a posteriori evidence
for the view that all deliberation is done in the individual mode if we
acknowledge explicitly that the preferences of an agent are not to be con-
structed on the basis of factual choice behaviour. This acknowledgment is
what the move from the revealed preference hypothesis to the discovered
preference hypothesis consists in. This is also the direction in which, for
example, Plott (1996) and Ken Binmore (1999) want to take experimental
economics.
In short, the discovered preference view holds that “each individual has
coherent preferences, but these preferences are not necessarily revealed in
decisions” (Cubitt, Starmer, and Sugden 2001, 386). The individual must
discover what her preferences in fact require her to do in a given situation.
This may require a lot of time and effort (information gathering, deliberation,
learning from experience, and so on), but only when this process is complete
will the behaviour of the agent reveal the true (discovered) preferences. In
particular, it would be a crucial mistake to think that behaviour prior to the
completion of the discovery process reveals real preferences.
A characteristic example of the discovered preference approach is experi-
mental work on the Ultimatum Game. It is a well-known fact that real agents
typically act irrationally in the Ultimatum Game. Instead of rationally seek-
ing to maximise their own benefits, people all over the world tend to follow
an implicit norm of distributive justice (Henrich et al. 2001) that is rational
only from the collective point of view. Binmore, however, argues—basing on
the empirical experiments he has conducted (see, in particular, Binmore,
Shaked, and Sutton 1985)—that if the Ultimatum Game is repeated in an
experimental setting such that agents are allowed enough time to think about
the activity and learn from experience, it will be noticed that the agents move

5. I can, of course, concentrate only on a small aspect of experimental economics. For a gen-
eral overview of the field, see, for example, Kagel and Roth (1995). From the perspective of my
argument, very helpful articles are also, e.g., Cubitt (2005), Hausman (2005), and other articles in
the special issue “Experiment, Theory, World: A Symposium on the Role of Experiments in Eco-
nomics,” Journal of Economic Methodology 12 (2), published in 2005.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


46 Philosophy of the Social Sciences

closer to the behaviour predicted by the theoretical models built on the


assumption of self-directed individual-mode motivations (Binmore 1999,
F20).6
It might be tempting to interpret Binmore’s results as suggesting that our
actions are indeed governed by individual-mode deliberations. Maybe some
actual situations just are so complex that real agents find it difficult to figure
out what rationality, considered in the individual mode, requires them to do.
Repeated games in the “pure” circumstances of Binmore’s laboratory would
then show that agents indeed gradually learn what individual-mode rational-
ity calculations actually dictate in a given situation and behave accordingly,
although their actions have all the time—even when manifestly different
from theoretical predictions—been motivated by the same type of consider-
ations. It is just that sometimes discovering the true preferences may take its
time. This interpretation would save the standard rational choice interpreta-
tion of instrumental rationality behind social action as individual-mode ben-
efit maximisation from the apparent falsification by the empirical fact that
real people tend to cooperate even when it is not rational from their individ-
ual point of view (not even from the perspective of their long-term considera-
tions). This, of course, is precisely the goal of Binmore’s argumentation.
However, I think this is not a decisive case for the present problem of the
nature of human psychology. It is quite possible to redescribe Binmore’s
observations to fit the theory of collective intentionality. The collective
intentionality theory does not deny the existence of individual-mode ratio-
nality considerations. The claim is rather that in social life, agents are often
capable of overcoming the individualistic perspective and acting in accor-
dance with collective rationality, that is, in the we-mode. Binmore’s experi-
mental settings could be interpreted as special circumstances that lead
agents, contrary to “normal” social surroundings, to give up collective-mode
considerations and to follow individual-mode rationality. Indeed, the empiri-
cal data suggest that people rather universally tend to approach social situa-
tions such as Ultimatum Games as the collective intentionality theory would

6. Binmore’s (1999) interpretation of the results of Binmore, Shaked, and Sutton (1985) is,
however, controversial (see Binmore 1999, F20, where Binmore explains his interpretation,
paraphrased above). Some theorists think that unless the players are given active guidance
(which Binmore, Shaked, and Sutton in effect did), the phenomenon of actual behaviour in the
Ultimatum Game moving towards the theoretical predictions is never observed (see, e.g.,
Henrich et al. 2001). I cannot pass judgment on this. However, if Binmore’s description of the
results is not warranted, the results do not challenge the theory of collective intentionality at all.
In this section I aim to show that even if Binmore’s description is accurate, it does not count
unambiguously against the theory of collective intentionality. I thank Joseph Henrich and Natalie
Gold for discussions on this point.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 47

predict, but with sufficient training they can be taught7 to overcome their
social tendencies and act in the individual mode. A fitting example of this is
Robert H. Frank, Thomas Gilovich, and Dennis T. Regan’s (1993) well-
known analysis of how studying theories of (individualistic) economic ratio-
nality tends to transform the behaviour of students of economics to corre-
spond to the theories. This is what Hacking (e.g., 1999, 32) calls the looping
effect of social scientific theorising—social scientific theories do not simply
describe mind-independent reality. Rather, since social reality consists
(partly) of the beliefs and attitudes of individuals, social scientific theories
may transform their own object by modifying the beliefs and attitudes of
individuals (Taylor 1985).
Thus, it seems that experimental economics cannot provide the kind of
unambiguous empirical evidence against the collective intentionality theory
that I am looking for. First, the instrumentalist framework of the revealed
preference hypothesis is unsuitable for providing empirical arguments con-
cerning the true nature of human psychology. Second, when experimental
economists seek to move beyond instrumentalist model-building of the
revealed preference view to the discovered hypothesis, their results can typi-
cally be interpreted as compatible with the pluralism of the theory of collec-
tive intentionality. Indeed, it seems that to defend the individualistic interpre-
tation of his experimental results, Binmore must return to the a priorism of
the revealed preference view and simply postulate that the process of discov-
ering the true preferences may be regarded as complete only when observ-
able behaviour complies with the individualistic theory. But then, of course,
the individualistic theory is not discovered empirically to be true. More
straightforward empirical evidence is needed.

3. The Social Identity Approach in Social Psychology

In order to find empirical studies that target the particular problem at


hand, the psychological processes present in social action, it is quite natural
to turn to social psychology. After all, social psychology by definition stud-
ies precisely this issue. In this article I concentrate on one tradition within
social psychology, the so-called social identity approach,8 which can be seen

7. Thus, the crucial question is whether the process Binmore describes is a process where an
agent discovers her true (independent of the process by which they are discovered) preferences or
where she is being taught new preferences (cf. note 6). The main argument of my article is, in a
sense, that this question may not be answerable empirically but requires us to return to the philo-
sophical nature of intentional actions.
8. For a concise history of the approach, see, for example, Brown (2000).

