Classifying Natural and Social Theories (Szostak, 2003)

Current Sociology
http://csi.sagepub.com
Classifying Natural and Social Scientific Theories

Rick Szostak
Current Sociology 2003; 51; 27
DOI: 10.1177/0011392103051001779
The online version of this article can be found at:
http://csi.sagepub.com/cgi/content/abstract/51/1/27
Published by:
http://www.sagepublications.com
On behalf of:
International Sociological Association
Additional services and information for Current Sociology can be found at:
Email Alerts: http://csi.sagepub.com/cgi/alerts
Subscriptions: http://csi.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Downloaded from http://csi.sagepub.com by Alberto Conde-Flores on January 23, 2008

2003 International Sociological Association. All rights reserved. Not for commercial use or unauthorized distribution.
03 Szostak (to/d)
12/17/02
1:05 PM
Page 27
Rick Szostak
Classifying Natural and Social

Scientific Theories
lassification has been said to be the essential first step in science. We

cannot analyse chemical reactions without first distinguishing one
chemical element from another, comprehend atomic reactions without some
sense of subatomic particles, nor identify the characteristics associated with
creativity without some sense of personality dimensions. Yet we seldom turn
our classificatory zeal towards the study of science itself.
In this article, a simple classification of scientific theories is developed,
and it is shown how theories in natural and especially social science can be
placed within this classification. The insights that can be gained from such an
endeavour are then discussed. One of these involves guiding us to a more
precise definition of what a scientific theory is. We, in the meantime, take a
pragmatic approach and treat as theories what is generally thought to be a
theory. In the larger research project of which this article forms a part, classifications of scientific methods, the phenomena that scientists study and
(critiques of) scientific practice, have been developed. These classifications in
combination allow an exhaustive typology of types of science.
A Simple Classificatory Guideline

One attribute to be sought in any classification system is that it aid recall: we
should ideally be able to remember diverse classes and where they fit in the
broader scheme. This article uses perhaps the simplest of all classificatory
guidelines: it asks of any theory the 5W questions: Who, What, Where, When
and Why.
In addition to the 5W questions, scientists often ask How?. But
philosophers often distinguish scientific approaches as to whether they
are focused on Why? or How?. This suggests that How? is actually a
Current Sociology, January 2003, Vol. 51(1): 2749 SAGE Publications
(London, Thousand Oaks, CA and New Delhi)
[00113921(200301)51:1;2749;031779]

03 Szostak (to/d)
28
12/17/02
1:05 PM
Page 28
Current Sociology Vol. 51 No. 1
combination of (some of) the other four Ws. When interviewers ask a How?
question, they are generally rewarded with a chronology of When and Where
a process occurred (Goldenberger, 1992: 118). Nevertheless, we briefly
consider in the following whether answers to How? have been adequately
subsumed within the 5W analysis.
Who?
Most theories deal with how one or more phenomena influence one or more
others. Less commonly, but importantly, theories discuss the nature of, or
internally generated transformations in, one phenomenon. In either case,
scholars investigate how changes, perhaps just relative to some hypothesized
alternative, in one (set of) phenomenon affect either itself or other phenomena. In asking Who?, then, we are asking Who is effecting change?. Any
theory must grapple with agency at this level. If we say racist attitudes affect
X, we will have a vague theory indeed unless we specify what type of agent
holds such attitudes. If we speak of the effects of aggressive behaviour, we
should specify whose behaviour. We must also, in most instances, concern
ourselves with the type of agent that responds to the initiating change.
Four broad categories of agency can be identified:
Non-Intentional Agent This is, in a sense, the easiest case to deal with. In
studying the effects of tornadoes on social cohesion, the question of why tornadoes choose to strike a particular locale can be ignored. Tornadoes can be
treated as exogenous, or the reasons for their spatial and temporal distribution explored. Along with natural phenomena, various non-intentional
societal aggregates such as level of income or population distribution or
gender divisions exert influences on diverse phenomena. While the scientific community as a whole should be curious about the generation of these
phenomena, and intentionality obviously plays some role here, particular
scientists may well wish to posit theories focused on only their effects,
though there may again be intentionality in how people react.
Two types of non-intentional agency are particularly important.
Though institutions, the formal rules by which our societies are governed,
are put in place intentionally (to at least some degree), they are intended
thereafter to place constraints on human behaviour. An especially important type of institution is the organization: a formally constituted body of
people that collectively pursues particular goals. While the intentional acts
of those involved in organizations are an important influence on their behaviour, if this was the whole story people would not bother to form organizations in the first place.
Individual Agents Theories here commonly focus on a typical individual.
They may also explore how different types of individual will act in a

03 Szostak (to/d)
12/17/02
1:05 PM
Page 29
Szostak: Classifying Natural and Social Scientific Theories
29
particular situation. Theories may emphasize the effects of individual acts,

though these theories should engage the possibility of feedback effects, such
that the individuals actions or beliefs are themselves conditioned by the
effects these are deemed likely to have. To understand why an individual
behaved or believed as they did, it is necessary to look within the individual.
Various theories thus look at genetic tendencies, abilities, personality dimensions and so on. While theories of individual agency must cope with intentionality, they can embrace both subconscious motives of which the agent is
consciously unaware and also accidents in which neither the agents conscious nor subconscious minds imagined the outcome.
Groups of Individuals Talcott Parsons hypothesizes four types of social
group worthy of scientific analysis (see Freidheim, 1982). The first is the
primary group or family. While some theories may reduce familial behaviour
to some weighted sum of the actions and wishes of individual family
members, many theories posit that families are not just the sum of their parts.
Parsonss second type is the organization. As hinted earlier, an understanding of organizations may require a mix of non-intentional and intentional
theorizing; again, organizations are not just the sum of their parts. Parsons
also speaks of community and society. The first of these refers to groups
operating in a non-formal fashion, as when a group develops clear cultural
guidelines. In the study of society, which operates through formal institutions, intentional analysis must be merged with non-intentional insights.
Relationships Some interactionist scholars argue that what happens
between people is more important than what happens within them: individuals are reconstituted through their interaction with others (Barnes, 1995).
Interactionists thus theorize about the effects that various types of human
relationship have. While there are overlaps with the study of groups, the
focus is on personal interactions rather than group processes.
Many scholars would wish to simplify this classification by arguing that only
some of these types of agency are legitimate. This is generally done by arguing
that only individuals or, alternatively, societal aggregates play an independent
causal role.1 The goal of this article is to classify scientific theories, not pass
judgement on their relative utility. Moreover, it is quite possible for both individuals and society to be important. Individuals may generate organizations,
cultural attitudes and technologies that have effects quite different from any
they imagined. In turn, while individuals operate within a complex web of
societal and interpersonal influences, we each have personalities and abilities
that influence our actions and beliefs. While individual scholars may emphasize one sort of agency within particular research projects, they need not
decree that the entire scientific enterprise should do likewise.

