James Davis
General Comments for All Finalists
All six finalists did a good job of conveying the overall meaning of the contest passage, and we wish all of them success in their translation careers. This section will be devoted to specific words and phrases in the passage that proved to be difficult for multiple finalists.
Translators often wonder when to translate literally and when to adjust the “dictionary” meaning of a word or phrase to fit the context. One example is the word 関係 in this sentence from the first paragraph of the contest passage: 違いは、外界からの情報に対して発動する振る舞いの種類が圧倒的に多いことと、外界からの情報に対してどの振る舞いを発動させるのか、の関係も極めて複雑であるところにある。 The word 関係 most commonly refers to a relation, a relationship, or some type of connection. However, in this instance 関係 appears in the context of a question: “which action will a human take in response to information from the outside world?” E21 and E26 chose “links” and “link,” respectively, E45 opted for “relationship,” and E54 went with “correlation.” If we recognize that the question embedded in this sentence involves a choice among a large number of possible actions, we could consider this 関係 to be the “process” by which that choice is made. If so, it is reasonable that for humans such a process would be 極めて複雑, as stated in the text. (The fact that this sentence pertains to humans can be gleaned from the overall flow of information in the passage—i.e., the context.) The entire sentence could then be translated as follows: “The differences lie in the fact that for humans the number of different types of behavior we may choose in response to the information we receive from our surroundings is overwhelming larger than the number of different types of behavior available to insects and the fact that the process through which humans decide which behavior to adopt is extremely complex.” E07 took a similar approach for the second difference: “determining which course of action to take can be incredibly complicated.” E07’s approach focuses on a “determination,” rather than a “process,” but this approach is equally valid.
The second half of the final sentence of the contest passage (さらに言語道断である「機械が人の命を絶つ」ということが現実とならないため、まさに人類の知恵が試されている状況にある) is challenging for several reasons. First, translation of the word 言語道断, which appears twice in this sentence, requires some thought. Second, the relationship between 戦争 and 「機械が人の命を絶つ」ということ lies at the heart of this portion of this sentence, and that relationship must be expressed in precisely the right way. Otherwise, this sentence, which is the climactic sentence of the paragraph, will fall flat. In this regard, 戦争 should be understood as a general concept; the writer is clearly not referring to any specific war. The word 言語道断 carries a very strong negative meaning. Generally speaking, options such as “outrageous,” “inexcusable,” “abominable,” and “despicable” are all possible. The translator’s task is to select the option that is most suitable for the context—one that involves war and death. The presence of さらに before the second occurrence of 言語道断 indicates that 「機械が人の命を絶つ」ということ is an even greater manifestation of this attribute than is 戦争. Here are the translations presented by the six finalists:
E07: “..., whether these inexcusable machines capable of ending human lives become a reality in an already senseless war will surely be a test of humanity’s intelligence.”
E21: “..., preventing a future where war—already monstrous enough—becomes even more monstrous through machines being the ones that decide who lives and dies, rests on humanity's judgment.”
E26: “..., preventing the use of deadly autonomous weapons (and war, which should be avoided at all costs in the first place) is truly a situation putting human judgement and intelligence to the test.”
E27: “..., as the absurd horror of ongoing wars will only be further exacerbated if machines begin killing humans. This must be prevented; truly, we are at a point where the wisdom of humanity is being tested.3”
E45: “..., to ensure that the senseless possibility of machines taking lives does not become a reality in already senseless wars, it is precisely the wisdom of humanity that is now being tested.”
E54: “..., whether or not we can avoid absolutely unacceptable situations where ‘machinery kills humankind,’ during a war which is abominable itself, really depends on the wisdom of humanity.”
The portions thought to represent 言語道断 have been underlined. Given the context—war and the killing of human beings—words such as “senseless,” “monstrous,” and “abominable” are excellent choices. Each of these six translations captures a portion of the writer’s intended meaning. Another option would be the following:
“..., we find ourselves at a crossroads whereby the intelligence of humankind is being severely tested in order to avoid the occurrence during a war, which itself is fundamentally immoral, of an action in which a machine takes the life of a human being—an action that is even more immoral.3”
In this instance the presence of the word そもそも in the phrase そもそも言語道断である戦争 suggests that considering 戦争 to be 言語道断 is an idea that has been discussed and broadly agreed upon by a wide swath of society. Thus, translating そもそも as “fundamentally” seems appropriate. The issue at hand (i.e., the context) is the decision made by a machine to kill a human being. Rendering 言語道断 as “immoral” in this instance is consistent with widely held views on war in general and on the killing of a human being in particular.