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


48 Philosophy of the Social Sciences

as an attempt to provide empirical support for the theory of collective


intentionality.
Michael A. Hogg and Dominic Abrams characterise the social identity
approach as follows:

The central tenet of this approach is that belonging to a group (of whatever size
and distribution) is largely a psychological state which is quite distinct from
that of being a unique and separate individual, and that it confers social iden-
tity, or a shared/collective representation of who one is and how one should
behave. It follows that the psychological processes associated with social iden-
tity are also responsible for generating distinctly ‘groupy’ behaviours, such as
solidarity within one’s group, conformity to group norms, and discrimination
against outgroups. (1988, 3)

The three core elements of this theory are captured accurately by the title
of an article by John Drury and Steve Reicher, “Collective Action and Psy-
chological Change: The Emergence of New Social Identities” (2000). The
social identity approach holds that truly collective action is such that it
involves a psychological change in the agents performing the action. The
change is taken to be that of social identification. Hogg and Abrams (1988, 7)
emphasise that an individual’s identification with a collective results in a psy-
chological state that is “very different from” or even “quali[ta]tively distinct
from” standard (individual-mode) psychology behind individual actions. In
social action, agents identify themselves with a collective, forming thus a
social group, and “[t]hese processes create identity and generate behaviours
which have a characteristic and distinctive form, that of group behaviour”
(Hogg and Abrams 1988, 17). Moreover, as Hogg and Abrams put it, “social
identity theorists posit a switch of identity in the group (from personal to
social)” (p. 153). Drury and Reicher (2000, 581) specify this by emphasising
that “individuals in crowds do not lose their identity but rather shift from
behaving in terms of disparate individual identities to behaving in terms of a
contextually specified common social identity.”
My point is, of course, that when the social identity theorists explain what
their view amounts to, their descriptions are almost identical with the stan-
dard characterisations of the collective intentionality theory. Accordingly, in
my view, the switch that the social identity theorists conceptualise fashion-
ably in terms of altering identities and identifications is nothing else than the
capacity to act and deliberate irreducibly both in the individual mode and in
the collective we-mode. Social identity theorists think that social identifica-
tions (and thereby collective we-mode intentionality) and the corresponding
truly social behaviour are real phenomena that are qualitatively different

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 49

from individual-mode phenomena and irreducible to them. It seems plausi-


ble to think that if the social identity theory is well confirmed by empirical
evidence, this evidence speaks also for the theory of collective intentionality.
Recall that collective action dilemmas provide a major theoretical motiva-
tion for the collective intentionality theory. Similarly, Marilynn B. Brewer
and Sherry K. Schneider (1990), for example,9 argue that such social dilem-
mas provide also important empirical motivation for the social identity
approach (and thus, I should add, for the theory of collective intentionality).
In particular, Brewer and Schneider argue that the dogma of individualism
“makes it difficult to account for the high level of apparently voluntary social
co-operation evidenced in both field and laboratory studies of social dilem-
mas” (Brewer and Schneider 1990, 170). Their conclusion is that coopera-
tion in collective action dilemmas is due to individuals identifying them-
selves with their collective (cf. Anderson 2001)—or to adopting the we-
mode. Correspondingly, if the agents hold on to their individualistic perspec-
tives, the mutually beneficial cooperative outcome remains unreachable.
The connection between (social) identification and the tendency to coop-
erate in collective action dilemmas is argued to be empirically well docu-
mented by the so-called minimal group studies characteristic of the empirical
experiments of social identity theorists. For the present purposes, the follow-
ing rough summary of the results of minimal group studies suffices. If indi-
viduals are provided with conceptual tools, no matter how trivial,10 with
which to classify agents to “us” and “them,” this is sufficient for promoting
cooperative attitudes towards the members of the we-group (and competitive
attitudes towards the nonmembers) even in tasks completely unrelated to the
underlying classifications:

What emerges from the minimal group studies is that social categorization—
the discontinuous classification of individuals into two distinct groups—is suf-
ficient to generate intergroup competition. From the social identity perspec-
tive . . . we would wish to argue that individuals in minimal group studies are
categorizing themselves in terms of the social category provided by the
experimenter. (Hogg and Abrams 1988, 51)
Once group identification has been established, intragroup orientations are
characterized by the best of human motivations: perceived mutuality, co-oper-
ation and willingness to sacrifice individual advantage for the sake of group
goals. (Brewer and Schneider 1990, 178)

9. See also Klandermans (2000) and Kerr and Park (2001).


10. For example, the experimenter may claim, perhaps counterfactually, that the individuals
can be divided into two groups according to their musical taste or something similar.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


50 Philosophy of the Social Sciences

I take it as obvious that these results are highly compatible with the theory
of collective intentionality in general and Searle’s (1990, 406) assertion that
adopting the we-mode is a precondition of all true cooperation in particular
(see also Tuomela 2000).

4. The Empirical Approaches: A Preliminary Conclusion

So what do the empirical research programmes discussed in this article


imply vis-à-vis the theory of collective intentionality in general and its com-
mitment to anti-individualism in particular? In the case of experimental eco-
nomics, the answer appears to be clear. Intentional action is not found out a
posteriori to be in the individual mode. Instead, intentional actions are
defined a priori to be individual-mode actions. Possible counterexamples are
defined to be irrelevant, for to the extent that they cannot be redescribed in
terms of individual-mode instrumental rationality, they either are not consid-
ered to be intentional actions at all or are claimed to misrepresent the true
preferences.
Obviously this kind of definitional ostrich strategy of defining a priori all
action to be in the individual mode cannot be used as an empirical argument
against the theory of collective intentionality. On the contrary, it could be
argued that by offering a more sophisticated conceptual framework for
understanding intentional actions, the collective intentionality theory can
assist experimental economists to overcome the artificial conceptual restric-
tions of their research programme.
The social identity approach in social psychology, in contrast, is obvi-
ously highly compatible with the collective intentionality theory. Compati-
bility, however, is not enough. We are interested in whether or not the social
identity approach can actually offer independent empirical support for the
theory of collective intentionality. This is where problems emerge. After all,
the social identity approach is a theoretical approach in social psychology.
Many of the examples that Hogg and Abrams discuss are such that, say, a
mainstream rational choice theorist could—and no doubt would—interpret
the examples as compatible with and indeed supportive of the individualist
rational choice paradigm. In particular, where social identity theorists see indi-
vidual and social identifications as instinctive moves prior to any rationalisable
choices, a mainstream individualist would interpret a social identification as
based on a more fundamental (individual-mode) pro-attitude towards the
benefits of the group the agent sees herself as belonging to.
Thus, I think one cannot say that the social identity theorists have discov-
ered collective intentionality to be a real phenomenon. Rather, similarly to