03 Szostak (to/d)
30
12/17/02
1:05 PM
Page 30
What?
Having identified the agents of change, a theory must then cope with the
question of What do change agents do?. In the case of non-intentional
agency, agents usually cannot actively do anything, though tornadoes provide
an obvious exception. They are thus generally restricted to a passive form of
action wherein they provide constraints and/or incentives. It should be
emphasized that these effects can be of huge importance despite their passive
nature.
Individuals and especially groups can also act passively. They also share
with natural phenomena such as tornadoes the capacity to act actively: to do
something in common parlance. They can also act in a way that by their
nature non-intentional agents cannot: they can form beliefs or attitudes or
intentions (which in turn influence actions). And these thoughts can be
analysed at the level of individuals and at the level of groups; in the latter case
we can speak of culture or public opinion or ideology or nationalism.
Relationships are often posited to affect mostly attitudes, but can also induce
actions and provide constraints.2
In the literature, a focus on actions is often associated with a positivistic approach to science that emphasizes the analysis of only that which we
can observe, and the focus on attitudes with an interpretive perspective that
emphasizes how agents interpret their situations. The critical role of passive
action is largely ignored in this dichotomy, at least explicitly, though advocates of both approaches may examine various constraints and incentives. As
in almost every scientific debate among opposites, there are legions of positivist scientists and philosophers who condemn the practice of interpretive
science, and legions of interpretivists who argue the opposite.3 Again, the
purpose here is to classify all theories, not seek to exclude some from consideration. I would, though, join Potter (2000) in hoping that this longstanding debate is lessening in the face of recent developments, including the
emergence of critical realist philosophy which advocates a mix of interpretivist and positivist approaches.4
Why?
Why did the agent(s) act (react) as they did? In the case of a non-intentional
agent, analysis must proceed in terms of the constraints and incentives
inherent in its nature. In the case of intentional agents, the decision-making
process involved must be investigated. There are five ways in which agents
might make any decision.
Rational/Consequentialist Agents focus on desirable consequences, and try
to rationally calculate the best way to achieve those goals. Their ability to do
so depends on their ability to accurately perceive how the world works.
Agents may be selfish or altruistic, and focused on a diverse set of basic drives

03 Szostak (to/d)
12/17/02
1:05 PM
Page 31
31
or personal desires, and these latter may reflect sociocultural influences to

varying degrees.
Intuitive Agents act on what their gut tells them, without much conscious
thought. Their intuition may guide them towards satisfying various drives. It
may act upon various subconscious schemas about how the world works,
the accuracy of which depends on the agents lifetime experience, and particularly of traumatic events. It may thus guide agents towards acts they would
not consciously countenance.
Traditional/Cultural Agents are guided to do what is done around here.
This can at times occur intuitively, if they have internalized their groups attitudes and will feel guilt for breaking them, but generally has some conscious
element.
Rules/Norms Agents follow a set of rules or norms, which they have previously determined to be socially desirable. These may or not accord with
cultural guidelines. Agents have devoted some thought to them (which in
turn may be influenced by any of the five decision-making processes).
Common examples are the Golden Rule, Kantian Imperative, or a belief in
Rights. But a host of more specific rules (regarding, say, how to treat
strangers) are also possible.
Process/Virtues Agents again act consciously, but may be concerned more
with how they act (the process) than what they achieve. Agents will generally evaluate their behaviour in terms of what are commonly called virtues.
They may for example be determined to behave honestly, or responsibly, or
punctually.
These decision-making methods are complementary: in judging what is
responsible behaviour, one may have recourse to cultural guidelines, evaluation of likely consequences, the Golden Rule, and/or gut feelings. Theorists
focusing on any one of them should be attentive to (and ideally speak to) the
possibility of other influences. The five-way classification system captures a
variety of distinctions made in the theoretical literature, such as conscious vs
subconscious, and process vs outcome-oriented. Most importantly, it has
room for both individual-level and societal-level influences. And while each
of the five broad types can be disaggregated, each is nevertheless characterized by a well-defined decision rule.
With respect to group or relationship agency there are additional
concerns. Is decision-making democratic, or do some individuals exert disproportionate influence? Are there formal or informal guidelines for decisionmaking? Is decision-making consensual, or are there dissatisfied minorities?

03 Szostak (to/d)
32
12/17/02
1:05 PM
Page 32
Where?
Where does whatever the agent does (and the reactions to this) occur? This
question can be interpreted both literally and figuratively. Literally, one can
wonder where in the real world it happens. There are two broad possibilities: that it can happen anywhere, or that it can only happen in certain
specified situations. Figuratively, one can think of a map of the potential set
of links among the diverse phenomena studied by scientists (see Szostak,
2000, 2003) and ask where on that map it occurs. Again, a theory that speaks
to a very small set of links can be distinguished from one that implies that the
same effect is felt along a wide array of links. Since different sorts of phenomena exhibit different sorts of agency, no theory could posit that any one sort
of agency was directly reflected along all links; other links might nevertheless encompass reactions to the particular type of agency being studied.
Both in the literal and figurative cases, a search for general regularities
can be distinguished from a search for particularities. Philosophers have
since the turn of the last century used the words nomothetic and idiographic to refer to these two types of analysis. There has been a fair bit of
confusion about what exactly these terms mean, what it is that they refer to,
and where the boundary lies between the two. The apparent answer to the
second query is that they represent different sorts of answer to the question
of Where? the relationship under examination occurs.
As for the first query, literal and figurative readings of the question
provide different interpretations. The literal interpretation refers to realizations of other phenomena. That is, if we say that it can only happen in
home or office, we need to carefully specify what aspects of relationships,
family structure, occupation, or several other phenomena are in fact
crucial. Nomothetic theory can thus be defined as that which posits a general
relationship among two or more phenomena, while idiographic theory is that
which posits a relationship only under specified conditions.
The figurative reading refers instead to links among phenomena. A
nomothetic theory is one that suggests a relationship that operates along
many links. Idiographic theory refers instead to a relationship that affects a
small number of links: a perfectly idiographic theory would embrace only
one link.
Defining a boundary between nomothetic and idiographic theory is
problematic in both cases. There is instead a continuum: in the first case
running from specifying no external conditions to specifying these in terms
of all other phenomena, in the second case running from application to one
set of phenomena to application to all sets. Fortunately, well-defined boundaries are not essential for a classification to be useful. Nomothetic and idiographic can be helpful classes, even if every theory cannot be flawlessly placed
into one class or another. Theories of both sorts should struggle to be as
explicit as possible about their breadth of applicability. Indeed, this is one of

