Saturday, April 8, 2017

Interview with a Translator, Part 2: Words Words Words

From Holy Kaw
Wrapping up with some reflections on language.  

Kate: What accounts for the excessive passive voice and vague pronouns in poorer translations?
Eugene: It's mostly from translating Japanese too quickly and too literally.

Japanese advantages its close integration with the culture and society to "compress" the grammatical structure whenever possible, shifting most of the heavy lifting to the verb and a myriad of agglutinative conjugations at the end of the sentence.

Consider as well that the shadow of feudalism lasted into the 20th century. Along with it came the lexical complexity of marking status and using honorifics. Thus dropping the subject of a sentence became a desired efficiency. (Along with titles taking the place of pronouns.)

But the "compression" in Japanese is often "lossy," which is difficult to reverse because of lost information. Unlike English, which tries to pack all the available data into self-contained sentences (and uses subject placeholders like "it" to keep the structure intact), Japanese can scatter information all across the page.

From the perspective of English grammar, Japanese favors "passive" formations that skip the subject ("Mistakes were made"), and sees no problem in failing to mention the subject for another several paragraphs. A Japanese writer can easily create a page of third-person narrative that fails to clarify the sex of the POV character. That's hard to reproduce in English.

One translation "shortcut" is to have a native Japanese speaker do a rough translation and then have a native English speaker do the cleanup. The problem here is that the cleanup editor may have no way figuring out the antecedent to one of those vague pronouns.
Purple Prose,  Prather-style
Kate: Some light novels have what is sometimes referred to as "purple prose"--it varies considerably from poetic to explicit. Do translators make a conscious choice which approach to take? Does the original text make the decision for the translator?
Eugene: I'd say the original text pretty much dictates the final product. There's always leeway in tone and word choices, but the explicitness of the terminology pretty well controls the explicitness of the prose.

Agatha Christie, Raymond Chandler, and Elmore Leonard wrote detective novels, but their use of "vocabulary," shall we say, is quite different. It mostly comes down to a matter of discerning the sociolinguistic milieu and the genre, and then deciding who the audience is.

 Harlequin novels--rights likely obtained for cheap--turned
into manga by Japanese artists.

Or rather, figuring out who the author pictured as his readers. Once you get all those variables adjusted properly, so that you are writing in the same mindset for the same readers, you don't have to think about it that much.

Although there is always the challenge of making purple prose not sound so purposely purple.
Kate: In the previous interview, we discussed colloquialisms—the difficulty/necessity of translating figurative language between cultures versus letting the phrases/references stand. Some translators seem to fall back on clichés due to lack of imagination. Sometimes, however, the original writer appears to deliberately use a cliché. How does a translator recognize and handle clichés?
Eugene: In a very real sense, all language is a cliché or we couldn't understand each other. Like continents and species, language drifts and mutates. Before long, the past and the present (and the here and there) are miles apart and have adapted to quite different environments.

Language is thus a moving window that attempts to pin down usage within a certain time-frame in order to maximize comprehensibility. Most usage is effectively transparent. We process it without paying undue attention to the semantic and syntactical structure.

When we do start paying attention, that window starts moving. Some usage, like the subjunctive, dwindles away over the protests of a few stubborn grammarians. A lot is like fashion. Some usages never go out of fashion, and others can't go fast enough.

Stock Phrase
So there are expressions that last for centuries, while others, like bell-bottoms, get shipped off to the Salvation Army with a roll of the eyes. And maybe some creative soul will find a totally self-aware use for them that brings the cliché back to life again.

In Japanese, there is a whole category of what are called four character idioms, often adapted from Chinese. They are expressions compressed to their essence, like saying "Two birds one stone." A couple dozen would qualify as cliches. The rest can get quite arcane.

And as in English, Japanese has stock phrases. For the non-native speaker, it can be difficult to identify an ironic usage when it comes into play. Luckily, Japanese tend to avoid irony. But contemporary references can be just as tricky. You can at least look up historical allusions.
Kate: Speaking of allusions, they can crop up unexpectedly. As P.J. O'Rourke mentions, when Senator Kennedy mocked Vice President Bush during the 1988 Democratic Convention by asking, “Where was Bush [during Reagan’s scandals]?” the reporters watching immediately responded with, “At home, in bed, with his wife.”

