Article body

1. Introduction

News-reporting texts are one of the genres that translator trainees encounter most frequently but often find difficult to translate well. News reports are considered appropriate to deal with in the early stages of a translation program (Nord 1991: 20; Kelly 2005: 123) and are placed in the category of general language translation as opposed to special language translation (e.g. legal and medical texts) (Snell-Hornby 1988: 32). Similarly, in the translation program of an educational institution, news article translation tasks are usually offered at an early rather than late stage (e.g. Macquarie University’s Translation and Interpreting Program). Newspaper articles are a typical type of text: that is, they are extremely familiar – so familiar that we assume that they cannot be mistranslated. However, these types of texts often betray such expectations.

One reason for this conundrum relates to genre. Genre points to the generic identity of a text, and one aspect of the meaning of a text is its relationship to text types (Eggins 2004: 55). The genre of news reporting has its own generic structure and wording that is distinct from other genres. The typical patterns in the structure of different genres are realized in different lexico-grammatical choices (lexis and grammar, Halliday 1994: 15). Thus, the genre structure of a news-reporting text and its correlation with its lexico-grammar are, together, a meaningful choice in a specific context.

This paper seeks to answer the following two questions. Firstly, what makes it difficult for students to produce an adequate translation of a hard news-reporting text, one of the most common types of texts? Secondly, are there any patterns in translation errors caused by lack of knowledge of the generic structure? In answering these questions, a systematic approach is attempted: the idea is not to fix all individual translation problems, but to point to a direction for possible remedies by drawing on the genre theory and lexico-grammar of Systemic Functional Linguistics (SFL). The paper will look at the concept and configuration of genre (Martin 1997), particularly those of news reports (White 1997), as a means of grounding the ensuing discussion. There will be discussion and interpretation of general features of translation errors, followed by specific examples of translation errors; these errors form certain patterns in the student translations at the three levels of above-sentence (genre), sentence and below-sentence. In this process, attention will be given to how the genre structure of news reports and logical meaning embedded in the texts together comprise “structures of meaning” (Candy 1989: 108). These structures, in turn, may serve as a criterion for analysing the error patterns of student translations of news-reporting genre.

2. Genre

2.1. Halliday’s register and Martin’s genre theory

Genre has become an important part of translation studies, but the concept has yet to be clearly defined for application in the discipline. In the 1990s, the concept of genre was popular across various disciplines including translation studies. This popularity is reflected in Trosborg (1997), Bhatia (1993) and Hatim and Mason (1990), to name a few. But a consensus on the notion of genre has not yet been reached (Trosborg 1997: 3): genre is still a slippery concept that is often used interchangeably with others, such as register and text type (Lumeras 2010: 216). This situation begs for clarification of the concept of genre before trying to apply it to translation education. The concept of genre is important for translation education because it serves to configure meaning in a text.

In Systemic Functional Linguistics, a functional variety of linguistics theory, the notions of register and text type are used in the concept termed the “cline of instantiation.” At one end of the cline is the language system, identified as the underlying potential. At the other end is a text, an actual instance of the potential. Halliday draws an analogy to the relationship between climate and weather to illustrate the concept (Halliday 1992: 8). We can visualise the two ends of the cline as the system pole and the text pole. Between the two poles, recurring patterns of language are found in the form of text type from the perspective of the text pole; from the perspective of the system pole, with the addition of the concept of context, they can be interpreted as register (Halliday and Matthiessen 2004: 26-27). Thus register and text type can be considered the same phenomenon viewed from either end of the cline.

In SFL, the concept of register is extensively researched. Employing the ‘context of situation’ and the ‘context of culture’ coined by Malinowski (1935), Halliday spent a considerable portion of his work on register, a slot set aside for the immediate ‘context of situation.’ He defines register as a functional variety of language with situational configuration of field (what is going on), tenor (who is taking part) and mode (role assigned to language) (Halliday 1978; Halliday and Hasan 1985).[1] However, his work mostly concentrated on the ‘context of situation,’ with little attention given to the ‘context of culture.’ In fact, he treated genre as part of mode (register), as a result of which the ‘context of culture’ remains largely undefined.

Genre was separated from register by Martin and his colleagues. In particular, they placed genre “at the stratum of culture, beyond register” (Martin and Rose 2008: 16) in the ‘context of culture.’ By putting genre beyond register in the context levels, genre was realized in “a pattern of field, tenor and mode patterns” (Martin and Rose 2008: 16) in their model (see Figure 1 below). In this model, register and genre do not stand in a relationship of part and whole, as in Halliday’s model, but in a relationship of complementarity at the context levels.

