Ann Taves[*] []


There is no way to study "religious experience" as such scientifically because what constitutes "religion" is highly contested. We are as a result restricted to studying stipulated aspects of experience, some of which may be deemed by some to be religious. These stipulated aspects of experience will naturally cut across the boundaries of disciplines and cultural categories, such as "religion". Stipulating the precise aspect of experience to be studied is the central act in the construction of an object of study. Stipulating a precise point of comparison is complicated by several factors: (1) that experience can be studied from the outside and described from the inside, (2) that comparisons always involve the use of categories, and (3) that the categories readily at hand often undercut our ability to work across disciplinary lines.


Não há como estudar a "experiência religiosa" de maneira científica porque os elementos que podem constituir "religião" são objeto de grande disputa. Somos, assim, limitados a estudar aspectos da experiência previamente estipulados, que podem ser considerados religiosos sob de certos pontos de vista. Naturalmente, o estudo destes aspectos iria além dos limites de disciplinas e categorias culturais como "religião". A determinação do aspecto exato da experiência a ser estudada é o ato central na construção de um objeto do estudo. Esta conduta discricionária, porém, é complicada por vários fatores: (1) a experiência pode ser estudada do exterior e descrita do interior; (2) a tarefa de comparar sempre implica o uso de categorias; e (3) as categorias disponíveis muitas vezes abalam a nossa capacidade de trabalhar além de linhas disciplinares.

The central methodological issue facing scholars interested in the empirical study of religious experience or, more broadly, the experiential dimension of religion is one of data selection. How do we know that an experience is religious? How do we distinguish religious experiences or the experience of religion from experience more generally? Douglas Allen clearly articulated the problem when he observed apropos of Mircea Eliade that if the historian of religion's "point of departure is the historical data which expresses the religious experiences of mankind[,] … how does one know what documents to collect, which phenomena to describe and interpret?" (1972, 171-72). Wayne Proudfoot identified two options: an experience can be designated as religious by the scholar (who must then supply a definition of religion) or by the subject of the experience (1985). In practice, however, matters are not so simple. Many texts that the scholar might intuitively want to consider were written by followers or observers rather than by the ostensible subject of the experience. More crucially many of those same texts do not explicitly refer to either "religion" or "experience". Moreover, when scholars turn to definitions of religion for assistance, they often find that they employ referents (e.g., ultimate reality, the sacred, the numinous) that are so vague that they offer little assistance in selecting texts. The problem of data selection, thus, leads directly to the underlying question of how (and to what extent) we as scholars constitute our objects of study.

If we want to move beyond simply paraphrasing the words of those we are studying, we must take responsibility for our role in constructing an object of study. Three options are particularly pertinent with regard to what has traditionally been construed as religious experience, each with its own advantages and disadvantages, depending on the scholar's aims (Taves, 2005).

The scholar can limit him or herself to third person texts that make explicit use of the concept "religious experience" and its near neighbors and competitors, always taking care to note which terms are actually being employed. This approach is particularly useful for tracing the use of concepts and the history of ideas. As a discursive strategy, it does not attempt to get at experiences per se.

The scholar can focus on first person accounts of experience that the subject specifically describes as religious. Scholars may find this approach unduly limiting, in which case they can decide in advance what they will count (i.e. define) as "religious experience" by developing a definition that is sufficiently specific that it can actually be used to select appropriate textual sources for consideration. While this approach allows the scholar to get at experiences, it runs the risk of telling us more about the scholar's understanding of religion (substantively or heuristically) than that of his or her subjects.[1] Selecting only texts that the scholar designates as religious limits comparisons, as has typically been the practice in religious studies, to phenomena designated as religious at the outset. It precludes non-religious explanations of the phenomena under consideration by definition and in that sense is tacitly protective of "religion".

The scholar can identify an aspect of experience that is not necessarily co-extensive with what scholars or their subjects take to be religion and analyze the way in which it has been or is understood by subjects and observers, including scholars. This approach allows scholars to compare experiences deemed religious and non-religious by practitioners and/or observers and to recognize and examine boundary issues (e.g. between experiences deemed psychical, psychopathological, religious, mystical and so on) typically suppressed by scholarly definitions of religion. This is the only one of the three approaches that is compatible with the scientific study of consciousness.

