More on Monkey Talk

To the Editors:

In her reply to Martin Gardner’s review of my book, Nim [NYR, March 20], Francine Patterson [NYR, October 9] questions my negative conclusions regarding an ape’s linguistic competence. She does so without coming to grips with the facts which led me to reverse my original interpretation of Nim’s multi-sign sequences as sentences.

Nim’s signing, and that of the other signing apes as well, appears to be motivated more by a desire to obtain some object, or to engage in some activity, than a desire to exchange information for its own sake. First the ape tries to obtain what it wants directly—without signing. When reminded by its teacher that it must sign, the ape often signs until the teacher complies with its request. The critical question is whether the ape is generating sentences or simply running on with its hands until it gets what it wants.

Careful scrutiny of the ape’s utterances favor the latter interpretation. Consider, for example, a typical exchange in which the teacher signed you play cat?, and Nim responded by signing me Nim cat play. Two of the teacher’s signs are combined with two general purpose signs, signs I refer to as “wild cards” because of their universal relevance. These and other features of an ape’s discourse with its teachers were discovered by painstaking frame-by-frame analyses of videotapes. Unlike the sentences of a child, Nim’s combinations amounted to unstructured mixtures of signs. Some are imitative of the teacher’s prior utterance; others are selected unsystematically from a small group of “wild card” signs.

Patterson rejects this interpretation of an ape’s sequences of signs on the grounds that three and a half hours of videotape of Nim signing with his teachers provided too small a sample and that the data we did collect were artifacts of the pressure Nim’s teachers exerted in getting him to sign. Both of these arguments are at odds with statements and data that appear in Patterson’s dissertation (Stanford U., 1979), a document that impresses me as the most thorough of Patterson’s publications about Koko’s signing.

Psycholinguists are in general agreement that a child’s production, as opposed to his comprehension, of language provides the most telling evidence of grammatical competence. Consider Patterson’s assessment of Koko’s production of signs. “The majority of Koko’s utterances were not spontaneous, but solicited by questions from her teachers and companions. My interactions with Koko were often characterized by frequent questions such as ‘What’s this?’ ” (p. 153).

Patterson’s dissertation contains five one-hour transcripts of videotapes of Koko signing with her teachers (the only such transcripts that have been published to date). I found no evidence that Koko’s use of sign language differed from Nim’s. Just as Nim was prone to produce long utterances such as give orange me give eat orange me eat orange give me eat orange give me you, Koko produced utterances of structurally unrelated signs such as mess red thirsty mouth thirsty (p. 339) and please milk please me like apple bottle (p. 345).

Until Patterson publishes data to support her view that my videotapes of Nim signing with his teachers were obtained under “artificial and high-pressure conditions which very likely contributed to the high levels of interruption and imitation,” I see no way to evaluate that claim. I also suggest that Patterson perform a discourse analysis of Koko’s signing as shown in a documentary film in which she participated, Koko, the Talking Gorilla (New Yorker Films). The scenes of that film, in which both Koko and her teacher are visible, left me (and many other viewers) with the clear impression that the teacher initiated most of the signing and that Koko’s signing was highly imitative of the teacher’s utterances.

That Koko could perform at better than chance levels on a comprehension test of novel utterances in sign language and in spoken English is not evidence of grammatical competence. As I have argued elsewhere (J. Exp. Anal. Beh., 1979, 31, 161-175), the type of problem that was administered in this test can be solved by applying non-grammatical strategies.

Much of Patterson’s letter is devoted to defending what Martin Gardner describes as her “exceptional claims” of Koko’s signing. While not directly germane to Patterson’s comments concerning my conclusions, they do raise questions about her criteria for characterizing Koko’s signing as language. Totally absent from Patterson’s description of Koko’s ability to rhyme, to use metaphors and such abstract concepts as because and imagine is the kind of training needed to establish such linguistic skills. No mention is made of just how a gorilla, who can’t produce human phonemes, learns to identify English words that rhyme with one another. Claims that a gorilla is as competent to produce metaphors as a seven-year-old human child cannot be evaluated without knowing what kind of training was used to get the gorilla to appreciate a metaphorical use of language. Without such information, one is left with the impression that Patterson is simply projecting onto her gorillas’ hand movements what a human child might do in similar circumstances. These and other aspects of the superficial assessments Patterson makes of the “linguistic” achievements of her gorillas can be overcome only by a rigorous description of their training and testing histories.

H.S. Terrace

Columbia University

New York City