Is Automated Translation Good Enough?:
(Or: Take the Best of Both Worlds)

I stated in an earlier blog that English as lingua franca across peoples and geographies has enabled travel (the movement of people), commerce (the movement of goods and services), and education (the movement of ideas) to flourish and has thus hugely contributed to human understanding, knowledge and well-being (“The Importance of Good English”).

Access to English for everyone

It is a good thing that the availability of data and computer power have given rise to automated translation software.  By providing cheap and fast access to English (from a foreign-language text), automated translation vastly expands the number of people who can benefit from English as the lingua franca.  It allows everybody, from students and small businesses to professional business persons and large organisations, to gain some access to the vast economic world outside their own linguistic environment.

The richness and complexity of human language and human living

Data and computer power have rendered automated translation tools surprisingly accurate.  However, they are still no match for the richness and complexity of human language or human living that is mastered by every human brain.  There is a limit to the ability of automated translation tools to transfer a human thought, concept or idea from one language to another.

Human language contains words in structured sentences, to express ideas in real-life contexts.  Humans use words and sentence structure to express meanings that reflect the richness and complexity of the human condition :

  • The sharp mind:  Very fine differences between meanings, also known as nuance
  • The emotional being:  Tone, emphasis, attitude, humour, emotion
  • The thinking being:  Literal, figurative or abstract expression
  • The associative mind:  Meanings related to specific real-life contexts

Computers are not thinking or feeling beings:  they do not understand the meaning of a word, or a real-life context.  They are simply programmed to make associations between words, and between words and contexts.

What Automated Translation is good at

Automated Translation is very efficient at providing very quickly and easily a first draft translation of large texts with meanings that are relatively simple.

What Automated Translation can’t do

Automation Translation will never be able to associate all the richness and complexity of the human condition with all the available words in a language. 

It will never be able to identify with certainty all the very precise expressions that reflect the human condition—fine differences in meaning, differences in emotions, differences between literal, figurative and abstract expression, and the endless differences in real-life contexts.

It will not always be able to identify with certainty the specific real-life context from the words of a text.

The best of both worlds:  The easy 80% and the difficult 20%

Automated translation does what all toos do: They help the human do the easy parts fast, cheaply and reliably. That allows humans to focus on the difficult parts. In the case of translation, the difficult part is adapting it to the precision, complexity and feeling of human expression.

The easy 80% :  Use automated translation tools to translate your text, especially if it is long and doesn’t have complex structures or meanings.  This gives you a very rapid and cheap path towards a starting draft, and can represent 80% or more of the effort required for an accurate and complete translation.

The difficult 20% : Use a native-speaker, a human being, to review, correct and improve the starting draft produced by the automated translation, to adapt it to make sure that it reflects correctly fine differences in meaning, that it expresses accurately and fully the “emotional being” and the “thinking being” in the message, and that it corresponds to the correct context.

Stay Updated

Be the first to know about our latest blog publication.

Make It Native © 2024. All rights reserved.Privacy Policy | Terms of Use