Archive for the ‘language’ Category

Gods and kami: etymology and semantic drift

Tuesday, January 18th, 2005

When I wrote in an earlier post about Japan’s “Batting God”, “God” was a direct translation of the Japanese kamisama . The nuances line up perfectly. Both Gods and kamisama are unrivaled, omniscient, and omnipotent entities—in their particular contexts, batting in this case.

So it’s not surprising that when Christianity reached Japan, the native word kamisama was immediately adopted as the translation for “God”. (African tribesmen also used native words for supernatural beings to refer to the new white God the missionaries introduced.)

“God” itself derives from the Proto-Indo-European “ghut”, meaning “that which is invoked”, “that which is called upon”, although some scholars think the origin is a word meaning “to sacrifice”. The alternative, deva/theos/deus/dios/dieu, derives from a different Indo-Germanic root meaning “to shine”.

Originally “god” and its etymological forebears indicated any spiritual force or influence present in a thing, a phenomenon, an animal, or a place. This nuance is extremely close to that of the Latin “numen” (whose original meaning, by the way, is “a nod of the head”).

Interestingly, this is also precisely the original nuance of the Japanese word kami. Kami are the gods or spirits worshipped at each Shinto shrine, and indeed the name “Shinto” means “Way of the Kami”. A kami could be the spirit of a well-known local figure, a more traditional historical/mythical god-type being, or, just as easily, a river, mountain, animal, or even rock. I remember visiting one shrine where the object of veneration, located on the back wall and symbolizing the kami being worshipped, was simply a hole in the wall—letting you gaze through at the mountain behind the shrine. The mountain was the kami.

[Note that many students assume that kami meaning God is the same word as, or shares an origin with, a different kami meaning “up”/”above”. But linguists tell us that this is not the case—the “mi” of the kami meaning God was originally a different sound than the “mi” of the word meaning “up”. So just as English “god” and “good” are not related (God is not necessarily good?), neither are Japanese kami/god and kami/up (God is not necessarily up?).]

We thus have the odd historical parallel that words which originally pointed to a diversity of spirits, immanent in objects and places, were borrowed to refer to a monotheistic entity in both hemispheres: the West (“god”) and, much later, the East (“kami”).

(Calligraphy by Kanjuro Shibata Sensei.)

Philosophy of translation

Tuesday, January 18th, 2005

We’ll take a look at the first line of Genjo Koan to explore some aspects of the translator’s assumptions and tasks.

The first phrase is:


Got that? Or not? Maybe it’s because you cannot read Japanese. Note—I’m assuming that. I’m assuming the reader cannot read Japanese characters and that therefore I need to do something about it. If I did not make that assumption, there would be no translation task. We want to provide something you can understand, at some level, and clearly given its unfamiliar script the original Japanese is not that, for you the Western reader. We’ve now made our first assumption. I’m highlighting this obvious point to emphasize that assumptions run throughout the translation process. They are assumptions about our readership, their capabilities, their interests.

To address this first problem, we’ll perform “transliteration”: converting one writing system to another. A standard transliteration would then give us:

shohou no buppou naru jisetsu

Got that? Or not? Maybe it’s because you cannot understand the words. There’s another assumption—that the reader cannot understand Japanese words and that therefore I need to do something about it. This is the second assumption necessary to provide something usable to the Western reader, another assumption about their capabilities.

To address this second problem, we’ll perform “translexicalization”—converting one set of words to another. That would yield:

all-law subject-particle Buddha-law be time

Got that? Or not? Maybe it’s because you cannot understand the grammar (syntax). Again, I’m assuming the reader cannot understand Japanese syntax and that therefore I need to do something about it. This is our third assumption. To address this third problem, we’ll perform “transsyntaxification”—converting one grammar to another. That gives us:

When all-law be Buddha-law

Aha, we now see something that resembles English. Some people would already call this a “translation”, although perhaps Nabokov’s term “transposition” would be more appropriate. We could actually publish this and plausibly claim it was English, and that we had “translated” it. But do you really get this? Or not? I’ll assume that you cannot understand the terms “all-law” and “Buddha-law”, which we went to such trouble to “translate”.

To make this “understandable”, we need to do something with these words. Leaving aside “all” for the time being, the “law” is the translation of Japanese 法 (hou), which we know comes from the Chinese, which we know is a translation of the Sanskrit dharma. So we can translate this as “all dharmas”, and the Buddha-law part as “Buddha-dharma”. Now we get:

When all dharmas are Buddhadharma

which actually looks like real live English as written by a real Zen master and seems really deep too. It’s a translation. But do you get that? Or not? Maybe it’s because you still cannot understand the words. This is yet another, fifth assumption—that such words are meaningless to the typical English reader. I’m assuming the reader cannot understand “dharma” and that therefore I need to do something about it. This assumption should be no more controversial than the first four, but yet it’s one which many scholars refuse to make.