Figure 1

Genre as an additional stratum of analysis beyond, tenor, field and mode

Genre as an additional stratum of analysis beyond, tenor, field and mode
From Martin and Rose 2008: 17[2]

-> See the list of figures

Martin’s model is reflected in his definition: genre “refers to the staged purposeful social processes through which a culture is realized in language” (Martin and Rothery 1986: 243). This means that genre is used by the members of a discourse community to express its goals (Swales 1990: 52) in a way that is coded in generic structure and “stages” (Martin and Rose 2003: 8) (in social institutions such as school, government and legal systems). Martin’s model is useful for translation education because it offers a theoretical framework for identifying meaning structures of a text in a professional context (e.g. news stories and legal texts), a framework that register alone cannot provide.

Martin suggests four types of genre structure: ‘orbital,’ ‘serial,’ ‘prosodic’ and ‘periodic’[3]. Of these four, the ‘orbital’ structure can be identified in news story texts (White 1997). It comprises a configuration of one nucleus with multiple numbers of satellites that depend on the nucleus as in the relationship between the sun and its satellite planets. The ‘orbital’ structure in news story texts will be discussed in more detail below.

2.2.White’s hard news-reporting genre

Orbital’ structure, one of the types of genre structure developed by Martin as fully applicable to different genres, is adopted and employed by White (1997) for news story texts. The news story genre analyzed in White (1997) refers to ‘hard news’ in journalistic terms. Hard news reports are carried generally in front pages of newspapers with the purpose of communicating newsworthy reports (Feez, Iedema and White 2008). According to White (1997), a journalist-turned-linguist, hard news reports are normally associated with events and statements that threaten the social order. These events and statements that destabilize the social status quo take the form of “eruptive violence, reversals of fortune and socially significant breaches of the moral order” (White 1997: 101). The sources of the social disruption include natural disasters, accidents, diseases, economic downturn, political/military power relations and breaches of moral obligations (see Feez, Iedema and White 2008 for soft news reporting such as human interest stories as well as editorials and features).

Hard news reports come in two subtypes. One is an event story that mainly describes what happened at first-hand as if the author was at the scene, and the other is an issues report that contains statements of authorized sources such as politicians, experts and researchers. An issues report[4] is particularly anchored in communicative events such as a speech, an interview or a press release. It is grounded in direct and indirect speeches much more frequently than an event story, which in turn means different strategies in lexico-grammatical choices between the two subtypes are required.

In terms of meaning structure, however, an event story and an issues report share the same generic configuration. Metaphorically, they have the same ‘orbital’ structure where the nucleus (headline and the opening sentence) comes at the centre, and is surrounded by a number of satellites (the remaining text coming after the headline and the opening sentence in the text). The ‘orbital’ structure of the news story genre is shown in Figure 2 below.

Figure 2

An example of orbital structure

An example of orbital structure
Adapted from White 1997: 121

-> See the list of figures

More specifically, the headline and opening sentence(s)[5] (also called ‘lead’) comprise the nucleus, and the rest of the text makes up the satellites. Each satellite comprises one paragraph or more. What is noteworthy in the nucleus is that the content of the headline is closely reflected in the lead. For instance, the headline is echoed in the opening sentence of the text below. The numbers 1 to 5 in the headline correspond to the same numbers in the opening sentence.

If we turn to the relationship between the nucleus and its satellites, the nucleus does not represent a summary or an announcement of what is to come in the satellites, but stands as the maximal peak of the social disorder being covered in the text (White 1997: 128-129). This often makes the lead sound abrupt, posing one of the most serious translation problems in a news-reporting text. In such a text structure, the satellites do not add or develop totally new meanings as the texts unfold, but keep referring back to the headline/lead nucleus to make sense through such relationships as ‘elaboration’ and ‘explanation’ as explained below. Thus the satellites do not have a chronological relationship with their adjacent satellites, nor are the satellites necessarily linked through, for example, a ‘causal relation.’ White (1997:116) dubs this phenomenon ‘radical editability’ to reflect the fact that satellites have a certain amount of freedom to move within the satellite phase. It reveals that, by retaining, omitting or prioritising satellites in a certain order, differing ideological orientations can be displayed over the same communicative event. But this will not be pursued any further here because it goes beyond the scope of this paper (refer to White 1997 for details).

