The Impact of Digital Technologies on the Structure of Translation Activities

Gao Tszyai Postgraduate Student at the Higher School of Translation and Interpretation, Lomonosov Moscow State University 119991, Russia, Moscow, Moscow, Leninskie Gory str., 1, p.51,, of. 1150 jiayi.gao@yandex.ru

Speaking about the structure of translation activities, first of all it is necessary to consider in more detail the concept of "activity". According to the Large Psychological Dictionary, "activity (Eng. Activity, it. Te tigkeit) is an active interaction with the surrounding reality, during which a living being acts as a subject purposefully influencing an object and thus satisfying its needs" ^{[3, p. 135]}. Activity is defined as "a way of reproducing social processes, self-realization of a person, his connections with the surrounding world" in the Modern Philosophical Dictionary ^{[17, p. 186]}. From the above definitions it can be seen that activity is the activity of a person, a certain way of self–realization of a person, his interaction with the world.

Translation is one of the most important types of human activity in the social sphere. The question of the social purpose of translation is directly related to the question of determining its essence. As O. I. Kostikova writes, "translation is the transfer of information contained in the original message from one communicant to another, involving the transition from one code to another" ^{[14, p. 18]}. O. I. Kostikova notes that "the transfer of information presupposes the presence of a certain channel, i.e. a certain material basis for the transfer of information O. I. Kostikova argues that the use of one or another channel of information transmission significantly affects all aspects of translation activities, including translation technology and strategy. In addition, if we consider the history of translation from the point of view of the information transmission channel, it is possible to see a number of separate periods, each of which was dominated by one or another channel" [Ibid.].

Nevertheless, emphasizing the importance of studying the experience of the history of translation activity, O. I. Kostikova believes that "no matter how the translation technology is improved, no matter what new actors are included in the translation process, each translator again and again solves the tasks that Cicero, Jerome, Dole, and Chukovsky thought about, and Benjamin, and Berman, and many other writers, translators, philosophers who turned to translation" [Ibid., p. 21].

We can consider a person's need to influence a certain object as a motive. A translator satisfies a social need through his work, in which he is guided by socially prescribed motives. Translation researcher A. F. Shiryaev notes that "the social environment through social sanctions usually compels the translator to return to the framework of his social role" ^{[20, p. 154]}.

Translation can be considered both as a translation activity and as a result of such activity. "The very term "translation" in linguistic theory correlates with two independent understandings-the translation process as an activity and the translation text (as a result of the process and activity," L. S. Barkhudarov notes ^{[2, p. 5]}. As V. N. Komissarov writes, "translation", in the sense of the product of human activity, is a text in oral or written form; "translation", in the sense of the translator's activity, is the process of creating such a product ^{[12, p. 135]}. In this article, "translation" is mainly considered as the activity of a translator aimed at translating.

Modeling of translation activity is an actual object of research in the field of translation theory. Thus, N. K. Garbovsky and O. I. Kostikova proposed a model of translation activity based on the actant Greimas model (Fig. 1).

Figure 1. The actant model of Greimas [Greimas, 1966: 180]

According to the researchers, the translator's activity can be represented in the form of the following scheme: the subject is the translator; the object is a dual subsystem consisting of the original text and the translated text as a transforming substance; the addressee is the audience of the recipients of the translation; auxiliary means are favorable factors such as digital technologies, information technologies, social professional networks, communication technologies the opponent is negative factors of various origins that hinder translation activities, for example, censorship, limited access to information, technical failures, cultural differences, etc.; the customer, publisher, company, social institutions, individuals are the addressee who orders and evaluates the result (Fig. 2) ^{[8, pp. 29-30]}.

Figure 2. Actant model of translation activity N. K. Garbovsky and O. I. Kostikova

As mentioned above, translation can be considered as a certain type of activity.

And any activity consists of "the goal, the means, the result and the process of activity itself" ^[15].

