Action and Institution, Network and
Function: The Cybernetic Concept of Social
Structure
Thomas J. Fararo
University of Pittsburgh
John Skvoretz
University of South Carolina
A program of research on the formal representation and analysis
of institutional structures is taken a step further by integrating it
with recent developments in the formal representation of hierar-
chical levels of inclusion or part-whole relations. We begin by re-
viewing a cybernetic conception of action and show how this re-
lates to the construction of production system models of institutional
structures. Thereafter, we treat the inclusion hierarchy to show
how the production rule constitutes the conceptual unit integrat-
ing social knowledge and social action upon which are built two
hierarchies, involving institutional entities and social networks, re-
spectively. We indicate some of the detailed forms of control in-
volved in these hierarchies and then show how a form of func-
tional analysis can be undertaken on this basis. Finally, we provide
a lengthy discussion of the promise and problems of this mode of
structural analysis.
INTRODUCTION
Tenure, the employment contract, the Presidency of the United
States, the motel, and the teller-customer relation in a bank are all
examples of institutional structures of American society. Each case of
tenure, each employment contract, each Presidency, each motel, each
teller-customer interaction in each bank are "embodiments" of these
structures. Thus, each institution has the property of potential or actual
multiple embodiment, either in time only (the Presidency of the United
States) or in both space and time (the tenure institution), with respect
to the specified social system in which the institution is a structural
component. For some years now, we have been working toward the
development of a formal mode of representation of institutions based
on two key ideas: that an institution has multiple embodiments in time
© 1986 by The Eastern Sociological Society. All rights reserved.
0884-8971/86/0102-0219 $1.50
SF Volume 1 Number 2 219
Sociological Forum
or space and that an institution is an intrinsically processual structure.
In addition, we have attempted to represent these structures in such a
way that they are a type of interpenetration of cultural symbol systems
and concrete social interaction (Parsons, 1951); that they constitute or
define social reality via types of actors, acts, and objects (Berger and
Luckmann, 1966); and that they include "if situations" (Nadel, 1951) in
which action is contingent on the matching of a concrete occasion to
a type of situation.
In a series of papers (Axten and Fararo, 1977; Skvoretz, 1978;
Skvoretz, Fararo, and Axten, 1980; Axten and Skvoretz, 1980; Skvoretz
and Fararo, 1980; Fararo, 1981; Skvoretz, 1984; Fararo and Skvoretz, 1984),
we develop the idea that institutions can be represented formally as
control structures. Our starting point is a cybernetic conception of ac-
tion. We use a mode of formal representation found in cognitive science
and artificial intelligence, namely, "production systems," which are
structures of "if, then" action rules defined in relation to a space of
symbols designating action-relevant events, persons, information and/
or activities. From a methodological point of view, these production
systems function in the description of action processes as do differential
equation systems in the description of physical processes.
As adapted to the analysis of institutional action, elements of the
symbolic space of a production system model designate types of social,
physical and cultural objects, types of action and goals, and types of
actors. The production system itself consists of rules or "productions"
assigned to various types of actors, individuals and collectivities. Each
production, in turn, is of the form:
situation-type --* action-type
meaning that if the situation-type is satisfied--in a comparison with cur-
rent situational knowledge, made by the actor embodying the produc-
t ion -an action of the indicated type is executed. The symbolic space
defines important dimensions that distinguish different situation-types
and action-types.
The next section reviews the "core elements" (Wagner, 1984) of
this program of theoretical research. We show how the program, as de-
veloped, focuses on action systems and their analysis. Then we generate
a further development, based on hierarchical inclusion or part-whole
relations, to show how the approach relates to the analysis of concrete
social systems. In particular, we discuss their structural analysis in terms
of networks of induced relations among and between actors and their
functional analysis in terms of "function" as a relation on production
systems. Thereafter, in a long final section, we provide an extensive dis-
cussion of certain conceptual and formal problems and some tentative
220
Action and Institution
remarks on their solution. Relative to our previous publications, then,
our first section is a capsule review of the ideas, our second section
reports some new developments, and our final section discusses the re-
search program as it now stands.