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 51

experimental economists, they have managed to construct a theoretical, con-


ceptual framework in terms of which social action can be relatively success-
fully understood, explained, and predicted. The experimental economists
and other mainstream rational choice theorists offer an alternative frame-
work. It seems that direct evidence does not distinguish either as more faith-
ful to the incontestable facts. One could, of course, argue that social identity
theorists have the stronger case, if only because their pluralism can accom-
modate the successes of the individualistic camp (since they recognise also
actions performed under individual identification) while being able to
acknowledge also cases of collectively rational behaviour (performed under
a social identification, or in the we-mode) without postulating metalevel pro-
attitudes. However, dedicated individualists would hold that it is precisely
their ability to account for collectively rational behaviour in terms of group-
directed but individual-mode preferences, that is, without an appeal to two
different modes of behaviour, that tips the scales in favour of the economists’
approach.
In conclusion, since my willingness to write off the empirical weight of
the arguments in favour of individualism given by mainstream economists
was due to the discovery that the economists simply construct a possible the-
oretical framework for conceptualising human actions, it seems that in the
name of intellectual honesty I must admit that precisely the same objection
can be directed to the social identity theorists. It is difficult to even imagine
what kind of empirical evidence could count as a decisive test between the
two frameworks. Instrumentalist as it may sound, one cannot but feel that it is
almost a mere matter of taste or practical applicability that determines which
model one prefers.11
Direct, empirical methods fail to settle the issue unambiguously. In a
sense this conclusion, even if not fully satisfactory, might nonetheless be suf-
ficient for the purposes of the philosophical theory of collective
intentionality. We already have strong philosophical arguments in favour of
the collective intentionality theory (given by Gilbert, Searle, and Tuomela),
and in the social identity approach in social psychology we have at least one
interesting empirical scientific research programme that could bring the
philosophical theory of collective intentionality in touch with empirical
social sciences in a fruitful way. My interpretation of the social identity
approach offers, I believe, a promising route for the theory of collective

11. This is indeed often the case in social scientific research. As Barnes (1995, 32) observes,
the dominant practice in contemporary social science seems to be that a social scientist simply
chooses to apply whichever theoretical framework she considers to be most useful in light of the
interests of her empirical research.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


52 Philosophy of the Social Sciences

intentionality to break free from the abstract realm of armchair philosophy


into the harsh world of empirical social science.
However, it would be somewhat disappointing to end here. We want to
understand why direct empirical attempts to verify or falsify the theory of
collective intentionality are unsuccessful. The easiest answer would be to
appeal either to the alleged immaturity of the empirical methods in the social
sciences or to some general philosophical problem such as the underdetermi-
nation of theory by data. These lines of thought are no doubt important, but in
this article I wish to examine a different possibility. In short, let us return to
philosophy and look at the philosophical nature of intentional action and its
explanation. Perhaps it will turn out that the question of the mode of inten-
tional action is not the kind of question that even in principle can be answered
by empirical12 studies. Maybe meaningful events, such as motivations, rea-
sons, and actions, are not description-independent events that have intrinsi-
cally the collective or individual mode independently of the theoretical
frameworks we use for talking about them.
In this article I cannot discuss all aspects of contemporary action theory.
Rather, I concentrate on one particularly influential tradition in philosophy
of action and intentional explanation, namely, the programme rising from the
works of G. E. M. Anscombe and Donald Davidson. I aim to show that if one
subscribes to the Davidsonian line of thought,13 one will also understand why
it is indeed inevitable that empirical approaches to intentional action and its
explanation end up simply creating different conceptual frameworks relative
to which we can understand actions. As Davidson likes to put this (recall the
title of Davidson [(1974) 1980], “Psychology as Philosophy”), the nature of
intentional action is always a philosophical and normative question rather
than a scientific problem. Somewhat surprisingly, however, I also show how
this line of thought actually forms a novel way of defending the theory of col-
lective intentionality. However, in this article I cannot discuss the philosophi-
cal foundations of the Davidsonian programme. Thus, my conclusion must

12. To anticipate my argument, perhaps it is somewhat misleading to say that the line of
thought I am about to introduce holds that there is no empirical fact of the matter concerning the
nature of intentiongal action. The suggestion is, rather, that we ought not to search for a naturalis-
tic (nonnormative) fact about theory-independent preferences of individuals but a social fact con-
cerning the norms governing rational action. The question of the nature of intentional action
would be shown to be an empirical question that belongs to the field of interpretative sociology
rather than to naturalistic psychology. Thanks to Philip Pettit and Lorenzo Bernasconi for helpful
discussions on this matter.
13. My interests are systematic rather than exegetical. Consequently, what I argue to be the
proper understanding of the lasting core of Davidson’s argumentation may well be different from
how Davidson himself understood the implications of his work and what people often refer to as
Davidson’s view (Saaristo forthcoming).

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 53

remain conditional: if one’s views on intentional action follow the reasoning


of Anscombe and Davidson, then the failure of the empirical approaches is
precisely what one must expect.

5. Back to Philosophy: Davidson on Meaningful Action

Let me begin with the platitude that the object of social science is action in
the sense of meaningful behaviour as opposed to mere bodily movements
(Weber [1922] 1947, 88). This is customarily taken to mean that the social
sciences study behaviours to the extent they embody intentions, that is, to the
extent that agents aim to achieve something with their behaviours. In short,
the social sciences study actions, and a behaviour is an action if and only if
reasons can be given and asked for it. This means, as Anscombe (1959) puts
it, that actions are essentially actions “under a description.”
To use Anscombe’s example, when we observe a sequence of physical
movements of a man and ask, “What was he doing?” there may well be more
than one true answer. The man was moving a lever up and down. He was
manually pumping water into the cistern of a house. He was pumping poi-
sonous water into the house where a meeting was held. He was poisoning the
evil men in the house. Following Davidson (1980), we can say that acting
intentionally is acting with some intention in the sense that there is a descrip-
tion under which one intends to act—a description of what one intends to do
with one’s performance. Hence, one may act intentionally under description
A so that one performs also an action under description B, even though one
does not intend to act under description B. In Anscombe’s view (see also
Davidson [1971] 1980, 44), for one’s behaviour to be an intentional action,
there must be a description such that one intended to act under that descrip-
tion. The central role of descriptions in action theory highlights the concep-
tual element inherent in all intentional actions. Conceptual descriptions are
largely constitutive of intentional actions qua actions.
This insight plays a crucial role also in Hacking’s line of thought when he
develops his dual thesis of the looping effect of the social sciences and how
we “make up” people. Hacking’s (e.g., 1995, 235) central observation is that
if intentional action indeed presupposes a description under which the act is
performed, then it follows that one can intentionally perform only the kinds
of actions one can contentfully describe. Consequently, “When new descrip-
tions become available . . . then there are new things to choose to do” (Hack-
ing 1995, 236). Conceptual constructions constitute new possibilities of
action. Moreover, one ought not to think that the actions (the very possibility
of which is due to such conceptual constructions) are somehow less real,

