Space in Language and Cognition_ Explorati - Stephen C. Levinson2

  Beyond language: frames of reference in wayfinding and pointing In the prior chapters,we have seen that language,together with other semiotic systems,seems to have a decisive impact on the choice of an in-ternalcodeforspatialmemoryandreckoning.Inthischapter,weexplore otherpossibleramificationsofframe-of-referencespecialization.Thelit-eraturereviewedinChaptersuggeststhatthereareamyriadofinternal representations of space for different sensory modalities and purposes. If language can be shown to influence the choice of frame of reference for the spatial memory of small-scale arrays,what about larger-scale arrays and mental models of the world around us? After all,spatial cognition must centrally be concerned with locating ourselves in a ‘mental map’ of the environment,and finding our way around in it. This chapter pur-suesdifferencesinthecognitionofwayfindingandorientationthatseem deeply linked to specializations in frames of reference in language. But the chapter also pursues another theme,the cross-modal nature of these frame-of-reference specializations. For the best evidence for wayfind-ing abilities and the nature of mental maps comes from pointing and gesture – that is from the motoric output driven (at least proximately) by kinaesthetic representations. Unreflective gesture gives us insight into another level of mental life,representations of space that are at least par-tially independent of language,and that seem close to the very heart of ourspatialthinkingandspatialimagery.Wecanthereforelookatgesture as a special window on underlying spatial cognition. .         .. The nature of wayfinding abilities There is a large literature on the navigation of non-human species (see, e.g.,Sch¨one ,Waterman ,Gallistel  for references),which  The role of language in everyday human navigation  reveals wondrous dead reckoning mechanisms in the simplest of ants through to the miracles of the arctic tern which flies from one pole to the other and back again every year. Some of these mechanisms appear to employ decidedly specialized sensory equipment,for example the ability to get fixes directly from polarized light or the earth’s magnetic field (Hughes ),mechanisms that seem to be denied to humans (but see Baker ). When we turn to the study of routine human navigation,we find no correspondinglyrichliterature.Ofcourse,thereisawealthofknowledge about marine navigation,but in our own culture this has been largely formalized since at least the fifteenth century,and is the province of ex-perts. Such expert knowledge is not necessarily in the navigator’s own head,for it is an accumulated lore and science,built into maps,instru-ments and procedures for their use (Hutchins ). Something is also known about the traditional marine lore of a few Polynesian societies, where navigation is purely mental rather than using the external plot-ting and calculating devices of western navigation,but nevertheless also constitutes expert knowledge rather than everyday practice (Gladwin ,Lewis ). Naturally,psychologists know a large amount about human abilities to estimate distance and angle on the small scale,and have shown exactly how able we are to glance at a scene and then steer our way through it without vision (see,e.g.,Lee and Thompson ). They have shown,for example,how the blind are also able to extrapo-late from experience of a route to a short-cut from one spot to another (Landau et al. ),and a few have carried out experiments outside the lab (see,e.g.,Baker ). Geographers have done much to elucidate for us the kinds of mental maps and other constructs urban-dwellers use to find their way around (see,e.g.,Golledge ,Golledge et al. ). Still,I think it must be conceded that in many ways we know much less about navigation in our own species than amongst birds,bees and ants. Apart from the efforts of the geographers,there are simply relatively few studies of how humans actually find their way around real novelenvironments,orcalculateangleanddistanceandcurrentlocation in moving around on a scale larger than the psychological laboratory. One might have expected anthropologists to have had a keen interest in wayfinding amongst,especially,hunter-gatherer groups. But,on the whole,the information available is extremely disappointing. Work on Australian Aboriginal wayfinding,for example,reduces to a few notes by a seconded Indian policeman,some notes by David Nash,a paper by theexplorerDavidLewis,andtheworkreportedbelow.Theonlyworkof  Beyond language: wayfinding and pointing anysophisticationisthatdoneonEskimogroups(seeMacDonald). The reason that so little information exists is that wayfinding knowledge is mostly implicit and difficult to extract by explicit questioning: Inuit navigational skills are learned experientially rather than formally and it is perhaps for this reason that Inuit elders,invited by the uninitiated to talk about their wayfinding practices,never quite give a satisfactory account. Snowdrifts, wind directions and stars are all mentioned,but how these and other external markers translate into that comprehensive ability that enables Inuit to excel as wayfinders,seems to elude complete description. (MacDonald : ) And one reason why there is still so much to know is that,unlike many other animal species,human groups vary enormously in their navigational systems and abilities: navigation is quite largely a cultural matter,as shown not only in the history of European or Austronesian expansion,but also in the details of everyday life,as I hope to explain. Consequently one cannot talk of ‘human navigation’ in the same breath as one might talk of the navigation of the arctic tern (Waterman : –) or the desert ant (Gallistel : ff.). Rather,varieties of human implicit navigation may exceed the range of types to be found across a wide range of animal species,as we shall see. For obvious reasons,knowing where you are with respect to other places has a fundamental biological and cognitive importance. Even for speciesthatarehomeless,optimalforagingrequiresbeingabletogetback to places earlier located. And for animals that have bases,being able to forageandthenmakea‘bee-line’homeisclearlyessential.Observations show that when a desert ant makes such a bee-line home,it heads off in therightdirection,andthenwhenithastraversedtheestimateddistance to base,circles around to pick up final landmark cues to guide it home (Collett and Zeil : ). There seem to be two modes of operation – a system that can calculate an approximate heading and distance to base fromanynovellocation,andasecondsystemforhomingattheendofthe trajectory. Following our own culturally developed systems of nautical navigation,we can distinguish these two kinds of cognitive operation as ‘deadreckoning’vs.