ACT ION AND INST ITUT ION
In this section we briefly review how we describe control struc-
tures in social life. "Iaae core elements are: first, action is a cybernetic
process; second, institutions are stable designs for action representable
(because of the first point) by production systems; third, there are two
modes of institutionalized action analysis, dynamic and grammatical, made
possible by virtue of the second point; and, fourth, institutions have an
inherent constraint/freedom duality revealed by- the analysis of the third
point.
Act ion is a cybernet ic process . The reference here is to action
as composed of two steps, testing/comparing and acting/operation. The
TOTE model of Miller, Galanter and Pibram (1960) is one of the earliest
formulations of this idea. The work of Newell and Simon (1972), which
was developed even earlier and over a period of decades, is another
example. The elegant cybernetic hierarchy for behavioral analysis pro-
posed by Powers (1973) should also be mentioned. In these cybernetic
models, the action phase may be a behavioral output or action evoked
in response to a difference to be reduced. If there is no difference, then
nothing is done. This is negative feedback control. Alternatively, the
action phase may be that of an action being "called" when a certain
condition is satisfied. Here the test or comparison determines whether
a condition is satisfied, ff not, control shifts to another comparison,
"seeking" a match. This is instructional control.
In sociology, the Heise (1979) model of affect control is of the
negative feedback type. The subject matter is essentially symbolic in-
teraction. Its basic idea is that actors' actions, from moment to moment,
are efforts to minimize deflections from "fundamental sentiments," i.e.,
multi-dimensional semantic evaluations of social objects and acts. The
cybernetic control of action is internal to the actor-individual. Our work
exhibits the instructional type of cybernetic process. Applied to insti-
tutionalized social interaction, it also assumes that human interaction is
symbolic interaction. The basic idea is that institutions are composed
of systems of "production rules," each rule a "particle" of instructional
control. Control of action is conceived as internal to the actor, whether
individual or collectivity, and the production systems are regarded as
221
Sociological Forum
"embodied," operative social software comprising institutional roles. 1
Ins t i tu t ions are stable des igns for ac t ion that are repre -
sentab le by product ion systems. This principle, discussed at length
elsewhere (most recently, Fararo and Skvorctz, 1984), involves three
points: first, institutions are designs for social action; second, they are
representable by product ion systems; and, third, they are stable. To
elaborate the first point: as a design, an institution has the capability for
"multiple embodiment," that is, the institution can be instantiated in an
indefinite number of concrete collectivities through time and space even
if, as a matter of fact, it is present in only one. On this score, an insti-
tution is like a symphonic composi t ion- -p layable by multiple collectiv-
ities, even simultaneously. The capacity for multiple embodiment is in-
herent in the view of an institution's production system as social software,
an organization of symbolic declarations and instructions.
On the second point, we have in Table 1 an outl ine with illustra-
tions of the two main components of a product ion system model: first,
the social space of symbolic declarations which designate types of so-
cial, physical and cultural objects and types of actors, actions and goals;
and second, the instructional, social action system component. 2 Both
components have several corresponding levels. In the space component ,
the designations vary from general declarations of types of action, types
of social objects, etc., down to the much more specific declarations of
a particular institutional embod iment - - f rom action being "declared" as
either instrumental-type or expressive-type to a beverage being "de-
clared" as either coffee, tea or milk. At all levels, the social action com-
ponent consists of product ion rules, but the specificity of these varies.
General action rules apply or are available in all institutional settings;
in fact, the il lustration given in the table of a general action rule refers
to a common procedure by which deviation from a design may be
countered in any institutional setting. At the other extreme, the "per-
sonal" action systems hold only for the particular embodiment. They
1 As a practical matter, the execution of institutional roles no doubt involves instructional
and negative feedback controls. We do not follow up this point since it would lead us
to a behavioral analysis and away from the structural and functional analyses we seek to
describe. In addition, "reduce difference" operations can be called for under certain sit-
uational conditions represented within an instructional control system. Both the General
Problem Solver of Newell and Simon (1972:ch. 9) and the Deviation Counteraction Gen-
erator we have sketched (Fararo and Skvoretz, 1984:142) are examples.