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


54 Philosophy of the Social Sciences

since all intentional actions depend on the availability of conceptual contents


under which the actions can be intentional.
However, in Davidson’s picture it is not only the contents or types of
intentional actions that are conceptual constructions. Rather, the very event-
type of intentional action itself is bootstrapped into existence in language
games. Intentional actions do not belong, as it were, to the natural furniture of
the mind-independent world. In the physical world as such, there are only
non-intentional events that feature in naturalistic relations of causes and
effects, governed by the causal laws of nature. However, some such natural
events can be redescribed in terms of the intentional types of reasons for
action (especially desires, beliefs, and intentions) and actions (events for
which reasons can be asked and given). Such intentional types receive their
distinctive character from standing in normative, rational relations to one
another. The rules of rationality are constitutive of reasons and actions (so
described). In the Davidsonian picture, physical events can be intentional
events, such as reasons and actions, only if the events can be embedded via
intentional redescriptions into a web of normative (rational) relations
essentially different from causal laws.
Davidson puts this as follows:

My point is that if we are intelligibly to attribute attitudes and beliefs, or use-


fully to describe motions as behaviour [intentional action, in my terminology],
then we are committed to finding, in the pattern of behaviour [action], belief,
and desire, a large degree of rationality and consistency. (Davidson [1974]
1980, 237)

In short, “the content of a propositional attitude [including intentions that


turn mere behaviour into action] derives from its place in the pattern [of ratio-
nal relations]” (Davidson [1970] 1980, 221), and “the satisfaction conditions
of consistency and rational coherence may be viewed as constitutive of the
range of applications of such concepts as those of belief, desire, intention and
action” (Davidson [1974] 1980, 237). Finally, the paradigmatic example of
the conditions of rationality constitutive of actions qua actions is an inten-
tional explanation of action governed by the rules of rationality connecting
reasons and actions. This allows Dagfinn Føllesdal to sum up the
Davidsonian picture as follows: “to be a desire or a belief is just to be a factor
that can figure appropriately [i.e., in accordance with the normative rules of
rationality] in reason [intentional] explanation” (Føllesdal 1994, 306), that is,
in the game of giving and asking for reasons. Similarly, an intentional expla-
nation of action makes explicit a normative framework within which there
can be reasons and actions in the first place.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 55

To recapitulate, in Davidson’s view it is the rules of rationality, such as


“the satisfaction conditions of consistency and rational coherence,” that con-
stitute, as it were, the rational game of giving and asking for reasons relative
to which there can be beliefs, desires, intentions, and actions as opposed to
mere physical brain states and nonmeaningful bodily behaviours (see Kim
2003, 119). Hence, a crucial question is where the rules of rationality come
from. Davidson’s strictly nominalistic and monistic naturalism implies that
he can appeal neither (i) to timeless universal Platonist principles of rational-
ity nor (ii) to hypothetical Cartesian laws applicable to a nonphysical sub-
stance. Thus, Davidson is left with two alternatives. Either the rules of ratio-
nality are in fact (iii) (causal) laws of nature applicable to rational beings
such as humans, or (iv) the rules are ultimately our own constructions.
Contrary to what is often claimed, Davidson rejects (iii).14 There are two
major, closely connected reasons for this. First, causal laws apply to entities
that exist independently of the law and that are connected only contingently.
Contrast this with Davidson’s view that the rational connections between
reasons and actions are constitutive of the very reasons and actions (so
described) that they connect. The rational connections between reasons and
actions are not contingent relations between self-sufficient events. Rather,
the rational interrelations between reasons and actions are, on conceptual
grounds, necessary for them to be the reasons and actions they are (Davidson
[1970] 1980, [1974] 1980).15 Second, in Davidson’s view the very feature
that distinguishes intentional events from physical events is that they follow,
so to speak, a different kind of logic. Intentional events belong to the logical
space of reasons, whereas physical events occupy the logical space of causes
and effects. Thus, the rules of rationality constituting reasons and actions
cannot be causal laws, for this would eliminate the very essence of mental

14. Elster (1985), for example, appears to think that Davidson accepts (iii). However, as I
show (below), it is a central aspect of Davidson’s thought that (iii) is not accepted (Saaristo
[forthcoming] examines Davidson’ view in more detail and argues explicitly against Elster’s
interpretation).
15. Thus, the first reason why Davidson rejects (iii) is essentially the (in)famous logical con-
nection argument (see, for example, von Wright 1971, 93ff.) against the possibility of seeing the
reasons for an action as the causes of the action. This is surprising, since Davidson is commonly
credited as the philosopher who refuted the logical connection argument (e.g., Mele 2003). This
view, although mistaken, is understandable, for (a) it is how Davidson himself saw his own work,
and (b) in Davidson’s most famous paper ([1963] 1980) on action theory, Davidson still did think
that mental entities (so described) are autonomous entities of some sort. When he developed his
mature view on the nature of the mental, however, Davidson had to admit explicitly that his 1963
view was crucially mistaken: reasons cannot be causes qua reasons but only qua physical events
(Saaristo forthcoming). In other words, rational connections connecting mental types cannot be
or even “have echo” (Davidson [1974] 1980, 231) in causal laws.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


56 Philosophy of the Social Sciences

events (so described) by placing them to the logical space of physical events,
since causal laws are the constitutive principles of that space.
In other words, when we treat behaviours as actions,

we necessarily impose conditions of coherence, rationality, and consistency.


These conditions [i.e., the rules of rationality] have no echo in physical theory
[i.e., in causal laws of nature], which is why we can look for no more than rough
correlations between psychological and physical phenomena. (Davidson
[1974] 1980, 231; italics added)

Moreover, since the normative rationality principles are constitutive of the


mental realm, “[s]hort of changing the subject, we cannot escape this feature
of the psychological; but this feature has no counterpart in the world of phys-
ics” (Davidson [1974] 1980, 79). This fundamental incommensurability of
normative psychology and nonnormative natural science, that is, the “no
echo” thesis or the rejection of (iii), lies at the very core of Davidson’s philos-
ophy of mind (mental anomalism) and his compatibilist defence of human
freedom (Davidson [1973] 1980).
Thus, Davidson is left with (iv), namely, the Wittgensteinian, con-
structivist view of seeing the rules of rationality that constitute the very pos-
sibility of actions as grounded in social practices of treating certain inferen-
tial steps as rationally acceptable, certain states as reasons and certain
behaviours as actions.16 To avoid circularity, we must say that such practices
are not ultimately based on intentional actions but, as Wittgenstein (1953,
e.g., §§199, 202, 219) puts it, on blind (pre-intentional) dispositions and a
social system of monitoring the dispositions (Saaristo 2003, 311ff.; Saaristo
2004a, 2004b). The rules of rationality are, then, seen as internal to the prac-
tice of tacitly accepting certain moves as rational. This tacit practice is essen-
tially prior to any explicit rules of rationality. Explicit formulations of the
rules are attempts to make the implicit practices explicit.