Specific Comments on Individual Translations
Robin Griffin (E07)
One feature of E07’s translation was the addition of information that was not explicitly stated in the source text but was necessary in order to fully express the writer’s intended meaning. This is an important skill for any successful translator. For example, E07 began the second paragraph with the phrase “This principle of autonomy.” The source text reads この自律性, and several finalists simply expressed this phrase as “This autonomy” or “The autonomy.” However, it is clear from the context that in this paragraph the writer is discussing 自律性 in the general sense, not in the specific sense in which it was introduced in the previous paragraph. Whether we choose to call it a “principle,” an “attribute,” or a “concept,” recognition that in this paragraph the writer is thinking of autonomy in the abstract makes the translation more effective.
In the final sentence of the same paragraph the writer refers to 自律性や汎用性. Several finalists translated this portion simply as “autonomy and versatility,” without tying these general characteristics to anything in particular. E07 made it clear that the writer was referring specifically to “autonomy and versatility in AI.” Those final two words add value to the translation.
However, E07 missed a few significant points in the passage. In the middle of the first paragraph, after discussing the motivation for various human actions, the writer comments on a similarity between insects and humans: しかし、元を正せばすべては生きるためであり、生きるという目的を達成するために、状況に応じて適切と思われる行動を選択し続けているシステムという意味では昆虫も人も同じであろう. E07 translated this sentence as follows: “Indeed, the human and insect are similar in that their actions can be boiled down to the desire to live, and in order to do so are equipped with a perpetual system that chooses what is considered to be the most apt course of action for a given situation.” The phrase “a perpetual system that chooses” is troubling. E07 seems to have treated 選択し and 続けている as two distinct actions. However, 選択し続けている should be regarded as a verb that describes a single action that is repeated again and again. Also, the first clause in this sentence refers to the actions listed in the previous sentence, which focuses only on human behavior. The remainder of this sentence encompasses both insects and humans. If so, this sentence could read, “However, if we look deeply into human behavior, all of those actions are taken in order to continue living. Both insects and humans function as systems whereby time after time the action that is thought to be most suitable for the situation at hand, in order to achieve the goal of remaining alive, is the action chosen.”
In the final sentence of the first paragraph the writer considers the relationship among certain automated devices (such as the Roomba vacuum cleaner and the AIBO robot dog), insects, and certain other animals. The second half of the sentence reads, 大雑把に見れば昆虫もこれに含まれるものの、学習する動物などは自律性をもつシステムということになる. E07 stated, “These automated machines, however, might be roughly categorized closer to the insect rather than autonomous animals that are capable of learning.” E07 identified two categories—insects and autonomous animals—and placed the automated devices closer to one category (insects) than the other (autonomous animals). However, the writer’s use of the particle も following 昆虫 indicates that “insects” is the topic of the first Japanese clause. The category for this topic is これ, which E07 correctly understood to refer to automated devices. There are, indeed, two categories, but the focus is on insects, which the writer suggests lie closer to one category (automated devices) than the other category (autonomous animals). In essence, E07 has the relationship between insects and automated devices reversed. Also, the writer uses the word システム to describe the second category. (It is worth noting that in this instance the writer regards these animals themselves as systems.) If we put everything together, this portion of the sentence could read, “Broadly speaking, insects fall into the same category as these automated devices. In contrast, animals that possess the ability to learn can be regarded as autonomous systems.”
In the second paragraph of the passage the writer discusses Japan’s potential for research related to AI. Specifically, the writer refers to Japan’s 確固たる立ち位置. E07 stated that Japan has the potential to “stand at the forefront” of future AI research and development. This is an overstatement. Phrases such as “solid position” (E21) or “firm position” (E45, E54) better express the writer’s intent.
Scott Wilson (E21)
E21 did a good job on the final sentence of the passage, as mentioned in the general comments, and accurately described the relationship (among insects, automated devices, and autonomous animals) that was presented in the final sentence of the first paragraph. (This sentence was discussed in detail under the comments for E07.)
