Ausgrenzung liegt im Wesen sozialer Strukturen

Das Phänomen der Marginalität und der Marginalisierung

Gunther Grewe

Games for Criminal Status

Justice as Order
Structured Social Inequality               

– edited 2022 –

Peter Lang


Frankfurt am Main • Bern • Las Vegas


(European University Papers: Series 2, Law; Bd. 210)
ISBN 3-8204-6480-8


The study proposes that we live in a world of social inequality where social status determines our behavior in the observer’s perception.

Depending on the relative social position of the actor and the observer, „saintly behavior“ of the high-status actor may be phenomenologically identical to „deviant behavior“ of the social miscreant but carries rewards for the „saint“ and sanctions for the „deviant.“

Such a view of social texture is sobering. Is Justice „Order through Structured Social Inequality“? The study develops a model of the processes leading to the differential distribution of immunity in society. These processes are a sequence of status degradation ceremonies. In a social psychological model based on game-theoretical conceptualizations, the status degradation ceremonies play a series of games where the grand prize is criminal status.

The study illustrates that the processes leading to criminal status parallel everyday life if we understand social life as a sequence of encounters, as games for social status. Given this understanding, the study of criminology attains a new meaning. It is no longer the study of some marginal, exotic, and esoteric group, be it criminals or criminologists, but as a part of social science, the study of social differentiation in general. Whatever we learn about the dynamics of obtaining criminal status clarifies the criminalization process and holds the properties for a novel understanding of the processes of reality construction in everyday life – of becoming prominent, an outsider, or simply of being a plain man.

For criminological research, the model of games for criminal status conceptualizes the labeling approach and the principle of marginality (i.e., the phenomenon of ubiquity, scarcity, and relativity of marginal positions in social groupings). Based on this model, we can reach a new understanding of justice and, especially, of criminal justice, which would allow us to develop the labeling approach into a theory from which we derive hypotheses whose validity is open to empirical investigation and validation.

The author, a sociologist, uses symbolic interactionist modeling and simulations from game theory. His publications deal with the interdependence of behavior and status. In essence, he observes that marginal positions are ubiquitous, rare, and relative in any social structure. A chilling aspect of the model is that removing someone from a marginal position pushes someone else into the margins of society. Even more chilling is that these processes could be socially engineered and misused.

He previously published „Straßenverkehrsdelinquenz und Marginalität“ (Lang: 1978), a study on the possibilities and limitations of regulating social behavior through law and law enforcement


.. 6

1.1. Perspective and
Assumptions. 6

1.2. The Concept of
Marginality. 7

1.3. The Model of “Games
for Status”. 9


2.1. Criminology as
Caretaker of Rationalizations. 14

The Controversy on Definitions of „Crime“ and „Criminal“

2.2. Value Neutrality in
Social Science. 17

Gouldner’s Complaint and the „1′ Art pour 1′ Art“ Approach to Social
. 18

The Problem of Common Sense: Social Science as a Social Activity
. 20 The Gross Cycle of Social Science. 20 The Petite Cycles of Social Science. 22
Value Neutrality as an Attempt to Avoid Ought-Worlds

2. 3. Towards a
Scientific Activity of Criminology. 37

.. 39

3. 1. Decision Theory as
Complementing Behavioral Theory. 39

The Social Exchange Concept

The Game-Theoretical Bargaining Model for Social Status

3.2. Games for Social
Status. 59

Bargaining for Esteem
.. 61

Selecting Partners

Status Maintenance and Status Change
. 76

3.3. Behavior as
Negotiated Reality. 77

.. 80

4.1. The Attrition of
Justice. 82

4.2. Games for Criminal
Status. 85

Distributing Negative Esteem
.. 85

Choosing Conflict Points

4. 3. Criminal Status
and Criminal Behavior. 94

4.4. Propositions on the
Construction of the Social Reality of Criminal Behavior and Criminal Status. 95




for Criminal Status:


1.1. Perspective and

Criminology has been a layman’s science,
traditionally incapable of transcending insights beyond common sense and
traditionally occupied with rationalizing societal irrationalities. It is the
caretaker of the common-sense notion that life chances depend on social
conduct, i. e., the notion that the good things in life are for the ones who
are well behaved and that the bad things in life are for the ones who violate
social conduct norms. Future historians and sociologists of knowledge might see
traditional criminological theory in the same way as the functional theory of stratification
is presently seen, namely as an attempt to justify and maintain structured
social inequality existing in advanced industrial society.

Common to many criminologists, regardless of whether
they aim at adjusting „the criminal“ to the prevailing social system,
or whether they strive in addition to adjusting the social system to their
ideas of appropriateness, seems to be the belief in the American Dream which
asserts that „all men are born free and equal“ and that
„everyone . . . has the right and often the duty, to try to succeed and to
do his best to reach the top“ (Warner et al., 1949: 3). Common to these
criminologists is a vague idea of justice according to which wrongdoers are or
should be punished only based on their conduct and are or should be assigned
criminal status regardless of their previous social status. Common to these
criminologists, therefore, is their refusal to see the contradictory character
of the American Dream, which disturbs its pastoral idyll of social equality.
They refuse to see that the two fundamental themes of the American Dream,

that all of us are equal and that each one of us
has the right to the chance of reaching the top, are mutually contradictory,
for if all men are equal, there can be no top-level to aim for, no bottom one
to get away from; there can be no superior or inferior positions, but only one
common level into which all Americans are born and in which all of them will
spend their lives (Warner et al., 1949: 3).

They refuse to recognize the possibility that a
common sense ought-world might exist, structuring the is-world and making it
inhabitable, which, for social scientists, is part of the is-world to be
investigated. They refuse to admit the reality-descriptive quality of the well-known
Latin proverb, „Quod decet Jovi, non decet bovi!”, which corresponds
closely to the phrase, „All of us are equal, but some more than
others“. They refuse to recognize that phenomenally similar behavior is
not necessarily cognitively similar.

The problem of these criminologists, and perhaps of
the entire activity of criminology, as we shall show in the course of our
study, lies in their attempt to solve the ancient discord between
„is“ and „ought“ by postulating the transformation of the
is-reality into an ought-reality (Gouldner, 1970: 489) as the primary goal of
criminology, without much concern for and perhaps even fear of the presently
existing is-reality.

Though this approach is understandable, since “it is
as extremely painful and threatening for a man to believe that what is powerful
in society is not good, as it would be for the religious believer to feel his
God was evil“ (Gouldner, 1970: 486), we do not consider it to be a viable
approach for engaging in criminology as a science.

We understand science as comprehending and reducing
complexity (Luhmann, 1967: 107), as mapped reality. We regard it as necessary
to distinguish between a pure and applied science in any science and therefore
also in criminology. By criminology as pure science, we mean an activity, which
is restricted to analysis, far away, disconnected from and unrestricted by any
ought-world of change proposals. Based on this perceived is-world encompassing
social ought-worlds, we believe an applied science of criminology is possible.

With this perspective and these assumptions in mind,
we shall now present a central concept of our study, the concept of

1.2. The Concept of

We propose to understand the is-world of criminal
justice as being governed by what we shall call marginality. By marginality, we
mean (1) the phenomenon that marginal positions exist in all social groupings, (2)
the phenomenon that the marginal positions in all social groupings are limited
in number, and finally, (3) the phenomenon that marginal positions are marginal
relative to the ranking within the particular social grouping. Marginality,
therefore, is characterized by (1) ubiquity, (2) scarcity, and (3) relativity
of marginal positions in social groupings.

Indirectly, we encountered marginality when we
discussed the internal inconsistency of the American Dream. There we noted that
the dream of equality conflicts with the reality of social ranking. The notion
of marginality is nothing but a consequence of structured social inequality in social

By introducing our concept of marginality, we
propose to effect „a shift in the problems noticed and investigated and a
change of rules of scientific practice, comparable to the switch in perceptual
gestalten in psychological experiments, when, e. g., the same figure may be
seen as two faces vs. cup, or as duck vs. rabbit“ (von Bertalanffy, 1968:
18). Instead of directing our attention to the inequalities in the social
system, we shall redirect it to its margins. The existence of these marginal positions
is implied by the very fact that we can assume the universality of rank
differentiation in social groupings. The redirection of our attention from the
social ranking to the fact that ranking implies the existence of marginal
positions is especially appropriate if we perceive the Criminal Justice System
as a sub-system of the general social system of structured inequality, which
has the task of deciding whether or not to assign somebody to the social status
of „being a criminal“, i. e., to a marginal status.

By understanding criminal status as a particular
case of marginal status, we take two steps, which have often been postulated.
Firstly, we implicitly recognize that criminal status is a „normal“
social status. Secondly, we free criminology from the „shackles“ of
criminal law definitions (cf., Sellin, 1938: 41).

The normality of criminal status follows from the
assumption of the ubiquity of social differentiation. In contrast, the
assumption of relativity and scarcity of marginal status accounts for the fact
that, for an observer whose standpoint is within a social system, i. e., from
the perspective of his socially constructed ought-world, the marginal status of
being defined criminal appears to be a non-normal social status.

However, our concept of marginality also has the
property of freeing a science of criminology from the shackles of criminal law
definitions. This follows from the assumption that criminal status is only one
type of marginal status and that its allocation procedure can be assumed as
being non-distinct from the allocation of other types of social statuses,
whether they be marginal or not. In our concept, the „criminal“, the
„mentally ill“ individual, the „weird“ person, the
„prominent“ man, Becker’s (1963) „outsider“, as well as the
„leader“, are all defined as marginal persons, but without the
„reality“-restrictive stigma of being good or being evil. Their
position is perceived as the result of a social evaluation process, or more
precisely, of a sequence of social evaluation processes, each not necessarily
consisting of the same evaluators as the previous one, except for the
(changing) self of the evaluee.

Therefore, the proposed concept of marginality is a
way to synthesize several perspectives of the „observed“ socially
constructed realities into a more complexly constructed reality. It takes
Durkheim’s hypothetical „society of saints“ (Durkheim, 1885: 12)
seriously, but goes beyond it; for it claims that corresponding to the
negatively evaluated „deviant“ saints, this society will also have
its “weird”, its “mentally ill”, and its “saintly” saints. It extends Quinney’s
(1970:23) elaboration of Durkheim’s (1885:13) theory of the social reality of
crime to the acquisition of social status, providing us with possible insights
into the process of assignment to a marginal position in a social system, as
well as insights into the process by which the status assignment is confirmed
or altered.

Certainly, however, the mere substitution of the
term “criminal” by the term “marginal” will have the effect of de-metaphysicizing
the study of “crime”.

The theory of the social reality of crime is where
we are today. It is a significant advancement over where we were before, as
this theory can integrate much of the factual information gathered by
generations of criminologists. By having introduced the concept of marginality,
we intend to indicate where we might go from there because the stubborn search
for a science of criminology is not yet over.

1.3. The Model
of “Games for Status”

Corresponding to the nature of concepts – – they are
neither true nor false, but tools designed to capture relevant aspects of
reality and thus “constitute the definitions (or prescriptions) of what is to
be observed” (Merton, 1968: 143) – – our task will be to construct within the
novel reality created by the concept of marginality a theory explaining the
mechanism of status allocation within the multidimensional social evaluation

We shall build two models of status allocation. The
first one is a general model of status change, where the essential factors and
the dynamics of status change are unfurled. The second model, describing the
dynamics of attaining criminal status, is a special case of the general mode.
It is characterized by the fact that criminal status is one type of marginal
status and that its assignment occurs only after an institutionalized ritual,
which is performed by the candidate for criminal status and by various agents
of the Criminal Justice System.

In constructing our models, we shall employ a technique
we often observe in economic literature. There we are used to watching the
growth of a tender theory-seedling through a succession of imaginary worlds.
These are the imaginary worlds of perfect markets, rational behavior, perfect
certainty, two-period worlds, etc. In a sequence of hypothetical tests, where
each one takes place in a world harsher (more “realistic”) than the previous
one, the theory-seedling weathers to a mature theory that can exist and survive
in the complex “real” world. Adopting this technique seems to avoid a significant
problem often confronting behavioral sciences: The Sacrifice of the “Gestalt”,
the complexity, on the altar of reductionism for the sake of exposition.

After the final construct is non-contradictory to
empirical findings and has also been shown to order these findings into a
general system, we shall consider this construct a mapping of reality, reduced
in complexity. On this map, we shall be able to recognize previously
disregarded or unknown interrelations within social systems, and we shall be
able to clarify the relationship between the is-world and the ought-world of
criminal justice.

Our first model deals with status change and status
maintenance in social groupings. It attempts to penetrate the dynamics
effecting a change of the social status X at time t0 to the new
social status Y at time t1 where X and Y can be either the same
status position, or where X ranks higher or lower than Y, and where X and Y are
status positions within the same social grouping. By grouping, we mean any
social entity consisting of at least two persons. Therefore, a grouping could
be a dyad as well as a social system with the high complexity of the
interrelation between its members.

Defining social status as perceived esteem (cf.,
Homans, 1961: 149), we understand the members of social groupings as being
constantly engaged in social exchange activities. Their exchange object is
esteem in various „esteem-currencies“. Exchange theory, as it was
developed for the social sciences by Thibaut and Kelley (1959), Homans (1961),
and Blau (1964), describes and explains this process. It seems, however, that
we can further clarify the notion of social exchange if we take a look over the
fence and observe how another area of science, namely economics, handles the
problem of exchange.

Economists have tried to deal with the problem of
exchange in various ways, one of them being decision and game theory. For our status
change and status maintenance model, we intend to borrow from economics the
descriptive elements in the theory of bargaining games and a modern version of
the rationality postulate. Correspondingly, we choose as the label for our
basic model: „Games for Social Status“.

As a unit of social organization to be analyzed, we
select the encounter, which we define as a sequence of interactions consisting
of at least one action and one response. Unfolding the status change and status
maintenance model, our attention is directed towards the factors that influence
the allotment of esteem by and among the parties of an encounter. We shall
develop the idea that life chances do not depend primarily on social conduct;
instead, as we shall show, the dominant factor for the allotment of esteem in
an encounter is the status quo ante of the participating „players of the
status game“. We shall also show that phenomenally similar behavior is not
necessarily conceptually similar; instead, it is the social status, i. e., the
perceived esteem, which confers the conceptual quality to a behavior. Finally,
our model will confirm the factual information about status movement, namely
that any social status is typically a viscous, „sticky“ state, which
is more likely to be enhanced rather than changed by an encounter.

Our second model is an extension of the basic model
to the process of criminalization. We shall call it „Games for Criminal
Status“. Through this model, we shall solve the contradiction, which seems
to exist between the statements that social status usually is a
„sticky“ state and our assumption that the criminal status is a
marginal status, which typically does not correspond to the status quo ante of
the first encounter with an agent of the Criminal Justice System.

We shall solve the seeming contradiction by
understanding the criminalization process as a sequence of encounters conceived
as bargaining games and by distinguishing four steps in the process of the
allotment of criminal status, i. e., in the process of becoming a person who
has been adjudicated as having committed a „crime“.

The „offender“ is processed, as a rule,
sequentially in four encounters with an agent of the Criminal Justice System.
These encounters, ending with the assignment of criminal status, occur between
the actor and the general public and the „victim“, the police, the
prosecutor, and the court. Each encounter, compared with everyday encounters,
is highly ritualized. Only one party in the encounter, the agent of the
Criminal Justice System, has both the freedom and, sometimes, the duty to
choose partners. This is because he has, on the one hand, more partners and
games available than his time would permit, and on the other hand, he has to
choose at least a few partners and play a few games if he is an agent of type
2, 3 or 4.

We propose, therefore, to understand the four steps
towards criminalization as a recruiting process by which it is sequentially
determined who is eligible for criminal status and who is not. Hence, the
criminalization process can be analyzed as a sequence of encounters, of
bargaining games for criminal status, systematically connected as status
degradation ceremonies. The label „criminal“ is awarded after a
succession of encounters, starting with encountering the victim and ending with
encountering the court. At any given time, the subsequent encounter takes
place, if and only if the previous encounter permits the entrance into the subsequent
encounter, that is if in the previous encounter a behavior has been defined as
criminal behavior and the status of the actor has been downgraded, e. g., to
the new status of „being a suspect“, „being an accused“, or
„being a defendant“.

The details and the results of the model agree with
the factual information about the criminalization process because the model
predicts an inverse relationship between social status quo ante and the chance
to attain criminal status. As our model is non-contradictory to the empirical
findings, we can accept it as an adequate mapping of the process by which
criminal justice is distributed. According to the model, the social reality of
criminal justice is part of a general structured social inequality. Contrary to
the ought-world of criminal justice, it perceives the chance to attain criminal
status as differentially distributed in social groupings.

Our model presents us with an uncomfortable reality
because this reality conflicts with the ought-world of criminal justice. Within
our perceived is-world, conduct, which can be defined as a crime, according to
the relevant (criminal) code, is only a necessary but not a sufficient
condition for attaining criminal status. Just as in our analysis of everyday
status change and status maintenance, the status, which exists before the first
encounter, is the dominant factor for any new status. The role of the
phenomenal conduct is secondary, as its conception depends on the perceived
status quo ante.

Through our model, we try to dissociate ourselves
from the traditional approach to criminology. In contrast to this approach, we
regard criminal status as one type of marginal status, thereby implying that
criminal status positions are ubiquitous, scarce, and relative in any social
grouping. We also imply that criminal status is attained in processes similar
to the ones by which social status in general is acquired. We understand the
model as a concretization of the labeling approach, which focuses on the
offender dealing with the Criminal Justice System.

However, with our model, we also try to present a
„radical“ alternative to the „radical“ neo-romanticism in
American social science as represented, for example, by Alvin W: Gouldner
(1970). We see Gouldner’s approach as reactive and inconsistent with its

We have in common with this approach the discomfort
with a system of structured social inequality, administered and maintained in
part by the Criminal Justice System. We share the attitude that the idea of
„justice“, reified in the Criminal Justice System, is a functional
device for maintaining structured social inequality, a device that obscures the
selection of marginal persons from relatively powerless status groups.

However, the radical-reactive approach expresses its
discomfort by constructing an ought world. Its favorite subject is the
political and economic suppression of the weak by the powerful. Structured
social in-equality is seen as a characteristic of particular social systems,
not of social systems in general. The reactive component of the approach
invalidates the radical one because the rules and the setting of structured
social inequality are implicitly acknowledged. Contrary to its own explicit
intentions, the radical-reactive approach is implicitly as system-conservative
as the traditional approach to criminology.

In contrast to the radical-reactive approach, we
attempt to communicate our discomfort through an analysis. Our model of the
criminalization process shows that there is an is-world of criminal justice
where justice is bargained, protected by an ought-world, where justice is
equally distributed. Through our analysis, we attempt to penetrate the
protective screen of the ought-world, which, if successful, has repercussions
on the is-world. We try to conceptualize social differentiation and, therefore,
social marginality in general, whether in the social system of American society
today, in a social system of a tribe of headhunters, in a small group, or a
society of saints.

The proposed analysis of the Criminal Justice System
as „Games for Criminal Status“, then, could be relevant through its
implications for our perception and administration of „justice“.
These very implications bear the possible necessity of a reorientation of
criminology from a caretaker of rationalizations toward a critical science.

In the past, when presenting the ideas of the
concept of marginality and games for criminal status as a possible approach to
map the reality of the „familiar chaos: our ordinary everyday social
behavior“ (Homans, 1961: 1), traditional criminologists have misunderstood
the proposed approach as „value-laden“. From the radical-reactive
side of the fence, the response is even more perplexing at first: What is your
idea good for? What are the implications for social policy? The criticism was
stunning until realizing that the disagreement went deeper than content and
form. It concerns a disagreement in what social science is about.

Our first task, and our next chapter, will therefore
clarify the position from which the basic concept and the models giving it
substance were developed. This will be necessary to delimit the position taken
from the traditional and the now fashionable radical-reactive approach.


The „Concept of Marginality“ and the model
of „Games for Criminal Status“, as we shall develop them in the
course of this study, are an adequately reduced reality, if understood from a
position of value-neutrality in social science. By value-neutrality in social
science, we mean an approach which neither proclaims to be free of any values,
as traditional criminology is an accused of doing, nor consciously infuses
values into social science research, as the radical-reactive approach professes
to do, but instead tries to avoid values.

In this chapter, we shall show that the endeavor to
do criminology as a science necessitates stepping beyond common sense. This, in
turn, implies value-neutrality, a pure science approach, and a redefinition of
criminology as a science of social differentiation. By stepping beyond common sense,
we mean stepping beyond „What Anyone Like Us Necessarily Knows“
(Garfinkel, 1967: 54), or in other words, perceiving as social
reality-to-be-analyzed the is-world (phenomenal world) and the everyday
ought-world (conceptual world). Criminology understood as a science, i. e., as
an activity trying to comprehend and reduce complexity, has to take this step
because „constructs of the social sciences are, so to speak, constructs of
the second degree, that is, constructs of constructs made by the actor on the
social scene, whose behavior the social scientist has to observe and to explain
in accordance with the procedural rule of his science“ (Schutz, 1953: 59).

We shall begin our discussion by demonstrating that
the controversy over the „warmed-over stock issues: the difficulty of
defining what is crime and who is the criminal“ (Blumberg, 1967: ix)
characterizes traditional criminology as a caretaker of rationalizations. We
shall argue that „criminology“ is a misnomer, or at least,
inadequately and misleadingly describes the activity of a social science
dealing with crime, the criminal and criminal justice in society. Then, to
delimit our approach from the currently fashionable radical-reactive approach,
we shall discuss the necessity of value-neutrality and a 1’art pour 1’art
attitude in social science. There we shall explain our notion that social
science is a social activity, i. e., that a metascience of social science again
is social science, and we shall demonstrate value-neutrality as an attempt to
avoid everyday ought-worlds. Furthermore, in the third section of this chapter,
we shall revisit our concept of marginality. We shall point out the normality
of „crime“ and „criminals“, and we shall argue based on the
earlier discussion in this chapter that social science „criminology“
has to be redefined as a science of social differentiation. Only after the
problem of common sense in social science and specifically in criminology is
extensively presented do we believe our model of „Games for Criminal
Status“ as a mapping of criminal justice can be made intercommunicable.

2.1. Criminology as
Caretaker of Rationalizations

Turning our attention to the traditional approach in
criminology, we perceive its present state still adequately described by
Sutherland (1947):

Much factual information regarding crime has been
accumulated over several generations. In spite of this, criminology lacks full
scientific standing. The defects of criminology consist principally of the
failure to integrate this factual information into consistent and valid general

Sutherland made this statement in the preface to the
fourth (1947) edition of Criminology.
One generation or 23 years later, in the eighth edition of the same work by
Cressey, we read: „Criminology is not a science, but criminologists hope
it will become a science“ (Sutherland and Cressey, 1970: 20).

These are sardonic, however, acceptable summaries of
the work of the elders in criminology made by two of its most highly respected
members. We contend that the principal reason for the failure of criminology as
a science is the commonly shared research bias, which prescribes the delimitation
of criminology by examining „crime“ and „criminal“, however
widely defined. We shall demonstrate that the battle over definitions is a
battle of irrelevancies, fought on the battleground of first degree constructs,
i. e., „constructs made by the actors on the social scene“ (Schutz,
1953: 59). The controversy over definitions shows that traditional criminology
parallels the everyday conceptual reality rather than supersedes. This
controversy characterizes traditional criminology as a caretaker of common-sense

2.1.1. The Controversy on
Definitions of „Crime“ and „Criminal“

There is not much agreement among the contestants in
the controversy except on two points:

a)     A delimitation of the field of criminology is
necessary and feasible. What has to be „found“ is a definition of
„crime“ and „criminal“.

b)     Given a definition of „crime“, this
automatically defines as „criminal“ individuals who
„commit“ these „crimes“; and, given a definition of
„criminal“, this automatically defines as „crimes“ the acts
committed, if they are connected with the evaluation of an individual as

Before showing why we disagree even on these points
with the traditional approach to criminology, we shall briefly consider the
aspects of this controversy as they are presented by four of the chief
participants: Thorsten Sellin (1938), Edwin Sutherland (1945), Paul Tappan
(1947), and Richard Quinney (1970).

Like many searchers in criminology, Thorsten Sellin
(1938) expected to arrive at generalizations, which state that if a person of
type A is placed in a life situation of type B, he will violate the norm
governing that life situation. Postulating that general propositions of
universal validity are the essence of science, Sellin asserted that the
variability in the legal definitions does not permit the formulation of the
universal categories required in all scientific research. He, therefore,
proposed the study of conduct norms in general by isolation and classification
of norms into universal categories rather than only the study of crimes as
defined by the criminal law. For social science criminology, then,
„crime“ is the violation of conduct norms.

Like Sellin, Sutherland (1945) was dissatisfied with
the use of the legal definition of „crime“ in social science research
because it arbitrarily excludes behavior, which is similar to behavior
classified as „crime“ according to the legal definition. Sutherland’s
approach to the problem of the definition of „crime“ and
„criminal“ was to include within the scope of theories of criminal
behavior all behavior that is legally sanctioned and socially injurious. The
legal sanctions he had in mind were not restricted to criminal law but included
those in civil law.

Paul Tappan (1947) found the proposed sociological
definitions of „crime“ and „criminal“ perturbing. He
interpreted them as arising from the „desire to discover and study wrongs
which are absolute and eternal rather than mere violations of a statutory and
case law system which vary in time and place; this is essentially the old
metaphysical search for the law of nature“ (Tappan, 1947: 41). He
criticized Sellin‘ s and Sutherland’s universal concepts for the absence of
criteria defining such terms as „injurious“. Tappan contended that
crime, as legally defined, is a sociologically significant province of study,
whereas the „white collar criminal“, the violator of conduct norms,
and the antisocial personality are not criminal in any sense meaningful to the
social scientist unless they have violated a criminal statute.

Richard Quinney (1970) has provided us with one of
the most recent definitions of „crime“. He defines „crime“
as „a definition of human conduct that is created by authorized agents in
a politically organized society“ (Quinney, 1970: 15). „The social
reality of crime is constructed by the formulation and application of criminal
definition, the development of behavior patterns related to criminal
definitions, and the construction of criminal conceptions“ (Quinney, 1970:
23). Quinney uses the term „crime“ in a somewhat unusual sense.
Sellin, Sutherland, and Tappan employ the term abstractly when they discuss the
rules according to which the social reality is constructed. Quinney, however,
talks about the process concretely when he implies that the rules about how
social reality should be constructed might not necessarily correspond to how it
is actually constructed. Equivalent to Quinney’s definition of „crime” is
that crime results from an evaluation of human conduct. Instead of attributing
the result of concrete evaluation processes to the person adjudicated, the
„criminal“, Quinney attributes it to a conduct, the

2.1.2. Criminology
— A Misnomer?

Common to all four participants in the controversy
is a basic assumption of traditional criminology. This assumption presupposes a
fundamental difference between „crime“ and „criminal“ on
the one hand and behavior and status not evaluated as „crime“ and
„criminal“ on the other hand. The difference from everyday life is
perceived to be so drastic that a separate social science is justified. A
powerful ally for the assumption seems to be the „common sense“, the
„What Anyone Like Us Necessarily Knows“ – grasp (Garfinkel, 1967: 54)
of everyday life.

However, this powerful ally does not appear to be
mighty at all from a phenomenological perspective. On the contrary, from this
perspective, it seems that traditional criminology, because of this alliance,
falls victim to one of the cruelest laws of science, which says that the scope
of the research question determines the range of the possible research results.
The history of sciences is full of examples where paradigms proved to be
inadequately reality-descriptive because of their reality-restrictive a-priori
assumptions (see, e.g., Conant, 1947; Kuhn, 1962).

