The architecture at this point thus looks like Fig. 8.3, in which the new generative syste mof phonology has been
interposed into the mapping between meaning and the auditory/motor levels. This is a parallel generative system
without a level of syntax.
The interface between phonology and meaning includes on one hand lexical items and on the other hand principles
that map phonological linear order to semantic relations. At the same time, the older interfaces straight from meaning
totheauditory/motorsystems need nothave goneaway. Infacttheyare stillused inmodern language intoneofvoice
and perhaps (modulated by phonology) in onomatopoeia.^126
Afinalnote:Bickertoninsists that protolanguageisnotlanguage, whileKlein and PerdueclaimthatBVislanguage. Yet
the two phenomena are incredibly similar. How do they arrive at these opposed positions? Bickerton wants to stress
the difference between protolanguage and modern language (in particular Creoles), and hence wants to distance
protogrammar fromUniversalGrammar,theessentialpartofmodernlanguage. Kleinand Perdue,incontrast, wantto
stress the role of UG in late second-language acquisition, so they want UG to be involved in BV. Therefore, they
tentatively assert that BV represents the default settings of all the parameters in a Principles-and-Parameters type of
Universal Grammar. But, as Bierwisch (1997) and Meisel (1997) point out, this either attributes too much
sophistication to BV or not enough to UG.
I suggest that Bickerton and Klein and Perdue are each forced into their position because they assume a discrete
“grammar box”witha syntactocentricarchitecture. UG for bothofthemissyntax, and you either haveit or youdon't.
Moreover, they forget that phonology is part of UG too. The present approach allows us to make the appropriate
compromise: we can say that protolanguage and BV both havepartof UG—and approximately the same part.
At the same time, protolanguage/BV is still a long way from the expressive possibilitiesof modern language. We now
progress through some further steps.
Fig. 8.3 Architecture of protolanguage/Basic Variety
AN EVOLUTIONARY PERSPECTIVE 251
(^126) The standard lineon onomatopoeia is that it is illusory. After all, dogs gobow-wowin Englishbutgnaf-gnafin French, and roosters gocockadoodledoo in Englishbutkikiriki
in German. Still, there is some degree of sound symbolism here. I doubt there's a language where dogs gokikiriki and roosters gothud.