Types, dimensions, and limitations

Abstract

Constantine's (2025) theoretical framework has been shown to provide a useful set of orienting ideas, and the set of papers cited show where other authors have employed it. Therapists have clearly gained useful insights from these orienting ideas to help them perceive and analyze family processes. Constantine has done an exemplary job of recognizing important empirical patterns. It is clear that when trying to theorize, he can derive some of the observed patterns from quite simple base structures.

The Paradigmatic Framework has a grandiose title. It is certainly a Kuhnian paradigm that describes some ways systems may be classified (Kuhn, 1962). Although this framework can be seen as a paradigmatic framework to be sure, every other framework that describes types of systems or families (like Baumrind's, 1971) is also a paradigmatic framework of the same kind.

Any theoretical framework only explains what it explains and is inherently a partial view of the world. All theorizing is limited in this way. “It is about what it is about,” as Constantine states (p. 22). Theorizing depends on a careful specification of concepts and delineating of relationships among these concepts. As Constantine points out, “Critical analysis of theory as theory—its assumptions and concepts, implicit biases and limitations, core ideas and hypotheses—is an important but sometimes overlooked route to advancement” (p. 2).

One of the issues with any theoretical typology is that when you divide the world into clearly theorized groups, there are lots of rough edges and fuzzy boundaries when you apply the typology in the empirical world. The types that Constantine describes are ideal types, a term that refers to multidimensional concepts simplified into a typology (Weber, 1949). An ideal type is designed to reduce the innumerable dimensions that describe real life to a few core concepts. An ideal type description is assumed to be generally but not exactly descriptive of any empirical system.

Constantine notes that his framework has three fundamentals (pp. 5–6): (a) mechanisms of system theory; (b) system “issues” (which seem to be functional prerequisites: Parsons, 1951); and (c) the dialectic. Constantine's Table 1 describes his framework, but the language of criteria (a) and (b) do not appear within the table. The table lists multiple other criteria. However, it is not clear how the items in the table are logically derived from the dialectic criteria or other sources. Constantine may have provided such linkages elsewhere, but they are not available here.

The dialectic has an outsized role within his typology of general system types, although he has little to say about it. Constantine does not provide a source for his version of the dialectic. I am aware of two versions of the dialectic. In one (which seems to follow Plato's example: Straus, 1987) an argument (thesis) followed by another argument (antithesis) followed by agreement (synthesis). I argue A; you argue some B; and then we agree that C, which may incorporate elements of A or B or both (Strauss, 1987). This version does not seem to correspond to Constantine's dialectic pattern. The second example, derived from Marx (Avineri, 1968) specifies that the first system (thesis) contains a logical contradiction (antithesis), which is resolved in a new system (synthesis). Constantine does not seem to be following this pattern either. Constantine proposes that each of his ideal typical systems can be identified as part of a dialectical sequence. He describes the types as forming a “dialectical progression” (p. 6) and argues that the four types of systems that he describes are emergent from such a dialectical process (Note that Constantine and I both ignore his fifth ideal type, the unified system).

My judgment is that this typology successfully describes the four system types. However, his dialectic system describes a chronological sequence of system types where each one (at least up to the open system) proposes improvements on the previous system. I did not find in the paper or the earlier book (Constantine, 1986) whether he had ever tested his hypothetical process. Do persons in random systems complain about one-person rule so they have consciously chosen this system? Do open members appreciate their escape from the chaos of the random system? If not, perhaps this is just a typology and not a dialectic progression. As a typology, this framework has demonstrated its worth. This framework had demonstrated its worth as a typology.

It is difficult in examining Table 1 whether each of the listed elements of the table is derived from theoretical analysis and which are empirical summaries from observation. Most entries are not described or explained in the text. In the end, Constantine does not seem concerned about his four-type framework, because he is going to analyze it as a two-dimensional scheme. This appears to be what I have elsewhere described as an “epicycle” (Bell, 2009, pp. 48ff), a fundamental change in logic while maintaining continuity with an original theoretical framework. Ptolemaic astronomy used the logic of a perfect (divine) universe built on perfect spheres to describe the motion of planets. In astronomers' observations of Mars, for example, increasing accuracy eventually required that the orbit of Mars be described by seven circles (epicycles), each centered on the previous epicycle.

The critical elements in Constantine's plan to compare his framework with other frameworks are (a) his transformation of a four-type framework to a four-dimensional framework; (b) his transformation of four dimensions to two dimensions; (c) his creative re-analysis of the resulting two dimensions. Constantine's first step is to interpret his four ideal types as four dimensions. This step requires the assertion that each of the four types of systems represents an independent dimension. This change is not trivial. Now, instead of hierarchy being one of several criteria for identifying a closed system (e.g., stability), hierarchy becomes a single dimension on which the closed system is evaluated.

Constantine's second step is to reduce his framework from four dimensions to two dimensions. He transforms the four taxonomic dimensions of the original framework (hierarchy, divergence, convergence, consentience) into the three process dimensions of Constantine's Figure 2 (connection, variability, synergy). Next this figure is collapsed into two dimensions by deleting synergy. This move requires him to throw out his “rigorously derived and comprehensive” (p. 5) ideal types and related dimensions and then to engage in a brand new de novo conceptualization of his two new dimensions. Although the original theoretical framework was about oppositions or contradictions (the dialectic), the new theoretical framework is about dimensions. None of the insights of his four-type framework are preserved. After learning that the four types are distinguishable along four dimensions, Constantine states that now they are distinguished along two dimensions. And these two dimensions do not appear to share any theoretical meaning with the original four dimensions.

