The Metonymics of Translating Marginalized Texts

Maria Tymoczko

Comparative Literature, Vol. 47, No. 1, On Translation. (Winter, 1995), pp. 11-24.

Stable URL:
http://links.jstor.org/sici?sici=0010-4124%28199524%2947%3A1%3C11%3ATMOTMT%3E2.0.CO%3B2-9

Comparative Literature is currently published by University of Oregon.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/uoregon.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic
journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,
and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take
advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org
Thu Sep 27 08:36:43 2007
MARIA TYMOCZIKI

The Metonyrnics of
Translating
Marginalized Texts

I T IS A CURIOUS FACT of coiltemporary literary studies that

very different branches of literary theory have coilverged on
the same insight: every telling is a retelling. In studies of folklore
and oral epic it is agreed that the content, form, and performance
patterns of ally given song or tale all belong to established tradi-
tions that the teller or singer inherits and in turn passes o n to suc-
ceeding tellers or singers. Albert Lord, follo~vingMilman Parry,
the framer of the theory of oral composition of epics such as the
Iliad, NeowulJ; a n d I,a Chanuon de Roland, has summarized suc-
cinctly, "the picture that emerges is not really one of coilflict be-
tween preserver of tradition and creative artist; it is rather one of
the preservation of tradition by the constant re-creation of it. The
ideal is a true story well and truly retold" (Lord 29, cf. 99 ff.). Every
creation is a re-creation.
Folk tellers themselves acknowledge their own indebtedness to
the tellers who have gone before them, as n7esee in the stories of
some of the most famous twentieth-century Irish storytellers. Peig
Sayers ends one of her tales, "That's my story, and if there's a lie in
it, let there be. 'Tis long ago I heard it from my father. He had the
world of stories" (Sekn O Suilleabhkin, Folktales 204). Similarly, af-
ter telling a version of the Deirdre story, ~ a m o n na Biwc con-
cludes, "That's the way I heard that story being told by my on7n
father, William Burke of Aird Mh6rn (0 Suilleabhkin, Folklore 29).
And again from ~ a m o n na Burc, "That's a true story, the way I
heard it! If 'tis a lie, it wasn't I made it up" (Folklore 119).
Deconstruction, as well as its critical progenitors, has also been
at pains to point o u t that writers d o not simply create original
texts: to a great extent ally literary text is dependent on literary
CObIPAR4TIYE LITERATURE

texts that have gone before and, moreover, literature is as much

about literature as about life. There are not only text and context,
but a fabric of intertextuality that links texts to other literary
works, both textual predecessors and contemporaries. A literary
n7ork like a translation depends o n previous texts: neither is an
"original semantic unity," both are "derivative a n d heteroge-
neous."' Every writing is a rewriting.
Polysystems theorists, notably Andre Lefevere, have stressed
that translation is a form of rewriting. Though translations are
"probably the most radical form of rewriting in a literature, or a
culture" (Lefevere, "M'hy Waste our Time" 241), they are to be
grouped with other modes of processing primary texts, including
film versions, children's versions, criticism, literary histories, an-
thologies, and the like, all of which shape the evolution of litera-
ture a n d culture.' Texts do not exist simply in their primary form;
rather texts are "surrounded by a great number o f . . . refracted
texts" (Lefevere, "Literary Theory" 13; cf. "I\llotl~erCourage's Cu-
cumbers" 4-8, 16-19). Processed for various audiences or adapted
to a particular poetics or ideology, refracted texts are responsible
in large measure for defining, maintaining, a n d redefining a
canon. Translation is one form of refraction, a form of writing that
is rewriting.
This discourse of retelling and rewriting is a particularly potent
framework for the discussion of the translation of a noncanonical
or marginalized literature. Since there are many types of non-ca-
nonical o r marginalized literatures, I should explain that here I
am talking about translating literature that is marginalized be-
cause it is the literature of a marginalized culture; the intent is to
consider texts that have b e e n excluded o r omitted from the
canon-or, more properly speaking, canons-of world literature
as defined by a Western perspective. There are often, in fact usu-
ally, massive obstacles facing translators who wis11 to bring the
texts of a marginalized culture to a dominant-culture audience:
issues related to the interpretation of material and social culture
(including law, economics, a n d so f o r t h ) , history, values, a n d
world view; serious problems with the transference of literary fea-
tures such as genre, form, performance conventions, and literary

' See Lawrence \'ennti 7-8, 68-69, 161, and sources cited
See Lefevere's discussions in "\Thy \Yaste our Time" 232-41; "Literary Theory"
12-20; 7-ran~lation,Ke~uriting,chs. 9-12. In "\Thy Waste our Time" (234), he notes
that translation is generally also accompanied by other sorts of rewriting, notably
"by an introduction, which is a form of criticism cum interpretation."
MARGINALIZED TEXTS

allusions; as well as the inevitable questions of linguistic interface.

