Forthcoming in Socializing
Metaphysics, ed. Frederick Schmitt, Rowman and Littlefield.
Social
Constructionism, Social Roles, and Stability
Ron Mallon
University of Utah
Social
constructionist claims are ubiquitous in the social sciences and humanities. Even a cursory scan of the titles of
academic books or journal articles will suggest that, among other things, race,
gender, sexual orientation, the emotions and mental illness are socially
constructed.[1] While many philosophers and academics
associate the term ‘social constructionism’ with radical anti-realism, a
careful study of at least some constructionist claims suggests a different
interpretation. Here I have in
mind claims about what Ian Hacking calls human kinds: “kinds of people, their
behaviour, their condition, kinds of action, kinds of temperament or tendency,
kinds of emotion, and kinds of experience”[2] On this alternative interpretation,
many of these constructionist claims about human kinds are attempts to
undermine what I will call human nature explanations of behavior or other facts about persons by
appeal to what I call social role
explanations. Understood in this
way, constructionist explanations are a variety of causal explanation.[3] Roughly, human nature explanations
appeal to biological facts about persons to explain other facts about
them. In contrast, social role
explanations appeal to the social role that a person occupies to explain the
same facts. For example, a social
constructionist might try explaining racial difference by simultaneously
undermining explanations of racial difference that appealed to biological
differences among members of different races, and instead explaining what
differences there are in terms of the different social roles occupied by
members of different races in a racialized society.
We can juxtapose
human nature and social role explanations of individual behaviors and social
facts by considering Hilary Kornblith’s broader question, “what is the world,
that we might know it?”[4] That is, what must the world be like
such that we are able to have knowledge of it? The answer Kornblith gives is that it is a world made up of
natural kinds – mind-independent
property clusters in the world that support our attempts at induction and
explanation.[5] Such clusters support these inductive
enterprises because they are stable:
they recur across instances and through time. They therefore allow us to induce from observed instances to
others. But what about the social
world? Our day-to-day world is a
social world, a world made up of people, groups of people, conventions, and
institutions with which we interact, and must interact, in order to survive and
flourish. Such interactions are
exquisitely complex, and so it is remarkable that we are able to interact with
such astonishing success. Echoing Kornblith, we can ask, ‘what is the social
world that we might know
it?’ The answer provided by human nature explanations allows a straightforward
extension of Kornblith’s answer since human nature explanations may proceed by
invoking natural kinds that human beings instantiate.[6] Stable features of the social world may
be explained as the products of stable features of human nature. For example, the evolutionary
psychologists Margo Wilson and Martin Daly (1992) explain human sexual jealousy
(and differential jealousy patterns between the sexes) as the result of evolved
mechanisms that serve to promote particular reproductive strategies. Sexual jealousy, they claim, is exhibited
cross-culturally, and the need to manage such emotions also provides a partial
explanation of the existence and structure of other social institutions
concerned with marriage and child care.
So, if we want to understand the stability of the social world, human
nature advocates have a ready answer: the stability of the social world is
explained by stable features of human nature.
What can the social
constructionist say? One variety
of social constructionist wants to explain various features of social life by reference
to differential social roles. But,
the question is, do such social roles have the stability to explain the
constancy and predictability of the social world and our success within it? One possibility – the one I will pursue
in this paper – is to claim that social roles are natural kinds, or are
analogous to natural kinds in the ways they structure the social world, and
reference to them figures in our successful inductive and predictive
enterprises.[7] If this is correct, then the stability
of the social world (and our knowledge of it) might also be explained by the
presence of stable, causally efficacious social roles.
But this approach
runs contrary to long-running
strands of the philosophy of social sciences and social theory which says that
social roles do not have such stability.
One source of doubt about the stability of social roles arises from the
camp of constructionist social theory itself. Constructionist social theorists
routinely assert the instability of social categories. For example, in their widely read book Racial
Formation in the United States,
Michael Omi and Howard Winant argue that the "effort must be made to
understand race as an unstable and 'decentered' complex of social meanings
constantly being transformed by political struggle.” [8] Omi and Winant are not speaking here
only of the meanings
or concepts of race
but of the social roles structured by such meanings or concepts. And the reason such social roles are
unstable, they believe, is that they are not grounded in biological facts, but
rather are created via a “social and historical process.”[9] And the point of constructionist work
(or at least one point) is to destabilize the phenomena further, by drawing
attention to this instability.
Another source of pessimism about social roles is found in the general
view that all regularities of the social sciences are constituted partly by the
intentional states of actors and that because of this, phenomena such as social
roles will be highly volatile.
This worry takes many forms.
First, that human choices constitute an open-ended causal system that is
susceptible to many forms of interference and may exhibit a highly sensitive
dependence on initial conditions.[10] Or that, in particular, because of
conceptual innovation and conceptual diversity, types structured by human
intentional states will be highly volatile and local.[11] A more recent source of such skepticism
– and the one on which I will be focusing – is Ian Hacking's thesis of
“the looping effect of human kinds.”[12] This phenomena occurs when persons
classified in a certain way come to change in response to the labels placed
upon them. According to Hacking,
this looping effect dramatically destabilizes knowledge of human kinds, and it
marks a fundamental difference between the social sciences and the natural
sciences. If these philosophers
and social theorists are right, then it would seem that social roles do not
have the stability to ground our inductive, predictive and practical success in
the social world. And insofar
as we take our social success to require such a ground, some other alternative
– perhaps that provided by human nature explanations – must be found.
While human action
may be usefully individuated by the intentional states of the actors, and human
institutions may be structured by the meanings of the actors that comprise
them, neither of these facts undermines the possibility of understanding social
roles on the model of stable natural kinds. So, I’ll maintain that social roles of the sort invoked by
some constructionists may be property-cluster kinds of the same sort Kornblith
invokes to explain our cognitive success in understanding the natural world.
One part of the explanation of our social success is that we live in a world
made up of social role kinds that we simultaneously know and maintain by our
epistemic and practical activities. If this is correct, a scientific approach
to social life may also take account of social roles in the explanation of
individual behavior or group difference. Note that social role kinds are explanatory kinds in two
senses. First, they are kinds in
that there is a kind of thing, the social role, that has a variety of manifestations
but can be characterized in a general way. Second, they are kinds in that particular sorts of social
role may be explanatory kinds within a social milieu. So for example, if gender is a social role, then the property of having one or another
gender may be an explanatory kind.
A few caveats are in order.
First, to call social role kinds “natural” kinds would be, for many, to stress
the meaning of ‘natural’ beyond its breaking point. So, I will alternately call them property-cluster kinds,
explanatory kinds, or relevant kinds.
Second, to claim that social roles may be property-cluster kinds is not
to say that every social role amounts to such a kind. Because the account of social roles is independent of the
account of cluster-kinds presented, the discussion leaves open which – and to
what extent – particular social roles count as explanatory kinds. So, for example, to say that race or
gender is a social role is not by itself to say that race or gender counts as a
robust enough social role to be a relevant kind. Third, even if, for example, race or gender social roles are
robust enough to be relevant property-cluster kinds, this should not be taken
to deny that there are important differences between human nature and social
role explanations of racial or gender difference. Finally, my explanation and defense of social role kinds
does not imply that I hold that human nature explanations are bankrupt or seldom appropriate. In fact, the simplistic division
between human nature and social role explanations I have employed cannot be
sustained across a great variety of behavioral phenomena. Instead, most behavioral phenomena will
require complex explanations invoking both biological (or psychological) and
socio-cultural causes, and both situational and developmental factors.[13] I take it that what explanation (or
combination of explanations) is true of any particular case will amount to a
set of difficult empirical questions.
My present aim is not
to explore these complexities, but rather to come to a better understanding of
social roles and social role explanations as understood by social
constructionists and other social theorists. This project is important both because social roles are
implicitly invoked in widespread talk of social construction, but also because
they are claimed to have an important role to play in both social and
psychological explanation. An
adequate theory of social roles also ought to have application in a variety of
other discourses. For example,
among moral and political philosophers, social roles are also invoked with some
frequency and are widely held to raise questions about moral obligation, and
social and political identity. I
won’t delve into these accounts here, except to say that a complete working out
of these accounts ought to provide a theory of social roles, and I hope that
the present account can be adapted and extended to suit these purposes.
Here’s how I will
proceed. In Section 1, I will set
out an account of social roles.
The account will be general so as to cut a broad swath across a variety
of apparently very different sorts of social roles. Then, in section 2, I will argue that social roles may count
as property-cluster kinds, and thus figure in our successful inductive and
explanatory projects. In section
3, I will focus on the challenge to this idea posed by Hacking’s claim that
knowledge of human kinds is fleeting because of the instability created by a
‘looping effect.’ My strategy will be to show that the very mechanisms that
Hacking suggests lead to instability, may instead be sources of stability for
the social role in question. By
sketching a plausible account of stability, I hope to answer Hacking and
provide a partial answer to other critics of stable social roles as well.
1. A Preliminary Account of Social Roles
Let us distinguish a
social role from a niche. A niche is simply a causal role that may be
occupied by a thing. This causal
role is distinguished by the causal effects on the occupant and by the causal
effects of the occupant. The
number of niches is indefinitely large (as large as the number of ways of
carving up causal roles), but some niches may have characteristics that make
them particularly interesting or explanatory within a given investigation. One example of a niche is what Frank
Sulloway calls functional birth order. Sulloway
defends the claim that birth order is a predictor of personality traits,
including the “big five” personality traits – openness to experience,
conscientiousness, agreeableness/ antagonism, neuroticism (or emotional
instability), and extraversion.[14] His explanation of the source of these
differences is developmental: younger siblings must pursue different strategies
than older siblings to acquire resources since they are smaller and less
capable of achieving their ends by force.
A social role, as I use the term, is a particular sort
of niche, distinguished by its occupant being acted upon by special sorts of
conceptual or linguistic causes.
Social roles in this sense are part of the apparatus that coordinates
and structures social life within a community. Such social roles may be characterized as follows:
(a)
Social role niches are associated with a term, label or mental representation
that picks out a class of persons, and a conception of the role – a set of beliefs about the
persons so picked out.
(b)
Many elements of the conception of the role are widely shared by members of the
community in which the role exists.
(c)
At least some of these beliefs are typically action-guiding, specifying particular forms of behavior
and courses of action for one so labeled.
(d)
The beliefs are also typically action-guiding for members of the community who
engage in labeling.
(e)
The actions structured by the conception give the role its causal power.
Social
roles, but not necessarily important niches, are conceptualized by those whose
lives the social roles order, and it is because they are so conceptualized that
the role has causal power. Notice
that a shared label or conception of the niche is no part of the causal
explanation of Sulloway’s birth-order effects. Thus, birth-order niches are not social roles in the sense
considered here. This
characterization of social roles is general enough to encompass several
importantly different cases.
First, sometimes persons may be labeled because they instantiate a
biological natural kind, for example, diabetic.
On the account in question, such persons may also count as occupying the
social role of the diabetic. Other social roles operate without the
associated concept or conception picking out a biological natural kind. As we noted at the outset, putatively
natural categories like race and certain mental illness are sometimes analyzed
as social roles. The account also
encompasses a third sort of role: institutional social roles like being a U.S.
Senator or being a licensed bass fisherman. Clearly the question of whether social roles can be invoked
to ground our knowledge of the social world is concerned primarily with the
second and third sorts of cases.
In these cases, according to the constructionist, the differences among
occupants of varying social roles are to be explained at least primarily by
reference to the fact that they occupy the social role. In the second sort of case,
differential features of category members are widely believed to result from
natural (for example, biological) differences. Call such roles covert roles. In contrast, roles of the third sort –
institutional social roles – are overt since the differentiation of members of these roles is
widely recognized within a community to be the result of that community’s
practices.[15] While it is a necessary feature of
social roles in the sense considered here that they be locally conceptualized,
it is not necessary that they be known to be social roles (i.e. they may be
believed to be natural categories).
It might seem as
though the conditions on being a social role are too weak. It is counterintuitive and to some
extent misleading to think of diabetic as a social role.
Moreover, the conditions on offer are so loose that even mildly
explanatory properties like being tall, wearing clown make-up, or having a last name that begins with
the letter ‘a’ would be
accompanied by social roles on the characterization offered above. In each of these cases, there is a
conception of appropriate behavior toward the person and of the person that
structures a social role.[16] Perhaps we should add additional
conditions that restrict the class of social roles further, ruling out
instances like diabetic
and being tall.
I prefer the
weaker characterization of social roles.
It is useful to think about recognized biological kinds of persons like diabetic as occupying social roles, because in
some cases there may be a question as to whether a feature typical of such a
person (e.g. a characteristic symptom) stems from their biological make-up or
from the social role they occupy.
Such is often the case in disputes over whether particular gender
differences result from biological sex differences or social role differences. Allowing that recognized biological
natural kinds that persons instantiate may be accompanied by social roles thus
avoids prejudging particular explanatory questions. It is also useful to offer a characterization of social
roles that includes weakly explanatory social roles like being tall, wearing
clown make-up, or having
a last name that begins with the letter ‘a’. While we could
strengthen the requirements on being a social role to require that bona fide
social roles must be explanatorily important, that would be a step towards
confusing the requirements on being a social role with those on being a
relevant, explanatory kind. I
prefer to retain a weaker conception of social roles in order to keep these
issues separate. This is in part
because I think the conception of social roles may characterize interesting
phenomena even if those phenomena are not explanatorily important enough for us
to consider them genuine kinds within a particular investigatory enterprise.[17]
Remember that social
roles are explanatory kinds in two senses. First, there is a sort of thing – the social role – with a
variety of manifestations. In this
section, I’ve characterized this sort of thing in a general way that draws
together a great deal of social theory and the metaphysics of social
institutions. It is a feature of
the account that it provides a unified story about the structure of overt and
covert roles by appealing to widely held beliefs, and actions structured by
those beliefs, to explain both sets of phenomena. In the next section, I will argue that particular sorts of
social roles may be property-cluster kinds, and such arguments apply both to
covert and overt social roles.
2. Social Roles as Property-Cluster Kinds
Paradigmatic natural
kinds for philosophers are species and elements. But over the last twenty years philosophical work in, for
example, the philosophy of mind and psychology has lead to relatively wide
acceptance of the view that categories from the special sciences that are
useful in explanation and prediction count as natural kinds from the point of
view of that science (e.g. Fodor
1981, Griffiths 1999). Richard
Boyd (1988, 1992, 1999a) develops an account of the metaphysics of such natural
kinds that captures the idea that such kinds support our attempts at induction
and explanation. Boyd suggests
that in our attempts to understand the world we want concepts or terms that
pick out causally homeostatic property clusters, the elements of which are (as
a matter of contingent fact) instantiated in the world.[18] Boyd, and as I noted at the outset,
Kornblith, build upon this foundation to provide a general explanation of our
success in understanding the natural world. In this section, I will set out Boyd’s account of natural
kinds as homeostatic property-cluster kinds. I will then argue that social roles of the sort considered
in Section 1 may be property-cluster kinds of this sort.
2.1. Expanding the Franchise: The
Liberalization of Natural Kinds
Central to Boyd’s
account of natural kinds is what he calls causal homeostasis:
“Either the presence of some of the properties . . . tends (under appropriate conditions) to favor the
presence of the others, or there are underlying mechanisms or processes that
tend to maintain the presence of the properties . . . or both” (1999a,
143). Such a kind is characterized
by both,
(1)
the properties in the property cluster, and
(2) the mechanism of causal homeostasis –
the mechanism that is the source of the properties' continued co-occurrence in
the cluster.
