HPEN7004
First Name: XIAO
Last Name: CHEN
Student No. 1904593

Critique on Application of Systemic Functional Linguistics to Translation Evaluation
1.	Introduction
Translation practice can date back to times immemorial even to ancient Greece and Rome. Many principles regarding translation practice have been proposed by translators over the ancient time. However, it is not until the last century that translation study becomes an academic discipline. Since the 1950s there appeared more and more publications devoted to translation study: by J.C Catford, Roman Jakobson or Vinay & Darbelnet. Later years saw the seminal works by Eugene Nida, Gideon Toury, Peter Newmark, Lawrence Venuti, Mona Baker, or Susan Bassnett and a great many others. Apart from the publications, every year there graduated numerous students specialized in translation in universities all over the world. Translation has gradually developed into an independent discipline.
Translation evaluation is the crucial issue for translation study as it serves as the fundamental guideline for translation. Since translation is so avidly discussed and in need, it is not surprising that diversity is the trend and consensus on this issue can hardly be achieved. However, it is commonly agreed that “equivalence” is the basic criterion. However, as Halliday (2001: p.15) has questioned, “equivalence with respect to what?”, the notion equivalence is still notoriously controversial. Should it be “word-for-word” or “sense-for-sense” equivalence? Should it be formal or dynamic equivalence? To better explore the nature of translation and to provide an ideal model for translation evaluation, theorists and practitioners attempt to seek inspirations from other disciplines such as linguistics, literature, philosophy, etc.
Against such a backdrop, Halliday’s Systemic Functional Linguistics (SFL) which foregrounds meaning in contexts draws attention from translators and translation theorists since the 1990s when the nature of translation as a cross-cultural communication has been highlighted. A great many of scholars turn to SFL for theoretical and methodological parameters to explore the implications of the long-debated “equivalence” as the criteria for a good translation, which contributes greatly to the development of both translation study and SFL per se.
2. Halliday’s Systemic Functional Linguistics
Devised by Michael Halliday, SFL is an approach that considers language as a social semiotic system which enables people to exchange meaning and therefore act socially. Contrasting with linguistic formalists who view language from psychological perspective, functionalists view language as a device designed for accomplishing communicative ends and hold that function supplies the organizational basis for the description of language. There are two central theoretical features in SFL, namely, “systemic” and “functional”. 
For “systemic”, any act of communication involves choices. The choice is “not a conscious decision made in real time but a set of possible alternatives” (Halliday, 1994: xiv–xxvi). SFL maps the choices available in any language variety using its representation tool of a “system network”. 
For “functional”, Halliday believes any function is achieved via metafunctions, which is the “organization of the functional framework around systems”. Metafunction consists of ideational metafunction, interpersonal metafunction and textual metafunction. The three general metafunctions are related to field, tenor and mode respectively which are three variables in his Register Theory. 
Halliday underscores the importance of context in meaning making and develops the Register Theory. The term register refers to the context of situation and “is the set of meanings, the configuration of semantic patterns, that are typically drawn upon under the specific conditions, along with the words and structures that are used in the realization of these meanings” (Halliday, 1978: p.123). According to Halliday, field refers to the subject matter; tenor focuses on the social relation and role relationships of the participants; mode describes the way the language is being used such as the channel or medium as well as the rhetorical mode (Ibid.: pp.142-44). The three variables are interrelated, any change of them may result in the change of register. All the three taken together enable people to characterize the situational context specifically.
Halliday notices that texts derive their meanings not only from the situational context within the discourse, but also from the conventionalized social occasions, thus he introduces the notion “genre” to denote the cultural context.
Apart from the notions metafunction and register, more core dimensions are covered in SFL, such as “stratification”, “rank”, etc. which are to be discussed in the forthcoming sections as they are all important parameters to be employed to frame the translation models.  
It is noticeable that SFL serves as both general linguistics and as appliable linguistics. And the past forty years has witnessed a robust development of the theory itself and a great diversity of its applications in various fields such as semiotics, multimodality, stylistics, teaching, lexicography, register variation and translation. 
3.	Application of Systemic Functional Linguistics to Translation Study
Halliday is a linguist; however, he is not just interested in linguistics, he also explores translation study based on his own linguistics theory. He (Halliday, 2001: pp.13-18) wrote a paper titled Towards a Theory of Good Translation to address the core concern on evaluating translation. 
