An Investigation of Word-Internal Syllabification in Moroccan Arabic
Amanda Rysling
New York University
ajr399@nyu
Amanda Rysling
New York University
ajr399@nyu
In Moroccan Arabic (MA), clusters of one, two, and three consonants are permitted in every position of a word. The syllabification of such clusters is a matter of great theoretical interest because of its potential to revise previous conclusions about the cross-linguistic typology of syllable structure (Dell & Elmedlaoui 2002 here). The present investigation follows previous work by Honorof and Browman (1995), Shaw, Gafos, Hoole, and Zeroual (2009) and others in attempting to discern the syllabic structure of the consonants in a cluster using experimental methods which allow one to assess the temporal organization of these phonological strings or clusters. Because previous work has thus far focused only on consonants at word edges, the patterns discovered may be interpreted as properties of word edges, and not onsets or codas themselves. This investigation is thus an important next step of inquiry in attempting to discern temporal patterns in word-medial consonant clusters. Data were collected from four MA speakers via Electromagnetic Articulometry (EMA). EMA tracks the movement of small pellets attached to articulators of interest during speech. The key articulatory events in the unfolding of each speech gesture were identified by analyzing the motion of the primary articulator for that gesture. These corresponded to the onset of movement, the peak velocities of the articulator to and away from the constriction, the target and release of the constriction, the maximal constriction of the articulation, and the offset of movement. Three measurements from each consonant cluster were taken to be representative of the qualities of that cluster: the left edge, C- center, and right edge. The left edge was identified as the gestural onset of the first consonant in a cluster, the right edge as the gestural offset of the last consonant in a cluster, and the C-center of the cluster was calculated as the mean of the mid-points of each consonant in the cluster, where mid-point for a consonant refers to the timepoint halfway between the gestural target and release of that Two hypotheses about the prospective structural nature of the onset in MA were contrasted in their predictions about patterns of temporal organization. These differed in stating either that the MA onset may be composed of a single consonant only, with clusters syllabified as [CC.CV], or that multiple consonants can participate in a complex onset, yielding the syllabification of [C.CCV] or [CCCV] (where two or all three of the prevocalic consonants join with the vowel in one syllable). These were designated the simplex onset hypothesis and the complex onset hypothesis, respectively. The simplex onset hypothesis, wherein only one consonant can participate in an onset, predicts that (1) the times from the left edge or C-center of the consonant cluster to the gestural anchors would vary as a result of the number of consonants in the cluster, and (2) the times from the right edge of the consonant cluster to the gestural anchors would not vary with the number of cluster consonants. The complex onset hypothesis predicts that (1) the times from the left or right edges of the consonant cluster to the gestural anchors should vary with the number of consonants in the cluster, and (2) the times from the C-center to the gestural anchors should not vary with the number of consonants in the cluster. One-way ANOVAs and Pearson's Correlation Coefficients were calculated to determine the significance of the patterns observed and the strength of the correlations discovered between the number of consonants in a given cluster and the timestamps of pertinent measurements. The intervals from the left edge and C-center to gestural anchors were found to increase with the number of consonants in the cluster, while the intervals from the right edge to the gestural anchors did not vary with the number of consonants in the cluster. On the basis of these, it was determined that MA follows a structural organization consistent with single-consonant onsets only. This confirms the conclusions of previous work about syllable onsets and codas for the case of word-medial consonant clusters.