For all these reasons the information load of translations of such
marginalized texts is often very high-in fact it is at risk of being
intolerably high. Because neither the content nor the intertextual
framework of such texts is familiar to the receiving audience, the
reception problems posed by marginalized texts in translation are
acute. Another way of putting this point is to say that while a
marginalized text is a retelling or re~vritingfor its original audi-
ence, it is neither for the target audience. T h e translator is in the
paradoxical position of "telling a nen7 story" to the receptor audi-
ence, even as the translator refracts and rewrites a source text-
and the more remote the source culture and literature, the more
radically nen7 the story will be for the receiving audience. Early
Irish literature is an example of such a marginalized literature,
a n d translators moving Irish literary works to another culture have
frequently been in the position of "telling a new story."
It has become a commonplace these days to say that literary lan-
guage is defamiliarized language-but it is also generally agreed
that if language becomes too strange or too defamiliarized, it can-
not be comprehended. T h e information load becomes too heal;
for comprel~ension,and, in the case of translations, the receiving
audience cannot understand the translated text."t is also the case
that human beings are not very good at hearing new stories: we
have the tendency to reinterpret them, to reshape them so that
they become versions of stories we already know, as Laura
Bohannan has epitomized wit11 her now classic article about tell-
ing the story of Hamlet in West Africa. Having chosen Hamlet to tell
because she felt the story was "universal," Bohannan discovered in
the course of narration numerous f ~ ~ n d a m e n t incompatibilities
al
between the tale and the expectations of her audience. As any au-
dience will in a traditional oral culture, her listeners soon inter-
vened, "corrected" her narration, and adapted the tale to their
own context with the result that they were satisfied to have heard a
good story, only to leave B o l ~ a n n a ndoubtful that it was "the same
story" after all.* In general, cognitive science suggests that n7etend
to assimilate new and unfamiliar information to patterns that are
already recognized and that have already become familiar, a n d
there is some evidence from studies of the brain that there is a
"sing a coder-decoder model, Eugene Nida (120-40) discusses the problems of
translating the heavy infornlation load of an unfamiliar text or text type for a re-
ceptor audience.
I See also Slaria Tyn~oczko,"Translation."
COMPARATIVE LITERATURE

biological basis to this tendency."

The power of this discourse about rewritings a n d retellings as a
framework for the discussioil of translations is illuminated by
some observations about the characteristics of the rewritings and
retelliilgs that are most familiar to literary scholars and traditional
audieilces alike: mythic retellings." Every telling is a retelling:
there are n o stories for which this is more true than myths, for
which there are n o "originals." Myths descend from the depths of
time: this is what it means to be a traditional story. And even if we
suppose that there was a single moment of creation for any spe-
cific myth (rvhicl~most theorists of oral tradition would agree is
the case), behind that moment of creation lies a vast body of ar-
chetypal patterns which the myth reworks and reanimates.'
John Foley (192 ff.) has argued that when a traditional oral tale
is told, the telling is metonymic. For a traditional audience each
telling evokes metonymically all previous telliilgs of the tale that
the audience has participated in and, f ~ ~ r t l ~the
e r ,telling instanti-
ates and reifies metonymically the entire tradition that the audi-
ence a n d teller share. T h e traditional audience hearing Avdo
Mededovik sing a song about Marko Kraljevik in Y~~goslavia had
evoked for them all other occasioils of Avdo singing the same
song, but also all other singers' versions of the same epic, and be-
yond that all of I\/larko's various adventures, a n d the Serbo-
Croatian epic tradition as a whole with all its many epic tales, as
well as the relation of that epic tradition to the culture and the
history of the community.': At the same time, the form of the epic
was also metoilymic of the formulas, the meter, the genre, and the
methods of oral composition in Serbo-Croatian tradition. I n
France an oral rendering of "La Belle e t la Beten not only called u p
all previous renditions of this tale, but all versions of tale type AT

