"I AM TRANSLATING" IS FALSE
 

The translator is anonymous

It has been astutely lamented that, in accordance with the principle of ideal equivalence, the translator remains "nobody in particular" (Belitt 1978). Of all the symbols and saints used to represent the profession - Janus or Jerome, forked tongues or true interpreters - , the figure of "nobody" is of particular theoretical profundity.

Some translators have of course expressed and exerted strong personal identities. Yet there must remain doubt as to whether their particularity was not in conflict with their work as translators. Reading a Hölderlin version of Pindar probably has more to do with reading Hölderlin than with establishing any strict relation of equivalence with Pindar. Or again, active appreciation of the subjectivity and work of a Jerome or a Luther effectively blocks the reading position necessary for ideal equivalence to what might be projected as "the" Bible. In principle, if translated texts are to be received and believed as ideal equivalents of their antecedents, translators themselves must remain anonymous and their work must remain unevaluated as individual labour.

This means that, although equivalence is certainly the result of work, its social function depends on the practical anonymity of this work; it can only function for as long as the receiver is indifferent to the translator's subjectivity. Equivalence itself may well be analysed in terms of exchange value; its false naturalness may be reduced to mere assurance of potential use; but no aspect of applied or misapplied economics gives the slightest indication that the principle of equivalence will allow translators to be appreciated in terms of any individualised labour value. A labour-value theory of equivalence would be a contradiction in terms.

Translation might thus be described as a potentially scandalous activity in which people work to produce an output which is ideally thought to have the same value as the input, leaving their labour without value in itself.

Or is this view merely a projection of bad theory? After all, physical translators are individuals, with individual bank accounts which should be individually affected to the extent that the value of their work is at least financially quantified. Should the anonymity posited as a correlative of ideal equivalence then be seen as no more than the way in which certain translations are read by certain people?

Careful consideration should be accorded to the two economies at work here. On the level of material production, translators are no different from most workers, transforming previous products to produce new products, receiving remuneration for the labour expended in the transformation process, and sometimes receiving a certain percentage in proportion to something called "value added", the difference in value between input and output. The exact financial remuneration varies widely according to the social context in which the work is carried out, since standard rates and product qualities vary considerably according to the local labour market and the presence or absence of effective professional organisations. On the semiotic level, however, translation is defined by a relation of equivalence which denies the very possibility of any such value added , since the output is supposed to be directly exchangeable for the input. As we have noted, this direct exchange situation is moreover indifferent to the actual material location of translators in one market or another. There are then two distinct economies at work, the first organising the production and distribution of textual material, the second governing the semiotic representation or reproduction of this process in terms of value.

The distinction between these two levels would present no problem if they were really separate. Unfortunately, since ideal equivalence on the semiotic level denies value added, it necessarily hides translational labour value or converts it into non-value, thereby exerting habitually nefarious influences on translators' financial remuneration and professional psychology. Because the two economies are interconnected, the anonymity of the translator should be taken seriously.

Several quite fundamental questions need to be raised here. For example, if ideal equivalence means that the translator must remain anonymous, what does it mean for other potential first and second persons? How do receivers recognise a text as a translation if they can only recognise the translator as anonymous? What might be the discursive form of such reception? And what are the conditions, if any, under which the anonymity associated with ideal equivalence can be challenged so that translators might have something of value to say?
 

The utterance "I am translating" is necessarily false

How does one know that equivalence is pertinent to a given text? A simple answer would be that the text "says so". But not all texts say so in the same way, and it is not at all clear that the saying actually belongs to "the text". In some cases the indicator is material or situational in others, it is properly semiotic.

Examples of material and semiotic economies in a Spanish museum:

a) In an eight-page brochure for an art museum, the title page is in Spanish ("Centro Atlántico de Arte Moderno") and the in-text is in English (describing the "Atlantic Centre of Modern Art"). There is no explicit indication that the English text is at all translational. It reaches very exactly the bottom of each page; it contains no visible linguistic interferences; it bears no mention of a translator. If I had not translated it, I could not guarantee its status as a translated text. And yet it might reasonably be assumed that, since the museum is in a Spanish-speaking community, its Spanish name is logically anterior to its English name and this particular text is therefore likely to be a translation addressed to English-speaking museum-goers. Although there are certain semiotic clues, the only decisive indicator here is situational and derives from the material economy of art and tourism in Spain.

b) An English-language catalogue produced for the same museum contains an absurdly lyrical text on the art of collecting, preceded by the name of its Spanish author and followed by a little note saying "Translated from the Spanish by...". Here there is no reasonable doubt. Independently of factors like its material location or the possible non-existence of either author or translator, the text is translational. The decisive indicator is in this case properly semiotic, although its function also depends on non-contradiction from situational determinants.

