Language assets are real assets in the sense that they require constant and meticulous management. There are several distinct language assets:

  • Terminology glossaries
  • Translation memories (TMs)
  • Styleguides
  • Computer assisted translation (CAT) or machine translation

There are numerous benefits from using these assets, but we wouldn’t be honest and fair if we didn’t acknowledge some of the problems, too.

  1. Management

Language assets should have centralized management. This usually means that there is a dedicated person who maintains, updates and otherwise manages your translation memories, glossaries, styleguides. Training this person takes time and resources, and experience with one TM software does not guarantee a smooth transition to other TM software. Language asset management is vital if you want to avoid other problems in this list, but it brings with itself costs, time and resource constrains.

  1. Translation constraints

Even though language assets like TMs increase consistency, they also constrain translators. TMs do not handle lexical and contextual ambiguities. Segmentation also abandons the guiding principle of translation, stating that the translator should translate the text/message as a whole, not individual sentences. Some translators need to review even the 100% matches to make sure that no contextual mismatches occurred. Fuzzy matches and CAT/machine translations also constrain the translator by providing a foundation for that segment’s translation and therefore guiding his thought process.

  1. Short-lived content

Information decay is a real problem that any content manager must face in our lifetimes. Languages and ideas evolve, change and morph. This constant change has a direct impact on content that might be saved in your language assets (and be out-of-date!). Updating and refreshing your content and information is a partial solution for this problem that expands into the next problem on this list:

  1. Perishable TM

Some segments in your translation memories become obsolete over time. The most common ways how this occurs are: terminology changes, new terms, style changes, different expectations or target audiences. Errors also accumulate in TMs if they are not properly maintained and cleaned up. Large TMs face most problems, as we will see in the next problem, because some segments might be translated in different styles and automated 100% matches might generate a text that has no consistent style.

  1. Large TMs might be more trouble than they’re worth

We get it – it’s easier to have all your TMs in a huge TM, but this causes all above mentioned problems and makes TM management a nightmare. Stylistic inconsistencies, accumulating errors, outdated segments and terminologies require your language service provider to pay exponentially more attention to revision and proof-reading, sometimes the translators themselves insist on reviewing 100% matches out of concern for their accuracy. Your company TM should be segregated by products or types, and maintained regularly. Glossaries should have defined types: medical, general, technical, marketing, general or core terms.

  1. Problems with terminology, precision and context

If a translation element has only one solution – it’s dealing with terminology. On the other hand, you might have more than one way of translating an element, and a previous 100% match that’s not perfect or completely precise could slip through. Glossaries are sometimes filled with conjugated or declined verbs and nouns – this decreases the usability of the resource. Keep terminology as simple and precise as possible!

Conceptual mismatches are unavoidable between languages and sometimes a glossary-based translation could convey a slightly different meaning. External, official termbanks might be used to alleviate this problem. Your LSP should keep glossaries and internal termbanks up to date.

Some clients do not provide as much context as they could, thusly decreasing the usability of glossaries: terms translated out of context make little sense. Compound entries (e.g. “pituitary gland”) should be incorporated into glossaries, not just base words (“gland”).

  1. Increased revision time, costs, and quality concerns

All these problems go hand in hand with minor mistakes: e. g. small changes in the source text segments might bring completely different meanings (e.g. a comma in a different place). Taking into account these problems, the review process might become very costly in life science translations or other industries that require very precise translations.

The solution

All these problems can be solved through investment of time and resources (you have to spend money to make money, right?). Proper maintenance of your language assets is essential if you want to avoid higher costs for revision down the line.

  • Keep your glossaries up to date.
  • Make sure your language service provider has their termbanks up to date, too.
  • Provide your LSP with as much information as possible.
  • Have a dedicated person to manage your language assets and avoid double entries.
  • Make sure you have distinct instructions on how to translate product names vs. feature names.

Yes, having language assets brings management, training, maintenance and software costs, but it pays for itself tenfold, if cared for properly.

#Language_Assets #Translation_Memory #Term_Glossary #Med_Terms #Translation_Styleguide #Machine_Translation #Translation_Problems #Lifescience_Translations #Medical_Translations

How We Helped Accuray Reduce Their Megaproject Cost 17 Percent