Another example of human/AI collaboration

93 views
Skip to first unread message

Tom Gally

unread,
Dec 20, 2024, 7:45:09 AM12/20/24
to hon...@googlegroups.com, swe...@groups.io
It’s been nearly twenty years since I left freelance translation, but it seems that translators and editors still encounter a problem that I often faced: a Japanese client revising an English text in bad ways. When this happened to me, I would do my best to explain why my version was better, but I was often frustrated by my inability to provide a good explanation. Just saying “my version is more natural” seemed like a cop-out and probably annoyed some of my clients. I was happiest when I could explain the grammar or vocabulary in a way that the client could understand and accept.

A few days ago, I faced a similar situation in my current in-house job. A Japanese text that I translated last month had been revised slightly, and a colleague came up with an English translation for the new portion on her own and asked for my opinion. I can’t give the example here, so let me just say that she had translated a Japanese verb using the first gloss that would appear in a Japanese-English dictionary. The problem was that the English verb normally uses both a direct object and a complement (“we do B for C,” where “for C” is the complement), no complement was present in the Japanese, the lack of a complement in the English made the sentence seem strange, and adding a complement to the English would have narrowed the meaning unnecessarily.

I couldn’t think of a better translation for the verb in that context and I would have had trouble explaining that complementation issue in Japanese, so I asked Claude for help. I gave it the full Japanese original text and the full English translation, I pointed out the verb that I didn’t like, and I asked for suggestions for a better one. It came up with a wonderful suggestion that was perfect for the context. I then asked it for an explanation in Japanese of why that verb was better than the one my colleague had proposed, and it produced a good explanation as well. I revised the explanation slightly and sent it to my colleague. Judging from her reply, I think she was happy with it  (「ありがとうございます・・・!大変勉強になります。それで行きたいと思います!」). I didn’t mention that I had been assisted by AI.

This experience was another case of productive collaboration between a human and AI. I made my native-speaker judgment about the naturalness of the English, Claude suggested a better solution than I could have, I evaluated the appropriateness of Claude’s suggestion and explanation, and I decided how to communicate the suggested change to my colleague. The result was a better translation and a better translator-client interaction than if it had been handled only by me or only by AI.

Since I can’t show you that example, I asked Claude to come up with similar examples of where a native English translator or editor might need to explain their language judgment to a Japanese client. Here are a few that it invented:

1. Context (JP): 従業員の安全を配慮した作業環境を整えています。
Initial translation: "We are creating a work environment that considers employee safety."
Better translation: "We are creating a work environment that prioritizes employee safety."

English explanation: "Consider" implies just thinking about something, while "prioritize" conveys active commitment to making safety a key focus. In business English, "prioritize safety" is a more standard and stronger expression.

日本語説明:「配慮する」の直訳として「consider」を使用すると、単に考慮するだけという印象を与えます。安全対策への積極的な取り組みを示すには「prioritize」の方が適切です。ビジネス英語では "prioritize safety" が標準的な表現として定着しています。

2. Context (JP): この製品は優れた耐久性を実現しています。
Initial translation: "This product realizes excellent durability."
Better translation: "This product delivers excellent durability."

English explanation: While "realize" (実現する) is commonly used in Japanese business writing, in English product descriptions, "deliver" or "offer" are more natural choices. "Realize" in English usually refers to becoming aware of something or making something real for the first time.

日本語説明:「実現する」の直訳として「realize」を使うのは一般的ですが、製品説明の文脈では不自然です。英語の「realize」は主に「何かに気づく」「初めて何かを現実のものにする」という意味で使われます。製品の特徴を説明する際は「deliver」や「offer」の方が自然な表現となります。

3. Context (JP): お客様の声を活かした商品開発を行っています。
Initial translation: "We are doing product development that makes use of customer voices."
Better translation: "Our product development is driven by customer feedback."

English explanation: "Customer voices" is a direct translation that sounds unnatural. "Customer feedback" is the standard term. Additionally, "driven by" is more dynamic and professional than "makes use of."

日本語説明:「お客様の声」を「customer voices」と直訳すると不自然です。「customer feedback」が標準的な表現です。また、「makes use of」より「driven by」の方が積極的で専門的な印象を与えます。

As you can see by the above, Claude sometimes fell into the trap of justifying its preferences as “natural” or “standard,” though it also was able to provide more substantive explanations. You can, of course, ask Claude to expound on its explanations, give more examples, avoid resorts to “naturalness,” etc., until you get an explanation that is acceptable to you.

