The data listed in Tables (58) to (64) below are 71 Southern Min loanwords in Mandarin

(1)

Chapter 5

Southern Min Loanwords in Mandarin

In this chapter, we attempt to investigate Southern Min loanwords in Mandarin

and account for the phonological adaptations from LS to LT through the constraint

ranking that is achieved in Chapter 3. The LS in this chapter is Southern Min, which is

also a dialect of Chinese. In such an inter-dialectal borrowing process, we generalize

four adaptation strategies, in which morphological and phonological factors are

involved. A schema of each is sketched for a better understanding of the borrowing

paths. In the OT analysis, we focus on adaptations through phonological factors. The

segmental alternations of this type, compared with the loanwords from English, are

simpler in terms of changing scale, due to similar sound systems of the two Chinese

dialects.

5.1 Four Paths

Mandarin and Southern Min are the two major dialects used in Taiwan society,

with the former being the official language and the latter the ruling dialect in a

colloquial conversation. Due to intensive contact, word-loaning between the two

(2)

dialects has taken place to a large scale, especially from Southern Min to Mandarin¹.

The data listed in Tables (58) to (64) below are 71 Southern Min loanwords in

Mandarin. A single Southern Min input can have more than one Mandarin output, as

in entries 56, 57, and 66. Loanwords from Southern Min to Mandarin are found to

have richer variations in terms of adaptation strategy, and are divided into four types,

each type featuring a peculiar adaptation path. The data sources include newspapers

(e.g. Apple News), magazines (e.g. Next Magazine), the internet (e.g. BBS stations),

Southern Min dictionaries (e.g. Wei 1992), literatures on Mandarin loanwords (e.g.

Chou 2004), as well as the author’s recording of people’s speech. It should be noted

that all the data are what people really say, excluding the printed forms frequently

appearing in print media, TV news titles, or karaoke subtitles. Some of them are the

Mandarin correspondent morphemes, while even a lot more are simply the substitute

morphemes to represent their Southern Min pronunciations when the actual

correspondent morphemes are unknown. Mandarin speakers hardly say these

substitute words. Examples of this kind include [tai.] (歹勢, ‘sorry’), [mu.] (目屎,

‘gum’), and [min.ti] (免驚, ‘do not be afraid’), to name just a few.

1 Most Taiwanese people speak both Mandarin and Southern Min, and therefore it is not difficult for them to identify which lexical items in Mandarin are borrowed from Southern Min. On the contrary, it is not an easy task to tell which in Southern Min are imported from Mandarin, since when a term which is usually used in Mandarin is pronounced in Southern Min, it will be taken as ‘formal Southern Min’.

For example, Southern Min [twi53 put5 ki53] (對不起, ‘sorry’) is taken as a more formal expression than the common [pai55 se11] (歹勢, ‘sorry’). Besides, even if Mandarin does donate terms to Southern Min, they are simply the outcomes through morphological correspondence, with no phonology involved, and thus they are temporarily not the main issue in our research.

(3)

5.1.1 Path 1: Morphological Correspondence Table (58) Southern Min Loanwords: Path 1

Southern Min Mandarin Gloss

1. [au] (拗²) [au] (拗) to argue speciously; to take advantage of someone

2. [kai] (蓋) [kai] (蓋) very

3. [ba.ka] (肉腳) [ou.tjau] (肉腳) loser

4. [ju.ki] (幼齒) [jou.t] (幼齒) young, minor 5. [kak.tau] (角頭) [tjau.tou] (角頭) boss of a gang 6. [ke.kjam] (加減) [tja.tjen] (加減) more or less

7. [ke.po] (雞婆) [ti.pwo] (雞婆) nosy

8. [ke.sen] (假仙) [tja.jen] (假仙) to fake

9. [ko.ku] (槓龜) [ka.kwei] (槓龜) to lose money in lottery 10. [ku.mo] (龜毛) [kwei.mau] (龜毛) irresolute or too picky 12. [kan.i] (牽絲) [tjen.s] (牽絲) stringy

