One of many methods is to look at the code page used if it is a text document or something similar. I guess if you wanted to there is nothing stopping you from making a document look like it came from a foreign country based on the document encoding. Then you have a ton of other metadata on files now that would have to match. Not exactly impossible.