See, for example, Partha Mitter's discussion in 8-14, as well as sources cited.
The tendency to assinlilate the unfamiliar to the familiar is a principal factor be-
hind the phenomenon of the simplification of conlplex and innovatory literary
models discussed by Itanlar Even-Zohar (21-22). Edward Said 0bser.i-es, "It is per-
fectly natural for the human mind to resist the assault on it of untreated strange-
ness; therefore cultures hare always been inclined to impose complete transforma-
tions on other cultures" ( 6 7 ) ; grids and codes are imposed "upon raw reality,
changing it from free-floating objects into units of knowledge" (67).
" I am using 1nyt1~
i n the broadest sense of "a traditional tale."
See, for example, Joseph Campbell's study of the archetypal patterns behind
hero tales, as well as Eric11 Neumann's treatment of archetypes related to narra-
tires about female figures.
' On these issues see Lord chapter T, and pnss~rn.
MARGINALIZED TEXTS

425C (or even ,4T 425, the entire animal groom cycle), as well as
the traditional lore of France." hloreover, its form was metonymic
of the various narrative conventions of wondertales in France,
from the opening and closing signals of the genre to the medieval
ambience of the settings. The Irish audience hearing the story of
Fill11 trapped in the bruiden (hostel, large banqueting-hall, house,
fairy-palace) in which Conan adheres by his posterior to a bench
has evoked metonymically all previous versions of the same tale, as
well as the genre of bvuiden tales as a whole, the entire corpus of
Fenian lore, and Irish traditional literature in general. At the same
time other aspects of the oral tradition in Ireland-such as the
narrative form (including the "runs" of Irish wondertale tradi-
tion), the relatioilship of the Fenian ballad tradition to the narra-
tive tradition, a n d the hierarchical prestige of various sorts of tales
with Fenian tales at the summit-are also evoked."'
Foley's a r g u m e n t can be extended to t h e case of written
retellings of myth, the most familiar examples of which in M'estern
tradition are reworkings of classical and biblical myths, though
rewritings of the Artl~urianlegend and the stories of Don Juan or
Faust could also be used to illustrate the same principles. T h e
rewritings of classical myths have been a staple of Western litera-
ture, from the Old French Enens and the Middle Englisl~Sir O~feo,
through Shakespeare's Troilus a n d Racine's Phe'dre, to Joyce's
Ul~sses,Anouil11's Antigone, a n d Camus's ~VIjthede Sisyphe. Ally
single version of these myths calls u p in a reader all other versioils
of the same story: Joyce's 171~ssesevokes in the reader not only the
O d ~ s s eb~u t Dante's Ulysses a n d Tennyson's Ulysses a n d even
Charles Lamb's version for children entitled The Adventures of'
LTlysses. Indeed 171~ssesis a perfect example of the metonymic as-
pect of literary rewritings of myths, for in order to understand it,
the reader must already knon7 other versions of the same myth or
come away with a very strange conception of Odysseus indeed and
have absolutely n o clue a b o u t the classical architectonics of
Joyce's work. Such rewritings stand metonymically also for the tra-
dition of Western written literature all the way back to the Greeks,
as well as the earlier oral Indo-European heritage. It is a paradox
that one must already know a myth in order to recognize a mythic
tale and to a p p r e l ~ e n dthe import of ally particular version of the
myth, yet the m y t l ~itself does not exist apart from specific ver-
"runo Bettelheim (277-310) discusses the animal groom cycle.
'"ames Delargv discusses these and other aspects of Irish oral tradition. A ver-
sion of the tale is included in Jeremiah Curtin 148-56.
COMPARr\TIVE LITERATURE

