Interactionism and The Non-Obviousness of Scientific Theories

Robert N. McCauley

Department of Philosophy

Emory University

Atlanta, Georgia 30322

philrnm@emory.edu

E. Thomas Lawson

Department of Religion

Western Michigan University

Kalamazoo, Michigan 49008

e.thomas.lawson@wmich.edu

I. Introduction

Levine's discussion of Rethinking Religion (1990) and "Crisis of Conscience, Riddle of Identity" (1993) includes some rash charges, some useful comments, and some profound misunderstandings. The latter, especially, reveal areas where we need to clarify and further defend our claims. In the second section we shall discuss the epistemological and methodological issues that Levine raises. Then we shall turn in the third section to theoretical and substantive matters. In fact, Levine remains almost completely silent on substantive matters (except to say that our claims are "obvious" and "trite.")

Levine claims, in effect, (1) that religion is outside of the scope of scientific analysis, (2) that our competence approach to theorizing is not necessary for generating the theoretical claims that we make, and (3) that the substantive consequences of those theoretical claims are obvious and trivial. We unequivocally reject the first and third claims and, Levine's profound misunderstandings about the competence approach to theorizing notwithstanding, completely agree with the second. Identifying the confusions in Levine's discussion that inform item (3) will clarify our position. We turn first, though, to matters of epistemology and method (as these bear on items (1) and (2)).

II. Epistemological and Methodological Issues

We welcome a discussion of epistemological and methodological issues, because they are so often ignored in the academic study of religion. We have argued that the history of religions is in crisis--a fact most recently acknowledged by Catherine Bell (1996). If our diagnosis is correct (as Levine, himself, concedes), then it behooves the discipline to come to terms with the underlying issues.

It is by now no secret that we have strong views about these matters. Oddly, Levine repeatedly cites passages from our work that straightforwardly and emphatically belie his most strident charges. Examples include Levine's suggestion that we trust (naively) in informants' pronouncements. In fact, all that we argue is that informants' judgments are part of the data to be employed in our investigation. Nowhere do we regard them as determinative. (See Lawson and McCauley 1990, pp. 183-184--which Levine cites on p. 9!)

More serious examples, however, include Levine's charge of scientism, which, presumably, entails his further accusation that we regard the study of religion within the humanities as a "waste of time." These charges are false. We nowhere even hint that we hold that the sciences exhaust our ways of knowing. We only hold that as an activity for gaining knowledge, science is second to none. Furthermore, although we have challenged the standard rationales for situating religious studies within the humanities, we have never denied that humanistic approaches frequently reveal a great deal about religion. We have only rejected the claim that such approaches exhaust the subject.

Exclusively humanistic approaches fail to address important epistemic goals--the formulation of general[1], testable, explanatory theories, in particular. Part of the reason for this failure concerns the preoccupation of scholars in the humanities with interpretive pursuits. They either subordinate explanation to interpretation or (as Levine frequently seems to advocate) they exclude explanatory endeavors entirely. We have argued that it is ill-advised to propound any form of exclusivism or subordination. We repudiate all forms of exclusivism--hermeneutic, phenomenological, and explanatory. We have consistently held that the interaction of interpretive and explanatory projects is a necessary condition for the growth of knowledge. Both Rethinking Religion and "Crisis" explicitly aim to redress an imbalance in research on religion, which is, in fact, overwhelmingly devoted to interpretive pursuits. It is only in this sense that "we have reversed the usual order of significance between explanation and interpretation in the study of religion."[2]

We do not deny the possibility that in specific instances scientific methods may generate findings that conflict, at least prima facie, with the assumptions of various non-scientific approaches to religious materials, but this does not mean that scientific and non-scientific approaches are necessarily "at odds." Such matters must be decided on a case by case basis.