3. Logical meaning

So far the discussion has concentrated on the contextual aspect of a text, but the levels of context are connected to the levels of language (Halliday and Matthiessen 2004). As shown in Figure 3 below, genre and register (context levels) are encoded in language. At the language levels, meaning is then expressed in wording (lexico-grammar). In this model, patterns of wording express patterns of meaning, and patterns of the language levels manifest patterns of context. Thus, context, meaning and wording closely relate with one another in a text.

Figure 3

Language and its semiotic environment

Language and its semiotic environment
From Martin 1992: 496

-> See the list of figures

Meaning plays a central part in a text. In SFL, there are four modes of meaning (or metafunction) often simultaneously operating in a text: experiential (what happened), interpersonal (social interactions), textual (coherence of the text) and logical (relations between experiential meanings). Of these, logical metafunction is realized by ‘combining’ or ‘complexing,’ which concern the relationship between clauses, groups/phrases and words: it is related to the decision to choose between, for example, a simple clause vs. a combination of clauses; and connectors that link clauses with such relations as ‘cause and effect’ and ‘condition’ (hereafter called logico-semantic relations).

More specifically, the same message can be put in a single sentence or multiple sentences in a written text, but the logical meaning of the two forms may not be the same. Thompson aptly mentions that “the difference between expressing messages in separate simple clauses and in complexes is a meaningful choice” (Thompson 2004: 195). Also, in a combination of clauses (hereafter called clause complexes), connectors play a vital role between clauses. They take the form of many different logico-semantic relations such as ‘elaboration’ and ‘extension.’ These relations are tabulated in Table 2 under 6.1. Methodology (refer to Munday 2000; Kim 2007a; 2007b; 2009 for studies on textual meanings of SFL).

Logical meaning is, however, closely related to the context levels as well. The concept of clause combining should not be restricted to clauses only, but can be used to explain linguistic phenomena above sentence-level (conjunction) (Martin and Rose 2003: Chapter 4). In fact, this is in line with White (1997) and Feez, Iedema and White (2008) in extending the notion of clause combining and logical meaning to the relationship between the nucleus and its satellite paragraphs. Thus, this paper does not limit clause combining and logical metafunction to the sentence level and below, but extends them to cover logical relationships at the context (genre) level as well.

4. Correlation between genre and logical meaning, and its translation implications

As logical meaning extends to genre, it highlights the ‘structures of meaning’ that are realized by paragraphs in a hard news-reporting text. These paragraphs form a meaningful relationship at the context level as the headline/lead or its satellite, where the context level (genre) and lexico-grammar (combining/complexing) are correlated. This interconnection enables what genre alone cannot address (local issues such as relations between groups/phrases or words) and what lexico-grammar cannot achieve on its own (most global issues affecting the overall structure of a text).

In Table 1 below, a local issue that might occur in translating the relation between Rod Blackmore, and senior children’s officer for 17 years in Specification 1 cannot be resolved with genre: they need to be handled with lexico-grammar (the logico-semantic relation of ‘elaboration’ as displayed in Table 2 below). Likewise, the generic structure of a hard news text cannot be covered by grammar, but needs to be dealt with by the genre structure of nucleus and satellite as specified in Figure 1. However, where lexico-grammar and genre converge, they can together reach the areas each of them cannot cover on its own. In the example below, the genre structure of a hard news text is intertwined with lexico-grammar: the headline/lead and its satellites are correlated with logico-semantic relations of ‘elaboration’ and ‘justification’ (cause and effect) (for detailed explanation of these relations, refer to 6.1. Methodology).

The association between genre and logical meaning has important translation implications. Apart from the nucleus paragraph with “abrupt” lead sentence(s), a hard news-reporting text has satellite paragraphs that have no relation with one another. Inherently, they are not chained by temporal or logico-semantic relations, but keep going back to the headline/lead paragraph for meaning-making. This characteristic of satellites can inform translation decisions such as whether to divide or combine paragraphs, or add a conjunction between them.

Table 1

Knives Ban (issues report): specification of Headline/lead analysis

Knives Ban (issues report): specification of Headline/lead analysis
From White 1997: 117

-> See the list of tables

Particularly in an issues report text, the correlation between genre and logical meaning is complicated by the high frequency of direct and indirect quotations (e.g. the one-sentenced satellite under Specification 3 above). As part of logical meaning, direct and indirect quotations play a crucial role in developing the text because they comprise the foundation of the communicative events-oriented news report. In translating an issues report text, therefore, these three elements (genre, logical meaning and quotations) should be considered as integral parts of the text.

