Strange Loops of Translation[1]

Over the course of the first decade of the new millennium, Douglas Hofstadter worked on two topics that had long been close to their heart: strange loops and translation. They had first come up with the strange-loops model in the late 1970s, for Gödel, Escher, Bach (1979/1989), where they had also recurred several times to the problematics of interlingual (372-73, 379-80) and intersemiotic translation (83), but also, and above all, to translations between and among mathematical systems in Kurt Gödel’s engagement with Principia Mathematica (209-13, 215, 267-72, 417, 441-46, 533).[2] In Le Ton Beau de Marot (1997), a 600-page book entirely about translation, they had dealt in passing with many of the cognitive quirks that they had previously linked to strange loops - especially analogy and indexicality/self-reference - but without theorizing translation as a strange loop, or the strange loops of translation.

Our purpose in this loopy article is to remedy that oversight.

 

In a preface to the twentieth-anniversary edition of GEB published in 1999, then, the DHP complained about readers missing the book’s strangely loopy point - and that was the beginning of the decade-long almost-convergence on which we focus here. In response to that 1999 preface, in the spring of 2003 Ken Williford and Uriah Kriegel asked them to write a continuation of their strange-loops theory for a philosophical essay collection they were editing, titled Self-Representational Approaches to Consciousness ; when that book came out in the spring of 2006, it contained a double-length article by the DHP, who felt that it still wasn’t long enough; their further work on the topic resulted in I Am a Strange Loop (2007).

While they were working on the article for Williford and Kriegel (2006), they decided to translate Françoise Sagan’s 1966 novel La Chamade into English - a labor of love, not a commission from a publisher. Some time between finishing that translation in early 2005 and publishing it in 2009, they wrote their second extended treatise on translation, titled “Translator, Trader” - only 100 pages long, considerably shorter than Le Ton Beau de Marot. They published “Translator, Trader” in an omnibus volume with the Sagan translation; their main focus in it is how they translated Sagan, with extended discussions (culminating in defenses) of some of their choices (the proliferation of Americanisms, most notably, by their own count at a rate of around 20 per page).

What strikes us, however, and inspires this article, is that, despite the intensification of their thinking about both strange loops and translation throughout that first decade of the new millennium, they persisted in what we take to be the obvious oversight in Le Ton Beau de Marot: they never raise­d the possibility of the strange loops of translation. In I Am a Strange Loop (2007: 101-2) they had defined a strange loop as “a paradoxical level-crossing feedback loop”; the subtitle of “Translator, Trader” (2009) is “Essays on the Pleasantly Pervasive Paradoxes of Translation.” That suggests a potential theoretical convergence - but one, unfortunately, that is never even broached, let alone explored. In the big strange-loops book (Hofstadter 2007) they referred in passing to translating in a broad sense, especially between Gödel’s theorem and English (145) and as an analogue of the replicability of patterns (224, 257) - a theme that had helped organize the cognitive science of their big translation book (Hofstadter 1997) - but never as strangely loopy. And just as it had not occurred to the DHP in 1997 to explore the self-constitutive effects of the analogies and indexicalities of translation, so too did it never occur to them in the little translation book (2009) to track the ways in which their paradoxes of translation twist around into strange loops.

Those twists are the loopy (in-and-out-of-)focus of this strange little essay.

 

“Paradoxical Level-crossing Feedback Loop”

Logically, as the DHP defined it, a strange loop is not just any old paradox but specifically a cyclical paradox:

What I mean by “strange loop” is … not a physical circuit but an abstract loop in which, in the series of stages that constitute the cycling-around, there is a shift from one level of abstraction (or structure) to another, which feels like an upward movement in a hierarchy, and yet somehow the successive “upward” shifts turn out to give rise to a closed circle. That is, despite one’s sense of departing ever further from one’s origin, one winds up, to one’s shock, exactly where one had started out. In short, a strange loop is a paradoxical level-crossing feedback loop.[3] (2007: 101-2)

The logic, however, is still not the main point - not what the DHP complained that readers of GEB had mostly missed. The main point that they returned to underscore in I Am a Strange Loop was that these strange loops in human consciousness are constitutive of consciousness, and thus of the “I” - that ultimate symbol(izer/-ized) of consciousness.