However, E21’s translation includes a few awkward phrases. In the discussion of the differences between the behavior of insects and that of humans, the writer identifies one of the features of human behavior: 外界からの情報に対して発動する振る舞いの種類が圧倒的に多い. E21 expressed this as, “humans have an overwhelming amount of possible actions that can be taken in response to external stimuli.” Actions are countable, so “an overwhelming number of possible actions” would be a better choice.
In the final sentence of the second paragraph the writer compares the R&D efforts of Japan with those the U.S. and China: そして、特に米中に大きく研究開発において遅れをとってしまっている機械学習分野と異なり. E21 rendered this portion as follows: “While Japan is particularly lagging greatly behind the U.S. and China in terms of research and development in that field, ....” E21 placed three adverbs in a row (“particularly lagging greatly”), which makes the sentence a bit clumsy. In fact, 特に is associated with 米中, while 大きく modifies 遅れをとってしまっている. Thus, these modifiers should be placed separately in the English translation. Based on this analysis of the sentence structure, this portion could read, “With regard to R&D in the field of machine learning Japan has fallen far behind, particularly in comparison with the U.S. and China.”
The next clause in the same sentence reads, 自律性や汎用性については現時点では日本が立ち後れているという状況にはないと認識しており. E21’s translation reads: “at present I have identified that there is no such lag when it comes to autonomy and versatility.” The writer uses the verb 認識しており, but there is no evidence that the writer is expressing his personal view. In section 4 (page I-50) of this paper the writer uses the phrase 私見であるが when he injects his own opinion into the discussion. He does nothing like that in the sentence at hand. Thus, we may assume that 認識しており represents the consensus view regarding R&D in this field. Options such as “at present there is no such lag” or “Japan does not currently lag behind other countries” would be adequate. Also, as mentioned in the comments for E07, the final portion of this sentence should read, “autonomy and versatility in AI.” Although the writer does not say so explicitly, everything in this paragraph is focused on AI. In this context any discussion of a lag is only relevant to R&D on AI.
After introducing the term “LAWS” in the final paragraph, the writer explains, 文字通り、自律型 AI を搭載した兵器が、自らの判断で最後のトリガーを引く兵器の開発である. E21 expressed this sentence as follows: “As the name implies, this is the development of weapons equipped with autonomous AI that uses its own judgment when deciding to pull the final trigger.” Use of the term “final trigger” suggests that there are multiple triggers, but that is not what the writer means. In this context “最後のトリガーを引く” indicates that “pulling the trigger” is the final step in the process of deciding what (if any) action the weapon should take in the situation at hand. If so, the translation could read, “Just as the name indicates, a weapon—such as a LAWS—that incorporates autonomous AI exemplifies the development of a weapon that takes the final action—pulling the trigger—on the basis of its own judgment.” Including the name “LAWS” in the translation retains the focus that the writer clearly intends. In the general literature the term “LAWS” is sometimes regarded as singular and sometimes regarded as plural. The inclusion of the article “a” in the translation indicates that LAWS is intended to be singular in this instance.
Ayami Kan (E26)
E26 generally captured the writer’s intended meaning, and correctly expressed several points that caused difficulty for other finalists. However, this translation contains a number of awkward phrases or incorrect word choices that interfere with the effectiveness of the overall translation.
In the fifth sentence of the first paragraph E26 included the phrase, “even humans are the same of insects.” Surely, E26 meant to write “are the same as insects.”
In a discussion of human behavior the writer states, 空腹だから食べる行動を発動させ. E26 wrote, “we are driven to act because of hunger.” The source text clearly indicates that the specific action is “eating,” rather than, for example, “hunting” or “shopping”—both of which are possible actions in this context—but that information is absent from the translation.
