I would suspect that a text field or document would be smart enough to know the charset being used and run LtR or RtL as appropriate. English written RtL wouldn't read correctly no matter what, it's not a normal state for it to be. Arabic would be different.Does it not put the text correctly?