Some of its suggestions of poor English were a bit comical, though:

4. Context (JP): 社内で情報を共有化することで、連携を強化します。
Initial translation: "We will strengthen cooperation by sharing-ization of information within the company."
Better translation: "We will strengthen collaboration through improved information sharing across the company."

I doubt if even the most obtuse client would propose “sharing-ization” for 共有化.

Tom Gally

P.S. I use Claude in this example, but the latest versions of ChatGPT and Gemini would probably have worked as well. I have been particularly impressed with the latest experimental releases of Gemini, available through the Google AI Studio.

Tom Gally

unread,
Dec 21, 2024, 4:28:09 AM12/21/24
to SWE...@groups.io, hon...@googlegroups.com
Thank you for the thoughtful comments, Lynne.

I agree about the importance of experience and judgment when using AI for practical purposes.

The same issue is a recurring theme in the debates I read online—usually on Hacker News—about whether AI-powered coding tools will put—or are already putting—software developers out of work. Many programmers new to the industry seem discouraged about their employment prospects, as AI is becoming able to write and edit code better and faster than they can. In contrast, senior software engineers often feel safe, because their jobs require more higher-level negotiation, planning, and coordination. Midlevel programmers seem at mixed minds about the usefulness and threat of AI.

A selling point for human translators, writers, and editors should continue to be that people respond emotionally to words in a way that AI does not. One hopes that some clients will continue to trust and pay for human judgment about the appropriateness and quality of texts, even if that human judgment is much more expensive than AI’s.

AI continues to advance, though. Just yesterday, Anthropic posted an informative piece about a recent trend in AI development: agents that plan and carry out complex tasks largely autonomously. In passing, the article mentions two applications relevant to us:

Examples where prompt chaining is useful:
Generating Marketing copy, then translating it into a different language. ...

Examples where evaluator-optimizer is useful:
Literary translation where there are nuances that the translator LLM might not capture initially, but where an evaluator LLM can provide useful critiques. ...


In the literary translation case, I expect that human translators’ sensitivity and readerly response to literary texts will enable them to resist the AI competition for quite some time, especially if they are able to link their names and personal identities to the published translations. AI composition and translation of marketing copy will be harder to compete with.

Tom Gally

John Stroman

unread,
Dec 21, 2024, 7:18:02 AM12/21/24
to hon...@googlegroups.com
Tom,

As I posted here yesterday, the article in Slator reviewing research on the suitability of LLMs in legal translation comes to this conclusion: "Given the sensitive nature of legal translation tasks, privacy and data security must also be addressed." I think by using the passive voice this statement (possibly translated from a European language into English)  says that "legal translations" (which I assume includes patents and other intellectual property documents) involve not only privacy and data security, but also the kind of accuracy that only a human translator (or perhaps, an experienced, field-specific MT/AI post editor) can provide.

My experience with GPT and Claude has shown that both AI engines tend to default to generic verbs such as 認識されている = is recognized, grasped, perceived, etc., based on frequency of use. However, the generic translations of many Japanese verbs, especially when written formally in passive voice, can have a wide variety of connotations that do not capture the essence of the author's intent in English 

You provided many examples in your previous post about prompting Claude to provide more context-appropriate, active voice alternatives to such generic translations, which does improve the translation in most cases. This may be helpful if you have time to spare, but it is a task that an experienced human translator who understands the context can perform almost instantaneously. In many cases, the human translator who knows the specific context can also do online research, especially of academic articles that have yet to be included in an AI database, to come up with the exact target-language terminology that the author/client may have used in the past.

I also think that the cultural/linguistic aspect of marketing copy clearly comes into play. From an academic standpoint (see the lesson on nuclear disarmament in Hibbett and Itasaka for example), persuasive arguments in Japanese have been described as funnel shaped, i.e., first providing all the background information and examples, and finally pulling them all together at the end to reach a logical conclusion, whereas English is an inverted funnel that begins with a topic sentence (a conclusion) and then provides examples and other supporting information, which may or may not be pulled together at the end with a closing sentence that echoes the topic sentence. Therefore, effective marketing in a target language must suit the proper order of presentation in the culture of the target language. I'm not sure whether AI can be prompted well enough to manage that, but a human translator who is a native speaker of the target language already has the ability to rearrange and present the message in the most effective way to reach the target audience.

Please understand that I am not trashing AI in any way. I use it myself every day for generating rough drafts. But I take the AI output with a large grain of salt because it frequently chooses the wrong term for the specific context, and I often have to edit my source language input because AI has trouble when continuing a topic from a previous sentence (or two, or three) if the Japanese author does not repeat the English equivalent of a grammatical subject. By its very nature, commercial translation is time sensitive, and the AI rough draft gives me a greater opportunity to check for accuracy online and rewrite as needed rather than spending my time on the busywork of typing my own rough draft. Does that make me an AI post-editor? I think not. Like translation memory software, AI is just another tool that improves my productivity.
 
