An application of autosegmental morphology to some nonconcatenative phenomena in germanic languages

This study is an attempt to apply the theory of autosegmental morphology, as proposed by McCarthy (1979, 1981), to some nonconcatenative phenomena in three languages with otherwise concatenative morphological systems, namely Dutch, German and English.

The main features of the theory of autosegmental phonology, which provides the basis for McCarthy's theory, are described. The literature on autosegmental studies of tone and harmony is reviewed, with particular attention paid to the Well-Formedness Condition, Association Conventions, and the Obligatory Contour Principle. The theory of autosegmental morphology is then introduced, and an application of autosegmental morphology to reduplication, as put forward by Marantz (1982), is detailed.

After a brief discussion of some major differences between the morphological systems of Dutch, German and English on the one hand, and the languages to which autosegmental morphology has so far mainly been applied on the other, attempts are made to apply the theory of autosegmental morphology to the Dutch and German past participle morphemes, a circumfix in Dutch, and reduplication in English and German.

It is found that neither the past participle morphemes nor the Dutch circumfix benefit from an autosegmental treatment. Rather more success results from the reduplication data, and some interesting theoretical issues are discussed in the light of this data.

It is concluded that autosegmental morphology is of use only for those languages, or those parts of languages, which manipulate the skeletal tier. It is further concluded that as the morphologies of Dutch, German and English overwhelmingly do not manipulate the skeletal tier, and as the description of the entire morphologies of these languages autosegmentally would moreover result in some disadvantages, this should not be attempted. Doubt is cast on any implied universality for the theory of autosegmental morphology.