Seminar from Jane Wottawa (LIUM)


Date: 15/03/2024
Time: 10h15
Localization: IC2, Boardroom
Speaker: Jane Wottawa

To [h] or not to [h]? L2 production of /h/ in semi-spontaneous speech by French leaners of English and of German.


The present study investigates the production of /h/ or lack thereof in semi- spontaneous L2 speech amongst two distinct groups of French native speakers learning either English (FE) or German (FG). Second languages that differ in phonotactics and prosodic organization from the first language (L1) present a challenge to the articulatory and perceptual habits of the learners (Flege, 1995, Weber & Cutler, 2006). Both English and German have the voiceless fricative /h/ in their inventory, unlike modern French, whose is only present orthographically. It is therefore considered as a “new” phone using Flege’s (1995) terminology.


Previous studies have shown that French learners of English and of German often drop the /h/ or resort to h-epenthesis in hiatuses or vowel-initial words (English: Kamiyama et al. 2011, Exare 2021, Capliez forthcoming; German: Quiehl 1906, Wottawa 2017). John and Cardoso (2009) remark that h-deletion is a “stigmatized” feature of francophone speech in English. As a result, “francophones devote considerable effort to overcoming the problem” (2009, p. 120). Quiehl’s early twentieth-century book on the pronunciation of French accented German also comments on French learner’s difficulty in pronouncing the /h/ in German due to the unpronounced in their L1 (Quiehl 1906, pp. 100-101).

While German rarely has instances of h-deletion in grammatical words, English speakers often reduce the auxiliary have to /ǝv/ and pronouns his, her, him to [ɪs], [ǝ], [ɪm]. This may impact French learners’ pronunciation and their underlying representations of /h/ in English vs. German. In addition, John and Cardoso (2009) for English L2 and Wottawa (2018) for German L2, both indicate that task has an impact on the pronunciation of /h/ amongst learners.


This study draws on Wottawa’s methodology (2018) using a picture description task by first-year university French learners of German (n=15) to measure proficiency in pronunciation and extends it to learners of English (n=12). Picture description is language independent, so the same picture was used for both learner groups. The key elements in the picture elicited words containing word-initial /h/ in both languages, e.g hopscotch in English and Himmel un Hölle in German. Several studies selected reading tasks to measure the pronunciation of /h/ in L2 speech (Kamiyama et al. 2011, Capliez forthcoming), while others used both reading and informal interviews (John & Cardoso). Fortkamp (2000) and Sample & Michel (2014) use picture description tasks to measure fluency but to our knowledge, little has been made to compare segmental pronunciation difficulties in two L2s in semi-spontaneous speech.


Preliminary results of our study show that FE and FG produce about 75% of word-initial /h/ as [h]. A more careful examination of the production indicates that some learners tend to avoid choosing words containing a word-intial [h]. The above-mentioned high percentage may be misleading: the degree of lexical diversity and the strategic selection of words not containing /h/ to describe the picture must also be factored in. The study investigates whether this avoidance is linked to low vocabulary size and thus low L2 proficiency (Ovtcharov et al., 2006) or if it is an individual choice of avoiding segments that are linked to production difficulties, being the result of potential trade-off effects on the working memory (Skehan 2009, Bygate 2009).