5. Translation brief and comparable texts

In this paper, the translation brief and comparable texts are not considered as criteria for error analysis because their roles in a news story are relatively small. Admittedly, the translation brief and comparable texts are complimentary to genre analysis in the translation of a news-reporting text. The translation brief provides specific guidelines including the functions and target readership, while comparable texts offer ideas of how the target text should look. Both of them certainly enrich the final product of translation. However, their functions are small in comparison to the role of genre structure in a hard news report. Student translations in this paper show a minimum number of translation errors caused by not paying attention to the translation brief, with a much larger number of translation errors coming from genre structure.

Using comparable texts for translation of a hard news report is more complicated for two reasons: 1) the ‘radical editability’ (White 1997: 116) mentioned above, and 2) different hard news writing conventions between Korea and the English speaking world. Firstly, utilizing ‘radical editability,’ journalists or news companies often edit the same news material differently in a way that serves their ideological interests. In the process of editing, paragraphs may be combined or divided differently from the originating text, and certain portions may be adopted or removed. This would make it pointless to refer to a comparable text written in Korean as an original text.

Secondly, news-reporting conventions in Korea do not seem to display as much ‘radical editability’ as in English news-reporting conventions. Anecdotal analysis conducted by the author suggests that more conjunctive linkers are found between paragraphs in news stories in Korean newspapers than in Australian newspapers (see Lee 2011 for Korean news writing). This discrepancy could make student translators confused as to which convention to follow for their translations. In this connection, Lee (2006: 319) argues that Korean news reports follow the American inverted pyramid model where the most important information comes at the top of the text and the least substantial is at the bottom. In this argument, however, the focus of the inverted pyramid model is basically on the lead, not the satellites phase. Moreover, the last paragraph of a news story text is not necessarily the least important (Feez, Iedema and White 2008: 118). Thus, a separate study needs to be conducted on the different genre features covering not just the lead but the whole of a news story between Korean and English.

6. Methodology and data

6.1. Methodology

In analysing student translations of the news texts, two tools are used: the nucleus and satellites relations proposed by White (1997), and the clause combining table put forward by Halliday and Matthiessen (2004: 380). The former provides a criterion in the dimensions of context, specifically genre, and the latter, in the dimensions of lexico-grammar. The nucleus and the satellites are related by logico-semantic relations such as ‘elaboration,’ ‘contextualization,’ ‘cause and effect,’ and ‘justification’ as below (only ‘appraisal’ concerns interpersonal meaning). These relations are used as guidelines for determining paragraph combining as well as for making meaning between the nucleus and the satellites.

  • Elaboration: one sentence or a group of sentences provide more detailed description or exemplification of information presented in the Headline/lead, or acts to restate it or describe the material in the Headline/lead in different terms;

  • Cause-and-effect: one or more sentences describe the causes, the reasons for, the consequences or the purpose of the ‘crisis point’ presented in the Headline/lead;

  • Justification (issues reports): one or more sentences provide the evidence or reasoning which supports the newsworthy claim presented in the Headline/lead nucleus. This justification could be seen as a text internal cause-and-effect in that it explains why a particular claim has been made and hence could be included within a single broader ‘Cause-and-effect category;

  • Contextualization: one or more sentences place the events or statements of the Headline/lead in a temporal, spatial or social context. The geographical setting will be described in some detail or the ‘crisis point’ will be located in the context of preceding, simultaneous or subsequent events. Prior events of a similar nature may be described for the purpose of comparison;

  • Appraisal: elements of the Headline/lead nucleus are appraised, typically by some expert external source, in terms of their significance, their emotional impact, or by reference to some system of value judgement.

White 1997: 115

In Systemic Functional Grammar (SFG), there are two subcategories of clause- combining types: degree of interdependency (whether or not one clause depends on the other) and logico-semantic relations (‘expansion’ and ‘projection’). ‘Expansion’ is subdivided into three subtypes: ‘elaboration’ (meaning is specified or described), ‘extension’ (addition, replacement or alternation of meaning) and ‘enhancement’ (meaning is qualified by time, cause and effect, condition, etc.). ‘Projection’ covers both direct and indirect speeches and thought. The concept of ‘projection’ can be better understood as a balloon or a cloud being projected above a character speaking or thinking in a cartoon (Halliday and Matthiessen 2004: 454). Being part of logical meaning, ‘projection’ is one of the translation problems in an issues report text students find difficult. The clause-combining (or clause complex) types are as shown below.

