Exploring "Levels of Explanation" Concepts 
Part II: Levels in Science-Religion Dialogue 

The King's College  
Edmonton, Alberta Canada

From: PSCF 42 (March 1990): 23-33.

This paper explores some of the ways "levels of explanation" and related concepts are used by students of religion-science dialogue. Hierarchical structures of levels help formulate broad frameworks for relating religion and science. Sorting out connections between spirituality and mind-brain relations also requires well-developed levels models. Evaluating the adequacy of hierarchy theories is made more difficult and also more interesting by the divergent use of similar concepts by scholars with different world view commitments. I conclude that science-religion dialogue can benefit from paying attention to interconnections among levels of organization, levels of explanation, and levels of functioning. 

In Part I of this discussion (McDonald, 1989), I suggested that scholars interested in the relations between religion and science can benefit from digging into the notion of "levels of explanation" more systematically than has been typical. By exploring work in evolutionary biology, genetics, and epistemology the applicability of levels concepts was illustrated for broad, comprehensive contexts as well as for questions focused on a boundary between levels. However, introducing literature on "hierarchy theory" absorbed the full attention of the first paper. In this paper, therefore, I will directly examine some uses of levels notions by scholars of science-religion relationships.

The central distinction from the literature on hierarchies emphasized in Part I dealt with interactions among ontic (ontological) and epistemic (epistemological) hierarchies. Though one cannot meaningfully isolate epistemic and ontic processes, focusing on their interdependence can be fruitful. Specifically, I emphasized distinctions among levels of explanation or description (epistemic focus), levels of organization or composition (ontic levels of things), and functional levels (ontic levels of processes). The interdependence of these different kinds of hierarchies is closely connected to observation processes, reflexivity, and complementarity patterns.1

In this paper, I will start by briefly describing the status of levels notions in discussions of science and religion. Then, Richard Bube's use of hierarchies will be examined to illustrate the broad, integrative roles of levels structures. Third, Donald MacKay's notions of levels will provide an example relating spirituality to a specific boundary between levels. Finally, I will outline a few possible directions for elaborating and applying hierarchy theory in religion-science "compatibility systems."2 

Levels in Science-Religion Dialogue

Although there might not be a consensus on the matter, levels of explanation and related notions are popular among a wide variety of authors studying relationships between science and religion. Barbour's classic text (1966) makes use of the ideas, as do various popularizations (Bube, 1971; Capra, 1982; MacKay, 1979). Scholars from various backgrounds and predilections toward religion find valuable places in their compatibility systems for these notions (Arbib & Hesse, 1986; Campbell, 1974; Peacocke, 1988; Polkinghorne, 1986). What significance is there to an emphasis on hierarchy ideas by such a wide diversity of authors?

Perhaps the central reason for the widespread thinking about hierarchies is that levels patterns become important in the organization and functioning of complex systems.3 The phenomena under study in science-religion dialogue definitely reflect a great deal of complexity. That complexity reflects both: (a) the broad scope of many considerations falling under the rubric "religion-science dialogue," and (b) the intricacy of specific problems like the nature and origins of life. The literature on hierarchy theory is a valuable resource for studying such patterned complexity.

Another significant feature of levels literature is that the diversity of people interested in hierarchy notions sometimes leads to quite different applications of similar ideas. Consider for the moment two cases where hierarchy theory is employed by scholars of science-religion relations working within diverse world view commitments. Campbell's (1974) classic paper uses hierarchy theory in what can be described as a "functionalist" account of religion (because the social functions of belief systems are seen as primary). Based on his account, Campbell describes one alternative to traditional values:

Evolutionary biologists and others who are confused about the meaning of the name God can worship their Creator by worshipping the selective system that produced man, man's ecological niche. Properly approached, this could succour many aspects of man's spiritual needs now being starved under the self-worship that too often comes with loss of traditional values, and without compromising scientific scruples. (p. 184)

Campbell's naturalist commitments illustrate one framework within which the significance of levels notions is recognized, including their value in understanding "spiritual needs," values, and science.4

Capra (1982) works within a different framework when he considers levels concepts as fitting well with both Eastern religion and Western science. He examines several aspects of

the notion of multiple levels of reality which differ in their complexities and are mutually interacting and interdependent. These levels include, in particular, levels of mind, which are seen as different manifestations of cosmic consciousness. Although mystical views of consciousness go far beyond the framework of contemporary science, they are by no means inconsistent with the modern systems concepts of mind and matter. (p. 303)

This statement reflects Capra's summary after having reviewed the levels of organizational complexity in biological systems. For him, compositional levels tie in clearly with levels of consciousness as understood within a mystical perspective. In general, it can be instructive to juxtapose Capra's views with Campbell's and both of theirs with the positions of Bube and MacKay discussed below. Such a contrast clearly points out the diversity of ways that levels concepts are used by scholars of science-religion relations, while also highlighting different world view commitments among these scholars.

This point is worth developing briefly. Since hierarchy notions are recognized as important by scholars with widely differing world view commitments, levels seem to reflect basic patterns of complexity in creation and/or human knowledge systems. That is, if recognition of levels is robust under a diversity of assumptions, levels structures may well be fundamental. The presence of basic patterns does not imply, however, that hierarchy models are necessarily employed in the same way by different scholars. Rather, the examination of relationships between world view commitments and one's manner of application of levels models promises to be a valuable project for students of religion-science relations.