Today’s most avid victims appear to be the social
scientists, especially the religious believers in empiricism, the centipede leg
counters, as their critics sometimes call them. They and the other social
scientists must face the problem that, while they spend their lives within the
conceptual everyday reality, i. e., within constructions of the reality of the
first degree, in their work, they must build second-degree constructs of the
social reality. Traditional criminology, it seems, has run aground because of
this problem.

The controversy on definitions shows that the
traditional approach to criminology takes the etymology of the term
„criminology“ rather literally. Etymologically,
„criminology“ means the scientific study of „crime“ and
„criminals“. The term was introduced one century ago by Rafaele
Garofalo (1885), who used it as the title of his book. It was at that time an
appropriate description of the activity of criminologists.

Today, six academic generations later, we have to
ask ourselves whether the term „criminology“, which might have been
appropriate at one time, is a misnomer and a mistake because of its implicit
a-priori assumptions.

It is no longer self-evident that there can be a
separate area of science, restricting itself to a behavior and its evaluation,
whether it is called „crime“ and „criminal“ or otherwise.
On the contrary, it seems a more reasonable approach to ask a preliminary
question. Before discussing which definition of „crime“ and
„criminal“ provides an adequate delimitation of criminology, we must
ask whether a social science of a behavior is possible at all.

Quinney’s work, and before him the labeling
approach, did cast the first shadow of doubt on the appropriateness of the name
of the science. Though Quinney does not seem to relinquish the tacit assumption
of traditional criminology, he takes the first step towards it by postulating
that what used to be perceived as a phenomenal reality, namely
„crime“ and „criminal“, are both a conceptual and a phenomenal

The consequence of this view is that to understand
the person and the act, we have to trace the process which led to the
evaluation. Our attention has to be redirected from the result of the
evaluation process to the evaluation itself. However, suppose we are willing to
take this step. In that case, there is no reason to accept, as Quinney still
does, an a-priori assumption of a fundamental difference between
„criminal“ and everyday social evaluation processes, such as to
justify the maintenance of a separate science dealing exclusively with
„crime“ and „criminal“. Instead, we must either relinquish
criminology as it is presently understood to another social science, or we must
imbue it with new meaning.

Therefore, Quinney, one of the youngest elders in
criminology, legitimates with his compilatory theory of social reality the
revolt against the oldest and most tenaciously held dogma of criminology.

He himself seems to be only dimly aware that his
work facilitates the death blow to the traditional approach to criminology. In
his book The Problem of Crime, he
states somewhat astounded, somewhat proud: „I have found, to my own
satisfaction at least, that an attention to crime can illuminate many of our
problems. Crime has something to do with the most profound of all sociological
problems – the relationship of the individual to his society“ (Quinney,
1970a: vi).

Given our perspective, however, we can no longer say,
„crime has something to do with
the most profound of all sociological problems“. Instead, we have to say
that „crime“ i s one of the possible relationships of an individual
to his society. The controversy on the definitions of „crime“ and
„criminal“ to delimit the subject matter of criminological
investigation, then, is not only controversy on irrelevancies, but also a controversy,
which is misleading by its problem formulation. Like the medieval scientific
controversy within alchemy, it has and will continue to make us lose sight of
some basic assumptions on which the arguments are founded.

2.2. Value Neutrality in
Social Science

In the discussion up to now, we have attempted to
point out how questionable it is to perceive „criminology“ as the
scientific study of „crime“ and „criminals“. We used some
recognized elders, Quinney, Sutherland, and Cressey, as our star witnesses in
our case against the defendant „traditional criminology“.
Nevertheless, we assumed the existence of a scientific activity
„criminology,“ which we, in contrast to the traditional approach, no
longer perceive as a separate social science studying „crime“ and
„criminal“, but understand as a part
of social science. In this new concept, „crime“ is seen as a socially
constructed reality, which can be studied, only as one of the possible
relationships of an individual to his society.

Given this conception of criminology, we have to
take issue with the currently fashionable radical-reactive approach in social
science. We shall contrast our position, which can be described as a position
of value neutrality, from the position of the radical-reactive approach.

We shall accuse the radical-reactive approach of
confusing social science problems with social problems, thereby confusing
constructions of reality of the first and the second degree. Furthermore, we
shall try to show that the social policy implications of a particular research
project are inadmissible in social science. Arguing that an „applied“
science necessitates a „pure“ science, we shall demonstrate that
value-neutrality and a l’art pour l’art approach are indispensable in social
science and, therefore, in criminology.

We shall begin our discussion by illustrating the problems
which an „l‘ art pour l‘ art“ approach to social science presently is
encountering. This will be done by choosing Gouldner (1970) as representative
for a value-oriented social science, which deems itself as radical but is
understood by us as radical-reactive. We shall discuss Gouldner’s objections to
a value-neutral position and point out that to remove the inconsistencies in
his work, even his so-called radical stance would entail an initially value-neutral
approach to social science.

Expanding on this idea, we shall try to clarify in a
second subsection what we see as the problem of common sense in social science,
i. e., the problem that social science itself is a social activity. In the
third subsection, we shall postulate that a value-neutral approach to social
science is a way to avoid many of the ought-worlds, which presently entangle
social science with ought-worlds of the first degree, i. e., with common-sense
notions about the phenomenal world.

Only after this extended discussion of values in
social science can we go about filling the empty hull of the term
„criminology“ with meaning.

2.2.1. Gouldner’s Complaint
and the „l‘ Art pour l‘ Art“ Approach to Social Science

Alvin W. Gouldner has a problem: „It is as
extremely painful and threatening for a man to believe that what is powerful in
society is not good, as it would be for the religious believer to feel his God
was evil“ (Gouldner, 1970: 486). Gouldner attempts to solve his problem by
proposing a „radical“ sociology whose „historical mission . . .
is to transcend sociology as it now exists“ (Ibid: 489), and which helps
„to produce a new breed of sociologists“ (Ibid: 490), the MANLY ONES, who see the „traditional theories . . .
as timidity generating creations of timid men“ (Ibid.: 8).

Gouldner experiences and expresses in his poetic
aesthetic way the ancient discord between „is“ and „ought“,
i. e., in Gouldner’s terms, the tensions between permitted and unpermitted
worlds. Gouldner shares this problem with many latter-day social scientists
whose patron saint he became as the author of „the coming crisis of
western sociology“.

The large number of social scientists who share
these beliefs, coupled with their zealous intolerance against non-believers
(which they justify by the usurpation of the epithet „radical“),
forces the excavation of the long-buried discussion on values in social
science. As a result of this discussion, we hope to demonstrate the position
from which the concept of marginality and the model of „Games for Criminal
Status“ were developed.

We clarify our position with the very arguments
Gouldner uses to establish his „radical“ sociology.

A crucial issue in the discussion of values and
social science seems to be: Social science — what for? This question underlies
Gouldner’s complaint, or, rather, the sensation of impotence to
„discover“ a satisfactory answer to the question i s Gouldner’s
complaint as well as the „coming crisis of western sociology“.

Social scientists, who regard themselves as radical
and are recognized as such by other social scientists who also perceive
themselves as radical, have made the question of the purpose of social science to
their existential core problem. They seem to expect an answer of astral quality
and, being unable to discover the answer, discover instead the decline of the
sociological occident.

We shall propose an answer of more telluric than
astral quality. We shall contemplate the idea that social science is
„merely“ a form of consciousness, assuming „that human events
have different levels of meanings, some of which are hidden from the
consciousness of everyday life“ (Berger, 1963: 29). The purpose of social
science, then, is the satisfaction of a playful curiosity on the part of the
social scientist (Weber, 1918: 137).

According to some people’s standards of manliness,
such an approach certainly has to appear unmanly. Manly, according to their
standards, is to change the world „for the better“. To avoid part of
their harsh criticism and make the approach at least initially more palatable,
let us rename our approach from „playful curiosity approach“ to
„l‘ art pour l‘ art approach“.

Our endeavor then is to show that even the
„manly“ intention of changing this world for the better necessitates an
„l‘ art pour l‘ art approach“ to social science. From the standpoint
of either approach, we contend that it is well worthwhile to insist on a
division between pure and applied social science, even if the demarcation line
is difficult to draw and, as Gouldner’s complaint exemplifies, is also painful
to draw.

Looking backward to Gouldner’s complaint, then, we
shall have seen and will continue to see that jolting collective reality are a
painful activity and that bewailing this fact does not live up to Gouldner’s
own „radical“ standard of manliness in social theorizing — that
Gouldner‘ s complaint is the „timidity-generating creation of a timid man“
(Gouldner, 1970: 8).

2.2.2. The Problem of
Common Sense: Social Science as a Social Activity

Gouldner suggests „that a significant part of
social theorizing is a symbolic effort to overcome social worlds that have
become unpermitted and to readjust the flawed relationship between goodness and
potency, restoring them to their ’normal‘ equilibrium condition, and/or to
defend permitted worlds from a threatened disequilibrium between goodness and
potency“ (Gouldner, 1970: 486). He scorns the traditional sociological
theories as wrong or irrelevant, and postulates a „radical“ social science,
which recognizes the existence of the fact that a social scientist exists in a
field of social gravitation and works toward a solution of the dilemma of
circularity of social theorizing.

As a first step to refute Gouldner’s argumentation,
let us discuss the implications of the notion that social science is a social
activity. Based on the idea that our social life, as social scientists and as
human beings, happens in a socially constructed reality, where the personal
reality of the „normal“ individual is embedded in some collective
reality, and where collective realities of different generality are congruous
with each other, we shall investigate the interaction between some of these
realities and the theorizing of social scientists.

Let us start with a very general collective reality,
the Z e i t g e i s t, and observe how Gouldner’s „radicalism“ is the
consequential development of the interaction between the Zeitgeist and social science in the United States.
We call this type of interaction the gross cycle of social science and contrast
it to other, smaller cycles in which the social scientist finds himself, e.g.,
in the interaction of his theorizing to that of his colleagues in the community
of recognized scientific practitioners. The Gross Cycle of
Social Science

American social science is characterized by
unparalleled provincialism. Reading any book on social science generates the
impression that social science is non-existent outside the American Empire. The
provinciality of American social science makes the task of describing its
development up to the present day neo-romanticism relatively easy.

We can distinguish roughly three stages in the
development of American social science: The Chicago School, the time of
empiricism and grand theory, and the latter-day romanticism. Each of the three
stages reflects the political and economic consciousness of its time and
reflects the gross cycle American social science is subjected to.

The once dominant Chicago
School is perhaps the first
faint-hearted budding of social science in America, worth mentioning for our
purposes. As an heir to the muckraking tradition, this school replaced
newspaper exposes with monographs. It was a school of „social
issues“, of „social problems“, with an outlook of the
„reform-minded Midwest and its dislike
for the industrial and financial centers, with their capitalist exploitation,
boss rule, and religious ’superstitions‘ “ (Roth, 1971: 47). It had its
apex between the two world wars, when the underdog, the poor, the derelicts,
the hoboes, the prostitutes, and other socially weak persons were the overdogs
of social science.

The second significant development of American
social science replacing the Midwestern isolationism of the Chicago School
started with recognizing the possibility of employing social science for policy
decisions. It was the period of grand theory and empiricism. An immense amount
of funds started pouring in for those demanding money as social science became
a respectable activity. It was the time of empire-building, both by the
empirical-minded social science and the Parson-type grand theory, which today
is the favorite object of latter-day sociologists‘ scorn.

In this time, the empirical method developed into empiricism
of sometimes grotesque extent. A description of the typical activities of this
period, self-reported and without self-consciousness, is Hammond’s (1967) collection of Sociologists at Work: Essays on the Craft of Social Research. In Hammond’s anthology, we
can observe how some of the most illustrious names in social science make fools
out of themselves. The perversity of the situation is emphasized because even
the most sardonic critic of empiricism could not have written a better satire
on the movement than the immediate participants.

Since the beginning of the sixties, we can observe a
decline of empiricism as government support for social research decreased. The
much-abused grand theory remained, changing its content but not its form.
Mills‘ T he Sociological Imagination (1959)
is the classic example. The decline of empiricism and the rise of
„radicalism“ in social science came when the fiasco of the American
foreign policy with its ill-concealed neo-imperialism could no longer be
masqueraded behind a helping-brother image. Inside and outside the United
States, it became apparent to the on-lookers that the United States
was an ill-tempered Big Brother in its tantrums. As a consequence,
„radical“ sociology is introverted sociology.

„The severe domestic crisis in the United
States since the mid-sixties in general, and fiascoes such as the abortive
Project Camelot in particular, have resulted in a refocusing on domestic issues
and a retrenchment of foreign area studies“ (Roth, 1971: 48). The
attention of social science is concentrated on domestic social problems and the
political struggles within the ghettos and the universities. “It revolves
around the poles of legitimacy and usurpation, analyzing the perennial triangular
struggle among rulers, staff, and subjects throughout history“ (Roth,
1971: 48).

The tottering of the colossus had its repercussions
on its parasites. The once so almighty empiricist social science and the
traditional grand theory today are subjected to an increasingly harsher
critique. Its critics are the self-professed radicals. The new social science
is of an introverted, brooding kind. The young ones dislike the world their
elders have left them. In a typical generational conflict, the elders are
attacked for their principles, which are perceived to have led to the inherited
quagmire. The reason for the elders‘ failure is sought in their
„irresponsible“ attitude towards values.

The elders profess to be value-free—their offspring
revolt against this. We, therefore, can understand the present reaction against
value-free social science as the reaction of disappointed former believers and
their sons under the guise of the epithet „radical“.

Gouldner has made himself into the spokesperson of
the latter-day sociologists and has won their recognition by acclamation.
However, in the acclamation lies the irony of Gouldner’s situation. When
Gouldner propagates a „reflexive sociology“ and proclaims it as being
radical, the very success of his ideas is a counter-indication of the claimed
extremeness of thought. Instead, as the wide acceptance of his ideas shows,
Gouldner attempts to keep pace with the rapid and sweeping changes in the Z e i
t g e i s t, instead of setting the pace.

Unlike C. Wright Mills, he merely joins the chorus, chanting about the
„emperor without clothes“ — the traditional social science. This
makes him no different from the gullible emperor. He, too, is a follower,
someone who, in his quest for understanding, moves through the world of men with respect for the usual lines of
demarcation. Gouldner’s case, summarized in his own words by what we called Gouldner’s
complaint, exemplifies the case of the social scientist today just as well as
yesterday. A social scientist is inevitably caught in the gross cycle of social
science. Even if he perceives himself as a pacesetter, he is likely to be such
a person because „that’s where it’s at, man“. He is likely to be
„radical“ if it is permissible to be so, or as it appears nowadays
because it seems indecent not to give this appearance.

Like his brother, the „man on the street“,
the social scientist is restricted in the realities he constructs by the
collective reality in which he lives. The „radicalism“ of Gouldner’s
type is a permitted reality embedded in the very general reality, which we call
Zeitgeist. The very fact of its wide
acceptance shows this interrelation.

Gouldner, who thinks that in his quest for
understanding he moves „through the world of man without respect for the usual lines of demarcation“ (Berger,
1963: 18), actually cringes in front of the Zeitgeist
which presently demands „radicalism“ from the youthful adept in
social science — and Gouldner has not much choice to act differently because
social science, according to his confession, is a social activity, and as such
takes place within some form of collectively constructed reality. The Petite Cycles
of Social Science

We now turn our attention to the interaction of
social science with collectively constructed realities of lower generality,
which are perhaps less controversial in respect to their existence than the
reality Zeitgeist.

We now discuss the realities of groups of scientists
adhering to and defending a particular school of thought within their
discipline, and also another type of realities of still lower generality, the
personal reality of the individual scientist. We call the interaction of these
two realities with the research of the social scientist the petite cycles of
social science.

As in the previous section, we start with the basic
assumption — which we share with Gouldner — that the imposition of a
conceptual order upon data is a social activity (cf. also, Conant, 1947;
Beveridge, 1957). Social science as a social activity, then, exhibits
characteristics common to other social behavior. The irony of the situation is
that we investigate social behavior and that even a study of the activity of
the social scientist as an object of investigation is no escape from this

In traditional sociology texts, the problem of the
petite cycles of social science is usually evaded by a chapter on objectivity
in social science or, more elegantly, by a chapter on what is called
„methodology“. C. Wright Mills was one of the first influential
writers in social science to knock this Holy Cow, and we intend to do our

Christian Morgenstern, a German satirist, has
caricatured the circularity of social analysis in his Palmstroem- poem,
„The Impossible Fact“ (Morgenstern, 1963: 34f.). The philistine hero,
Palmstroem, is run over by a car in this poem. Ignoring his death, Palmstroem
conducts an analysis of the situation:

Tightly swathed
in dampened tissues
he explores the legal issues,
and it soon is clear as air:
Cars were not permitted there!

And he comes to
the conclusion:
His mishap was an illusion,
for, he reasons pointedly,
that which must not, can not be.

Like Gouldner in his complaint, Morgenstern’s
philistine experiences the discord between „is“ and
„ought“, between permitted and unpermitted worlds. Like Gouldner,
Palmstroem solves his dilemma in favor of his idiosyncratic normative reality.
Of course, compared with Gouldner, the phantasy-creature Palmstroem is
immensely less sophisticated and openly egotistical; however, he is
refreshingly illustrative: that which must not, cannot be.
Palmstroem illustrates a petite cycle of social science, where Gouldner’s petit
mal prevents him from realizing it.

Let us start with the observation that the
Palmstroem- problem if it were the object of a social scientific investigation,
would be a problem of Palmstroem’s cognitive balance. Were Palmstroem a social
scientist, however, we could surmise that both he and his social scientist colleagues
would study Palmstroem’s problem with „reality“ as a question of
scientific objectivity. What would be the dimension of inquiry if a
psychiatrist or psychoanalyst did it is superfluous to mention. If we disregard
the soul-doctors, we are still left with the paradoxical situation that the
dimensions of inquiry into „reality“ are dissimilar depending on who the
inquisitor is and who the subject is.

The usual argument for maintaining the different
standard is that the activities of the „scholar“ and the „man on
the street“ are fundamentally dissimilar, insofar as the one balances his
cognition in a way, which is intercommunicable, whereas the „man on the
street“ orders his cognition in a non-methodical way.

This argument brings the discussion into the
disputes raging among the philosophers of science. In the context of this
paper, our counter-arguments will necessarily appear scanty and only scrape the
surface of the ontological problem involved. However, so are the other side’s
arguments, generally, if they are proffered in a context other than the philosophy
of science.

As a response to the argument, which attempts to
justify different dimensions of inquiry depending on who is the object of
inquiry, the scholar or the „man on the street“, we distinguish
between the context of discovery and the context of demonstration.
Traditionally, the context of discovery is left to psychological analysis. In
contrast, the logical analysis of thought is concerned with the context of
demonstration, „i. e., with the analysis of ordered series of thought
operations so constructed that they make the results of thought
justifiable“ (Reichenbach, 1947: 2).

However, Hans Reichenbach (1938: 311), who originated
the distinction of two realms of analysis, naming them context of discovery and
context of justification, already provides at the same time a word of warning
not to overindulge in methodology:

When we call logic analysis of
, the expression should be interpreted so as to leave no doubt
that it is not actual thought, which we pretend to analyze. It is rather a
substitute for thinking processes, their rational r e c o n s t r u c t i o n,
which constitutes the basis of analysis. Once a result of thinking is obtained,
we can reorder our thoughts in a cogent way, constructing a chain of thoughts
between point of departure and point of arrival; it is this rational
reconstruction of thinking that is controlled by logic, and whose analysis
reveals those rules, which we call logical laws (Reichenbach, 1947: 2).

Methodology and discussions of scientific
objectivity with their ceremonial formalism, therefore, tend to obscure the
fact that social-psychological dynamics are essential not only in the context
of discovery but also in the context of demonstration. The same tendency to
obscure more than elucidate the problems of the context of demonstration can be
ascribed to those philosophers of science, who, in a logico-philosophical jargonese, write
learned tractates to lay bare the simple-mindedness of formalistic methodology.

Before we get entangled in argument and
counter-argument in their language and on their „level“, let us leave
the tract of science, which philosophers claim as theirs, and take the trophy
of our sneaky and short hunting-excursion with us into the relative security of
a social science discussion. Our trophy is the notion that undisputed by the
philosophy of science, cognitive elements play a role in the discovery and the
demonstration of scientific theories in the same way as they do in the everyday
life of the „man on the street“. For empirical research, this is
confirmed by Rosenthal (1966), who shows by extensive examples that different
observers of the same experiment can come up with results widely differing from
each other. Rosenthal’s study was predominantly concerned with the influence of
the experimenter’s expectations on the subject of investigation, i, e., changes
in reality by expectations. However, the notion that cognitive elements are
present in the discovery and the presentation of a theory implies that
perceptions used to justify a theory are themselves firmly formed and guided by
this theory. Perceptions are dependent on the context in which they are made. Parts
of the context are cognitive factors like the set, expectancies, and
hypotheses. This means, among other things, that under the influence of
theories, there is a tendency towards a disposition and a habit to observations,
which favor theory-conforming observation results (Kuhn, 1962).

We show this through a simple example from the
history of science. According to an excellent account by Conant (1947), for
centuries, the principle that „nature abhors a vacuum“ served to
account for various phenomena, such as the action of pumps, the behavior of
liquids in joined vessels, suction, and the like. The strength of everyday
evidence was so overwhelming that the principle was seldom questioned. However,
it was known that one could not draw water to the height of more than 34 feet.
The simplest solution to this problem was to reformulate the principle to read that
„nature abhors a vacuum below 34 feet“. This modified version of the
horror-vacui theory was again satisfactory for the phenomena it dealt with
until it was discovered that „nature abhors a vacuum below 34 feet only
when we deal with water“. As Torricelli has shown, „nature abhors a
vacuum below 30 inches“ when it comes to mercury. Displeased with the
crudity of a principle, which must accommodate numerous exceptions, Torricelli
formulated the notion that the pressure of air acting upon the surface of the
liquid was responsible for the height to which one could draw liquid by the
action of a pump. The 34 feet limit represents the weight of water, which the
air pressure can maintain, and the 30-inch limit represents the weight of
mercury that air pressure can maintain. This was an entirely different and
revolutionary (Kuhn, 1962) concept and its consequences had a drastic impact on

Let us analyze the development of the concept of air
pressure for its cognitive elements. We see that at each stage of its unfolding
process, the available perceptions of reality were structured to a consistent construct,
which was able to withstand a methodological inquiry — until new facts were
introduced and perceived as such. Not at any point in the concept’s development
can we point out a methodological mistake. However, we can observe how a
cognitively consistent reality is constructed at each step.

The same process is the object of our study when we, as social scientists, observe the
„man on the street“ building his cognitive world out of perceptions
of his physical and psychical environment. The process is used in the
Rohrschachtest, Murray’s Thematical Apperception Test, and IQ-test.

Solving problems in an IQ-test can be regarded as
perhaps the closest model to constructing scientific theories. There we find,
put to the extreme, Reichenbach’s notion that a methodologically correct
construction is a cognitive construct that is rationally reconstructable. The
„man on the street“, i. e., the subject of an IQ-test, therefore, is
not so fundamentally different in the ways he arrives at the construction of
his everyday world from the scientist, as to warrant the examination of his
cognitive world with two different standards: one methodology, and the other a
test of his cognitive world by „someone who knows“. Neither the
scientist nor the „man on the street“ are persons „who
know“, but both are persons who deal with a socially constructed reality.

If the scientist takes it as warranted to perceive
the constructs of the world of the „man on the street“ with
professional suspicion, there is no reason not to regard his own activity in
the same way. He, too, is caught in a petite cycle of science where social
science is a science and simultaneously a meta-science. Therefore, social
science as a meta-science stands side-by-side with methodology in the
examination of the context of demonstration.

However, let us not stop and assume there is merely something,
which we termed the gross and the petite cycle of social science, the latter
standing for the interaction of the cognitive world with the individual
scientist. That would mean oversimplifying our initial model in which we
claimed the existence of many socially constructed worlds related to each
other, similar to the Chinese boxes buried within each other, as collectively
constructed realities of different generality.

The next world we want to discuss stands in close
relationship to the world we outlined as constructed due to the universal
tendency to cognitively balance our perceptions. This is the world of normal

Historians of science like Conant (1947) and Kuhn
(1962) have consistently pointed out that there are beliefs that are held by
the majority of scientists until a „scientific revolution“ replaces
them with new concepts. Some of the concepts appear hilarious today, although
they were not so to our ancestors. As an example, for a long time, people
believed that the universe revolved around the earth. Moreover, it is not so
long ago that it was a scientific dogma to hold that the earth was flat, with
an edge off of which the incautious might fall; or, to take our previous
example, there was a time when the action of a pump was explained by the
„horror vacui“.

Instead of following the historians of science who
attempted to find the regularities governing the change of concepts through a
historical analysis, we shall use their presentation to demonstrate that
science, and of course, social science is a social activity.

The very persistence with which concepts were
defended against a new school of thought, which had a differing concept, is
perhaps the strongest indication that the imposition of conceptual order onto
data is a social activity. As such, it exhibits characteristics common to other
social behavior.

Merely the intercommunication with his professional
colleagues makes even the lone investigator someone who is involved in a group
activity. His research is in some way the product of others’ prior work and the
precursor of subsequent work (Nisbet, 1970: 28). Within the group of scientists,
there is a powerful control mechanism, which guards the fact that the
personally constructed reality of most individual scientists falls within the
collectively permissible reality.

Kuhn’s (1962) definition of normal science implies
the existence of a powerful defense against non-believers. This can easily be
observed in today’s scientific life. There are a vast number of screening
procedures, formal and informal, guarding the threshold of the community of
recognized scientific practitioners of the respective discipline, barring
heretics from the entrance into or continuance of membership in the community.
The procedures and effects of these control mechanisms are similar to the controls,
which hold the „man on the street“ within the framework of a
collective reality, be it a small group reality or a national reality.
The screening functions, yesterday taken care of by the Holy Inquisition, are
today effectively and efficiently taken care of by the universities and
grant-dispensing agencies (Kerr, 1966). The admission procedures range from admission
to graduate study, acceptance of a project for agency funding, acceptance of a
publication, and acceptance as a tenured faculty member (certified scientist).
There are ingenious control and guidance systems, like the grading system
guaranteeing constant surveillance and intellectual bondage in the early stages
of membership in the community; there are introductory courses that make sure
that the new pledge starts along with the „right“ path to wisdom;
there are degree programs; and last but not least there are the bestowals of
degrees of adjustment.
If the budding savant finally identifies with and is accepted by the community
of recognized practitioners of his choice, this usually means that he also
identifies with and accepts the basic concepts (collectively constructed
realities) in this science. At that time, when his „revolutionary“
spirit seems to be broken, he is allowed to help build and mend the
collectively constructed realities, now that he is fully conscious of the
likelihood of disastrous results for his „career“, should he propose
a nonconformist reality.
Therefore, we can only agree with Lemert’s observation:

Novel paradigms most often are created by youthful
scientists, primarily because they are less committed by prior practice to the
traditional rules of normal science, they are freer to conceive new images of
the world, new sets of rules for problem-solving, and to sympathetically
entertain new classes of facts. By the same reasoning, resistance to new
paradigms is strongest among older scientists, who have long-standing practical
commitments to the established ways of perceiving their world of study (Lemert,
1970: 7).