In the eighth sentence of the first paragraph the writer describes some of the advantages that humans possess, relative to insects. This sentence reads, 昆虫は予め生得的な行動ルールに基づく反応しかできないが、われわれは教えられることによる学習や、自ら経験することで学習を通して、蓄えた経験・知識を活用することで、反応の仕方自体を動的に変更するだけでなく、振る舞いのバリエーション自体も動的に増やすことができる. E26 expressed this information as follows: “Insects can only react according to their innate behavioral rules. In contrast, by making use of what we learned from experience and the knowledge we acquired from others, we are able to both dynamically change the way we respond and create variation in our behaviors.” Most of this sentence is very well done. The issue here is the final clause. It is certainly tempting to translate the noun バリエーション as “variation.” However, in this instance the verb that goes with バリエーション自体 is 増やす. This suggests that the writer is describing an increase in the “range” of human behavior over time. If we look back at previous clauses in this same sentence, the writer tells us that this increase results from 蓄えた経験・知識を活用すること. Thus, as humans gain knowledge and experience, we develop a wider range of behavior, from which we may select a particular action in a particular situation. (This expansion in the number of options over time is distinct from variation or rotation among a static number of options.) In contrast, the writer makes it clear that insects are limited by “innate behavioral rules.” There is no corresponding increase in the number of behavioral options available to them. Based on this understanding, this sentence could read, “Insects can only respond on the basis of the rules of behavior with which they were born. However, we humans can not only change—in real time—the type of behavior with which we respond, we can also increase—in real time—the number of different types of behavior at our disposal. We accomplish these tasks by making use of the knowledge and the experience we accumulate through the process of learning. This learning could be based on what we were taught by others or it could be based on what we have experienced in our own lives.” It is also worth noting that the adverb 動的に appears twice in the source text. Thus, we must include corresponding language (“dynamically” or “in real time”) with respect to both changes in the type of response and changes in the number of different responses.
E26 attempted to describe “AIBO,” which was a very good decision, but E26 referred to AIBO as a “dog robot.” The standard description for AIBO is “robot dog.”
In the first sentence of the final paragraph E26 stated that “R&D on autonomous AIs are being actively pursued ...” It is true that research and development are two distinct activities, but the abbreviation “R&D” is normally treated as singular. Use of the word “AIs” may have been a reference to different forms of AI, or it may have been a typo. In any case, the writer treats “autonomous AI” as a single concept, so this portion should read, “R&D on autonomous AI is being actively pursued.”
In the discussion of LAWS E26 stated, “As the name implies, LAWS are weapons guided by autonomous AI that, based on its own judgement, has the ability to release its own trigger.” The final clause is confusing, and one piece of information is missing. In the context of firing a weapon, the reader expects to encounter some variation of “pull the trigger.” (The release of a trigger normally happens after the weapon has been fired.) Also, the two occurrences of “its” in the English sentence refer back to “AI,” but AI does not possess a trigger. Rather, it is the LAWS that is equipped with a trigger (or some other firing mechanism). The idea that “pulling the trigger” is the final step in a decision-making process—as indicated by the presence of 最後の in the source text—does not appear in the translation at all. The translation could read, “Just as the name indicates, a weapon—such as a LAWS—that incorporates autonomous AI exemplifies the development of a weapon that takes the final action—pulling the trigger—on the basis of its own judgment.” In the general literature the term “LAWS” is sometimes treated as singular and sometimes treated as plural. The inclusion of the article “a” indicates that LAWS is intended to be singular in this sentence.
Ciarán Doyle (E27)
E27 included several well-crafted phrases (e.g., “sensory input” and “tactile sensors”) and was the only finalist to include the footnote at the end of the final sentence of the passage. However, E27 made significant errors in conveying the content. The focus in the second paragraph lies on the relative position of Japan with respect to R&D on two different generations of AI. The writer states that Japan lags behind the U.S. and China with respect to machine-learning-based AI but does not lag behind when it comes to autonomy-based AI. Unfortunately, E27 reversed these two points. E27 rendered the final sentence of this paragraph as follows: “In contrast to the field of machine learning, where Japan has the lead over the major research and development work (of the US and China in particular), autonomy and versatility are areas where we must recognize that Japan is lagging behind the others. Japan must act to secure its place at the head of future AI research and development.” Mixing up these two points would be a fatal error in the world of commercial translation. In this discussion the attributes of “autonomy and versatility” are only relevant to R&D on AI, and that fact should be stated in the translation. In connection with Japan’s ability to secure a solid position (確固たる立ち位置) with respect to future AI-related R&D the writer also states, 可能性があると考えている. E27 wrote that “Japan must act” to secure its place “at the head of” such R&D. The first point is a misstatement of the writer’s wording, and the second point is an overstatement of the writer’s wording. (E27 may have attempted to make the content of the second sentence consistent with the content of the first sentence. Unfortunately, the content of the first sentence is incorrect. Forcing the second sentence to be consistent with the first sentence compounded the damage.) This portion could read, “With regard to R&D in the field of machine learning Japan has fallen far behind, particularly in comparison with the U.S. and China. However, Japan does not currently lag behind other countries with regard to attributes such as autonomy and versatility in AI. Thus, it is certainly possible for Japan to secure a solid position with respect to future R&D related to AI.” If we consider these three English sentences as a group, the content of the second sentence represents a contrast to the information presented in the first sentence, and the content of the third sentence is a logical outcome of the assessment presented in the second sentence. Thus, these three sentences form a consistent package.