sions.
T h e metonymic aspect of literary rewritings and retellings is an
important aspect in cultural continuity. It permits the adaptation
of traditional content and form to new circumstances, allowing
change while still maintaining the predominant sense of preserva-
tion of larger elements of tradition. It enables traditional audi-
ences to correct (and forgive) the mistakes o r omissions of tradi-
tional tellers, to take enjoyment in tales told in an abbreviated o r
cryptic manner, to fill gaps in narrative textures, to understand
literary allusion. It enables young tellers to learn from others, and
correct or improve upon their teachers' versions, and go on to be-
come even greater masters themselves. Performance theorists are
coming to understand how allusions-including references to the
sites of traditional tales-are metonymic tellings of the tales them-
selves (Richard Baumann and Charles Briggs 7.5; cf. Earl Miner 94,
151-54).
Literary artists use the metonymic aspect of mythic retelling in
powerful-though different-ways as 1i7ell. Authors commonly use
a "baseline" version of the myth as an implicit standard of com-
parison, against which the audience measures the author's own
vision. When the twelfth-century author of the Eneas foregrounds
Aeneas as lover, rather than as the heroic and dedicated (Dido
might even have said monomaniacal) founder of Rome, h e is
speaking to his contemporaries about the relative importance of
love and war in a man's life, using the Aen,eid as the implicit stan-
dard for his own 1i7ork. And for that message to make its f ~ ~impact
ll
o n the audience, they must have as a baseline the canonical ver-
sion of the myth. When Giraudoux wrote La Guerre de Troie n'aurn
pas lieu, he counted o n the fact that the audience knew that the
Trojan M'ar 7oould take place, that they were familiar with the Iliad,
if only through refractions. Though the metonymic aspect of
mythic rewritings is particularly clear because the content of the
texts represents larger wholes (whole families of texts), all litera-
ture works this n7ay and the metonymic aspects of texts are not re-
stricted to content. Aspects of poetics, specifically literary form,
are also metonymic. Thus, for example, any single English sonnet
evokes all the sonnets of Shakespeare and Petrarch, as well as the
entire tradition of sonnet writing. This is so because any writing is
a rewriting.
T h e special group of rewriters called translators grapples wit11
the metonymic aspects of literature all the time. In translations of
works from literary systems that are related to the receptor literary
MAKCINALIZED TEXTS

system-for example, literary systems that with the receiving lan-

guage system form a megasystem, such as French and English lit-
erature do-most of the metonymic aspects of the source text are
transparent to the target audience. Modern English-speaking au-
diences understand fairly ell the generic signals of nineteenth-
century Russian ilovels such as Anna Ka,renina; the plotting and
character types are familiar, even though certain aspects of the
culture such as elements of the law or the use of nicknames o r the
symbolic significance of samovars may not be. Thus, such works in
translation are able to be integrated into canons of world litera-
ture-or at least canons defined within the framework of domi-
nant cultures-wit11 relative ease. But what happens when the met-
onymic aspects of the story are opaque rather than transparent to
the receptor audience? How is a translator to translate a n7ork
whose characters, plot, genre, and literary allusions, just to name a
few parameters of the literary system, are unfamiliar and "unread-
able" by the intended receptor audience?
T h e way in which a literary text metonymically represents fea-
tures of its literary system and ultimately features of its whole cul-
ture is what makes translating a text of a marginalized culture so
difficult. A translator assumes a large responsibility in undertaking
to produce a text that ill become representative of the source
literature and, indeed, of the entire source culture for the recep-
tor audience. As Xorman Simms has shown, the politics of such
represeiltatioils are not to be underestimated, and it is the coiltrol
of the image of the source culture and the source culture literary
tradition that has made the translation of certain Asian texts so
controversial and "touchy" in the twentieth century. But beyond
the question of politics, it is in large measure the lack of familiarity
wit11 the metonymic aspects of the literary texts of marginalized
cultures that makes it difficult for the audieilces of domillant cul-
tures to integrate marginalized texts into their canons, irrespec-
tive of any linguistic barrier. M'llat happens when the audience
doesn't uilderstaild the metonomies-when the audience doesn't
understand the literary signals, the form, the genre, the story?
What happens, in short, when a translator has to tell a new story?
Leaving aside wider cultural questions for the moment, in a case
like this the amount of literary information to be conveyed to the
receiving audieilce is excessive; the translator must either make
some decisive choices about ~ v l ~ iaspects
cl~ to translate-that is, d o
a partial translation of the literary information in the text-or
seek a format that allows dense information transfer through a va-
COMPARATIVE LITER4TURE