 In short, we have to acknowledge that many authors in specialized fields are not good writers. I believe that my job as a translator involves first capturing authors' intent and then, if possible, making those authors better writers in the target language than in the source language. 

John Stroman

Tom Gally

unread,
Dec 21, 2024, 11:07:20 PM12/21/24
to hon...@googlegroups.com
John,

Thank you for the thoughtful comments. They prompted some thoughts from me. Apologies in advance for the length.

[P]ersuasive arguments in Japanese have been described as funnel shaped, i.e., first providing all the background information and examples, and finally pulling them all together at the end to reach a logical conclusion, whereas English is an inverted funnel that begins with a topic sentence (a conclusion) and then provides examples and other supporting information, which may or may not be pulled together at the end with a closing sentence that echoes the topic sentence.

In second-language writing pedagogy, this is called “contrastive rhetoric.” The term was coined by a linguist named Robert Kaplan in the 1960s. His original diagram showing supposed rhetorical patterns in various linguistic cultures ( https://scottthornbury.wordpress.com/wp-content/uploads/2012/04/kaplan-patterns-of-written-discourse.png ) was oversimplified and probably wrong—why is argumentation in English shown as a straight line, when often the conclusion comes at the beginning?—but it did help to focus people’s attention on different ways texts are organized in various languages.

Therefore, effective marketing in a target language must suit the proper order of presentation in the culture of the target language.

Yes. I had second thoughts yesterday after writing that translation for marketing is more susceptible to displacement by AI than literary translation is. I used to do advertising copywriting and translation for Japanese clients myself, and I felt that it was as challenging and creative its own way as literary translation. I guess yesterday I was thinking of more prosaic catalogue copy and the like.
 
I'm not sure whether AI can be prompted well enough to manage that, but a human translator who is a native speaker of the target language already has the ability to rearrange and present the message in the most effective way to reach the target audience.

One issue is whether the human translator has the remit to rearrange the message. In my experience, it depends on the purpose of the job and the relationship between the client and translator. When I was freelancing, I always preferred writer-driven jobs where I had the freedom to rewrite the text so that it flowed naturally and effectively in English, even if that meant reorganizing paragraphs or the entire text. But often, including in my present job, my task has been to produce a translation that reads and flows reasonably naturally in English but that follows the paragraph-by-paragraph, often the sentence-by-sentence, organization of the original Japanese. Most of my translations now are English versions of official statements and speeches and get published side-by-side with the Japanese; any major rearrangement would be taking unwarranted liberties with my employer’s official positions. I assume that much legal translation is subject to similar restrictions.

As to whether AI can be prompted to adapt the order of presentation and other cultural components of a text—well, let’s try some experiments. Here is yesterday’s 天声人語 column from the Asahi Shimbun:

くねりとした、柔らかい文字だった。石川啄木は、謹賀新年と毛筆で書き始めている。明治41年だから1908年、同郷の金田一京助に送った年賀状である。東京にある森鴎外記念館の特別展「111枚のはがきの世界」で、目にした▼「さすらひ来し北の浜辺の冬は寒く候、御無沙汰の罪は幾重にも」と文章は続く。21歳の啄木は北海道の寒風の地で、新聞社を転々としていた。その侘(わ)びしげな姿が、黄ばんだ一枚のはがきから、じんわりと立ち上がってくるかのようだ▼どうして手書きの文字というのは、こうも不思議な力があるのだろう。印刷された賀状でも、ペンで書き足された「お元気ですか」の一文があれば、差出人の顔が浮かぶ。達筆もいいし、悪筆もまた個性。温かみを持つ肉筆のよさである▼それだけに、先日の本紙の世論調査には、驚いた。郵便で出す年賀状の数を尋ねたところ、「出さない」という人が57%と、半数を超えたという。5年前には33%だったというから、かなりの駆け足で、紙の年賀状離れが進んでいるらしい▼年始のあいさつはSNSやメールで十分、という人は6割近くいた。18~29歳の若い層に限れば、8割だそうだ。はがき文化そのものが、このまま無くなってしまうのだろうか。寂しいなあ、とひとりごつ▼きょうはもう冬至。年賀状を書かれるという少数派の方は、お早めに。〈かの時に言ひそびれたる/大切の言葉は今も/胸にのこれど〉。啄木の歌が、妙に沁(し)みいる年の瀬である。

Like many such columns, its overall organization is 起承転結, the 起 section being from the start to 立ち上がってくるかのようだ; the 承 up to 肉筆のよさである; the 転 from there to とひとりごつ; and the 結 from there to the end. That structure, when maintained in an English translation, can be confusing to English-speaking readers who are used to other forms of textual organization. When I read those translations in the Asahi Evening News forty years ago, just as I was starting to learn Japanese, I found them confusing, too. I couldn’t figure out what the author was trying to say.

