At least the raft, if not the shore

Hi, I’m Dung Kai-cheung, a Hong Kong fiction writer. I still remember the excitement I felt when I published my first short story in a local newspaper supplement thirty years ago. Honestly I am having similar feelings now when I am going to publish my first post in English, this time on an electronic publication I run on my own.

Yes, I have written only in Chinese, including fiction, essays and literary criticism. My last extended piece of writing in English was my M. Phil thesis in Comparative Literature when I was in my twenties. I have spent more time reading English than writing it, so it takes me a little effort to do so. But I am happy to finally giving it a go.

The purpose of this publication is simple: To share my work, both new and old, in serialized form, with non-Chinese-reading fiction lovers of whatever nationalities who can read English. To do so I have to translate them into English in the first place, and that’s where the tricky part lies.

How do I translate? The way to do it needs some explanation and justification. I am going to use AI translation machine, preferably GPT-4, to produce the first draft, then work on it to bring the translation as close as possible to the original. I may have the advantage of being the author, but admittedly I have little experience, not to say expertise, in translation, which requires professional training. That’s why I enlist the help of technology.

I’m aware of the reservations about using machine instead of human translation. The major concerns are not just technical, as to the accuracy and reliability of AI models, but also aesthetic, or even ethical. We all know that word-for-word equivalence simply doesn’t exist between any languages. What’s involved is nothing less than cross-cultural exchanges which result, to the best of efforts, in only a close approximation in another language. The choices a translator makes is artistry itself, and the best translations are works of art in their own right. It seems that machine generated products can never achieve such status. Leaving aside the question of whether AI computation can out-perform human interpretation, there is also the doubt on the legitimacy of an AI generated “work.” Whose work is it? Who can claim the right to such work? Such issues are still open to debate.

If it is so, why not find a human translator? I am lucky enough to have three of my works translated into English, two by eminent scholars and translators Bonnie McDougall and Anders Hansson, another one by my old schoolmate, now a professor of translation, Yau Wai-ping. Yet many of my other works have been sitting on the sidelines for quite some time, waiting for the opportunity to descend from nowhere.

Obstacles to translating Hong Kong literature are formidable. First of all, of all literature written in Chinese, works of Hong Kong writers are low in priority for Western publishers. The reasons are perhaps lack of knowledge and interest, and above all, marketing concerns. Even if there are publishers willing to take the risk, one needs to apply for special funding, without which it’s simply beyond the means of an ordinary writer, or even a small publisher, to pay the translator. Unless there are translators offering to work for free, which is the case of Bonnie and Anders, the most generous souls on earth! But this is not to be expected and taken for granted. Such unfavourable conditions apply not just to Hong Kong but to every non-Western community which would love to share its literature with the rest of the world.

Why not translate them myself then? Suppose I have the competence, or train myself up to be so, I still need the time to do it. I have written and published more than thirty books, many of which are full-length novels. Translating any one equals to rewriting it all over. If I am to translate a quarter of them by hand, it will probably take me another lifetime. Not to say that I have to quit writing for translating.

It is upon this background that new innovations in technology come as a possible solution, at least in reducing the costs and labour involved, endowing the author with increased autonomy on the translation, publication and distribution of his or her own works. Rather than wanting to replace or rule out translators, it is more a question of empowering the author.

While there are certainly differences between machine and human translation, the gap is not as great as that between machine and human creation. In translation by whatever means, the original remains stable in its status as a human-created artwork. Machine translation may not be artistic but it can be useful. It is more often literal than idiomatic, yet faithful to the original. In spite of its limitations, it is still worth trying, so long as we keep the differences and their implications in mind, and being honest about when and how it is used. Employed with moderation and under proper supervision, it serves the author and the original work. I believe that it can facilitate more meaningful exchanges between cultures.

As the Buddhist saying goes: the raft is not the shore; or in another version: the finger pointing at the moon is not the moon. Allow me to twist the metaphor a bit to my purpose: all translations are the raft that transports or the finger that points, but not the destination itself. Translations by AI machines are perhaps even more steps removed. Yet it is certainly better to have the raft or the finger than not, so that at least we are on the way to the shore, or can behold it in some distance.

The translations given here may not be definitive, but they are good enough indications of what the original works are like. I hope they will provide more than just a glimpse of what I have written over the years. If in case any professional translator find them inadequate and is interested in bringing the raft closer to the shore, please contact me to figure out the chance for working on further improvements together.

The rule here is simple: All original works of fiction are translated with the help of AI which generates the first draft, then corrected, refined and edited by the author. All communicative passages like the present one come directly from the hand of the author himself, without any collaboration with or polishing up by machine.

In spite of the apparent efficiency and ease machine translation has made possible, no time or effort on revision and proofreading can be spared. To share more of my works with international readers will be a long term project. This attempt is so new that I can’t foresee how far it can go. All I can promise at the moment is the following plan:

  1. Short stories from my V City Series first published in the late nineties ( belonging to the same period as two of my already published translations Atlas: the Archaeology of an Imaginary City and A Catalog of Such Stuff as Dreams Are Made On)
  2. Recently written short stories from the Bodhisattva Series.
  3. Serialization of my 2021 novel Hong Kong Type: A Love Letter Late for 150 Years.

You are welcome to subscribe for free to have a look at what this project offers. If you appreciate my works and want to show your support, please consider pledging for paid subscription.


Share DKC in Translation

For more information about myself and my works, please visit my bilingual personal website :

For those who read Chinese, you may be interested in my Chinese Substack:

董富記快報 Dungfookei

You may also be interested in the NFT book version of Works and Creations, As Vivid as Real, (天工開物.栩栩如真), the first ever Chinese novel published in NFT form.

天工開物.栩栩如真 NFT書

Published English translations of my works:

Atlas: the Archaeology of an Imaginary City, trans. Anders Hansson, Bonnie S. McDougall and Dung Kai-cheung , Columbia University Press, 2012. (Chinese original first published in 1997)

Cantonese Love Stories: Twenty-five Vignettes of a City, trans. Anders Hansson and Bonnie S. McDougall, Penguin Australia, 2017. (A selection from the fourth item, A Catalog)

The History of the Adventures of Vivi and Vera, trans. Yau Wai-ping, Hong Kong University Press, 2018. (Chinese original, literally translated as Works and Creations, As Vivid As Real, first published in 2005)

A Catalog of Such Stuff as Dreams Are Made On, trans. Bonnie S. McDougall and Anders Hansson, Columbia University Press, 2022. (Chinese original first published in 1999)

Atlas has also been translated into Japanese by Professor Fuji Shozo and my writer friend Nakajima Kyoko in 2012.



發佈留言必須填寫的電子郵件地址不會公開。 必填欄位標示為 *