Investigating the relationship between production and perception of epenthetic glottal stops in the case of Maltese

Abstract

The Maltese language, like many other languages, marks vowel-initial words with a glottal stop. In the case of Maltese, this epenthetic segment occurs separately as a separate phoneme, with the potential for ambiguity due to words becoming homophonous. In a series of two experiments, we tried to investigate whether the underlying functions of such epenthetic glottal stop is constrained by matters of audience design. Our first experiment laid important grounds upon which we were to interpret the findings from the second experiment, because it provided us with descriptive evidence that Maltese listeners do not possess an a-priori preference to words with or without an underlying glottal stop, hence implying that the result from the production experiment occurred was not constrained by a baseline preference for one of the forms. This was done by a simple procedure, where participants had to rate the auditory stimuli they were exposed to, on a 7-point scale. The stimuli contained an even frequency of words with and without glottal stops. Our second experiment found a stark difference between the two main conditions, accented and unaccented. This was done by having participants participating in a simple question-answer experiment. Participants had to answer written questions by reading another written answer, placing strategic emphasis as indicated in the trials. More glottal stop productions occurred in the accented condition than in the unaccented condition. Furthermore, as opposed to listener-oriented paradigms, we found more glottal stop productions in the phonological contrast of the accented condition than in the lexical, suggesting that speakers do not modulate their efforts in favor of the listener when producing the epenthetic glottal stop.