New Dialect Formation: the Case of Taiwanese Mandarin
A thesis submitted for the degree of Doctor of Philosophy in
LANGUAGE AND LINGUISTICS
Department of Language and Linguistics
University of Essex
I dedicate this dissertation to my parents
Kuo Fen Ching and Kuo Ku Mei Yu
and my grandmother
Wu Yu Hsin
1.1 Introduction 1
1.2 The Sociolinguistic Settings of Taiwan 3
1.3 Aims 4
1.4 Research Questions 10
1.5 Organisation 11
2 Previous Studies of Taiwanese Mandarin
2.1 Introduction 12
2.2 Background to the Transplantation of Mandarin to Taiwan 12
2.3 Studies of Taiwanese Mandarin 15
2.4 An Alternative Explanation for the Formation of Taiwanese Mandarin 41
2.5 Summary 42
3 Taiwanese Mandarin as the Outcome of Koinéisation Processes through
Dialect Contact and Mixture during New Dialect Formation
3.1 Introduction 43
3.2 Dialect Contact and Koinéisation 43
3.3 Koineisation and New Dialect Formation 47
3.4 New-dialect Formation 53
3.5 The Important Role of the Socio-Demographic and Linguistic
Characteristics of Dialects in New-dialect Formation
3.6 Taiwanese Mandarin and New-Dialect Formation 61
3.7 Summary 64
4 Sociohistorical Linguistic Setting of Taiwan
4.1 Introduction 65
4.2 Background of Taiwan 65
4.3 Immigration History and Languages Transplanted 68
4.4 The Geographical Origin, Languages, and the Size of the Mainland
Immigrant Population, especially that of the Original Mandarin
4.5 Background of Keelung 85
5.1 Introduction 90
5.2 Data Collection in Keelung 90
5.3 Selection of Linguistic Variables 101
5.4 Method Used for Data Analysis 101
5.5 Statistical Package 106
5.6 Summary 106
6 The Retroflex Initials /t/, /th/, and //
6.1 Introduction 107
6.2 Description of [t], [th], and  in Standard Mandarin 113
6.3 Reviews of Previous Explanations for the Lack of the Retroflexes /W/,
/WK/, and // in Taiwanese Mandarin
6.4 Analysis of the Use of the Retroflexes in Keelung Taiwanese Mandarin
and the Source of Variants
6.5 Further Evidence from Keelung Taiwanese Mandarin Questioning the
Southern Min Population as the Sole Contributor for the Use of
Alveolar Variants of /t/, /th/, and //
6.6 Conclusion 151
7 The Retroflex Initial //
7.1 Introduction 153
7.2 Review of Previous Studies 155
7.3 Description of  in Standard Mandarin 162
7.4 Analysis of the Use of the Retroflex // in Keelung Taiwanese
Mandarin and the Source of the Variants of the Retroflex //
7.5 Further Evidence Questioning the Southern Min-origin Hypothesis 171
7.6 Evidence for the Three-Stage Model of New-Dialect Formation 177
7.7 Conclusion 180
8 Further Evidence Demonstrating the Likelihood of Taiwanese Mandarin as
the Outcome of Koinéisation
8.1 Introduction 182
8.2 Further Evidence Demonstrating Taiwanese Mandarin as a Koineised
Variety in a Dialect Contact and Mixture Situation during New-Dialect
8.3 Can the Processes of New-Dialect Formation Help Us Resolve the
Mystery of Two Controversial Variants?
8.4 Conclusion 205
9 Conclusion 206
Appendix 1 Background of Informants 209
Appendix 2 Questionnaire 211
Appendix 3 GoldVarb 2001 212
Following the end of the Second World War, when the Nationalists (Kuomintang, KMT)
took over Taiwan from the Japanese, bringing many migrants from the Mainland with them,
Beijing Mandarin (commonly known as ‘Guoyu’) was promoted as the national language of
the island. One of the salient features of Beijing Mandarin is the use of retroflex initials (/t/,
/th/, //, and //) However, the Mandarin dialect spoken today in Taiwan mostly lacks these
salient features. Common explanations for the lack of retroflex features suggest that they
merged, on Taiwanese soil, with Southern Min dental and alveolar equivalents (/t/ Î /ts/,
/th/ Î /tsh/, and // Î /s/). The merger is often regarded as a result of sound substitutions
often found in contexts of language contact or of second language acquisition failure
deriving from first language (Southern Min) interference. The lack of the retroflex initials in
Southern Min is also commonly said to have caused alveolars to replace the retroflex initial
// through deretroflexion.
I will demonstrate that such claims have, in fact, oversimplified the situation and, indeed,
that the so-called merger and deretroflexion may have never taken place at all in Taiwan for
a number of socio-historical and linguistic reasons. An investigation of the socio-
demographic composition and dialect use of the original Mandarin population in Taiwan
leads us to question the traditional merger and deretroflexion hypotheses. Rather than
arguing, as most traditional explanations do, that retroflex initials merged with non-retroflex
ones in Taiwan after 1945, I suggest here that processes of dialect contact - koinéisation -
better explain the eradication of the retroflex initials. I propose that the relatively small
number of retroflex-using dialects in the original Mandarin population of Taiwan led to the
retroflex initials being levelled away and that the great number of alveolar-using dialects
safeguarded the survival of the alveolars, hence the lack of retroflex initials and the presence
I will first demonstrate how koinéisation processes have been at work in shaping
Taiwanese Mandarin by analysing the distribution of the variants of four retroflex variables
/t/, /th/, //, and // in spoken Taiwanese Mandarin in Keelung, a northern city of Taiwan
where Mandarin immigrants first landed.
Further evidence is then presented in support of my hypothesis that the original Mandarin
population indeed, may have played a much more important role in the formation of
Taiwanese Mandarin than previously suggested and that Taiwanese Mandarin is highly
likely to be a koineised variety following dialect contact. I investigate four types of dialect
variant: (i) those absent both in Southern Min and in early Taiwanese Mandarin but present
in Taiwanese Mandarin today; (ii) those absent in Southern Min but present both in early
Taiwanese Mandarin and in Taiwanese Mandarin today; (iii) those reported to be likely to
disappear but have been present in Taiwanese Mandarin today; and (iv) those present in
Taiwanese Mandarin but were reported to be either absent or present both in Southern Min
and in early Taiwanese Mandarin.
By scrutinising the use of four retroflex variables and four types of variant, I will
demonstrate that some previous paradoxes may be resolved by acknowledging the crucial
role of the demographic, geographic, and linguistic make-up of the original Mandarin
population in the formation of Taiwanese Mandarin.
I conclude that the development of Taiwanese Mandarin is highly likely to be the result
of koinéisation processes, in particular the levelling process, which is, to quote Trudgill
(2004:149), ‘a matter of simple calculation’.
Chapter 1 ~ Introduction
Mandarin in Taiwan has been spoken for 59 years since Guoyu1, the standardised
Mandarin2 ‘based on the dialect of Beijing’3 (known as Pekinese4 or Pekingese5, which also
serves as the base of the Standard Mandarin6 of the Chinese Mainland, known as Putonghua7)
was promoted as the official language of Taiwan during the National Language Movement
that commenced in 19458 and since the influx of immigrants, known as mainlanders9, fled to
the island from the Mainland between 1945 and 1956. 60 years on, the variety10 of Mandarin
spoken on Taiwan is quite distinct from the Pekingese spoken on the Chinese Mainland and
indeed, ‘differs considerably from Pekingese in phonology, syntax, and lexicon’11. For
example, the most salient feature of the Beijing sound system – the retroflex12 – is lacking in
the Mandarin variety spoken on Taiwan (hereafter, Taiwanese Mandarin13). Other
1 Guoyu, equivalent to ‘national language’, was implemented by Republic of China in 1911 (Tsao 1999a:333).
2 See Tse (1986:25) for ‘Standardization of Chinese in Taiwan’.
3 Kubler (1985a:85).
4 ‘Beijing’ is also spelled as ‘Peking’ or ‘Peiping’; the former is based on Hanyu Pinyin (Chinese Romanisation),
while the latter is based on Luoma Pinyin (Romanisation). Kubler (1985a) uses ‘Pekinese’ (p85) and ‘Beijing
Mandarin’ (p145) to refer to ‘the dialect of Beijing’.
