Skip to main content

How do I handle glossary conflicts?

Updated this week

Glossary conflicts happen when two or more glossary rules overlap, contradict each other, or cause inconsistent translations during a job. This can lead to incorrect terms in your output, especially if you’re translating into multiple languages or using both AI and traditional providers.


1. How glossary conflicts occur

  • Same source term, different target translations in different glossaries.

    • Example: “Spring” → “Lente” in Glossary A but “Spring” → “Veer” in Glossary B.

  • Overlapping phrases where one glossary term is part of another.

    • Example: “Apple” and “Apple Watch” may create unexpected replacements.

  • Duplicate terms in the same glossary — this can confuse the provider’s logic.

  • Provider-specific limitations — DeepL applies strict glossary rules, while ChatGPT interprets them as instructions, meaning conflicting instructions can result in unpredictable output.


2. Detecting conflicts

  • Before running a job – Review glossary entries when creating or editing them. Elovate does not automatically detect duplicates, so a manual check is important.

  • After running a job – If translations look inconsistent, check the glossary that was applied to that job and see if there’s a term clash.


3. Resolving glossary conflicts

  • Merge and clean up – Keep only one definitive glossary for each language pair to avoid overlap.

  • Use precise, non-overlapping terms – For example, replace “Apple” with “Apple (brand)” if needed, to differentiate from the fruit.

  • Prioritize one glossary per job – Avoid assigning multiple glossaries unless you’re certain they don’t overlap.

  • Test with a small job – If you must use two glossaries, test a few items before running large-scale translations.


4. Best practices to avoid future conflicts

  • Maintain a master glossary per language pair and update it regularly.

  • Assign one person in your team as glossary owner to ensure consistent changes.

  • Document terminology decisions in your team so all members know why certain translations are used.

  • For AI providers like ChatGPT, use the “additional instructions” field in the job to reinforce glossary priorities.


5. If conflicts still cause errors

If you notice a provider consistently ignoring certain glossary entries:

  • DeepL – Check if the glossary is linked to the correct source and target language.

  • ChatGPT – Adjust your prompt to explicitly state “Always use glossary terms exactly as listed.”

  • If issues persist, contact Elovate Support and provide the glossary file plus an example of the incorrect output so we can investigate.


Did this answer your question?