Along these lines, Levine asks for an example of how cognitive science might conflict with phenomenology. Fair enough. Although space limitations preclude extensive discussions, it seems to us that the discoveries of blindsight (Weiskrantz 1986) and Anton's syndrome, i.e., blindness denial (Churchland 1983), are two examples. Blindsight victims have cortical scotomas--lesions in their primary visual cortices. Although they claim to be blind in parts of their visual fields, they prove to have a kind of conscious access to information from those fields when requested to speculate. Conversely, victims of blindness denial, who have suffered trauma or lesions (of a different sort) in their visual cortices claim to have conscious access to visual information that they demonstrably do not have. When the condition is only temporary, patients are unaware that they have recovered from anything! Critically, the intellectual capacities of the victims of these disorders are in the case of blindness denial largely unimpaired or in the case of blindsight essentially normal. Nothing bars them from carrying out the intellectual exercises characteristic of the phenomenological method, however, in both cases they are systematically mistaken about the quantity and character of information to which they have a type of conscious access. Prima facie, these persons pose problems for those who hold that carrying out the appropriate phenomenological analyses renders a descriptively perspicuous account of the contents of consciousness accessible.

We cannot leave this matter, however, without noting a further puzzling feature of Levine's claims on this front. Insisting that the cognitive sciences do not address the "subjectivity and 'architecture of consciousness' that phenomenology investigates," Levine, nonetheless, maintains that phenomenological proposals "may be informed" by the cognitive sciences. (p. 4) It is difficult to see, however, how the findings of the cognitive sciences can "inform" phenomenological claims yet be incapable of conflicting with them (ever?).

Levine's comments reveal both genuine disagreements with and significant misunderstandings of our positions. A disagreement first. Levine asserts that the methods of the physical sciences do not apply to religion. To pursue them "prevents useful inquiry and knowledge." Although we do not hold out any great hopes along these lines, we are pragmatists about such matters. If some modelling strategy from the physical sciences offered insights about human or religious materials, we see no principled reason for rejecting such a strategy. Recent attempts to apply dynamical systems theory to human cognition (for example) are a good deal more suggestive than Levine might imagine.[3] (Van Gelder 1995) Levine insists that the history of religion is "too multi-faceted to be subject to reductionistic programs." We agree that religion is multi-faceted and no more subject to reductionistic explanation than anything else of comparable complexity. On the other hand, we see no reasons (beyond theological ones) for holding that religion is any less susceptible to reduction than comparably complex systems either!

Note, however, that Levine does not even consider whether scientists other than physicists might have developed methods and strategies appropriate for human affairs. His says little or nothing about the long history of social scientific investigation that does not attempt just to mimic the physical sciences. And what about the cognitive sciences' attempts to account for aspects of human mentality and behavior?

Levine makes repeated pronouncements (without a shred of argument that we can find) about religion's insulation from scientific inquiry. He holds that when we study religion we are "dealing with a different beast." Religious phenomena are " . . . outside the methodological, theoretical, and practical scope of such sciences. This is . . . due . . . to the nature of religion." (p.4)

Levine accuses us of attacking figures who justify such defensive postures concerning religion with arguments that are no longer current. He asserts that contemporary figures in the study of religion no longer privilege either their data or their methods. (We hasten to add that concerning the former, at least, Levine himself is a glaring exception to this pronouncement, as the citations in the previous paragraph reveal.) When we cite an example from a contemporary historian of religion, Charles Long, Levine summarily dismisses him as "basically a theologian, not an historian of religion." This might prove surprising not only to Professor Long but also to the various programs in the history of religions in which he has served. But let us look at another example.

William Paden in his Interpreting the Sacred (1992, p. 10, some emphasis added) says,

. . . while causal explanation is the foundation of science, it has limitations with regard to accounting for the subject matter of religion. . . . The antecedent condition of every human act is an activity of neurons. All religious life is influenced by social contexts. Do these fully explain . . . religious experience; or do they just explain preconditions of these phenomena? While such explanations may themselves have meanings in the world of the explainer they are clearly a different kind of interpretive approach from one which tries to show what religious objects signify to people or even what they signify religiously.

His unfounded fears about "fully" explaining something, notwithstanding, Paden does not completely insulate the "preconditions" of religious phenomena from scientific inquiry and, in fact, very graciously credits us (1992, p. 139) with taking explanatory theorizing about religion seriously without reducing it to the positivistic rules of empirical verification. Nevertheless, Paden does want to protect some core of religious phenomena from scientific explanation. This still looks to us like privileging some of the data from the prying eyes of science.