In this paper, upon offering some thought-provoking comments on the notions system, equivalence and value, Halliday proposes the criteria of assessing translation quality within the frame of SFL by referring to three vectors “which are probably the most relevant”, namely stratification, metafunction and rank (Ibid.: p.15). According to Halliday, stratification is the organization of language in ordered strata: phonetic, phonological, lexicogrammatical and semantic — and one or more contextual strata outside of language proper. Rank is the organization of the formal strata (phonology and lexicogrammar) in a compositional hierarchy (ibid.: p.15).
Halliday argues that equivalence at different ranks carries differential values. Normally the higher the stratum, the higher the value. For example, semantic equivalence is valued more highly than lexicogrammatical, and contextual equivalence perhaps most highly of all. He claims there is no ordering among the different metafunctions except that the ideational carries by far the highest value overall as “translation equivalence is defined in ideational terms (Ibid.: p.16)” and the text which failed in ideational match does not qualify as a translation at all. In terms of register, the field of discourse is supposed to be adequately construed in the target language but there can be considerable variation in the value that is accorded to equivalence in tenor and mode. He also admits in some contexts, a good match of tenor or mode may be very highly valued in the translation, “to such an extent as even to override the demand for exact ideational equivalence” (Ibid.: p.16). He further explains such situation usually occurs when the highest value in stratal terms is being placed on contextual equivalence and overrides the requirement for semantic equivalence. Halliday suggests that the equivalence at the higher rank be given priority while the features at lower ranks be allowed to vary. He concludes by answering the “notoriously difficult” question on what is a good translation (Ibid.: p.17): 
A “good” translation is a text which is a translation (i.e. is equivalent) in respect of those linguistic features which are most valued in the given translation context...and perhaps also in respect of the value which is assigned the original (source language) text.
Halliday’s paper establishes a style of theorizing which makes constant contact with important practical issues such as why some texts are much more highly valued than others, and how it is that a translation can be judged good. He expands the application of SFL to translation study by introducing the three vectors as criteria for translation quality assessment, which undoubtedly provides a much more accurate yardstick than merely judging with the subjective ambiguous notions such as “faithful” and “expressive”, etc. However, it is still not clear enough on the question as by what standard we should rank the value of different strata. As he himself admits, “what this problematizes, of course, is the notion of value itself” (Ibid: p.17). Different text types may require different hierarchy of ranking. For example, should the formal strata at phonology level value the same in poem translation and prose translation? Should the ideational equivalence weigh the same in translating a piece of humor or an instruction manual? And there does exist situations where no ideational equivalence of certain texts can be found in the target culture, can it still be considered as translation?
Juliane House, the German linguist and translation studies scholar, believes that “the core concept of translation criticism is translation quality” (House, 2001: p.156) and has been devoted to establishing a model of translation quality assessment since 1977 when she was doing her PhD. The model was later revised in 1997. She names her model “functional-pragmatic model” for it is mainly based on Hallidayan systemic-functional theory. 
In 2001, House took up Halliday’s questions by writing a thesis titled as How do We Know When a Translation is Good? In this thesis, she first gives a brief review of how translations have been evaluated inside different traditions and schools of thoughts. After elaborating on their valuable contributions, she points out the limitations of these approaches for translation quality evaluation: the mentalist views undertaken since time immemorial are just too “subjective and intuitive”; the “equivalent response” in the response-based behavioral approaches pioneered by Nida is too tangible to be measured; the empirical-descriptive approaches represented by Toury is oriented squarely towards the translation text and put the source text as subordinate; the functionalistic or “skopos”-related approaches advocated by Reiss and Vermeer focus on how the target culture norms are heeded and the original text is reduced to a simple offer of information; the post-modernist and deconstructionist approaches represented by Venuti, Robinson and others view translation primarily on the shifts and skews stemming from ideologically motivated manipulations, which is not adequate to assess translation; Linguistically-oriented approaches take the relationship between source and translation text seriously, but they differ in their capacity to provide detailed procedures for analysis and evaluation.
