Commentaries from the Judges （Japanese-to-English Division）
Suggestions for All Contestants
Producing a good translation requires much more than language skill. The translator’s goal is to convey to the target-language reader the same meaning that the original writer conveys to the source-language reader. In order to produce a meaning-based translation the translator should look beyond individual words and focus on units of meaning. A unit of meaning may be a word, but it may also be a clause, a phrase, or a figure of speech. It may require fewer words or more words to express the same unit of meaning in the target language than it does in the source language. When working from Japanese into English it is quite common for the translator to add words that do not appear in the Japanese document, because Japanese writers often assume or imply information that English writers are expected to state explicitly. Deciding exactly how much information to include in a translation may be difficult, but this is an important skill to develop. The units of meaning may appear in a different order in the target language than they do in the source language. In many cases the translator will rearrange clauses or other units of meaning to present the information in a way that seems natural to the target-language reader. These are some of the challenges that confront the translator as (s)he works to produce a translation that is complete, accurate, and natural.
Another important element in translation is the concept of the “mirror image.” To the casual observer the translation should look as much like the source document as possible. Ensuring that the translation has the same paragraph structure as the source text is essential in order to produce a “mirror image.” Some contestants combined the first two paragraphs on page 6 of the source text into a single paragraph, and several contestants—including three finalists—combined the last two paragraphs on the same page into a single paragraph, and. Counting the number of paragraphs under each heading or sub-heading within a document is a convenient way to avoid this type of error. Another aspect of the “mirror image” is the treatment of headings and lists. If the source document includes numbers (such as “1” or “(1)” or “①”), the translation should include the same type of number that appears in the source document. The contest passage contains several sections, which are identified by different types of numbers. Some contestants used letters (“a” and “b”) or Roman numerals (“i” and “ii”) instead of “①” and “②.” The instructions for this contest indicated that it was not necessary to duplicate the exact formatting of the Japanese text, but making it a habit to duplicate the source-text formatting as closely as possible will be helpful in the world of commercial translation.
It is not uncommon for a translator to confront a portion of a document that could be understood in multiple ways. In such situations it is important to make use of all available information, including information that appears in other parts of the same document. It is tempting to focus like a laser on the words or the sentence at hand, but many ambiguities in a Japanese text can be resolved by remembering what content appeared in previous sentences or previous paragraphs. In some instances a sentence or paragraph that appears later in the document may help to confirm or reject a possible meaning for the portion in question. Content that appears in figures or tables may also provide a basis for deciding which among several possible meanings is intended by the writer for the current portion of the document. These are excellent reasons for reading--or at least skimming--the entire document before beginning a translation.
General Comments for All Finalists
This year’s passage dealt with autonomous driving, which lies on the technical horizon for the near future. All five finalists developed a good sense of the overall structure and purpose of this document, and all five translations are well done. Consequently, these comments focus primarily on specific words and phrases.
In many instances a translator needs to produce names for multiple categories of objects or actions. This document includes the names of several types of roads. Regardless of what a dictionary or glossary might suggest, the translator should employ names that when taken separately make clear the distinctions among the different categories but when taken together represent a coherent, logical, and natural-sounding set of terms within the given context. Here are the types of roads mentioned in the contest passage and the translations adopted by the finalists:
高速道路 expressway (E24/E34/E37/E40), highway (E35)
一般道路 general roads (E24/E35), local roads (E34), public roads (E37/E40)
主要幹線道路 main arterial roads (E24), major arterial roads (E34)
main roads (E35), major highways (E37), major roads (E40)
国道 national roads (E24), national highways (E34/E35/E37/E40)
地方道 local roads (E24/E34/E37), regional roads (E35/E40)
In the U.S. the term “expressway” often refers to a limited-access road that is found in or around an urban area. A limited-access road that connects one urban area with another could be a “highway” but may not necessarily be an “expressway.” For this document we need a general term that covers both a road within a single urban area and a road that runs between urban areas, so “highway” is probably the better choice for 高速道路. In this document an 一般道路 is not a 高速道路 but it encompasses 主要幹線道路, 国道, and 地方道. Keeping in mind these three sub-categories, a good option for 一般道路 would be “ordinary roads.” (These roads are “ordinary” in the sense that they are not “limited-access” and there are numerous intersections.) It is clear from the text that 主要幹線道路 includes both 国道 and 地方道. Thus, 主要幹線道路 could be “main thoroughfares” or “main roads,” 国道 could be “national roads,” and 地方道 could be “local roads.” (None of the last three categories could be “highways,” because “highways” are 高速道路, which is a separate category.) When selecting translations for any set of categories, the translator should keep in mind all of the categories that are mentioned in the document, so each individual term will provide the proper scope within the overall hierarchy—in this case, a hierarchy of roads.
