criticisms of grammaticalization

jess tauber phonosemantics at
Sun Feb 26 16:12:18 UTC 2006

As one might also say of DNA, great thread!

'Mutation' is a pretty squishy notion- all the really interesting (from the evolutionary standpoint) alterations happen at higher levels than classical point mutations in DNA- involving rearrangements of blocks of materials, sometimes coding for protein or intervening segments which are then edited out (introns and exons), regulatory elements which direct replication or transcription, re-emplacements of entire sections of chromosomes (with same or inverted orientation), or even splits and fusions of entire chromosomes, not to mention deletions or multiplications at every level (including entire genomes).

Most point mutations are either rather neutral in effect, or relatively deleterious (even fatal) (so black and white), but these higher level rearrangements often produce viable results due to backup plans (systemic degeneracy?) that can pick up slack (grays). The bigger the system chunk, the more plastic its responses, in many cases.

Years ago I pointed to higher level parallelisms between genes and language on LINGUIST- flexibly edited split genes in eukaryotes versus analytical/isolating type, appositional bacterial genes resembling agglutinating languages, and rigidly maintained overlapping genes in smaller viruses as against certain polysynthetic systems.

In the analytical genetic/linguistic type one sees a lot of external modification of the stored information depending on context (thus pragmatically organized). Duplications and deletions are used extensively. Lots of many to one and one to many mappings for gene parts. Combinatoric heaven. Eukaryotes are the result of organismal fusions, so a little like cosmopolitan empires. The regulatory/transcriptional machinery has to be generalized to deal with all comers and all situations. Somewhat like similar linguistic environments. The 'ecology' becomes internalized.

At the other end of the continuum, the overlapping/polysynthetic genetic/linguistic type seems more characterized by overall system architectural loss, increased dependency on fuller systems elsewhere, and idiosyncratic (i.e. specialized) behavioral rigidification (given less material to play with, in more automatic and fixed combinations). This is the inverse of the analytical 'type'

Just as linguistic systems can evolve between types, so can genetic ones. Many parasitic organisms have been demonstrated to have simplified genetically, which is alright given that they just let their host carry out functions the lost genes would otherwise cover. The process can become extreme, even nearly total, but by that time such 'parasites' have become nearly completely integrated with their hosts (such as in organelles, or certain viral elements now known to aid in element movement within the genome).

Thus we have the makings of a genetic cycle similar to the linguistic one (which is usually thought of out of the social/historical/areal context).

Grammaticalization/grammaticization (the geneticists have their own version of this- eucaryote/eukaryote) should probably be seen as part of this bigger overall picture in a complex systems perspective- opposing poles of behavioral flexibility versus rigidity (or automatization) reflecting environmental flux as against predictability needing to be responded to.

Given creeping loss of lexical resources as one moves towards polysynthesis (Michael Fortescue has suggested that the root inventory is reduced), one might expect a certain narrowing of possibilities for grammaticalization along the line (and perhaps longer staying power for existing grams as one 'fills in the dots'). Reminds one a little also of tissue specialization in multicellular organisms. Fewer degrees of variational freedom.

Anyway, just a couple of slightly off-focus thoughts for a very interesting discussion.

Jess Tauber

More information about the Funknet mailing list