The Corpus and Patterns
4.3 Simplex onsets
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
112
4.3 Simplex onsets
As has been discussed earlier in Chapter 2, TM, like most other languages, allows most consonants in the onset. On account of this, it may well be predicted that most Englishconsonants in the [.CV] sequence will be faithfully retained in their TM output forms, insofar as the consonant as well as the [.CV] combination are accepted by both languages. For one thing, for those consonants and combinations that English has but TMlacks, such as the interdental ['] and the structure of [.si.], substitution for the onset consonant is obligatory, yet the replacing consonants are limited to those with the greatest auditory similarity to the English input. For another, it will be claimed that unexpected mappings of the onset consonants between English and TM are largely due to social (e.g. family names, genders) or semantic (e.g. commercial terms) considerations. The patterns, along with examples (transcribed in broad IPA symbols) and descriptive generalizations, are given in the subsections that follow, where obstruents are discussed in subsections from 4.3.1 to 4.3.3, and sonorants in subsections from 4.3.4 to 4.3.5.
4.3.1 Stops
English stops include bilabial [p, p , b], alveolar [t, t , d], and velar [k, k , (], whereas those in TM are [p, p ], [t, t ], and [k, k ], respectively. Voiced stops are absent in TM. A comparison between the stop inventories of the two languages is given below. To give a clearer picture, an English stop is listed in italic on the left and a TM one on the right (if any) in the cell that corresponds to their places and manners of articulation.
‧
Since English voiceless stops are aspirated by default syllable-initially, voiceless unaspirated stops are absent in this position (an unaspirated stop preceded by [s] is discussed later in 4.4.1). Before we look at the facts, accordingly, we plausibly expect that English aspirated stops are mapped to the identical aspirated stops in TM, while English voiced stops should be mapped to TM voiceless unaspirated stops, given that a voiced stop differs from a voiceless unaspirated one solely in the feature of [±voiced]. In total, 876 stop onsets are collected from the corpus, and the results are shown in (75). reasons will be omitted throughout.
‧
corresponding borrowing forms (the only exception is “[.se .bu.] Cebu → [.su.u.] 宿‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
115
霧”). Furthermore, despite the very few exceptions (as in “[b] → [m]” in (75d)), the manners of articulation of the English consonants are faithfully retained in most of the correspondent consonants in TM. By and large, consistent to our expectations, English aspirated stops are mostly interpreted as the matching TM aspirated stops, and English voiced stops tend to be interpreted as voiceless unaspirated stops in TM. In what follows, we give a more detailed discussion on the patterns.
Adaptation patterns of voiceless aspirated stops appear in (75a-c). In (75a), as expected, the labial [p ] is interpreted as the identical [p ] in the most cases (64.55%) and as the unaspirated counterpart [p] in the second most cases (33.64%), altogether accounting for over 98%. Interestingly, the other two alternatives, though extremely rare (0.91% each), are both labials, [f] and [m], too. In (75b), likewise, the coronal [t ] is mapped to [t ] (53.38%) or the unaspirated [t] (45.27%) over 98% of the time. The interpretations of the dorsal [k ] in (75c), however, reveal a more diverse pattern. First, its mapping to the identical [k ] makes up the majority of all adaptations (67.03%), similar to the patterns of [p ] and [t ]. The secondary mapping, surprisingly, is to the palatalized [t$] (18.68%), rather than the expected [k], which comes in the third (9.89%). The least likely option is to the aspirated alveolo-palatal [t$ ], which is found in 4.95% of the data.
Patterns (75d) to (75f) illustrate the adaptations of voiced stops. Since TM does not have voiced stops, phonemic substitution is obligatory in this category. In (75d), a vast majority of [b]-onsets are interpreted as the unaspirated [p], with the frequency of 88.59%. Only 9.24% of them are adapted into the aspirated [p ]. What’s more, 3 out of the 184 [b]’s are adapted into the nasal [m], another labial sound in TM. The interpretations of the coronal [d] in (75e) shows a similar pattern to (75d): it is mapped to the voiceless [t] up to 93.65% of the time, with only 6.35% of them
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
116
mapped to the aspirated [t ]. Finally, like [k ] in (75c), the pattern of the voiced dorsal [(] demonstrates a greater diversity. The primary mapping is also the unaspirated [k], with the frequency of 65.67%. The mapping to the alveolo-palatal [t$] constitutes 29.85%, coming in the second most. The rest two minor adaptations are [k ] and the velar glide [w], each of which accounts for only 1.49% of all.
