purely by context. For example,Fred apple(imagine this uttered by an eighteen-month-old or a signing chimp) might
express any number of connections between Fred and apples, expressible in modern language in sentences such as
That's Fred's apple, Fred is eating an apple, Fred likes apples, Take the apple from Fred, Give the apple to Fred, or evenAn apple fell
on Fred. Though still vague, then,Fred appleis far more precise than justFredorapplein isolation. Moreover, it isn't
totallyvague: itprobablywouldn'tbe used to expressFred has incorrect beliefs about the color of applesorApples frighten Fred's
sister. That is, although there are many possible connections, the pragmatics are not unlimited.
Concatenatingmore thantwosymbolsmultipliesthenumberof pragmaticpossibilities. Much depends on thesymbols
in question.Bread cheese beermight well expressI want bread, cheese, and beer.Bread cheese Fredis less obvious,Bread Fred
cheeseeven less so.
This is clearlya differentkindofcombination thanthatdiscussed intheprevious section. Phonologicalgenerativityis a
way of analyzing meaningful symbols and producing new ones in terms of a repertoire of smaller meaningless units.
The present sort of combination puts together meaningful symbols to form larger utterances whose meanings are a
function of the meanings of the constituent symbols. The two kinds of combination could have evolved
simultaneously or in either order.
This sort of combination has not been attested in the ethological literature. As mentioned above, the units of bird
songs, cetacean songs, and primate“long calls”are not meaningful on their own, and/or different combinations are
not distinctively meaningful. (As Hauser 1996 points out, however, this may be for lack of means to assess such
combinations.) On the other hand, the language-trained apes do show this capability, at least to some degree, on some
assessments.
To see if this is where apes' capability stops, it is most revealing to look at the less controlled cases, in which free
utterances were possible: the experiments with sign. Terrace (1979) claims that his chimp Nim reached this stage and
this stage only, producing large numbers of concatenated (and repeated) signs in an utterance, but without any further
organization. He claims that a careful look at the full data from the other signing experiments reveals similar results
(see also Seidenberg and Petitto 1978; Ristau and Robbins 1982; Kako 1999), though more enthusiastic researchers
have claimed greater organization.
8.7 Using linear position to signal semantic relations
Concatenating symbols opens up many opportunities for enhancing expressive power and precision. Two important
classes of innovations are orthogonal: