premises. We simply fail to follow the instructions – we fail to allow Mary to know
everything physical there is to know about colour – because the instructions tell us
to imagine something preposterously immense.
Dennett gives an alternative ending to the story. Mary’s captors release her into
the colourful world and, as a trick, present her with a blue banana. Mary is not
fooled at all. ‘Hey!’ she says. ‘You tried to trick me! Bananas are yellow, but this one
is blue!’ (1991, p. 399). She goes on to explain that because she knew everything
about the physical causes and effects of colour vision, she already knew exactly
what impressions yellow and blue objects would make on her nervous system,
and exactly what thoughts this would induce in her. This is just what it means to
have all the physical information about colour vision. When we readily assume
that Mary will be surprised, it is because we have not actually followed the instruc-
tions – because it is really hard to imagine knowing absolutely everything physical
about anything. And so, we have succumbed to ‘Philosophers’ Syndrome: mistak-
ing a failure of imagination for an insight into necessity’ (Dennett, 1991, p. 401).
To make it easier for us to imagine Mary having all the physical facts, Dennett
(2005) invents ‘a standard Mark 19 robot’ with hardware equipped for colour
vision, but with black-and-white video cameras installed instead of colour ones.
While waiting for her cameras to be replaced, RoboMary learns all the physical
information about the colour vision of Mark 19s:
She has all her hard-won knowledge of that system of color vision, but
she can’t use it to adjust her own hardware so that it matches that of her
conspecifics. This doesn’t faze her for a minute, however. Using a few
terabytes of spare (undedicated) RAM, she builds a model of herself and
from the outside, just as she would if she were building a model of some
other being’s color vision, she figures out just how she would react in every
possible color situation.
(Dennett, 2005, p. 126)
And so, Dennett demonstrates that ‘What RoboMary knows’ leaves no space left
over for her to be startled, delighted, or surprised.
British philosopher Michael Beaton retaliates with ‘What RoboDennett still doesn’t
know’. Beaton argues that RoboMary cannot have a perfect model of herself, any
more than RoboDennett can. And even if she could, she would be modelling the
state of ‘knowing what it is like’ only as a state of the model, not of herself. Objec-
tive knowledge of oneself cannot necessarily be used as a simulation of oneself,
and knowing all the facts about what one would say and how one would react is
not the same as knowing all the facts about what it’s like. Even if physicalism is
true, Mary really could learn something new when she comes out (Beaton, 2005).
The imaginary Mary has led to many philosophical tangles like this (see e.g. Lud-
low et al., 2004), and along the way her inventor even changed his mind about
the argument against physicalism, suggesting that when we feel sure that Mary
will learn something new we are under an illusion about our own experience
(Jackson, 1998). Mary does not learn anything new, but merely finds herself in a
different kind of representational state from those she was in before – that is, she
now has the ability to recognise, imagine, and remember the state of seeing this
colour (Jackson, 2003).