13. [kan.twa] (牽拖) [tjen.two] (牽拖) to find excuses

14. [kau.pe] (哭爸) [ku.pa] (哭爸) (curse) to speak offensively 15. [ke.hja] (客兄) [k.o] (客兄) the wife’s lover in an

extra-marital affair 16. [lau.kwi] (漏氣) [lou.ti] (漏氣) to discourage

17. [le.tsan] (犁田) [li.tjen] (犁田) to fall off a motorcycle 18. [lja.pau] (掠包) [twa.pau] (抓³包) to catch someone doing

bad things on the spot

19. [p.bak] (白目) [pai.mu] (白目) foolish

20. [pet.po] (撇步) [pje.pu] (撇步) tip of success

21. [ti.ko] (豬哥) [tu.k] (豬哥) horny (male)

22. [to.twa] (倒彈) [tau.tan] (倒彈) to feel disgusted 23. [tau.ke] (頭家) [tou.tja] (頭家) boss

24. [tau.l] (頭路) [tou.lu] (頭路) job

25. [ti.ki] (鐵齒) [tje.t] (鐵齒) stubborn

2 The characters of the Southern Min morphemes in this thesis are looked up in Wei (1992), Yang (1993), and Chen (1994).

3 Though apparently the Mandarin morpheme 抓 is not the same as the Southern Min 掠, this entry is sorted in this type because 抓 is mistaken for the Mandarin morpheme of [lja11] (‘to catch’) with a similar meaning. Several other such cases can occur with reasonable phonetic correspondence. For example ‘吃人’ are taken as the Mandarin morphemes of Southern Min ‘食郎’ in entry 39.

(4)

26. [twa.ta] (脫窗) [two.twa] (脫窗) unable to see clearly 27. [tja.ka] (正港) [t.ka] (正港) genuine

28. [tja.pja] (車拼) [t.pin] (車拼) to severely compete 29. [tsai.te] (在地) [tsai.ti] (在地) native

30. [tsau.pwi] (臭屁) [tou.pi] (臭屁) cocky 31. [wa.ko] (碗糕) [wan.kau] (碗糕) weird thing 32. [ko.la.wan] (顧人怨) [ku.n.en] (顧人怨) annoying

33. [.pe.ko] (烏白講) [xei.pai.tja] (黑白講) to talk incorrectly 34. [so.wai.wai] (爽歪歪) [wa.wai.wai] (爽歪歪) cheerful

35. [tsau.tau.tau] (走透透)

[tsou.tou.tou] (走透透) to walk around thoroughly

36. [tau.tja.po] (偷食步) [tou.t.pu] (偷吃步) to take a snap course 37. [.i.sa] (五四三) [u.s.san] (五四三) nasty remark

38. [sa.put..i] (三不五時)

[san.pu.u. ] (三不五時) every now and then

39. [tia.la.kau.kau] (食郎夠夠)

[t.n.kou.kou] (吃人夠夠)

to hector hard

Mandarin and Southern Min are two dialects in Chinese language family, and

have the same morpheme system. A borrowing process of this type is clarified through

the schema below.

(59) Path 1 (Morphological Correspondence) Entry 23: [tau.ke] → [tou.tja]

I = /tau.ke/, O = [tou.tja]

I LS

O LT

(5)

The schema in (59) reveals two levels in the borrowing process, namely LS and

LT. When a Mandarin speaker borrows a Southern Min word (the input, I), they

recognize its corresponding morphemes in Mandarin (the output, O).

Loanwords through this Path 1 outnumber those through other paths (the

numbers are 39, 22, 5 and 5 from Path 1 to Path 4 respectively). Thus, morphological

correspondence as the repair strategy should take precedence over other

considerations.

Adopting this path, loanword phonology between two dialects would appear

straightforward if the lexicon of each has an equally complete writing system. Not

every Southern Min morpheme, however, has its written character; quite a few are

unknown to the public. What are the strategies for words without characters?

5.1.2 Path 2: Homophone

The writing system of Southern Min has been under reconstruction with the

movement of ‘localization’ in Taiwan in the past few decades. Despite this, still a

number of characters of Southern Min morphemes have not been confirmed.

