Chapter 2 Literature Review
2.4 Research on French Loanwords
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
21
position. As for sonorants and glides, they tend to be faithfully adapted. Deviations of features such as [lateral] and [sonorant] are found. In sum, Miao suggests that perceptual similarity is the key factor of phoneme substitutions.
Wu (2008) discusses the phonological adaptations of English loanwords in Mandarin. The focus is on consonant substitutions. The analysis shows that the adaptation patterns are influenced by the positions of consonant segments.
Consonants in word-initial and word-medial positions tend to be retained or adapted.
On the contrary, consonants in word-final position are mostly deleted. In terms of perceptual saliency, onset consonants are more prominent than the coda counterparts.
Therefore, onset segments tend to be preserved. Some salient segments like [s] are bound to be preserved in both onset and coda positions. Similar to Miao’s (2005) analysis, Wu stresses the perceptual factors in loanword adaptation.
2.4 Research on French Loanwords
This section reviews some studies of French loanwords in different languages, including German, Japanese, and Fon.
Itô and Mester (2001) propose a type of stratum-specific faithfulness constraints to account for French loanwords in German. They argue that markedness constraints are fixed in the ranking hierarchy. By putting faithfulness constraints in different places of the ranking hierarchy, sublexicons (i.e. variations) appear. These sublexicons differ in the degree of nativization. For example, both [ʒ] and nasal vowels are absent in German inventory. Long lax vowels and coda [ʀ] segment are also non-native in German. When the French word ‘jongleur’ [ʒɔ̃glœ:ʀ] (juggler) is borrowed, there are only five forms allowed. The reason is that the fixed ranking
*Ṽ>>*LaxLong>>*Coda [ʀ]>>*[ʒ] could only provide five places for inserting the
‧
faithfulness constraints5. Through the insertion of faithfulness constraints, loanword variations are accounted for in terms of nativization degree and entailment relationship.
Shinohara (2001) argues that the emergent patterns in loanword adaptation reflect universal grammar. By examining the loanwords, one could observe some aspects of UG. For example, in Japanese, segments [t] and [d] tend to become affricates when they precede high vowels, such as /tuki/ [tsuki] (moon) and /kaNduki/ [kandzuki] (cold mouth). In Japanese, voiced affricates [dz] are dispreferred. Shinohara proposes the ranking *[dz]>>*[ts] in the native lexicon6. Such ranking explains the dispreference of affricates, and voiceless affricates are more favored than the voiced counterparts. The ranking is also active when French loanwords enter Japanese. For example, the loanword ‘Toulouse’ [tuluz] (place name) is substituted by [tsuuruuzu]. ‘Pompidou’ [pɔ̃pidu] (person name) is adapted as [poNpidoo] rather than [poNpidzu]. The native lexicon and the loanwords both exhibit the markedness of voiced affricates, which is shown in the ranking
*[dz]>>*[ts]. Shinohara shows that a part of Universal Grammar is observed in loanword adaptations. The constraint rankings are likely to be ‘inherited from the initial state of UG and remain latent in L1’.
Kenstowicz (2003) provides a perceptual account of French loanwords in Fon, a language spoken in West Africa. In Kenstowicz’s analysis, an adapted form is chosen if it has minimal change in perception. For example, segments [m] and [b] are in complementary distribution in Fon, and nasal vowel [ĩ] is present in Fon inventory.
5 Each position of the faithfulness constraints is indicated by a vertical line:
|*Ṽ >>|*LaxLong>>|*Coda [ʀ]>>|*[ʒ] |). There are five vertical lines in total. Hence five variants occur.
6 The constraint *[dz] show that voiced affricates are disfavored. ‘*[ts]’ means that voiceless affricates are dispreferred.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
23
When the French word ‘ministre’ [ministʁ] (minister) is borrowed, the first syllable is adapted as [mĩ] instead of [bi]. Kenstowicz suggests that changing the nasality of consonant is perceptually more salient than the alternation of vowel nasality.