In the regular day-to-day study of social life, the
cognitive world of the social scientist is subjected to a control system whose
presence does not necessarily penetrate the scientist’s consciousness. This
system is built around him, keeping his cognitive world within the boundaries
of some collective reality. As social scientists, they are not different from
the „man on the street“. In the community of scientific
practitioners, everybody is the victim and at the same time the enforcer of the
control system, which guides us through „reality“. The controls
become visible, they enter our consciousness only when the system penalizes a
deviation overstepping the boundaries of the permissible or tolerable.

„The activity in which most scientists
inevitably spend almost all their time“, therefore, is normal science.
Normal science, however, „is predicated on the assumption that the scientific
community knows what the world is like. Much of the success of the enterprise
derives from the community’s willingness to defend that assumption, if
necessary at considerable cost“ (Kuhn, 1962: 5). Value Neutrality
as an Attempt to Avoid Ought-Worlds

Our operating premise was that social science is the
endeavor of conducting a rational activity to construct second-degree realities,
which successively approximate the „real“ reality, intending to
construct a reality capable of integrating at least one more observed fact than
the presently accepted reality. In the previous section, we showed that the
social scientist is inevitably caught in his time and place’s social reality.
We demonstrated that this is the case because there exists an interaction
between the constructed reality of the social scientist, the second-degree
construction, and the conceptual reality of his social environment. The
situation of the social scientist is more difficult when compared with the
situation of his colleague in „hard“ science since he is directly
faced not only with the constructed realities of his field but also with the
constructed realities of the objects of his studies, namely the social
environment, of which he is himself a part.

In this section, we shall deal with ways of
attenuating the circularity of social research. We shall do this by
distinguishing in social science between „pure“ science and
„applied“ science and postulating that only a pure science of the
social environment will construct a social reality of the second degree. By
pure science, we mean an activity which builds its constructs, not like
Gouldner and the radical-reactive approach to social science do in order to
„transform as well as to know“ (Gouldner, 1970: 489), but in order to
satisfy a playful curiosity about the everyday construction of social reality.
In pure social science, so we claim, we have to evaluate first-degree
constructs. However, we have to avoid using valuations, which in this context,
we understand as „the practical rejection or approval of a phenomenon
capable of being influenced by our actions“ (Weber, 1923: 475). In
contrast, we mean by „evaluation“ in social science the appraisal of
first-degree constructs of reality.

Like Dahrendorf (1968: 16ff. ), we might distinguish
six points in the process of acquiring knowledge in social science at which
value judgments are at issue. These points are the choice of the subject, the
selective formation of theories, values as the subject of inquiry, ideological
distortion either by over-extending specific propositions or by presenting
untestable and thus speculative propositions as scientific, the application of
scientific results to practical problems, and finally the social role of the

So much has been written in the past about the last
four points that we forgo their discussion. Instead, we turn our attention for
the rest of this chapter to what Dahrendorf sees as the first two points in the
process of acquiring knowledge in social science: the choice of the subject and
the selective formation of theories.

Dahrendorf, Weber, and generally the rest of the
literature on values tend to disregard these two points as unproblematic in the
discussion of values. They acknowledge that value-judgment may come in at both
points, but they designate the influence there as „pseudo-problems“
(Dahrendorf, 1968: 10) because „clearly . . . the choice of subject is
made in what may be called the antechamber of science, where the sociologist is
still free from the rules of procedure that will later govern his
research“ (Dahrendorf, 1968: 7), and because:

A selective point of view, such as the conservative
bias of the sociologist described above, does cause a scholar to see what he
wants to see and be blind to other things. However, this merely tells us how
the scholar has come to formulate a given hypothesis X; it does not tell us
whether hypothesis X is true or false, tenable or untenable. Neither the values
nor the thought processes of a scientist determine the validity of his hypotheses;
their validity is determined only by empirical tests. Nor can empirical tests
as such affect the values and thought processes of the scientist in any way
(Dahrendorf, 1968: 10).

These two quotes from Dahrendorf represent a very
general position in the literature. We shall try to contrast our position that
value judgments influencing the choice of a subject and affecting the selective
formation of theories are highly significant disturbances in the scientific
construction of reality. From our position, the existence and influence of
value judgments are not „pseudo-problems“ that can be swept under the
carpet with an „even if“ argument but are problems we have to deal

Characteristic of the position of our opponents is
the devout belief in the power of methodology. Methodology is the Holy Cow
around which social scientists were „disciplined“ to dance,
constantly encouraged, and if necessary forced by its High Priests, the
methodologists. The hierarchy of the methodologists is founded on the belief
that there is an empirical world out there in which we can test our hypotheses
and theories for their truth. Reichenbach’s exhortation (1947: 2) that the
analysis of thought consists in its rational reconstruction is dismissed, just
as is Popper’s (1935) conclusive argumentation that verification can never be a
proof for the correctness of a theory or hypothesis, but that falsification can
only be a proof for their incorrectness.

In our last section, we demonstrated that
methodology is not the only meta-science to social science in the context of
demonstration. Social science, like any science, is a social activity and, as
such, is subjected to the patterns we are used to observing with the objects of
social scientific investigation. Therefore, any formation of social scientific
theories occurs within the circle: social science is a social activity. Because
of our position as human beings, we are by definition incapable of breaking out
of this circle. Methodology fails to be of any help (cf., Merton, 1962: 19).

Regardless of at what point of the scientific
inquiry they seep in, value judgments are not automatically eliminated if we
follow the rules of empirical social science. On the contrary, due to the
inherent deficiency of methodology, which can examine only whether the
presented construction of thought is rationally reconstructable, value
judgments made in connection with the selection of the subject of the
investigation and the formation of the theory can be expected to have a continual
influence on the further course of the inquiry.

Regarding the influence of value-judgments at these
points of scientific inquiry as pseudo-problems, or as negligible as Dahrendorf
does explicitly and most social scientists do implicitly, reminds one of the
tricks that toddlers sometimes use to protect themselves against an unpleasant
reality: They cover their face with their hands in order to extricate
themselves from the presence of an uncongenial actuality. Moreover, the result
of this course of action is just as dangerous for the social scientist as it is
for the toddler because refusing to accept reality does not necessarily
eliminate this reality. Not always are we so fortunate that the situation
changes through a redefinition of a situation.

Therefore, we cannot avoid examining the problem of
values and their influence on the choice of subject, the selective formation of
theories, and their effect on the scientific results. Instead, we had better
admit that the existence as a human being keeps the social scientist
unalterably within the circle: Social science, a social activity. He lacks the
Archimedean point of observation and never will attain it. The circle is there.
We might just as well acknowledge its existence and admit that we are incapable
of breaking out of this circle. However, we might try to attenuate it.

We can perceive the history of science as the
history of successful successive attenuations of this circle. Consecutively
accepted collective realities were replaced by newly constructed realities,
each better fitted to the observations of the world than the previously valid
one. Therefore, the attempt to attenuate the circle of social science is also
an attempt to step up the progress we make in approximating our collectively
accepted realities to some „real“ reality.

Value judgments in scientific investigation, however,
have the opposite effect. They lead us to Gouldner’s permitted and unpermitted
worlds, and lastly, to Gouldner’s complaint. They amplify the vicious circle
which crowds the accepted reality into an „ought“-world, pushing it
away from the unknown and therefore intimidating construction of a more
encompassing collective reality towards the secure womb of a world where
„things are as they should be“.

We can distinguish between two types of value
judgments, the ones we are conscious of and the others we do not know and which
we usually call bias. Both types tend to lead us to an
„ought“-reality which theoretically might, but does not necessarily
have to, coincide with a more encompassing reality. Both types of value
judgments constrain the attempts to approximate the „is“-reality.

Weber and other social scientists accused of
postulating a value-free social science restricted themselves to the postulate
that pure science, which wants to be a rational activity, has to eliminate
conscious value judgments. The subconscious value judgments (the bias), while
obstructing the quest for reality, were seen as something which, because of its
character as the subconscious, cannot be avoided as long as they remain

This opinion is voiced by Dahrendorf as far as the
choices of a subject of scientific investigation are concerned:

. . . the choice of subject is made in . . . the
antechamber of science . . . It is probably unrealistic to insist that value
judgments be eliminated from the choice of subjects . . . (Dahrendorf, 1968:

Concerning the selective formation of theories,
Dahrendorf states: „… many sociologists, in dealing with their subjects,
see only what they want to see“ (Dahrendorf, 1968: 9). Nevertheless, he
later negates the necessity of eliminating value judgments from the formulation
of scientific theories. He accepts Popper’s argumentation „as
convincing“ (Dahrendorf, 1968: 9):

All scientific descriptions of facts are highly selective…
It is not only impossible to avoid a selective point of view, but also wholly
undesirable to attempt to do so; for if we could do so, we should get not a
more ‚objective‘ description, but only a mere heap of entirely unconnected
statements. But, of course, a point of view is inevitable; the naive attempt to
avoid it can only lead to self-deception, to the uncritical application of an
unconscious point of view (Popper, 1952, II: 260f.).

Popper’s and Dahrendorf’s points might be well taken
insofar as natural science is concerned. There we find a dichotomy between
applied and pure science, which can look back onto a long tradition. The social
scientist’s situation differs from that of the natural scientist. Social
science has a comparatively short history, and even more importantly, a social
scientist is himself inevitably part of the investigated matter, regardless of
the formulation of his subject.

The natural scientist can minimize his bias if he
strives to understand a system. In his thinking, he is probably socially
constrained to the area of „normal science“; but he neither has an ax
to grind with the bacteria of his choice nor an existential vested interest in the
composition of a molecule.

The social scientist might have the same goal as his
colleague in natural science, namely system understanding. However, inevitably
the social scientist experiences in his search for reality the interference of
an „ought“-world, where things are as they should be. This
„ought“-world does not only influence the way the social scientist
wants to transform things but also the way he perceives these things. Therefore,
the system understanding is hampered by the inescapable fact that the social
scientist is a system participant.

Let us summarize our discussion up to this point: We
have stated so far that we social scientists are inescapably entangled in the
circle: Social science is a social activity. The circle prevents us from reaching
our goal, the approximation of the „real“ reality. We argued that the
inclusion of value judgments into social scientific inquiry amplified the
circle: uncorrected and uncorrectable by methodology, whereas their exclusion
attenuated it in a way that we can hope to construct in successive attempts a reality,
which is better, fitted to our observation of facts than the presently accepted
one. We decided conscious value judgments should be eliminated from social
inquiry whenever possible. Our problem now is to devise a way to diminish the
value judgments, which subconsciously flow into social inquiry at the two
critical points, the choice of the subject and the selective formation of

We know that our bias (subconscious value judgments)
can stem from at least two sources: our own value make-up and the conceptual
make-up to which we were disciplined when we were trained to become adept in
social science. Both sources are certainly not independent from each other, but
for the sake of presentation, we shall treat them as such.

The conceptual makeup, which the social scientist
acquired when learning his discipline, is the more controversial source of
bias. The irony of the situation is that, on the one hand, this bias is
necessary to do social science at all, but on the other hand, it tethers the
social scientist to the area of „normal social science“. Therefore,
it is a two-horned problem where the decision for either alternative is no

A clue to where we might look for an understanding
of the problem was given by Lemert in the quote we presented above (p. 34):

Novel paradigms most often are created by youthful
scientists, primarily because they are less committed by prior practice to the
traditional rules of normal science, they are freer to conceive new images of
the world, new sets of rules for problem-solving, and to sympathetically
entertain new classes of facts (Lemert, 1970: 7).

Perhaps when he wrote this sentence, Lemert was
looking backward to the time when, as a youthful scientist, he introduced the
concept of secondary deviance (Lemert, 1951). Moreover, perhaps he was ruefully
comparing himself to his youthful pastime when he contrasted the older
scientist in the following sentence:

By the same reasoning, resistance to new paradigms
is strongest among older scientists, who have long-standing practical
commitments to the established ways of perceiving their worlds of study
(Lemert, 1970: 7).

Another clue is given by a quote from Max Weber:

An anarchist can surely be a good legal scholar.
And if he is, then indeed the Archimedean point of his convictions, which is
outside the conventions and the presuppositions which are so self-evident to
us, can equip him to perceive problems in the fundamental postulates of legal
theory which escape those who take them for granted. Fundamental doubt is the
father of knowledge (quoted in Roth, 1971: 50).

If we combine both clues, Lemert’s „youthful
scientist“ and Weber’s „legal scholar who is an anarchist“, with
the idea we presented earlier, namely that a scientist is part of a community
of recognized practitioners in his discipline, it seems that we have moved
closer to indicating where a problem solution might be located. Both
scientists, the youthful scientist, and the anarchistic legal scholar, occupy a
somewhat marginal position in respect to their more recognized colleagues in
the community of scientific practitioners.

Lemert’s „youthful scientist“ is someone
who still has to prove his ability as a scientist. Fundamental doubt in what he
has learned is his prerogative of youth. He cannot yet look back on his
achievements, nor does he have any to defend. He can only establish himself by
voicing his fundamental doubt and then, if successful, by attempting to defend
his newly acquired position. He stands with his back against the wall, and his
only chance of proving himself lies in front of him.

Weber’s anarchist is in a similar situation. He
cannot lose much more status in the community of scientific practitioners by
advancing unconventional ideas, as long as he makes them rationally
reconstructable. His sense of self-preservation makes him sensitive to the
presuppositions taken as self-evident by most of his professional colleagues.
Both scientists, therefore, are in a position where they cannot lose much more.
However, the presentation of unconventional ideas has inherent the small
likelihood of a significant gain in scientific status if these ideas are
accepted. This situation, then, makes these marginal scientists more perceptive
of professional bias and its weakness.

The empirical method in social science uses this
notion to select the initial informants for a social system under

Being not fully integrated among colleagues or
somewhat alienated from the bureaucratic system may have made these officials
more critical of their social
environment, less restrained by feelings of loyalty from sharing their
criticism with an outsider, and more interested in the approval of the observer
than were those who received much social support and approval within the
organization (Blau, 1967: 34).

Suppose we now transpose our notion that the
marginal person is more wary of and resistant to prevalent biases than the
integrated one to the field of social science. In that case, we arrive at the
seemingly paradoxical result that a „good“ social scientist is one
who is professionally not (yet) successful, or stated reversely, a
professionally successful social scientist can no longer be a „good“
social scientist. „Good“ stands in this context for „critical of
professional bias“, as we defined the goal of the social scientist to be
the attempt at constructing realities that successively approximate a
„real“ reality.

Of course, like any statement in social science, we
have to take both statements with a grain of salt. They certainly do not mean
that every marginal social scientist, due to his position in his professional
world, is critical and therefore „good“, and that every
professionally successful social scientist is ipso
uncritical to professional bias. However, they do mean that the
chances to escape professional bias are greater for the one who holds a
marginal position in his professional community. The professional bias of the
social scientist, being to some extent the precondition for the social activity
social science, enhances the social position of a scientist in the community of
scientific practitioners on the one hand; on the other hand, however, we
observe that it obstructs the very goal of scientific activity —

The examples, which justify our observation, are
ample. We had already mentioned Lemert as someone who made his most important
contribution when he was a youthful scientist. We can cite many others as
examples, e.g., Lewis A. Coser, Kai T. Erikson, David Matza, etc. However, in
contrast to these social scientists, there are others who, although recognized
for their contributions, remained marginal persons in their professional
community and thereby conserved a critical stance to the collective
professional bias. Their marginal position enabled them to produce works of
professional eminence over a more extended period than the youthful scientists
we mentioned above.

We find, e.g., Max Weber, C. Wright Mills, and
Erving Goffman in this category. Weber spent the zenith of his creative years
outside the university and, disregarding a short stint in the „Deutsche
Gesellschaft für Soziologie“, outside the professional organizations of
his time. C. Wright Mills never attained more professional recognition than an
associate professorship. Moreover, Goffman tenaciously refuses any public
involvement with his professional societies, whatever his reasons might be.
Compared with other academicians, he leads a „shy“ professional life,
secluded and immensely productive.

Whether consciously or unconsciously accomplished,
the attempt to avoid professional bias is undoubtedly a painful activity. Using
Gouldner’s terminology, the avoidance of professional bias requires the
scientist to endure the tensions existing between permitted and unpermitted
worlds and to stand the ambiguity, which is involved without ever trying to
attain a final adjustment between the two worlds for the sake of conserving the
ability to constantly construct new realities.

As we indicated at the beginning of the discussion
of professional bias, there is no solution to the paradox of social science.
Not every social scientist can be marginal since marginality presupposes, by
definition, centrality. However, being conscious of the dynamics involved might
prove to be a corrective factor for the inescapable professional bias.

Similarly, if we now discuss the other source of
bias influencing the professional work of the social scientist, the bias
originating from his personal value make-up, we shall see that a
„solution“ in the sense of a concrete instruction is not attainable.
Here, too, we shall fall back on what was developed within the empirical
method. We shall use the idea of a „self-obscuring methodology“.

The basic notion is that the social scientist, just
as in the empirical method where he professionally distrusts reasons given by
his subjects as explanations of their empirical world, he should also distrust
himself, his own subconscious bias. Our basic assumption for this discussion is
that there is a personal reality embedded in collective realities of different
generality, which provides a subconscious value make-up. This value makeup
flows unconsciously into the subject selection and the formation of theory. It
restricts the number of possible results to results, which do not necessarily
constitute the area where more encompassing social realities can be

If we assume the universality of personal realities,
our problem can also be characterized as the problem of diminishing the
influence of the subconscious value make-up on our reality construction as
social scientists. Analogous to the idea of the „self-obscuring
methodology“, what we can try to avoid is only the consistency of
distortion in our attempt at reality construction, not, however, the distortion
altogether. Then, our problem solution is searching for a way of
„value-randomization“ and not of value-elimination.

For this purpose, let us again look at the process
and the steps of scientific inquiry. Previously we had accepted Reichenbach’s
notion that methodology is the rational reconstruction of thought. Reichenbach’s
notion, however, did not imply that the steps of scientific inquiry are taken
sequentially, starting with the choice of the subject and ending with the
result of the inquiry. We should note that Reichenbach merely stated: If the
inquiry wants to withstand a
methodological examination, then it has to be rationally re-reconstructable. Like
the „man on the street“, the scientist frequently has a desirable
result in mind at the very start of his inquiry, desirable according to his
personal makeup. He then presents it in a way that makes it rationally
reconstructable. He builds back from the result to the subject selection and
fills in the rest. The methodology does not have any safeguard against this
kind of procedure. The traditional value-free sociology demands that the
scientist be conscious of value judgments in his inquiry. However, how can he
build in a safeguard against the influence of his subconscious value makeup if
he cannot even identify them due to the character of these value judgments?

The search for a solution has to start at the cause
of the problem, which we identified as the existence of desirable results of
scientific inquiry. They are desirable as these results have an
action-implication consistent with the personal value makeup.

It is our claim, then, that abdicating as admissible
results of social science the results which have any action-implication, and
substituting them with the goal system-understanding, bears in it the
opportunity to randomize the inescapable bias which flows into the selection of
a subject and the formation of a theory. Separating social science from its
application increases the chances that we enter scientific inquiry as a
sequential problem-solving process. At the same time, this decreases the
chances that we build back from a desirable result so that the process is
rationally reconstructable as if it were a sequential problem-solving approach.
Our postulate for a social science, which is not a problem-oriented purposeful
activity, coincides with the old postulate that we should distinguish between
applied and pure science. It is based on the idea that we presented at the very
beginning of this chapter, namely, that applied science necessitates the
existence of pure science.

Indeed, the distinction between the two is hard and
painful to draw. However, as we understand it, so is science because it is the
recurring, however by definition unsuccessful, attempt to create a constructed
reality approximating the „real“ reality. It is constantly saying
goodbye to realities, which we have to like, where we knew our way around, in
favor of new intimidating and more encompassing realities, which are at the
time of their becoming still foreign to us.

Gouldner’s complaint and the demands of the
sociologists who like to see the values conserved are reactionary in character.
We find the same kind of nostalgia and romanticism of the past, which we found
so amusing with the sociologists of the Chicago School,
whose idyll was the „pure“ country-life where they came from.

Our accusation against the new radicals and their
intellectual Boy-Scout movement is not so much against their goals, as against
their attempt to declare as science what is rationalization, and against their
attempt to use the halo of science for their anabaptist movement. From history,
we know that the worst suffering was caused by people who believed that they
knew the standards for good and bad. This statement holds for Moses, Jesus, and
Marx, just as it does for Hitler and presently ruling military juntas. The
people who knew“ and their followers fell for two fallacies: the fallacy
that there are standards for goodness and the fallacy that „good“
parts produce — even according to their own standards — a good system.

2. 3. Towards a Scientific
Activity of Criminology

At this point, we want to summarize our discussion
and to take stock of what we can use to build our model of „Games for
Criminal Status“.

Traditional criminology, as we saw, found it
„convenient to assume that the deviant person is somehow ‚different‘ from
those of his fellows who manage to conform, but years of research into the
problem have not yielded any important evidence as to what, if anything, this
difference might be“ (Erikson, 1966: 5). From our phenomenological
standpoint, we argued that while we agree with the result of Erikson’s
analysis, we consider him and other labeling theorists as still adhering to the
traditional approach to criminology.

Though the labeling approach already recognizes that
„deviance is not a property inherent in any particular kind of behavior;
it is a property conferred upon that behavior by people who come into direct or
indirect contact with it“ (Erikson, 1966: 6), it still seems to cling to
the a-priori assumption that the processes, by which the quality of deviance is
conferred upon behavior, differ from everyday life processes which confer upon
behavior any other quality. This conclusion can be drawn from the fact that
there is no serious attempt to free the labeling approach from the shackles of
traditional criminology, i. e., the study of „crime“ and
„criminal“. Even Quinney (1970), whose approach can be described as
phenomenological, upholds the a-priori assumption.

We proposed to drop the a-priori assumption, joining
Durkheim in his notion that there is „no phenomenon that presents more
indisputably all symptoms of normality since it appears closely connected with
the conditions of all collective life“ (Durkheim, 1885: 11). The immediate
consequence is that we can no longer consider the study of „crime“
and „criminals“ as an area of social research, which can be
undertaken by itself, as „criminology“ in the traditional sense of
the word. The object of a scientific study, which might still be called
„criminology“, can only be social evaluation processes of behavior in
general, where evaluation processes leading to the label „criminal“
might be used as an example that clarifies a hitherto unknown part of the
dynamics of social evaluation, and where the knowledge about everyday
evaluation processes might be employed to reach an understanding of the
dynamics of the more ritualized evaluation processes leading to „criminal“

Given such an understanding, criminology would be
studying social differentiation as part of social science. On this background,
we developed the „Concept of Marginality“. There we perceive the
criminal status as one type of marginal status. Marginal status, in turn, is
seen as one type of social status which is of particular interest for social
science because it demarcates the boundaries of structured social inequality
and distinctly illustrates the three characteristics of a status position in a
social grouping, namely (1) universality, (2) scarcity, and (3) relativity.

These three characteristics of social status and the
notion that criminal status in one type of marginal status will give our study
its conceptual structure. How we go about building a model within this
structure is the second question. We attempted to find a part of the answer in
the dispute with the direction in social science, which professes to be

There we claimed that a metascience of social
science again is social science. We illustrated this view by showing the dependence
of social research on the social environment in which the research takes place.
Specifically, we pointed out that the type and scope of the research question
depend on Zeitgeist, personal value makeup of the scientist, and „normal
science“. Methodology, another metascience of social science, is of no
help for intervening in the circularity of social research, as it only examines
whether a research result is rationally reconstructable.

To attain second-degree constructions of social reality,
which are less influenced and determined by the everyday conceptual reality, we
proposed to distinguish between applied social science and „pure“, i.
e., not directly applicable, social science. We claimed only a „pure“
social science, taking an attitude of playful curiosity towards the conceptual
reality within which it takes place, has a chance of avoiding valuations and of
evaluating the social construction of reality.

We shall make use of this „l’art pour l’art“
approach to social science, as we call it, to construct our model of
„Games for Criminal Status“. This approach will help us break the
social shackles which have prevented criminology from becoming a scientific
activity. It will also help us break out of some traditional ways of thinking
in social science.

„Specialties have been part of the problem and
if we are going to think in new ways, we have to break out of traditional
thoughts and established ways. We have to break out of the disciplines that
perpetuate these ideas“ (Quinney, 1971: 46).


The elements of interactional bargaining and social
decision-making have been mentioned sporadically in criminological literature
(cf., „plea bargaining“), but they have never been perceived to
provide a framework within which criminological findings can be interpreted new
research can be undertaken.

In our introductory chapter, we proposed to extend
the notion of interaction as a bargaining process to all stages of the criminal
justice process. We suggested that the encounter be chosen as the unit of
analysis and that the succession of encounters of an „offender“ with
agents of the Criminal Justice System be perceived as an institutionalized
ritual of successive degradation ceremonies. In the second chapter, we
indicated that — once criminology is freed of common-sense notions and
recognized as part of a social science whose objective is the building of
constructs of the socially constructed reality — the pivotal problem for the
comprehension of how criminal status is awarded is the analytical apperception
of the dynamics of attaining and retaining social status in general.

To understand these dynamics, we present a model that
conceptualizes status acquisition and status maintenance in everyday life. We
call it a model of „Games for Social Status“. The model of
„Games for Criminal Status“, which we develop in the following
chapter, is a particular, however not dissimilar, case of the basic model.

The ideas, which will be used to build the model,
are not only the traditional social science ideas. In order to explore and
extend the concept of social exchange, we shall fall back on ideas that are
used in modern economics: decision theory and its offspring game theory.

In the first section of this chapter, we shall
demonstrate that modern decision theory complements the behavioral theory.
Based on the ideas developed there, we shall present a theory of elementary
social behavior, the „Games for Social Status“ model in the second
part of this chapter.

In the third section, we shall show that behavior is
a negotiated reality, socially constructed on the social construct of status.

3. 1. Decision Theory as
Complementing Behavioral Theory

John C. Harsanyi recently turned his attention to
the problem of whether the models of normative decision theory developed in
game theory provide instruments that allow the analysis of elementary social
behavior. In his two papers, „A Bargaining Model for Social Status in
Informal Groups and Formal Organizations (Harsanyi, 1966) and
„Individualistic and Functionalistic Explanations in the Light of Game
Theory: The Example of Social Status“ (Harsanyi, 1968), Harsanyi attempts
to analyze social status in terms of some primary human motivations by asking
the questions why all people seek high social status and why some people are
granted high social status by others. Based on a game-theoretical bargaining
model, the resulting theory is very similar to the theory of social status
arrived at by Peter M. Blau (1964), who based his work on extensive empirical
research and did not use game-theoretical considerations.

In our study, we shall follow Harsanyi in his
attempt to reconcile normative and descriptive decision theory. We shall show
how the (normative) game-theoretical approach complements the approaches
traditionally used in social science. The title of this chapter,
„Non-Monetary Cost-Benefit Analysis as a Basic Behavioral Pattern“,
was chosen to illustrate the presently existing terminological divergence
between the two approaches. Translated into social science terminology (social ex-change
concept), it would be Social Exchange as a Basic Behavioral Pattern.

Like Harsanyi and Homans (1961), we shall be mainly
concerned with the exchange object „esteem“ in a multitude of
„esteem-currencies“, and the perceptual recognition of the amount
accumulated in the form of social status, that is, „perceived
esteem“. Using Harsanyi’s game-theoretical bargaining model, we shall
attempt to widen its explanatory scope by assuming a status allocation as given,
and instead, we shall focus on status change and status maintenance. This leads
us to the new questions: what is the process affecting a change in social
status, and why is the status allocation stable on the „positive“ and
„negative“ margins of the multi-dimensional social evaluation space?