16. I think the unacceptability of (i)-(iii) and the adequacy of (iv) can be demonstrated firmly
in the context of the Wittgensteinian problem of rule-following—see, in particular, Brandom
(1994), Esfeld (2001), or, if you read Finnish, my account in Saaristo (2004b). I have also
sketched the main line of thought shortly in English in Saaristo (2004a). However, for the present
purpose of constructing a Davidsonian diagnosis of the confusing results of the empirical
approaches above, it suffices to see why Davidson rejects (i)-(iii). Nonetheless I should add that
my formulation of (iv)—and especially the way in which I use (iv), below—is admittedly essen-
tially Wittgensteinian rather than Davidsonian (cf. Williams 2000).

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 57

6. Practical Reasoning and Social Constructivism:


Making Up People
A behaviour is an action if and only if reasons can be given for it by citing a
practical inference such that the inference is treated as an acceptable infer-
ence within the social game of giving and asking for reasons. To use
Brandom’s (2000, 87) rather surprising example, in our everyday practices
we de facto accept the following inference as a valid practical inference:

A. It is raining.
B. Therefore, I shall open my umbrella.

The fact that it is raining is accepted as a good reason for opening one’s
umbrella. By contrast, according to the orthodox individualistic (which
many see as Davidsonian!) tradition in action theory, we ought to insist that
such an inference is crucially incomplete. According to this tradition, we
must assume that at least the following premises are implicitly present in the
inference:

C. I desire to stay dry.


D. I believe it is raining.
E. I believe that if it is raining, I must open my umbrella in order to stay dry.17

However, when we understand that the form of intentional explanation is


determined by our own (collective) requirements, it becomes clear that the
insistence of adding the extra premises (C)-(E) is justified only if it is our
practice to require them to be added. Consequently, it is possible that
although it is an empirical fact that “It is raining” is in general accepted as a
good answer to the question concerning my reasons for opening my umbrella
when I leave the comfort of the philosophy department and face the streets of
London, forcing all the undergraduates to study individualistic rational
choice theory and not enough philosophy may in the long run transform our
practices such that an explicit formulation of (C)-(E) is required after all

17. Note that even the orthodox individualistic picture faces the problem of an indispensable
ceteris paribus clause: we must require that I do not have any desire overriding (c), that I know
how to do (b), and so on. It seems that no matter how many such premises we add to the inference,
a clever enough philosopher could come up with a counterexample. The indispensability of the
ceteris paribus clause is a serious problem for the standard view of seeing the inference as
describing a naturalistic, causal process, since it seems that we can complete the inference only
by at some point agreeing that it is not reasonable to take into account further complicating possi-
bilities. This, of course, is highly compatible with the Wittgensteinian, social acceptance view of
practical reasoning favoured in this article.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


58 Philosophy of the Social Sciences

(recall Hacking’s looping effect). The point is, however, that in practice the
inference from (A) to (B) suffices, and since ontologically speaking the rules
of rationality reside in our practices, the inference from (A) to (B) is a rational
move. When an individualist insists on expressing the propositions (C)-(E),
she is not explicating implicit premises, but an implicit, personal (individual-
mode!) commitment to the inference from (A) to (B) (cf. Brandom 2000, 89).
However, the construction of the scope of actions (possible contents of
actions) and the very category of intentional actions itself is still only the
starting point of the constructivism inherent in the Davidsonian action the-
ory. The performers of intentional actions (so described), namely, intentional
agents qua agents, are also constructed in the social practice of giving and
asking for reasons. Even the more naturalistic self-proclaimed Davidson-
ians, such as Jon Elster, admit this explicitly: “A successful intentional expla-
nation establishes the behavior as an action and the performer as an agent”
(Elster 1985, 60, italics added). In other words, intentional agents (so
described) are partly constituted by the normative web interdefining reasons
and actions, on the one hand, and possessors of mental states and performers
of intentional actions, on the other. The practices of giving and asking for
reasons make it possible to see certain physical states as propositional atti-
tudes, certain bodily behaviours as intentional actions, and certain bodies as
intentional agents. Or, as Wittgenstein might have put it, practices establish-
ing the patterns of acceptable practical inferences constitute the human form
of life that essentially consists of intentional actions and rationally responsi-
ble agents. To be an agent is to participate in the game of giving and asking
for reasons that constitutes the human form of life.
This aspect of Davidsonian action theory forms the radical core of Hack-
ing’s theory of the looping effect of the social science. Hacking’s interesting
case study in Hacking (1995) is the multiple personality disorder. Hacking
argues that the multiple personality disorder was bootstrapped into existence
some 20 years ago. Of course, the physiological basis of the disorder has
most likely been around for ages, but the point of Davidsonian action theory
is that a physical condition is not sufficient for any type of agency, including
the type of an agent with multiple personalities. The category of agency is
conceptually tied to the game of giving and asking for reasons. Just as it is not
possible to perform intentionally an action for which there is no conceptual
description, it is also not possible to have a multiple personality disorder
unless there is a conceptual framework explicating what it is to have a
multiple disorder.
Thus, Hacking explains, when multiple personality became an officially
defined diagnosis of the American Psychiatric Association in 1980, cases of
multiple personality disorder were suddenly found in the thousands (Hack-