This means, to make several obvious points:

  1. There is no way to study "religious experience" as such scientifically simply because what constitutes "religion" is highly contested. We are restricted to studying stipulated aspects of experience, some of which may be deemed by some to be religious.[2]
  2. These stipulated aspects of experience will naturally cut across the boundaries of disciplines and cultural categories, such as "religion".
  3. It is for this reason that, when we need to be precise, it is more accurate to refer to our object of study as "experiences some of which are deemed by some to be religious" than as "religious experience".
  4. Stipulating the precise aspect of experience to be studied - that is, establishing an exact, stipulated point of comparison (Paden 2005: 1880) - is the central act in the construction of an object of study in relation to a varied data set.

Stipulating a precise point of comparison is complicated by several factors: (1) that experience can be studied from the outside and described from the inside, (2) that comparisons always involve the use of categories, and (3) that the categories readily at hand often undercut our ability to work across disciplinary lines. I will address each difficulty in turn.

Stipulating a precise point of comparison is complicated by the fact that experience can be studied from the outside, i.e., through the observation of behaviors and the measurement of brain states, and can be described from the inside, i.e. phenomenologically, through first-person accounts. The study of "the qualitative, subjective dimension of experience" falls under the heading of consciousness studies, a rapidly burgeoning subfield within the cognitive neurosciences (Zeman, 2001, 1265). As Frederic Peters has noted, "advances [in neuroscience] provide an opportunity to marry objective explanation with phenomenological descriptions of the view from the inside" (2000, 379). Although debate continues over the best way to incorporate internally-generated (phenomenological) with externally-generated (behavioral and neurophysiological) data in the context of experimental design (Chalmers, 2005, 1111),[3] the issues faced by neuroscientists are parallel in some respects to those faced by historians when attempting to reconstruct experiences based on first person narratives and third person accounts. In both cases, whether dealing with experimental or historical data, it is clear that if we want to understand the subjective dimension of experience, we must begin with the internally-generated accounts, that is, the first-person narratives, of those who had the experiences. Doing so, however, does not mean that scholars must assume the metaphysical commitments or explanatory perspectives of their subjects. It only means that the starting point for any discussion of subjective experience must be the experience of the subjects, either as described by the subject or as redescribed by the scholar in terms that would be recognizable to subject (Proudfoot, 1985; Dennett 1991, 2003).[4]

Constructing an object of study from first-person data, i.e. at the phenomenological level, becomes more complicated if we want to work comparatively across types of experiences. Jeffrey Carter defines "comparison [as] a method of clumping and splitting based upon stipulated similarities and differences". Comparison, Carter argues, "is essentially related to categorization". Thus, he writes: "distinguishing between entities, detailing to some degree how they are similar and different, and thereby collecting and dividing them in response to the desire for, and the consistency of, a (invented) cognitive/theoretical order, is effectively the work of categorization" (2004, 5). If, as Carter argues, "comparison is the method by which categories are formed and manipulated", then the crucial question is not whether we are going to utilize categories in order to compare but what categories we are going to use. While the phenomenological description of the subject's experience should be constituted in terms that could be ascribed to the subject, the way that researchers compare these experiences and, thus, the categories that are utilized in the act of comparing them are the responsibility of the researcher (Satlow, 2005).

One of the chief difficulties surrounding the use of academic or intellectual categories in comparative work lies in the fact that they have emerged and been utilized within disciplines that are not in the habit of communicating with one another. Conceptual categories across disciplines are not necessarily compatible with one another, making research across disciplines difficult. While compatibility or "vertical integration" is taken for granted in the natural sciences, this is not the case in the social sciences or the humanities (Barkow, et. al., 1992, 4).