In searching for an answer to these questions, we
shall use a corollary to our assumption that social reality is a social
construction. We assume that the social reality of behavior is a social construction,
which is inter-personally determined through the exchange of esteem. Our goal,
the re-construction of the construction processes of the social reality, is
attained if we follow the construction of the social reality of behavior from
the set of possible constructions to the actual construction by the allocation
of esteem.

Our presentation method will explain the
„reasonableness“ of the theory of non-monetary cost-benefit analysis
as a basic behavioral pattern.

There will be no attempt to „prove“ the
theory since the currently available methodology of social science is perceived
as wanting in terms of its own goal: „reality“-approximation.
Therefore, we shall follow Harsanyi and Homans in their method of presentation
as well, preferring analytical elaboration to the traditional farce of

On our way, we shall meet the economic man who
maximizes the utility of his actions. We shall have to deal with the problem of
utility as it was developed from Bernoulli (1738) to Hakansson (1970) and with
the problem of interpersonal comparability of utility. Finally, using the
game-theoretical bargaining model will lead us to the realization that status
or „perceived esteem“ largely determines our behavior because it
provides the social bargaining position of an actor relative to the other
actors on the social scene. This does not mean that there is no interrelation
between status and behavior. Instead, the result we shall obtain reverses the
commonly accepted sequence of behavior leading to status and arrives at a
concept of elementary social behavior somewhat similar to the one presented by
the role theory.

That is as far as we shall go in this chapter. However,
the proposed approach, combining the game-theoretical bargaining model for
social status with the social exchange concept as presented by Homans, is
open-ended enough to leave all kinds of problems to be solved. One of these
problems is the subject of the fourth chapter, the understanding of
„crime“ and „criminal“.

Corresponding to the model presented in this
chapter, we shall assume that „criminality“ is a status assigned
after a process in which a behavior is evaluated, i. e., socially constructed
as a „crime“.

Turning now to the elaboration of our topic,
„Non-Monetary Cost-Benefit Analysis as a Basic Behavioral Pattern“,
we start with a synopsis of the social exchange concept as presented by Homans
and a summary of Harsanyi’s game-theoretical bargaining model for social
status. From Harsanyi’s and Homans‘ work, we shall develop a theory of the
dynamics of status change, our basic model of the „Games for Social

3.1.1. The Social Exchange

A central theme, which will accompany us throughout
this study, is the proposition that social behavior, analyzed as a process of
exchange, provides a conceptual framework for social science. Exchange is
conceived as a social process of central significance in social life, derived
from more simple processes and from which more complex processes are derived.

For a long time, exchange was considered almost
exclusively an economic process; in contemporary sociology, however, it is now
recognized that

exchange is in fact part of the social bond. Even
in the smallest and most intimate of social relationships – the relationship,
for example, of close friends or lovers – we find the exchange principle not
infrequently at work. This does not say that one person loves another merely to
receive the other person’s love. Such a characterization would be hedonistic
psychology carried to a ludicrous degree (Nisbet, 1970: 63).

Instead, it means that there are „times when
the act of one member is predicated upon his confidence that from that act
there will flow a reward – in the form of some desired expression of love,
gratitude, or simple recognition“ (Nisbet, 1970: 63).

The social exchange concept was formulated in the
books of Thibaut and Kelley (1959), Homans (1961), and Blau (1964). Thibaut and
Kelley (1959) draw heavily on the representation of social interaction as
matrixes of possible interactions and outcomes similar to the ones used in
basic game theory. As the first comprehensive formulation of the social
exchange concept, their discussion remains rudimentary, partly because of their
non-use of existing parallel conceptualizations in decision theory. The most
serious shortcomings of their study are the omission of the notion of
subjectively expected outcomes allowing the parties to randomize their
strategies, and the omission of a discussion on „social rationality“.

Thibaut and Kelley’s work was soon superseded by
Homans‘ analysis in Social Behavior: Its Elementary Forms (1961). Homans put
together a set of general propositions envisaging „social behavior as an
exchange of activity, tangible or intangible, and more or less rewarding or
costly, between at least two persons“ (Homans, 1961: 13).

Blau (1964) finally, to distinguish himself from
Homans‘ and Thibaut and Kelley’s work, postulates that „two conditions
must be met for behavior to lead to social exchange. (1) It must be oriented
toward ends that can be achieved through interaction with other persons, and
(2) it must seek to adapt means to further the achievement of these ends“
(Blau, 1964: 5). Blau contends that an extension of the notion of exchange to
all social behavior produces a tautology. The fallacy of his argumentation lies
in not differentiating between concept and theory (for this distinction, see, e.g.,
Conant, 1947: 102f.). Further, going more into the detail than both Thibaut and
Kelley, as well as Homans, Blau tends to run aground at the „Charybdis of
reductionism“ in his attempt to avoid „the Scylla of abstract
conceptions too remote from observable empirical reality“ (Blau, 1964: 3).

The problem of what „distance“ to choose
from the object of investigation is a general one in science. We are used to
classifying sciences according to their object of investigation and their
distance from it, e.g., there is molecular biology, physiology, and medicine —
or psychology, social psychology, and sociology. Social psychology seems especially
prone to fall into extremes. Indicative is how it is usually taught, either in
the psychology or the sociology department. Compared to the
„ram-ta-ta“ method of experimental social psychology (problem
formulation, research design, and procedures, discussion), Homans‘ choice of
distance to the object of investigation (human behavior) appears to be optimal:
it enables him to formulate propositions about elementary social behavior which
have a certain analytical if not predictive value for the further analysis of
social processes. It is Homans‘ distance from the object of investigation,
which we chose for our study, and, of course, Homans‘ formulation of the social
exchange concept.

Homans‘ methodological strategy in Social Behavior: Its E1ementary Forms is the strategy of explanation. From a set
of propositions, from the general, he builds back to the empirical. He
„tries to explain . . . why empirical propositions about elementary social
behavior, both those cited in ‚The Human Group‘ and others, should take the
form they do“ (Homans, 1961: 10). Homans ‚ set of propositions and
corollaries gives an initial comprehensive presentation of the concept of
social exchange and its elaboration in the form it will be used in our study.

Proposition (1): If in the past the occurrence of a
particular stimulus-situation has been the occasion on which a man’s activity
has been rewarded, then the more similar the present stimulus-situation is to
the past one, the more likely he is to emit the activity, or some similar
activity, now (Homans, 1961: 53).

Proposition (2): The more often a man’s activity
rewards the activity of another within a given period of time, the more often
the other will emit the activity (Ibid.: 54).

Proposition (3): The more valuable to a man a unit
of the activity another gives him, the more often he will emit activity
rewarded by the activity of the other (Ibid.: 55).

Corollary to (3): The more cost Person incurs in
emitting an activity, the less often he will emit it. Cost defined: the cost of
a unit of a given activity is the value of the reward obtainable through a unit
of an alternative activity, forgone in emitting the given one (Ibid.: 60).

Corollary to (2) and (3): The frequency of interaction between Person and Other
depends on the frequency with which each rewards the activity of the other and
on the value to each of the activities he receives (Ibid: 55).

Proposition (4): The more often a man has in the
recent past received a rewarding activity from another, the less valuable any
further unit of that activity becomes to him (Ibid. : 55).

Proposition (4) (a): The more valuable to Person is
the activity Other gives him, the more valuable is the approval or liking
Person gives Other and the more often
he emits activity, including sentiment, to Other (Ibid. : 182).

Proposition (4) (b): The larger is the number of
members in the group that like other members, the more frequent is the overall
interaction between them (Ibid: 188).

Proposition (4) (c): The larger the amount of social
approval received by a single member from other members (that is, the higher
his esteem), the more frequent the interaction he receives from other members (Ibid:

Corollary to (4): The more often Person has emitted
a costly activity, the more costly he finds any additional unit of that
activity (Ibid.: 60).

Corollary to (3) and (4): The more help Person has
recently received from Other, the less, for the time being, he needs any
further help and the less often he will ask for help or give thanks (Ibid. : 55).

Proposition (5): The more to a man’s disadvantage
the rule of distributive justice fails of realization, the more likely he is to
display the emotional behavior we call anger (Ibid. : 75). Distributive justice
explained: The more often in the past an activity emitted under particular
stimulus-conditions has been rewarded, the more anger they will display at
present when the same activity, emitted under similar conditions, goes without
its reward: precedents are always turning into rights (Ibid. : 73).

A central position in Homans‘ discussion of the
social exchange concept occupies what he calls „esteem“ and
„status“. He defines them as follows:

The greater the total reward in expressed social
approval a man receives from other members of his group, the higher is the
esteem in which they hold him. But this does not mean that we propose to
abandon status. Social approval is an actual reward, but any activity (or
sentiment) may be stimulus as well as reward, and we shall use status to refer
to the stimuli a man presents to other men (and to himself) (Ibid. : 149).

Analyzing interaction, Homans observes that there
are consistently in any group „members who can emit rare and highly
rewarding services, and who command in return much approval from many
others“ (Ibid.: 145). Highly rewarding services characteristically are services
perceived to be in short supply, where short supply is a relative matter, i. e.,
it is a question of the perceived relation between supply and demand.

Homans substantiates his observation by reporting a
classic study of Jennings
(1950: 46) with the results of a sociometric test of 133 girls responding to
the question of with whom they liked living and working together. Although
Jennings did not get the responses according to the rank-order given (this
would have had the effect of increasing the differences in approval, because
girls who got many choices also got many first choices [Jennings, 1950: 887]),
the shape of the plotted whole distribution is not what the statistician calls
a „normal“ distribution, forming a symmetrical bell-shaped curve.
(See Figure 1, next page) Instead, it tapers toward the top and bulges toward
the bottom like a beet or radish. This means that choice tended to focus on a few
very popular girls, but that indifference (or failure to choose) did not focus
nearly so sharply: more girls were chosen little than were chosen much. Similar
distributions have been found in other sociometric studies.

The result of studies on the distribution of
sociometric choices, therefore, are consistent with Homans‘ less structured
observation that there are always in any group „members who can emit rare
and highly rewarding services, and who command in return much approval from
many others“ (Homans, 1961: 145). Unfortunately, however, perhaps due to
his „positive“ orientation toward „esteem“ and
„status“, Homans omits an equally thorough discussion of the
corresponding distribution of sociometric rejections given by Jennings.

data on the distribution of sociometric rejection, plotted in the same way as
the data on the distribution of sociometric choices does not produce the same
kind of curve. Similar to the distribution of positive choices, only a few
girls were each rejected by a large number of others. Then the curve changes.
The rejection curve does not bulge in the low middle like the curve mapping the
choices. Instead, the numbers increase fairly steadily and are the greatest at
the bottom, with the greatest number of girls being those rejected by no one or
by only one other girl. Thus, while relatively few girls found no one to choose
them, a very large number found no one to reject them.

Homans‘ principal interpretation of differences
between the two distributions is:

It is clear that rejection was not the same thing
as low positive choice. If it had been, many girls would have received many
rejections, whereas only a few did. To be positively rejected, a girl must have
conspicuously violated whatever norms were current in the school, and in her
peculiar way we must have been as rare a creature as a very popular girl
(Homans, 1961: 158).

Here we encounter a weakness in Homans‘ discussion
of elementary social behavior for the first time. His treatment of low status
tends to be comparatively superficial. His analysis tends to use commonly
accepted clichés about low status.

Being predominantly concerned with the attainment of
status, with the dynamics of the acquaintance process, Homans neglects the fact
that generally, we are already assigned a status in social life. In the social exchange
concept terminology, this initial status is the bargaining position based on
which a new status is attained, or the old one is retained. While mentioning
social bargaining, Homans never discusses it extensively, overlooking, like
Thibaut and Kelley (1959) and Blau (1964), the development of modern decision
theory available when he wrote his book.

We intend to pick up the discussion of the social
exchange concept where Homans left off. The bargaining position in social
exchange, i. e., the social status at the beginning of an exchange, will be of
central importance in this discussion.

Distribution of Sociometric Choices

3.1.2. The Game-Theoretical
Bargaining Model for Social Status

To be equipped with the „tools“ for our
analysis of the dynamics of attaining and retaining social status, our next
step is the presentation of a synopsis of modern decision theory, especially of
the game-theoretical bargaining model developed by Harsanyi. After a short
history of the decision theory, we shall introduce the bargaining problem and
then discuss the notion of utility and the rationality postulate underlying
both normative and descriptive decision theory. Finally, we shall present Harsanyi’s
game-theoretical bargaining model for social status.

When introducing a game-theoretical bargaining model,
we have to start by clarifying the objectives of modern game theory and its
possible implications for social science.

Game Theory and Social

Game Theory deals with game rules per se and rules
for playing games intelligently. It has been rules of the latter sort which
have claimed by far the larger share of the game theorists‘ attention — and
quite properly so, given the motivation that guides most game-theoretical

The research in modern decision theory was
essentially initiated by von Neumann and Morgenstern’s famous book, Theory of Games and Economic Behavior
(1944), where they laid the foundation for the new science, game theory. In
some ways, the name „game theory“ is unfortunate, for it suggests
that the theory deals only with the socially unimportant conflict found in
parlor games, whereas it is far more general than that. It is the study of c o n f l i c t of interest, both among
individuals and institutions.

In the thirty years since von Neumann and Morgenstern’s
book, the game theory has been developed into a complex construct by scientists,
especially in mathematical economics and finance theory. While von Neumann and
Morgenstern „attempted to write their book so that a patient scientist
with limited mathematical training could absorb the motivation, the reasoning,
and the conclusions of the theory“ (Luce and Raiffa, 1957: 3), game theory
today has developed into an area largely accessible only to the specialist, the
mathematician, and econometrician. Social scientists had an initial „naive
band-wagon feeling that game theory solved innumerable problems of sociology
and economics, or that, at the least, it made solutions a practical matter of a
few years work“ (Luce and Raiffa, 1957: 10), have generally become
disillusioned with game theory, as it turned out that their expectations were
not fulfilled. Consequently, the social science literature in game theory is

However, we must keep in mind that game theory,
presented today, is predominantly a (conditionally) normative theory. Mostly it
prescribes, rather than describes,

for given assumptions courses of action for the
attainment of outcome having certain formal ‚optimum‘ properties. These
properties may or may not de deemed pertinent in any given real world conflict
of interest. If they are, the theory prescribes the choices, which must
be made to get to that optimum (Luce and Raiffa, 1957: 63).

In social science, we are predominantly interested
in a descriptive decision theory, that is, in system-understanding instead of
system-planning. Therefore, a direct „application“ of game theory,
which the social scientists mentioned above had in mind, is out of the
question. Instead, the use of game-theoretical concepts has to focus on the
descriptive elements of game theory.

Hence, the predominantly descriptive
game-theoretical bargaining model developed in the last fifteen years,
especially by Harsanyi, is of particular interest for the social scientist. The
general purpose of the bargaining model is to predict the outcome of a
bargaining process for given initial bargaining positions.

The bargaining model formalizes the processes, which
we usually refer to as „bargaining“ in our everyday life. Often we
speak of bargaining when an economic and monetary exchange is involved.
However, that is not the only way we use the term. We also speak of bargaining
in our everyday lives if we describe social exchange. For instance, we use the
term „bargaining“ to describe the social process between parents and
children attempting to postpone bedtime, just as we use it when we describe the
process of the exchange of goods in the marketplace.

In the explanation of the social exchange concept by
Thibaut and Kelley (1959), Homans (1961), and Blau (1964),
„bargaining“ was used to describe the process of social exchange.
Unfortunately, these authors either misunderstood (see Thibaut and Kelley,
1959: 24f.) or did not know about the parallel conceptualizations of exchange
processes developed in game theory. As we later show in this chapter, using
these conceptualizations as a vehicle for presenting the social exchange
concept will lead to a more concise and, at the same time, more comprehensive
understanding of social exchange, social behavior, and social status.

The Bargaining Problem

From basic economic theory, we know that the following
two properties define the range of practicable bargains:

<![if !supportLists]>1)
<![endif]>It is on
the Edgeworth contract curve, i. e., the locus of situations, where neither party’s
position can be improved without at the same time worsening the position of the
other party.

<![if !supportLists]>2)
<![endif]>It is on
this curve between two limits, which may be called the „maximum-concession
points“ of each party and which are determined by the fact that neither
party would accept an agreement that puts him in a worse position than the
conflict situation (not reaching an agreement). In other words, each party’s
maximum-concession point is at that point at which all the net gain resulting
from an agreement and the cooperation agreed upon would go wholly to the other

Then, the game-theoretical bargaining model explains
how the position of the actual agreement is determined within this range of practicable bargains; that is, how the
division of the net gains is determined between the two bargaining parties.

A graphical representation of the bargaining problem
looks as follows:

The Bargaining Problem 


The Notion of Utility

In order to determine the actual agreement point, the
game theory uses standard utility as measurement and makes assumptions about
the rationality of the bargaining parties. The utility theory (see, e.g.,
Savage, 1954: 69ff; Luce and Raiffa, 1957: 12ff; Anscombe and Aumann, 1963)
employs axioms for rational behavior to establish a non-tautological notion of
utility as a measurement standard.

The concept of „anticipation“ is essential
in this theory. For example, if A and B are the only two anticipations, then
there is an anticipation which we can represent by pA + (1 – p) B where 0 ≤p
≤1, which is a probability combination of the two anticipations where
there is a probability p of A and (1 – p) for B.

By making the following assumptions, we can develop
the utility theory for an individual (from Nash, 1950):

<![if !supportLists]>(1)    <![endif]>An individual offered two possible
anticipations can decide which is preferable or that they are equally

<![if !supportLists]>(2)    <![endif]>The ordering thus produced is transitive; if A
is better than B and B is better than C, then A is better than C.

<![if !supportLists]>(3)    <![endif]>Any probability combination of equally
desirable states is just as desirable as either.

<![if !supportLists]>(4)    <![endif]>If A, B, and C are as in assumption (2), then
there is a probability combination of A and C, which is just as desirable as B.
This amounts to an assumption of continuity.

<![if !supportLists]>(5)    <![endif]>If 0 ≤p ≤1 and A and B are equally
desirable, then pA + (1 – p) C and pB + (1 – p) C are equally desirable. Also,
if A and B are equally desirable, A may be substituted for B in any
desirability ordering relationship satisfied by B.

These assumptions are
sufficient to show the existence of a satisfactory utility function, assigning
an actual number to each anticipation of an individual. This utility function
is not unique; that is, if u is such a function, then au + b is also such a
function, provided a > 0. Letting capital letters represent anticipations
and small ones real numbers, such a utility function will satisfy the following

<![if !supportLists]>a)       <![endif]>u (A) > u (B) is equivalent to A is more
desirable than B, etc.

<![if !supportLists]>b)       <![endif]>If 0 ≤ p ≤ 1, then u [pA + (1 –
p)B] =

pu (A) + (1 – p) u (B).
This is the important linearity property of a utility function.

Of course, the question with which we now have to
deal is whether we can accept the rationality postulates used to establish the
utility theory in describing social behavior. Is man really a utility

The Rationality Postulate

Rationality itself is a concept that encompasses the
assumptions of the social exchange concept. However, what is rationality?

If the term „rationality“ is mentioned in
social science circles, one can almost see the hair bristling at the neck of
the social scientists, and almost inevitably, an emotionally laden discussion
starts. Therefore, we shall attempt to clarify some common misunderstandings of
the concept of rationality. We shall discuss „rationality“ in terms
of its applicability as a set of assumptions in the context of social science.

The subject of our investigation is the exchange
object „esteem“ in a multitude of „esteem-currencies“ as
presented in the exchange concept, and this concept stands and falls with the
acceptance of the assumption that man is basically rational. Therefore, the
question is whether this assumption can „reasonably“ be made.

Economic theory is built on the assumption that man
can basically be regarded as rational. Many of the misconceptions held by
social scientists are based on a misunderstanding of the economic man. The
economic man indeed and necessarily is a generalized, conceptual man (Idealtyp). However, he is also a complex

Social scientists frequently remember only the
pithecanthropus-stage of the rational man, the monster we all have met in our introductory
economics classes who is willing to sell his mother in slices as long as the
marginal return per slice is higher than the marginal cost of slicing. That is,
unfortunately, the first and last impression many social scientists have of the
economic man.

However, the new economic man, the homo sapiens of
the rational man, looks somewhat different. The creature sired by modern
economic theory, although by definition an Idealtyp
in Max Weber’s sense, has become considerably more complex and sophisticated. Bernoulli
(1738) considered a man maximizing the utility of money, whereas Friedman and Savage’s
(1948) individual maximized the utility of income. Markowitz (1952), Hakansson
(1970), and other modern theoreticians finally talked about a man maximizing
the utility of wealth, where wealth is the probability estimate for the
attainment or retainment of tangible and intangible goods.

The access to information and the computational
capacities of this new economic man is limited (Simon, 1957). Therefore, the
new economic man is a man with „approximate“ rationality. Empirical
studies on utility maximization in a social context (cf., Ofshe and Ofshe,
1970; Rapoport and Chammah, 1965) confirm the existence of this rational man as
it was developed by modern economic theory and decision theory.

The trouble with the „old“ economic man

was not that he was economic, that he used his
resources to some advantage, but that he was antisocial and materialistic,
interested only in money and material goods . . . What was wrong with him were
his values: he was only allowed a limited range of values; but the new economic
man is not so limited. He may have any values whatever, from altruism to hedonism, but so long
as he does not utterly squander his resources in achieving these values, his
behavior is still economic. Indeed, if he has learned to find reward in not
husbanding his resources, if he values not taking any thought for the morrow,
and acts accordingly, his behavior is still economic. In fact, the new economic
man is a plain man (Homans, 1961: 79f.).

Therefore, the assumption underlying the use of the
game-theoretical bargaining model for the analysis of human behavior from a
decision-making perspective (namely, that although decisions vary in content,
importance, and social context, the abstract principles which guide behavior in
these choice situations are basically the same) was verified and is verifiable
through empirical investigations. The economic situation for which the
bargaining model was developed is just one type of social situation, and the
assumption of rationality is valid there as it is in non-economic social

Given the concept of social exchange, social
psychology and sociology differ from economics not in their perspective but
only in their history. The title of a „dismal science“ for economics
(Berger, 1963: 93) is just as appropriate, or inappropriate, for social
psychology and sociology. In all social interactions, an individual is forced to
choose a particular set of alternatives, and it is assumed that he makes his
decision in a manner that maximizes his expected utility.

The Nash-Solution

An excellent description of how the complex economic
man maximizes his expected utility, that is, of how a bargain is struck,
assuming rationality of the bargaining parties, is given by Schelling:

In bargaining, each party is guided mainly by his
expectations of what the other will accept. But with each guided by
expectations and knowing that the other is too, expectations become compounded.
A bargain is struck when somebody makes a final, sufficient concession. Why
does he concede? Because he thinks, the other will not. ‘I will concede because
he won’t. He won’t because he thinks I will. He thinks I will because he thinks
I think he thinks so . . . ‚ (Schelling, 1963: 21f.).

Given that this description of the bargaining
process is adequate, we can say that the bargaining problem has an obvious,
determinable solution for fixed conflict points, in at least the case where the
bargaining situations for both parties are entirely symmetric. In this case, it
is natural to assume that the two parties will tend to equally share the net
gain since neither would be prepared to grant the other better terms than the
other would grant him.

Nash (1950) generalized this principle for the case
of asymmetric bargaining situations by postulating that the general solution of
the bargaining problem is at that point on the Edgeworth contract curve at
which the product of the utilities, which each party ascribes to his net gain
over the conflict situation, is maximized (Nash-solution). Nash assumed (in
addition to the usual assumptions of utility theory) that there is a set of
points R representing all possible bargains, where R is bounded, convex and
closed. If there is within the set R a point (uc, vc)
directing the conflict pay-off, we know that the players enter a bargain so
long as there are any points (u, v) in R both above and to the right of (uc,
vc). The Nash solution to the bargaining game is the unique point (uo,
vo) on the Edgeworth contract curve such that

(uo -uc) (vo -vc)
≥ (u-uc) (v – vc)

for all (u, v) belonging to R and such that u ≥uc
and v ≥ vc.

It is worth noting that this solution, where both
parties‘ utilities are measured by von Neumann and Morgenstern cardinal utility
indexes, does not presuppose the interpersonal comparability of utilities,
since the index for each party’s utility can be chosen independently of the other’s,
because the solution is invariant to changes to the zero point or the unit of
either party’s utility index. In effect, this invariance is one of the
postulates used by Nash in deriving his solution.

A Numerical Example for the

Luce and Raiffa (1957: 129) give us a simple
numerical example for the Nash solution:

Consider the game where two players are to divide $
100 between them if they can reach an agreement, or where they will receive
nothing if they do not agree . . . Suppose (player) 1 is rich and selfish and 2
is poor. Player 1 may then make a good case for (75, 25), let us say, on the
grounds that the utility increment to 2 of $ 25 is at least as great as the
gain to 1 of $ 75. We would argue that an ethically fair division of the
proceeds is a fair division of utility, not of money. For example, suppose that
(part of) the utility functions were these:


Harsanyi’s Solution

The Nash-solution of the bargaining problem was
primarily designed as a prescriptive solution. However, as stated before,
social science is essentially interested in description rather than
prescription. Luce and Raiffa’s example of the Nash-solution, which we cited
above already, provided us with the idea that at the basis of Nash’s
prescriptive model lays a descriptive analysis of social situations. We shall
now present another solution of the bargaining problem, which corresponds in
the outcome to the Nash-solution, but which arrives at the solution through
descriptive analysis.

This way of solving the bargaining problem was
proposed by Harsanyi (1956, 1961), who based his approach on an earlier
analysis by Zeuthen (1930). Harsanyi’s approach to the bargaining problem is to
determine, for a given stage of the bargaining process, which of the players is
likely to (in a normative formulation: should) make the next concession.
Basically, therefore, Harsanyi’s approach corresponds to the way we analyze a
situation in social science and, as its result is the solution of the
bargaining problem, which is mathematically equivalent to the Nash-solution, it
illustrates our thesis that the social exchange concept can be extended through
the use of parallel conceptualizations in game theory.

Harsanyi shows that „the Nash-solution can also
be obtained from weaker and, I believe, more fundamental postulates if one uses
the Zeuthen model of the bargaining process“ (Harsanyi, 1961: 183; cf.
also, 1956). He uses the following „strong“ postulates of rationality
(1961: 183f.):

(1) Individual utility maximization; (2) efficiency;
(3) acceptance of higher payoffs; (4) symmetry; (5) restriction of variables;
and (6) mutually expected rationality.

He then assumes that each player knows his opponent’s
subjective probability of conceding, so it becomes meaningful to consider the
(expected) utility of conceding or not conceding. In a descriptive bargaining
theory, the knowledge assumed in Harsanyi’s model has to be weakened, of
course, to a reasonably accurate estimate where accuracy is partly attained
through self-fulfilling prophecies.

Given the same general bargaining situation we
described in the exposition of the bargaining problem and Nash’s approach,
Harsanyi starts his analysis of the bargaining problem at a stage of the
bargaining process where player 1’s offer is A1 with the coordinates
(u11, v11) and player 2’s last
offer is A2 with the coordinates (u21, v21)
and where A1 is different from A2, and both points are
Pareto optimal. The conflict point is, as before, C with the coordinates (uc,
vc). The offer A1 then would yield the utility u1 or
U1 (A1) to player 1 and the utility v11
or U2 (A1) to player 2. Correspondingly, the utilities of
A2 are U1 (A2), and the utilities of C are U1(C)
and U2(C).

Harsanyi’s resulting concession rule is that player
1 „should“ concede if and only if


Moreover, player 2 „should“ make it if the
inequality is reversed. In case both expressions are equal to each other and A1
≠ A2 both players „should“ yield.

