Having devoted so much space to Descartes's Two Tests for animal mind, we wish to close by proposing an empirical test for animal consciousness that falls completely outside this Cartesian framework. Hardly any philosopher pays attention to olfaction, yet it is the source of remarkable phenomena that enable one to discriminate conscious from unconscious perception. We appeal in this section to olfactory phenomena to formulate a test of consciousness for a restricted class of animals. The test we propose is neither a language test nor an action test. We believe our proposal to be the first serious proffer of a test for animal consciousness that breaks out of the straightjacket of Descartes's Two Tests. Our olfactory test is not air-tight, of course. A skeptic might plead, however implausibly, that the differential physiological response central to the test is mediated by something other than consciousness. Still, we believe that our olfactory test creates a strong presumption of consciousness in any species that passes it. Unlike Descartes's Two Tests, our olfactory test is asymmetrical in the sense that failure to pass it in no way tells against the consciousness hypothesis.
 

We turn to the psychologists Piet Vroon and Hans de Vries, and to the biologist Anton van Amerongen, for an account of the relevant phenomena:

"Many people turn out to be anosmic for androstenes; for example, almost half the population cannot perceive the smell of androstenone. There are indications that genetic factors may be connected with this inability. Anosmia for androstenone, however, does not mean that there is never a reaction to the odor. It has been shown that the electrical resistance of the skin (the galvanic skin response, or GSR, an indicator, for example, for alertness) can drop sharply after an encounter with androstenone, particularly when the odor is not consciously perceived. This fall is an indication of a general rise in activity in the emotional field, too. If the smell is experienced consciously and is found to be pleasant, such a fall also occurs; if the odor is found to be unpleasant, however, then the GSR does not change perceptibly. These are remarkable phenomena, as they show a difference between the effects of conscious and unconscious perceptions of smell. With unconscious perception the general bodily level of activity increases, but if the perception penetrates consciousness, the reaction then becomes highly dependent on the appreciation of the smell. It follows from this that such a smell can influence the physical and mental condition in various ways, depending on the concentration." Piet Vroon with Anton van Amerongen and Hans de Vries, Smell: The Secret Seducer, Farrar, Straus & Giroux, New York, 1994, pp. 134-135 (our italics).
 

Our olfactory test for conscious perception turns on three of the phenomena reported by Vroon, van Amerongen, & de Vries, namely: (1) that unconscious perception of androstenone gives rise in humans to a sharp GSR drop; (2) that conscious perception of androstenone as pleasant also gives rise in humans to a sharp GSR drop, but (3) that conscious perception of androstenone as unpleasant effects no perceptible GSR change in humans. Our test is applicable only to an animal species that exhibits these three phenomena. Whether there are such species is an empirical matter, of course, but it does not seem farfetched to suppose that these or similar phenomena characterize our closest animal relatives.

Suppose, then, that we've found a group of bonobos who are sensitive to the smell of androstenone Suppose further that some of them exhibit a sharp GSR drop in the presence of androstenone when their attention is focused elsewhere, and that they also experience a sharp GSR drop when their attention is directed toward odors, which they appear to find pleasant, that emanate from an androstenone source. Suppose finally that the other bonobos in the group also exhibit a sharp GSR drop in the presence of androstenone when their attention is focused elsewhere, but that these other bonobos do not exhibit a sharp GSR drop when their attention is directed toward odors, which they appear to find unpleasant, that emanate from the androstenone source. We would then have good grounds for saying -- with considerable plausibility -- that the unpleasant olfactory perceptions of androstenone had by the second set of apes were in fact conscious, i.e., that the apes were aware of their olfactory perceptions.
 

Our olfactory test for consciousness is really a schema for a family of similar tests on many species of animals. Variants of our test might utilize different odor sources, different affects, and different physiological responses to the odors when detected consciously and when detected unconsciously. Some variants of our test might even incorporate other sensory modalities. For example, we can envision a taste test for canine consciousness along the following lines. Suppose that we humans exhibit some distinctive physiological response to the taste of a certain substance when we consciously perceive the substance as sweet as well as when our gustatory perception of the substance is unconscious. Suppose further that this distinctive physiological response does not occur when we humans consciously perceive the substance as bitter. We can readily imagine that dogs might resemble us in their physiological and gustatory responses to this substance, thereby affording us a means to test whether some canine perceptions are conscious.