Search Engine Interfaces

Donny Smith

Analysis

Since I work in a library, I get to observe people looking. They come to the library looking for something, sometimes clumsily, sometimes nimbly. They wander the stacks, stare bleary-eyed at shelves or computer screens, type misspelled (or poorly chosen) search terms into the computers. They're usually searching for some thing, some piece of information that will complete an assignment or a task, but almost always we give them a list--a conglomeration they must convert by various mental processes into something they can use.

Now, librarians love lists (we love searching, sifting, classifying; we read bibliographies and footnotes; what students consider a chore is a joy to us), but items in a list often have no obvious connection other than being on the same list. Keyword-in-context (KWIC) lists (Fig. 31), especially, fragment information and disorient the searcher--reducing intellectual processes to piecework and words and sentences to bearers of fact only. Argument becomes accumulation. The rhythm and emotion of prose are replaced with machine patterns. A results list is the opposite of an annotated bibliography or a reference interview, and the writing that comes from a results list is itself often more accretion than argumentation (Randy Bass calls this "shifting the relationship between argument and artifact," 276).

Yet searchers are eager for this. They avoid books, libraries, and librarians. They love Google. (We love Google.) Why? Maybe it's the speed of getting thousands of potential answers in a split second (Fig. 11). Maybe it's the seeming simplicity of the list, or the delight in the reduction of a complex question to a series of puzzle pieces to be (apparently) easily assembled into something new. Maybe it's the appearance of neutrality in the list, its lack of point of view. Or maybe, hatred of reading, hatred of learning, fear of librarians, fear of looking stupid, the absence of a teacher's or a librarian's interference. Or, a recognition of the limits of print sources; hurriedness; desire for anonymity or for "interactivity." Or maybe it's the joy of accumulation, the reassurance of hoarding. In the end, probably, the glut of information itself is a reward. As Kenneth Burke wrote, "information ... is intrinsically interesting" (43; adding, however, that "by intrinsically interesting I do not necessarily mean intrinsically valuable," 43).

I think we persuade ourselves that information itself will grant us knowledge. (We never thought books as books were persuasive, but books didn't compile themselves at our command.) And we eagerly open the floodgates.

For fifty years at least, researchers have complained of this inundation: "… we are being bogged down … . The investigator is staggered by the findings and conclusions of thousands of other workers--conclusions which he cannot find time to grasp, much less to remember, as they appear" (Bush 101). And our vision of how the "modern" searcher will navigate has not changed much since the mid-twentieth century: "You will be able to dial [some words] into the catalogue machine … . There will be a flutter … in the machine. Soon it will stop, and, in front of you on the screen will be projected … the names of three or four books …" (Edmund C. Berkeley, qtd. in Buckland 286). In other words, a searcher enters text into a machine, the machine whirs a bit, then presents the searcher with a list of possibly relevant items (see Buckland for a survey of pre-digital devices).

As we've mechanized information retrieval, we've also tried to mechanize the creation of meaning, since information ceases to be interesting if the searcher cannot make sense of it. But machines have not shown great promise in taking part in the conversations that construct sense. For John Searle, a conversation does not happen because of rules about turn-taking or such, rules a machine might master, but because of "shared intentionality" (15). When people speak with one another, each has knowledge about the world. This knowledge remains unspoken, but leads to the conversation and supports its interplay. The knowledges of the speakers do not match perfectly, but overlap enough to make interaction possible. Searle calls this background. The background will vary from conversation to conversation, but will never be laid out in full and does not need to be. As Searle says, "all meaning and understanding goes on against a background that is not itself meant or understood" (19).

We've pretended that the forms in which we present information have little value in themselves and add nothing to information. (But what they do add is not necessarily meaning.) When Burke predicted in 1931 that "Atrophy of form follows hypertrophy of information" (183), he was discussing the glut of information within a literary work, but I think his statement also applies to the information within the body of writing in the world. Too much makes us want to surf; we can't linger over one because there's always another; it wants to surfeit. The Meet Me at Hot or Not (Fig. 14) interface exemplifies this. The searcher is always immediately presented with output, even before providing any input, and each output entry asks for further input: "Do you want to meet me?" If the searcher clicks on no, the interface immediately provides another output entry (Fig. 29). To me, no single graphics file in Hot or Not is particularly inviting, but I always feel compelled to click to see the next one; there's always some possibility that the next one will be interesting. This gorging makes coherent, intelligent argument difficult and makes eloquence impossible. It leads to a passive negation of the true interactivity of an engaging presentation; in Burke's words, "The contemporary audience ... is content to have facts placed before it in some more or less adequate sequence" (48). In fact, if I read Burke correctly, the only beauty to come from this surge of information is silence, which he suggests is "the 'flowering' of information" (49).

Much of Walter Benjamin's massive Arcades Project reads like a KWIC list:

And much of Benjamin's work leads back to silence. He compares study and research to a hunt on a chaotic, oceanic savanna of information, to tracking and killing prey; he says, "The experiences of one who attends to a trace ... have no sequence and no system. They are a product of chance ..." (801-802); a student now is "as free intellectually as hunter and herdsman were free sensually" (Oswald Spengler, qtd. in Benjamin 806). What was once a physical activity is now mental; industrialism has separated information from narrative and so destroyed part of truth (Benjamin 804). Once the animal or object or fact is captured, its disturbing "presence" can be silenced, with "its integration into a new, expressly devised historical system ..." a vast, silent web in which "every single thing becomes an encyclopedia of knowledge ..." (Benjamin 204-205; for further discussion, see Smith 13).

The trends of technology seem inexorable. If we think of information as a sea whose level is ever rising, there is little we can do. For now, one dike we can maintain is old-fashioned talking: classroom teaching and the library reference interview, for instance. In a reference interview, the searcher often does not really ask a question, but rather ends up answering questions put by the librarian. Sometimes several minutes of discussion take place before a hand is even placed on a computer keyboard or a book is opened. And when output does appear on the screen or the proper page is turned to, the librarian is there to interpret and to question further.