‘piloting’,wheredeadreckoninginvolvesestimation of position by calculating distance on each course,and piloting involves usingobservablelandmarkstohelponelocateone’spositiononamental orphysicalmap,andthustocurrentlyunobservablelandmarks(Gallistel : ). Dead reckoning is the computationally more intensive process,in-volving a procedure for calculating current position from estimates of The role of language in everyday human navigation  distance and direction travelled from a previous known position,and is the process most animals use for long-distance navigation. There are four essential ingredients: places,distances,directions and time – time comes in as a factor both in estimates of distance through velocity,and in the use of many directional cues (e.g. a compass based on the sun must allow for its daily and seasonal variations). Dead reckoning is sup-plemented by piloting (especially,as with the desert ant,in locating the precise goal towards the end of a journey),and in turn involves headings calculated directly from landmarks,for example by lining them up on an approach towards a harbour,or triangulating with bearings from a number of visible landmarks. These processes have been the subject of much human conscious rumination and are explicitly formulated in the westernmarinenavigationtradition–itistheseexplicitformulationsthat have effectively guided research on animal navigation (see,e.g.,Gallistel ). Of these,dead reckoning is especially interesting,because it in-volves implicit computation of arithmetic functions: distance along each heading must be estimated by integrating velocity with respect to time, and vectors summed to give a current location. Gallistel (: –) offers us a fully explicit computational model,arguing that accurate dead reckoning requires Cartesian rather than polar coordinates. He hasgoneontoarguethat,sinceevenanimalswithsimpleneuralsystems like the desert ant appear to instantiate such computational devices,a connectionist model of even such simple neural systems must be wrong since connectionist models cannot hold the values of variables constant justuntiltheyareneeded(Gallistel).Thusdeadreckoningpromises to be an important test area for theories of the fundamental nature of computation in organic systems. The nature of the cognitive representations involved in navigation among the different species is not well understood at the current time. The fully trigonometric model outlined by Gallistel () does not cap-turethesystematicerrorpatternsobservableinanimalwayfinding,which are better modelled in a succession of vector estimations (Mu¨ller and Wehner ). But regardless of that,the two input variables,estima-tions of angle and distance,are likely to involve multiple modalities. Angularestimationsarebasedonvariousmeasuresaccordingtospecies, desert ants,for example,using the direction of polarized light and other measures of sun position,coupled with presumably in-built ephemeris tables (i.e. expectations of the sun’s position across the day,allowing for season),whilehumansrelycruciallyonthevestibularsystemformeasur-ing rotations. Distance estimations are probably largely based on optical  Beyond language: wayfinding and pointing flow (measuring rate,yielding distance over time),kinaesthetic informa-tion (number of steps),and measures of effort (Etienne et al. : ). Thusmanydifferentsourcesofinformationhavetobecombined,under varying environmental conditions (like darkness),to yield a current esti-mation of homeward direction and distance. .. The fall from grace: why are we such bad wayfinders? Compared to other species,human natural abilities in the navigational field can only be described as extremely poor – so impoverished that we really need an explanation. Cultures have slowly attempted to recreate culturally what we lack natively,developing especially in the west over the last  years an elaborate structure of prosthetic ideas and devices for working out where we are,culminating in GPS navigational aids that at last let us rival the skills of migratory birds. In Chapter ,I will suggestthatwecanonlyunderstandthisatrophyofnativeabilitiesinthe context of a theory about the co-evolution of the human genome and culture. Naturalistshavebeenawareforalongtimeofthenavigationalfeatsof animals and insects. But it is only in relatively recent years that we have gained knowledge about how some of these feats are achieved. Many of themrelyonexoticsensesthathumansarepresumedtolackentirely–the senseoftheearth’smagneticfield,specializedsensorsforpolarizedlight, sonar systems and so forth (see Waterman ,Hughes ). What we have learnt is truly astonishing. Consider,for example,the moustache bat’s echo-location system: such a bat sends out a high frequency signal (fundamental frequency  kHz,with four formants and most energy at the second formant of  kHz) and compares the echo. From the speed at which the echo is received,the bat determines the range to a target object (at  metres distance the interval will be only  milliseconds), from the range plus loudness it estimates the target’s size,and from the differential loudness in each ear it estimates the location with respect to its own heading. From Doppler shifts (i.e. bunching or stretching of frequencies in the echo) as small as  Hz,it can calculate the speed of approachtoamovingtarget.Inshort,itcanpaintanentiresoundpicture of its spatial environment,distinguishing edible from inedible insects on the fly. It does this in the left hemisphere of a brain the size of a peanut, using neurons specialized to the harmonic frequencies of the echo and the temporal delay between call and echo,with elaborate circuitry and biomechanics to assure that the bat processes the echoes of its own calls ... - tailieumienphi.vn