Dialog Cards (Can You Guess The Mistake Here?)

Machine vs.
Human Translation – Course presentation
Introduction
Up until the late 90s when Machine Translation burst onto the translation scene, translations were
mostly performed by human translators either in pen-and-paper fashion or using Word processors on
computers. Translation is a time-consuming and creative task in itself as it involves the transference of
meaning from one language to another. A human translator can at best translate a few thousand
words on an average in a day. Even as technology came to be introduced into translation in the form
of computer-aided translation tools, electronic dictionaries etc., it could only ‘lend a hand’ to a
translator in his work.
Whereas, when Machine Translation was introduced to the public in 1996 by Systran and later
Babelfish, it was the beginning of a revolutionary trend in the field of translation. Although the results
of the translation were nowhere comparable to that of a human translator, the novelty of translating
millions of words at the click of a button quickly saw MT catch the imagination of technology
companies and translation providers.
As of today, Machine Translation is no longer just a trend, but an active and necessary player in the
translation workflow. It has its critics and its faults, but no stakeholder of the industry can now afford
to ignore MT and would do so at their own peril.
Dialog cards (Examples of typical errors made by MT - Can you guess the mistake here?)
1. MT cannot understand context.

Source - vielen Dank für das konstruktive Gespräch in Ihrem Hause
MT - Thank you for the constructive conversation in your home
Human translation - Thank you very much for the constructive discussion in your office
2. MT has no sense of style! MT reuses or rebuilds a translation.
Source -15 Minuten stehen lassen
MT - Leave for 15 minutes
Human translation - Set aside for 15 minutes
3. MT cannot make sense of new words or those which are not in its database.
Source- Krisen sind in der Regel auch nicht monokausal erklärbar
MT - Crises are usually not monokausal explainable
Human translation - Crises cannot typically be traced to a single cause
4. MT produces word-to-word or literal translation if it cannot find a match for the sentence.
Source - fertig ist Ihr Dessert
MT - Finished is your dessert
Human translation - your dessert is ready
A COMPARISON
MACHINE TRANSLATION HUMAN TRANSLATION

GT translates more than 100 An experienced professional
billion words per day translator can translate 3000
words per day on average
One tool is enough to A multilingual human translator
translate between multiple can at best translate between 3-4
language pairs language pairs
Cost of translation is low Cost of translation is more as
as several free tools are translators usually charge a fee
available and even paid MT
(often per word)
comes cheaper
Cannot translate context Can understand the context and

translate in many different ways
Cannot interpret puns, Can understand and translate
jokes, word play idiomatic usage
Inconsistent and Part of a translator’s job is to
inaccurate ensure accuracy and consistency
The Debate
Machine Translation is now a reality of our everyday life. However, like
most other things, it has admirers and critics. The advantages and
disadvantages, costs and benefits of machine translation over human
translation is constantly being debated not just among translators, but also in
the technology circles. Though the constant evolution of MT technology also
has some in the translation industry worried, MT is very unlikely to replace a
human translator, at least in the foreseeable future. Just for the plain reason
that language is a subjective matter and is evolving constantly in inorganic
ways. On the other hand, using MT as a technology tool to enhance
translation processes and workflows instead of shying away from it will bring
great benefit for translators and customers alike.
The below links will lead you to more discussions on this debate:
http://www.k-international.com/blog/google-translate/
http://www.globaltimes.cn/content/1010980.shtml
http://news.bbc.co.uk/2/hi/technology/8575526.stm

Dialog Cards (Can You Guess The Mistake Here?)

Uploaded by

Copyright:

Available Formats

Dialog Cards (Can You Guess The Mistake Here?)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dialog Cards (Can You Guess The Mistake Here?)

Uploaded by

Copyright:

Available Formats

Machine vs.

Human Translation – Course presentation

1. MT cannot understand context.

MACHINE TRANSLATION HUMAN TRANSLATION

Cannot translate context Can understand the context and

Machine Translation is now a reality of our everyday life. However, like

the technology circles. Though the constant evolution of MT technology also

has some in the translation industry worried, MT is very unlikely to replace a

that language is a subjective matter and is evolving constantly in inorganic

ways. On the other hand, using MT as a technology tool to enhance

great benefit for translators and customers alike.

You might also like