So let’s see what Gemini 2.0 Flash Thinking Experimental—released just a couple of days ago—can do with it.

First, a straightforward prompt:

Prompt: The Japanese text below is yesterday’s 天声人語 column from the Asahi Shimbun. Please translate it accurately into natural English.

Gemini:

The Curving, Gentle Script

The characters were curved and gentle. Ishikawa Takuboku began with “Happy New Year” written with a calligraphy brush. It was Meiji 41 (1908), a New Year's card sent to his fellow native, Kyōsuke Kindaichi. I saw it at the Mori Ōgai Memorial Museum in Tokyo, at the special exhibition “The World of 111 Postcards.”

▼ The message continues, “Having wandered, I find myself on the cold northern shore this winter, and I am deeply remiss for my silence.” The 21-year-old Takuboku was drifting between newspaper companies in the frigid lands of Hokkaido. His somewhat forlorn figure seems to rise up gently from the yellowed postcard.

▼ Why is it that handwritten words possess such a mysterious power? Even on a printed New Year's card, if the sender has added a single phrase in pen, like "How are you?", their face comes to mind. Skillful calligraphy is good, and less refined penmanship is also a unique characteristic. It's the warmth of handwriting.

▼ That's why I was startled by the recent Asahi Shimbun public opinion poll. When we asked about the number of New Year's cards sent by mail, over half, 57%, said they "don't send any." Five years ago, the figure was 33%, suggesting a rather rapid shift away from paper New Year's cards.

▼ Nearly 60% feel that New Year's greetings via SNS or email are sufficient. When limited to the young demographic aged 18 to 29, that number climbs to 80%. Will the postcard culture itself disappear entirely like this? "That's sad," I muttered to myself.

▼ Today marks the winter solstice. For those of you in the minority who still write New Year's cards, please send them early. 〈Words of importance / I failed to say back then / Still linger in my heart.〉 Takuboku's poem feels particularly poignant as the year draws to a close.

Let’s try giving it more freedom:

Prompt: The Japanese text below is yesterday’s 天声人語 column from the Asahi Shimbun. Please translate it so that reads like an essay originally written in English. Do not delete any significant information, but add brief explanations within the text where necessary to make it understandable and enjoyable to English-speaking readers not familiar with Japan or Japanese customs, such as those surrounding New Year’s cards.

Gemini:

The flowing, gentle script caught my eye. It was the handwriting of Takuboku Ishikawa, beginning a New Year's greeting in elegant calligraphy: "謹賀新年" (Kinga Shinnen), a formal way of saying "Happy New Year." The year was Meiji 41, or 1908 in the Western calendar, and this particular nengajo, as these Japanese New Year's cards are known, was addressed to his friend and fellow linguist, Kyosuke Kindaichi. I encountered it at a special exhibition titled "The World of 111 Postcards" at the Mori Ōgai Memorial Museum in Tokyo.

Below the auspicious greeting, the message continued: "さすらひ来し北の浜辺の冬は寒く候、御無沙汰の罪は幾重にも" (Sasurai kishi kita no hamabe no fuyu wa samuku sōrō, gobusata no tsumi wa iku e ni mo). Translated, it reads something like: "Wandering, I've come to the cold winter shores of the north, my apologies for this long silence are many." Takuboku, only 21 years old at the time, was moving from one newspaper job to another in the frigid landscapes of Hokkaido, Japan's northernmost main island. His somewhat wretched circumstances seem to subtly emerge from that single, slightly yellowed postcard.

It's remarkable, isn't it, the power held within handwritten words? Even on a printed New Year's card, if someone adds a simple "お元気ですか" (O-genki desu ka?), meaning "How are you?", penned by hand, the sender's face immediately comes to mind. Skilled calligraphy is beautiful, but even less practiced handwriting possesses a unique personality. That's the warmth inherent in the personal touch of a handwritten note.

That's why I was so taken aback by a recent public opinion poll conducted by this very newspaper, the Asahi Shimbun. When people were asked about the number of New Year's cards they send via postal mail, a staggering 57% responded that they send none at all. Just five years ago, that figure was only 33%. It seems the shift away from traditional paper New Year's cards is happening at a considerable pace.

