Machine Trans EN 15
Machine Translation: Advantage or Disadvantage for the Human Translator
Mariam Touré 2020GBJ002301 Hunan Normal University
1-Abstract
Breaking language barriers has been one of the ultimate goals to achieve in terms of globalization. The scarcity of human translator as compare to the demand for it is considerable hence introducing machine translation (MT) services was a logical step to take because one can not always get hold of a translator when needed. That being said, one must ask himself if the introduction of Artificial intelligence in this case applied to translation is an advantage or a disadvantage to the human translators. This paper intends to discuss whether machine translation is going to enhance the efficiency of human translators or deprive them of their job.
2-Key words
Machine translation, globalization, human translator, artificial intelligence
3-Introduction
By definition translation is a written or spoken expression of the meaning of a word, speech, book, etc. in another language. (The Concise Oxford English Dictionary. Machine translation is the process of using artificial intelligence (AI) to automatically translate content from one language (the source) to another (the target) without any human input. Translation is not a new concept to mankind, but there has been a tremendous change with globalization, which has been defined as 'the widening, deepening and speeding up of worldwide interconnectedness in all aspects of contemporary social life' (Held et al., 1999: 2).
The technological advancements allowed the introduction of machines to the existing discipline aiming to solve the problem of accessibility, time efficiency and cost. This advancement of technology like any other was welcome, as it appeared to be what human translation needed. But looking at the definition of machine translation, it is only fair to ask oneself few questions like: is machine translation going to erase human translation? What could the inconveniences of having machine convey our messages? And is there any way to create a balance between human translation and machine translation for the greater good?
4- Brief History of machine translation:
The origin of machine translation also known as automatic translation in non-English speaking countries traces back to the seventeenth century and it only became a reality in the twentieth century. In 1933 with two patents issued in France and Russia to Georges Artsrouni and Petr Trojanskij respectively. Artsrouni's patent was for a general-purpose machine, which could also function as a mechanical multilingual dictionary. Artsrouni's patent was for a general-purpose machine, which could also function as a mechanical multilingual dictionary. Trojanskij'upgraded the concept of multilingual dictionary by making proposals for coding and interpreting grammatical functions using universal' (Esperanto-based) symbols in a multilingual translation device.1946 and 1947 Andrew Booth (a British crystallographer) and Warren Weaver teamed up and put forward the first tentative ideas for using the newly invented computers for translating natural languages without having an insight on the precursors mentioned above. 949 Warren Weaver (a director at the Rockefeller Foundation) put forward specific proposals for tackling the obvious problems of ambiguity (or 'multiple meanings'), based on his knowledge of cryptography, statistics, information theory, logic and language universals. The decade of the 1950s and 1960s saw the beginnings of the three basic approaches to machine translation (MT). The first was the 'direct translation' model, where programming rules were developed for the translation specifically from one source language (SL) into one particular target language (TL) with a minimal amount of analysis and syntactic reorganization and this approach was rather ambiguous. The second approach which was the 'interlingua' model, based on abstract language-neutral representations (codes or symbols independent of both SL and TL), where translation would then be in two stages, from SL to interlingua and from interlingua to TL attempted to get rid of those ambiguities. The third one or the 'transfer approach', where conversion was through a transfer stage from abstract (i.e. disambiguated) representations of SL texts to equivalent TL representations; in this case, translation comprised three stages: analysis, transfer, and generation (or synthesis) was a much simpler approach but the direct approach was still used by most people.
The Automatic Language Processing Advisory Committee (ALPAC) was introduced in 1964 to examine the situation of the machine translation problems and its 1966 report titled " Languages and Machines: Computers in Translation and Linguistics." Was undeniably one of the most important events in the history of MT. The report highlighted the failures of the tentative around MT since its introduction and seemed to have been a follow to an influential survey by, Bar-Hillel (1960) in which he criticized the prevailing assumption that the goal of MT research should be the creation of fully automatic high quality translation (FAHQT) systems producing results indistinguishable from those of human translators. He argued that it was not merely unrealistic, given the current state of linguistic knowledge and computer systems, but impossible in principle. His pessimistic observation would be questioned due to the advancement of Artificial intelligence and the fact that are getting closer and closer to the dream of a fully automated translator. In the 1970s the third approach stated above received a lot more attention for the purpose of a full-automated translation. The first approach to MT was known as rule based approach until 1989, when the corpus based methods were introduced. The corpus-based methods are the origin of the current evolution that is used by almost every one.
