Report A, by Amaia Montes Garate


In this document I will provide information to show how difficult translation is. In the report I will be explaining from a personal point of view the difficulties human translators face and I will also let the reader know about Machine Translation. This view about machines is not so personal and that is why I have added several opinions of different people both to support and contrast my ideas about MT, Language Engineering... I will also give several examples and, therefore, use more languages apart from English in order to show the difficulties these kind of systems or translators themselves face.


Language, which is the most effective way we have to express ourselves to each other, lets us explain concepts, negotiate, persuade, argue, express our feelings, narrate stories, record our culture for future generations and even create beauty in poetry and prose. All of as agree up to this point but, is it easy to translate all that language implies? As far as I am concerned, it is not; that is why I have chosen this subject for my report. Not only do I have the objective of showing the difficulties translators (meaning human translators and also Machine Translators) face but I also mean to provide more information about Machine Translation, Language Engineering and Translation Workstation, which, I believe, are still something many people do not know about. Besides, I also want to give different opinions about the use of Machine Translation and I have inserted some quotations for this purpose. In this quotations different people give their opinion about these new technologies.

Among all I want to point out this quotation of Martin Kay (1992):

There is nothing that a person could know, or feel, or dream, that could not be crucial for getting a good translation of some text or other. To be a translator, therefore, one cannot just have some parts of humanity; one must be a complete human being.

The reason why I have chosen such a relevant quotation to write in my introduction is that it represents very well all my project; first, it gives a little explanation of how qualified a translator should be and, secondly, we can guess the underlying ideia that even if machines are good, real translators have some abilities that machines will not gain.

According to methodology, I have not followed the question-answer method we did in class but, as it can be seen in the reference section, I have borrowed a lot of material from these answers, combining them with several web pages that refer to translating both Basque and Spanish into English or vice versa.


English has been considered one of the easiest languages according to structures; however, translating anything into English is not an easy game, neither is it easy to change something that has been written in English into another language.

Translating used to be something that only human beings could do but it is no longer like this; from the very beginning of the course we learnt the society where we live is considered an information society; that is to say, a kind of society where it is very important the creation, distribution and manipulation of information. All this lets us know translation must also be relevant in this kind of society which has led to machine translation and other ways of translating by means of machines that I will be explaining later on.

First of all, I am going to focus on the main problems translators, meaning human translators, face. The fact that the translator is supposed to know both languages well does not mean we will always get a good translation; this is sometimes due to ambiguity. In fact, there are to different types of ambiguity; lexical and structural ambiguity; anyway, this will have more relevance when referring to Machine Translation.

Something else translators have to be careful with are tenses; for instance, in English we use the past perfect to speak about something that happened this morning (I ate an apple this morning) and in Spanish we use the past participle (Esta mañana he comido una manzana). This example shows how important it is to know when to use each one.

Registers are something translators have to be aware of too, as, depending on the language, we will have to change the register even if we are in an identical environment. As far as many Translation and Interpretation students are concerned the use of prepositions is rather difficult because they cannot expect to have a clear correspondence between the two languages.

Another issue that carries problems with it even for the better qualified translators are the use of sociolects and also idiolects in their texts. These sociolects and idiolects always show a variety of details that are very personal, in the case of the idiolect and, details that belong to a certain group or social group in the case of the sociolects. As a consequence translators have to face many difficulties when translating these kinds of texts into other languages as they seem to be changing the sense the writer meant to provide. Related to this, translators also find it difficult the translation of swearwords because they always seem to loose expressiveness.

A good example of a difficult problem to solve is the translation of certain words that refer to things or objects that only exist in the native language. For instance, how would we translate what a "talo" is into English as native speakers of Basque we are? There are certain cases where a simple explanation could be provided; for example, in the middle of a narration we could explain that a "talo" is what Basque rural families used to cook by means of cornmeal mixed with water and got a kind of paste which they usually combined with another product. This explanation might be good or, at least, not bad in a text but if we happen to be translating poetry the translator finds himself or herself lost in certain circumstances like those. Some other good examples would be how to translate these terms into other languages: English cottage, French châteu… All these might be similar but the translation very often looses the identity the author meant to provide. This is called a problem of lexical mismatches as we can check in the following page by the title Multiword units: Idioms and Collocationsand it also has influence when according to Machine Translation:

