“Winner, Winner, Chicken Dinner” originally comes from the movie 21. In the old Las Vegas casinos, a chicken meal cost $1.79, and a single bet was $2, so winning once would buy you a chicken meal. It was later used in games like PUBG, where the phrase “Winner, Winner, Chicken Dinner” appears on the screen when a player wins. Netizens call it “winner winner chicken dinner,” which has quickly become a popular internet catchphrase meaning “great luck, chicken for dinner tonight.”
“Eat chicken,” “give it up,” “oily,” “Buddhist-style”...... Whenever new things emerge, people create new words to describe them or extend the meanings of existing ones. The charm of language lies in using “one-dimensional characters” to express an entire scene. Translation, even more so, requires a comprehensive consideration of the original context and underlying connotations—being both concise and thorough. Only such translation can truly resonate.
Thus emerged the “Holographic One-Dimension Theory” of He Enpei, a veteran in the translation industry and chairman of Transn: Translation is the true high ground of human wisdom. Language exists because of humans, and language is the “one-dimensional expression” of humanity’s “holographic wisdom.” In the long run, AI may surpass human intelligence, but human wisdom is hard to surpass.
He Enpei believes: “Machines’ ‘one-dimensional training’ based on text cannot replace the mapping between the ‘holographic information’ in the human brain. The holographic information in the human brain is a synthesis of all information from the eyes, ears, nose, mouth, tongue, body, and mind, formed based on shared cultural experiences and consensus. Language is just a one-dimensional expression of the brain’s holographic information, and it relies on the mapping of that holographic information to be truly understood. The understanding of the mapping between two languages is based on the holographic information behind those two languages. Therefore, translation robots trained on large amounts of one-dimensional textual data can basically only mechanically ‘map’ and find it difficult to achieve a full holographic conversion, naturally failing to truly convey the original meaning into another language. Take a translation scenario for example: the translator synthesizes a holographic scene based on the speaker’s emotions, tone, expressions, and cultural background, then expresses it in language—but machines cannot do this. Therefore, language and translation exist because of humans and are expressions of human wisdom. In the long run, artificial intelligence may surpass human intelligence, but human wisdom is hard to surpass.”
Indeed, translation grounded in holographic information is complex. This complexity is not only reflected in the multidimensionality of the original scene but is also closely related to the nature of language itself.
First, language is racial. Different languages emerge from different cultural backgrounds and reflect how different ethnic groups perceive and approach things. Take the word “sofa” as an example—it is a European term for seating furniture and carries racial uniqueness. The racial nature of language is supported by cultural background; without culture, language cannot survive. There was once a constructed international language called Esperanto, which humans hoped would enable barrier-free communication. At its peak, it was promoted to 30 million users, but now there may be fewer than 2,000 speakers. A language not rooted in cultural heritage cannot survive.
Second, language is a consensus among a group of people, characterized by being non-demonstrable and non-derivable. In other words, the name of something is essentially a matter of common agreement, which is especially evident in dialects. For example, phrases like “I don’t know” (meaning “I don’t know”) and “go for a few days” (meaning “go for a few days”)—for AI, it would consider the word order of “I don’t know” incorrect and would think “go for a few days” literally means two days, not an approximate number.Consensus changes over time sporadically, and machines cannot accurately express this in a short time.
The reason Go could be quickly surpassed by AI is that it has fixed rules and definite outcomes, so AI could train day and night by playing against itself. But language is different—it is racial, non-demonstrable, and non-derivable, and it is almost impossible to improve like ALPHA does by generating its own games and playing against itself. He Enpei stated: “I’ve seen some reports exploring ‘self-play training’ for translation. In my personal opinion, that would at most produce a kind of ‘machine dialect,’ like a dialect used only within a closed tribe, and could never reach the level of translation needed by modern society. The synchronization and collaboration between the iTransSmart intelligent translation robot and translators aim to resolve the conflict between machine dialect and the holographic wisdom of language, expanding translation capacity.”
In today’s world where the “AI threat” is a hot topic, He Enpei’s “Holographic One-Dimension Theory” opens up a new train of thought for the industry. He also reminds the industry to be less impetuous and more sincere, to view new technologies rationally, and to gradually achieve industry transformation in a more practical way.