There are several problems with the behaviourist arugment. Firstly, behaviour can be faked, as in the case of actors in a movie, or preprogrammed automata. An entity could be in pain and not necessarily show it, as in a stomach ache, or minor toothache. Additionally, we can show pain without having it (to get a doctor's note off of school, for example.)
We can extend the behaviourist argument problems to computers which are programmed to respond to various input. Another example may be made of a maple tree, whose syrup pours out after a cut in its bark. It's responding to a violation of it's ``body.'' But it begs the question: does it really have a ``mind.'' So what we have is a condition where behaviour is neither a necessary nor sufficient condition to demonstrate pain, and inversely the existance of ``other minds.''