The first sign of strange loopiness, they suggested, is the presence of indexicals, words pointing to the work being done by the speaker (or thinker) to point to their own current symbols, such as “this” or “here,” but also “I” (160); but these words are only indications that symbols are registering symbols, and especially that a higher-level symbol, called “I,” is thinking not only about symbols, but about its own thinking about those symbols. What makes this process strangely loopy is the pandemonial proliferation of analogies (148-52) or isomorphisms (243) between lower-level symbolization and higher-level symbolization: whatever an “I” says about the symbols it is/they are manipulating can easily be applied to the “I” saying those things, so that any attempt to idealize the “I” as a god-given glory that is inserted into the human organism at birth, or at conception, and that may even survive the death of the organism into which it was so inserted, is everywhere undermined by its own continuous self-creation.

The second important thing to note there is that the analogical parallels, the isomorphisms, that give the game away are only logical in secondary idealizations. First and foremost they are audience-effects. They are situated interactive performativities. And the implication of that realization is that the “I” itself is also a performativity, an audience-effect. It is a perceptual multiplicity (the pandemonial “they”) presented and ultimately experienced as a unity (the “I”). It/they come/s to feel like a “real person” inside the human organism only gradually, as it/they internalize/s how it is/they are perceived by others. The “I,” in other words, is a rhetorical construct. Aristotle would have called it ēthos - “character,” but also authority, and identity - and argued that it is a joint production of the speaker and their audience.[4]

 

“Pleasantly Pervasive Paradoxes”

The subtitle of “Translator, Trader” (2009), you’ll recall, is “Essays on the Pleasantly Pervasive Paradoxes of Translation.” the DHP’s first paradox is what they called “The Wrong-Tongue Paradox” (9), which they glossed with the rhetorical question “How can Dante Alighieri have written a book in English, a language that didn’t even exist when he was alive, or William Shakespeare a book of sonnets in Russian, when he knew not a word of that language?” Yes, good questions. But now take it up a notch. What happens when Dante doesn’t just write a book in English, but tells us in English that they are writing their book in Italian (rather than Latin)? Enter the strange loop. (The DHP came closer to hinting at that possibility in 1997: 445-47 - but there too hovered just shy of making the connection with strange loops.)

A small difference, you say? Maybe. But the paradox the DHP invoked in 2009 is a curiosity. It makes the reader scratch their head and say “huh!” It is odd, but easy to ignore. Most readers of translations never notice it. The DRP’s souped-up strangely loopy paradox is in the reader’s face. But even that is not the main point to be made about it. Even if no one notices the strange loops of the Dante Alighieri Pandemonium saying in English that they’re speaking Italian, they do cognitive work that “The Wrong-Tongue Paradox” doesn’t do. Translational self-reference is cognitively productive. Like all strange loops in the DHP’s account, it is constitutive of the self. And because it is collectively self-constitutive, because it constructs the self as a shared audience-effect , it is constitutive of culture.

That culture-constitutive effect is especially clear in another paradox mentioned by the Hofstadter Pandemonium. They called it “The Wrong-Place Paradox”:

When characters in a story speak, the language they use is informal and colloquial, so their lines are inevitably peppered with colorful words and idioms that come from the streets of their land and the history of their people. And any other people has a different history, different traditions, and scads of different colorful ways of expressing itself. Therefore, when idioms of culture B are placed in the mouths of individuals from culture A, the result is incoherent. (2009: 9-10)

This is another curiosity, of course, one that is well-known to translators and translation scholars, but apparently unknown to ordinary target readers: “And yet as readers, we are expected to skip right over that - and the curious thing is that we usually do! The incoherence passes invisibly right through our filters, and we often see nothing wrong at all” (10). Any Literary Translation Pandemonium must inevitably engage this paradox, and somehow work around it; and since the DHP did some literary translating, and commented in this 100-page booklet on the 200-page literary translation with which it shares an omnibus volume, they knew it quite well. It is no longer invisible to them, and they wanted to make it visible to their readers as well. A salutary desire!

They are, however, not at all interested in how that paradox twists around into a strange loop that is constitutive of cultures. Why do we think of “culture A” and “culture B” as discrete entities? Why is French culture separate from English culture? Why does it seem so odd and even deliciously absurd to us in Monty Python and the Holy Grail when King Arthur and their knights arrive at a castle in England where everyone not only speaks French but seems to think they are in France? In Critical Translation Studies (Robinson 2017c) the DRP demons track the work of Sakai Naoki (1997) showing historically how translation practice constructed “Japan” and “the Japanese language” and “Japanese culture” as national entities; what we don’t track there is how that constructionist history proceeded through strange loops. We will be exploring the strange loops of that history in one of the chapters of the book that will eventually emerge out of these early ruminations.