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 59

ing 1995, 8). Hacking argues that it is a mistake to think that there had always
been undiagnosed cases of multiple personality disorder among us, for the
existence of the description of the disorder is constitutive of the very possi-
bility of having the psychological disorder (although no doubt there had been
individuals with the same physiological condition). As Hacking puts this, the
construction of a relatively well-defined concept of multiple personality dis-
order has “provided a new way to be an unhappy person” (Hacking 1995,
236). Importantly, Hacking emphasises that this does not mean that the disor-
der is somehow not real, since all forms of intentional agency are conceptual
constructions in this sense.18 According to the Anscombe-Davidson tradi-
tion, to be an agent is to have a normatively defined status in the game of giv-
ing and asking for reasons, which means that a social practice that constitutes
the status must on conceptual grounds be seen as prior to agency.19
In other words, in Hacking’s picture, agency is not based on a metaphysi-
cal self, a Cartesian ego that exists prior and external to social practices.
When Hacking (2002b) says that in our social practices we “make up peo-
ple,” his message should be read quite literally to the extent that we mean
“intentional agents” by “people.” “My claim is that we ‘make up people’ in a
stronger sense than we ‘make up’ the world,” says Hacking (2002a, 40),
which means, I think, that we can causally transform the physical world that
ultimately exists independently of and prior to us, whereas in constructing
(in our social practices of giving and asking for reasons) the implicit rules
defining acceptable forms of practical inferences, we not only make it possi-
ble to perform new kinds of actions but also (noncausally) constitute the very
possibility for a physical body to be an intentional agent. To be a person is to
be assigned the status—a social construction par excellence (see Searle
1995)—of an agent (Saaristo 2004a).
The point that I wish to drive home is, of course, that within the
Davidsonian tradition (properly understood; see Saaristo forthcoming) in
action theory, there is no true, original form of intentional agency and inten-
tional action independent of conceptual constructions and explanatory prac-
tices interdefining what will count as intentional actions, contentful mental
states, and indeed intentional agents. If this tradition is correct, we must say
that there is no theory-independent fact of the matter of what is the true form
of intentional action and agency (so described). The essence of intentional
agency (qua agency) is largely constituted by the way in which we talk about

18. Besides Anscombe and Davidson, in the background of Hacking’s arguments lurks the
towering figure of Michel Foucault and his arguments on how new ways of being a person can be
constructed in (scientific) discourse.
19. Thus, this line of thought forms a strong argument in favour of methodological holism:
social practices are conceptually prior to individual agency.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


60 Philosophy of the Social Sciences

actions and reasons for actions. Thus, theories of such issues are not (only)
descriptive but in a fundamental sense prescriptive. They can function as
self-fulfilling prophecies: if a large enough number of people accept, say, the
economists’ account of reasons, actions, and agency, and internalise the
account sufficiently to be disposed to act accordingly, the behaviour of peo-
ple will make the theory true. This, of course, is precisely the lesson of Hack-
ing’s doctrines of the looping effect of the human sciences and our theories
making up people.

7. The Nature of Collective Intentionality:


A Constructivist Account
What does all this imply vis-à-vis the theory of collective intentionality?
It is not obvious how we can get hold of the mode of actions and agency as
opposed to the other aspects I have mentioned. Recall that the paradigmatic
example of collective intentionality and we-mode action is seen to be cooper-
ation in social dilemma situations. Individual-mode agents conceptualise a
social dilemma situation in terms of their individual preferences (that can, of
course, be benevolent) and aim to maximise their individual utility functions
by acting strategically, that is, by taking into account the (expected) behav-
iour of others when aiming to maximise personal benefits. Collective we-
mode agents, in contrast, address the situation from a collective perspective
and frame the situation in terms of what the collective must do in order to
reach the best outcome for the group, dissolving thus the whole setting of a
game, and then derive their individual roles in the collective task. Thus, the
issue concerns the difference between cooperation based on we-mode con-
siderations and cases where cooperation is based on an individual-mode pro-
attitude towards cooperation.
My suggestion is, roughly, that the status of collective we-mode action is
comparable to the case of opening the umbrella for the reason that it is rain-
ing. Suppose an agent faces a situation she recognises to be a social dilemma
situation (and where cooperation is obviously the collectively, although not
individually, optimal thing to do). The question of the status of the theory of
collective we-mode intentionality can now be formulated as the question of
what we ought to say concerning the following practical inference:

A. This is a social dilemma situation where cooperation is collectively optimal.


B. Therefore, I shall cooperate.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 61

A champion of the theory of collective we-mode intentionality will hold


that the inference from (A) to (B) captures accurately the psychological pro-
cess underlying social action. An orthodox individualist, in contrast, insists
that such an inference must be completed by adding at least the following
premise:

C. I desire to realise collectively optimal outcomes.

Or, alternatively,

C´. I desire to do what is right.


D´. I believe that in a social dilemma situation, the right thing to do is to realise the
collectively optimal outcome.

It seems to be a robust empirical fact that people normally do cooperate


and expect others to cooperate in social dilemma situations all over the world
(e.g., Henrich et al. 2001). In other words, the inference from (A) to (B)
makes explicit an implicit normative practice highly characteristic of human
societies. Many philosophers would even say that analyses of cultural evolu-
tion should make us expect that societies will generally evolve to include an
implicit norm or practice that could be made explicit as a commitment to the
practical inference from (A) to (B) (e.g., Skyrms [1996] argues for precisely
this view). And since the ontological home of the rules of rationality govern-
ing acceptable forms of practical inferences lies precisely in social practices
(i.e., the rejection of (i)-(iii)), the theory of collective intentionality appears
to be well confirmed.
However, as we saw earlier, arguably also the individualist can concep-
tualise coherently the empirical findings. The inference from (A) to (B),
when completed with (C) or (C´) and (D´), seems to be an alternative way of
making explicit the practice of cooperation in social dilemma situations and
the individual commitment to the practice. Adding the extra premises would
reduce collective intentionality to more fundamental individual-mode
intentionality and individual-mode preferences.
Which conceptualisation of the implicit reasoning behind cooperative
action is ultimately correct—the direct move from (A) to (B) or a detour via
(C) or (C´) and (D´)? The Davidson-Hacking perspective to this problem
suggests that maybe there is no practice-independent answer to our question.
According to this tradition, practical inferences are governed by the rules of
rationality that constitute brain states as intentional states, bodily behaviours
as meaningful actions, and physical bodies as rational agents. Ontologically
speaking, these rules reside in our collective practices of simply treating