In support of this claim, consider the process of identifying weaknesses or strengths in levels models. Any principles developed to evaluate hierarchy models promise broad implications given the widespread interest in levels.5 Moreover, evaluating the adequacy of levels theories will necessitate distinguishing between ideas which are consensually adequate (or inadequate) and those which are adequate within some frameworks but not others. For example, in evaluating Campbell's hierarchy models, one can be alert to the possibility that certain features of his hierarchy model would be acceptable by most scholars of science-religion dialogue, while other features would not be congruent with alternative world view commitments. Features of levels models that are either world view-specific or world view-independent provide important clues about the function of hierarchy theory in the compatibility systems being studied. 

Efforts at adequately grounding levels theories can contribute 
both to specific compatibility systems and to comparisions of 
scholars' assumptions with their hierarchy models. 

By way of illustration, Campbell describes himself as a reductionist, while MacKay (discussed below; see also MacKay, 1979) sees his work as combating reductionism (called "nothing-buttery"). How are these two positions related to their respective hierarchy models and their views on science and religion? If nothing else, examining these positions suggests that levels models do not exclusively reinforce one side or the other in the classic debate over reduction-emergence. Identifying interactions between epistemic and ontic levels is one important direction to pursue in sorting out questions relating reduction and hierarchy.6

In short, examining relations between scholars' world views and their hierarchy models promises to strengthen the rigor of conceptualization. Consequently, one improves one's capacity to benefit from the work of other scholars, especially if there is disagreement on basic issues.

To summarize, the popularity of hierarchy models stems largely from their value in dealing with complexity. The different perspectives of authors studying the same complexities offer an important opportunity: the chance to develop significant evaluative principles for levels notions applied to models of science-religion relationships. Thus, efforts at adequately grounding levels theories can contribute both to specific compatibility systems and to comparisons of scholars' assumptions with their hierarchy models.7

Having briefly noted a few implications of the plurality of frameworks among scholars using hierarchy theory, I now focus our attention on two authors working within explicitly Christian world views. 

Bube: On the Levels of Nature 

Richard Bube (1971; cf., 1985) has discussed "levels of description" and a classic compositional hierarchy ("the structure of the world") as central to relations between science and religion. As a popular work, his main application of hierarchy notions is to organize the broad scope of his discussion of "the structure of the world." Given the advantage of perspectives gained by recent work in hierarchy theory and the philosophy of science, how can Bube's early work be evaluated as we further develop the dialogue between science and religion?

First of all, several aspects of Bube's 1971 presentation can be identified as strengths within the framework of contemporary levels models. He does in fact draw upon the "hierarchy theory" available at the time through an emphasis upon systems theory. Systems theory was then, and continues to be, an important source of levels models (e.g., Bunge, 1977a). Bube's description of part-whole relations in terms of "patterned interactions" fits with one way of describing systems properties that can clarify relationships between levels (cf., Barbour, 1966; 1988; Wimsatt, 1986). In one sense, then, revising Bube's work at this time is a task of updating.

A second strength of Bube's presentation is that he points out that God is not merely another level in the structure of the world: God indwells all creation. Thus, God is not simply present in one level in the structure of the world, but God sustains all of creation and provides the context for ultimate meaning of the world.8

Paying attention to levels notions can also identify areas needing additional elaboration in Bube's approach. For example, Bube addresses both descriptions (epistemic concerns) and compositional levels (ontic concerns) in his discussion, but his presentation implicitly assumes a simple one-to-one correspondence between compositional levels, levels of explanation, and disciplinary boundaries. With an additional 20 years of work addressing the interactions of epistemic and ontic processing in hierarchies, we have the vantage point to improve our consideration of these issues (e.g., Beckner, 1974; Bunge, 1956, 1977a; Salthe, 1985). For example, we can now point out that different disciplines often deal with the same levels and sometimes even with the same categories, while most disciplines deal with multiple levels within their areas of study.9

A second area where updating Bube's approach would be of benefit deals with reflexivity. Acknowledgement of observer reflexivity in a multi-leveled context requires an exploration of ontic-epistemic interactions (Salthe, 1985). The broad scope of a general religion-science compatibility system yields multiple threads of self-reference through both the scientific and the theological aspects of the enterprise. Most directly, such a compatibility system implicitly accounts for the personal integration of science and religion in the person or persons conducting the scholarly project.10 Likewise, the nature of one's personal integration has significant implications for the construction of a compatibility system (cf., my discussion above; Van Leeuwen, 1988; Evans, 1989). In a popular presentation like Bube's, one would not want to detail all interconnections and circularities, but it would be misleading to ignore the issues altogether. At the very least, the reflexive element between compatibility scholarship and personal integration needs to be addressed, especially in popular discussions. 

Levels concepts help to identify strengths and weaknesses 
of [Bube's] work for contemporary discussion, thus showing 
the applicability of hierarchies to broadly synthetic compatibility models.

 Third, more work is needed on Bube's discussion of levels related to God, theology, and human spirituality (esp. chapters 1 & 7). What does it mean for theology to be a "higher" (or subsequent) level in an epistemic hierarchy of scientific disciplines? Bube's discussion emphasizes "inclusiveness" of theology in the sense that it deals with both the cosmos and the creator, but this form of "inclusion" is not why biology is "higher" than chemistry and physics: biology does not include all of physics and chemistry within its purview.11 In effect, his comments about God not fitting into nature's compositional hierarchy needs to be expanded and applied further to theology's relationships with epistemic levels of scientific theory. It seems to me that a well developed hierarchy model will be needed to formulate the relations of spiritual domains with "familiar" levels of organization and explanation.12

In summary, Bube's discussion of science-religion relations emphasizes an inclusiveness of perspective and a breadth of scope. Levels concepts help to identify strengths and weaknesses of his work for contemporary discussion, thus showing the applicability of hierarchies to broadly synthetic compatibility models. For further illustration of hierarchy theory's applications, I now turn to the work of Donald M. MacKay. 