As we’ll see in what we think will be Chapter 1, in fact, the DHP tended to resolve these paradoxes by urging the translator to translate sense for sense. We will show there that their resolutions not only fail to resolve the paradoxes, but twist and loop them even more tightly, and even more strangely. But they did not always even recognize the paradoxes in which translation is mired. Consider for example this passage:

The pithy slogan “Translator, traitor” shows very clearly that a translator need not be a betrayer or traitor, for it beautifully preserves the key quality that makes the original Italian phrase so memorable - namely, its catchiness, which is due to the fact that the two nouns inside it sound so much alike. There is no important aspect of the phrase Traduttore, traditore that is missed by “Translator, traitor,” and so this English translation is a checkmate in response to the strong-seeming check tendered by the Italian opponent. (Hofstadter 2009: 18)

We would argue that it’s not actually a checkmate, because a checkmate ends the game, while the elegant translation of traduttore, traditore as “translator, traitor” keeps looping back around strangely, potentially without end. By calling the translator a traitor in a nontraitorous translation, the target text “translator, traitor” both confirms and disconfirms the Italian slander, and does so cyclically, very much like the liar paradox. The two words of the Italian dictum (source text) say to the English translator (target author), “you can’t translate us elegantly, because in translating, you invariably commit treason against us”[5]; and the two words of the target text, “translator, traitor,” which implicitly contradict that accusation (“See, I can translate you without treason!”), also explicitly reiterate the accusation, returning us back to the beginning: “Ha, see,” the source text word-pair retort, “you agree with us!” Just as the source text operates simultaneously on two levels, the propositional (“you can’t translate without treason”) and the stylistic (the eloquent consonance of “trad(u/it)tore”), the target text, in attempting to operate equivalently on the same two levels, proves the Italian proposition both wrong and right, wrong stylistically (by mimicking it) and right propositionally (by reiterating it). Every attempt the target text makes to rise above the propositional accusation through eloquent propositional and stylistic equivalence returns it to the starting point; any potential attempt the target text might make to deviate from propositional equivalence would simply prove the source text right.

 

The Strange Loops of Translation

So where might we begin to look for the strange loops of translation?

The initial answer is quite easy: If the strange loop of consciousness is generated by analogical parallels between the symbols an “I” is thinking about and the “I” itself thinking about them, the strangeness of that loop is only compounded when the translator inserts their “I” in between a source author and a source reader and deflects or redirects the rhetorical situation from the source culture to a target culture.

The DHP’s reflections on “fidelity” in their translation of Sagan into highly Americanized English are rife with examples of this. When Sagan’s character Antoine says “et il ne dit rien,” for example, Hofstadter bristles at the notion that they should be “faithful” to Sagan by translating that “and he said nothing,” as the novel’s previous English translator did. “I, by contrast,” they noted, “stray a bit further from Sagan, and write ‘and he bit his tongue’” (64).

What we find striking there is that the DHP imagined themselves straying further not from Antoine’s words, nor from the narrator’s report of Antoine’s words, nor even from Sagan’s creation of the narrator’s double-voicing of Antoine, but from Sagan: from the authorial “I.” The Hofstadter demons have most likely not read Bakhtin on double-voicing, so that’s a bit of a red herring; they are also not enough of a literary scholar to distinguish between authorial voice and narratorial voice (cf. 2009: 85). But surely they recognized the analogical shifts that give us both level-jumps and significant parallels between Sagan’s “I” and Antoine’s “I”? Certainly they seemed to be fervently aware of the analogical jumps/parallels between Sagan’s authorial “I” and their own translatorial “I”: When they asked, rhetorically, “But do I, a mere translator, have the right to turn up the clarity and vividness knobs?” (64), their response came very close to theorizing the rhetorical construction of the translatorial “I,” but fell short of that strangely loopy goal, and in falling short arguably provided tacit support for what the 2007 DHP repeatedly derogated as the “hoax” or “hallucination” of the stable ontological “I”: “Well, the fact is that I’m naturally inclined to turn these knobs up high no matter what I’m writing, because clarity and vividness are, in some sense, my religion. I would be betraying myself if I didn’t allow myself to be as clear and as vivid as possible when I translate” (64).

But, regarding that “arguably”: Let’s not jump to conclusions here. Let’s give the DHP the benefit of the doubt. The “naturally” is probably a colloquialism that shouldn’t be read as implying the DHP’s belief that their “I” is a naturally occurring entity with specific stable qualities implanted there by “nature.” The author of Gödel, Escher, Bach and I Am a Strange Loop would certainly not want to be caught objectifying or essentializing (or even singularizing) their own “I”! (It may be “naturalized,” in the sense of coming to feel natural through frequent repetition and resulting habitualization, but it is never “natural.”) And the assertion that “clarity and vividness are, in some sense, my religion” is saved from the opprobrium of objectivism and essentialism not only by the hedge “in some sense” but the very notion that it is a “religion” - something that Hofstadter, too, would agree is not so much a stable set of ontological truths as it is a set of conventionalized and normativized social practices, beliefs, and values. (Calling clarity and vividness their “religion” doubles up the rhetorical constructedness undergirding “religion”: as a social construct and as an analogue of the “I.”)

But even the normative invisibilizing of the translatorial “I” is loopily constitutive of the translatorial “I.” The more closely we study the actual verbal practicalities of translation, in fact, the stranger the loops become. As we’ve noted, raising the DHP’s “Wrong-Language Paradox” to the status of strange loopiness would mean not only tracking the impossible loops as a curiosity, but following the 2007 DHP themselves in tracking the constitutive force such pandemonial strange loops wield in the generation and maintenance of personal and cultural identities. More than that, too, it means envisioning and experimenting with alternative translational strategies for mobilizing that constitutive force for new cultural expressions, orientations, behaviors, norms and values, and so on.

 

First Strange Loop of Translation: Self-reference

For example, think of the complications that arise whenever the source author uses indexicals: “this,” “here,” “I,” “now,” and so on. When Martin Luther writes in their “Sendbrief vom Dolmetschen” (1530/2004)/“Circular Letter on Translation” (Robinson 1997/2015: 88), “Wenn ich nun den Buchstaben nach, aus der Esel Kunst sollt des Engels Wort verdeutschen, müsste ich so sagen: Daniel, du Mann der Begierungen, oder Daniel, du Mann der Lüste,” by which of these three logics should we translate that into English?

1. German-German: “If I now Germanize the angel’s word, should I say: ‘Daniel, du Mann der Begierung’?”

2. German-English: “If I now Germanize the angel’s word, should I say: ‘Daniel, you man of desires’?”

3. English-English: “If I now Anglicize the angel’s word, should I say: ‘Daniel, you man of desires’?”

Only the a-model there functions logically in terms of Luther’s choices - after all, Luther and their assistants translated the Bible into German - but “du Mann der Begierung” is Luther’s (mocking) German translation, which the target reader presumably can’t read. Only the c-model functions logically in terms of the choices made by target readers - after all, they read the text in English - but Luther didn’t translate into English. There is no logical solution. The “preferred” (normative) translation would follow the b-logic, which of course is completely illogical, because “you man of desires” is not German. The conceptual “workaround” (actually more of a “let’s just not think about it”) that is tacitly invoked to justify this illogic in normative translation practice is that this is just one of the unresolvable paradoxes of translation, sigh. And, after all, if a scene in an American movie is set in Germany, where Germans are speaking to Germans, and the American actors all speak English with German accents, we accept that they are “actually” (imaginarily) speaking German. So what if it’s not “real”? We can live with the paradox. We can pretend. We can suspend disbelief. (This would be more or less the approach championed by the DHP’s “solution” to the paradox.)

Luther’s letter in German is a Sendbrief , literally a send-letter: you send it around to people. In English it is a circular letter, a letter that circulates. In the first strange loop, what circulates is layers of the translated translations of the letter, problematically distributed among Luther’s author-function, Luther’s translator-function, and the DRP’s translator-function (those Foucauldian concepts to be defined and explored by the DRP’s author-function in more detail in one of the book’s chapters). The imagined hierarchical upward-chain of inference is:

Since the DRP’s translator-function is also assigned the task of rendering that move from 2 to 4 into English, the target reader’s pleasant passage up those latter steps in the idealized hierarchy is disrupted. All of a sudden we find Luther’s German author-function describing the German activity of their translator-function as if it had all happened in English: “Now if I wanted to translate this like the jackasses insist, literally...” (the DRP’s actual translation on 88). The DRP’s c-translation in step 1, in other words, returns the analogical stair-stepping up the hierarchical chain of inference from the implied higher levels (Luther doing this or that in steps 2-4) to the DRP’s own English-speaking translator-function back in step 1. And if we try hard enough not to be disruptive in that strangely loopy way - if we try to preserve the logic of Luther’s author-function by writing, “Now if I wanted to Germanize this like the jackasses insist, literally...” - we still have to decide whether to give Luther’s suggested translation in English (again, disrupting the hierarchy) or in German (foreclosing on our translator-function and not serving the target reader); and if we take that latter course, the first step up the hierarchical chain of inference is missing, and the target reader trips and falls.