5 Barnes (1977:255) uses ‘Pekingese’ to refer to the dialect of Beijing.
6 Kubler (1985a:18) defines ‘Standard Mandarin’ as ‘the dialect of Beijing minus extremely colloquial local
features’ and states that ‘[i]t is the speech form which, in theory at least, both mainland China and Taiwan are
promoting as their national language’. In his earlier work, Kubler (1979:27, 28, 38) uses the term ‘standard
Peking’ and ‘Peking Mandarin’ to refer to Standard Mandarin, representing ‘the speech of the pre-1949 Peking
speakers’. Zhang (1974:53-4) stated that Standard Mandarin is based on Peking Mandarin that is spoken daily in
Peking by people who have received middle school education.
7 ‘Putonghua [literally, ‘common language’] is defined as a speech form based on the grammar and lexicon of
North Chinese dialects and the phonology of Pekingese’ (Barnes 1977:255) and it has been promoted by the
People’s Republic of China since 1956 (Barnes 1983:292).
8 Tsao (1999a:339), but 1946 in Tse (1986:25).
9 Most came with the Nationalist Party (Kuomintang, KMT) and some came independently during the Chinese
Civil War (1945-1949). They are collectively referred to as mainlanders. To be detailed in §4.
10 Following Chambers and Trudgill (1998), ‘variety’ is adopted throughout the study as a neutral term for any
particular kind of language, when it is considered as a single entity. Throughout the study, ‘dialects’ refers to
regional varieties spoken in China, which are often mutually unintelligible. Detailed description of the
ethnolinguistic environment is given in §4.
11 Kubler (1985a:85).
12 More details in §1.3.1.
13 Mandarin spoken in Taiwan is commonly known as ‘Taiwan de Guoyu’ or ‘Taiwan Guoyu’. The former
literally means ‘Mandarin of Taiwan’, while the latter means ‘Taiwanese-accented Mandarin’. The variety of
Mandarin spoken in Taiwan was termed ‘Taiwan Mandarin’ by Kubler (1979, 1980, 1984, 1985a, b), which
covers both meanings mentioned above and is restricted to ‘the Mandarin spoken by the younger generation of
both native Taiwanese and mainlanders who were born on Taiwan after 1945 and have received at least a high
school education’ (Kubler 1979:28). ‘Taiwanese Mandarin’ in this thesis refers to the Mandarin variety spoken
by ‘the people of Taiwan’ regardless of accents, age, education, or ethnic groups. The term is used to cover all
Mandarin varieties spoken in Taiwan, including but not limited to the Mandarin varieties spoken by Taiwanese
Aboriginal, Hakka, Mainlander, and Southern Min, four major ethnolinguistic groups in Taiwan. More details in
differences14 between Taiwanese Mandarin and Standard Mandarin include, but are not
limited to, lesser use of the neutral tone and no rise of the third tone in pronunciation, the use
of <yԁu/méiyԁu15> ‘have/not have’ as auxiliaries in grammar, and the use of lexemes such as
<jiӽotàchƝ> as opposed to <zìxíngchƝ> ‘bicycle’ in vocabulary. This thesis questions how
Taiwanese Mandarin originated. Previous theories16 have suggested that Taiwanese Mandarin
was formed primarily through language contact interference or second dialect acquisition
failure resulting from the lack of corresponding sound systems and from differences in
grammar and vocabulary in Southern Min17, the native dialect spoken by the majority of the
population. Here I suggest, contrary to these theories, that the formation of Taiwanese
Mandarin can be explained by processes of dialect contact.
In what follows, I will first briefly describe Taiwan's sociolinguistic setting which has
provided a fertile environment for the cultivation of sociolinguistic studies before proceeding
to state the aims of the thesis. Based on these aims, I propose five sets of research questions.
Finally, the organisation of the thesis is outlined.
§1.2 and §4. Taiwanese Mandarin is also used to cover regional varieties of Mandarin in Taiwan, e.g. Taiwanese
Mandarin spoken in Taipei, Taichung, or Tainan.
14 Examples here cited from Kubler (1985a:106, 142, 160, 166-8). See also Kubler (1979:28-39; 1984:6-9;
1985b:157-173). More details in §1.3.1.
15 The spellings <yԁu/méiyԁu> are romanised transliterations of the characters <ڶ/ڶ> ‘have/not have’. The
spellings are based on the standardised Chinese romanisation system, known as Hanyu Pinyin, ‘promulgated in
1958 by the government of the People’s Republic of China’. Hanyu Pinyin is ‘based exclusively upon the
pronunciation of the Beijing dialect’ and ‘provides the standard pronunciation of pԃtǀnghuà’ (Chen
1999:187-188). It ‘adopts the Roman alphabet and assumes a phonemic representation’ and borrows the
diacritics from <Zhuyin Zimu>-an earlier version of the sound-annotating system for the standardised
pronunciation of characters’-promulgated in 1918-to indicate four tones of Modern Standard Chinese
(Putonghua, Guoyu) (Chen 1999:180, 186-7). For example, in Zhuyin Zimu, four tones are superimposed next to
vowels, e.g. 1st tone: <ა> ԃԖ ‘mother’, 2nd tone: <> ԃԖԨ‘numb’, 3rd tone: <್> ԃԖԩ‘horse’, 4th tone:
<ᒼ> ԃԖԪ‘scold’, Neutral tone: <ma> ԃԖԦ ‘question marker’, while in Hanyu Pinyin, the same characters
are romanised with added diacritics as in mƗ, má, mӽ, mà, ma. Hanyu Pinyin is a ‘phonetic scheme of Chinese’
and intended to be used as ‘a sound-annotating tool to facilitate learning and use of characters…[and] an aid in
promoting Modern Standard Chinese among the native speakers of other dialects, in much the same way as
Zhuyin Zimu’ (Chen 1999:188-9).
16 E.g. Kubler (1979, 1980, 1985a, b, c); Tung (1995); Tsao (2000). More details in §1.3.1.
17 Southern Min or Southern Fujian is known as <Minnan-yu> in Mandarin Chinese. Min is the short name for
Fujian. Min consists of Northern Min and Southern Min, the local dialects of northern and southern Fujian
province, China. Southern Min is also know as <Hokkien>, especially in Southeast Asia. Within Fujian province,
Southern Min has three major varieties, Amoy (Xiamen), Changchew (Zhangzhou), and Chinchew (Quanzhou).
Outside Fujian province, Southern Min varieties include Hainan, Teochew (Chaozhou), and Tai-gi (Taiwanese).
1.2 The Sociolinguistic Settings of Taiwan
Taiwan, also known as Formosa18, is an island separated from the southeastern coast of the
Chinese Mainland by the Taiwan Strait, with an area of 36,006 square kilometres and a
population of about 23 million. Two major ethnic groups make up the society: the
Aborigines19 and the Chinese20. The Chinese group is composed of three ethnic subgroups: the
Hakkas21 (12% of the population), the mainlanders22 (13%), and the Southern Mins23
(73.3%).24 The Aborigines speak Austronesian languages, while the Chinese speak varieties
of Chinese dialects. Each ethnic group speaks a distinct variety of its own: Hakka varieties of
Chinese, Southern Min varieties of Chinese, mainlander varieties of Chinese, and varieties of
Austronesian languages.25 Varieties spoken by each group are mostly mutually unintelligible.
