普通话连续话语中轻重音的层级性和韵律表征

The Hierarchical Nature and Prosodic Features of Stress in Continuous Speech in Mandarin

本文讨论普通话连续话语中轻重音的韵律表征和以韵律结构为基础的重音的层级性,根据声调实现域的确定,提出区分带轻声的和不带轻声的两类韵律词,分别确定各自的词重音位置。在连续话语中,轻声和轻读出现的频率远远高于词典中标注的轻声,说明前重后轻的重音格式在口语中是非常活跃的韵律模式。不带轻声的韵律词的重音位置受到韵律和结构因素以及上下文语境的影响。通过分析朗读语料,我们发现,连续话语中实词获得重音的概率一般要高于虚词,发音人在大多数情况下倾向使用前重的格式。

The paper is primarily concerned with the prosodic hierarchy of stress and prosodic features of strong and weak stress,as manifested in continuous speech in Mandarin.Since Mandarin used tones to distinguish lexical meanings,it is necessary to determine how tone and stress interact at some level of the prosodic hierarchy,especially given that they share similar acoustic cues such as pitch and duration variations on the syllable that carries a tone or stress.We establish that a tonal domain is also a stress domain in Mandarin,in line with Duanmu’s(1995)analysis of Shanghai and Southern Fujian dialects.We distinguish,prosodically,two types of words in Mandarin.The stress pattern in the first type is S-W pattern,where the first syllable is in one of the four lexical tones(S)and the second in the neutral tone(W).The stress pattern in the second type is S-S,where both syllables are said in one of the four tones.Between the two strong syllables,which one has the greater stress has been the focus of debate.The S-W pattern is first motivated by minimal pairs of disyllabic words distinguished by the second syllable in examples such as“m☒i mài”(buying and selling)and“m☒i mai”(business).Although such pairs are small in number,the S-W pattern is widely adopted at the phrasal level.For example,the directional complement after the verb is always in neutral tone.The pattern is even more extensively attested in continuous speech.According to Li(2021),the percentage of syllables in neutral tone is as high as 16%–18%in recorded scripted and unscripted conversations.The percentage is much higher than the percentage of neutral-tone syllables in entries recorded in dictionaries.What this fact entails is that the S-W pattern is a strong,productive prosodic pattern in continuous speech.The alternation of S and W can also be used as a prosodic means of disambiguation.In the example below,the two sentences carry different meanings as a result of the metrical status of the syllable“guo”,in the S-W pattern in a.and in the S-S pattern in b.(17)a.W☒ k☒o guo le.I test par par“I took the test.”b.W☒ k☒o guò le.I test pass par“I passed the test.”The S-S pattern distinguishes itself from the S-W pattern in that both syllables bear one of the four lexical tones.The controversy centers around whether word-level stress pattern can be unequivocally established as in the S-W pattern.If the answer is yes,then what it is.Three sub-patterns have been proposed in the literature,S-S(i.e.,equal-level stress),M-S and S-M,where M refers to the stress level between W and S(Cao,2008;Wang&Feng,2006).It is worth pointing out that since both syllables in this pattern carry a lexical tone,stress does not play a contrastive role.For example,disyllabic words can be distinguished by varying tones in the first or second syllable,as in liùjí(level 6)and liújí(fail to move up to the next grade),or sìjì(four seasons)and sìjí(level 4).In these examples,tones are the only distinctive element.To determine which syllable is said with more stress,the judgement has to be based on auditory signals that contain pitch and duration variations,which are more subtle to differentiate in syllables bearing one of the four tones.We propose in the paper that the stress pattern of S-S combinations,which can be word entries in a dictionary or phrasal structures,is more readily defined in metrical terms in a prosodic hierarchy.At the level of prosodic word,prosodic position and tone type are found to impact the surface stress pattern.In a disyllabic word,the final syllable tends to be longer and the extra length is conducive to stronger stress perception.Meredith(1990)put forth the Stressability Hierarchy of Tones in Mandarin.Yin(2011)studied the effect of tone types on the judgement of stress in disyllabic words.They both identified Tone 4,the high falling tone in Mandarin,as the strongest tone,and the half Tone 3,the low tone with the final rise,as the weakest tone.Combining tone type and prosodic position,previous research(e.g.,Chen,Li&Wang,2003)found that Tone 4 in the first syllable is more likely to be stressed in Mandarin,as in“měi shù”(art).At the level of prosodic phrase,the Non-head Stress Principle(Duanmu,2007)places the phrase-level stress on the non-head in a phrasal structure.For example,in a verb-object(VO)structure the object receives more stress while in a modifier-noun(MN)structure,the modifier is more prominent.Disyllabic phrases such as ch☒o fàn are structurally ambiguous.When it means“to cook rice”it is a VO structure;when it means“cooked rice”,it is a MN structure.However,more people would agree fàn is more stressed than ch⛝o in both readings.We suggest that tone type and prosodic position are involved in the determination of stress here.In the half Tone 3 and Tone 4 combinations,Tone 4 carries more stress than the half Tone 3.Another important factor that contributes to the surface stress pattern of the S-S combinations is the placement of the nuclear stress.At the level of the intonational phrase,placement of the nuclear stress is also dependent upon contextual information such as focus structure of the sentence.The judgement of stress in a disyllabic S-S word can be enhanced greatly when the nuclear stress falls on the word.In our analysis of a read speech corpus,we found that morphosyntactic structure,tone type and prosodic position are relevant in determination of the stress position in prosodic words.In addition,lexical words are more likely to receive stress than function words in continuous speech.There is also a strong tendency to use the S-W pattern in disyllabic prosodic words in the data.