It is perhaps a little perplexing to find that these material and/or semiotic indicators are strictly neither internal nor external to the text proper. They belong instead to a border region between the material and the semiotic, to the published thresholds which comprise objective cover pages and become meaningful though limit-signs. They belong to the "paratext" (to follow the terminology of Genette 1987, who nevertheless overlooks the paratextual place and role of the translator).

The first and most important principle of this paratextual status is that these indicators should be excluded from the part of the text that is presented as "ideally equivalent" or of exchangeable value. Although the indicators point to the translational nature of the value concerned, they are placed outside the equivalence relation between TT and Y. The translational paratext thus functions as a kind of instruction for use, saying "...the absent Spanish text translates as...", thus necessarily excluding the possibility that it is itself a TT. If the indicator of equivalence is to function, it must itself be a non-translation.

That is why the discursive person who says "I am translating" cannot be translating at the moment of utterance. This paradox has consequences not only for the first person of translated texts, but also, as we shall see, for the implicit second and third persons of translation as a discursive act.

But this phenomenon should be analysed step by step.
 

Can interpreters say they are frightened?

If, in the consecutive or simultaneous situation, the text "J'ai peur" (de Gaulle?) is rendered as TT1 "I am frightened", it could be misunderstood as the translator talking about his or her immediate condition. One way of avoiding ambiguity would perhaps be to jump straight to the third-person TT2 "The speaker is frightened". But this could equally be misunderstood as the translator commenting on the speaker's immedate condition. In this case, the one way to make sure that the fear is unambiguously the speaker's is to frame it as indirect speech, as in TT3 "He says that he is frightened". The indirect speech of TT3 might thus be seen as an elaborate form of the particularly unsuccessful translational utterance TT1, making explicit a certain kind of paratextual operator - a discursive form or "instruction for use" - that is otherwise normally implicit ("He says that..."). Although it is clear that successful translation need not be transformed into indirect speech in order to be understood, the relation between the two discursive forms remains of at least analytical interest.

John Bigelow (1978) has suggested that translation is indeed a peculiar case of indirect speech whose specificity is based on the form "translates as", analysed as an operator which is at once quotational and productive of new information. The following are simplified versions of two of the steps by which he attempts to generate the discursive form of translation:

(1) Ludwig said, "I must tell you: I am frightened".

(2) Ludwig's words translate as, "I must tell you: I am frightened".

The paratextual shift from (1) to (2) represents the basic progression from a text received as direct speech to one received as a translation. Although Bigelow draws attention to the fact that whereas (1) refers to a person, (2) refers to a linguistic entity, he then adds that "this departure from the general format for other hyperintensional sentences is of minor significance and introduces no new problems of principle". I beg to differ.

It is of considerable importance and interest that the translational operator "Y translates as TT" refers to things (words) rather than to discursive subjectivity (a person). This is for two main reasons. First, although a spoken version of (1) would remain as fundamentally ambiguous as the translational "I am frightened" - one or both first persons could be attributed to the reporter - , the properly translational operator of (2) successfully removes this ambiguity. Second, the same ambiguity could partly be resolved by explicit third-person substantivisations such as "The situation is frightening", thus indicating that the translational operator is associated with the third person not only in its verbal form but also in its specific discursive function. The transition from (1) to (2), from subjectivity to things, represents a move from unsuccessful to successful translation. This principle is moreover coherent with the reasons why translation studies should accord priority to objects rather than subjects.

Why translators seek refuge in the third person:

A student attempting consecutive translation had some difficulty locating appropriate pronouns. The speaker was making repeated references to "notre Centre" - in fact the Centre Pompidou in Paris of which she was one of the directors - ; the translator consecutively grasped this as "our Centre" (adding "I am speaking for her") and "her Centre" (adding "the Centre where she works"), before settling on "the Pompidou Centre" and then "the Centre" (without addition). The problem of first-person pronouns was happily resolved by retreat to the neutral space of an unattributed third-person term, a name. The TT "Pompidou Centre" belongs neither to the ST speaker, nor to the translator, nor indeed to Georges Pompidou.