03 Szostak (to/d)
12/17/02
1:05 PM
Page 33
33
the primary goals of science, and one area in which natural science excels over
social science: social scientists can be quite careless about specifying where
their theory will hold and where it will not.
Answers to the What and Why questions were related to answers to
the Who question in the earlier section. Do nomothetic and idiographic
theories differ in importance depending on who the change agent is? If nomothetic theory is defined in terms of covering laws, the answer would surely
be yes. For as Little (1998) forcefully argues, covering laws are best suited to
non-intentionality. A covering law stipulates a necessary relationship: if A
then B. This can be observed in the natural world, as in the Laws of Thermodynamics. But if human agency is involved anywhere in the process, there will
almost certainly be exceptions to any rule, and thus it can only be said that
A usually or almost always results in B. But such a regularity would still
qualify as nomothetic under any but the most narrow of definitions. Alternatively, Bunge (1998) notes that there are idiosyncrasies even among atoms,
though much less than among complex organisms. There is thus a place for
both types of theory across types of agent, though the scope for idiographic
analysis will tend to increase with the role of intentional agency.
When?
When does it happen? Interpreted literally, this question can yield the
same answer as the Where question, for it would guide scholars to look at
what conjunction of realizations of other phenomena must be in place for a
given relationship to hold.5 Instead the time path of the process needs to be
investigated. There are four broad possibilities:
There are negative feedback effects such that the original impulse is
exactly negated and the system of phenomena as a whole is unchanged.
The original change induces changes in a variety of other phenomena, but
the system of phenomena moves towards a new equilibrium where the
system will only change in response to further shocks.
There are positive feedback effects such that certain phenomena continue
to change in a particular direction.
The effects are stochastic, and thus quite different outcomes are possible.
We could at the level of a comprehensive cluster of theories which encompassed all phenomena suggest that some at least of these types of theory must
be misguided: a world of positive feedbacks would spin out of control, a
theory which posits no change seems inherently unrealistic, and even a series
of equilibria may seem an unrealistic description of reality. But individual
theories almost inevitably deal with a limited system of phenomena. Even at
that level certain types of theory might seem suspicious. Indeed, all but the
last time path can arguably only be posited by ignoring links between the
system in question and all other phenomena. Still, all types of time path can

03 Szostak (to/d)
34
12/17/02
1:05 PM
Page 34
likely shed some light on particular questions. We should, yet again, be open
to insights from more than one type of theory.
Theorists should carefully specify which type of time path they envision.
Freidheim (1982), for example, discusses eight group processes. Some of
these, such as socialization, coordination and the response to individual deviations, appear to fit best within the first time path, but only if it is thought
that these processes are successful. Conversely, differentiation within the
group and stratification may be essential to maintaining the group as is, or
could lead to group fragmentation, which may in turn be perceived as a new
equilibrium, or part of a dynamic or even chaotic process. Group formation,
innovation and social change are processes that would fit poorly within the
first time path, but might characterize any of the remaining three.
While theorists should specify which of the four types of time path they
intend, they should not be forbidden from appealing to combinations of
these. A theorist might think that one of the first three time paths is the likely
outcome, but admit that the more uncertain fourth type is a possibility. Or
they might imagine circumstances in which two or more time paths are
possible. Of course, they should strive to identify the circumstances in
question.6
How?
As noted earlier, the How? question can be used as a check on the exhaustiveness of our 5W questions. How does something happen? In answering
such a question, scholars would wonder who instigated the change, what they
did, and perhaps why they did what they did. They would also wonder what
other conditions had to be in place for it to have happened as it did. And
how precisely did it occur through time and space? All elements of the
How? question have been captured in asking our 5W questions.
Types of Theory and the Five Types of Causation/Influence

Szostak (2003) argues that all links between phenomena can be understood
in terms of five types of causation or influence: strict causation, as when
sodium and chlorine react; intentional causation, as when a person or group
acts purposefully; hermeneutic causation, when meaning is transmitted;
functional/structural causation, when a particular realization of one
phenomenon is necessitated by a particular realization of another phenomenon; and evolutionary causation, involving some transmission mechanism
and selection criteria. Szostak (2003) notes that the relationship between these
is unclear: some but not all scholars treat intentional as a subset of strict
causation, or hermeneutic as a subset of intentional, and some question
whether functional explanations are not reducible to intentional arguments.