Many translators struggle with the concept of 整備. The term 整備 is often associated with maintenance or repair—keeping a piece of equipment or a system in its original condition and avoiding deterioration. However, 整備 can also be associated with improvement or the provision of additional capabilities beyond the original condition of that same piece of equipment or system. The specific translation depends on the context. The word 整備 appears three times in the contest passage. In the second paragraph under the heading “一般道路における自動走行” we encounter the phrase “一部の整備された主要幹線道路” in the context of autonomous driving at Level 3. In this context this phrase refers to “certain main roads/thoroughfares that have been suitably modified.” “Specially equipped sections of main roads” (E35) is another good choice. Either option makes clear that these roads have been improved (in some way) beyond their original condition in order to make autonomous driving at Level 3 possible, and that is the intent of the Japanese writer. Translations such as “(roads) eligible for (autonomous driving),” “well-maintained (roads),” or even “well-developed (roads)” lack the sense of specialized “improvement” that accompanies 整備された in this context. The word 整備 appears twice in the final paragraph of the passage. In those instances 整備 refers to “improvements (in infrastructure)” (インフラ整備) or “enhancements/modifications (to the legal framework)” (法的な制度の整備), respectively.
The single sentence that constitutes the final paragraph of the contest passage presented the greatest challenge for all five finalists. When confronted with such a long sentence, the first task for the translator is to analyze the structure of the sentence. Once we have identified the main clause or clauses and we understand the relationships among the various clauses, we can decide how to arrange the clauses in English and we can determine whether it is better to present all of the information in a single sentence or to divide up the information between or among multiple sentences. The most important segment (i.e., the main clause) in the first portion of this sentence is “2020 年頃、社会ニーズが強い地域や経済性の成立し易い地域において、 レベル4、 レベル4の実現が見込まれ.” The phrase “その地域に必要なインフラ整備を行うことで、” and the phrase “法的な制度の整備に合わせて、” provide information that supports the main clause. These phrases explain how the 実現 of レベル4 will occur. There seem to be two different “forms” of Level 4: “実現が技術的に比較的容易な低速走行の移動サービスや無人宅配等のレベル4” and “事業成立性に鑑みた都市部の他の車両との混合交通下でも走行が可能な速度での移動サービスや無人宅配等のレベル4,” respectively. The final clause of this sentence (順次、レベル4 が可能な地域のエリアの広さや数を増やすことで導入地域が拡大していく見通しである) indicates that the total area in which service at Level 4 is provided is expected to increase gradually over time. In this clause it is important to recognize that the phrase “地域のエリアの広さ” refers to a single location where autonomous driving has been introduced, while the word 数 refers to the number of such locations. Now that we have identified the major “pieces” of this sentence and we understand the relationships among the various pieces, we are in a position to put those pieces together. The translation of the sentence could read as follows: “Around 2020 it is expected that level 4 will be implemented in those regions/locales with a strong/pressing social need for autonomous driving and in those regions/locales in which the economic value of autonomous driving is easily established/verified. This (implementation) will be accomplished through necessary improvements in infrastructure in those regions/locales and in accordance with enhancements/modifications to the legal framework. Level 4 will provide transportation/mobility services and driverless delivery services at a low speed; this is relatively easy to implement from a technical standpoint. Level 4 will also provide transportation/mobility services and driverless delivery services at a speed that will permit driving in traffic with other vehicles in urban areas, in recognition of the commercial viability of such services. It is expected that the total area in which such services will be introduced will gradually expand as both the number of regions/locales that are capable of implementing Level 4 and the size/area of each region/locale increase.” In this translation the main clause (of the Japanese sentence) appears in a sentence by itself. This sentence is followed by a separate sentence that includes the two phrases that were thought to contain “supporting” information. The next two sentences describe the two different forms of Level 4, and the final sentence contains the discussion of future growth. In this translation we consciously present the most important information first, so the reader will understand the context for the information that follows. By introducing the “big picture” first and then gradually presenting more detailed information, we ensure that the reader will not become lost in the details. This approach not only yields a translation that is more accessible to the English-language reader, it also allows the translator to confirm his/her understanding of the source text at key steps during the translation process.
