phonemes in affixes

[Adam Werle]

Dear Joan,

Heine & Reh 1984 is the earliest reference I’ve been able to find that
discusses “erosion” (pages 21-25), the general loss of phonological content --
segmental and contrastive -- in bound morphemes:

Bernd Heine & Mechthild Reh. 1984. Grammaticalization and Reanalysis in African
Languages. Hamburg: Helmut Buske Verlag.

page 21: "Erosion is a process by which the phonological substance of a
morpheme is reduced, usually in accordance with its new evolutionary status."
page 24: "The more grammaticalized a unit is, i.e. the more processes it has
undergone, the more susceptible it is to Erosion."

Also, Adam Ussishkin posted a summary to the Linguist List a couple of years
ago about root-affix asymmetries in segmental markedness. The summary mostly
just contains examples from different languages -- including some where
segments are not just lost but become less marked -- but few references:

http://www.linguistlist.org/issues/12/12-2430.html

Adam Werle.
nrwan at linguist.umass.edu
University of Massachusetts, Amherst

Joan Bybee wrote:

> Dear typologists,
>
> I am trying to find out who has written about the hypothesis that affixes
> tend to have fewer phonemic contrasts than lexical roots or stems. This is
> an idea that has floated by me several times, but never with a definite
> reference attached. I assume some revered typologist or historical linguist
> of yesteryear first discussed the idea, but I don't know who that was. The
> OT phonologists have picked it up recently and used it to justify a certain
> ordering of constraints. Any help will be greatly appreciated.
>
> Joan