The Corpus and Patterns
4.4 Complex onsets
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.
4.4 Complex onsets
It has been introduced in Chapter 3 that TM syllable structures, compared to English, are rather simple, with no consonant clusters permissible in syllable margins.
As TM speakers treat onset clusters as illicit, adjustment is obligatory when
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
128
borrowing of words with onset clusters takes place. What constitutes the core issues of cluster adaptation is the exploration the extent to which internal cues (e.g. the acoustic properties of the segment) and external cues (e.g. positive/negative cues given by the neighboring sounds) exert their influence on word-loaning by observing whether a consonant embedded in a cluster is rescued (retention) or sacrificed (deletion). To investigate the effects of perceptual salience on adaptations as such, it is rendered redundant to look into the TM mappings on a segment-by-segment basis.
Because of this, in the foregoing subsections, we will collect all consonant combinations in the onset position of English syllables based on the sequential order of the consonants. Each sequence is followed by a list of all correspondent TM surface structures with percentages of occurrence. Presumably, each consonant embedded in the English cluster should have a correspondent in the TM borrowing forms, given that syllable-initial onsets enjoy the perceptual privilege over syllable-final codas.
4.4.1 The “[s]-stop” sequences
The only fricative that can be followed by a stop [p], [t], or [k] syllable-initially in English is the alveolar [s]. It is a well-known phonological process in this language that an aspirated stop preceded by [s] will lose aspiration, hence “[.p .] par → [.sp .] spar”, “[.t .] tar → [.st .] star”, and [.k .] car → [.sk .] scar”. To reach a clearer picture of how English voiceless unaspirated stops are interpreted in TM in later discussion, we specifically divide the mapped structures into aspirated and unaspirated stops in the column of L1 structure, next to which is the percentage of occurrence of each TM structure.
‧
(85) Adaptation patterns of the “[s]-stop” sequences
L2 seq. L1 struct. Percentage Example (L2 → L1) clusters in onset are treated like simplex onsets by TM adapters: 97.37% (43/44) stay intact in TM through inserting a vowel for the initial [s], while the prevocalic consonant remains the onset consonant in TM, except for only one item where [s] is deleted and the stop is retained (2.63%). Second, a large number of the pre-stop [s]’s are mapped to a less similar [. !.], in preference to the phonetically closer [.s!.]. A plausible account for this has been offered previously: the character “史” ([. !.]) is a well-known family name, and is thus widely used for foreign name transliterations.
For instance, the [s] in Sparta, the name of a city-state in ancient Greece, is thereby faithfully mapped to [.s!.] (斯, ‘well-mannered’), whereas that in Stanford and Scott, both being people’s names, are mapped to [. !.] (史, a family name) to invoke more semantic reference to the source words. Third, crucially, 61.36% (27/44) of the unaspirated voiceless stops following [s] are mapped to the corresponding unaspirated stops, as in “[.spa .t .] Sparta → [.s .pa.ta.] 斯巴達”, or an unaspirated affricate, as
22 Note that the TM correspondent of English [s] is not necessarily [s], yet for convenience it is shown as the numbered [.s1…] in the column of L1 structure, since TM mapping of English [s] is not the focus of current discussion. For example, English personal name [.spa k.] (Spike) is mapped to L1 [. .p ai.k .] (史派克), while the L1 structure is still classified into the category of [.s1V1.p 2].
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
130
in “[.ski.p .] Skippy → [.t i.pi.] 吉比”, and the remaining 38.64% (17/44) are mapped to their aspirated counterparts, such as “[.spa k.] Spike → [. .p ai.k .] 史 派克”.
4.4.2 The “[t s]” sequence
In the corpus, the only English sequence where a stop can be followed by a fricative as an onset cluster is [t s], consisting of two homorganic obstruents (a rare exception is “[.pf .n (.] pfennig → [.f n.ni.] 芬尼”). The pattern is given in (86) below.
(86) Adaptation pattern of the “[t s]” sequence
L2 seq. L1 struct. Percentage Example (L2 → L1)
[t s] [ 2] 87.5% (7/8) [.w .t s n.] Watson → [.xwa. .] 華生 [ts 12] 12.5% (1/8) [.n .t s .] Nazi → [.na.ts wei.] 納粹
Against our presumption, the primary adaptation is deletion of [t ] and retention of [s], as found in 87.5% of the data. There is only one instance of faithful mapping to the TM affricate [ts ], making up the remaining 12.5%.
4.4.3 The “[s]-nasal” sequences
Observing the regulation of SSP, a nasal, being more sonorant, can follow [s], which is less sonorant, to form rising sonority before the nucleus. In English, [ ] is not allowed to be (part of) the onset, and hence the two possible “[s]-nasal” sequences are [sn] and [sm]. The patterns are shown in (87).
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
131
(87) Adaptation patterns of the “[s]-nasal” sequences
L2 seq. L1 struct. Percentage Example (L2 → L1)
[sn] [.s1V1.n2] 100% (5/5) [.sne p.] Snape → [. .nai.pu.] 石內卜 [sm] [.s1V1.m2] 100% (3/3) [.sm t.] smart → [. .m u.] 時髦
As expected, both [s] and the nasal are retained via vowel epenthesis for the syllable-initial [s] without exceptions, regardless of the place of articulation of the nasal.
4.4.4 The “obstruent-[l]” sequences
Following SSP, once again, English stops and fricatives can precede a liquid in the onset position, excluding the homorganic stops [t ] and [d]. As stops and fricatives behave in the same way when preceding the lateral liquid [l] syllable-initially, we do not separate them apart in looking into the patterns of sequences of this type. This is shown in (88).
(88) Adaptation patterns of the “obstruent-[l]” sequences L2 seq. L1 struct. Percentage Example (L2 → L1)
[p l] [.p 1V1.l2] 100% (2/2) [.p le .t o .] Plato → [.pwo.la.t u.] 柏拉圖 [k l] [.k 1V1.l2] 100% (21/21) [.k ln.t n.] Clinton →
[.k .lin.tw n.] 柯林頓
[bl] [.p1V1.l2] 100% (2/2) [.bl (.] blog → [.pu.lwo.k .] 部落格 [(l] [.k1V1.l2] 100% (4/4) [. l n.] Glen → [.k .lw n.] 葛倫 [fl] [.f1V1.l2] 100% (6/6) [.flo . .] Flora → [.fu.lwo.la.] 芙蘿拉 [sl] [.s1V1.l2] 100% (7/7) [.slæ.v k.] Slavic → [.s .la.fu.] 斯拉夫
It comes as no surprise that amongst onset clusters of this sequence, both the L2
obstruent and the lateral are retained in the L1 mappings. No exceptions are found.
Also, like those as the simplex onsets, aspirated stops within this sequence ([p ] and [k ] only) are 100% (23/23) realized as the identical aspirated stops in L1, as in
‧
100% (6/6) interpreted as their unaspirated counterparts ([p] and [k], respectively), as in “[.bl (.] blog → [.pu.lwo.k .] 部落格”.4.4.5 The “stop-[ ]” sequences
The other liquid in Englishthat may follow an obstruent to form an onset cluster with increasing sonority is the retroflex [ ]. Unlike 4.4.4, however, we discuss stops and fricatives in separate subsections, considering they act distinctively in onsets of this sequence. The patterns for stops are shown in (89).
(89) Adaptation patterns of the “stop-[ ]” sequences
L2 seq. L1 struct. Percentage Example (L2 → L1)
‧
Overall, stops with the same place of articulation act similarly and hence should be paired up in the description of the patterns. To begin with, for the [p ] and [b ] onset clusters, both the bilabial stop and [ ] are retained in a majority (92% combined, 23/25), with two exceptions. In one of the two exceptions, English [ ] is “partially”
preserved in that the feature [+round] is salvaged while [+coronal] is not, yielding the velar glide [w], as in “[.e .p l.] April → [.ai.pwo.] 愛波”. The partial preservation of the secondary rounding feature in [ ] with [w] is termed “gliding” throughout our research. The other case is deletion [ ], as in “[.p .(o .] Prego → [.p u.k .] 普格”.
As has been discussed in 4.3.5, moreover, the lateral [l] is still the primary mapping for both of the liquids [l] and [ ], which remains true when [ ] is embedded in the cluster.
