vowel features and tone assignment - university of michigancwlin/arabic loanwords.pdfvowel features...

1
Vowel Features and Tone Assignment: A Corpus Study of Arabic Loanwords in Mandarin Chinese Cheng-Wei Lin [email protected] University of Michigan, Department of Near Eastern Studies The present study reports empirical data from Arabic loanwords in Mandarin Chinese and investigates the correlation between vowel features in Arabic and tone assignment. This study follows Wu’s (2006) previous investigation into tone assignment, who examined English loanword in Mandarin. In her study, the author looked into tone adaptation in certain syllabic and prosodic contexts and showed specifically 1) monosyllabic words in English are adapted to Falling tone 2) initially stressed syllable tend to have Level or Rising tone 3) reliance on acoustic cues in non-initially-stressed syllable 4) salient consonants undergo epenthesis and receive a Low tone. These observations address the influence of prosodic and syllabic constraints on tone adaptations, whereas no further discussion in the effect of vowel properties is available. To fill in this gap, the present study considers possible effects brought by stress to word borrowing, but mainly focuses on the interplay between the vowel qualities from the donor language and the assigned tones in the borrowing language. Through an examination of a corpus built from a variety of resources, the study presents less-explored loanwords and indentifies other factors contributing to tonal adaptation. The following table gives some examples of Arabic loanwords, e.g., names, places, religious terms and animals, and offers tones assigned to each syllable of the loanword. Following the examples is a description about how the corpus was constructed. To investigate the effect of vowel length on tonal assignment, I built a corpus of 316 lexical items, consisting of a total of 937 syllabic tokens. The lexical items are in simplified Mandarin Chinese, collected from four sources: two dictionaries of foreign words in Mandarin Chinese (Liu 1984, Cen 1990), an Arabic-Chinese dictionary (2008) and Xinhua news agency. Each loanword is coded binarily based on the height and the length of the vowel, the tone assigned to is coded as well. Under the circumstances where the same Arabic word is borrowed differently by two or more sources, I incorporated all of them into the corpus. Loanwords were carefully examined and consultation was sought for the ones with questionable origin. If a loanword is found to have a false origin, then it is removed. 217 tokens were removed from the corpus since they are epenthesized syllables which have no vowel value in Arabic. As seen in table (1), the vowel inventory of Arabic is straightforward, i.e., three phonologically distinctive vowels, [i], [u], and [a], with contrastive feature [length]. Three independent variables are considered in the corpus : tone assignment in Mandarin, vowel length and height in Arabic. Due to a strong correlation found between the feature [back] and [high], which is the result of a small vowel inventory in Arabic, the feature [back] was not considered to be an independent variable. Diphthongs, [ai] and [au], were considered to carry the place feature of the first vowel and a length feature, that is, all diphthongs are [-high] [+long]. INTRODUCTION Cen, Q. (1990). Han yu wai lai yu ci dian =: Dictionary of foreign words in Mandarin. 1st ed.. Beijing: Shang wu yin shu guan . Department of Arabic Language of College of Foreign Languages at University of Beijiin. (2008). A la bo yu han yu ci dian =: Arabic-Mandarin Dictionary. Revised ed. Beijing: University of Beijing Press Liu, Z. (1984). Han yu wai lai ci ci dian =: A dictionary of loan words and hybrid words in Chinese. 1st ed. Shanghai: Shanghai ci shu chu ban she. Wu, Hsiao-hung Iris. (2006). Stress to tone: a study of tone loans in Mandarin Chinese. Feng-fan Hsieh and Michael Kenstowicz, eds. Studies in Loanword Phonology. MIT Working Papers in Linguistics 52, 227-53. METHODOLOGY CONCLUSIONS BIBLIOGRAPHY DISCUSSION Tone Category Tone Contour Pitch Value Arabic Mandarin Glossary T1 Level 55 ['babʁat] [ba55.gə55] Crested Myna T2 Rising 35 [ʔis.'laːm] [ʔi55.sɨ55.lan35] Islam T3 Dipping 214 [ʔal.'ya.man] [ye214.mən35 ] Yemen T4 Falling 51 [mu.'ħam.mad] [mu51.han213.mo51.də35] Muhammad Nonetheless, a statistical significance would not necessarily be a true indication of the effect of the variables in discuss on the tendency of tone assignment before a fundamental issue is resolved. That is, what is the distribution of tones in all possible phonotactics and syllabic structures? Table (2) presents simply possible tonal-phonotactical combination and (3) is a possible tonal-phonotactical combination weighted by the number of different orthographic variations. Neutral tone is ignored in the table due to the fact that it accounts for less than 0.1% of the overall counts. Tone Category Frequecy Percentage T1 335 26.3% T2 253 19.8% T3 325 25.5% T4 362 28.4% 1275 100.0% Tone Category Frequency Percentage T1 2864 24.4% T2 2963 25.3% T3 2012 17.2% T4 3887 33.1% 11726 100.00% From (2) and (3), we notice that the tones are actually unevenly distributed; therefore, a statistical analysis might be biased since the variable is not randomly distributed but has a predisposition. That is to say, in order to access the tendency of tone assignment, alternative approach is needed. I propose here a comparison between the percentage of the default distribution and the observed distribution in corpus to determine the preferred tone under each vowel feature as shown in (4) – (7). To be conservative and to avoid bias that might be introduced from the frequency of the orthographical representations, I adopted the distribution in (2) for comparison. 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 1 2 3 4 High Vowel Default Observed 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 1 2 3 4 Low Vowel Default Observed RESULTS A Chi-Square analysis was conducted, and a significant effect of vowel height from the donor language (p< 0.001) as well as the vowel length (0.01) was found on tone assignment in Mandarin Chinese. (4) (5) 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 1 2 3 4 Long Vowel Default Observed 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 1 2 3 4 Short Vowel Default Observed T2 preferred T1, T3 preferred (6) (7) T2, T3 preferred T1 preferred (1) (2) (3) As illustrated in the charts, although a dominating preference is not available, there are indications of preferences for tone assignment in specific vowel contexts. These indications are partially in harmony with Wu’s (2006) findings, e.g, an initially stressed syllable with short vowel. On the other hand, the vowel context would counterbalance the predisposition established by syllabic and prosodic contexts. These competing forces towards tone assignment confirm Wu’s conclusion that only all different factors are considered and integrated can we reach a comprehensive understanding of tonal adaptation.

