SOV original word order?

[jess tauber]

>From the standpoint of iconicity there might be something to this. The languages with the largest ideophone inventories are verb-medial and/or verb-final. I've never found the largest-sized inventories in V-initial types. This doesn't mean there can't be iconicity- Mayan languages, in their expressive verbs, as an example, or Salish and Wakashan, Austronesian, etc., where one can often break down lexical roots into smaller primitives. The issue for me is how lexicalized iconicity becomes. For the V-initial types, phonological-scale iconicity seems more heavily lexicalized, and rich. Consider Chinookan and Sahaptian with their etensive augmentative/diminutive marking on normal lexicon, but smallish ideophone sets. 

In languages with SVO order ideophones seem to be more often somewhat distinct from other parts of speech, but still quite lexical, as in Mon-Khmer and other SEA Austroasiatic. This is the case for ideophones in more analytic varieties of Niger-Congo. Many Bantu languages on the other hand show some flux between lexical and more prototypically ideophonic status.

So broadly, as one travels from V-final to V-medial to V-initial the degree of lexicalization of large iconic word-form inventories appears to increase. It would be interesting to know whether supposed (relative) immunity to historical change shifts along the way. Size of iconically (more) transparent inventories may have more to do with the degree of elaboration and/or fusion of derivational and grammatical morphological systems.

Jess Tauber
goldenratio at earthlink.net