BEHAVIORAL SYSTEMS ANALYSIS:

METHODOLOGICAL STRATEGIES AND TACTICS

by Roger D. Ray

 Rollins College

and

 Dennis J. Delprato

 Eastern Michigan University

A general-systems model stressing the inherent organizational aspects of continuous behavioral-environmental interaction is presented as a natural extension of 20th century developments in scientific worldviews. The cultural context of scientific inquiry has often been depicted as a dynamic process which has evolved through at least three phases of philosophical orientation, or paradigmatics. Each paradigm has carried its own special implications for research strategies and tactics. Beginning with a substance-property view emphasizing self-action, scientific philosophy gradually shifted to a cause-effect orientation with emphasis on the lineal-mechanical aspects of that presumed process. Developments within this century have brought science to the integrated-field/systems view of natural phenomena, with an associated emphasis on organizational properties and dynamics. However, psychology has lagged considerably in accommodating to this latter stage of scientific philosophy, both theoretically and methodologically. The model presented identifies with modern integrated-field/systems theory and offers empirical illustrations of its implications. Emphasis is placed on practical distinctions between structural, functional, and operational analysis, with special attention to such model-specific phenomena as behavioral velocity, kinematic syntax and variability, systemic coherence, and dynamic stability versus perturbed oscillations in systemic operations. Compatibilities between the behavioral systems model and various other contemporary efforts to expand psychological paradigmatics are reviewed in conclusion.


KEY WORDs: behavioral systems, individual organism, scientific philosophy, systems methodology.


CRITICAL OBSERVERS of behavioral science (e.g., Giorgi, 1970; Kantor, 1953, 1969; Overton & Reese, 1973, 1981; Reese & Overton, 1970) reveal that issues apart from events under investigation impact all phases of the scientific enterprise, including problem selection, methods and procedures, and interpretations. The present paper introduces a relatively new methodological orientation which stems from such concerns. Our fundamental thesis is twofold: first, scientific methodologies are inseparable from cultural and philosophical traditions (or worldviews) which translate into philosophies, strategies, and tactics of science (i.e., generally accepted rules and guidelines for scientific behavior); and second, these rules and guidelines gradually change across time and cultures and today call for revised theory and methods in the behavioral sciences. In response to this need, we herein offer a systems-based methodology that comports with significant developments in scientific thinking in a variety of disciplines outside of psychology and which offers unique descriptions of behavioral-environmental interactions that are veridical to events under investigation.

The first part of this paper describes specific developments in scientific thinking by overviewing briefly major changes in how scientists have approached their subject matter over the years. We make a case that modern science has evolved to an integrated-field (i.e., systems) perspective, and that this perspective calls for unique methodological guidelines and developments to be embraced by the behavioral sciences. The next part of the paper addresses a few major orienting assumptions, derived from the integrated field/systems perspective, that guide the construction of a specific theory of behavioral systems and generate a specific research methodology. Next we outline the major strategic components of this behavioral systems methodology in the abstract and demonstrate the potential value of the components by sampling particular descriptive and analytic projects illustrating their special character. In the final section, we take up the theoretical and methodological implications of the integrated-field/systems perspective for a much overlooked aspect of scientific research in general, i.e., its descriptive-phenomenological foundations. Addressed here are tactical issues revolving around the fact that all forms of description involve the replacement of the primary events by symbolic representational systems.

EVOLUTION AND FOUNDATIONS OF

INTEGRATED-FIELD PHILOSOPHY

Several authoritative historical analyses (e.g., Dewey & Bentley, 1949; Einstein & Infeld, 1938; Handy & Harwood, 1973; Kantor, 1946, 1969) agree on three general stages in the evolution of scientific thinking. Thinkers first assumed that natural events acted under self-contained powers. As noted by Dewey and Bentley (1949), to the time of Galileo the learned view was "that there exist things which completely, inherently, and hence necessarily, possess Being, that these continue eternally in action (movement) under their own power--continue, indeed in some particular action essential to them in which they are engaged" (p. 110). Theorists invoked various substances with unique, inherent properties to account for heat (caloric), combustion (phlogiston), light (the ether), biological functioning (vital force, entelechy), and human psychological behavior (soul, mind). Thus this initial stage is referred to as substance theory (Einstein & Infeld, 1938), substance-property stage (Kantor, 1946, 1969), and self-actional stage (Dewey & Bentley, 1949).

