Next: Final Remarks Up: Event-based Phonotactics Previous: Phonological Events and Event

German Phonotactics

German phonotactics is concerned with constraints on the permissable combinations of phonological segments found in the standard language. It is somewhat difficult to classify which syllables are considered to belong to the standard language since combinations from certain loan words have now become part of the standard language. The phonotactic description of German presented in this section is a close approximation to this, based on standard pronuncation lexica (cf. bibliography) and generalisations are made on the basis of structural similarities. No reference is made to nonstandard forms such as fast speech variants or regional variation as these are within certain limits considered to be part of the allophonotactics. Below, only the phonotactics of a nonreduced syllable is described; the reduced syllable with short vowels and schwa is a restricted version of this. These syllable types together describe phonological words.

The definition of the event-based phonotactic network is based on melodies and phonological event structures as specified in section 4. The relevant relations for the autosegmental phonotactic description are overlap and immediate precedence. The temporal relations of precedence and inclusion are required at the phonetic level for the construction of complex events and will not be discussed any further here. A event-based phonotactic network is defined in terms of:

declarations of initial and final nodes
a phonological event set
a set of generalisation devices
a collection of arc description

The declarations relevant for the event-based phonotactic net for German are the initial node ``0'' and the final node ``end''.
A phonological event is defined as an interval and a property <i:p>. Since the property is relevant to a particular phonological tier, this is marked as an event type using a subscript. The relevant phonological event set is:
{ , , , , , , ,, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ...}
Generalisations are expressed in the network by:
- underspecification: the classification of phonological segments of the language is in terms of partial information describing classes of sounds with structural similarity with respect to syllable structure
- variables: variables (represented here by upper case letters e.g. A) are used in the sense of Greek letter variables in phonology in order to abbreviate event descriptions
- pseudo-events: in order to reduce the size of the phonotactic network some generalisations over arcs have been undertaken by introducing pseudo-events which refer to one phonological event. Two adjacent arcs may refer to a single event although in the phonotactic description the interval is associated more than once. For example both the arcs from ``0'' to ``peak'' and from ``peak'' to ``end'' may describe an event of phonation type with property voiced; it may be the case that this is a single voicing event which is consistent throughout the syllable. Pseudo-events are thus constructed sub-events of a phonological event.
An arc in the phonotactic net is defined as a quintuple:
Q = < f, E, , , n >
where
- f : the name of the first node of the arc (e.g."2", "peak", "15")
- E : set of C-Events relevant for the arc (e.g. {PHONATION: voicelessMANNER: plosivePLACE: {labial, velar}. The sub/superscripts are used to indicate unique intervals in T with respect to tiers and arcs. The variable is specified together with the set it abbreviates.
- : set of overlap constraints expressing relations (e.g. {} which states an overlap constraint between an event on the MANNER tier and an event on the PLACE tier in the intervals I and I
- : set of immediate precedence constraints expressing relations (e.g. {} which states an immediate precedence constraint between events on the MANNER tier in the intervals I and I)
- n : name of next node (e.g. "7", "end")