Making concessions continues until the common point
A on the Edgeworth contract curve is reached where the product of [U1 (A)
– U1(C)] and [U2 (A) – U2 (C)] takes its
maximum value. This point is unique, and it is the Nash-solution of the
bargaining game.

We can restate Harsanyi’s concession rule more
tersely if we use our assumption made to establish the utility theory, namely
that the utility function is not unique; that is, if u is such a function, then
au + b is such a function provided a › O. Then we can define the point C with
the coordinates (uc, vc) as (0, 0). The concession rule
for player 1 changes to


which is obviously equivalent to


obviously, measure the relative losses incurred when
players 1 and 2 conceded. Harsanyi’s concession rule then predicts that the player whose relative
loss is the smaller will concede.

It is the predictive value of Harsanyi’s
game-theoretical bargaining model, which makes it possible to use it to explain
elementary social behavior if we accept the concept of social exchange. Harsanyi’s
bargaining model bridges the gap which seems to have existed between the conceptualizations
of exchange in economics and in social science. Thereby it opens avenues of
thinking that may prove generative not only for new ideas in social science but
also, in a reciprocal process, for innovative approaches to economics and the
economic „W e l t b i l d “.

Harsanyi’s Model of Status

Harsanyi realized the implications for the analysis
and the explanation of elementary social behavior if the game-theoretical
bargaining model is combined with the concept of social exchange. In his
papers, „A Bargaining Model for Social Status in Informal Groups and
Formal Organizations“ (Harsanyi, 1966) and „Individualistic and
Functionalistic Explanations in the Light of Game Theory: The Example of Social
Status“ (Harsanyi, 1968), Harsanyi analyzes social status in terms of some
more basic human motivations by asking the questions why all people seek high
social status and why some people are granted high social status by others. Understanding
Harsanyi’s concise analysis will significantly help our own analysis of the
closely related problem concerning the dynamics of status change and status

Harsanyi starts by proposing a simplifying
assumption that a given individual’s services (and/or disservices) have the
same importance to all group members so that the analysis can be based on a
two-person bargaining model. Therefore, one bargaining unit consists of the
individual A and the other bargaining unit of the other (n – 1) members of a
social grouping.

From bargaining theory, we know that the outcome of
the bargaining process depends on players‘ bargaining position. The bargaining
position, in turn, „will depend on the costs of a possible conflict to
each side. Each side will be in a stronger position the smaller its own
conflict costs would be and the greater those of the other side“
(Harsanyi, 1966: 363).

As the net conflict costs of either side, Harsanyi
specifies the following items:

<![if !supportLists]>(1)    <![endif]>loss of rewards usually provided by the other

<![if !supportLists]>(2)    <![endif]>damages resulting from the penalties imposed by
the other side;

<![if !supportLists]>(3)    <![endif]>the costs incurred by actively imposing
penalties on the other side; less

<![if !supportLists]>(4)    <![endif]>the costs of the rewards usually provided for
the other side (in case of a conflict, these rewards would be withheld so their
costs would be saved).

Restating these relationships in mathematical
symbols, Harsanyi denotes the rewards received and the penalties or costs
incurred by individual A by the small letters r, p, and c, and the rewards
received and the penalties and costs incurred by each of the other (n – 1)
members of the group by the

capital letters R, P, and C.

Individual A’s net conflict costs are then:


If, in the computation of these conflict costs, we
take into account the alternative arrangements available to each side in a
conflict situation (individual A might be able to join other groups, and his
losses would consist only of those payoffs that these other groups could not
provide for him; or, the group might be able to replace him by other
individuals, and the group’s losses would consist only of those services that
these other individuals could not provide for the group), „the strength of
individual A’s bargaining position vis-a-vis the other members of the group
will depend on the quantities c* and C* of the equations (1) and (2). The
social status we can achieve within the group will be higher the smaller the
quantity c* and the greater the quantity C*“ (Harsanyi, 1966: 364).

Harsanyi’s bargaining model (see also, Harsanyi,
1962: 77) then predicts that the amount of esteem e that individual A will
receive within the group will be given by the expression:


where:    s             
denotes the utility that individual A would assign to the very highest social
status position in the group, and

                C(s)        – denotes the disutility the other group members would incur, on
average, by granting him such a high social position.

Two principal comments on Harsanyi’s
game-theoretical bargaining model for social status seem to be appropriate. The
first one concerns the scarcity and relativity of status positions, and the
other is Harsanyi’s method to determine the social bargaining position.

We know that rank differentiation in social systems
is a universal phenomenon. Further, we know that social status in any social
system is determined by the relative position of the individual in the system
(Svalastoga, 1964). Generally speaking, a change in the social status of an
individual, therefore, means not only a change in the social status of this
individual but in the case of an upward movement, also a downward change of
status of others (and their allies) who previously had a higher status. According
to the interpretation of social status as perceived esteem, these others are
part of the body providing the esteem. Assuming an equilibrium situation of the
social system under observation, any status change threatens their status
position and the general equilibrium of the system. From Harsanyi’s equation
(3), however, it is by no means evident that the value of e in a static model,
i. e., a model dealing only with one interaction changes, is bounded from above
and below by the present social status.

The second comment is based on the same idea.
Whenever we enter a social interaction, or, in our terminology, a social
bargaining situation, we enter it with a status already assigned. If this
status is not apparent to the partner in the interaction, the initial part of
the encounter will consist in its determination. However, most of our life is
spent in situations where our relative status is known to the partner of the
interaction. Being like Homans predominantly concerned with attaining status,
Harsanyi presents a more helpful model for mapping the dynamics of the
acquaintance process (Newcomb, 1961) than the dynamics in an ongoing group.

While we do not doubt that the variable c* in
Harsanyi’s equation (1), defining the actual bargaining position of individual
A through his net conflict costs, is adequately described through r, p, and c,
it appears that we have more information available about the value of the
variables. We know that these variables are all a function of the social status
(perceived esteem) before the bargaining process begins. Our problem
formulation, therefore, differs slightly from that of Harsanyi. Harsanyi is
asking why some people are granted high social status. We shall attempt to
develop this problem formulation to: What is the process affecting a change in social status? Harsanyi’s
game-theoretical bargaining model will provide us with the vehicle for analyzing
elementary social behavior, picking up the discussion of where Homans left off.

3.2. Games for Social

Homans‘ propositions about the concept of social
exchange have provided us with the building blocks for a theory of elementary
social behavior. Harsanyi’s game-theoretical bargaining model gives us the
instructions about how we should fit them together. In the previous section, we
laid out the building blocks and the instructions, and now we are ready to

We saw that both Homans and Harsanyi were
predominantly concerned with processes to attain status, with the dynamics of
the acquaintance process. In discussing their models, we pointed out that
generally, we are already assigned a status in social life. With this initial
„capital”, a person enters an interaction. Therefore, in our analysis of
elementary social behavior, we will consider the status quo ante, as we call it,
to determine the bargaining position for each encounter. In this respect, our
analysis will attempt to supplement the work of Harsanyi and Homans.

In another respect, we will try to extend their work
to attain a more comprehensive theory of elementary social behavior. Both
authors primarily directed their research to the investigation of high status.
It appears, however, that only a theory, which also incorporates low status, its
attainment, and retainment, can claim to be a theory of elementary social
behavior. Therefore, the extension of Harsanyi’s and Homans‘ work to low status
is the other area where we shall attempt to supplement their work.

This section will try to explain the
„reasonableness“ of the thesis: If a basic behavioral pattern is the
non-monetary cost-benefit analysis, then the game-theoretical bargaining model
using the social exchange concept predicts that social status is a
„sticky“ state. It further predicts that social status is the major
explanatory factor for social behavior, i. e., in the interrelation of social
status and social behavior, the socially constructed reality, social status,
limits the range of possible constructions of social behavior.

The „result“ of this study, therefore,
does not provide any revolutionary new knowledge on the ordering of the
„familiar chaos: ordinary, everyday social behavior“ (Homans, 1961:
1). Indeed, our results conform to the claims made by role theory, which states
that social behavior conforms to the social position of the actor. What is new
is the method by which we attain this result. Suppose we accept the legitimacy
of the proposed method (the combination of Harsanyi’s game-theoretical model
with Homans‘ propositions to the social exchange concept) to analyze social
contexts in general. In that case, it leads us to a novel understanding of
social systems, which makes us doubt the foundations of some cornerstones of
our present social system.

One example of this is our notion of justice. If we
analyze the criminal status as obtained through a succession of bargaining
games with agents of the Criminal Justice System whose basic behavioral pattern
is non-monetary cost-benefit analysis, the analysis is bound to make us
question whether the commonly accepted reasons for the existence of the
Criminal Justice System are not mere rationalizations, covering the fact that
immunity is differentially distributed.

Another example is our notion of the educational
system. In fact, we shall start the explanation of our thesis with a more
anecdotal than analytical quote concerning an educational game, the graduate
school game, as seem from the perspective of transactional analysis. After this
introduction, we shall analyze social interaction as a bargaining game in
general, discussing the social correspondent of bargaining position, feasible
set, and utility.

3.2.1. Bargaining for

Joen Fagan (1967: 103) describes the graduate school
game as follows:

Graduate school as a sub-section of the educational
enterprise generally fosters Parent-Adolescent interaction on the part of professors
and graduate students. Most graduate students live on a relatively meager
stipend (allowance) provided by the department or a professor’s research
project and in return help keep the house running by providing various low
level services. Any involvement outside the department that would make for
greater financial or personal independence is discouraged. Students are
expected to act like responsible adults yet still defer to the professor’s
superior judgment and wisdom. The professor overtly communicates to the
student, ‚Grow up and get wise,‘ but this may hide messages that say, ‚ But if
you end up different from me, you’re wrong, ‚ or ‚ But if you do, I’ll zap you
for showing me up, ‚ or ‚ But if you do, I’ll make you feel guilty for leaving
me behind. ‚ As a result, graduate students often feel perplexed about how
smart to be, what to do if they either do or don’t know more than the
professor, whether they are supposed to do it their way or his, whether they are
actually supposed to finish their degrees and leave. This atmosphere powerfully
supports games involving hanging on and phobic avoidance of the outside world.

The graduate school game is a particular case of the
general bargaining game of social interaction. The professor, a high-status
person, interacts with a low-status person, the graduate student. The initial
bargaining position of the professor is so much stronger than the position of
the student that for nearly any offer A p 1 (last offer
of professor) better than the conflict solution, and for any offer A s
(last offer of the student), where Up (A p 1)
› UP (A S 1) and Us (A S
) > Us (A P 1), Harsanyi’s
concession rule predicts that

(A p 1) x Us (A P 1) › UP
(A S 1) x Us (A S 1)

which means the student makes the next concession As2.

This result follows Harsanyi’s formula of net
conflict costs for the two individuals. The initial bargaining position of the
student, given by

C*S = Rs + Ps + Cs
(Pp) – Cs (Rp), is obviously weaker than
the bargaining position of the professor, given by

C*P = RP + PP + CP
(PS) – CP (RS), because we can assume

RS > Rp, PS >
PP, CS (PP) > CP (Ps)

CS (RP) < CP (RS),
especially as the number of possible bargaining partners in the graduate school
game is smaller for the student than for the professor.

Let us now leave the professor and his graduate
student and present the first part of our model of „Games for Social
Status“, where we map interaction as bargaining for esteem. Perceived
esteem, as we defined it following Homans, is social status. It has the
property to diminish over time if it is not constantly reconfirmed. According
to Homans‘ exposition of the exchange concept, we acquire esteem in social
interactions as an exchange object in various „esteem-currencies“.

If we postulate approximate rationality in social
interaction (i. e., if we postulate that more is always preferred to less,
where individual preference determines what is more and what is less, and where
individual preferences can be generalized to some extent), we can represent the
social exchange of esteem as a bargaining game in which the social reality of
behavior is constructed from the feasible set of constructions. To keep the
representation simple, we collapse the n-person games on the social scene into
two-party bargaining games.

Let us then introduce three categories of players:
H, belonging to a high-status category (a category of players who enter a game
with a large „capital“ of perceived esteem), M, belonging to a medium
status category, and L, belonging to a low-status category. Let us for now
assume that the players of all three categories are equal in bargaining skill,
i. e., any bluff about the initial bargaining positions will be discovered, and
we, therefore, can assume the bargaining positions as given and known to the
players. Finally, let us assume the following simple bargaining situation,
where the conflict point with the coordinates (uc, vc) is
„normalized“ at (0, 0) and the line bounds the feasible set through
the „normalized“ points (1, 0) and (0, 1) on the utility scales of
the players. What, then, are the outcomes predicted by bargaining theory for
all possible games between players H, M, and L?

We can represent these types of games by the
following figure:

The Normalized Bargaining Game

From our discussion of the bargaining problem, we
know that the solution lies within the feasible set. Further, we know that if
the bargaining game were played perfectly rationally, the solution would lie on
the Edgeworth contract curve, which connects in this particular type of game
the points A (0, 1) and C (1, 0). For social games, where we assumed
„approximate“ rationality, the predicted outcome then will lie within
the feasible set approximately on the Edgeworth curve. Finally, we know that the
predicted outcome of the bargaining game for perfectly rationally played games
is the unique point (uo, vo) on the Edgeworth contract
curve such that

(uo – uc) (vo – vc)                                                                    (u-uc) (v-vc)

for all (u, v) belonging to the feasible set R and
such that u ≥ uc and v ≥ vc (Nash-solution). Again,
as we assume approximate rationality for social games, the predicted outcome
will lie approximately at the point (uo, vo). To predict
the outcomes for all possible games between players H, M, and L, we, therefore,
have to know their utility function for esteem.

From Homans‘ propositions (see pp. 51 f. above),
especially propositions three, four, and its corollaries, we can determine the
general shape of the utility curve for units of esteem to be exchanged. It
corresponds to the Markowitz hypothesis of the utility of wealth (Markowitz,
1952; see, however, Hakansson, 1970), which is described as follows:

The utility function has three inflection points.
The middle inflection point is defined to be at the ‚customary‘ level of
wealth. Except in cases of recent windfall gains and losses, customary wealth
equals present wealth. The first inflection point is below, the third
inflection point is above, customary wealth. The distance between the
inflection points is a nondecreasing function of wealth. The curve is
monotonically increasing but bounded; it is first concave, then convex, then concave,
and finally convex. We may also assume that
lU (-X) l > U(X), X     0 (where X = 0
is customary wealth) (Markowitz, 1952: 155).

The Markowitz Utility Function 

If our assumption about the general shape of the
utility function for esteem approximates adequately the empirical reality, that
is, if Homans‘ propositions can be accepted as a mapping of reality, we can now
predict the outcome of a simple bargaining game for esteem between the players
H, M and L. We distinguish two types of encounters. In the first type, two
players with approximately the same social status quo ante play a game for
esteem currencies. These are the games where player H1 plays with H2,
M1 with M2, and L1 with L2. In the
second type of game, players with different social status quo ante encounter
each other. These are the games where H meets M or L and M meets L.

As an illustration, we again use the figures from
the example given by Luce and Raiffa (1957: 129):

Luce and Raiffa Utility

If the same status group members meet to play a
bargaining game, we can expect their utility function for increments of esteem
to be approximately comparable. Taking Luce and Raiffa’s example, this might
mean that there exists the following symmetrical bargaining situation:

Bargaining Situation

Alternatively, the following situation might exist:

Bargaining Situation

Our model predicts that in both cases, the outcome
of the bargaining game will lie at or approximately near the halfway point B,
which has, in our normalized bargaining game (figure 3, p. 71 above), the
coordinates (1/2, 1/2). The predicted outcome corresponds to what we are used
to observing in everyday life, namely, parties of equal social status tend to
share the net gain in esteem from encounters with each other.

If we now turn to the second type of encounter, the
meeting of players with differing social statuses, we can use the original figure
combination in Luce and Raiffa’s example. There, the utility of the bargaining
parties is no longer symmetrical; instead, we can assume that the first and the
third inflection points of the utility curves of the players are the closer to
the u – axis (see figure 4, p. 73 above), the lower the social status quo ante
of the player. In Luce and Raiffa’s example, we, therefore, can take as utility
function of the player with the higher status, e. g., of player H, the one of
Player 1, and can use the utility function of Player 2 for the person with the
lower status quo ante in the encounter, for instance, for player L.

The game can be represented in two ways: first, in
terms of the payoff in esteem units (EU) of some esteem currency, where the
bargaining outcome is determined by the point of tangency of the highest
possible indifference curve on the Edgeworth contract curve, or, second, in the
utility space of the players. (See figures 5a and 5b on the following pages).

The first representation (figure 5a) is the
normalized bargaining game (figure 3, p. 71 above), where we filled in the
values for our specific bargaining situation. The second representation (figure
5b) assumes comparability of the utility scales and extends the outer right
boundary of the feasible set to the right of the straight line through points A
(0, 1) and C (1, 0).

In the game, the halfway point B of the normalized
game (figure 3, p. 71 above) has the coordinates (. 5, 9), which is obviously
not the Nash-solution of this bargaining game. Instead, the outcome lies to the
right of the point B at B* on the Edgeworth contract curve, where the product
of the utilities is a maximum (see figures 5a and 5b).

The esteem available for distribution among the
parties, which would be evenly divided at the halfway point B in the
distribution space (see figure 5a), is divided evenly at the halfway point B*
in the utility space (see figure 5b). The division of esteem is made so that a
greater part goes to the player whose utility function for esteem increases
slower. According to our assumption about the utility function for esteem, this
person is the player who entered the encounter with the higher status quo ante.

We attain the same solution with Harsanyi’s
concession rule (see pp. 63ff. above), because, as illustrated already in the
„Graduate School Game“, the net conflict cost of a high-status person,
Chigh-status is considerably smaller than the net conflict costs of
a low-status person, Clow-status.

Bargaining Game - Payoff Space of the Players

Bargaining Game - Utility Space of the Players

Using the figures from Luce and Raiffa’s example, the steps for attaining a
bargaining agreement according to Harsanyi’s concession rule are the following:

<![if !supportLists]>(1)    <![endif]>Assuming as last offer by H, the division of
the available esteem of 75 esteem units for himself and 25 esteem units for the
other player (75/ 25) gives us B11 (. 75,  .73). If the last offer by L is 25 / 75, i. e.,
if B21 has the value of (. 25, .98), Harsanyi’s
concession rule results in 2/3 › -25/73, which means that L has to make the
next offer.

<![if !supportLists]>(2)    <![endif]>Assuming as next offer by L the division 50 /
50, B22 has the value of (. 5,  .9). Harsanyi’s concession rule results in 1/3
> -17/73, which again means that L has to make the next offer.

<![if !supportLists]>(3)    <![endif]>Only when L offers 75/25, that is, when B23
has the value (. 75,  73) and, therefore,
equals B11, an agreement is reached. However, B23
= B11 also equals B*, the Nash-solution of the game.

The predicted outcome of the second type of game,
the game for esteem among parties with differing social status, corresponds,
like the outcome of the first type of game, to that which we are used to
observing in everyday life, namely, that parties of differing social status
tend to distribute the net gain in esteem from encounters with each other
according to their bargaining position. The lion’s share in the second type of
game falls to the party with the higher social status quo ante. His share is
the larger the more significant the status difference. Among other things, this
construction of reality in its conflict with another construct, the common
sense notion of justice, is the subject of our fourth chapter, the chapter on
„Games for Criminal Status“.

We can strengthen our argument about the
differential distribution of esteem by removing a restrictive assumption, which
we implicitly upheld in the preceding analysis. Until now, we assumed that any
player in any encounter plays with the same approximate rationality, i. e.,
that any player knows his bargaining situation, the one of his opponent, and,
therefore, also the optimal bargaining strategy available to him. Now we want
to weaken this assumption of equal bargaining skill, as it seems reasonable to
assume that the bargaining skill of a low-status individual is likely to be
less adroit than the bargaining skill of a high-status person.

Under this modified assumption of a differentially
distributed approximate rationality, the predicted outcome of the first type of
encounters, the games of persons with equal status, does not change. The
outcome will be approximately a division of the net gain in esteem from the encounter,
that is, the outcome of all possible games will variate around the halfway
point B (see figure 3, p. 71 above, and examples on p. 74 above). However, the
outcome changes in the second type of game, namely the encounter between
persons with differing status quo ante. Instead of a variation around the
predicted outcome B* (see figures 5a and 5b, pp. 76f. above), as we assumed
with the strong postulate of approximate rationality, the outcome with our
weakened assumption is likely to be a distribution on the Edgeworth contract
curve which is centered to the right of B*. Therefore, weakening the
rationality assumption further supports our analysis: the net gain of esteem
from an encounter is differentially distributed by the parties such that a
larger share falls to the party with the higher status quo ante.

3.2.2. Selecting Partners

Our analysis up to now has assumed that there always
exists a feasible set of bargaining outcomes. We can maintain this assumption
if we admit sets where the outcomes are not necessarily positive. Then, every
player could play a bargaining game for esteem with any other person as well as
with any combination of persons. However, as we know that the bargaining games
actually played are limited in number, we now have to answer the question of
how games and partners are selected.

The fact that partners select each other for specific
encounters has consequences for the simple bargaining model presented in the
previous section. There, we assumed a uniform set of feasible interpartes
constructions of social behavior for all games, and we assumed the conflict
point as being located within the feasible set and as given. In this section,
we shall explain that the feasible set is a function of the social status of
the players and that the position of the conflict point depends on each
partner’s available alternative choices. Therefore, we shall modify our simple
model to account for the varying size and the varying location of the feasible
set and the variability of the conflict point. Further, we shall supplement our
model by explaining the dynamics of the choice of partner and game.

An inroad towards a solution to the problem of how
games and partners are selected is given by Homans‘ propositions about
elementary social behavior and by the results of the sociometric study by Jennings (1950).
According to these studies, we can observe a pattern of choices. The
high-status person typically is a person chosen by many and rejected by few. He
himself preferably chooses persons who have a high-status position. On the
other hand, the low-status person does not receive many choices.

Given our assumption of approximate rationality, we
can deduce from this observation that the outcomes of encounters, that is, the
gain in esteem through the encounter and the social definitions of the reality
of the behavior, are directly and positively related to the status of the
players. Moreover, we can deduce that the higher the social status of the
players, the more esteem is likely to be available, or, in other words, the
more positive inter-partes definitions of the behavior are attainable. This
deduction illustrates the Thomas theorem, which states: „If men define
situations as real, they are real in their consequences“ (Thomas, 1928:
572; cf. also, Merton, 1968: 475ff.).

To illustrate the notion that high status positively
correlates with the availability of choices, and to describe its effects on the
model of „Games for Social Status“, we give an example in which we
shall show that a dynamic conception of social bargaining necessitates the
modification of our initial simple bargaining model to a more complex one. We
shall see that, if there is the possibility to choose an alternative partner or
game, the outcome depends on a game played with threat strategies, the choice
of which is interdependent with the process of choosing a partner and a game.

Let us assume a group of games represented in the
distribution space, where each game is played as if the conflict point between
any player were the point (0, 0). Further, let us assume the participation of a
player of the category H in any of these games has the effect that the net gain
in esteem distributable is α = 200 esteem units (EU), such that the total
esteem distributable in a game with another player of the category H is (α
+ α) = 400 EU. Correspondingly, let us assume that the increment of net
gain in esteem distributable is β = 100 EU if a player of the category M
participates and that it is γ
= 50 EU if a player of category L participates.

From our previous analysis, we know the outcomes of
these games, given the utility distribution in Luce and Raiffa’s example. In
the games of the first type, that is, in games between H1 and H2,
between M1 and M2, and between L1 and L2,
the total esteem available is shared equally. There, H1 and H2
receive 200 EU each, M1 and M2 receive 100 EU each, and L1
and L2 receive 50 EU each. Taking, as an illustration, the encounter
of H with L for the outcomes of games of the second type, our predicted outcome
is that H will receive 187.5 EU and L will receive 62.5 EU of the total
available esteem of (α + γ) = 250 EU.

For the values given, a player of the category H is
always better off if he plays a game with a player of the same status category.
This is true as long as α ≥ β ≥ γ, which we assume.
On the other hand, it is likely that a player with lower status than his
bargaining partner can attain higher outcomes in encounters with him than he
could attain in encounters with partners of the same status category, even
though the esteem available for distribution is not divided into equal parts. Consequently,
we can expect a high-status person to be chosen more often than he himself can
choose. At the same time, we can expect that the situation of the low-status
person is the reverse: He has fewer choices available to him than he actually
could make. A further consequence is that we can expect a high-status person to
preferably choose other high-status persons.

We now leave our example and turn our attention to
the cases where a person chooses somebody with lower status. For this purpose,
we introduce a new element into our model, the cost of choosing, which
diminishes the outcome. We assume that the cost of choosing can vary between
zero and infinite, whereas infinite costs describe the impossibility of
choosing a specific partner. Because of the rationality postulate, only the
introduction of the costs of choosing makes any choice outside the status
category of the chooser conceivable.

Under this assumption, anybody might choose anybody.
We are then faced with the two interdependent problems of how choices are made
and the effect of the availability of alternative choices on the outcomes of
the encounters chosen. We start with an analysis of the second problem. There
we shall show how the differential availability of choices of alternative
partners modifies our initial simple bargaining model. Then we shall extend the
modified bargaining model by a supplement describing how choices are made.

John F. Nash discussed the effect of the
availability of choices on the outcomes in his analysis of optimal threat
strategies (Nash, 1953). Extending his model of the solution for bargaining
games for a fixed conflict point, the Nash-solution (Nash, 1950), to a broader
class of situations where threats can play a role, Nash presents a
predominantly prescriptive theory on the choice of optimal threat strategies
for two-party games. Of relevance for our study are the descriptive elements of
this theory.

Nash draws attention to the fact that bargaining can
be understood as a process consisting of two stages. In the first stage, the
players determine whether the feasible set consists of any outcomes that have a
positive value. Whenever there are no expected positive outcomes for one party,
the encounter is not further pursued by this party — if this party can choose
to abandon the game. We can call this stage the fact-finding stage. The
decision to continue the encounter leads to a second stage, the negotiation
stage. Here, the players point out to each other their alternative games,
thereby determining the conflict point. As we have seen in our simple
bargaining model, knowing the conflict point and the utilities of the parties
means that the outcome of the bargaining game can be predicted.

Therefore, we can represent the bargaining process
in a more complex version, which also considers the availability of alternative
choices by the following figure:

Conflict Point 

This figure represents a possible bargaining game
between the players H and L. In the fact-finding stage, both players have
determined that there are outcomes (u, v) such that u > 0 and v > 0. As a
result, they decide to enter the negotiation stage. Had there been no
alternative choice for either player, the game’s outcome would have been B*
with the coordinates (u o*, v o*). In the negotiation
stage, however, player H claims that he has at least one alternative choice available,
which would give him the utility uc > 0. As player L accepts the
claim of player H as true and cannot make the claim that he also has
alternative choices available, the conflict point for the game is (uc,
0). The outcome of the game, then, is B** with the coordinates (u0*+,
v0**), such that u0** > u0* > 0 and v0*
> v0** > 0. Our modified bargaining model, which considers the
availability of alternative choices, therefore, predicts that in games between
players of differing status, not only does the player with the higher status
quo ante receive the lion’s share of the esteem distributable but also that he
still increases his share by pointing out his alternatives.