In addition to these varieties, each group also speaks a common variety of Chinese, Guoyu,
though the mother tongue of each individual undoubtedly colours actual realisations of it26.
Japanese is also commonly spoken among those who received Japanese education during the
occupation between 1895 and 194527. English is taught as a second language at secondary
schools. Since 2000, many have received English lessons in kindergartens. Since 2001,
Southern Min, Hakka, and Austronesian languages have been taught in primary schools. Such
a multilingual society creates a live laboratory for linguists who are interested in
multilingualism28, language maintenance and language shift29, language contact30 and
18 Literally, ‘beautiful island’, named by the Portuguese explorers who sailed past the island in the 17th century.
19 <Yuanzhumin> in Guoyu and ‘Formosans’ in English. The aboriginal people came from the Chinese Mainland
to Taiwan several thousand years ago. More details in §4.
20 Known as <Han-ren> ‘ethnic Chinese’ in Guoyu. More details in §4.
21 Literally, ‘guest people’. <Kejia-ren> in Guoyu and <Hak7 ka1> in Hakka. Although the origin of the Hakkas is
still not well understood, Taiwanese Hakkas are known to have come from Guangdong and Fujian provinces of
China in the 17th century. More details in §4.
22 Mainlanders can be referred to as <waisheng-ren> ‘external-province person’ or <dalu-ren> ‘residents of
Mainland China’ in Guoyu. The former refers to people emigrating from Mainland China between 1945 and 1956
and their descendants, as opposed to the Taiwanese local residents, known as <bensheng-ren> ‘original-province
person’, who have been in Taiwan since the 17th century. The latter refers to immigrants from Mainland China
after 1987. The use of the term ‘mainlanders’ in this thesis refers to the first sense. More details in §4.
23 <Minnan-ren> in Guoyu and <Hokkiens> in Southern Min. The Southern Mins came from Southern Fujian
province in the 17th century. More details in §4.
24 The proportion of the ethnic population is based on S.F. Huang (1995). See also Chaffee et al (1969:35),
Lamley (1981:291-3), and Kubler (1985a:26).
25 The history of the ethnolinguistic make-up of Taiwan is discussed in §4.
26 It has been claimed that by 1991, at least 89.97% of the population were able to communicate with each other
in Guoyu (Ke 1991:5).
27 Tsao (1999b:321).
28 S.F. Huang (1995:319-353).
29 Chan (1994); Young (1989).
language typology31, language and ethnic eco-political power32, language ecology and
vitality33, language consciousness, attitudes and identity34, language choice and language
death35, language policy and preservation of mother tongues36, language planning and
language use37, language conflict and language competition38, and language variation and
change39. As J.H. Hsu (1987) rightly observed, ‘since almost all Chinese dialects can be found
on this island, Taiwan might be linguistically regarded as a miniature version of China.’
1.3.1 Challenging previous theories in accounting for various characteristics of
Much literature to date has reported one distinct characteristic of Taiwanese Mandarin: the
lack of retroflex initials /W/, /Wh/, //, and //. It has suggested that the absence of retroflex
initials in present-day Taiwan is a result of the lack of sound correspondences in Southern
Min40; thus, Southern Min speakers had difficulty either in articulating retroflexes41 or
acquiring them42 and as a result, they ‘tend(ed) to substitute’ retroflexes with non-retroflexes
‘from their native language’43. Specifically, some scholars44 have reported that the retroflex
consonants have merged with the non-retroflex ones in Taiwanese Mandarin due to the
influence of Southern Min, based on a comparison between standard Beijing Mandarin and
Southern Min sound systems (table 1.1).
30 Kubler (1985a, b, c); J.L. Hsu (1994).
31 R.L. Cheng (1997).
32 S.F. Huang (1995:19-49).
33 Ang (1997); S.F. Huang (1995:128-166); Liao (2000).
34 S.F. Huang (1995:174-226); K.P. Hsieh (1997); J.H. Hsu (1987); S.F. Huang (2000); L.J. Lu (1988); Tse
35 S.F. Huang (1995:257-294); Tsao (1995).
36 Ang (2002); Cheng and Yu (1997); M.J. Huang (1997); Tsao (1997); Tse (1987); Y.H. Zhang (1997).
37 Berg (1986); Han (1996); S.F. Huang (1995:88-127); Tsao (1999a).
38 S.F. Huang (1995:55-87); Tsai (1997).
39 Chan (1984); Fu (1999); H.H. Huang (1995); M.C. Li (1995); Yue (1992).
40 Chien (1971:13); Kubler (1985a:90); M.C. Li (1995:90); C.S. Tung (1995:19).
41 Wei (1980:30); C.C. Li (1983:263); Tsao (2000:284).
42 Kubler (1985a:91).
43 Kubler (1985a:90).
44 C.C. Li (1983:263); Kubler (1985a:92); Dai (1997:79).
Table 1.1 Comparison between Standard Beijing Mandarin and Southern Min45
Standard Beijing Mandarin Standard Southern Min
(a) Retroflex affricates and fricatives: /W, Wh, / None exist
(b) Dental affricates and fricatives: /ts, tsh, s/ /ts, tsh, s/
(c) Retroflex fricative: // Doesn’t exist
(d) Dental continuant: /l/ /l/
The lack of sound correspondence is often regarded as a reason for Southern Min speakers’
failure to acquire standard Beijing Mandarin46 and for their replacement of (a) and (c) by (b)
and (d)47. For example, Kubler (1985a)48 claimed that in Taiwan, the retroflexes /t/, /th/, and
// have merged into the alveolars /ts/, /tsh/, and /s/ due to the lack of retroflexes in Southern
Min49. Similarly, in his linguistic study of Taiwan and Beijing Mandarin, Dai (1997:79)
claimed that speakers ‘whose dialect is Taiwanese tend to speak Mandarin with vernacular
colour’ and that ‘the Minnan50 dialect speakers cannot articulate the retroflex zh-, ch-, sh-
initials; he then concluded that ‘as such, the retroflex initials zh-, ch-, sh- have gradually
merged with the dental sibilants z-, c-, and s-.’ C.C. Li (1983:263-4) reported that
‘the Southern Min speakers cannot articulate the retroflex zh-, ch-, sh- initials…It takes
time and effort for them to master some phonetic features which are very different from
those in their Taiwanese or Southern Min dialect. In fact, the Southern Min has had a
great influence on Mandarin speakers in Taiwan. Even today, standard or official
Mandarin as commonly spoken in Taiwan differs considerably from standard Peking.
Demographically, those who were born and raised in Peking have been reduced in
numbers thirty-one (1949-1981) years after they settled down in Taiwan. The new
generations who were born to Peking-Mandarin speakers tend to adopt the Mandarin
gradually detached from PM in pronunciation and grammar. As such, the retroflex
initials zh-, ch-, and sh- [/t
h/, and /
/] have gradually merged with the dental
sibilants z-, c-, and s- [/ts/, /tsh/, and /s/] so that <zhǌ> ‘pig’ sounds like <zǌ> ‘rent’,
<chǌ > ‘out’ like <cǌ> ‘coarse’ and <shǌ> ‘book’ like <sǌ> ‘vegetable’…deviation of
those sounds mentioned above…have become more and more acceptable in Taiwan. The
Taiwanese pronunciation has had a great impact on the original PM, since about 71% of
the population in Taiwan still speak Taiwanese.’51
45 See Chien (1971:13) and Kubler (1985b:157) for contrastive analysis of this kind.
46 See Wei (1980:30) and C.C. Li (1983:263) for this position.
47 See Chien (1971:16) and Kubler (1985b:157) for this treatment.
48 It should be noted that Kubler’s emphasis was on ‘the influence of Southern Min on Taiwan Mandarin’ (p15)
and his data were collected during the 1970s (1971-1972; 1974; 1975-1978; 1980). Perhaps, because Kubler
(1985a) was trying to describe a general picture of Southern Min’s influence on Taiwan Mandarin, he did not
consistently include background information on his informants, such as age, ethnicity, gender, and education.