Upload: others

Post on 24-Mar-2020

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Vowel Features and Tone Assignment - University of Michigancwlin/Arabic loanwords.pdfVowel Features and Tone Assignment: A Corpus Study of Arabic Loanwords in Mandarin Chinese . Cheng-Wei

Vowel Features and Tone Assignment: A Corpus Study of Arabic Loanwords in Mandarin Chinese

Cheng-Wei Lin [email protected] University of Michigan, Department of Near Eastern Studies

The present study reports empirical data from Arabic loanwords in Mandarin Chinese and investigates the correlation between vowel features in Arabic and tone assignment. This study follows Wu’s (2006) previous investigation into tone assignment, who examined English loanword in Mandarin. In her study, the author looked into tone adaptation in certain syllabic and prosodic contexts and showed specifically 1) monosyllabic words in English are adapted to Falling tone 2) initially stressed syllable tend to have Level or Rising tone 3) reliance on acoustic cues in non-initially-stressed syllable 4) salient consonants undergo epenthesis and receive a Low tone. These observations address the influence of prosodic and syllabic constraints on tone adaptations, whereas no further discussion in the effect of vowel properties is available. To fill in this gap, the present study considers possible effects brought by stress to word borrowing, but mainly focuses on the interplay between the vowel qualities from the donor language and the assigned tones in the borrowing language. Through an examination of a corpus built from a variety of resources, the study presents less-explored loanwords and indentifies other factors contributing to tonal adaptation. The following table gives some examples of Arabic loanwords, e.g., names, places, religious terms and animals, and offers tones assigned to each syllable of the loanword. Following the examples is a description about how the corpus was constructed.

To investigate the effect of vowel length on tonal assignment, I built a corpus of 316 lexical items, consisting of a total of 937 syllabic tokens. The lexical items are in simplified Mandarin Chinese, collected from four sources: two dictionaries of foreign words in Mandarin Chinese (Liu 1984, Cen 1990), an Arabic-Chinese dictionary (2008) and Xinhua news agency. Each loanword is coded binarily based on the height and the length of the vowel, the tone assigned to is coded as well. Under the circumstances where the same Arabic word is borrowed differently by two or more sources, I incorporated all of them into the corpus. Loanwords were carefully examined and consultation was sought for the ones with questionable origin. If a loanword is found to have a false origin, then it is removed. 217 tokens were removed from the corpus since they are epenthesized syllables which have no vowel value in Arabic. As seen in table (1), the vowel inventory of Arabic is straightforward, i.e., three phonologically distinctive vowels, [i], [u], and [a], with contrastive feature [length]. Three independent variables are considered in the corpus : tone assignment in Mandarin, vowel length and height in Arabic. Due to a strong correlation found between the feature [back] and [high], which is the result of a small vowel inventory in Arabic, the feature [back] was not considered to be an independent variable. Diphthongs, [ai] and [au], were considered to carry the place feature of the first vowel and a length feature, that is, all diphthongs are [-high] [+long].