Moreover, Southern Min itself includes terms of foreign origins, such as Austric,

Miao, and Tai-Kadai in early times, and Austronesian, Japanese, and English in later

(6)

times. When words of such are borrowed into Mandarin, they do not have

corresponding morphemes.

Table (60) Southern Min Loanwords: Path 2

40. [ka] (軋) [ka] (尬) to compete

41. [kwe] (詼) [kwei] (虧、窺) to ridicule

42. [tjau] (撨) [tjau] (喬) to adjust or organize 43. [an.tswa] (按怎) [an⁴.tsn] (安怎) how

44. [au.ke] (?⁵客) [au.k] (奧客) nasty customer

45. [ha.jo] (噓燒) [x⁶a.au] (哈燒) to blow to make cooler 46. [ho.lan] (虎膦) [xu.lan] (唬爛) (curse) to bluff

47. [ho.ka] (好孔) [xau.ka] (好康) advantage

48. [ko.tswi] (婟婎) [ku.twei] (古錐) cute

49. [ka.ju] (揩油) [ka.jou] (卡油) to take advantage of someone

50. [ka.ho] (較/恰好) [ka.xau] (卡好) better

51. [lan.tjau] (膦鳥) [lan.njau] (爛鳥) (curse) penis 52. [pe.lan] (?膦) [pai.lan] (白爛) (curse) moron 53. [sa.jau] (啥潃) [san.jau] (三小) (curse) what 54. [su.pwe] (四配) [su.pei] (速配) well-matched

55. [sut.la] (屑仔) [su.la] (俗辣) coward

56. [tu.lan] (?膦) [tu.lan] (賭爛、杜爛) (curse) to hate 57. [ti.bai] (膣屄) [ti.wai] (雞歪) (curse) cunt

58. [ut.tsut] (鬱悴) [y.tsu] (鬱卒) depressed

59. [se.se.ljam] (詍詍唸) [swei.swei.njen] (碎碎唸)

to grumble endlessly

60. [ts.jau.e] (裝痟?) [twa.jau.wei] (莊孝維)

to play the fool

61. [tja.tja.kun] (沖沖 [tja.tja.kun] (強 mirthfully

4 When a syllable is underlined in the Mandarin output form, it means only this underlined morpheme is used through that path, and the other is simply through Path 1, namely morphological

correspondence.

5 A question mark means this morpheme is still unverifiable.

6 The velar fricative [x] and the glottal fricative [h] have been treated as the allophones of /x/ in Taiwan Mandarin, while we ignore all the [h] alternatives in Mandarin transcriptions.

(7)

滾) 強滾)

The borrowing thus operates on the phonological level. For example, in entry 50

of Type 2, ‘[ka ho] → [ka xau]’, ‘[xau]’ is the Mandarin correspondent of

Southern Min ‘[ho]’ (好, ‘good’), but ‘[ka]’ (卡, ‘to get stuck’, ‘card’, etc.) is not

that of ‘[ka]’ (較 or 恰, comparative marker). How do we have ‘[ka]’ as the

output? The operation is illustrated in (61).

(61) Path 2 (Homophone)

Entry 50: [ka.ho] →[ka.xau]

I = /ka/, R = /ka/, O = [ka]

I LS

R O LT

Not knowing the Mandarin correspondent of Southern Min ‘[ka]’⁷, the speaker

preserves the input’s syllable structure as much as possible, simply deleting the glottal

stop in the coda position. The modified /ka/ may have four possible tones in

Mandarin. The speaker looks for a homophone with the tone that is more commonly

used. Thus the choice is the ‘[ka]’ with the low tone (卡, ‘card’, ‘to get stuck’, etc.).

7 The character form of Southern Min [ka] is still under debate among scholars. Chen (1991) and Yang (1993) use 較 ([tjau]), but Yang claims it can only be used in the comparison of two things, so he created a character for it in other situations; Wei (1992) uses 恰 ([tja]) for it. Apparently the real one is not known to the general public.

(8)

5.1.3 Path 3: Near Homophone

Chinese has a lot of compounds, i.e. words consisting of two or more morphemes

(syllables). When the speaker borrows a Southern Min compound of which the

characters are unknown, and which has no homophone in Mandarin, they are unable

to adopt Path 1 or Path 2. In this case, near homophones will be chosen.

