I don't think of it as a failure. The LLM figured out a reasonable interpretation for a term it did not know. It might be better to come back with "do you mean pure ptolemaic major third?" instead of assuming, but the way it does it now might be better for programmatic requests.