Specific Comments on Individual Translations
E24 captured the intended meaning in most instances. However, there were a few awkward phrases.
The second paragraph under the heading “一般道路にける自動走行” includes the following text with reference to autonomous driving at Level 3 on suitable roads without turning: “システムによる作動継続が可能な限り、システムが安全運転に係る監視、運転操作を行い、運転者の一部のセカンダリーアクティビティを許容する.” It is clear that there is a constraint (システムによる作動継続が可能な限り), it is clear that the system will handle certain tasks (システムが安全運転に係る監視、運転操作を行い), and it is clear that the driver will be “permitted” or “allowed” to do certain things (運転者の一部のセカンダリーアクティビティを許容する). In the translation it is important to separate the constraint on the system from the actions of the system and the actions of the driver. Keeping this goal in mind, this portion of the text could read, “so long as continued operation of the vehicle by the system is possible, the system will monitor the driving to ensure safety and will operate the vehicle, thereby permitting the driver to engage in certain secondary activities.” Strictly speaking, the words in italics do not appear in the source text but have been included to improve the readability of the translation.
The following paragraph reads “その後、全道路におけるレベル2 の実現や一部のセカンダリーアクティビティを可能とするレベル3 の対象道路、対象車両の拡大が見込まれる。” In this sentence a distinction is made between expectations related to Level 2 (全道路におけるレベル2 の実現) and expectations related to Level 3 (一部のセカンダリーアクティビティを可能とするレベル3 の対象道路、対象車両の拡大). However, E24 shifted some information from Level 3 to Level 2. A more accurate translation would be the following: “In subsequent years it is anticipated that level 2 will be implemented on all roads and that the range of roads and vehicles suitable for the implementation of level 3, which will permit the driver to engage in certain secondary activities, will expand.”
E34’s translation was generally smooth, but E34 had difficulty with some specific words and phrases.
In the final sentence of the second paragraph of the passage (そのためには、走行環境の複雑性を車両側の性能により如何に上回るかが重要であることから、走行環境の複雑性とハード・ソフトの性能を類型化・指標化した上で、その組合せから、地域の抽出、必要な性能を定めて実現していく。) the word 性能 appears three times, the word 重要 appears once, and the word 必要 appears once. E34 correctly translated 必要 as “necessary,” but rendered 重要 as “essential”, rather than “important.” The phrase ハード・ソフトの性能 became “performance from both hardware and software sides.” The translation of 性能 is fine, but the inclusion of the word “sides” is unnatural. On the other hand, 車両側の性能 was translated as “vehicle efficiency,” rather than “vehicle performance,” and 必要な性能 became “necessary efficiency,” rather than “necessary performance.” In some situations the “performance” of a piece of equipment or a system may be closely tied to “efficiency,” but in this context 性能 is clearly “performance.”