Of course, many knowledgeable Buddhists in the West will understand “dharma”, or at least think they do. So if I was translating for them, this might be an OK translation, Problem is, they represent about 0.001% of all the people in the world that are potentially interested in what this is saying. What about the other 99.999%? We don’t care about them? Or maybe they should just learn Buddhist terminology first? Dogen was writing for 99% of Japanese people. Is it OK that I should randomly decide that it’s OK to arbitrarily change Dogen’s assumption about his audience? No, I should also come up with an English translation that addressed 99% of people, like Dogen was.

To stop at this point and say that the “dharma” translation is “correct” or “loyal” or “strict” is a copout. It’s just being lazy. By doing so, I would be violating Dogen’s implicit audience assumption. Until I do something about “dharma”, I still don’t really have a “translation”, but a “trans-literal-posi-syntactico-lation”. Of course, if that’s all I’m trying to represent it as—fine. If I lack confidence in what Dogen might have been trying to say and am trying to cover my ass—fine. But do we really want “cover-your-ass” translations?

To make this into a true translation—an expression that maps to the mental images and behavioral impact of what Dogen said—we have to go deeper. This is where some people get cold feet, saying this is going beyond “translation” and entering the realm of “interpretation”. Come on. Every single person that reads Dogen’s words is interpreting them. It’s certainly not unreasonable to ask the translator, presumably well-informed, to participate in this interpretive process.

So to make “all-law” meaningful to Westerners, what should we do with the “law” (法 dharma)? Some Buddhist dictionaries list as many as several dozen meanings of the term. But it’s a fair guess that in this case the meaning is “phenomena” or “things”. So we have “many things”, which is indeed how Tanahashi translates this.

But what about “Buddha-dharma”? Tanahashi translates this as is—a major copout; he might as well have left it in the original Japanese characters. Cleary gives us “Buddha-teachings”, which seems to be going too far. Neither translation reflects the 法 common to 諸法• and 仏法. That would seem to be a major oversight. Whatever we want to do with Dogen, we should respect his style, and in this crucially important first phrase of the first sentence in the first chapter of his magnum opus Shobo Genzo he explicitly chose wordings which shared the word 法 (hou, dharma). Certainly this is something we should respect.

Are we there yet? Certainly, Dogen intended for his writings to “mean” something, even if only for himself, as one Dogen scholar I recently met indicated the possibility of. “Mean” for whom? For the people that read them. Dogen was certainly sophisticated enough that he knew that a wide range of people would read his works, and presumably wrote them so that all could “understand” them. It would not surprise me if Dogen intuited that people from the 21st century would read his essays, including people from other cultures—given his experience in China, he was certianly aware of cross-cultural issues. Our duty, then, in translating Dogen, is to realize his vision and produce a version of his thoughts which is meaningful across centuries and cultures. And translating literally using terms such as “buddha-dharma” clearly fails that test.

Here in the West, we have the concept of “God”. No-one knows exactly what it means, but in a way everyone does. It refers to something external, if you prefer, or something internal, if you prefer, an unknown essence. This is precisely the sense of the “buddha” in Dogen’s “buddha-dharma” phrase. In other words, “buddha-dharma” refers to God’s law, or things of God. As such, that is exactly how it should be “translated”. That is why I insist that “shohou no buppou naru jisetsu” should be translated exactly as

when all things are God’s things

That is what Dogen “meant”. It is not “interpretive”. It is the precise expression of Dogen’s intent, to the extent possible, in modern English.

There is one additional step which is possible and desirable: to fine-tune the English style. Again, Dogen’s Japanese was beautiful, flowing, almost poetic. Presumably he adopted this style for a reason—to give his writings greater impact and make them more memorable. By producing clunky English and trying to pass it off as a translation of Dogen, we are denying this prominent aspect of Dogen’s prose.

One aspect of English is that it makes verbs play a more central role in the overall semantics, preferring sentences with an active feeling. In that spirit, in our final, stylistic step, we will move to a verb-centric, English-like syntax, while also reading a bit more into what Dogen is trying to say with his use of “jisetsu” (when):

Sometimes, God shows us a world of things…

Or, perhaps evne

This world of things in godly terms…

Now, we just have to go through this process for the remaining 999 phrases of Genjo Koan.

Nabokov on translation

Sunday, January 9th, 2005

Vladimir Nabokov is fond of literal translations—and he is certainly someone we should pay attention to.

The Nov. 7, 2005 issue of the New Yorker, in a great article by editor David Remnich entitled “The Translation Wars”, from which much of the content of this post is lifted, talks about Nabokov’s ideas on translating, from Russian to English in particular.

