The Translator's Task
Walter Benjamin

When seeking knowledge of a work of art or an art form,
it never proves useful to take the receiver into account. Not only is
every effort to relate art to a specific public or its representatives
misleading, but the very concept of an "ideal" receiver is spurious
in any discussion concerning the theory of art, since such
discussions are required to presuppose only the existence and
essence of human beings. Art itself also presupposes man's
corporal and spiritual essence — but no work of art presupposes his
attention. No poem is meant for the reader, no picture for the
beholder, no symphony for the audience.

Is a translation meant for readers who do not understand
the original? That would seem sufficient to explain the differing
status of original and translation in the domain of art. In any event,
it appears to be the only possible reason for saying "the same
thing" over again. What does a poem "say," then? What does it
communicate? Very little, to a person who understands it. Neither
message nor statement is essential to it. However, a translation that
seeks to transmit something can transmit nothing other than a
message —- that is, something inessential. And this is also the
hallmark of bad translations. But what then is there in a poem —
and even bad translators concede this to be essential — besides a
message? Isn't it generally acknowledged to be the
incomprehensible, the secret, the "poetic"? That which the
translator can render only insofar as he — also writes poetry? This
in fact leads to another distinguishing mark of bad translation,
which can be defined as inexact transmission of an inessential
content. And we never get beyond this, so long as translation
claims to serve the reader. If it were intended for the reader, then
the original would also have to be intended for the reader. If the
original is not created for the reader's sake, then how can this
relationship allow us to understand translation?

Translation is a mode. In order to grasp it as such, we have
to go back to the original. For in it lies translation's law, decreed
as the original's translatability. The question of a work's
translatability has two senses. It can mean: will it ever find, among
the totality of its readers, an adequate translator? Or, more
pertinently, whether by its very essence it allows itself to be
translated, and hence — in accord with the meaning of this mode
— also calls for translation. In principle, the first question can be
answered only in a problematic manner, the second apodictically.
Only superficial thinking will, by denying the independent sense
of the second question, declare them to have the same meaning. In
opposition to this it must be pointed out that certain relational
concepts gain their proper, indeed their best sense, when they are
not from the outset connected exclusively with human beings.
Thus we could still speak of an unforgettable life or moment, even
if all human beings had forgotten it. If the essence of such lives or
moments required that they not be forgotten, this predicate would
not be false, it would merely be a demand to which human beings
had failed to respond, and at the same time, no doubt, a reference
to a place where this demand would find a response, that is, a
reference to a thought in the mind of God. The translatability of
linguistic constructions would accordingly have to be taken into
consideration even if they were untranslatable by human beings.
And mustn't they actually be untranslatable to a certain degree, if
a rigorous concept of translation is applied? In that case we must
ask whether the translation of certain linguistic constructions is
required. For this proposition is relevant here : if translation is a
mode, then translatability must be essential to certain works.

Translation is properly essential to certain works : this
does not mean that their translation is essential for themselves, but
rather that a specific significance inherent in the original texts
expresses itself in their translatability. It is clear that a translation,
no matter how good, cannot have any significance for the original.
Nevertheless, it stands in the closest connection with the original
by virtue of the latter's translatability. Indeed, this connection is all
the more intimate because it no longer has any significance for the
original itself. It can be called a natural connection, and more
precisely a vital connection. Just as expressions of life are
connected in the most intimate manner with the living being
without having any significance for the latter, a translation
proceeds from the original. Not indeed so much from its life as
from its "afterlife" or "survival" [Überleben]. Nonetheless the
translation is later than the original, and in the case of the most
significant works, which never find their chosen translators in the
era in which they are produced, indicates that they have reached
the stage of their continuing life [Fortleben]. The notion of the life
and continuing life of works of art should be considered with
completely unmetaphorical objectivity. Even in ages of the most
prejudiced thinking it has been suspected that life must not be
attributed to organic corporeality alone. But there can be no
question of extending its dominion under the feeble aegis of the
soul, as Fechner attempted to do; not to mention defining life on
the basis of still less decisive aspects of animal life such as
sensitivity, which betokens life only occasionally. Rather, it is only
when life is attributed to everything that has a history, and not to
that which is only a stage setting for history, that this concept
comes into its own. For the range of the living must ultimately be
delimited on the basis of history and not of nature, without
mentioning such unstable notions as sensitivity and soul. From this
arises the philosopher's task, which is to understand all natural life
on the basis of the more comprehensive life of history. And isn't
the continuing life of works incomparably easier to recognize than
that of creatures? The history of great works of art knows about
their descent from their sources, their shaping in the age of the
artists, and the periods of their basically eternal continuing life in
later generations. Where it appears, the latter is called fame.

