metathesis defined?

[Dennis R. Preston]

This is not even technically right if you consider that some rhotic dialcts
(Scots) show clear evidence of metathesis.

dInIs

>Strictly speaking, I suppose that Coye is technically right about the
>relationship between syllabic [r] and [r] + schwa, even in words such as
>"bird", at least in "surface" structure. It is doubtless slightly stretching
>the term METATHESIS to cover this situation. However, it is such a small
>reach that it I am comfortable with it--as are all the standard textbooks on
>the history of English that I am familiar with--indeed they often give "bird"
>as a garden-variety example of metathesis.
>
>In a message dated 2/3/2000 8:07:26 AM, aaron at LING.ED.AC.UK writes:
>
><< On Wed, 2 Feb 2000 Dfcoye at AOL.COM wrote:
>
>}Also metathesis does not really apply when
>}we're dealing with the hooked schwa in a word like 'performance' -
>}'preformance' in rhotic speech-- the first vowel is not schwa plus /r/.
>
>
>Isn't it?  Phonetically, it is clearly one sound, hence hooked schwa or
>alternatively syllabic r.  But is the second syllable of 'performance'
>also just one sound phonetically, a rhoticized O? At least it is for me,
>and I suspect for many other rhotic north Americans.  Yet, there would be
>no argument in transcribing the sequence as [Or] two sounds- or maybe even
>O plus hooked schwa. So why couldn't the first syllable be schwa+r and
>coalesce on the surface.  Or would you argue that hooked schwa is a single
>sound throughout the whole of the phonology?  Positing 'er' as two sounds
>somewhere certainly allows for the metathesis in the 'per/pre' prefixes,
>as well as in 'modren' and 'southren'. >>