Thus, this model of translation activity can be interpreted as the fact that the addressee (speaker or writer) creates a text, and the subject of activity (translator) transforms the text into a translation text that the addressee can listen to, read and understand, but in the process of transformation, the subject (translator) is influenced by auxiliary means and obstacles, which either help or hinder him in this activity.

The purpose of translation activityAccording to Ya. B. Emelyanova, the purpose of translation activity is "the production of speech utterances according to a certain social order" ^{[9, p. 94]}.

Sharing this point of view, we note that translation is not carried out "by itself"; any single translation has a clearly defined purpose.

Depending on the place, topic, language pair (source language and target language), the translation can be performed in full, abbreviated, abstract, in the form of a retelling, etc. The translated text contains not only superficial information, but also extra-textual (extralinguistic), cultural, ideological and artistic information.

The subject of translation activityAs it was shown above, the subject of translation activity is the translator.

Today, digital technologies are being introduced into the field of both written and oral translation, giving energy for updating the traditional model of translation activity. New forms of translation activity are emerging, driven by technology and supported by big data. Therefore, we consider it necessary to add a new element to the traditional model of translation activity (Fig. 2) – "digital technologies". However, a logical question arises: where exactly should this element be added? The question immediately arises: will artificial intelligence be able to become an independent actor of the actant model of translation activity? One hundred percent subject? What role will artificial intelligence play?

In our opinion, in order to try to answer this question, first of all it is necessary to consider some characteristics of the translator as a subject of translation activity.

According to Y. B. Yemelyanova, "the personal characteristics of the translator are focused on:

1) flexible and plastic mental organization (the ability to quickly switch attention, switch from one language to another, the ability to concentrate, to mobilize their full potential);

2) breadth of interests, the ability to constantly enrich their knowledge;

3) the moral and ethical component associated with responsibility for the quality of their work; objectivity, reliability, tact, the ability to keep a secret;

4) the ability to hold on confidently, perform in front of a large audience of listeners" ^{[9, p. 95]}.

Such translator characteristics as "the ability to quickly switch attention from one language to another" and "the ability to constantly replenish their knowledge" are now inherent in the machine; online translators, especially those working on the basis of neural networks, keep up with the times, constantly developing and modernizing, and each subsequent version of online translators translates everything more precisely. However, the machine does not correspond to the third point from the list of characteristics proposed by Ya. B. Yemelyanova: there is a risk of data leakage and there is no responsibility for the quality of translation. In this aspect, a person is more reliable than a machine.

In addition, according to E. R. Porshneva, the global function of translation activity is an "interlinguocultural (interlanguage and intercultural communicative) function" ^{[16, p. 28]}. Therefore, it is important for the translator "not just to convey the main idea of the utterance, but to penetrate into the deep meaning of the word, often hidden behind the formalization of use, its national and historical development fund" ^{[4, p. 17]}. In this regard, as Y. B. Yemelyanova notes, the following set of skills comes to the fore: "a) the translator's ability to analyze the context; b) the ability to interpret the text; c) the ability to "see" and take into account in translation what is implicit in the original; d) the ability to compare the pre-information reserves of communicants and when the need to adjust the translation taking into account knowledge about their culture; e) the ability to compare their culture with a foreign language" ^{[9, p. 94]}. By implicit meaning, according to N. A. Talalaeva, we mean "a part of the integrated meaning that has not found expression with the help of formal linguistic means" ^{[19, p. 19]}. In order to determine the implied meaning and understand the components of the meaning of a word, it is necessary to know the culture of the corresponding language. The machine is not capable of this.

Overcoming differences between cultures is one of the most important tasks of the translator. N. K. Garbovsky and O. I. Kostikova, describing translation strategies in this case, when the translator was faced with this kind of task, indicate that "... the translator was offered either to bring the reader closer to the author and introduce him into the world of "alien", alien reality, or invite the author to the reader and make him understandable and accessible, i.e. "his own"" ^{[8, p. 30]}.