Personally, I concur wholeheartedly with Cervantes, who is quoted in the article:

Reading a translation is like looking at the Flanders tapestries from behind; you can see the basic shapes but they are so filled with threads that you canot fathom their original lustre.

I myself have viewed many such tapestries from the rear.

Now on to Nabokov. He says of all the sins of a translator

The third, and worst, degree of turpitude is reached when a masterpiece is planished and patted into such a shape, vilely beautified in such a fashion as to conform to the notions and prejudices of a given public. Ths is a crime, to be punished by the stocks as palgiarists were in the show-buckle days.

He goes on to say that when translating his intent is to provide the reader with a literal-minded “crib, a pony. And to the fidelity of transposal I have sacrificed everyhting: elegance, euphony, clarity, good taste, and even grammar.”

So he’s “transposing” now, instead of “translating”? He captured his philosophy of translation in a pithy poem, also published in the New Yorker, way back in 1955:

What is translation? On a platter / A poet’s pale and glaring head, / A parrot’s speech, a monkey’s chatter, / And profanation of the dead. / The parasites you were so hard on / Are pardoned if I have your pardon, / O Pushkin, for my strategem. / I travelled down your secret stem, / And reached the root, and fed upon it; / Then, in a language newly learned, / I grew another stalk and turned / Your stanza, patterned on a sonnet, / Into my honest roadside prose—/ All thorn, but cousin to your rose.

The noted critic Edmund Wilson (Wikipedia), Nabokov’s erstwhile friend, begged to differ. Reviewing Nabokov’s translation of Pushkin’s Onegin (Wikipedia ), he accused him of “bald and awkward language”, a desire to “torture both the reader and himself”, “sado-masochism”, “actual errors in English”, an “unnecessarily clumsy style”, “vulgar” phrases, inaccurate transliteration, a “lack of common sense”, and “serious failures of interpretation”. Hmmm. Sounds like some Dogen translations I know. (Later Wilson admitted that these criticisms were “more damaging” than he had intended.)

This tension between literal and interpretive translations pervades the entire translation world as it applies not just to novels but even computer manuals. (I don’t think I’d like my Japanese digital camera to come with a manual translated by the Nobel laureate author of “Lolita”.) But one can’t help feeling that the people having this war of words are missing the point—or at least failing to clearly state, or more likely understand, the groundrules: the nature and intent of translation itself.

I’m deeply interested in this discussion because I’ve translated, and am still translating, the 13th-century Japanese Zen master Dogen. The more I work on Dogen, the more I move toward the interpretive side. In the next post, I will go through a specific passage from Dogen as a means to shed light on the true nature of the interpretive decisions the translator must make, across the entire lexical/syntactic/semantic spectrum, and how they relate to the purpose and meaning of the translation.

William Safire's muddled orthographical Esperanto

Thursday, January 6th, 2005

Transliteration refers to conversion between phonetic alphabets. In Japanese I write ��, and I can transliterate this into the Roman alphabet as hana. In English I write cherry, and I can transliterate this into the Japanese katakana syllabary as �ェリー.

Transliteration tries to accomplish two quite different things. The first is to write a word in another alphabet so that when it is pronounced according to the rules of the language using that alphabet, it sounds as much as possible “like” the original word in its original language. The second is to provide a unique, bidirectional orthographical mapping from one alphabet (that I don’t know) to another (that I do). Among other reasons, when entering text into a computer this lets me use a keyboard mapping I am more familiar with. Even many Japanese prefer to input Japanese content using the Roman alphabet keyboard mapping.

(There is a third, less important goal in some transliteration systems: to reproduce structural aspects of the original alphabet. The example I’m familiar with is Japanese. The syllabary is organized into rows (vowels) and columns (consonants). The “ha” column contains ha, hi, hu, he, and ho. The “hu” sound is perceived by most English speakers as being closer to “fu”. Thus, a transliteration system which emphasized phonological fidelity would represent this syllable as “fu”, whereas one emphasizing source-alphabet structural integrity would represent it as “hu”. Does this problem exist in other transliteration systems?)

The above is just a basic introduction to transliteration; another is at Wikipedia. What motivated me to post about the topic is the horribly garbled discussion that recently appeared in William Safire’s column in the New York Times, our national newspaper of record.

Safire starts off on the wrong foot, revealing a weak understanding of the distinction between orthography and phonology, making absurd statements such as “The closest I can get in Roman spelling [he means English spelling] to the sound of [Putin’s] name is…”. He then lapses into bemusement at the fact that if for some unknown reason the French were to use the English-style transliteration of Russian President Putin’s name, it would come out sounding like the French word for “prostitute”, and so gee, that must be why they adopted their own weird transliteration. How confused this all is is analyzed in detail by our friends at Blogos.