Translations that are more than transmissions of a message are
produced when a work, in its continuing life, has reached the age
of its fame. Hence they do not so much serve the work's fame (as
bad translators customarily claim) as owe their existence to it. In
them the original's life achieves its constantly renewed, latest and
most comprehensive unfolding.
As the unfolding of a special, high form of life, this
unfolding is determined by a special, high purposefulness. Life and
purposefulness — the connection between them seems easily
accessible but nevertheless almost escapes knowledge, disclosing
itself only where that purpose, toward which all the particular
purposes of life tend, ceases to be sought in its own sphere, and is
sought instead in a higher one. All purposeful phenomena of life,
as well as life's purposefulness itself, are in the final analysis
purposeful not for life, but for the expression of its essence, for the
representation of its significance. Thus translation ultimately has
as its purpose the expression of the most intimate relationships
among languages. Translation cannot possibly reveal or produce
this hidden relationship; however, translation can represent this
relationship, insofar as it realizes it seminally or intensively. In
fact, this representation of the intended object by means of an
incomplete form or seed of its production is a very special mode
of representation seldom to be encountered in the domain of nonlinguistic
life. For in analogies and signs non-linguistic life has
types of reference other than intensive, that is, anticipatory,
intimating realization. — This imagined, inner relationship among
languages is, however, a relationship of special convergence. It
consists in the fact that languages are not alien to each other, but
a priori, and independently of all historical connections, related to
each other in what they want to say.

With this attempt at an explanation the discussion seems
clearly to have come out, after futile detours, at the traditional
theory of translation. If the relationship among languages is to
demonstrate itself in translations, how could it do so except by
conveying the form and sense of the original as accurately as
possible? Of course, the traditional theory would scarcely be able
to define this concept of accuracy, and thus could give no account
of what is essential to translation. In truth, however, the
relationship among languages shows itself in translations to be far
deeper and more definite than in the superficial and indefinable
similarity of two literary texts. To grasp the true relationship
between original and translation, we must undertake a line of
thought completely analogous, in its goal, to those taken by critical
epistemology in demonstrating the impossibility of a reflection
theory. Just as in critical epistemology it is shown that there can be
no objective knowledge, or even the claim to such knowledge, if
the latter consists in reflections of the real, so here it can be shown
that no translation would be possible if, in accord with its ultimate
essence, it were to strive for similarity to the original. For in its
continuing life, which could not be so called if it were not the
transformation and renewal of a living thing, the original is
changed. Established words also have their after-ripening. What
might have been the tendency of an author's poetic language in his
own time may later be exhausted, and immanent tendencies can
arise anew out of the formed work. What once sounded fresh may
come to sound stale, and what once sounded idiomatic may later
sound archaic. To seek what is essential in such transformations,
as well as in the equally constant transformations of sense, in the
subjectivity of later generations rather than in the inner life of
language and its works, would be — even granting the crudest
psychologism — to confuse the ground and the essence of a thing;
or, putting it more strongly, it would be to deny, out of an
impotence of thought, one of the most powerful and fruitful
historical processes. Even if one were to consider the last stroke of
the author's pen the work's coup de grâce, that would not suffice
to save this dead theory of translation. For just as the tone and
significance of great literary works are completely transformed
over the centuries, the translator's native language is also
transformed. Indeed, whereas the poetic word endures in its own
language, even the greatest translation is destined to be taken up
into the growth of its language and perish as a result of its renewal.
Far from being a sterile similarity between two languages that have
died out, translation is, of all modes, precisely the one called upon
to mark the after-ripening of the alien word, and the birth pangs of
its own.