Arguing that these approaches are not adequate for handling the question of evaluating a translation for one reason or another, she puts forward her own model ---the functional-pragmatic model of translation evaluation which “explicitly links text and context, and at the same time take account of the human agents involved in text reception and production operating”. In this model the source text and the translation text are compared at three levels, namely Language/Text, Register (field, mode, tenor) and Genre.
According to House, translation is essentially an operation in which the “meaning” of linguistic units is to be kept equivalent across languages, and “equivalence” is the fundamental criterion of translation quality. However, the meaning should cover three aspects, i.e. semantic, pragmatic and textual aspect which sometimes she would term them as ideational, interpersonal and textual dimensions within Halliday’s frame, and equivalence cannot be just related to formal, syntactic or lexical similarities alone but to context of situation where many factors may interplay as well. Overall, translation is viewed as the “recontextualization” of the source text by a semantically and pragmatically equivalent target text. In her model, the broad notion of “context of situation” is referred to as the more manageable parts: field, mode and tenor. 
Moreover, as register captures the connection between texts and their “microcontext” on the linguistic surface, the category of genre is introduced to connect texts with the “macrocontext” of the linguistic and cultural community in which texts are embedded.
Translation quality is then examined under the above-mentioned four-tiered analytical model (Function-Genre-Register-Language/Text). However, House further argues that a distinction between overt translation and covert translation is essential because these two types of translation make qualitatively different demands on translation evaluation. And the application of a “cultural filter” is introduced when evaluating covert translation to differentiate between a covert translation and a covert version. For overt translation, the equivalence is supposed to be achieved at the levels of Language/Text, Register and Genre, while for covert translation, equivalence is to be achieved at the levels of Genre and the Text Function with levels at the Language/Text and register exempted.
House further tests her model with an English-German translation of a children’s book Peace at Last. She gives detailed account of the assessing process referring to her four-tier functional-pragmatic model. By comparing the original and the translated texts in terms of the dimensions of field, tenor, mode, genre, differentiating the lexical, syntactic and textual means, she examines to what degree the equivalence in terms of ideational, interpersonal and textual functions is achieved and finally presents a statement of quality.
To summarize, House’s translation evaluation model with the four-tier dimensions (register, genre, metafunction) drawn from SFL (ideational, interpersonal, textual) is a much more objective approach to examine the translation quality than those by taking the complex psychological categories of translation receptors’ intuitions, feelings, beliefs as a cornerstone, which maximally avoids facile generalization although it is still not totally free from the inevitable subjective part of any translation assessment by a human evaluator.
We see the strong logic behind House’s basing her translation evaluation model on SFL. Translation is essentially a translation of meaning. However, meaning can be meaningless beyond specific context. As Malinowski has claimed, the meaning of a linguistic unit cannot be captured unless one takes account of the interrelationship between linguistic units and the context of the situation (see House, 2006: p.343). Context thus is crucial for translation. SFL foregrounds the significance of context (register as the situational context and genre as the cultural context). House notes this perfect match and holds that “functional-systemic approach provides one of the most fruitful bases for analyzing and evaluating source and target texts” (House, 2001:p.134). 
She views translation as re-contextualization as suggested in another paper Text and Context in Translation, in which she admits “The notion of context of the situation --- developed in systemic functional theory by Halliday and his collaborators --- is of fundamental importance for a theory of translation as re-contextualization, and indeed for the theoretical possibility of translation”(House, 2006: p.343). In the conclusion part of this paper she further claims (Ibid.: p.356): 
Functional approaches to language, functional pragmatics and Hallidayan systemic functional linguistics were given preference over philosophical, psychological, pragmatic, sociolinguistic and conversation analytic approaches because their notion of context was found to be more suitable for written text and thus for a theory of translation as re-contextualization. Re-contextualization was defined as taking a text out of its original frame and context and placing it within a new set of relationships and culturally conditioned expectations.