As
Robert Wilson (1999) notes, Boyd’s account is a cluster account twice
over. First, because an individual
does not have to instantiate all the properties in the cluster in order to be a
member of the kind.[19] (Manx cats do not have tails, but they
are cats all the same.) Kind
membership, on this view, may be a partial affair. Second, because the properties of a cluster kind actually
‘cluster’ in the world. The world
is lumpy in that properties are not instantiated evenly throughout space and
time, but instead occur in clumps.
Let’s try to spell
out Boyd’s account more carefully by employing a paradigmatic philosophical
natural kind: water. Instances of
water share a variety of superficial properties (e.g. liquidity, freezing
point, etc.) and these superficial properties are clustered because they are
caused by the microstructural properties of H20. The mechanism of H20’s microstructural properties thus
explains the clustering of the superficial properties. This picture of natural kinds leads
Kripke and Putnam to view the chemical structure of water as giving its
essence.[20] On this view it is sufficient for some
stuff to be water (in any possible world) that it be an instance of H20, whether
or not it has the superficial properties.[21]
And conversely, nothing is water unless it is H20. H20, it seems, is a property-cluster kind in Wilson’s second
sense only, since instantiating the property of water looks to be an all or
nothing affair.[22]
Such property
clusters support our inductive enterprises because they allow us to draw
successful conclusions about all instances of a kind on the basis of examining
a few instances. For example, on
the basis of examining very few instances of water, we can successfully infer
lots of things about other instances of water. And this success is supported by the fact that the
properties of water are tightly clustered and causally homeostatic.
But, as Boyd points
out, natural kinds in some sciences may be causally homeostatic categories that
lack the kind of simple “essence” that chemical compounds like water have. By
way of example, he writes,
The
appropriateness of any particular biological species for induction and
explanation in biology depends upon the imperfectly shared and homeostatically
related morphological, physiological and behavioral features which characterize
its members. (1991, 142)
In
biological species, the instances of a kind instantiate (more or less) property
clusters of various sorts of features, but such instantiation is imperfect
(remember manx cats). Nonetheless,
kind terms picking out species can also figure in successful inductive
enterprises. Because species
exhibit a variety of properties that are clustered and causally homeostatic, we
can – imperfectly – induce facts about all the instances of a cluster kind
from the few that we actually examine.
Boyd’s
account is thus a prinicipled liberalization of the idea of natural kinds. This is first, because the
property-cluster account of kinds moves from all or nothing accounts of kind
membership to a cluster view on which kind membership need not involve
satisfying interesting necessary and sufficient conditions. But there is also a second sense in
which Boyd’s account liberalizes the notion of natural kinds: it expands the
sorts of properties that may be contained in the property cluster. While examples like water might lead us
to believe that properties that characterize natural kinds or their mechanisms
of causal homeostasis must be intrinsic features of members of the kind, reflection
on, for example, species suggests that relational properties may be included as
well. [23] While disputes over the correct species concept or concepts are
contentious and on-going, it is widely accepted that an adequate species
concept will take into account relational or historical features of the
organisms to be classified. So for
example, according to Mayr’s classic biological species concept, species are groups of populations that
“respond to one another as potential mates and seek one another for the purpose
of reproduction” (1984, 533).[24] The stability of species property
clusters is thus sustained by (a variety of) barriers to genetic flow. In contrast, according to phylogenetic
accounts, species depend for their identity on relations to historically
situated speciation events (on their location in a phylogenetic tree). Thus
these accounts, too, incorporate relational features into species definitions.
Boyd’s account thus incorporates his insight that these developments in the
species debate have more general implications for the theory of natural
kinds. Allowing relational
features to figure as components of causally homeostatic property clusters is a
principled liberalization
of the notion of natural kinds because it is guided by attempts to pick out
kinds of great explanatory and predictive importance. What is relevant to the ability of the kind to support
induction is that there is some stable mechanism or set of mechanisms that
causes properties to cluster in a regular way, not whether the properties in
the cluster or the mechanisms of causal homeostasis are intrinsic to the kind
members.[25] To recap: what emerges from Boyd’s
discussion of causally homeostatic kinds is that properties in a cluster may be
imperfectly shared, and also that there need be no restriction on the sorts of
properties that may be included in the cluster or the mechanisms of
homeostasis, as long as the clusters are doing the inductive and explanatory
work required of them.
Opening the door to
relational properties in property-cluster kinds also opens up the door to
properties that result from human convention. Boyd says as much:
It . . . follows that there should be
kinds and categories whose definitions combine naturalistic and conventional
features in quite complex ways. .
. . . It follows that extensions of the traditional account of natural kinds
should be appropriate just to the extent that the kinds in question are
employed for induction and explanation. (1992, 140)
There
is simply no a priori
reason why social roles of the sort discussed in Section 1 cannot figure among the important
property-cluster kinds that structure our social world. Instead, whether a
particular social role type does count as explanatory is an empirical
question. We may not want to call
such kinds ‘natural’ kinds, but insofar as social kinds are property-cluster
kinds, they may support induction and explanation, and thus deserve a place in
our best theories of the social world.
2.2.
Social Roles as Homeostatic Property Clusters
Boyd’s account of
property-cluster kinds allows us to see quite clearly why social roles may be
explanatory, relevant kinds.
Occupants of robust social roles will share a variety of interesting and
important properties worthy of notice in social life and social theory, and the
mechanism by which they share these properties is the social role that they
occupy. These clustered properties
are of at least four kinds:
(i)
Social role occupants may have greater propensity to act in certain ways, in
virtue of their currently being in a particular social role.[26] (These propensities will be
intentionally mediated by an occupant’s beliefs about their situation.)
(ii)
They may also share various causal effects of occupying the social role that do
not depend on the concurrent existence of the social role. For example, if
social role occupation causes individuals to develop skills, capacities,
proclivities or character traits, these would presumably persist in those
individuals even after the social institutions that caused them to develop have
passed away. [27]
(iii)
Social role occupants may also share dispositional properties to cause
responses in others (i.e. they may share secondary qualities). Consider a dispositional property to be
one that is manifested in certain conditions, thus
x has disposition P iff x manifests P in situation C.
A
person occupying the social role of “man” may have a disposition to cause
others to respond to the person in particular ways. The property is dispositional, because it is manifested only
in a particular social milieu, and it is a response-dependent property in that
being in a “man” social role depends on the dispositions of others to treat one
as a man. In general, social role
occupants may have lots of dispositions to produce various kinds of reactions
(judgings, approvings, disapprovings, etc.) in members of the community that
participate in the practices that constitute the social role.
(iv)
Finally, social role occupants will share various properties that comprise the
basis of ascription to the category.
For example, Senators will share the property of having been
appropriately elected, and occupants of a masculine social role will (by and
large) share characteristic male sexual features.
These
sorts of properties may be relevant to understanding social life. They are clustered in individuals
occupying social roles precisely in virtue of their occupation of those social
roles. Thus, social role occupants
instantiate property-cluster kinds where the mechanism controlling the clustering
of those properties is the social role they occupy.[28]
We noted above that homeostatic
property-cluster kinds support induction.
Because various mechanisms cause explanatorily relevant properties to
cluster in the world, and because our own conceptual apparatus leads us to pick
out these kinds with concepts and seek to reason inductively about them, we can
achieve extremely useful knowledge about the mind-independent world. For similar reasons, social role kinds
also support induction. To the extent
social roles involve a thick cluster of interesting properties that are tightly
correlated, you can successfully infer things about one social role occupant
from facts about another. Among
the properties you can engage in induction about are even those properties that
are conventionally assigned (iii above). For example, if you know that Senator
Lenkiewicz served six years between elections (and you treat ‘Senator’ as a
kind term), you can infer from that that Senator Fitzpatrick will probably
serve six years as well. Thus,
knowledge of facts about overt social role occupants comes to us in at least
two ways: first, by checking the rules to which we collectively defer (e.g. the
U.S. Constitution says that Senators serve six year terms), but also, and importantly,
by induction. Acquiring knowledge of the social world is, in this second way,
much like knowledge of the natural world.