Nearly 60% of respondents felt that New Year's greetings via social media or email were sufficient. When looking specifically at the younger demographic, those between 18 and 29 years old, that figure rises to a striking 80%. Is it possible that the very culture of sending nengajo, a tradition deeply embedded in Japanese society, will simply vanish? A touch of loneliness creeps in, I find myself thinking.

Today marks the winter solstice, toji in Japanese. For the dwindling number of people who still embrace the tradition of writing New Year's cards, it’s a reminder to get them sent soon. A tanka, a traditional form of short Japanese poetry, by Takuboku comes to mind: 〈かの時に言ひそびれたる/大切の言葉は今も/胸にのこれど〉 (Kano toki ni iisobiretataru / taisetsu no kotoba wa ima mo / mune ni nokoredo). This translates to: "Those precious words I failed to say at that time / even now / remain in my heart." As the year draws to a close, these words strike a strangely resonant chord, perhaps a reminder of the enduring value of heartfelt communication, whether spoken, written, or sent with a New Year's greeting.

Let’s try to cut it loose even more:

Prompt: The Japanese text below is yesterday’s 天声人語 column from the Asahi Shimbun. Please translate and adapt it so that it reads like an original literary essay in English. Present basically the same information as in the Japanese, but adapt it freely so that it is interesting and enjoyable to read in English. Do not use the 起承転結 organization of the original; rather, create your own overall structure for the essay in a way that would be familiar to English-speaking readers. You may also add brief explanations within the text where necessary to make the content understandable to English-speaking readers not familiar with Japan or Japanese customs, such as those surrounding New Year’s cards.

Gemini:

The Ghost in the Postcard: A Meditation on Connection in the Digital Age

The ink, a dark, unwavering line against the aged paper, possessed a peculiar grace. Ishikawa Takuboku, the poet whose life burned with the intensity of a fleeting spark, began his New Year’s greeting with the formal characters for "Happy New Year." The year was 1908, Meiji 41 by the Japanese calendar. This particular card, now part of a special exhibition at the Mori Ōgai Memorial Museum in Tokyo titled "The World of 111 Postcards," was addressed to his friend and fellow linguist, Kyosuke Kindaichi.

My gaze lingered on the photograph of the postcard, a window into a past both familiar and distant. Below the cheerful greeting, Takuboku's words unfolded: "Having wandered, I find myself on a cold northern shore this winter, my apologies for my long silence." He was just twenty-one then, adrift in the biting winds of Hokkaido, hopping between newspaper jobs. From that single, yellowed piece of cardstock, a faint yet palpable sense of his isolation, his precarious existence, seemed to rise and mingle with the hushed air of the museum.

What is it about handwriting that holds such power? Even with the crisp perfection of printed New Year’s cards – known as nengajo in Japan, a tradition akin to sending Christmas cards but exchanged at the start of the new year – a handwritten addendum, a simple "How are you?" scrawled in pen, conjures the sender’s face with startling clarity. Whether the script is elegantly calligraphic or endearingly awkward, the human touch, the physical act of inscription, imbues the message with a warmth that no digital font can replicate.

And yet. A recent survey conducted by the very newspaper whose editorial this is – the Asahi Shimbun – delivered a disquieting statistic. When asked about the number of New Year’s cards they planned to send by mail, a staggering 57% of respondents answered "none." Just five years ago, that number stood at 33%. It seems the practice of sending paper nengajo is fading with remarkable speed, swept away by the tide of digital communication.

Nearly sixty percent of respondents felt that New Year’s greetings via social media or email were sufficient. Among those aged 18 to 29, that figure soared to eighty percent. Is this then the inevitable twilight of the postcard? Will this cherished cultural ritual, deeply intertwined with the Japanese New Year, simply vanish? A touch of melancholy whispers the question.

Today marks the winter solstice, the shortest day of the year. For those among us who still embrace the tradition, a gentle reminder: the time to write and send your New Year's cards is now. As Takuboku himself wrote in a poignant tanka poem, a verse that resonates with particular force at this year-end juncture: “The crucial words I failed to speak / in times now gone / remain within my heart.” Perhaps the handwritten New Year's card, a tangible piece of connection sent across the miles, is one way to ensure those crucial words don't remain unspoken, lost in the ether of the digital world.


It didn’t rearrange it as much as I hoped it would. Let’s break the task into two independent prompts:

Prompt 1: The Japanese text below is yesterday’s 天声人語 column from the Asahi Shimbun. Please convert this column into bullet points, with each bullet point representing one fact, thought, or point mentioned in the column. Each bullet point should be in both Japanese and English. Ensure that your bullet points cover every point mentioned in the column. Label each bullet point according to its function in this essay, such as “Main point,” “Illustrative point,” “Side anecdote,” etc. Create your own categories for those labels. After you have prepared your bullet points, rearrange them in random order.

