Reframed narrativity in literary translation: an investigation of the explicitation of cohesive chains

3.1 Cohesive chains investigated in a small pilot study

To illustrate explicit analysis of cohesive chains formation in the ST and TTs, a small extract from Chapter 1 has been provided and analysed below.

Extract 1

Note: taken from Chapter One of the novel, Extract 1 provides a mythical account of the origin of a precious stone, which is essential in driving the development of the story, as the stone descends to the human world and bears witness to the life of the main character, Baoyu, throughout the story. The Chinese ST has been tabulated below and divided into clauses, each supplemented with Latin transliteration using Pinyin^[3]; an English translation is provided by the first author of this article. The parallel TT1 and TT2 are provided afterwards.

By tying the cohesive devices and their interpretative sources, i.e. the referent of a cohesive device (Hasan 1984: 190), or by linking similar lexical items, cohesive chains are identified and visualised throughout the text, as shown in Figures 1, 2, and 3 below. To be specific, the pronominal references such as the definite article the and demonstratives such as this and that all refer to specific characters, entities, events and so on alike in the text, therefore, they form identity chains that are the realization of “topical” unity of a text through those pronominal references (Hasan 1984: 184). On the other hand, similarity chains “are a realization of particular portions of semantic fields [;…] they reflect the generic status of the text, and […] contribute to its individuality” (Hasan 1984: 206). In Extract 1, some references and semantically equivalent items form identity chains by referring to the same referent, whereas some other items from the same category or class form similarity chains, which represent the generic features of the text and associate the text with its context of culture. In Figures 1, 2, and 3, the identity chains are visualised in —, while the similarity chains are visualised in ……

Figure 1:

Cohesive chains analysis of the ST.

Figure 2:

Cohesive chains analysis of the TT1.

Figure 3:

Cohesive chains analysis of the STT2.

The process of chain formation can be shown with a few examples. In the ST, the —— identity chain 11e ^[6] –10e–8e–7e–6–4s refers to a stone which can be traced back from the elliptical subject of 此石 cǐshí ‘this stone’ in clause 11 to the same subject ellipsis in clause 10, 8 and 7, to 此石cǐshí ‘this stone’ in clause 6 and to 一块 yīkuài ‘one stone’ in clause 4. Elliptical Subjects had been recovered prior to the chain formation for visualization. All the cohesive devices in these clauses form an IC, referring to 顽石 wánshí ‘one block of stone’ in the text.

In the TT1, the specific block of stone has been referred to as the block in clause 5, as the demonstrative this in clause 6, as the pronominal its in clause 7, and as it in clauses 8 and 9, thus forming an IC 9–8–7–6–5. Apart from this specific block, there are other stones mentioned in the TT1 in chain 4–3–3, referred to as these in clause 4, each in clause 3 and blocks of stone in clause 3. Moreover, the reference of goddess can be traced from the elliptical subject in clause 5 to she in clause 4 and 3, and then to the goddess Nu Wa in clause 1, forming another IC 5e–4–3–1.

Similarly, in the TT2, the block in clause 6 is referred to throughout the text using pronouns and demonstrative reference, from its in clause 7, to this in clause 8, to it in clause 10, and subject ellipsis of it in clause 11, to it in clause 11, 12, 13 and another subject ellipsis in clause 14 till the pronominal its in clause 14. This chain is represented as 14–14e–13–12–11e–10–8–7–6. The goddess in the TT2 is referred to in the chain 5–5–3e–2–1 with pronominal reference she and possessive reference her.

All these identity chains form strings that concern specific entities in the text, the topical development of which contributes to the unity of a text. The investigation of whether these ICs are maintained or shifted in translation will thus be indicative of the development of the general topics in different texts, and of readership design in the translations.

On the other hand, formed out of lexical cohesive categories, similarity chains can also be traced within Extract 1. For instance, in the ST, 石 shí ‘stone’ in clauses 1, 2, 6, 8 are linked via the relation of semblance, referring to the general solid mineral material of stone, not necessarily to the specific blocks used for repairing the sky or the leftover block in this context. In the TT1, the simple repetitions of the sky in clause 2 and 7 refer to the same sky, which is applicable in any text. In the TT2, Green sickness Peak and mountains in clause 7 are related to the Incredible Crags of the Great Fable Mountains in clause 3 via meronymy.

Referring respectively to the same entity within a specific text or a similar entity within a specific genre, these identity and similarity chains together construct the “logical skeleton” (Butt et al. 2010: 273) of meanings within a text. By connecting cohesive devices, cohesive chains can link items of any size at any distance (Halliday and Matthiessen 2014: 607)”. The implicitness or explicitness of this logical skeleton shows the way in which the text realizes its meaning. The denser these links, the more explicit a text is. Considering typological differences between Chinese and English, we measure the density of cohesive chains per clause rather than per certain number of words. Impressionistically, Figures 1, 2, and 3 show denser chain formation in the TTs. This has been confirmed in Table 1 below. First, both TTs have more cohesive chains than the ST. Second, both TTs display higher numbers of chains per clause. They suggest a trend of explicitation in translation.

3.2 Relevant tokens

Apart from the number of cohesive chains, which does not reflect the length of chains, the explicitness of each text is also reflected by the number of cohesive devices that form cohesive chains. According to Halliday and Hasan (1989 [1985]: 93), “all tokens that enter into identity or similarity chains” are defined as relevant tokens (hereafter referred as RT). For example, in Figure 2, there are four RTs that form the IC of the goddess: the goddess Nu Wa, she, she, (she). The RTs in chain formation serve as “patterns that gradually appear in the course of the unfolding” (Halliday and Matthiessen 2014: 601) of a text can be viewed as explicitness indicators dotted throughout the text. They contribute to readers’ perception and interpretation of the text (Morris and Hirst 2006: 41). The more RTs there are, the more explicit the text can be considered. As text sizes vary, the ratio of RTs to the number of words and clauses in a text will be considered a direct exhibition of the explicitness of cohesive chains.

Take ICs in Figures 1, 2, and 3 as an example. In Figure 1, the IC 11e ^[7] –10e–8e–7e–6–4s referring to “the stone” contains 6 relevant tokens in driving “the development of the content of the text” (Hasan 1984: 199). In Figure 2, the referents to “the stone” form an IC 9–8–7–6–5, which contains 5 relevant tokens: block in clause 5, this in clause 6, its in clause 7, and it in clause 8 and 9. In Figure 3, the referents to “the stone” make an IC 14–14e–13–12–11e–10–8–7–6, with altogether 9 relevant tokens: block in clause 6, its in clause 7, this block of stone in clause 8, It in clause 10, (it) and it in clause 11, it in clause 12, it in clause 13 and (it) and its in clause 14. The different numbers of relevant tokens in the ICs referring to the stone shows that TT2 here has a higher degree of explicitness than the ST and TT1.

The total numbers of RTs in the three texts are shown in Table 2 below, which shows higher numbers of RT in both translations than in the ST.

As shown in Tables 1 and 2, the different degrees of explicitness of cohesive chains in the ST and TTs may suggest that there is a trend of explicitation in translation. This is tested in the larger corpus in the next section.

4.1 Cohesive chains in the texts

To compare how cohesive chains are manifested in the translations, all identity and similarity chains are identified in ST and TTs to examine how different texts realize their semantic meanings, as shown in Table 3.

The numbers of ICs and SCs in the texts exhibit clear differences between the Chinese source text and the two English target texts as well as between the two target texts. The ST has the least overall number (671) of cohesive chains as well as IC (351) and SC (320). Firstly, while supposedly conveying the same ideational meaning as do the ST, both TTs have more cohesive chains than does the ST: 803 cohesive chains in the TT1 and 968 in the TT2 respectively. Accordingly, the three texts also show discrepancies in the percentages of the two types of cohesive chains. The ST has the lowest percentage of ICs, whereas the TT1 has the highest. This demonstrates that the ST has the strongest preference for SCs over ICs in developing semantic sequence, whereas the TT1 has the weakest. The two TTs show only less than 1% discrepancy in the ratio of ICs to SCs, but both differ considerably from the ST in the ratio. Secondly, the clauses that are linked by cohesive chains also show a trend of explicitation. From Table 3, we see that each chain on average connects 0.83 clause in the ST, while this number is 1.19 and 1.35 clauses respectively in TT1 and TT2. The explicitness displayed by the connection among clauses through cohesive chains in the three texts is in accordance with the different degrees of explicitness shown by the number of cohesive chains.

The considerably larger numbers of cohesive chains and chain per clause ratio in the two TTs show a trend of explicitation in both translations. As stated before, explicitation in this research is measured in terms of the explicit semantic strings, that is, the cohesive chains. However, the two translations also vary in the degree of explicitation. In general, the Chinese source text is the most implicit one, while Hawkes’ translation (TT2) is the most explicit one, and Yang’s version (TT1) is in between the ST and TT2 in terms of the instances of cohesive chains.

4.2 Relevant tokens in the texts

Cohesive chains reflect degrees of explicit semantic realization in translation, while the RTs that form these chains also contribute to this trend of explicitation. Table 4 below shows that the ST has the least number of RTs while the TT2 has the most.

In accordance with the increasing numbers of RTs from the ST to the TTs, the percentages of them also show a slight increase from 20.41% in the ST to 21.40% in the TT1 and 22.33% in the TT2. By linking the RTs into a network, these cohesive chains form a “grid” (Hasan 1984: 210), which entails the textual coherence of a text. In addition, as the size of texts vary, these RTs are further measured per clause. In the ST, every clause contains the least RTs, 3.3, while the TT2 contains the most RTs per clause, 5.64, and TT1 falls in between, with 4.59 relevant tokens per clause. Thus, the number of RTs per clause also displays a trend of explicitation from the ST to the TT2. If we view the connections among RTs as a measurement of textual structure, the Chinese ST is looser in structure; while the two English TTs are tighter, with more threads of meaning concerning the topical development.

5.1 Cline of explicitation

As stated in the above sections, the use of cohesive chains and relevant tokens in the three texts shows varied degrees of explicitness in translation, with the ST displaying the most implicit patterns of cohesive chains, while the TT2 the most explicit ones, and the TT1 falls in between. The varied degrees of explicitness and explicitation in the process of translation from Chinese into English can be visualized as a cline of explicitation (see Figure 4).

Figure 4:

Cline of explicitation of cohesive chains.

This cline of explicitation implies that the Chinese ST and the English TTs display different preferences in creating semantic continuity in the text, as cohesive chains are parameters of the realization of semantic meaning. In translation from ST to the two TTs, more explicit cohesive devices and chains are required in reconstructing the semantic meaning of the original text and construing the context of situation in the target language.

The varied extents of explicitness suggest differences in the socio-cultural backgrounds of each text and the translators’ personal styles. Kieras (1980) argues that highly cohesive texts are easier to process, in that one may say that it is easier for the TT2 readers to perceive the text and its meaning due to the highest extent of cohesive explicitness displayed.

The trend of explicitation from Chinese to English may reflect the “writer and reader’s responsibility” of different languages for effective communication (Hinds 1987: 146). In writer responsibility languages, “it is the writer’s task to provide appropriate transition statements so that the reader can piece together the thread of the writer’s logic which binds the composition together” (Hinds 1987: 146). In comparison, in reader responsibility languages, the readers need to play “a more active role” in determining the textual relationships (Hinds 1987: 146). From the ST to the TTs, English translations display a higher degree of writer’s responsibility while the comprehension of the Chinese text relies more on readers.

5.2 Cohesive chains and the context

Explicitation reduces the possibility of the target text from being ambiguous and makes more tangible the relation between the target text and the context, as a text “construes a context of situation, in which cultural norms and assumptions are in play” (Lukin 2013: 527). Context of situation can be described from contextual configurations, which consist of field (what the text is about), tenor (the interpersonal relationship involved), and mode (the particular part the language is playing: written or spoken) (Halliday and Hasan 1989 [1985]; Hasan 2009).

In the case of translation, where a ST and a TT are expected to be identical in terms of field, it is mode and tenor that reveal more differences, as the texts are produced by different authors targeting new readers. The Chinese ST initially targeted its contemporary Chinese readers in the eighteenth century, although it has, since then, been read by generations of Chinese readers as a classical novel. On the other hand, both English translations target Anglophone readers after the mid-1970s. Therefore, the translators would have taken into consideration not only the linguistic and cultural gaps between Chinese and English but also the considerable temporal gaps. Moreover, the two translations can be said to have been created with slightly different readership design. The Yang couple worked for the Beijing Foreign Languages Press, which is in charge of promoting Chinese culture in the world by publishing books and journals in non-Chinese languages; their translation is more suitable for readers with high cultural attainments in Chinese. In comparison, Hawkes, a sinologist, translated the novel out of his interest, and his version targets more general Anglophone readers (Cao 1973: 18). This difference suggests that Hawkes may have adopted the explicitation of cohesive devices more actively as a strategy to reduce the “responsibility” for his readers who may be less informed about Chinese culture.

The factor under mode that makes a difference across the texts is Medium, which refers to the “patterning the wordings themselves” (Halliday and Hasan 1989 [1985]: 58). The use of cohesive chains construes the patterns in different texts by driving text development through the narrative existents (characters, events, and subject), which remain the same from one event to the next in a narrative structure (Chatman 1980: 30). As ICs are text-bound, they contribute to the creation of the context of situation of each text. The higher percentages of ICs than those of SCs in both English TTs than in the ST indicate that the perceptions of the TTs rely more on the context of situation. On the other hand, as the SCs are not text-bound, they relate the text with its context of culture. According to Table 3, the two TTs use fewer SCs than the ST, which suggests that the perception of the ST relies more on the context of culture. In short, as the target text readers are temporally and spatially distant from the source context of culture, the explicitation of cohesive chains and the increased use of ICs in translation help the TT readers understand the source context of culture within the target context of situation.