How Machine Translation Can Harm and Help Your Business

Page Content:

1. Drawbacks
2. Evolution
3. Useful Books

There’s no doubt that if we compare the quality of today’s machine translation and the one of a few years ago, there will be a dramatic difference between them. The earlier versions of machine translation tools were too inaccurate and delivered unsatisfactory results. The best they could do is to translate your text word by word without paying any attention to the context or its topic. It is a computer trying to understand human language, after all. But can the machine translation be useful for your business now?

You would spend a considerable amount of time trying to figure out the meaning of the text that ignores the important details your translation tool was unable to give you. But the progress of every aspect of digital development couldn’t leave this flaws behind without necessary attention. The tools we have nowadays are much more effective in giving us an overall understanding of a translated text.

They use an improved algorithm to analyze the context and choose the most suitable meaning of a translated word. Professional translation tools include this feature to help the experts work faster, so you can’t say that machine translation is useless. With its help, translators get to focus on the process without checking some of the words’ interpretations themselves.

1. Drawbacks of Machine Translation

The biggest problem with a translation tool is its inability to cope with the multiple meanings a word can have. Almost every language in the world is full of metaphors enriching our speech and making communication more vibrant and interesting. Machines simply can’t make sense of those and thus, give us ridiculously meaningless translations. Also, there are words that have completely different meanings when used in various contexts. In English the examples of such words would be:

To bear – to carry; a bear – an animal

can – to be able to; can- container

to go – to turn; to go – to leave

There are words that have multiple meanings that are very different from one another. Here are some more examples of possible translation mistakes:

Spanish:

Italian:

German:

French:

The list goes on and on. And if some of them will be easy to spot as a mistake in translation, the others can be not that obvious. The worst part is that they can alter the meaning of a source text drastically. But there’s yet another problem with machine translation that doesn’t allow it to reach the level of a human translation. It’s the order of words in a sentence. In most of the European languages, the order of words is incredibly important. Changing it can result in a completely different meaning of the initial phrase and huge misunderstandings.

2. Evolution of the Machine Translation

All of these factors are the main obstacles for a machine translation replacing professional translators. The masterminds of computer development will still be trying to come up with a universal solution for improving the process, but it is rather weak at the moment. There might come a time when only a professional translator will be able to notice some inaccuracies in the machine translation but for now, they continue to be quite obvious for everybody.

There was an Israeli philosopher named Yehoshua Bar-Hillel who berried any possible expectations of a machine translation in 1960. He wrote a phrase that no machine translator could cope with: “John was looking for his toy box. Finally, he found it. The box was in the pen”. The word pen, having a meaning of an object you can write with and a small enclosure for a baby, turned to be a problem with no solution for all of the known translation machines at the time. You’ll be surprised to know that modern MT tools still can’t translate this phrase correctly.

All You Need to Know About Machine Translation

If you want to use machine translation tools for your work or are planning to write a research paper on machine translation, there are some important points you should be aware of. First of all, it’s vital to understand the definition of a machine translation and the way it functions. As you might already know, it has a lot of disadvantages in comparison to a human translation. You might want to find out about their existence and some of the possible ways to fix them. But let’s start with a few historical facts first.

It was in early 1960-s when the humanity understood there was no bright future for a machine translation. The flaws of a word-to-word translation those tools could provide were too obvious. The multiple meanings of the words were a huge obstacle on the way of MT progress. The low effectiveness and accuracy level were then expressed in the report of the Committee of Applied Linguistic of 1966 and was supposed to put a period to its further development.

New Era

A new era of machine translation started with the introduction of a statistical translation during the second half of 1990-s. This was a new approach to the translation process made by computers which allowed them to use the previous knowledge.Its functioning was based on the analysis of previously translated pieces of information. This made it possible to create a rough understanding of the language structure and find the meanings of the words using comparative analysis. The biggest disadvantage it had, was the necessity of uploading a huge amount of information for a computer to work as accurately as possible.

And, of course, it still made a lot of errors. The computers of those times didn’t have enough energy to process those enormous information files. This problem is still being solved. It brings positive results from time to time but remains to be an obstacle. Its biggest advantage, on the other hand, was that it allowed people all over the world to read articles written in different languages while surfing on the web. It also played the leading role in deciphering the Rosetta stone.

3. Useful Books on Machine Translation

Our recommendation is to start with Statistical Machine Translation Philipp Koehn. It’s a comprehensive text that will help you find out about the vital details of the statistical translation processes that are developing pretty rapidly. Foundations of Statistical Language Processing by Christopher Manning and Hinrich Schutze will also be a great introduction to this amazing translation tool. If you need a 360-degree view on the phenomenon of machine translation, check Machine Translation by Yorick Wilks. The book by the Professor of Artificial Intelligence at the University of Sheffield and knows all the ins and outs of the MT functioning.

There’s a chance for a machine translation to become a highly effective tool and even surpass the possibilities of a human translation. We’re just humans after all, and our abilities are limited, unfortunately. Unlike us, computers can keep in their archives terabytes of information without missing out any tiny bit of it. Let’s hope for the best and wait for these times to come. this would definitely ease the life of professional translators. Meanwhile, you can use their help at Translateshark.com. Their talent and accuracy are something modern MT tools can only be dreaming of.