The advent of the mechanical view (as termed by Einstein & Infeld, 1938), statistical correlational stage (Kantor's 1946, 1969 term), or interactional stage (Dewey & Bentley's 1949 term) is marked by the work of Galileo. This second general scientific approach retained substances, but now thinkers interpreted natural phenomena in terms of forces acting between unalterable objects. According to Einstein and Infeld (1938), Newton's gravitational law connecting the motion of the earth with the action of the distant sun exemplify the second stage. "The earth and the sun though so far apart, are both actors in the play of forces" (p. 152). It was in the mechanistic stage that theorists advanced the energy construct as a new substance and used it as the basis for transformational descriptions expressed in statistical-correlational laws. In biology, mechanists such as Sherrington (1906) and Loeb (1912) countered vitalism by reducing integrated biological activity to physicochemical causal chains. And many contemporary psychological descriptions can be seen to have foundations both in mechanistic biology and in Fechner's (1860) statistical formula purporting to correlate the mental (sensations) with the physical. This second stage of scientific thinking was the era of the world machine, mechanism, materialism, causal determinism, and reductionism. Workers acknowledged experimentation by an impartial experimenter manipulating "truly" independent variables (causes) and measuring dependent variables (effects) as the ideal scientific model for empirical and conceptual analysis.

Contemporary physical scientists no longer compare the world to a machine. The world-machine notion has gradually faded (Frank, 1955), although with complexity and some continued influence (Holton, 1973). According to Einstein and Infeld (1938), the transition from classical mechanics (e.g., Newton's gravitational laws) to Maxwell's equations was a critical development in the evolution of a third stage of thinking in physics. Now there are no material actors, the mathematical equations "do not connect two widely separate events, they do not connect the happening here with the conditions there" (Einstein & Infeld, 1938, pp. 152-153). Maxwell's theory introduced the field construct, according to which events here or now are not connected to conditions there or then. As Einstein and Infeld (1938) put it, "The field here and now depends on the field in the immediate neighborhood at a time just past" (p. 153). Furthermore, although the mechanical theorist attempted "to describe the action of two electric charges only by concepts referring to the two charges,... in the new field language it is the description of the field between the two charges, and not the charges themselves, which is essential for an understanding of their action" (Einstein & Infeld, 1938, p. 157). The field construct has taken physics far away from the mechanistic stage with its bifurcation of nature (e.g., mass and energy, matter and force, gravitational mass and inertial mass) to the inertial-energy concept and the equivalence of mass-energy and gravitational-inertial mass.

Field thinking has also directed explanatory efforts in physics away from mechanism and its implied search for ultimate causes. Modern physical scientists no longer approach their science from the cause-effect framework (e.g., Feigl, 1953; Holton, 1973; Russell, 1953). According to Feigl (1953), the field alternative to the terms cause and effect in ordinary language is "an entire set of conditions [event-fields]" (p. 410), and this set represents the explanation of an event. Kantor (1959) further clarifies the field construct and makes the same point in discussing the field alternative to conventional causal construction:

All creative agencies, all powers and forces, are rejected. An event is regarded as a field of factors all of which are equally necessary, or, more properly speaking, equal participants in the event. In fact, events are scientifically described by analyzing these participating factors and finding how they are related (P. 90).


In retrospect, Kantor (1959, 1969) and Kantor and Smith (1975) seem to be the to have most fully developed the modern field conception per se within the realm of psychology. According to Kantor (1969), the psychological field is:

... the entire system of things and conditions operating in any event taken in its available totality. It is only the entire system of factors which will provide proper descriptive and explanatory materials for the handling of events. It is not the reacting organism alone which makes up the event but also the stimulating things and conditions, as well as the setting factors (p. 371).

Further, the integrated-field perspective rejects heredity, environment, mind, cognition, stimuli, reinforcers, independent variables, and so on, as creative forces in behavior. In fact, behavior itself must be totally redefined as an event. Kantor (1959) felt so strongly about this fact that he substituted the term "interbehavior" to describe the psychological event. Such interbehavioral events involve not only the actions of an active organism, but also the stimulating objects, media of contact between organism and environment, functional stimulus and response attributes, and attending setting factors, all of which define the psychological field (Kantor, 1959). It is in this spirit that we herein apply the term "behavioral system," and we assert the fundamental units of such a system as being synonymous with the organization of multiples of Kantor's psychological (i.e., interbehavioral) events. As such, behavior must be analyzed in terms not just of organismic reactions, but rather in terms of all participating components: organismic actions, stimulating objects, response and stimulus functions, media of contact, and attending setting factors. These comprise the total field of conditions for a continuous flow of psychological events which ultimately define an individual psychological system. The remaining challenge is the development of appropriate methodological strategies for describing the most salient characteristics of such a system.