MacKay: On Mind-Brain Relations

 Puzzles associated with mind-brain connections provide a rich context within which to explore hierarchies in science and religion. The intertwining of religious and scientific concerns in brain-mind questions has been noted by Roger Sperry (1980), a prominent neuroscientist:

At stake [in one's views on mind-brain problems] are central key concepts that directly involve fundamental convictions regarding the nature of man's inner being, physical reality, the meaning of existence, and related matters of ultimate concern.... Perspectives in this area profoundly shape human value systems and societal decision-making and hence human destiny.13

To exemplify the significance of levels concepts in sorting out this domain, I will briefly discuss Donald MacKay's work on hierarchical models in neuroscience. His prominence in the areas of neuroscience, philosophy of science, and science-religion dialogue make his uses of levels notions particularly appropriate for the present discussion. My first step is an overview of selected papers to illustrate the role played by levels notions in MacKay's formulation of mind-brain relations.14 Then, I will highlight ways that hierarchy concepts can contribute to understanding and evaluating MacKay's views.

MacKay, calling his position on mind-brain "comprehensive realism," claims that human beings are "a unity with logically complementary mental and physical aspects, which must be held together and reckoned with as equally real" (1982, p. 293). Basing his analysis on "information engineering" categories, MacKay distinguishes between different "levels of determination," variously labeled "levels of causal analysis," the "categorical level of [one's] question," or "levels of explanation" of a "conceptual hierarchy." He does not develop a complete hierarchy model in these discussions. Nevertheless, by pulling together scattered comments, one can identify some details of the hierarchical patterns to which he refers.

When considering human agency (the topic under discussion in mind-brain problematics according to MacKay), a physical level of brain activity precedes an informational level. The latter describes the form of a physical activity or a state of affairs as influences on the form of other physical activities.15 MacKay's example of forms is the relation between the firing pattern of a cortical neuron and the spatio-temporal form of the retinal image. The informational level precedes a mental level (identified by processes of conscious human choice, for example), and a spiritual level is subsequent to the mental. These levels are characterized by mutually exclusive standpoints or categories, yielding a hierarchical version of logical complementarity between levels (cf., Haas, 1983; MacKay, 1958). So, at the heart of comprehensive realism is a hierarchical model which includes informational and spiritual functioning as well as biological and mental activity.

MacKay's contrasts of his own views with those of other scholars express the heart of his approach.16 He sees his position as a "middle way" between the dualist (or pluralist) interactionism of Popper and Eccles (1977) and the "emergent materialism" of Bunge (1977a). He also sees his views as closely aligned with those of Roger Sperry (1980; MacKay, 1980). What distinctives come to the fore from this set of contrasts? MacKay likes to describe conscious mental activity as "embodied in" brain activity, thereby suggesting that the interaction between levels is "more intimate" than is the cause-and-effect view of interactionism, and is more complex than is the identity relation of brain and mind proposed by materialism. Although a detailed theory of embodiment was not presented in the papers reviewed here, MacKay does describe a higher level activity influencing lower level processes through constraints which shape the form of lower level activity (i.e., through "formal causality").17 Likewise, lower levels influence upper levels largely through material or efficient causes. For example, mental activity directs (formally causes) the pattern (form) of growth in the developing brain and certain drugs (efficiently) cause certain aspects of mental activity.

The advantage of an embodiment view, then, is that "we can find a place for all the existing evidence without having to deny the completeness-in-principle of physical explanations in their own categories, or the determinative efficacy of mental activity" (1982, p. 292). MacKay sees dualism as inappropriately sacrificing the completeness of physical explanations and materialism as ignoring the causal efficacy of mental activity per se. MacKay's discussion clearly makes use of hierarchy concepts to distinguish embodiment from interactionism and materialism. 

Close attention to levels structures in MacKay's work 
highlights both their central place in his approach 
and their value in critiquing and extending his model

 MacKay's emphasis on the "equally real" status of mental and physical activities provides one starting point for exploring the approach he offers. His discussion suggests that levels of determination form a functional hierarchy.18 His language clearly emphasizes "mental activity" and "informational functioning" in the context of "causal analysis." In this interpretation of MacKay's views, he is championing a functional hierarchy which: (a) dualists have confused with a distinction between entities or with compositional hierarchies, and (b) materialists have collapsed into a unilevel ("reductionistic") ontology.19 In this view, the logical and epistemological concerns emphasized by MacKay reflect this basic ontic pattern of processes. This interpretation of his stance has the advantage of fitting what MacKay might have intended when calling his stance "realism." As opposed to traditional instrumentalist or idealist emphases, MacKay probably wants to say that our explanatory systems reflect ontic patterns in some important ways, thus qualifying as a "realist" view.

The identification of MacKay's levels as a functional hierarchy becomes especially interesting when one recalls that there is a spiritual level of determination. In a well-formed hierarchy, the interactions between levels are of similar nature regardless of which levels are involved.1 Presumably, then, all levels of determination are "embodied" in the unitary entity: a person.20 So the point of MacKay's oft-repeated examples about computer software being embodied in hardware is not that mental activities are at the same level as software (which is at an informational level), but that the embodiment relation is analogous or identical. Presumably, he is positing further details of embodiment when he describes the formal causal control of lower levels and the material/efficient causal control of higher levels in electronic information systems. By implication, the same principles hold for any cross-level interaction in MacKay's determination hierarchy. It is meaningful, then, to ask about spiritual activity shaping mental processes and neurological processes influencing spiritual functioning. By formulating MacKay's key notions into a more systematic hierarchical model, I claim we can see some implications of the position more clearly and are thus in a better position to critique, defend, or develop his work. 

