A century ago, in “Religion and Neurology,” the opening chapter of The Varieties of Religious Experience, William James argued against a “medical materialism” that would reduce religious experiences to their neurological causes for the purpose either of dismissing them or confirming them. Since that time, many have tried to understand religion through the study of religious experience and, like James, many have given special attention to mysticism. New techniques for the study of the brain have brought great advances, but David Brooks’s New York Times column “The Neural Buddhists” and the work of Andrew Newberg, to whom he refers, stand squarely in the tradition James was criticizing.

Religious experience can’t be described or explained in biological terms alone. An experience is constituted, in part, by the way it is interpreted by the person experiencing it, and it is the interpretation that makes an experience religious. Two people might have experiences that are indistinguishable biologically (slower heart rate, decrease in body temperature) or phenomenologically (sense of oneness, calm), but one might experience what is happening to him in religious terms while the other does not. Study of any experience requires attention to historical and cultural contexts that inform a person’s interpretation of what is happening to her and to the conditions under which she comes to identify her experience in particular terms. Though he criticized medical materialism, James didn’t sufficiently appreciate the need to elucidate the concepts and practices that enter into a person’s experience and to study them as historical products. He was too intent on trying to identify a common core in religious experience that would be universal across cultures.

As psychologists examine the extent to which people can control attention, emotion, and other mental and even autonomic processes, it is not surprising that they would be drawn to practitioners of spiritual exercises in various religious traditions that have developed techniques for achieving this kind of control. Buddhism in particular has a long history of sophisticated reflection on and experimentation with the control of mental and physical states, though here Donald Lopez’s comments about Buddhist Modernism are apt.

In The Mystical Mind: Probing the Biology of Religious Experience, Newberg and his colleague Eugene d’Aquili, psychiatrists at the University of Pennsylvania Medical School, propose a “study of theology from a neuropsychological perspective” that they claim would explain all elements of religion. They describe their book as the culmination of almost 25 years of research on the relationship between the brain and religious experience. This suggests that their theory carries the prestige and credibility of science. In fact, the model they offer is based on speculation, not laboratory science. There is nothing wrong with this. They are as entitled to speculate as are other theorists of religion, but their conclusions ought not to be given special scientific status.

D’Aquili and Newberg maintain that there are core elements in religious experience that “appear to be universal and can be separated from particular cultural matrices.” They suggest that sensations of awe and a unitary experience may be caused by deafferentation, the cutting off of incoming information to a brain structure. From speculation about localization of functions in different areas of the brain, they develop a structural model that includes a holistic operator that “might allow us to apprehend the unity of God or the oneness of the universe.” This structuralism, combined with deafferentation, constitutes their theory of religion.

In the one experimental result cited in the book directly relevant to the theory and in another paper published in 2003, Newberg and his collaborators report experiments using brain imaging (functional neuroimaging) on experienced meditators in the Tibetan Buddhist tradition. They found that meditation was associated with mildly increased blood flow in the frontal areas of the brain, consistent with studies that show increased activity in the prefrontal cortex associated with concentration. They had hypothesized that there would be decreased flow in parietal areas of the brain associated with altered experience of space, but results here were mixed. In one kind of analysis of the data they found no significant decrease, though they did find it using another kind of analysis.

These results are consistent with Newberg’s theory, but they don’t provide experimental support for it. In addition, by ignoring the concepts by which their subjects understood what was happening to them they fail to ask what makes these experiences religious. Newberg’s comment about apprehending the unity of God and Brooks’s statement in his column that people are equipped to experience the sacred and overflow with love include interpretations that go way beyond anything measured in these experiments. “God,” “the sacred,” and “overflowing love” each have a culturally specific provenance that can only be elucidated historically.

Richard J. Davidson, a neuroscientist at the University of Wisconsin, and his colleagues have undertaken a project to study mental abilities achieved by long-term meditative practitioners that is more sophisticated than Newberg’s. Here, again, the subjects are Tibetan monks, but the researchers are trying to study the ways in which these monks pursue and attain compassion through meditation. That requires them to attend to the content of Buddhist doctrine and practice.

