Description’s an element, like air or water.
—Charles Wright, “Black Zodiac”

An adequate theory of consciousness requires being able to describe it. This isn’t a grand metaphysical point, just an obvious fact about writing. Theories exist in discourse, and to bring mental phenomena like conscious experience into contact with theory, they too have to be described. How they are described makes all the difference to how they are theorized. So it is worth reflecting on the descriptions of experience that philosophers habitually turn to.

Many of these descriptions are embedded in the brief vignettes that are a staple of contemporary philosophical writing. Some of these belong to the unfortunately named genre of “thought experiments,” but they can include illustrations of how conceptual distinctions operate, accounts of facts to be explained, veiled attempts at persuasion, and more. Here I am interested in the role of description in these vignettes. I suggest that the conventions of philosophical description narrow the range of admissible facts and thereby make many phenomena invisible to our theorizing.

These vignettes or sketches of mental life have a few common qualities. First, they are brief, typically not longer than a few sentences, sometimes just one. Second, they are plainly narrated in a dry and literal style that remains light on details. Third, they tend to draw from a familiar range of everyday scenarios, which are themselves often short episodes. In sum, they are characteristically highly schematic and stereotyped in form and content.

Consider a canonical vignette, Armstrong’s (1968, pp. 92-3) “distracted driver”:

Case 1. This is something that can happen when one is driving very long distances in monotonous conditions. One can ‘come to’ at some point and realize that one has driven many miles without consciousness of the driving, or, perhaps, anything else. One has kept the car on the road, changed gears, even, or used the brake, but all in a state of ‘automatism’… In Case 1 one must in some sense have been perceiving, and acting purposively. Otherwise the car would have ended in a ditch. But one was not conscious of one’s perceptions and one’s purposes. It may be surmised that animals spend much of their life in that state of automatism enjoyed by the long-distance driver.

This little sketch has borne enormous theoretical weight in the subsequent literature, despite the fact that it elides any number of details that are relevant to assessing its phenomenological aptness. Is the distinction between consciousness and “automatism” really so sharp? Or do we more often exist in an intermediate state that drifts back and forth according to its own rhythms? Much could be said about the experiential microworlds that blossom in these seemingly deserted periods of boredom and disengagement. By contrast, Armstrong’s case gives the sense not of a genuine attempt to say what it feels like when one is driving slightly inattentively, but rather an illustration of the idea of sensation and action taking place without consciousness. Once this idea is on the table, the example expires. It is description manqué.

This tradition of the exiguous vignette being used as dressing for a pre-established conceptual distinction continues; see, for instance, the opening pages of Michael Tye’s Consciousness and Persons, where no fewer than 12 such cases are presented seriatim. The reluctance to produce or engage with more highly elaborated descriptions of experience allows a quick pivot from the presentation of a vignette to its restatement in technical language. Such vignettes serve as pegs to hang theoretical terms and distinctions on, not genuine attempts to descriptively capture or evoke states of mind.

The recycling of these stereotyped forms results in a sort of phenomenological tunnel vision. It encourages the notion that all experiences must be analyzable in our familiar contemporary terms. Over-reliance on these sorts of descriptions, then, is fundamentally conservative, since it privileges those experiences that can easily be dissected using the pre-established categories of the field.

Moreover, by focusing mainly on atomized fragments of experience, and thinking of them as reports of first-person data, we reinforce theories that assume this is the natural structure that experience has. As Paul Livingston (2004) points out, the quasi-scientific terminology of calling these descriptions “data” and “reports” also links them rhetorically with earlier constructs such as the positivists’ protocol sentences. What is lost in all of this is not just the microstructure of sensation, but its continuous dynamics as well as its large-scale and temporally extended features, which do not come so conveniently named and readily packaged.

The issue here is not just the old debate about the reliability of introspection. Even if introspection were an inner measuring device that produced data in the form of canonically encoded readouts or recordings, description is not data. Description is reflective and unmechanical. It is a product of attention and judicious choice, both of what to talk about and what to say about it. Its worth can’t be weighed like that of a gauge, solely in terms of correct indication.

In defense of the conventional vignette, one might argue that experiences and their salient qualities are so obvious that all we need are such thin descriptions, which are really no more than pointers or reminders. Even if this were so, there are many less obvious qualities of experience that can only be recovered through more engaged descriptive efforts. For evidence, see Hurlburt and Schwitzgebel’s (2007) detailed interrogation of introspective diaries elicited using descriptive experience sampling, as well as the phenomenological interview method of Høffding and Martiny (2015) and Høffding’s extensive work with the Danish String Quartet. The texts that these methods generate are more unruly than our little artificial case-studies; they resist easy incorporation precisely because they are not produced just to illustrate a pre-established theoretical point.

Above all, I suspect that the habit of vignettes encourages an unearned sense of ease or complacency about our grasp of the actual shape of consciousness. The clamor for explanations blurs the fact that we don’t yet properly understand what it is we hope to explain. To get a sense of how arduous arriving at correct descriptions can be, consider Lorraine Daston’s (2016) study of the long attempt to stabilize a language and a taxonomic scheme for clouds. Here too the phenomena seemed hopelessly fluid and lawless, with few obvious cues as to what types might be concealed within the seething mass of details. To describe these everyday atmospheric phenomena in international atlases of clouds took over a century’s labor. Nor are there purely technical solutions to problems of description: photographs actually made the task harder because they capture detail irrespective of whether it is relevant or not, and are therefore more of a distraction than an aid. How much more complex is the world within us than the one above us? The vast gulf between consciousness and clouds is a measure of the descriptive work that remains undone.