FROM INTEGRATED-FIELD TO SYSTEMS

METHODOLOGY

Unfortunately, the mechanistic stage of science left a methodological legacy in behavioral sciences that even Kantor did not overcome in practical terms. He, as well as others (see Giorgi, 1970; Kantor, 1953, 1969; Overton, 1984; Overton & Reese, 1973, 1981; Reese & Overton, 1970), recognized the need to: (a) place even the experimental method in a purely descriptive stance; (b) replace the notion of the unbiased experimenter with a recognized field-participating observer; (c) abandon cause-effect deterministic thinking, with its emphasis on independent/dependent variable relationships, in favor of field descriptions and interdependent components; (d) ultimately replace notions of purely reactive organisms and simple antecedent-consequential analyses of control; and much more. Kantor, Giorgi, Overton, and Reese have all addressed these needs philosophically, but none have yet accomplished their implied ends methodologically. Nevertheless, Kantor and others have outlined methodological requirements compatible with solutions which already exist, at least in part, within modern systems science. Yet there are major points to be discerned to assure that only the appropriate components of this approach are utilized. To this end, we first review the concept of a system in general and then attend to some of the special strategic and tactical considerations needed for describing the psychological system in particular.


What Are Systems?

The systems orientation presumes that natural events derive from localized organizations. Organization itself implies an interrelationship among constituent elements of the system. Thus a system is defined as a set of interrelated (i.e., organized) constituent elements, as may be illustrated by a very simple geometric example. Figure 1A shows 20 geometric points (elements) which are mutually exclusive of one another, and thus have no discernible organization (i.e., no descriptive rule allowing specification of any single point from knowledge of any other points). Thus, if any point were to disappear, one could not determine which one had been removed by knowledge of even all remaining points. Systems theorists refer to such a state as one of complete entropy (cf., Bowler, 1981). On the other hand, Figure 1B illustrates an array of 20 similar points with a high degree of organization. Each point thus implies the location of others. In this case, knowledge of only a few such points' locations allows accurate prediction of virtually all others. And this mutual implication is specified by a simple descriptive organization principle, or structural rule, which invokes no concept of causation among points (i.e., no single point causes another point to be located where it is, it merely implies its existence). Systems theorists refer to such highly organized systems as having a state of

coherence, or negentropy (Bowler, 1981).



FIG. 1. Illustration of two extremes of systemization between the elements of a simple geometric system. The top (1a) illustrates a random array of points which define a state of entropy, while the bottom (1b) illustrates a coherently organized array where each point shares a mutual implication for others.

At least three descriptive advantages of the systems approach are readily apparent at this point, once we add that naturally occurring complex systems rarely, if ever, achieve complete coherence and are thus virtually never static in their organization. First, such systems are invariably less than fully deterministic--they are best described probabilistically. Second, like integrated-field and kindred thinkers (see Feigl, 1953: Giorgi, 1970; Holton, 1973; Kantor, 1953, 1969; Overton, 1984, Overton & Reese, 1973, 1981; Reese & Overton, 1970), systems theorists assert that simple lineal mechanics models, with their assumptions of cause-effect, do not provide satisfactory foundations for systemic analysis. Third, unlike lineal mechanics which stresses one-way causality between independent and dependent variables, mutual implications incorporate a bidirectionality among systemic elements which is mutually shared, thus requiring analysts to avoid labeling one as independent and another as dependent--they are, rather, interdependent elements.


Organization in Behavior

Virtually all philosophies of science take organization in the world as axiomatic. Therefore, psychologists uniformly have agreed that behavior is organized, and they have maintained an interest in behavior organization over the years. However, because researchers typically use mechanist principles to study and account for organization, this characteristic of behavior has received limited treatment following the strategies derived from lineal mechanics. Consequently, psychologists have not experienced much success to date in describing the character of multibehavioral or environmental-behavioral organizations (Delprato, 1986).


In contrast to the scant treatment of behavioral organization that follows from the mechanistic stage of science, the integrated-field and system perspective takes organization as a given property. The treatment of organization is much the same as that of change. That is, the integrated-field view takes development or change as basic, not derived, and therefore as being explained by interpretative or constructive descriptions (Kantor, 1959; Overton & Reese, 1981). In line with the most recently evolved scientific thinking, there is no formal division between description and explanation (Frank, 1955; Holton, 1973; Kantor, 1953; Overton & Reese, 1981; Schafer, 1976). It is substance-property and mechanistic thinking that promoted the traditional views that explanation is to identify higher level substantive causes of described phenomena and that the essence of the distinction between description (observation) and explanation (law) is that "explanation somehow transcends events, [and] provides certain and a priori knowledge concerning their nature" (Kantor, 1953, p. 33-34).