03 Szostak (to/d)
12/17/02
1:05 PM
Page 35
35
The 5W typology of theory allows these five types of causation or influence to be identified much more precisely. Strict causation applies to nonintentional agents. Intentional and hermeneutic causation apply to
intentional agents: the former deals with their actions, the latter with their
thoughts. The first three types of causation, then, can be defined in terms of
answers to the Who and What questions. It would, in this formulation,
clearly be a mistake to view any one of these as a subset of the others.
Elster (1989) and others have worried that it is too easy to assume functional relationships. The theorist must go a step further, and ask how the
functional relationship emerges: how, for example, do intentional agents of
the state establish an army? The question is thus raised of whether the only
sensible part of a functional relationship is the intentional part. But why is an
army necessary for the maintenance of a state? This can only be so if the army
creates appropriate constraints and incentives. Functional arguments, then,
must combine the passive action of a non-intentional agent with the intentional acts of intentional agents.
What sorts of agents evolve? The 5W typology suggests this unusual
question. And the answer is that only non-intentional agents evolve. Individuals and groups do not evolve, though elements of these, such as genes and
beliefs, may evolve. Evolutionary theories focus not on what particular nonintentional agents do but on how they become what they are. The analysis
proceeds in terms of the inherent nature of the agent itself (the possibilities
for mutation) and the effects of a selection environment that comprises other
agents. These in turn may be non-intentional, as when genetic evolution is
conditioned by climatic change, or intentional, as when technologies are
selected by individuals or groups. Since a theory of evolution should specify
the relevant selection environment, evolutionary causation may involve
intentional as well as non-intentional arguments.
Classifying Individual Theories (see Table 1)

We can ask of any theory five questions: Who are the agents? What sort of
action is involved? What decision-making process is employed? How generalizable is the theory (literally and figuratively)? What sort of time path is
followed? Most theories in natural science outside biological science involve
non-intentional agents, passive reaction and no active decision-making
process. They can be distinguished in terms of where they occur (in both
senses of where) and what sort of time path is involved. Evolutionary theory
in biology, as discussed earlier, also involves non-intentional agency and no
active decision-making. Here, though, there is non-passive action, for
mutation is an inherent quality of the organisms under study. Evolution is
thought to apply to all organisms. While nominally nomothetic, evolutionary

Symbolic
interactionist
Rational choice
Phenomenology
When?
Time Path
Where?
Generalizability
Non-intentional
Passive
Inherent
Various
Various
Non-intentional
Active
Inherent
Nomothetic
Individual
(group)
Individual
(relationship)
Various
Active
Various
Action
(attitude)
Action and
attitude
Individual
(look within)
Relationships
emphasized
Individual
Relationships
(individuals)
Attitudes
Various;
often rational
Various;
emphasizes
constraints
Intuition;
others possible
Various
Not same
equilibrium
Not same
(any) equilibrium
Various
Attitudes
Action
Attitudes
(actions)
Rational
Various
New
equilibrium
Various
Stochastic
Usually equilibrium
Various
Nomothetic
Generally
idiographic
Generally
nomothetic
Implicitly
nomothetic
Idiographic;
some generalizing
Nomothetic
Various
Page 36
Psychoanalytic
Why?
Decision-Making
Process
1:05 PM
Systems theory;
functionalist
What?
Action
12/17/02
Most natural
science
Evolutionary
biology
Evolutionary
social science
Action theory
Who?
Agency
Theory
Type
03 Szostak (to/d)
36

Table 1 Typology of Selected Theories
03 Szostak (to/d)
12/17/02
1:05 PM
Page 37
37
theory has idiographic elements in that every mutation and selection environment is unique. Evolution is a dynamic process, with organisms changing
through time (though in unpredictable directions); change likely occurs
slowly enough that some semblance of ecosystem equilibrium is achieved.
Evolutionary theory can also be applied to various elements of social science,
notably culture, institutions, technology and art. Intentional agency thus
becomes important, and considerations of decision-making more complex.
As well, mutations are much more frequent, and thus equilibrium a less likely
outcome.
Theories in social science are much more diverse than those in natural
science. Indeed, some theory in social science could likely be found that
would fit every possible combination of agent, action, decision-making
process and time path (though that is not shown here). With respect to the
Where? question, though, there are as many possibilities as there are links
among phenomena (a finite but huge number). It is thus unlikely that every
type of theory has been applied to every possible link. The typology can then
guide us to ask new questions. Note in this regard that in investigating a link
XY, it is useful to know something about WX and YZ, and thus understanding of a link that involves one sort of agency can be enhanced through
appeals to theory types focused upon other sorts of agency.
Given the infinite number of possible theories, this article could not place
every social science theory within this five-dimensional typology. But a selection of major theories can be classified. To make the task both more manageable and less subject to authorial bias, the article engages those theories
surveyed in Turner (2000) and the analyses contained therein.7
Action Theory (Including Theories of Praxis)
Who? The focus of these theories is on individual-level behaviour, though
the analysis of interpersonal relationships has come to be emphasized by
many recent action theorists. As Cohen (2000) suggests, this could easily be
viewed as a different type of theory.
What? The very title of action theory would lead us to suspect that this
theory is primarily concerned with action. Yet in distinguishing theories of
action from related theories of praxis, Cohen (2000) identifies the latter as
concerned with what we do when we act and the former as analysing the
subjective meaning of our actions. Action theory proper is concerned with
the intersection of attitude and action. And, as Cohen notes with respect to
Talcott Parsons, sometimes meaning is emphasized so much that little is said
about acts themselves.
Why? Some action theorists emphasize utilitarian calculations, while others
emphasize social meaning, culture and personal passions. Utilitarian

03 Szostak (to/d)
38
12/17/02
1:05 PM
Page 38
calculations are, of course, the most common type of consequentialist

analysis. Personal passions in turn suggest a role for intuition. And social
meaning and culture imply an appeal to tradition, but also raise the possibility of appeals to virtues or rules.8 Weber emphasizes the importance of
value-oriented decision-making where one strives to achieve a value. Action
theory, then, neither dictates nor excludes any single decision-making
process, though individual examples of action theory generally do. Indeed,
Cohen (2000: 74) applauds the recognition that there is no one right answer
here (but is critical of the fact that action theorists have ignored certain
internal and external influences on behaviour, such as power relations).
Theories of praxis discuss how individuals both receive and transmit
signals that allow them to interact habitually. By recognizing the importance
of habit, praxis theorists open the door to passive reaction. But theories of
praxis realize and in general applaud the exercise in free will inherent in the
fact that people do not always act out of habit. While the decision-making
process may be subconscious, and the outcome imperfectly predictable, some
scope for decision-making remains.
Where? Very little mention is made of either the literal or figurative spatial
dimension in Cohen (2000). Giddens is hailed for recognizing that spatial
context matters: it is hard to talk on a crowded bus, for example. But less
attention is paid to what sorts of actions or decision-making might occur in
some places, or between certain participants or phenomena, but not others.
This in turn reflects the tendency of theorists in this tradition to speak as if
one answer applied to all situations, rather than seek the limits of particular
theories. Cohen (2000) reasonably suggests that different types of action
theory likely apply to different situations.9
When? Talcott Parsons posits that social order results from individuals
pursuing integrated ends: there is some process through which individuals
choose largely complementary goals, and then act to achieve these. This is
clearly an equilibrium process. Giddens too implies some sort of equilibrium
process in maintaining that practices support structural patterns. But Giddens
like Dewey celebrates those times at which individuals break free from habit,
and recognizes that in doing so they may encourage cultural change. Giddens
and Dewey thus envisage a more dynamic process than Parsons: pursuit of
habit may maintain social stability, but individuals making conscious decisions
will induce societal change. If these acts are rare, the system may move
between equilibria; otherwise it may be constantly changing.
Comments While Who? and What? are fairly well specified, action
theories embrace a very wide range of answers to the other three questions.
It is perhaps not surprising that many action theorists believe that it is the