Therefore, the form [bi] is disfavored and [mĩ] is chosen. Apart from the restrictions of native phonology, the principle ‘Minimal Saliency’ is important when French words enter Fon phonology.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
24
Chapter 3 Corpus
This chapter builds a corpus of French loanwords in Mandarin. There are 719 words in the corpus. Based on the data, segmental substitution of each French loanword is calculated. This statistical analysis provides the tendency of segment substitutions. Adaptation patterns can be derived based on this evidence.
3.1 The Corpus
This thesis focuses on the discussions of consonant substitutions and nasal vowel adaptations. The French loanwords were collected from a loanword dictionary (A Loanword Dictionary published by Guo Yu Ri Bao, 國語日報外來語辭典), academic journals, France-related books, and websites containing information of tourism in France. The loanwords were transcribed with the help of two native speakers of French, who are experienced teachers of French in Taiwan. The transcription follows the IPA system. Adaptation variants were obtained through Google. The corpus contains 719 loanwords.
If a loanword has two variants, the shared adapted form between them is only counted once. Take the substitutes of ‘Rennes’ ([ʁɛn]) as an example. It has two parts for loanword adaptation, the onset segment [ʁ] and the coda [n]. There are two substitutes in Mandarin: ‘雷恩’ ([lej.ən]) and ‘漢恩’ ([xan.ən]). Since the two variants both adapt the nasal coda [n] as [ən], the preservation of nasal coda is only counted once. However, the onset segment [ʁ] has two variations [lej] and [xan]. Therefore, two different types of adaptations are counted.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
25
Several factors lead to the adaptations of French loanwords. Because this thesis only deals with phonologically adapted forms, loanwords influenced by non-phonological factors will not be analyzed. For example, orthography plays a part in translating ‘Aisne’ ([ɛn]). The Mandarin adaptation is 埃斯內 ([aj.sɨ.nej]). The fricative segment [s] is absent in the input form but present in the adaptation. The [sɨ]
adaptation is likely to be influenced by orthography. Syllable like this will not be included in the corpus.
Variant forms are common in the present corpus. Sometimes the adaptation percentages between dominant outputs and deviations are very close. However, the uneven distribution between the dominant output and the other possible outputs are also observed. This raises a question: Why is one variant form counted as a possible output, while another variant form is left out? In general, a variant form is counted as a possible output if its percentage in the corpus is close to that of the dominant output.
For example, the foreign onset [p] has two adaptations [ph] and [p]. The former occupies 57.65% of the data, and the latter accounts for 42.35% in the corpus. There is no huge discrepancy between the two percentages, thus both variant forms are possible outputs. If one variant form presents a great percentage gap between itself and the dominant form, it could still be counted as a possible output if it occurs in a different phonological environment from the dominant output. For example, the foreign onset [g] has two adaptations [k] and [tɕ]. The former occupies 64.58% in the corpus, while the latter accounts for 31.25% of the data. Both forms are possible outputs because [tɕ] substitutes for syllables [ga], [gi], and [gy], whereas the adaptation [k] replaces other syllable types.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
26
3.2 Statistical Analysis-Onset Adaptations
This section discusses the adaptation of French onsets in Mandarin. The onset segments in the corpus include consonants [p, t, k, b, d, g, s, f, ʃ, z, v, ʒ, m, n, ɲ, l, ʁ].
Onset segments that constitute the major portion of the data will be presented first, followed by the segments of fewer occurrences.
3.2.1 Alveo-Dental Lateral [l] and Uvular Fricative [ʁ]
Segments [l] and [ʁ] are of the majority in the corpus. As (1) shows, [l] segments are faithfully adapted most of the time (95%). Onset deletion or other adaptations are relatively uncommon, only occupying 5% in the data. Mandarin speakers tend to replace [ʁ] with [l] or [x] (59.69% and 24.69%, respectively). Occasionally, Mandarin speakers delete an onset [ʁ] segment. The fricative [ʁ] may be replaced by [ɕ], [ʐ], or other segments. The substitutions of [ɕ], [ʐ], and other segments are very limited in the corpus. They are considered the minority and excluded from the discussion.
Examples are provided in (2).
