Two major themes appear among the recent critiques of the attempts to teach languages to great apes: (1) whether apes can use the symbols (words) of a language as true surrogates for the objects and events they reference, including objects and events remote in time and space, and (2) whether apes can construct or understand sentences by using
combinations of words as information. The first theme, recently pursued most vigorously
by Savage-Rumbaugh and colleagues (e.g. Terrace et al., 1979), addresses the syntactic aspect – the information that is added to the string of words comprising a sentence by the structure of the string, such as the ordering of the words. In human languages it is most often necessary to account for both the semantic and the syntactic components of a sentence to interpret a sentence correctly. Semantics and syntax have been described as the ‘indispensable core attributes of any human language”.
The contention of the Savage-Rumbaugh group is that the use of symbols by apes to name objects or to make requests for foods or other desirables is not in itself evidence that the symbols function as surrogates for objects and events, as do words in human languages. Careful analysis of the majority of work with apes reveals that the symbols used are often constrained to limited contexts and are principally a learned means of obtaining a desired outcome. Typically, the words are not used to ‘refer’ ( cf. Terrace, 1985) but with special training may eventually take on that function ( Savage-Rumbaugh and Rumbaugh, 1978).
The syntactic issue, as raised in the review by Terrace et al. (1979) ( also see Rastau and Robbins, 1982), asked whether the language produced by apes was grammatical, showing evidence of the use of structure, such as word order, as information. Or, as the question was put directly by Terrace et al. In the title of their 1979 paper, “ Can an Ape Create a Sentence?” It was acknowledged that apes occasionally generated long strings of words but that successive words added little or no information, being mainly repetitions, synonyms, or the like, without elaboration or expansion of meaning. For example, Terrace et al. Described one sequence produced by Nim, the chimp they tutored in sign language. To request an orange, Nim produced the following 16 signs: “give orange me give eat orange me eat orange give me eat orange give me you.” Nim probably did get the orange.
In other cases of apparently well-formed word sequences produced by apes additional problems were sited by Terrace et al. (1979) and in other reviews ( e.g., Bronowski and Bellugi, 1970; Fodor et al., 1974; Petitto and Seidenberg, 1979; Ristau and Robbins, 1982; Savage-Rumbaugh et al., 1980; Seidenberg and Petitto, 1979, 1981; Terrace et al. 1981). These problems included insufficiencies or deficiencies in the reporting of data; the presence of context cues to guide responding, prompts, or other paralinguistic cues from the trainers that may have led the ape into a prescribed series of signs; the probable underreporting of examples that did not conform to grammatical sequence; the practice in some cases of deleting extraneous or redundant signs in reports of the sequences produced by the apes; and the overinterpretation of a string of signs or unique combination of words produced by the apes. Thus, when the chimp Washoe signed “water bird” on seeing a swan in a lake for the first time, did she create a novel combination of words and thereby name the swan , or was she simply giving two independent ‘ WATER-----BIRD?” These criticisms cast grave doubt on the evidence for the sentence-processing abilities of apes and, by implication, on all of the claimed linguistic skills of these animals.