Erich Steiner celebrates House’s providing “insightful and creative exploration of the concept in translation quality assessment” (Steiner, 2001: p.163) and explores the notion of translation by a detailed examination of “intralingual” and “interlingual” versions of a text. In his paper Intralingual and Interlingual Versions of a Text – How Specific is the Notion of Translation?, he makes a detailed explication of the register theory in SFL in order to address the question of what the specific properties of translated texts are with a set of examples from advertisements. He combines the broad perspective of register and variation with careful attention to the details of wording. Apart from discussion on the common versions of the notions of field, tenor and mode of discourse, he assumes a few sub-dimensions for each of these three variables, which are experiential domain, goal orientation and social activity under the notion field, agentive roles, social roles, social distance and affect under tenor, the language role, channel and medium under mode, with more sub-sub-dimensions of each displayed in a figure.
He further claims that “an almost unique characteristic of the concept of register is to be found in the theoretically motivated realization relationships between register and the lexicogrammar of languages, most clearly as relationships between field, tenor, mode of discourse on the one hand, and the general language functions ideational, interpersonal, textual in the grammar” (Ibid.: p.163). 
Steiner suggests that translation be modelled within the SFL. He even differentiates “paraphrase” from “translation” in terms of the register theory, saying by paraphrase only important aspects of the ideational (experiential and logical) semantics are preserved while the textual and interpersonal meanings are ignored. He compares the three levels where features and structures of the source text can be preserved in translation, i.e. lexicogrammar, semantics and register. Arguing that modeling translation as preservation on lexicogrammar would likely to yield a relatively literal translation while preservation on register a relatively free one, he suggests semantic-based approach. For these three levels, Elke Teich has more detailed elucidation. He claims that according to SFL, the three levels are the descriptive responsibility of the strata, among which context (register) is realized by semantics, semantics are realized by grammar (Teich, 2001: p.195).
Steiner then makes a clarification of the status of the semantic level of linguistic organization. He emphasizes that in this connection that semantics in SFL is a multifunctional semantics without privileging any one of the dimensions. He argues that “translation would be the preservation — or maximally close preservation — of experiential, logical, interpersonal and textual meanings in the relationship of translation between texts, or in the process of translation by the translator” (Steiner, 2001: p.187). 
Steiner admits in the beginning that “we are also much indebted to efforts in translation studies using concepts from Systemic Functional Linguistics” (Ibid.: p.162). He expands the level of delicacy of the three dimensions and explores the relationships between register and the lexicogrammar of languages, which is a contribution to the development of the SFL. He also explores the nature of translation and makes differentiation between register-variation, paraphrase and translation based on their realization of the metafunctional meanings. He models translation in Systemic Functional Linguistics, assessing translation products by comparing different translation versions in detail in terms of the preservation of the ideational, interpersonal and textual semantics. He concludes his paper by claiming that “modelling of translation will force the systemic linguistic community more than hitherto — and give them the tools — to face the interesting challenge of language typology and comparison generally” (Ibid.: p.187), which highlights the significance of SFL and presents exciting possibilities for its application to translation, with promises to better develop translation theory as well as SFL per se.
Elke Teich in her Towards a Model for the Description of Cross-linguistic Divergence and Commonality in Translation attempts to scheme a translation theory based on her model of contrastive-linguistic resources that takes into account multiple dimensions of cross-linguistic divergence and commonality. These dimensions are all crucial notions from SFL. Unlike House and Steiner who focus mostly on register, she employs a broader scope of categories in SFL, which includes theme, rheme, transitivity, grammatical metaphor as well as metafunctional diversification, stratification, axiality, rank and delicacy. She holds that a model of language comparison comes first before translation. This comparison is a process to accommodate both cross-linguistic commonality and divergence. She illustrates how the above-mentioned systemic functional dimensions are identified and then generalizes the fact that as a tendency, languages shows more similarity on the semantic plane and more differences on the grammatical plane, which serves the basis of translation. A model of contrastive-linguistic resources is developed following her analysis of the linguistic categories of linguistic descriptions in SFL, and then she attempts to locate translational concepts and categories in this model. She elaborates on how the four translation notions, namely translation type, translation strategy, equivalence, and translation procedure are contextualized in her proposed model of contrastive-linguistic description, with each of them being refined using the dimensions according to which the model is organized (stratification, axiality, delicacy, metafunction and instantiation). 
