Ruth McCreery
To translate requires reading the source text, thinking about what it means, and writing to communicate the content of the source text in an appropriate manner, using the correct technical terms and writing in a suitable style. As Fred Uleman has said, read a little, think a little, write a little. Or think a lot: producing a translation as accurate and at least as compelling as the original can take time, especially when the text moves from insect behavior to machines making decisions in human warfare.
E21
E21 produced a largely accurate and readable translation. Engaging the reader by breaking up the first sentence was an excellent strategy. The description of insects’ actions as resulting from “selecting behaviors from a pool of innate behavioral sequences, following rules that activate certain sequences in response to input from external stimuli” is clear, and the use of “behavioral sequences” for 行動 系列 both is succinct and makes one visualize what is happening with the insect. Unfortunately, the following phrase (in parentheses) is both a sentence fragment and unnecessarily wordy. (“The rules are also innate.”)
A key issue in dealing with this text is to sort out the difference between automatic and autonomous, which E21 does well, translating 極めて自動に近い自律性を持つ生物 as “an organism whose level of autonomy is extremely close to automatic.” Then, sadly, the translator imitates the use of 鉤括弧 around 「生きること」by adding scare quotes: “survival.” Translating the punctuation is essential.
Surprisingly, the concepts in the last sentences of this paragraph have become rather muddled. I would take 意志をもっている as “having volition/a will of their own” rather than “consciousness” and 学習する動物などは 自律性をもつシステムということas “any being capable of learning is a system with autonomy” not “can be said to be a system of autonomy.”
The next paragraph contains rather clumsy and confusing English. In the first sentence, the final clause’s relationship to the rest is unclear. Making it a separate sentence would help. In the next sentence, “in that field” should logically refer to the last field mentioned, “machine learning,” as it presumably does, but the sentence could also be read to refer to “AI technology” in general. More precise English is needed.
The last paragraph has more awkward phrasings. “Progress is actively being made all over the world in this field, and as we've seen before with the issue of deep fakes arising along with machine learning, such has the issue of LAWS (Lethal Autonomous Weapon Systems) been identified along with autonomous AI technology” is tediously long, and “such has” probably should have been “so has.” The final sentence is, however, forceful and effective.
I suspect that if E21 had set the translation aside for another day before producing the final version, many of these problems would have been cleaned up, since the translator has a good grasp of both languages.
E07
E07 also produced a well written and largely accurate translation, with a few more outright errors and chunks of awkward English than did our first-place winner, but also some thoughtful solutions. This translator also adopted the strategy of breaking up the first sentence, but immediately introduced an odd phrase: “An insect’s goal is to live (or to keep living).” (Did E07 mean “to keep on living”?) “Equipped with” is good, but “a series of action patterns” implies some sort of organized relationship between those action patterns; that notion is not in the Japanese. The 系列 are the patterns (or sequences, to borrow from E21), not how they are organized. If it is necessary to add a collective noun for 系列, “set,” “group,” or E21’s “pool” would be less distracting. Describing the rules as “innate” is precise and succinct. Unfortunately, the next sentence left me wondering what the final “its” refers to. Similarly, with “a human’s goal of living,” I was not sure whether that should be “goal in living” (which would be a major divergence from the Japanese) or an unsuccessful attempt to avoid repeating “to live.”
“Can be boiled down to” is a great solution for 元を正せば、but “perpetual” seems an overstatement for 続けている (at least when talking about living creatures). “Survivability” for 生存確率 may be literally correct but skewed in nuance. We usually talk about survivability of a category or group under certain, usually hostile, circumstances; “chances of surviving” seems more appropriate in this context.
While E07 described Roomba and AIBO as having “a will of their own,” the following sentence loses the contrasts in the original and thus obscures the author’s point. The author is stating that while we think of them as having a will of their own, they are actually automatic machines, and, broadly speaking, insects are also included in that category. They are not autonomous. But a being capable of learning is a system with autonomy.
The next paragraph has only two problems. The second sentence is confusing due to the unnecessary “that of”; just “fallen behind the US and China” is sufficient. “Stand at the forefront” seems over the top for 確固たる立ち意置; “secure a firm position,” or even “be a competitive figure” would work.
The final paragraph is quite clean, except that “Similar to discussions” ends up modifying “many”; a rewrite is needed. The same sentence also, however, contains “denouncing” for 問題が指摘される--a smooth switch from passive to active and excellent word choice. Had E07 thought a little further and made wars plural or generalized it instead of specifying “an already senseless war,” the translation would have finished on a high note.
E26
E26 began painfully literally, with “In the case of insects,” followed by “their action.” (Why not “their actions”?) At the same time, the translator managed to eliminate the notion of insects’ acting to achieve a goal. The first paragraph continues by stating that insects are autonomous, if similar to automatic systems, rather than that their degree of autonomy is close to automatic. Awkward English follows. E26 has skipped rendering the significant contrast between insects and humans in the sentence that follows, and scare quotes are added to “survival,” paralleling the Japanese punctuation. “Analogous systems” is an excellent solution further down, and “system’s likelihood for survival,” if changed to “likelihood of survival,” would work.
E26 continues with largely effective English plus some awkward bits until the very end, where the translator has over-interpreted the text. E26 is of course right in implying that “preventing ... war” is an admirable goal, but the author’s stated concern is the more limited one of preventing the use of LAWS in war.
E45
E45 produced a strong beginning. “Behavioural repertoire” is a brilliant solution. But then we hit a sentence with grammar problems and confusing rendering of the Japanese: “Therefore, since the behaviour of these insects are [should be “is”] controlled by rules for [should be “for responding to”] input data detected by sensors, such as antennae, insects can appear to be autonomous organisms that are extremely close to being automated [what about “level of autonomy”?]. Similar problems continue down to the last sentence of the first paragraph, in which E45 puts insects in the same category as animals that learn instead of with automatic machines.
In the next paragraph, E45 introduces “artificial general intelligence” for 汎用性, instead of “versatility,” as well as some grammatical errors. The final paragraph turns around the relationship between LAWS and global research on AI; unfortunately, that research is not addressing the problem but causing it. The final sentence, however, is strongly and accurately worded.
E54
E54 ran with keeping the first sentence intact in the translation, not unsuccessfully, and also used the first person effectively. When we get down to “And humans are the same,” a little more literalness would not have hurt: これに対して is setting us up for a contrast, which the translator missed. “Differences lie where humans have a far wider variety” would imply that those differences exist only in some areas; probably true, but not what the author is saying. In that sentence, “correlation” seems a poor choice for 関係. In the next sentence, “higher survival probability” is a good solution for 生存確率, but “the better animals can adapt” is a slight miss. Adaptation usually refers to long-term, even evolutionary, changes in behavior. 適応度 suggests response, instead: “the better animals can respond.” The rest of the paragraph, while somewhat clumsy, proceeds well until the last sentence. Those animals do not have autonomous systems; they are systems with autonomy.
The second and third paragraphs proceed somewhat awkwardly, but without major problems, until the last sentence: “machines taking lives" has been happening for centuries; the new threat is machines choosing to take lives.
E27
This translator produced the most translation errors and the clumsiest prose. Many of the errors were shared with other finalists, but “We act in accordance with the people around us" for 「人と会うために移動する」was new, as was introducing “Only two factors separate our two species.” The text is indeed talking about two factors, but does not imply that they are the only differences. (Also, insects are not a species.) Another major error was reversing the lead and lag relationships between Japan and the US and China in terms of machine learning and autonomy and versatility.
Nonetheless, only E27 translated the footnote to the text; that attention to detail must be commended.