62. [hjau.pai] (囂掰) [jau.pai] (搖擺) arrogant 63. [hjo.hjo] (雄雄) [jo.jo] (熊熊) temporarily 64. [kau.kwai] (狡怪) [kau.kwai] (搞怪) freaky

65. [tin.tsai] (凊采) [tsi.tsai] (青菜) sloppy; at will

66. [ti.bai] (膣屄) [ti.pai] (雞排) (euphemism) genitals

The operation of this type is sketched below.

(63) Path 3 (Near Homophone)

Entry 65: [tin.tsai] → [ti.tsai]

I = /tin.tsai/, R = /tin.tsai/, O = [ti.tsai]

I LS

R O LT

In this example, the speaker does not know the Mandarin correspondent of

(9)

Southern Min [tin.tsai], nor is there such a homophone for the disyllabic

compound as a whole. Therefore, the speaker looks for a near homophone,

[ti.tsai] (青菜, ‘vegetable’), as the output, in which only one segment in the coda

position of the first syllable differs from its source word. When Mandarin [ti.tsai]

refers to the Southern Min loanword [tin.tsai], its original meaning ‘vegetable’ is

replaced by ‘sloppy; at will’.

5.1.4 Path 4: Homophone of the Source

The three paths discussed above have the priorities of adoption within

themselves—morphological correspondence takes precedence; once the strategy of

morphological correspondence fails, the search for homophones or near homophones

in Mandarin is triggered.

67. [ten] (?) [tjen] (電) to defeat

68. [a.pa] (壓霸) [ja.pa] (鴨霸) bully

69. [ka.tai] (佳哉) [tja.tsai] (家在) fortunately

70. [wai.ko] (歪咼) [wai.k] (歪哥) to corrupt

71. [ho.ja.la] (好額郎) [xau.je.n] (好也人) rich people

The fourth path has nothing to do with the above three. When the source word

has no written form, the speaker, instead of matching it with a homophone or near

(10)

homophone in LT, looks for a homophone of it in LS, and takes its Mandarin

correspondent as the output. Thus there are two operations involved in this process.

(65) Path 4 (Homophone of the Source) Entry 69: [ka.tsai] → [tia.tsai]

O = [ka], R = [ka] I = /ka/, O = [tja]

Within Southern Min Loanword grammar

I O R

O

In this path, the Mandarin correspondent of the source word (O), the [ka] in

[ka.tsai] (‘fortunately’), is unknown to the public. The speaker thus finds a

homophone of it in Southern Min, the [ka] as in [ka.ti] (家庭, ‘family’). Through

morphological correspondence (Path 1), the output [tja] is obtained.

5.1.5 Summary

We have suggested four strategies with which a Mandarin speaker borrows

Southern Min words. In Path 1, the output is reached through morphological

correspondence between two dialects. When the character of the Southern Min input

cannot be found or confirmed, the speaker looks for a homophone in Mandarin (Path

(11)

2) or a disyllabic near homophone (Path 3). The speaker may also find a homophone

of the source word in Southern Min, which is then imported to Mandarin through

morphological correspondence the Mandarin output is obtained (Path 4).

5.2 OT Analysis

In this section, we analyze Southern Min loanwords in Mandarin from an OT

perspective. As we have seen in the previous section, the four adaptation strategies are

influenced by morphological (Path 1 and Path 4) and phonological (Path 2 and Path 3)

factors. In the OT analysis, however, we are to deal with data formed simply through

phonological adjustments.

We first take a look at the sound inventories of Southern Min and Mandarin for a

better understanding of the similarities and differences between the sound systems of

the two dialects. With the background knowledge, we then introduce and rank the

relevant markedness and faithfulness constraints. Finally we illustrate the established

constraint ranking with Southern Min loanwords in Mandarin.

