making use of the lexigram boards to communicate with human laboratory workers,
showing, as Savage-Rumbaugh puts it, that he had been ‘keeping a secret’ (Savage-
Rumbaugh et al. 1998: 22), concealed by his indifferent progress in prior trials with the
boards. On the day before Matata’s departure, he used the lexigram board on 21
occasions, asking for 3 different foods. On the following day, he produced 120 lexigram-
acts exploiting 12 different symbols (Savage-Rumbaugh et al. 1998: 22), twice what
Matata had mastered in two years. Savage-Rumbaugh claims that the sudden change
suggested that what had changed was not ‘his knowledge but [...] his motivation’
(Savage-Rumbaugh et al. 1998: 22). Consequently ongoing study of Kanzi focussed less
on repeated trials, and more on interactions between him and human laboratory workers.
An aspect of this shift which we regard as especially important is that in the resulting
environment there was a great deal for Kanzi to gain from working out how to
manipulate his generally attentive, co-operative, and often indulgent human companions,
and to do so with increasing sophistication and precision. Kanzi, then, led a life far
closer to that of human infants than most ALR subjects.
Both of the features of Savage-Rumbaugh’s research just highlighted (the lexigram
boards as part of an extended mind, and Kanzi’ s own biography) suggest that standard
features of debates over the poverty of the stimulus should be re-evaluated. Such debates
generally share commitment to the notion that the infant learner is a solitary
epistemologist, attempting to make sense of external data on the basis of internal
processing, and that it does so with a strikingly scholarly disinterest, or a bare appetite
for generalisations. This results in undervaluing or ignoring the ways in which non-
neural resources can augment and transform cognitive capacities, and the ways in which
social interaction can provide both powerful incentives and mediating structures that
support the learning process. If these commitments are conjoined with the tendency,
noted above, to focus on the abstraction-amenable aspects of language, the result, we
argue, is a grievous misconstrual of the nature of the stimulus and the learning problem,
but most strikingly of all, of the nature of the learner.
In the second part of this paper, we present a largely descriptive account of a selection of
key episodes - one involving an infant and its mother, one with a child and its father, and
one with three interacting adults. We aim, in so doing, to show what it is possible to say
about, and identify in, the behaviour of interacting humans when unencumbered either by
identification of language with only its abstraction-amenable aspects, or by the view of
infants and children as disembodied, or solitary, epistemologists. The re-evaluation of
the nature of the learner and of language that this descriptive work suggests, is a further
elaboration of the ways in which minds can be extended.
The ‘How’ Question
We call the question which we want to put at centre stage the ‘how’ question: How can
anything come to count as a symbol?7 We don’t say be a symbol because, like Clark (e.g.
1993), we are wary of many of the associations carried by the notion of symbols in
debates about cognition and language. Any reference to a symbol is too likely, on our
view, to suggest some kind of token with fairly precise individuation criteria,
determinate intrinsic syntactic properties, and capacities for being more or less literally
moved around, operated upon, and combined with other symbols, often in the head. Of
course, whatever is in (and around) the head, it is undeniable that a great deal of what
goes on with people can be described in terms of symbols, and structured arrangements
of symbols, as well as rules for operating on and with symbols. We want to remain
tactically agnostic about what actually goes on under the cognitive hood, so as to try and
get a better handle on a particular set of phenomena that we think would be possible
assuming too much about symbols.
6 / 23