‧
(1) Statistics of lateral substitutions7 French onset
(2) Examples of [l, ʁ] substitutions10
a. Louvre [luvʁ] 羅浮 [lwɔ.fu] ‘name of a museum’
Three onset segments [p, t, k] are discussed in this part. Mandarin has two kinds of voiceless stops, the unaspirated stops and the aspirated counterparts. In (3), the aspirated stops are favored over the unaspirated counterparts. In the case of [p]
substitution, [pʰ] and [p] are the only possibilities. The most frequent substitutions of [t] are [tʰ] and [t]. Other adaptations are also possible. Mandarin speakers may substitute [t] with segments [tsʰ], [tʂʰ], [s], [ɕ], [tɕ], or even deletion. Yet the chances
7 As the adaptations of simplex and complex onsets are similar, there is no category for complex onsets.
8 ‘Others’ refers to tokens numbering less than three, including [i, ɥ, n].
9 ‘Others’ refers to tokens numbering less than three, including [n, əɹ].
10 The minor adaptations are shaded in gray.
‧
voiceless stops are given in (4).(3) Statistics of voiceless stops substitutions11 French onset
(4) Examples of [p, t, k] substitutions
a. Picardy [pi.kaʁ.di] 皮卡底 [pʰi.kʰa.ti] ‘place name’
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
29
stops with the voiceless counterparts. This is evidenced in the computation of (5).
Every category of voiced stops is substituted by the voiceless stops. There are only two possibilities of [b] adaptation: [p] and [pʰ]. The former is the dominant substitute (98.36%). Such pattern is similar to the adaptation of [d]. The voiceless [t] is the majority (95.41%), while the remaining substitutes are very few in number. The situation is more complicated in [g] adaptation. Besides the adapted form [k], the segment [g] could also be replaced with the affricate [tɕ]. The two variant forms occupy 64.58% and 31.25% of the corpus, respectively. However, the aspirated adaptation [kʰ] has limited occurrences. Examples of voiced stops are provided in (6).
(5) Statistics of voiced stops substitutions French onset
structure
Mandarin structure
Number Total Percentage
bV p_ 120 122 98.36%
pʰ_ 2 1.64%
dV t_ 104 109 95.41%
tʰ_ 2 1.83%
tɕ_ 2 1.83%
tɕʰ_ 1 0.93%
gV k_ 31 48 64.58%
tɕ_ 15 31.25%
kʰ_ 2 4.17%
‧
(6) Examples of [b, d, g] substitutions
a. ballet [ba.lε] 芭蕾 [pa.lej] ‘a type of dance’
French has three voiceless fricatives [f, s, ∫]. First, consider the adaptations of [f].
The computation in (7) shows that 95.74% of the adaptations are faithful, while the velar fricative [x] has few occurrences.
Compared with [f] substitution, more variants are observed in the adaptation of [s] and [∫]. As (7) indicates, [s] is faithfully preserved most of the time (56.83% of the data). The consonants [ʂ] and [ɕ] have lower frequency in the adaptation (25.18% and 15.83%, respectively). Here we observe that the feature [continuant] is preserved, even though the place of articulation may alter. Because the substitute [tɕʰ] only occupies 2.16% of the data, it is excluded from the analysis.
The palato-alveolar fricative [∫] is mostly replaced by the alveolo-palatal fricative [ɕ] in Mandarin (70.15% of the occurrences). The post-alveolar fricative [ʂ], though more deviant from [∫], is the second-most adapted form (22.39%). Other substitutes, such as [s], [ts], and [tʂʰ] are very rare in the data. They are treated as minorities in the corpus and excluded from the discussion. Examples of each category can be found in (8).
‧
(7) Statistics of voiceless fricatives substitutions French onset
(8) Examples of [f, s, ∫] substitutions
a. morphine [mɔʁ.fin] 嗎啡 [ma.fej] ‘name of a drug’
[mej.ti.tɕʰi] ‘name of a literary award’ labio-velar fricative [w]. Sometimes they may devoice the [v] segment to obtain [f], which occupies 31.07% of the occurrences. Rarely do Mandarin speakers delete the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
32
onset [v] segments.