riety of commentaries o n the translation. This is why initial trans-

lations of unfamiliar texts are so often either popular o r scholarly:
the former are usually severely limited in their transfer intent a n d
minimally representative of the metonymic aspects of the original,
while the latter allow a good deal of metatranslation to proceed,
presenting quantities of information through vehicles such as in-
troductions, footnotes, appendices, parallel texts, and so forth. In
a scholarly translation the text is e m b e d d e d in a shell of
paratextual devices that serve to explain the metonymies of the
source text, providing a set of contexts for the translation. In the
case of a popular translation, by contrast, the translator typically
focuses on a few aspects of the literary text which are brought to a
broad segment of the target audience.
Thus, o n the one hand, in his ground-breaking versions of early
Irish narratives published in 1878-80, Standish O'Grady sought to
convey the plots, the characters, and the general historicized tex-
ture of the Iris11 stories to his Anglo-Irish audience. O'Grady sacri-
ficed the genres, the character types, the linguistic texture, and
even the names found in his source texts; but his works were acces-
sible to any competent reader of English, and they became widely
known, popular, and influential, leading to a demand for more
translations and refractions about Ireland's gods and heroes like
Cti Chulainn. Later translations in turn evoked more adequately
the metonymic relationship of the texts to other features of the
Iris11 linguistic and literary systems for people who were already
familiar with the content of the tales. Ernst Windiscl~,on the other
hand, in his 190.5 German translation of Ta'inDo' CLiailnge was able
to present such features as genre and character type and even the
peculiar character of the Irish heroic tradition; but his translation
is a scl~olarlyone, accompanied by an edition of the text as well as
by enormous commentary and footnotes providing masses of con-
textual and intertextual information. Windiscll's n7ork n7as clearly
aimed at a very narrow scholarly population within the target audi-
ence.
Not all the information serving metonymic functions in a text
from a n unfamiliar literature o r culture can be realized in transla-
tion: the information load is too great, as I have already said, and
the information is coded in textual features that make inconsis-
tent a n d irreconcilable demands o n the translator." Scholarly
l 1 There is an analogue o n the linguistic level. Because of incongruities in
obligatory aspects of language, translation inevitably involves linguistic loss and
gain, and it is not possible to capture every linguistic feature of the source text,
bIARGINALIZED TEXTS

translations with their metatranslation devices are able to convey

more information to the reader, but all translators, including
scholarly ones, select specific aspects of the metonymic relation-
ship between text and literary system or text and culture to realize
and to privilege. They ask, implicitly or explicitly, what larger
wl~olesthe translation text will point to, stand for, be related to.
M'ill the translation text, for example, be metonymic of the lan-
guage of the source culture, of a generic convention in the source
culture, of the value structure of the source culture, or of other
cultural patterns conveyed by and t l ~ r o u g literature?
l~ A translator
n7ho foregrounds the translation as a ~vindo~v into a new language,
importing or transferring lexis, syntax, and the like into the recep-
tor language, is choosing to hare the translation be metonymic
primarily of the source language as a ~ v l ~ o (as
l e any text is in part)
and will do~vnplayother metonymic aspects-such as those of
genre-so as to make the information load manageable." Another
translator, who is interested in poetics, may privilege the generic
codes of the source text, preserving these metonymies, while on a
linguistic level playing fast and loose wit11 the text so as to adapt
the text thoroughly to the linguistic norms of the receptor lan-
guage; Pound's translations of Chinese texts might be described
this way. The choice of which metonymies to preserve has much to
do with the translator's purpose, and the translator who wishes to
challenge elements of the privileged center of the target system of
poetics, will probably privilege metonymies of genre or poetics
over those of content or language, while a translator who wishes to
challenge the value structure, say, of the receptor audience ill
make different cl~oices.'"

either in its paradigmatic or its svntagnlatic levels. See, for example,J.C. Catford
o n these points.
'"ire should note the importance of translations for the linguistic expansion of
any language, but particularlv of minority languages; such privileging of language
per se is, therefore, not to be dismissed as a useless, foolish, or tril-ial strategy.
Latin translations of Greek works served this function as Cicero notes explicitly
(see the discussion in Susan Bassnett and Leferere 23-24); cf. also the arguments
of Els Oksaar and of Norman Denison.
A similar translation strategy is often also chosen by philological translators who
use translations as a sort of extended linguistic commentarv.
' X e f e r e r e , l ' r a n ~ l a f i o 4n 1, argues that the image cast by a translation is a func-
tion of the ideology and the poetics of the receptor culture. Thi? is to sketch the
situation in broad strokes; the issue is rather more complex.
Jifi Lev? discusses the multiplicity of choices initially open to a translator, and
the ways in which any single value will dictate a path of translation with a concomi-
tant narrowing of options thereafter. Cf. Dinda Gorlee, ch. 4, who applies game
COMPARATIVE L I T E M T U R E

