5.  Chapter 5: Approaches to the Translation of Poetry and Poetic Prose

The starting point for the poetics of translation usually begins with the supposition that what is being dealt with is ‘the art of the impossible’ as in the notorious Robert Frost dictum that what great literature consists of, is what is lost in translation. This entrenched position assumes that the untranslatability of literature is an incontrovertible truth and indeed, there are many eminent proponents for the absolute impossibility of this activity including Jakobson (2000):

[. . .] - paronomasia reigns over poetic art, and whether its rule is absolute or limited, poetry by definition is untranslatable. Only creative transposition is possible. [. . .] (Jakobson 2000: 118. My emphasis.)

Like most linguists, Jakobson’s rejection of any form of paronomasia as translatable reflects the lexically bound view of the translation process, even though he does allow for “creative transposition”. Interestingly, Hatim and Mason (1988: 13) also use the phrase creative transposition to show the impossibility of dialect translation and it is also significant that this most excellent of strategies is qualified by the adverb only. Yet, it will be seen that creative transposition is a frequent strategy that must be employed not only with regard to literary texts but also often for commercial translations, particularly in the field of publicity and advertising.

Many contemporary linguists such as House (1997) support Jakobson’s belief in the untranslatability of poetry:

In a poetic-aesthetic work of art, the usual distinctions between form and content (or meaning) no longer hold. In poetry, the form of a linguistic unit cannot be changed without a corresponding change in (semantic, pragmatic and textual) meaning. And since the form cannot be detached from its meaning, this meaning cannot be expressed in any other way, i.e. through paraphrase, explanation or commentary, borrowing of new words etc. In poetry the signifiers have an autonomous value and can therefore not be exchanged for the signifiers of another language, although they may in fact express the same signified concept or referent. Since the physical nature of signifiers in one language can never be duplicated in another language, the relations of signifiers to signified, which are no longer arbitrary in a poetic-aesthetic work, cannot be expressed in another language. (House 1997: 48)

It can be shown that there is a certain circularity (petitio principii) in this seemingly watertight argument which is based merely on the self-evident principle of identity as in the well-known Bishop Butler proposition: “Everything is what it is and not another [page 87↓]thing.” Basically, House’s argument states that you can either reproduce form or content, but not both. The fallacy of this argument consists in the covert assumption that translation means academic translation.

However, House makes the valid point that the linguistic use of a word, phrase or formulation which is bound to a specific culture is at one level untranslatable. House gives many examples contrasting the exaggerated politeness of English with German directness. She also shows that German signs and requests tend to require ‘scientific’ justification which is usually omitted in English. One example will suffice to make this point:

(4) Sign in a hotel bathroom

Lieber Gast! Weniger Wäsche und weniger Waschmittel schützen unsere Umwelt. Bitte entscheiden Sie sich selbst, ob Ihre Handtücher gewaschen werden sollen. Nochmals benutzen: Handtücher bitte hängen lassen. Neue Handtücher: Handtücher auf den Boden legen.

Vs

Dear Guest, will you please decide for yourself, whether your towels shall be washed. Use again: please leave your towels on the towel rack. Clean towels: please put your towels on the floor.

In the German original, but not in the translation, an explicit justification for the request is offered in the first sentence. Further, the German original seems slightly less polite than the translation, i.e. mentioning “bitte” twice may have seemed too much for the German writer, whereas the English translation inserts a “please” in each of the requests. (House 1997: 87)

Similarly, the concept bread is different even in different European languages: let us take a period such as the nineteen fifties as opposed to the present multi-cultural world, the French pain may well be a baguette or a bread roll, the German version could vary from Graubrot and Schwarzbrot to regional varieties whereas the English concept may evoke a traditional brown or white loaf or even white sliced bread depending on the social context²⁸. This line of argument, that languages are unique and are therefore fundamentally untranslatable, has many variations from the famous [page 88↓]Sapir/Whorf hypothesis to the latest edition of Steiner’s After Babel including Jakobson (2000: 113-116) and Quine (2000: 98) .

Steiner (1998: 252-253) is also a vociferous proponent of the ultimate untranslatability of poetry. Ironically, to clinch his argument he quotes the Rumanian poet, Marin Sorescu who, in his poem called “Translation”, claimed that his translation of a classical poem “utterly failed/At the soul”, implying that this is always the case with poetry translations. Incredibly and yet without any reference to the irony of the situation, the poem about the untranslatability of poetry quoted by Steiner is itself an English translation by T. Cribbs. This contradictory attitude is at the heart of the purists’ argument against the translation of poetry. Of course, it is better in one sense to read the poem in the original, if possible, but this is often not possible and so translations of poetry abound alongside original works. A cursory glance at the poetry section of any continental bookshop, particularly in Germany, will reveal that about half the titles are translations; the same is true for the many popular compendia with themes such as love and marriage where thoughts and light poems are chosen from a great variety of international sources. The purists themselves, as is obviously the case with Steiner, must often read poetry translations if they are interested in poetry on a world-wide basis unless they are extraordinary polyglots.

The purists whether Benjamin, Heidegger or Steiner seem to base their arguments on Judaeo/Greco philosophy and ‘myths’, which is well summarised by Barnstone (1993):

After the expulsion from Eden and the Flood, translation was initiated with the third diaspora, the Babelean linguistic dissemination, as an endeavour to return to that Edenic state when Adam gave names to all cattle, and to the birds of the air, and to every beast of the field. Translation sought to regain the universality of that earthly knowledge that was ours before the fall, when we were a single people with a single tongue. God’s dispersal offered an implicit injunction against that knowledge, yet at the same time it hurled mankind into the necessity of translation and the eventual restoration of that single tongue. (Barnstone 1993: 135)

This summary applies particularly to certain aspects of Benjamin’s theories concerning the relationship of the translation to its original.

Die Übersetzung aber sieht sich nicht gleichsam wie die Dichtung im innern Bergwald der Sprache selbst, sondern außerhalb desselben, ihm gegenüber, und ohne ihn zu betreten, ruft sie das Original hinein, an denjenigen einzigen Ort hinein, wo jeweils das Echo in der eigenen den Widerhall eines Werkes der fremden Sprache zu geben vermag. (Benjamin 1961: 63-64)

There are many such statements in Benjamin’s work which refer to a sacred hierarchy of meaning. The phrase im innern Bergwald der Sprache selbst together with the verb betreten implies entering into the holiest of holies (of language) whilst the translation remains “außerhalb”, in the outer darkness. The original is sacred text (whether as a work of art or scripture) which echoes something of the “divine” (and which can be understood in a secularised post-Nietzschean world as a reflection of truth via “reine Sprache”). Thus the translation is at best merely an echo of an echo. It is fortunate that there are two words for echo in German (Widerhall, Echo) to illustrate his point. Yet, as with Steiner, there is a fundamental contradiction at the heart of this despairing attitude because this whole untranslatability thesis is expounded in the introduction to Benjamin’s own translation of Baudelaire’s Tableaux Parisiens. A similar attitude is reflected in Steiner’s main book on translation theory with its title After Babel. In interpreting Benjamin, Steiner resorts to mystical and religious (Cabbalistic) terminology as highlighted by the added emphasis:

A genuine translation evokes the shadowy and yet unmistakable contours of the coherent design from which, after Babel, the jagged fragments of human speech broke off. (Steiner 1998: 67)

The Babel theme is a recurrent motif in Western literature on translation. Derrida (1985) has written in depth on this theme as have many modern critics such as Barnstone as quoted above. Steiner goes on to refer to a pre-Babel Ursprache which is to be understood more in chimerical than philological terms.

The main fallacy lies in the purists’ implied downgrading of the status of the translation as summarised by Nabokov’s poem “On Translating Eugene Onegin” quoted by Steiner in defence of his purist thesis:

What is translation? On a platter

A poet’s pale and glaring head,

A parrot’s screech, a monkey’s chatter,

And profanation of the dead. (Steiner 1998: 252)

This fallacy also ignores the fact that a translation can be an improvement of the original. This is often the case in technical and commercial translation for the simple reason many translators (of the highest standards) are language experts and writers with a good stylistic sense whereas for some engineers or commercial writers language is of secondary importance.

One of the arguments against untranslatability accepts that the translation does not claim to be the same as the original, but that its validity depends on its function in [page 90↓]the target language. This approach is sometimes referred to as “Skopos-theory” as in the exposition of Vermeer (1996) who claims that the main value of the translation is based on its purpose or “skopos” and on its function in the target culture rather than its closeness to the original in the SL. Similarly, Toury (1985) used the term polysystem(coined by Even-Zohar (1978)) for his equally target-oriented approach in which the value of a translation depends on its interaction with other genres within the complex system (polysytem) of the target culture. His definition of translation illustrates this point:

A ‘translation’ will be taken to be any target-language utterance which is presented or regarded as such within the target culture on whatever grounds. (Toury 1985: 20)

This definition is so broad that it would also include ‘pseudo-translations’ such as McPherson’s Ossian, the notorious ‘translation’ of a non-existent text which fooled writers such as Schiller and Goethe and yet which was influential in its time as an inspiration to poets and literati. Even-Zohar (1990), Holz-Mänttäri (1984) and Kußmaul (1995) have also contributed to target-oriented or “functional” theories of translation. In this approach, there is no more searching after a chimerical ideal translation which finds the set of perfect equivalents in L2 for L1, no more striving after the often mythical yet always elusive mot juste. Instead, there are many possible translations so that criteria such as coherence, readability and acceptability assume a new importance. A translation can be assessed as a work in itself, almost or even absolutely independently from its source text. This view is not as radical as it might seem at first sight. Many people in English-speaking countries are only vaguely conscious that the King James Authorised Version of the Bible is only a translation because this text has acquired the status of a ‘holy text’. This point has been made very forcefully by Barnstone (1993) with regard to the New Testament:

So the New Testament, most of which is translated from lost sources, is presented as original gospel, not translation; so the Authorized Version or King James Version of the Bible is popularly perceived to be God’s words, delivered by the Creator in English and sacredly original. (Barnstone 1993: 9)

Luther’s translation of the Bible has a similar status in Germany. The same principle applies to the Schlegel-Tieck-Baudissin translations of Shakespeare. Whilst not being Shakespeare, they are great literary works in themselves and the German literary tradition would have been very different without them. The point is not controversial. It has been forcefully made by Barnstone (1993) among many others:

[. . .] so Geoffrey Chaucer’s Troilus stands alone, without reference to Chaucer’s genius in revising versions from Boccaccio and from French epic love poetry; so Richard Crashaw’s close translation [page 91↓]of Saint Teresa’s famous “Vivo sin vivir in mí” (itself an intralingual glosa of a traditional anonymous poem) goes unrecognized in all editions of Crashaw’s writings; so even W. B. Yeats’s “When You Are Old” (a close version of Pierre de Ronsard’s most famous sonnet)[. . .] (Barnstone 1993: 9)

Numerous other examples can be adduced where translations have gained the status of original works. Indeed, the question arises when is a translation a translation. The great medieval German poets Wolfram von Eschenbach, Gottfried von Strassburg and Hartmann von Aue described their activities as translations, insisting even that theirs were more accurate than other versions, even though their works have a much higher status within the literary canon than their source texts. To suggest to a patriotic French scholar that some of the works of Racine could be regarded as adaptations of their classical sources would be greeted with horror, so high is the canonical status of Racine in the French literature and so low is the status of translators and writers of adaptations. Yet many passages in Racine closely parallel their sources. The polysystem school breaks down these barriers and divisions, thus liberating the translator from the tyranny of the source text.

According to Gentzler, deconstructionists go one step further than the polysystem theoreticians by dethroning of the primacy of the source text even to the point of questioning whether the original could not also be regarded as being dependent on the translation rather than vice versa:

Questions being posed by deconstructionists include the following: What if one theoretically reversed the direction of thought and posited the hypothesis that the original text is dependent on the translation? What if one suggested that, without translation, the original text ceased to exist, that the very survival of the original depends not on any particular quality it contains, but upon those qualities that its translation contains? What if the very definition of a text’s meaning was determined not by the original, but by the translation [. . .] What exists before the original? An idea? A form? A thing? Nothing? (Gentzler 1993: 144-145).