49 Although Kubler (1985a:174-5) also acknowledged that the lack of retroflexes is also a feature of ‘Southern
Mandarin…spoken as a first or second language by the people of Southern China’ and thus ‘it is very well
possible that Southern Mandarin may have had a reinforcing effect,’ since his work was mainly on the influence
of Southern Min, he did not elaborate further how such an effect might have operated.
50 Literally, Southern Min.
51 The use of brackets < > to enclose spellings (e.g. <cǌ>) is mine.
S.M. Wu (1990:70) claimed that due to a lack of corresponding retroflexes, /t/, /th/, //, and
// were replaced by phonetically close-matched sounds in Southern Min: /ts/, /tsh/, /s/, and /l/.
Dai (1997), Kubler (1985a), and S.M. Wu (1990) were describing a type of Mandarin spoken
by speakers whose dialect is Southern Min (or Taiwanese), while C.C. Li (1983) was
describing a type of Mandarin spoken by both the Southern Min population and the new
generations who were born to Beijing Mandarin speakers. However, we see that they appear to
share a similar view – that the retroflex consonants /t/, /th/, //, and // had once existed on
Taiwanese soil but later merged into the non-retroflex ones /ts/, /tsh/, /s/, and /l/ in Taiwanese
Mandarin52 due to the great impact of the Taiwanese pronunciation on the original Beijing
The above view takes for granted the pre-existence of retroflex sounds on Taiwan. But, is it
possible that the retroflex sounds were never actually brought to Taiwan by Beijing Mandarin
speakers? Or, if they were introduced to Taiwan, is it possible that the number of Beijing
Mandarin speakers was too small for retroflex initials to survive? No systematic research has
been conducted to investigate this possibility or to consider other sources of influence on
Taiwanese Mandarin, probably because of the overwhelming consensus about the enormous
influence of Southern Min on Taiwanese Mandarin. This thesis is devoted to addressing these
holes in the literature by embarking on a preliminary investigation of the validity of the
Southern Min origin of Taiwanese Mandarin. Most importantly, I attempt to demonstrate that
Taiwanese Mandarin is likely to be a koineised variety that emerged from dialect contact and
mixture and that the original Mandarin population played an important role in shaping modern
My findings indicate that the majority of the original Mandarin population used
non-retroflex variants but not retroflex ones, hence the levelling of the retroflex variants and
the victorious survival of the non-retroflex ones. So, rather than arguing, as most traditional
explanations do, that retroflex initials merged with non-retroflex ones on Taiwan due to the
52 Or ‘Taiwan Mandarin’ to use Kubler’s (1985a) terminology; ‘Taiwan-based Mandarin’ to use C.C. Li’s (1983)
influence of Southern Min, I suggest here that the original Mandarin population played an
important role in the development of the dominant use of non-retroflex variants in Taiwanese
Mandarin and that processes of dialect contact and koineisation53 better explain the
eradication of retroflex variants and the arrival of non-retroflex variants. That is, Taiwanese
Mandarin is very likely to have been a koineised variety that emerged through dialect contact
and mixture during ‘new-dialect formation’54.
To further support my hypothesis that Taiwanese Mandarin may indeed be a koineised
variety, I investigated why some variants55 that are absent in Southern Min have survived into
both early Taiwanese Mandarin and contemporary Taiwanese Mandarin and why certain
variants that are absent in Southern Min but either absent or present in early Taiwanese
Mandarin56 have survived into Taiwanese Mandarin today. My findings suggest that the
variants under investigation were the majority variants in the original Mandarin population,
hence their survival into both early Taiwanese Mandarin and contemporary Taiwanese
Mandarin or into contemporary Taiwanese Mandarin only. The results support my hypotheses
that Taiwanese Mandarin is highly likely to derive from the original Mandarin population and
to have been the result of koineisation.
Investigating the source of variants in Taiwanese Mandarin reveals the crucial role of the
original Mandarin population in the formation of Taiwanese Mandarin. Echoing the ‘founder
principle’57 proposed (under various labels) by various scholars58, the present study again
demonstrates that a careful analysis of the sociolinguistic history of the community, especially
an appreciation of the ‘ingredient’ dialects used by the founding settlers in communities that
have undergone demographic upheaval, leads us to a greater understanding of the formation of
53 Trudgill (1986, 2004) describes a series of linguistic processes, together called “koineisation”, which occurred
as a result of dialect contact (mixing, simplification, levelling, and reallocation). See §3.2 for further discussion.
54 Trudgill (2004).
55 Diphthongs /ie/ and ye/ as well as the triphthong /uei/. See §8.
56 The /n/-/l/ distinction, the allophonic distribution of [n] and  in /1/, the free variation of [)]- in /)/, the
functional distribution of the weak stress, the falling tone variant in the third tone /214/, the monophthongs /y/
and //, and the voiceless labial-dental fricative /f/. See §8.
57 See §3.5 for further discussion.
58 Mufwene (1996, 2001); Britain (2001a, b).
1.3.2 Lack of Chinese evidence in the dialect contact model proposed by Trudgill (1986,
As mentioned, the purpose of the thesis is to demonstrate that Taiwanese Mandarin is a
new variety that emerged through the processes of dialect contact described in Trudgill (1986,
2004). In the dialect contact and new-dialect formation model proposed by Trudgill (1986,
2004), most evidence is drawn from Indo-European languages and particularly Germanic
languages, while Chinese, as the world’s most populous language, is not drawn on for
evidence. The study therefore hopes to contribute to evidence profiles of the dialect contact
and new-dialect formation model by investigating whether we could explain the development
of Taiwanese Mandarin as the output of the processes of dialect contact working on the many
varieties of Mandarin that came to Taiwan between 1945 and 1956.
1.3.3 Challenging the common practice of understanding the development of
Taiwanese Mandarin by comparing standard Mandarin to Taiwanese Mandarin
In existing studies of Taiwanese Mandarin, standard Mandarin is often compared to
Southern Min to find out the linguistic differences between the two, so that teachers may
understand the learning difficulties Southern Min speakers may have in learning Mandarin.
Their exploration into the various difficulties faced by Southern Min learners of Mandarin
often leads them to believe that the differences between standard Mandarin and Taiwanese
Mandarin come from the influence of Southern Min and that Taiwanese Mandarin has
deviated from standard Mandarin since the promotion of standard Mandarin in 194559.
Sometimes, Taiwanese Mandarin is compared to standard Mandarin to examine how
Taiwanese Mandarin has differed from standard Mandarin since the promotion of standard
Mandarin60. In other cases, standard Mandarin, Taiwanese Mandarin, and Southern Min are
compared to understand the development of Taiwanese Mandarin61, in terms of how it has
departed from standard Mandarin (under the influence of the dominant dialect - Southern Min)
since the promotion of standard Mandarin. Variationist studies have also compared standard
59 E.g. Chien (1971); Kubler (1979); S.M. Wu (1990); Ye (1995).
60 E.g. Dai (1997).
61 E.g. Kubler (1985a, b); Wei (1980).
Mandarin and Taiwanese Mandarin to examine whether there has been linguistic change62.