But this is not the e n d of the question of metonymy in the trans-

lation of marginalized texts. The human tendency to assimilate
the unkno~vnto the closest known pattern must also be reckoned
with. Even as metonymic aspects of the source text are stripped
away in the translation process, the translation of the marginalized
text gets assimilated to existing metonymies in the receptor sys-
tem. T h e translator consciously o r unconsciously picks other me-
tonymies to evoke than those of the marginalized text, specifically
the metonymies of the receptor-language literary system and lan-
guage. This is what happens when a translator looks for a "dynami-
cally equivalent" meter, for example. O r the plot of the source text
may be altered so as to facilitate assimilation between the plot of
the translation a n d plots in the target culture; one early transla-
tion of The D r ~ a mof the Red Chambe1; for example, assimilated it to a
love story. O r again there may be a generic shift such as the shift
that occurs when the branches of the M'elsh lWabinogi are assimi-
lated to the genre of romance, or when the earliest Iris11 version of
the Deirdre tale is interpreted as a love story (cf. Tymoczko, "Ani-
mal Imagery"). O r the vocabulary of the translated text may evoke
passages in the literature of the target language and set u p new
intertextual resonances. Thus the marginalized text gets assimi-
lated to existing structures in the receptor literary system; it is es-
sentially presented as a "rewriting" of elements of the receptor lit-
erary system, even as it brings with it some aspects that challenge
the receiving system and that remain eccentric.
In a rationale for a descriptive study of translation, Gideon
Toury has suggested that the investigator needs to look for the
norms governing the translations being described. I n the
translator's in,itial norms Toury makes the distinction between ad-
equate translations, which tend "to adhere to the norms of the
original work," the source text, and acceptable translations, which
adhere "to the linguistic a n d literary norms of the target system";
Toury views these strategies as polar opposites, though h e ac-
knowledges that in practice there ill generally be some combina-
tion of o r compromise between the two extremes (55). Once this
initial norm is selected, Toury posits that the translator picks opera-
theory to translation; the strategy or heuristic ~vithwhich the game is approached
(as well as Lei-!'s idea of an initial choice which determines others thereafter) can
be coiupared to the choice of metonymies I ha\-e discussed here. O n e aspect that
Lev? and Gor1i.e underestiiuate, however, is the extent to ~vhichchoices made.
heuristics adopted, or metonymies p r i d e g e d are shaped by the ideology and po-
etics of the context of the translation. Constraints in the interpretive process are
notjust internal, linguistic, or expedient, but also public, cultural, and political.
MXKGIXALIZED TEXTS

tional norms wrhich govern the specific choices of his work. While
this all sounds good in the abstract and in fact often offers useful
tools for thinking about translation as a process, when we turn to
actual translations, particularly translations of texts from
marginalized cultures, the situation turns o u t to be much more
messy: a translation may be radically oriented to the source text in
some respects (hence in Toury's terms the translation is adequate),
but depart radically from the source text in other respects so as to
assimilate it to a norm of the receiving culture (hence the transla-
tion is acceptablein Toury's scheme).
T h e same sort of ambiguity is found in the translations of
marginalized texts when we try to apply the other sorts of polari-
ties that are commonly put forward to categorize translations: po-
larities such as literal and free, o r djnamic equivalence and formal
equi~alence.'~ Again, with respect to these typologies, translations
seem to have self-contradictory elements in their specific transla-
tion strategies: a text that is "formally equivalent" in language will,
for example, by virtue of the translation process lose the humor of
the text and fail to convey the formal qualities of the humorous
genre being translated, as occurs in the Jones and Jones transla-
tion of the Welsh tale Culhwck and Olwen. Cecile O'Rahilly's trans-
lations of the Irish Tuin Bd Cziailnge are painfully literal in their
transfer of the syntax of the early Irish texts, but she chooses eu-
phemisms for the lexis (these tone down the humor of the text,
hence shifting the genre, and elide the content, including sexual-
ity and the grotesque) and she homogenizes the register as well.
Mrhat are tve to d o with this seeming breakdown of theory?
In translation studies o n e sometimes despairs of being a
Mrittgensteinian fly trapped in a fly bottle, waiting for the right
philosopher to point the way out. In darker moments o n e suspects
that the vessel is in fact a Klein bottle. T h e ways in which me-
tonymy operates in retellings and refractions can be extended to
translation and help to explain the seeming inconsistencies that
appear in translation strategies. In an attempt to avoid informa-
tion overload while at the same time honoring the fact that a
marginalized text does represent its culture and literary tradition,
the translator makes significant compromises, orienting the target

I" O n the latter see Nida 139-78.