Wang Xinyuan, agreeing with this point of view, tries to model it by referring to the story of two foreign missionaries, Lin Lezhi and Jia Yuehan, who lived at the end of the Qing Dynasty. They translated "Education in Japan" as "The rise of the country through literature", changing "In Japan" to "the rise of the country". If the translators had chosen a different version, it would have been difficult for Chinese readers to show a desire to get acquainted with the situation in Japan at this special time for China (the Second Opium War was going on in China at that time) (The Second Opium War was the war of Great Britain and France with the Qing Empire, which lasted from 1856 to 1860. October 24-25 , 1860 was signed The Beijing Treaty, under which the Qing government agreed to pay Great Britain and France 8 million liang indemnity, open Tianjin to foreign trade, allow the use of Chinese as labor (coolies) in the colonies Great Britain and France. The southern part of the Jiulong Peninsula was transferred to the UK from that moment on.

The given example shows that "human contribution is of a cultural nature and has a vitality that goes beyond the existence of individual lives" ^{[6, p.11]}

According to Wang Xinyuan, there are two structures of the social role of a person: firstly, it is consciousness and role psychology of a person, and secondly, it is the role behavior of a person. "Role behavior is the actual activity in which a person is involved during the performance of role—playing duties. The fulfillment of a person's social role is always accompanied by thinking, emotion, etc. And artificial intelligence, no matter how developed it is, is only a product developed by a person, and its ultimate goal is to serve a person" [Ibid., pp. 12-13].

Thus, we believe that digital technologies and artificial intelligence will not be able to become an independent actor of the model of translation activity, and also cannot become a full-fledged subject. In the structure of translation activities, they can only be assigned the place of an auxiliary actor.

An objectDefining the object of translation activity, we present the point of view of N. K. Garbovsky.

The researcher writes that "the object of translation is speech works of different content, different speech genres, differing in the degree of the author's individual originality of speech forms." ^{[7, p. 69]}.

Motive

According to the definitions we gave earlier, any activity arises from a certain need. The translator satisfies the needs of society through his activities, and "at the same time he is guided not by a personal motive, but by a motive prescribed to him by society. Translation satisfies the constantly emerging need for communication between people separated by a linguistic-ethnic barrier" ^{[5, p. 63]}. In the era of digitalization, the Internet and information and communication technologies are developing rapidly, making translation practice more convenient. Intelligent machine translation has radically changed the system of translation technologies, stimulating the transformation and modernization of the translation industry, as well as the development of models of translation activities in the direction of automation. Automation of translation practice models has led to increased interchange between multiple languages and cultures, and digitalization has further strengthened the process of global integration.   

According to the US news agency CISION PR Newswire, by 2021 the value of the global linguistic services market reached 62.6 billion US dollars. In the future, according to the organization's forecasts, by 2027 the market will reach 96.1 billion US dollars, demonstrating a cumulative average annual growth rate of 7.5% in 2022-2027 ^[27]. As the demand for translation services increases, the variability of information content storage also increases, including, in addition to plain text, a high proportion of hypertext containing audio, video, images, websites, programs, video games, graphics, animation and other types of information that cannot be stored on paper. In our opinion, translation is no longer limited to working with text, and translation tools no longer function by themselves, but are integrated into the customer's global content management system using technology.

In addition, new business principles and simultaneous launch of products around the world impose requirements on the timing of translation, and real-time translation allows customers to simultaneously access information in different languages. The changing requirements of the translation community force the translation industry to constantly look for new modes of translation work that meet modern requirements, for example, such characteristics as large volume, technification, real-time work. The translation automation model has become the dominant mode of operation in the translation industry due to its unique advantages in terms of efficiency, cost and scale.

According to statistics, in 2016 the volume of the machine translation market amounted to 400 million US dollars, and by 2024 it is expected that it will reach 1.6 billion US dollars and will grow at an average annual growth rate of about 19.0% over the next few years ^[28].