This methodology can be seen as flawed however. It assumes that standard Mandarin was
spoken by the Taiwanese population and that today’s Taiwanese Mandarin diverged from that
standard. However, there is very little evidence that standard Mandarin was once used as a
vernacular form of Taiwanese peoples. Thus, studies that propose there have been changes in
spoken forms from standard to a Taiwanese Mandarin need to demonstrate first the existence
of that standard form in the community.
Treating language varieties as diverging from standard forms is a common practice among
some sociolinguists and such a practice has been criticised by others. For example, Mufwene
(2001:75) raises his concern over the practice of some creole studies to compare standard
British English with African American Vernacular English (AAVE):
‘[o]ne of the flaws of genetic creolists lies in the all-too-common comparison of creoles’
structural features with those of the standard varieties of their lexifiers. The social
histories of the relevant colonies suggest that the varieties to which the makers of creoles
were exposed and which they restructured were nonstandard. Thus it is with them that
comparisons must be made to develop an adequate picture of what was restructured and
In the study of the origins of New Zealand English, Britain (2001a:40-41, 72) comments that:
‘[i]n the study of post-colonial varieties of English, such as those spoken in New Zealand
and Australia, the need for social realism in dialectological analysis…applies to the very
origins of those dialects…differences between present-day New Zealand English (NZE)
and British English have, in the past, frequently been analysed as if the New Zealand
forms necessarily must be innovations, and often appear to take RP as their baseline for
analysis…Yet,…we must pay much more attention to the social makeup of the settler
speech community – as heterogeneous, with diverse geographical, social and linguistic
origins… – in order to fully understand the nature and course of linguistic change…the
history of New Zealand English is not a history of divergence from Standard English, but
one that has in many ways been shaped by the contact and koineisation of predominantly
southern English, Scottish, Irish and Australian varieties.’
Crucially, he argues that if one wishes to claim that a sound has changed from A to B in a
particular community, the claim must be wrong if A has not been confirmed to have ever
existed in that community.
The present study will show that as in the study of the origins of New Zealand English and
African American Vernacular English, the common practice of studying the development of
post-colonial or immigrant varieties by comparing standard forms to the varieties under
62 E.g. G.C. Lu (1997); M.C. Li (1995); Yue (1992).
investigation has impeded our understanding of the origins of Taiwanese Mandarin,
resulting in misleading conclusions on the formation of the variety.
1.4 Research Questions
The investigation was carried out in Keelung, one of the northern cities of Taiwan, where
speakers from the Mainland concentrated and which also served as a transition city before
most migrated to other parts of Taiwan. I postulate five sets of research questions here:
(1) Common explanations for the lack of retroflex initials /t/, /th/, and // in Taiwanese
Mandarin suggest that they merged, on Taiwanese soil, with Southern Min alveolar
counterparts /ts/, /tsh/, and /s/. The merger is often regarded as a result of sound
substitutions often found in language contact or of second language acquisition
failure deriving from first language (Southern Min) interference. But, was it really a
merger? Is it possible that the alveolars had already existed in the dialects brought by
the original Mandarin population who arrived between 1945 and 1956?
(2) Why are the retroflex initials /t/, /th/, and // missing in Taiwan? Is it possible that
they were never actually brought to Taiwan in the first place? If they were introduced
to Taiwan, is it possible that the number of speakers who did use them was too small
for retroflex initials to survive into Taiwanese Mandarin?
(3) Common explanations for the lack of retroflex initial // in Taiwanese Mandarin
suggest that // was “substituted”, on Taiwanese soil, by alveolars /dz/ or /l/ that are
present in Southern Min. The substitution is often regarded as a result of second
language acquisition errors made by Southern Min learners of Mandarin. But, was it
really a “substitution”? Is it possible that the alveolars already existed in the dialects
brought by the original Mandarin population?
(4) What are the origins of the variants in Taiwanese Mandarin today that Southern Min
lacks? Is it possible that the variants were already in use among the original Mandarin
(5) Previous studies have found that the variants that are present in standard Mandarin
but absent in Southern Min are present in Taiwanese Mandarin, while others have
reported that the same variants are replaced by “non-standard” variants in Taiwanese
Mandarin. Both the presence of standard variants and of non-standard variants have
been claimed to have come from the Southern Min population. But, is it possible that
both the “standard” variants and “non-standard” variants came from the original
Mandarin population and that the mixed use of both types of variants reflects a
diffuse early Taiwanese Mandarin speech community?
Discussion of the above questions aims to verify my hypotheses that Taiwanese Mandarin is
likely to be a koineised variety in dialect contact and mixture situations and that the original
Mandarin population may have played a significant role in shaping Taiwanese Mandarin.
The thesis is organised as follows. I provide in chapter two, a critical review of previous
studies of Taiwanese Mandarin. In chapter three, the theoretical background of the thesis is
given. Chapter four outlines the sociolinguistic and socio-demographic settings of Taiwan.
Chapter five discusses the methodology applied in the thesis. Chapter six describes, analyses,
and discusses retroflex initials /t/, /th/, and //. Chapter seven describes, analyses, and
discusses the retroflex initial //. Chapter eight provides further evidence supporting my
hypothesis that Taiwanese Mandarin is a koineised variety. Finally, a conclusion is drawn in
Chapter 2 ~ Previous Studies of Taiwanese Mandarin
In this chapter, I aim to provide a critical review of the previous studies of Taiwanese
Mandarin in order to provide a background to the present study. I will first briefly introduce
the background of the transplantation of Mandarin onto Taiwan before proceeding to the
review section. Finally, based on the studies reviewed, I propose an alternative view of the
formation of Taiwanese Mandarin.
2.2 Background to the Transplantation of Mandarin to Taiwan
At the end of the Second World War, the Japanese government returned Taiwan to the
then Chinese government that was under the rule of the Nationalist Party (Kuomintang,
KMT). In the same year, the National Language Movement that began in 1911 in China
aiming to propagate and promote Standard Mandarin was brought to Taiwan1. When the
National Language Movement was transplanted to Taiwan, the lingua franca in Taiwanese
society was either a native language2 of the Taiwanese people (in private domains) or
Japanese (in public domains) and if without a lingua franca, the communication between
Taiwanese people and the mainlanders relied on translation3. Tsao (1999a:340) describes the
situation at the time:
‘in the Taiwanese society, there existed a ruling class of Mainlanders, most of whom
could speak some form of Mandarin and a lower class of people comprising Southern
Min, Hakka, and Austro-Polynesian speakers, and there was no way for these groups to
communicate with each other except through translation. The situation was extremely
delicate and needed to be handled with care.’
With the aim of eradicating Japanese, during the early stages4 of the promotion of
Standard Mandarin in Taiwan, linguistic studies on Taiwanese dialects5 were greatly
1 Tsao (1999a:339).
2 Native languages of the Taiwanese people included ‘Southern Min, Hakka, and Austro-Polynesian’ languages
3 Tsao (1999a:340).
4 According to Tsao (1999a:342), the provincial level promotion committee was abolished in 1959 and replaced
by a low-level committee under the Ministry of Education after ‘the Taiwan Provincial Committee of
Propagation and Promotion of the National Language’ (CPPNL) felt that they had achieved the objective set to
propagate and promote Guoyu. The early stages of the promotion of Guoyu thus covered the period 1946-1959.
encouraged6 and one of the criteria7 in the selection of officers of the Guoyu Promotion
Committee8 was knowledge9 of Southern Min, since the committee hoped that officers with
Southern Min expertise could teach Standard Mandarin in Southern Min10. As such, by 1946,
about 173 teachers were recruited from Fujian Xiamen and Guangdong Chaozhou in China
(mainly Xiamen), where Southern Min was spoken, and 31 teachers11 were recruited from
Sichuan Chongqing, Gansu Lanzhou, and Shanghai12, where varieties of Standard
Mandarin13 were spoken.
