The technology of translation - 6 Minute English

Summary

In this episode of BBC Learning English, the hosts discuss the evolution of language technologies, particularly how translation software like Google Translate works and its limitations. While technology has revolutionized language learning and removed communication barriers, it struggles with translating complex language, idioms, and metaphors. The program highlights the historical mistranslation of Braniff Airlines' slogan and emphasizes how human creativity and understanding still outpace machines. Viewers also learn key vocabulary related to language translation, including 'metaphor,' 'chunk,' and 'rote learning,' while exploring the challenges of translating creative expressions like poetry.

Takeaways

• 😀 Language technologies, such as translation apps, are helping bridge communication gaps across languages.
• 😀 Modern software like Google Translate has revolutionized language learning and communication.
• 😀 Despite advancements, translation software can struggle with complex language and idiomatic expressions.
• 😀 One example of a translation mistake occurred in 1987 when Braniff's 'fly in leather' slogan was mistranslated as 'fly naked' in Spanish.
• 😀 Traditional computer translation systems rely on rules-based methods, translating word-for-word based on grammar rules.
• 😀 Translating idiomatic expressions, like 'it’s raining cats and dogs,' requires context, as literal translations often don't make sense.
• 😀 Apps like Google Translate use predefined chunks to translate idioms into equivalent expressions in other languages.
• 😀 Machines can handle repetitive tasks well but struggle with creative or novel expressions, such as poetry.
• 😀 Computers 'learn' languages by memorizing patterns from millions of translations, not by truly understanding the meaning behind them.
• 😀 Rote learning, which involves memorizing without understanding, is a key feature of machine translation but not human learning.
• 😀 Even the most sophisticated machines can be confused by creative language, such as poetry, which remains a human advantage.

Q & A

What technology is discussed in this program?

The program discusses language translation technologies, particularly software like Google Translate that helps translate between languages.

How does the language translation app Rob uses work?

Rob uses an app that scans a word and translates it. For example, it translates 'it’s raining' to 'está lloviendo' in Spanish.

What historical example is used to highlight the challenges of machine translation?

The example of Braniff Airlines' 1987 advertisement 'fly in leather' being mistranslated into Spanish as 'fly naked' is used to show how machine translation can go wrong.

What did Braniff’s advertising slogan 'fly in leather' translate to in Spanish?

The slogan was mistranslated into Spanish as 'fly naked' because 'en cuero' in Spanish can be slang for 'naked'.

Why is translating metaphors challenging for computers?

Metaphors can be difficult for computers to translate because they require understanding of cultural nuances and context, rather than just literal translation.

What is the example given for a metaphor in translation?

'It’s raining cats and dogs' is an example of a metaphor that would not make sense if translated literally. Instead, the equivalent Spanish expression 'lueve a cántaros' is used.

What does Lane Greene argue about poetry and machine translation?

Lane Greene argues that poetry, with its creativity and novelty, is untranslatable by machines, as they rely on memorized patterns rather than creative thinking.

What is the main difference between human and machine learning, according to Lane Greene?

The main difference is that machines learn by rote, memorizing patterns without truly understanding them, while humans use creative thinking to learn and understand new concepts.

What is the concept of 'rote learning' in the context of machine translation?

Rote learning refers to the process where machines memorize information to repeat it, but they do not understand it in the way humans do.

What vocabulary terms are highlighted in the program?

The program highlights the following terms: 'automated translation', 'metaphor', 'chunk', 'rote learning', and 'flummoxed'.