Is the creation of contemporary allusions/slogans easier or harder to see in another culture? How “current” do you have to stay in order to “get” other cultures’ allusions?
Eugene: The most recent Godzilla movie apparently makes veiled references to Fukushima and the subsequent political storms. Those are easy enough as long as you keep up on the news. Harder are trends that truly are "socially constructed," that come and go like mayflies.

On the other hand, language that is to subjective would probably not be accessible to a foreign audience either, so translated too literally you could end up with translated language that isn't any more comprehensible. 
Kate: Different countries use different punctuation. For example, American quotations are double (“) on the outside, single (‘) on the inside; the reverse is true in much British literature. And when I was taking French literature, many of the books used <> to indicate a speaker speaking.

What do the Japanese do? Do you “translate” punctuation?
Introduction to Japanese Punctuation
Eugene: I've always found Japanese punctuation to be logical and comprehensible. Perhaps because there is no interference from the familiar conventions I already associate with Latin scripts, my brain maps punctuation marks pretty much on a one-to-one basis.

Japanese has adopted several punctuation marks directly from Latin script, including the exclamation point, question mark, parentheses, and the comma. And increasingly uses smart quotes (“…”) alongside the traditional kagi kakko (「…」 and 『…』).

Emphasis (italics) is indicated with a dot or comma next to (or above) each character (bouten, meaning "side mark").

NHK in particular likes using smart quotes rather like "air quotes." Kagi kakko remain the standard in narrative fiction and the usage is almost the same, although it is quite common for any dialogue enclosed in kagi kakko to be separated into its own paragraph.

Yes, this can at times make it easy to lose track of dialogue tags.
Kate: Is there any grand unifying theory that explains how language works? And does a grand unifying theory help the translator?
Eugene: Language universals do exist, but it's tricky getting from there to the "universal grammar" concepts pioneered by Noam Chomsky, that tie language to structures in the human brain that work exactly the same for everyone everywhere.

As a result, a "linguistic theory of everything" remains as elusive as it does for physicists, who end up with compelling explanations and neat ideas and no way to empirically test them.

Unfortunately, Chomsky was still all the rage when I was in graduate school so I had to study transformational grammar. This was Chomsky's attempt to create a calculus of language.

It is a useful tool for analyzing language but not necessary for creating real-world
Language is a grass-roots thing.
functionality or for describing how language actually works in the minds of the human beings using it.

But in the 1980s, Moore's Law was taking off. The revolution in computer technology
triggered much wishful thinking that rules-based computing could solve all the difficult algorithmic problems that had eluded the more mechanical processes to date.

One of the goals of the Fifth Generation Computer project, initiated by Japan's Ministry of International Trade and Industry in 1982, was machine translation. It pretty much completely failed.

Simply consider the imprecision of rule-based grammar checkers. They're useful only when paired with human beings who can weed out all the false positives.

A chess or go program based on algorithms alone can play a pretty good game. But beating a smart human requires pattern recognition based on massive real-world data sets and machine learning systems. Saying "Oh, this resembles that" a billion times a second.

Pattern recognition is the key. It's at the core of all modern machine translation systems. It's what the human brain does best (so well we eagerly perceive patterns where they don't exist).

But, again, we can't confuse explanation with application, descriptions of how language works with prescriptions of how it ought to work. What's of actual use to a translator also involves universals but at a much higher level. I'm talking about story universals.

In other words, Joseph Campbell instead of Noam Chomsky. Less universal grammar and more monomyth. (Well, and you do need a good copy editor.)

Granted, art can get so abstract at one extreme, and so culturally-bound at the other, as to defeat reasonable attempts to identify the shared patterns. But neither is there a point in translating stories without universal appeal.
Ah, words are not enough . . . except, Thanks, Eugene!