Gemini: (Came up with a list of labeled bullet points in random order.)

Prompt 2 in new chat: Write a first-person literary essay in English based on the bullet points shown below in Japanese and English. The role of each bullet point is mentioned in the labels before the bullet points. Create your own overall structure for the essay in a way that would be familiar to English-speaking readers. You may also add brief explanations within the text where necessary to make the content understandable to English-speaking readers not familiar with Japan or Japanese customs, such as those surrounding New Year’s cards.

Gemini: 

The Fading Ink of "Kinga Shinnen"

The yellowed paper felt brittle beneath my fingertips, a fragile relic from a bygone era. It was a reproduction, of course, but the sentiment resonated nonetheless. The card, sent in 1908, Meiji 41, bore the bold, brushed strokes of "謹賀新年" (Kinga Shinnen – Happy New Year), penned by the poet Ishikawa Takuboku to his friend Kyosuke Kindaichi. Seeing those characters, imbued with the weight of history and the personal touch of a struggling young artist, sparked a familiar pang of melancholy. I fear for the very future of this tradition, this tangible exchange of well wishes – the postcard culture itself. [TG: Note the conclusion mentioned at the beginning!]

Lately, this concern has solidified into something more akin to a quiet dread. The digital age, while offering unparalleled convenience, seems to be steadily eroding the simple beauty of handwritten communication. We’ve witnessed this shift acutely with New Year’s cards, or nengajo as they are known in Japan. These postcards, often adorned with zodiac animals or festive imagery, are traditionally sent to express gratitude for the past year and hopes for the new one. They are a ritual, a small act of connection that transcends mere obligation.

There’s a peculiar magic held within the strokes of a pen. Whether the characters are elegant and flowing, showcasing practiced calligraphy, or slightly clumsy and uneven, revealing the writer’s hurried hand, each is undeniably unique. Good handwriting and bad handwriting, both possess a distinct personality, a reflection of the individual who wielded the pen. This physicality, this tangible link between sender and receiver, imbues handwritten text with a certain warmth that digital fonts, however artful, can never replicate. The ink bleeding slightly into the paper, the slight indentations left by the nib – these imperfections are what make it perfect.

But the tide is turning. It seems the movement away from paper nengajo is happening at an alarming pace. Just the other day, our national newspaper, the Asahi Shimbun, published the results of a public opinion survey. A staggering 57% of respondents declared they would not be sending New Year's cards this year. This news hit with a quiet thud, a confirmation of a trend I’d sensed but hoped wasn't quite so stark. Compare this to a survey from just five years ago, where the "no send" group was only 33%. The acceleration is undeniable.

Returning to Takuboku’s card, the context of its creation amplifies its poignancy. The inscription continued beyond the formal greeting: 「さすらひ来し北の浜辺の冬は寒く候、御無沙汰の罪は幾重にも」 ("Having wandered and come to the northern beach, the winter is cold, and I apologize deeply for my long silence.") At just 21 years old, Takuboku was drifting through the chilly landscapes of Hokkaido, moving from one precarious newspaper job to another. That yellowed postcard, fragile as it is, conjures an image of a solitary figure, wrestling with hardship and reaching out across the distance to connect with a friend. His loneliness, his wabishisa, seems to rise from the paper itself.

Perhaps this yearning for tangible connection is becoming a minority sentiment. According to recent data, nearly 60% of people believe that New Year's greetings via social media or email are perfectly adequate. Among younger generations, those between 18 and 29, that number jumps to a startling 80%. They have grown up in a digital world, where instant communication is the norm. The delay inherent in sending a physical card, the anticipation of its arrival – these are not experiences they necessarily value.

Yet, even in our hyper-connected world, the human desire for genuine connection persists. I’ve seen it myself. Even on a mass-produced, printed New Year’s card, the simple addition of “お元気ですか” (O-genki desu ka? – How are you?) scrawled in pen brings the sender vividly to mind. That small, handwritten inquiry bridges the gap between formality and personal care.

"寂しいなあ," I mutter to myself. That’s lonely. Perhaps it’s the looming end of the year, the shiwasu bustle giving way to quieter reflection, but a particular verse from Takuboku comes unbidden to mind: 〈かの時に言ひそびれたる/大切の言葉は今も/胸にのこれど〉 ("The important words I failed to say then / Still remain in my heart.") This sentiment, this lingering regret for unspoken feelings, resonates deeply as the year draws to a close. Are we losing the opportunity for these heartfelt, tangible expressions, these carefully chosen words sent across distances?