What is the essential difference between (1) "Ludwig said" and (2) "Ludwig's words translate as"? It is not just a difference between a person and the words said by that person. In referring to Ludwig's action in the past tense, (1) positions a first person (non-Ludwig) in the present; it positions its I-here-now in relation to the reported speech. By contrast, (2) uses an eternal and subjectless present ("translates as") to project its action as necessary and valid at all times and places, in the absence of the first person and perhaps even of all persons. Its I-here-now has no relation to the reported speech. The essential difference between the two operators is thus that (1) positions an I-here-now for a semiotic reporter-translator to stand on, whereas (2) completely eliminates the possibility of situating the translator. This is as it should be, since if the translating subject could manifest itself in relation to a discursive I-here-now, the utterance "I am translating" could be true. Since I believe the utterance to be necessarily false, I accept the operator "translates as" as the discursive form of the way ideal equivalence functions in reception.

Second persons can be anonymous

If (1) positions a first person and (2) does not, it is to be expected that this difference will have consequences for the position and role of a macrostructural second person, the implied receiver (cf. Iser 1978). It is perhaps useful to refer here to two of Bigelow's initial examples:

(3) Ludwig said, "Du musst wissen: ich fürchte mich".

(4) Ludwig said, "I must tell you, I am frightened".

Bigelow points out that the operator "said" is purely quotational in (3) but partly productive of new information in (4) since it does not quote Ludwig's words exactly but "reports them in a fashion closely resembling that of indirect speech". The move from (3) to (4) is a further step towards successful translation. But why should this transformation have been introduced? We could suppose that Ludwig wished to address a friend who did not know German (several hundred other situations would be equally plausible). Before the translation is carried out, the friend cannot occupy the position of the implied receiver. The purpose of the translation is to enable him to do so. In this sense, every act of translation must be for an implied receiver - a discursive second person - in some way excluded from the non-translational text.

I should stress the major distinction being made here between the macrostructural level of discourse and the level of linguistic form. The above move from "Du musst wissen" (literally "You must know", with the linguistic second person) to "I must tell you" (linguistic first person) is certainly interesting as an indication of the way equivalence can respect local conventions, but the differences between these two forms have nothing to do with translation itself as a discursive act. It is enough that, on the level of discourse, both these tags call the attention of an implied receiver, quite independently of their linguistic first-person or second-person status.

Now, what is interesting here is that translational status itself functions as a similar tag. The non-translational nature of (3) indicates that its implied receiver understands both English and German. The implied receiver of (4), on the other hand, understands English but not necessarily German. That is, the translational choice of a language or languages always implies the minimal profile of a translational second person. The status of the text as a translation thus functions in a way similar to a pronominal utterance, but without the linguistic pronoun. Nobody is going to specify that the translated text is for "dear reader, you who understand English but perhaps do not understand German". There is no need to mention this paratextual "you". It is enough that the text be presented and initially received as a translation.

But is it not strange that translational discourse, like a bad riddle, should function like a personal pronoun and yet have no personal pronouns of its own? Why is this so? On the one hand because, once the text has been adequately adapted to a new second person, any naming of that person is entirely superfluous: the second-person position has been integrated into the translational status itself. But it is surely also because, if a specifically translational second person is named, it must be distinguished from the specifically non-translational second person of an untranslated text. Such a distinction cannot be manifested because mention of a new second person would necessarily activate the distance from a previous second person, bringing to the surface the problematic realities of transfer and the discursive intervention of a first-person translator who is supposed to have no voice. For this reason, translation gives the profile of a second person but negates the deictics associated with linguistic pronominal forms. The choice of a certain language or code alone implies a certain directionality and opens space for a certain kind of receiver, but must do so in silence.

More elegantly, one could say that the utterance "I am translating" is no truer than is its extended form "You are reading the translation I am now doing". The translational second person must be as anonymous as the translator.

This analysis should further explain why translational discourse is happier referring to words rather than to subjectivity, to neutral objects rather than to positioned first or second persons. If the translator's position is to be without linguistic manifestation, so must the first and second persons of properly translational discourse.