Having updated our basic bargaining model for esteem
by including the effect of alternative choices, we see that our conclusion from
the simple bargaining game is once again confirmed. There we reasoned that any
individual entering an encounter with much perceived esteem is likely to
receive a disproportionately large share of the net gain in esteem when
interacting with an individual entering the encounter with less perceived
esteem. Assuming differential bargaining skill, i. e. , skill „to
propagandize each other into misconceptions of the utilities involved“
(Nash, 1953: 138), this result was reinforced, as we could assume the
low-status individual to be comparatively less skilled. In the following pages,
where we shall analyze the process of choosing, we will find further support
for our claim that esteem is unequally distributed in encounters.

In building our supplementary model of the process
of choosing a partner, we have to return to the observation that a high-status
person has many choices available, as he is chosen more often than his capacity
for encounters would allow him to return the choices received. We can assume a
limited capacity for encounters since each encounter requires a certain amount
of the scarce good — time. Because of the rationality postulate, we can expect
that the high-status person will tend to allocate his resources so that the
most rewarding interactions are assigned priority. Therefore, the problem of
how partners and games are chosen is the same as when a decision-maker in
monetary exchange is confronted if he has to allocate his scarce resources to
investment proposals exceeding his resources. This problem is known as the
„capital budgeting problem“. Weingartner’s (1967: 17) basic solution
to this problem, although designed as a prescriptive model, can describe the
process of choosing partners and encounters if we weaken the underlying
assumptions of perfect information and unlimited computational ability with our
assumption of approximate rationality.

We can then supplement our modified bargaining model
by the following model describing the selection of partners and games: Under
the assumption of approximate rationality („social“ rationality), a
high-status person will tend to husband his scarce resource — „time for
social exchange“, and he will tend to choose interactions so that he

Modified Bargaining Model Supplement

where bj is the perceived net-present
value which he assigns to interaction proposal j (as the total utility may be a
non-linear function of esteem [see figure 4, p. 73 above, this value is
interdependent with the ordering of the proposals), xj is an amount between 0 and 1, ctj is the perceived net outlay
(cost) required for interaction proposal j in the period t.

Although a shortcoming of this model is the implicit
simplifying assumption that the utility of esteem is a linear function of
esteem, we can find consolation in the fact that we share this shortcoming with
Weingartner’s solution to the capital budgeting problem. In other respects, the
model can be seen as an adequate and concise mapping of the process of choosing
partners and encounters in social life.

According to this model, the high-status person
maximizes his utility gain from encounters within his capacity constraint if he
prefers encounters with persons similar to him in social status. He will enter
an encounter with a person with lower status only if there is no better choice
available to him.

As the same model applies to the choices of persons
in all status categories, the model maps as a reality that choices are
preferably made within the status category of the chooser and that the
likelihood of a choice decreases with the perceived status difference. Low-status
persons of any social grouping are then left with a smaller number of
bargaining games (encounters) than persons of other status categories, and,
especially, they are excluded from games whose feasible esteem set includes
particularly large outcomes.

3.2.3. Status Maintenance
and Status Change

We have now reached a point in our discussion where
we can summarize the results. With Homans, we assumed that social status or
perceived esteem has the property to diminish over time if it is not constantly
re-confirmed. Then, our research question was how esteem is acquired in a
social system.

According to Homans‘ exposition of the social
exchange concept, we acquire esteem in social interaction as an exchange object
in various „esteem-currencies“. In our discussion of Homans‘
conceptualization of the acquisition of esteem, we pointed out that although it
parallels conceptualizations in modern economics, it does not make use of them.
As a result, the discussion of the social exchange concept omits such fundamental
conceptions as, for example, probability, utility, or probabilistic decisions,
which in economics by now have allowed the conceptualization of exchange to go
beyond the purely monetary exchange. Therefore, we proposed extending the
analysis of social exchange by using a game-theoretical approach to

As a basic unit of analysis, we chose the encounter.
Within an encounter, we argued, esteem in various esteem-currencies is
exchanged. Our game-theoretical analysis of interaction as a bargaining game
showed that individuals who enter an interaction with much perceived esteem, i.
e., high-status individuals, are likely to receive as „due amount“
half of the esteem distributable when interacting with „equals“.
However, they are likely to receive as due amount a „lion’s share“
when interacting with individuals entering an encounter with less perceived
esteem. On the other hand, low-status individuals are also likely to receive
their due amount of esteem. However, regardless of whether they interact with
persons of higher status or with „equals“, our analysis showed that
their „due amount“ of esteem from an encounter is likely to be
considerably smaller as compared with the amount likely to be received in any
encounter by a person belonging to a high-status category.

Given this unequal replenishment of esteem, any
encounter, therefore, is likely to enhance the present social position of the
actors on the social scene, and we can say that social status is a
„sticky“ state. The term „sticky“ means that we can expect
status changes in social life to take place rung by rung up or down the
„status ladder“, and, further, that the probability of a status
change towards a new status in any time period is the higher, the closer the
new status ranks to the old status, the highest probability being that a person
retains his present social status.

These results, which we obtained, by deduction from
our analytical model, are in accord with the findings of empirical studies on
social differentiation (see, e.g., Svalastoga, 1964: 530ff.). We find them
confirmed, regardless of whether we look at the larger society and observe
intragenerational or intergenerational mobility, or whether the object of observation
is the status stability in a smaller social grouping as, for instance, in an
organization or a small group.

What is new is the method by which we attained these
results. We saw the social reality as a social construction and observed the
actors on the social scene in their construction process. Our model was built
as a construct of constructs, i. e., as a construct that attempts to comprehend
the construction of social reality and reduce its complexity, then gives us
more than merely the results. At the same time, the model constitutes a
conceptualization of the object of the construction processes, namely,
elementary social behavior, where the results check the adequacy of the

3.3. Behavior as Negotiated

In the previous section, we used our bargaining
model for esteem to make Homans‘ social exchange concept more encompassing. By
laying this foundation, we now have support for the statement made in the
introductory chapter that the social status quo ante limits life chances and
life styles in a probabilistic way.

We conclude from our analysis that, because the
outcome of a bargaining game is determined by the initial bargaining position
of the parties involved, in the interrelation between social status and social
behavior, perceived behavior is a negotiated reality. The major explanatory
factor for an understanding of elementary social behavior is social status,
which, in turn, solidifies the results from previous negotiations of social
reality. In other words, we can conclude from our analysis of social
interaction as a bargaining game for esteem that behavior is limited to
status-specific behavior. Any behavior is likely to be perceived as such
regardless of its phenomenality. Because the actor participated in the construction
process, any behavior is also likely to be enacted so that it is phenomenally
similar for persons from the same status category.

As an illustration for our claim that behavior is a
negotiated reality in which one and the same behavior is differently
constructed depending on the social positions of the actor and the evaluator,
we propose to consider the quote mentioned in the introductory chapter:

In the bright glow and the warm presence of the
American Dream all men are born free and equal. Everyone in the American Dream
has the right, and often the duty, to try to succeed and to do his best to
reach the top. Its fundamental themes and propositions, that all of us are
equal and that each of us has the right to the chance of reaching the top, are
mutually contradictory, for if all men are equal there can be no top level to
aim for, no bottom one to get away from; there can be no superior or inferior positions,
but only one common level into which all Americans are born and in which all of
them will spend their lives.

Let us stop here for a self-administered
Garfinkel-experiment and look back to the last paragraph, observing ourselves in
our construction of the everyday world in terms of „What Anyone Like Us
Necessarily Knows“ (Garfinkel, 1967: 54). Let us be for a moment our own
experimenter and our own subject for the experiment and successively
interchange the persons to whom the quote is ascribed. We might, for example,
assume the quote originated from a speech of the American president; or the
talk of a French tourist to a local rotary club; or the term paper on
stratification by an undergraduate sociology major; or, finally, from the
introduction of W. Lloyd Warner’s famous work, Social Class in America (Warner et al., 1949: 3), from which it
was actually taken.

Each time we observe within ourselves a change in
the perception of credibility, quality, and acceptability of the communication.
(Pertinent experiments can be found especially in the studies constituting the
Yale series: Hovland et al., 1953; Hovland et al., 1957; Janis et al., 1959;
Rosenberg et al., 1960). Depending on his perceived status, the amount of
esteem we are willing to give to the imaginary originator of the communication
differs. His behavior, the communication, is primarily evaluated according to
the amount of esteem we allot to him, though the esteem apportioned to some
extent is also dependent on the perception of behavior enacted. In our
self-administered experiment, the evaluations might range from „immature
utterance“, to „impertinence“, to „sophisticated
statement“, depending on whether the undergraduate, the French tourist, or
Warner is seen as the originator of the communication, and, of course,
depending on our own, self-perceived status. If we, the readers, were American
presidents or high school juniors, our evaluation of the behavior of the
various communicators would again differ.

Our self-administered Garfinkel-experiment chose a straightforward
situation. We had complete information about the status of our imaginary
partner in the negotiation and our own status. No interference in our
perception of his or our status could be expected from our imaginary partner.
Negotiating the social reality of his behavior, the communication, then, was a
process consisting of only one step, our evaluation based on the amount of
esteem we were willing to give to him.

The situation in face-to-face encounters differs in
complexity. There, our willingness to give esteem to the other party is
interdependent on the willingness of the other party to allot esteem to us.
There, we can no longer play the dreamlike Walter-Mitty role of our
Garfinkel-experiment. There, our negotiation on how to evaluate a behavior is
interdependent on the outcome of a bargaining game for esteem.

In conceptualizing the differential patterns of
social behavior as a consequence of social differentiation, we have to be aware
that these patterns, in turn, affect the system of social differentiation. The
pattern of mutual dependence is operating here, as it does with social
phenomena in general. (For the classical criticism of the concept of one-sided
causation in the study of social phenomena, see Pareto, 1935: §§ 245ff.). In
social life, the effects of conditions react on the conditions themselves, or,
as Homans expressed it, „they wax and wane together“ (Homans, 1950:

In summarizing our construction of the process by
which reality is socially constructed, we can say that since esteem is a
function of the social status of the bargaining parties, the process of
constructing a behavior as a social reality is again a function of esteem. This
process, in turn, influences the amount of esteem to be distributed and therefore
indirectly influences the social status of the bargaining parties. Social
differentiation has a double implication: different evaluations for the same
behavior in differing social conditions („Quod decet Jovi, non decet bovi!
„) and differing behavior in differing social conditions. Both are implications
that must be considered simultaneously if we want to avoid „finding“
rationalizations for unequal evaluations, i. e., for a system of structured
social inequality.

Now that we have presented our notion that behavior
is a negotiated reality and that status is both the result and cause of social
negotiations, we would like to apply the „Games for Social Status“
model in the following analysis of the criminalization process. In this analysis
of „Games for Criminal Status“, we have to keep in mind that whatever
we learn about the dynamics of obtaining criminal status not only clarifies the
criminalization process; it is „not simply the study of some marginal,
exotic or esoteric group, be they criminals or criminologists“ (Gouldner,
1973: x), but contains the properties for a new understanding of the everyday
life, of becoming prominent, of becoming an outsider, or of simply being a
plain man.


In much of the literature on social science,
especially in that portion which terms itself criminological, we observe that
behavior is not seen as negotiated reality but is investigated as if it could
be divorced from the social conditions in which it is enacted. Although this is
done in many cases after paying initial lip-service to interdependence, the
emphasis is on the differences of social behavior in different social
conditions rather than on the differences in evaluation of phenomenally similar
social behavior in differing social conditions.

This social-psychological reductionism detaches
behavior from the social conditions in which it takes place and the context in
which it is evaluated. By its implicit claim that parts add up approximately to
the whole, it goes beyond merely sacrificing complexity for the sake of
analysis. This reductionism approach distorts the perception of the social
phenomenon of structured social inequality and thereby creates a socially acceptable
perception. One of the most gruesome products of the reductionist approach is
the functionalist explanation of social differentiation (Davis and Moore,

The so-called radicals and tolerance of them by the
community of scientific practitioners provide the figleaf for the scientific
cementation of societal rationalizations by the reductionist approach. The very
direction of the radicals‘ attack and their preoccupation with the conduct model
make them tolerable and even useful for perpetuating social systems based on
social inequality. In this respect, Sutherland’s treatise on
„White-Collar-Criminality“ (Sutherland, 1945), which was regarded as
radical at the time it was written, lies on the same level as The
Schwendingers‘ recent attempt (The Schwendingers, 1970; 1972) to rename as
criminal the phenomena which displease them. Perhaps contrary to their
self-image, these authors give validity to the objects of their attacks and,
therefore, present system-sustaining schemes.

In contrast to these reactive and reactionary
attempts at redefining crime, we endeavor to identify the basis for active
approaches. In our analysis, we shall show that the Criminal Justice System is
a sub-system of the general social system of structured social inequality and that
it constructs its reality in the same fashion as the general system. We shall
follow the intricacies of the construction process, which builds a reality such
that, although an individual is punished as a criminal by virtue of personal
condition or by virtue of belonging to a particular class (Turk, 1969: 9f.),
this fact can be obscured by a criminology-sponsored construction of reality
which confers on the Criminal Justice System „an artificial character of
humanness, of goodness and, sometimes of scientific seriousness“
(Bertrand, 1973: 37). Through our analysis, we shall demonstrate that both
traditional and radical-reactive criminologists are „engaged in a
collective fraud“ (Bertrand, 1973: 37). As a result, we shall be able to
subscribe to Peter Macnaughton-Smith’s recent harsh characterization of the criminologist’s
role in society as a caretaker of rationalizations:

The criminologist is . . . a government-paid
obscurantist. He is a person who studies the causes of crime, that is to say,
why people that society calls criminals did certain things; he ignores the more
basic question why, in the face of certain circumstances, certain members of
society respond by criminalizing other members (Macnaughton-Smith, 1973)

Our next chapter will deal with this fundamental
question. To provide an answer, we shall break with a tradition in criminology,
namely „the insulation of criminology from sociology in general“
(Taylor et al., 1973: 268). However, we shall not join the chorus of latter-day
meliorists who demand that „the new criminology must . . . be a normative
theory“ (Taylor et al., 1973: 280). Instead, we will try to solve the problem,
which Macnaughton-Smith pointed out by our conceptualization of the process of
criminalization as a process of status attainment, that is, as „Games for
Criminal Status“ which are played by the actors in the social scene not
differently, though in a more ritualized form, than „Games for Social

We begin our analysis of the construction process of
the social realities „crime“ and „criminal“ by considering
what has been called the attrition of justice. In the attempt to identify the
persons who construct the social realities of crime and criminal, that is, in
the attempt to identify the agents of the Criminal Justice System, we will
trace the career of an individual whose behavior could be defined as violating
a criminal law through a succession of encounters. This first section will show
that the traditional analysis of the criminalization process has left out the
most important agent of the Criminal Justice System, the victim and/or the
general public. It has thereby severely distorted the attainable perceptions of
the construction process.

The following section will demonstrate that these
successive encounters with the Criminal Justice System agents can be
conceptualized as successive status degradation ceremonies, systematically
connected as „Games for Criminal Status“. In each one of these
encounters, a part of the social reality crime and criminal is sculptured and
is eventually handed over to the next agent until the construction is completed
in a final ceremony at the court stage.

In the third section of this chapter, we will
summarize the results of our conceptualization of the criminalization process
as „Games for Criminal Status“. There we shall see that criminal
status and criminal behavior are two separate, though interconnected,
constructions of the social reality whose construction process are, in its fundamental
principles, not different from processes by which our everyday reality is

Then, the final section will illustrate the model of
„Games for Criminal Status“. There, we shall expand on the
interrelations within the model by developing propositions that can be deduced
from the model.

This will bring us to the concluding chapter of our
study. At this point, we shall have reached an understanding of the
distribution of criminal justice as a process in which a part of the social
reality is constructed, not however determined from „facts“ as common
sense postulates. We shall see that cowed behind common sense rationalizations,
justice is an artifice that maintains structured social inequality.

4.1. The Attrition of

We know that only a few known offenses lead to a
conviction from criminal and delinquency statistics. Data indicate, for
example, that of 2,780,000 index crimes known in the United States in 1965, only 3 %
resulted in offenders being sentenced to prison (Task Force Report: Science and
Technology, 1970). This process of screening offenders in the Criminal Justice
System has been called the attrition of justice (Ennis, 1967: 78).

Since we are interested in the construction of the
social reality „crime“ and „criminal“, we turn our
attention in this section to the functioning of the screening process in the
Criminal Justice System. We propose to understand the construction of the
social reality as consisting of four stages. These stages are (1) the victim
and/or general public stage, (2) the police stage, (3) the prosecutor stage,
and (4) the court stage. In each of these stages, the „offender“ and
an agent of the Criminal Justice System create a part of the social reality of
the behavior and the status leading to criminal status. Depending on whether or
not the status degradation and the definition of a behavior as criminal
behavior was successful at one stage, the „offender“ is handed over
to the next stage for further processing.

Our general model of the Criminal Justice System as
a four-stage process in which a status degradation ceremony between two parties
is successively enacted at each stage is, of course, a simplification both of
the multitude of interrelations between the stages and of interrelations within
the stages (cf., Reiss and Bordua, 1967: 38). As the further analysis will
show, however, it is a simplification, which is an adequate reduction of the
complex reality, corresponding to the one we proposed in the previous chapter
when we conceptualized the „Games for Social Status“ as two-party

Using a general model of screening as a four-step
process enables us to consider a forgotten agent, the victim and/or the general
public. As the study by Ennis (1967a) shows, it is at this stage that the primary
screening, the construction of „good“ and „bad“, is
enacted. Therefore, including the victim and/or the general public in a model
of the screening process has a double effect: Firstly, it clarifies that most
of the time, the police“ constitute only the second screen, that is, they
deal with an already prestructured reality (cf., Wilson: 1968, 85). Secondly,
it establishes a link between the general societal process of status change and
status maintenance and the process by which criminal status is attained. Therefore,
our perception of the victim and/or the general public as the first screening
agent provides the basis for our statement that „Games for Criminal
Status“ are special rather than unique forms of „Games for Social
Status“. In other words, it means that criminology is a part of social
science and not a science, which can be undertaken by itself.

This new understanding of the screening process in
the Criminal Justice System and of the identity of screening agents allows us
to bring some order into the present confusion about the screening process. As
it is now,

we know very little about the properties of this
screen, but we do know that it takes many factors into account which are not
directly related to the deviant act itself: it is sensitive to the suspect’s
social class, his past record as an offender, the amount of remorse he manages
to convey, and many similar concerns which take hold in the shifting mood of
the community. This may not be so obvious when the screen is dealing with
extreme forms of deviance like serious crimes, but in the day-by-day, filtering
processes, which take place throughout the community this feature is easily
observable. Some men who drink too much are called alcoholics and others are
not, some who act oddly are committed to hospitals and others are not, some men
who have no visible means of support are hauled into court and others are not –
and the difference between those who earn a deviant label and those who go
their own way in peace depends almost entirely on the way in which the
community sifts out and codes the many details of behavior to which it is
witness (Erikson, 1962: 11f.).

In the past, criminology only paid attention to the
screening procedure within a narrowly defined Criminal Justice System. Agents
of this system who do the screening were perceived to be the official agents — „the police“,
„the prosecutor“ and „the court“.

It is important to note here that many well-intentioned
studies of the screening process focused primarily on the police as the most
visible and, at the same time, the weakest component of the system. Oblivious
to the fact that one of the functions of the police in the social structure is
to act as „a social lightning rod“ (Blumberg, 1967: xxxi; cf. also,
Silver, 1967: 11), that is, to attract and neutralize criticism of the Criminal
Justice System, some of these studies fell into the trap prepared by a system
which protects its existence by a smoke-screen of justice. An example of this
„injudicious“ critique of the workings of the Criminal Justice System
is Skolnick’s Justice without Trial

For some time, at least, it seemed that an
understanding of the screening process in the Criminal Justice System could be
reached if a system approach is chosen where not only the screening by the
police is studied, but where at the same time the screening by the prosecutor
and by the court is investigated, that is, where the career of an individual is
followed after he has been handed over to the prosecution and from there to the

This avenue was chosen, for instance, by the
American Bar Foundation’s Survey of the Administration of Criminal Justice in
the United States
(Tiffany et al., 1967; LaFave, 1965; Miller, 1969; Newman, 1966; Dawson, 1969).
This series focuses

on two basic aspects of criminal justice
administration: (1) the study of the process of criminal justice as a system,
one which must be studied as a whole if individual decisions and agencies are
to be adequately understood; (2) a study of the exercise of discretionary power
by criminal justice agents and agencies (Remington, 1969: xvii).

Although it was able to collect and order a plethora
of information about the workings of the screening procedure in the Criminal
Justice System, the editor, in a final assessment, admitted that the survey
remained incomplete (cf., Remington, 1969: xx ff.). As a result of insufficient
funds, overtaken by the rapid development in the study of crime in the last
decade (e. g., the conception of „criminality as a status rather than
behavior“, Turk, 1969:8), and ailing from some of its theoretical and
empirical foundations, the series had fallen short of its goal to provide a
basis for continuing research of the overall process of criminal justice. However,
its conceptual approach, the understanding of the criminal justice
administration as a system of agents with discretionary power, remains.

The relative failure of the series of the American
Bar Foundation typifies the situation of traditional criminology.
Characteristics of both are the presuppositions that there is a science of criminology,
which is not part of social science, and that sufficient legitimation for its
practitioners is „goodwill“ and a „sound“ common sense.

If we perceive the social reality as socially
constructed, both presuppositions are erroneous. In order to reach an
understanding of the construction of social reality, we have to reconstruct
this reality, that is, seek to attain a construction of the second degree. Common
sense, however, by definition does not allow us to transcend common sense,
which is the first-degree construction of reality. Therefore, for the study of
the screening processes in the Criminal Justice System, our assumption means
that we have to transcend common sense. We have to study the construction
processes of social reality following construction principles, which are not
dissimilar from the principles used for other constructions of social reality.
This means we have to do criminology as social science.

We know that, although the police are on principle
the first agents of the Criminal Justice System acting in an official capacity
in the career of an individual who is suspected of having violated a criminal
law, they only uphold or reject definitions of reality conveyed to them by the
victim and/or the general public in many, if not in most, cases. A narrow
definition of the Criminal Justice System, then, does not consider that the
police as first official agents deal in a similar way with a prestructured
reality, just as the prosecutor does with the prestructured reality conveyed to
him by the police, and just as the court does with the reality handed to it by
the prosecutor. Therefore, the American Bar Foundation’s survey’s system
approach is incomplete. In a social science study, we have to widen the
definition of the Criminal Justice System and include as one of its agents the
initial constructor of the social reality. We have to include the victim and/or
the general public as a player in the „Games for Criminal Status“.

4.2. Games for Criminal

Having indicated in the previous section the
identity of the principal actors on the social scene who collectively construct
the social reality of „crime“ and „criminal“, this
section’s task is to reconstruct the construction process by which these actors
build the social reality. We shall do this by building a „Games for
Criminal Status“ model. This model, which is non-contradictory to
empirical findings, is a second-degree construction of the social reality of
the criminalization process. It orders the findings into general propositions
based on understanding social reality as we have developed it in our

Our method of presentation will be similar to the
one used in developing the model of „Games for Social Status“. We
introduce our model in a simplified world and, gradually relinquishing the
simplifying assumptions, guide it through more and more complex worlds until we
can accept it as a model for describing and explaining processes of

We begin the explanation of our model by showing how
negative esteem is distributed in successive encounters with agents of the
Criminal Justice System. Next, we discuss the effect of the capacity constraint
of the agents on the outcome of the „Games for Criminal Status“.

4.2.1. Distributing
Negative Esteem

Our model of the criminalization process corresponds
in nearly all details to the model of the social construction of behavior and status,
which we presented in the third chapter. In both models, we make the
simplifying assumption of encounters as two-party games, where the inter-partes-possible
constructions of their behavior constitute the feasible set of outcomes. We
assume that the bargaining object is esteem in various esteem-currencies and
that the bargaining outcome, the distribution of the available esteem among the
parties, is an inter-partes construction of their behavior as well as a
replenishment of their perceived esteem, that is, a determinant of status
change and status maintenance.

In contrast to the general model of „Games for
Social Status“, the special model of „Games for Criminal Status“
is, as far as the „offender“ is concerned, characterized by the
depletion of his perceived esteem, i. e., the shift from the status quo ante to
the marginal status of „being a criminal“. In successive encounters
with agents of the Criminal Justice System, the „offender“ is
processed if, and only if, he loses a game, which means if he acquires negative
esteem in this encounter, while his opponent, the agent, acquires positive
esteem through the encounter.

A general representation of a game for criminal
status, then, is given by the following figure:

Distribution of Negative Esteem 

In this general representation of a game for
criminal status, we make the following simplifying assumptions in addition to
the ones implicit in the conceptualization of the construction of the social
reality of behavior and status as a two-party bargaining game:

<![if !supportLists]>(1)    <![endif]>The players have equal status quo ante and
corresponding utility functions for esteem.

<![if !supportLists]>(2)    <![endif]>The feasible set is bounded by the lines
through the points A, B, C, and D in the utility space of the players.
This means:

<![if !supportLists]>a.
the players, at least four inter-partes constructions of their behavior are
possible such that each construction results in an outcome represented by the
points A, B, C, or D.

<![if !supportLists]>b.
probability mix of these four constructions is possible.

<![if !supportLists]>(3)    <![endif]>The constructions A and D and their mix are
connected such that they are situated on the outer right boundary of the feasible
set (the Edgeworth contract curve), which runs at a 90o angle with
the 45° line through the point of origin.

<![if !supportLists]>(4)    <![endif]>At no point is the feasible set of outcomes
(the feasible set of constructions of social behavior) positive for both

Of these four assumptions, assumptions number (2)
and number (3) are chosen to attain desirable mathematical properties. Their
removal would make our representation more complex without adding much to an
understanding of the reality construction of behavior. Therefore, we retain
these two assumptions.

Assumptions number (1) and number (4), while also
providing us with desirable mathematical properties which allow a simple
representation, are assumptions about the complexity of the social world which
have a significant bearing on our understanding of how this world operates.
After demonstrating how our model functions in a simplified world, we shall
drop these two assumptions.

We start our explanation of the model in a simple
two-period world where there exists only one agent of the Criminal Justice
System, player 1, who either chooses or does not choose to enter an encounter
with a possible „offender“, player 2. If the agent chooses not to
enter into the encounter, we assume for now that he has no available alternative
games for esteem. Player 2, the prospective „offender“, cannot choose
his partner after showing behavior that can be interpreted as violating a
criminal law. He cannot improve or impair his bargaining position by a good or
bad defense.

In this simple two-period world, our „Games for
Criminal Status“ consist of a single game with a fixed conflict point. If
we choose as a conflict point the point (u2c, v2
), where u2c= 0 and v2 c
< 0, and assume a feasible set of the type given in figure 7 (p. 95 above),
we know that player 1 will choose to play the game as long as there are
feasible outcomes which are larger than u2c = 0.
According to our assumption, player 2 has no choice. Therefore, from our
reasoning in chapter three, we know that the game’s solution is the reality construction
(u2, v2) of the inter-partes behavior. This means that
player 1 acquires in the example the esteem of (u2 – u2c)
> 0 and that player 2 acquires the esteem of (v2 – v2c)
< 0.

If we assume that any esteem outcome below the value
vx, where 0 > vx > v2, means that a
behavior is evaluated as criminal, the behavior of player 2 has been
constructed in this encounter as behavior that violates a criminal law.
However, if we remember our definition of status, this does not necessarily
mean that player 2 also attained criminal status. According to our assumptions
of status as accumulated perceived esteem and of criminal status as marginal
status, player 2 has attained criminal status only if, as a result of the
encounter, the amount of negative esteem received there sufficiently decreased
his „capital“ of accumulated perceived esteem.