03 Szostak (to/d)
12/17/02
1:05 PM
Page 39
39
emphasis on praxis or meaning that is central to their theory, but many

especially those who emphasize relationships instead argue that the What?
is quite secondary to the Who? (see Cohen, 2000; those who emphasize
relationships thus borrow pragmatically from theories of action and praxis).
The flexibility with respect to Where? has not led theorists to identify complementary ranges of applicability for different theories within this family, or
encouraged a general appreciation for theories outside the family.10 The
typology developed here would encourage the identification of range of
applicability.
Systems Theory and Functionalism
What? Lechner (2000) describes the core element of systems theory as a
recognition that patterns in social life are not just accidents. When people act,
they are enmeshed in relationships, institutions and/or societies that maintain
a distinct unity.11 While systems theories must thus deal with actions, these
are generally seen as guided by attitudes.
Why? The emphasis in systems theory is on constraints that encourage individuals to act in complementary ways. There is, though, considerable room
for disagreement over the form that these constraints take. Talcott Parsons
argued that American capitalism depended upon cultural values which supported economic activity but also a sense of community. Many were sceptical of his emphasis on actors and actions as if individuals can simply carry
structure with them (Lechner, 2000). Sceptics doubted that internalized
norms alone could maintain a system, and emphasized instead such phenomena as institutional constraints.
Where? Talcott Parsons thought of society as composed of subsystems
which served different purposes in the maintenance of societal stability.
Within this view, all phenomena were part of one and only one subsystem.
Other systems theorists have been less inclusive, suggesting that certain
phenomena or links may serve only to effect change in systems of which they
are not part. And still others have emphasized systems that may embrace only
a handful of phenomena. Thus, while systems theory is often (figuratively)
all-encompassing, it need not be so.
Systems theorists often use a language that suggests that the systems they
discuss are (literally) universal. Parsonss treatment of American capitalism,
however, stressed the uniqueness of that system. The common emphasis on
culture in systems theory suggests that a system observed in one society
cannot function in a society with quite different values.
When? The very idea of a system implies that multiple causal links interact
in such a way as to ensure system stability. This does not necessarily mean

03 Szostak (to/d)
40
12/17/02
1:05 PM
Page 40
that the system as a whole is unchanging. Some theorists explicitly recognize

that external forces may force change. A few note that dynamics internal to
the system might also generate change. Systems theory can be characterized,
then, as involving movement between equilibria. In the hands of some
theorists, external shocks are downplayed such that it seems that a stable
equilibrium is being examined. In the hands of others, either external shocks
or internal dynamics may be such that some sort of dynamic change is
implied. From the time of Herbert Spencer in the 19th century, some theorists
have spoken of systemic evolution through a process of differentiation and
integration (this is also a key component of Boynes [2000] discussion of
structuralism). Yet while systems theorists can recognize conflict and change,
they will agree that change can only be appreciated after the system has been
identified (Lechner, 2000).
Who? This question is hardly engaged by systems theory. Parsons emphasized individuals (though also relationships), but he was in turn criticized for
downplaying institutions. Since systems can potentially embrace any
phenomenon, systems theorists could potentially deal with every type of
agency. Arguably, a common flaw in systems theories is that one or two types
of agent are emphasized, even when the system clearly embraces others. If
these others are simply reactive, this should be clearly stated and justified
within the theory.
Comments As with action theory, systems theory is well specified in terms
of a couple of our 5W questions, and poorly specified with respect to others.
Interestingly, many scholars have questioned whether systems theory is
really a theory or just a frame of reference. We can join such critics in urging
greater attention to the Who? and What? questions. It is noteworthy,
though, that similar denigration has not occurred with respect to theories that
specify Who? and What? but not When? and Where?.
Psychoanalytic Theory
Who? Psychoanalytic theory looks within the individual for desires, fantasies and repressed feelings from our dependent childhoods. Still, some
theorists emphasize how relationships or institutions affect our subconscious. Habermas, Lacan and many feminists stress relationships, while
Marcuse emphasized the effects of capitalist institutions (Elliott, 2000).
What? By focusing on the internal workings of our minds, psychoanalytic
theorists are naturally concerned primarily with how individuals develop
attitudes.

03 Szostak (to/d)
12/17/02
1:05 PM
Page 41
41
Why? Given that none of us can know our subconscious perfectly, and yet
this influences our decisions, at least some of our decisions cannot be strictly
rational. Intuitive decision-making will dominate in at least some circumstances.
When? Althusser defines ideology as a discourse which leads individuals
to see themselves and others in a way that supports the existing social structure. His version of psychoanalytic theory suggests an equilibrium orientation. Others such as Deleuze and Lyotard have objected to the implication
that human beings are creatures of social forces to such an extent; they have
wondered, albeit pessimistically, if by changing people society can be transformed (Elliott, 2000).
Where? While there is little explicit discussion of this question, there seems
a presupposition that our subconscious is widely influential. Moreover,
psychoanalytic theorists have generally eschewed empirical investigation for
broad theorizing.
Comments Psychoanalytic theory gives very firm answers to the Who,
What and Why questions. Implicit consensus on Where may unfortunately reflect a preference for theorizing over empirical analysis. There is
scope for considerable disagreement on When.
Symbolic Interactionism
Who? As the phrase interactionism suggests, the focus here is on relationships. Plummer (2000) identifies four key themes of symbolic interactionist
theory. The first celebrates the human ability to create symbols, and through
these define both ourselves and our situation; these meanings evolve through
relationships. The third theme is interaction itself: The focus of all interactionist work is neither with the individual nor the society per se; rather its
concern is with the joint acts through which lives are organized and societies
assembled (Plummer, 2000: 195). While theorists may ask questions about
individuals or groups they will strive to understand these in terms of relationships. Detractors have criticized symbolic interactionism for having little
place for the essence of the self: subconscious, emotions and so on. Some
symbolic interactionists have attempted to incorporate a more detailed vision
of the self.
What? The first of Plummers themes stresses the negotiation of meaning.
Symbolic interactionism can thus be seen to emphasize attitudes. Especially
in the hands of Blumer, symbolic interactionism could involve analysis of
how attitudes influence actions.