5-The positive impact of machine translation
As stated in the introduction as well as the abstract machine translation is a must, because of the lower cost, accessibility, flexibility and speed. It is helping in man ways with the worldwide use of Internet. One of the most common use of it is Google translation which is defined as an automatic machine-translation service provided by Google Inc. It translates one written source language to another directly or with English as a medium (Boitet et al. 2009) And has proven to be a life savior. For instance if one goes to another country and have difficulties finding anyone to communicate with, the person can get to any place where there is a Wi-Fi (Wi-Fi is the same word everywhere so you can read on the door of a Cafe or any place that might have it) and Google what he wants to say even just a key word and get a result within seconds and receive assistance. Nowadays people are keen to travel because machine translation made it easy to break the language barrier.
Content creations have a greater use for it, they can easily translate their content to allow people from many places to interact and share with them. This is a great example of globalization as people have different talent, tips and thoughts to share. Most content creators don't make enough money at the begging of their carrier to hire a human translator anytime they need one. It is not cost efficient to hire a human to translate a content into more than three languages in the least considering that most human translator don't speak more than three languages therefore aren't really flexible. MT researchers are upgrading the list of languages on different applications and devices continuously and it is not easy for a human to learn a new language and master it enough to translate at the same speed.
Human translation is influenced by the characteristics of source-target language transfer, cultural context and individual translator’s translation ability (Bassnett and Lefevere 1992; Wong and Shen 1999) while MT isn't affected by these factors.
6-Comparison of the limitations of both human and machine translation:
6.1- translation affected by culture and emotions
Human translation is influenced by the characteristics of source-target language transfer, cultural context and individual translator’s translation ability (Bassnett and Lefevere 1992; Wong and Shen 1999). Human translation is still required for the good functioning of MT and is still better suited for literary text translation, and sign language. Understanding the cultural background of bothe target language and source can minimize translation errors and avoiding offending one party or the other. Additionally, MT tools have problems translating cultural context such as idiomatic expressions.
6.2- internet dependence of MT unlike Human translation
All the limitations related to MT could not possibly be spoken about this essay hence the author will stick to the relevant ones and will take the Chinese language as an example to provide a clearer view on this. We must start with its dependence on the internet, in underdeveloped countries the access to internet is still challenging, MT depends exclusively on a good network except for the dictionary apps that are available on phones and of course those can only translate words and cannot help much to communicate efficiently while human translator can do the job without the internet. Here are some few limitations of both in relation to specific factors:
Aiken et al. (2009) and Seljan et al. (2011) determine that MT is shown to be domain specific and is especially useful for gist translation, i.e. yielding an understandable meaning even if the grammar was garbled. MT has a gap in translatability when translating untranslatable words. It's is nearly impossible to separate literary translation from humans participation because it it affected by the emotions, the cultural background thus a human translator can accurately give an authentic translated material. Machines can not take into account the differences of background from the source language to the target language and they can't convey the authors message with the emotions and perceptions he or she wants to share with the reader without twisting the meaning. In the article entitled machine translation of literature: implications for translation pedagogy, the authors Abdul Fattah Omar and Yasser A. Gomaa in their case study aimed to evaluate the usefulness of applying machine translation systems to literature with the purpose of identifying the challenges that may have negative impacts on the reliability of machine translation systems. In order to do this they used to stories: Harry Potter by j.k .Rowlings and the black cat by Edgar Allan by comparing their own translation and that of other humans. They concluded that MT was somehow adequate for both prose translation and literary translation, but argued that the use of human intervention, specifically, in regard to translation, can provide, evidently, the most accurate form of translation, with respect to correct translation of idioms, grammar and colloquial understandings. But it still important to point out that human translation highly depend on personal ability thus getting a little assistance from MT might be the solution in terms of high accuracy and completeness.