Teich develops the contrastive-linguistic model by grounding its theoretic frame on Halliday’s SFL, and then roots her translation theory to the model she schemes, which provides translation practice with theoretical supports. However, her comments in her conclusion are inspiring (Teich, 2001: p.219): 
Multilingual studies in linguistics should therefore consider translations as interesting data, use translation as a test bed and consult translation studies for valuable contrastive-linguistic insights. For Systemic Functional Linguistics in particular, investigating translations offers the chance of making some of its notions more precise – in particular that of instantiation. 
As agreed by Steiner, Teich also holds that translation and linguistics (SFL in particular) are interrelated and a bi-disciplinary research can yield win-win result for development of both translation and linguistics. However, with so many parameters involved, the model is still not so operatable on which degrees equivalence can be established. 
Not only the above-mentioned scholars endeavour to apply a systemic functional perspective to translation study, many others, for example, Marco, Hatim & Mason, and Baker also employ register analysis as part of their overall account of translation study (See Liu, 2003). 
Moreover, China also sees a variety of research on SFL and its application. For example, Huang Guowen (2014), Huang Zhonglian (2017), Chen Xiao (2014), Si Xianzhu (2007, 2014), to list only a few. They explore translation either by employing some of the dimensions from SFL or the theory as a whole and yield considerable contributions. 
In his paper Metafunctional Equivalence in Translation Studies, Huang Guowen believes that the metafunctional equivalence is of vital importance regarding evaluating translated texts by analyzing the realization of the three metafunctions in some English-Chinese translation pairs. However, he further challenges Halliday’s assertion that “if a text does not match its source text ideationally, it does not qualify as a translation (Halliday, 2001: p.16)” by raising the question “Is translation which is not metafunctionally equivalent necessarily bad translation? (Huang, 2014: p.101)”. He compares a few Chinese versions of an English poem, where he observes several of them do not match in terms of experiential function yet are still well accepted even welcomed by Chinese readers. He then concludes by agreeing with Halliday that equivalence in terms of metafunction is the fundamental criterion for translation evaluation but adding that it is not the only criterion in specific contexts. If inequivalence occurs, the evaluator may explore other parameters (e.g. cultural context or ideology) to justify whether it is translation. 
Huang Zhonglian (2017) views register as a social symbol and attempts to detect the reasons for and methods of register transformation in textual translation. Assuming that textual translation is in essence a register transformation from the source text to target text, Huang, with detailed illustration, provides a three⁃step method to achieve this transformation, i.e. restructure of register’s referent, readjustment of register’s signified and reproduction of register’s signifier. Such a method can better serve to yield good translations.
Chen Xiao (2014) discusses the application of register theory to the translation of addressing terms with examples from the Chinese novel A Dream of Red Mansions and its two English versions (by David Hawkes and Yang Hsienyi respectively). Translation of addressing terms is considered to be challenging and problematic due to the huge lexical gap between the addressing systems in China and the Anglophone countries. Based on the assumption that “no meaning of an addressing term can be determined out of its register”, she elaborates on how equivalence at the dimensions of field, tenor and mode can be achieved regarding the translation of addressing terms. A good translation of addressing terms should at least show conformity in terms of their ideational, interpersonal and textual metafunctions. 
Si Xianzhu who focuses on applying SFL in translation evaluation published a series of journal articles in this field as well as a book “Translation Studies: A Functional Linguistics Approach --- Constructing A Translation Quality Assessment Model”. In his paper A Decade of Systemic-functional Linguistic Approach to Translation Studies in China: Retrospect and Prospect, he reviews the major literature of this field during the years 2004-2014 and lists a variety of perspectives being employed in translation study such as Hallidayan theory as a whole and its subordinate theories metafunctions, verbal contexts, appraisal analysis, and grammatical metaphor, etc. The theoretical and practical significance of SFL in translation is highlighted. He also points out that despite such a broad coverage more empirical researches regarding non-literature texts in particular are expected.
4.	Conclusion 
From discussed above, we see a lot of efforts made for better evaluating translation. All the analyses above illustrate that it is beneficial to introduce SFL as the theoretical frame for translation criticism. The SFL perspective narrows down the level of equivalence to more tangible aspects rather than the macro mappings in terms of for example formal equivalence or dynamic equivalence. And it avoids the subjectivity of judging the quality of translation in terms of the receptor’s intuitive feelings on whether it is faithful or communicable. 