I have argued that there is a lot of work to do to adequately 
ground levels notions as they are used in religion-science compatibility models.

 Further, focusing on levels notions helps us explore ways that MacKay's views compare or contrast with those of other authors. If MacKay envisions a functional hierarchy, it is conceivable that the functions may well expand the view of hierarchy discussed by Bube. Even if we elaborate the complexity of Bube's compositional hierarchy, he does not provide an explicit integration of functional levels. Also, Pattee's interrelation of informational and structural levels at the biomolecular levels displays suggestive parallels with the biological and informational levels discussed by MacKay (see Pattee, 1979; McDonald, 1989). From philosophical literature, it would be of interest to compare MacKay's hierarchy with functional hierarchies in process thought or Dooyewerdian philosophy (e.g., Barbour, 1966, 1988; Hart, 1984). In any event, the formulation of MacKay's views within a more systematic hierarchy model does lend itself well to the exploration of relationships between his work and that of other scholars, including those interested in science-religion dialogue.

To summarize, several points emerge from an examination of levels notions in MacKay's views on mind-brain relations. First, it is clear that hierarchy concepts are central to his formulation, particularly when he contrasts his views with those of other scholars. Second, close examination of his model of levels and of embodiment is a valuable way to elaborate and/or critique his views. One possible direction for such an analysis, emphasizing a functional hierarchy, was briefly described. Third, his view of levels in mind-brain relations is one key part of a larger picture which relates spiritual, mental, and biological realms. Unfortunately, his levels notions still require significant "fleshing out" to clarify these broader implications (cf., Cramer, 1985; Orlebeke, 1977). So, close attention to levels structures in MacKay's work highlights both their central place in his approach and their value in critiquing and extending his model.21 

Summary and Implications

 In concluding this discussion of levels in science-religion dialogue, it is helpful to highlight the nature of the task undertaken. I pointed out at the beginning of Part I that many open questions remain in hierarchy theory. Throughout the discussion I ignored important complexities for the sake of delimiting a clear focus. Such a strategy is characteristic of the issues under examination. As Beckner noted in his paper on reduction in biology:

Discussions of the philosophical issues of...the relations between levels of organization...are always conducted in a context of insufficiently clarified ideas. And not only are they insufficiently clarified; they are ideas whose connections involve the most fundamental controversies of metaphysics, the theory of knowledge, and the philosophy of science. This means that when we look at the details of some reasonably definite question...we find that the ideas we need are located at the nodes of a ravelled background web. (1974, p. 163)

Since one cannot unravel or display the entire web at once, I have had to be satisfied with a narrow spotlight focusing on a few nodes of the web. That is why I have insisted that this discussion makes sense primarily in the context of many scholars addressing levels and related notions like complementarity and reflexivity. Furthermore, the continued development of adequate foundations for religion-science compatibility systems will require programmatic attention to this web of concepts. The detailed elaboration of specific research topics (e.g., neuroscience) and the comparison among topics is one essential means of elaborating the web.

Given this backdrop, then, one can summarize the present focus on levels structures. First of all, the diversity of world view commitments among scholars like Campbell, Capra, Sperry, Bube, and MacKay illustrates that hierarchy models serve many different approaches to compatibility systems for religion and science. Evaluating hierarchy models is a complex task which needs to take into account basic assumptions of scholars. That is, attempts at enhancing the rigor of hierarchy models involve exploration of potential connections between world view commitments and hierarchy concepts. One cannot expect to find simple, direct implications between a scholar's levels models and his or her world view commitments, even in the domain of science-religion dialogue. Nevertheless, elaborating foundations for hierarchy notions can facilitate the identification of any connections with world views.

A second summary topic for this paper can be drawn from Richard Bube's work. Evaluating his work illustrates the potential of levels notions for formulating comprehensively-oriented compatibility systems in religion and science. Further development of broadly focused compatibility models can be aided by tools from recent hierarchy theory. Even in popular discussions, one needs to connect spirituality, human reflexivity, levels of biological, social, and physical organization, and epistemic hierarchies of language and scientific theory. Systematically identifying interactions among functional, compositional, and epistemic hierarchies seems especially necessary when presenting comprehensive compatibility models for scholarly audiences.

A third set of lessons emerges from Donald MacKay's work on mind-brain relationships. His neuroscience research provided a focus for an embodiment model which was also used in a theory of science-religion relationships. His concepts of levels, complementarity, and the logic of scientific explanation generate a framework which deals with both mind-brain and spiritual functioning. Digging into the grounding of MacKay's levels notions not only identified ambiguities, but it also provided means for elaboration and critique of his views.18 In addition, hierarchy notions helped highlight some aspects of disagreements among different scholars in the area. The pattern of complexity in these issues warrants the application of levels models. For a focal problem involving interlevel boundaries (mind-brain), elaborating our understanding of hierarchies is of value especially when the principles developed in a limited context are related to broader issues in religion-science dialogue.

In the two parts of this discussion, I have touched on basic conceptual issues in evolutionary theory, molecular biology, cognitive science, and neuroscience, in addition to science-religion relationships. 

It seems likely that efforts at unraveling the complex web of issues associated 
with levels will continue to be worked out in the context of 
isagreements among scholars.