The following paragraph (なお、レベル3 以上の実現性、時期については、更なる法的、技術的な議論が必要なため、記載は目安である。) is a bit tricky. E34’s translation reads, “The feasibility and the time schedule of Level 3 and higher autonomous driving is subject to change, as more legal and technological discussions are required.” Although this statement is probably true, that is not the intended meaning of the sentence. In this instance the word 記載 refers to the text and figures that constitute the remainder of the passage, and the word 目安 indicates to the reader that the content of the following text and figures is merely a rough estimate or a crude forecast of future events, rather than a detailed plan. Based on this understanding of these key words, a translation of this sentence could read, “With regard to the implementation and timing of autonomous driving at level 3 or higher, further discussion of legal and technical matters is necessary. Thus, the following represents (merely) one possibility/scenario.”
E35’s translation includes several well-crafted phrases. However, E35 stumbled over several patterns.
The first paragraph of the passage includes a description of 事業用車. E35 recognized that the “driving area and conditions can be controlled,” but omitted the phrase 企業側で (“by companies”). The fact that companies, rather than individual drivers, exercise such control with respect to 事業用車 is an important element in the distinction between 自家用車 and 事業用車.
The second paragraph of the passage includes the phrase “技術を制度やインフラで補いつつ.” This phrase appears in the translation as “through ongoing improvements of technology systems and infrastructure.” The verb 補う often means “to supplement (something) (with something else).” The particle を indicates that in this instance 技術 is the “something,” while the particles や and で indicate that 制度 and インフラ represent two kinds of the “something else.” (The 制度 could be the 法的な制度 that is mentioned in the phrase “法的な制度の整備,” which appears in the second line of the final paragraph of the passage.) Based on this understanding of the grammar, this phrase could be translated as “by supplementing/supporting technology with a (suitable) system and (suitable) infrastructure, ...”
The third paragraph of the passage (なお、レベル3 以上の実現性、時期については、更なる法的、技術的な議論が必要なため、記載は目安である。) contains the linking word ため. The word ため usually indicates that the preceding clause represents a cause, a reason, or an explanation for an action or statement that appears in the following clause, but in certain instances it can indicate a goal or a purpose. E35 rendered the sentence as “Furthermore, rough milestones are given here for the feasibility and timing of realizing level 3 and above in order to facilitate further discussions about the legal and technological aspects.” The presence of 必要な immediately before ため is a hint that the first clause represents an explanation—not a goal or a purpose. One possible translation would be this: “With regard to the implementation and timing of autonomous driving at level 3 or higher, further discussion of legal and technical matters is necessary. Thus, the following represents (merely) one possibility/scenario.”
E37 produced a very solid translation. Comments here are limited to particular words and phrases.
In the first paragraph of the passage E37 rendered 限定区画 as “restricted areas.” This translation may be too severe for the context. An alternative such as “defined areas” (E24), “limited areas” (E34), “a limited area” (E40) or “a limited domain” would be a better choice.
The first sentence in the second paragraph of the passage ends with the phrase “世界最先端を目指すことが求められる. This phrase appeared as “it is required to aim to be at the forefront.” This translation is very literal, but it is awkward. One way to avoid an awkward translation in English would be to insert a subject for the verb 目指す. If we adopt this approach, this phrase could read, “we must aim to be at the forefront” or “Japan must aim to be at the forefront.” Many translators and technical writers are reluctant to insert the word “we” into a technical document. However, this document was produced by a panel commissioned by the Japanese government, and it consists of a collection of forecasts and policy recommendations. The insertion of either “we” or “Japan” eliminates the need for a passive verb, thus making the sentence much smoother, but it does not detract from the impact of the sentence in the overall document.
E40 grasped most aspects of the passage very well but missed a few points.