It could equally be a mistake to imply that the translation is more important than the original, but the deconstructionists have the useful function of demythologising translation theory. Whether poetry is translatable or not, there is an enormous literature of translated poems presumably with an even greater readership. At this point, it is relevant to examine the various strategies undertaken by translators of poetry.

It can be argued that the whole field of poetry translation is still in its infancy at the theoretical level despite three millennia of practice²⁹. The past and present states of the theory regarding the translation of poetry are well summarised in The Encyclopaedia of Literary Translation (1998) under the headings The Poetics of Translation and Poetry Translation. There is no need to repeat these excellent summaries written by Gentzler and Venuti respectively, but instead, it will be of greater relevance to examine the language of discourse in this field. In short, it can almost be said ‘anything goes in the theory of poetic discourse translation as there are distinguished theorists, literati and poets who represent more or less every conceivable stance on this most difficult of topics. Based on Lefevere (1975), Bassnett (1991) list of the various possible approaches still applies:

1. phonemic translation (imitation of ST sounds);
2. literal translation (cf. Nabokov);
3. metrical translation (imitation of metre of ST);
4. prose translation (rendering as much sense as possible);
5. rhymed translation (added constraints of rhyme and metre);
6. blank verse translation (no constraint of rhyme but still one of structure);
7. interpretation (complete change of form and/or imitation).
   (Abridged from Bassnett. 1991: 81-82)

More detailed examples of these various stances will be given in the course of this introduction.

There has been much written about poetry translation by poets, translators and literary critics, but there has been little written in a systematic way. The wide range of stances on this issue is also well summarised by Holmes (1978) who also reflects some of the vehemence with which these views are held by the various parties involved:

[page 93↓] What should the verse form of a metapoem be? There is, surely, no other problem of translation that has generated so much heat, and so little light, among the normative critics. Poetry, says one, should be translated into prose. No, says a second, it should be translated into verse, for in prose its very essence is lost. By all means into verse, and into the form of the original, urges a third. Verse into verse, fair enough, says a fourth, but God save us from Homer in hexameters. (Holmes 1978: 94)

In the history of translation and literature, each school of thought has distinguished representatives. It could also be added that the language of discourse has both a moral and absolutist tone which excludes open debate on these matters. It will be useful to begin with the first category mentioned by Holmes (1970) which refers to those poets and theoreticians who are convinced that all poetry in all cases (such is the universalist form of their discourse) should be translated into prose.

The literary critic and translator, John Middleton Murry (1923) is a vigorous supporter of the ‘poetry-into-prose’ school:

Poetry ought always to be rendered into prose. Since the aim of the translator should be to present the original as exactly as possible, no fetters of rhyme or metre should be imposed to hamper this difficult labour. Indeed they make it impossible. (Murry 1923: 129. My emphasis.)

The argument is based on moral exhortations as illustrated by the emphasis. Similarly, the more recent critic, writer and translator Nabokov, whose essay “Problems of Translation: Onegin in English” originally published in 1955, quoted in full in Venuti (2000), takes an equally extreme and absolutist position on this topic. His justification of this stance is based on an uncompromising literalist view of translation:

The term “free translation” smacks of knavery and tyranny. It is when the translator sets out to render the “spirit” - not the textual sense - that he begins to traduce the author. The clumsiest literal translation is a thousand times more useful that the prettiest paraphrase. (Nabokov 2000: 71. My emphasis.)

By his use of the verb traduce, Nabokov implies a severe moral condemnation for the ‘free’ translator, possibly as an echo of the well-known Italian dictum to the effect that traduttore (to translate) equals traditore (to betray).The same tone of moral indignation concerning ‘free’ translators pervades the whole essay:

The person who desires to turn a literary masterpiece into another language has only one duty to perform, and this is to produce with absolute exactitude the whole text and nothing but the text. (Venuti 2000: 77. My emphasis.)

The phrase “the whole text and nothing but the text” is redolent of the oath to be sworn before a jury: “the whole truth and nothing but the truth”. This is to imply that free translation is not only betrayal but is also a form of perjury.

It is, however, not very well known that the poet Robert Browning’s views on poetry anticipate those of the ‘literalist’ school³⁰. Pound and Benjamin also tend towards this approach to translation where the target language is sometimes violated to preserve the rugged and raw nature of the original.

In between the two extremes of translation into prose versus translation into verse, there are, however, other opinions which include grey areas such as those of Matthew Arnold (1909), whose essay “On Translating Homer” originally appeared in 1861, is a slightly less categorical supporter of the poetry-into-prose school since he restricts his dogmatic ban only to the ‘great works’ of literature on account of the variety entailed in such literary monuments:

There are great works composed of parts so disparate that one translator is not likely to have the requisite gifts for poetically rendering all of them. Such are the works of Shakespeare and Goethe’s Faust; and these it is best to attempt to render in prose only. (Arnold 1909: 274)

Although Arnold’s arguments are consistent in theory, they are rather weak in practice as they involve preferring an obscure French prose version of Shakespeare to the universally acclaimed Schlegel-Tieck translations³¹. Similarly, he supports a very weak English prose version of Goethe’s Faust. ³²

At the other extreme, Alexander Fraser Tytler (1791), who was one of the early theoreticians to discuss the problem of poetry translation into English, takes a [page 95↓] diametrically opposite stance to both the translation-into-prose school with an equally confident dogmatism. Tytler asserts:

To attempt, therefore, a translation of a lyric poem into prose, is the most absurd of all undertakings; for those very characters of the original which are essential to it, and which constitute its highest beauties, if transferred to a prose translation, become unpardonable blemishes. (Tytler 1791: 111. My emphasis.)

Again as with Nabokov, opprobrium is supported by ethical threats with Tytler’s use of the adjective unpardonable. Tytler also adds the threat of ridicule to possible opponents of stance by his use of the phrase most absurd. Sometimes, even national prejudices are invoked to support extreme views on poetry translation as in the case of the poet Coleridge:

I do not admit the argument for prose translations. I would, in general, rather see verse in so capable a language as ours. The French cannot help themselves, of course, with such a language as theirs. (Selver 1966: 13)

Entertaining though it may be to consider the diverse opinions of poets and scholars from the past on the topic of translating poetry, it has already seen to be not very illuminating as there are few arguments other than oracular pronouncements based on the supposed authority of the writer or there are dire moral threats for those who dare to disagree. There have, however, been some dispassionate analyses a classic example of which will be treated in 5.3.

With the advent of machine translation from the 1940s, scientific and mathematical approaches dominated linguistic discourse on translation theory from this period up to the end of the 1980s. The elusive concept of equivalence was the key concept that has almost as many definitions as theorists as noted by Gallagher:

Übersetzungsäquivalenz ist bekanntlich ein schwer fassbarer und kein einheitlicher Begriff (vgl.. Koller 1979: 176; Stein 1980: 33-34; Reiß/Vermeer 1984: 124; Nord 1986: 30; Snell-Hornby 1988: 13-22; Gallagher 1993c: 150). Deshalb versuchen viele Forscher Missverständnissen vorzubeugen, indem sie verschiedene Äquivalenztypen unterscheiden. So wird in der übersetzungstheoretischen Literatur von denotativer, konnotativer, inhaltlicher dynamischer, formaler, kommunikativer, pragmatischer und wirkungsmäßiger Äquivalenz gesprochen, um nur acht Beispiele wahllos herauszugreifen. (Gallagher 1998: 1)

Similarly, Koller defines five types of equivalence most of which are included in Gallagher’s list which, but the impression created by both authors is that the list could well be endless.

Although it has been argued that equivalence theories have limited application in the field of literature, it is clear that in other areas such as science, technology and commerce, they can be useful strategies. To give an obvious example, the German noun Spannung can have many different meanings and differing contexts. It could mean tension, stress, voltage, pressure, strain and potential - to name but a few examples. The simplest and most practical definition of equivalence involves finding the correct meaning of the word in the appropriate context, which can also be a matter of life and death. If a notice such as Vorsicht Hochspannung in a context of where Danger High Voltage would be an appropriate translation is wrongly translated as ‘Be cautious - there is a lot of stress about’, this could have fatal consequences for even a wary wanderer on an electrical installation! Obviously such a crass mistake rarely occurs even in the field of technical translation where less dangerous errors abound. Anecdotal evidence alone suffices to make this point. Even here, however, for the experienced translator, stating the necessity for equivalence is merely a case of stating the obvious.

In this Section the ‘classical’ concept of equivalence is connected with its use in mathematics and formal logic. As there is not space to deal with all the various forms of equivalence, a formal refutation of Holmes’ attempt to formalise the process of literary translation will suffice as an example.This is to illustrate the basic theoretical approach of this dissertation which argues that a non-dogmatic and pragmatic use of the notion translation strategies is more fruitful than following the blind alley of the supposedly scientifically based equivalence.

Van den Broeck (1978) defines translation acts in terms of equivalence with sub-categories such as ‘synonymy’ or ‘semantic equivalence.’ He quotes Mates’ definition to describe these terms:

Two expressions are synonymous in a language L if and only if they may be interchanged in each sentence in L without altering the truth value of that sentence.’ (Mates 1950: 209).

It is surprising that Mates does not go one step further and give mathematical form to what is already a mathematical definition. A possible formulation of the above could be as follows: where E refers to any translation act which is defined as equivalent for an item (a) in the source language L1 translated into the target language L2 (b) so that, using standard formal logic notation, the Mates’ definition could be expressed as follows:

E = (L1 (a) ↔ L2 (b))

It must also be noted that (a) ↔ (b) ≠ ((a = b)) or, in other words, equivalence must never be confused with identity even though their truth values may be the same. House rightly expresses a sense of outrage when as distinguished a theoretician as Snell-Hornby fails to make this distinction:

Given the relative nature of ‘equivalence’ and the fact that it has nothing to do with ‘identity’ it is more than surprising that a polemic attack should have been directed against the concept of equivalence, in the course of which an analysis of the English and German dictionary meaning of the term ‘Equivalence’ was presented. Snell-Hornby singles out one dictionary entry, which supports her claim that equivalence basically equals identity and promptly proceeds to dismiss equivalence as ‘an illusion’ in translation studies. She writes that equivalence means ‘virtually the same thing’. By contrast, I found the following dictionary entries for ‘equivalent’ and ‘equivalence’ in my own dictionary searches. ‘having the same value, purpose [. . .] etc. as a person or thing of a different kind (Longman Dictionary of Contemporary English 1995), and having the same relative position or function; corresponding [. . .]’ (Shorter Oxford English Dictionary 1995), as well as ‘equivalence is something that has the same use or function as something else’ (Collins Cobuild 1987). And in German, too, ‘Äquivalenz’ is not only a term in the ‘exakte Wissenschaften’ as Snell-Hornby claims: in my Brockhaus I read: ‘das, was in gewissen Fällen gleiche Wirkung hervorzubringen vermag’. (House 1997: 26)

This is, however, not a debate which should be solved by an appeal to lexicographers because what needs to be made clear is whether equivalence is defined as in mathematical logic, or, as in ordinary language, or again whether a stipulative definition has been made for this term.

In technical translation, there are, however, occasions when formal equivalence and identity are identical such as when dealing with measurements or describing machines, but in literary translation, this is rarely the case. Van den Broeck (1978) also makes this point with his example of the two sentences I am an orphan and I am a child and I have no father and mother as a case of equivalence of reference, but not of sense as ‘orphan’ has all kinds of connotations which would be missed by the mere reference to a child without parents. Literary equivalence is completely different from scientific or logical equivalence. To make this point even more clearly, the following example adapted from Frege (1892) should suffice: from a logical point of view the planet Venus, the Morning Star, the Evening Star and the second nearest planet to the sun within the solar system refer to the same object and are identical and thus the reference (“Bedeutung”) is identical. Particularly from a [page 98↓]literary perspective, the various expressions referring to this particular planet are by no means identical with regard to their sense (“Sinn”). In a hypothetical poem referring to a very amorous poet or even philanderer, a line such as:

(a) My heart leapt for joy when I saw Venus flood the evening sky

is certainly not equivalent with regard to sense to:

(b) My heart leapt for joy when I saw the Evening Star flood the evening sky

as the romantic or in some contexts, erotic connotations are totally lost.