Following the recommended teaching methodology - using Southern Min as the medium
for teaching Standard Mandarin - a series of books was published on how to use Southern
Min to learn Standard Mandarin14. These publications were characterised by studies of
(Xiamen) Southern Min15 or Standard Mandarin and contrastive studies of the two in aspects
5 The use of the term ‘dialects’ (‘Fangyan’ in Guoyu), in Taiwan, varies depending on the context where it is
used. Dialects can be used to refer to the regional varieties of China (in studies of Chinese dialectology), the
local varieties of Taiwan (in studies of Taiwanese dialectology), or the non-standard varieties, as opposed to the
national standard variety, Guoyu. The thesis follows the use of the term in studies of both Chinese and
Taiwanese dialectology. Taiwanese dialects here refer to Southern Min and Hakka dialects in Taiwan.
6 Guoyu promotion officials assumed Southern Min dialects were widely spoken in Taiwan (B.Y. Zhang
7 See B.Y. Zhang (1974:132) for other criteria.
8 The Guoyu Promotion Committee covers both ‘the Taiwan Provincial CPPNL’, established in April 1946,
and ‘a lower-level committee under the Taiwan Provincial Department of Education’ that replaced the Taiwan
Provincial CPPNL in 1959 (Tsao 1999a:340).
9 Knowledge of Southern Min: ‘could speak and understand Southern Min or were capable of conducting
research on dialect pronunciation and grammar’ (B.Y. Zhang 1974:132).
10 This was based on the assumption that ‘Southern Min was still commonly used in all walks of life in
Taiwanese society and thus could be used to replace Japanese before Taiwanese local residents learned Guoyu,
though later, the Guoyu Promotion Committee found that it was necessary to revive Southern Min [from low
social status during the Japanese rule]’ (B.Y. Zhang 1974:129).
11 See B.Y. Zhang (1974:135-136) for the number of teachers recruited from individual places, e.g. Sichuan
(23), Gansu (6), and Shanghai (2).
12 The reason why the teachers had to come from the Chinese mainland was that it was difficult to find trained
Guoyu teachers in Taiwan at the time and promoting Guoyu was an urgent task (B.Y. Zhang 1974:135-136).
13 “Varieties of Standard Mandarin” were commonly referred to as Lanqing Guanhua (lit. ‘blue-green official
language’) ‘mixed official language’ or ‘non-standard Mandarin’, e.g. ‘Jiangsu residents speak Jiangsu Guoyu,
Sichuan residents speak Sichuan Guoyu; actually, they were all Lanqing Guanhua but not standard Guoyu’
(B.Y. Zhang 1974:45). Here, varieties of Standard Mandarin include Sichuan, Gansu, and Shanghai varieties of
Standard Mandarin. The dialectology literature tells us, however, the so-called non-standard varieties of
Mandarin were in fact native languages of speakers from Gansu, Jiangsu, and Sichuan as well as second
dialects of speakers from Shanghai (cf. Norman 1988; Yuan 1989; Zhan 1991).
14 B.Y. Zhang (1974:134). B.Y. Zhang (1974:88) listed fifteen books published between 1946 and 1955 that
reported the experimental results of teaching Guoyu in Southern Min.
15 B.Y. Zhang (1974:88) listed twelve publications written by five authors. Seven of the publications were
written by Zhu Zhao Xiang, a native Xiamen Southern Min speaker, who used the Xiamen sound system in
China as the representative variety of Southern Min in his publications, regardless of variation within Southern
Min spoken in Taiwan (Quanzhou, Zhangzhou, and Chaozhou Southern Min). It is unclear whether other
publications also used Xiamen as the Standard variety of Southern Min, though it is likely, as the authors of
these publications were officers of the promotion of Guoyu who were recruited from Fujian, China (B.Y.
Zhang 1974:135), where Xiamen or Amoy has long been taken as the representative variety of the Fujian
dialect. See (R.L. Cheng 1988:104-106) for a number of publications that followed the Xiamen sound system.
of spelling systems (romanisation and phonetic alphabets), pronunciation, vocabulary, and
conversation. The methodology of teaching Standard Mandarin via Southern Min was also
used in broadcasting the model Standard Mandarin pronunciation on Taiwan Radio Station,
where Standard Mandarin textbooks were read out first by a Guoyu promotion officer,
followed by a Southern Min translation16. The contrastive methodology ceased to be
recommended by the Guoyu Promotion Committee towards the later stages17 of the
promotion of Standard Mandarin and the use of dialects in public domains was advised to be
The contrastive methodology (comparing alphabets, pronunciation, vocabulary, etc. of
Southern Min and Standard Mandarin) was later used as a way for teachers to understand
difficulties that students whose native dialect was Southern Min had in learning Standard
Mandarin as well as a way for sociolinguists to understand how Taiwanese Mandarin has
come to its present being since the promotion of Standard Mandarin on the island. As we
will see later, the contrastive analysis established in the field of Mandarin language teaching
had a great impact on the studies of the development of Taiwanese Mandarin, in that by
comparing Standard Mandarin to Southern Min and Taiwanese Mandarin, much of the
existing sociolinguistic literature has claimed that Southern Min is the source of the
characteristics of Taiwanese Mandarin, which can be referred to as the Southern Min-origin
hypothesis. Since the Southern Min population constitutes the majority of the Taiwanese
population, there has been little doubt among sociolinguists about the validity of the
Southern Min-origin hypothesis.
16 B.Y. Zhang (1974:134).
17 Began in 1959 when the Taiwan Provincial CPPNL was abolished and ended in 1987 when Martial Law was
lifted together with the dialect ban during the promotion of Standard Mandarin.
18 The Ministry of Education announced on 30th May, 1956 that ‘Secondary schools should use Guoyu
wherever possible, avoiding the use of dialects’ (B.Y. Zhang 1974:173).
2.3 Studies of Taiwanese Mandarin
In this section, I first review studies of the differences between Standard Mandarin and
Southern Min by researchers who observed that Taiwanese Mandarin is a deviated form of
Standard Mandarin and believed that the Southern Min sound system is the cause of the
deviation. I will then review studies of the development of Taiwanese Mandarin that applied
contrastive analysis as a way to locate the source of the characteristics of Taiwanese
2.3.2 Taiwanese Mandarin as a deviated form of standard Mandarin
Based on the belief that Taiwanese Mandarin is a deviated form of standard Mandarin
and that Southern Min is the cause of such a deviation, a number of studies19 have focussed
on analysing the differences between (Standard20) Southern Min, Standard Mandarin, and
Taiwanese Mandarin with the aim of helping teachers better understand the difficulties
Southern Min learners of Standard Mandarin have or helping foreign students better
understand the differences between Standard Mandarin and Taiwanese Mandarin.
Chien (1971:1), for example, provided ‘a contrastive linguistic analysis of Mandarin
Chinese21 [Standard Mandarin] and Taiwanese [Southern Min]22…in order to help teachers
better understand why students whose native language is Taiwanese [Southern Min] face
problems of linguistic interference in learning the pronunciation of Mandarin and help them
teach the pronunciation of Mandarin more effectively’. Similarly, having recognised that ‘in
Taiwan’s Guoyu teaching environment…among dialects that may influence the learning of
Guoyu, the Southern Min dialect is used most commonly and thus has the greatest influence
19 Chien (1971); Ye (1991, 1995); S.M. Wu (1990); C.S. Tung (1995).
20 E.g. Chien (1971) used the Amoy [Xiamen] dialect as the Standard Southern Min.
21 Chien (1971:3), citing T.H. Tung (1970:15), defined ‘Mandarin Chinese’ as ‘the language spoken mainly in
the city of Peking (Peiping) and its vicinity…[and] is the standard language of China, being based on the
language spoken by the educated Pekinese.’
