|Title:||Coronal epenthesis and markedness|
|Comment:||to appear in Phonology|
|Abstract:||Recent work on the special properties of coronal consonants (e.g.
Paradis and Prunet 1991) has shown that coronals behave in some
ways as unmarked consonants. In this paper I will address the
claim that epenthetic segments tend to be coronal. This behavior
has been used to argue that coronals are Placeless. However,
laryngeal consonants (?,h) are also common epenthetic segments
and have also been argued to be Placeless. Obviously it is
problematic if both laryngeals and coronals need to be Placeless,
since we must be able to distinguish them by place of articulation
within a language. The empirical basis for the claim about
coronal epenthesis has also been unclear, with only a single
example commonly cited.
In Optimality Theory, it is possible to analyze
markedness phenomena without underspecification by the use of
markedness constraints: in this case, the proposal of Prince and
Smolensky 1993, Smolensky 1993 that there is a universally ranked
hierarchy *Dor, *Lab >> *Cor. I will argue first of all that by
extending this hierarchy to include the laryngeals, we can
account for their appearance as the 'unmarked' epenthetic
segment. I will then show that coronals do sometimes occur
as epenthetic segments, but that this is only the case in limited
situations, showing the classic signs of constraint conflict.
For example, in some cases of coronal epenthesis higher-ranked
constraints demand that the consonant be sonorant; therefore the
lowest-marked Place that is possible will be Coronal.