So, when Levine asks where contemporary figures claim special privilege for religion, the answer lies in reading what they have said. As we have noted (Lawson and McCauley 1993, pp. 207-208 and McCauley and Lawson (1996)), the new way that many scholars of religion make this move is to appeal to even more ambitious assertions by humanists for insulating human matters generally from the analytical scalpel of science by employing phenomenological, hermeneutic, or post-modern maneuvers. Paden's work, for example, appeals to the uniqueness of interpretations. Although their comments may seem less parochial than those of their quasi-theological predecessors, these scholars have just substituted a special core of human subjectivity for what Levine calls "a special core of religious subjectivity." (p. 3)

Because of the sciences' alleged inability to capture religion's "nature," historians of religion generally have and, according to Levine, should " . . . look for interpretive categories that can help make sense of religion." His undefended claims about the "nature" of religion aside, surely, Levine must recognize how much his own position ends up resembling the hermeneutic exclusivism that we criticized in Rethinking Religion. Yet with dazzling aplomb in the very next paragraph Levine states that we " . . . are right in claiming that the exclusive adoption of the hermeneutic method is a mistake" (p. 4). It is from making just this mistake that Levine's principal motive emerges for his repeated protests that our theory does not provide insight about the nature, meaning, or function of ritual. It is fairly clear from his comments that he thinks that these are the only questions about ritual that are worth asking. For example, he says, what our theory "does not tell us about is the nature of ritual, its function, or what it means. Simply put, it does not tell us anything about ritual that is of interest." (p. 8) Levine not only presumes to know what must be of interest in advance, he is, apparently, incapable of imagining anyone having legitimate interests that differ from his own.

Closely related, we suspect, are Levine's numerous puzzling comments about the cognitive sciences generally[4] and about the competence approach to theorizing in particular. For example, he (p. 6) seems to deny linguistics a place among the cognitive sciences--a surprise, no doubt, to those linguists who have been directing cognitive science programs (for example, at SUNY Buffalo and Ohio State)! Levine also provides ample evidence that he does not understand competence modelling. Consider his statement that "the 'competence approach' . . . involves simply asking participants about the acceptability of certain performances" (p. 12). This is not true about either linguistic theories or our own.

What we are modelling is participants' tacit knowledge of their religious ritual systems that their judgments about a wide range of features of ritual acts reveal. Those features include not only a ritual act's general acceptability, but its comparative centrality[5] in the religious ritual system, its possible repeatability, its possible reversibility, and (we now believe) its ability to include ritual substitution. Ultimately, the model accounts for and makes fairly precise predictions about participants' judgments on all of these fronts that have, so far as we know, never been connected systematically in a single theory before. (To anticipate a bit, these hardly seem "mundane" or "obvious" matters!) Moreover, in work in progress, we argue that our theory provides insights about the transmission of religious systems over time, about the place of emotional stimulation in religious ritual, and about conditions sufficient for the outbreak of ecstatic movements within religious systems. (McCauley and Lawson (in progress))

Typically, this knowledge participants' possess is tacit and unconscious. What piques our interest is the possibility that such knowledge might be rule governed (or, at least, perspicuously organized by a system of generative rules) without participants actually following such rules consciously (or, perhaps, even unconsciously). As John Searle (1969) has noted, following rules and being governed by rules are not the same things. Contrary to Levine (p. 7), showing that participants' judgments about their ritual systems are rule governed does not require the presumption that those participants are following rules.

Levine also asserts (p. 12) that the system of rules (characterizing the action representation system) does not generate the structural descriptions of rituals but follows from them. We hardly know what to say here. As a matter of both historical fact and formal procedure (depending upon what sense of "follows from" Levine intends), this assertion is patently false. Nor does it follow, contra Levine (especially p. 9), from the fact that we do not adopt Chomsky's strong claims concerning task specificity and innateness for the system of representation we describe that a competence approach will be unhelpful in characterizing general structural principles underlying religious rituals. Levine is just wrong to think that either task specificity or nativism are necessary accoutrements of a competence approach to theorizing.