Figuring out how to deal with competing conceptual categories was one of the chief methodological hurdles I faced in writing Fits, Trances, and Visions. Not only did Protestants variously characterize the experiences I wanted to discuss positively as "religious experiences" and negatively as "enthusiasm", but academic disciplines designated similar sorts of experiences as "dissociation" (psychiatry); "trance", "spirit possession", and "altered states of consciousness" (anthropology); and "visions", "inspiration", "mysticism", and "ecstasy" (religious studies). These designations are not simply descriptive, but reflect the various historical and explanatory commitments of the disciplines themselves. Adopting any one of these categories as my stipulated point of comparison would have tacitly positioned me in relation to disciplinary subject matters (e.g., religion, culture, or psychopathology) and explanatory commitments (Taves, 1999, 7-9). As Luther Martin has observed with respect to the comparative study of religion, the overarching problem is that we have begun our "inquiries from the 'top down', i.e., from the perspective of categories and ideologies prelabeled by cultural interests" (Martin, 2004, 38-39).

The key to enhancing compatibility and ultimately integration across disciplines, I want to suggest, is to avoid ready-made categories and to focus instead on identifying the stipulated point of analogy between the things we want to compare with more extended statements. Practically speaking, this means that we need to move away from single terms to more extended descriptive statements that identify common features in a way that is simultaneously intelligible across disciplines and workable in terms of designating comparable subject matters at the phenomenal level of lived experience. This stipulated point of analogy then defines the data set - that is, the set of things that are similar in that stipulated regard - clumping them together and splitting them off from those things that do not share that stipulated feature (Taves, 1999, 8; Paden, 2005, 1880).

The key in terms of setting up Fits, Trances, and Visions was to find a language that allowed me to compare phenomena across disciplinary lines -- religious, anthropological, and psychotherapeutic -- without unduly violating the experience of persons within those traditions. In the end, I characterized the experiences I wished to discuss as "involuntary experiences", that is, experiences in which the subjects did not feel that they were the agent or cause of their own experience. I specified that such experiences might include the loss of voluntary motor control, unusual sensory perceptions, and/or discontinuities of consciousness, memory or identity. Experiences that fit this description can be found in numerous places, including the New Testament (see, for example, Paul's descriptions of life in Christ [Gal. 2:20, 2 Cor. 12:2] and the account of his conversion in Acts 22:6-9), and the American Psychiatric Association's Diagnostic and Statistical Manual (DSM-IV).[5] My intent in subsuming such disparate experiences under one descriptive heading was not to equate them, but to specify the feature that these experiences held in common in order to say what the book was about, define the parameters of a comparative field, and identify relevant instances in the context of historical research.

Although none of my historical subjects explicitly described their experiences as "involuntary", this term highlighted an aspect that all the experiences in question shared in terms that were compatible with their various self-representations. In specifying the three types of involuntary experience with which I would be concerned, I did not simply accept the categories offered by the DSM-IV. Instead, I tested each against my historical subjects' self descriptions looking for hidden explanatory assumptions. In deference to Jonathan Edwards' desire to root authentic religious experience in a new "supernatural sense" entirely distinct from the natural senses, I referred to unusual sensory experiences (that presumably could be understood as supernatural) rather than to altered (and thus presumably still natural) sensory perceptions or experiences. In a parallel fashion, some neuroscientists have been working with living subjects to develop generic categories that make sense to their subjects. Shaun Gallagher, for example, discusses the way that some cognitive scientists are training their research subjects to shift their attention from what they are experiencing to how they are experiencing it in order to surface subjective accounts of processes that underlie differences in content (2003, 88).

This approach to the study of experiences some of which are deemed by some to be religious allows us to make two important methodological moves. First, articulating a stipulated point of analogy relatively unencumbered by disciplinary presuppositions allows us to construct a narrative space in which we can analyze the contentious dialogue between the various interpretive positions. As one reviewer noted, "this theoretical emphasis upon intersection and interstitial space provides an excellent model for cultural historians tracking multiple systems in conversation with one another, particularly those who take seriously the ideological systems in place" (Westercamp, 2002, 96-97). Second, constructing our object of study in generic terms compatible with the phenomenological self-descriptions of our subjects allows us to move in multiple directions in terms of explanation. It allows us to move "downward" in terms of connecting first-person phenomenological data with the third-person data of the cognitive neurosciences, "upward" in terms of connecting first-person phenomenological data with socio-cultural dynamics, and "backward" in terms of connecting it with the past on multiple levels (Kallio and Revonsuo, 2003, 137-39).[6]


ALLEN, D, Mircea Eliade's Phenomenological Analysis of Religious Experience. In: Journal of Religion 52 (1972): 170-186.

