SQL "near()" Operator - How does it handle invalid characters

Legal department has database application, and they search fields for words that are close to each other using the SQL near() operator.  The fields get data from pdf files and images and some characters are unrecognized and come into the database as invalid characters.

'near('XXXX','YYYY','ZZZZ')'

Here is their scenerio and questions to be answered:

Pre-Conversion: XXXX[InvalidChar]YYYY[InvalidChar]ZZZZ

Question 1: If we’re running a proximity search, would that currently be treated as 3 separate words or 1 word?

Post-Conversion: XXXX YYYY ZZZZ

Question 2: If all converted to spaces, should be represented and proximity searched as 3 separate words.  Can you confirm?