But what about meaning? In the fifth chapter of Rethinking Religion, we are quite explicit that, in formulating the formal system for the representation of action, we do not address (directly, at least) issues of meaning or function (in contrast to issues of structure).[7] Although our theory focuses primarily on regularities in religious ritual systems that hold largely independently of the meanings attached to them by participants (or scholarly interpreters), it is not as if we do not discuss meaning at all. Levine seems to have overlooked the entire sixth chapter of Rethinking Religion. There, in response to the semantic eliminativism of Staal and Sperber, we make considerable use of research on concepts and categorization in cognitive psychology over the past two decades to explicate respects in which religious symbols might be said to mean. Our purpose in attending to those resources was to analyze the role of religious meaning(s) in our larger cognitive economies. Levine's further claims (p. 20, fn3) that we "do not think that the study of religion should be neutral with regard to religious truth claims" and that we see the "falsity of religious claims" as "integral to . . . [our] scientific approach" are also completely false. We state explicitly on the first page of Rethinking Religion that "most issues concerning the explanation of religious behavior are completely orthogonal to the truth values of religious beliefs."

The deeper issue, though, is whether questions of meaning and function exhaust "what is interesting about ritual." We think not. Unlike Levine, we make no claims about the "nature" of religious ritual. The structural relations our theory uncovers, however, do reveal much about its character. The evidence for that is the theory's continuing success at making sense of religious ritual materials from a wide array of religious systems. (See, for example, Maguire 199X)

Levine's astonishing claim that explanatory principles can be "theoretically inspired, systematically related, empirically testable, and general" and yet "still be uninformative" (p. 8) discloses just how narrow the range of questions about religion he deems legitimate is and who really holds "an extraordinarily narrow view of religion." Because our theory does not directly address Levine's worries about meaning(s), it certainly does not follow that it is uninformative. (Actually, consistent with our interactionism, we do not concede that our theory is completely irrelevant to interpretive matters, but, alas, space limitations again require that that must remain a battle for another day.) In short, Levine advocates a hermeneutic exclusivism first class where only interpretation can deliver informative goods.

We find all of this quite puzzling in the light of Levine's concessions that we are correct not only about the riddle of identity that plagues the history of religions but also about that riddle's origins in confusions about method. It seems, after all, that Levine is advancing the same sort of exclusivistic hermeneutic defense of the standard approaches to religion that we have repeatedly criticized in our work. The only difference, so far as we can see, is that, unlike his predecessors in the study of religion who offered covertly theological grounds for such conclusions, Levine offers none.

III. Theoretical and Substantive Matters

Our most basic substantive disagreement with Levine concerns the relative importance of presumptions about culturally postulated superhuman (CPS) agents in religious systems. We maintain that the single most important variable predicting participants' judgments about the features of religious rituals that we discuss is the position of CPS agents in the structural descriptions of ritual acts. The importance our theory accords CPS agents certainly coincides with pretheoretic intuitions--which, surely, ought not be held against it! The more fundamental consideration, however, is the theory's on-going empirical success.

By contrast, Levine holds that "[a]part from its role in a particular symbolic-cultural system embodying a worldview and ethos" (hardly an insignificant qualification, we might add!) "belief in superhuman agents is relatively inconsequential to religion" (p. 5). The only support for this assertion that Levine offers is that it is "the dominant view in religious studies." (p. 5)

We strongly suspect that such liberality about the extension of "religion" causes more problems than it solves. Construing religion as "ultimate concern," for example, will eventually force scholars to wonder whether playing the piano passionately, playing politics cunningly, or even playing golf habitually constitute religion. Indeed, the challenge becomes specifying what might not ever qualify as religion!

To say, as Levine does (p. 5), that religion is a cultural system is progress. But inquiry surely cannot stop there, for one of the most important questions remaining concerns what is distinctive about religious systems as a particular type of cultural system. What, after all, is the justification for departments of religious studies separate from anthropology, if it is sufficient to describe religious systems as nothing more than cultural systems?