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


62 Philosophy of the Social Sciences

some inferences as valid and other inferences as invalid (recall Carroll 1895).
Fundamentally, the question boils down to this: do we in practice require
agents to cooperate for the sake of common good, or do we require them to
contemplate whether such sociality is what really promotes their personal
projects in the situation at hand? Thus, which rationalisation of cooperative
action is correct and, therefore, whether the theory of collective
intentionality is true depend largely on whether stating (A) is in practice
accepted as the final answer in the game of giving and asking for reasons for
(B), or is stating (C) or (C´) and (D´) required as well? Thus, if the
Davidsonian line of thought is correct, it is indeed up to us to make either the
individualistic or the collective we-mode theory of social action true by
accepting one or the other rationalisation in our practices.
Thus, also in the Davidson-Hacking framework the question of the reality
of collective we-mode intentionality is, in a sense, an empirical question after
all. Nevertheless, it is not a question about self-sufficient preferences or
description-independent mental states behind observable behaviour. Rather,
it is a question about normative social practices: is the inference from (A) to
(B) generally accepted as a valid practical inference, or do we in fact require
further premises to be added? Although I think the collective we-mode the-
ory does pass this test in most real-life cases (and empirical research seems to
support this; see, e.g., Kerr and Park 2001), the choice between the compet-
ing approaches to social action is not merely a question of descriptive accu-
racy. Rather, we must also ask which form of agency we ought to promote
and, perhaps, make true. In other words, we should take very seriously Hack-
ing’s dual thesis of the looping effect of the social sciences and our social
practices making up people. Since individuals are very capable of internalis-
ing the rules of rationality, becoming thereby disposed to behave in accor-
dance with them, the choice of fundamental psychological theories is also a
choice of what kind of agents we would like the society to consist of. Do we
value agents acting essentially qua group members, or more autonomous,
disparate individuals who, as it were, always look after number one? Do we
want to live with neighbours who ask, “What’s best for us?” or “What’s in it
for me?”20 The human sciences—indeed, the moral sciences—are essen-
tially normative sciences.

20. I am not suggesting that one or the other is the morally and politically obviously superior
line to take. Indeed, history teaches us that both approaches have the capacity for devastating as
well as encouraging manifestations. Rather, the point is to acknowledge that social scientific
practice is never neutral regarding such political and moral issues (cf. C. Taylor 1985).

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 63

8. Conclusion

The theory of collective intentionality that builds on the notion of we-


mode psychology paints a picture of intentional human action and agency
very different from the mainstream individualism that dominates the social
sciences and that builds its model of social action on strategic individual-
mode actions. However, empirical approaches to social action have repeat-
edly failed to settle unambiguously the debate between holism and individu-
alism in this sense.
In this article, I have searched for a philosophical explanation of the dead-
lock. I have shown that if one takes seriously the very prominent line in philo-
sophical action theory running from Anscombe, via Davidson, to Hacking,
the situation is precisely what one ought to expect. In this picture, actions qua
actions and agents qua agents exist only relative to a normative conceptual
framework—the game of giving and asking for reasons—interdefining what
counts as reasons, actions, and agents. The theory of collective intentionality
and mainstream individualism constitute different rules for playing the game
and, therefore, different ways of being an agent, performing intentional
actions, and possessing mental states. Action and agency are always consti-
tuted within the framework of giving and asking for reasons, and thus in the
Davidson-Hacking view it makes no sense to ask which approach captures
the original (independent of our practice of playing the reason game) form of
agency.21 Constructivist and relativist as it may sound, we must choose which
one we prefer to make true in our practices. This is the core of the looping
effect of social kinds that Hacking emphasises; social scientific theories are
partly constitutive of their own objects. The choice between them is largely
normative or even political.

21. Therefore, the strong constructivism of Davidson and Hacking implies that any internally
sufficiently coherent way of interdefining reasons, actions, and agency could in principle boot-
strap new ways of being an agent and performing intentional actions into being, as Hacking’s
analysis of the multiple personality disorder indeed testifies. In the context of this article, this
kind of constructivist coherentism can be appreciated, but as Davidson is well aware of (see espe-
cially Davidson 1983), coherentism threatens to lead to rather unattractive consequences in other
contexts. However, addressing the consequences of Davidson’s coherentism in other contexts
goes well beyond the scope of this article. Thus, I must restrain myself and simply refer the reader
to Davidson’s own later works and, in particular, to McDowell (1994), which gets its inspiration
precisely from this problem (see also Brandom 1994).

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


64 Philosophy of the Social Sciences

References
Anderson, E. 2001. Unstrapping the straitjacket of ‘preference’: A comment on Amartya Sen’s
contributions to philosophy and economics. Economics and Philosophy 17:21-38.
Anscombe, G. E. M. 1959. Intention. Oxford: Blackwell.
Barnes, B. 1995. The elements of social theory. Princeton, NJ: Princeton University Press.
Binmore, K. 1999. Why experiment in economics? Economic Journal 109:F16-F24.
Binmore, K., A. Shaked, and J. Sutton. 1985. Testing noncooperative bargaining theory: A pre-
liminary study. American Economic Review 75 (5): 1178-80.
Brandom, R. B. 1994. Making it explicit: Reasoning, representing, and discursive commitments.
Cambridge, MA: Harvard University Press.
———. 2000. Articulating reasons: An introduction to inferentialism. Cambridge, MA: Har-
vard University Press.
Bratman, M. 1999. Faces of intention. Cambridge: Cambridge University Press.
Brewer, M. B., and S. K. Schneider. 1990. Social identity and social dilemmas: A double-edged
sword. In Social identity theory: Constructive and critical advances, edited by D. Abrams
and M. A. Hogg. Hemel Hempstead, UK: Harvester Wheatsheaf.
Brown, R. 2000. Social identity theory: Past achievements, current problems and future chal-
lenges. European Journal of Social Psychology 30 (6): 745-78.
Carroll, L. 1895. What the tortoise said to Achilles. Mind 4 (14): 278-80.
Cubitt, R. P. 2005. Experiments and the domain of economic theory. Journal of Economic Meth-
odology 12 (2): 197-210.
Cubitt, R. P., C. Starmer, and R. Sugden. 2001. Discovered preferences and the experimental evi-
dence of violations of expected utility theory. Journal of Economic Methodology 8 (3): 385-
414.
Davidson, D. [1963] 1980. Actions, reasons, and causes. In D. Davidson, Essays on actions and
events, 3-19. Oxford: Oxford University Press.
———. [1970] 1980. Mental events. In D. Davidson, Essays on actions and events, 207-27.
Oxford: Oxford University Press.
———. [1971] 1980. Agency. In D. Davidson, Essays on actions and events, 43-61. Oxford:
Oxford University Press.
———. [1973] 1980. Freedom to act. In D. Davidson, Essays on actions and events, 63-81.
Oxford: Oxford University Press.
———. [1974] 1980. Psychology as philosophy. In D. Davidson, Essays on actions and events,
229-39. Oxford: Oxford University Press.
———. 1980. Essays on actions and events. Oxford: Oxford University Press.
———. 1983. A coherence theory of truth and knowledge. In D. Davidson, Subjective, inter-
subjective, objective, 137-53. Oxford: Clarendon.
Drury, J., and S. Reicher. 2000. Collective action and psychological change: The emergence of
new social identities. British Journal of Social Psychology 39:579-604.
Elster, J. 1985. The nature and scope of rational-choice explanation. In Actions and events: Per-
spectives on Donald Davidson, edited by E. LePore and B. McLaughlin, 60-72. Oxford:
Blackwell.
Esfeld, M. 2001. Holism in philosophy of mind and philosophy of physics. Dordrecht: Kluwer.
Frank, R. H., T. Gilovich, and T. Regan. 1993. Does studying economics inhibit cooperation?
Journal of Economic Perspectives 7 (2): 159-71.
Føllesdal, D. 1994. The status of rationality assumptions in interpretation and in the explanation
of action. In Readings in the philosophy of social science, edited by M. Martin and L. C.
McIntyre, 299-310. Cambridge, MA: MIT Press.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


Saaristo / No Escape from Philosophy 65

Gilbert, M. 1989. On social facts. Princeton, NJ: Princeton University Press.


