The basic assumption of laymen concerning translation is that every word or meaning can universally be translated from one language to another. Their idea of translation is that of a straightforward mechanical process which simply replaces source language (SL) items with target language (TL) items. Some conceptions in translation studies seem to encourage this view, and debates suggest that it is only a matter of the right scope, focus or technique to create perfect translations.
The paper at hand will refute this notion. It will prove and exemplify the facts that not everything is translatable, and that a transfer of meaning necessarily involves changes entailing loss or gain of linguistic, cultural and stylistic features (cf. Harvey 2001, 38; Pym & Turk 2001, 274). Translation cannot create an identical TL copy of the SL text, but only permits a relative equivalence to it. A maximal approximation, however, can never be achieved, due to the complexity of language, its dependence on constantly changing cultural norms, and because the human factor. Especially the aspect acceptance by the audience will show that perfection is just an abstract evaluative term, which largely depends on individual taste.
According to the Longman Dictionary of Contemporary English, perfection is the state of being perfect, which in turn can be described as the absence of mistakes and faults, or as good as possible (Summers 2003, CD-ROM entries). But what does this term imply for translation?
Flawlessness could be regarded as total sameness in the sense that the translated text is an identical one-to-one copy of the source text. In this vein the notion of a literal or gloss translation assumes that an ideal translation can be achieved by a segmentation of the SL text into individual words and by transferring these word-segments one at a time (cf. Robinson 2001, 125). This might work with simple grammatical patterns, as the translation of the English sentence Peter took the book into the German Peter nahm das Buch illustrates. But already slightly more complex structures lead to ungrammatical or at least unnatural or clumsy results. The translation of Peter has taken the book from the desk into * Peter hat genommen das Buch von dem Schreibtisch shows that a syntactic modification is necessary to form the grammatical sentence Peter hat das Buch von dem Schreibtisch genommen. Thus, “translation is no formal procedure of substitution on the basis of simple one-to-one correspondences” (Wilss 2001, 58) which could yield an identical TL text. Due to this, sameness cannot constitute perfection in terms of translation.
As a relative norm perfection can be seen as a maximum convergence of a translation to the original. In this case the notion of equivalence of SL and TL text is the main criterion for translation. Equivalence is the relationship between a source text (ST) and a target text (TT) which allows the TT to be considered a translation of the ST (cf. Kenny 2001, 77). Equivalence can exist on different levels, for instance as linguistic equivalence, referring to lexis, grammar and pragmatics, as cultural or as stylistic equivalence. Basically, a translation must be adapted to the TL and its cultural norms to use equivalent pragmatic means which ensure that the TT creates the same response in TL receptors like the ST did in SL receptors. (cf. Kenny 2001, 77 et seq.; House 2001, 197 et seq.). Perfection would assume a maximum of equivalence on all levels, but such a maximum can never be achieved due to the complexity of language, its dependence on constantly changing cultural norm, and because of the human factor in form of translators and receptors. These factors inevitably lead to modifications of the SL text including loss or gain of lexical, grammatical, cultural or stylistic features.
Language itself it very complex, but difficulties increase exponentially in translation, since it has to cope with two languages. The degree of translatability basically depends on the structural differences between SL and TL (cf. Pym & Turk 2001, 274). If equivalence should be achieved, obligatory shifts have to be made in order to allow for language constraints (cf. BAKKER, KOSTER, VAN LEUVEN-ZWART 2001, 228 et seq.).
Certain lexical items or grammatical structures may only exist in one language but not in the other. Certain words have no TL equivalent, since the concept that they contain is unknown in the TL culture. For instance, a TL culture that does not know the concept ´ cheese ´ will have no correspondent word for it. This lack can be compensated by paraphrasing or by using calques, or just by using the SL word and describing it in a footnote. Cheese for example can be described as ´ coagulated milk curds ´ (cf. Pym & Turk 2001, 275). Nevertheless, do paraphrasing and the use of calques entail a loss of SL precision and make the meaning of words more elusive. Furthermore, the usage of SL words is no true translation, even if it may lead to the creation of TL loanwords. Finally, an explanation in footnotes is a deviation from the SL text in form of surplus information. Examples of words in German that demand some kind of elucidation are Einwohnermeldeamt or Personenregister, since the British and Anglo-American culture does not know such institutions. On the other hand words like Schadenfreude, Rucksack, and Kindergarten are examples of German words which became loanwords in English due to the former uniqueness of their concepts.
Also grammatical structures may exist in the SL, but not in the TL. The gerund in English for instance has no German counterpart. Therefore a translation can only approximate it by using nouns or infinitive constructions. For instance, Reading is fun can be translated Lesen macht Spaß and It is hard getting up in the mornings as Es ist schwer, morgens aufzustehen. Also English progressive tenses do not have a grammatical counterpart in German. Thus, a translation has to take recourse to a lexical solution to express the special tense meaning of the verb, e.g. I’m watching the news, can wait for five minutes? turns into Ich schaue mir gerade die Nachrichten an, kannst du fünf Minuten warten? Even if meaning can be conveyed more or less equivalently by different patterns, there is still a loss of grammatical originality and of SL authenticity.
Other untranslatable features of languages are dialects and accents, which exhibit phonetic, lexical and grammatical peculiarities. For instance Irish, Scottish or Welsh cannot be translated into Bavarian, Hessian, or Saxon in German. Even if TL dialects are used for stylistic reasons, this method is only a compensation which is, despite all possible underlying parallels, an arbitrary decision connected to a massive loss of uniqueness and authenticity. For instance, the adaptation of the Eliza Doolittle’s Cockney accent to a Berlin or Viennese dialect in German versions of the musical My Fair Lady (cf. German Wikipedia entry) may be a good equivalence, but it only renders some aspects of Cockney.