Today, as the winter solstice arrives, marking the shortest day of the year, I find myself thinking of those who still value this tradition. To those of you who are part of this dwindling minority who still diligently write New Year’s cards, please do so early. Let your ink flow, carrying warmth and connection into the coming year. Let's not allow the personal touch to fade entirely, like the ink on an old, cherished postcard.


This new Gemini model also shows its “chain of thought” for each response. Here is what it was “thinking” when it produced the final essay above.

Here's a breakdown of the thinking process to construct the essay:

  1. Understand the Core Request: The goal is a first-person literary essay in English based on the provided Japanese and English bullet points. The essay should have a familiar structure to English readers and explain necessary cultural context.

  2. Identify the Central Theme: The recurring theme is the shift away from handwritten New Year's cards (nengajo) and the author's nostalgic feelings about it, particularly in the context of Ishikawa Takuboku's historical New Year's card.

  3. Determine a Suitable Essay Structure: A common structure for this type of essay is:

    • Introduction: Hook the reader, introduce the topic (declining nengajo), and subtly bring in the Takuboku anecdote.

    • Body Paragraphs: Develop the theme with evidence and personal reflection, incorporating the provided bullet points. Group related points logically.

    • Counterarguments/Nuances: Acknowledge the convenience of digital greetings.

    • Return to Core Theme: Reiterate the value of handwritten communication.

    • Conclusion: Summarize and offer a final thought, potentially connecting back to the opening or introducing a lingering emotion.

  4. Map Bullet Points to the Structure: This is the key step for organizing the essay.

    • Introduction: Start with the Takuboku anecdote (Historical Context) to immediately ground the essay in a tangible example. Introduce the author's concern about the disappearing postcard culture (Author's Concern).

    • Body Paragraph 1 (The Power of Handwriting): Focus on the positive aspects of handwriting (Positive Aspects of Handwriting, Reflection on Handwritten Text). Include the illustrative point about the handwritten addition to a printed card.

    • Body Paragraph 2 (The Shifting Landscape): Introduce the observed trend of moving away from paper cards (Observed Trend) and the statistical backing for this (Surprising Survey Result, Comparison to Previous Survey, Reason for Not Sending Paper Cards, Demographic Detail).

    • Body Paragraph 3 (Takuboku's Card and Connection): Detail Takuboku's condition (Description of Takuboku's Condition), the quote from his card (Quote from the Postcard), and the sensory detail/impression the card evokes. This section allows for deeper emotional connection and justifies the author's sentimentality.

    • Conclusion: Bring in the concluding thought/quote from Takuboku's poem, connect it to the author's feelings (Author's Reflection), and end with the call to action and seasonal reminder.

  5. Draft the Essay, Weaving in Cultural Context: As writing, ensure to explain elements like "Kinga Shinnen" and the significance of New Year's cards in Japanese culture. Use transition words and phrases to create smooth flow.

  6. Refine and Polish: Review for clarity, conciseness, and emotional impact. Ensure the first-person voice is consistent. Check for any awkward phrasing or logical gaps.

Self-Correction/Refinement During the Process:

  • Initial Thought: Maybe start with the survey results to grab attention.

  • Correction: Starting with the historical anecdote about Takuboku is more evocative and literary, better fitting the essay's tone. The survey results can be more impactful later as evidence.

  • Initial Thought: Separate paragraphs for each bullet point.

  • Correction: Grouping related bullet points creates a more cohesive and thematic flow within each paragraph.

  • Initial Thought: Focus heavily on statistics.

  • Correction: Balance the statistical information with personal reflections and the powerful imagery of Takuboku's card to maintain the literary feel.

  • Initial Thought: Briefly mention nengajo.

  • Correction: Provide a slightly more detailed explanation of what nengajo are and their cultural significance for readers unfamiliar with the custom.

By following these steps and iteratively refining the structure and content, the final essay effectively addresses the prompt and creates a compelling narrative around the decline of handwritten New Year's cards.


This was an interesting exercise! Thanks, John, for giving me the idea to try it.

Tom Gally

--
You received this message because you are subscribed to the Google Groups "Honyaku E<>J translation list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to honyaku+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/honyaku/CAJVTx96%2BJBoUPjw6wNdZ68wWdb7YEbYYD9QY2poHeyrY%2Br-f6g%40mail.gmail.com.