2.3. Are Social Role Kinds Scientific Kinds?
The sorts of
mechanisms that underlie many paradigmatic natural kind property clusters are
not the same sorts of mechanisms underlying social kind property clusters. For
many important natural kinds, the mechanisms that result in property clusters
are themselves independent of human actions and thoughts. But the present point is that this
difference, as important as it is, makes no difference at all to the capacity
of social kinds to be explanatory and to support induction about the social
world, or to the capacity of the concepts of social kinds to be projectable
within a certain social context that includes such social role occupants. But there remain questions about whether social role kinds of
the sort suggested here are appropriately scientific to figure in genuine
law-like regularities. Paul Griffiths
draws a useful distinction between scope and force. Scope indicates the size of the domain
over which a law-like generalization is applicable. Force, in contrast, is a measure of the “reliability of
predictions made using that generalization.” (1999, 217). Social
role kinds might be thought to fail on both counts. Their scope is limited, since they apply only within a
particular cultural milieu (or subcultural milieu), and their reliability in
prediction is weak.
Law-like
generalizations employing social role kinds do lack the sort of scope that laws
of the basic sciences have. Such
social role kinds are what Ruth Millikan (1999) has called “historical
kinds.” They are kinds the
“predicates for which are non-accidentally projectable: there are good reasons
in nature why one member of an historical kind is like another, hence why
inductions are successful over the kind” (55). But such inductions will be limited in scope because the
“good reasons” will obtain only in a cultural setting that is structured by a
conception of the kind in question.
This limitation is at least one reason why humanist social scientists
and philosophers of social science have come to view the project of finding
social kinds as bankrupt. Taylor,
for example, writes:
The
success of prediction in the natural sciences is bound up with the fact that
all states of the system, past and future, can be described in the same range
of concepts . . . This conceptual unity is vitiated in the sciences of man by
the fact of conceptual innovation, which in turn alters human reality. (1971,
209)
The
problem, according to Taylor, is that since human reality differs fundamentally
from time to time, place to place, the concepts that accurately describe kinds
in one cultural location will be inadequate to another. But such limitations do not rule out
social role kinds any more than they rule out other historical kinds like
biological species. The test of
whether a particular social role amounts to an important kind is not whether it
obtains universally, but whether making reference to the kind is useful in
explanation and prediction. And if
what you are trying to explain or predict is itself local to a historical or
spatial social setting, then limitations on scope are unproblematic. There is also the additional
possibility that scientists or critical theorists may be able to recognize
social role types that exist across various cultural milieus. In order to be ‘nonaccidentally’
grouped together as members of the same projectable kind, such social roles
would have to have a common cause.
Such a cause could take the form of an older social role that caused two
modern descendents. Alternatively,
social roles of the same type might come about as the result of constraints
imposed by human psychology,[29]
or of some more general facts about human relations. To the extent that such cross-cultural roles exist, we
can expect the scope of generalizations involving such role kinds to be larger.
As for the
reliability of predictions made using law-like generalizations employing social
role terms, they will vary according to the strength of the causally
homeostatic mechanisms binding together different elements of the property
cluster. As with any special science,
the generalizations can be expected to be at best true, ceteris paribus.
But there seems no reason to believe, a priori, that such predictions that rely on such
generalizations must be especially weak.
We should expect to find causally homeostatic mechanisms that exhibit a
whole continuum of different strengths.
At one extreme, we can find the property clusters of the basic sciences
that have extraordinary explanatory power and that may be characterized as
having simple necessary and sufficient conditions. And at the other end of the continuum we can find clusters
with quite local mechanisms whose
explanatory power is limited to one or a few instances, and we find mechanisms
that maintain weak and quite imperfect property clusters. The variability in strength of these
mechanisms will mean variability
in the reliability of predictions making reference to them. What I have been suggesting is that
social role kinds may be much further towards the explanatory end this
continuum than is normally assumed.
Recall that both philosophers of social science and social theorists
routinely emphasize the instability of social kinds and implicitly cast doubt
on the possibility of reliable generalizations. Interpreting such instability within the idiom of Boyd’s
account of kinds, we can say that these theorists maintain that the properties
of social role kinds are not causally homeostatic – that is, the mechanisms
that are responsible for such clustering are themselves insubstantial or
transient, and so the reliability of predictions employing social role terms is
very weak. In the next section, I
will scrutinize one such claim more closely.
3. Hacking and Instability
Understanding social
roles as stable property-cluster kinds helps us understand both how constructionists
could be right that our social world is structured by social roles in ways we
do not always recognize and how it could be the case that our social world is
nonetheless stable. But the view
that social roles could be stable kinds is seen as problematic by both
philosophers of social science and many social theorists. In fact, such claims are so widespread
so as to make a survey of all of them impossible here. Instead, I will consider
one influential account of instability: Hacking’s account of the ‘looping of
human kinds’ (1995b). Hacking’s
account has the virtue of being both philosophical and social theoretical –
concerned both with the metaphysics of kinds and the nature of social
identities. I will argue, contra
Hacking, that the looping of human kinds shows how the mechanisms involved in
creating social roles may act to stabilize those social roles. Thus, my aim is both to critique
Hacking’s account of the looping effect of human kinds and to sketch an account
of the stability of social roles. To the extent the account of stability I
provide is viable, it may also provide an answer other critics of the stability
of social roles.
3.1 The Looping Effect and the Instability
Thesis
Hacking’s ‘looping
effect’ occurs when persons interact with the systems of classification they
fall under. The effect of this
looping, according to Hacking, is that knowledge of human kinds is difficult to
get and harder to keep. He writes,
“People classified in a certain way tend to conform to or grow into the ways
that they are described; but they also evolve in their own ways, so that the
classifications and descriptions have to be constantly revised.”[30]
Writing of child abuse, he is even less sanguine: “The concept of child abuse
may . . . be so made and molded by attempts at knowledge and intervention, and
social reaction to these studies, that there is no stable object, child abuse,
to have knowledge about.”[31] While Hacking’s discussion is often
framed in terms of particular kinds (e.g. child abuse, multiple personality
disorder, etc.), he frames his thesis quite generally, as one about ‘the
looping of human kinds’. Recall
that he defines these quite broadly as “kinds of people, their behaviour, their
condition, kinds of action, kinds of temperament or tendency, kinds of emotion,
and kinds of experience”.[32] This broad characterization thus
includes many things we might think of as biological kinds. For example, having an O positive blood
type or being diabetic counts as a human kind on this characterization. But if all human kinds are as unstable
as Hacking says child abuse is, then his position amounts to a general
skepticism about the possibility of successful medical science, psychiatry,
psychology, social science, and basic social knowedge.
A more charitable interpretation of
Hacking limits the claim that human kinds (and our knowledge of them) are
unstable to social role kinds – kinds of person whose differentiation is the
result of the application and internalization of culturally local conceptions
of the kind.[33] This would
include both what we earlier called ‘overt’ and ‘covert’ social roles. But the kind of instability of most
interests us is that which threatens the constructionist claim that our social
world is to be explained by reference to covert social roles rather than
biological nature – the instability of covert social roles. In fact,
Hacking’s own work focuses upon covert social roles (without explicitly
limiting himself to them) in making his case about the looping effects of human
kinds. For example, Hacking
believes that most or all of the symptoms of most sufferers of multiple
personality disorder are to be explained by appeal to the social roles occupied
by multiples and not by some underlying biological dysfunction. So, for the purpose of focusing
the discussion, I will limit the discussion of the ‘looping effect’ even
further to include just covert roles.
In doing this, I am considering the very class of roles that are of most
interest to the constructionist, and the very class of roles that Hacking
utilizes to make his case.
On this more limited
interpretation of Hacking, attempts at knowledge of covert roles are doomed to
inadequacy because the study of the phenomena will cause them to change.[34] On this reading of Hacking’s view, the
social world is a Heraclitean river, and knowledge of it is fleeting at
best. Even though Hacking limits
his discussion of looping effects to covert kinds, he clearly sees his view as
radical. He writes that looping
effects mark “a cardinal difference between the traditional natural and social
sciences” because “the targets of the natural sciences are stationary” while
“the targets of the social sciences are on the move” (1999, 108). For present purposes, if looping
effects are as pervasive and destabilizing as Hacking suggests, then it
suggests that social role kinds cannot be part of the metaphysical basis that
supports our knowledge of the social world. At the outset, I said that social
roles were explanatory and relevant kinds in two senses. First, I said there is a general kind
of thing, the social role, that may be manifested in a variety of particular
ways, either overtly or covertly.
Second, I said that particular sorts of social roles such as being
mentally ill or being a man may be explanatory kinds within a social
setting. Hacking’s view threatens
the second of these claims, since it suggests that the differential features of
occupants of social roles will not remain stable enough for terms associated
with social roles to play a useful part in explanatory and predictive theories.
Let’s try to get more
clear on just what is the ‘looping effect of human kinds’. The looping occurs when a theory held
by a community (I’ll call them the labelers) interacts with the properties of those
described by the theory (I’ll call them the labeled).
Since we are all theorists of our social world, the labeled may be
labelers as well. We can visualize
the loop like this:
The
‘b’ arrow – the arrow from those who are labeled to the labelers is an
arrow of epistemic constraint. As
the labeled class changes, the labelers will attempt to track those changes in
their theories of the class (on pain of having a false theory). What about the
other, ‘a’ arrow that links the
labelers to the labeled? This
arrow represents the labeling and differential treatment of the labeled class. Call this broad set of actions
undertaken by labelers vis-à-vis
the labeled class the regime of labeling. The ‘a’ arrow
represents the regime of labeling directed at the labeled class. We can now more carefully state
Hacking’s view as the Instability Thesis:
(IT) For (covert) human kinds, changes in
the labeled class caused by the regime of labeling cause the beliefs guiding
the regime of labeling to become untrue.
If Hacking is correct, human kinds may be
relatively stable across instances at a particular time, but over time the kind
will be destabilized by changes caused by the regime of labeling. What sorts of changes are
relevant? In discussing the
looping of human kinds, Hacking has in mind changes in persons' intentional
states and behaviors (or changes that are mediated by changes in intentional states
and behaviors). Changes in the
regime of labeling cause intentional reactions on the part of those who are
labeled. (This is what
distinguishes looping effects from just any old causal effects.) But what is the mechanism of this
causation, and how does it occur so rapidly? Work in social psychology, cognitive science, anthropology,
and sociology review a variety of ways in which persons may be affected by
their social roles, and this is hardly the place to review this literature.
Instead, I want to concentrate on the set of mechanisms that are central to
Hacking’s own account – and much
work in philosophy of social science and social theory as well – those
mechanisms involving intentional action and practical agency.[35]
3.2. Identifications
A key element of the mechanisms in
question is what K. Anthony Appiah calls identification: “the process through which an
individual intentionally shapes her projects – including her plans for her own
life and her conception of the good – by reference to available labels,
available identities” (1996, 78).
By ‘label’, Appiah means here a term that is associated with a concept
and a conception of the thing picked out by the term – just the elements I
suggested characterized a social role.
So identification is an avenue via which social role memberships may
shape persons’ projects and behaviors.
And crucially, that avenue operates via the construction of intentions
from available concepts.
Appiah’s notion of identification involves
two unclear ideas: the idea that labels play a ‘shaping role’ in intentional
projects, and the idea that some labels or identities are ‘available’. Let me
say something first about ‘availability’. In developing his account of
identification, Appiah follows earlier work by Hacking (1986, 1995), in which
Hacking notes that Elizabeth Anscombe’s insight that “all action is action
under a description” has the important consequence that,
When new
descriptions become available, when they come into circulation, or even when
they become the sorts of things that it is all right to say, to think, then
there are new things to choose to do.
When new intentions become open to me, because new descriptions, new
concepts, become available to me, I live in a new world of opportunities.
(Hacking 1995, 236)
So, for example, the possibility of
acting “as black” or “as a man” or “as a plumber” presupposes the availability
of the concepts black,
man, and plumber.[36]
Geoffrey Chaucer could not have chosen to write a beat poem at least in part because the concept of a
beat poem was not available to Chaucer.
Since these concepts are only available in certain cultural milieus, the
actions that they play a role in describing are also available only those same
milieus. The meaning of
‘available’ here is something like conceptually available, and a concept is conceptually available
if it is part of the conceptual repertoire of the actor.[37]
What does it mean for available concepts
to shape our intentional projects?
One way in which the available terms or concepts shape our intentional
projects is by constituting the descriptions under which our intentional
actions occur. Thus, if James
chooses not to cry in public, because it is a instance of his general project
of being a macho man, the concept macho is a proper part of the description being a macho man under which his action occurs. Call this sort of shaping minimal
shaping. Minimal shaping occurs whenever there
is an action, because all actions are under some or another description and
thus minimally shaped by that description.
Both Hacking and Appiah seem to have more
than minimal shaping in mind, though.
Both are concerned not just with minimal shaping, but with the way in
which a conception associated with a concept can affect one’s choice of
projects and behaviors. So, for
example, Hacking discusses both the ways in which people may act according to
the label and conception they fall under.
How do such effects occur? I’ll consider two sorts of avenues via which
the conception may act on labeled persons. First, the conception may make some behaviors causally
salient. Second, the conception may structure
the social world so as to make some behaviors strategically salient.
Earlier in his discussion of identification, Appiah alludes to both
sorts of effects when he writes,
Once
the racial label is applied to people ideas about what it refers to, ideas that
may be much less consensual than the application of the label, come to have
their social effects. But they
have not only their social effects but psychological ones as well; and they
shape the ways people conceive of themselves and their projects. (78)
Labels express concepts that figure in a
causal theory of the world. And,
insofar as the concepts pick out explanatory kinds in the world, the labels one
falls under will have some explanatory force, and this force constrains one’s
choice of projects. How does this
constraint occur? According to the
causal salience account of shaping, some labels pick out putatively natural
categories, and if one falls under that label, engaging in the ‘natural’
behavior is likely to meet with less resistance than engaging in ‘unnatural’
behavior. Appiah seems to have
something like this in mind in his discussion of psychological effects. Explaining this idea, Sally Haslanger
writes:
in
practical decision making we ought to be attentive to things’ natures. It won’t do to try to fry an egg on a
paper plate; there’s no point in trying to teach a rock how to read. Because the world is not infinitely
malleable to our wants and needs, reasonable decision making will accommodate
“how things are,” where this is understood as accomodating the natures of
things, the background conditions constraining our actions. (1993, 105)
Haslanger is writing about the way
conceptions of gender serve to regulate behavior because of the putatively
natural character of gender. And
it’s easy to see how other human kind concepts and labels might work the same
way. For example, if when someone
falls under a racial concept it suggests that some projects are natural and
others unnatural, then practical reason (together with Haslanger’s suggestion
about avoiding conflict with the natures of things) dictates that one should
take one’s natural category memberships into account. In short, a concept and conception of a kind of person makes
some sorts of projects and behaviors causally salient to the labeled. Note that the claim is not that a
person’s projects are directly shaped by their intrinsic nature, but rather
that their projects are shaped by the local theory of what their intrinsic
nature entails – a local theory embodied in a conception that makes certain
projects or behaviors salient.
Thus the shaping of intentional projects is itself intentionally
mediated.
By itself, such causal salience is
an incomplete account of the way in which regimes of labelling affect
individual projects. It’s
incomplete, because it is not merely the perceived natures of things
(understood as presocial features of the world) that guide our identifications,
but all features of
the world which seem relatively intransigent, including social features. To put it in a Durkheimian way, we need
to understand social facts as things – as exhibiting coercive power over
individual choice. It is a central
feature of everyday human life that we attempt to coordinate our actions with
others. In other words, we attempt
to decide what others will do, given various things we might do, and we choose
our actual actions accordingly.