As his third step is to make sense of these new dimensions, Constantine has to conduct a new (contextual, idiographic) analysis to come up with names for the new dimensions. Naming empirically derived dimensions is a creative process (abduction: Bell, 2022, p. 27). Basically, hierarchy, which used to uniquely define the closed type of system, has disappeared and Constantine has to discover some concept that will well describe the closed (hierarchy) and open (convergence) systems and not describe random (divergence) and synchronous (consentience) systems. He calls this dimension “connection,” which he defines as “the degree of active engagement between participants” (p. 11). I find this confusing, as it seems that a random system where every member expresses their own ideas seems to imply a high degree of active engagement (“a free-wheeling family of creative types”, p. 7).

The second dimension he names “variability.” Thus he recognizes that both random (divergence) and open (convergence) systems share the characteristic of high behavioral variability. Of course, variability in a random system comes from chaotic processes of individual motivation of expressing one's own ideas, while variability in an open system comes from mutual processes of collective interdependence. He decides that these are close enough. He gives us only the briefest explanations of his two new dimensions: connection and variability. The issue here from the point of view of theory is that Constantine now has to engage in a new process of theorizing what his new theoretical framework is about. The sketchiness of his descriptions of connection and variability is a symptom of his lack of coherence in this direction.

Constantine's project is to compare the multiple theoretical frameworks of his own with Olson's (1993) and Baumrind's (1971). He sees similarities of form but differences of content in his framework and the other frameworks. His stated goal is to “clarify connections among these models and to anchor them together” by identifying “systems-theoretic fundamentals” (p. 3).

After his transformations, Constantine has reduced his framework to two dimensions. His first dimension is labeled “connection,” by which he means “active engagement.” This dimension is mapped onto cohesion, which represents emotional bonding for Olson (2000). Having defined his first conceptual dimensions as “active engagement” when comparing it with Olson's framework, Constantine renames it as connection/cohesion when analyzing demandingness or control in Baumrind's (1971) framework. Thus, Constantine seems to be incorporating Olson's concept into his. Now the connection dimension means both active engagement and emotional bonding. I propose that the very flexibility of his new dimensions shows that Constantine has not yet completed his conceptualization of his dimensions.

Constantine also maps variability, meaning “variation over time” onto Olson's (2000) flexibility, which represents an openness to change in structure over time and Baumrind's (1971) responsiveness, meaning to “intentionally foster individuality, self-regulation, and self assertion” (Constantine, p. 21). This leads Constantine to expand his variability dimension to become variability/flexibility.

Remember that Constantine has promised to “clarify connections among these models and to anchor them together” (pp. 2–3). In the end, I conclude that each of the three frameworks has two dimensions, but neither Constantine nor I seem to know what his two dimensions mean.

Every framework is limited in what it can explain. Constantine's analytical decision to generalize his typology of families to a general systems framework is limiting. In formulating his theoretical framework, Constantine intends to describe and potentially explain many types of organizational systems, with family systems as one example. The theoretical framework has been formulated to apply to any system, including, for example, work-based organizations or violent street gangs. As noted above, every framework attempts to explain a small part of the world. Constantine's framework has no room for emotion. None of the systems in Constantine's typology invokes closeness or caring. In fact, these are generic, unemotional systems. Thus, this is a framework that excludes emotions that would be inconvenient for many organizations. However, to formulate types of families or family dimensions while eschewing emotions, seems likely to be excluding what makes a family.

Although Constantine can accurately claim to have a framework of systems, his framework seems to be incomplete when applied to family systems. There may be empirical families that do not feel love for one another (families of sociopaths for one: Stout, 2005). Constantine's typology does not distinguish such families. His framework has omitted their emotional base. There is no mention in his analysis of whether people love or respect each other. This means that when he tries to compare his scheme to Olson's and Baumrind's, both of whom consider emotions, what this comparison highlights is that his original four family types seem sterile, at least in this formulation.

To be sure, there are interesting comparisons that can be made within and across frameworks, which is a strength of Constantine's framework. Flexibility is similar to variability, but flexibility refers to structures and variability to outcomes. Connection and cohesion sound similar until you realize that connection refers to actor engagement (without emotion) while cohesion is defined by emotional bonding.

Although the comparisons are thought-provoking, the level of the analysis is not particularly illuminating. Constantine's Figure 3 shows the comparison of Constantine with Olson, and Figure 4 compares with Baumrind, represent an isomorphic which is perfectly true in the sense that all have two dimensions, but I do not find the comparisons compelling. I am left with the awareness that there are a large number of concepts that can be used to describe families and other systems. Each theorist has provided compelling arguments for why their concepts and dimensions make sense. In the end this appears to be Constantine's opinion as well.

I agree that the typology of systems shows that the paradigmatic theoretical framework has “advantages as a theoretical framework for spread of continued research and theory development” (p. 23). Exactly the same can be said for Olson's circumplex model and Baumrind's parenting models. All are very successful as bases for research and theorizing. Each is limited in that each ignores a lot about families. It would be helpful to the field to analyze the explicit concepts of connection, control, and cohesion and the separate concepts of variability, responsiveness, and flexibility. I would like to see Constantine give more focused attention to the theoretical comparison of these concepts.