More on language and genome

[Brian MacWhinney]

Jess,

      I agree that there are strong parallels between the genetic and linguistic codes.  In an article in "Connection Science" from 2005, I used the example of the four structural levels in protein-folding as an illustration of the emergence of properties in biological systems.  Like you, I view language as an emergent biological system.  I also understand and agree with the parallels you are drawing between viral diffusion in eukaryotic DNA and the diffusion of sound patterns in language communities.  However, there is a certain semiotic slippage in some of the attempted parallels.  In particular, the effects of a given virus on gene expression seem much more accidental than the parallel effects in linguistic memes.  As you note, there is no known way to predict quaternary structure in proteins.  To the extent that a given virus would embed itself in the regulatory "junk" of the DNA, its effects on protein folding would not be systematic or intentional, but rather just fortuitous.   In this way, viruses would be unlike ideophones or other mimetic patterns in human language.  Mimetic patterns have meaning, at least initially, whereas the only meaning of a virus is self-replication.  By the way, I am assuming here that the full range of linguistic mimetic patterns is much wider than just segmental ideophones, and that it includes tone patterns, vowel shifts, communicative markers, trendy lexical items, and higher-level constructions.
   You might respond that the initial transparency of mimetic patterns eventually gets lost in human communication systems.  I would agree with that.  But then I would bump into my second problem with your analysis, which is that you are relying on "feature geometry" to constrain the effects of ideophones.  Probably you are also relying on something like morphosyntactic geometry too, since you suggest that mimetic expression is clearer in analytic languages.  But, this seems to me to be too narrow.  If memes really operate in an isomorphic way to viruses, then we would expect interactions throughout the linguistic system, just as viruses have their effects through all the levels of protein-folding.   The diffusion of a particular linguistic meme across language communities and through the whole of linguistic structure should be constrained by many forces apart from simple feature geometry.  For example, a communicator form such as a new intonational pattern on the old communicator "well" is going to be constrained by intonational structures from other lexical items, particularly other communicators, specific conversational patterns within which it is used, specific sentential constructions to which it attaches, and so on.  
    I don't think readers of this list fail to appreciate the interactive quality of linguistic expression.  So, I think it makes sense to push the viral analysis beyond just a defense of the role of ideophones.  And I also think it important to recognize the extent to which mimetic patterns have meaning initially, although this meaning can eventually be lost as the viral memes become incorporated into the DNA of language.

-- Brian MacWhinney