Recognizing that the social world places strong constraints on our actions
allows us to understand a second mechanism by which social roles shape their
occupants’ projects. Insofar as a
person is a union leader, garçon de café, gang member, prom queen, or is gay, white, a man, joyful,
or mentally ill, that person will have certain permissions and expectations as
well as restrictions and prohibitions placed upon her by others. What’s more, some of these social
strictures may be enforced, with social disapproval or even violence. Because such categories often come with
a policy on the part of a community to treat certain people in a certain way,
they create a range of socially available options from which labeled persons
may choose. (Such social
transformation is what Appiah refers to as the ‘social effects’ of labeling.) Thus, behaviors may also be strategically salient.
In many contexts, strategic salience radically affects one’s choices of
intentional projects. If, for
example, you are a member of a race in a highly racialized society, or a member
of a sex in a highly gendered society, you may have little choice but to choose
your intentional projects from among those that are socially permitted.
So, social
roles affect individuals projects and intentions by delimiting the space of
what is conceivable (because one can only act in ways one can conceive), and
constraining the space of causally and strategically salient actions. How does this lead to instability? While Hacking discusses a number of
forces that may result in the instability of a social role, only he discusses
only two that bear on the instability thesis.[38] The first possibility he considers is
that members of a labeled class will organize in response to a regime of
labeling and attempt to alter it.
Those classified as homosexuals, for example, have organized and
systematically agitated to alter the regime of labeling (including the
practices of discriminatory treatment) that go with the label (e.g. Hacking
1986). Hacking documents a similar
phenomena occuring with multiple personality sufferers (1995a), and sufferers
(and the relatives of sufferers) of childhood autism (1995b, 1999). We must be
careful here. A conception can
change not for epistemic reasons, but because of political pressure placed upon
those who count as experts about a category. But the instability thesis posits a particular sort if
problem that occurs when a labeled class actually changes, rendering the old
conception obsolete. So, insofar
as the agitation of a labeled class takes the form of exhibiting alternative
sorts of behaviors (acting against the dominant conception of the type), such
agitation has the effect of undermining the accuracy of the conception of the
labeled class and is a genuine source of the sort of instability relevant to
assessing the looping effect of human kinds. Michael Walzer provides an example of such instability from
the caste system of ancient India, “A certain kind of collective mobility is
possible, for castes or subcastes can cultivate the outward marks of purity and
(within severe limits) raise their position in the social scale” (1983,
27). A second possibility is that
a new conception of a labeled class may alter those conditions that gave rise
to the behavior described by the conception. This is Hacking’s view of child abuse. As new theories of abuse become
disseminated, they create a new social context in which agents may choose
different actions (1986; 1995a, Chp. 4; 1999, Chp. 5). Abuse statistics may then rise or shift
as new people (and sorts of people) decide to abuse, or as new forms of abuse
take hold. These two possibilities
form the core of what Hacking calls the looping effect of human kinds, and both
are genuine sources of instability for a social role.
3.3. Instability and Stability
Although the
ways in which conceptions of a kind of person may change the causal and
strategic salience of particular sorts of act are undoubtedly only part of the
causal story of social role effects on persons, they seem the best
interpretation of what Hacking thinks is creating instability. But now we are in a position to see the
Instability Thesis as unmotivated and probably wrong. Hacking provides little reason to think
that such regimes of labeling must always be so causally efficacious as to
undermine the associated conception or theory as an instrument of explanation
and prediction. Hacking gives us a picture of how instability might work, but
no illumination to why it should always work, or in what cases it does work. To
recognize that social roles may be unstable is not even to show that they
usually are.
Begin with causal
salience. Suppose the going
conceptions of x do
make some actions and life options causally salient for those who fall under
the associated concept for x. Then a person who falls under a
particular concept might choose to pursue one project rather than another
because they believe it would be more ‘natural’ for someone like them. But this should have the effect of
making those who fall under such a label conform to the conception, rather than deviate
from it. And thus, the conception
might come to describe accurately a world that is of its own making. Thus, the possibility that conceptions
make certain actions causally salient provides no reason to believe the
Instability Thesis.
What about strategic
salience? Hacking’s own arguments
for instability resulting from the looping of human kinds stem from shifts in
the strategic salience of certain actions. But strategic salience may also
result not in deviance but conformity. This mechanism affects labeled individuals via the
community’s behavior towards the labeled class. And this behavior will be guided by the conception of the
labeled class. To choose a simple
example, if Norton were to wear a skirt to his workplace at the bank, instead
of trousers, he would likely face snickers, jokes, jeers, and perhaps even
disgust on the part of some. This
communal behavior would be guided by a conception of gender appropriate
clothing for a bank teller. And on
the street outside the bank, he might face worse yet. These are facts about the social world Norton inhabits, and
they conspire to keep him in trousers.
Because Norton plans his dress against the background of his other
preferences and the stable social world he inhabits, he seldom seriously thinks
about deviating from even this
minor norm. Moreover,
because deviance from this minor norm is rare, false or unreflective beliefs
about this gender-marked behavior (and the gender category it is associated
with) face little pressure for revision. Such
characteristics may amount to elements of property-cluster kinds to the extent
that whole sets of such behavioral (or behaviorally mediated) characteristics
are associated in a conception, and are reliably preferred under prevailing
social conditions. For example, if
this one gender norm is associated with many others, including other norms for
how to dress, speak, walk, think and more generally act, and if these norms are
enforced on those who fall under gender labels, then they could have the effect
of causing a variety of properties to cluster in social role occupants.
The idea is that
social roles may be maintained by the strategic individual choices of actors in
a community. This may seem surprising,
but once we consider that the relevant mechanisms operate via individual
actors’ rational choices, it should not be. For what I am suggesting is that social roles might be
sustained by a sort of equilibrium common in examples from game theory. Game theory makes clear the way
elaborate, multi-person games may be sustained by the rational choice of
individuals. The present
suggestion is that we can understand certain effects of social roles (e.g.
systematic behaviors on the part of social role occupants) as being in
equilibrium with the actions of labelers that sustain the social role. Each party participates in the acts as
they do because to deviate unilaterally from doing so would result in lower expected
utility, given what every other person is doing.[39] That such game theory can be drawn upon
to provide a theory of stable conventions is something David Lewis’s pioneering
work established decades ago.[40] Unfortunately, there is not room here
to consider in detail how game theory or Lewis’s theory can be applied social
roles, but we can at least state a sufficient condition for the stability of a
particular feature of the social world caused by the presence of a social role:
F
is a stable feature of members of labeled group L in a community C that employs
a regime of labeling R if:
1)
Members of C prefer to employ R as long as they believe the conception that
figures in R picks out L and correctly ascribes F to members of L.
2)
Members of C do believe that the conception that figures in R correctly
ascribes F to members of L.
3)
Members of group L prefer to act so as to maintain F as long as members of
community C employ R.
Thus,
to return to our earlier example, trouser wearing among men is stable as long
as:
T1)
Members of the community prefer to maintain that men wear trousers and punish
exceptions, as long as they believe it is the case that men generally wear
trousers and exceptions are unusual and punishable.
T2)
Members of the community do believe that men generally wear trousers and exceptions
are unusual and punishable.
T3)
Men prefer to wear trousers as long as deviance from this regularity is
punished.
Strictly
speaking, this sufficient condition isn’t met, since not all men wear trousers,
and not all nontrouser-wearing by men is treated as abnormal. (I assume the condition could be
modified to be a more plausible description of an equilibrium in the real
world.) But what the condition
does is show the way that a social role and its causal effects may be kept in
equilibrium by the preferences and approximate rationality of participants in
the society. It shows how
conceptions of particular kinds of persons may stabilize the behavior they
describe, even when behaviors in question have no pre-social link to the kinds
of person in question. And this shows a quite general way in which even
relatively superficial effects of labeling practices may be stable, and thus
gives us good reason to doubt the Instability Thesis.