03 Szostak (to/d)
42
12/17/02
1:05 PM
Page 42
When? Plummers second theme stresses change, flux and the emergence of
new properties from interaction. This indicates a stochastic perspective on
dynamics.
Where? Plummers fourth theme expresses a concern with empirics.
Symbolic interactionists may explore any aspect of the social world, but are
expected to do so with close attention to detail. A great deal of modern
research was presaged by the work of Simmel, who sought richly textured
vignettes rather than abstract generalizations (Plummer, 2000: 199). While in
some sense symbolic interactionists believe that their type of theory has universal application (at least among intentional agents), their efforts tend to be
solidly idiographic. There is, though, some discernible effort by some
scholars to draw generalizations across studies.
Why? There is little explicit discussion of how meaning is negotiated. The
stress, though, seems to be on the intuitive: humans have inherent abilities to
create and interpret symbols.
Comments Symbolic interactionism (or perhaps Plummer) does a better job
of identifying answers to each 5W question than was the case with previous
theories. Still, there is room for disagreement among individual theorists with
respect to both Where and Why.
Plummer (2000: 196) suggests in passing that since symbolic interactionism holds that meaning is constantly being negotiated, scholars cannot hope
to pin down precisely what symbolic interactionism means. He is, of course,
correct in noting that any theory will evolve in subtle ways through time.
Note that the 5W typology can serve to identify such changes. But Plummer
nevertheless manages to identify fairly precise answers to the 5W questions.
A general lesson can be drawn here: the scholarly community can and should
essay to identify the characteristics of any theory at any point in time.
Plummer also remarks that symbolic interactionism, though a minority
view in modern social science, nevertheless has wide but generally unrecognized influences on writers as diverse as Bourdieu, Giddens and
Habermas. An effort to identify theories typologically should make it easier
to recognize when elements of particular theory types are being applied
elsewhere.
Rational Choice Theory
Who? Individualism is the key assumption of rational choice theory (Abell,
2000: 231). Abell, though, sees game theory as an offshoot of rational choice
theory that investigates strategic interaction among rational actors. Game
theorists thus emphasize relationships, though in a way that these are understood in terms of the individuals involved.

03 Szostak (to/d)
12/17/02
1:05 PM
Page 43
43
Why? As the name suggests, emphasis is upon rational decision-making.

Abell (2000) identifies two other key assumptions self-regard and optimality of decision-making but notes that all are relaxed on occasion. Discussion
of cultural norms or altruism is usually in terms of explaining these as
outcomes of rational decision-making. Abell suggests that if cultural norms
effectively limit us to one choice, rational choice theory collapses into functionalism. He also recognizes that in some strategic situations the best
strategy for an individual to pursue is to mimic others; this bears some similarity to decision-making according to tradition.
What? The general focus is upon actions. Individuals are assumed to
proceed rationally from preferences and schemas to decisions about actions.
Abell worries that rational choice theory is generally unclear on where ones
preferences come from. He suggests that insights from other theories regarding learning and emotions could be useful here. Recently rational choice
theorists have begun to investigate the tricky question of how individuals
decide whether it is worthwhile to obtain additional information, when by
definition they cannot know how useful it will be.
When? Economists have found it easier to generate equilibrium outcomes
in models with rational agents than have sociologists such as Parsons who
investigate multiple influences upon decision-making (Abell, 2000). Still,
equilibrium outcomes are not inevitable. Game theory in particular has long
sought to identify equilibria, though not always successfully.
Where? Rational choice theorists have not only striven to explain all individual-level behaviours, but also to reduce social behaviours to individual
actions.
Comments Like symbolic interactionism, rational choice theory provides
answers to all 5W questions. In the case of What? and Why? this precision
may be achieved by ignoring the genesis of preferences and schemas. This is
no great problem if, as Abell recommends, rational choice theory is viewed
as complementary to, rather than a substitute for, other theories. If so,
rational choice theorists would have to revise their answer to Where? as
well.
Phenomenology
Those chapters in Turner that embrace philosophical perspectives rather than
theories have been eschewed here. Phenomenology is an intermediate case.
According to Vaitkus (2000), phenomenologists celebrate the fact that there
has never been a satisfactory answer to the question of what phenomonology
is. This must render classification difficult. Moreover, in urging scientists to

03 Szostak (to/d)
44
12/17/02
1:05 PM
Page 44
directly explore individual phenomena and to free themselves from presuppositions unjustified by prior examination, phenomenology takes a
(laudable) position on philosophical issues, rather than a position on theoretical questions. There is an emphasis on the problem of intersubjectivity, and
thus the Who? question can tentatively be answered in terms of relationships and What? with attitudes, though there seems to be much concern
with actions (and at least one modern phenomenologist, Natanson, has
explored the reflective ego contemplation of the world). Answers to the other
three questions are even more tenuous.
What Is a Theory?
The vexed question of defining the word theory has been avoided to this
point. Trying to give a basic comprehensive account of the concept of a
theory is an invigorating but fruitless walkabout in metaphysics (Ziman,
2000: 117). Ziman goes on to assert that scholars generally know a theory
when they see one, and that theories are social institutions that conform to
the standards of scholarly communities. Such a definition hardly indicates
what a theory should contain. By identifying the various key dimensions
along which theories can differ, a more precise definition of theory can be
developed. Theories, in other words, are attempts to specify in a logically
consistent manner agency, action, decision-making processes, location and
time path, with respect to interactions among phenomena. As we have seen,
many theories fail to specify answers to each of the 5W questions (though
variants of these often do). Such theories could and should do so; they thus,
at least potentially, qualify as theories within this definition.
Grand theories especially, but many narrower theories as well, combine
theoretical analyses of different types. These might usefully be thought of as
theory clusters. Such theory clusters could be evaluated theory by theory.
The similarities and differences between theory clusters can more readily be
appreciated at the level of individual theories. Identifying which theories
within a theory cluster accord best with empirical reality, or involve arguments that are hard to examine empirically, is also possible. So too is identifying theories within a theory cluster that are poorly specified in terms of one
or more of the 5W questions.
Advantages of a Typology of Theory

Turner (2000) identifies four major problems with social theory. The
typology developed here can provide a partial or complete solution to each
of these. First, Turner worries that there is no obvious progress in theory.