the main limitation of human translators is irrevocably the problem of time and space. In a globalized world people wish to move freely and communicate efficiently anywhere they go. Human translators are practitioners in their field, they are often working in structures or even if their offering freelance services it is highly focused on official documents translation unlike machines that would translating anything at any given time as long as one has a machine in hand and an internet connection. In some places there aren't many skilled translators so the scarcity of it makes the machines more reliable despite their inability to give a high quality translation specifically in third world countries where freelance isn't even known. For instance if a student goes abroad for studies, there is a content need for translation, firstly because the person might have challenges with the new language and secondly because of the new technical world employed during lectures. It is impossible for this student to get anyone at their disposal to help with communication. Humans have to rest therefore they are not available 24/7 while machines don't need to rest, one is always at ease using them on a phone, a laptop or a desktop in a Cafe. It's only fair to talk about speed, humans cannot translate as fast as machines, they translate within seconds, if a person is given a document to translate same as a computer, there is no way he is going to finish before the computer does. When I comes to flexibly in terms of time and space MT is definitely a more sustainable approach to translation.
6.5-Machine translation for sign languages
There is approximately more than 200 different sign languages commonly use by disabled people like deaf people or the ones that can not speak correctly or completely mute. With the help of neural computing and applications, researchers have compiled sign language or depositories comprising of gestures. Similarly, algorithms have been proposed to translate the natural language into sign language, which is subsequently converted into gestures using avatar technology. The avatar technology has been developed to meet the needs of this minority (around five percent of the global population) to access translation services and have a semi- normal life experience. An avatar is an image that represents one party in a collaborating argument. The avatar may represent a human being since it operates like an agent of any application system used for this purpose. It is a build in element of interactivity and it responds to the user's need and requests by providing a clear insightful rapid link to a database in a simple manner. For example Microsoft has created multilingual avatar-based translations system that translates natural language to 8 different sign language avatars. This shows the strength of MT because many human translators only deal with natural languages and cannot translate into a single sign language let alone many. Although there is a lot to improve in MT in order to help this minority, the possibility of MT meeting their translation needs is far more realistic than humans improving to meet their special needs. The majority of the world populations are able to talk and hear and there isn't enough translators to satisfy the normal needs for translation, hence it will be absurd to put that responsibility on the humans to meet the special need of minority groups. The limitations of MT in language of sign language are the Lack of sizeable data sets and low accuracy. Humans are still pretty much on the top of their game when it comes to translating sign language even if they don't master all the 200 and above models of it.
Machine translation is cost efficient compare to human translation. With Google translate one only need to pay for Internet but the amount of words or sentences translated doesn't affect the cost. Most full automated apps used for translation are paid monthly and do not account for the size of documents whereas human translators charge according to the size and complexity of the document. When a documents contain technical words and calculations people charge more because they are paid by the time and amount of effort spent in doing the work. Nevertheless Human are irreplaceable in infrastructures and official business because MT still has a lot of improvement to do. When official document are translated, they need to be stamped to show where the job has been done. Human translators are better at doing the official jobs because they can adapt the content to grammatical, vocabulary differences between 2 languages. The average price per word translated according to statistics is 0.10$ (us DOLLAR) and the average translation app run by humans cost 20 to 150 dollars per hour. Also when translating a degree there is a need to stamp the document to show how reliable the translator is, the reason being that it is still human that exclusively translate them. Because of the stamp and the certifications and regulations attached to that human translators has to be expensive for they are held responsible for quality and eligibility of their work on an official level. So far machines haven't been eligible to translate official documents or books(which falls into the context of literature). It explains the cost of bestseller books that are translated in different languages to give access to a wider and diverse audience. An author can surely afford a translator but a content creator could not possibly pay 0.10$ per word for his work which lead them to contempt with the poor translation services offered by platforms like Google translation on their literary work.
6-7-Limitations of MT in relation with chinglish7-Limitations of MT in relation with chinglish
There is a huge gap between Chinese language and English when it comes to translating, Chinese is spoken by 1.6billion people around the world and the economical and technological impact of China as a country makes machine translation important for the purpose of buisiness relationships that China entertains with all the other countries around the globe. English and Chinese are currently the most powerful and spoken languages in the world which makes it important to upgrade the translation tools used for the two languages in order to make communication easier. The errors made by MT related to this are: sentence structure ( example: the place of the verb in the sentence),the non-existence of some words like "the " and "is" in Chinese or Chinese vocabulary like dragon, subject-verb mistranslation , difference in grammar and cultural sensibility.
