Since the need for communication and understanding between different cultures and nations has been an imperative for mankind since the dawn of time, many people claim that translation profession is the oldest profession in the world. However, it is increasingly common in the industry to say that this profession is moving on to the past and is about to disappear completely. So are the translators doomed to be forgotten? Is there no longer room on the market for newbies? And what alternatives exist for educated, experienced professionals?
The future of the translation industry
One of the American magazines at the turn of the 19th and 20th century published an article, the author of which proved that in 1950 American cities will have a huge problem with removing horse droppings from their streets. The journalist based his deliberations on an analysis of the increase in the number of horse-drawn carriages in American cities, extending it over the following years until 1950. However, he was unable to predict that the American market would be conquered by the diesel cars, and in 1950 the only carriages drawn by horses will be those with newlyweds.
In the language services field, such a milestone was launch of the Google Translator with a static machine translation mechanism in 2006. Due to misuse and many abuses, over the years the programme has become synonymous with poor quality translation and a source of industry anecdotes. However, starting from 2016, it slowly began to change. This is when Google Translate switched to the GNMT system, which supports a large, artificial neural network. The translation industry split into two camps: sceptics proclaim the unquestionable supremacy of human translations, and those who see their professional careers in machine-translated texts post-editing.
To answer the question, which of these two groups is right, one must first explain concepts such as machine translation, neural network and post-editing. Let's get to work!
Machine translation and neural network
Machine Translation is one of the fields of linguistics, focusing on the study and analysis of the possibilities of using computer software in translation. In practice, this term is used for texts that have been translated from one language to another by means of a dedicated program or application.
This process is based on very popular recently concepts of algorithms with specific rules, neural network technology and deep learning. But what do they actually stand for? The neural network is in other words a very simplified model of the human brain, and it works by processing the input signals, summing them up and referring to appropriate weights and parameters that determine degree of influence of individual signals on the process outcome.
In other words, when translating, the program recognizes linguistic nuances when the user enters the text and tries to reflect their meaning in the translated text. Although the accuracy of this solution is constantly increasing, thanks to the continuously improved architecture of neural networks, which is supervised by teams of TM suppliers all over the world, the quality of the target document on many levels depends on the quality of the input. This is why machine translation is the most effective when it comes to translation of legal and technical texts of a repetitive nature, such as treaties, conventions, regulations, provisions, agreements and statutes, MSDS and simple manuals.
Post-editing – what it is and how it differs from translation
Machine translation of content that require more creativity, such as marketing and advertising texts, localization, poetry or literature, as well as technical documents containing specific industry vocabulary, may require human intervention, i.e. the so-called post-editing. This process consists of a set of actions aimed at adapting the language of a machine-translated message in such a way to make it understandable and natural to native speakers.
It is also important to check grammatical structures, correct punctuation and terminology. It can be said that this is advanced verification of a document combined with its proofreading. However, the MTPE output is not expected to be perfect, as is required for a translation made by a human translator – it should be understandable and accurate, but less sophisticated language and reduced expression are allowed.
Due to lower requirements, post-editing is usually evaluated at about 70-80% of the translation rate or about 150% of the standard proofreading rate. So is this a profitable business? Absolutely! Especially if we take into account the fact that demand for this type of services is growing dramatically.
Increasing number of orders
Only a few people realize that in the 1940s and 1950s term “computer” did not refer to a device, but to a person professionally operating and programming counting machines. This illustrates how, over the years, circumstances can affect perception of different phenomena. Will history go in the same way for the language services industry, and soon the term “translator” will refer only to applications and programs based on neural network? Experts agree that it will take a long time before this happens.
Meanwhile, machine translation stimulates demand in the translation market by boosting appetite for communication in the native language among users who previously accepted certain types of text in the source language, such as websites’ privacy policies, FAQ, site rules and interfaces. Not so long ago, it was common to buy equipment from abroad with a user manual only in a foreign language. Now, thanks to the machine translation post-editing, practically every device can have user manual translated into many languages. Thanks to the increasing efficiency of machine translations, today we can read manuals for small domestic appliances, low-budget electronics and personal use devices in mother tongue. This increases not only the comfort of use but also safety of operation, which in turn directly translates into additional benefits for consumers.
Based on the speech of PayPal's Localization Manager, Michał Antczak, language service providers despite the most sincere efforts do not seem to speak the same language as their potential customers. Offering excellent quality, cost optimization, speed of service and the possibility of conquering foreign markets, they forget about what is most important for the client – business indicators, key objectives and the implementation of the company's mission.
Looking at the world of yesterday, at what is happening today and imagining on this basis what tomorrow will look like, we make a fundamental mistake - we do not take into account new technologies and changing forms of dialogue. And as the Chinese thinker Sun Zu wrote, “A man (...) who ignores his opponent will inevitably end up as a prisoner”. In order not to be enslaved, it is worth to meet the needs of the market and make modern tools our allies, not enemies.
In conclusion, it is up to us whether we use the latest technologies as a weapon to increase the efficiency of our work or as a tool to self-destruct the entire profession.