Hacking, I. 1995. Rewriting the soul: Multiple personality and the sciences of memory. Prince-
ton, NJ: Princeton University Press.
———. 1999. The social construction of what? Cambridge, MA: Harvard University Press.
———. 2002a. Five parables. In I. Hacking, Historical ontology, 27-50. Cambridge, MA: Har-
vard University Press.
———. 2002b. Making up people. In I. Hacking, Historical ontology, 99-114. Cambridge, MA:
Harvard University Press.
Hausman, D. M. 2000. Revealed preference, belief, and game theory. Economics and Philoso-
phy 16:99-115.
———. 2005. Testing game theory. Journal of Economic Methodology 12 (2): 211-23.
Henrich, J., R. Boyd, S. Bowles, C. Camerer, H. Gintis, R. McElreath, and E. Fehr. 2001. In
search of Homo economicus: Experiments in 15 small-scale societies. American Economic
Review 91 (2): 73-79.
Hogg, M. A., and D. Abrams. 1988. Social identifications: A social psychology of intergroup
relations and group processes. London: Routledge.
Kagel, J. H., and A. E. Roth, eds. 1995. The handbook of experimental economics. Princeton, NJ:
Princeton University Press.
Kerr, N. L., and E. S. Park. 2001. Group performance in collaborative and social dilemma tasks:
Progress and prospects. In Blackwell handbook of social psychology: Group processes,
edited by M. A. Hogg and R. S. Tindale, 107-38. Oxford: Blackwell.
Kim, J. 2003. Philosophy of mind and psychology. In Donald Davidson, edited by K. Ludwig,
113-36. Cambridge: Cambridge University Press.
Klandermans, B. 2000. Identity and protest: How group identification helps to overcome collec-
tive action dilemmas. In Cooperation in modern society: Promoting the welfare of communi-
ties, states and organizations, edited by M. van Vugt, M. Snyder, T. R. Tyler, and A. Biel,
162-83. London: Routledge.
McDowell, J. 1994. Mind and world. Cambridge, MA: Harvard University Press.
Mele, A. R. 2003. Philosophy of action. In Donald Davidson, edited by K. Ludwig, 64-84. Cam-
bridge: Cambridge University Press.
Plott, C. R. 1996. Rational individual behaviour in markets and social choice processes: The dis-
covered preference hypothesis. In The rational foundations of economic behaviour: Pro-
ceedings of the IEA Conference held in Turin, Italy, edited by K. J. Arrow, E. Colombatto, M.
Perlman, and C. Schmidt, 225-54. Houndmills, UK: Macmillan.
Saaristo, A. 2003. On the objectivity of social facts. ProtoSociology 18-19:291-316.
———. 2004a. Personhood as a social status. In Personhood, edited by H. Ikäheimo, J.
Kotkavirta, A. Laitinen, and P. Lyyra, 192-98. Publications in Philosophy 68. Jyväskylä, Fin-
land: University of Jyväskylä Press.
———. 2004b. Merkitys sosiaalisena käytäntönä: kohti ihmistieteiden filosofisia perusteita
(Meaning as social practice: Towards the philosophical foundations of the human sciences).
Ajatus 61:81-114.
———. Forthcoming. Intentional action and the limits of scientific naturalism: Davidson’s
alleged refutation of the logical connection argument. In Science: A challenge to philoso-
phy?, edited by S. Pihlström, H. J. Koskinen, and R. Vilkko. Frankfurt: Peter Lang.
Searle, J. R. 1990. Collective intentions and actions. In Intentions in communication, edited by
P. R. Cohen, J. Morgan, and M. E. Pollack, 401-15. Cambridge, MA: MIT Press.
———. 1995. The construction of social reality. London: Penguin.
Schmid, H. B. 2004. Personhood and the structure of commitment. In Personhood, edited by H.
Ikäheimo, J. Kotkavirta, A. Laitinen, and P. Lyyra, 199-211. Publications in Philosophy 68.
Jyväskylä, Finland: University of Jyväskylä Press.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016


66 Philosophy of the Social Sciences

Skyrms, B. 1996. Evolution of the social contract. Cambridge: Cambridge University Press.
Taylor, C. 1985. Social theory as practice. In Philosophy and the human sciences. Cambridge:
Cambridge University Press.
Taylor, M. 1987. The possibility of cooperation. Cambridge: Cambridge University Press.
Tollefsen, D. P. 2004. Collective intentionality. Internet encyclopedia of philosophy. http://
www.iep.utm.edu (accessed August 19, 2004).
Tuomela, R. 1995. The importance of us: A philosophical study of basic social notions. Stanford,
CA: Stanford University Press.
———. 2000. Cooperation: A philosophical study. Dordrecht: Kluwer.
———. 2002. The philosophy of social practices: A collective acceptance view. Cambridge:
Cambridge University Press.
Weber, M. [1922] 1947. The theory of social and economic organization, vol. 1 of Wirtschaft und
Gesellschaft, translated by A. M. Henderson and T. Parsons. New York: Free Press.
Williams, M. 2000. Wittgenstein and Davidson on the sociality of language. Journal for the The-
ory of Social Behaviour 30 (3): 299-318.
Wittgenstein, L. 1953. Philosophical investigations, translated by G. E. M. Anscombe. Oxford:
Blackwell.
Von Wright, G. H. 1971. Explanation and understanding. Ithaca, NY: Cornell University Press.
Zaibert, L. A. 2003. Collective intentions and collective intentionality. In John Searle’s ideas
about social reality: Extensions, criticisms and reconstructions, edited by D. Koepsell and
L. S. Moss, 209-32. Oxford: Blackwell.

Antti Saaristo is a researcher in the Department of Social and Moral Philosophy at the Univer-
sity of Helsinki, where his research concentrates on the philosophical foundations of the social
sciences. He is also currently writing his doctoral thesis on social action in the Department of
Philosophy, Logic and Scientific Method at the London School of Economics and Political
Science.

Downloaded from pos.sagepub.com at Universitas Gadjah Mada on July 2, 2016

Das könnte Ihnen auch gefallen