The concept of Chinese English shows the impact of the two influential language. Chinese English can be defined as the mistranslation of the English language that is influenced by the Chinese culture and language. According to Wang (2012), the causes of Chinese English include syntactic transfer from Chinese, influences of the Chinese thought patterns, and inadequate cultural awareness. An simple and accurate example of Chinese English can be found in Jing (2007): “Pulling someone's leg” in English means you are joking with someone, while a Chinese person who is not familiar with the idiom thinks that it literally means someone is trying to pull someone else's leg. In the previous example it's obvious that if one uses MT, the machine will not take the cultural background of the two languages. The competence of professional Chinese translators is required because MT is limited and ultimately people users have to decrease their expectations. You Dao dictionary is the most used MT tool in China by student and few of them have access to Google translation. Students in China cannot use you Dao or even Google translate to translate their academic work without falling into the mistranslation trap or sentence structure trap. The awareness around the cultural background in Chinese language is very important because of the mindset; the lack of cultural awareness is the major problem in English Chinese translation.
7- Conclusion:
Machine translation is not ready to erase humans simply because it needs humans to be updated and meet the needs of users. The data sets used for MT Devices lack vocabulary givin8g that languages change a lot. The above limitations stated include both humans and Machines limitations hence demonstrate that they are complementary and a good way to create a balance between the two would be to have more competent human translators specially regarding the cultural aspect of languages, and have them consistently update the database in order to offer a more accurate database. Human translators must put machines at their service to speed up their rate of work to help content creators and freelancers to touch a wider audience. Additionally by using machine to speed up their work human translators could be more accessible in terms of price and ultimately attract more people, and maintain their role in the world.
8-refrences
Abdulfattah Omar , Yasser A. Goma,.June 2020. The machine translation of literature : implications for translation pedagogy. International Journal of Emerging Technologies in Learning (ijet) 15(11):228
Aiken, M., Ghosh, K., Wee, J., & Vanjani, M. (2009). An evaluation of the accuracy of online translation systems. Communications of the IIMA, 9(4), 67.
Bassnett, S., and Lefevere, A. 1992. Translation, History and Culture. London: Cassell.
Bar-Hillel, Y. 1960. “The present status of automatic translation of languages.” Advances in Computers 1, 91-163.
Booth, A. D. (ed.) 1967. Machine translation. Amsterdam: North-Holland.
Boitet, C., Hervé B., Mark S., and Valérie B. 2009. Evolution of MT with the Web. In Proceedings of the Conference Machine Translation 25 Years On, 1-13. Cranfield: Bedfordshire.
Brazill, Shihua, "CHINESE TO ENGLISH TRANSLATION: IDENTIFYING PROBLEMS AND PROVIDING SOLUTIONS" (2016). Graduate Theses & Non-Theses. 71.
Held, D., 'Democracy, the Nation‐state and the Global System', in Held, D. (ed.), Political Theory Today, Cambridge, Polity Press, 1991, p. 202.Google Scholar
Hutchins, W.J. 1988. “Recent developments in machine translation: a review of the last five years.” In: Maxwell, D. et al. (eds.) New directions in machine translation, 9-63. Dordrecht: Foris.
Hutchins, W.J. 1994. “Research methods and system designs in machine translation: a ten-year review, 1984-1994.” In: Machine Translation, Ten Years On, 12-14 November 1994, Cranfield University. 16pp.
Uzma farooq et al,.June 2020.I Advances in machine translation for sign language: approaches, limitations, and challenges. under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2021
Seljan, S., Brkić, M., & Kučiš, V. (2011, January). Evaluation of free online machine translations for Croatian-English and English-Croatian language pairs. In Proceedings of the 3rd International Conference on the Future of Information Sciences: INFuture2011Information Sciences and e-Society (pp. 331-345).
Wong, D., and Shen, D. 1999. Factors Influencing the Process of Translating. Meta: Journal des Traducteurs [Meta: Translators' Journal] 44(1): 78-100.