The justification for such rambling is the desire to trace a "thread" across diverse terrains. Specifically, I have argued that there is a lot of work to do to adequately ground levels notions as they are used in religion-science compatibility models. The thread running through the various examples involves interdependencies among epistemic, compositional, and functional hierarchies. If we accept the challenge of grounding our hierarchy notions, scholars of religion-science dialogue will work toward being consistent and systematic in addressing interlevel interaction, intra-hierarchy dynamics, and criteria for hierarchy identification. Most importantly, our models will not focus merely on multiple-level models, but we will strive to ground our understanding in the dynamic articulation of multiple hierarchies bridging human knowledge systems and fundamental patterns of creation (cf., Barbour, 1966; Simon, 1973). These issues are only a handful of nodes in the web of topics which interweaves science and religion, yet levels issues are central enough to merit our patient attention.

The reasons for working so hard to ground hierarchy models basically reflect the original motives for using these notions in the first place. Broad, integrative discussions in science and religion require tools for synthesizing diverse perspectives. Other questions (like mind-brain relations) seem to reflect a knot of difficulties associated with an interface between levels. The lesson of our use of levels notions over the past few decades is that simplistic hierarchies are inadequate to the task of understanding complexity in religion-science relations. In that regard, levels concepts join the rank of notions like complementarity, reflexivity, and observation. In closing, I would like to highlight a few directions for possible elaboration of hierarchy models in religion-science compatibility systems.

As noted above, hierarchy notions are closely intertwined with basic ideas like complementarity, observation, and reflexivity. MacKay and Salthe (1985) both emphasize this set of issues. These authors' motivation for foundational work is partly fueled by disagreements among scholars in their fields. MacKay uses a formulation of levels to describe the relations of his work to that of other neuroscientists, while Salthe's work addresses debates in evolutionary theory. Disagreements among experts can reflect either the complexity of the topic or the consequences of world view commitments, among other things. It seems likely that efforts at unraveling the complex web of issues associated with levels will continue to be worked out in the context of disagreements among scholars. Such a prospect can be exciting and beneficial as well as challenging. The work of Arbib and Hesse (1986; briefly discussed in Part I) is one instance of a challenging collaboration while studying religion-science relationships. Their work shows both commonalities and differences between one scholar with atheistic and one with theistic commitments. Work of this kind suggests one viable form for cooperation among scholars with different world views. Levels notions and associated concepts are one resource to facilitate collaboration without capitulation in the context of world view diversity. One might even be so bold as to consider disagreements like these as opportunities for the pursuit of foundational studies.22 

Multidisciplinary views are essential. 

A second direction for further work implied by digging into levels notions might possibly have gone out of focus in the meandering path presented here, but it is virtually common sense among students of religion-science dialogue. That is, multidisciplinary views are essential. Most of the concerns raised here are largely designated as the domain of philosophers, and with good reason. The resources of the philosophy of science in particular play prominent roles in the issues raised here. Nevertheless, many of the authors drawn upon here are scientists who are addressing the conceptual and methodological foundations of their work. Hierarchy issues are as much substantive scientific concerns as they are problems of philosophy; or, rather, philosophical issues are as much the concern of scientists as of our colleagues in the humanities. In my opinion, everyone will be hindered if such issues are somehow delimited as the domain of only a few disciplines or scholars. For example, the full elaboration of hierarchy models in religion-science dialogue certainly requires the benefit of extensive theological reflection. And we will continue to need scholars like Donald MacKay with multidisciplinary competencies. Most importantly, the topics raised in this paper require the corporate commitment and collaboration of many scholars with diverse backgrounds and orientations and with overlapping fields of inquiry.23

As a final example of potential contributions from hierarchy theory, I want to mention one specific debate currently receiving attention from Christian students of science. I have focused primarily on topics in biology in these papers, largely because much of the work on levels notions has addressed that area. However, levels structures are also important for compatibility models focusing on physical or social sciences (e.g., Arbib & Hesse, 1986; Polkinghorne, 1986; Manicas & Secord, 1983). As Mary Stewart Van Leeuwen (1988) recently pointed out to readers of this journal, MacKay's work has been influential among social scientists as well as among natural scientists. MacKay's views on levels notions are, along with his complementarity notions, central to defining a "perspectival" approach to compatibility systems. It seems to me that both perspectivalists and hermeneutically-oriented scholars (Van Leeuwen, 1988; Evans, 1989) will benefit from more elaborate grounding for hierarchy models. For example, Salthe's (1985) work on observer reflexivity in hierarchical systems reflects concerns of both hermeneutic approaches (interdependence of observer and observed) and perspectival views (levels of explanation). Evans' (1977) suggestion that both persectivalist and hermeneutic ("humanizer") approaches need to be taken seriously does encourage such explorations. As seems to be the case in other debates (e.g., the point about reduction noted above), levels notions may not favour either perspectival or hermeneutic assumptions. Indeed, one can hope that grounding our hierarchy models will enhance the viability of both alternatives in addition to highlighting the nature of choices involved in selecting or synthesizing these approaches.

Levels notions are certainly no intellectual panacea, but the literature of hierarchy theory is nonetheless a valuable resource for students of constructive dialogue between science and religion.



This paper is a revision of part of a paper presented at the 1988 annual convention of the ASA in Malibu. I want to acknowledge the contributions of colleagues to this paper. Vaden House provided critique and conversation which significantly strengthened both my thinking and my writing. I cannot distinguish at this point between lessons learned in conversation with him from insights gained elsewhere. For repeated encouragement and material support, I thank Harry Cook and also Henk Bestman. Comments by several people at the 1988 conference were stimulating as well as encouraging.


1 For additional background on definitions see McDonald (1989), especially the references. It is helpful to recall that a hierarchy of levels is identified by the nature of the relationships between levels (composition is the relation for traditionally designated biological levels of organization), and by the units of analysis which define the levels (cells, organisms, etc. are units for biological levels). Thus organisms are made up of cells, etc. 