Another reason why it is so important for the translator to have a very good knowledge of the languages is idiomatic expressions. The problem with them is that it is not usually possible to translate them as we translate the rest of the text, as we are told in the same web page I have already mentioned, one has to treat idioms as single units in translation. Generalizing can always be dangerous as there are some cases where the translation can be done without much difficulty:

These three mean ´face and tackle a difficulty without shirking` and there is very little difference among them if we translate them literary. However, this hardly ever happens; here we have another example where it is clear what a mess it would be if we translated idioms literary:

These three examples have the same meaning; to pull ´someone´s leg`but the third example, that is to say, the one in Basque would be literary translated into English like ´to hit someone´s horn` and the Spanish example; the second one, would be similar but in reference to the ´hair`. As a consequence we can say that the use of normal rules in order to translate idioms will result in nonsense.

There are still so many difficulties translators have to face in their everyday life that my project would become endless were I to point out all of them. To sum up, the potentially good translator must be a sensitive, wise, vigilant, talented, gifted, experienced, and knowledgeable person, as we are said in the web page that follows:

Let us then hear about machines. First of all I must point out that when I refer to machines not only do I mean the machine tools of the Information Society, that is to say, computers and telecommunications, but I also imply Machine Translation, Language Engineering…To sum up, language technologies; information technologies that are specialized for dealing with a complex information medium in our world: human language.

To begin with, Language Engineering can improve the quality of information services by using techniques which give more accurate results to search requests and also increase the possibility of finding all the information we are looking for. Language engineering is the application of knowledge of language to the development of computer systems which can recognise, understand, interpret, and generate human language in all its forms. In practice, Language Engineering comprises a set of techniques and language resources.

These are the main techniques used in Language Engineering: speaker identification and verification, speech recognition, where we often find difficulties in dealing with accents, dialects and language spoken; character and document image recognition, natural language understanding… The last one is not always essential; understanding is not always a requirement as combinations of analysis and generation with semantic model help and even allow texts to be translated.

To finish with Language Engineering I must add that language resources are very important components because they are one of the main ways of representing the knowledge of language which leads, I underline, to recognition and understanding. These are some of the resources: first, lexicons, as lexicons are needed for every language of application; secondly, specialist lexicons where we find proper names, terminology, wordnets, grammar…

Machine Translation, often known as MT, normally means fully automatic translation. Machine Translation helps a user to transform or just transforms a written text from one language into another by means of a computer-based process. Inside Machine Translation we can differentiate several branches. Fully Automated Machine Translation or FAMT is basically the same of Machine Translation or MT but without the intervention of a person during its process. Human-Assisted Machine Translation (HAMT) is a style of translation where the one who does most of the translation is the computer system itself. Machine-Aided Translation (MAT) is the translation in which the human does nearly all the work but uses one or more computer systems in order to be more efficient. For example, dictionaries and spelling checkers.

These Machine Translation systems have to deal with rather difficult tasks; an example would be ambiguity, which I have already mentioned. There are two types of ambiguity; lexical ambiguity, which we get when a word is ambiguous and structural ambiguity; when a phrase or a sentence has more than one structure. All this makes the work of machine translation harder. Besides, idioms, collocations and all difficulties that people have as translators are bound to be difficulties when according to machines.

In the evolution of Machine Translation, the main thrust of research was based on the explicit or the implicit assumption that the aim of MT would be automatic systems producing translations as good as those made by human translators. And, perhaps, being so ambitious did not help very much because professional translators, who naturally became anxious and antagonistic were against the idea of Machine Translation because they foresaw the loss of their jobs. Yet, they did not know many of the systems would enquire human involment.

As language, knowledge and thought are represented in the human brain language technology had to create formal representation systems that would get to link language to concepts and tasks in the real world. All this is connected to knowledge technologies, opposed to what many people think of the machine´s not owning knowledge at all.

But humans are not so limited; we combine speech with gesture and also facial expressions and, as a consequence digital texts have been combined with pictures and sounds. The idea or dream of translating human voice automatically has been present for a long time (Locke, 1955) but it is still a dream.

And why do we hear about Language Engineering, Machine Translation… so often? This has an easy answer; there has been an increase of information in electronic format which is linked to advances in computational techniques for dealing with it. Together with the proliferation of informational webs in Internet, we can also see a number of search and devices which is becoming bigger and bigger, some of which integrate translation technology. Technical documentation is becoming electronic as well, in the form of CD-ROM, on-line manuals, intranets, etc. Internet has become popular and its consequence is that the access to information is now global and the demand for localizing Web sites is growing fast.