If the kinship of languages manifests itself in translation,
it does so otherwise than through the vague similarity of original
and copy. For it is clear that kinship does not necessarily involve
similarity. In this context the notion of kinship is in accord with its
narrower usage, to the extent that in both cases it cannot be
adequately defined by similarity of origin, although the concept of
origin remains indispensable in defining the narrower usage. —
Wherein can the kinship of two languages be sought, apart from a
historical kinship? No more in the similarity of literary texts than
in the similarity of their words. All suprahistorical kinship of
languages consists rather in the fact that in each of them as a
whole, one and the same thing is intended; this cannot be attained
by any one of them alone, however, but only by the totality of their
mutually complementary intentions : pure language. Whereas all
the particular elements of different languages — words, sentences,
structures — are mutually exclusive, these languages complement
each other in their intentions. To gain a precise understanding of
this law, one of the most fundamental laws of the philosophy of
language, it is necessary to distinguish, within intention, the
intended object from the mode of its intention. In "Brof and
"pain" the intended object is the same, but the mode of intention
differs. It is because of their modes of intention that the two words
signify something different to a German or a Frenchman, that they
are not regarded as interchangeable, and in fact ultimately seek to
exclude one another; however, with respect to their intended
object, taken absolutely, they signify one and the same thing. Thus
whereas these two words' modes of intention are in conflict, they
complement each other in the two languages from which they
stem. And indeed in them the relation between the mode of
intention and the intended object is complemented. In the
individual, uncomplemented languages, the intended object is
never encountered in relative independence, for instance in
individual words or sentences, but is rather caught up in constant
transformation, until it is able to emerge as pure language from the
harmony of all these modes of intention. Until then it remains
hidden in the various languages. But if languages grow in this way
until they reach the messianic end of their history, then it is
translation that is ignited by the eternal continuing life of the work
and the endless revival of languages in order to constantly test this
sacred growth of languages, to determine how distant what is
hidden within them is from revelation, how close it might become
with knowledge of this distance.

To say this is of course to admit that translation is merely
a preliminary way of coming to terms with the foreignness of
languages to each other. A dissolution of this foreignness that
would not be temporal and preliminary, but rather instantaneous
and final, remains out of human reach, or is at least not to be
sought directly. Indirectly, however, the growth of religions ripens
into a higher language the seed hidden in languages. Thus
translation, although it cannot claim that its products will endure,
and in this respect differs from art, does not renounce its striving
toward a final, ultimate, and decisive stage of all linguistic
development. In translation the original grows into a linguistic
sphere that is both higher and purer. It cannot, however, go on
living indefinitely in this sphere, since it is far from attaining it in
all parts of its form; but it nevertheless at least points, with
wonderful penetration, toward the predetermined, inaccessible
domain where languages are reconciled and fulfilled. The original
does not attain this domain in every respect, but in it lies that
which, in a translation, is more than a message. This essential
kernel can be more precisely defined as what is not retranslatable
in a translation. One can extract from a translation as much
communicable content as one wishes, and this much can be
translated; but the element toward which the genuine translator's
efforts are directed remains out of reach. It is not translatable, like
the literary language of the original, because the relation between
content and language in the original is entirely different from that
in the translation. In the original, content and language constitute
a certain unity, like that between a fruit and its skin, whereas a
translation surrounds its content as if with the broad folds of a
royal mantle. For translation indicates a higher language than its
own, and thereby remains inappropriate, violent, and alien with
respect to its content. This fracture hinders any further translation,
and at the same time renders it superfluous. For every translation
of a work at a specific point in the history of language represents,
with respect to a specific aspect of its content, translation into all
other languages. Thus translation transplants the original into an —
ironically — more ultimate linguistic domain, since it cannot be
displaced from it by any further translation, but only raised into it
anew and in other parts. It is not for nothing that the word
"ironically" reminds us here of Romantic modes of thought. The
Romantics, more than any others, gained insight into the life of
works of art, to which translation bears the highest witness. The
Romantics, of course, hardly recognized the significance of
translation, turning their attention instead entirely toward criticism,
which also represents a genuine, though narrower, element in the
work's continuing life. But even if their theory was not much
inclined to focus on translation, their great translation work itself
was accompanied by a sense of the essence and dignity of this
mode. This feeling — everything points to this — need not be at its
strongest in the poet; it may in fact play the smallest role in him
qua poet. History certainly does not suggest that major translators
are poets and minor poets are mediocre translators, as is generally
believed. Many of the greatest, such as Luther, Voss, and Schlegel,
are incomparably more important as translators than as poets, and
others, such as Hölderlin and George, cannot be adequately
described solely as poets when the whole range of their work —
and especially their translations — is taken into account. Just as
translation is a distinctive mode, the translator's task may also be
conceived as distinctive and clearly differentiated from the poet's.

