Modal markers, aspect and light verb constructions in literary texts as testing ground for the machine translationese hypothesis

PurposeHow closely does the translation match the meaning of the reference has always been a key aspect of any machine translation (MT) service. Therefore, the primary goal of this research is to assess and compare translation adequacy in machine vs human translation (HT) from Arabic to English. The study looks into whether the MT product is adequate and more reliable than the HT. It also seeks to determine whether MT poses a real threat to professional Arabic–English translators.Design/methodology/approachSix different texts were chosen and translated from Arabic to English by two nonexpert undergraduate translation students as well as MT services, including Google Translate and Babylon Translation. The first system is free, whereas the second system is a fee-based service. Additionally, two expert translators developed a reference translation (RT) against which human and machine translations were compared and analyzed. Furthermore, the Sketch Engine software was utilized to examine the translations to determine if there is a significant difference between human and machine translations against the RT.FindingsThe findings indicated that when compared to the RT, there was no statistically significant difference between human and machine translations and that MTs were adequate translations. The human–machine relationship is mutually beneficial. However, MT will never be able to completely automated; rather, it will benefit rather than endanger humans. A translator who knows how to use MT will have an opportunity over those who are unfamiliar with the most up-to-date translation technology. As MTs improve, human translators may no longer be accurate translators, but rather editors and editing materials previously translated by machines.Practical implicationsThe findings of this study provide valuable and practical implications for research in the field of MTs and for anyone interested in conducting MT research.Originality/valueIn general, this study is significant as it is a serious attempt at getting a better understanding of the efficiency of MT vs HT in translating the Arabic–English texts, and it will be beneficial for translators, students, educators as well as scholars in the field of translation.

Most local public health departments serve limited English proficiency groups but lack sufficient resources to translate the health promotion materials that they produce into different languages. Machine translation (MT) with human postediting could fill this gap and work toward decreasing health disparities among non-English speakers. (1) To identify the time and costs associated with human translation (HT) of public health documents, (2) determine the time necessary for human postediting of MT, and (3) compare the quality of postedited MT and HT. A quality comparison of 25 MT and HT documents was performed with public health translators. The public health professionals involved were queried about the workflow, costs, and time for HT of 11 English public health documents over a 20-month period. Three recently translated documents of similar size and topic were then machine translated, the time for human postediting was recorded, and a blind quality analysis was performed. Seattle/King County, Washington. Public health professionals. (1) Estimated times for various HT tasks; (2) observed postediting times for MT documents; (3) actual costs for HT; and (4) comparison of quality ratings for HT and MT. Human translation via local health department methods took 17 hours to 6 days. While HT postediting words per minute ranged from 1.58 to 5.88, MT plus human postediting words per minute ranged from 10 to 30. The cost of HT ranged from $130 to $1220; MT required no additional costs. A quality comparison by bilingual public health professionals showed that MT and HT were equivalently preferred. MT with human postediting can reduce the time and costs of translating public health materials while maintaining quality similar to HT. In conjunction with postediting, MT could greatly improve the availability of multilingual public health materials.

The present descriptive study aimed at investigating the human and machine Persian translations of The Kite Runner by Khalid Hosseini, and comparing the applied translation strategies in the translated texts for culture-specific items (CSI). To this end, based on Newmark’s (1988) category, the applied strategies were identified in the two translations and compared. The obtained results showed that Naturalization and Transposition strategies were the most frequently-used strategies by both human translators and machine translation. The results also showed that machine translation could not present a comprehensible translation due to overuse of these strategies (75%). It was further revealed that the spirit of the original text was not lost in the translated versions due to the closeness of Iranian and Afghan cultures. In fact, the translated versions kept the real beauty and creativity of the original work. However, the remorseful theme of the source text was kept intact to a great extent in the human translation of the novel, while machine translation lost it. Thus, the general impression is that culture-specific terms make it difficult for the machine translation to achieve complete word-for-word and semantic equivalence, and that the human translator must have a broad knowledge of the literature and traditions of both the source and target languages.