According to Wang Xinyuan, machine translation will continue to dominate in such areas as for general texts, everyday communication, product specifications, etc. ^{[6, p. 12]}. These areas do not require exclusively professional and accurate translation, and from the point of view of quality, modern machine translation is sufficient. In the future, the translator will focus on market segments that require greater accuracy, quality or literary translation.

All things change with time, and both Eastern and Western philosophers have talked about this idea. Confucius said “("Longyu·Zihan" (*)” (Our translation: Gao Chiayi: Time runs non-stop like water, day and night.) Chuang-Tzu also says in "Zosui” that “(Chuang Tzu · Qiushui (*) (the famous book of proverbs, written in the III century BC and named after the author; one of the fundamental texts of Taoism, from the Tang era is known as: "Nanhua Zhenjing")" (our translation by Gao Jiayi: Everything is born and develops like a horse galloping, or a cart rushing forward. Movement is impossible without changes, just as time is impossible without movement). Both of these statements are similar to the words of Heraclitus of Ephesus - Everything flows, everything changes (From ancient Greek: Panta rhei. Literally: Everything is moving).

We can imagine that there is a two–dimensional coordinate system (Fig. 3), where the horizontal axis is space and the vertical axis is time, and place the aforementioned model (the actant Greimas model) in this system, presented in Fig. 4.

                                                         Figure 3. Two-dimensional coordinate system

 Figure 4. Translation activity: actant model

Auxiliary tools

Auxiliary tools also change over time, due to the increase in the volume and complexity of the translated content (object of activity).

In various fields, from the initial everyday communication, documentation, instructions to software localization, translation of large contracts and regulations, these tools have also evolved, going from the emergence of papermaking technologies, printing, thick and heavy paper dictionaries, the emergence of research centers, to information and communication technologies, "reservoirs of knowledge" (electronic libraries, electronic resources), automated translation, artificial intelligence, etc.

The process of translation activityDigitalization has profoundly influenced translation activities.

As N. K. Garbovsky writes, "increasingly, written text is perceived, including by a translator, not from the pages of a paper book, but from digital sources, i.e. in the form of text on electronic media after its digital processing" ^{[7, p. 69]}.

These two heterogeneous principles of translation are inextricably linked: "creative" and "automatic". V. N. Komissarov, speaking about the translation process, emphasizes that "from the point of view of the translator's behavior, translation is a heuristic process during which the translator solves a number of creative tasks using a certain set of techniques" ^{[10, p. 57]}. In the process of developing the professional skills necessary for translation, according to V. N. Komissarov, "a kind of linguistic personality is formed, which has a number of differences from a "normal", non-translational personality. These differences are revealed in all the main aspects of speech communication: a) linguistic; b) text-forming; c) communicative; d) personal; e) professional-technical" ^{[11, p. 275]}. L. S. Barkhudarov notes that "the creative nature of translation activity consists in the ability to find individual, single, interlanguage correspondences"not provided for by the theory" ^{[2, p. 78]}.

Hutchins and Summers (Hutchins and Somers H.) use Figure 5 to represent the degree of automation of translation technology. In the initial technological vision, attempts were made to create a completely high-quality automatic translation that would completely replace a person in the center of translation activities. However, technical and theoretical difficulties led to the rejection of this idea and the transition to a computer-assisted translation model in which a person and technology cooperate, and which can be divided into machine translation using a person and machine translation, depending on whether the text is edited before or after translation ^{[21, p. 148]}. In this model, a person remains a subject, and technology is used as an auxiliary tool for facilitating practical activities.

Figure 5. Degree of translation automation according to Hutchins and SummersIf a person makes decisions in the process of translation, although translating using digital technologies (such as information and communication technologies, automated translation, post-machine editing, etc.), then this process is still a creative activity.