The translator's task consists in this : to find the intention
toward the language into which the work is to be translated, on the
basis of which an echo of the original can be awakened in it. Here
we encounter a characteristic of translation that decisively
distinguishes it from the poetic work, because the latter's intention
never is directed toward language as such, in its totality, but solely
and immediately toward certain linguistic structurings of content.
However, unlike a literary work, a translation does not find itself,
so to speak, in middle of the high forest of the language itself;
instead, from outside it, facing it, and without entering it, the
translation calls to the original within, at that one point where the
echo in its own language can produce a reverberation of the
foreign language's work. Its intention is not only directed toward
an object entirely different from that of the poetic work, namely
toward a language as a whole, starting out from a single work of
art, but is also different in itself : the poet's intention is
spontaneous, primary, concrete, whereas the translator's is
derivative, final, ideal. For the great motive of integrating the
plurality of languages into a single true language is here carrying
out its work. In this integration individual propositions, poetic
structures, and judgments never arrive at agreement (since they
remain dependent on translation); it is rather the languages
themselves that agree, complemented and reconciled with each
other in their mode of intention. If there is nevertheless a language
of truth, in which the ultimate secrets toward which all thinking
strives are stored up, at peace and even silent, then this language
of truth is — "the true language." And in fact this language, in the
anticipation and description of which lies the only perfection
philosophy can hope to achieve, is concealed intensively in
translations. There is no muse of philosophy, and there is also no
muse of translation. They are not, however, philistine, as
sentimental artistic folk would like to think. For there is a
philosophical genius, whose essential characteristic is the longing
for the language that is announced in translation. "Les langues
imparfaites en cela que plusieurs, manque la suprême: penser étant
écrire sans accessoires, ni chuchotement mais tacite encore
l'immortelle parole, la diversité, sur terre, des idiomes empêche
personne de proférer les mots qui, sinon se trouveraient, par une
frappe unique, elle-même matériellement la vérité." If what
Mallarmé conceives in these words is rigorously applied to the
philosopher, then translation, with its seeds of such a language,
stands half-way between poetry and doctrine. Translation's work
is less prominent than doctrine's, but it puts its mark on history no
less deeply.

