Chapter 3 Corpus and Generalization of Loanword Substitutions
3.2 Segmental Substitutions
3.2.4 Coda Consonants
3.2.4.1 Obstruents
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
become aspirated (making up 17.39% of all occurrences) in Taiwanese.
As for the segments that are in the word-medial position, the distributions are quite uneven. The chance for the Japanese labial [p] to be aspirated (about 57.43%) in the word-medial position is slightly higher than for it to be unaspirated (about
42.57%). As for Japanese alveolar [t], it remains unaspirated mostly, accounting for 91.16% of all occurrences. Regarding the Japanese velar [k], it is more likely to be aspirated than to be unaspirated in the word-medial position. The percentage of the occurrences of aspirated [kʰ] is 83.41%, which is higher than the 16.59% of the occurrences of unaspirated [k]. For Japanese voiceless affricates [ʧ] / [ʦ], nearly all of them are unaspirated, making up 98.61% of all.
Finally, though not consistently, aspiration is likely to be assigned to a
word-medial voiceless non-continuant obstruent if there is an aspirated counterpart in the previous syllable, as in thathamih [tʰatʰamiʔ] ‘straw mattress’. Based on the corpus data, this accounts for 63.64% of all occurrences.
3.2.4. Coda Consonants
Consonants that are allowed in the coda position are limited in both Japanese and Taiwanese. This section discusses how Japanese coda consonants are substituted by Taiwanese segments.
3.2.4.1. Obstruents
In Japanese, obstruents such as the stops [p], [t], and [k] and the fricative [s] can occupy the coda position. However, the four segments show different substitution processes in Taiwanese. First, consider the substitutions of stops. The stops will remain unchanged, as shown in (53).
‧
As illustrated in (55), almost all of the Japanese stops [p] and [t] are borrowed without any adjustment, and account for 85.71% and 93.62% of all occurrences, respectively. As for the stop [k], it corresponds to the same segment in Taiwanese.
Normally, Japanese coda stops will remain intact in Taiwanese, since those segments are allowed in the coda position in Taiwanese. In addition, as in Japanese, the place of articulation of the word-medial coda stops usually will be the same with their following consonant in Taiwanese. However, counter examples are observed in the present corpus. As shown in (54a) and (54b), the labial [p] that is followed by the same segment changes to [t] and [k], respectively. In addition, alveolar [t] that is followed by the same segment changes to [k], as in (54c), or is deleted, as in (54d).
(54) Exceptions of coda stops
a. chippu [ʧip.pɯ] → tsitpuh [ʦit.puʔ] ‘tip’
b. koppu [kop.pɯ] → khokpuh [k*ok.puʔ] ‘cup’
c. torikotto [to.ɾi.kot.to] → thoolikhootooh [t*o.li.k*o.toʔ] ‘tricot’
d. torotto [to.ɾot.to] → tolokto [to.lok.to] ‘foxtrot’
The change of the coda segments, according to Chang (1993), is due to the possibility of the occurrence of certain rime patterns in Taiwanese. He suggested that rime patterns such as im, ip, om, and op are occasionally substituted by in, it, ong, and ok, respectively since the former are less common than the latter.
As for the *ot sequence in (54c) and (54d), it is not a legal CV syllable in
Taiwanese, and adjustment will occur to prevent the formation of illegal structures. To repair such illegitimate syllables, the coda [t] is either deleted or substituted by [k].
‧
(55) Statistics of stop substitutions Japanese
Let us now consider the substitution of the fricatives [s] and [ʃ]. The fricative is not a licit coda in Taiwanese. There are two strategies that Taiwanese speakers use to repair the problematic coda [s] and [ʃ]: a) consonant deletion, as in (56a) and (56b); b) segment substitution, as in (56c) and (56d).
(56) Fricative [s] substitution
a. massaaji [mãs.sa:.ʤi] → masajih [mã.sa.ziʔ] ‘massage’
b. komisshoɴ [ko.miʃ.ʃõɴ] → khoomisiong [k*o.mi.sjoŋ] ‘commission’
c. bussouge [bɯ
s.so:.
ɡe] → butsooge [but.so.ɡe] ‘hibiscus’d. kusshoɴ [kɯʃ.ʃõɴ] → khutsiong [k*ut.sjoŋ] ‘cushion’
Since both of the fricatives [s] and [ʃ] that are in the coda position are perceived as [s] by Taiwanese speakers, we combine the numbers and percentages of the [s] and [ʃ] in (57).
