Working with Machine Translation

Like CAT tools, machine translation facilitates humans in the translation process. Learn about how to read results and understand the functions of machine translation.

Machine Translation: Human-assisting Technology

The process of Spanish machine translation is quite similar to the process of Spanish human translation. There are three essential steps: extracting the true meaning of the source text; rendering the true meaning in the target language; and giving form to the true meaning by expressing it in the words and syntax of the source language. There are two programs that are used in the process of Spanish machine translation. The first program is the active one as it analyzes the source text in order to extract the true meaning, it renders the true meaning in the target language and it follows it up by expressing it in the target language. The second program is the passive one as it simply comprises dictionaries. The dictionaries are used to locate the corresponding words in the target language. Apart from extracting the true meaning of the source text, the process of Spanish machine translation also involves analyzing the grammar of the source text. Grammar is analyzed through the method known as parsing. Through this method the function of each phrase is determined. Each phrase is then classified according to the known types of phrases, for example, noun phrases (NP), verb phrases (VP), prepositional phrases (PP), etc. Each phrase is then broken down further into the component words in order to determine their function. The transferring process can begin only after the function of each word in the source text has been determined.

Two Different Strategies for Machine Translation

Spanish machine translation is used to follow either of the following two strategies: The first one was the direct approach and was mostly in vogue till the 1960s. It worked on a word—for-word basis between any two languages, for example, English and Spanish. The syntax in both the source and target languages was kept simple to avoid ambiguities. This translation strategy worked best with standardized materials. The second one was more complex because it involved multiple languages, but the approach was more or less similar to the one used in the direct approach. Spanish machine translation requires over-sized dictionaries. They are of three types: monolingual dictionaries in the source language, bilingual dictionaries in the source and the target languages, and monolingual dictionaries in the target language. The three dictionaries have three distinct functions: the first one assists in the analysis of the source text, the second one helps to transfer meaning from the source to the target language, and the third one helps re-structuring in the target language. Some of the dictionaries contain only the infinitive forms of verbs, while some contain words with all the possible inflections. In spite of the existence of these dictionaries, Spanish machine translation finds it difficult to render idiomatic expressions and homographs. Recent developments that have enhanced the appeal of Spanish translation machine include disks and the optical character recognition technology that has eliminated the need to type and retype data into the computer. Editing does not need be done manually either, but it can be done on screen with the help of word processors.