TY - JOUR
T1 - Contextual phonological errors and omission of obligatory liaison as a window into a reduced span of phonological encoding
AU - Lange, Violaine Michel
AU - Pellet Cheneval, Pauline
AU - Python, Grégoire
AU - Laganaro, Marina
PY - 2017/2/1
Y1 - 2017/2/1
N2 - Background: The question of how much speakers plan ahead before they start articulating their message is essential to understand how fluency is ensured during speech production. This question has been largely investigated in healthy speakers. Surprisingly, this remains unexplored for brain-damaged speakers, even though a reduced span of encoding might account for the fact that those impaired speakers often produce scattered speech. Aims: In this study, we examine whether the span of encoding is reduced in some left hemisphere brain-damaged speakers by taking advantage of two linguistic phenomena which provide an insight into ahead phonological planning. Methods and procedures: First, we elicit the production of French sequences involving obligatory liaisons (e.g., /mõ/ and /ami/ produced /mõ.nami/), for which the correct production requires ahead planning, at least up to the first phoneme of the following word of the utterance. Second, we use tongue-twister-like sequences in order to elicit contextual phonological errors, as phoneme anticipation errors (e.g., bureau vert—green desk—produced /vyRo.vER/) suggest that the speaker has planned ahead before articulating. If brain-damaged speakers do present a reduced span of encoding, they should both produce a high rate of liaison consonant omissions and a low rate of anticipation contextual phonological errors. Outcome and results: The results on a group of 13 speakers with aphasia and/or apraxia of speech overall show few contextual (syntagmatic) errors despite a high rate of segmental errors, whereas the majority of phonological errors produced on the same utterances by healthy speakers were syntagmatic. The speech/language impaired participants also presented a high rate of obligatory liaison consonants omission. Crucially, a negative correlation was observed between the rate of phoneme anticipation errors and the rate of liaison consonant omission. Conclusion: These results suggest that some brain-damaged speakers present a span of phonological encoding limited to single words and that the use of inter-word sandhi phenomena, such as French liaison and the analysis of phoneme anticipation errors, are valid linguistic tools to inform on the span of encoding.
AB - Background: The question of how much speakers plan ahead before they start articulating their message is essential to understand how fluency is ensured during speech production. This question has been largely investigated in healthy speakers. Surprisingly, this remains unexplored for brain-damaged speakers, even though a reduced span of encoding might account for the fact that those impaired speakers often produce scattered speech. Aims: In this study, we examine whether the span of encoding is reduced in some left hemisphere brain-damaged speakers by taking advantage of two linguistic phenomena which provide an insight into ahead phonological planning. Methods and procedures: First, we elicit the production of French sequences involving obligatory liaisons (e.g., /mõ/ and /ami/ produced /mõ.nami/), for which the correct production requires ahead planning, at least up to the first phoneme of the following word of the utterance. Second, we use tongue-twister-like sequences in order to elicit contextual phonological errors, as phoneme anticipation errors (e.g., bureau vert—green desk—produced /vyRo.vER/) suggest that the speaker has planned ahead before articulating. If brain-damaged speakers do present a reduced span of encoding, they should both produce a high rate of liaison consonant omissions and a low rate of anticipation contextual phonological errors. Outcome and results: The results on a group of 13 speakers with aphasia and/or apraxia of speech overall show few contextual (syntagmatic) errors despite a high rate of segmental errors, whereas the majority of phonological errors produced on the same utterances by healthy speakers were syntagmatic. The speech/language impaired participants also presented a high rate of obligatory liaison consonants omission. Crucially, a negative correlation was observed between the rate of phoneme anticipation errors and the rate of liaison consonant omission. Conclusion: These results suggest that some brain-damaged speakers present a span of phonological encoding limited to single words and that the use of inter-word sandhi phenomena, such as French liaison and the analysis of phoneme anticipation errors, are valid linguistic tools to inform on the span of encoding.
U2 - 10.1080/02687038.2016.1176121
DO - 10.1080/02687038.2016.1176121
M3 - Journal article
SN - 0268-7038
VL - 31
SP - 201
EP - 220
JO - Aphasiology
JF - Aphasiology
IS - 2
ER -