much generative linguistics, well-formedness constraints would underdetermine the
options for contour realizations. A different goal, and one congenial with a musical per-
spective, is to ask whether a grammatical contour is normative or not. One need then
derive only normative contours from stress grids, letting the other grammatical contours
be seen as nuanced variants.
To pursue this question, let us initially restrict contour features to a binary classification
of high and low, and apply it to the two most stressed syllables in an intonational phrase.
Contour theory in linguistics has represented this classification by ‘H’and ‘L’, with accom-
panying diacritics to indicate hierarchical function.12,13However, it is in the spirit of the
approach taken here to employ musical notation, which after all is designed to show pitch
height. As is common in musical reduction theory, hierarchy can be represented by note
values that stand not for durations but for relative structural importance. In Figure 27.11a
the first syllable is high and the second low, with the half-note first syllable acting as super-
ordinate to the quarter-note second syllable. The high–low pattern continues in Figure
27.11b but with a superordinate low syllable. Figure 27.11c and d reverses the high–low
pattern. These four cases exhaust the possibilities. The prototypical declarative contour is
that in Figure 27.11a and b, a rise and fall whose musical analogue is the tensing–relaxing
pattern of a standard tonal phrase. The difference between the two is whether the high or
the low syllable dominates. In Figure 27.11c and d, the high syllable is to the right of the
low one; this is the typical shape of an interrogative.
The claim that a higher pitch causes stress would seem to disqualify Figure 27.11b and d
as normative, since in these cases the low syllable dominates. As a consequence of nuclear
stress, however, a superordinate low syllable can indeed function in the position of the last
clitic host of an intonational phrase. In other words, this syllable’s location at the close of a
major phrase boundary compensates for its lower pitch. Hence Figure 27.11b is normative.
By the same token, Figure 27.11d is not normative, because in this case the dominating low
syllable is not in a location to benefit from major nuclear stress. As there are no interrogat-
ives in our poem, we can also ignore Figure 27.11c. This leaves Figure 27.11a and b as the
remaining options. This very restricted repertory will provide the framework for mapping
stress importance onto contour peaks and valleys.
In this presentation we take for granted but do not notate that there is sliding from one
pitch level to another (in musical terms, there is a glissando from one pitch to another). We
also posit that a given syllable has a single pitch level, ignoring cases in which a syllable
slides between two pitch levels, especially at a phrase boundary. For example, in the utter-
ance ‘He has gone!’,‘gone’normally begins high because it is stressed but ends low in order
to convey its declarative status. In ‘He has gone?’,however,‘gone’begins low but ends high
in order to convey its interrogative status. Finally, we bypass the possibility that an extended
sequence of highs and lows can gradually ramp up or (more commonly) down, so that the
421Figure 27.11Four normative contour frameworks.
ABC D