[sdbrown]

Or, can it (like humans?) adapt its responses to suit the style of the question? Like if you start asking it lots of antagonizing questions, will it become more or less antagonistic itself?