Once we have theories on the table, squabbling about definitions is a nearly useless enterprise. Whether CPS agents turn out to be central to an account of religious systems will depend upon the relative empirical success of the theories, like ours, that accord them prominence. No other theory that we know of has so systematically explored the various consequences of this claim about the prominence of CPS agents for either ritual form or ritual systems. Our theory will only contribute to the "correctness" (cf. Levine p. 15) of this commitment about religion, though, through testing both more and more penetrating empirical consequences of the theory that that commitment motivates. In light of those pretheoretic intuitions and the range, depth, and quantity of the empirical evidence we and others (e.g., Abbink 1995) have cited in support of our claims, it seems to us that the burden of evidence must be born by our liberal critics within religious studies. To undo our proposal completely, though, those critics also owe the discipline a comparably coherent, theoretical account of their own enterprise! Objections are never enough to justify dumping a theory. What is also required, throughout the sciences, is an alternative theory.

Nonetheless, if it would placate our critics, we would happily surrender the claim that we are theorizing about religious ritual systems, conceding, instead, that we are only talking about those systems of ritual acts that are connected to conceptual schemes that include CPS agents among the class of possible ritual participants. Regardless of what they are called, though, these ritual systems are not adequately characterized by merely declaring that they are cultural systems. We maintain that such systems constitute a distinct class of ritual systems for which our theory discloses all sorts of systematically connected properties that largely turn on the positions and roles of these CPS agents in the structural descriptions of the particular rituals in question. Therefore, on our view clear grounds exist for a separate discipline devoted to their study.

Of course, adopting that interpretation of our theory will not placate our critics, because it still does not address what most annoys them. That is that we presume to explain at least some features of religious systems while employing a theory that makes no essential appeal to any interpretive details about anything other than the roles attributed to CPS agents in rituals.

This seems the most likely explanation for Levine's willingness to take what are, in effect, contradictory stands in his assaults on our position. For he holds at one point that our hypothesis concerning the prominence of CPS agents is false ("belief in superhuman agents is relatively inconsequential to religion" (p. 5)) yet later he asserts that it is trivially true (it "illustrates the obvious" to insist, as we do, that "commitments to supernatural beings . . . are significant for ritual" (p. 15)). Overall, it looks as if the latter complaint prevails. Levine's repeated assertions that our theoretical and substantive claims are "obvious," "mundane," "superficial," and "trite" indicate that, finally, what irks him is not that our theory makes false claims but rather that it makes insipid ones.

In formulating our reply we should note another of Levine's complaints that is closely connected with his charges that the truths our theory captures are obvious. Levine maintains that the structural descriptions of rituals that our formal apparatus generates "obfuscate instead of clarify" these allegedly mundane truths about rituals' structures so much so that those diagrams might be mistaken as "a parody" (fn 6).

It was certainly not our intention to obfuscate. We should, however, note from the outset that the vast majority of the information that particular diagrams capture is commonplace and that understanding the substantive import of those diagrams (even taken collectively) does not require mastery of the formal apparatus we developed. But it helps! More importantly, no one who understands our theory would be tempted to mistake those diagrams as a parody, for collectively they perspicuously illustrate all of our claims about how structural considerations generally and those pertaining to CPS agents especially determine a wide range of features of the rituals in question and participants' judgments about them.

As a matter of fact, the formal system and the diagrams it generates introduced a precision to our descriptions that enabled us to see clearly how rituals' general action structures and the roles attributed to CPS agents' suggest systematic principles for predicting a number of those rituals' features as well as participants' judgments. We were not the first to notice that some rituals presuppose others or that some rituals get repeated while some do not or that some rituals but not others are reversible or that some rituals but not others permit substitutions or even that rituals differ in their perceived centrality to a religious system. So far as we know, however, our theory is the first to provide a unified account of all of these (and other) patterns, and it is the diagrams that permit us to elucidate the underlying theoretical principles precisely.