AMERICAN PSYCHIATRIC ASSOCIATION, Diagnostic and Statistical Manual of Mental Disorders, 4th edition, Washington, D.C., 1994.

BARKOW, J.; COMIDES, L.; and TOOBY, J. The Adapted Mind: Evolutionary Psychology and the Generation of Culture, New York: Oxford University Press, 1992.

CARTER, J., Comparison in the History of Religions: Reflections and Critiques. In: Method and Theory in the Study of Religion 16, 1 (2004): 3-11.

CHALMERS, D., How can we construct a science of consciousness? In: Michael Gazzaniga, Ed. The Cognitive Neurosciences, 3rd ed., Cambridge: MIT Press, 2004.

DENNETT, D. Consciousness Explained, Boston: Little, Brown, & Co., 1991.

DENNETT, D., Who's On First? Heterophenomenology Explained. In: Journal of Consciousness Studies 10, 9-10 (2003): 19-30.

GALLAGHER, S., Phenomenology and Experimental Design: Toward a Phenomenologically Enlightened Experimental Science. In: Journal of Consciousness Studies 10, 9-10 (2003): 85-99.

KALLEO, S. and REVONSUO, A., Hypnotic Phenomena and Altered States of Consciousness: A Multilevel Framework of Description and Explanation. In: Contemporary Hypnosis 20, 3 (2003): 111-64.

MARTIN, L., "Disenchanting" the Comparative Study of Religion. In: Method and Theory in the Study of Religion 16, 1 (2004): 36-44.

PADEN, W., Comparative Religion. In: Encyclopedia of Religion, 2nd edition. New York: Macmillan, 2005, vol. 3, 1877-1881.

PETERS, F.H., Neurophenomenology. In: Method and Theory in the Study of Religion 12, 3 (2000): 379-415.

PETERS, F.H., Neurophenomenology of the Supernatural Sense in Religion. In: Method and Theory in the Study of Religion 16, 2 (2005): 122-48.

PROUDFOOT, W. Religious Experience, Berkeley: University of California Press, 1985.

PYYSIAINEN, I, How Religion Works: Towards a New Cognitive Science of Religion, Leiden: Brill, 2003.

RUDRAUF, D., et. al., From autopoiesis to neurophenomenology: Francisco Varela's exploration of the biophysics of being. In: Biological Research 36 (2003): 27-65.

SATLOW, M.L., Disappearing Categories: Using Categories in the Study of Religion, Method and Theory in the Study of Religion 17, 4 (2005): 287-98.

TAVES, A. Fits, Trances, and Visions: Experiencing Religion and Explaining Experience from Wesley to James, Princeton: Princeton University Press, 1999.

TAVES, A., Religious Experience. In: Encyclopedia of Religion, 2nd edition, Macmillan, 2005, vol. 11, 7736-7750.

WESTERCAMP, M.J., Review of Fits, Trances, and Visions. In: Journal of Interdisciplinary History 33, 1 (2002): 96-97.

ZEMAN, A, Consciousness: Invited Review. In: Brain 124 (2001): 1263-1289.


[*] ANN TAVES is Cordano Professor of Catholic Studies at the University of California at Santa Barbara and a member of the Executive Committee of the International Association for the Cognitive Science of Religion (IACSR). She is the author of Fits, Trances, and Visions: Experiencing Religion and Explaining Experience from Wesley to James (Princeton, 1999).

[1] This holds true even for the newer understandings of religion proposed by cognitive scientists of religion. Thus, to quote I. PYYSIÄINEN (How Religion Works, p. 227): "Even if counter-intuitive representations serve as an especially good defining characteristic of religion, the presence of counter-intuitiveness along is not a sufficient criterion for something being an instance of religion. Nor is it possible to say what more is needed, as it seems that there are no jointly sufficient and singly necessary criteria according to which [we might] judge something as an instance of religion".