Table 2

Basic types of clause complex

Basic types of clause complex
From Halliday and Matthiessen 2004: 380

-> See the list of tables

As for the procedure of analysis, translation errors for the two news-reporting texts were broken down into four types: experiential, interpersonal, textual and logical. Then, logical meaning errors were subcategorized into translation errors at above-sentence level, at sentence level and at below-sentence level. Here, the above-sentence level corresponded to genre structure level because most of the paragraphs in the texts are one sentence in length. After this, logical meaning errors involving direct and indirect quotations (‘projections’) were counted to see if there was any relation between the number of ‘projections’ and the number of errors at genre structure level. Finally, relevant examples were identified and developed to demonstrate the cases in point of logical meaning errors at the three levels of above-sentence, sentence and below-sentence. In discussing the qualitative analysis, more instances were used from the issues report text than from the event story text, where a higher number of translation errors were found.

6.2. Data

Two English source texts (STs) and 18 Korean target texts (TTs) along with their learning journals were analyzed. As part of weekly assignments, the translations were produced by nine students enrolled in the unit entitled Translation Practice of the Postgraduate Translation and Interpreting Program at Macquarie University, Sydney, during the second semester of 2009, when I worked as marker (English-to-Korean translation). Most of the students did not have formal training in translation before joining the program, and this was their second semester in the postgraduate translation program. At the end of the semester, students were to take their final exam, which was at the same time a test for accreditation as a professional translator. The Accreditation is awarded by the National Accreditation Authority for Translators and Interpreters (NAATI), the Australian accreditation organization. As the NAATI standards for professional translators focus on “correspondence, reports, standard text material in the general field of scholarship” (Accreditation Standards for Translators and Interpreters in Australia[6]), a majority of texts used in class were news articles and government information materials.

The source texts for this paper are news reports extracted from the Sydney Morning Herald and the Financial Times, with the word count ranging between 250 and 300 words. The first text is an event story (Smith 9 April 2009; titled Australia’s jobless rate surges to 5.7%)[7] and the second one is an issues report (Patty 4 August 2009; titled Dud teachers caned by union).[8] The source of the social disequilibrium of the event story is an economic downturn, while that of the issues report concerns a breach of moral obligations (the two texts are attached as appendices 1-2 with analysis at genre level).

7. Discussion and interpretation

A majority of translation errors are identified as logical meaning errors in both texts. Of the logical meaning errors, the above-sentence (genre level) translation errors account for a significant portion in both texts, though to differing degrees. In this process, direct and indirect quotations seem to have played an important role in the issues report text, complicating the correlation between genre structure and logical meaning. These findings indicate that the interconnection between genre structure and logical meaning, as well as each of these two elements on its own, are the main reasons underlying difficulty in translating a news-reporting text. Particularly, lack of knowledge by the student translators of the triad (genre, logical meaning, and direct and indirect quotations) in the issues report may have led to translation errors including combining unrelated satellites and, in doing so, often inserting misleading logical connectors between them.

7.1. General features of translation errors related to correlation between genre and logical meaning

Table 3 below shows translation errors from each of the nine students (S1…S9) across four modes of meanings in the event story text (Australia’s jobless rate surges to 5.7% Smith 9 April 2009; see note 7). In this table, logical meaning errors are by far the most common. The number of logical meaning errors (58.7%) is more than twice as high as experiential meaning errors (27%), and six and eleven times higher than interpersonal (9.5%) and textual meaning errors (4.8%), respectively. Thus, the dominance of logical meaning errors highlights the fact that logical meaning constitutes a significant part of the student translation errors in the event story text.

Table 3

Share of logical meaning errors out of four meanings in the event story text

Share of logical meaning errors out of four meanings in the event story text

-> See the list of tables

In Table 4, logical meaning exhibits the highest share of translation meaning errors (52.3%) in the issues report text (Patty 4 August 2009; see note 8) as well, though there is a discrepancy in the number of translation errors between the two texts with respect to the projection category. Thus, this table confirms that logical meaning plays an important role in analysing the student translation errors in both subtypes of the hard news story genre.

Table 4

Share of logical meaning errors out of four meanings in the issues-reporting text

Share of logical meaning errors out of four meanings in the issues-reporting text

-> See the list of tables

However, there is a notable difference in the total number of translation errors between the two tables. In Table 4, the number of translation errors is 63, whereas it is 111 in Table 5, or nearly twice as many. The difference is considerable, given that there is only a 12% difference in word count between the event story and the issues report texts (264 words and 295 words, respectively). This gap leads to an assumption that the issues report text has posed more challenging translation problems to the students.