From our simple two-period world, we can then take
as an intermediate result that social status quo ante has a double effect:

<![if !supportLists]>(1)    <![endif]>As the utility function for esteem is dependent
on social status (pp. 72f. above), social status quo ante influences the
construction of the social reality of a behavior, such that it becomes more
unlikely that a behavior is constructed as criminal behavior, the higher
ranking the status of a type 2 player.

<![if !supportLists]>(2)    <![endif]>The more the distance between the ranking of
the social status quo ante of a player 2 and the ranking of criminal status,
the more unlikely it is that the construction of a behavior as violating a
criminal law has the effect of demoting the social quo ante to criminal status.

We can illustrate these two results by an example
from everyday life. Let us imagine a law firm in which there are three people
employed. One is a lawyer, another is a secretary, and the third is an office boy.
Each one of these people takes stationery from the office to use for writing
personal letters. Phenomenally, therefore, each of these three people does the
same thing: he appropriates property that does not belong to him and, assuming
that there is neither an explicit nor implicit permission by the owner of this
property, thereby „technically“ violates criminal law. Conceptually,
however, the reality of their behavior is different. The reality of the
behavior of the employed lawyer might be perceived as „taking“, the
action of the secretary as „swiping“ and, finally, the action of the
office boy might be perceived as „stealing“. In our example, then,
only the action of the lowest ranking individual, the office boy, is
constructed as criminal behavior. Because of the value of the objects
appropriated, we can assume that the construction of the behavior of the office
boy as „stealing“ is not per se
sufficient to demote his social status to criminal status.

If we now increase the value of the objects
appropriated, we can observe that the reality construction „stealing“
next reaches the secretary and, finally, the lawyer. At one point in our
succession of appropriations, we might construct the reality of the behavior of
both the office boy and the secretary as „criminal behavior“ while
the lawyer’s action is not perceived as such. At this point, the status quo
ante of the office boy might be already sufficiently lowered so that he has now
conceptually reached „criminal status“.

Our example, therefore, illustrates the two
intermediate results, namely (1) that the chance of attaining the reality
construction of „criminal behavior“ and the chance of being demoted
to „criminal status“ correlates positively with the actor’s status
quo ante, and further, (2) that the reality construction of a behavior as
criminal behavior does not necessarily demote the social status quo ante to
criminal status.

With these two intermediate results, we now enter a
more complex world, a three-period world in which a second agent of the
Criminal Justice System exists. In this world, player 2 has a chance to play a
second game with the second agent (who is then player 1) if he leaves the first
game with the first agent with at least the new status of „being
suspect“. Having attained the status of „being suspect“ is
analogous to our previous reasoning that player two’s behavior has been defined
in the first game as violating a criminal law a n d that, in the perception of
the agent, the actor’s status has been sufficiently demoted so that a further
degradation in the second game could lead to the attainment of criminal status.

From our reasoning about game-playing in the
two-period world, we know that the likelihood for a status degradation to the
status of „being suspect“ is a function of the social status quo
ante, given the same feasible set of outcomes for all players of type 2. Therefore,
not all players of type 2 will have a chance to enter a second game with a
second agent as player 1.

For the ones who do enter the second game, we know
that since they have undergone a status degradation, their utility function for
esteem has changed to one which differs from the one in the first game in that
the utility for esteem increases more rapidly (see our assumptions on the shape
of the utility curve, p. 72 f. above). Further, we know that the feasible set
of outcomes is now prestructured. This means that, although the second agent
brings his own interpretation of the behavior of player 2 into the game, the
feasible set in the second game is likely to have changed for the worse for
player 2, because a second game is about to be played influences the range of
the possible inter-partes constructions.

The result, however, is similar to the one we attained
in our considerations about the two-period world: The higher the status before
the second game of a player of type 2 (i. e., the higher the remaining status
of a player who is a suspect), the less likely it is, for the same set of
feasible outcomes, that behavior is constructed as violating a criminal law and
the less likely it is that the status is degraded to the marginal status of

Increasing the number of periods, that is, the
number of successive games with agents of the Criminal Justice System and the
gradations of status loss necessary to qualify for another game, reinforces the
general result attained in the analysis of a game for social status in a
two-period world.
Any screen works to the disadvantage of low-status persons, and successive
screening compounds this effect.

4.2.2. Choosing Conflict

So far, our model of „Games for Criminal
Status“ has included three assumptions, which we would now like to remove.
These assumptions are: (1) the players have equal status and corresponding
utility functions for esteem. (2) At no point is the feasible set of outcomes
positive for both parties. (3) The point of conflict, the minimum outcome for
each player, is given.

By removing the first assumption, we know from
previous discussions that the outcome is changed in favor of the player whose
utility function for esteem increases more slowly. A high-status person,
whether a player of type 1 or type 2, then, can always expect better outcomes
for a game with the same structure than a low-status person. Therefore,
discarding the first assumption does not change the results of our previous

Elimination of the second assumption also does not
change our results. In effect, it supports our claim that „Games for
Criminal Status“ are merely a particular form of „Games for Social
Status“, as moving the Edgeworth contract curve in our example (figure 7,
p. 95 above) to the right given us the „normalized“ bargaining game
(figure 3, p. 71 above) from which we developed our model of „Games for
Social Status“.

Discarding the third assumption, the assumption of
fixed conflict points has more complex consequences for our model. If we assume
that each player has one possibility of impairing and one possibility of
improving his conflict outcome, we have nine conflict points, each with a
possibly different predicted outcome. As in our example in figure 7 (p. 95
above), these conflict points might be the points determined by u1c,
u2c, u3c, and vlc,
v2c, v3c. Since we are concerned
with a descriptive theory of the criminalization process, we have to
characterize the situations in which each of these points might be the conflict
point and, then, by predicting the outcomes for each conflict point, make the
necessary changes in our model of „Games for Criminal Status“ such
that it encompasses these situations.

We begin by considering the circumstances under
which player 1 has conflict outcomes, better or worse than the conflict outcome
u2c. From our discussion of „Games for Social Status,“
we know that there are two situations that can influence the conflict outcome.
One of them is the availability of other games, which guarantee at least the
conflict outcome of the game under consideration, and the other one is the
successful bluff pretending an alternative of this kind.

Let us first assume the possibility of a successful
bluff. Then, we can assume that in the succession of the Criminal Justice
System agents, an agent is the more sophisticated in bluffing the later is his
turn in playing the games for social status. This means that we assume that the
police are more sophisticated in bluffing than the victim and/or the general
public, that the prosecutor is more sophisticated than the police is, and that
the court is more sophisticated than the prosecutor is. If we then introduce
the possibility that a player of type 2, an „offender“, also can
bluff about the availability of alternative games, which would move the
conflict outcome, we can say that the introduction of a bluff supports our
argument. As high-status persons are likely to be screened out earlier, this is
to say that the chances of the reality construction of criminal behavior and
criminal status for the same feasible set of constructions are the higher, the
lower the social status quo ante of a type 2 player.

The same is true, as can be shown easily, for the
case of the actual availability of alternative games for players of type 1. If
we first consider the case where the actual conflict outcome is u1c
< u2c = 0, which means that player 1’s conflict
outcome is below his no-alternative-choice outcome u2c =
0 because he is dependent on player 2 — either because of family relations or
because of community relations (cf., Stinchcombe, 1963: 154; Whyte, 1943: 138f.
; Miller, 1969: 6), or because player 1 has accepted graft money, for instance,
— we know that no reasonable player 1 will enter into an encounter as long as
the outcome, the inter-partes construction of the behavior, is smaller or equal
to the outcome attainable by simply not playing the game, that is, as long as u1
≤ 0. According to our assumption, only player 1 has the choice.
Therefore, through our analysis, we have excluded a group of games and, as the
access to these types of games is differentially distributed according to
social status, we have at the same time confirmed our intermediate results.

If we now remove our assumption that no other games
are available to a player of type 1 and instead take the more realistic
assumption that any agent of the Criminal Justice System has more games
available than he actually could or would play (Reiss, 1971: 14; Miller, 1969:
157), we have a second group of cases. As we explained in chapter three (p. 84
above), the agent chooses encounters so that he

Modified Bargaining Model Supplement

He will tend to husband his scarce resources,
„time for encounter“, and he will tend to maximize the sum of the
outcomes. He can attain this in two ways.

First, he might perceive his minimum conflict
outcome to be at u3c > 0. Given the values in our
example in figure 7 (p. 95 above), this means that regardless of the conflict
outcome of player 2, the agent will never enter into an encounter, as all
values in the feasible set are below his minimum expected outcome. Examples for
this situation are (a) the individual in the general public who observes a
„crime“ but does not report it because he shuns the inconvenience,
(b) the victim who does not notify the police because he doubts police
effectiveness or fears a cancellation of or increase in insurance rates (Ennis,
1967a: 44), (c) the policeman or the prosecutor who does not pursue a case,
although there is sufficient probability of guilt, because he regards it as a
minor case (Wilson, 1968: 83f. ; Miller, 1969: 5ff.).

The second alternative involves a two-step
procedure. After the agent has decided that, on principle, he is not averse to
entering into the encounter, after the agent has perceived a feasible set of outcomes,
which might make the encounter interesting to him, he first plays a game for
the conflict point. This could be described as a game to determine the optimal
threat strategies in game-theoretical terms. The objective of this game is to
test the opponent’s strength. If an opponent turns out to be strong in this
game for the conflict point, the agent might be inclined to concede to him and,
thus, save his resources for playing other games in a similar fashion.

For instance, in a given case, an agent might be
able to reach a conflict point (u2c + δ, v1c)
if he uses all his resources in this
case. He will thereby attain in the second game for esteem the outcome of
approximately (u2, v1). However, if he decides to use
only a part of his resources on this
given case, thus saving the remainder for other encounters to be played in a
similar fashion, he might, for example, reach in the game for the conflict
point the point (u2c +
δ, v2c). This means that in the subsequent
game for esteem, he will reach the outcome of approximately (u2, v2).
The outcome (u2, v2) is, of course, less than the outcome
(u2, v1), but the agent will seek to attain the former
outcome because, in the long run, he will, by playing games in this fashion,
use his scarce resources more efficiently. By conceding to an opponent in an
individual case and thus saving the resources, „time for encounter“,
he might, for example, have the opportunity to pursue nine other cases in a
similar fashion, which, in an overall evaluation, would give him a better
outcome for his limited resources. This theoretical deduction from our model
has been documented empirically (cf., Wilson, 1968: 49; Miller, 1969: 151ff.).

If we now look back at the changes in our model of
„Games for Criminal Status“ effected by eliminating the assumption of
fixed conflict points, we find the general result derived from the simplified
two-period model confirmed. We found that for given agents of the Criminal
Justice System –although we assume that for players of type 2 there are no or
only negligible opportunities for selecting agents, or, in other words, that
there is no „court-shopping“ — player 2 can influence his conflict
outcome by more or less astutely playing the game. Astutely playing a game for
criminal status means playing a game with all available resources, be it one’s
own resources like intelligence and/or the ability to master social games, or
be it the help of someone else who contributes his ability to play games for
criminal status, such as a defense counsel. From our general analysis of
„Games for Social Status“, we can conclude that these resources are
differentially distributed according to social status (cf., Stinchcombe, 1963:
157). Using this conclusion in the model of „Games for Criminal Status‘.‘,
we arrive at two general results. Firstly, social status negatively correlates
with the chance that a behavior is socially constructed to the social reality
„criminal behavior“. Secondly, the process of screening of
„offenders“ in successive encounters with various agents of the
Criminal Justice System gives an „unjust“ advantage to the
„offender“ with high social status, so that the chances of attaining
criminal status are inversely related to the ranking of the status before the
first encounter.

4. 3. Criminal Status and
Criminal Behavior

The results we attained in our analysis of the
criminalization process as „Games for Criminal Status“ are not novel.
They are in accord with the conceptions of most criminologists, with the
empirical findings, with the official statistics, and with the results of
studies on the „dark-figure“ of crime. Had we been primarily
concerned with reaching these results, our study would have been as inane as
the attempt to carry water to the sea.

Instead, our primary interest was to understand the
processes by which these results become a reality. For this purpose, we
developed a model which is non-contradictory to findings in social science but
integrates the information we have about „crime“ and
„criminals“ into the information we have about social behavior and
social status in general, ordering it into consistent and valid propositions. Therefore,
it is the method that allows us to build models of „Games for Social
Status“ and „Games for Criminal Status“ and whose validity is
indicated by the results that are of interest and importance.

Guided by the notion of social reality as
construction by the actors on the social scene, we followed the construction
processes using the simplifying assumption of encounters as two-party games. In
the discussion of the attrition of justice, we showed that the first game for
criminal status in most cases is played with an agent of the Criminal Justice
System who is often not perceived as such, namely the victim and/or the general
public. In the encounter with this first agent, most incidents, which could be
regarded as violations of the criminal law, are screened out. This was demonstrated
for the case of the victim in a study done by Ennis (1967 a: 41ff.).

In this first encounter, the social reality of the
behavior is constructed. As our analysis showed, there are on principle four
outcomes to such an encounter: Either (1) the agent abstains from the game, or
(2) he plays it and the result is a construction of the behavior as
non-criminal, or (3) he plays it and constructs the behavior as criminal but
does not perceive the status as sufficiently downgraded so that a next agent can
be notified to play another game, or (4) he plays the game, regards the
behavior as criminal, and perceives the status as sufficiently downgraded for a
game with the next agent.

Only if the first agent decides on a construction of
the behavior as criminal, and a l s o considers the status of his opponent as
sufficiently down-graded, that is, considers his opponent as a result of the
encounter as a „suspect“, will he notify the next agent, „the
police“ or „the prosecutor“, who then starts a similar game, but
now with a partially prestructured reality of the behavior and the status of
the opponent. In the final stage of the succession of games, then, „the
court“ constructs its reality of behavior and status, evaluating, like
other agents before it, the behavior as criminal or non-criminal. In contrast
to the other agents, who only have the opportunity of conferring upon the
„offender“ less finalized constructions of the reality of his status,
like the status of being „suspect“ or of being „accused“,
the court, however, decides on the social reality of the criminal status of the

Therefore, according to our analysis, we have to
distinguish between criminal status and criminal behavior in the same way we
make such distinctions in everyday life situations. Criminal behavior as one
form of social behavior is a construction of the social reality, which any
individual can make in a social system. The social reality of criminal status,
as one form of social status, is a construction of the social reality made by
an authorized agent, „the court“. It is the result of multiple
encounters in which, at each encounter, the sum of the perceived esteem of the
„offender“, his social status, is depleted or replenished. In the
Criminal Justice System, this process has been ritualized so that only one of
its agents, „the court“, can change the reality of a social status to
the marginal status of being „criminal“.

The method of reconstructing the construction
processes of social reality enables us to show that our analysis of the
criminalization process as „Games for Social Status“ is a corollary
to an analysis of status change and status maintenance. The result of the
latter analysis was the seemingly innocuous statement that social status is a
„sticky“ state, as the status quo before any encounter
probabilistically determines the construction of social behavior and social
status after the encounter. For the special case of criminal status and
criminal behavior, we can now state as a corollary that the chances that
behavior is constructed as criminal behavior and the chances that an actor on
the social scene attains criminal status are probabilistically determined by
the actor’s status before any game for criminal status.

With this understanding of the workings of the
Criminal Justice System, derived through our method of analyzing social life
through the concept of social exchange, we would be able, in an
„applied“ criminology, to evaluate change proposals and to influence
or manipulate the process of reality construction at each stage of the
criminalization process and thereby alter the differential distribution of
immunity. Corresponding to our argument in the second chapter, however, we have
to refrain, for the sake of a system understanding, from such an application of
our model within the „real“ world, and we have to leave this more
pleasing task of adjusting the social is-world to a social ought-world for
others, for the system-planners.

4.4. Propositions on the
Construction of the Social Reality of Criminal Behavior and Criminal Status

The task remaining is to illustrate our model of
„Games for Criminal Status“ in the form of propositions.

By propositions, we mean statements in the form of theorems,
which can be deduced from our model. The propositions to be developed in this
section serve a double purpose. Firstly, by expanding on the interrelations
within the model, they further our understanding of the construction of social
reality by the Criminal Justice System and clarify what might have remained
arcane within our model. Secondly, by stating the interrelations within the
model in the form of theorems, they open the model for empirical investigation
— i. e. Hypotheses, as operationalized propositions, could be developed from
the propositions and the validity of these hypotheses investigated in the

The Basic Proposition

At the end of the previous section, we stated – as a
result of the inquiry into the criminalization process – a corollary to a game-theoretical
analysis of status change and status maintenance based on the concept of social
exchange. Focusing on the construction of the social reality of criminal
behavior and criminal status, this corollary is at the same time the first and
most basic proposition which can be deduced from our model of „Games for
Criminal Status“. It can be formulated as follows:

Proposition 1: The chances that behavior is
constructed as criminal behavior and that an actor on the social scene attains
criminal status are probabilistically determined by the actor’s status before
any game for criminal status.

We can clarify and supplement this basic proposition
further through a series of propositions if we analyze the effect of
(hypothetical) changes in existing conditions within a Criminal Justice System
on the outcome of bargaining games (encounters) for negative esteem. Then, to
generate other propositions from the model of „Games for Criminal
Status“, we shall be analyzing in this section the complex function of the
model by varying the state of the elements constituting the bargaining

Varying the State of Elements in a Bargaining Situation

We must first identify the elements characterizing a
given bargaining situation within a game for criminal status. Furthermore, we must
look at the conceivable combinations, which these elements can assume if their
state is varied.

If for the moment, we leave the elements in their
abstract form, we can distinguish the following six elements within a given
bargaining situation (see figure 8 on the next page):

<![if !supportLists]>A.      <![endif]>Feasible set
a. Contract curve
b.Conflict-line of Player 1, the uc-line
c.Conflict-line of Player 2, the vc -line

B. Utility functions
d.Utility function of Player 1
e.Utility function of Player 2

C. f. Number of successive games necessary to
obtain criminal status

As our model is based solely on these six elements,
it might appear that we can draw our propositions only from a few straightforward
situations. However, we have to consider that the elements themselves are
abstractions from complex situations. Furthermore, we have to take into account
that, even if we restricted ourselves to the analysis of only three states for
each element (that is, to x0, x1, and x2,
where xo is a given state of an element x and where x1
and x2 are chosen such that, e. g., in the case of the conflict line
of Player 1 b1 > b0 and b2 < b0),
we would have to derive the propositions from a matrix of the following type:

Elements of a Bargaining Situation

States of the Elements - Combination

This matrix means that the number of combinations,
each of the type (ai, bi,… fi), is:

(3) x (3) x . . . (3) – (1) = 36 – 1 =

At a second glance, we see that even the small
number of elements in the bargaining model provide for a larger number of
combinations from which propositions could be drawn than it is feasible to
analyze within this study. This problem becomes even more evident if, instead
of three states for each element, we admit two gradations within each direction
of change such that each element could assume the states x0, x1a,
x1b, x2a, and x2b. The number of possible
combinations from which propositions could be drawn would be much higher,
namely 56 – 1 = 15, 624. Alternatively, if there were just five more
states admitted for each element, the number of possible combinations would
rise to the astronomical amount of 10,000,000.

Based on the theoretical reflections presented in
chapter 2, we already excluded an evaluation of the different conceivable
combinations in favor of a system understanding. At this point, then, we have
to further restrict our discussion to a few of these combinations. Given the
matrix presented above, it seems reasonable to confine our discussion to only
those combinations (ai, bi, … fi) where all
but one i equal 0. In other words, it appears to be reasonable to restrict the
further discussion to combinations where the state of only one element of a given
bargaining situation is altered.

This will still leave us with twelve combinations
from which we shall draw six propositions belonging to a large family of
propositions. These exemplary propositions are the following ones:

Proposition 2: If it were mandatory for any
„offender“ to enter a coalition with a person capable of
demonstrating additional interpretations of given facts at least at the court
stage, then the influence of social status quo ante of the „offender“
on the construction of the social reality of a behavior and a status would be

Proposition 3: If a player of type 1 knows about his
opponent’s „criminal record“, this will increase the chances that
Player 2’s behavior is constructed as criminal behavior and his status as
criminal status.

Proposition 4: The clearer the „evidence“
(i. e., the fewer the possible interpretations) for the same phenomenal
behavior, the higher the chances for a conviction.

Proposition 5: The larger the number of successive
encounters necessary to obtain criminal status, the more likely it is that one
would find a disproportionately large number of low-status persons among the
persons convicted.

Proposition 6: One’s peers are likely to come to
less favorable constructions of reality for the same phenomenal act than agents
of a Criminal Justice System whose status is below
the one of a prospective „offender“ and to more favorable
constructions than agents with higher

Proposition 7: The greater the number of
alternatives for an agent, the more he will be inclined to choose from the
number of available encounters those encounters in which the return in esteem,
compared to the effort to attain this esteem, is the highest.


The first three of these propositions we intend to
illustrate through hypothetical changes in the situation of a
„real“-life case, namely: PEOPLE v. GORSHEN (51 C. 2d 716 (1959); 336
P. 2d 492 (1959)). The analysis of this case at the trial court stage will give
us at the same time also the opportunity to demonstrate once again the
explanatory and analytical powers of the model of „Games for Criminal

PEOPLE v. GORSHEN is a criminal case that an
appellate court decided. We shall only be interested here in the presentation
of the case and the decision-making process at the trial court stage.

In this case, a fifty-five-year-old longshoreman had
shot and killed his foreman in the presence of police officers after having
announced that he would do so. Two hours before the shooting, the foreman had
told the defendant that he was drunk and was not doing his work properly and
had directed the defendant to go home.

The defendant was charged with first-degree murder
on the theory that the homicide was willful, deliberate, and premeditated. The
theory of the defense was that GORSHEN’s mental condition had impaired his
mental facilities to such an extent that, although not insane, he lacked the
capacity to deliberate and premeditate. This case is unique because prior to
GORSHEN, evidence of mental condition could not be received in a California
Court except upon a plea of not guilty by reason of insanity. The decision of
the Supreme Court of California, in this case, introduced the doctrine of
limited or diminished capacity.

At the trial court, a psychiatrist, Dr. Bernard L.
Diamond, diagnosed the defendant as suffering from chronic paranoic
schizophrenia. Dr. Diamond testified that the defendant became concerned about
the loss of sexual power a year before the shooting. With this concern, the defendant’s
sexual hallucinations, which he had had for the previous 20 years, occurred
with increased frequency, and his ability to work became increasingly important
to him as proof of manhood. On the night of the shooting, the foreman’s
statement that the defendant was drunk and should leave his work was to the
defendant the psychological equivalent of the statement that „You’re not a
man, you’re impotent . . . you’re a sexual pervert.”

In the doctor’s opinion, at this point, the
defendant felt threatened by the loss of his sanity and was compelled to act in
a way that would serve as a defense against this. At the time of the shooting,
he acted almost like an automaton; „even the fact that policemen were
right at his elbow and there was no possibility of getting away with this,
still couldn’t stop the train of obsessive thoughts which resulted in the
killing . . . (H) e did not have the mental state which is required for malice
aforethought or premeditation or anything which implies intention, deliberation
or premeditation.”

The Bargaining Situation in
the GORSHEN Case

In terms of our abstract model of „Games for
Criminal Status”, this situation at the trial court can be represented by the
following figure (see figure 9 on the next page):

The Gorshen Case 

In this representation of the GORSHEN case, u stands
for the utility for esteem of Player 1 (trial court) and v for the utility for
esteem of Player 2 (defendant). Both the utility function u and the utility
function v are assumed to be given and not subject to change. uoc
is the conflict outcome for Player 1, i. e., which he can expect as a minimum
should the opponent refuse to agree to any other bargaining solution within the
feasible set or on the Edgeworth contract curve. voc is
the corresponding conflict outcome for Player 2. Finally, the line through the
points A and B on the conflict lines uoc and voc is the Edgeworth contract

Then, the feasible set of bargaining outcomes is
described by the reality constructions A, B, and C and any mix thereof. From
the previous discussion, we know that the reality construction to be chosen (i.
e., the bargaining outcome for this game) will be a point Xo with
the coordinates (uo, vo) on the Edgeworth contract curve.

As this case takes place at the court stage within
the sequence of encounters in the Criminal Justice System — that is, as the
finalized status determination (disregarding appellate proceedings, for our
purposes) for Player 2 is made at this stage — we know that the construction
of the reality of Player 2’s behavior means at the same time the reconstruction
of the reality of Player 2’s status. If for Player 2 the minimum esteem loss
necessary for a judgment of first-degree murder is e1 (corresponding
to the value ve1 on his utility-scale), and if e2
(corresponding to the value ve2) is the minimum esteem loss
necessary for a judgment of second-degree murder and, finally, if e3
(corresponding to the value ve3) is the minimum esteem loss
necessary for a judgment of „other homicide“, we also know that the
judgment in the GORSHEN case of second-degree murder, i. e. the bargaining
outcome X0, has in our model the coordinates (uo, ve2
≥ vo > ve1).

The Decision Making Process
in the GORSHEN Case

Before we make our first change in the bargaining
situation, observe its results on the outcome, and draw from it the first of a
series of propositions, which explain the basic Proposition 1, we shall let the
trial court state the matters considered in reaching its decision. Beyond
preparing the basis for later hypothetical changes in the situation, this will
give us a concrete example of the analytical value of the „Games for
Criminal Status“ model.

According to the trial court, „up till the time
that Dr. Diamond testified in this case, there was no explanation of why this
crime was committed. . . . (The doctor is) the first person that has any
reasonable explanation. Whether it’s correct or not, I don’t know . . . (I)f I
would follow Dr. Diamond’s testimony in toto, I should acquit this man. .. .

„I‘ m willing to go on record that in all
probability his theories are correct . . . that he (the defendant) had no
particular intent to commit this crime.

„I like to be advanced. But it seems to me that
my hands are tied with the legal jurisprudence as it stands today, and that’s
why I‘ m saying this for the purpose of the record. The Appellate Court might
say that my hands are not tied, but I think they are. . . . (E)ven accepting in
part the testimony of Dr. Diamond, I still feel that this man is guilty of
second degree murder. „

In reply to defense counsel’s assertion that
„There is not one scintilla of malice,” the court said, „it all
depends on how you view it. . . . Some other person or another Judge, might
say, ‚ Malice, why, it’s full of it. He planned it. He said he was going to do
it, he went home, he had an hour. ‚ „

Again defense asked, „Does your Honor feel that
there is malice (afore-thought) here?“ The court replied, „there was
some intent. Now, whether you have free will or not free will, that’s so
advanced, we’re not prepared for that . . . There’s plenty, plenty of malice as
far as statements are concerned, and plenty of malice as far as actions are
concerned. Now, whether he was compelled to do this because of some mental
condition, that is so advanced and so far from us that we don’t understand it.
. . . (I)t would be a perfect first degree if it wasn’t for the fact that he’s
never been in trouble and because of the statement of the Psychiatrist. „

 Three Propositions Exemplified on the GORSHEN

If we now begin our changes in the state of elements
of the bargaining situation by considering the very last statement of the trial
court, we can establish as a first proposition supplementing the basic
Proposition 1:

P r o p o s i t i on 2: If it were mandatory for
any „offender“ to enter a coalition with a person capable of
demonstrating additional interpretations of given facts at least at the court
stage, then the influence of social status quo ante of the „offender“
on the construction of the social reality of a behavior and a status would be

From the GORSHEN case, we arrive at this proposition
by hypothesizing that the same case was to be decided by the same court,
however with the difference that no psychiatric testimony was available. Here
we have recourse to the trial court’s statement.