The principal reason that Levine offers for why our theory captures only mundane and obvious truths is that he thinks that the religious conceptual scheme and the contexts of ritual performances jointly account for religious rituals' unique features. He asserts (p. 15) that "since the only thing special about ritual as action is its conceptual scheme, any explanation of ritual action will have to focus on that rather than on action." Levine claims, for example, that the interpretation of the Christian blessing we analyze in Rethinking Religion turns on details of the Christian conceptual scheme that are wholly "independent" of our structural description of rituals.

At one level all of this is surely true. Our theory, however, aims to establish (1) that many aspects of participants' understandings of their religious rituals (for example, their representations of their underlying structures) do not turn on any uniquely religious considerations and (2) that many features of religious ritual systems (for example, systematic distinctions between various types of religious rituals) depend fundamentally upon interactions between a religious conceptual system and thoroughly ordinary features of cognition (including the possession of resources sufficient for the representation of action). Levine, therefore, is correct to note that we see religious ritual as "action structured like any other" (p. 10) and that, for example, the object agency filter does not explain "ritual per se" but applies to the representation of action generally. The action representation system we describe does pertain to the cognitive representation of any action, however, the resulting representations helped us to see the systematic links between at least a half dozen, previously unconnected features of religious ritual.

What we are trying to formulate is a theory of religious ritual materials that does not depend on interpreting the myriad, unique details of each religion's conceptual scheme. That claim, though, does not deny that the details of religious conceptual schemes (or the contexts of ritual performances) will illuminate a good deal about religious rituals. That, pace Levine, is a point that we have asserted repeatedly and at length! (Lawson and McCauley 1990, pp. 93-95 and 157-158 especially) Details always matter (Lawson and McCauley 1995), but behind knowing which ones count and when stands all the differences between undisciplined wandering and theoretically informed investigation.

The principles we have identified can account for much about both a ritual's form and its status within the larger ritual system, regardless of interpretive variations across religious conceptual systems. Making the case for religious ritual as action--structured like any other--is pivotal to supporting our theory's generality. Knowing no more than how CPS agents (that are characterized within the religious conceptual scheme!) figure in a structural description, we can explain participants' conceptions of (1) various constraints on well-formedness and possible variation, (2) the role of ritual presuppositions, (3) rituals' comparative centrality (as variously measured), (4) ritual repetition, (5) ritual reversal, and (6) (we now think) ritual substitution. Although our theory does not explain a lot of things about religious ritual, Levine is just wrong when he says (p. 13) that it cannot explain the fact that water plays the role that it does in the Christian blessing. It is explained by the fact that the entry for the water in the structural description dominates the highest level entry for a CPS agent. Less formally, the water is critical for the blessing, because (in this case) it has the most direct connection to a CPS agent.

Levine affirms our claims that ritual systems, like languages, are rule governed, that participants demonstrate mastery of a shared body of cultural knowledge, and that participants can easily form a wide array of intuitive judgments about many features of their rituals. Levine again notes, though, that all of this is obvious.[8] "To identify ritual activity one must already know these things, and those who have sought to explain ritual simply assume them." (p. 8) These observations are "no news" (p. 8).

Again, at one level Levine's claims here are true. But what is to be made of his question "how could ritual possibly not be rule governed?" (p. 7) in the face of his apparently total lack of interest in proposals about the specific rules involved (let alone the universal constraints on them)? But, perhaps that is unfair. Maybe it is just our proposal that Levine disdains--because of its putative obviousness. Levine's repeated charges about both the dispensability of our method and the reputed superficiality and obviousness of our theory as well as his comment about our alleged analogy "between ritual systems and linguistics" (p. 7) all indicate that confusions on two matters have led him to seriously overestimate exactly what is and is not obvious about our position.

The first confusion concerns a failure to distinguish between participants' judgments and the theoretical principles that account for those judgments. The obviousness of the first does not entail any obviousness about the second. Just because the judgments are (sometimes) manifest (at least to other participants) does not mean that the underlying explanatory principles are! Typically, the grammatical acceptability of a sentence (or the lack thereof) is transparent to native speakers of a language, even when they cannot articulate a single principle of their grammars. So, while it may be true that it is "no news" that well-formed rituals are well-formed and that this is clear to both participants and researchers, the question of what is to be made of these facts remains. How are they to be explained? To answer that question requires the development of a (non-obvious) theory.