As examined above, logico-semantic relations organize the overall meaning structure of hard news story genre. Also, as demonstrated here, logical meaning (including genre structure and lexico-grammar as argued in Section 3) comprises an essential part of the student translation errors in both event story and issues report texts. To investigate more closely the seemingly entangled correlation between genre structure and logical meaning, logical meaning errors in student translations were further broken down into three subcategories: above-sentence[9]-level (genre level) translation errors, sentence-level translation errors and below-sentence-level translation errors.

With logical meaning regarded as extending to genre and thus covering the lexico-grammar up to the context levels, these three subcategories can be classified depending on the gravity of a translation error. If a translation error affects the genre structure of a text, it is regarded as an above-sentence-level translation error. It is a sentence-level translation error if the impact it has is limited to sentence(s) while it is a below-sentence-level translation error if its impact does not go beyond the boundaries of groups/phrases, words or below. In this vein, above-sentence-level translation errors are considered most serious, whereas below-sentence-level translation errors are the least serious unless their impact surpasses the limit of groups/phrases or words. For this reason, more examples of above-sentence-level translation errors are presented than the others below.

In Table 5 below (event story text), sentence-level errors show the highest percentage, whereas above-sentence-level errors top the subcategory in Table 6 (issues report text). With no special motivation identified for such different results, the discrepancy in number of ‘projections’ (quotations) indicated in tables 3 and 4 above (0 vs. 10) may provide an explanation, because it is supported by student learning journals. While no student mentioned ‘projection’ as a translation problem in their journals for the event story text, seven out of nine students expressed puzzlement at the ubiquitous presence of ‘projections’ in the issues-reporting text. One of them was even worried that too many ‘said’ and ‘say’ might make her translation sound “too boring” (Appendix 3) (say or said appear nine times plus one called for in 12 sentences of the issues report text).

Table 5

Share of sub-categories of logical meaning errors in the event story text

Share of sub-categories of logical meaning errors in the event story text

-> See the list of tables

Table 6

Share of sub-categories of logical meaning errors in the issues report text

Share of sub-categories of logical meaning errors in the issues report text

-> See the list of tables

As discussed above, however, ‘projections’ comprise the backbone of an issues report text because it is grounded in opinions of others. In this sense, ‘projections’ are intertwined with meaning structure of an issues report text more closely than with that of a news story text. Given this, the higher number of projection-related translation errors in the issues report text may have contributed to the high number of mistranslations at the above-sentence level (genre level) in the text. In turn, this may have led to a higher number of translation errors. A detailed explanation is provided on this correlation with an example in Section 7.2.3.

Also, more translation errors with wrong conjunctions are found in the issues report text (11 vs. 2). Conjunction-related errors are observed as translation errors committed at the above-sentence level because they are inserted between satellites. These errors come from the failure to understand that satellites have no logical relationships with each other, but instead refer back to the nucleus to make sense of the text. This and the previous points discussed above will be dealt with specifically with reference to the examples below.

7.2. Above sentence (genre level)

7.2.1. Headline and the opening sentence translation errors

The headline is one of the trickiest parts to translate in non-literary information-oriented texts such as a news article, particularly in the English-Korean language combination. English titles tend to be concise and allusive while Korean titles are “direct and explanatory” (Kirk 2001: 191). In the headline of the issues report text below, the word caned may invoke an appropriate imagery of punishment in the school environment in English, but the word-to-word translation of the allusion, as in TT 1 below where it is rendered as is (회초리 [cane]), does not work in the Korean target text. The literal translation could engender an interpersonal side-effect in such a way that the union is higher in status to the point of humiliating underperforming teachers by using a cane meant for children in the target culture where socio-linguistic aspects (e.g. politeness and honorification) are particularly important.

To deal with the issue, it would be useful to refer to the nucleus of which the headline is part, together with the opening sentence. These two elements are considered a subset of the text as the headline signals what is to come in the opening sentence. In the first sentence of the text, caned is reworded as should leave the profession. Undoubtedly, criticized could be one candidate for caned because both words do not involve any action. But, given the convention of information-centred text in the target language that headlines are direct and explanatory of what is to come in the text, should leave the profession can be a more viable option. One more option is to go further down to paragraph 7, where should leave the profession is paraphrased as should be dismissed. Coupling the headline and the first sentence can thus give an anchor in translating the headline of a hard news-reporting text.

7.2.2. Headline/lead and satellite translation errors compounded by a conjunction

Familiar with other types of genres with chronological sequence or transitional signals, students try to establish unnecessary logical relationships between paragraphs in their translations, when there is none in the source text (news reports). This unsolicited addition of conjunctions is a translation error in itself as well as tending to confirm translation errors on other fronts.