It might be objected that our analysis has unfairly extrapolated from peculiarly oral cases like "I am frightened", where the ambiguities are perhaps due not to the nature of translation in general but to the specific situation of consecutive work, since the translator's physical presence induces a special risk of misattribution. It might be argued that the proper situation for translators is, after all, to be invisible; unlike children, they should be heard but not seen. Yet there are many cases where purely written translation requires the same suppression of first-person and second-person positions. In EU reports on the Spanish economy, for instance, terms like "nuestro país" (our country) or "nuestra economía" (our economy) are usually rendered as "Spain" or "the Spanish economy", restricting first-person pronouns to expressions of personal opinion where it is clear that the "we" is exclusive and not inclusive. Is this an isolated stylistic norm? In noting that the Jerusalem Bible is a christianisation of the Hebrew text, Meschonnic observes that the Tetragrammaton formulae "Yahve my Lord" and "Yahve your Lord" are translationally transformed into "the Lord" (1973, 419). The divinity that belonged to a people has become a Christian God available to all, just as the economy that once belonged to Spaniards is now exposed to the neutral nouns of EEC policymakers. According to Meschonnic, "the religion of the Son has always wanted to kill the Father". But surely there is a spirit common to them both, middle ground for negotiation through translation?

Third persons allow translators to talk

The discursive form or operator "translates as" is clearly as much a fiction as is the translational equivalence it proposes. Words do not translate themselves; they are translated by humans in society. And since humans in society are historical, words are translated differently according to different times, places and situations. In suppressing the I-here-now of its first and second persons, the translational operator attains a neutrality manifestly devoid of concrete correlative. Indeed, a stubborn realist might doubt that "translates as" is the form of equivalence in reception. Surely what we normally see in paratexts is more like the following:

(5) WORDS BY LUDWIG

Translated by Bigelow

Is this title-form any better as a translational operator? Its status is no doubt just as peculiar, but it is more explicitly a paratext, a threshold situating a translated text. It is composed of three names - Title, Author and Translator - and the connector "by", which is not the same in its two appearances: "BY LUDWIG" is defining; "by Bigelow" might be relative (preceded by a comma) but is possibly also defining (not necessarily preceded by a comma). That is, the text thus introduced can be approached as the classic WORDS in its eternal form, such that the work of the translator is merely "relative" value added, without consequence for the receiver who begins from an initial "Let's see what Ludwig has to say". Or the text can be approached as Bigelow's particular version of a known original, this second "by" then being defining and the particular receiver asking something like "Let's see how Bigelow translates (as opposed to some other actual or possible translator)". On the level of the title-form, there will always be doubt as to the relative or defining nature of the translator's presence. That is why the properly translational operator, which continues to be operative within the text itself, should not be confused with the common form of translational paratexts. Reading a title page and reading a translation are fundamentally different activities.

As the physical receiver enters the text and ideally conforms to the profile of the implied receiver, the named persons - author and translator - are immediately transformed into either a first person - in the case of certain modes of authorship - or absent third persons. The ambiguity of the translator's relative or defining presence disappears. If the "relative" reception strategy has been selected and the translator is a wholly unwanted presence merely disturbing access to the eternal WORDS, the operator "translates as" becomes a wholly necessary fiction enabling the reader to forget the uncertainty of the paratext, inducing the classical "willing suspension of disbelief" typical of all fiction. And if, alternatively, the reader has selected the "defining" option and is interested in seeing how Bigelow translates, then the third-person status of the operator serves at least to remove linguistic doubts about personal pronouns: in accordance with ideal equivalence, such a reader is free to look for Bigelow's voice behind every linguistic element except the persons "I" and "you" in the text (we shall test this principle immediately below). In both cases, the third person is the discursive mode best suited to the reception of the translated text.

The operator "translates as" thus makes third persons out of the subjectivities involved in translation. And as Benveniste (1966) has argued - after Arabic grammarians - , the third person is the one who is absent; it is a non-person, a thing and not a subjectivity.
 

Does anyone speak Redford's language?

The above analysis has strongly associated ideal equivalence with suppression of the translator's first person and with various retreats to the neutral space of third-person terms. But can the translator's presence also be hidden behind the first person?

In film dubbing from English into Spanish, code-referring utterances like "Do you speak English?" are commonly rendered as "¿Hablas mi idioma?" ("Do you speak my language?"). Here the problem of self-reference in translation is solved by avoiding the third-person name ("English"/"inglés") and retreating to the neutral if highly ambiguous space of a first-person pronoun. No one can really say if the "I" of the resulting "my language" speaks Spanish, English or, in a vaguely utopian projection, all languages at once. Reference to the translator's situation would appear to have been avoided by the construction of a very peculiar first-person pronoun. Does this then contradict the general tendency for translators to seek refuge in third-person terms?

