An-Najah National University Faculty of Graduate Studies Machine Translation Limits of Accuracy and Fidelity By Shaymaa Yousef Ibraheem Abdulhaq Supervised by Dr. Nabil Alawi This Thesis is Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Applied Linguistics and Translation Program, Faculty of Graduate Studies, An-Najah National University, Nablus, Palestine. 2016 iii Acknowledgement The researcher acknowledges and appreciates the incredible efforts and guidance of Dr Nabil Alawi. The researcher would also like to seize this opportunity to thank her family, relatives and friends for their unconditional love and support. The researcher would also like to thank Dr. Roqayia Herzallah and Dr. Omar Najjar for their valuable suggestions and advice. v Table of Contents Content Page Acknowledgement iii Declaration iv Table of Contents v Abstract vi Chapter One: Introduction 1 Statement of the Problem 2 Questions of the Study 3 Significance of the Study 3 Limitations of the Study 3 Methodology 4 Chapter Two: Literature Review 5 Chapter Three: Data Analysis 30 Errors and Reasons 31 1. Capturing the Different Senses 32 2. Everyday Language and Culture 36 3. The Implied Meaning vs Literal Meaning 40 4. Fixed and Context Free Expressions 42 5. Unit of Translation 46 6. Grammatical Structure 48 Suggested Translations and Strategies 50 1- Capturing the Different Senses 51 2- Everyday Language and Culture 53 3- The Implied Meaning vs Literal Meaning 55 4- Fixed and Context Free Expressions 57 5- Unit of Translation 59 6- Grammatical Structure 62 Best Uses of Machine Translation 63 1- Hints to the Implied Meaning 64 2- Conveying the Message as it is 66 3- Scientific Expressions 67 4- Unit of Translation 68 Chapter Four: Conclusions and Recommendations 72 Works Cited 75 ﺹﺨﻠﻤﻝﺍ ﺏ vi Machine Translation: Limits of Accuracy and Fidelity By Shaymaa Yousef Ibraheem Abdulhaq Supervised by Dr. Nabil Alawi Abstract While we grow to be more dependent on machines in all our walks of life, using machines in translation has become a remarkable challenge. In this study, the researcher has analyzed some texts and their translations which are produced by Google Translate. The researcher has categorized the translations according to a group of markers and indicators: the implied vs literal meaning, capturing the different senses, everyday language and culture, fixed and context free expressions, grammatical structure and unit of translation. The researcher has reached a number of recommendations on how to enhance the use of machine translation. The translation movement is moving towards utilizing the use of machine programs in translation effectively; there is no turning back even though the human intervention in translation is still fundamental. 1 Chapter One Introduction 2 Chapter One Introduction When one is sitting before his/her computer and wants to translate a letter from one of his/her friends from another continent or perhaps one is assigned a homework to translate a long text or maybe one works as a professional translator; one is commissioned to translate an article or a report of many pages, the first thing that occurs to one's mind if s/he is in any of these situations is to use Google Translate. Of course, it would be absurd to use a dictionary in these situations because one will be wasting minutes, perhaps hours, to find the proper word especially since wasting those minutes is considered something one cannot really afford to lose in this age of unstoppable force of speed. Statement of the Problem The thesis addresses the problem of the mistranslation of expressions especially idiomatic expressions from English to Arabic while using online machine translation programs such as Google Translate. While the work of the machine is still far from being adequately accurate, the task of translators is to find ways to boost the role of the machine in translation by giving certain recommendations on the procedural aspects since the technical side is not within the domain of translators. By highlighting their needs and expectations, translators can alert programmers to such needs so that they can be improved. 3 Questions of the Study The thesis tackles and answers three main questions. First, what are the main errors that occur while using machine translation to translate cultural expressions from English to Arabic? Second, what are the factors that lead to these errors? Third, what solutions or strategies can be used to help solve these problems? Finally, how can the use of machines in translation be optimized on the side of the translator? Significance of the Study The topic of this thesis is important for many reasons. First, it would give an insight on the most common errors that occur when we as translators use machine translation to translate cultural expressions. Second, it would shed a light on the important suggestions or advice on how to deal with such errors. Third, it would enhance the translators' skills and abilities on how to deal with machine translation, by doing so, it would also enhance their translations, as well. Limitations of the Study The thesis has certain limitations and one of those limitations is the focus on Google Translate since it fits the purpose of the thesis which is to help benefit not only novice translators but also professionals. Both of them use machine translation especially Google Translate; the reason why it is used a lot is because it is an easily accessed webpage; it doesn’t cost a thing and it is always available and under the disposal of translators 24 hours. It 4 is probably the most widely used tool for it’s the first thing that crosses the translator's mind when s/he wants to know the meaning of a word quickly especially if s/he is reluctant to use other means or if s/he is in a hurry. By doing this research and reaching the answers to the research questions, the researcher hopes to guide the translators on how to deal with machine translation and what to do in cases where there are errors in the translation of idiomatic expressions while using machine translation. This study is limited to the translation between Arabic and English. Methodology The method follows a descriptive and analytical method of the source text (English cultural expressions) and the target text (the translated expressions into Arabic on online machine translation by using Google Translate). The method includes human intervention; when the human translator interferes in cases where machine translation can be useless in producing certain translations or the wanted effects that could help locate errors, and predict their reasons.