3

A while ago I remember reading a blog post about the (in)famous dw > (e)rk sound change in Armenian. This post discussed it in terms of feature metathesis, i.e. the stop feature and the continuant feature swapping places. It drew comparisons to examples in many Iranian languages (which have been in close contact with Armenian since before its earliest attestation).

Unfortunately, since then I have forgotten what these examples were, and unable to locate this blog post.

The closest I've been able to find is Jessica DeLisi's 2013 paper *Feature metathesis and the change of PIE dw- to Classical Armenian -rk, which discusses Iranian parallels, but ultimately settles on a phonotactic requirement that in CC clusters the first element must be more sonorous, rather than (unmotivated) feature metathesis.

Does anyone know of any examples of fairly uncontroversial feature metathesis in consonant clusters (especially in the Iranian languages), especially those that leave place unchanged?

7
  • related, but specifically for place, and not so focussed on Iranian: linguistics.stackexchange.com/questions/15227/…
    – Tristan
    Commented Mar 7, 2022 at 10:25
  • Since actual feature metathesis doesn't exist, synchronically or diachronically, in Iranian or elsewhere, in consonant clusters or other contexts, finding an example of a subset of that non-existent set is even harder. A problem with your question: you don't explain what "feature metathesis" means, especially given the existence of privative features, and the distinction between "metathesis" and "shift". A feature can move across an unspecified position.
    – user6726
    Commented Mar 7, 2022 at 16:22
  • I'd have thought the term would be self explanatory: the metathesis of features, rather than entire segments. Regardless, the term is certainly in use in the literature, see the cited paper (which whilst uploaded to academi.edu is from Diahronica)
    – Tristan
    Commented Mar 7, 2022 at 16:33
  • Are you familiar with the theory of privative features? Metathesis is where the order i, j changes to j, i: but that assumes that the objects in question are actually ordered. "voice" is not ordered w.r.t. "nasal" at the feature level –- which is why I ask, and why the question is not self-explanatory.
    – user6726
    Commented Mar 7, 2022 at 16:45
  • the features on a given segment need not have any order within segment. Consider a segment A followed by a segment B. Segment A has features i j k, segment B has features m n o. Feature metathesis would be if they exchange two features such that we now have segment C followed by segment D where segment C has features i j o, and segment D has features m n k. Obviously the loss of the k & o features respectively may entail the introduction of other features in order to produce a well specified phoneme. As examples: pd > bt, km > ŋp, fk > px all have one or more features metathesise
    – Tristan
    Commented Mar 7, 2022 at 17:00

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.