INTRODUCTION

Cen, Q. (1990). Han yu wai lai yu ci dian =: Dictionary of foreign words in Mandarin. 1st ed.. Beijing: Shang wu yin shu guan . Department of Arabic Language of College of Foreign Languages at University of Beijiin. (2008). A la bo yu han yu ci dian =: Arabic-Mandarin Dictionary. Revised ed. Beijing: University of Beijing Press Liu, Z. (1984). Han yu wai lai ci ci dian =: A dictionary of loan words and hybrid words in Chinese. 1st ed. Shanghai: Shanghai ci shu chu ban she. Wu, Hsiao-hung Iris. (2006). Stress to tone: a study of tone loans in Mandarin Chinese. Feng-fan Hsieh and Michael Kenstowicz, eds. Studies in Loanword Phonology. MIT Working Papers in Linguistics 52, 227-53.

METHODOLOGY

CONCLUSIONS

BIBLIOGRAPHY

DISCUSSION

Tone Category

Tone Contour

Pitch Value Arabic Mandarin Glossary

T1 Level 55 ['babʁat] [ba55.gə55] Crested Myna

T2 Rising 35 [ʔis.'laːm] [ʔi55.sɨ55.lan35] Islam

T3 Dipping 214 [ʔal.'ya.man] [ye214.mən35 ] Yemen

T4 Falling 51 [mu.'ħam.mad] [mu51.han213.mo51.də35] Muhammad

Nonetheless, a statistical significance would not necessarily be a true indication of the effect of the variables in discuss on the tendency of tone assignment before a fundamental issue is resolved. That is, what is the distribution of tones in all possible phonotactics and syllabic structures? Table (2) presents simply possible tonal-phonotactical combination and (3) is a possible tonal-phonotactical combination weighted by the number of different orthographic variations. Neutral tone is ignored in the table due to the fact that it accounts for less than 0.1% of the overall counts.

Tone Category Frequecy Percentage T1 335 26.3% T2 253 19.8% T3 325 25.5% T4 362 28.4%

1275 100.0%

Tone Category Frequency Percentage T1 2864 24.4% T2 2963 25.3% T3 2012 17.2% T4 3887 33.1%

11726 100.00%

From (2) and (3), we notice that the tones are actually unevenly distributed; therefore, a statistical analysis might be biased since the variable is not randomly distributed but has a predisposition. That is to say, in order to access the tendency of tone assignment, alternative approach is needed. I propose here a comparison between the percentage of the default distribution and the observed distribution in corpus to determine the preferred tone under each vowel feature as shown in (4) – (7). To be conservative and to avoid bias that might be introduced from the frequency of the orthographical representations, I adopted the distribution in (2) for comparison.

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

1 2 3 4

High Vowel

Default Observed

0.0% 5.0%

10.0% 15.0% 20.0% 25.0% 30.0% 35.0%

1 2 3 4

Low Vowel

Default Observed

RESULTS

A Chi-Square analysis was conducted, and a significant effect of vowel height from the donor language (p< 0.001) as well as the vowel length (0.01) was found on tone assignment in Mandarin Chinese.

(4) (5)

0.0% 5.0%

10.0% 15.0% 20.0% 25.0% 30.0% 35.0%

1 2 3 4

Long Vowel

Default Observed

0.0% 5.0%

10.0% 15.0% 20.0% 25.0% 30.0% 35.0%

1 2 3 4

Short Vowel

Default Observed

T2 preferred T1, T3 preferred (6) (7)

T2, T3 preferred T1 preferred

(1)

(2) (3)

As illustrated in the charts, although a dominating preference is not available, there are indications of preferences for tone assignment in specific vowel contexts. These indications are partially in harmony with Wu’s (2006) findings, e.g, an initially stressed syllable with short vowel. On the other hand, the vowel context would counterbalance the predisposition established by syllabic and prosodic contexts. These competing forces towards tone assignment confirm Wu’s conclusion that only all different factors are considered and integrated can we reach a comprehensive understanding of tonal adaptation.