ICPhS XVII Regular Session Hong Kong, 17-21 August 2011 2062 ALIGNMENT OF THE “EARLY” HL SEQUENCE IN MALTESE FALLING TUNE WH-QUESTIONS Alexandra Vella University of Cologne, Germany; University of Malta, Malta alexandra.vella@um.edu.mt ABSTRACT This study investigates two factors related to the falling tune found in wh-questions in Maltese. It examines the alignment, of the H(igh) and L(ow) tone targets found in the early part of this tune in an attempt at providing a more definitive phonological analysis of the tune. Carefully controlled laboratory speech is used in an attempt at quantifying the alignment details. The study also addresses the issue of the accentability or otherwise of wh- or Q(uestion)words in Maltese. Evidence is provided for alignment of the H tone target on or around the beginning of the Qword rather than at the accented syllable for Qwords having a stressed syllable removed by at least one syllable from the left edge of the phrase. Results are as yet unclear for monosyllabic or initially-stressed Qwords, different strategies for dealing with tonal crowding possibly being at work. The analysis suggests that the Qword in Maltese does in fact receive an accent. Keywords: intonational phonology, tonal alignment, wh-questions, accenting of Q(uestion)words, Maltese 1. INTRODUCTION This study aims to further analysis of the intonation of Maltese interrogative word questions (henceforth referred to, for ease, as “wh-q(uestion)s”). Earlier analysis of the intonation of such questions reported in [22] was inconclusive on two main counts. First, the analysis raised various questions about the realisation, particularly the alignment and scaling details, of the intonational form – a form which may also extend to other structures and contexts – of this type of question in Maltese. The primary aim of this study is therefore to use carefully controlled laboratory speech to re-examine the phonetic implementation details of the wh-q falling tune in Maltese. The study aligns with work in the A(utosegmental-)M(etrical) theory of intonational phonology, e.g. [10, 12, 17]. An interpretation of the implementation details for this tune in Maltese should allow for a more definitive analysis than that available to date. Second, it was not clear whether the Q(uestion) word in such questions is accented or not in Maltese. The issue of the accentedness or otherwise of the Qword in Maltese, at least in the highly constrained kind of data analysed, is also addressed. 2. BACKGROUND 2.1. Wh-questions and intonation A strong tendency is reported for the global shape of wh-qs to be a (rise-)fall rather than a rise. A rising overall movement, as in other types of questions, has however also been observed, e.g. Bartels [2] p.172, suggests a rise as the canonical form for English wh- qs which involve a request for “further specification of an element that occurred in a prior utterance”. A number of studies, e.g. [8], examine the issue of whether the choice of the less commonly used rising intonation in such qs correlates with a difference in function as compared to that in their falling intonation counterparts. The possibility of wh-qs with rising intonation has been attested for Maltese, [4], as for other langauges. However, the focus of the present study is on wh-qs involving a fall. Experimental work has shown that information on phonetic details such as those relating to tonal alignment can be used to inform a phonological analysis of the tune in question. E.g. [13] show that alignment is relevant to the distinction between two rising pitch accents in Dutch. Similarly, [1] provide an analysis of the alignment and scaling of the intonational rise-fall in wh-qs in Greek. The analysis shows that the position and number of syllables before the stressed syllable crucially determine the tonal alignment and scaling details. The present paper forms part of a study of wh-qs in Maltese whose broader aim is that of providing an analysis of the alignment and scaling details of the tune which, although falling, is distinct from the falling tune found in declaratives. This paper reports on the part of the pilot study which deals with alignment. The picture which has begun to emerge on the matter of the accentedness or otherwise of wh-words suggests that languages differ in this respect. Some languages, notably English, e.g. [12, 19], accent a word other than the wh-word; others accent the wh- word itself. Amongst the latter are Greek [1, 15], Hungarian [12], Korean and Japanese, cited in [7], and Tamil [11]. Ladd [12] suggests a greater