For the adaptations of [z] and [ʒ], at least four different variants could be observed. In the case of [z], the segment [s] is most favored (27.66%), followed by the dental affricate [ts] and the alveolo-palatal fricative [ɕ]. The latter two substitutes both occupy 14.89% of the data. Occasionally, [tɕʰ] replaces [z], which accounts for 10.64% of all occurrences. The segments [tɕ, tʂ, tsʰ, ʐ, ʂ] and deletion are minor substitutes. None of them appears more than three times. Therefore, they are excluded from the analysis.
The alveolo-palatal affricate [tɕ] is dominant in the adaptations of [ʒ]. It accounts for 34.21% of all occurrences. The second-most adapted form is the fricative [ʐ]. It occupies 26.32% of the data. The percentages of [ʐ] and [tɕ] are very close. The third-most chosen substitute is the post-alveolar affricate [tʂ] (14.47%). It is followed by the post-alveolar fricative [ʂ] (7.89%). Other segments [tɕʰ, ts, j, l, ɕ, ɥ] and onset deletion have limited occurrences (17.11% in total). They are excluded from the discussion. Examples of dominant adaptations are listed in (10).
‧
(9) Statistics of voiced fricatives substitutions French onset
(10) Examples of [v, z, ʒ] substitutions
a. Avignon [a.vi.ɲɔ̃] 亞維儂 [ja.wej.noŋ] ‘place name’
The nasal segments [m, n, ɲ] present a rather consistent pattern in the adaptation.
They are mostly faithfully preserved. As shown in (11), both [m] and [n] have faithful
12 ‘Others’ refers to tokens numbering less than three, including [tɕ, tʂ, tsʰ, ʐ, ʂ] and deletion.
13 ‘Others’ refers to tokens numbering less than five, including [tɕʰ, ts, j, l, ɕ, ɥ] and deletion
‧
adaptations. This may result from the shared segments [m] and [n] in Mandarin. The nasal [ɲ] is replaced by the closest substitute [n]. The number of [ɲ] adaptation is not very huge, yet the tendency of [n] substitution (75%) is observed. Examples are listed in (12).
(11) Statistics of nasals substitutions French onset
(12) Examples of [m, n, ɲ] substitutions
a. massage [ma.saʒ] 馬殺雞 [ma.ʂa.tɕi] ‘massage’
3.3 Statistical Analysis-Coda and Nasal Vowel Adaptations
This section discusses the adaptations of French codas and nasal vowels. The coda segments in the corpus include consonants [p, t, k, b, d, g, s, f, ʃ, z, v, ʒ, ts, m, n, ɲ, l, ʁ]. The nasal vowels include [ɛ̃, ɔ̃, ɑ̃]. Segments that constitute the major portion
14 ‘Others’ refers to tokens numbering less than three, including [l, an, ən].
‧
will be presented first, followed by the segments of fewer occurrences.
3.3.1 Alveo-Dental Lateral [l] and Uvular Fricative [ʁ]
Mandarin speakers tend to replace coda [l] with [əɹ], null adaptation, or [lɤ]. The syllable [əɹ] is the most dominant substitute (46.21%). Coda deletion is the second-most adopted form (28.03%), followed by the substitute of [lɤ] (23.48%).
Other adapted forms [lwɔ] and [li] have limited occurrences (2.28%). They are excluded from the discussion.
Similar substitution patterns can be found in the adaptations of [ʁ]. Coda deletion and [əɹ] replacement are the major substitutes. The former occupies 58.68% of the data, and the latter accounts for 34.38% of all occurrences. Sometimes Mandarin speakers replace [Vʁ] with [xɤ], [lɤ], or [lin]. Yet these substitutions are minor in the corpus (6.94%). Therefore, they will not be included in the discussion. Examples are provided in (2).
(1) Statistics of [l, ʁ] substitutions French coda
15 Complex codas containing [l] tend to be preserved rather than deleted (6 cases of the former; 2 cases of the latter).
16 No huge discrepancy is observed in complex codas of [ʁ] (adaptation: 20 cases; deletion: 21 cases), therefore they are treated as the same.