If the Morning Star is substituted, there is a paradoxical effect, but quite different from the original (a) and also, interestingly, from (b):

(c) My heart leapt for joy when I saw the Morning Star flood the evening sky.

If we use the scientific equivalent (d), the effect becomes absurd:

My heart leapt for joy when I saw the second planet nearest to the sun within the solar system flood the evening sky.

The distinction thus needs to be made whether equivalence refers to the sense (Sinn) or whether it is a case of reference (Bedeutung). Most European languages have exact equivalents for the various aspects of ‘sense’ in this case such as in German with the names Morgenstern, Abendstern and Venus, but the problem arises with cultures in which such equivalents are lacking, particularly those of the southern hemisphere where Venus does not appear either at all or at least in the same way. These problems are dealt with by some theoreticians such as Koller (1979: 187-191, 1979: 100-104) who does distinguish between denotative (Bedeutung) and connotative (Sinn) meaning, but more as a matter classification than of strategy. Nida’s concept of “dynamic” equivalence is relevant for the problem of translating for languages in the southern hemisphere:

In contrast, a translation which attempts to produce a dynamic rather than a formal equivalence is based upon “the principle of equivalent effect” (Rieu & Phillips 1954). In such a translation one is not concerned with matching the receptor-language message with the source-language message, but with the dynamic relationship, that the relationship between receptor and message should be substantially the same as that which existed between the original receptor and message. (Nida 2000: 129)

As this involves discovering or finding a strategy which would have the same effect in the target language, the notion of equivalence is again strained to its limits.

Although Van den Broeck’s recognises the difficulty of always finding a translational equivalent his narrowly scientific approach³³, he does admit to the inherent contradiction in this pursuit with the following reasonable concession:

Unfortunately, it will be difficult to find any pairs of expressions in natural languages which meet the very stringent requirements this criterion of semantic equivalence seems to impose. (Van den Broeck 1978: 36)

In a similar vein, Van den Broeck refers to the scepticism of both Mates and Leech with regard to the possibility of ever finding true equivalents, particularly with regard to stylistic aspects:

If we take into account the fact that expressions in context not only have conceptual meanings but also convey connotative, stylistic, affective, reflected, and collocative meanings, it will in fact be difficult to discover any pair of expressions in actual speech which are really equivalent. (Van den Broeck 1978: 36)

Van den Broeck is well aware of the limitations of what has been defined in this dissertation as the academic approach:

In view of the semantic gap between languages and the fact that any text communicates more than mere ‘cognitive’ (or ‘conceptual’) meanings, it is impossible to maintain that, for example, the problem of translating a book from German into English simply amounts to ‘the problem of producing an English version which faithfully reproduces the sense of the original, that is, of producing a book which contains, for every meaningful expression in the German original, a synonymous expression in English, and conversely’ (Mates 1950: 202). (Van den Broeck 1978: 37)

However, his hoped-for solution for an explanatory “elaborate” theory, based presumably on scientific grounds, will be revealed in the next Section to be a chimera:

It would seem to be quite possible to achieve a very elaborate and quite useful theory about literary translation and yet have to admit that we do not know a single law, in the ordinary sense of the word, which it obeys. (Van den Broeck 1978: 45)

An extreme and extraordinary example of the scientific or mathematical approach can be found in Holmes (1970) who attempts to give scientific definitions in [page 100↓]mathematical form for the literary translation process. He uses the term mimetic ³⁴ to describe the approach of using the same verse form of the original and gives Lattimore’s version of Homer’s Odyssey (the opening of Book XI) as an example of this form which retains the hexameter form of the original:

No verse form in any one language can be identical with a verse form in any other, however similar their nomenclatures and however cognate the languages. What in reality happens is that, much as one dancer may perform a pattern of steps closely resembling another’s, yet always somehow different, in the same way the translator taking this first approach will imitate the form of the original as best he can. (Holmes 1970: 95)

This is another example of an argument based on Butler’s maxim: “Everything is what it is and not another thing.” The status of the original is given a transcendental authority like the musical score to the conductor or, to quote Holmes’ analogy, like the choreography to the individual dancer. On the contrary, a great dancer might well make even a mediocre choreography seem brilliant simply by the individual and interpretative manifestation of the choreography. Similarly, a good translator may well translate dull conventional mechanical verse into something brilliant and natural as has already been argued in 5.1.

Holmes then goes on to give a mathematical definition of ‘mimetic’ form:

(1)F_P S ³⁵F_MP

where F_P designates the verse of the original poem, F_MP that of the metapoem (i.e. the translation) and S denotes fundamental similarity.

The purpose of putting these ideas into logical form is unclear. The symbol S ‘is fundamentally similar to’ would seem to be an arbitrary invention of the author on no mathematical basis. It is not clear if logical transitivity rules would apply for the variables x, y and z to produce the following argument: ((x S y & y S z) →x S z). From Van den Broeck’s (1978) definition of equivalence in a collection of essays co-edited by Holmes, it would appear that equivalence as defined by the linguists’ school does not imply transitivity:

[page 101↓] The properties of a strict equivalence relationship (symmetry, transitivity, reflexivity) do not apply to the translation relationship. (Van den Broeck 1978: 33)

In terms of the formal aspects, the argument would appear to be valid if contrary to Van den Broeck as quoted above, equivalence is understood as a transitive relationship. In metalanguage³⁶, however, it would appear to be so in some cases, but not in others. A beloved may be compared with a red rose as in the famous line from the Robert Burns’ poem, “My love is like a red red rose” with associations of beauty, freshness, symmetry, fragrance, ruddiness of lips or cheeks whereas a martyr’s death may also be compared to a rose with different associations such as red blood, the odour of sanctity and the thorns of suffering, but in no way is the similarity logically transitive, because it would imply that the beautiful young woman resembles a martyr undergoing torture and death! Thus, the logical relation ‘is fundamentally similar to’ is not necessarily transitive and yet the symbol S is given the function of a constant.

A secondary point is, however, that fundamental similarity cannot be effectively used in mathematical notation without defining more clearly how ‘fundamental similarity’ differs from ‘superficial similarity’ or what logical constants are used to determine the continuum between ‘identity,’ ‘similarity’ and ‘dissimilarity’. With the same casual disregard for the rules of formal logic, Holmes goes on in the next paragraph to invent another new constant: : : to express the relation of being “analogical to”. Holmes then attempts a mathematical definition of analogy:

The principle underlying this approach, is that of ‘analogical form’, which might be formulated:

FP : PTSL : : FMP : PTTL

Where PTSL indicates the poetic tradition of the source language and PTTL that of the target language. (Holmes 1970: 95-96)

However, to provide a mathematical constant for analogy would introduce the same objections as have already been applied to similarity³⁷ except that, in this context, to [page 102↓]give precise definition of analogy is an even more arduous a task. Even if there were a precise definition of analogy, how this should work as a logical constant is not only not proven by Holmes, it is not even mentioned by him. His introduction of three hitherto unknown constants is doubly confusing because he uses them in the context of traditional constants within formal logic such as his use of brackets and both the implication → and the equivalence ↔ signs and yet he sometimes uses them in a different way from their usual signification as implication and equivalence. The implication sign is sometimes used to mean ‘goes into’ or ‘translates into’ and yet seems to have the force of transitivity by producing derivable arguments. In short, Holmes’ symbolism remains unconvincing at the formal level, but the whole enterprise of trying to find a formal symbolic schema to represent poetry translation would seem to be questionable in the light of the difficulty of finding sufficient consensus at the common sense level of ordinary language. It is one of the major goals of this thesis to try to open debate on these issues and to find some clarity amidst the whole confusion of conflicting ideas concerning literary translation, and specifically the translation of poetry. The refutation of the Holmes’ approach is important in this context to show that at least at the moment mathematical theories produce more confusion than light on this matter.

Outside the garbled formal aspects, the content of Holmes’ article is useful because he makes the following three distinctions: 1) of mimetic form to reproduce the same metrical pattern as the original 2) analogue form which tries to achieve an equivalent effect in the target language 3) form determined by content which implies that the content shapes its own suitable form. Category (3) could be better expressed as appropriacy, i.e. that a form is used which in some ways reflects the content rather than, as according to Holmes, invoking the “mysterious process” of form determining of itself the content. He uses Ezra Pound’s adaptation of the Andreas Divus 1538 translation of the same passage into Latin as an example of ‘content-derivative’ form.

It is a great pity that Holmes does not subject the three passages to detailed analysis and it is perhaps an indictment of the whole ‘mathematical’ school that, despite all the formulations, the article reaches the rather feeble conclusion that a normative approach will not produce the best results, but that this area is in need of further study:

As these three quotations emphasise, (i.e. the three verse translations or adaptations of the Homer passage) there is an extremely close relationship between the kind of verse form a translator chooses and the kind of total effect his translation achieves. It is, in fact, a relationship so central to the entire problem of verse translation that its study deserves our utmost attention - study, not in order to arrive at normative dicta. So it must be, and not otherwise; but to come to understand the nature of the various kinds of metapoem, each of which can never be more than a single interpretation out of many of the original whose image it darkly mirrors. (Holmes 1970: 101-102)

In his later work of Holmes (1978) sees translation as a decision procedure when he discusses the translation of Baudelaire’s poem “La géante” in terms of a hierarchy of correspondences involving ‘homologues’ (SL-bound form) and anologues (TL-bound form):

To return to my hypothetical translator Mr X. Should he, in his English translation of ‘La géante’, ‘retain’ such features as syllabic verse, the twelve- and thirteen -syllable line, the continental rhyme scheme, all of them homologues, that is to say in the English setting parallel in form to the French, but clearly not in function? Or should he choose analogues: syllabotonic verse, ten-syllable lines, the rhyme scheme of the English sonnet? These are obviously momentous choices, and which ones he is to make and which to reject will be determined by the correspondence rules which the translator has consciously or unconsciously chosen on the basis of his confrontative knowledge of the French and English languages, literatures and cultures. (Holmes 1978: 75-76)

Holmes, however, offers no answers other than suggesting directions for translation theory. He proposes that a ‘repertory’ of criteria should be set up to assess translations involving several axes with features such as microstructure, mesostructure, macrostructure on one axis, for example and on another axis, form, meaning, function (morphologue, semasiologue, analogue) and on yet another, this time third-dimensional axis, criteria such as contextuality, intertextuality and situationality. The language and methods like many in the equivalence school resemble those of mathematics and the natural sciences. There are formulations, as has already shown with regard to Mates, which could be given a mathematical form; there is a liberal use of block diagrams and charts; rules are formulated using symbolism or the language [page 104↓]of mathematics such as ‘if and only if’; Cartesian geometrical models are suggested and Linnaeus is explicitly quoted as a model for classification:

The task of working out such a repertory would be enormous. But if scholars were to arrive at a consensus regarding it, in the way for instance, that botanists since Linnaeus have arrived at a consensus regarding systematic methods for the description of plants, it would then become possible, for the first time, to provide descriptions of original and translated texts, of their respective maps, and of correspondence networks, rules and hierarchies that would be mutually comparable. (Holmes 1978: 80-81)

Despite some successes in the field of machine translation which is still only in its infancy as far as sophisticated translation is concerned, it is not surprising that this school has had a relatively minor effect on literary translation because literature cannot easily be reduced to mathematical models. The project to give a precise scientific description of literary translation is doomed from the start. It is even very difficult sometimes to give an imprecise, ad hoc description of a literary text or translation.