22 The use of the term ‘Taiwanese’ to refer to Southern Min varieties spoken in Taiwan has gradually become
inadequate through time, as ethnolinguistic groups residing on the island have gradually regarded themselves as
Taiwanese, thus their dialects or languages have become eligible to be named as Taiwanese. Given linguistic
and historical considerations, the thesis uses ‘Southern Min’ to refer to ‘Southern Fujian dialects’ and
‘Southern Mins’ to refer to ‘Southern Fujian immigrants who arrived in Taiwan between the 17th century and
1895’ to reflect historical origins of dialects and people in Taiwan.
on the learning of Standard Mandarin’23, S.M. Wu (1990:71) compared phonemic
differences between Southern Min and Standard Mandarin in a bid to help ‘Guoyu
teachers…understand the knowledge of Southern Min so as to transform the interference of
Southern Min on learning of Guoyu into assistance’24. Ye (1991, 1995), in order to help
foreign students learn Guoyu, described differences between ‘Taipei Guoyu’25 (1991) and
Standard Mandarin as well as between ‘Taiwanised Guoyu’26 and Standard Mandarin (1995)
by contrasting Taipei Mandarin and Taiwanised Mandarin with Standard Mandarin in
pronunciation, grammar, and vocabulary. With the purpose of helping American students
studying Chinese language in Taiwan understand ‘to what degree [the] “Textbook
Mandarin27” taught in America represents Chinese as it is spoken’ in Taiwan, Kubler
(1979:27-28) illustrated, in aspects of ‘pronunciation, grammar, and vocabulary’, ‘some
differences between Taiwan Mandarin and Textbook Mandarin’ by contrasting Taiwan
Mandarin with ‘[Standard] Peking Mandarin28’ ‘as represented in three common Chinese
language textbooks29’ and showed that these differences are ‘some of the changes that have
taken place’ in Taiwanese Mandarin after 1945 ‘because of influence from the native
Southern Min dialect and from Southern Mandarin in general’. Since Chien (1971) provided
the most comprehensive comparison between sound systems of Standard Southern Min and
Standard Mandarin, in what follows, I will first summarise Chien’s study, as in tables 2.1a, b,
c, and d30, followed by a comparison between Chien (1971), Kubler (1979), Ye (1991, 1995),
23 S.M. Wu (1990:71), original text in Chinese, my translation.
24 Original text in Chinese, my translation.
25 Original text in Chinese, my translation.
26 Original text in Chinese, my translation.
27 Mandarin as written in textbooks, representing ‘Peking Mandarin’ (Kubler 1979:28).
28 Standard Peking=Peking Mandarin=Textbook Mandarin=Standard Mandarin, as in Kubler (1979).
29 DeFrancis’ Beginning Chinese  Fenn and Tewkbury’s Speak Mandarin , and Huang and
Stimson’s Spoken Standard Chinese  (Kubler 1979:28).
30 I made tables (2.1a, b, c, and d) based on Chien’s (1971) contrastive analyses. Some notations used in Chien
are modified to keep an overall consistency of notation in the thesis. Columns under ‘Taiwanese Mandarin’ are
based on Chien’s (1971) discussions of difficulties Southern Min learners of Standard Mandarin faced. In the
tables, “+” or “-” refers to the presence or the absence of the standard variant (either Standard Mandarin or
Standard Southern Min) in Taiwanese Mandarin. For example, the phoneme /b/ in Standard Southern Min is
non-existent in both Standard Mandarin and Taiwanese Mandarin, hence the placement of “-” in the box under
Taiwanese Mandarin corresponding to /b/. The note ‘lacking’ indicates that Southern Min lacks the Standard
Mandarin variant in the list. The note ‘unmentioned’ indicates that Chien did not give information as to
whether the phonemes in either Standard Mandarin or Southern Min are also present in Taiwanese Mandarin.
When other variants are used at the expense of the standard variant, Chien often give examples to illustrate
such use; thus, the note ‘no examples’ indicates that Chien did not provide examples. The note ‘no clash’
and S.M. Wu (1990) in aspects of phonology, and a comparison between Kubler (1979)
and Ye (1991, 1995) in aspects of grammar and vocabulary.
indicates that Chien suggests that sometimes, the use of other variants may result in one word having two
pronunciations. For example, Chien reported that in Taiwanese Mandarin, the word <fang fa>: [fa1 fa]
‘method’ is pronounced as [ua1 ua], with no clash with other words. As such, [ua1 ua] could be used
readily as an alternative pronunciation of <fang fa>. The note ‘clash’ indicates that Chien reports that other
times, the use of a pronunciation non-existent in standard Mandarin may result in two words having the same
pronunciation. For example, Chien reported that in Taiwanese Mandarin, the word <ming yu>: [mi1 yu]
‘fame/reputation’ is pronounced as [min ji], which clashes with the word <ming yi> ‘name’ [min ji]. As such,
[min ji] could not be readily used as an alternative pronunciation of the word <ming yu> without causing
Table 2.1a Comparison between Phonemes and Allophones in Standard Southern
Min, Standard Mandarin, and Taiwanese Mandarin
Standard Mandarin Standard Southern Min Taiwanese Mandarin
/S/: [p] /S/: [p] +
/Sh/: [ph] /Sh/: [ph] +
/m/: [m] /b/: [m] +
Lacking /b/: [b] -
/f/: [I] Lacking
/f/: [f] Æ , e.g. <fang fa>: [fa1 fa] ‘method’
Æ [ua1 ua] ‘method’
/t/: [t] /t/: [t] +
/Wh/: [th] /th/: [Wh] +
/n/: [n] /l/: [n]
/n/: [n] Æ /l/: [l], e.g. <ni tai tai>: /ni tai tai/ ‘your wife’
Æ <Li tai tai>: /li tai tai/ ‘Mrs. Lee’
/l/: [l] /l/: [l]
/l/: [l] Æ /n/: [n], e.g. <nao>: /nao/ ‘brain’
Æ <lao>: /lao/ ‘old’
/ts/: [Ws], [tsh], [tà], [tàh] /ts/: [Ws], [tsh], [tà], [tàh] +
/s/: [s], [à] /s/: [s], [à] +
/th/: [th] Lacking
/t, th/Æ/ts, tsh/
<chao ren>: /thau n/ ‘to disturb others’
Æ <cao ren>: /tshau n/ ‘scarecrow’
<zhang ren>: /ta1 n/ ‘father-in-law’
Æ <zang ren>: /tsa1 n/ ‘The Tibetan’
//:  Lacking
<shi lao shi>: / lau / ‘Teacher Shi’
Æ <si lao shi>: /s lau s/ ‘dead teacher’
//:  Lacking +
Lacking /dÞ/: [dÞ] -
Lacking /J/:  -
Lacking //:  -
/k/: [k], [Nh] /k/: [k], [Nh] +
Lacking /J/: [J] -
/x/: [x], [h] /x/: [x], [h] +
Lacking /J/:  -
Lacking //:  -
/k/: [k], [Nh] /k/: [k], [Nh] +
Lacking /J/: [J] -
/x/: [x], [h] /x/: [x], [h] +
Lacking /-S/, /-t/, /-k/:
[p], [t], [k] -
/-m/: [m] /-b/: [m] -
/-n/: [n] /-l/: [n] +
/-1/:  /-J/:  +
Lacking /-/:  -
/-r/:  Lacking -
Lacking /-S/, /-t/, /-k/:
[p], [t], [k] -
/-m/: [m] /-b/: [m] -
/-n/: [n] /-l/: [n] +
/-1/:  /-J/:  +
Lacking /-/:  -
/-r/:  Lacking -
Table 2.1a Comparison between Phonemes and Allophones in Standard Southern
Min, Standard Mandarin, and Taiwanese Mandarin (cont.)