3.4. Other Sources of Stability
The two mechanisms
that I have examined are central to Hacking's own account of the looping
effect, but I have argued they may lead to stabilized rather than destabilized
social roles. Of course, the
discussion here is far from complete, and there are other sources of both
stability and instability for social role kinds. In particular, I limited my discussion of instability to
Hacking's looping thesis, and thereby concentrated on mechanisms of stability
and instability that are intentionally mediated. There is, however, another important model of social
constructionist thought that emphasizes social role explanations that are not
necessarily intentionally mediated.
I have in mind what Paul Griffiths (1997) has called a reinforcement
model of social
construction.
I
will briefly sketch this alternative model and consider a way in which it too
could give rise to a stable relationship between the conception of the social
role and the properties of the occupants of the role.[41]
The reinforcement
model of social construction is a diachronic, developmental account. It emphasizes nonintentionally mediated
changes in a person brought by social role occupation, explaining variant
behaviors as resulting from variations in patterns of developmental reinforcement. Griffiths suggests that the conceptions
of appropriate behavior in the social role establish norms that create
"something akin to a pattern of reinforcement which shapes people's
behavior so that it conforms to the norms" (143). The patterns of reinforcement in
question don't cause the agent to represent propositionally certain patterns of behavior as
desirable. Rather, once the person
is ascribed to a particular social role, the reinforcement shapes particular
cognitive mechanisms such that the individual components of the ascribed social
role become nearly reflexive, and together these components comprise the role
behaviors.[42]
Griffiths provides an
interesting example of this model that emerges from research on basic
affects. Over the last 30 years,
an impressive body of research has accumulated documenting the existence of
near universal correlations between certain facial expressions and
emotions. Paul Ekman, the leading
researcher in this field, summarized the state-of-the-art in 1992:
There
is consistent evidence, across investigators, of universal facial expressions
for at least five emotions [happiness, surprise/fear, sadness, anger, and
disgust/contempt]. More research
is needed to resolve questions about whether there are three or four more.
(550-1; cf. Griffiths, Chp. 3 for a review)
These
robust correlations lead Griffiths to remark that the emergence of these
expressions in infants "conforms to the classical biological determinist
model, in which almost any environment that supports survival to adulthood
supports development of the trait" (156). But drawing on work by Ekman (1972), Griffiths notes that
even these "biologically determined" motor responses can be shaped by
cultural reinforcement:
In
studies of facial expression in Japanese and American students, Ekman and his
collaborators found that the Japanese suppressed their facial expressions in
the presence of authority figures.
They superimposed voluntary muscle movements so as to produce a polite
smile. These voluntary movements
were initiated so quickly that the initial emotional expressions could be
detected only by using frame-by-frame analysis of videotapes. (156)
In
this case, a pattern of reinforcement in the social environment of the Japanese
students leads them to develop an alternative, reflexive expression in the
presence of authority figures. The
pattern of reinforcement was likely the result of the community's norms about
what counts as appropriate behavior towards others in the social hierarchy. And these norms are presumably part of
the community's conceptions of the social role of subordinate in a hierarchy.
How do such
effects created by reinforcement bear on the instability thesis? Once a regime of labeling actually
creates new properties of persons, those effects are really there in the
world. They are a stable part of
the world that we are epistemically pulled to describe in our theories of
it. In fact, such effects are more
stable than the strategic responses of role occupants discussed in the last
subsection, since reflexive dispositions will continue to be manifested – at
least for a time – even when the
regime of labeling is removed. Any
conception of hierarchical behavior among Japanese college students that failed
to mention differential facial expressions would be importantly
incomplete. But a complete
conception may itself contribute to the continued reinforcement of
hierarchically 'appropriate' behavior.
As with intentionally mediated responses, reinforced behaviors may come
to be in equilibrium with the conceptions that created and describe them.
Hacking’s general
statements about the looping of human kinds are too broad to be plausible. But the more limited Instability Thesis
is still a threat to viewing social roles as cluster kinds that underwrite our
knowledge of the social world.
What I’ve tried to show here is that the Instability Thesis is still
much too general, and that there’s every reason to expect the forces that
contribute to the ‘looping of human kinds’ to result in stability as well as
instability. If what I have said is correct, then we
also have an answer to social constructionist critics of stability. Such critics, as I noted in the
introduction, often emphasize the instability of social kinds because such
kinds are not rooted in biological facts.
Such categories have, in the words of the critical sociologist Stuart
Hall “no guarantees in
Nature" (1987 [1996], 166).
But having no guarantees in nature, is not the same as having no
guarantees at all. A particular
social setting, structured by the conception in question, may be guarantee
enough for a particular social role to create a stable, explanatorily relevant
property-cluster kind.
Conclusion
The preceding
discussion offers an interpretation of a broad class social constructionist claims
about human kinds. On this
interpretation, to say that x
is socially constructed (where x
is replaced by the term for a human kind) is to offer a substantive empirical
hypothesis that the differential properties of instances of x
are produced by the occupation of a social role. When a social role differentiates its
occupant to the extent that a cluster of important properties regularly
co-occur, I have suggested we ought to regard those occupants as instances of
social kinds. Moreover, I have suggested
that we do and should employ those kinds in reasoning about our social
world. Because the theory of
social roles set out makes only quite general assumptions about human nature
(e.g. that humans may act in ways that are approximately rational), it is
compatible with humanist approaches to the philosophy of social science. But it
differs sharply from hermeneutic, critical, and constructionist approaches to
the social sciences in emphasizing the stability of social kinds. It is no surprise that the social world
can be extremely volatile. But by
excessively emphasizing this volatility, theorists in philosophy and the social
sciences have threatened to make it a mystery how we manage to negotiate our
social world successfully every day, and it leaves us without a clear
understanding of the intransigence of social roles. Understanding mechanisms of stability, in contrast,
not only offers an explanation of our social knowledge, but also presumably
focuses our attention on those mechanisms that must be addressed in the attempt
to disrupt social roles and effect the transformation of social life.
Bibliography
Appiah, K. A. 1996. Race, Culture,
Identity: Misunderstood Connections. . In Color Conscious: The Political
Morality of Race, ed. K.
A. Appiah, A. Guttmann, pp. 192. Princeton, NJ: Princeton University Press.
Atran, S. (1998). “Folk Biology and the
Anthropology of Science: Cognitive Universals and Cultural Particulars.” Behavioral
and Brain Sciences 21:
547-609.
Averill, J.R. 1980a. "A
constructivist view of emotion." In Emotion: Theory, Research and
Experience., vol. 1: Theories
of Emotion. ed. R. Plutchik and H. Kellerman.
Random House. New York.
Averill, J.R. 1980b. "Emotion and Anxiety:
sociocultural, biological, and psychological determinants." In Explaining Emotions,
A. Rorty. (ed.) University
of California Press. Berkeley.
Boyd, Richard. 1999a.
“Homeostasis, Species, and Higher Taxa.” In Species: New Interdisciplinary Essays.
Robert A. Wilson, (ed.).
MIT Press. Cambridge. pp. 141-186.
Boyd, R. 1999b. “Kinds, Complexity and
Multiple Realization.” Philosophical Studies 95: 67-98.
Boyd, Richard. 1992.
Constructivism, Realism, and Philosophical Method. ed. Earman, John. Inference, Explanation, and Other Frustrations: Essays in
the Philosophy of Science. University of California Press:
Berkeley and Los Angeles. pp. 131-98.
Boyd, Richard. 1991. Realism,
Anti-Foundationalism and the Enthusiasm for Natural Kinds. Philosophical
Studies 61:127-148
Boyd, Richard. 1988. “How to
Be a Moral Realist.” In
Sayre-McCord, G. (Ed.) Essays
on Moral Realism. Ithaca, NY. Cornell University Press. pp. 181-228.
Burge, T. (1986). “Individualism and
Self-Knowledge.” Philosophical Review XCV(1):
3-45.
Darley, J. M. and C. D. Batson (1973).