If a translation is made entirely by a machine without human intervention (including without post-machine editing), we cannot consider it a creative activity. Since the machine (artificial intelligence) uses the same algorithms that do not depend on place and time, unlike a human translator, it cannot make decisions and make choices.

This fact can also be used to explain the difference between the work of a machine and a person during the processing or perception of the source text. Any act of translation begins with the perception of the original message. "The process of "entering" into the translation process begins with the student reading, since at this stage vocabulary, structural and semantic features of individual sentences, the semantic structure of the entire text are recognized," writes E. G. Korotkov ^{[13, p. 52]}. The perception of the initial test is carried out by a person using linguistic competence and the amount of background knowledge. These factors directly affect the quality of translation. "The higher the level of the target language of the translator, the freer he gets the translation" ^{[22, p. 35]}, and the machine (AI), performing translation using algorithms or neural networks, analyzes the source text, while the quality of the translation product is affected by the quality and quantity of parallel language corpora.

In addition to "perception", the translation process includes "transformation" (decoding) and "generation".

Within the framework of the traditional model of translation activity, a person independently passes these phases without the help of technology. However, according to the new translation model, the translation process looks different. The transfer is carried out in three stages:

"1. getting the text and determining the volume of the text;

2. preparation of the text for translation by means of automatic or automated translation;

3. post-machine editing.

The translator must take into account all the circumstances when developing a translation strategy. At the same time, for each translator, the share of the use of digital technologies (automatic translation and automated translation) will be a purely individual indicator" ^{[18, p. 73]}.

I. S. Alekseeva defines "strategy" as "an algorithm of translation actions consciously chosen by the translator when translating one particular text (or group of texts)" ^{[1, p. 323]}. "At the same time, translation actions are understood as the whole "set of possible actions for the implementation of translation"" [Ibid., p. 324]. Translators need to develop a strategy when working with software or other tools.

In other words, when translating a text or a group of texts using automatic or automated translation tools, the translator must decide which software or other digital tool is most suitable for translating a text or a group of texts. It should be remembered that a wide variety of electronic resources not only do not help, but on the contrary, they complicate the translation process. Digital technologies exist only to help the translator make a choice, the final decision is always in the hands of the translator.

As we noted earlier, with the development of digital technologies, the model of translation activity is developing towards automation and the emergence of new forms of translation activity. We will outline possible ways of using translation software to form a translation strategy based on the logical sequence of actions of the translator.

Form 1?Automated Translation (CAT) + Neural Network Machine Translation + Post-machine editing

 According to Li Zheng, "whether it is automated translation or machine translation using neural networks, the ultimate goal is to implement translation automation" ^{[26, p. 70]}. The rapid development of artificial intelligence technologies provides support for technologies to achieve the ultimate goal, and there is a tendency to merge them on the way to translation automation. The largest translation service providers are creating platforms that combine machine translation technologies with their own accumulated corpus to provide low-cost, high-quality and fast translation services for the market. For example, the Directorate General for Translation of the European Union, which is based on neural network machine translation in combination with its own Euramis translation memory and the TATE terminology database, has created the E-Translation 2017 platform, which allows translating into 600 languages, including 24 official languages of the European Union. Currently, the main neural network machine translation systems, such as Google, Baidu, IBM, WorldLingo, etc., provide open APIs (Application programming interface) (Application Programming Interface) of machine translation tools that can be combined with automated translation systems to facilitate direct use of machine translation results. According to Man Delyan?Pi Qiang?Mo Aipin, "workplace translator now is a system that incorporates a variety of tools and resources, for example, terminology databases, translation memories and neural machine translation systems on-premises systems and cloud computing systems" (", " (our translation Gao C.) ^{[25, p. 80]}. In translation practice, if the translation memory cannot provide a suitable translation, the system automatically activates the machine translation mechanism to quickly provide a translation option that can be accepted immediately or edited, and the approved translation can be automatically stored in the translation corpus memory for reuse. Translation practice has moved from isolated machine translation to a combination of automated translation + machine translation + post-editing, while the task of professional translators, first of all, is to evaluate and control the quality of machine translation.

