1. Abstract
The primary aim of this research was to determine whether full post-editing of raw machine translation output, and the subsuming machine translation, is quicker compared to human translation. A group of 10 professional translators (N=10) were randomly selected for the project and interviewed on their experiences and views on bothy machine and human translation especially with focus to the primary aim of the research. It was predicted that, indeed, machine translation is faster, but not always efficient given that not all contexts of translation require speed. Rightly so, 8 out of the 10 participants contended that machine translation is faster, but acknowledged that this was mostly laudable when dealing with text translations like books. This can be interpreted to mean, that faster does not always imply quality especially with regards to oral or conference translation. The 2 two remaining respondents, all human translators, had no sufficient experience with machine translation thus could not offer any comparative insight between the two. Regardless, all the 10 participants agreed that quality of translation matters more than its duration. It is also noteworthy that this assignment was done in a group of four students. Since we conducted thorough research on the topic, we chose a group of 4 so as to assign a specific part of the assignment to a team member. In this manner, we would induce diversity of opinion and inculcate the culture of teamwork, thus, easing the entire research and assignment completion.
2. Introductory underpinning of the project
Translation entails transmitting source language text (or information) into target language text with close attention to both the linguistic and cultural differences in the languages. Translation dates back to the 1940s during the World War II when translators were mainly required for translation of spying documents. After this epoch, translation continued to gain momentum and its importance increased especially in the field of economics. All this time, human translation was the most dominant form of translation before technological advancements of the 1940s facilitated creation and adoption of machine translation. Research proposals that paved way for machine translation were founded on the successes of World War II code breaking as well as projections about the universal principles that undergird natural language. Factors that fuelled the quest for, and subsequent adoption include time, accuracy, and general translation expenses. The emergence of machine translation led to speculations that human translation may soon subside and machines will dominate the field of translation and interpretation. Yet human translation persists on. In fact, it is the most preferred translation technique in conferences-conference translation. On this note, it is important to understand the chief tenets of both human and machine translation with a view to gaining a comprehension of the reasons that have made human translation persist despite the high efficiency and effectiveness of machines and machine translation. Therefore, the current study serves to help in the understanding of the dynamics of both machine and human translation with a focus to the reason why both stay on despite the efficiency often associated with machine translation.
3. The Undergirding literature framework
Li, Graesser and Cai (2014) conducted a research comparing Google translation with human translation. It is important to note that Google Translate, a translation machine, provides multilingual translation services through automatic translation of a source language to a target language. The major focus of the study was to investigate the accuracy of the machine with respect to Chinese-to-English translation and consideration of cohesion and formality which are the two primary elements of a quality translation. The text sample consisted of 289 written and spoken texts from particular works of Mao Zedong both in Chines and English version with the machine meant to translate the sample to Chinese from English. The texts were then examined by the use of such instruments as the Chinese and English Coh-Metrix as well as the English and Chinese LIWC. The research found a high correlation between Google English translation products and both the Chinese texts and human English translation in terms of cohesion and formality. Martinez (2003) conducted a relatively similar research earlier and drew the same conclusions- machine translation is generally faster than human translation. However, the researcher noted that, besides speed, the application of machine translation is constrained given the existence of a wide range of conditions required for the effectiveness of post-editing (PE) of raw machine translation output. Some of the dominant conditions included features and translation environment of PE, the machine translation approach, the dictionary preparation time, and the entire translation process. According to the researcher, a high percentage of efficiency of machine translation is only possible through strict observation of these factors. The crowning finding of the research, however, was that machine translation systems are faster especially for translation of marketing brochures, but the efficiency was relatively low in the translation of scientific documents and technical manuals that were highly repetitive.
Ulitkin (2011), on the other hand, differs from the two previous studies by comparing outputs from two machine translation instruments: Google and PROMPT. On the overall, Google output showed a high correlation with the text of reference as compared to the output from PROMPT. However, the researcher also notes that Google showed the worst grammatical results compared to PROMPT. These findings illustrate the differences between machine translation systems and the variations in their operations and functionalities. Ulitkin (2011), however, notes that the Google’s translation potential is constantly improved with a view to improve the translation results. While he fails to downplay the need for a fully automated machine translation system, he concludes that despite the ongoing progress, fully automated translations will still be way less than perfect. All these three studies undoubtedly agreed with the hypothesis of this research.