2 See Wiebe (1978), for example. For the purposes of this paper, the term "compatibility system" will be used generically for any model of relations between science and religion. Thus, attempts to demonstrate an inherent conflict between religion and science would provide a model for low compatibility. Some broader convention of this kind seems important if scholars working with different assumptions about the nature of science and religion are to clearly communicate with one another. 

3See Simon (1973), for example. Clearly there are many conceptual tools for dealing with complexity other than levels concepts. Compare Davies' (1988), overview of complexity in contemporary science. My assertion of the value of hierarchy theory is by no means an exclusivist claim. 

4Campbell's paper is a classic discussion of ways that higher levels influence phenomena at lower levels ("downward causation"). His brief discussion of religion is one arena in which he sees levels notions and evolutionary theory to be of value. For more discussion of his theory of religion, see Campbell (1975). As for the gender exclusive language in this and later quotes, I have not revised other authors' words even though I strive to maintain inclusive language in my own writing. 

5In my reading, some applications of hierarchy concepts seem vague, misleading or vacuous. Moving beyond impressions to critique requires the identification of evaluative principles which have at least a minimal degree of generality. Also, questioning specific applications of levels ideas can arise as part of substantive disagreements. Debates over the contributions of MacKay's work is one example of levels concepts intertwining with substantive disagreements (cf., Cramer, 1985; MacKay, 1978b, Orlebeke, 1977; Van Leeuwen, 1988). I anticipate that systematic efforts at critiquing hierarchy formulations will bear fruit both in clarifying some debates and in sorting out complex topics under investigation. 

6Compare, for example, Barbour (1966), Beckner (1974), Bunge (1977a), and Wimsatt (1986). In fact, one benefit of distinguishing different kinds of hierarchies is the illumination of false dichotomies in the classic formulations of reduction-emergence debates. 

7 To the extent that this work draws one into formulating principles of evaluation for compatibility systems, a metatheory of science-religion dialogue is being developed. These are among the issues which I call "foundational." Most scientists are familiar with methodological studies which are often "metascientific" or foundational in an important sense. My intention here is to simply point out that these foundational concerns are not esoteric or irrelevant. Although research scientists do not spend all of their energy on methodology, such work is an important component of programmatic research. Likewise, when we sort out relations between religion and science, the devotion of significant energy to foundational work is both appropriate and necessary for programmatic development. 

8 These comments are offered as qualification of diagrams used for illustration which simply include a "God-level" above the society level (though graphically distinguished by a double line to indicate the uniqueness of the level). Clearly pointing out that God does not merely define another level is an important starting point in exploring spirituality. It is important to go beyond the denial, however, and offer constructive formulations of the relations between God and hierarchies of creation. 

9 Beyond the disciplinary question, it is tempting to assume a strict correspondence between epistemic categories and ontic levels. After all, category errors and similar logical problems reflect one's ontology quite directly. However, the simplicity or complexity of relations between knowing and ontology depends very specifically on the epistemology and ontology in question. A one-to-one correspondence would not be possible with all ontologies or epistemologies. In my view, one of the key challenges facing hierarchy theorists at this time is to develop tools for specifying in greater detail the dynamics of ontic-epistemic interaction. In this regard, hierarchy theory requires substantive input from philosophers of science. 

10 That is, one of the questions a general compatibility system can address is whether one can participate fully as a scientist and as a member of a faith community without sacrificing the integrity of either enterprise. I submit that each scholar already embodies at least an implicit answer to that question in her or his own life. The neutrality myth surrounding scholarship is untenable, in this area of scholarship most clearly. 

In a similar vein, one sometimes finds attempts to circumvent the self-reference of scholarship on science-religion relations by positing a third arena of scholarship (often philosophy) which short-circuits the circularity. For example, a simple reflexive circle might go like this: (a) science/religion (S/R) scholarship accounts for, explains, S/R relationships by developing a model of the two enterprises, including the scholarly components of their activities; (b) however, S/R scholarship is itself a scholarly enterprise; (c) thus we have scholarship developing a (delimited) model of scholarship, i.e.,. reflexive scholarship. The attempt to short circuit the self-reference involves the assertion that science/religion scholarship is neither theology nor scientific theorizing, but instead is a realm of philosophy, or some other "metascholarly" activity. This strategy harkens back to Russell's theory of types: the rule is to simply disallow certain forms of self-reference. By contrast, consider the readily established enterprises of social scientific study of the science-religion relationship on the one hand, and of the theological study of science/relgion compatibility on the other. Philosophical study does not "break" or "cut short" this loop of circularity. An infinite regress of meta-disciplines is a "solution" I find to be highly unsatisfying and artificial. In addition to the personal-professional reflexivity of religion-science scholarship, I prefer to examine the enterprise of reflexive scholarship directly instead of defining it as illegitimate in an a priori manner. It is not excessively difficult to start tracing some of the many threads of reflexivity in dialogue between religion and science. I earnestly invite others to join in this valuable area of study. 

11 Compare McDonald, 1989, and note 1 above regarding definitions of hierarchies. The coherence of the relationship between levels is precisely the source of power in hierarchical models. One might say that they operate as extended metaphors where, for example, the chemical-physiological level boundary is similar to the brain-mind boundary, the person-organization boundary, and so forth for each boundary in a hierarchy. Thus, the issue of consistency in interlevel relationships strikes at the heart of hierarchy theory. Clarity and detail will be central, then, to advances in this area. 