Standard Mandarin Standard Southern Min Taiwanese Mandarin
/a/: [a], [$] /a/: [a], [$]
Further back +
/i/: [i], [,] /i/: [i], [,] Longer vowel length +
/e/: [e], [(] /e/: [e], [(] +
/u/: [u],  /u/: [u], 
Longer vowel length +
/e/: [o] /o/: [o] +
Lacking /o/: [o] -
/e/: , , , [o ] Lacking Unmentioned
/y/Æ/i/, e.g. <ming yu>:
/mi1 yu/ ‘fame’
Æ <min yi>: /mi1 i/
/)/Æ/o/, e.g. <ke ren>:
/kh) n/ ‘guest’
Æ <kuo ren>: /khuo
n/ ‘rich man’
/ai, au/ /ai, au/ +
/ia/ /ia/ +
Lacking /io, iu/ -
Lacking /ui, ue/ -
/ua/ /ua/ +
/ie/ Lacking +
/eu/ or /ou/ Lacking
/eu or ou/Æ/o/
/ue/ or /uo/ Lacking
/ue or /uo/Æ/o/
/ye/ Lacking +
/iau, uai/ /iau, uai/ +
/uei/ Lacking +
/ieu/ or /iou/ Lacking
/ieu or iou/Æ/io/
- /i, a, u, e, o, ia, iu, io, ai, au, ua,
Nasalise all vowels after
nasals, e.g. <man man>:
/man man/: [man man]
Æ [mãn mãn] ‘slowly’
Table 2.1b Comparison between Syllable Structure and Distributions of Phonemes
(Phonotactics) in Standard Southern Min and Standard Mandarin31
Syllable Structure Standard Mandarin Standard Southern Min
31 ‘C’ = Consonant; ‘V’ = Vowel.
Tone, sonority One tone, one peak of sonority. One tone, one peak of sonority.
V /u/ ‘five’ /u/ ‘to have’
VV /u/ ‘I’ /ui/ ‘stomach’
CV /tha/ ‘he’ /ti/ ‘pig’
VVV /iao/ ‘waist’ /iao/ ‘hungry’
VVC /ia1/ ‘a surname’ /uan/ ‘winding’
CVC /t1/ ‘lamp’ /him/ ‘bear’ (animal)
CVV /mau/ ‘cat’ /kau/ ‘monkey’
CVVV /khuai/ ‘quick’ /khiau/ ‘adept’
CVVC /tien/ ‘spot’ /hio1/ ‘fierce’
VC /an/ ‘peace’ /ut/ ‘to iron’
C Absence /1 CC/ ‘yellow’
CC Absence /p1 C/ ‘cooked rice’
Vowel cluster 9 diphthongs, 4 triphthongs 15 diphthongs, 4 triphthongs
‘C’ in CV Any consonants except /1/ Any consonants
‘C’ in VC /-1/, /-n/, /-m/, /-r/ /-p/, /-t/, /-k/, /-m/, /-n/, /-1/, /-/
Vowel Monophthongs, diphthongs, or
Monophthongs, diphthongs, or triphthongs
Table 2.1c Comparison between Suprasegmental Phonemes in Standard Southern
Min and Standard Mandarin
Standard Mandarin Standard Sout hern Min
Normal Stress Syllables that have neither weak nor contrastive stress
/©xao ªu/ ‘good book’
Syllables pronounced in
/ªtài/, /ªn1/, /ªsã/, /ªài/
‘one, two, three, four’
Secondary Stress Sequences of normally stressed syllables without
intermediate pause; the secondary stress falls on the first
syllable of the following example:
/©xao ªu/ ‘good book’
Three-syllable phrase: 1st and
/©pan ©tǊa ªla1/ ‘a lazy man’
Tertiary Stress Three-syllable phrases/compound words: 3rd > 1st > 2nd
syllables. Tertiary stress falls on the intermediate
/©uo ©©mei ªtu1/ ‘I did not understand.’
Contrastive Stress Emphasis.
/ªuo ©©ài1 ªªua1 ©pu ©©ài1 ªªxuang/
‘My surname is Wang, not Huang.’
/©Jua ªbe ªkhi ªTaipak/,
‘I want to go to Taipei’; /©Jua
ªªbe ªkhi ªTaipak/
‘I do want to go to Taipei.’
Weak Stress /©uo ˙t)/ ‘my’ Lack of tone
Plus Juncture /lio/ ‘to flow’ - /li/+/io/ ‘reason’
/àio/ ‘to repair - /ài/+/o/ ‘Western Europe’
Sustained Juncture At the end of non-final clauses.
/àia y t) xuaÆuo tàio pu thu tàhy/
‘If it rains, I won’t go out.’
At the end of non-final
/na ài lo hoÆua do muai khi/
‘If it rains, I won’t go out.’
Rising Juncture 1. Asking a question
/ni pu tàhyu ma®/ ‘Are you not going?’
2. Casting doubt
/ni mei tàhien®/ ‘You don’t have money?’
1. Asking a question ---
2. Casting doubt
/i ài-khi®/ ‘He died!’
Falling Juncture 1. Statement
/ta mei lai¯/ ‘He didn’t come.’
/Ni iao tha xai iao kha fei¯/
‘Which do you prefer, tea or coffee?’
/Jua ài hak-ài1¯/ ‘I am a
/li ai kam-a a si ai kim-tàio¯/
‘Which do you prefer, orange
3. Simple question
/li tsai bo¯/ ‘Do you know?’
High level /55/ -- /fei55/ ‘to fly’ /55/ -- /to55/ ‘a knife’
Mid rising /35/ -- /fei35/ ‘fat’ Lacking
Falling rising /214/ -- /fei214/ ‘bandit’ Lacking
High falling /51/ -- /fei51/ ‘lung’ /51/ -- /to51/ ‘to fall’
Low falling - /21/ -- /to21/ ‘to pour out’
Mid abrupt - /3/ -- /to3/ ‘table’
Rising - /45/ -- /to45/ ‘to escape’
Mid - /33/ -- /to33/ ‘doctrine’
Low abrupt - /2/ -- /to2/ ‘to kindle’
Tone Sandhi +Æ+
/tao214=>35 xuo214/ ‘look for fire’
2Æ1 3 & 4 Æ 2 7 & 8 Æ 3
1 & 5 Æ 7
Table 2.1d Comparison between Morphophonemics in Standard Southern Min and
32 In Southern Min, ‘the initial consonant of a syllable is sometimes dropped in compound morphemes or
characters…occur only in rapid speech’ (Chien 1971:63).
Morphophonemics Standard Mandarin Standard Southern Min
/san/+/ba/ Æ /samba/ ‘foolish’
/àin/+/khu/ Æ /ài1khu/ ‘body’
/tàit/+/pai/ Æ /tàippai/ ‘once’
Loss of Phonemes32 Loss of the final of the root morpheme
/phai/+/r/Æ/phar/ ‘small card’
Loss of initial consonants
/tshut khi/ Æ /tshut i/ ‘go out’
- [-p, -t, -k] + suffix a
Æ [-b, -l, -J]
/ap/+/a/ Æ /aba/ ‘small