In example (3), the TT combines two paragraphs in the source text into one, with the conjunction 왜냐하면 [because] added. Though short, these two paragraphs are not related in the ST. Rather, the first paragraph is related to the nucleus by ‘elaboration’ (of more needs to be done) and the second paragraph by ‘justification’ (to point to a case in point to explain more needs to be done). Moreover, the added conjunction in translation (because) seems to have supported mistranslation in part of the first sentence below. The TT renders the segment in the ST calling for additional professional learning time in all schools into 모든 학교에 확대, 시행되어야 한다고 [calling for expanding (the program) to all schools]. The conjunction placed in the satellites is thus a mistranslation, compounding misunderstanding, and leading to mistranslation of its co-text.

7.2.3. Satellite relationship translation errors compounded by projection

The example below shows how a satellite relationship compounded by quotations affects the student in combining paragraphs and sentences. As shown below, two paragraphs in ST (4) are combined into one in TT (4). In the source text, the first paragraph comprises one sentence, and the other paragraph is composed of two sentences. The first paragraph makes up one satellite that refers back to the headline/lead through the relation of ‘elaboration’ (of more needs to be done in the lead). The other paragraph is analyzed as comprising a separate satellite of the nucleus by the ‘justification’ relation (of more needs to be done in the lead). This second paragraph might be seen as related logico-semantically to the first paragraph, but considering that it can move anywhere within the satellite phase independently of the first paragraph (‘radical editability,’ White 1997: 116), it is analyzed as one separate satellite in this paper (refer to Appendix 2 for analysis of the whole text). Both satellites are complete with an indirect quotation.

However, TT (4) combines these two paragraphs into one long sentence. In doing so, it renders the content of the two indirect quotations (projected clauses) as one lengthy direct quotation. In this process, it replaces the introductory clause of the second quotation (projecting clause he said) found in ST (4) with a connector (concessive but). Moreover, TT (4) adds a projecting clause ([Deputy President Zadkovich] backed up) toward the end of the sentence. This restructuring in TT (4) leads to a completely new structure composed of one sentence where three projection clauses (from saying to [approximately 100 million KRW],”) are dependent on the last two clauses (‘Deputy President Zadkovich […] backed up that only 50 public school […]), as shown below.

The reason why the student has arrived at this structure is not clear from the translation, but her learning journal tells how the high frequency of projection clauses affects her in devising her translation strategies. One of the strategies is to diversify the expression of say by using different expressions such as convey, and the other is to eliminate subjects (e.g. Mr Zadkovich, the union). The details are as follows:

In this week’s text, every sentence has quotation except for only one sentence. If I translate as it is written in English, it would be like this – 노조는… 말했다 (the union… said). 부회장은… 말했다 (the deputy president… said). 부회장은… 말했다 (the deputy president… said). 노조는… 말했다 (the union… said). The original text itself doesn’t seem awkward at all because it is a mixture of direct and indirect quotations, or someone or a group’s name and pronouns, and because of the order of words or clauses. However, when reproducing in Korean, somehow it seemed a little awkward, so I use different words for 말했다 (said) such as 전했다 (conveyed), 밝혔다 (made clear), 덧붙였다 (added) etc., and I omitted the subject – names of a person or group…

Excerpt from Student 1’s journal as reproduced in Appendix 3; only the content in the parentheses is translated by the author, and the last clause is underlined by the author

In the underlined part of the learning journal, the student writes that she omits subjects as part of her translation strategies. However, she omits not only subjects, but their accompanying verbs (he said) as well in TT (3). Then, she collapses the remainders into one long sentence which is unlikely in a hard news text. In this process, she unwittingly combines two different satellites into one. Then she combines the collapsed sentence with the last sentence of the source text, and adds a projecting clause ([Deputy President Zadkovich] backed up) to it. By doing so, the student changes the voice of the ST author to that of Zadkovich. But, in ST (3), it is not Zadkovich but the author of the news report who utters the last sentence (Only 50 public schools out of more than 2200 in NSW had been selected to drive the improvements) (the issue of the reporter’s voice will be discussed in more detail in Section 7.3.1.).

Taking all these together, the student seems disoriented in rendering the meaning structure of the issues report text, affected by the high frequency of projecting clauses. If she was aware of the genre structure of a hard news text and its correlation with lexico-grammar, plus the roles projection plays in the issues report text, she would have better rendered the ST author’s intentions.

7.3. Sentence (clause complex) level: sentence combining

Thus far, representative examples of student translation errors have been analyzed at genre level. In this section, the discussion will focus on translation errors related to combining of sentences. These sentence-level translation errors are found between two sentences, but what is interesting here is that they are also associated, albeit indirectly, with the nucleus-satellite (genre) structure.