In a follow-up article, Blogos expresses shock that “there are still people out there, writing columns in some of the most influential newspapers in the world, who think that computers and the Internet can only work with roman alphabets”.

But that’s not exactly where our famed pundit is confused, if you read his closing paragraph closely:

Here’s the problem for globocrats: most computer operating systems are based on the Roman alphabet, Maybe the United Nations will find a new raison d’etre (that’s ray-ZON DET-ra) in standardizing a system to encode Roman and Cyrillic letters and Chinese and Japanese characters to make them computer-friendly on all the world’s screens.

Now he’s started talking about “encodings”, something else he plainly does not understand. It turns out there is a widely-implemented encoding making all the world’s characters “computer-friendly”, called Unicode. Clearly Safire has no idea what is going on in multilingual computing, and one must certainly question his judgment in writing such nonsense in a national newspaper without a minute’s worth of checking. The problem is not that these characters cannot be displayed on “all the world’s screens”, since they can; it’s that, once displayed, they still cannot be read by people that don’t know the alphabets. He continues:

…For users of tomorrow’s Internet to accurately cross cultures, experts in phonetics and transliteration will first have to create and agree on a standard system.

Ignoring the fact that Safire now is confusing “cultures” with languages and writing systems, he’s apparently saying that there could be, or should be, some type of orthographical Esperanto that would magically meet the two conflicting objectives of transliteration systems: to be faithful to the original orthography while also being pronounced by native speakers of any world language, according to their language’s phonological rules, in a way which is close to the phonology of the original word. Sorry, all the “transliterati” in the world won’t be able to pull off that trick.

Only then will President Poutine get his real name back.

Bill, he doesn’t need his name back, he never lost it. It’s a Russian name written in Cyrillic. The French didn’t “take it away”, they just tried to write it in their alphabet so people can read it.

So much for our reigning language maven, of whom, I should add, I am a great fan and faithful reader of his column.

Kanji Topology

Sunday, January 2nd, 2005

Every Westerner exposed to Kanjis immediately senses their topological nature. But this inherent aspect of Kanjis is still not reflected in any fontographical computing model. Bob has now put on-line his unique, if dated, survey of research into models of Kanji topology (PDF, 612K).


Saturday, October 9th, 2004

Is “performant” a word? I came across it in an article about Microsoft Visual Studio .NET 2005:

C++ is the easiest language to use for native interop and is often the most performant.

Stamping out the loan-word disease in Japanese

Tuesday, June 29th, 2004

The “Foreign Loan-word Committee” has issued recommendations for replacing 33 common katakana-isms with “native” Japanese.

Thank God they backed off on some of their worst proposals, like replacing “online” with “kaisen-setsuzoku”.

Of their new proposals, I especially like “setsumei sekinin” for “accountability”. In other words, the Japanese view accountability as the question of who has to explain something.

A lot of the proposed replacements are to just use the obvious Japanese, such as “dougu” for “tool”. Ditto for replacing “stance” with “tachiba”, or “conference” with “kaigi”.

But that begs the question: why did people start using “tool” in the first place, when they already had “dougu”? That’s a critical question of linguistic philosophy which the grayhairs on the committee didn’t even try to answer. I know the answer. The centuries-old Chinese compounds have been rounded and smoothed like rocks in a river-bed by the forces of linguistic nature over time. The English words are young, agile, opinionated, angular, with a personality (make that PA-SONARITI). In that sense, they have a different semantic profile. Simply put, they mean something different. That’s why people started to use them and will continue to use them.

But what’s really weird is that what they’re proposing to replace the 30-year-old borrowings with are themselves borrowings into Japanese, just much older ones!

Distros and convos

Thursday, June 10th, 2004

The Japanese shorten long English loanwords by the simplest of expedients—simply chopping off the last half of the word. So “convenience” becomes “conveni”. That always seemed kind of crude to me, albeit cute in a way.

But now I’ve noticed a trend in English to do the same thing (instead of, or in addition to, the old acronym approach). Two recent examples are the teen-age “convo”, for “conversation”, and the geekian “distro”, for “distribution” as in a Linux distribution. And of course there’s the old stand-by “combo”, corresponding to the Japanese “kombi”.

Why is it that in English we tend to want to end these words with an “o” sound?

Any more examples out there?

Weird examples of spoken English grammar

Sunday, May 30th, 2004

Today on TV I heard the following:

That stuff was the guy who lived there before his.

This sounded very natural so it took me a second to figure out what was happening. The stuff belonged to the guy “who lived their before him“, and then to make that phrase possessive, it was easy enough to make the basic him to his transformation.