• No results found

Sonesson (1989) has convincingly argued for the picture being a sign, en par with language symbols in terms of referential potential. Often pictures are predominantly iconic signs, but this is far from true for all objects that we call pictures. Sonesson reserves the term pictures to icons based on primary iconicity. He discriminates be-tween two types of iconic signs, depending on whether likeness precedes the sign relation, and even is the reason for appreciating it, or whether likeness is only dis-covered as a result of an appreciated sign relation. The first type of likeness has been called primary iconicity, and the second type has been dubbed secondary iconicity.

The main argument for the existence of primary iconic signs is that a naïve viewer is able to decode such signs without instruction, and that cross-cultural and child data supports this case (e.g. Sonesson, in press b). However, the data for this argu-ment is inconclusive. The Hochberg and Brooks’ (1962) child study is just not enough, and cross-cultural data paints a mixed picture. Recognition of non-photographic material in picture naïve subjects has only been demonstrated after instruction, or in a setting that scaffolds the subject’s view of the stimuli as informa-tive. There are several instances in the literature of complete failure to recognise de-pictions, even photographs. If instruction or scaffolding is necessary the presence of secondary iconicity cannot be precluded. The extant data rather reserve primary ico-nicity to photographs, if even that. However, it should be said that the subspecies of primary and secondary iconicities in turn can be numerous (Sonesson, in press a), and if this is indeed the case one might expect that some of them are not easily as-cribed to a pre or post sign existence.

However, rather than focusing on the possible interactions between primary and secondary iconicities in pictures, Sonesson (e.g. 1989) argues that there is a hierar-chy in the real world of suitable and unsuitable mediums for signs, which accounts for the problems of recognising pictures as primary icons. That which is more prominent always serves as comparison to that which is less prominent, not the other way around. The reason picture-naïve people need guidance is because they struggle with this hierarchy. If they only get to grips with the surface of a picture, iconic interpretation will follow naturally. Beyond this point iconicity will more easily precede the sign relation and more types of pictures than before can qualify as primary iconic signs.

The process that is here proposed to account for the fact that iconicity can be appreciated before the sign function, even in non-realistic pictures, is successive ap-proximation (section 3.3), or resemanticisation (Sonesson, 1989) (see fig. 6).34 This implies that parts and wholes define each other in continuous feedback. Correspon-dence, and thereby recognition, occurs on the level of relationships rather than dis-crete features. An illustrative case in point is the caricature face drawing where

34 Sonesson himself would perhaps rather invoke the ecological optics of Gibson (e.g. 1979).

vidu and can preci conv icons

Figur resem point critica the lin

Rese powe tions relat iconi expre too s itself proa clue secon preta relies time The press T cause simil trans a pic ject.

a sub perce categ perce func mod emp N and whic

al features swift recog continue to ision where veying mor

s (Sonesson

re 6. Gregory manticisation

ted out by th al position. T ne as a cigare

emanticisati erful in a p s. As ment

ion has bee icity predo ession. The scant, or to f. However ached first a

has been fo ndary iconi ation next t s on interp e. The cont

viewer mu s a) agrees t The discrim e it entails larity and slate well to cture only i One could bjective pic eption of b gorical. Per eption of ( ction being de and is th irical work Now that w picture-naï ch pictures

can be gro gnition (e.g

o induce e e left out p e informati n, in press a

y’s tree/smok where nose he simple add The facial fea ette. From So

ion is of c post-sign si ioned abov en establish minantly e e source for oo ambival r, many suc as a conven ound that h

ic sign. (Th time one se pretation, a tent of the ust therefor that a sign c mination be

s two very the sign re o two types

in one of t d perhaps sa cture in onl both prima rceiving sec (objective)

present. Th he reason fo in Chapter we know wh

ïve cultures seem to be

ossly distor g. Perkins &

expectations parts should ion than go a).