The third example in the writer’s discussion of human behavior reads, 人と会うために移動する. E27 translated this portion as, “We act in accordance with the people around us.” The verb 会う typically means “to meet” or “to encounter.” The phrase “in accordance with” seems out of place. E27 may have been thinking of the verb 合う, but the writer uses 会う. The verb 移動する generally signifies motion, movement, or a change of location, but E27 seems to have translated this verb simply as “to act.” Better options for this sentence include phrases such as the following: “traveling to meet with others” (E07); “we move somewhere in order to meet with someone” (E45); “we go somewhere in order to meet someone” (E54).
In the last clause of the final sentence of the first paragraph the writer states: 学習する動物などは自律性をもつシステムということになる. E27’s translation reads, “The result is that we classify only organisms capable of learning as having autonomous systems.” From the source text it is clear that in the writer’s view such organisms do not “have” autonomous systems. Rather, for this writer such organisms “are” autonomous systems. (Thinking of an organism as a system may not be obvious, but that is the approach adopted by the writer throughout the article.)
Catherine Xu (E45)
E45 handled a number of concepts very well (e.g., “genetically-encoded behavioral rules” and “make the final decision to pull the trigger”). However, E45 stumbled in connection with the writer’s thoughts about automated devices, insects, and autonomous animals. These thoughts appear in the final sentence of the first paragraph. As mentioned in the comments for E07, the focus here is on insects, which the writer suggests lie closer to one category (automated devices) than the other (autonomous animals). In this half of the sentence the writer states, 大雑把に見れば昆虫もこれに含まれるものの、学習する動物などは自律性をもつシステムということになる. This occurrence of the pronoun これ refers to the noun 自動機械 (“automated devices”), which appeared in the first half of the same sentence. This is the category into which the writer places insects. The second category, 自律性をもつシステム, is the category for 学習する動物など. The use of ものの highlights the contrast between the two categories. E45’s translation of this sentence reads, “Under this criterion, however conscious AI robots such as Roomba vacuums or Sony’s AIBO dogs may appear to be, they are automatic machines, and it is animals that learn, and roughly speaking, insects too, that fall under the umbrella of autonomous systems.” The first half of this sentence is well done, but the second half is problematic. E45 may have assumed that insects and animals that learn would reside in the same category, but that is not what the text tells us. Recognizing that for this writer insects have more in common with automated devices, the second half of this English sentence could read, “Thinking in very broad terms, insects fall into the same category as these automated devices. In contrast, animals that possess the ability to learn can be regarded as autonomous systems.”
With regard to Japan’s place in the world of R&D related to AI, E45 stated, “Japan may be lagging far behind the US and China in terms of research and development in the field of machine learning, but at present, this is not at all the case in terms of autonomous intelligence and artificial general intelligence.” The corresponding portion of the source text reads, そして、特に米中に大きく研究開発において遅れをとってしまっている機械学習分野と異なり、自律性や汎用性については現時点では日本が立ち後れているという状況にはないと認識しており. E45 is correct to tie both 自律性 and 汎用性 to AI, but the terms “autonomous intelligence” and “artificial general intelligence” miss the mark. If 自律性 is the noun “autonomy,” then 汎用性 indicates that the resulting AI can be used or applied in a wide variety of situations. “Versatility” would be a good choice. Taking note of the presence of 特に (“in particular” or “particularly”) near the beginning of the sentence, this portion could read, “With regard to R&D in the field of machine learning Japan has fallen far behind, particularly in comparison with the U.S. and China. However, Japan does not currently lag behind other countries with regard to attributes such as autonomy and versatility in AI.”