Even though Holmes’ goal to describe literary translation in terms of mathematics may be rejected, his final appeal for a more precise and rigorous methodology in translation theory would be welcomed by most, if not all translation theoreticians:

Such goals, of course, the scholars of our generation have tended to reject: they seem to us unattainable, and so outside the range of our less-than vaulting ambition. It is in any case certain that they exceed the grasp of the subjective, largely intuitive and impressionist methods still so often being applied today. And only a more explicit, a more precise, a stricter intersubjective approach holds any promise of greater things to come. (Holmes 1978: 81)

In this Section, it will be shown that other approaches can be more relevant such as Levý’s (1969) semiotic or structural analysis within the Prague school of linguists which reveals how many significant features are hidden even in one sentence of children’s or ‘nonsense’ poetry. The concept of semiotics is used in a special sense with regard to literary translation which is not so much concerned with semiotics as strictly defined by classical linguists such as De Saussure (1959) in which texts are studied as linguistic entities, but more as second-order semiotic systems as defined by Hatim and Mason (1998):

Roland Barthes, particularly in his work on myth, pioneered investigations into what came to be known as second-order semiotic systems. These are systems which, in order to signify, build on other systems. Literature is an ideal example of such systems in that, primarily [page 105↓]through the element of ‘creativity’, it provides an alternative to the real world. (Hatim and Mason 1998: 112)

Thus a text can be seen as a system of signs with its own internal dynamics rather than as a string of lexical items for which equivalents need to be sought. The semiotic analysis is fruitful because it represents a radical break from the traditional hide-bound stances using the outworn faithful/free terminology. Fidelity is now seen in terms of fidelity to a certain semiotic process or language game rather than mere semantic fidelity to a string of lexical items. Levý adopts Klemensiewicz’s (1955) definition of the semiotic approach, which is used specifically in a translation context:

Das Original sollte als ein System und nicht als eine Summe von Elementen betrachtet werden, als organische Ganzheit und nicht als eine mechanische Ansammlung von Elementen. Die Aufgabe des Übersetzers besteht weder darin zu reproduzieren, noch darin, die Elemente und Strukturen des Originals umzuformen, sondern darin, ihre Funktion zu erfassen und solche Elemente der eigenen Sprache anzuwenden, die, soweit wie möglich, deren Ersatz und Gegenwert mit der gleichen funktionalen Eignung und Wirksamkeit sein könnten. (Levý 1969: 21-22. My emphasis.)

Levý then goes to make the inherent semiotic approach in Klemensiewicz’s definition explicit:

Die strukturelle Linguistik findet ihre logische Fortsetzung in der Semiotik, der allgemeinen Theorie von Zeichensystemen, die die Sprache als Code auffaßt, d. h. als einen Komplex von sprachlichen Elementen (z. B. Wortzeichen) und Regeln für deren Kombination. (Levý 1969: 21-22)

Levý’s (1969) attempt at evolving a semiotic theory for poetry translation is particularly interesting because he illustrates the problem of translating poetry in a clear and concrete fashion by taking his examples from Christian Morgenstern whose clever but charming nonsense rhymes may, in fact, seem untranslatable. Levý proves the opposite is the case both at the theoretical level and by concrete examples taken from Max Knight’s translations of Morgenstern. At one level, his examples could be criticised as trivial, but they reveal how a gifted poet such as Morgenstern conceals a multitude of subtleties in a three-line poem. In their very simplicity, they provide a paradigmatic example of the essential problem of literary translation in general and poetry translation in particular. At the same time, they show that even light humorous poetry can conceal a number of subtle language games which are deciphered in Levý’s semiotic analysis.

Invidious though the division between form and content may be and even though the greatest poetry is such a subtle blend of both that form and content can [page 106↓]hardly be distinguished, the translator needs to make this distinction before work is begun. In narrative verse such as popular ballads, content would seem to be the major factor as long as a basic ballad form is maintained whereas with humorous and nonsense poetry, the content could in certain cases almost be said to be the form itself. This point is illustrated by Levý with regard to Christian Morgenstern’s non-sense poem in the following example:

Ein Wiesel

saß auf einem Kiesel

inmitten Bachgeriesel. (Levý 1969: 103)

together with Max Knight’s inventive version:

A weasel

perched on an easel

within a patch of teasel. (Levý 1969: 104)

Levý rightly notes that in such verses the form is far more important than the content:

In Christian Morgensterns Gedicht Das ästhetische Wiesel ist das Reimspiel wesentlicher als die zoologische und topographische Genauigkeit, denn Morgenstern selbst fügt hinzu:

Das raffinier te Tier

Tat’s um des Reimes willen. (Levý 1969: 104)

Knight offers several alternatives claiming that they are equally acceptable and is supported by Levý in this opinion as quoted above. Knight’s ingenious inventions are as follows:

A ferret

nibbling a carrot

in a garret.

Or

A mink

sipping a drink

in a kitchen sink.

Or

A hyena

playing a concertina

in an arena.

Or

A lizard

shaking its gizzard

in a blizzard.

(Levý 1969: 104)

In an academic translation with the stress on semantic equivalence, the ‘poem’ would be absurdly flat and dull:

A weasel

sat on a pebble

in the midst of a ripple of a brook.* (Translation from Levý 1967)

Levý’s semiotic approach attempts to identify the semiotics of the poem and presents these factors diagrammatically. As this approach will be subjected to critical analysis, it is worth quoting Levý in full:

Die Varianten der Übersetzungen von Morgensterns Wortspielen drängen uns die Frage auf, was alle diese Substitutionen eigentlich bewahren, welche Invariante ihnen allen mit dem Original gemeinsam ist. Wenn wir die allen Lösungen gemeinsamen Zügen abstrahieren, können wir folgendes sagen: allen Übersetzungen bleibt gemeinsam die Konfrontation der Reimübereinstimmung von 1. Dem Namen des Tieres, 2. Dem Objekt, zu dem seine Tätigkeit hinstrebt, 3. Dem Schauplatz. In allen fünf Übersetzungen sind gerade nur diese abstrakten Funktionen der drei einzelnen Verse in der Gesamtheit des Wortspiels erhalten und keineswegs die konkreten Bedeutungen der einzelnen Wörter. Anders ausgedrückt haben einige Wörter in Morgensterns Text zwei semantische Funktionen: 1. Eine denotative eigene Bedeutung, 2. Die Funktion in einer Struktur höherer Ordnung (eben diese blieb in den Übersetzungen gewahrt):

Levý rightly claims that there is a hierarchy of priorities³⁸:

[page 108↓] Das literarische Werk ist ein System von sprachlichen Zeichen, von denen einige neben ihrer konkreten denotativen Bedeutung noch eine allgemeinere Aussagefunktion höherer Ordnung haben, d. h. ein Bestandteil von Zeichensystemen höherer Ordnung sind. (Levý 1969: 105)

In this case, the höhere Ordnung would refer to the humorous poetry or Kalauerstil. It is clear that if a translator were commissioned to translate a book of Morgenstern’s lighter verses, possibly to entertain children (and parents!), a prose translation would be absurd as has been shown in the example of literal translation. This also applies to any attempt to stick rigidly to the content in the vain search for equivalents. Versions such as those offered by Knight would be far more acceptable. Even so, objections could be made that they are not really translations, but such objections are easily refuted because they would represent a misunderstanding of the whole translation process. The Morgenstern poem provides the counterargument to the untranslatability school with utmost clarity because Levý’s semiotic analysis shows that certain elements of semantic content are either secondary or irrelevant and that other formal features comprise the essence of the poem, and thus, in a certain sense, provide the content. Levý’s clear analysis shows that, if the poems are conceived as semiotic systems involving a hierarchical structure, then parallel semiotic systems (semiotically though not semantically equivalent) can be produced with similar material to produce a similar effect, thus, from a semiotic point of view, satisfying adequacy criteria for functional³⁹ or dynamic⁴⁰ equivalence. The concept of ‘equivalence’ is now being stretched to the limit in comparison with the other precise scientific definitions in the previous chapter so that, at this stage, the use of this term must be questioned. Again, the question arises concerning the scope of the equivalent, which, in this example, refers to the whole of the three-line verse.

Despite his brilliant analysis, Levý fails to differentiate sufficiently with regard to qualitative matters by blithely categorising the five poetic translations as all equally valid. He quotes Knight’s comments to the same effect:

[. . .] und fügt im Vorwort richtig hinzu, daß anderslautende Übersetzungen ebenso möglich wären. (Levý 1969: 104. My emphasis.)

To regard each version to be equally good would seem to be a fallacy which will be demonstrated by detailed analysis. It is a pity, however, that, in his analysis, Levý seems to be satisfied with functional equivalence as adequate for literary translation. It is, however, the qualitative differentiation which should be the essential activity of [page 109↓]literary translation criticism, but with many linguistic theoreticians, the debate remains merely at the ‘adequacy’ or, in this case, the functional equivalence level. Just as the translator of poetry has to be something of a poet, the literary translation theoretician has to become involved in literary criticism. Barnstone (1993) goes one step further by claiming that literary criticism and translation are identical in that both are concerned with interpretation or “reading”:

Translation theory and literary theory come together in the act common to them both: reading. Reading is an act of interpretation, which is itself an act of translation (an intralingual translation from graphic sign to mind). [. . .] Hence reading is translation and translation is reading. (Barnstone 1993: 7. Author’s italics.)

Certainly, from Levý’s semiotic analysis all the five poems are formally equivalent in that they fulfil the criteria defined by Levý’s analysis, but his analysis is by no means exhaustive. There are other factors such as the naturalness of both the picture painted together with the rhyme, the coherence of the whole picture and the whimsical nature of the humour.

The first version “A weasel/perched on an easel/within a patch of teasel” obviously depends on the very few rhymes for weasel, if this subject of the poem is to be retained semantically. However, the idea of a weasel perched on an easel is awkward in comparison with the weasel sitting on a pebble in a brook. In addition, the noun teasel is obscure in contrast to Bachgeriesel which blends semantically, sonically and even scenically with Kiesel together with its echo of both Geröll and Geräusch (and the idea of rieseln).

The second translation, “A ferret/ nibbling a carrot/ in a garret” would seem to be much weaker, partly because of impure feminine rhyme (ferret/carrot) in a poem where felicitous use of rhyme is paramount and partly because of the inappropriacy of garret which has too poetic and inappropriate literary associations in this context and which result in an incomplete picture so that the last line is something of an anti-climax.

The third version “A mink/sipping a drink/in a kitchen sink” has something of the naturalness and simplicity of the original. The incongruity of the refined activity of sipping a drink contrasting with the banality of kitchen sink strikes a humorous note which compensates for the lack of unity in the original German version where a full natural picture is conveyed as if in three brush strokes with the three very short lines. If this ‘mink’ version were to be further ‘translated’ into a picture, an amusing [page 110↓]scene could be provided by the illustrator such as a very refined mink sipping from a cocktail glass in a very sordid kitchen sink. I would go one step further in Levý’ hierarchy, which places Kalauerstil at the top as the most important element, to maintain that the final ‘court of appeal’ within the semiotic hierarchy for deciding the success of the poem is the very combination of its humorous and aesthetic impact. Thus, this version would seem to have succeeded the most by producing an almost equally humorous and pleasing effect.

The fourth version “A hyena/ playing a concertina/ in an arena” has now moved entirely away from the rodent world and thus is much semantically ‘freer’ than the others. A vivid picture is portrayed as in the original although the absurd effect of playing a concertina would have been best left to the last line if possible to avoid the slight anti-climax of the conventional context in an arena. The last ‘poem’ would seem to be the weakest, A lizard/ shaking his gizzard/ in a blizzard where there seems to be rhyme only for rhyme’s sake without any compensating literary effects such as humour.

There are numerous possibilities which would fulfil Levý’s semiotic criteria. One more example is offered to illustrate this point:

A stoat

Almost afloat

In a castle moat.