Earlier studies have corroborated that human translation exhibits unique linguistic features, usually referred to as translationese. However, research on machine translationese, in spite of some sparse efforts, is still in its infancy. By comparing machine translation with human translation and original target language texts, this study aims to investigate if machine translation has unique linguistic features of its own too, to what extent machine translations are different from human translations and target-language originals, and what characteristics are typical of machine translations. To this end, we collected a corpus containing English translations of modern Chinese literary texts produced by neural machine translation systems and human professional translators and comparable original texts in the target language. Based on the corpus, a quantitative study of discourse coherence was conducted by observing metrics in three dimensions borrowed from Coh-Metrix, including connectives, latent semantic analysis and the situation/mental model. The results support the existence of translationese in both human and machine translations when they are compared with original texts. However, machine translationese is not the same as human translationese in some metrics of discourse coherence. Additionally, machine translation systems, such as Google and DeepL, when compared with each other, show unique features in some coherence metrics, although on the whole they are not significantly different from each other in those coherence metrics.

Poetry is essentially the most condensed and pragmatically loaded form of art. The daunting task of effectively translating poetry requires avant-garde translating acumen. With the advent of neural machine translation, the debate on Human vs. Machine Translation and the speculation of Machine Translation replacing humans went rife. The current study aims to gauge the efficacy of Google Translate’s (MT) rendering of Iqbal’s poem La Illaha Ilallah as compared to Bashir Ahmed Dar’s Human Translation (HT) of the same poem. Bahir Ahmed Dar published his English translation of the chosen poem in his book ‘Rod of Moses’. This study examines the impact of Antoine Berman’s deforming tendencies on the Machine and Human English translations of Iqbal’s poem La Illaha Ilallah. Moreover, the study aims to find out how lexical elements change in the translations of the source language to make it suitable for the target language audience by adopting the deforming tendencies. It is a qualitative type of study employing Berman’s Model of Twelve Deforming Tendencies to analyze which strategies were used in both the English translation of Iqbal’s Urdu poem. The data of the Source Urdu Text was collected from the Internet Archive, and that of Human Translation was taken from Bashir A. Dar’s book Rod of Moses. To generate the Machine Translation, Google Translate was used to generate the MT output. The results indicated that the MT of the work suffered from lexical mismatches, destruction of rhythm and destruction of vernacular networks, rendering an erroneous and vague translation, while HT managed to retain the cultural and contextual essence of the source text. The findings also revealed that the Human translation also faced many deformative tendencies but successfully maintained the genre and social stance of the author. Overall, it appears that poetry translation may benefit from a human translator's profound attention to cultural and contextual detail, which is ignored by NMT tools such as Google Translate, leading to inconsistencies in the translation.

Systematic comparison between machine translation (MT) and human translation (HT) is mostly limited to the evaluation of MT output with HT as reference, as opposed to a more general study of the properties of MT and HT output texts. We present preliminary experiments investigating dierences between MT and HT in terms of readability and language complexity. We perform both quantitative and qualitative comparison of the outputs of machine and human translation, using samples of English text across multiple domains and genres and their Hungarian translations created by humans and by the state-of-the-art machine translation system deepl. Our results show that machine translation produces somewhat simpler text than human translation on 3 out of 4 samples, and on 2 samples this eect is caused primarily by human translators using a higher number of complex words. We release all software used in our experiments to facilitate further studies on larger samples, additional languages and domains, and using alternative MT systems.

There seems to be a consensus that register-specific, informative texts are more suitable for automated machine translation, while form-focused texts are less so. Since there are limitations in machine translation in terms of communicative and translation competence, texts in which linguistic form, pragmatic meanings, connotations, and culture-specific elements play an important role, in addition to content, are more difficult for machine translation programs to cope with. In this paper, I will attempt to demonstrate the relevant differences that arise in the process of machine and human translation by comparing a neural machine translation (by DeepL) of a literary text (“Fatelessness” written by Imre Kertész) with the human translated text (by Tim Wilkinson). The ultimate goal of my research is to gain more insight into the quality of Hungarian–English machine translation, how corpus linguistic analysis of the source and target languages can be of further use, and what are the limitations (if any) of the use of machine translation in the translation and post-editing of literary texts.