We can also arrive at the same proposition from our
abstract model. The reasoning is as follows: If we accept the claim made
through our model that the model is an abstraction of the situation in the
criminalization process or, in other words, if we accept the claim that the six
elements of the bargaining situation are both necessary and sufficient
conditions for the description and analysis of any game for criminal status,
then we know that any change in the situation corresponds to a change in the
state of one or more of the six elements in the model. Therefore, if we want to
have a probabilistic prediction of the construction of the social reality of
behavior and status in a given case, we must „translate“ the change
in the „real“-life situation into the parameters of the model. We can
accomplish this by identifying the element of the bargaining situation, which
has been changed, and the direction in which it has been changed. Then we can
observe the result in the model, compare it to the result without the change,
and, finally, we can „translate“ the observed parameter back into the
„real“-life situation. On the other hand, if we want to arrive at a
proposition, the procedure to attain it differs only insofar as we
„translate“ the parameter which results in our model after a
hypothetical change into the general statement: a is a function of b (a = f (b)
). Furthermore, of course, we can also arrive at such a proposition if we do
not have a specific case in mind and generate propositions directly from the

Returning now to the GORSHEN case, we know that if
the psychiatric testimony is omitted, we certainly have not changed the status
of the Players (and, therefore, their utility functions for esteem), the number
of successive games necessary to obtain criminal status or the conflict outcome
of Player 2. Therefore, four elements of the bargaining situation are untouched
by the change in the situation. Also, since the „evidence“ in the
case has remained the same and only the number of possible interpretations of
the same „evidence“ has been decreased, we know that the position of
the Edgeworth contract curve (which we also could call
„evidence“-line in a game for criminal status) has remained
unchanged. This leaves us with the last of the six elements in a bargaining
situation, namely the conflict outcome of Player 1.

It becomes immediately apparent that this is the
parameter of the model, which has been changed by the hypothetical omission of
the psychiatric testimony in the GORSHEN case, if we envision the case of a
„conflict“, i. e., a situation where both parties refuse to deal with
each other and to come to an agreement. Then, the conflict outcome of Player 2
would remain the same, whether with or without the psychiatric testimony. However,
Player 1’s outcome in a conflict would be better if there was no psychiatric
testimony, as his esteem would not additionally and negatively be influenced
for not considering this testimony.

In terms of our model, then, the hypothetical
omission of the psychiatric testimony in the GORSHEN case would mean that we
moved the conflict line ucc from its previous position uoc
to the new position u1c , such that uoc  < u1c. This has
the effect that the feasible set of reality constructions becomes smaller than
before. From the rules developed for the bargaining game for esteem, we further
know that the new outcome X1 has the coordinates (u1, v1),
such that compared to the previous outcome X0 with the coordinates
(uo, vo), we have u1 > uo and v1
< vo. This means that the outcome for the defendant is more
disadvantageous than before the change of state of the element. Generalizing
this result leads us to Proposition 2.

However, Proposition 2 is only one of the propositions
which can be developed based on this analysis. It has to be remembered here, as
well as with the propositions brought forward later in this section, that each
one belongs to a family of propositions and that these families are
interrelated. On the one hand, the same proposition could be stated more or
less expressly (e. g., in Proposition 2, we also could say: „. . . to be
represented by a counsel, then . . . „); on the other hand, coalitions are
not the only way by which the conflict line of Player 1 can be changed. Other
ways by which it could be changed would be by procedural rules or even simply
by the ability of players of type 2 to play a game for criminal status.

Another proposition belonging to the family of
proposition 2 would be, then: If every player of type 2 has to be told his
legal rights at the time of arrest (cf. MIRANDA v. ARIZONA, 384 U. S.
436 (1966)), this will tend to equalize the chances of members
of different status groups that the social reality of their behavior will not
be constructed as criminal behavior and the chances that their status after the
final encounter at the court stage will not be constructed as criminal status.

If we now leave the psychiatric testimony and the
position of the conflict line for Player 1 and turn to the other part of the
trial court’s candid final statement („It would be a perfect first degree,
if it wasn’t for the fact that he’s never been in trouble . . . „), we can
develop another proposition concerning the state of a second element of the
bargaining situation. This proposition also seems at odds with the traditional
understanding of „equal justice before the law“. However, it follows
from the proposed approach in the „Games for Criminal Status“ model
and provides an ordering of seeming incongruities in the administration of

First, we have to „translate“ this
statement of the court into the parameters of our model. Here, the state of two
elements in the model could be involved: One is the conflict line of Player 2,
the other is his utility function for esteem.

Though cases are imaginable where a previous
criminal record might influence the position of Player 2’s conflict line, this
certainly is not so in the GORSHEN case, since here Section 190 of the California Penal
Code provides an absolute conflict line through the death penalty in cases
where one of the possible reality constructions is first-degree murder. Instead,
the court’s statement concerns the utility-function for esteem of Player 2. If
we bring the court’s perception of the bargaining situation into the form of a
proposition, it could be restated in the following way:

Proposition 3: If a player of type 1 knows about his
opponent’s „criminal record“, this will increase the chances that
Player 2’s behavior is constructed as criminal behavior and his status as
criminal status.

Proposition 3 follows directly from our model if we
remember our previous discussion of the effect of the utility curve on the
bargaining outcome in the example between a rich man and a poor man (see above,
pages 62 f. and 72 ff.).

As in the previous proposition, we have to remember
that this proposition is only one of a family of propositions concerning the
state of the utility function of Player 2. Another one belonging to this family
would be, e.g.: The same amount of status information (e. g., on occupation, on
sex, on childhood) about a prospective „offender“ works to the
advantage of a „high“-status „offender“ and the
disadvantage of a „low“-status „offender“.

Turning now to the „evidence“ in the
GORSHEN case, we come to the fourth proposition:

Proposition 4: The clearer the „evidence“
(i. e., the fewer the feasible interpretations) for the same phenomenal
behavior, the higher are the chances for a conviction.

If we regard the Edgeworth contract curve as the
„evidence“ line and if we define „evidence“ as being the
„clearer“, the smaller the angle of the lines through CB and AB (see
figure 9, page 110), this follows directly from our model.

This proposition seems very obvious and almost too
superfluous to mention. However, as „evidence“ is regarded as a
function of privacy and privacy as differentially distributed among status
groups (Stinchcombe, 1963), propositions belonging to the same family are not
quite so obvious and even contradict prevailing beliefs. E.g., another
proposition belonging to this family would be: A law protecting the privacy of one’s
home has the effect that in criminal cases, the chances for a conviction are
higher for „low“-status persons in a social system.

At this point, we propose to leave the GORSHEN case,
as there is a danger of overinterpretation given the limited information on it.
At any rate, it was meant to be used mainly as an illustration for both the
explanatory and analytical powers of the model of „Games for Criminal
Status“. We have to keep in mind that even though this case, through the
candid language of the trial court, demonstrated the criminalization process as
a „Game for Criminal Status“, it had at least two serious shortcomings.
Firstly, it is a specific case where a decision has already been made. That
means that propositions drawn from it run the danger of not being sufficiently
abstract and valid only for this and related cases. Secondly, there is more
danger that our illustration of the model with the GORSHEN case, just like the
traditional approaches in criminology, concentrates too much on the official
agents of the Criminal Justice System, thereby losing sight of what we perceive
to be the most important agent, the victim and/or the general public.

Another Three Propositions

There are still three elements of the bargaining
situation which have to be discussed and the parameters of which have to be
illustrated through propositions. These are (1) the number of successive games
necessary to obtain criminal status, (2) the utility function for esteem of
Player 1, and (3) the conflict line of Player 2.

Starting with the element, „number of
successive games necessary to obtain criminal status“, we can set up the
following proposition:

Proposition 5: The larger the number of successive
encounters necessary to obtain criminal status (e.g., the determination of
guilt separated from sentencing would mean an additional game), the more likely
one would find a disproportionately large number of low-status persons among
the persons convicted.

We can derive this proposition from the results
theoretically obtained through our model. We found that a low-status person
automatically has a disadvantage in any game for criminal status. If that is
true, then a sequence of games compounds this disadvantage, and our Proposition
5 follows.

This proposition also belongs to a family of
propositions. Another one would be: The free choice of counsel (which itself is
a subgame for coalition) results in a disadvantage to the bargaining position
of low status „offenders“.

If we turn now to the next element, the utility
function for esteem of Player 1, we can derive from our model the following

Proposition 6: One’s peers (e. g., lieutenants in a
court-martial for a lieutenant) are likely to come to less favorable
constructions of reality for the same phenomenal act than agents of a Criminal
Justice System whose status is b e l o w the one of the prospective,
„offender“ (e. g., privates) and to more favorable constructions than
agents with higher status (e. g.,

The reasoning here is similar to the reasoning used
for Proposition 3. A member of the family of this proposition is, e.g.: A
„war on white-collar crime“ is more likely to produce results if high-status
agents are in charge of detection, investigation, prosecution, and judgment of
white-collar „offenders“.

Coming, finally, to the last element in the
bargaining situation, namely the conflict line of Player 2, a proposition would

<![if !mso]>



<![if !mso]>


Proposition 7: The greater the number of alternatives for an agent,
the more he will be inclined to choose from that number those encounters in
which the return in esteem, as compared to the effort to attain this esteem, is
the highest (i. e., encounters in which a player of type 2 is not likely to be
able to engage in a „hard“ bargain for the position of his conflict
line guaranteeing his minimum outcome).

This proposition directly follows from the
rationality postulate that more is always preferred to less. A more specific
proposition within the same family is, e. g.: If the outcome level of
alternatives for an agent is reduced (as this can be the case with a special
law enforcement group, e. g., IRS-agents),
then the likelihood increases that previously unacceptable games (tax fraud
would be in most cases an unacceptable game in a non-specialized law
enforcement group) might now be accepted.

The Model of „Games for
Criminal Status“ and the Propositions Deduced from It

The propositions in this section are all examples of
the descriptive and analytical powers of the „Games for Criminal
Status“ model. As we repeatedly emphasized in this section, many more
propositions on different levels of generality could be deduced from the model.
The ones chosen here were selected more or less arbitrarily from the universe
of possible propositions to delineate the model’s scope.

Propositions such as those demonstrated in this
section could be empirically investigated once they are brought into an
operationalized form. They could also be used for social policy decisions
concerning the Criminal Justice System, that is, they could be used for system planning.
However, as we explained previously, both these steps are beyond the scope of
this study.

What we tried to achieve instead was the
construction of an alternative model capable of providing an understanding of
the workings and the dynamics of a Criminal Justice System based on fewer
assumptions than traditional approaches in criminology. Using the conception
that „crime“ is one of the possible relationships of an individual to
his society, we designed such an alternative model and illustrated it by

The model of „Games for Criminal Status“
might not be a very pleasing conceptualization of the criminalization process
to traditional preconceptions of „crime“ and „criminal“ or
of „behavior“ and „status“. However, it is a conceptualization
that provides an understanding of the criminalization process beyond the one,
which can be reached with traditional preconceptions. It explains what seemed
to be not explainable: The Criminalization of Man in Society.


We began our study with the statement that
„criminology has been and still is a laymen’s science, traditionally
incapable of transcending insights beyond common sense and traditionally
occupied with rationalizing societal irrationalities.” In the course of our
reasoning, we attempted to step beyond common sense. Now, we can sum up the
results of our study with the statement that justice is order through
structured social inequality.

When we use the term justice, we mean more than
merely criminal justice; what we have in mind is the justice of social
differentiation in social groupings in general. We see the principles of
justice at work in the evaluation of the behavior and the status of the
„leader“ and the „outsider“ in a small group, just as we
observe them in the „making of a president“ and the making of a
criminal in the larger society. As we understand it, justice is a functional
device, an artifice, which maintains structured social inequality in general.
It is a complicated, intricate cover for the existence of social
differentiation. The idea of justice could be seen as making the social world
livable. It could be understood as supporting constructions like the American
Dream, just as it could be regarded as providing reasons for differential
social evaluation and, in the case of criminal justice, for the differential
distribution of immunity.

Social science has done its share in justifying the
existing social structure. Only in recent years has social science come to
recognize its problem of being a serf to a social structure, which distributes
the good things in life in a way, which is blatantly inconsistent with society’s
own maxims. Gouldner’s book The Coming Crisis in Western Sociology (1970) and Bertrand’s
characterization of the criminologists as being „engaged in a collective
fraud“ (1973: 37) are jarring testimonials to the growing awareness of today’s
problematic situation in social science.

In the light of these circumstances, the current
reorientation of social science towards a social meliorism becomes
understandable — and indeed predictable from a position which perceives, as we
do, social science as a social activity and, therefore, social science as a
metascience of social science. It was predictable because we know from the
findings on the behavior of individuals or social entities in situations of
diffused reality, of feeling “ u n b e h a u s t „, that there is a
tendency to restructure a shattered environment, actually and cognitively. For
instance, in Festinger’s study, When Prophecy Fails (Festinger et al., 1956), a
group of religious fanatics stepped up proselyting after their prophecy,
predicting the end of the world, failed. Similarly, we know from the study of
group problem-solving that groups experiencing severe difficulties in the stage
of problem definition are likely to skip this stage and become
„actionistic „.

In our own study, we considered the comparable
crisis in social science. We proposed, as an attempt to avoid the ought-worlds
of the traditional and the radical-reactive approach, a 1′ art pour 1′ art
approach to social science, that is, an approach which postulates the
restriction of social science to problem definitions, to a „pure“
social science, and which abdicates the claim that a scientific endeavor of
social investigation can simultaneously provide problem definitions and problem
solutions. Through the example of „Gouldner’s Complaint“, we showed
that such a distinction between „pure“ and „applied“ social
science is both challenging and painful to draw. However, we also showed that
this distinction is necessary if we want to improve the chances that our
constructions in social ’science are reconstructions of the construction of
social reality, instead of merely constructions of the first degree which run
parallel and in competition with the common sense structuring of the social
environment. Furthermore, we showed that this distinction between
„pure“ and „applied“ social science is a precondition for
restructuring (changing) the presently existing social reality.

Within this theoretical framework, we developed our
„Concept of Marginality“ and the models of „Games for Social
Status“ and of „Games for Criminal Status“. Perceiving the
second model as a particular case of the first model, we attempted to integrate
criminology into an overall concept of social science as the study of social
differentiation. As a result, we could state that justice is order through
structured social inequality. This result is bleak and much less pleasing than
a construction of a social ought-world. It does not show us a world as we would
like it to be. However, it is a second-degree construction of the social
reality that is non-contradictory to empirical findings, which orders them into
general propositions, thus providing a basis for restructuring the social
is-world. Then, our models of „Games for Status“ are a problem
definition, not a problem solution.

Whether we have to or want to accept a social
reality of structured inequality, as it exists now, was not at issue. This is a
separate question for T h e Sociology of the Possible (Ofshe, 1970a) and a significant
challenge for social science and the sociological imagination.


F. J. and R. J. Aumann

1963     „A Definition of Subjective
Probability.” The Annals of Mathematical Statistics 34 (1963): 199ff.

Bernard and Walter Hirsch, editors

1962     The Sociology of Science. New York: The
Free Press of Glencoe

Howard S.

1963     Outsiders: Studies in the Sociology of
Deviance. New York: The Free Press of Glencoe

Howard S., editor

1964     The Other Side: Perspectives on Deviance.
New York: The Free Press of Glencoe

BENDIX, Reinhard
and Guenther Roth

1971     Scholarship and Partisanship: Essays on Max
Weber. Berkeley: University of California

Peter L.

1963     Invitation to Sociology: A Humanistic
Perspective. Garden City, N. Y.: Doubleday


1738     „Exposition of a New Theory on the
Measurement of Risk. Econometrica 22 (1954): 23ff.

Ludwig von

1968     General System Theory: Foundations,
Development, Applications. New York: Braziller

Marie Andree

1973     „Dialogue with Marie Andree Bertrand.
An Interview by Virginia Enquist Grabiner. “ Issues in Criminology 8
(1973): 31ff.

Wilbert E.

1957     The Art of Scientific Investigation. (3rd
ed.) New York: Norton

Peter M.

1964     Exchange and Power in Social Life New York:

Peter M.

1967     „The Research Process in the Study of
‚ The Dynamics of Bureaucracy‘.” In: Hammond, 1967: 18ff.

Abraham S.

1967     Criminal Justice. Chicago: Quadrangle

David J., editor

1967     The Police: Six Sociological Essays. New
York: Wiley


1947     On Understanding Science. An Historical
Approach. New Haven: Yale University Press


1968     Essays in the Theory of Society. Stanford,
Calif.: Stanford University Press

Kingsley and Wilbert E. Moore

1945     „Some Principles of Stratification.”

Sociological Review 10 (1945): 245ff.


1885     „The Normal and the
In: Wolfgang, et al., 1970: llff.

Robert O.

1969     Sentencing: The Decision as to Type,
Length, and Conditions of Sentence. Boston: Little, Brown

Philip H.

1967     „Crime, Victims, and the Police“.
In: Wolfgang, et al. 1970: 74ff.

Philip H.

Criminal Victimization in the United States: A Report of a National Survey.
Washington, D. C.: U.S. Government Printing Office

Kai T.

1962     „Notes on the Sociology of Deviance.” In: Becker, 1964: 9ff.


1966     Wayward Puritans. A Study in the Sociology
of Deviance. New York: Wiley


1967     „Graduate School Games. “ Transactional
Analysis Bulletin 6 (1967): 103ff.

Robert E. L., editor

1964     Handbook of Modern Sociology. Chicago: Rand

FESTINGER, Leon, Henry W. Rieken, and
Stanley Schachter

1956     When Prophecy Fails. Minneapolis:
University of Minnesota Press

Milton and Leonard J. Savage

1948     „The Utility Analysis of Choices
Involving Risk.” The Journal of Political Economy LVI (1948): 249ff.


1967     Studies in Ethnomethodology. Englewood
Cliffs, N. J.: Prentice-Hall


1885     Criminology. Boston: Little, Brown. 1914.
(first published in Italian)

H. and C. Wright Mills, editors

1958     From Max Weber: Essays in Sociology. New
York: Oxford University Press

Alvin W.

1970     The Coming Crisis of Western Sociology. New
York: Avon

Alvin W.

1973     „Foreword to: The New Criminology.”
In: Taylor, et al., 1973: ix ff.

Nils H.

„Friedman-Savage Utility Functions Consistent with Risk Aversion.” The
Quarterly Journal of Economics LXXXIV (1970): 472ff.

Philip E., editor

1967     Sociologists at Work: Essays on the Craft
of Social Research. Garden City, N. Y.: Anchor

John C.

„Approaches to the Bargaining Problem before and after the Theory of
Games: A Critical Discussion of Zeuthen’s, Hicks’, and Nash’s Theories.”
Econometrica 24 (1956): 144ff.

John C.

1961     „On the Rationality Postulates
Underlying the Theory of Cooperative Games.” Journal of Conflict Resolution 5
(1961): 179ff.

John C.

1962     „Measurement of Social Power,
Opportunity Costs, and the Theory of Two-Person Bargaining Games.” Behavioral
Science 7 (1962): 67ff.

John C.

1966     „A Bargaining Model for Social Status
in Informal Groups and Formal Organizations.” Behavioral Science 11 (1966):

John C.

1968     „Individualistic and Functionalistic
Explanations in the Light of Game Theory: The Example of Social Status“.
In: Problems in the Philosophy of Science. Amsterdam: North-Holland, 1968:

George Caspar

1950     The Human Group. New York: Harcourt, Brace

George Caspar

1961     Social Behavior: Its Elementary Forms. New
York: Harcourt, Brace & World

C. I., I. L. Janis, and H. H. Kelley

1953     Communication and Persuasion. New Haven:
Yale University Press

C. I., W. Mandell, Enid H. Campbell, T. Brook. A. S. Luchins, A. R. Cohen, W.
J. McGuire, I. L. Janis, Rosalind L. Feierabend, and N. H. Anderson

1957     The Order of Presentation in Persuasion.
New Haven: Yale University Press

L., C. I. Hovland, P. B. Field, Harriet Linton, Elaine Graham,

A. R.
Cohen, D. Rife, R. R. Abelson, G. S. Lesser and B. T. King

1959     Personality and Persuasibility. New Haven:
Yale University


Helen Hall

1950     Leadership and Isolation. (2nd ed.) New
York: Longmans, Green


1966     The Uses of the University. New York:
Harper and Row

Thomas S.

1962     The Structure of Scientific Revolutions. Chicago:
University of Chicago Press

Wayne R.

1965     Arrest: The Decision to Take a Suspect into
Custody. Boston: Little, Brown

Edwin M.

1951     Social Pathology. A Systematic Approach to
the Theory of Sociopathic Behavior. New York: McGraw – Hill

Edwin M.

1970     Social Action and Legal Change. Revolution
within the Juvenile Court. Chicago: Aldine

Duncan and Howard Raiffa

1957     Games and Decisions. New York: Wiley

LUHMANN, Nikolas

1967     „Soziologische
Aufklarung“. Soziale Welt 18 (1967): 97ff.


„The Role of the Criminologist in Society“. Center of Criminology,
University of Toronto, Contribution to Round Table Discussion, 14th Annual
Congress of the Quebec Society of Criminology, Montreal, March 15, 1973


1952     „The Utility of Wealth“. The
Journal of Political Economy 60 (1952): 151ff.

Robert K.

1962     „Foreword to: The Sociology of
Science“. In: Barber and Hirsch, 1962

Robert K.

1968     Social Theory and Social Structure. (1968
enlarged edition) New York: The Free Press

Frank W.

1969     Prosecution: The Decision to Charge a
Suspect with a Crime. Boston: Little, Brown


1959     The Sociological Imagination. London:
Oxford University Press


1963     Galgenlieder; a Selection. Translated, with
an Introduction by Max Knight. Berkeley: University of California Press

NASH, John
F., Jr.

1950     „The Bargaining Problem“.
Econometrica 18 (1950): 155ff.

NASH, John
F., Jr.

1953     „Two-Person Cooperative Games“.
Econometrica 21 (1953): 128ff.

NEUMANN, John von, and Oskar Morgenstern

1944     Theory of Games and Economic Behavior.
Princeton: Princeton University Press

Theodore M.

1961     The Acquaintance Process. New York: Holt,
Rinehart and Winston

Donald J.

1966     Conviction: The Determination of Guilt or
Innocence without Trial. Boston: Little, Brown

Robert A.

1970     The Social Bond. An Introduction to the
Study of Society. New York: Knopf

Lynne and Richard Ofshe

1970     Utility and Choice in Social Interaction.
Englewood Cliffs, N. J.: Prentice-Hall

Richard, editor

1970a    The Sociology of the Possible. Englewood
Cliffs, N. J.: Prentice -Hall


1935     The Mind and Society (Trattato di
Sociologia Generale). New York: Harcourt, Brace

Karl R.

1935     Logik
der Forschung. Zur Erkenntnistheorie der modernen Naturwissenschaft.

Karl R.

1952     The Open Society and Its Enemies. (2nd ed.)
London: Routledge & Kegan Paul


1970     The Social Reality of Crime. Boston:
Little, Brown


1970a The
Problem of Crime. New York: Dodd, Mead


„Dialogue with Richard Quinney. An Interview Conducted and Prepared by
Eileen Goldwyn“. Issues in Criminology 6 (Summer 1971): 41ff.

Anatol and Albert M. Chammah

1965     Prisoners‘ Dilemma. Ann Arbor: University
of Michigan Press


1938     Experience and Prediction. An Analysis of
the Foundations and the Structure of Knowledge. Chicago: University of Chicago


1947     Elements of Symbolic Logic. New York:

Albert J., Jr., and David J. Bordua

1967     „Environment and Organization: A
Perspective on the Police“. In: Bordua, 1967: 25ff.

Albert J., Jr.

1971     The Police and the Public. New Haven: Yale
University Press

Frank J.

1969     „Editor’s Foreword“. In: Dawson,
1969: xvff.

M. J. , C. I. Hovland, W. J. McGuire, R. P. Abelson, and J. W. Brehm

1960     Attitude Organization and Change. New
Haven: Yale University Press


1966     Experimental Effects in Behavioral
Research. New York: Appleton-Century-Crofts


1971     „Value-Neutrality in Germany and in
the United States“. In: Bendix and Roth, 1971: 34ff.

Leonard J.

1954     The Foundations of Statistics. New York:

Thomas C.

1963     The Strategy of Conflict. Cambridge, Mass.:
Harvard University Press


1953     „Concept and Theory Formation in the
Social Sciences“. In: Schutz, 1962: 48ff.


1962     Collected Papers I: The Problem of Social
Reality. The Hague: Martinus Nijhoff

SCHWENDINGERS, Herman and Julia R.

1970     „Defenders of Order or Guardians of
Human Rights?” Issues in Criminology 5 (1970): 123ff.

SCHWENDINGERS, Herman and Julia R.

1972     „The Continuing Debate on the
Legalistic Approach to the Definition of Crime“. Issues in Criminology 7
(1972): 71ff.


1938     „A Sociological Approach“. In:
Wolfgang, et al., 1970: 3ff.


„The Demand for Order in Civil Society: A Review of Some Themes in the
History of Urban Crime, Police, and Riot“. In: Bordua, 1967: 1ff.

Herbert A.

1957     Models of Man: Social and Rational.
Mathematical Essays

Rational Human Behavior in a Social Setting. New York:


Jerome H.

1966     Justice without Trial. Law Enforcement in
Democratic Society. New York: Wiley


1963     „Institutions of Privacy in the
Determination of Police Administrative Practice“. American Journal of
Sociology 69 (1963): 150ff.

Edwin H.

1945     „Is ‚ White Collar Crime‘ Crime?” In: Wolfgang, et al. , 1970: 32ff.

Edwin H. and Donald R. Cressey

1970     Criminology (eighth edition). Philadelphia:


1964     „Social Differentiation“. In:
Faris, 1964: 5 3 Off.

Paul W.

1947     „Who is the Criminal? “ In: Wolfgang, et al. , 1970: 41ff.

REPORT: Science and Technology, President’s Commission on Law Enforcement and
Administration of Justice, Washington, D. C.

1967     „Measuring Crime“. In: Wolfgang,
et al., 1970: 101ff.

Ian, Paul Walton and Jock Young

1973     The New Criminology: For a Social Theory of
Deviance. London: Routledge & Kegan Paul

John W. and Harold H. Kelley

1959     The Social Psychology of Groups. New York:

William I., and Dorothy Swaine Thomas

1928     The Child in America. Behavior Problems and
Programs. New York: Knopf

Lawrence P., Donald M. McIntyre, Jr., and Daniel L. Rotenberg

1967     Detection of Crime. Boston: Little, Brown

Austin T.

1969     Criminality and Legal Order. Chicago: Rand

Lloyd, Marchia Meeker, and Kenneth Eells

1949     Social Class in America. A Manual of
Procedure for the Measurement of Social Status. Chicago: Science Research


1918     „Science as a Vocation“. In:
Gerth and Mills, 1958: 129ff.


1923     Gesammelte
Aufsatze zur Wissenschaftslehre. Tübingen: Mohr (Siebeck), 1951


1967     Mathematical Programming and the Analysis
of Capital Budgeting Problems. Chicago: Markham

William Foote

1943     Street Corner Society. The Social Structure
of an Italian Slum (2nd ed., 1955), Chicago: University of Chicago Press

James Q.

Varieties of Police Behavior. The Management of Law and Order in Eight
Communities. Cambridge, Mass.: Harvard University Press

Marvin E., Leonard Savitz, and Norman Johnston

1970     The Sociology of Crime and Delinquency (2nd
ed.). New York: Wiley


1930     Problems of Monopoly and Economic Warfare. London: G. Routledge