This version⁴¹ also fulfils Levý’s criteria of functional equivalence whilst supplying an element of humour with the qualifier almost which could imply that the poor rodent is having trouble keeping its head above water. This version also satisfies more semantic element with the close relative of the rodent world, i.e. the stoat, (even though the semantic elements are relatively minor importance, they are not to be wholly ignored). However, there is the loss of the natural surroundings which is only partially compensated by the relatively exotic medieval background.

From this analysis it can be concluded that being ‘true’ to the original does not necessarily consist in finding a string of semantic equivalents to correspond to each of the original elements but involves the semiotic features included in Levý’s analysis and also other factors which lie outside his analysis, i.e. those more elusive though no less real features such as the mood, the spirit, the diction, the humour and the [page 111↓]naturalness of the original. Thus a new definition of fidelity is emerging. The analysed examples have made this point clear.

It will be seen in this analysis that Gentzler is justified when he asserts that the insights of deconstructionists offer many valuable insights to literary translation theory:

I would like to suggest, however, that the deconstructionists’ entire project is intricately relevant to questions of translation theory, and their thinking is seminal to any understanding of the theoretical problems of that translation process. (Gentzler 1993: 146)

Derrida’s use of his coined word différance as opposed to différence to imply a deferring of meaning in a twilight zone of non-existence, an area between the original writer’s conception of an idea to the infinity of possible translations is certainly at the opposite end of the spectrum from the formalist and ‘scientific’ schools of translation which search after ‘the’ equivalent or, at least, after a restricted number of possible equivalents. It has, however, already been shown that the precision of the scientific schools of translation is illusory. The surface vagueness of Derrida is, by no means, meaningless or too obscure because Derrida has an immediate liberating effect for the translator. It has already been shown that great translations can be great works of art. The translator is now invited to enter into the world of différance, “this bottomless chessboard on which Being is put into play” (Derrida 1982: 22), where the original has no automatic priority and where the translator is free to join the eternal game of deferring meaning, of creating his or her own forms and meaning. The difference and status of original and translation, of author and adapter are now blurred.

To illustrate this theory, it would be helpful to look at a theme in European literature such as the ‘Faust’ legend. The status of the original legend is secondary. The first known compilation of the Faust legend in 1587 was Die Historie des Dr. Faust by an unknown author and was no more than a series of entertaining anecdotes. Christopher Marlowe’s masterpiece, The Tragical History of the Life and Death of Dr. Faustus (1997), first published in 1593, was the first great literary work based on the legend and in a post-Derridean sense could be said to be a ‘translation’ of the original German set of anecdotes into an English tragedy. Goethe’s ‘translation’ of the legend is yet another step in this direction. These literary transformations which could serve as examples of Derrida’s idea of translation as διαφερειν or ‘transporting [page 112↓]across’ can now be taken one step further so that the ‘chasm’ separating ‘original’ work from ‘translation’ is now crossed:

Différance is never pure, no more so is translation, and for the notion of translation we would have to substitute a notion of transformation: a regulated transformation of one language by another, of one text by another. We will never have, and in fact have never had, to do with some ‘transport’ of pure signifiers from one language to another, or within one and the same language, that the signifying instrument would leave virgin and untouched. (Derrida 1981: 20)

Now the Derridean paradox of the source being also dependent on the translation is becoming clearer.

Thus, the highest level of literary translation involves a process akin to the highest level of literary creation. This point will be amply illustrated by the two examples in Sections 5.5.1 and 5.5.2. The medium may change in the process of διαφερειν as is the case with the many musical versions of the Faust legend such as Gounot’s Faust, Berlioz’s La Damnation de Faust, Busoni’s Dr. Faust, Liszt’s A Faust Symphony and Spohr’s Faust to name just a few of the more well-known works. In a post-Derridean context, Visconti’s (1971) film Death in Venice could be said to be a ‘translation’ of Mann’s work, but again into a different medium. These examples of transformation constitute examples of Derrida’s notion of survival in the sense of Fortleben (continuing to live as a work of art) rather than Überleben (merely surviving as manuscript rather than text):

Just as the manifestations of life are intimately connected with the living, without signifying anything for it, a translation proceeds from the original. Indeed not so much as from its life as from its survival (Überleben). For a translation comes after the original and, for the important works that never find their predestined translator at the time of their birth, it characterises the stage of their survival. (‘Fortleben,’ this time sur-vival as continuation of life rather than life as post-mortem.) (Derrida 1985: 178)

This discourse has taken translation theory to its extreme limit, but it is a debate which could well be pursued further. An interesting study could be to show, for example, how much hidden translation there can be in an ‘original’ work and conversely, how much creativity there can be in a successful literary translation. Thomas Mann’s Der Tod in Venedig is a prime example of this. ‘The Phaedo Dialogues’, for example, are a translation of a translation:

Plato may have been the source for both writers. Rilke was a friend of Rudolf Kaßner, whose versions of Phaedrus and the Symposium Thomas Mann used. (Reed 1994: 118)

Another example is Mann’s quotation and translation of the Liebestod scenes from Tristan und Isolde in his novella Tristan – again, the Wagner version itself is a [page 113↓]translation as διαφερειν of Gottfried von Strassburg’s Tristan which in turn was a translation (and improvement on) Chrétien de Troyes’ Tristan et Yseut. To take the post-Derridean discourse one step further, it can be argued that part of Mann’s originality is his ability to ‘translate’ (διαφερειν) his own everyday experiences into story and myth reflecting the fundamental themes that were a constant part of his life. Translation can thus be understood as one of the most creative intellectual activities, but, like great literary writing, the translation process involves a creativity which combines craftsmanship with a perspicacious interpretation of the source text. It is thus not surprising that, according to Gentzler, translation is ‘a’, if not ‘the’ central theme in Derrida’s philosophy:

According to Derrida, all of philosophy is centrally concerned with the notion of translation: “the origin of philosophy is translation or the thesis of translatability.” (Gentzler 1993: 146)

Inspiring though Derrida’s analysis may be, he has little to offer the translator in concrete terms other than advice to the effect that the translator of a great work of literature should simply produce another great work of literature on the same theme in the target language. It is for this reason it will be more useful to illustrate this approach with two case studies: Joyce’s own translation of Finnegans Wake into Italian and two versions of Hölderlin’s translation of twenty lines taken from the chorus of Sophocles’ Antigone.

Poetic discourse can have many levels of richness, ambiguity, density and complexity, but perhaps one of the most complex examples of poetic prose in the twentieth century is James Joyce’s Finnegans Wake. If ever a work has been untranslatable, this must be a prime example. Yet, perhaps precisely for this reason, this text has been tackled by a great number of translators including several German translators such as Ulrich Blumenbach, Reinhard Markner, Dieter Stündel, Friedhelm Rathjen and Arno Schmidt among others whose works appear in Reichert’s (1988) collection of Finnegans Wake translations into German. Reichert makes this point on the inside sleeve of his translation collection:

1998 jährt sich zum 50. Male das Erscheinen von Finnegans Wake, des unverständlichsten Werkes der Weltliteratur. Das Werk gilt als unübersetzbar, und dennoch, oder gerade deshalb, hat es immer wieder Übersetzer und Schriftsteller, Außenseiter und Fachleute gereizt, Übersetzungen zu probieren. Joyce selbst hat dazu den Anstoß gegeben, als er Beckett und andere Freunde ermunterte, eine längere Passage ins Französische zu übertragen. Inzwischen [page 114↓]gibt es größere Auszüge aus dem Werk auf französisch und italienisch. (Reichert 1988: Frontpiece inside sleeve. My emphasis.)

It is unfortunately beyond the scope of this dissertation to compare any of the various attempts to translate sections of Finnegans Wake, other than to mention the fact that some versions of Finnegans Wake can, in my opinion, only be described as brilliant. It would be equally interesting to compare Samuel Beckett’s and Philippe Soupault’s French translation of the “Anna Livia Plurabelle” passage of Finnegans Wake with Joyce’s own Italian translation, but this would be good material for another dissertation. What is relevant in the context of this dissertation is to disprove the untranslatability school and the best counter-example is Joyce’s own translation of a few pages of the “Anna Livia Plurabelle ” chapter into Italian, which will act as the case study for this Section. This translation can also be regarded as a classic illustration of what is meant by the post-Derridean approach.

The first article written on Joyce’s translation into Italian was by Risset (1985) to which Gentzler (1993) then referred when placing Joyce’s translation within a post-Derridean context:

In circles, much of the discussion of deconstruction, translation and the nature of language centers around writing by James Joyce, and strategies preferred by his translators. Perhaps the best example of the practice of “affirmative productivity” as preferred by deconstructionists is James Joyce’s own translation of two passages from Finnegans Wake. (Gentzler 1993: 169)

Whether this is the ideal context or not, it is convenient to discuss this work within post-Derridean and deconstructionist discourse. It attempts to describe the process behind what in the cases under discussion can perhaps only be referred to as ‘genius-level’ translation. The basic meaning of this term is that a great work of art in one language is transformed into a great work art in another language whilst paralleling the original at the deepest level. The second aspect of this definition is that the original does not necessarily have a higher status than the translation, which is fundamentally opposed to the ‘canonised’ status of the original as in the theorising of Benjamin, for example. Joyce’s own Italian translation of Finnegans Wake is a transformation (διαφερειν) of a multilingual text into a monolingual but ‘multi-dialectal’ text. Risset (1985) maintains that Joyce managed to fuse various Italian dialects with the supporting base of Dante’s own rich dialectal usage:

This out-and-out Italianisation, this liberty in the emendation of the text, is based upon the exact and simultaneous use of different areas and levels of Italian: in particular, dialects (above all Venetian, Triestian, Tuscan), literary archaisms (drawn from Dante, from Florentine comedy or [page 115↓]from the poetry and drama of D’Annunzio), specialised idiom (that of the opera for instance). These layers are not juxtaposed, but mixed and fused. Lofty discourse is assimilated, absorbed in the context; one needs a second process of analysis to distinguish under the seemingly homogeneous level of the spoken discourse. (Risset 1984: 12)

This is perhaps one of the reasons why Risset does not like the word translation in this context even though this is the word used by Joyce and his collaborators in this enterprise as she herself does in the title to this essay “Joyce Translates Joyce”:

But this Italian text from Finnegans Wake cannot really be called - in the usual sense of the word - a translation at all; for what takes place is a complete rewriting, a later elaboration of the original, which consequently does not stand opposite the new version as ‘original text’, but as ‘work in progress’. (Risset 1984: 3. My emphasis.)

The italicised phrase a later elaboration of the original is a concrete example of Derrida’s idea of différance or, in other words, deferring the final version, deferring meanings in a flux of reworkings. What is generally not understood is that literary translation at the highest level is both ‘translation’ and rewriting at the same time. As has been argued, the term translation includes a wide variety of interpretative activities as listed from Wilss in 1.

At this point, it is appropriate to analyse one sentence of the text in detail to illustrate Joyce’s translation technique as a form of post-Derridean translation. The full text of Joyce’s Italian translation of the “Anna Livia” chapter is included in the Risset article and the following sentence taken from the “Anna Livia Plurabella” section of Finnegans Wake has been selected for analysis because it has been rightly quoted by Gentzler (1993) as an example of post-Derridean translation although Gentzler does not undertake a detailed analysis:

Annona gebroren aroostokrat Nivia, dochter of Sense and Art, with Spark’s pirryphlickathims funkling her fan [. . .]. (Joyce 1993: 199. My emphasis.)

Joyce’s Italian translation is also taken from the same page in Gentzler:

Annona genata arusticrata Nivea, laureolata in Senso e Arte, il ventaglio costellato di filigettanti [. . .]. (Gentzler 1993: 170. My emphasis.)