Bube's discussion of "soul" and spirituality can be further extended by being placed in a more extensive levels model. Looking at his definitions of life, soul, and mind illustrate the point here: "life should be considered as a systems property in a system where the parts interact according to an appropriate pattern" (p. 145); "the word soul refers to the particular systems property of the totality of the life-system then active" (p. 147, original italics); and, the "property of mind is a systems property of the totality of the subsystems that make up the thinking creature" (p. 150). The importance of distinguishing and coordinating a set of definitions like these is clear. What I think needs to be accomplished in a reasonable way even for popular presentation is to briefly outline the nature of systems properties, how they interact across levels, and how different hierarchies articulate with one another. 

For a recent discussion of human spirituality, see Benner, 1988. Although he does not develop a hierarchy model, his discussion of psycho-spiritual interaction and relationships fits readily into such a formulation. Many of his points in fact parallel Pattee's discussions of structure and information in molecular biology. 

Levels structures are also important for compatibility models focusing on physical or social sciences. 

13 (The quote is from p. 197.) At this point in his paper, Sperry is contrasting his work specifically with other neuroscientists. However, in other work, Sperry goes much further and claims that we can "get ethico-religious values from science in a prescriptive sense" (1988, p. 610). Given the similarity between MacKay and Sperry in: (a) their neuroscience models (MacKay, 1980), (b) their convictions that these models facilitate rapprochement between science and religion (Sperry, 1988; MacKay, 1979), and (c) the importance they attach to hierarchical systems, it would be particularly instructive to explore the relations between their world view commitments and their levels models in the spirit presented in a previous section of this paper. Unfortunately, such questions take us beyond the scope of the present discussion. 

14 See MacKay (1978a, 1980, 1982). I do not make an attempt to exhaustively review his ideas partly due to the volume of his work and partly due to the narrow focus of this discussion. For other evaluations of his work, see Orlebeke (1977), Cramer (1985), and Evans (1977). 

15 For some of MacKay's statements about levels structures, see pp. 601f (1978a), p. 1390 (1980), and p. 289 (1982). His discussion of the form of activity sounds similar here to Pattee's views on genetic information: see Pattee (1979) and McDonald (1989). Also, see note 17. 

16MacKay frequently uses levels concepts in formulating disagreements with other scholars. For another example which expresses disagreement on a different topic, see MacKay (1978b). 

17Formal, material, and efficient causality are used by MacKay in Aristotle's sense: e.g., Book 2 of Physics, esp. chap. 3ff. The view of upper levels influencing lower ones through constraints or boundary conditions (shape?) is similar to Polanyi's (1968), Campbell's (1974), and Salthe's (1985), discussions of cross-level interactions, among others. 

18 The description of MacKay's views in terms of a functional hierarchy is not the only way of interpreting his use of levels notions. For example, in light of his "underspecification" notion (MacKay, 1985), he may have viewed "levels of causal analysis" as epistemic levels reflecting the cognitive limitations of human beings, or of information systems more broadly. In this case, MacKay's position is ontic monism (thus involving "equally real" mental and physical activities with descriptions partitioning reality in "artificial" ways; i.e., in ways that do not reflect ontic patterns). Of course, he might also envision both epistemic and functional features in the levels of determination<197>possibly while assuming a 1:1 correspondence between functional and epistemic levels. His emphasis on logic may well reflect a tight interconnection between epistemic and ontic levels structures. Also, his labels for the levels imply both ontic and epistemic features. Other possibilities exist as well. Any complete portrayal of his views in this area would necessitate broader discussion than is feasible in this paper. Nevertheless, one's interpretations of mind-brain relations can borrow or modify ideas from MacKay and benefit from his levels concepts without necessarily doing justice to the entire corpus of his work. For the present, it is sufficient to recall that the purpose of discussing MacKay's work here is to illustrate applications of hierarchy theory, not to exhaustively exegete MacKay's work. 19 19I do not imply here that Bunge (1977a, 1977b) reduces the cosmos to a unilevel structure. MacKay's objection to Bunge's materialism seems to be that calling a hierarchy model a "materialism" results in one level being assigned a stronger ontic status in some way, thus violating his principle of equal reality of levels. (I am thankful to Vaden House for highlighting these issues to me). MacKay's reaction to Bunge's work illuminates MacKay's views, even if his descriptions of Bunge are incomplete or misleading. Bunge's work is strongly steeped in systems theory, and he is in fact an important writer in contemporary hierarchy theory. Detailed description of Bunge's views would take this discussion far afield, but the interested reader is referred to his 1980 work on mind-brain relations. 

20 It is important to be clear that the term "embodiment" cannot, in MacKay's view, imply some sort of privilege or priority on physical levels. Compare note 19. 

21 MacKay's work focuses in important ways on the interfaces between physiological, informational, and mental functioning. The perspectives developed in that context also frame his work on science-religion relations. As scholars interested in religion-science dialogue work to evaluate MacKay's hierarchical models, I think his work stands as a valuable example. He clearly highlighted levels notions, but he did not, as far as I can tell, give the same systematic attention to levels as he did to complementarity, for example. Since, of course, no one can do everything, I see this point as less a criticism of MacKay than as an invitation for scholars to pursue the implications of his work, critically and constructively. For that task, greater elaboration of levels conceptualizations seems necessary. 

22 Are we ready to nurture discussions among scholars with diverse faith commitments? For example, are we willing to dedicate journal pages to this kind of dialogue? Over the long run, a series of such articles could become very instructive. ASA has already engaged in similar activities through invitations to scholars at conferences. Perhaps we can expand contacts like these into systematic collaboration along the lines of Arbib and Hesse's work while involving multiple scholars over time. 

23Close attention to epistemic levels structures may also contribute to a framework for fruitful interactions among disciplines. Such a metatheory of interdisciplinary dialogue is another species of what has been called here a "compatibility system." After all, religion-science dialogue is primarily an attempt to explore what scientific theory and theology have to say to one another. Generalizing this notion slightly leads to a model of interdisciplinary dialogue.