(57) Statistics of [s] substitution Japanese
As can be observed, the percentages of the two strategies in substituting coda
‧
fricatives are very close, and they sometimes do not contrast with each other. For example, [as.sa.ɾi] ‘frankly’ can become [a.sa.liʔ] as well as [at.sa.liʔ] in Taiwanese.
The two substitutes are in free variation.
3.2.4.2. Nasals
The coda nasal is realized as uvular [ɴ] in the word-final position. Such uvular nasal, which does not exist in Taiwanese, will become a labial [m], an alveolar nasal [n], or a velar nasal [ŋ], as illustrated in (58).
(58) Dominant substitutes of word-final [ɴ]
a. kiriɴ [kiɾĩɴ] → khilin [k*ilin] ‘giraffe’
b. oobuɴ [o:bɯ̃ɴ] → oobun [obun] ‘oven’
c. keireɴ [ke:ɾẽɴ] → khelian [k*elen] ‘convulsion’
d. kyahaɴ [k,ahãɴ] → khiahang [k*jahaŋ] ‘leggings’
e. kaatoɴ [ka:tõɴ] → khatong [k*atoŋ] ‘carton’
The substituted nasal of the word-final uvular [ɴ] is determined by the preceding vowel. The uvular nasal [ɴ] that follows the vowels [i], [u], or [e] will be substituted by the alveolar nasal [n], as in (58a) to (58c). No exceptions are found in the present corpus. In addition, as shown in (58d) and (58e), the uvular nasal [ɴ] will change to a velar nasal [ŋ] mostly when it is preceded by the vowels [a] or [o], making up 97.06%
and 97.37% of all occurrences, respectively.
(59) Minor substitutes of word-final [ɴ]
a. napukiɴ [nãpkĩɴ] → napukhing [nãpuk*iŋ] ‘napkin’
b. kaɴ [kãɴ] → kam [kam] ‘warship’
As shown in (59), the velar [ŋ] can also substitute for the uvular nasal [ɴ] that follows the vowel [i], and the labial [m] can also substitute for the nasal [ɴ] that
‧
follows the vowel [a]. However, the two substitutions are extremely rare (only one case of each), as shown in (60).
(60) Statistics of [ɴ] substitution Japanese
source
Taiwanese
loans Number Total Percentage
Vɴ a ŋ 33 34 97.06%
Next, consider Japanese word-medial nasals. In Japanese, the realization of a word-medial coda nasal depends on the consonant that immediately follows it.
Likewise, a word non-final nasal in Taiwanese will become phonetically similar to its following consonant in the place of articulation. Since Japanese and Taiwanese both have similar phenomena of coarticulation, Japanese word-medial nasals remain unchanged in Taiwanese. However, substitutions are still observed.
First, consider the labial nasal [m]. Although the labial [m] is allowed in the coda position in Taiwanese, and Japanese and Taiwanese both have the same assimilation phenomenon, the coda [m] seems uncommon in loanwords. The substitution of the Japanese labial [m] after the vowels [i], [e], [u], and [o] is very similar to that of the Japanese uvular [ɴ], as shown in (61).
(61) Dominant substitutes of word-medial [m]
a. pimpoɴ [pĩmpõɴ] → phinphong [p*inp*oŋ] ‘ping-pong’
b. rumba [ɾɯ̃
mba] → lunba
[lunba] ‘rumba’c. sembei [sẽmbe:] → sianbe [senbe] ‘a rice cracker’
d. kombu [kõmbɯ] → khongbu [k*oŋbu] ‘kelp’
‧
In (61a) to (61c), the [m] to [n] change occurs when the preceding vowels are [i], [u], or [e]. In (61d), a velar [ŋ] substitutes for a labial [m] after the vowel [o]. No
exceptions are found in the present corpus, as shown in (63).
However, the labial [m] display variations in substitution after the vowel [a]. The nasals [n], [ŋ] and [m] all can be substitutes of the Japanese [m], as in (62).
(62) Word-medial [m] after vowel [a]
a. kampyou [kãmp,o:] → khanpio [k*anpjə] ‘dried gourd’
b. shampuu [ʃãmpɯ:] → siangphu [sjaŋp*u] ‘shampoo’
c. rampu [ɾãmpɯ] → lamphu [lamp*u] ‘lamp’
As shown in (63), the alveolar [n] is the dominant substitute of the word-medial [m]. About half of the occurrences of the Japanese coda [m] become alveolar [n].