As this is such a wonderful example of something that is both translation and rewriting by a great literary and linguistic talent, it might be interesting to begin by dwelling on the translation of one word, the translation of the noun dochter by laureolata. Dochter here is a pun, based on the German nouns Tochter (daughter) and Doktor implying that Annona is both a specialist (Doktor) in “Sense and Art” as well as being formed by them in the most natural way possible (Tochter) whereas the Italianate participle laureolata is a quite different pun but with similar import whilst [page 116↓]being more suited to Italian culture. There is not only the idea of Annona being given the (Roman) ‘laurel’ crown (alloro) by “Sense and Art” because of her accomplishment, but also that art has both beautified and beatified her with its own ‘aureole’ or ‘halo’ (l’aureola). The translation conveys both ideas, or, in other words, she is both very knowledgeable and gifted or ‘blessed’ by ‘sense and art’. The original involves a heavier more Germanic pun suitable for the Anglo-Saxon world of Northern Europe whereas laureolata not only echoes the idea of the l’aureola of ancient Italy but also evokes the saints and painters of medieval Italy such as Giotto’s famous painting of St. Francis of Assisi with his head surrounded by a very powerful aureola or halo. Also laureata is the Italian word for some one who has received a doctorate, which is, at the same time, a close translation of the term doctor in this context.

In the original version, the rhythmical effects are hard and alliterative with the repetition of fricatives, consonant clusters and plosives so that it has an ‘Anglo-Saxon’ or ‘Germanic’ ring. This ‘Anglo-Saxon’ element is also compounded by a hint of humorous obscenity in the phrase funkling her fan taken from the same sentence:

Annona gebroren aroostokrat Nivia, dochter of sense and Art, with Spark’s pirryphlickathims funkling her fan . . . (Emphasis added and underling to highlight the alliterative phonic effect.)

In contrast, the Italian version is sonorous and musical depending on the rich play of the endings such as the repetition of the feminine ending ata echoing the idea and sound of Arte even though this is not a case of perfect rhyme:

Annona genata arusticrata Nivea, laureolata in Senso e Arte, il ventaglio costellato di filigettanti. (My underlining to emphasise the rhymed endings.)

The process can be described in terms of a multi-dimensional analysis or deconstruction of the source text and a radical reconstruction in the translation so that the two texts closely parallel each other as translations whilst, at the same time, reading as profoundly original works of literature - a multi-lingual but basically European text is transformed into a musical and multi-dialectal Italian creation. It involves semiotics, deconstruction and reconstruction and so is, at a deeper level ‘faithful’ to the original, i.e. at the semiotic level as defined in 5.4, whilst, at the same time, reading like an original work with a similar richness, density, ambiguity, profundity and musicality to these elements in the original, which, in turn, is seen as différance or work in progress. Risset maintains that it is the multi-[page 117↓]dialectal language of Dante that holds the work together, but it is not a case of Dante as traditionally understood but more a resurrection and ‘re-creation’of Dante’s language:

Joyce evokes, in other words, something very different from the traditional Dante: not the corpus of culture, not the Bible, but the living root of the language, beyond sense - yet in a way which takes up the same direction experienced by Dante. What is captured is precisely a movement (generally congealed even in the boldest literary ventures) between ‘tongue’ (lingua) and sentence (sentenza), a movement along a line of extreme tension between two levels (also between ‘langue’ and ‘parole’), the opposite of the ‘normal’ movement: not the word which rises from the language and then forgets it, but the word which turns towards the language and ‘excavates’ it. [. . .] The project of translation, as the analysis illuminates, has finally as its deep aim the ‘re-creation’ in the Italian language of the experience of Dante. (Risset 1984: 7. My emphasis.)

However, to describe Joyce’s translation as re-creation is not totally accurate as the Italian version is in some ways a very close translation and to describe it as translation in the narrow academic sense of the word is also not totally accurate because the reformulation is, at the surface level, semantically so different that another process is taking place. This is the area where fidelity is more to the density, musicality and diction (in this case, wit) of the text rather a mere search for semantic ‘equivalents’. This point is recognised by Risset:

At the same time this total immersion in the world of idiomatic Italian is far removed from simple mimesis, from a reduction to the spoken level of discourse. On the contrary, analysing the language of the passage one finds the text organized entirely according to the rules of a poetic language within three levels: rhythm, syntactic structure, phonic texture. (Risset 1984: 7.)

The semiotics of the text involves the aspects “rhythm, syntactic structure, phonic texture” as well as semantic richness (paronomasia with multifarious connotations) and cross-cultural language games. The semantic surface is still retained with Joyce so that it is still a translation, which is also a literary work in itself. It is at this interface (translation versus literary creation) that the surface vagueness of many of Derrida’s pronouncements over the status of the translation vis-à-vis the original begins to make sense.

Risset concludes her analysis with a clear rejection of the traditional equivalence approach in translation theory even though her article, at times, reflects the language of the linguistic approach with frequent reference to concepts such as equivalence itself:

[page 118↓] Every translation which becomes fixed on the problem of semantic and local phonetic equivalence is doomed to fail in its purpose, to miss Finnegans Wake, the stream of Finnegans Wake. But a translation able to rediscover and extend this stream was perhaps only to be achieved in this formula: Joyce translator of Joyce, under the aegis of Dante. (Risset 1984: 3. Risset’s emphasis.)

To translate like Joyce, you have to write like Joyce, or at the very least be a brilliant imitator of his style. Although Risset does not make any explicit reference to Derrida or to other deconstructionists, her article uses similar language to describe this translation process so that there is a tension between the two poles of literary creation on the one hand and close translation on the other, which explains her reluctance to regard the work as a translation, at least “in the traditional sense”. This is another reason why Gentzler’s coining of the term post-Derridean would seem appropriate in this context:

The Italian version affords a special perspective on Joyce’s work, permitting us to analyse in another language what Joyce termed ‘the technique of deformation’, showing how the text is worked and transformed. Moreover (and it is perhaps what emerges most strikingly and fruitfully), in this translation one can catch the complexity and boldness of Joyce’s technique of linguistic arrangement as it were in the very act, to reveal a very rich process, one perhaps unique in this field: namely an exploration of the furthest limits of the Italian language conducted by a great writer; a writer who was not Italian, but, according to his collaborators, ‘italianista unico’. (Risset 1984: 3. My emphasis.)

Risset rightly sees this ‘translation’ as transformation, another Derridean term. This area of creative twilight beautifully illustrates the Derridean notion of différance with the idea of deferring meaning until the whole text is both deconstructed and reconstructed in a seemingly endless Heraclitean flux of parallel but constantly shifting meanings. This is not to imply that Derrida had another hidden agenda of intending to produce a Joycean translation theory, but rather that his ideas, like those of the polysystem and skopos theories, liberate the translator from the surface-level semantic ‘tyranny’ of the original, and, more than this, allow the translator to participate in the process of creative transformation.

I also agree with Gentzler’s coinage post-Derridean, which is not the same as Derridean. Derrida’s ideas provoke, stimulate and often infuriate the reader. Searle (1977), one of the many critics of Derrida as philosopher, has a point with his assertion:

Derrida has a distressing penchant for saying things that are obviously false. (Searle 1977: 203)

Derrida had no blue-print for a new translation theory or, even more so, could be said to be hostile to such an approach so that there is no embarrassment to the post-Derridean literary translation theorist in being labelled in some respects as anti-Derridean because the post-Derrideans, as defined by Gentzler, try to make sense out of the deliberate Derridean chaos, by finding meaning within the Derridean vacuum.⁴² A post-Derridean approach by no means involves abandoning many of the useful insights of the linguists with regard to translation theory, but merely finds the linguists’ categories break down when confronted with the highest literature and that many of the paradoxical assertions of Derrida can, in the hands of a post-Derridean theorist, act as a stimulus for a more radical and successful approach to the translation of the highest literature, as will also be seen in the next case study.

According to Schadewaldt (1970) basing his ideas on an earlier study by Beißner, the translations of Hölderlin fall into three phases, which in themselves reflect a progression from academic translation to ‘foreignising’ or ‘semantic’ translation and finally, to “erneuerndes Nachgestalten” or, in other words, what might be described as post-Derridean ‘transformation’:

Wie die eindringlichen Stilbeobachtungen Friedrich Beißners gezeigt haben, lassen sich an dem Übersetzungswerk Hölderlins vor allem drei Stilstufen unterscheiden: eine umsetzende Übersetzungsart, in der nach der herkömmlichen Weise des Übersetzens der im Zusammenhang erfaßte Sinn in freierer Form im ganzen wiedergegeben wird, eine genau hinhörende, nachformende, in der jedes einzelne Wort ernst genommen und vor allem auch die Wortfolge ernst genommen wird, und eine aus einer neu erreichten tiefen ‘Innigkeit’ des Wort- und Sinnverständnisses geschöpfte frei deutende dichterische Art des ‘Übersetzens’, das [page 120↓]nun kaum mehr ein Übersetzen, sondern ein erneuerndes Nachgestalten ist. (Schadewaldt 1970: 768)

Phase one falls within the scope of what has been defined as academic translation. Phases two and three are well illustrated by Hölderlin’s two translations of the same verses taken from the Greek chorus lines in Sophocles’ Antigone. Both translations will be analysed in this Section because they are eminently suitable for study being amongst the very best creations and poems of the great poet. They reveal a progression from a foreignised translation to an equally close translation, which becomes poetry at the highest level. The context for these magnificent verses concerns Antigone’s tragic decision to bury her brother, Polyneikes against the orders of Creon and the town authorities. Creon has just threatened the death penalty for any one who disobeys his orders. The mood is of extreme tension caused by impending doom and the chorus looks at the situation from the point of view of extreme detachment wondering at the cleverness and power of mortal man who is also capable of great evil and great evil is sure to follow as a result of the conflict:

πoλλα τα dεινa κoυδεν αν-

θρωπoυ δεινoτερoν πελει

τoυτo και τε πoλιoυ περαν

πoντoυ χειμεριω νoτω

χoρει περιβρυχιoισιν

περων υπ oιδμασιν, θεων

τε ταν υπερταταν, Γαν

αφθιτoν, ακαματαν απoτρυεται,

ιλλoμενων αρoτρων ετoς εις ετoς,

ιππειω γενει πoλευων.

κoυφoνoων τε φυλoν oρ-

νιθων αμφιβαλων αγρει

και θηρων αγριων εθνη

πoντoυ τ’ ειναλιαν φυσιν

σπειραισι δικτυoκλωστoις,

περιφραδης ανηρ κρατει

δε μαχαναις αγραυλoυ

θηρoς oρεσσιβατα, λασιαυχενα θ’

ιππoν υπαξεμεν αμφιλoφoν ζυγoν

oυρειoν τ’ ακμητα ταυρoν.(Sophokles 1985: 215-216: lines: 331-352. Accentuation omitted.)

Constantine (1990), however, appropriately sets the context at a much deeper level, expressing the situation in terms of catastrophe and immanence, using Hölderlin’s own vocabulary to interpret the spiritual and philosophical background to the play:

Antigone is set, according to Hölderlin, like his own Empedokeles, at a time of upheaval and change, and is a document of it. The quarrel between Creon and Antigone is in that sense emblematic. A new order is being brought about, a republican one (v. 272). Creon and Antigone struggle in the meantime, at the turning point, as two principles: law and (in Hölderlin’s sense) sobriety versus pure fire, ‘lawlessness’. Creon is ‘förmlich’, she is ‘gegenförmlich’ (v. 272). Antigone pits herself against Creon with an ecstatic violence; she is as bent on conjuring up catastrophe as Oedipus is. Both figures, in Hölderlin’s view, ‘force God to appear’, they bring about immanence precisely in the moment of their tragedy. This hubristic, coercive tendency is present in Empedokeles too, and in Hölderlin’s poetics. The ground of feeling in Hölderlin’s work was always the longing for immanence, and his persistent preoccupation with these two holy texts and with the mechanics of tragedy has undertones of an increasing desperation. Steiner detects in Hölderlin’s Sophocles ‘a solicitation of chaos’, rightly, I think. (Constantine 1990: 295)

Using Hölderlin’s language, law and order are defined as das Organische in contrast with the “ecstatic violence” (das Aorgische) with which Antigone embraces her fate. It is tempting to re-interpret these in the post-Nietzschean sense as das Apollonische and das Dionysische, but something different is meant by Hölderlin’s concept of das Organische as this repressively structuring principle tends towards rigid order, thus towards tyranny, punishment and death rather than to Apollonian beauty creating form, order and harmony (music). Although das Aorgische like Dionysian tendencies is ecstatically destructive, a moral harmony is reached in death through das Aorgische whereas the Dionysian destruction destroys all those who have the misfortune to fall within its orgiastic wake. Das Aorgische is less a “solicitation of chaos” as Steiner and Constantine maintain, but more a solicitation of God within a ‘moral’ self-destructive frenzy culminating in a manifestation or ‘epiphany’ that leads to tragic death and, ultimately to a solemn peace. Hölderlin’s translation reflects the ‘metaphysical’ dimension of these themes.

