First, it’s important to realize that there are different types of translation software. The two main differences are Translation Memory Software and basic Translation Software. Translation Memory memorizes translations produced by a human and is able to assist translators by speeding up the translation of repeating texts in documents. Software translation is built by using algorithms that draw on a database of words – similar to a dictionary. They use these algorithms to translate the syntax (or rules) of one language to another.
How Does Translation Software Work?
When you type in a sentence it uses a rule formula to access certain components of the database and produces a translation that is as accurate as possible in a fraction of a second.
This tool allows the translation of simple sentences that are predictable (every time they are the same) like basic phrases, to be done quickly and perfectly.
Ex: almost every language has a translation for “hello!” , “how are you?”, “what is your name?” etc.
Some of problems or disadvantages when using Translation Software:
Problems exist when literal “word for word” translations do not translate because the strings of words carry a specific meaning based on their proximity and their context. Idioms are especially difficult for translation software as many of them require more than just a few words to describe and generally do not translate from one language to another because they simply do not have the same cultural references. A few very common idiomatic expressions dependant on context are:
- Don’t count your chickens before they hatch
Meaning: Don’t make plans for something that may not happen; alternatively, don’t make an assumption about something that does not have a definitively predetermined outcome
- Hit the nail on the head
1. To describe exactly what is causing a situation or problem; 2. To do exactly the right thing; 3.To do something in the most effective and efficient way; 4. To say exactly the right thing or to find the exact answer; 5.To be accurate or correct about something. Often abbreviated as HTNOTH all over the web
- Take with a grain of salt
To not take what someone says too seriously; to treat someone’s words with a degree of skepticism.
There are other areas that also pose great difficulties when translating. Synonyms are especially difficult to translate as they are highly contextual based. The more abstract the presiding idea is, the harder it is to bring the meaning to the word in the sentence.
- An easy example is “going to the ball”. If you’re talking sports, it’s easy to know that you have a player or number of players going to the ball or even going after the ball. This can be quite confusing if you’re actually talking about the ball being a large party. Or, we could simply “have a ball” when we go out. In context, it’s easy for an English speaker to realize this but in a translation when written, it can prove extremely difficult to depict.
- Language is loaded with homonyms (words with different meanings). Take the following quick list for example and while you do, think about how each of these words may not have a direct translation:
Reservation; ground; check; pupil; suit; bolt; draft;
Ambiguity is another area that poses extreme problems when translating text and is sometimes used purposely in writing and speech, or is sometimes accidental. A quick example is the following: ‘Jefferson knows a richer man than Gates’. It has two meanings. 1. Jefferson knows a man who is richer than Gates; and 2. Jefferson knows man who is richer than any man that Gates knows.
How about slang? This one poses even greater problems as many cultures, and subcultures including English speaking ones, create new terms based on their context almost daily. According to the all too familiar site buzzfeed, you could have a lot of fun trying to figure out what teen slang from 2016 actually means.
My personal favourite is “lit” as in, that translation is lit.
Put into context, a young witness in court may say “she triggered after he threw shade on Instagram”.
Acronyms are another common difficulty when it comes to software and are extremely difficult to translate. This sort of thing, from a translator’s perspective, requires actual consultation with the creator of the acronym. Although some are quite common and most people use them, and therefore could possibly be placed into a translation program and translated successfully, there are many acronyms that are simply industry specific. A few examples of acronyms that appear quite frequently in business are: MOU, M&A, E&O, EOD, CAD, QC, NDA, SEO, and EBITDA just to name a few. Of course, these may seem easy for some, but to the untrained eye, or to a software program, they simply don’t exist in their lexicon.
Poetry is the final icing on the cake and has scholars debating currently and has had them doing so for centuries. In the end, the translation is really what the translator understands/interprets from the text. As we may remember from our poetry classes, we can derive many meanings of what poetry means to us. Translating this type of text is truly an art. Scholars also argue, therefore, that Translation in general is an art.
The more we study translation and its functions, the more we realize that language is innately human. Furthermore, given the rapid evolution of our language and the creation of new concepts every day, we can conclude that unless a software program has the ability to truly understand all of the source context and all of the target context and exactly who will be reading it, it is in its essence, impossible… for now.
In my opinion, the “singularity” is the only way that translation will ever reach full human competency; this is, when computers become conscious and when Artificial Intelligence becomes as real as we are and begins to think for itself and become aware of its surroundings in order to perceive context. We’ll always need human translators as long as humans are around though, as we’re the ones that make up the language and will end up explaining it to the AI program responsible for translating the text and making the appropriate updates on both ends.