Through trial and error, since the researchers cannot tell them what to
And punishment is far greater when we try to shape an animal, espe-
Cially in research projects that have to deliver results in short time
According to controlled and reportable procedures.
P 32-33:
Savage-Rumbaugh’s sister and colleague Elizabeth Pugh, who initiated
the apes into many new aspects of language by exploiting how the
moment attracts their attention, describes to us how laborious her work
was before the discovery of young Kanzi’s language. In those times she
Used the simple mechanics of reinforcement, and she explains how
Deciding when to give the ape a food reward soon became so compli-
Cated that she felt that food became a problem in her interactions with
the apes, rather than a means of developing them. Food tended to come
in the way, and it made the ape’s attention to the situation one-
dimensional. It was a relief to leave this dull and laborious method
Behind and raise apes as children are raised, and talk about whatever it
was that for the moment attracted their attention. Instead of using food
as a reward, it became a way of making social life more pleasurable. This
opened the gates of life. When the bonobos got the opportunity to
develop their interests and inclinations in an environment that
responded sensitively to their responses, the process of learning new
forms of language no longer had to be fuelled or pushed forward: ‘learn-
ing tends to be much more comprehensive than suggested by the
traditional stimulus–response framework’ (Rumbaugh and Washburn
2003: 234). Experiencing this self-supporting process has been akin to
witnessing a force of nature, since we did not have to control it behav-
iouristically in the traditional sense, but could as it were stand beside
and watch its effects. Let us call this the design feature of spontaneity.
P 40-41:
When experimental psychologists test cognitive skills in apes who
have not acquired human language, the apes have to discover the task
through trial and error, since the researchers cannot tell them what to
do. If the apes accidentally discover what the experimenters want them
to do, they immediately obtain food. Their ability to solve the same or
related tasks can thereafter be tested more systematically. Kanzi’s and
Panbanisha’s tests, however, are mediated by language. We describe new
tasks for them and often they succeed immediately. (We sometimes buy
computer games in toyshops and bring them to the apes, who often
understand what to do simply by listening to the computer voice; see
NHK documentary Kanzi II.) One reason why Kanzi and Panbanisha
succeed so much better than other apes in psychological tests is the
design feature of immanence: the test situation becomes meaningful for
them in continuous conversation. This resembles how we constantly
talk with children when we take them to the doctor. Conversation
makes it easier to carry out the tests together with the apes: we achieve
a common understanding of the task. Such a meaningful psychological
test is filmed in Kanzi II, where Panbanisha shows that she under-
stands that a person’s beliefs may differ from how things in fact are.
P 44:
Language-trained apes use artificial symbols in a mechanical fashion,
rapidly and according to static patterns. Language is reduced to a gram-
matical steeplechase course towards obtaining food. Even though the
apes enjoy the challenge of the task, it is difficult to recognize the char-
acteristic pace and expressive rhythmic variations of human language
in the often tense and one-dimensional movements they produce. The
simple fact of the matter is that they are running a racetrack to earn a
gratification – that is what they are invited to do – rather than trying
to communicate freely with a fellow creature.
Kanzi uses the keyboard in a strikingly thoughtful manner and self-
willed Panbanisha often communicates while lying down. Lazily she
reaches out for the keyboard, pulls it towards her and slowly searches
for the symbols that express what she wants to say. Language use at the