John Stroman

unread,
Dec 22, 2024, 11:43:19 AM12/22/24
to hon...@googlegroups.com
Tom,

Thank you for your timely and considerate response. For the patient readers of this thread, here is the same essay published in the English edition of Asahi Shimbun.

VOX POPULI: Time to post New Year’s greeting cards, if you cling to custom

In 1908, Takuboku Ishikawa (1886-1912), a poet born in Iwate Prefecture, sent a brush-written “nengajo” (New Year’s greeting card) to the linguist Kyosuke Kindaichi (1882-1971), a fellow Iwate native. “Best wishes for a happy new year,” Ishikawa began in a twisty hand with a gentle touch. I saw it at the Mori Ogai Memorial Museum in Tokyo, where a special exhibition titled “111-mai no Hagaki no Sekai” (literally, the world of 111 postcards”) is being held until next month. Ishikawa went on: “The winter is cold in the northern coastal town where my wanderings have brought me. I am guilty, many times over, for not writing sooner.” The poet was 21 years old at the time, drifting from one newspaper company to another in the northernmost main island of Hokkaido. It felt to me as if his sense of desolation was wafting slowly from the yellowed postcard. Why is it that handwritten words can be so powerful? Even with a bland, printed New Year’s greeting card, all it takes to make it intimate is just one handwritten line from the sender, like “How are you doing?” Fine handwriting is nice, but unformed scrawls are so full of personality, they are irreplaceable. Such is the charm of handwritten messages. And precisely for this reason, I was stunned by the results of a recent opinion survey by The Asahi Shimbun. Asked how many New Year’s greeting cards they were sending out, a majority of respondents--57 percent--said they “aren’t sending any.” Since the corresponding number was 33 percent five years ago, the move away from physical greeting cards must have progressed rapidly in just a few years. In the same survey, nearly 60 percent of respondents said “emails and social media posts are good enough” for New Year’s wishes. And among young people aged 18 to 29, that number was a whopping 80 percent. Is the culture of exchanging postcards on its way to extinction? “How sad,” I mumble to myself. Today is the winter solstice already. Let me remind my fellow non-digital New Year’s greeting card senders--who have become a minority--that the sooner they get it done, the better. A poem by Ishikawa goes to the effect, “The precious words that I failed to tell you the last time/ Are still in my heart.” With the year winding down, I am feeling oddly sentimental.

--The Asahi Shimbun, Dec. 21

It would be interesting to learn whether the published version above was initially generated by AI and then enhanced by the editors with additional information such as "a fellow Iwate native" for 同郷 .  Personally, I found all of the versions very touching. I think that reflects the expert prompting on your part as well as my greater tolerance over the years for 起承転結, which left our whole class of college students in America completely confused when we encountered the essay on nuclear disarmament in Hibbett and Itasaka in 1972. The native Japanese professor spent most of the hour explaining the structure.

With regard to marketing copy, in my work I frequently visit the Japanese websites of certain non-prescription concoctions such as herbal supplements, beauty products, and so on. It's merely an impression on my part, but I think that Japanese websites advertising these products have been influenced by English-based marketing strategies in the way they are organized and presented. In some ways they actually resemble the infomercial come-ons we still see on late-night TV here in the U.S. 

I agree with your statement about how much leeway the translator may or may not have to rearrange and supplement marketing copy. I think Japanese customers prefer J>E translators to translate as literally as possible, especially when they have an in-house NJS proofreading the translation. I suspect American companies trying to market in Japan simply cross their fingers and hope that the translators they hire are presenting their products in the best possible light, but if they farm their work out to agencies, most likely the agency will choose an NJS translator (hopefully).

Thanks again for your detailed response.

John Stroman
----------------

Tom Gally

unread,
Dec 22, 2024, 9:47:40 PM12/22/24
to hon...@googlegroups.com
John wrote:

It's merely an impression on my part, but I think that Japanese websites advertising these products have been influenced by English-based marketing strategies in the way they are organized and presented.

Another example of cross-cultural rhetorical influence is in academic writing. Japanese text has not traditionally been organized into long thematic paragraphs, each with a topic sentence, as much formal writing in English is, and a lot of serious Japanese writing still does not use that structure much. But Japanese academics now often write Japanese in paragraphs with topic sentences. I don’t know to what extent that has been influenced by translations from English, researchers reading papers and books written in English, the teaching of English academic writing, or the incorporation of English-like paragraph structures into the teaching of Japanese academic writing. Probably some combination of all four.

One of the many surprising things to me about ChatGPT when it was first released two years ago was its ability to write original essays consisting of well-formed paragraphs with topic sentences. That can be hard for novice human writers.

Tom Gally
Reply all
Reply to author
Forward
0 new messages