Let us approach through a slightly simpler example. In the film Butch Cassidy and the Sundance Kid, two bank-robbing heroes, played by Redford and Newman, move to Bolivia where they are faced with the problem of having to learn enough Spanish to exercise their profession, robbing banks. But in the Spanish version of the film they already speak very good Spanish. Why should they now have to learn this same language? Perhaps the translators could have made them learn English, but this part of the film is very definitely set in a visual Bolivia. So the Spanish adapters hijacked the storyline for two or three sequences: Redford and Newman now decide to learn French in the hope that they will not be recognised as Americans. When they enter a Bolivian bank, the rudimentary Spanish of Y is replaced by rudimentary French in TT. The neutral space of the third term "French" conceals their identity, not from the bank officials - who take about two seconds to recognise them as Americans - , but from the viewer of the dubbed film, who is supposedly spared an upsetting reference to the fact of translation and thus the presence of a translator.

Use of the apparently neutral translational world "French" thus conceals two conflicting first persons: the "I" who is the English-speaking bankrobber and the potential "I" who is the dubber into Spanish. The solution "French" thereby avoids locating the specifically translational second person who would otherwise be identified as a Spanish-speaker unaddressed by the original film.

Why might such a reference have been upsetting? After all, translated versions of Pygmalion and My Fair Lady make prolonged and repeated references to the English language, setting up the acceptable - since accepted - convention that the TT language is to be treated and named as if it were really English. The fictional "as if" of this contradictory self-reference can assume conventional status when structurally prolonged, when it becomes part of a fictional world where two "I"s can become one. The Butch Cassidy example, however, concerns a language with only transitory status in its fictional context, without sufficient narrative space to set up conventions of its own. The difficulty has more to do with a moment of code-switching than with actual content. It concerns not a language, but a frontier between languages.

Until now we have assumed that the operator "translates as" has as its correlative the homogeneous content "words", ascribed to one sole author and thus implicitly to one sole language. But this is a dangerous assumption. Derrida (1987) has correctly criticised translation theory for too readily assuming texts to be monolingual, but one needs no cabbalism to see the point: Maurice Blanchot some time ago pointed out the translational status of Hemingway's characters who, by speaking Spanish in English - inserting the occasional Spanish term and adopting Spanish syntax - , create a "shadow of distance" that can then be translated as such (1949, 186). Does translation only deal with words in one language? I suggest that the internal distance described by Blanchot and picked up by Derrida could be applied to our present example as follows:

(6) Redford's code-switching translates as code-switching from Spanish into French.

Interpreted in this way, the solution "French" is really no more astounding than the fact that the English "It's Greek to me" translates as "Das kommt mir Spanisch vor" ("It's Spanish to Germans"). The linguistic content of Redford's first person is not in any one language, nor in all languages. It is an asymmetric interlingual frontier.

If we now go back to the original example "Do you speak my language?", it is clear that the content of the utterance is not whatever language the first person speaks, but whether or not the situation calls for code-switching. Indeed, since the question "Do you speak English?" can be rewritten as "Is this a situation requiring code-switching?", the transition from Y to TT does not necessarily throw up any problematic first person at all: there is no translational schizophrenia, only cultural disjunction.

Once again, the translational operator overrides problematic first and second persons in order to set up a world of third-person terms.
 

Third persons can conflict

 In the Spanish version of the BBC series Fawlty Towers, the waiter Manuel is not from Barcelona - as he is in the English version - but from Mexico. The translators obviously strove to avoid an embarrassing reference to certain English preconceptions about Iberian culture (although they were apparently unconcerned about a negative image of Mexican culture). Problems of self-reference in translation do not just concern languages but entire cultural codes, the way different cultures see each other and themselves. Faced with this kind of problem, translational discourse tends to prefer neutral worlds where there are no first or second persons, and no marked terms that might belong to such persons. But do mere names always allow such neutrality?

Quine's treatment of potential synonyms borrows from Schrödinger the example of a mountain-climber who has learned to apply the name "Chomolungma" to a peak seen from Nepal and "Everest" to a peak seen from Tibet (1960, 49; corrected in the French translation, 1977, 87). The mountain-climber believes these names refer to different peaks until the day his explorations reveal that they are in fact one and the same. This equation presumably solves all problems of reference. Moreover, since the two names continue to exist, "Everest" translates as "Chomolungma" in the south, "Chomolungma" translates as "Everest" in the north, and no strictly semantic problems should be expected to ensue. However, beyond the Himalayas, translators have to choose between the peak as named from the north and the same peak as named from the south. The name itself cannot be neutral. The choice of "Everest" implies that the Tibetans were authors and the Nepalis bad translators, whereas "Chomolungma" places the authors in the south and the bad translators in the north. As I too often discover when deciding the English names of Catalan/Spanish towns - are there still German names for towns in the west of Poland? - , the use of certain third-person terms necessarily positions the author of the term, and therefore its translator.

