• No results found

8.6 Koko

8.6.1 Koko’s naming of drawings

More promising than the naming of photographs, as indicative of a fuller pictorial competence in Koko, is her ability to adequately respond to the content of non-photographic pictures, like line-drawings and cartoons, especially if the pictures can be shown not to have been used in conversation before. These requirements are un-fortunately not fulfilled for the case presented in Patterson (1978b) or Patterson and Linden (1981) where Koko is commenting that the cat mother is angry, the kittens are crying and that something is “bad” in connection to a story read to her by Pat-terson. The story is illustrated with colour drawings. But we do not know if Koko is commenting the pictures or the telling of the story, and we do not know if this is the first time the book is ever read to her.92

In Patterson and Linden (1981) a second anecdotal illustration of Koko respond-ing adequately (from the point of view of a scientist) to a line-drawrespond-ing of unknown history is given. “As part of a vocabulary test, Penny asks Koko to find ‘crying’ and Koko points to a picture of a child crying” (p. 56). The point of this illustration is to demonstrate Koko’s linguistic capabilities. If the pictorial part of this or other events would have been unusual for Koko, a point would probably have been made about it.

Koko is subjected to non-photographic pictures (as well as photographs) in the regular testing of her vocabulary. The only readily available published data for such a testing session that involves pictures, where the types of picture are specified, and where they furthermore might be suspected to be novel to Koko, is a test using the Assessment of Children’s Language Comprehension (ACLC) material (Patterson &

Linden, 1981). This test was given to Koko at the age of 4.5, with potentially im-pressive results from a pictorial perspective. It is reported that Koko did not receive any drilling or training before the test, which can be taken to mean that also the test material, i.e. the pictures, were novel to her. However, it can also mean that only those aspects intended for testing, i.e. linguistic comprehension, were not preceded by drilling, which excludes the pictures as such. This is even likely since one would like to make sure that the subject understands the testing material. A couple of the pictures used are shown in Patterson & Linden (1981) (see fig. 11). They described different objects, attributes and relationships between objects. In the vocabulary part of the test Koko had to point to the appropriate item in a picture, and in the com-prehension part she had to choose which scene among four or five that depicted a specified relation between objects. For example: “point to the bird above the house”

or “point to the broken sailboat on the table.” The latter example thus implies, pic-torially, recognising a broken sailboat, a table, and the relation between the two.

Interpretation in relation to the non-matches had to occur. The complete test con-tained 40 picture cards of with 30 depicted scenes like this.

92 In A Conversation with Koko (Brennan & Visty, 1999) the very same book is described as one of her favourites.

Figur drawi the g recep From (1981

Koko cantl and rately 103) ily a elem enter error task, it is were cond cond we d A same disto prob of wh A Peab peop orde not i beca son &

A from (Sch the “ outsi

93 Vik plete

94 Thi teract

95 She invest

re 11. An exa ing stimuli gorilla Koko’s

tive English m Patterson


o perform ly better th

“slightly y than edu ). The resul across prob ments in the rtain the id rs. She attri although K not stated e only used ditions. Info ditions is n do not know Also at age

e – differen orted drawi blems includ

hiskers.93 And when sh

body Pictur ple in action

r to acquir increase her

me worse w

& Linden, A trial from m the origi

roeder, 197

“bird in the ide of a bas

ki, the chimpa

from incomp is documenta tion.

e performs he tigation.

ample of the used to test s (4.5 years) vocabulary.

and Linden

ed signifi-han csignifi-hance less accu-ucationally lts were dif

lem difficu e request), b dea that pe

ibuted the Koko did r whether th d in one con

formation o ot given. I w if there w

4.5, Koko nt judgemen

ings” (Patt ded drawin he was 6 ye re Vocabula n. Patterson re repeated

r scores, wh with repetit


a task simi nal set, ca 78/2006).94 e basket” w sket, somet

anzee in Hay plete drawing ary is a valuab er indication b

e t



y handicapp fficult to in ulty from a but she mad rhaps the p performan eceive moti he pictures ndition, bu on how the f pictures w were learnin

seemingly nts, as well erson & L ngs of hand ears old Ko ary Test,” w n admits th

measures o hich were v tion, seemin

ilar to the A an be seen

4 At the tim without a fli

hing in a ca

yes and Nissen gs in a similar

ble record of by kissing the

ped childre terpret, tho a linguistic de several e pictorial sti ce to attitu ivating edib

were reuse ut the rema e pictorial s were reused ng effects fo

y outperfor l as detectin Linden, 198 ds missing a

oko receive which (to m hat Koko w of her perfo very good fr ngly due to ACLC desc

in the do me Koko is inch.95 The

art, and las

n (1971), also r manner (see

Koko becaus e image, an im

en” (Patter ough, becau c perspectiv errors on th imuli as su ude factors,

ble treats fo d or not. T aining 30 w stimuli wer d, trial one or the pictor

rmed huma ng “flaws in 81, p. 127) a fingernail, ed, among o

my knowle was given th ormance, b from the sta o boredom

cribed abov ocumentary s about 6 y

three non-stly an emp

o showed tha e table 2, secti e it contains mportant rem

rson & Li use Koko p ve (the num he whole. P

ch were th such as bo or her work The 10 voc

were used i re distribut e data is ne rial content an 6 year o

n a series o ). The pict , or a cat w other tests, edge) conta he same test ut also tha art. On the with famili ve, probabl y Koko: A years old an -matches se pty basket. W

at she can disc ion 7.3).