Second, the adaptations of the [t ] and [d ] sequences reveal the greatest diversity. In the majority, the adapter maps the feature [–continuant] of [t ]/[d] and the [+retroflex] feature of [ ] to the affricate [t ] or [t ], a process termed “coalescence”
(segment fusion) in that both being coronal obstruents, [t ]/[d] and [ ] tend to merge into one but meanwhile each retains a feature specifically pertaining to the stop and
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
134
the retroflex. Cases of coalescence account for 38.64% (17/44). Another equally frequent strategy is gliding, as just elaborated, where only the feature [+round] of [ ] is retained and mapped to the labial (/velar) glide [w], accounting for 31.82% (14/44).
Third, in clusters of the [k ] and [( ] sequences, the primary scenario is again vowel insertion for the velar stop and lateralization for the retroflex, as collected in 89.66% of all such items. Another observation that is worth noting, however, is that deletion of [ ] is only spotted in mappings of the two sequences with the “stop-[ ]”
combination, though the number is rather small (3 instances, 10.34% of all).
In respect to the interpretation of the stops that precede [ ], of all the 58 aspirated stops, 51 (87.93%) are mapped to either the identical aspirated stops, such as
“[.p a s.] Price → [.p u.lai.s!.] 普萊斯”, aspirated affricates, such as “[.t e .s .]
Tracy → [.ts wei.$i.] 崔西”, or unaspirated affricate, such as “[.sæ.k .m n.t o .]
Sacremento → [. a.t ja.mjen.tu.] 沙加緬度”. The remaining 7 (12.07%) are realized as their unaspirated counterparts, as in “[.p æ(.] Prague → [.pu.la.k .] 布拉格”, or unaspirated affricates, as in “[.p i.t s.] Pietrus → [.p i.t wo.s!.] 皮卓斯” (under the influence of coronal coalescence). As for voiced stops, 97.5% (39/40) of them are interpreted as thier unaspirated counterparts, as in “[.b o .k .] Broca → [.pu.lwo.k a.] 布洛卡”, or unaspirated affricates, as in “[.d æ.k o .] Draco → [.t wai.k .] 跩哥” (again, under the influence of coronal coalescence), and only 5%
(2/40) are realized as their aspirated counterpart, as in “[.b d#.ston.] Bridgestone → [.p u.li.s!.t o .] 普利斯通”, or an aspirated affricate, as in “[.d ks.l .] Drexler → [.ts wei.s!.l .] 崔斯勒”.
4.4.6 The “fricative-[ ]” sequences
Some of the fricatives in English may precede [ ] to form a complex onset, yet
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
135
only the [f ] sequence occurs in a high frequency in the corpus. The patterns are shown in (90).
(90) Adaptation patterns of the “fricative-[ ]” sequences L2 seq. L1 struct. Percentage Example (L2 → L1)
[f ] [.f1V1.l2] 94.12% (16/17) [.f o .do .] Frodo → [.fo.lwo.two.] 佛羅多 [.f1] 5.88% (1/17) [.f st.] Frost → [.fo.s!.t .] 佛斯特
[v ] [.w12] 100% (1/1) [.æ.v l.] Avril → [.ai.wei. .] 艾薇兒 [' ] [.s1V1. 2] 66.67% (2/3) [.( . .] Guthrie → [.k .s . wei.] 葛瑟瑞
[.s1] 33.33% (1/3) [.d a v. u.] Drive Thru → [.t .lai.su.] 得來速
[s ] [.s1V1.l2] 100% (1/1) [.s .læ .k .] Sri Lanka → [.s .li.lan.k a.] 斯里蘭卡
[% ] [. 1V1. 2] 100% (1/1) [. k.] Shrek → [. . wei.k .] 史瑞克
Overall, retention of both segments through vowel insertion for the fricative and lateralization for the retroflex approximant remains the primary strategy, accounting for 86.96% (20/23) across the board. A small 8.7% (2/23) of them undergo retention of the fricative but elision of [ ], as can be observed in the [f ], [v ] and [' ] sequences. Notably, in “[.æ.v l.] Avril → [.ai.wei. .] 艾薇兒”, the labial [w] in the output can be said to be the mapping of the labiodental [v], the retroflex [ ] through gliding, or both.