I have so far claimed that the anonymity of the translator is generally protected by the third-person nature of translational discourse. But here, in cases of conflict between third-person terms, it must be admitted that ideal anonymity can be successfully challenged. This form of disclosure is moreover of some importance for the ethics of translation, since it is in the choice between third-person entities, in the selection of one external term or another, that translators can hope to exert influence on the way cultures perceive each other.

This poses obvious problems for the workings of ideal equivalence.

 A curious third person left in Yugoslavia:

"The accused driver called Zof, which was evidently short for his funny surname understandable in this part of the world, walked barefoot on Mostar cobbles bearing visible traces of torture." (Dragoslav Janjic, The Moving Target / Pokretna Meta, 1989, 20, italics mine).

Ideal equivalence should be challenged by the question, "Who says 'this part of the world'?"

Ideal equivalence can be challenged

Should we be surprised that, even within translation, there are limits to ideal equivalence? Not really. After all, the kind of challenged anonymity found in third-person conflicts has long been reflected in the way certain theories describe the way a translation can signify its antecedents. When Levy (1969) distinguishes between "illusionary" and "anti-illusionary" translation, or House (1977) develops the parallel notions of "covert" and "overt" translation, the object of theorisation is in fact the tension between equivalence as an ideal - ideally "illusionary" and "covert" to the extent that it hides real work - and certain ways in which actual translational work can be manifested without necessarily entering the realm of non-translation. The very existence of such categories indicates that the anonymity of the translating translator is not always complete. This suggests that the translational operator is not the only formal determinant on legitimately translated texts. There must be other factors, other ways in which translational subjectivity can slip through the grasp of equivalence and yet remain translational.

Rather than simply assume the validity of Levy's or House's categories, I find it more interesting to ask how it is possible that some translations can become "anti-illusionary" or "overt" without losing their status as translations. What aspects of transfer and translation can become significant without necessarily destroying the qualities of ideal equivalence?

But before answering this question, let me briefly summarise the discursive significance of equivalence and its relation to translational anonymity. 

Over the last few pages I have described equivalence as a fact of reception, as a mode of the exchange relation TT:Y proposed by the operator "translates as". As such, its material location is between TT and the receiver willing to adopt a certain suspension of disbelief. Equivalence exists for a receiver who is willing to believe that, to all intents and purposes, TT is ST and this voice saying "I" is the voice that said the same "I" in another place and time, in another culture, even though it is otherwise clear that the text concerned is really the work of a translator and that ST is really only available in its transferred form Y. Thanks to the operator "translates as", innumerable flesh-and-blood receivers have momentarily forgotten about their real place in the world; they have willingly occupied the similarly anonymous position of the implied receiver of translated texts; they have accepted TT as the ideal equivalent of its antecedents.

But we have also seen that the only way to maintain such illusions is to hide all the specific mediations involved - the particular translator, the particular receiver, the times and places marked by the transfer from ST to Y - , and to do so through the falsely universal operator "translates as". In this way, equivalence should suppress not only the translator but also the material fact of transfer across distance.

Let us now go back to the general formula for the way translation represents distance:

 For ideal equivalence to function, the value of TT:Y should be 1 and the value Y­ST should be rigorously non-significant. That is, the translated text should represent no distance and the implied receiver should ideally be unconcerned by the actual distance covered by the act of transfer. Any attempt to attribute other values above or below the line will break this fiction of ideal equivalence.

As a kind of general signpost to the path we have followed and the destinations awaiting us, let me now draw up an initial list of circumstances under which ideal equivalence is likely to be challenged:

- When the translator manifests a position through third-person choices (as we have seen above)

- When the quantitative ratio TT:Y deviates significantly from the value 1 (analysed in chapter 4, on quantity)

- When the distance Y­ST becomes significant for the receiver as a positive value or as zero (as suggested above and developed in chapters 5 and 6, on belonging and transfer).

- When TT transgresses a norm or becomes an object of translational analysis or theorisation (dealt with in chapter 7, on historical ethics, and chapter 8, on theory).

Despite the apparent restrictiveness of the operator "translates as", these four possible circumstances - there may be more - open up a spectrum of relative equivalences and contradictory equivalences, a range of phenomena which surround ideal equivalence and yet do not make the utterance "I am translating" true.
 

Next chapter