“From Jerusalem to Jericho: A study of situational and dispositional variables
in helping behavior.” Journal of Personality and Social Psychology 27(100-119).
Ekman, 1992. Are There Basic Emotions? Psychological Review 99: 550-553.
Fay, B. (1983). General Laws and
Explaining Human Behavior. Changing Social Science. D. R. Sabia and J. Wallulis. Albany,
SUNY Press: 103-128.
Fiske, Alan Page. 1993. The Structures of Social Life: The Four Elementary Forms
of Human Relations. New York. Free Press.
Fodor, Jerry. 1981. “Special
Sciences.” Representations:
Philosophical Essays on the Foundations of Cognitive Science.
Cambridge, MA. MIT
Press. pp. 127-145.
Foucault, Michel. 1978. The History of Sexuality, Vol. I: An Introduction.
Robert Hurley, trans. Pantheon. New York.
Griffiths, Paul E. 1999. “Squaring the Circle: Natural Kinds with Historical
Essences.” In Species: New
Interdisciplinary Essays. Robert A. Wilson, (ed.). MIT Press. Cambridge.
pp. 209-228.
Griffiths, Paul E. 1997. What Emotions Really Are. University of Chicago Press. Chicago.
Hacking, Ian. 1999. The
Social Construction of What? Harvard University Press. Cambridge.
Hacking, Ian. 1998. Mad Travelers: Reflections on the Reality of Transient
Mental Illnesses. Charlottesville, VA. University Press of Virginia.
Hacking, Ian. 1995a. "The looping effects of human
kinds." in Causal
Cognition, ed. by Dan
Sperber, et al. Clarendon
Press. New York.
Hacking, Ian. 1995b. Rewriting the
Soul: Multiple Personality and the Sciences of Memory Princeton, NJ: Princeton University
Press
Hacking, I. 1986. Making Up People. . In Reconstructing
Individualism: Autonomy, Individuality, and the Self in Western Thought, ed. T. C. Heller, M. Sosna, D. E.
Wellbery, pp. 222-236. Stanford: Stanford University Press.
Hall, Stuart. 1987.
[1996]. New Ethnicities. In
Black British Cultural Studies. Ed. H.A. Baker, Jr., M. Diawara, and R.H.
Lindeborg. Chicago: University of Chicago
Press. Pp. 163-172.
Haslanger, Sally. 1993. On Being Objective and Being Objectified.in A Mind of
One's Own: Feminist Essays on Reason and Objectivity, ed., Louise Antony and Charlotte
Witt. Boulder, CO: Westview Press. pp. 85-125.
Hirschfeld, L. A. (1996). Race in the
Making: Cognition, Culture, and the Child's Construction of Human Kinds. Cambridge, MA, MIT Press.
Hirschfeld, L. A. (1994). Is the
acquisition of social categories based on domain-specific competence or on
knowledge transfer? Mapping the Mind: Domain Specificity in Cognition and Culture. L. A. Hirschfeld and S. A. Gelman. New
York, Cambridge University Press: 201-233.
Jackson, Frank and Pettit, Philip. 1988. “Functionalism and Broad Content.” Mind. 97: 381-400.
Kessler & McKenna 1977. Gender: An
Ethnomethodological Approach. Chicago: University of Chicago Press.
Kornblith, Hilary. 1993. Inductive
inference and its natural ground : an essay in naturalistic epistemology.
Cambridge, MA. MIT Press.
Kripke, S. 1972. Naming
and Necessity. Harvard University Press. Cambridge, MA.
Latanè, B. and J. M. Darley (1968).
“Group inhibition of bystander intervention in emergencies.” Journal of
Personality and Social Psychology
10: 215-221.
Lewis, David. 1969. Convention:
A Philosophical Study. New York: Blackwell.
MacKinnon, C. A. (1989). Toward a
Feminist Theory of the State.
Cambridge, MA, Harvard University Press.
Mallon, Ron, and Stich, Stephen P. 2000. The Odd Couple: The Compatibility
of Social Construction and Evolutionary Psychology. Philosophy
of Science 67: 133-154.
Mayr, Ernst. 1984. Species Concepts and Their Application. Conceptual Issues in Evolutionary
Biology: An Anthology. ed. Elliot
Sober. Cambridge, MA. MIT Press. pp. 531-40.
McIntosh, Mary. 1968. “The
Homosexual Role.” in Social
Problems
16:182-193. Reprinted in Forms
of Desire: Sexual Orientation and the Social Constructionist Controversy, ed. by Edward Stein. Routledge. New York. pp. 25-42.
McCulloch, Warren. 1965. What is a Number that a Man May Know It, and a Man, that He
May Know a Number? Embodiments
of Mind. Cambridge, MA: MIT Press.
Milgram, S. (1963). “Behavioral study of
obedience.” Journal of Abnormal and Social Psychology 67: 371-378.
Miller, R. (1978). “Methodological
Individualism and Social Explanation.” Philosophy of Science 45: 387-414.
Millikan, R. G. (1999). “Historical Kinds
and the "Special Sciences".” Philosophical Studies 95: 45-65.
Mills, C. 1998. Blackness Visible:
Essays on Philosophy and Race
Ithaca: Cornell University Press.
Omi, Michael and Winant, Howard. 1994. Racial
Formation in the United States: From the 1960s to the 1990s. New York: Routledge and Kegan Paul.
Omi, Michael and Winant, Howard. 1986. Racial
Formation in the United States: From the 1960s to the 1980s.
New York: Routledge and Kegan Paul.
Outlaw, L. 1996. On Race and
Philosophy New York:
Routledge.
Padgug, Robert. 1979. “Sexual Matters: On Conceptualizing Sexuality in
History.” Radical History
Review. 20:3-23. Reprinted in Forms of
Desire: Sexual Orientation and the Social Constructionist Controversy, ed. by Edward Stein. Routledge. New York. pp.43-68.
Putnam, Hilary. 1975. The meaning of ‘meaning’. Mind, Language and
Reality: Philosophical Papers,
Cambridge University Press. New
York. pp. 215-271.
Ross, L. and Nisbett, R.E. 1991 The Person and the Situation.
McGraw-Hill.
Scheff, Thomas. 1984. Being
Mentally Ill: A Sociological Theory.
Aldine Publishing Co. New York.
Showalter, Elaine. 1996. Hystories: Hysterical Epidemics and Modern Media.
Columbia University Press.
New York.
Sperber, D. (1996). Explaining Culture. Cambridge, MA, Blackwell Publishers.
Stalnaker, R. (1989). On What's in the
Head. Philosophical Perspectives, 3: Philosophy of Mind and Action: 287-316.
Sulloway, Frank. 1996. Born To Rebel. New York. Pantheon Books.
Taylor, C. (1971). “Interpretation and
the Sciences of Man.” Review of Metaphysics 25: 3-51.
Tooby, J., & Cosmides, L. 1992. The
psychological foundations of culture. The Adapted Mind In J. Barkow, L. Cosmides, & J.
Tooby (Eds.), Oxford University Press. New York. Pp. 19-136.
Walzer, M. (1983). Spheres of Justice:
A Defense of Pluralism and Equality.
New York, Basic Books.
Watkins, J. W. N. (1957). “Historical
Explanations in the Social Sciences.” British Journal for the Philosophy of
Science 8: 104-117.
Wilson, M. and M. Daly (1992). The Man
Who Mistook His Wife for a Chattel. The Adapted Mind: Evolutionary
Psychology and the Generation of Culture. J. H. Barkow, L. Cosmides and J. Tooby. New York, Oxford
University Press: 289-322.
Wilson, Robert A. 1999. “Realism, Essence, and Kind: Resuscitating Species
Essentialism?.” In Species: New
Interdisciplinary Essays. Robert A. Wilson, (ed.). MIT Press. Cambridge.
Wollstonecraft, M. (1995). A Vindication
of the Rights of Woman. A Vindication of the Rights of Man with A
Vindication of the Rights of Woman and Hints. S. Tomaselli. New York, Cambridge University Press: 65-294.