Contrasting treatments of organization by the integrated-field perspective and by the mechanistic approach are readily apparent if one examines how each handles those field factors centered on the actions of the organism (responses). The mechanistic approach typically assumes an independence between response classes. Thus the integrity of this assumption is routinely maintained by the use of single-response class methodologies, i.e., researchers measure only one type of response, or when they do measure more than one response class, their methodologies rarely focus on interdependencies among the classes. The result is that response organization, when noted, is attributed to organizing factors external to the responses (causes, independent variables). These hypothetical organizing factors may be placed in the external environment or inside the organism (e.g., biological states and processes, mental structures). On the other hand, the integrated-field orientation recognizes the inherent (probabilistic) organization of the constituent elements in natural systems. This means that both response interdependence and stimulation- response interdependence is a given property, hence, responses are not dependent variables subject exclusively to external control (Henton & Iverson, 1978; Honig, 1959; Ray & Brown, 1975, 1976; Ray, 1977; Upson & Ray, 1984). Clearly, single-response and isolated stimulus-response sequence methodologies are insufficient to reveal complex organism-environment interdependencies; a multiple-component methodology is required.


Multiple Component Methods

Multiple component methods should recognize a variety of levels of composition and organization. First, given the holistic view of organismic behavior, adequate description calls for measurement over multiple domains of organismic action systems and environmental stimuli. Although there is not yet a generally recognized classification of organismic action domains, biological and behavioral researchers have found it useful to regard as measurement domains the overt somatic, transduced muscular (electromyographic potentials), cardiovascular, central nervous, glandular, and experiential. Measurement of multiple response and environmental element classes and (i.e., categories) is then accomplished within domains (for example, within the overt somatic domain we might isolate categories such as walking, standing, sitting, lying, etc.). Likewise, environmental stimuli may be separated into physical, chemical, biological, or social domains and subsequently categorized via relevant taxonomies, much like the table of elements categorizes atomic levels of environmental constituents and taxonomies in biology identify categories of organic systems. Unfortunately, psychology has accomplished little toward a suitable category system for describing either behavior or environment in a systematic way. Thus most publications still refer to physical measurement criteria, such as sound frequency and loudness levels, lux levels for lighting, etc. In brief, multiple-component class analysis within and across multiple domains is a necessary beginning for eventual efforts to reveal inherent spatiotemporal organization of behavioral-environmental interaction systems. As we will soon note, such efforts may focus on structural attributes for classification, or may stress functional rules in the interaction as the classifying criteria.


Time Series Requirements

The above discussion makes explicit the assumption that organization requires the existence of more than one systemic element. But it also tends to focus exclusively on the concept of multiple events as if they were independent of time. Clearly, this is not the case for living systems. Living systems experience not only a multitude of event components in selected psychological moments, but also are time-bound in their integration of multiples of such moments. That is, living systems exist across time well as within time, thanks to historical field factors. Thus as researchers, we are required not only to focus on multiple domains and categories, but must also focus on multiple observations of these events across time. As such, all systems research requires at least implicit time series technologies. Organizational properties such as multiresponse interdependencies may be summarized as somewhat static states for a given temporal "window," but are nonetheless derived from multiple observations across the time represented within the window. As we shall soon see, such implicit collapsing of time-series observations is typical of structural/functional analysis. As we will also soon discover, systems methodology also stresses the explicit use of time series descriptions in its conduct of operations research.

To sum up the major implications of the integrative-field focus, behavioral systems methodology needs to be sensitive to (a) the continuity of organismic-environmental interactions across time and circumstance; (b) the interdependencies of these interactions; (c) the interactional context, or setting conditions attending the interaction; and (d) multiple classes of events within each of many possible domains. These impose requirements on a behavioral systems methodology which depart notably from conventional behavioral methodologies. The latter tend to (a) focus on the organism; (b) take measurements over instants of time or over narrow time windows; (c) ignore the settings; and (d) measure single response classes within one domain as a dependent variable driven by relatively singular environmental (independent) variables.

Next section.... BEHAVIORAL SYSTEMS ANALYSIS: STRATEGIES AND ILLUSTRATIONS