Barriers to Translation, English to Japanese

(Note: This document uses Unicode(utf-8) encoding.)

The objective of this document is to explain some of the unique problems encountered when translating from English into Japanese, and to introduce the approach taken by Passage Co., Ltd. in dealing with these problems.

When translating into Japanese, the translator is faced with unique challenges. The manner in which Japanese sentences are put together differs radically from that of Indo-European languages -- so much so that grammatical terms used to describe languages such as English have little relevance to Japanese. For example:

  • Japanese does not have definite or indefinite articles (a/the), and special measures are often required to convey the nuances of such articles
  • There are no plural forms of words in Japanese, and special measures are often required to indicate plurality. This applies to adjectives (my/our) and demonstrative pronouns (this/these, that/those), as well as nouns.
  • Japanese does not have equivalents of English prepositions (at/by/in/to/from/with). Instead, the role of prepositions is performed by inserting functional (but lexically meaningless) structure particles.
  • In Japanese, verbs come at the end of sentences, clauses, or utterances. Whereas in English the verb gives a clue as to a sentence's direction, in Japanese the clues must be provided by the structural particles.
  • Japanese grammar does not require that a sentence have a subject. Inclusion of a subject in every sentence would form a barrier to comprehensibility.
  • Japanese does not have a future tense, and massive restructuring is often necessary to produce meaningful equivalents to English sentences that use the future tense.
  • In many cases, Japanese does not have equivalents to English words. (This is not so much a problem in translations that are only concerned with pushing buttons, but becomes a major headache when the original English deals with abstract matters.)

The problems encountered do not end there.

For most languages, the goals of translation are to maintain equivalence between the source document and the translation in the following areas:

  • Semantic invariance- Preserving the meaning of the source text.
  • Pragmatic invariance- Preserving the intent of the source text. (The manner in which propositions are stated, conveying intent, urgency, politeness, etc.)
  • Structural invariance: Preserving as far as possible the syntactic structure of the text under translation.
  • Lexical invariance: Preserving a one-to-one mapping of words or phrases from source to target text.
  • Spatial invariance: Preserving the external characteristics of the text, such as its length, location on the page, etc.

The problem is that it is usually not desirable, and often not possible, to achieve the third, fourth, and (sometimes) fifth objectives when translating from English into Japanese.

We have already mentioned some of the issues that stand in the way of structural invariance. The problem with lexical equivalence is that translations attempting to preserve one-to-one correspondence between English and Japanese words are likely to read as gibberish. As an example, take the following English sentence:

Computers have not yet been applied to all the uses that will be found for them.

It is possible to produce a collection of Japanese word equivalents to this that have the form of a grammatical sentence. It might go as follows (with color used to show equivalence to the original English):

Computershavenot yet been applied to all the uses that will be found for them.

In producing this, the translator has:

  • Completely rearranged word order to produce a sentence that conforms to Japanese grammar
  • Added structure particles to indicate the sentence subject (computer), and the roles played by the words "all" and "applied" in modifying "uses".
  • Added the word "future" to indicate the original sentence's use of the future tense.

The problem with this translation is that it makes no sense in Japanese. A sense of how meaningless the translation is can be gotten by looking at this reverse translation, which was rendered by a popular translation program.

Computer is applied to no use which in the future is found for the reason yet.

In order to achieve semantic and pragmatic equivalence with the original English, the translator has to sacrifice lexical equivalence, as well as structural equivalence. Here is a Japanese translation of the above that might be produced by a more competent translator. Again, colors are used to indicate lexical equivalents between the Japanese and the original English.

Computers have not yet been applied to all the usesthat will be found for them.

Here, lexical equivalence between the source text and translation is reduced to just three words, and even one of those ("applied") has been transformed by changing it from a noun into an adjective.

Once again, reverse translation gives a sense of the nature of the transformation, with a result that reads something like the:

Turn off the power immediately if the device produces smoke, abnormal noises, or strange odors.
  • 煙が発生するとき
  • 異常な音がするとき
  • 異様な臭いがするとき

What all of this points to is that Japanese translators have to resort to a different mix of strategies than European translators.

European translators generally use a strategy consisting of two steps:

  1. Semantic analysis of the source sentence, and
  2. Syntactic transformation to produce a translation

Japanese translators generally must resort to a more complex process.

  1. Semantic analysis of the source sentence
  2. Parsing of the sentence into English units of meaning
  3. Translation of the individual units of meaning into Japanese
  4. Rearrangement of the individual translated units into a Japanese sequence
  5. Revision of the re-sequenced units to produce natural-sounding Japanese

Although European translators sometimes use the latter approach, because of the vastly different nature of English and Japanese, the Japanese translator must take the second approach much more frequently.

Given the challenges, even the most competent translators usually cannot produce a fully fluent translation in a single pass. In practical terms, the best we can expect is that they will produce a translation that is essentially free of errors, and which provides a solid base for review and revision. This is the work of our reviewers. They do it well, and make a better job of it than can be achieved by the translators themselves.