5.2.1 Sound Inventories

Southern Min and Mandarin have similar vowel and consonant inventories, with

a few distinctions in place of articulations and segment. Before the OT discussions,

(12)

we need to consider their diversions. The sound inventories of Southern Min and

Mandarin are listed in Tables (66) to (69). The Southern Min vowels and consonants

inventories are based on Chung (1996).

Table (66) Southern Min Vowels (based on Chung 1996)

a. Oral Vowels b. Nasalized Vowels i u i

e o e

 

a a

Table (67) Mandarin Vowels (based on Cheng 1973, also (28) in Chapter 3) i y  u

e    o a

In addition to what listed in the tables, Southern Min has two diphthongs, [ai]

and [au], while Mandarin has four, [ai], [ei], [au], and [ou]. Mandarin has the front

rounded [y], the schwa [], the retroflex vowel [], the back unrounded [], which

Southern Min does not have, but Southern Min has the back lax [], which Mandarin

does not have. Another critical difference is that Southern Min has nasalized vowels,

while Mandarin does not.

Table (68) Southern Min Consonants (based on Chung 1996)

bilabial alveolar alveopalatal velar glottal

(13)

+vcd b 

-asp p t k 

stop

-vcd

+asp p t k

+vcd

fricative

-vcd s  h

-asp ts t

affricate -vcd

+asp ts t

nasal +vcd m n 

liquid +vcd l⁸

glide +vcd j w

Table (69) Mandarin Consonants (based on Cheng 1973, also (29) in Chapter 3)

bilabial labio-dental alveolar alveopalatal retroflex velar +vcd

-asp p t k

stop

-vcd

+asp p t k

+vcd 

fricative

-vcd f s   x

-asp ts t t

affricate -vcd

+asp ts t t

nasal +vcd m n 

liquid +vcd l

glide +vcd j w

Comparing the consonant inventories in Tables (68) and (69), we observe the

following differences. First, in terms of place of articulation, Mandarin has

labio-dental ([f]) and retroflex ([t], [t], [], and []), which Southern Min does not

have, but Southern Min has glottal consonants ([] and [h]), which Mandarin does not

8 The alveolar voiced fricative [z] is listed in Chung’s (1996) inventory. However, it is later stated that the liquid [l] and [z] are no longer distinctive to some people. We adopt [l] only since it is becoming the mainstream variation in Taiwan (in Footnote 8 in Chung 1996, a test further verifies this

observation). Besides, [z] is not found in our data, thus irrelevant.

(14)

have. Second, simply in comparison of segments, Southern Min has voiced stops [b]

and [], while Mandarin does not have voiced stops.

5.2.2 Constraints

Several constraints have been proposed in the precious chapter, and the relevant

ones are reproduced here for convenience. We look at the markedness constraints

first.

(70) CODACONDITION: Syllables must have no coda, except that the coda is an

alveolar nasal, a velar nasal, or a retroflex in the [] syllable. (also (32) in

Chapter 4)

CODACONDITION is crucial when the loanword input is Southern Min. Besides

alveolar and velar nasals, Southern Min also allows the bilabial nasal [m] and

unreleased stops [p t k ] as the coda. *C^OMPLEXCODA is not used in the current

discussion, because Southern Min does not allow consonant clusters to be the onset or

the coda either, except for Cj and Cw in the onset position, and hence

*COMPLEXCODA is irrelevant when the LS is Southern Min.

As mentioned above, Southern Min has sounds that Mandarin does not have, and

(15)

vice versa. Thus we need a constraint that rules out outputs with legitimate segment(s)

in LS which are however illegitimate in the LT. The constraint is formulated below:

(71) POSSIBLESEG: A segment in the output must be a possible segment in LT.

We did not discuss POSSIBLESEG in Chapter 4, where we focus on the coda

adaptation of English loanwords. However, it should be noted that such a phonotactic

constraint is still undominant in the ranking in English loanwords as well.

The other type of constraints is the faithfulness constraints, as reproduced in

(72-75):

(72) MAX-IO(S): Every segment in the input must have a correspondent in the output.

(No deletion.) (also (33) in Chapter 4)

(73) DEP-IO(S): Every segment in the output must have a correspondent in the input.