king lady. A , chin, hair dition of a li atures in turn onesson (1989

course even ituation, w ve, some si hed. Such a exists in the

r similarity lent, and n ch pictures, ntional sign hints at ano his in turn ees it.) Sinc a particular

picture, ho re settle for

can mix pri etween prim

different t elation. Ho s of pictures two possibl ay that the ly one of th

ry and seco condary ico

pictures in his is perha for why I h

r 13.

hat picture s interact w

processed,

ted, but th

& Hagen, 1 s on the p d have appe

oes into th

case of etc. is ine in a n define

9).

n more where attent

imilarities c a picture is e mind of th y is in the e

no help com perhaps es n, but as an other way o

could perh ce iconicity r picture c owever, can r one view imary and s mary and s types of ca owever, as s. It would le views by picture is a hem. Howe ondary icon onicity is in n a pictorial aps the best have used d es and icon with them,

let us turn

heir interrel 1980). Foll

arts. This eared in the heir produc

tion is con can only be s more of a

he observer xpression, mes from c specially am

n ordinary of seeing it,

haps be the is not an o an have m n only be on

or swap be secondary i secondary ausal relati

suggested, mean that y the same

an objective ever, since nicity, we d n fact the m

l mode, sin t way to pre drawings as nicity are, a

and have n to the non

lations still lowing reco

allows us t e picture. T tion is a un

nstrained by e discovere a conventio r, rather th but the inf context wit mbivalent o iconic sign the picture new prima objective sta more than o ne at every etween them

iconicity.

iconicity is onship bet this distin

the same p pictorially e picture in

pictorial co do not nee most telling nce it depe

eclude proc s the critica and are no

a framewo nhuman pri

l allow effo ognition, w to imagine The potent

nique featu

y other exp ed after the onal sign, w han extant i

formation i thin the pi ones, are no n. Then, af e can turn i

ary iconic i ate of affair one nature y given mom

m. Sonesso s importan tween perc nction does physical obj

competent n both cases ompetence e

d to be qu g evidence nds on the cessing in r al stimuli in

t, how chi ork of mod

imates.

ortless wholes with tial of ure of

pecta-e sign where n the is just icture ot

ap-fter a into a inter-rs but

over ment.

on (in nt

be-ceived s not ject is t

sub-s, but ntails uite as for a e sign eality n my ildren des in

II

The program we have described avoids the question of whether an animal other than man can acquire language.

As comparative psychologists we must reject this question. It is like the question of whether an animal other than man can have thoughts. It depends on the definition of language rather than on the observations of what animals do.

Allen Gardner and Beatrix Gardner (1971, p. 181)

Chapter 5

Primates in picture tests

Given what we have learnt from the empirical work with humans presented in Part I, we can assume with some certainty that pictorial competence in humans is a de-velopmental matter, including both cognitive and cultural growth. It is then in its place to reflect upon what one can expect to find from scrutinising the use of pic-tures in animal research. We assume that animals do not share our pictorial culture, so what use is it to study this in animals? There are at least three areas, all connected to humans, that makes the endeavour worthwhile.

One obvious reason is to be able to say something about the onset of iconic com-petence in human development as a species. What are its requirements and early expressions? The comparative approach has always been a popular way to recreate human prehistory. This thesis will not be that ambitious, however, but I am sure that interested readers will be able to find useful information in this text for those kinds of inferences.35

An area where human and animal lives intercept, which involves pictures, is in experimental settings. Pictures are used as convenient stimuli in place of real objects, and far-reaching conclusions are drawn from such research. All three modes of view-ing pictures have different implications for how the results are best interpreted.

The claim made further above, that pictures are exclusive to human culture, is not really true. Pictures are also integrated as part of some animals’ everyday life, for example in ape language research. It is worth studying if these individuals perform differently with pictures than those where pictures are less integrated, say at an an-thropoid station in Tenerife. We can in this manner investigate the effects of lan-guage and culture on pictorial competence as such. But we can also investigate con-cepts and imagination in new ways, by looking at categorisation and interpretation of pictorial material. This is applicable also to language-naïve subjects, if the step to a pictorial mode can be made. In this way, working with pictures can replace lin-guistic symbols as a window to the ape mind.