‧
(2) Examples of [l, ʁ] substitutions
a. Albi [al.bi] 阿爾比 [a.əɹ.pi] ‘place name’
b. Bartholdi [baʁ.tɔl.di] 巴陶第 [pa.tʰɑw.ti] ‘name of a sculptor’
c. De Gaulle [də.ɡol] 戴高樂 [taj.kɑw.lɤ] ‘name of a politician’
d. Paul Meyer [pɔl.mɛ.jəʁ] 保羅 梅耶
[pɑw.lwɔ.mej.jɛ] ‘name of a clarinetist’
e. É mile Loubet [e.mil.lu.bɛ] 艾米利 盧貝
[aj.mi.li.lu.pej] ‘name of a president’
f. Quimper [kɛ̃.pɛʁ] 坎培 [kʰan.pʰej] ‘place name’
Three nasal consonants [m, n, ɲ] are observed in the corpus. Of the three nasals, only [m] and [n] are present in Mandarin, and [ɲ] is a foreign segment. As indicated in (3), coda [m] is mostly adapted as [mu] (61.53% of all occurrences). The remaining substitutions include [Vŋ], [ən], [mej], and deletion. There is a huge discrepancy between the percentage of [mu] adaptation and that of other substitutes. Therefore, only [mu] is included in the discussion. Examples of each substitution type are presented in (4).
(3) Statistics of [m] substitutions
French coda structure Mandarin structure Number Total Percentage
Vm mu 8 13 61.53%
Vŋ 2 15.38%
ən 1 7.69%
mej 1 7.69%
deletion 1 7.69%
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
37
(4) Examples of [m] substitutions
a. Nîmes [nim] 尼姆 [ni.mu] ‘place name’
b. Vendôme [vɑ̃.dom] 凡登/芳登 [fan.təŋ]/[fɑŋ.təŋ] ‘place name’
c. Bir-Hakeim [biʁ.a.kɛm] 畢哈肯 [pi.xa.kʰən] ‘name of a bridge’
d. Clamecy [klam.si] 克拉梅西 [kʰɤ.la.mej.ɕi] ‘place name’
e. Lancôme [lɑ̃.kom] 蘭蔻 [lan.kʰow] ‘brand name of a cosmetics product’
In (5), the top three adaptations of coda [n] are [Vn] (50.82%), [Vŋ] (24.59%), and [na] (19.67%). Deletion and [nej] substitutes have limited occurrences. They are excluded from the discussion. The substitution pattern reveals that the nasal feature tends to be preserved. Both the alveolar nasal and the velar nasal are licit codas in Mandarin. That is why Mandarin speakers tend to choose [n] and [ŋ] in the adaptation.
Examples of each substitution type are provided in (6).
(5) Statistics of [n] substitutions
French coda structure Mandarin structure Number Total Percentage
Vn17 Vn 31 61 50.82%
Vŋ 15 24.59%
na 12 19.67%
deletion 2 3.28%
nej 1 1.64%
17 No category of consonant cluster is singled out, as all [n] segments in coda are preserved:
‘La Liane’ [(la).ljan] [li.an.(xɤ)] 里安(河)
‘Cézanne’ [se.zan] [saj.ʂɑŋ] 塞尚
‘Marne’ [maʁn] [ma.ən.(xɤ)]] 馬恩(河)
‘Lanzmann’ [lanz.man][lan/lɑŋ.tsɨ.man] 蘭茲曼/朗茲曼
‧
(6) Examples of [n] substitutions
a. Guillotine [ɡi.jɔ.tin] 吉約丹 [tɕi.ɥɛ.tan] ‘name of an second-most adopted form is [ni] (25%). The substitute [Vn] and coda deletion only have one occurrence. They are excluded from the discussion. The adaptation of [njɛ]
and [ni] suggests that [nasal] and [high] features are preserved in the adaptation of [ɲ].
Examples of each substitution are shown in (8).