The essential thing to translation evaluation is that we must be equipped with a set of criteria in mind, otherwise any assessment is of no value. Provided that we accept equivalence as the criterion, we must then make clear it should be with respect to what equivalence and at what level of equivalence. In this regards, Halliday’s SFL is a straightforward and appropriate perspective. If we see translation as a process of re-construction of context, then register theory in particular serves as a tool for translation assessment: as different contexts of situation encode meaning in different ways, we have to compare the Source Text and Target Text in terms of field (to see whether there is a reconstruction of context through an analysis of what has taken place), tenor (to examine whether there is a match between the relationship of the participants in the Source Text and Target Text) and mode (to examine whether the medium for relaying message in Target Text is in consistent with that of Source Text). The other core dimensions such as stratification and rank help address the priority of equivalence the translator should lay on. 
However, as all texts happen in context rather than in vacuum, translation becomes such complex a task involving a great number of factors, any evaluation models cannot be all-comprehensive or absolute. Halliday suggests the term equivalence value, nevertheless, the weighting is hard to measure as genre varies, e.g. some genres (poem vs product instructional manual for example) may require a totally different priority on certain level of systemic correspondence or a particular metafunction. External factors such as ideology, cultural norms which House suggests putting aside as a secondary consideration can by no means be ignored when assessing a translation text. Absolute accuracy is rare mainly because translation is a subjective piece of work made by human agent. And it is noteworthy that almost no evaluation model based on SFL can be valid or reliable enough to decide the aesthetic value of the translation text.
Despite the fact whether a piece of translation is good or not is vulnerable to too many factors since it involves texts of a various genres in a particular community at a particular time and for a particular client, Systemic Functional Linguistics provides relatively good perspective and exciting possibilities.  

References:
Chen, Xiao, Jiayi Wang, and Congyi Zhang, ‘Register Conformity in Translation of          Addressing Terms in Literary Texts’, Foreign Language Research, 5 (2014), 81-84. 
Elke Teich, ‘Towards a Model for the Description of Cross-linguistic Divergence and Commonality in Translation’, in Erich Steiner & Collin Yallop (eds), Exploring Translation and Multilingual Text Production: Beyond Content. (Berlin: Mouton de Gruyter, 2001), pp.127-60.
Erich Steiner, ‘Intralingual and Interlingual Versions of a Text – How Specific is the Notion of Translation?’, in Erich Steiner & Collin Yallop (eds), Exploring Translation and Multilingual Text Production: Beyond Content. (Berlin: Mouton de Gruyter, 2001), pp.161-90.
Halliday, Michael, An Introduction to Functional Grammar, 2nd edn (London: Arnold, 1994).
Halliday, Michael, Language as Social Semiotic: The Social Interpretation of Language and Meaning (London: Edward Arnold, 1978).
Halliday, Michael, ‘Towards a Theory of Good Translation’, in Erich Steiner & Collin Yallop (eds), Exploring Translation and Multilingual Text Production: Beyond Content. (Berlin: Mouton de Gruyter, 2001), pp.13-18.
House, Juliane, ‘How do We Know When a Translation is Good?’, in Erich Steiner & Collin Yallop (eds), Exploring Translation and Multilingual Text Production: Beyond Content. (Berlin: Mouton de Gruyter, 2001), pp.127-60.
House, Juliane, ‘Text and Context in Translation’, Journal of Pragmatics, 38(2006), 338–358.
Huang, Guowen & Yang Chen, ‘Metafunctional Equivalence in Translation Studies’, Foreign Languages in China, 2 (2014), 97-102.
Huang, Zhonglian, and Yao Sun, ‘Toward a Three ⁃ Step Transformation of Register in Textual Translation’, Modern Foreign Languages, 2 (2017), 201-212. 
Liu, Zequan, ‘Register Analysis as a Tool for Translation Quality Assessment’, Translation Journal, 3(2003).
Si, Xianzhu, ‘A Decade of Systemic-functional Linguistic Approach to Translation Studies in China: Retrospect and Prospect’, Foreign Languages in China, 3(2014), 99-105.
Si, Xianzhu, Translation Studies: A Functional Linguistics Approach -- Constructing A Translation Quality Assessment Model (Beijing: Peking University Press, 2007).