One of the interesting features of the book Attention and Meaning is the way different authors have personal stories about how they found their way to interest in attention. Chapter 4 is by Kai-Uwe Carstensen, whose web page either boasts or confesses, "I am one of the few who believe that selective attention plays an important role in cognition, much more important than currently acknowledged." His chapter is titled, "A Cognitivist Attentional Semantics of Locative Prepositions," and states early on, "Attention-related phenomena… seem to have come to the fore only recently…. In this chapter, I will show that this is by no means warranted and that… attention must be regarded as a phenomenon at the heart of the field, and as an essential link in the relation of language and space" [p. 94].
He began his studies in the usual way, a graduate student with a distinguished professor as a teacher. He was interested in how locative prepositions work and began following an orthodox path. Prepositions like in, on, and above are particularly maddening because they seem so straightforward, but when you study them closely their logic seems to evaporate. Why does most of America say wait in line while New Yorkers wait on line? What is the common thread that unites a helicopter hovered over the house, clouds are over the sun, John lives over the hill, the game is over, etc.? Carstensen began taking some notice of attention's role in the early 1990s and in 1995 Gordon Logan published an important paper arguing that the discovery of space requires a shift in attention. Logan's experiments showed "spatial relations do not 'pop out' (i.e., are not directly consciously available as a whole) but always involve attentional shifts" . Sufferers from Balint's syndrome have, as one of their symptoms, an inability to shift attention between objects, and are, therefore, unable to perceive the spatial location of the object that does receive attention. It seems we do not see things in space by attending to an object but by looking at something and then another. As far as perception is concerned, space is the relationship between points of serial attention.
Carstensen developed as his hypothesis: "selective attention makes implicit spatial relations by imposing an order in the visuo-spatial processing of the involved objects" . In other words, space relations are specified by the order of shifting attention.
There might seem to be a very large number of ways attention can be shifted, but fortunately the gestalt psychologists have managed to reduce the number of things perceived to two categories: the figure and the ground. Carstensen calls the figure the LO (locative object) and the ground the (RO) reference object, but that has proved too confusing for me and I'm sticking with gestalt terminology. Gestalt psychology leaves us with only three possible attention shifts. (A fourth, ground to ground, does not work. One of the grounds becomes a figure):
- Figure to Figure (FF): We can ignore the ground entirely and shift our attention from Susan to a desk, as in Susan sits at her desk. We can visualize the two figures together, but without a ground the scene is a bit abstract.
- Figure to Ground (FG): Nancy flew over the ocean. We can visualize Nancy's airplane against an ocean backdrop.
- Ground to Figure (GF): It is less common to go from ground to object, but it is possible. Across the whole wide sea, only one sailing mast poked up.
Of course there are many more than three locative expressions, so changes in the focus of attention cannot be the whole story. Carstensen distinguishes between the linguistic and the conceptual roles of spatial location. The linguistic roles shift attention; the conceptual ones are the many culturally determined conventions for discussing space. As many as a third of all languages, for example, may not use right and left distinctions. These cultural details specify, so to speak, where to look for an object or ground. Examples,
- Nancy flew over the ocean: Nancy is above the ocean.
- Nancy swam through the ocean: Nancy is two-dimensionally surrounded by the ocean.
- Nancy sank in the ocean: Nancy is three-dimensionally surrounded by the ocean.
- Nancy went across the ocean: Nancy's relation to the ocean is uncertain.
A series of research questions present themselves.
- Do locative expressions predictably call for a particular type of attention shift? Notice that in those different examples using over -- a helicopter hovered over the house, clouds are over the sun, John lives over the hill, the game is over—all but one shifts from figure (helicopter, clouds, John) to ground (house, sun, hill). The exception is the game is over, but in that case over is not a locative at all. How reliable is the relation between locative and type of attention shift? It seems pretty clear that FG and GF shifts can take the same locative since the GF example can be revised as an FG: Only one sailing masked poked up across the whole wide sea.
- Do all languages have expressions for all three attention shifts? I am particularly interested in knowing whether creole languages express them all and how long it takes for all three to appear.
- How translatable are the cultural aspects of locatives? Presumably any speaker of any language can locate figure A in relation to a ground or figure B, but how literal can the translation be? If one language describes a person as in front of a mountain and a different language set the person behind the mountain, must the translator simply ignore all peculiarities of culture? And how much of this ignoring goes on? There are relations that are expressed in one language—e.g., Korean expresses a tight fit; German distinguishes between widths that are all the same to English speakers. What do non-Korean or non-German speakers say when they need to express differences that their languages ordinarily ignore?
This observation suggests that the universals of language and space may have to do with shifts of attention. All other aspects depend on the culture.