(7) Statistics of [ɲ] substitutions
French coda structure Mandarin structure Number Total Percentage
Vɲ njɛ 7 12 58.33%
ni 3 25.00%
_n 1 8.33%
deletion 1 8.33%
(8) Examples of [ɲ] substitutions
a. Gascogne [ɡas.kɔɲ] 加斯科涅 [tɕja.sɨ.kʰɤ.njɛ] ‘place name’
b. Bretagne [bʁə.taɲ] 布列塔尼 [pu.ljɛ.tʰa.ni] ‘place name’
c. Charlemagne [ʃaʁ.lə.maɲ] 查理曼 [tʂʰa.li.man] ‘name of a king’
d. Chassagne [ʃɑ.sɑɲ] 夏沙 [ɕja.ʂa] ‘place name’
‧
3.3.3 Voiceless Fricatives
Three consonants [s, f, ʃ] are discussed in this section. First, consider the adaptation of [s]. In (9), the dominant substitution of [Vs] is [sɨ]. Other substitutes include [s] deletion, [ʂɨ], [saj], and [ɕi]. They only account for a minor part of the corpus, so they are excluded from the discussion. Examples of each substitution type are provided in (10).
(9) Statistics of [s] substitutions
French coda structure Mandarin structure Number Total Percentage
Vs18 sɨ 49 61 80.33%
others
(deletion, ʂɨ, saj, ɕi)
12 19.67%
(10) Examples of [s] substitutions
a. Nice [nis] 尼斯 [ni.sɨ] ‘place name’ there are only two occurrences of coda [f]. The only adaptation form is [fu].
(11) Statistics of [f] substitutions
French coda structure Mandarin structure Number Total Percentage
Vf fu 2 2 100%
18 No category of consonant cluster is provided, as their behavior is similar to that of simplex coda:
‘Île-d'Aix’ [il.dɛks] [aj.kʰɤ.sɨ.(tɑw)] 埃克斯/艾克斯(島)
‘Manuel Valls’ [ma.ny.ɛl.vals] [wa.əɹ.(sɨ)] 瓦爾(斯)總理
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
40
(12) Example of [f] substitution
a. Roscoff [ʁɔs.kɔf] 侯斯寇夫 [xow.sɨ.kʰow.fu] ‘place name’
Finally, let us examine the adaptation of the [Vʃ] structure. Six substitutes are observed. The syllables [ɕV], [ʂɤ], and [tɕʰi] are the top three adaptations. The adaptation of [ɕV] is dominant (44.44%). Both [ʃ] and [ɕ] have the feature of [+distributed]. This may explain why [ɕ] constitutes the principal part of the substitution. The adaptations [tɕi], [tsɨ], and [ɕjɛ] only have one example each. They are regarded as minority forms and are left out of the discussion. Examples of each substitution type are given in (14).
(13) Statistics of [ʃ] substitutions
French coda structure Mandarin structure Number Total Percentage
Vʃ ɕV19 8 18 44.44%
ʂɤ 4 22.22%
tɕʰi 3 16.66%
tɕi 1 5.56%
tsɨ 1 5.56%
ɕjɛ 1 5.56%
(14) Examples of [ʃ] substitutions
a. Avranches [a.vʁɑ̃ʃ] 阿弗宏許 [a.fu.xoŋ.ɕy] ‘place name’
b. Avranches [a.vʁɑ̃ʃ] 阿夫朗什 [a.fu.lɑŋ.ʂɤ] ‘place name’
c. Avranches [a.vʁɑ̃ʃ] 阿弗洪奇 [a.fu.xoŋ.tɕʰi] ‘place name’
d Avranches [a.vʁɑ̃ʃ] 阿維蘭及 [a.wej.lan.tɕi] ‘place name’
e. Kouchner [kuʃ.nɛʁ] 庫茲內 [kʰu.tsɨ.nej] ‘name of a politician’
f. Loches [lɔʃ] 洛榭 [lwɔ.ɕjɛ] ‘place name’
19 V could be [i] or [y], the latter being the majority.
‧
3.3.4 Voiced Fricatives
There are three voiced fricatives [z, v, ʒ] in the corpus. The fricative [z] has four substitutions, [tsɨ], [sɨ], [tʂV], and coda deletion. The distribution in (15) shows that [tsɨ] and [sɨ] are the top two substitutes. The percentages of [tʂV] and coda deletion are relatively low. Therefore, they are excluded from the discussion. Examples of [z]
adaptation are listed in (16).
(15) Statistics of [z] substitutions
French coda structure Mandarin structure Number Total Percentage
Vz20 tsɨ 8 17 47.06%
sɨ 5 29.41%
tʂV21 2 11.76%
deletion 2 11.76%
(16) Examples of [z] substitutions
a. La Meuse [(la).møz] 謬茲(河) [mjow.tsɨ.(xɤ)] ‘river name’
In (17), the adaptation of [v] is dominated by [fu]. It accounts for 70% of the data.