Let us turn our attention to the mapping of aspiration and voicing between the two languages. In cases where the aspirated stop onsets in English are retained in TM (439 items), 64.24% (282/439) of them are mapped to aspirated voiceless stops (e.g.
[.p k.] punk → [.p .k .] 龐克) or, in very few cases, to fricatives (e.g. [.p æn.]
pan → [.fan.] 泛) or aspirated affricates (e.g. [.m .k .] Micky → [.mi.t$ i.] 米奇), and the remaining 35.76% (157/439) are mapped to their unaspirated voiceless counterparts ([.p i.t ] Peter → [.pi.t .] 彼得) or an unaspirated affricate ([.k .]
King → [.t in. n.] 金恩). In cases where the voiced stop onsets in English are retained in TM (438 items), on the other hand, 93.15% (408/438) are mapped to their voiceless unaspirated counterparts (e.g. [.d .n m.] denim → [.tan.ni .] 丹寧) or, in very few cases, to nasals ([.bam.be .] Bombay → [.mo .mai.] 孟買), affricates ([. l.]
Gill → [.t i. .] 吉兒), or glides ([. æs.] gas → [.wa.s!.] 瓦斯). The rest 6.85%
(30/438) are realized as their aspirated counterparts ([.d . .' .] Dorothy → [.t u.l .s!.] 桃樂絲).
4.3.2 Fricatives
There exists a greater divergence between English and TM in the inventory of fricatives than of stops. Like what is done in (74), the fricatives in the two languages are listed in (76).
‧
As can be seen, the only two fricatives that the two languages share in common are the labial-dental [f] and the alveolar [s]. English additionally has the voiced counterparts [v] and [z] respectively. In the remaining places of articulation, English has the inter-dentals [ ] and [ ], palatals [ ] and [ ], and glottal [h], while TM has the alveolo-palatal [ ], retroflexes [ ], and velar [x]. In addition to the expected faithful mappings of [f] and [s] from English to TM, it might as well be the case that a foreign segment should be mapped to a similar one with the nearest place of articulation. The patterns of fricatives are shown in (77), in a similar layout to (75), where (77a-e) present the voiceless fricatives and (77f-i) present the voiced fricatives.
(77) Adaptation patterns of fricative onsets
‧
TM, and hence is used to refer to the long strip of magnetic material contained in a cassette. A case like this should be classified as a hybrid, since semantic factors are involved in the formation of the loanword. The reader is referred back to Section 1.2 for more information on hybrid loanwords.‧
Upon comparing the patterns of fricatives with those of stops, it will be clear that in general, like stops, a vast majority of the fricatives that are absent in the TM sound inventory are mapped to similar sounds that have a close place of articulation, as in [']
→ [s], [%] → [$], [h] → [x], and [z] → [s]. In addition, also like stops, despite the few exceptions (e.g. ['] → [t] in (77b)), the manner feature [+fricative] is overwhelmingly retained. A distinctive case is the pattern of [v], where the primary mapping is the glide [w] (61.25%), which differs from [v] in place and manner, rather than the expected [f] (30%), differing in voicing only.
Like stop onsets, nearly 100% (99.1%) of the 447 fricative onsets are retained in the borrowing forms (with four exceptions of deletion), whereas the correspondent consonants are not always identical to the sources. As predicted, the majority of
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
120
English [f] and [s] onsets are faithfully mapped to L1 [f] and [s] respectively, as these are the sounds that the two languages have in common. Between [f] and [s], however, the adaptation pattern of the latter shows a greater variation, in which the primary mapping [s] accounts for only 40.85%.