[2] For another way of stating this point, see. W. PROUDFOOT, Religious Experience, p. 216.

[3] Although both Peters and Chambers use the term "neurophenomenology" to refer to the question of how first person data relates to third person data in the construction of experimental design, the term is associated somewhat more narrowly with the work of Francisco Varela and his students among the cognitive scientists (see RUDRAUF, D., et. al., "From autopoiesis to neurophenomenology, In: Biological Research, pp. 27-65.). D. Dennett, who advocates an approach he calls "heterophenomenology", is sometimes depicted as presenting a competing perspective, though he has recently commented that the differences are exaggerated (see D. DENNETT, Consciousness Explained and D. DENNETT, Who's On First? In: Journal of Consciousness Studies, pp. 19-30).

[4] The parallels are evident in the methods advocated by D. DENNETT (Ibid.) and W. PROUDFOOT, Religious Experience. Heterophenomenology, according to Dennett, "is a neutral path leading from objective physical science and its insistence on the third-person point of view, to a method of phenomenological description that can (in principle) do justice to the most private and ineffable subjective experiences, while never abandoning the methodological scruples of science" (D. DENNETT, Consciousness Explained, p. 72). Scientific research, in his view, demands a "metaphysical minimalism" relative to first person accounts of experience. In order to avoid "begging any questions about whether his subjects are zombies, computers, lying, or confused", he interprets the verbal accounts of consciousness produced by his subjects "as fictions of a sort, not as literature, but as generators of a theorist's fiction (which might, of course, prove to be true after all)".
Within these strictures, the subject has the right to constitute their own phenomenological world, thus avoiding what Proudfoot refers to as "descriptive reductionism": "[T]he heterophenomenologist lets the subject's text constitute that subject's heterophenomenological world, a world determined by fiat by the text (as interpreted) and indeterminate beyond. As in fiction, what the author … says goes" (D. DENNETT, Consciousness Explained, p. 81). In a parallel fashion, Wayne Proudfoot stresses that any explanation of an experience must start from "the identification of an experience under a description that can be ascribed to the subject" (W. PROUDFOOT, Religious Experience, p. 218). Dennett and Proudfoot would agree that (in Proudfoot's words): "The subjects identification of his experience in religious terms makes it a religious experience and is normative for describing that experience. But the subject's explanation may not be the correct one; it may be that the correct explanation requires no reference to religious realities" (W. PROUDFOOT, Religious Experience, p. 201). Or in Dennett's words:

[S]ubjects are unwitting creators of fiction, but to say that they are unwitting is to grant that what they say is, or can be, an account of exactly how it seems to them. They tell us what it is like … we grant that that must be what it is like to them, but then it follows that what it is like to them is at best an uncertain guide to what is going on in them. Sometimes, the unwitting fictions we subjects create can be show to be true after all, if we allow for some metaphorical slack (D. DENNETT, Consciousness Explained, p. 94).

[5] See, for example, the descriptions of Conversion and Dissociative Disorders.

[6] S. KALLEO, and A. REVONSUO (Hypnotic Phenomena and Altered States of Consciousness, In: Contemporary Hypnosis, pp. 111-64) identify five levels of description (social-psychological, personal, and sub-personal [phenomenal, cognitive, and neural]) and three types of explanation (constitutive [downward-looking], etiological [backward looking], and contextual [upward-looking]), all of which are relevant to the task at hand. Kallio and Revonsuo, like Proudfoot and Dennett, think "that the particular changes in subjective conscious experience experienced by the subject (the phenomenal level) form the core explanandum of any theory of hypnosis". A full explanation, they argue, would encompass all three types of explanation. The constitutive, which looks downward to the cognitive and neural levels, reflects the explanatory interests of the cognitive neuroscientists. The contextual, which looks upward to the personal and socio-cultural levels, reflects the explanatory interests of humanists and social scientists. The etiological explanation, which is backward looking and can be examined at any of the five levels, reflects the explanatory interests of historians, broadly conceived.