If the translator's task is regarded in this light, then the
paths to its fulfillment threaten to become all the more
impenetrably dark. Indeed, this task — that of bringing the seeds of
pure speech to ripeness in translation — seems impossible to
accomplish, determinable in no realization. And isn't the ground
cut out from under any such realization if the reproduction of
meaning is no longer the criterion? Viewed negatively, that is
precisely the import of all the foregoing. Fidelity and freedom —
the freedom of rendering in accord with the meaning, and in its
service, fidelity in opposition to the word — these are the old,
traditional concepts in every discussion of translation. They no
longer seem useful for a theory that seeks in translation something
other than the reproduction of meaning. Indeed, used in the
conventional way, they are perpetually caught up in an irresolvable
conflict. What precisely can fidelity actually contribute to the
reproduction of meaning? Fidelity in translating the individual
word can almost never fully render the meaning it has in the
original. For this meaning is fully realized in accord with its poetic
significance for the original work not in the intended object, but
rather precisely in the way the intended object is bound up with the
mode of intention in a particular word. It is customary to express
this by saying that words carry emotional connotations. In reality,
with regard to syntax, word-for-word translation completely rejects
the reproduction of meaning and threatens to lead directly to
incomprehensibility. For the nineteenth century, Hölderlin's
translations of Sophocles represented a monstrous example of this
kind of literalness. Finally, it is self-evident that fidelity in
rendering form makes rendering meaning more difficult. Hence
the demand for literalness cannot be deduced from the interest in
maintaining meaning. The latter serves the undisciplined license of
bad translators far more than it serves poetry and language.

Therefore this demand, whose justice is obvious and whose ground
is deeply concealed, must necessarily be understood on the basis
of more pertinent relationships. Just as fragments of a vessel, in
order to be fitted together, must correspond to each other in the
tiniest details but need not resemble each other, so translation,
instead of making itself resemble the meaning of the original, must
lovingly, and in detail, fashion in its own language a counterpart
to the original's mode of intention, in order to make both of them
recognizable as fragments of a vessel, as fragments of a greater
language. For that very reason translation must in large measure
turn its attention away from trying to communicate something,
away from meaning; the original is essential to translation only
insofar as it has already relieved the translator and his work of the
burden and organization of what is communicated. En arche hen
ho logos, in the beginning was the word : this is also valid in the
realm of translation. On the other hand, the translation's language
can, indeed must free itself from bondage to meaning, in order to
allow its own mode of intentio to resound, not as the intentio to
reproduce, but rather as harmony, as a complement to its language
in which language communicates itself. Hence reading a
translation as if it were an original work in the translation's own
language is not the highest form of praise, especially in the age
when the translation is produced. On the contrary, the meaning of
the fidelity ensured by literal translation is that the great longing
for the completion of language is expressed by the work. True
translation is transparent, it does not obscure the original, does not
stand in its light, but rather allows pure language, as if
strengthened by its own medium, to shine even more fully on the
original. This is made possible above all by conveying the syntax
word-for-word, and this demonstrates that the word, not the
sentence, is the original element of translation. For the sentence is
the wall in front of the language of the original, and word-for-word
rendering is the arcade.

While fidelity and freedom in translation have long been
seen as conflicting tendencies, it also seems that this deeper
interpretation of one of them does not reconcile the two, but on the
contrary denies the other any justification. For what can the point
of freedom be, if not the reproduction of meaning, which is no
longer to be regarded as normative? Only if it can be posited that
the meaning of a linguistic construction is identical with the
meaning of its communication, does something ultimate and
decisive remain beyond any message, very near it and yet infinitely
distant, hidden under it or clearer, broken by it or more powerful.
Beyond the communicable, there remains in all language and its
constructions something incommunicable which is, depending on
the context in which it is encountered, either symbolizing or
symbolized; symbolized however in the development of the
languages themselves. And what seeks to be represented and even
produced in the development of languages is that kernel of pure
language itself. But if this hidden and fragmentary kernel is
nevertheless present in life as something symbolized, it inhabits
linguistic constructions only as something symbolizing. While this
ultimate being, which is therefore pure speech itself, is in
languages bound up only with the linguistic and its
transformations, in linguistic constructions it is burdened with
heavy and alien meaning. Translation alone possesses the mighty
capacity to unbind it from meaning, to turn the symbolizing
element into the symbolized itself, to recuperate the pure language
growing in linguistic development. In this pure language — which
no longer signifies or expresses anything but rather, as the
expressionless and creative word that is the intended object of
every language — all communication, all meaning, and all
intention arrive at a level where they are destined to be
extinguished. And it is in fact on the basis of them that freedom in
translation acquires a new and higher justification. Freedom does
not gain its standing from the communication's meaning; it is
precisely truth's task to emancipate freedom from meaning.
Rather, freedom demonstrates in the translation's own language
what it can contribute to the service of pure language. To set free
in his own language the pure language spellbound in the foreign
language, to liberate the language imprisoned in the work by
rewriting it, is the translator's task. To this end he breaks through
the rotten barriers of his own language : Luther, Voss, Hölderlin,
George have all extended the frontiers of the German language. —