The second confusion concerns what, precisely, the critical explanatory principles of our theory are and whether or not they are non-obvious. It is critical to distinguish the action representation system (and the structural descriptions it generates) from the two functional universals we identify, viz., the Principles of Superhuman Agency and Superhuman Immediacy, and our comments about their implications for ritual systems (summarized in figure 17 of Lawson and McCauley 1990, pp. 128-130).

As we have already noted (see p. 16 above), the action representation system mostly captures relatively commonplace observations about the representation of action. Levine himself emphasizes that it says nothing about what is peculiar to religious ritual! It is not the locus of the critical explanatory principles of our theory, though the structural descriptions it generates provide the materials for their illustration.

Perhaps, our two functional universals were also transparent to Levine. If that is so, we admire him for seeing immediately what took us months to figure out. We freely admit that the basic import of these two principles is not all that difficult to state (though note that the clarity of a claim is not the same thing as the obviousness of a truth). In short, the roles attributed to and the relative proximity of CPS agents in religious rituals turn out to be the critical variables that explain participants' tacit knowledge of a wide array of their features.[9] But again, what Levine now suggests is obviously true about our position is what he had earlier seemed to regard as most obviously false about it!

The Principles of Superhuman Agency and Superhuman Immediacy are neither true nor false obviously. If they are true, as we suspect, their truth is utterly non-obvious. But because they are clear, we can trace their testable consequences, and because they have such consequences, we can collect evidence for their truth, their non-obviousness aside. This is the way things work with scientific theories. We are profoundly skeptical that the implications of those principles that we summarize in the final ten pages of the fifth chapter of Rethinking Religion were obvious to anyone. (Lawson and McCauley 1990, pp. 126-136) To repeat, although we did not discover the patterns we note, so far as we know, no other theory has systematically organized them, let alone in such an economical fashion. To repeat, the predictions that follow from the theory on these fronts are eminently testable. Levine basically ignores them.

On at least five occasions Levine does raise the possibility of finding counter-examples to various positions we endorse:

(1) "Perhaps with completely new rituals or religions, principles underlying older rituals will not be generative." (p. 8),

(2) "It would be useful to discuss cases--and I am sure that they can be found--where objects qua objects have agency." (p. 13),

(3) "There are probably exceptions" to the theory's predictions about particular rituals' "relative centrality in a religious ritual system." (p. 16),

(4) "Exceptions" to the priority of the Principle of Superhuman Immediacy "would . . . prove more informative than the generalization." (p. 17), and

(5) "Rituals that involve gods only indirectly may in some ways be more prominent and important to the system than those directly involving them." (p. 17, emphasis added)

We end with a few brief comments. First, by suggesting that such counter-examples are possible, Levine recognizes that our theory is empirically testable. Second, we have already addressed the second sort of alleged counter-example (see pp. 102f. in Lawson and McCauley 1990). Third, we have no doubt that the fifth claim is true. We simply deny that it has any relevance to our theory. (See footnote 5 above.) Finally and most importantly, for all of his talk of counter-examples, Levine nowhere supplies even one! Since it is the challenge of counter-examples that force a theory to either put up or adjust its tune, for anyone who is interested--as we are--in forging better scientific theories, that deficiency in Levine's comments is especially disappointing.

References

Abbink, J. (1995). "Ritual and Environment: The Mosit Ceremony of the Ethiopian Me'en People," Journal of Religion in Africa 25, 163-190.

Bell, C. (1996). "Modernism and Post-Modernism in the Study of Religion," Religious Studies Review 22, 179-190.

Churchland, P. S. (1983). "Consciousness: The Transmutation of a Concept," Pacific Philosophical Quarterly 64, 80-95.

Lawson, E. T. and McCauley, R. N. (1990). Rethinking Religion: Connecting Cognition and Culture. Cambridge: Cambridge University Press.

Lawson, E. T. and McCauley, R. N. (1993). "Crisis of Conscience, Riddle of Identity: Making Space for a Cognitive Approach to Religious Phenomena," Journal of the American Academy of Religion 61, 201-223.