(No insertion.) (also (34) in Chapter 4)

(74) IDENT-IO(F): Correspondent segment in the input and the output must be

identical in feature. (also (35) in Chapter 4)

(75) MAX-IO([+nasal]): A nasal in the input must have a correspondent in the output.

(No deletion.) (also (48) in Chapter 4)

(16)

In Chapter 4 it is claimed that the consonant type in the coda of an English input

may be determinant in the preservation or deletion of the consonant. For example,

fricatives, affricates, and nasals are salient in perception and tend to be preserved,

while retroflexes are perceptually less salient when it follows a nuclear vowel and

thus are apt to be unparsed. We adopt MAX-IO([+nasal]) here but not MAX-IO([+del

rel]) or ¬MAX-IO(r/V_) because an Southern Min input dose not have a fricative,

affricate or retroflex in the coda position either.

It may be questioned again that since Mandarin and Southern Min both allow

nasals to be the coda, MAX-IO(S) alone should be enough to deal with the problem of

nasal preservation; i.e. any nasal in the input should have a correspondent in the

output. It may appear redundant to propose MAX-IO([+nasal]). However, besides

alveolar and velar nasals, Southern Min allows the bilabial nasal [m] to be the coda

and it has nasalized vowels, both of which are against Mandarin phonotactics.

According to the data, the nasal property of a nasalized vowel in the input is found to

surface in the output by the substitution of a nasal consonant (entry 53 in Table (60)).

Thus MAX-IO([+nasal]) is still in play in Southern Min loanwords.

5.2.3 Ranking and Tableaux

(17)

The relevant constraints are ranked below, a basically the same constraint

ranking has been proposed in Chapter 4:

(76) CODACONDITION, POSSIBLESEG >> MAX-IO([+nasal]) >> MAX-IO(S),

DEP-IO(S) >> IDENT-IO(F)

We exemplify the constraint ranking with a couple of cases in loanwords of Type

2, which is largely concerned with phonology. It is noted that most of the Mandarin

forms of Type 2 are the ‘zero adaptations’ of their Southern Min inputs, because these

Southern Min syllable structures are also well-formed in Mandarin. For a clearer

examination of every constraint, we intentionally take data with phonological

adaptations for example.

We first examine the effect of CODACONDITION. Southern Min allows [p t k ]

to be the coda, but they are unreleased in this position. In English, however, [p t k]

can either be aspirated or unreleased in the coda position. For example, in entry 55,

the [t] in [sut la] (屑仔, ‘coward’) is unreleased, while the [t] as in ‘sit down’ can

either be aspirated in slow speech or unreleased in normal or fast speech⁹. Two

possibilities arise when the Sothern Min input has the unreleased [p t k ] as the coda:

9 It may be the reason for the fact that English stop codas are sometimes preserved and sometimes deleted in Mandarin adaptations.

(18)

the Mandarin speaker may be able to perceive the unreleased stops, or completely lose

it. Let us test the first possibility. Note again that a constraint can be ignored if it is

irrelevant to that case.

Tableau (77)

Input: /su□/ 屑, ‘small pieces’

Output: [su] 俗, ‘cheap’

CODACON MAX-IO(S) DEP-IO(S)

a. su□ *!

)b. su *

())c. su.ta *

The empty square stands for the unreleased stop, which may be a short period of

silence in the Mandarin speaker’s perception. In (77), candidate (a) violates

CODACONDITION with the illicit coda of the unreleased stop and is ruled out.

Candidates (b) and (c) each incurs one violation of MAX-IO(S) or DEP-IO(S), and

thus both are winners, though the real output is candidate (b). It is an outcome of

random selection.

The second possibility is that Mandarin speakers do not perceive the unreleased

stop, which is tested below:

Tableau (78)

(19)

Input: /su/ 屑, ‘small pieces’

Output: [su] 俗, ‘cheap’

CODACON MAX-IO(S) DEP-IO(S) )a. su

b. su.ta *!*

In (78), candidate (b) violates DEP-IO(S) twice and candidate (a) does not incur

any violation. Therefore candidate (a) is the winner.

