ALIES MACLEAN & HANS BENNIS - Variation in Inflection Dutch Dialects

At first sight, there appears to be a huge amount of variation within the verbal inflectional paradigm of Dutch. In this presentation, we will propose a preliminary feature organisation to structure and explain the variation in inflection found in the Dutch dialects. We have analyzed the paradigms of the verbs leven ('to live') and gaan ('to go') of 227 Dutch dialects. These paradigms were taken from the SAND database, the Syntactic Atlas of Dutch Dialects, which is developed at the Meertens Institute. The focus in this presentation is on the syntactic features that are minimally required in the system.

There are five generalisations that can be filtered from the data:

(1)     Plural forms of the verbal paradigm have one spell out.

We assume that plural forms represent the default spell-out of  verbs and have no phi-features. In this sense, they are similar to the spell-out of infinitives (leven, 'to live') and nominalisations (het leven, 'the life'). We thus assume that [person] and [number] should not be taken as syntactic features in the case of plural subjects.

As a first approximation of the organisation of features within the Dutch inflectional paradigm, a notation is used in which we make use of [1], [2] and [3] as referring to first, second and third person respectively:

         [1]. [2], [3], [-]

A paradigm with maximum spell out (four different inflections) is represented in the notation above. This feature organisation can be reduced in two ways, either by enlarging the default, or by syncretisms of [1], [2] and [3]. We will show that these strategies for reduction predict a number of paradigms, that are indeed found in the data. In addition to (1), we will argue that the following generalizations hold:

(2)     [3] has the morphological form -t in all Duch dialects.

(3)     [1] is being maintained as a unique form in the paradigm.

(4)     [2] only has a specific spell out in dialects that have the inflectional suffix -st.

(5)    Syncretism gives rise to the classes [2/3] and [1/2/3], but not to [1/2] en [1/3]

Finally, we will propose a feature hierarchy that predicts the natural classes following from generalisation 5 in terms of the features [speaker] and [addressee].

In the next stage of this project, we hope to strengthen this system by adding data from Flemish, to refine these generalizations and to rephrase them in terms of a theoretical model which allows us to explain the observed patterns in inflectional variation. In this latter phase we will also take into account diachronic variation and the variation that is found in (first and second) language acquistion (cf. Blom, Polisenska & Weerman).