Do SMEs and Linguists mix like Water and Oil?
by Steve Dept, cApStAn CEO
The idiom “these folks mix like water and oil” usually refer to an insuperable difficulty for certain people to blend and see eye to eye with one another. When we were asked to verify the translation of assessments for which it was obvious that linguists alone would not have the required competence, we set up translation verification by dyads, or dual verification: one linguist and one bilingual subject matter expert, or SME. In the feasibility study of the OECD’s Assessment of Higher Education Learning Outcomes (AHELO), we worked with engineers or economists or, to be more precise, with university lecturers in engineering or economics in the target countries, and with experienced translators in those countries. That is when we the ominous warning was fielded at us: SMEs and linguists mix like water and oil.
And just like the Scientific American suggests an experiment to actually mix water and oil, cApStAn set up a process, not only to make the interaction between SME and linguist merely possible, but to effectively combine the strengths of each into meaningful feedback about translation quality and relevance.
The first rule is about compartmentalizing competencies and instructions
Let us assume that the translators have experience in the respective disciplines, e.g., the engineering skills assessment is translated by a linguist who is well-versed in civil engineering because she has translated many assignment specifications for building contractors. Nevertheless, we have selected the translator because of her translation skills, not her engineering savvy. So, during her training session, she is asked to do her research and translate according to a checklist of criteria, and to insert queries for the SME each time she has a doubt. Now, the SME may possess expert writing skills in her native language, but we have selected her because she is an assistant lecturer in engineering in a technical college in the target country. We ask her to refrain from suggesting any non-technical changes. Her job is to ascertain that the questions make as much (or as little) sense in the target version as in the source version. Finally, the verifier is a linguist with expertise in quality control of translated and adapted tests. His job is to check linguistic equivalence (and to trust the SME’s judgement when it comes to technical terms and statements).
The second rule for SME/linguist interaction is the order of appearance
It won’t surprise anyone that the translator comes first. Before she begins, though, cApStAn project managers and translation technologists have done their magic, optimized the source, provided a bilingual glossary and item-by-item translation notes, validated by the test developers. Second comes the SME, to address the translator’s questions and check domain-specific relevance and clarity. Third comes the linguist-verifier, who checks the balance between faithfulness to the source version and fluency in the target language.
The third rule for SME/linguist interaction is all about UX
User experience (UX) is often disregarded: unwieldy Excel files with 20+ columns, reference documents in .PDF format, a time-tracking application can all make the task more tedious (and less effective) than it is meant to be. The document must be shared, but locked for the user working on it. There are enough good platforms nowadays, where online collaboration is made easy. The days of mailing Excel forms to successive users are long gone (besides, this would be a serious security concern, too). It is very good if the SME and the linguist can exchange comments in the work environment while they work, as this mostly eliminates the need for additional iterations.
There are other rules, but we don’t want to bore you with our trade secrets, or do we?
While we are aware that people who usually steam their vegetables are unlikely to fry them in olive oil, we eventually want a dish that is palatable for the entire target audience. It is not a matter of preference, but a matter of cultural appropriateness, suitable lexical choices and, of course, maintaining the same level of difficulty across language versions.