03 Szostak (to/d)
12/17/02
1:05 PM
Page 45
45
Theorists work on one problem for a while, and then abandon this to work
on something else. Often they merely reinvent the wheel, reproducing
research on a problem abandoned long before. To start with the last, the 5W
typology, combined ideally with a classification of methods, provides a
potential system of categorization such that a researcher could immediately
know if a particular type of theory (and method) had ever been applied to a
particular set of phenomena. More generally, the typology provides an
exhaustive set of what is possible. This can alert researchers to possibilities
they may have overlooked. And most centrally it guides scholars to appreciate that the scholarly community should be striving for a mix of different
types of theories. Necessarily each of these will be deficient in some ways.
Rather than abandoning a research agenda because of this realization,
scholars are guided to work towards an ever greater synthesis. From this
perspective, the goal is to identify the range of applicability of every theory.
And thus finding limitations to a particular theory should be seen as progress,
unless the discovery is soon forgotten.12
Questions of generalizability are rarely addressed by theorists. For many
theories, some of the other 5W questions are ignored as well. The typology
guides theorists and empiricists to specify answers to each question: it
thus directly encourages one important sort of theoretical progress.
The fact that the typology provides an exhaustive set of types of theory
deserves emphasis. While infinite permutations of detail may be possible
within theory types, there is a clear finite limit to the number of types. This
typology, in concert with finite lists of methods and phenomena, means that
science is essentially bounded: there are limits to the directions in which
scientific investigation can be taken. Rule (1997: 8) worries that, There is
simply no limit to the theoretical agendas that could conceivably serve to
animate our work. At least at the important level of theory types, our
typology shows this to be untrue.
Recognition that different theory types have different strengths and limitations since each is focused on different answers to the 5W questions has
a more immediate effect. Researchers should be guided to ask what theory
types have not been applied to a particular question. Note in this regard that
even if the link being analysed involves phenomena reflecting only one type
of agency, it may still be useful to employ other types. These can show how
the phenomena under study both react to and affect diverse other phenomena. For example, in asking why one institution has a particular effect on
another, considerable insight may be gained by exploring how individuals or
groups affect and are affected by the institutions in question.
Turners second problem is that it is not clear whether the goal of theory
is explanation (of actions) or interpretation (of attitudes). We have seen that
the answer depends on the type of theory, and suggested that scientific understanding requires both types.

03 Szostak (to/d)
46
12/17/02
1:05 PM
Page 46
Turners third problem is that there is no consensus on what a theory is,

or what would define progress. Is a theory a broad framework for organizing research? Or a collection of general concepts that direct research? Or a
specific orientation that guides researchers to ask particular questions?
Theory can be all of these things, but cannot be just any one of them. Theory
clusters, and perhaps some nomothetic theories, may serve as broad frameworks. But theory should do more than paint the big picture; it should also
fill in the details. Nomothetic theories can be thought of as general concepts;
research guided only by these would again miss important details. As for
specific orientation theories can and should guide the formation of research
questions. But they should in turn be informed by research. And since theory
must be broader than any of these definitions, progress is best evaluated in
terms of whether our overall understanding of the strengths and limitations,
and especially the range of applicability, of our theories has increased.
Turners fourth problem is that social theory has not come to grips with
various dichotomies, such as action vs praxis, agency vs structure, or individual vs societal causation. The 5W typology suggests that these are all not
only legitimate but necessary types of theory. The strengths and weaknesses
of each type of theory should be appreciated.
To Turners list can be added a more pragmatic concern that serves to
exacerbate those that he discusses. At present it is difficult to identify the
essence of particular theories, or to appreciate the differences and similarities
across theories. The 5W typology summarizes most/all of what the authors
in Turner (2000) found it important to say about the theories that they discussed. It thus readily identifies the essence of different theories, and the key
sources of theoretical disagreement both within and across theories. Communication across theoretical camps, and even disciplines, should be facilitated. Note here that a systems theorist who answers the Who and What
questions as an action theorist would can be indistinguishable from an action
theorist who answers the When and Where questions as in systems theory.
In sum, the typology serves two main goals: it defines what theory is
(and is not), and shows that there are a finite number of potentially complementary theory types, each of which may have an important role to play in
the scientific enterprise. It shows us that each theory type has strengths and
limitations. All existing theories can be situated within the typology; in doing
so it often becomes apparent that certain questions have been unanswered.
We are thus guided both to improve individual theories and to seek to
identify the range of applicability of each.
Notes
1 Barnes (1995) argues that only relational theory is valid. He suggests that individual-level theory is wrong because tastes are not fixed, and functional theory is