T h e r e a r e also o t h e r polarities used to describe translation types a n d translation
strategies. For example, Richard Jacquemond (131-35) discusses a dialectic i n
translations between what h e calls naturalization a n d exoticization, orientations
n o t unrelated to t h e points m a d e here.
C041PXKATIT'E LITERATURE

text in two directions. It is the selection of metonymies to preserve

and to relinquish, to assinlilate and to resist, that principally char-
acterizes the initial translation norms of marginalized texts, more
than the standard polarities that are usually discussed in transla-
tion theory. These decisions about the metonyrnics of the target
text in turn determine the operational norms. To use the termi-
nology of James Hollnes (73-81), the question of nletonynly struc-
tures the map determining the specific transfer rules between
source text and target text. The map is typically highly personal-
ized rather than a simplistic either/'or, thus often resulting in a
translation that does not neatly fit any of the convenient dichoto-
mies that figure so often in discussions of translation theory.'; :In
understanding of the nletonnlyics of translated texts makes it pos-
sible to grade more finely the sorts of larger and relatively inoper-
able classifications of translation strategies that are generally pro-
posed in the literature of Translation Studies. At the same time,
awareness of the metonymies of translation are a key to the con-
struction of representations that translations project-whether
representations of history, culture, values, or literary form.

University of il/la.ssachusett.s,Amlzerst

Works Cited

Bassnett, Susan a n d Andre Lefel-ere, eds. Trn?zslntion, Histo? a?zd Cnlture. Lon-
don: Pinter, 1990.
Bauiuan, Richard and Charles L. Briggs. "Poetics and Perforiuance as Critical Per-
specti\-es o n Language a n d Social Life." A?znual R~uiewoJ' Anthropologj 19
(1990), 59-88.
Bettelheiiu, Bruno. Tize C s ~ oJ'E~1rha~ztnzent:
s The ~Menningand Inzportnnce qfFair?
Tales. 1976. New 170rk:Vintage-Randoiu, 1977.
Bohannan, Laura. "Shakespeare in the Bush." ,\ratural Histo~y75.7 (1966): 28-33.
Campbell, Joseph. The Hero with n Thousa?zd Fares. 1949. Bollingen Series 16.
Princeton: Princeton Lrniversity Press, 1972.
Catford, J. C . .4 Linguistic Theorj of Tmnslation: An Essnj in rIpplied Linguistics.
London: Oxford Cni\-ersity Press, 1965.
Curtin, Jeremiah. myths nndFolli T a l ~ oflreland.
s 1890. New170rk: Dover, 1975.
Delargy, James H. "The Gaelic Story-Teller." proceeding^ of t h British
~ Acndemj 31
(194<5):3-46.

" Clearly the balance between source a n d receptor metonymies also changes
o\.er time in the translation histor! of n text; cf. Bassnett a n d Lefevere 3.
PyIARGINXLIZED TEXTS