Several of E45’s sentences are confusing. The second sentence in the second paragraph reads, “However, it involves different AI technology to machine learning.” E45 may have understood the content, but the grammar is awkward. It would be better to say something like, “However, it involves AI technology that differs from machine learning” or “However, it involves AI technology that is distinct from machine learning.” At the beginning of the third paragraph E45 wrote, “Of course, Japan is certainly not the only country with their attention on ...” If the writer had referred to “the Japanese people,” the use of “their” would be fine. However, Japan is a single country, so the underlined portion should read, “with its attention on.” (These days many people use “they” or “their” to avoid gender constraints, but that is not an issue here.) In connection with the discussion of LAWS in the third paragraph, E45 noted, “there already exists weapons ...” The word “weapons” is a plural noun, so this portion should read, “there already exist weapons.” Paying close attention to subject-verb agreement is important, particularly in a formal document, such as a journal article.
Mizuki Kitamoto (E54)
E54 demonstrated excellent comprehension of the source text (e.g., “the autonomy level of insects is quite close to automation”) but struggled in some places to express the resulting information smoothly. For example, in the first paragraph the writer compares the goal of insects with that of humans: 人の場合であっても目的は「生きること」であることは昆虫と同じである. E54 stated, “And humans are the same as insects in that their goals are ‘to live.’” It is true that two groups are mentioned in this sentence, but there is only one shared goal. Thus, the underlined portion should read, “their goal is to live.” (If the writer had identified two shared goals, a phrase such as “their goals are to live and to reproduce” would be fine.) (“Scare quotes” are not needed in English.)
Later in the same paragraph the writer discusses differences between the behavior of insects and that of humans. E54’s sentence reads, “Differences lie where ... and where ...” E54’s translation of the specific differences is very well done, but E54’s use of “where” is too colloquial. A better option would be, “Differences lie in the fact that ... and the fact that ...”
In the following sentence the writer uses the なればなるほど pattern to indicate a trend in a particular direction. E54 wrote, “The more ingenious reactions to the outside world get, ...” The use of “get” in this way is also too colloquial for a journal article. Replacing “get” with “become” would solve this problem.
In the first paragraph E54 included “I think that what we can call” to account for the expression と呼べるのではないだろうか, and in the second paragraph E54 used “I don’t think” to represent 。。。ないと認識しており. E54’s attempt to express the writer’s lack of certainty in the first instance is admirable. However, the phrase “I think” seems out of place in a journal article, and the resulting phrase is cumbersome. In fact, English is equipped with many “wiggle words” that can accomplish the same goal and maintain the appropriate tone and level of formality. For example, the phrase と呼べるのではないだろうか could be expressed with phrases such as “we could call ...” or “... could be called.” (The wording of the English sentence would need to be rearranged accordingly.) In the second instance it is not clear that と認識しており must be translated at all. As mentioned in the comments for E21, it is likely that認識しており represents the consensus view regarding R&D in this field, rather than the writer’s personal opinion. One option would be to simply ignore it: “Japan does not currently lag behind ...” If E54 really wanted to include verbiage to account for this verb, one option would be, “Japan does not appear to currently lag behind ...”
The final sentence in the second paragraph ends with 可能性はあると考えている. E54 began the translation of this sentence with a literal rendering: “I think there is a possibility that ....” Here we confront the same problem with the phrase “I think.” In this instance, however, the solution is different. Because the writer explicitly states that 可能性はある, the translator may ignore と考えている, and the translation could simply read, “There is a possibility that ....”
In the second sentence of the third paragraph E54 wrote, “Lethal Autonomous Weapons Systems (LAWS) is causing controversy in the field of autonomous AI, as deepfake is in that of machine learning.” Once again, E54 clearly understood the content. However, there are problems in the realm of expression. The term “lethal autonomous weapons systems” is plural, but the verb that follows this term is singular. Here, it is important to note that the text actually refers to the “LAWS (...) 問題.” If E54 were to rearrange the first clause to read, “The issue of LAWS (lethal autonomous weapons systems) is causing controversy in the field of autonomous AI,” the problem would disappear. Usage of the term “deepfake” vs. the term “deepfakes” in the general literature is a bit ambiguous. However, adopting the same strategy in the second clause—“just as the issue of deepfakes is in the realm of machine learning”—would resolve any potential tense issues and would add a sense of balance to the translation.