Form 2?Self-serviceThe Industrial Revolution, marked by the invention of the steam engine, allowed humanity to overcome physical limitations and use energy at its discretion.

The current information technology revolution is aimed at automating the "intellectual" aspect of humanity. This is an automated mode in which the demand for translation services is met by the customer of the translation service through the use of certain translation tools in accordance with certain rules. With the further development of modern information technologies, in particular with the breakthrough in machine translation technology using artificial intelligence, forms of self-service have come to the translation industry. Basically, such activities include:

1)Online translation. Online neural network machine translation based on a public cloud, for example, Google and Baidu, can support the translation of words, sentences and short chapters between any two languages. Today, Google provides online translation services to more than 200 million Internet users worldwide every day, translating a billion times a day and processing the equivalent of one million books per day ^[29].

2) Web translation. By providing clients with a special browser-based web translation and access to an intelligent online translation system, the system helps them choose the right language and thus overcome the language barrier when browsing the web.  In addition, more and more e-commerce platforms offer customers a choice of several language versions, for example, the Alibaba international website, which is available in 15 languages, and Booking.com , which is available in 42 languages.

3) Portable translation devices and mobile applications. Mobile applications including voice recognition technology, optical character recognition, intelligent machine translation (for example, Speak&Voice, iTranslate Voice, SayHiTranslate, etc.) and small translation devices (for example, portable translation devices, etc.) can help customers to carry out instant translation and self-service using voice, image translation and other actions, which can satisfy the demand for simple translation services for some fields of activity, such as travel, office work and international e-commerce.

Form 3?Neural Network Machine Translation + Crowdsourcing Post-editing + Crowdsourced human proofreading. In recent years, when intelligent machine translation systems, represented by neural network machine translation, have become most widespread, the crowdsourcing translation model has also acquired automation features.

The target material is first translated by a machine translation system, non-professional volunteer translators on the Internet carry out its post–editing, and, finally, the text is finalized by professional translators - this is a new crowdsourcing model. We see the formation of a new crowdsourcing translation model consisting of neural network machine translation + crowdsourcing post-editing + crowdsourcing (often professional) human proofreading, which has significantly increased the efficiency and quality of crowdsourcing translation.

ResultsCurrently, neural network machine translators already have some characteristics of human translators and are able to translate less texts with a small artistic component, such as user content, communication messages, internal company documentation, product specifications and technical support messages, but they do not yet have adaptive experience, which is the fundamental difference between human and machine translation.

For highly creative texts, such as user-oriented documents, user interfaces, marketing and branding materials, the adaptable experience of a translator is of great importance, and human translation is still required in this area.

At the level of intercultural communication, since AI does not yet possess the emotional and creative intelligence of a higher mind, as well as the biological and social characteristics inherent in human intelligence, automatic translation models are subject to limited cultural, contextual, historical and social norms, which leads to the identity of translations everywhere and always, in cases of the same source text and machine translation systems. Such a translation does not have the necessary communicative, emotional and creative characteristics. It is for this reason that translations made by machine translation tools require final editing, interpretation, adaptation and other cultural adjustments on the part of human translators in order to meet the needs of the dissemination of interlanguage information and intercultural communication.

Conclusion

Constant development in the field of intelligent machine translation has not only led to the trend of automation of translation practice, but also had an impact on the ecology of translation, primarily on professional translators, whose goals largely coincide with the goals of the translation industry. On the one hand, the trend towards automation of translation practice reduces the burden on human translators, freeing them from repetitive, tedious and low-tech tasks and freeing them up for more complex and creative tasks, on the other hand, it also "eats up" the territory of professional translators at the level of language transformation, putting significant pressure on their survival space.