About 33.33% of them will change to the velar [ŋ]. Only 16.67% remain intact.
(63) Statistics of [m] substitution Japanese
source Taiwanese
loans Number Total Percentage Vm a n 12 24 50.00%
Next, consider the substitution of the alveolar [n]. Like the uvular [ɴ] and the labial [m], the alveolar [n] occurs after the vowels [i], [u], and [e], as in (64). No exceptions are found in the present corpus.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
(64) Word-medial [n] after vowels [i], [u], and [e]
a. hinto [hĩnto] → hintoo [hinto] ‘hint’
b. unchaɴ [ɯ̃
n
ʧãɴ] → untsiang [unʦjaŋ] ‘driver’c. bentou [bẽnto:] → biantoo [bento] ‘lunch box’
In addition, the alveolar [n] will either remain unchanged or become the velar [ŋ]
after the vowel [a], as in (65). Statistics in (67) show that the percentage of the occurrences of the substitute [n] is higher than that of the substitute [ŋ]. This shows that the assimilation rule is still obeyed in most of the Japanese loanwords. It also suggests that the loans will be as similar to their sources as possible.
(65) Word-medial [n] after vowel [a]
a. kanji [kãnʤi] → khanji [k*anzi] ‘feeling’
b. ranjerii [ɾãnʤeɾi:] → langjelih [laŋzeliʔ] ‘lingerie’
Loanwords may sometimes violate the phonotatics of the native language, as illustrated in (66).
(66) Word-medial [n] after vowel [o]
a. hontou [hõnto:] → hontoo [honto] ‘sure’
b. pondo [põndo] → phongloo [p*oŋlo] ‘pound’
Taiwanese does not have an *[on] structure in the rime, at the same time, however;
[oŋ] is a legal rime structure. As in (66), the alveolar [n] will either remain unchanged or become the velar [ŋ] when following the vowel [o], which results in the illegitimate rime *[on]. As shown in (67), the percentage of occurrences of *[on] (about 57.14%) is slightly higher than that of the [oŋ] (about 42.86%). This further confirms the significance of the similarity between the loans and their sources.
‧
(67) Statistics of [n] substitution Japanese
source
Taiwanese
loans Number Total Percentage
Vn a n 33 35 94.29%
Finally, consider the substitution of the velar [ŋ]. All of occurrences of the velar [ŋ] remain unchanged when it is preceded by vowels [a] and [o], as in (68a) and (68b).
In addition, the preceding vowels still have an influence on the substitution. As in (68c) and (68d), the velar [ŋ] changes to the alveolar [n] when it occurs after the vowels [i] and [e].
(68) Word-medial [ŋ] substitution
a. dangou [dãŋɡo:] → langgoo [laŋɡo] ‘conference’
b. sukonku [sɯkõŋkɯ] → kongku [koŋku] ‘skunk’
c. ringo [ɾĩŋɡo] → lingoo [linɡo] ‘apple’
d. enka [ẽŋka] → ianka [enka] ‘traditional Japanese song’
e. ringu [ɾĩŋɡɯ] → lingguh [liŋɡuʔ] ‘ring’
The [ŋ] to [n] change does not necessarily happen, as in (68e). The Japanese word-medial [ŋ] remains intact even if it follows a front vowel [i]. However, this is the only case in the present corpus, as shown in (69). In addition, both [liŋɡuʔ] and [linɡuʔ] ‘ring’ are both observed in the corpus.
‧
(69) Statistics of [ŋ] substitution Japanese
source
Taiwanese
loans Number Total Percentage
Vŋ a ŋ 21 21 100%
i n 23 24 95.83%
ŋ 1 4.17%
e n 3 3 100%
o ŋ 8 8 100%
In sum, the preceding vowel has a significant influence on the substitution of both the word-final and word-medial coda nasals. Typically, the alveolar nasal [n]
appears after the vowels [i], [e], and [u], and the velar [ŋ] appears after the vowels [o]
and [a]. In addition, the onset segment of the following syllable needs to be taken into account when substituting coda nasals after the vowel [a]. The alveolar [n] is the substitute before an alveolar and a labial nasal onset, while the velar [ŋ] is the
substitute before a velar onset. As for the word-final nasal that follows the vowel [a], it is mostly substituted by the velar [ŋ]. The appearance of the *[on] structure suggests that the degree of similarity between sources and loans is crucial.