The substitutions of [fwɔ] and [wej] constitute 30% of the data. They are left out because of limited occurrences. The preservation of [labial] feature and [continuant]
feature are prominent in adapting [v]. Examples are shown in (18).
20 No category of consonant cluster is singled out, as their behavior is similar to that of simplex codas:
‘Toulouse’ [tu.luz] [tʰu.lu.sɨ] 土魯斯
‘Oise’ [waz] [wa.tsɨ] 瓦茲
‘Lanzmann’ [lanz.man][lan/lɑŋ.tsɨ.man] 蘭茲曼/朗茲曼 [lan.sɨ.man] 藍斯曼
21 The vowel could be [ɤ] or [ɨ]. There is one example of each.
‧
(17) Statistics of [v] substitutions
French coda structure Mandarin structure Number Total Percentage
Vv22 fu 7 10 70%
fwɔ 2 20%
wej 1 10%
(18) Examples of [v] substitutions
a. Nièvre [njɛvʁ] 涅夫勒 [njɛ.fu.lɤ] ‘place name’
b. La Sèvre [(la).sɛvʁ] 賽佛爾(河) [saj.fwɔ.əɹ.(xɤ)] ‘name of a river’
c. Hauterives [o.tə.ʁiv] 歐特希維 [ow.tʰɤ.ɕi.wej] ‘place name’
Variant adaptations of coda [ʒ] are observed in the corpus. The substitution of [ʐɨ]
and [tɕi] are dominant. The percentage is 46.15% and 30.77%, respectively. The rest of the substitutes [tɕʰi], [tsɨ], and coda deletion only have one example each. They are considered minor parts and excluded. Statistical analysis is presented in (19), and examples are given in (20).
(19) Statistics of [ʒ] substitutions
French coda structure Mandarin structure Number Total Percentage
Vʒ23 ʐɨ 6 13 46.15%
22 No category of consonant cluster is singled out, as their behavior is similar to that of simplex codas:
Catherine Deneuve [ka.tʁin də.nœv] [kʰaj.sa.lin.tan.ni.fu] 凱薩琳丹妮芙 Louvre [luvʁ] [lwɔ.fu] 羅浮
23 No category of consonant cluster is singled out, as their behavior is similar to that of simplex codas:
Serge Gainsbourg [sɛʁʒ.ɡɛ̃s.buʁ] [sɤ.tɕi.kan.sɨ.pwɔ] 塞吉甘斯伯 massage [ma.saʒ] [ma.ʂa.tɕi]馬殺雞
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
43
(20) Examples of [ʒ] substitutions
a. Pérouges [pe.ʁuʒ] 佩胡日 [pʰej.xu.ʐɨ] ‘place name’
b. Pérouges [pe.ʁuʒ] 貝胡吉 [pej.xu.tɕi] ‘place name’
c. montage [mɔ̃.taʒ] 蒙太奇 [məŋ.tʰaj.tɕʰi] ‘a technique in film editing’
d. Bourges [buʁʒ] 布爾茲 [pu.əɹ.tsɨ] ‘place name’
e Limoges [li.mɔʒ] 里摩 [li.mwɔ] ‘place name’
3.3.5 Voiceless Stops
The substitutions of [p] are listed in (21). Compared with the distributions of [t, k], the number of [p] codas is relatively small. The coda [p] can be replaced by [phu], [pwɔ], [pu] or deleted. Since there are limited occurrences of [p], generalizations are unlikely. Therefore, the adaptations of voiceless stops will focus on [t] and [k]. The substitutes of [p] will not be discussed. Examples of each category are available in (22).