We have even listened about the translation workstation, which would combine the following features to be the ideal work to get ideal translations, as we can see in

Full integration in the translator's general working environment, which comprises the operating system, the document editor (hypertext authoring, desktop publisher or the standard word-processor), as well as the e-mailer or the Web browser. These would be complemented with a wide collection of linguistic tools: from spell, grammar and style checkers to on-line dictionaries, and glossaries, including terminology management, annotated corpora, concordances, collated texts, etc.

But, are all these in fact reliable? Has any of these systems been put into practice yet? Of course it has; an example of a project which is successfully helping to improve communications in Europe is one which interconnects many of the police forces of northern Europe using a limited, controlled language which can be automatically translated, in real-time. Such a facility helps to prevent and detect international crime and it also assists the emergency services to communicate effectively during a major incident.

Besides, we also perceive that important managers such as the General Manager of LionBridge, Santi van der Kruk finds knowledge over translators essential as we can see in the quotation that follows:

The profile we look for in translators is an excellent knowledge of computer technology and superb linguistic ability in both the source and target languages. They must know how to use the leading CAT [computer assisted translation] tools and applications and be flexible. The information technology and localization industries are evolving very rapidly and translators need to move with them.

And to finish showing examples of the application of some of the terms that seem to be new for us, and, actually, which are not so new, I must add that ten years ago, for instance, Machine Translation was something that was being already used by large organizations such as the European Commission, the US Government, the Pan American Health Organization, Xerox, Fujitsu, etc.

After having studied different types of translating without needing a person translating all the text and after reading some of the difficulties they face, we see that these machines are used more and more which means fewer translators are needed. But, does not translation excellence go beyond technology? This is what Martin Kay (1987) thinks:

A computer is a device that can be used to magnify human productivity. Properly used, it does not dehumanize by imposing its own Orwellian stamp on the products of human spirit and the dignity of human labor but, by taking over what is mechanical and routine, it frees human beings over what is mechanical and routine. Translation is a fine and exacting art, but there is much about it that is mechanical and routine, if this were given over to a machine, the productivity of the translator would not only be magnified but this work would become more rewarding, more exciting, more human.

In fact, I agree with this statement; first of all the translator must be well qualified but, afterwards his or her skills can be improved by means of these new technologies and systems. The following quotation belongs to Douglas Hofstadter (1998):

A skilled literary translator makes a far larger number of changes, and far more significant changes, than any virtuoso performer of classical music would ever dare to make in playing notes in the score of, say, a Beethoven piano sonata. In literary translation, it's totally humdrum stuff for new ideas to be interpreted, old ideas to be deleted, structures to be inverted, twisted around, and on and on.

As Martin Kay, again (1992), argues, there is an intrinsic and irreplaceable human aspect of translation and I very much agree with him after having read his quotation. In my opinion, I repeat, technologies are important because they help you improve but what a translator feels is more like what a poet feels, though hardly ever getting to transmit all the real author wanted the audience to receive.

There is nothing that a person could know, or feel, or dream, that could not be crucial for getting a good translation of some text or other. To be a translator, therefore, one cannot just have some parts of humanity; one must be a complete human being.

Besides, I also agree with Laura Cruz García, from the University of Las Palmas de Gran Canaria when she says that words can be easily translated whereas cultures cannot be translated because there will always be some mistakes, no matter how tiny these are, which can lead to misinterpretation.

Before I begin with my conclusion I must add that the previous idea has sometimes been presented as a reason to prefer human translators rather than machine translation. Besides, the famous sentence Knowledge is power, but information is not strengthens this idea as, if what machines have is only information, the translation would be rather useless; knowledge is information that is transformed through reasoning and reflection into beliefs, concepts, and mental models and machines might not be so good qualified yet. However, after having read this report we should already know about ´knowledge technologies`.


After having read the report we see we have different choices for translation. As I have already said, it is not a case of being whether in favour of human translators or in favour of machines but considering a combination of both of them or just accepting the existence of both methods. Actually, we would get the best results if we were proffesional translators and if we were capable to combine our knowledge with the help of machines. Anyway, as always, everything depends on quality and that is why we should choose reliable sources for our translation.

I must add that a text that has been translated is never the same of the original; I understand there are cases where it is essential to translate as we can see in the methods used to look for criminals through Europe but even if a novel by Dickens will be understandable in any other language many of the feelings the writer wanted to trasmit and many of the significance of those times (the London of the Dickensian times, for instance) cannot be translated at all.

References h