The second paragraph of the passage includes the phrase “技術を制度やインフラで補いつつ.” This phrase appears in the translation as “supplementing technology and systems as things progress.” E40 handled the verb 補う correctly, but overlooked the word インフラ and lumped 制度 together with 技術, rather than treating 制度 as something to be considered separately from 技術. This phrase could be translated as “..., while supplementing/supporting technology with a (suitable) system and (suitable) infrastructure.” (The 制度 could be the 法的な制度 that is mentioned in the phrase “法的な制度の整備,” which appears in the second line of the final paragraph of the passage.)
A description of autonomous driving at Level 3 in the third paragraph under the heading “一般道路における自動走行” includes the clause “一部のセカンダリーアクティビティを可能とする.” E40 expressed this phrase as “allowing for some secondary activities.” In this translation it is not clear who or what will perform the secondary activities. However, in the previous paragraph the writer stated that these activities can—under certain conditions—be carried out by the driver. A translation such as “which will permit the driver to engage in certain secondary activities” more closely matches the intended meaning, even though the driver is not mentioned explicitly in this sentence.
All the finalists did a reasonably good job of making sense of a gnarl of bureaucratese. Among them, E40’s translation stands out for accuracy and readability, although the choice of “household vehicles” instead of the usual “private vehicles” to contrast with “commercial vehicles” was puzzling.
I was also surprised that E40 (and E24, E34, E37) translated 法的な制度 as “legal system.” The English term “legal system” refers to an overarching system for establishing, interpreting, and enforcing the law: civil law, statutory law, common law, or religious law, for example. My thought for 法的な制度 was something like “set of pertinent laws and regulations” until I read E35’s “framing legislation,” which is brilliant.
The use of “legal system” also made me realize that there are at least two types of regrettably literal translations. The ones that make no sense in the target language are painfully obvious. Here, however, we have an example that sounds like English and fits with the general tone of the text but fundamentally does not work: the silent but deadly literal translation.
Two of the finalists (E37 and E24) stuck with “car” instead of “vehicle,” even when it became obvious that the means of transport in question was not necessarily a passenger vehicle. Both also used “cars” in their titles. I found myself wondering if E37 and E24 had waded their way through the complexities of the second paragraph and on to the end without pausing to reconsider their earlier translation choices. It is easy for us translators to fixate on the details; we also need to learn to consider whether the solutions we have painfully worked out are effective at the paragraph and larger level. Titles can be a key indicator here: a title applied first thing, before developing a deeper understanding of the text, is often clumsy, lacking in nuance, or just plain wrong.
Japanese-to-English translation contestants once again faced an impenetrable morass of bureaucratic writing in the form of an excerpt from a report on government plans to implement automated driving. The difficulty was further compounded by the fact that the original contest passage was replaced when an English translation was found online in the middle of the contest period and a second contest passage was selected. But, alas, the replacement passage was also found to have an English translation after the contest deadline. However, the terminology and language usage in the “official” Ministry translation was so bad that contestants improved their translations by using other terminology and manners of expression. Unlike some other countries, Japanese copyright laws require the author’s permission for educational and other non-profit uses of copyrighted materials. The JAT Translation Contest therefore tends to use government publications. However, the Japanese government is providing English translations for an increasing number of websites these days. I would therefore like to express my admiration for all who tackled this obtuse piece of bureaucratic writing and perhaps even had to start their translations over.
All writing – original and translation – consists of brick and mortar. Obviously, the brick is the essential information that must be conveyed and the mortar includes the organization, transition elements, and vocabulary that make the information easy to understand and palatable (memorable?) to the reader. This year’s first and second place winners (E40 and E37) distinguished themselves by their use of mortar. E40 and E37 usually divided up long sentences and added subjects or direct objects to make the information readily understandable and pleasant-sounding in English. E40 transferred the meaning of the original text at an extremely high level of accuracy and E37 offered a tour de force in the use of English writing conventions.
Here are some considerations about the translations in general.
Some of the questionable vocabulary and English usage in the online English version of the contest passage highlight the need to carefully evaluate net resources. Online dictionaries are full of machine translations and English examples produced by native Japanese speakers. The English on Japanese websites is often written to be comprehensible to the Japanese website owner, but not the native English reader. These resources should be used with caution.