Lawson, E. T. and McCauley, R. N. (1995). "Caring for the Details: A Humane Reply to Buckley and Buckley," Journal of the American Academy of Religion 63, 353-357.

Maguire, J. F. (199X). Review of Rethinking Religion, Journal for the Scientific Study of Religion XX, 344-346.

Malley, B. (199X). XXXXXXXXXXXXXXXXXXXXXXXXXX cited in footnote 3

McCauley, R. N. and Lawson, E. T. (1984). "Functionalism Reconsidered," History of Religions 23, 372-381.

McCauley, R. N. and Lawson, E. T. (1993). "Connecting the Cognitive and the Cultural: Artificial Minds as Methodological Devices in the Study of the Sociocultural," Minds: Natural and Artificial. R. Burton (ed.). Albany: State University of New York Press, 121-145.

McCauley, R. N. and Lawson, E. T. (1996). "Who Owns 'Culture'?" Method and Theory in the Study of Religion 8, 171-190.

McCauley, R. N. and Lawson, E. T. (in progress). Ritual, Agency, and Memory.

Paden, W. (1992). Interpreting the Sacred. Boston: Beacon Press.

Searle, J. (1969). Speech Acts. New York: Cambridge University Press.

Van Gelder, T. (1995). "What Might Cognition Be If Not Computation?" Journal of Philosophy 91, 345-381.

Weiskrantz, L. (1986). Blindsight: A Case Study and Implications. Oxford: Oxford University Press.

[1] In one respect Levine's comment that "exceptions would . . . prove more informative than the generalization" (p. 17) is profoundly short-sighted. The interest and, finally, the very existence of exceptions (qua exceptions) depend upon the formulation and initial success of general theories!

[2] Levine faults us for not making more connections with existing interpretive materials. This charge is puzzling, because we do, in fact, discuss many standard interpretations of the ritual materials we treat in Rethinking Religion. Nonetheless, we should point out the comparative irrelevance of much interpretive work for our project. Our theory often requests answers to questions that field workers, exclusively in pursuit of meanings, have had no reasons to ask.

[3] Malley (199X--get the reference) has suggested that the resources of dynamical systems theory may prove helpful in understanding the historical development of particular religious traditions.

[4] Levine claims that the fact that the study of religion is interdisciplinary eludes us. But, in fact, we just disagree with Levine about the most desirable direction for its future interdisciplinarity. We have argued repeatedly (Lawson and McCauley 1990 and 1993 and McCauley and Lawson 1993) that the resources of the cognitive sciences promise to make a valuable contribution to our knowledge about religion.

[5] Levine's claims (p. 17) that alternative criteria for assessing a ritual's relative centrality will yield alternative assessments is true but irrelevant. The point is that--based on the structural representations (of the rituals) that they employ--our theory can predict what the predominant comparative assessments of participants will be.

[6] This is why, finally, Levine's inability to grasp the analogies between religious ritual systems and languages (not between "ritual systems and linguistics," p. 7) should not matter too much. (We should reiterate that we are by no means the first to argue for such analogies! See Lawson and McCauley 1990, chapter 3.) We discuss these analogies for two reasons: (1) they actually played an important role in our own process of discovery, and (2) we thought they would help the reader understand the sorts of issues a competence approach to theorizing highlights that bear on the interpretation of the resulting theory.

Levine asserts that they are unhelpful, but he neither says why nor discusses any of them individually. We are suggesting that these similarities are not random but rather ones that reveal commonalities in the cognitive representation and processing of the materials in question. We think those commonalities should help to motivate theories of religious ritual cast in the same style as theories in the study of language. Hence, we do not reject the theories we discuss in chapter 3 of Rethinking Religion, we simply emphasize that we are taking our inspiration from a different theoretical approach to natural language.

[7] We not only do not have anything against functional analyses, we have explicitly defended them. (McCauley and Lawson 1984)

[8] That, however, does not deter Levine from attacking the first of these three claims!

[9] It was the precision of the structural descriptions that enabled us to make those claims clear and to make those claims confidently.