Criticism: The Nim Chimpsky project (Terrace et al., 1979)

• Terrace et al. trained Nim Chimpsky to use ASL with training methods similar to those of Gardner & Gardner, although Nim was not raised like a human child.

• Terrace et al. found some evidence for syntactic regularity in Nim's signing.

2-sign combination	Types	Tokens
more X	47	974
X more	27	124
give X	51	271
X give	24	77

• However, Nim's signing differed from childrens' speech in that, in using three and four-sign combinations, Nim rarely added any new information:

• Most-frequent sign combinations:

2 signs	#	3 signs	#	4 signs	#
play me	375	play me Nim	81	eat drink eat drink	15
me Nim	328	eat me Nim	48	eat Nim eat Nim	7
tickle me	316	eat Nim eat	46	banana Nim banana Nim	5
eat Nim	302	tickle me Nim	44	drink Nim drink Nim	5
more eat	287	grape eat Nim	37	banana eat me Nim	4

Other problems:

• Nim's MLU (mean length of utterance) did not increase with age, as a child's would.

• Semantic limitations: 99% of beneficiaries in Nim's utterances were 'Nim' and 'me'

• Also, 76% of agents in agent-object combos were 'you'

• Nim did not take turns in the conversation the way children do

• 39% of Nim's utterances were imitations or reductions of what had just been signed to him; only 18% of a comparable child's utterances were imitations or reductions.

• Nim expanded on what had just been signed only 7% of the time; children do so 21% of the time.

• A full 71% of Nim's utterances were interruptions.

• Terrace et al. claimed that previous researchers had not convincingly shown that chimps' "linguistic" abilities couldn't be explained with reference to simpler, non-linguistic processes. For instance,

"The function of the symbols of an ape's vocabulary appears to be not so much to identify things or to convey information as it is to satisfy a demand that it use that symbol in order to obtain some reward."

(There were no "tacts"--there were only "mands".)