Arbib, M. A. & Hesse, M. B. (1986). The Construction of Reality. Cambridge: Cambridge University Press. 
Barbour, I. G. (1966/1971). Issues In Science and Religion. New York: Harper & Row. 
Barbour, I. G. (1988). On two issues in science and religion: A response to David Griffin. Zygon, 23, 83-88. 
Beckner, M. (1974). Reduction, hierarchies and organicism. In F. J. Ayala & T. Dobzhansky (eds.), Studies in the Philosophy of Biology: Reduction and Related Problems. Berkeley, CA: University of California Press. 
Benner, D. (1988). Psychotherapy and the Spiritual Quest. Grand Rapids, MI: Baker. 
Bube, R. H. (1971). The Human Quest: A New Look at Science and the Christian Faith. Waco, TX: Word. 
Bube, R. H. (1985). Reductionism, preductionism and hierarchical emergence. Journal of the American Scientific Affiliation, 37, 177-179. 
Bunge, M. (1956). Do the levels of science reflect the levels of being? In M. Bunge, Metascientific Queries. Springfield, IL: Charles C. Thomas. 
Bunge, M. (1977a). Emergence and the mind. Neuroscience, 2, 501-510. 
Bunge, M. (1977b). Levels and reduction. American Journal of Physiology, 233(3), R75-R82. @REFERENCES = Bunge, M. (1980). The Mind-Body Problem: A Psychobiological Approach. New York: Pergamon. 
Campbell, D. T. (1974). "Downward causation" in hierarchically organized biological systems. In F. J. Ayala & T. Dobzhansky (eds.), Studies in the Philosophy of Biology: Reduction and Related Problems. Berkeley, CA: University of California Press. 
Campbell, D. T. (1975). On the conflict between biological and social evolution and between psychology and moral tradition. American Psychologist, 30, 1103-1126. 
Capra, F. (1982). The Turning Point: Science, Society, and the Rising of Culture. New York: Simon and Schuster. 
Haas, J. W. Jr (1983) An assessment: 2. Logical complementarity. Journal of the American Scientific Affiliation, 35, 203-209. 
Hart, H. (1984). Understanding Our World: An Integral Ontology. Lanham, MD: University Press of America. 
McDonald, M. J. (1989). Exploring "levels of explanation" concepts, Part I: Interaction of epistemic and ontic levels. Perspectives on Science and the Christian Faith, 41, 194-205.
MacKay, D. M. (1958). Complementarity II. The Aristotelian Society Supplement, 32, 105-122.
MacKay, D. M. (1978a). Selves and brains. Neuroscience, 3, 599-606.
MacKay, D. M. (1978b). Responsible mechanism or responsible agent? A response to William Hasker. Christian Scholars Review, 8, 141-148.
MacKay, D. M. (1979). Human Science and Human Dignity. Downers Grove, IL: InterVarsity Press.
MacKay, D. M. (1980). The interdependence of mind and brain. Neuroscience, 5, 1389-1391.
MacKay, D. M. (1982). Ourselves and our brains: Duality without dualism. Psychoneuroendocrinology, 7, 285-294.
MacKay, D. M. (1985). Artificial intelligence? A Christian appraisal. In R.L. Brabenec (ed.), I. Wheaton, IL: Wheaton College and the Association of Christians in the Mathematical Sciences.
Manicas, P. T. & Secord, P. F. (1983). Implications for psychology of the new philosophy of science. American Psychologist, 38, 399-413.
Orlebeke, C. J. (1977). Donald MacKay's philosophy of science. Christian Scholar's Review, 7, 51-63.
Pattee, H. H. (1979). The complementarity principle and the origin of macromolecular information. BioSystems, 11, 217-226.
Peacocke, A. (1986). God and the New Biology. San Francisco: Harper & Row.
Peacocke, A. (1988). Science and theology today: A critical realist perspective. Religion and Intellectual Life, 5, 45-58. Polanyi, M. (1968). Life's irreducible structure. Science, 160, 1308-1312.
Polkinghorne, J. (1986). One World: The Interaction of Science and Theology. Princeton, NJ: Princeton University Press.
Popper, K. R. & Eccles, J. C. (1977). The Self and Its Brain: An Argument For Interactionism. Berlin: Springer International.
Salthe, S. N. (1985). Evolving Hierarchical Systems: Their Structure and Representation. New York: Columbia University Press.
Simon, H. (1973). The organization of complex systems. In H. H. Pattee (ed.), Hierarchy Theory: The Challenge of Complex Systems. New York: George Braziller.
Sperry, R. (1980). Mind-brain interaction: Mentalism, yes; Dualism, no. Neuroscience, 5, 195-206.
Sperry, R. (1988). Psychology's mentalist paradigm and the religion/science tension. American Psychologist, 43, 607-613.
Van Leeuwen, M. Stewart (1983). Reflexivity in North American psychology: Historical reflections on one aspect of a changing paradigm. Journal of the American Scientific Affiliation, 35, 162-167.
Van Leeuwen, M. Stewart (1988). North American Evangelicalism and the social sciences: A historical & critical appraisal. Perspectives on Science and Christian Faith, 40, 194-203.
Wiebe, D. (1978). Science and religion: Is compatibility possible? Journal of the American Scientific Affiliation, 30, 169-176.
Wimsatt, W. (1986). Forms of aggregativity. In A. Donagan, A. N. Perovich, Jr., & M. V. Wedin (eds.), Human Nature and Natural Knowledge (vol. 89 in the Boston Studies in the Philosophy of Science). Dordrecht: D. Reidel.