l i s mTranslation." T h ~ o ~and
Denison, Norman. "On P l ~ ~ r i l i n g ~ ~ a and y Prnctic~of
Translation. Ed. Lillebill Grahs, Gustal- Korltn, a n d Bertil Rlalmberg. Bern: Pe-
ter Lang, 1978. 313-29.
El-en-Zohar, Itarnar. Poly,~ystemStudies. Spec. issue of Poetics Torlaj 11.1 (1990): 1-
268.
Foley, John Xliles. "Reading the Oral Traditional Text: -4esthetics of Creation and
Response." Comparative Research on Oral Traditions: A LbIe~norial ]or milm man
Par?. Ed. John Rliles Foley. Coluiubus: Slavica, 1987. 185-212.
Gorlte, Dinda. Semiotics nlzd the Problenz of Trnnslntio?z. .4lblasserdarn: Kenters,
B.V., 1993.
Holiues. James S. "Describing Literary Translations: Xlodels and Methods." Litern-
i n Literaty Studies. Ed. Jaiues S. Holmes,
ture and Tra?zslatio?z:A'ew Per,~pecti?~~s
Josk Laiubert, and Kaymond van d e n Broeck. Lou\-ain:.4cco, 1978.69-82.
Jacquemond. Richard. "Translation a n d the Cultural Hegemony: .A Case a n d
French/.4rabic Translation." Venuti 139-38.
Jones, Gwyn and Thomas Jones. Trans. The Mnbi?zog.io?z. 1949. New York: Dutton.
1963.
Lefevere, Andri.. "Literary Theory a n d Translated Literature." The Art and Science
of Tm?zslation.Ed. Andri. Lefevere a n d Kenneth Da\-idJackson. Special issue of
Dispositio i ( l 9 8 2 ) : 3-22.
---. "Mother Courage's Cucumbers: Text, System and Refraction in a Theory
of Literature." ,Vfodrrn Language Studirs 12 (1982): 3-20.
--- T ~ u n ~ l a t ~Rezu~ztz?zg,
on, and t h Mn?zzpulat~on
~ of Lzteraq Famp London
Routledye, 1992
---. "I'Vhy I\'aste O u r Time o n Rewrites? T h e Trouble with Interpretation and
the Role of Rewriting in an Alternative Paradigm." The Mn?zipulation of Litern-
ture: Sturlies i n Litera7y Translation. Ed. T h e o Heriuans. New York: St. Xlartin's
Press, 1985. 21\5-43,
Le\-9,Jii-i. "Translation as a Decision Process." T o Honor R o m n n Jakobso?~:Essnjs
on the Occasion of his S e v ~ n t i e t hBirthdaj, 11 O c t o b ~ r1966. 2 vols. T h e Hague:
Xlouton, 1967. 2.1071-82.
Lord, -4lbert B. The Singer o f T n l ~ s 1. 960. Cambridge: Harvard University Press,
1964.
Miner, Earl. C o m p n r a t i ~ ~Poetics:
e A n Interculturn1 Essaj o n Theories of L i t ~ r a t u r ~ .
Princeton: Princeton C n i x r s i t y Press, 1990.
Xlitter, Partha. "Can We El-er Understand Alien Cultures?" Comparative Criticism9
(1987): 3-34.
Neumann, Erich. ?'he G r ~ a LbIother:
t i l n Analjsis o] the Archetype. Trans. Ralph
Manheiiu. 1955. Bollingen Series 47. Princeton: Princeton University Press,
1972.
Nida, Eugene A. Tozunrd n S c i ~ n c eof Tmnslating: IVith Special Refer~nceto P r i m
riplps and Procedur~sI n v o l v ~ di n Bible Translating. Leiden: E. J. Brill, 1964.
O'Grady, Standish. Histo7y oflrelnnd: The Heroic Period. 2 vols. London: Saiupson
Low, 1878-80.
O'Rahilly, Cecile, ed. a n d trans. " T h i n Bd Czinllzge" from the Book of L ~ i n s t e r .
Dublin: Dublin Institute for h d n n c e d Studies, 1967.
---, ed. a n d trans. Thin Bd C,tiailnge: Recension I. Dublin: Dublin Institute for
.id\-anced Studies, 1976.
Oksaar, Els. "Interference, Bilingualism a n d Interactional Competence." Theor)
nnrl Prnctic~o f Translation. Ed. Lillebill Griihs, Gustal- Korlkn, a n d Bertil
hfaliuberg. Bern: Peter Lang, 1978. 295-50.5.
0 Si~illeabhiin,S e i n [Sean O'Sulli\-anl, ti-;rns. T h ~ l ' b l h l o roflrelnnd.
~ London: B.
T . Batsford, 1974.
---, ed. and trans. Folktal(?s of'lr~ln~rzd. Chicago: Lrniversitr; of Chicago Press,
1966.
Said, Edward T.V. Orientctlism. 1978. New 170rk:Vintage-Randoiu, 1979
Siiums, N o r m a n . "Three Types of 'Touchy' Translation." Ninzrorl's Sin. Ed.
Norrnan Siiums. Pnclfic Q ~ ~ w t ~Llfonnn
r l j 8.2 (1983): 48-58.
Toury. Gideon. I n Search of a Theo? of Trun~lation.Tel Xviv: Portcr Institute for
Poetics and Semiotics, 1980.
Tymoczko, Maria. ".inimal Iiuagery in I,oing~sA,Inc nCislenn." Studin Cr4ticn 20-21
(1985-86): 145-66.
---. "Translation in Oral Tradition as a Touchstone for Translation Theory
and Practice." Bassnett and L e f e x r e 46-55.
Yenuti, Lawrence, ed. Rethinking T r a n s l a t i o ~ ~L)iscourse,
: Subjecti-i~itj,Irleology.
London: Routledge, 1992.
g ~ Bd C1inlng.e nnclz dem
Windisch, Ernst, e d . and trans. L)ie altirische H e l c l e ~ ~ r nThin
Buch von Leinstel; Text und 1J~bercetzungmit ~ i u e Einleitzrng.
r Leipzig: S. IIirzcl,
190.5.