The Four Eyes Principle in the Language Industry: How Does it Work?

Ciklopea 6 years ago 5 min read

Every serious language service provider has a quality management system in place that may include various steps depending on the project requirements. However, these quality management systems are always built around the two major phases of translation/localization project – translation and revision (also known as bilingual editing) that are performed by two individual linguists or linguist teams.

This is the four eyes principle (or the two-person rule) of the language business that is, just like in any other sector, used to assure that (at least) two persons confirmed the action, i.e. that the translated/localized materials had been checked by at least two professional linguists before the project was delivered to the client.

What is the purpose of linguistic revision?

Translations are performed by professional human linguists and, just like to all humans, errors are likely to happen even to the best among them. Tight deadlines and large volumes take their toll through an occasional slip of the keyboard, lack of concentration and, more often than not, poor source materials. When the specific technical, terminological and other aspects of the project (that may or may not be fully in line with the standard rules of the target language) are added to the equation, it’s easy to see that professional translating is not an easy job.

Linguistic revision is a quality check of the raw translated materials by a second linguist whose job can be somewhat less stressful only because he/she doesn’t have to produce the translation, though revising a poor translation is the very definition of stress in the language industry.

This check includes detection and fixing of standard linguistic errors such as typos, mistranslations, omissions, inconsistencies, as well as the suitability of the translated materials for the agreed purpose, taking into account any specific project requirements, terminology and references.

However, the style is the matter of linguistic revision only in cases when it is inappropriate (for example, use of informal style or slang in a legal or medical translation) and therefore, in most cases, it remains the responsibility of the translator, the reason being that it is a preferential aspect that cannot be objectively assessed.

Introduction of a second linguist doesn’t mean that every translation is bad by default or that the translator is not to be trusted – it is an industry practice based on the rational presumption that two persons can see more than one and thus take any appropriate action before it is too late and a substandard translation is released to the public. This is the reason why every translation used for commercial purposes must be (and usually is) checked at least by a second linguist.

Revision and LQA reports

Once again, every serious LSP uses some form of language revision reporting or language quality assurance (LQA) process to ensure documenting and traceability of all performed phases within a delivered project.

These reports demonstrate that a revision or QA check (or both) was performed, document the detected and fixed errors / error types and usually feature a general assessment / recommendation to help the translator avoid them in the future.

The tools

Depending on the type of project and the purpose and format of the translated materials, a variety of tools is used in the linguistic revision phase. The most common ones include spellcheckers (integrated into text editors and CAT tools), track changes, language quality and editing distance reports (reports showing how much work was done on a project during the language revision stage).

Another pair of eyes?

It is not uncommon that a third linguist is invited to proofread and review the revised materials once they have been technically processed (e.g. after the DTP process), or that the pre-final delivery is returned to translator and editor for the final check.

This is because the pre-release version with the edited layout may provide linguists with more contextual information and reveal any issues that might have passed undetected in CAT tools due to segmentation.

Wrap-up

Translation and localization are largely automatized and computer-assisted processes, but they are still performed and managed by people and people can make mistakes. Quality assurance processes have been developed (and they are continuously improved) with the aim of eliminating the human errors and the utilization of these processes means that the client receives the translated/localized materials that have been checked by at least two professional linguists.

Like This Article? Subscribe to Receive More Via Email

  • receive a digest with new articles
  • up to 2 emails a month

Related Articles

Why LLMs Aren’t Ready to Replace Human Translators

3 months ago

In today’s digital age, the landscape of language services has undergone a dramatic transformation. With the advent of free translation tools, many wonder about the future of language service providers (LSPs). At Ciklopea, we’re not only adapting to these changes, but have evolved to offer unparalleled value in this new ecosystem.

Continue reading

Understanding Linguistic Annotation: Enhancing Language Data for AI and ML

5 months ago

In the evolving landscape of artificial intelligence (AI) and machine learning (ML), the role of linguistic annotation has become more important than ever. As companies and developers work to build smarter systems, from chatbots to translation tools, the need for high-quality language data is paramount. But what exactly is linguistic annotation, and why is it so important?

Continue reading

The Crucial Role of Linguistic Validation in Global Research

5 months ago

In an increasingly globalised world, research and clinical trials often span multiple countries and cultures. This expansion necessitates the accurate translation of materials to ensure that all participants, regardless of their native language, can understand and respond appropriately. This is where linguistic validation comes into play.

Continue reading