In the translation below (Example [5]), two sentences are rendered as conjoined and attributed to the same speaker, the union. In contrast, in the ST, the first sentence is encoded as an utterance by the union while the second one is not. We can understand the source of this translation problem if we compare this paragraph with the headline/lead of the ST. These two paragraphs are in a relationship of ‘elaboration’ (underperforming teachers should leave the profession and more needs to be done to address teacher quality). In other words they share the same basic message, but express it differently. In the headline/lead, the two clauses are expressed as sourced from the same interlocutor (the union) while in ST (5) the clauses are expressed as coming from two different interlocutors (the union and the writer of this news report, respectively). This sort of variation in language use is often found in news-reporting texts. At first sight, it might not seem as serious a translation error as the ones discussed above, but if compounded could become a significant error. TT (5) fails to grasp this delicate distinction and mixes two different voices into one, the union. Thus, as shown in this example, translation errors at sentence level are not necessarily limited to the scope of sentence but may extend to genre level.

7.4. Below sentence level: elaboration

The most common type of translation errors identified at the below-sentence level in the student translations of the two texts arises from ‘elaboration’ (meaning is specified or explained). In the source text below, the group a smaller decline than in previous months has the ‘elaboration’ relationship with the preceding clause. In TTs 6(a) and 6(b), however, the logical relation is rendered as ‘extension’ (addition, replacement or alternation of meaning) and ‘enhancement’ (meaning is qualified by time, case and effect, condition, etc.), respectively. In TT 6(a), the connector of the first clause ‘-으며-umye [and] does not match the theme 이는-inun [which is] in the second clause because the first clause signals the relation of ‘addition’ or ‘simultaneous action’ while the second clause is used to explain the preceding clause. In TT 6(b), the connector ‘-어-u [and so]has the logical meaning of linking the preceding ‘reason’ clause and the ‘resultative’ clause coming after it. However, it is the relation of ‘elaboration’ that links the clause and the nominal group in the ST, not the relation of ‘cause and effect.’ (One possible solution would be the elaborating link –인데indey.’)

8. Conclusion

In this paper, it has been argued that genre, correlated with logical meaning, is at the heart of student translation errors in a news story text. The generic structure of the print media news-reporting text offers a unique ‘context of culture,’ as distinct from other genres such as narratives and historical recounts (Martin and Rose 2008) with which students are more familiar. It was found that the nucleus and satellite configuration is closely correlated with logical meaning at three levels (above sentence, sentence and below sentence) that govern translational decisions of whether to combine sentences and paragraphs. In an issues report, particularly, this correlation between the nuclear-satellite structure and logical meaning is further complicated by the frequent presence of direct and indirect quotations. Lacking this information, a large majority of student translations showed errors at the above-sentence level (genre level).

In the translation of a news-reporting text, genre structure and its correlation with lexico-grammar therefore have something to offer in the context of translation education. Genre structure and logical meaning, most importantly, provide the skeleton of meaning of the given text which students can use as an anchor for meaning-making. The nucleus-satellite configuration, in short, acts as a guide through the different phases and relationships of the meaning structure so that student translators will not lose a sense of direction.[12] Genre, in a hard news story text, therefore, plays a pivotal role.

Genre, however, is a concept with great development potential, particularly for translation education. Some of the development has already been achieved but many genres currently developed are for secondary school education (e.g. Veel 1997). Professional settings relevant to translation (e.g. economic, legal and IT) need full-fledged genre research. Those professional settings, whose genre structures are fully identified, are still limited in number (e.g. Bhatia 1993), and although identified, need further consideration specifically for translation. In order to see more development in professional settings for translation, as in the example of the hard news-reporting genre, close cooperation is necessary between translation studies and the experts from the aforementioned relevant professional discourse communities.

I would like to wrap up this paper by noting some students’ learning journals which caught my attention in the process of analysing student translations. In the short reflective notes, some of the students expressed deep frustration at the overall structure of the hard news text with many quotations abounding in the text (Appendix 3). For me, this indicates how important it is for a student translator to grasp the language of the discourse community concerned. The language of the relevant discourse community requires a focus not only on its background knowledge and register but also on the genre structure and grammar used to realize the generic configuration. Therefore, acquiring a given genre structure is an essential element that helps empower students to acquire much needed self-confidence and self-reliance which, in turn, will lead to the translator competence necessary to move into the community of professional translators (Kiraly 2000: 58).