long unedite minder that n

nden, 198 performed s mber of cr atterson did he cause of

oredom wit k. Unfortun cabulary pic n nine diff ted among eeded, othe


old childre of incomple ture-compl with only on

“form B o ains drawin t many tim at familiarit e contrary, K

iar tasks (P ly using pic

Talking G nd she pinp eem to be a When requ

criminate com ed episodes of not all mouth

1, p.


d not these th the nately ctures ferent these erwise en on

ete or letion ne set of the ngs of es, in ty did Koko atter-ctures Gorilla points a bird uested

m-f in-hing is

to find the alligator in another vocabulary trial she kisses, then points to, a line-drawing of an alligator. Non-matches are a mouse, a kangaroo and a goat. In a pro-ductive vocabulary (and pictorial) test96 she is asked to name line-drawings depicting a tree, a hat and a knife. She fails on the knife. The potential difference between

“receptive” and “productive” pictorial comprehension will be expanded upon in Chapter 13.

As previously, Koko’s history with the pictures in the above episode is unknown, but her behaviour towards pictures often suggests that it is not merely a case of memorising previous answers or conversations. An example from the above docu-mentary (Schroeder, 1978/2006): Koko, lying down by Patterson, browses a picture book that contains drawings. She starts out by viewing the first page for a while and taps a few times with her index finger at the sea lion enclosure in the middle of a zoo, an action that is difficult to interpret. She then turns the pages quickly with her lips, seemingly glancing at each page. She stops and points to a group of assorted flowers. As spectators we cannot know why she does this but the voice of Patterson hints to us that they look beautiful to her or Koko. Then there is an editing cut, and in the next frame Koko points to a red square and signs red, her favourite colour.

She seems to address this to Patterson. Then, on the same page, she points to an-other square-like image, but yellow, with a small spider depicted in its centre. The tapping of Koko’s finger is targeted at the spider. Koko emphatically hits her hands together which Patterson translates “that is a scary bug, scary bug spider” while Koko sits up.

It is difficult to judge what modes of pictorial processing are operating here, col-our is a feature easily processed in surface mode and it could have been colcol-our and not “flowers” that were extracted from the flower picture etc. Bugs and spiders are one of those things that often look quite bug-like even as drawings.97 However, there is an element of involvement, if not captivation, in all of the latter three actions to-wards the pictures, which hints to us that Koko actively acts toto-wards, and perhaps about, these pictures. Furthermore, she is in no way guided in her responses by an interaction with Patterson, other than perhaps by the memory of earlier rewarding interactions around the same or similar pictures. The script she is going through of pointing and naming and socialising with a caretaker might all be part of a drill she has developed through her countless sessions with picture material, but her actual choice of pictures to respond (adequately) to does not seem to be part of a stereo-typed drill (perhaps with the exception of the sea lion). Her reactions seem to unfold spontaneously as she browses through the book. Unfortunately, a few seconds of data from an edited documentary do not allow for the full necessary analysis. How-ever, the point is that how an animal, or human, behaves in a task, be it informal or experimental, is sometimes just as important as the figures describing the result. Had we seen the same actions in a human child we would not doubt that the child

96 The terms “receptive” and ”productive” are borrowed from language research where a difference is made between e.g. receptive and productive vocabulary. In this thesis I use ”receptive” when the investigated picture is on the choice-item side in MTS, and ”productive” when the picture is on the sample side in MTS, because that is how they appear in language testing.

97 A drawing of a bug encased on the backside of a laminated drawing of an apple received close visual inspection by a bonobo in a test session (see section 13.6). The apple received less attention.

tively interprets the pictures. Human children also develop habitual social responses towards specific pictures in picture books (i.e. “stable reading routines” in Fletcher

& Reese, 2005), but that does not mean that they cannot be stereotypical and rec-ognise the content of pictures at the same time. Habits do not need to be a sign of drilled, rote learned, behaviour.

Is it possible that Koko’s performance is due solely to a good memory for paired picture – referent associations? When attempts were made to introduce a voice com-puter to Koko she learned to use it quickly and transferred her words to the arbitrary geometrical coloured shapes on the keyboard (Patterson & Linden, 1981). She soon typed out requests like “want apple eat” and “want drink sip.” If Koko can learn to ascribe meaning to new symbols easily (unfortunately we do not have a learning curve on the ability) perhaps she can also learn the meaning of pictures in this way, i.e. is as symbols rather than icons, bypassing visual similarity. By applying iconicity between pictures, rather than to the real world, learned meaning has the potential to generalise to novel pictures as well. If this is a possibility the prediction would be that shapes that have a tendency to remain the same in different pictures would be more frequent in Koko’s comments on drawings, than shapes that have a large vari-ability.

I have for example noticed in photographs and movie clips of Koko on the pro-ject’s website (The Gorilla Foundation, 2007) that several comments of hers towards drawings on greeting cards, patterns on clothing, and even engravings on a cufflink (which was actually of a sun), are examples of flowers. A flower, as it is typically por-trayed in handmade pictures, might be one of those stable shapes that easily transfers and is recognised across depictions (as well as potentially becomes confused with certain suns). That said, looking for depictions in cufflinks and clothing, besides on paper, speaks for an ability of Koko’s to spontaneously look for iconically based meanings in objects, even though it might be limited to certain motifs, like flowers.

Another interpretation, though, is that when flower-like motifs hit Koko, she com-ments on them. This would be the very opposite of looking for faces in clouds. A targeted study of this ability is greatly needed.98