03 Szostak (to/d)
12/17/02
1:05 PM
Page 47
47
wrong because norms are manipulated rather than taken. As with all who argue
for only one type of causal agent, he downplays the simple fact that the world is
a complex web of causal links.
One might also make a distinction between whether agents are acting or
reacting. Yet every phenomenon influences and is influenced by hundreds of
others. Every act, whether passive or active, action or attitude, is necessarily
conditioned by the numerous influences upon the phenomenon in question. A
scholar studying reaction is explicitly taking these into account. A scholar
studying action is treating these as exogenous for the moment and focusing upon
effects.
I join Seale (1999: 22) in decrying the oversimplification involved in terms like
positivist, which tend to lump together scholars with diverse perspectives on
many issues. Nevertheless, certain common or average tendencies can be
identified.
Bunge (1998) provides numerous examples of the productive mixing of positivist
and interpretive approaches. Bunge emphasizes in particular how neuroscience
the analysis of physical changes in the brain has aided and will aid our understanding of how we think and act. Our human fallibility means that individuals
cannot attribute the results of their actions entirely to their intentions, and must
thus pay heed to the unforeseen consequences of their actions. But nor can they
ignore those intentions, and treat their actions as the unmediated responses to
certain stimuli.
While it may seem counter-intuitive that Where and When questions yield the
same answer, scientists and philosophers tend to address questions of time and
space in concert. An example is Urry (2000).
To claim that a certain result will inevitably follow from a set of causes is to assert
that this set of causes is sufficient. Ragin (2000: 103) urges scientists to seek
sufficiency, whether the posited result is discrete (such as a movement to a particular equilibrium) or continuous (such as movement in a particular direction). He
thus appears to underestimate the importance of the fourth type of time path.
Ragin (2000: 108) recognizes that especially in social science sufficiency is an
elusive quarry: he suggests that social scientists speak of almost sufficient in cases
where only a small number of exceptions occur but many successes are observed;
ideally circumstances would be identified that distinguish the exceptions. Ragin
notes that the strategy for identifying sufficiency is to look across a wide range of
cases in which the posited set of causes was in place, and ask whether the result
always followed; he regrets that scientists often ignore this simple strategy. The
reverse strategy for identifying necessity is equally neglected.
The selection was not entirely random. Chapters were chosen that seemed focused
on one type of theory (or a closely related family of theories). Chapters on anthropology, feminism and postmodernism were ignored, being judged to cover too
wide a range of theorizing. Note that Plummer (2000) stresses the similarity
between symbolic interactionism and much postmodern theory. The chapter on
structuralism overlapped significantly with that on functionalism. Marxian theory
is treated only in passing in Turner.
Action theories that emphasize cultural influences on individual behaviour can in
the extreme portray individuals as merely reacting to overwhelming cultural influences. Cohen (2000) thus details conflicts within action theory concerning both

03 Szostak (to/d)
12/17/02
48
9
10
11
12
1:05 PM
Page 48
the scientific and ethical implications (with respect to free will) of individual-level
vs societal-level agency.
Yet later Cohen (2000)expresses doubt that theories of action and praxis can be
linked. Why not view these as complements with different ranges of applicability?
Cohen (2000) notes that network theorists tend to ignore the fact that individuals
spend much time alone not worrying about relationships.
According to Boyne (2000), the essence of structuralism is similar: a belief in
structures in which every element is necessarily related to others in a similar way.
Saussure, for example, had stressed how words only exist in relationship to each
other within the structure of a language.
Rule (1997: 32) warns us against glorying in the simple application of a theory to
a new domain, and then calling this progress. It must be established that this application has some empirical validity.
References
ABELL, Peter (2000) Sociological Theory and Rational Choice Theory, in Bryan S.
Turner (ed.) The Blackwell Companion to Social Theory, pp. 22344. Oxford:
Blackwell.
BARNES, Barry (1995) The Elements of Social Theory. Princeton, NJ: Princeton
University Press.
BOYNE, Roy (2000) Structuralism, in Bryan S. Turner (ed.) The Blackwell
Companion to Social Theory, pp. 16090. Oxford: Blackwell.
BUNGE, Mario (1998) Social Science under Debate: A Philosophical Perspective.
Toronto: University of Toronto Press.
COHEN, Ira J. (2000) Theories of Action and Praxis, in Bryan S. Turner (ed.) The
Blackwell Companion to Social Theory, pp. 73111. Oxford: Blackwell.
ELLIOTT, Anthony (2000) Psychoanalysis and Social Theory, in Bryan S. Turner (ed.)
The Blackwell Companion to Social Theory, pp. 13359. Oxford: Blackwell.
ELSTER, Jon (1989) Nuts and Bolts for the Social Sciences. Cambridge: Cambridge
University Press.
FREIDHEIM, Elizabeth A. (1982) From Types to Theory: A Natural Method for an
Unnatural Science. Washington, DC: University Press of America.
GOLDENBERGER, Sheldon (1992) Thinking Methodologically. New York: Harper
Collins.
LECHNER, Frank J. (2000) Systems Theory and Functionalism, in Bryan S. Turner
(ed.) The Blackwell Companion to Social Theory, pp. 11232. Oxford: Blackwell.
LITTLE, David (1998) Microfoundations, Method, and Causation. New Brunswick,
NJ: Transaction.
PLUMMER, Ken (2000) Symbolic Interactionism in the Twentieth Century, in Bryan
S. Turner (ed.) The Blackwell Companion to Social Theory, pp. 193222. Oxford:
Blackwell.
POTTER, Garry (2000) The Philosophy of Social Science. Harlow: Prentice-Hall.
RAGIN, Charles C. (2000) Fuzzy Set Social Science. Chicago, IL: University of
Chicago Press.

03 Szostak (to/d)
12/17/02
1:05 PM
Page 49
49
RULE, James B. (1997) Theory and Progress in Social Science. Cambridge: Cambridge
University Press.
SEALE, Clive (1999) The Quality of Qualitative Research. London: Sage.
SZOSTAK, Rick (2000) Toward a Unified Human Science, Issues in Integrative Studies
11557.
SZOSTAK, Rick (2003) A Schema for Unifying Human Science: Interdisciplinary
Perspectives on Culture. Selinsgrove, PA: Susquehanna University Press.
TURNER, Bryan S., ed. (2000) Preface, Introduction, in The Blackwell Companion
to Social Theory, pp. 118. Oxford: Blackwell.
URRY, John (2000) Sociology of Time and Space, in Bryan S. Turner (ed.) The
VAITKUS, Steven (2000) Phenomenology and Sociology, in Bryan S. Turner (ed.) The
ZIMAN, John (2000) Real Science: What It Is and What It Means. New York:
Cambridge University Press.


Classifying Natural and Social Theories (Szostak, 2003)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Classifying Natural and Social Theories (Szostak, 2003)

Uploaded by

Copyright:

Available Formats

Current Sociology

Classifying Natural and Social Scientific Theories

International Sociological Association

Downloaded from http://csi.sagepub.com by Alberto Conde-Flores on January 23, 2008