Machine translation has witnessed great development in the recent decades and we have entered the era of neural machine translation (NMT). A review of MT is necessary for a better understanding of the relationship between MT and human translators and translation teaching in this era when MT has flourished. This paper first briefs the machine translation (MT) development in the past decades, focusing on the features, application, and drawbacks of each main paradigm of rule-based machine translation (RBMT), corpus-based translation (CBMT), and long-short term memory (LSTM), a main paradigm of NMT. It continues with a discussion of what MT means to human translators and translation teaching in universities. It concludes that MT should not and could not replace human translators which will always be vital in some fields and aspects; only a good integration between the two can ensure satisfying output with post-editing by human translators to meet the increasingly demanding market. This signifies that translation teaching in universities should embrace MT knowledge.

Translation has a significant impact on the process of intercultural communication, international business, and transfer of knowledge. In recent decades, major advancements in machine translations (MT) have been experienced especially with the emergence of neural machine translations (NMT), which has significantly enhanced significantly the levels of fluency, grammar, and accessibility. Whether or not MT can maintain contextual accuracy, particularly in areas where subtlety and idiomatic phrasing, emotional appeal and cultural sensitivity are more important, still hangs. In this paper, a comparative analysis of human and machine translation is provided mainly on the contextual fidelity. A combination of the mixed-methods development process included a corpus of legal texts, literary texts, medical texts and marketing texts in Arabic, French and Japanese, which were translated by professional human translators, and the most popular systems of machine translators (Google Translate, DeepL, Microsoft Translator). A five-criterion evaluation system, that is, semantic fidelity, cultural appropriateness, idiomatic accuracy, emotional tone, and grammatical correctness, was used to evaluate translations. Objective data also reveal that, although MT scores close to human functional levels of grammar and fundamental semantics faithfulness, it perpetually performs poorly when it comes to the expression of idioms, tonality, and cultural subtext. Qualitative results also support the fact that MT cannot process light contextual clues, including irony, formality hierarchies, and culturally associated metaphors. In comparison, human translators are better at cultural adaptation as well as cost-effective and scalable. The research finds that hybrid systems in which MT delivers initial translations that are further processed with the assistance of human post-editing are the way to go. The implications of these findings for the research of translation, AI ethics, and professional training are immense, as they can confirm that there is ongoing relevance of human judgment in a translation field that is becoming more automated.

Since Google introduced the Transformer model into natural language processing (NLP) in 2017, AI-aided translation has rapidly advanced. At the same time, translation is evolving from a solitary endeavor into a cooperative activity between human translators and machine translation systems, epitomized by the emergency of platforms with the Machine Translation Post Editing (MTPE) function. The advent of new translation modes also leads to increased research evaluating the effectiveness and quality of machine translation, for example, studies on the translation quality under the Multidimensional Quality Metrics (MQM) error typology framework. Involving AI-based translators and MTPE in their translation enables human translators to prepare the engineering documents efficiently. However, researchers notice that it is difficult for most machine translators to figure out the semantic and cultural differences in the source language and generate coherent structural translation in the target language. This research opens up ChatGPTs application in tender document translation under the MQM framework, hoping to cast light on assessment on ChatGPT's translation quality, identification of ChatGPT's errors in translating such documents and suggestions on human translators' performance throughout MTPE.

In this research paper, we investigated the viability of AI-supported translations of survey materials in intercultural and cross-cultural research, comparing the quality of machine translations to traditional human translations. Focusing on the HEXACO personality inventory, we translated the original English inventory using Google Translate and GPT-3.5 into 33 languages for which validated human translations exist. The statistical similarity between human- and machine-generated translations varied considerably between the target languages. It was highest for target languages from the same language family as the source language, arguably because this relatedness allowed for more direct machine translations. Consistent with this reasoning, the genetic similarity between languages largely explained the differences observed. GPT’s temperature setting determining how stringently or freely a text is translated had little influence on the similarity estimates, but very high levels tended to produce somewhat lower statistical similarity. To validate the quality of the machine translations, a group of social scientists rated the translation in a language for which the human and machine translations statistically converged strongly. Although the human translation was rated as being of higher quality than four out of five machine translations, these differences were relatively small. Crucially, the social scientists did not rate the human translation as significantly better than the GPT 3.5 translation with the lowest temperature setting. Based on these insights, we propose a framework outlining four recommendations for utilizing AI-supported translation in cross-cultural and intercultural research, involving AI to varying degrees in the forward-back translation process.