According to the data, Southern Min unreleased codas [p t k ] are all deleted in

the Mandarin adaptations, the other instances of which include ‘[ka] (較 or 恰,

comparative marker) → [ka] (卡, ‘card’, ‘to get stuck’, etc.) and ‘[tsut] (悴,

‘depressed’) → [tsu] (卒, ‘soldier’). The adaptation tendency compels us to adopt the

second possibility, in which the speaker do not perceive the unreleased stops. Owing

to the limited data, however, further experiments are needed to examine such an

assumption.

The case below illustrates the priority of a legitimate segment in the output:

Tableau (79)

Input: [bai] 屄, ‘cunt’

Output: [wai] 歪, ‘not straight’

POSSIBLESEG MAX-IO(S) DEP-IO(S) IDENT-IO(F)

a. bai *!

)b. wai *[+dorsal]

In (79), candidate (a) is eliminated for the illicit voiced stop [b]. Candidate (b)

(20)

wins despite a violation of IDENT-IO(F) for [+dorsal].

Let us turn to the issue of nasality preservation.

Tableau (80)

Input: [sa] 啥, ‘what’

Output: [san] 三, ‘three’

POSSIBLESEG MAX-IO([+nasal]) MAX-IO(S) DEP-IO(S) IDENT-IO(F)

a. sa *!

)b. san * *[-nasal]

c. sa *!

In (80), candidate (a) incurs a fatal violation of POSSIBLESEG for the illicit

nasalized vowel. Candidate (c) is ruled out by MAX-IO([+nasal]), since the nasal

value of the input is not preserved. Candidate (b) is the optimal output, though it

violates the lower DEP-IO(S) for the insertion of [n] and I^DENT-IO(F) for the [-nasal]

of [a].

5.2.4 Summary

In this section, we analyze the Southern Min loanwords in Mandarin from the

OT perspective. The data being discussed are limited to loanwords of Type 2, namely

loanwords formed through phonological adaptation. Before the OT discussion, the

sound inventories of both Southern Min and Mandarin are briefly introduced for the

(21)

purpose of comparing the differences in sound patterns between the two dialects. In

the constraint-based analyses, the output structure is constrained by Mandarin

phonotactics. For example, it bans voiced stops and nasalized vowels, and the coda

can only be an alveolar or velar nasal. These syllable structure constraints, i.e.

POSSIBLESEG and CODACONDITION, rank in the topmost position. Within the

faithfulness constraints, in addition to the basic faithfulness constraints on segments,

MAX-IO([+nasal]) requires that the nasal feature be preserved in the output.

5.3 Conclusion

In this chapter, we look into the loaning between two related dialects. We find

that word-loaning in such a situation has a richer variation in terms of adaptation

strategies than word-loaning between two unrelated languages.

In the first half of this chapter, we unravel four paths through which the

Mandarin speakers borrow Southern Min terms and make proper adaptations.

Southern Min loanwords through Path 1 are simply the different pronunciations of

their corresponding morphemes in Mandarin. Loanwords through Path 2 are the

homophones of Southern Min inputs in Mandarin when morphological

correspondence (Path 1) cannot be achieved. Path 3 is related to both phonology and

morphology; in addition to phonetic similarity, the output must be an existent

(22)

compound word in Mandarin lexicon. Path 4 is also related to both phonology and

morphology, but is completely independent from the other three paths and involves

two separate operations: the input is first substituted by a homophone in Southern Min.

Then the substitution is imported into Mandarin, through Path 1 its corresponding

morpheme is found to be the output.

In the second half of this chapter, we lay the focus on the data of Type 2 and

analyze them within OT framework. As the previous literatures propose, markedness

constraints outrank faithfulness constraints in loanword phonology, which still holds

for the inter-dialectal loaning process. Different from our discussion of the coda

adaptation of English loanwords, the only relevant markedness constraint in Southern

Min loanwords is CODACONDITION, since Southern Min does not allow complex

coda either. Apart from the basic faithfulness constraints, only MAX-IO([+nasal]) is

relevant in Southern Min loanword, since it has nasalized vowels and it allows the

bilabial nasal to be the coda, both of which are nonetheless unacceptable in Mandarin.