The role of phonology in the disambiguation of disjunctive questions

The disambiguation of similarly-worded alternative questions (altqs) and disjunctive yes-no questions (dynqs) has sparked a debate in English. The debate revolves around which prosodic feature can disambiguate them. In Arabic, little attention has been dedicated to how these two types of disjunction question are disambiguated. What adds to the complexity of the disambiguation in Arabic is that Arabic dialects, unlike English, use two disjunctive elements, equivalent to the English or, in altqs and dynqs. In order to replicate Pruitt and Roelofsen’s (2013) English perception study on Arabic, the disambiguating cues pertinent to Arabic need to be used in such a perception study. Hence, a thorough investigation of the general behaviour of disjunctive elements in the literature and in a corpus of eight Arabic dialects is run; based on this investigation, four dialects are selected for further investigation of the prosodic details of their disjunctive questions (Jordanian (JA), Egyptian (EA), Kuwaiti (KA), and Syrian (SA) Arabic) in two production studies. One is analysis of corpus production data in the four dialects, and the other is a production study dedicated to JA. The results of the two production studies indicate that both choice of disjunctive element (ʔaw vs. willa) and choice of contour (late-rise vs. rise-fall) seem to play a role in the disambiguation. So, two perception studies are run to investigate the relative role of each of the cues: one on JA, and one on all four dialects. The results reveal that the choice of contour contributes significantly to a dynq reading in all dialects, and the choice of disjunctive element contributes significantly to the disambiguation in three dialects (JA, EA, and KA). This finding shows that Arabic is ‘like English’ in employing choice of contour in the disambiguation, but it is also different from English in employing another disambiguating cue.