What now remains for the significance of meaning in the
relationship between translation and original can be easily summed
up in a comparison. Just as a tangent touches a circle fleetingly and
at only a single point, and just as this contact, not the point,
prescribes the law in accord with which the tangent pursues its
path into the infinite, in the same way a translation touches the
original fleetingly and only at the infinitely small point of
meaning, in order to follow its own path in accord with the law of
fidelity in the freedom of linguistic development. Without naming
or grounding it, Rudolf Pannwitz has characterized the true
significance of this freedom in certain passages of his book Die
Krisis der europäischen Kultur which, next to Goethe's remarks in
the notes to his Westöstlicher Divan, must be by far the best thing
published in Germany on the theory of translation. He writes : "our
translations even the best start out from a false principle they want
to germanize Indie Greek English instead of indicizing, graecizing,
anglicizing German, they are far more awed by their own linguistic
habits than by the spirit of the foreign work [...] the fundamental
error of the translator is that he holds fast to the state in which his
own language happens to be rather than allowing it to be put
powerfully in movement by the foreign language, he must in
particular when he is translating out of a language very distant
from his own penetrate back to the ultimate elements of the
language at that very point where image tone meld into one he
must broaden and deepen his own language through the foreign
one we have no notion how far this is possible to what degree each
language can transform itself one language differentiates itself
from another almost as one dialect from another but this happens
not when they are considered all too lightly but only when they are
considered with sufficient gravity."

To what extent a translation can correspond to the essence
of this mode is determined objectively by the translatability of the
original. The less value and dignity its language has, the more it is
communication of meaning, the less is to be gained from it for
translation, up to the point where the overpowering weight ofthat
meaning, far from being a lever for producing a translation fully in
accord with its mode, makes the latter impossible. The higher the
work's constitution, the more it remains translatable, in the very
fleetingness of its contact with its meaning. This is of course true
only of original works. Translations, on the contrary, prove to be
untranslatable not because meaning weighs on them heavily, but
rather because it attaches to them all too fleetingly. For this as for
every other essential aspect, Hölderlin's translations represent a
confirmation, particularly his translations of the two Sophoclean
tragedies. In them the harmony of languages is so deep that
meaning is touched by language only in the way an Aeolian harp
is touched by the wind. Hölderlin's translations are prototypes of
their mode; they are related to even the most fully realized
translations of their texts as a prototype is related to a model, as a
comparison of Hölderlin's and Borchardt's translations of Pindar's
third Pythian ode shows. For that very reason they, more than all
others, are inhabited by the monstrous and original danger of all
translation : that the portals of a language broadened and made
malleable in this way may close and lock up the translator in
silence. The Sophocles translations were Hölderlin's last work. In
them meaning plunges from abyss to abyss until it threatens to
become lost in the bottomless depths of language. But there is a
stopping point. It is, however, accorded only to holy scripture, in
which meaning has ceased to be the watershed dividing the flow
of language from the flow of revelation. Where the text belongs
immediately to truth or doctrine, without the mediation of
meaning, in its literalness of true language, it is unconditionally
translatable. No longer for its own sake, but solely for that of the
languages. With regard to this text such boundless trust is required
of translation, that just as language and revelation must be united
in the text, literalness and freedom must be united in the form of
an interlinear translation. For to some degree all great writings, but
above all holy scripture, contain their virtual translation between
the lines. The interlinear version of the holy scriptures is the
prototype or ideal of all translation.