Of course there is also the potential disruption caused by the fact that in all three of the DRP translations Luther is speaking English - what the DHP called “The Wrong Tongue Paradox.” This too might disrupt target readers’ idealized chain of hierarchical inference, if they begin to realize that what the DRP translator-function is leading them to is not the real Luther but Luther’s English-speaking author-function, a kind of author-function prosthesis as it were. But as the DHP themselves pointed out, most target readers never notice this problem. The disruptions that the DRP author-function called the first strange loop of translation are the more flagrant ones.

Then again, considering that the DHP’s author-function, which first noticed and theorized strange loops, is also a translator-function and attentive reader of their own translations, and they didn’t notice (let alone theorize) these supposedly “more flagrant” disruptions, perhaps the DRP author-function is just imagining their disruptive effect?

 

Second Strange Loop of Translation: The Incoherently Written Source Text

The incoherently written source text is of course a disturbingly common occurrence in technical translation - something that the demons of the DHP’s translator-function have either never tried their hand at or never thought interesting enough to theorize. The marketplace norm in this case is strangely loopy: the normatively idealized translator-function must both translate the text equivalently and fix it up, so that the translation gives target readers the impression that the interiority of the source author’s “I” has been clearly and faithfully communicated to them. Needless to say, a well-written target text is not equivalent to a badly written source text, and the “clarity and vividness” of the translator-function’s idealized “I” is not equivalent to the pandemonial murkiness of the sloppy source author’s unidealized “I”! But that’s the desideratum.

This second strange loop is akin to the Drawing Hands sketch of M.C. Escher, which the DHP reprinted and discussed as a classic example of the strange loop (Hofstadter 2007: 102-3). Just as the viewer’s response to the Escher drawing is to climb what appears to be a “natural” hierarchy from drawing to hand to arm to artist, but ends up at the point of another pencil and another hand holding it and drawing the first hand, so, too, is the target reader’s response to the “fixed-up” translation to climb from the coherent target text toward what they might take to be a coherent source text, and beyond that to the source author’s intended meaning, and beyond that to the full serene interiority of the source author’s “I” - only then to find themselves precisely where the translator started, at the point of creating a coherent text out of the incoherent original.

Ideally, of course, the translator is normatively expected to contact the source author and ask a series of clarifying questions: “What did you mean here?” and “Which interpretation of this syntax is correct?” This would engineer the construction of a coherent authorial “I” by means of an activist participatory audience-effect, which would in turn engineer the transformation of the audience-effect into a translatorial “I” in the target language.

But the authors of incoherent technical texts typically have no coherent public “I” that they can represent normatively in writing. Whether they are verbally incoherent in both their speaking and writing or only their writing, there is an inarticulacy in or about them that is constitutive of a fractal or fractured “I,” a disorganized and disruptive pandemonium whose demons are all pulling in opposite directions. These are most often nameless technicians buried deep in the bowels of some gigantic corporation, often in fact nameless teams of technicians who hate to write and do it on reluctant assignment, through gritted teeth. (Back in the mid-1990s, we had a good friend who was a tech writer and had worked in several of these companies before realizing that the cultures there were death for a writer’s professional competence and self-respect, and thus went freelance.) Sometimes the badly written source texts are themselves second- or third-generation translations from Japanese or Chinese, cobbled together too fast and with too little professionalism by those nameless technicians, or, these days, by free online machine translation apps with little or no post-editing. Even if by some fluke there is an identifiable source author, they are typically sequestered behind a translation agency’s firewall, lest the freelancer hired to translate their output try to “steal” their business (bypass the agency so as to maximize profits). And if, by some miracle, the translator does manage to contact the actual living, breathing source author, it is axiomatic that writers who submit incoherent texts for public use are incapable of even recognizing, let alone explaining, let alone resolving, syntactic confusions.

So what is the norm-obedient translator to do? How does the pandemonial but professionally responsible translator produce a pandemonial but coherent translation that is arguably faithful to the incoherent source text?

The answer is: they must invent it. They must rewrite the text coherently, based on their own best guesses.

But surely this is progress ? Surely, in the DHP’s terms, this is not a case where “one winds up, to one’s shock, exactly where one had started out,” and so not a strange loop?

No, because the marketplace requirement is that the translation be equivalent . If the translator invented the translation’s coherence, then it’s not equivalent, and by definition not a translation; but since the translator was hired to produce a translation, then the text they submit is a translation, and by definition also is equivalent; and so on, in potentially infinite recursions.

Gideon Toury’s (1980: 63-70) ingenious suggestion was that all texts regarded as translations are by default equivalent to their originals - even if there is no original, as when a pseudotranslation is “mistakenly” regarded as a translation - because equivalence is the normative assumption that attends the translation-designation. The presumptive translationality of a text makes it equivalent, which makes it a translation of its proclaimed source text - say, James MacPherson’s hoaxical collection of “Ossian” poems - which is admired greatly through the self-constitutive loop of the translation until a plausible enough hoax accusation creates a big enough backlash audience-effect to make the presumed source text vanish like smoke on the wind.

Thus no matter which way we try to straighten the loop, to iron out its strange twists, we keep ending up exactly where we started - even though that place makes us extraordinarily uncomfortable.

Again, though, the cognitively and affectively productive question is: what does our discomfort generate? If the strange loops of perception generate the “illusion” or the “myth” or the “hoax” of the authorial “I,” what does our discomfort with the strange loops of fix-it-up technical translation constitute? What is the constitutive effect of those strange loops on the translatorial “I,” which has been professionally formed to submerge itself in (and thereby loopily work to constitute) the putative authorial “I”? Does the translatorial “I” begin wanting more? More money, more recognition, more joy? Does the newly awakened (but still vestigially suppressed) translatorial “I” begin to assert its “rights” more aggressively, more egotistically, saying things like “I would be betraying myself if I didn’t allow myself to be as clear and as vivid as possible when I translate,” and “were I told that I had to adopt the principle of such rigid ‘faithfulness’ to the author, then I would just give up translating, for it wouldn’t allow me to use my own mind” (Hofstadter 2009: 64-65)?

 

Third Strange Loop of Translation: The Passage of Time

Finally, should the translator of, say, Homer’s Odyssey , or the Bible, archaize the target language, because the source text is archaic? Or should they modernize it, because (we’re told) the source text was not written to be archaic, and was originally heard by source listeners as ordinary contemporary speech? The marketplace norm in this case vacillates, but tends to lean more toward modernization - or rather, toward a cautious kind of modernization that is also just elevated enough to create the illusion (audience-effect) of being old and venerable. Nothing slangy; nothing impenetrably ancient; nothing in the target reader’s face: just enough hints in both directions to allow the target reader to feel comfortable projecting some kind of idealized normative coherence onto the text.

Even those modernizing loops are strange, in fact. Their strangeness has a lot to do with the notion, perpetuated among others by Walter Benjamin, that translations age but originals do not. Because this is a temporal loop, we might well apply to it what Umberto Eco in “The Myth of Superman” (1979) theorizes as a temporal “iterative scheme”: Just as Superman and the other characters in the strip undergo experiences that in the real world would age them but they never age, always starting back from scratch in the next strip or the next comic book or movie, so too does each modernizing retranslation seek to restart the original back at “scratch,” at some imaginary starting point that is to be understood as its “true” “origin.” The rhetorical construction of this restart is that it corrects for the “temporal distortions” caused by the aging of past translations - but then it, too, ages, and it, too, must be replaced by a new retranslation that restarts the loop. This retranslational iterative scheme seeks, through an audience-effect that is highly vulnerable to the passage of time, to stabilize the source text, to fix it in time, so that it is perceived as not aging - but it is perceived that way only through the rhetorical ministrations of each new retranslation. Because each new retranslation presents itself as capturing the true originary and stable essence of the source text, what the target readers who ratify that presentation experience, even for a fleeting moment, is the fortuitous convergence of source and target texts at a starting point in time that is surreptitiously (ideally) out of time. For those target readers who do not ratify the image of serendipitous convergence, the new retranslation is just another retranslation - and of course the numbers of these latter readers tend to swell over time.

The strange loops of archaizing translation work along similar lines; the translator approximates not some primordial moment at which the source text was modern, but rather its current archaic flavor. One fairly widespread opinion is that the Odyssey and the Bible are not only ancient texts but valuable specifically for their hoary antiquity, and therefore should be archaized in translation as well. What is then being rhetorically stabilized is not an imaginary past origin-moment at which the source text was “modern” but an imaginary present noble-rust-moment (see Schlegel 1795/1846; Robinson 1997/2015: 214 in English) at which the source text is always respectably “old.” A classic example would be Rudolf Borchardt’s 1923 Dante Deutsch , with its radically archaizing aim of reimagining German history had Dante written originally in German. According to George Steiner (1975/1998: 357), Borchardt’s archaizing project “is not antiquarian pastiche, but an active, even violent intrusion on the seemingly unalterable fabric of the past” - but of course that violence was inflicted not on some ontological “fabric of the past” but on an imaginary past as constructed in the present as an audience-effect. And to the extent that German readers in the mid-twenties ratified Borchardt’s archaic German as “German Dante,” they were participating in the stabilization not of some past German Dante but a present-day archaic German Dante - Dante in contemporary (twenties) German culture as old and in German. And like the “true” “originary” “essence” of the source text as rhetorically stabilized by modernizing retranslations, that present archaic German Dante is/was subject to temporal changes in German reading practices over time; anyone seeking to restabilize that present archaic German Dante would need to retranslate it, and somehow convince target readers that now , finally, the archaic German Dante had finally converged and merged with archaic Italian Dante.