Hölderlin’s first translation written in 1799 fits neatly into Schadewaldt’s second category, “eine genau hinhörende, nachformende, in der jedes einzelne Wort ernst genommen und vor allem auch die Wortfolge ernst genommen wird,” with the syntax of the very compact Greek compounds being retained:

[page 122↓] Vieles gewaltge giebts. Doch nichts

Ist gewaltiger, als der Mensch.

Denn der schweiffet im grauen

Meer’ in stürmischer Südluft

Umher in woogenumrauschten

Geflügelten Wohnungen.

Der Götter heiliger Erde, sie, die

Reine, die mühelose,

Arbeitet er um, das Pferdegeschlecht

Am leichtbewegten Pflug von

Jahr zu Jahr umtreibend.

Leichtgeschaffener Vogelart

Legt er Schlingen, verfolget sie,

Und der Thiere wildes Volk,

Und des salzigen Meeres Geschlecht

Mit listiggeschlungenen Seilen,

Der wohlerfahrene Mann.

Beherrscht mit seiner Kunst des Landes

Bergbewandelndes Wild.

Dem Naken des Rosses wirft er das Joch

Um die Mähne und dem wilden

Ungezähmten Stiere. (v. 42) (Hölderlin 1969: 792)

This version is more what is traditionally understood as translation even though it is at the extreme end of the foreignising spectrum. It is so close to the Greek that German syntax is strained but not broken. Thus, the compound past participial construction woogenumrauschten is used partially to translate the equally compact Greek construction: περιβρυκιoισιν. Similarly, very unusual compound constructions such as listiggeschlungenen for δικτυoκλωστoις and Bergbewandelndes for oρεσσιβατα reflect something of both the density and the feel of the Greek text. Despite the highly convoluted syntax, the poem works. Although the adjectival past participle listiggeschlungenen would normally be two words, the compound is extremely effective as it is itself ‘cunningly twined together’ so that the syntax reflects the sense, reinforcing, perhaps at a subliminal level, the idea of clever, complex nets being spun to trap even ‘the birds of the air’ (τε φυλον ορνιθων).

In the past participle phrasal construction woogenumrauschten, the strangeness of the archaic form of the noun woogen contributes both to the tone of alienation running through the whole text - man as controller and destroyer of nature - and to the remoteness of an ancient civilisation with a very different culture. Paradoxically, in Hölderlin’s case, it still has the effect of bringing that culture closer to us because it is a strangeness with which we can cope whereas a purely domesticating version arouses the suspicion of trivialisation taking place.

The even more unusual compound present participle Bergbewandelndes is also successful for similar reasons. Yet, in the hands of this consummate poet, it still reads like an original, though difficult poem. Despite the translation being so close to the Greek as almost to offend German grammar rules, the poem works as an original of the highest quality owing to the metre, rhythm and tone. The metrical control together with the solemn diction provides a deep coherence below the surface complexity. It is an absurdly close translation yet a work of original genius at the same time.

Hölderlin’s later version of the same extract written between 1803 and 1804 fits equally well into Schadewaldt’s third category, “eine aus einer neu erreichten tiefen ‘Innigkeit’ des Wort- und Sinnverständnisses geschöpfte frei deutende dichterische Art des ‘Übersetzens’, das nun kaum mehr ein Übersetzen, sondern ein erneuerndes Nachgestalten ist” or, in other words, it is more an example of what Schadewaldt calls Nachdichtung:

Ungeheuer ist viel. Doch nichts

Ungeheuerer, als der Mensch.

Denn der, über die Nacht

Des Meers, wenn gegen den Winter wehet

Der Südwind, fähret er aus

In geflügelten sausenden Häußern.

Und der Himmlischen erhabene Erde

Die unverderbliche, unermüdete

Reibet er auf; mit dem strebenden Pfluge,

Von Jahr zu Jahr,

Treibt sein Verkehr er, mit dem Rossengeschlecht’,

Und leichtträumender Vögel Welt

Bestrikt er, und jagt sie;

Und wilder Thiere Zug,

Und des Pontos salzbelebte Natur

Mit gesponnenen Nezen,

[page 124↓] Der kundige Mann.

Und fängt mit Künsten das Wild,

Das auf Bergen übernachtet und schweift.

Und dem rauhmähnigen Rosse wirft er um

Den Naken das Joch, und dem Berge

Bewandelnden unbezähmten Stier. (Hölderlin 1969: 748-749)

Even here, the categories break down because this version is pure Hölderlin and yet, whatever that phrase may mean to various generations, it is also pure Sophocles. In its style and diction, it has a similar effect to some of Hölderlin’s greatest original poems and in particular, the opening lines resemble his poem Andenken:

Der Nordost wehet

Der liebste unter den Winden

Mir, weil er feurigen Geist

Und gute Fahrt verheißet den Schiffern. Andenken (1-4)

The classical rhythms and structure echo those of the chorus as in this extract:

Denn der, über die Nacht

Des Meers, wenn gegen den Winter wehet

Der Südwind, fähret er aus

In geflügelten sausenden Haüßern. (Antigone) (Hölderlin 1969: 194)

As in the first version, the diction of the chorus is one of a high seriousness, a fine balance between tragic passion and philosophical detachment. Yet the language is slightly more fluent and less difficult than in the first version although there are some similar compounds such as leichtträumend, salzbelebt and rauhmähnig and some difficult constructions which stretch German syntax beyond its normal limits in phrases such as:

und dem Berge

Bewandelnden unbezähmten Stier. (v. 239-240).

When this high impassioned lyrical fluency is, however, coupled with the alien and fascinating Greek-based compounds, the effect is powerful: the poem works as a poem and it works as an impassioned and ‘religious’ interpretation of the Greek chorus. Both versions succeed because they display fidelity to the register of the original, to the density of the text and finally because of the mysterious quality of Hölderlin’s metre. Archaic usage such as the et endings in the third person of verbs as in wehet, for example, or the archaic spellings would normally act as an irritant, but with Hölderlin they work. It probably has something to do with the passionate sincerity of the poet, the high diction and the masterly metrical control. Whatever, as [page 125↓]Constantine rightly states, they are both works of “the highest genius”: the former being more alien, more Greek more tragic and densely philosophical whereas the second version is more lyrical, more purely Hölderlin writing at his best, more accessible and yet retaining many of the features of the former. If there has ever been a counterexample to the untranslatability school, then these two versions alone would suffice. To assert that one is ‘better’ than the other is invidious as they both work in different ways as outlined in this analysis. Constantine also rightly refers to the language of these translations as “an ultimate achievement.” He also adds:

But fully to appreciate the interplay of literal and interpretative translation we should have to take a passage word by word from Greek into German. (Constantine 1990: 296)

Even a superficial glance has been sufficient to establish at least that not only creative translation but also great poetry is taking place. Schadewaldt’s thirty-page analysis of Hölderlin’s translations of Sophocles from the point of view of a classical scholar shows in detail how Hölderlin’s translations work not only as translations but as great poetry in their own right.

There is a transition from the former version where Hölderlin chooses the adjective gewaltig, which is close to the Greek and expresses the violence of the despot whereas in the second version the horror becomes almost metaphysical with the adjective ungeheuer. In Hölderlin’s theory of tragedy, the key tragic event involves confrontation with the deity, a kind of negative epiphany which, in death, leads to a new resolution and unity with the absolute. In this ‘theology’ of tragedy, the concept of the ungeheuer is a key theme as illustrated by Schadewaldt:

Gesammelter, gedrungener und im tiefsten Sinn bestimmter hat Hölderlin dasselbe Verhältnis in den Anmerkungen zum Ödipus und zur Antigone ausgedrückt: “Die Darstellung des Tragischen beruht vorzüglich darauf, daß das Ungeheure (ungehiûre, Unheimliche), wie der Gott und Mensch sich paart und grenzenlos die Naturmacht (das Aorgische) und des Menschen Innerstes (das Individuellste, Organische) im Zorn (Streit) Eins wird, dadurch sich begreift (sich faßt, hält, bestimmt), daß das Grenzenlose-eines durch grenzenloses Scheiden sich reinigt.” (Schadewaldt 1970: 782)

The horror expressed by the chorus is not only tragic but is also metaphysical. There is something eerie and alien in the way they express fascinated horror at the human species and the impending confrontation of the finite human being with the Ultimate. The human beings dominate the seas, exhaust the ‘holy’ earth itself and not only tame the savage beasts, but even subject them to their own purposes.

For Hölderlin, the plays expressed his own view of life at the deepest level so that Hölderlin approaches the translation with a feeling of awe as if he is interpreting and re-creating a holy text, as implied by Schadewaldt who maintains that this comes from Hölderlin’s religious attitude to the plays and refers to Hölderlin’s translations as “Übersetzungen aus religiösem Geist.”:

[. . .] so mußte ihm auch das Wort des Sophokles als heilges, gottgesprochenes Wort erscheinen und das Geschäft des Übersetzens, wie überhaupt das Geschäft des Dichtens als ein heiliges Geschäft, dazu bestimmt, das ursprünglich Wort des griechischen Dichters neu zu verwirklichen. (Schadewaldt 1970: 805-806)

It is only since the twentieth century that Hölderlin’s translations have been recognised to be of an outstanding quality, having a similar impact as the original version, even though, in his time, they were treated with derision. Schadewaldt’s article offers at least some explanation as to why this was the case. In the first place, the version of Sophocles used by Hölderlin was faulty and secondly, there were numerous printing errors in the first edition of his translation:

Um mit dem Äußerlichen zu beginnen, so sei zunächst der Tatsache gedacht, daß der Text der Erstausgabe von 1804 - die Handschrift des Dichters selbst ist uns bisher verloren - durch Druckfehler auf gröbste entstellt ist. (Schadewaldt 1970: 770)

In addition to the numerous printing errors, there were frequent misinterpretations caused by Hölderlin’s relatively limited knowledge of Greek. After confronting the reader with several scholarly and eminent translations set side by side with those of Hölderlin, Schadewaldt rightly comes to this conclusion:

In all seiner mangelnden Wort- und Regelkenntnis, Kenntnis der üblichen Verstehensroutine, blieb Hölderlin auch vor aller jener übersetzerischen Routine gleichsam fromm bewahrt, die gängigen Übersetzungen seiner und späterer Zeit so korrekt und zugleich belanglos machen. Instinktkräftig ergriff er zumal den ‘Klang’ des Sophokleischen Wortes mit Ernst in seiner Sachlichkeit und Gründlichkeit, aus jener Verantwortung für die Sprache, die alle poetisierenden Unarten nicht erst abzutun braucht. Was ihm so gelang, ist ihm über die Maßen gelungen: Chorlieder wie auch die großen Reden in ihrer Härte, Dichte, Sachlichkeit des Worts. (Schadewaldt 1970: 777. My emphasis.)