Turning to the fricatives that are absent from the TM sound inventory, we see that a vast majority of them are mapped to similar sounds that have a close place of articulation, as in ['] → [s], [%] → [$], [h] → [x], and [z] → [s]. Moreover, once again, despite the few exceptions, the manner feature [+continuant] is overwhelmingly retained. An exceptional case is the pattern of [v], where the major mapping is the glide [w] (61.25%), which differs from [v] in place and manner, instead of the expected [f] (30%), differing in voicing only. The rest of this subsection addresses a closer observation on the patterns.
The patterns of voiceless fricative onsets appear in (77a) to (77e). In (77a), labiodental [f] is ideally mapped to [f] in TM massively, as found in 95.71% of the data. The other alternate, in the rest 4.29%, is [xw]. In a sense, the feature of [+fricative] is retained in [x] and that of [labial] in the lip-rounding glide [w], i.e., a process termed “reconfiguration” in literature. A similar process is observed by Hsiao (2011) in Taiwanese-accented Mandarin. This is exemplified in (78), where TAM stands for Taiwanese-accented Mandarin.
(78) Reconfiguration in accent formation and loanword adaptation
TM TAM Gloss
[.wei.fa.] → [.we.hwa.] ‘against the law’
English TM
[.we .f .] → [.wei.xwa.] ‘wafer’
The lack of [f] in Taiwanese, a member of Chinese language family, leads the
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
121
native speaker of this dialect to preserve the features of [+fricative] and [+labial] of the sound by “unpacking” this segment into two, i.e. [h] and [w] respectively (Taiwanese has no [x]). A critical difference between the two applications is that this process is obligatory in the formation of Taiwanese accent, seeing that [f] is illicit in Taiwanese, while it should be viewed as redundant, or at least optional, in the adaptation of English words by TM speakers, since [f] is the consonant that the two languages share in common. Reconfiguration of this sound seems to be unnecessary.
Pronunciation of the interdental ['] has been a long-lasting difficulty for most English learners in Taiwan. A typical example is that they say [.san.k ju.] for “Thank you.” Thus a witty written expression that is widespread in the younger generation when emailing or texting through smartphones is “3Q”, in which “three” is pronounced as [.san.] in TM. This phenomenon lends support to TM speakers’
adaptation of ['] in (77b), where its mapping to [s] holds a majority, as found in 43.48% of the cases. Nonetheless, if the following vowel is the high front [i] or [ ], it is palatalized to [$], since [$] is the only fricative that is allowed to precede a high front vowel in TM. Four such syllables are found in the sources (17.39%), and all of the [']’s are mapped to [$] with no exceptions.
The fricative that is found in the most simplex onsets is the voiceless alveolar [s], with 141 tokens. It follows that a more diverse pattern emerges in (77c). First, the faithful mapping [s], though being the primary mapping as expected, makes up only 41.13%. Furthermore, as has been discussed for (77b), it is palatalized to [$] in all the 30 items (21.83%) where the nuclear vowel or the prenuclear glide is high front.
Another 34.51% of them, somehow counter to our expectation, are mapped to the retroflex [ ].
In (77d), English [%] is mostly mapped to the more front [$] in TM in 71.79% of
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
122
the cases. The second choice is the more back [ ], with the percentage of 23.08%. The rest 5.13% is the coronal fricative [s], which is the farthest from [%]. The glottal [h] is mapped to the nearest velar [x] in 75.81% of the data. And like ['] and [s], when the following vowel or glide is high front (found in 20.97% of all), it is palatalized to [$]
due to the prohibition against a *[.xi.] syllable.
The adaptation patterns of voiced fricative onsets are presented in (77f) to (77i).
First, as shown in (77f), the mappings of the voiced labiodental [v] reveals the greatest diversity among all fricatives, where as many as six mappings are attested. A noteworthy observation is that the primary mapping of [v] is not the more expected [f]
(31.33%, the second most occurrences), which is distinctive from its voiced counterpart solely in voicing, but rather the glide [w] (60.24%), which may be perceptually more distinct from [v] for it is [+round] and [+velar]. Another three minor alternatives are the labial nasal [m], the “unpacked” [xw] that is also attested in (4a) and discussed, and deletion of it, each accounting for 2.44%. The voiced interdental [;] is hardly attested in the source, with only one example, where it is mapped to the aspirated coronal stop [t ].
