When syllabic l meets Hungarian Vowel Harmony

Abstract

In this paper, we examine an unexplored part of Hungarian loanword phonology, namely, the adaptation of words containing syllabic l, and the role that these words play in Hungarian Vowel Harmony (hereinafter: HVH). In Hungarian, there are no syllabic consonants, and word-final consonant–l clusters are almost completely absent. In a laboratory experiment, it was studied how Hungarian speakers pronounce loanwords ending with syllabic l and how they choose from suffix alternants.
Data was collected from 28 speakers. They were asked to pronounce 6 target words in syntactic contexts that enforced the nominative, superessive (SUE; -on/-ɛn/-øn) and allative (ALL;  hoz/-hɛz/-høz) forms. The target words included two surnames of German origin ([ibl̩], [ʃtroːbl̩]) and four recent loans ([duːdl̩], [ɡuːɡl̩], [lidl̩], [pazl̩]).
After the experiment, the vowel quality of the suffixes attached to the target word and the number of realised syllables were examined, thereafter, the formant structure and length of l and the inserted vowels were analysed by means of Praat and R.
Amongst all, our results were the following. Firstly, it has been showed that speakers were pronouncing the target words in two syllables. The intonation peak in yes/no questions served as evidence. Secondly, both forms with inserted vowel and syllabic l ([ɡuːɡVl] ∼ [ɡuːɡl̩]) are present. Insertion is more frequent in yes/no questions than in declaratives, in SUE forms than in ALL forms, further, in both suffixed forms than in the stem form.
Thirdly, according to the analysis, neutral stems select mainly front alternants ([ibl̩-høz], [lidl̩-øn]), while in case of back stems, both back and front suffixes occur ([pazl̩-øn], [ʃtroːbl̩-hoz]), even for a single speaker and target word ([ʃtroːbl̩-hoz], [ʃtroːbl̩-øn]). Therefore, syllabic l seems to be optionally transparent to palatal harmony.
However, the harmonic behaviour of the target words cannot be explained by the insertion of a reduced vowel. In spite of the tendencies seen earlier, only 60% of the forms with the front alternant contained a reduced vowel, and 78% of forms without a reduced vowel did select the front alternant, too. Regression methods reinforce this, as well.
Thus, in conclusion, other factors may play a role, such as the acoustic similarity of [ø] and the alveolar consonants.