2 We are adopting the notation of Newell and Simon (1972) to describe the components.
Our usage of it is described in detail in Axten and Fararo (1977) and in Fararo and Skvo-
retz (1984). Briefly, (x) is a kind of template or place-hoMer for a range of possible in-
stances. It can be thought of as a variable with the values it can assume listed after it,
separated by vertical bars. The symbol : : = separates the variable from the list of instances
and translates to "can be rewritten as."
222
Act ion and Ins t i tu t ion
TABLE 1. Two Components of Product ion System Models
Social Spaces Social Action Systems
Declarations of entities are rewrite rules
at several nested levels:
1. General action space
2. Institution space
3. Embodiment space
Examples
1. General action level (partial)
(situational object): : =(physical
object)l(social object)I(information}
(action): : =(instrumental
action)I(expressive action} I. . .
(instrumental action): : =OPERATE
[parameter](inputs)
(expressive action):: =EXPRESS
((attitude)l(sentiment))
2. Institution space (partial:restaurant)
(social object): : =(role): : =
(customer) l(waitress)[ , , ,
(physical object): : =(food)t . . .
(goal)::= get (food)] . . .
(information):: =(menu)] . . .
(instrumental action): : =SERVE[FIND-
PLACE]PAY]...
3. Embodiment space
(menu): : =(main course)(beverage)
(main course): : = fried chicken] . . .
(beverage): : = coffee I . . .
Production rules call for actions contingent
on situation knowledge satisfying situation-
type:
I. General social action systems
Deviation-counteraction generator
2. Institutional social action systems
Hierarchy of specification (subpro-
duction systems for action called)
3. Personal social action systems
a. Given for concrete situation
b. Emergent situationally
Examples
1. General action level
(action) by (role) due obligatory new
-->ESTIMATE-REASON [action] (role)
( = errorlignoranceldeliberate )
2. Institutional level
bill ((order)) unpaid ~ PAY (bill)
3. Personal
a. Given: PAY ( item) is accomplished
by the actor's own subproduction
system for deciding on and imple-
menting a mode of payment
b. Emergent: actor forgot wallet, re-
quires problem space and heuris-
tics
represent an actor ' s own s i tuat iona l ly g iven or s i tuat iona l ly emergent
way o f imp lement ing an ins t i tu t iona l i zed act ion or o f respond ing to nove l
events .
The midd le , ins t i tu t iona l leve l has been the focus o f our a t tent ion .
Because act ion is a cybernet ic p rocess at a / / leve ls , ins t i tu t ions are bu i l t
out o f p roduct ions , each product ion a "par t i c le" o f ins t ruct iona l cont ro l .
D i f fe rent combinat ions o f such par t i c les in to sys tems y ie ld the d i f fe rent
var ie t ies o f fo rm an ins t i tu t ion may exh ib i t . The hypothes is is that , fo r
example , all res taurants (o r schoo ls , t r ia l cour ts , banks , e tc . ) a re com-
posed o f the same set o f p roduct ions , w i th d i f fe rent types o f res taurants
( se l f - serv ice , fast - food, tab le -serv ice) p roduced by d i f fe rences in how
223
Sociological Forum
productions are aggregated into different production systems and as-
signed to roles in the overall action system. An institution type, there-
fore, takes the form of a specific collection of what we term rolegrams,
component programs of hierarchically ordered production rules that
control the activity of individuals assigned to the corresponding sta-
tuses. The rolegrams are coordinated definitionally--they are defined
with respect to the common institutional spacehand interact ivelyhthe
actions controlled by one rolegram can create situations in the domain
of the situation-types found in others, thereby indirectly evoking action
from these systems.
Production rules must be formulated in systems in order to prop-
erly represent the contextual aspects of social action. The writing down
of such a system is a highly empirical and difficult task of model-build-
ing. We have constructed and at least partially tested two such systems:
one for the system of customer-waitress interactions in restaurants and
one for aspects of the teacher-student-administrator system of interac-
tions in elementary schools (Axten and Skvoretz, 1980; Skvoretz, 1978,
respectively). For the purpose of this paper we do not try to illustrate
the construction of such models. Instead, the reader is asked to simply
think of "if-then" rules in complex combinations that contextualize each
rule. Here are a few sketches of possible rules that might appear in con-
trol structures, corresponding to types of actors. In each case, the pos-
sible rule is abstracted from its context of other such rules:
WAITRESS:
STUDENT:
MOTHER:
MOURNER:
(order)unserved --~ CHECK[kitchen]
registration-time ---> SELECT[courses] (course
schedule list)
(child)ego --~ CARE[child] (need-of-child)
(relative-of-deceased)present new
EXPRESS[condolence] (relative of deceased)
Finally, we turn to the third point of our general principle, that
institutions have a stability aspect. The concept of design stability is
meant in the usual dynamic systems sense, namely, that departures from
normal forms of action defined by the design tend to be counteracted.
