The Roles of Tonal and Segmental Information in Mandarin Spoken Word Recognition: an Eyetracking Study
Assignment 1 – Article Summary In their article, “the roles of tonal and segmental information in Mandarin spoken word recognition: an eyetracking study”, Malins and Joanisse investigate how suprasegmental features affect on-line auditory word perception. Current speech perception models have fallen short in accounting for suprasegmental features because all have been based on Indo-European languages. In this study, Malins and Joanisse extend the potential of suprasegmental features by examining the role of tone in auditory word recognition.
Preliminary studies on which Malins and Joanisse based their study show that features of stress alone could not account for word constraint activation. One of the reasons that stress was a poor candidate for word constraint during lexical priming tasks was that they were situated within polysyllabic structures.
Tone features, on the other hand, are situated within monosyllabic structures. Listeners of tonal languages are required to make distinctions based on these features alone. In addition, these preliminary studies showed that tonal information was processed after segmental information.
However, these propositions were brought into question as Liu and Samuel (2007) discovered that most of the preliminary studies in speech perception did not utilize contextual constraints. More recent studies seem to show that tonal features are more effective in constraining word activation when in a sentential context and segmental features more effective in word constraint in a lack of contextual information.
Malins and Joanisse set out to examine the role of tonal features in Mandarin word recognition by employing the visual world paradigm.
This would allow the use of eye tracking which would measure the on-line auditory processing among four competitor categories – segmental, cohort, rhyme and tonal. The segmental category shared all phonemic features except for tones, the cohort shared all but word final phonemes, the rhyme differed in onset consonant, and the tonal shared only tone. The following was hypothesized: tonal and segmental information would be accessed at very similar points in time and share equal potential for word recognition constraint.
Additionally, rhyme competitors would be viewed more than unrelated items and that tonal categories would produce little interference effect with target items.
Twenty-four native Mandarin speakers were chosen as participants in this study. The mean age was 28, although seven were excluded from the study due to excessive use of peripheral vision (lack of eye movement data). Experimental stimuli comprised of 27 monosyllabic Mandarin words. The unrelated distractor tones were comprised of phonologically unrelated segments.
A picture array was shown on screen for 1.
5s followed by a . 5 second transition at which point the central fixation point appeared. Participants focused on the fixation point while selecting the correct target picture for the spoken utterance. When the auditory stimulus was provided, the time course was measured through observing eye movements. If there were significant interference between tone and segment, then processing times would be increased therefore suggesting comparable roles in constraining word activation.
The pictures in the array provided conceptual constraint. This ensured that sub-lexical access was being observed. This visual word picture matching task was used to create an environment that demanded lexical access using conceptual constraint. This method would supplement the hypothesis by leading participants to utilize tonal information. In the tasks, the act of pressing a button demanded lexical access and the four competitor pictures provided the conceptual constraint.
There were 280 experimental trials.
The experimental setting was kept monolingual as to eliminate the potential for confounding effects of bilingual interference. The results of this study show that the competition between target words and competitors was greatest for the segmental and cohort categories. This is crucial because the segmental category diverged from the cohort in tonal information only and the cohort category diverged in segmental information. This supports the hypothesis that both tonal and segmental information show an equal role in word recognition.
Furthermore, the time course information for both cohort and segmental categories were similar. This seemed to indicate that segmental and tonal information are both accessed at the same time. There were no data supporting the claim that rhyme competitors would be looked at more than the distractors. This went against the second hypothesis and was likely a result of a discontinuity between rhyming features in Mandarin compared to English. This study could help improve existing models of speech perception by providing a firmer basis for tonal processing in auditory word recognition.
This might raise the question of how tonal information should be represented in a speech perception model. Ye and Connine (1999) have proposed the use of ‘tonemes’ since they act as contrastive units that can be categorically perceived like their segmental counterparts. Through summarizing, I have come to a better understanding of how this study would influence the TRACE model. Particularly, how different levels of word representations are shaped by spreading activation — tonemes possibly being the newest.
By accounting for suprasegmental features, a speech perception model would be much more robust and functional. One particular thought that has arisen as a result of this study is how tone recognition contrasts between the modalities of speech vs.
music. Do they share overlapping neurological locations and functions? Also of interest is whether individuals with amusia (tone deafness) are able to utilize tones as contrastive units in a language? References Liu, S. , & Samuel, A. G. (2007).
The role of Mandarin lexical tones in lexical access under different contextual conditions.
Language and Cognitive Processes, 22(4), 566–594. Malins, G. M. & Joanisse, M. F. (2010).
The roles of tonal and segmental information in Mandarin spoken word recognition: an eyetracking study. Journal of Memory and Language, 62(4), 407-420. Ye, Y. , & Connine, C. M. (1999).
Processing spoken Chinese: The role of tone information. Language and Cognitive Processes Special Issue: Processing East Asian Languages, 14(5–6), 609–630.