In Visions of Compassion, the volume he and Ann Harrington edited describing the project, Davidson surveys psychological research on emotion and notes that much has been done on negative emotions, but considerably less on positive affect. Psychologists have identified distinctive facial expressions recognizable across cultures for negative emotions, but not for their positive counterparts. The only basic classification of positive emotions that correlates with biological data, he says, is that between pre-goal attainment and post-goal attainment affect, between eager anticipation and satisfaction. For his research, Davidson writes, he will focus on this division rather than on distinctions within the emotion lexicon that don’t have biological correlates.

That may be a reasonable research decision, but it substitutes a simple bipolar classification for a much larger set of emotion terms, each of which has its own grammar and criteria of application. For example, within the category of post-goal attainment positive affect, we can easily distinguish between the meanings of contentment, pride, enthrallment, and relief. The choice of which is appropriate in a given situation requires knowledge of context, background assumptions, and reference to relevant norms.

Davidson and his colleagues compared Buddhist monks who had been training in Tibetan traditions for at least 15 and in some cases 40 years with control subjects who had undergone just one week of meditative training. During the study both long-term practitioners and control subjects were asked first “to let their minds be invaded by a feeling of love or compassion” toward someone they cared about, and later “to generate unconditional loving-kindness and compassion toward all sentient beings without thinking about anyone in particular.” The researchers recorded electroencephalographic (EEG) activity and found that the monks had induced in themselves high-amplitude EEG gamma oscillations and phase synchrony during the nonreferential compassion meditative state. This kind of synchronization, Davidson writes, may reflect attention and affective processes, and is consistent with the idea that these are flexible skills that can be trained.

This study may be valuable for training subjects to control their affective states, but it doesn’t tell us anything about the Buddhist notion of compassion or its pursuit. Traditional Buddhist texts on meditation distinguish between calming (achieving a stillness of mind and body) and discernment (insight). The first is considered preparation for the second. Davidson’s EEG study may measure calming, but it leaves discernment unexplored, in this case discernment of Buddhist teaching about compassion and its application to oneself or others.

Instructions to “generate a state of loving-kindness and compassion toward all sentient beings” must have been understood very differently by the Buddhist monks and by members of the control group. The monks had undergone years of training in debating skills that are a large part of Tibetan scholastic pedagogy. They had cultivated practices that include not only techniques for the control of mental and physical states, but also criteria for the proper use of terms like “compassion” and “sentient beings.” Their understanding of these terms would be quite different from that of the controls, who had trained for a week in an admittedly superficial way.

Georges Dreyfus, a scholar of Tibetan Buddhism, writes in Visions of Compassion that there is no Tibetan word for emotion and he asks whether compassion is an emotion. He cites a distinction drawn both in texts and practice between beginners who feel compassion (they are saddened by the sufferings of others and moved to wish them relief) and advanced bodhisattvas who attain compassion with equanimity without being themselves moved.

Is compassion then a trait? The difference between a state and a character trait is not only one of endurance over time. We call someone “loving” or “kind” not on the basis of what we take her to feel, either on the basis of her self-report or some other evidence, but because we take the term to capture something of her character. We ascribe kindness to her by observing how she acts in response to different situations over time. Similar observations and attempts at discernment are at work in my reflections on my own character, though usually clouded by bias in my own favor. To understand Buddhist compassion meditation one would have to explicate the criteria that govern the way the monks themselves use that term in their own self reflection. Were Buddhist practitioners to agree to conditions under which they would be given simultaneous feedback of their gamma-band oscillations and phase-synchrony during meditation, it might be possible for them to manipulate those variables more directly, but that would have no bearing on their judgment or anyone else’s about the extent to which they were on the path to compassion.

The study of religious experience requires attention to culturally specific language and social practices by which people come to understand themselves and their world in religious terms, as well as to the results of neuroscience and other inquiries into universal characteristics of mind and body.