Do LLMs trained on languages that treat any double (or more) negatives as one have a slightly different world model than those that treat negatives like separat

Do LLMs trained on languages that treat any double (or more) negatives as one have a slightly different world model than those that treat negatives like separate logical elements, like English? I wonder if that'd be one way to demonstrate what you're saying.