Counteraction is only secondary backup and applies at the embodiment
level, i.e., in particular situations realizing the institution. The primary
reason for stability is what reality construction theorists, in particular
Meyer (1977), stress as stabilization by constitution of social reality. As
Berger and Luckmann (1966) emphasize in less formal terms, elements
of an institution's symbolic space provide a constitutive definition of
social reality by their designation/declaration of the types of social,
physical and cultural objects, types of action, goals and actors to be
224
Action and Institution
found in the setting. For example, one of the declarations of a "trial
court" institution may be of the form:
(trial-role): : -(judge)l(juror)l(defendant)l(defense
attorney) l(prosecutor)
meaning that only the five actor types on the right are legitimate forms
of a trial-role. If this declaration is part of an institutional space, then
there are, in a constitutive sense, no other types of actors involved in
a trial and an)' individual legitimately involved in a trial must be in-
volved in one of these five capacities. Equally important, from the point
of view of an actor trying to achieve some objective through the trial
court institution, action must be in one of the constitutive forms of
involvement if the objective is to be attained. Action not in one of these
forms is "abnormal," outside normal forms and is either counteracted
by those acting in appropriate capacities and so brought back in line or
it receives no response and is thus ineffective (Zelditch, 1984).
There are complications connected with this notion of stability
that relate to the idea of the system or network of embodiments of an
institution and to that system's institutional environment (Fararo, 1981).
There may be a struggle, essentially political, to modify or overhaul the
social space of action; there may be diffusion of social innovations that
generate incessant pressures in local embodiments to modify their forms
of action; there may be important interconnections among a set of in-
stitutions that give rise to ripple-effect changes. When a society is highly
conscious rather than "unconscious" of its institutions (in the sense of
Levi-Strauss, 1963), the structural stability is less secure. All these and
other complications do not negate the idea that an institution, although
intrinsically processual, is also intrinsically stable; however, they require
that the application of the idea be done in an analytical way, so that
empirically-observed change is mapped into a representation in which
a system's stable state is changing over time (in analogy to analysis of
behavior manifolds associated with differential equation systems, with a
focus on structural stability and catastrophes; see Fararo, 1978).
There are two modes o f inst i tut ional ized act ion analysis:
dynamic and grammat ica l . The dynamic analysis task is to show
how the rolegrams of an institution generate particular situationally con-
ditional normal forms of action and interaction. It is analogous to in-
tegrating a system of differential equations to arrive at the trajectory of
the property modeled. For institutions the production system model is
constructed and tested to see if it implies the normal forms commonly
observed and does not generate any institutional nonsense. We have
done this taking a customer-waitress action system as the institution.
Consider a single waitress, for instance, and a sequence of temporally
225
Sociological Forum
overlapping customers arriving and departing for service. The problem
is to model both the symbolic space of declarations and the production
systems of waitress and customer such that the observed types of nor-
mal interactions can be generat
本文档为【Action and institution, network and function】,请使用软件OFFICE或WPS软件打开。作品中的文字与图均可以修改和编辑,
图片更改请在作品中右键图片并更换,文字修改请直接点击文字进行修改,也可以新增和删除文档中的内容。
该文档来自用户分享,如有侵权行为请发邮件ishare@vip.sina.com联系网站客服,我们会及时删除。
[版权声明] 本站所有资料为用户分享产生,若发现您的权利被侵害,请联系客服邮件isharekefu@iask.cn,我们尽快处理。
本作品所展示的图片、画像、字体、音乐的版权可能需版权方额外授权,请谨慎使用。
网站提供的党政主题相关内容(国旗、国徽、党徽..)目的在于配合国家政策宣传,仅限个人学习分享使用,禁止用于任何广告和商用目的。