Translating culturally loaded texts is considered a daunting challenge in the translation field. While machine translation is adept at linguistic aspects, it often struggles to convey the nuanced cultural features inherent in cultural texts mainly traditional Arabic proverbs and anecdotes under study in this endeavor. This emphasizes the vital role of collaboration between machine efficiency and human expertise in ensuring accurate and authentic translation. This study explores the intricacies of translating culture-rich content by comparing between machines flaws even when assisted by artificial intelligence, and human translators’ expertise limitations. The comparison has revealed weaknesses in both machine and human translations, prompting the call for collaborative approach between the two methods. This study measures the limits of human translators with computer assisted translation from Arabic to English; first comparing between them then combining machines and humans performances. The translated case studies have proven the efficiency and accuracy of collaborative translation method. The aim is to optimize translation of cultural sensitive content, preserve cultural integrity and elevate authenticity and quality. The findings inform the discourse in translation studies and cross-cultural communication, offering practical insights to enhance translation quality and authenticity based on collaboration among machine translation developers, linguists, and cultural experts. As AI progresses, the synergy between machines and human translators shows potential to exceed current capabilities, facilitating more effective communication in diverse contexts.

Recent advances in neural machine translation have significantly improved translation quality, yet its ability to handle syntactic complexity in literary texts remains underexplored. This study examines syntactic differences between human and machine translations of a Chinese literary text from the perspective of mean dependency distance. Drawing on one human translation and four machine-generated translations, the analysis compares dependency distance patterns and investigates how sentence length relates to differences across translations. The findings indicate that although both human and machine translations show a general tendency toward syntactic simplification, notable divergences persist between human and machine output. These divergences are unevenly distributed and are closely associated with sentence length, especially in longer sentences. The study suggests that sentence-level restructuring constitutes a key distinction between human and machine literary translation and remains a challenge for current machine translation systems.

This study aims to remark the differences between human translation (HT) and machine translation (MT) on linguistic, cultural, and stylistic levels when translating English literary texts into Arabic. To accomplish the goal of this study, a comparison between the Arabic HT and MT of Saki’s (1914) short story ‘The Open Window’ is conducted. The study focuses on comparing the two translations (HT and MT) on linguistic, cultural, and stylistic levels to identify the differences between HT and MT in translating literary texts. Throughout this comparison, it is found out that both HT and MT have their advantages and disadvantages on different levels. It has also been found out that MT is unable to identify cultural items and consequently mistranslate them. It is, therefore, concluded that MT can work proficiently on certain levels besides the intervention of the human mind. The findings of this study provide translators using MT with a clear vision on the points of strength and weaknesses in translating literary texts.

Technological advances have made the ability to translate no longer exclusively belong to humans. Today, machine translation has turned into a tool with superior performance to convert text between languages without the need for human intervention. One of the translation research foci is the studies of causative translation, especially from English to several other languages. Yet, it might be interesting to compare the translation of that topic by human and machine translation. This study investigates the comparison of Google Translate and humans in terms of causative translation from English into Indonesian. The data were obtained from six English novels and their translations in Indonesian. To analyze the data, 100 clauses with causative have and get were selected from English novels and translated by Google Translate into Indonesian. The result showed that the translation and strategies used between Google Translate of causative have and get had similarity with human translation in relation to causative-to-causative translation. Through the investigation, the result is expected to be beneficial for further studies in the translation of causative have and get related to their translations into Indonesian analytic or morphological causative. Furthermore, the result of strategies compared is expected to be beneficial to the translation study regarding machine and human translation in causative, especially from English into Indonesian.