There was some general confusion about including footnote numbers, and the contest instructions probably should have been clear about that. Disembodied footnote numbers with no attached notes are not that useful. In a case like this with no communication with the “client,” it is generally good to provide everything in the text and let the client decide what to do with it.
Also, there is a definite automobile culture, science, and art in car-driving countries, and they have their own conventions and a definite terminology. For example, with so many SUVs and vans on the roads these days, the term “(private) passenger vehicle” is probably preferable to any type of “car,” private, general, or otherwise. Also, in the United States, at least, there are limited access expressways and freeways that are contrasted with unrestricted-access “surface streets.” On a surface street, there are intersecting streets and opposing traffic, and vehicles can collide with you from any direction. On a limited-access expressway, everyone is ideally going the same direction and there is no intersecting traffic. Terms like “general roads” and “public roads” do not sharply distinguish surface streets from expressways.
Here are some comments about the individual translations.
E40 was this year’s winner. E40’s translation was the most accurate and well written translation overall with 25% fewer errors in meaning transfer and English usage than the other closest finalists. E40 did a masterful job on Paragraph 2. The contestant chose to infer a subject (developers) for the verbs and flushed out the paragraph fully with transition words and phrases. E40’s editorial stylings were at times a bit too ambitious, when the contestant moved “at an early stage” too far away from “simple scenarios” in the course of rewriting. In Paragraph 2, E40 clearly said “the extent to which” performance exceeds complexity while several finalists said “how” rather than “how much” performance exceeds complexity. E40 used the favored term “driving task.”
Unfortunately, E40 chose to use “well-developed roads,” the wording from the online English version, for “整備された,” which would be better translated as “equipped with [an autonomous driving system, etc.]. E40 also called surface streets “public roads” and used the somewhat absurd term “household cars.” This term is used to indicate the number of cars per household, but not a type of vehicle. There were also a few omissions. Most readers would probably have knowledge of the subject, so things were perhaps spelled out more clearly than necessary.
E40 added “autonomous driving” to the numbered levels (Level 2 autonomous driving); although probably not necessary for the intended readership, it’s a nice touch. Admirably, E40 also did not feel the need to account for 対象 literally (targeted, intended, etc.) every time it occurred (e.g., 対象環境). Articles and verbs usually perform the function of 対象 in English. However, E40 may have over-edited this sentence, replacing the word “environment” with something else. The Japanese authors apparently added the term “作動継続” to the SAE levels of autonomous driving, so an official term for this was not available in the original table of SAE levels. However, E40 deftly translated this as “uninterrupted system operation.” The entire last paragraph of the contest passage was a single enormous sentence, but E40 nicely divided it into four sentences.
E37 was this year’s second-place finisher. E37’s translation was not as accurate as E40, but E37 went to great lengths to write readable English, especially in Paragraph 2. However, E37 may have provided a little too much mortar as there was some unnecessary repetition in this paragraph. E37 aptly translated “類型化・指標化し” as “categorized to serve as metrics.” The contestant translated “レベル2 の対象環境” as “range of circumstances subject to Level 2.” However, implementing autonomous driving is considered a positive thing that governments choose to do and “subject to” is defined as “likely or prone to be affected by (a particular condition or occurrence, typically an unwelcome or unpleasant one). After exerting so much effort to write readable English, E37 chose to break the unwieldy last paragraph into only two sentences.
Reading the translation independent of the Japanese text, E35 sounded like professional writing. E35 used the term “commercial vehicles.” “Commercial” is almost always preferable to “business” or “business-use.” And E35 had the best-sounding translation of “法的な制度の整備” with “framing legislation.”
Once again, considering how confusing and difficult this passage was to decipher and get everything straight, all finalists, E24 and E34 included, did an admirable job trying to place things in sensible and pleasant-sounding English order.