(21) Statistics of [p] substitutions
French coda structure Mandarin structure Number Total Percentage
Vp phu 2 6 33.33%
pwɔ 2 33.33%
pu 1 16.67%
deletion 1 16.67%
(22) Examples of [p] substitutions
a. Guadeloupe [ɡwa.də.lup] 瓜德羅普 [kwa.tɤ.lwɔ.pʰu] ‘place name’
b. Le Procope [(lə).pʁo.kɔp] 波伯 [pwɔ.pwɔ] ‘shop name’
c. Le Procope [(lə).pʁo.kɔp] 波布 [pwɔ.pu] ‘shop name’
d. Le Procope [(lə).pʁo.kɔp] 波蔻 [pwɔ.kʰow] ‘shop name’
Mandarin speakers tend to adapt the coda [t] as [tʰɤ] (60.87% of the data). The second-most adopted form is coda deletion (30.43%), followed by the substitute of [tɤ]
‧
(8.70%). Only [tʰɤ] and coda deletion are included in the discussion. The substitute [tɤ]
is treated as a minority and is left out. Statistical analysis and examples are presented in (23) and (24), respectively.
(23) Statistics of [t] substitutions
French coda structure Mandarin structure Number Total Percentage
Vt24 tʰɤ 14 23 60.87%
deletion 7 30.43%
tɤ 2 8.70%
(24) Examples of [t] substitutions
a. Nantes [nɑ̃t] 南特 [nan.tʰɤ] ‘place name’
b. Lafayette [la.fa.jɛt] 拉法葉 [la.fa.jɛ] ‘name of a department store’
c. Comte [kɔ̃t] 孔德 [kʰoŋ.tɤ] ‘name of a philosopher’
The coda [k] is mostly substituted by the syllable [kʰɤ], which occupies 81.25%
of the data. Coda deletion is the second-most adopted form. It accounts for 14.58% of all occurrences. As for [kʰu] and [kɤ], they have limited occurrences (4.16% in total).
Therefore, they will not be counted in [Vk] generalization. Statistics and examples are found in (25) and (26), respectively.
(25) Statistics of [k] substitutions
French coda structure Mandarin structure Number Total Percentage
Vk25 kʰɤ 39 48 81.25%
deletion 7 14.58%
others (kʰu, kɤ) 2 4.16%
24 No category of consonant cluster is singled out, as their behavior is similar to that of simplex codas:
‘Proust’ [pʁust] [pʰu.lu.sɨ.tʰɤ] 普魯斯特
‘Le Nôtre’ [lə.notʁ] [lɤ.nwɔ.tʰɤ] 勒諾特
25 Out of the 6 words containing consonant clusters, only 1 of them deletes the coda [k]:
‘Jeanne d’Arc’ [ʒan.daʁk] [tʂən.tɤ] 聖女貞德
‧
(26) Examples of [k] substitutions
a. Chirac [ʃi.ʁak] 席哈克 [ɕi.xa.kʰɤ] ‘name of a
president’
b. Viollet-le-Duc [vjɔ.lɛ.lə.dyk] 維優雷 勒杜
[wej.jow.lej.lɤ.tu] ‘name of an architect’
c. Cocteau [kɔk.to] 柯庫德 [kʰɤ.kʰu.tɤ] ‘name of a poet’
d. Conques [kɔ̃k] 孔格 [kʰoŋ.kɤ] ‘place name’
3.3.6 Voiced Stops
There are three voiced stops in the present corpus: [b, d, g]. Mandarin speakers tend to adapt [b] as [pwɔ] or [pu], as indicated in (27). Since there is only one example each for [pej] and coda deletion, they will not be counted in the corpus.
Examples are given in (28).
(27) Statistics of [b] substitutions
French coda structure Mandarin structure Number Total Percentage
Vb26 pwɔ 6 12 50%
pu 4 33.33%
others (pej, deletion) 2 16.67%
(28) Examples of [b] substitutions
a. Fabre [fabʁ] 法伯 [fa.pwɔ] ‘name of an
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
46
Mandarin speakers tend to substitute the syllable [tɤ] for [Vd] (75%). The variant adaptation [tʰɤ] occupies 16.67% of all occurrences. Since there is only single example for [taj], it is excluded from the discussion. Examples of each variant are
Mandarin speakers tend to substitute the syllable [tɤ] for [Vd] (75%). The variant adaptation [tʰɤ] occupies 16.67% of all occurrences. Since there is only single example for [taj], it is excluded from the discussion. Examples of each variant are