These conventional prescriptions contain at least three underlying assumptions deserving critical attention:
   First, they recommend the native language as a tool, as a means to extract information. Somehow, what one seeks is thought to exist separately from language and the activity of speaking. To be sure, anthropologists have, before and after Whorf, maintained that the language of a people offers clues, perhaps even the key, to its culture. In one respect, however, the views of those who saw in the native language a mere vehicle of research, and others, who proclaimed it the depository of culture, converged: neither considered seriously that the "usefulness" of the native language might rest on the fact that it draws the researcher into a communicative praxis as a result of which metaphors such as tool, vehicle, or recepticle might be difficult to maintain... All these images encourage a manipulative use of language derived from visual and spatial conceptualizations whose long history will occupy us throughout this chapter.
   Second, the recommendations to use maps, charts, and tables signals convictions deeply ingrained in an empirical, scientific tradition. Ultimately they rest on a corpuscular, atomic theory of knowledge and information. Such a theory in turn encourages quantification and diagrammatic representation so that the ability to "visualize" a culture or society almost becomes synonymous for understanding it. I shall call this tendency visualism....  The term is to connote a cultural, ideological bias toward vision as the "noblest sense" and toward geometry qua graphic-spatial conceptualization as the most "exact" way of communicating knowledge....
   Visualism  may take different directions--toward the mathematical-geometric or toward the pictorial-aesthetic. In the latter case, its idolatrous tendency is often mitigated by the precept to approach culture not as a picture but as a text. Certainly there has been progress in anthropology from mere counting and mapping of cultural traits toward accounts of culture which are attentive to context, symbols, and semantics. Still, sooner or later one will come upon syntheses of knowledge whose organizing metaphors, models, and schemes are thoroughly visual and spatial. This is obvious in such terms as trait, pattern, configuration, structure, model, cognitive map; it is presupposed in notions such as system, integration, organization, function, relation, network, exchange, transaction, and many others which cannot be purified from reference to bodies, parts of bodies, ensembles, machines, and points in space; in short, to objects of knowledge whose primary mode of perception is visual, spatial, or tangible. Therefore it is not surprising that anthropologists of all persuasions have been in overwhelming agreement that their knowledge is based upon, and validated by, observation.
   Third, even the most simple and seemingly commonsense recommendations of the kind which served as a point of departure for these remarks carry notions of speed, or expeditiousness of procedure. In other words, they are aimed at instituting a time-economy for anthropological research. Not only is the total time for fieldwork conventionally fixed, it is also thought (and often said) that the field-worker "saves time" by learning the language beforehand; that he "gains time" through the use of techniques and devices. Advice may take a moral twist, when the student is told to make good use of time by never letting the sun set on untyped field notes. In all this it is the researcher's time which is thought to affect the production of knowledge. This observation is not invalidated by recommendations to take note of native ideas of Time either as explicitly formulated, or as inferred from the organization of ritual and practical activities. As an object of knowledge, the Time of the natives will be processed by the visual-spatial tools and methods invoked earlier.
   Anthropologists who have gone through the experience of field research, and others who are capable of imagining what happens to a stranger entering a society with the intention of learning something about it, are likely to be put off by this account. Why did extrapolations from simple and sensible advice regarding method result in a caricature of ethnography? Because these recommendations not only exaggerate (the visual), they omit dimensions of experience. No provision seems to be made for the beat of drums or the blaring of bar music that keep you awake at night; none for the strange taste and texture of food, or the smells and the stench. How does method deal with the hours of waiting, with maladroitness and gaffes due to confusion or bad timing? Where does it put the frustrations caused by diffidence and intransigence, where the joys of purposeless chatter and conviviality? Often all this is written off as the "human side" of our scientific activity. Method is expected to yield objective knowledge by filtering out experiential "noise" thought to impinge on the quality of information. But what makes a (reported) sight more objective than a (reported) sound, smell, or taste? Our bias for one and against the other is a matter of cultural choice rather than universal validity. It derives from a scientific tradition which was firmly established by the time J. Locke formulated the empiricist canons of modern social science. "The perception of the mind," he maintained, is "most aptly explained by words relating to the sight". Among all the tenets of empiricism this one seems to have been the most tenacious.
   Even if detached observation is regarded positively as a means to lift oneself above the immediacy of fleeting sounds, ineffable odors, confused emotions, and the flow of time passing, the anthropologist so inclined should give, at the very least, some thought to the cultural determinedness of his quest for distance.