Word-specific tonal realizations in Mandarin

TL;DR

This study investigates whether Mandarin tone realization for disyllabic words with the rise-fall (RF) pattern is partially determined by token meaning. It combines generalized additive models on a Taiwan Mandarin spontaneous corpus with a Discriminative Lexicon Model and contextualized embeddings to show that word type and sense predict F0 contours beyond segmental cues, and that token-specific pitch contours can predict word meaning with notable accuracy. The results reveal a robust link between meaning and tonal realization, including learnable, token-level mappings from meaning to pitch contour and vice versa, suggesting functional use in alignment of form and meaning. These findings have implications for theories of the mental lexicon, tone-language phonetics, and practical natural-language processing where meaning-phonetics links may enhance comprehension and production models.

Abstract

The pitch contours of Mandarin two-character words are generally understood as being shaped by the underlying tones of the constituent single-character words, in interaction with articulatory constraints imposed by factors such as speech rate, co-articulation with adjacent tones, segmental make-up, and predictability. This study shows that tonal realization is also partially determined by words' meanings. We first show, on the basis of a corpus of Taiwan Mandarin spontaneous conversations, using a generalized additive regression model, and focusing on the rise-fall tone pattern, that after controlling for effects of speaker and context, word type is a stronger predictor of tonal realization than all the previously established word-form related predictors combined. Importantly, the addition of information about meaning in context improves prediction accuracy even further. We then proceed to show, using computational modeling with context-specific word embeddings, that token-specific pitch contours predict word type with 50% accuracy on held-out data, and that context-sensitive, token-specific embeddings can predict the shape of pitch contours with 40% accuracy. These accuracies, which are an order of magnitude above chance level, suggest that the relation between words' pitch contours and their meanings are sufficiently strong to be potentially functional for language users. The theoretical implications of these empirical findings are discussed.

Figures (15)

• Figure 1: Toy dataset. The left-hand panel shows the F0 contours of single tokens of six Taiwan Mandarin words with the RF tonal pattern, produced in isolation by the same speaker. The right-hand panel shows the RF contour predicted by the a simple GAM, using a thin plate regression spline smooth for normalized time as predictor.
• Figure 2: The left-hand panel shows by-word adjustment contours from the toy model with only by-word factor smooth and normalized time as predictors. The right-hand panel plots the fitted contour for each word, with the predicted general contour (identical for all words) indicated by the dashed line.
• Figure 3: Partial effects in the baseline GAM. The upper left-hand panel shows the predicted base contours for speakers identified as female (red) and speakers identified as male (blue). The next four panels show, for female speakers, how the base contour is modulated by duration, utterance position, previous bigram probability, and following bigram probability, respectively. The final panel presents, again for female speakers, the effect of tonal coarticulation with the tone of the preceding word, when the following word has a high-level tone.
• Figure 4: The left-hand panel shows model fit improvement gauged by decrease in AIC units when a predictor (or set of predictors) is added to the baseline model for the word-type analysis. The right-hand panel shows the concurvity score of individual predictors in two models using the full dataset of 3,778 tokens: the omnibus-segment model (blue) with factor smooths for all segment-related control variables added to the baseline, and the word model (red) with only a factor smooth for word added to the baseline.
• Figure 5: Examples of the pitch contours predicted by the general smooth for time for female speakers, combined with the partial effects of the factor smooth for word. These partial effects do not include the general intercept, nor the differences in pitch between female and male speakers. As they represent the pure effect of word on the pitch contour, irrespective of other predictors, the curves are centered around the y-axis (indicated by a horizontal dotted line). The vertical dotted lines in the panels indicate the average (word-specific) syllable boundary.