During one of his seminars, management consultant W. Edwards Deming, referring to the communication of information, once aptly stated: “We know what we told him, but we don’t know what he heard.” In other words, if the information that we’re communicating isn’t being communicated in a way that is fully transparent and understood as it is intended, then the overall message is lacking. While reliable and easily understood information is important in all industries, it is of the utmost importance for the translation industry; therefore, it only makes sense that we take a closer look at today’s term, linguistic validation.
For the term itself, linguistic validation, broadly defined, is the process of examining the reliability, validity of content, and equivalence of concept of a translated work. Linguistic, being a back-formation of linguist, both of which find their root in the Latin lingua, meaning ‘language’, demonstrates that this term specifically “relates to language” and first appears in English from the short-lived periodical The German Museum in May of 1801, where it was written that: “The linguistic monuments become more numerous in the thirteenth century, and the low-Saxon language appears in some of them.” The act of validating or making something valid – our term validation – comes from the Latin adjective validus, and can be first found in the work of lexicographer (dictionary writer) Thomas Blount, who, writing in his 1656 Glossographia, recorded the term as: “Validation, a strengthening, inforcement, confirming; an establishing or ratifying.”
Linguistic validation is important because, unsurprisingly, all languages are not the same. Sure, they may all have the same basic building blocks but individual languages are also an expression of cultural identity, embodying different expressions, thoughts, concepts, and other intangibles. Attempting to translate 2 languages as though they were identical has been the cause of many well-publicised marketing fails, and, on a more individual level, it is the reason why most basic machine translation sites on the Internet fail to adequately express English thoughts into another language. At best, these language “hiccups” may only slightly skew the poignant understanding of what you’re trying to say; however, at its worst, this lack of clarity can seriously affect medical data and the specific communication of results, which is why linguistic validation is often used when looking at Clinical Outcomes Assessments (COAs), such as Patient Reported Outcomes (PROs), Clinician Reported Outcomes (ClinROs), Observer Reported Outcomes (ObsROs), and Quality of Life (QOL) questionnaires.
To best assure that no data gets (literally) lost in translation, linguistic validation incorporates 4 basic steps.
1. The data is first translated by at least 2 independent translators who are experts in the native language as well as specialists in the specific field.
2. The plural sets of data are then harmonised into a single work that contains the best of the aforementioned translations while staying as conceptually equivalent to the source.
3. The harmonised translation is then translated back into the original language.
4. If the translation of the harmonised data appears valid when compared to the original data, production then begins on a final corrected translation into the target language.
As has been said before, translation errors mean lost time and lost money; moreover, when dealing with the medical field, specifically in clinical trials and clinical settings, time is money. Much like with detail-oriented medical research itself, the translation of this data shouldn’t be left to chance either, otherwise, to reiterate Deming, we risk not knowing what we’re hearing.