Faced with two versions of highest genius, it is not surprising that even an eminent Hölderlin scholar and able translator of Hölderlin’s poetry such as Constantine himself, prefers simply to let them stand by themselves for the admiration of the reader. His only comment rightly summarises the situation:

Wherever that comes from, by whatever means, it is the highest poetry. (Constantine 1990: 298)

Steiner (1998) goes so far as to assert that commentary on Hölderlin’s poetry translation is an “impertinence”, which is, my opinion, an exaggerated view, but shows the extent of awe and reverence Hölderlin’s translation oeuvre inspires:

We find ourselves here at the far limits of any rational theory or practice of linguistic exchange. Hölderlin’s is the most exalted, enigmatic stance in the literature of translation. It merits constant attention and respect by virtue of the psychological risks implied and because it produced an intensity of understanding and ‘re-saying’ such as to make commentary impertinent. (Steiner 1998: 350)

However, we are here at the heart of literary translation theory and practice. Both versions have been compared from translation theoretical point of view, bearing in mind that clear-cut categories break down in the hands of a poet whose language is able to ascend to the very highest levels of genius.

Hölderlin could have chosen to translate many poets such as Homer or Virgil, perhaps, but he wisely kept his range limited to those poets who expressed his own most inner feelings, philosophy and poetry and so, Hölderlin qua Hölderlin can almost perfectly render what is regarded as Sophocles, the great tragedian, afresh to many generations. His translation may be described as re-inspiration. Whatever Sophocles may have felt or tried to express, it is as if Hölderlin felt and expressed the same kind of emotion anew, afresh and this is why his translations stand above all others in the German language. Other translators look at the words and merely render the same words either into felicitous or infelicitous formulations in the target language or “so korrekt und so belanglos” as Schadewaldt refers to the later translations. It is a great tribute to Hölderlin that a Greek scholar of the stature of Schadewaldt pays the poet such homage. If any one were to take exception at the great translations of Hölderlin, it could well be a Greek scholar who is only too aware of the linguistic limitations of the poet with regard to the Greek language and of missed nuances, even of his not too infrequent gross errors. There have been many tributes to Hölderlin as a translator, the most noteworthy being the encomium of Steiner (1998), but it is perhaps more appropriate to leave the final word with the sober scholarship of Schadewaldt:

Die Einheit indessen, in der die Extreme doch wieder miteinander verbunden sind, ist der Gott und das von Hölderlin mit tiefem Recht als Grund und Inhalt der Tragödie erkannte Gottesgeschehen. In ihr ist er dem Sophokles so nahe gekommen wie kein anderer Übersetzer.(Schadewaldt 1970: 805-806. My emphasis.)

Translators such as Hölderlin act as a contradictory to the ‘untranslatability’ school by showing that great poets can successfully translate great poetry. It may be the case that only great poets can translate great poetry, but even this opinion needs to be proven. The other main conclusion to emerge from this study is that serious literary translation is less a matter of being highly qualified in both languages (reference has already been made to Hölderlin’s relatively limited knowledge of Greek), but more a case of sharing the same inspiration, the same muse as the poet one is translating. Sager (1966) cites the Brazilian poet and translator Manuel Bandeira who illustrates this point with regard to his own translation methodology:

Moreover, I only translate successfully, those poems that I myself should like to have written, that is to say, those poems which express things that were already within me, although my “discoveries” in translations as in my original poems, are always the result of my intuitions. (Sager 1966: 198)

Hölderlin was certainly one of the greatest practitioners of poetry translation and so, he acts as a model for poetry translators, even if his example seems impossible to follow. Hölderlin, unlike many literary translators, seemed to know his limitations in that he translated only those works in harmony with his own muse. Similarly, it is important for a translator of Thomas Mann to display some affinity with the great author. It is regrettable that the two translations of Thomas Mann in this study fail even to indicate that poetry is taking place and that the elements of self-caricature, humorous decadence and literary virtuosity are missing and replaced by merely pretentious ungainly prose displaying here and there a slightly poetic moment, but with the general sense, tone, diction, form, rhythm and poetry being completely lost in translation. The example of Hölderlin shows that this need not be the case.

The next chapter will analyse some the classical poetry encoded in Der Tod in Venedig and compare not only the versions of the two translators in question but will also look at one French and three Italian versions which do capture something of the poetry of the original.

Footnotes and Endnotes

²⁸  De Man (1986) makes the same point with regard to French culture: “To mean ‘bread’, when I need to name bread, I have the word Brot, so that the way in which I mean this is by using the word Brot. The translation will reveal a fundamental discrepancy between the intent to name Brot and the name Brot itself in its materiality, as a device of meaning. If you hear Brot in the context of Hölderlin, who is so often mentioned in this text, I hear Brot und Wein necessarily, which is the great Hölderlin text that is very much present in this - which in French becomes Pain et vin. ‘Pain et vin’ is what you get for free in a restaurant, in a cheap restaurant where it is still included, so pain et vin has very different connotations from Brot und Wein. It brings to mind the pain français, baguette, ficelle, bâtard, all those things - I now hear in Brot ‘bastard’. This upsets the stability of the quotidian.” (De Man 1986: 87)

²⁹ 

It can be seen from Hatim and Mason (1990) who quote and translate a text taken from Badawi (1968: 33) that the same issues such as literalness versus free and equivalence versus the impossibility of equivalence were current even in the fourteenth century. The procedure quoted below provides a good illustration of the approach of the more ‘scientific’ wing of the present-day equivalence theoreticians: “The ‘literal’ versus ‘free’ controversy has been more or less a constant in translation studies, no matter how far back one goes. The extreme case is that referred to by the fourteenth-century translator Salah al-Din al-Safadi who, writing about earlier generations of Arab translators, complains that they look at each Greek word and what it means. They seek an

equivalent term in Arabic and write it down. Then they take the next word and do the same, and so on until the end of what they have to translate. Al-Safadi faults this method of translating on two counts:

1. It is erroneous to assume that one-for-one equivalents exist for all lexical items in Greek and Arabic.

2. The sentence structure of one language does not match that of another.” (Hatim and Mason 1990: 15-16)

³⁰  Browning’s notes are taken from the diary of John Addington Symonds as quoted by Selver (1966). Poets tend towards dogmatic extremes in their theoretical discourse as illustrated by the added emphasis in the following extract: “Browning’s theory of translation. Ought to be absolutely literal, with exact rendering of words, and words placed in the order of the original. Only a rendering of this sort gives any real insight into the original. Fitzgerald’s ‘Omar Khayam’ - a fine English poem but no translation [. . .]. Let it be said, then, that the translator of a poem is not entitled to tamper with the original. He should omit nothing essential. He should add nothing extraneous. It is primarily by unsubstantiated additions that the mediocre or slovenly translator betrays himself. Frequently he indulges in them merely to engineer a rhyme which would otherwise elude him. The adroit, inspired translator is never reduced to such a shift as that. His skill in this respect may be described as a knack, in the same way that juggling billiard balls is a knack.”(Selver 1966: 26. My italics.)

³¹  He states in main work on translation theory On Translating Homer: “People praise Tieck and Schlegel’s version of Shakespeare. I for my part would sooner read Shakespeare in the French prose translation, and that is saying a good deal; but in the German poet’s hands, Shakespeare so often gets, especially where he is humorous, an air of what the French call niaiserie and can anything be more un-Shakespearean than that? Again Mr Hayward’s prose translation of the first part of ‘Faust’ is not likely to be surpassed by any translation in verse.”(Arnold 1909: 274)

³²  A brief quotation from his translation of the opening lines of the Walpurgisnacht scenes will suffice to show that the quality of this prose translation can hardly be taken to be superior to the verse of Schlegel-Tieck: “Do you not long for a broomstick? For my part, I should be glad of the sturdiest he-goat. By this road we are still far from our destination.” (Selver 1966: 14)

³³ 

Interestingly, Van den Broeck’s more recent publications imply a shift of stance away from ‘hard-edged’ linguistics to a more literary approach with a tolerant attitude: “Contrary to what I thought some eight years ago, Derrida’s philosophical approach may offer a substantial theoretical basis for explaining and describing translational phenomena.” (Van den Broeck 1995: 4)

Even though this is a very valuable and courageous concession on behalf of Van den Broeck, it is interesting that he is still searching for what in my opinion is the chimera that a scientific theory of literary translation is possible. This approach is revealed by his choice of phrases such as substantial theoretical basis and explaining translational phenomena when the whole thrust of Derrida is to avoid being pinned down by a scientific approach.

³⁴  Steiner rightly notes: “This word has along and chequered history”. (Steiner 1998: 268) On pages 267-268, he discusses Dryden’s use of this term, which could be redefined in terms of functional equivalence. See (Nida 2000: 129) for a definition of this term and also the discussion in 5.3.

³⁵  The diagonal S is used to represent Holmes’ horizontal S which has a point or dot placed below the middle of the symbol. This is because Holmes’ symbol is an entirely new symbol which does not exist in any symbol index. If S is seen as a metasymbol, then the intended logic of the argument is not impaired.

³⁶  This term also used by Holmes is applied here in its strictly logical sense: i.e. the metalanguage of the formal logic in this context is ordinary language.

³⁷ 

There is neither an explanation nor a definition of the new, arbitrarily invented constant::: .He then ‘derives’ a further relationship from (1) and (2) already quoted:

where C_P indicates the ‘content’, the non-formal material, of the original poem, C_MP that of the meta-poem, and the translingual process. (1970: 96)

Holmes goes on to try to give a precise mathematical form to what he calls the “organic form”:

In the ‘derivations’ not only new terms CP for content but also even more new constants → (TR) for the translingual process are arbitrarily introduced and yet traditional constants such as the use of brackets, → (implication) and ↔ (equivalence) are used with the new nomenclature to reach this final (invalid) conclusion:

The reverse arrow is introduced without explanation. It is obvious that there is no formal logical validity between his various statements nor any clear logical relationship between those propositions which could theoretically stand on their own as descriptive symbolised statements. In such a case, the question arises as to what purpose is served by the use of this hybrid symbolism.

³⁸  Gutt (2000) questions the validity of setting up functional hierarchical criteria: “[. . .] it is not clear on what principles Levý’s hierarchy is constructed [. . .]. Thus the overall organization of this hierarchy remains unclear.” (Gutt 2000: 383-384). At the formal linguistic level, Gutt has a point, but he goes on to say that this is not of prime importance: “However, it seems more than doubtful anyway that ‘such functional hierarchies’ play any significant role here at all. What is actually being done here can be both accounted for and evaluated in terms of interpretive use within the relevance-theoretic framework.” (Gutt 2000: 384). As the scientific linguistic approach to these problems has been found to be inadequate, the methodology of this section has been from the outset “within the relevance-theoretic framework.” Although a critic from the cognitive approach, Gutt admits in conclusion that the interpretive approach is the more fruitful for this kind of analysis: “Thus it seems that an account of translation as interlingual interpretive use has much to commend it. In fact, it could be said to achieve what translation theory has been attempting to do for a long time - that is to develop a concept of faithfulness that is generally applicable and yet both text- and context-specific.” (Gutt 2000: 384).

³⁹  See Osers’ clear definition of functional equivalence in the footnote 26.

⁴⁰  See also (Nida 2000: 129) together with the discussion 5.3

⁴¹  This version was supplied by Andrew Gledhill.

⁴²  An example of the inherent ambiguity in the various Derridean approaches can be invoked to provide a theory which would seem to be the opposite of Gentzler’s, i.e. that words are fundamentally untranslatable as is argued by Delabastita (1997): “Deconstructionists have indeed a clear tendency to conceive translation in function of problematic words (especially proper names and polysemic words) rather than texts [. . .] as well as to promote untranslatability to an absolute principle or blanket rule. [. . .] These are certainly among the points that would need to be re-examined if deconstructionist critics and more empirically orientated translation scholars should one day attempt to meet halfway.” (Delabastita 1997: 226-27; Emphasis as in the original text.) Davies (1997), on the other hand, holds the middle ground between Gentzler and Delabastita: “If a text were totally translatable, it would exhibit no difference from some other text (its translation), and it would, therefore, disappear into that with which it would already be identical. Likewise, in order to be totally untranslatable, a text would bear no relation at all to the language system(s) in which other texts are written: irrevocably self-contained, it would die immediately. Both of these scenarios are unrealistic, of course. Derrida is, in part, pointing out what translation scholars well know: translation is always relative, and relative translation is always possible.” (Davies 1997: 33)