Where the strange loop of archaizing textuality becomes especially loopy, however, is in texts deliberately written to be archaic. In this sort of source text the archaizing strategy becomes an implicit self-reference: “I am archaizing this text,” we can imagine the Source Author Pandemonium (SAP) implicating by and through their writing strategy; or, more explicitly, “I am writing this text in Year X in Culture A, and imitating the style of Year/Decade/Century Y.” Authorial archaism, in other words, is a complex kind of culturally situated and loopily performative self-reference: “We are performing our ‘I’ as simultaneously existing in both this time and place in which we now stand and that earlier time and place in which our characters and/or imaginary source readers stand.” The complexity of that dually or multiply situated performativity is especially salient because, typically, the culture of the earlier time and place is performed (by both the source author and the source reader: a dual audience-effect) as historically constitutive for the writing “I” in the present time and place.

And all that of course applies only to the source side. What happens when the translatorial Target Author Pandemonium (TAP) insert themselves into the middle of that performativity, and partially hijack it into another invented (target-cultural) “past” as imaginary prologue for target readers in another “present” (this) time and place (now)? What authorial, narratorial, and translatorial “I’s” are constituted through this hijacking, and what do the demons of the Target Reader Pandemonium (TRP) naturalize as the “true” or “authentic” “self” or “core” of the source-becoming-target text, and how does that naturalization work? Or is it possible that the TRP naturalize a split “I”? The Pandemonium model posits not just the existence of fragments or fractals of the various “I’s,” but the ballistic launching of those fractal “I’s” as demons, all jostling for position and control of articulation. In translation, the question becomes whether the rhetorical audience-effect cohesion of those demons into a single coherent “I” would depend on the coherence of the authorial SAP’s “I,” or of the translatorial TAP’s “I.” Or would any relative coherence be a periperformative “hoax” or “illusion” built collectively out of the three-way rhetorical interaction, the circulation of evaluative audience-effects among SAP, TAP, and TRP? (We will be mobilizing the periperformative as theorized by Eve Sedgwick [2003] in one of the chapters in the book.)

This third strange loop is obviously a strangely loopy escalation of what the DHP called “The Wrong-Place Paradox,” in which, because the source text is saturated in the history of the source culture, and the target culture lacks that history, “when idioms of culture B are placed in the mouths of individuals from culture A, the result is incoherent,” as the DHP put it. We would demur on that “incoherence,” which is really a problem in the second strange loop, not this one. (“I” think what they mean by “incoherent” is not so much confusing and messy but anachronistic, anatopistic, illogical [mismatching] and therefore putatively incoherent from the perspective of eternity.) Escalating the DHP’s talk of “paradox” into a strange loop entails an exploration of the ways in which “the result is [not] incoherent” but merely mind-bogglingly and recursively complex; adding the strange loopiness of deliberate archaism further complicates that exploration.

A few years ago the DRP faced something like this problem in translating Finland’s greatest novel, Aleksis Kivi’s Seitsemän veljestä (1870), as The Brothers Seven (Robinson 2017a; see also Robinson 2017b). The Aleksis Kivi Pandemonium (AKP) had written the novel in the 1860s about their parents’ generation, using a stylized Southern Finnish dialect from the 1840s to flesh out a paleo lifestyle that by their day was almost defunct. What to do in English?

Archaic texts are notoriously difficult to read, and lectorial TRPs typically expect the translatorial TAP to make things easier for them; indeed some translation scholars call the lexical simplification of the source text - “the language is usually flatter, less structured, less ambiguous, less specific to a given text, more habitual, and so on” (Pym 2010: 79, summarizing Toury 1995: 268-73) - a universal of translation. This makes archaizing translation strategies counternormative, at least in a generalized “marketplace preference” sort of way.

The interesting tension in the market’s equivalence norm, however, is that it requires the translator to capture both the semantic content and the style of the source text. The commonsensical (low-end) response to that norm is that it is impossible to achieve both; hence the standard retreat from the full implications of the norm into dumbed-down sense-for-sense equivalence, which is entirely compatible with lexical simplification and stylistic flattening. The greatest literary translators have historically sought to achieve both - apparently mutually exclusive - goals, but typically with a greater willingness to fudge semantic equivalence where necessary in order to achieve something like stylistic equivalence (or perhaps just stylistic brilliance).