To what extent nonhumans are able to replicate human development depends on what similarities are there from the start, what is open to learning, and what is con-strained by species differences. However, the contrastive approach is not the only one. A second one, that I will tend to favour in the remainder of this thesis, is to

35 A recent attempt to relate iconicity to cognitive evolution by the present author and colleagues can be found in Zlatev et al. (2006).

investigate what nonhuman primates are actually able to do with pictures, irrespec-tively of the humanness of those behaviours. The three modes of picture compe-tence, i.e. surface, reality, and pictorial modes, can stand on their own as a frame-work, regardless of typical human performance.

Pictures have indeed been used extensively as stimuli since the early days of animal cognition research, both as substitute for real-life objects, individuals, and events, and as abstract stimuli. They have for example been used to assess questions of spa-tial representation, social cognition, viewpoint consistency, and serial list learning (Fagot et al., 2000), as well as all sorts of discrimination and categorisation (Bovet &

Vauclair, 2000). Pictorial stimuli seem to have been used with such success that few scientists have paused and asked the questions of why it works, and what it means.

Since animals of all kinds readily accept pictures, in the right circumstances, as ex-amples of real-world objects without fuss, it must mean that pictures are simple and intuitive phenomena.

This observation easily leads to the idea that there is only one way of viewing pic-tures: you either see what is in them, or you do not. The most common mistake in the literature is thus to lump performance in reality mode as performance in picto-rial mode, or neglect to control for surface mode processing. However, it should be noted that we do not know about all those unsuccessful and unpublished attempts at training animals in discriminating pictures. Bovet and Vauclair (2000) could only find ten published papers between 1953 and 1998 that demonstrate difficulties for animals in recognising pictures. Most certainly a vast number of pictures have been dropped from experiments because subjects have had difficulties decoding them.

And many subjects might have been dropped too.

Bovet and Vauclair (2000) cite more negative examples with birds than with pri-mates. Photographs and video images are made for human (i.e. primate) vision, while birds have a very different visual system (e.g. Delius et al., 2000). It might also be the case that primates can conceptualise pictorial stimuli in a way that birds can-not. Bovet & Vauclair (2000) concluded that “[...] picture recognition in animals is not obvious and is dependent on experimental factors” (p. 158). What those factors might be will hopefully be made clear in the chapters to follow.

There is a serious problem related to the ease by which we grant animals a complete pictorial competence by looking at their behavioural performance only, and disre-gard the underlying processes. If you do not ascribe referential abilities to animals, the option is to explain their competence with pictures based on a complete corre-spondence between pictorial stimuli and real stimuli. Experiments that substitute real objects or individuals for photographs, and then draw conclusions about the everyday functioning of the subjects’ perception or thinking, do exactly that. Con-clusions based on such experiments can be misguided if it turns out that animals perceive real objects and depicted ones differently. But not only should we avoid presuming that subjects lack a picture concept of some sort, we must also avoid the opposite. When we use pictures and presume that animals must appreciate that there is a difference, maybe none is perceived, at least not of the kind that we expect. As I will continue to suggest in the present part of the thesis, reality mode can be quite

broad and can encompass certain degrees of “magical thinking”36 and unusual in-stances of “reality.” After all, an object in a picture can differ markedly from ordi-nary reality in appearance and behaviour. But often animals still seem able to recog-nise such content, while at the same time not being able to use pictures in referential tasks. They even act out on pictorial displays that should, from an objective point of view, facilitate differentiation.

In the coming chapters I will review primate (and some pigeon) experiments that use iconic stimuli (pictures, replicas, scale-models, video, and mirrors) for one pur-pose or the other. I will also review some observations and anecdotal evidence from the literature. Such data is useful for painting a picture of the potential and variety of behaviours with pictures that can be expected from apes. But let me start, in this chapter, with a closer look at those few experiments that have been aimed directly at picture comprehension, regardless of which views one have had on the ability at the time.