>Looks like some sort of parsing/evaluating irregularity. Similarly:
>

>? NOT .F. # .T.
>? .F. # NOT .T.
>

This works.

? .F. # (! .T.)

It seems that negation after a # sign is somehow misinterpreted by the parser. The pre-parser function normalize() breaks with the same error in such cases:

? normalize(".not..T.")
? normalize(".t.=.not..T.")
? normalize(".t.<>.not..T.")
? normalize(".t.<>.not.(1=2)")
? normalize(".t.=.not.(1=2)")
? normalize(".t.=!(1=2)")
? normalize(".t.=! 1=2")
? normalize(".t.=.not.(1=2)")
? normalize(".t.=.not.1=2")
? normalize(".t.!=.not.1=2")
? normalize(".t.=.not.1=2")