The adaptation pattern of the voiced alveolar [z] in (77h) demonstrates the greatest variety among all English fricatives: it can be mapped to eight different consonants in TM. Not surprisingly, the majority is [s], which differs from [z] in voicing only, with the frequency of 44.12%. Like the pattern of [s] in (4c), the second and third most alternatives are [ ] and [$] respectively. Along the same lines, the mapping of [ ] involves social/semantic factors and that of [$] is the result of palatalization before high front vowels or glides. Finally, compared to the adaptation of [s], a higher percentage of [z]-onsets are mapped to affricates [t$ ], [ts ], [t ] and [ts], which in all constitute 20.58%. Finally, adaptation of the voiced [#] is attested in
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
123
only two instances, where it consistently maps to the voiceless [$] in TM.
4.3.3 Affricates
The last obstruent type to observe is affricates. The inventories of affricates in the two languages are combined in (79).
(79) Affricates of English and TM
dental alveolo-palatal palatal retroflex
[–vcd, -asp] ts t t
[–vcd, +asp] ts t t t
[+vcd] d
Contrary to fricatives, TM has a richer inventory of affricates than English does:
TM has alveolar [ts ] and [ts], alveolo-palatal [t ] and [t ], and retroflex [t ] and [t ], while English has palatals [t ] and [d ] only. Based on the generalizations we have achieved from (75) and (77), if once again manner is robustly perceived and hence retained in L1, we may well expect that in most items, English [t ]/[d ] should be interpreted as the neighboring [t ]/[t ] or [t ]/[t ] respectively but less likely as the farther [ts]/[ts ]. The adaptation patterns of English [t ] and [d ] in TM are presented below.
(80) Adaptation patterns of affricate onsets a. [t%]
TM Percentage Example (L2 → L1)
[t$ ] 56.67% (17/30) [.t .t .] cha-cha → [.t ja.t ja.] 恰恰 [t ] 16.67% (5/30) [.t .l .] Chile → [.t !.li.] 智利
[t ] 13.33% (4/30) [.t p.m n.] Chapman → [.t a.p u.man.] 查普曼 [t$] 6.67% (2/30) [.t% .t l.] Churchill → [.t$ jou.t i. .] 邱吉爾 [$] 3.33% (1/30) [.d .v n.t i.] da Vinci → [.ta.w n. i.] 達文西 [x] 3.33% (1/30) [.t m.ski.] Chomsky → [.x . !.t$i.] 杭士基
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
124
b. [d#]
TM Percentage Example (L2 → L1)
[t$] 60.71% (51/84) [.l .d k.] logic → [.lwo.t i.] 邏輯 [t$ ] 21.43% (18/84) [. n.d n.] engine → [.in.t i .] 引擎 [t ] 8.33% (7/84) [.d e mz.] James → [.t an.mu. !.] 詹姆士 [y] 4.76% (4/84) [.d n.] John → [. e.xan.] 約翰 [ ] 2.38% (2/84) [.d .m n.] German → [. !. .man.] 日耳曼 [ts] 1.19% (1/84) [.d o .d .n s.]
Jourdeness
→ [.tswo.tan.ni.s!.]
佐丹妮絲
[ ] 1.19% (1/84) [.mæ.d k.] Magic → [.mwo. u.] 魔術
As expected, a great majority of English affricates [t%] and [d#] are mapped to the neighboring affricates [t$], [t$ ], [t ], and [t ] in TM, as found in 91.15% (103/113) of all. Respectively, in (80a), [t%] is mapped to the aspirated alveolo-palatal [t$ ] in 56.67% of the cases. As noted in Chapter 3, within a syllable, alveolo-palatal phonemes [t$] (ㄐ), [t$ ] (ㄑ), and [$] (ㄒ) have to be followed by a front high vowel/glide [i]/[j] (both transcribed as “ㄧ” in Chinese phonetic symbols) or [y]/[ ] (both as “ㄩ”) in TM phonotactics. Accordingly, it is assumed that in adaptation forms such as “[.t% .t% .] cha-cha → [.t$ ja.t$ ja.] 恰恰” and “[.t% . .] Cherry → [.t$ 'e.li.]