That latter (or “higher”) approach would seem to require that the translator archaize, and thus create a text that TRPs - beginning with acquisitions editors - will typically find excessively difficult. While the AKP’s Finnish is hilarious, it is also quite difficult for the “ordinary” or “uneducated” or “untrained” demons of the Finnish Source Reader Pandemonium (SRP) to read, and the previous two (low-end) translations, from 1929 by Alex Matson and 1991 by Richard Impola, do make it simpler. In that way, arguably, they do the English TRP a great service. The novel is much easier for TRPs to understand in their English than it is for SRPs in the AKP’s Finnish.

Unfortunately, however, Matson and Impola also purge the novel of all humor, and indeed all other literary qualities as well. Maybe that, then, is not such a great service? If the “universal” of translation is to reduce a brilliantly funny but difficult novel to tedious, turgid, pedestrian simplicity - flatter and more habitualized than the original, with rhythms and other prosodics deadened, with alliterations and other euphonic effects dulled through a tin ear, with the delectable paleo vocabulary dumbed down through a deluge of high-frequency lexical items - maybe translation is “universally” not such a great thing? Maybe then the Comparative Literature Pandemonia who have long warned us against translations have always been right? Maybe the only answer for anyone who loves literature is to learn the source language and read the work in the original?

So what did the Douglas Robinson Pandemonium do with the AKP’s novel? We translated it into a playful reconstruction of Shakespearean English - from the early seventeenth century, much earlier than the AKP’s 1840s Finnish rural dialect. Why? Partly because the AKP was saturated in the William Shakespeare Pandemonium (WSP) in Swedish translation, had memorized several whole plays in Swedish, and often translated quotes from the WSP’s plays into their Finnish texts; and because their archaic Finnish was always playful, always designed to make the reader laugh out loud, and in every way strongly indebted to Shakespearean humor. (Also to Cervantean humor, but we were not translating them into Spanish. Had we been, we would certainly have used Cervantean Spanish.)

Partly, too, our strategy was based on the assumption that English readers would be at least slightly familiar with the WSP, and would enjoy a playful Shakespearean English more than some obscure 1840s Scots dialect, say.

Mainly, though, our strategy was motivated by the fact that we, too, love the WSP, much as the AKP did. Gearing up to tackle the considerable challenge of this translation, we assumed - perhaps projecting this onto the AKP - that they and we both felt the “same” thrill in the brilliant edge of the WSP’s writing, the restlessness of their intelligence, the explosive audacity of their humor. Prompted by that “shared” love, we blended the mutual self-constitution we experienced between “our” WSP and ourselves with the mutual self-constitution between “our” AKP’s WSP and “our” AKP, all of us creating and organizing the others’ “I’s” in playful interaction, or active interplay. We read the AKP reading the WSP, we read the WSP reading the AKP, they read us reading both, and gradually the strange loopiness of all that anachronism constituted the “I” of the AKP who shadowed us as the DRP was translating through the word-demons of a partly invented Shakespearean English.

Or, all right, we did it all. It all happened in our head (one head, many demons). Like the translator who invents a target coherence for an incoherent source text, we invented a playfully archaic style for our translation that radically exceeds the original - that gives the impression of archaizing equivalence only as a periperformative audience-effect. Progress? Not really a translation? Not, therefore, a strange loop? No, because again, it’s a translation, and therefore equivalent. Back to square one.

And again, the big question is not only how that strange loop shaped (reconstituted/collectivized) our translatorial “I,” but what effect it might have on target readers’ “I’s,” and thus on the target culture. (We will return to the question of archaizing translation strategies in one of the chapters, drawing on Johannes Göransson’s 2018 use of Tiffany 1995: 195-96.)

 

Conclusion

To the casual reader it may seem, from our discussions of Hofstadter on translation - especially “Translator, Trader” - that the article is set up as a hit job on Hofstadter’s thought. Not so. The DRP’s demons are great admirers of the DHP’s cognitivist theorizing - even when they/we disagree with it. Those demons also find the DHP’s remarks on translation in Gödel, Escher, Bach and I Am a Strange Loop cognitively exciting, and their remarks on literary translation in Le Ton Beau de Marot witty and artistically inventive. There are, to be sure, DRP demons that read “Translator, Trader” as somewhat unfortunate, yes; but in the aggregate the DRP demons are fans of the DHP’s work, and it is precisely as fans that they/we would like to make a contribution to its aggregate success, especially by building the bridges they so far neglected to build between their theorizing of translation and their theorizing of strange loops.

 

References

 

Notes