雀莉”, the addition of the glides [j] and [ ] should be the result of perceptual insertion, given that TM phonology is influential in loanword adaptation early in the level of perception. Detailed discussion on perception will be given as we move on to the next chapter. The secondary alternatives are the retroflexes [t ] and [t ], making up 30%
altogether. In (80b), where we present the pattern of the voiced [d#], 60.71% of the items are mapped to the unaspirated alveolo-palatal [t$], and 21.43% to the aspirated [t$ ].
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
125
4.3.4 Nasals
English and TM have the same set of nasal phonemes, namely the alveolar [n], bilabial [m] and velar [ ]. This is shown in (81).
(81) Nasals of English and TM
bilabial alveolar/dental velar m m n n
As introduced in Chapter 3, both languages allow [n] and [m] and disallow [ ] to be the syllable onset. In that event, we plausibly expect that English nasal onsets [n]
and [m] should be faithfully copied as [n] and [m] respectively in TM. The adaptation patterns of nasal onsets are given in (82).
(82) Adaptation patterns of nasal onsets a. [n]
TM Percentage Example (L2 → L1)
[n] 94.08% (143/152) [.na .l n.] nylon → [.ni.lo .] 尼龍 [l] 5.26% (8/152) [.s d.n .] Sydney → [.$ e.li.] 雪梨
[ ] 0.66% (1/152) [.flæ.n l.] flannel → [.fa.lan. o .] 法蘭絨21
b. [m]
TM Percentage Example (L2 → L1)
[m] 99.54% (218/219) [.mo .t l.] motel → [.mwo.t je.] 摩鐵 [w] 0.46% (1/219) [.m l.b . o .] Marlboro → [.wan.pau.lu.] 萬寶路
Seen from the above patterns, the result closely meets our expectation in that both [n] and [m] in English are faithfully mapped to the identical [n] and [m]
respectively in TM, with predominant frequencies 94.08% for the former and 99.54%
21 The last syllable (also a TM character) of the borrowing form [.fa.lan. o .] is plausibly attributable to semantic factors, since the character “[. o .] 絨” refers to a type of soft woven fabric, which may mean “down” or “velvet”. Both of the two fabrics are close to the texture of flannel.
‧
and retroflex [ ], as presented below.
(83) Liquids of English and TM
alveolar alveolar/post-alveolar
l l
Here again, English liquids [l] and [ ] are presumed to be faithfully mapped to the identical liquid phonemes [l] and [ ] respectively in TM. The observed adaptation patterns are presented below.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
127
retroflex. The percentage is 99.18% in (84a), where the input is the lateral [l], and it is 87.55%, which still remains high even though the input is the retroflex [ ] in (84b). By contrast, the occurrence rate of the faithful mapping [ ] in (84b), is the lower 12.09%, which runs counter to our expectation. A noteworthy observation in (84b), moreover, is that for the 33 syllables from the source words where the retroflex approximant is interpreted as [ ], there are 29 in which the nucleus is a front vowel, and in the remaining 4 the nuclear vowel is [+back].
On the whole, the patterns that emerge from the TM speakers’ adaptation of English simplex onsets have revealed two important tendencies that entail further elaborations from the perceptual points of view. First, over 99% of the simplex onsets across the board are retained and mapped to a correspondent phoneme in TM. Second, it is equally clear that the foreign onset consonant is either faithfully copied when the consonant is shared in common by the two languages involved, or minimally repaired into a phonetically similar sound so long as the foreign onset consonant is absent from the native sound inventory. A couple of exceptions, however, show that neither is obligatory. In (77f), an English [v], a foreign sound to TM speakers, is mostly interpreted as [w], rather than the expected [f], even though [w] deviates from the source in rounding and place of articulation, whereas [f] differs solely in voicing.
Moreover, in (84b), an English [ ] is mapped to the TM lateral [l] in most cases, rather than the supposed [ ].
In the forthcoming section, we will turn our focus to complex onsets.