Preprint of Shaun Gallagher, 2000. Reply to Cole, Sacks, and Waterman Trends in Cognitive Science 4, No. 5 (2000): 167-68. Please cite and quote from the original publication. This is a reply to Cole, Sacks, and Waterman. 2000. "On the immunity principle: A view from a robot." Trends in Cognitive Science 4 (5): 167, which was a reply to Shaun Gallagher, S. 2000. "Philosophical conceptions of the self: implications for cognitive science, Trends in Cognitive Science 4 (1):14-21

Reply to Cole, Sacks, and Waterman

Cole, Sacks, and Waterman offer a fascinating example that raises a number of questions about the immunity principle, and the related senses of agency and ownership. Whether there are exceptions to the immunity principle is still, I think, an open question. With respect to the example cited, there are at least two aspects in which I can not make an error in self-reference when I am moving the robotic arms. First, if in this situation I say, for example, "I am tying the knot," to whom does the "I" refer? It can refer to no one other than myself, even if I am wrong about who is tying the knot. Indeed, if some other person is actually working the robotic arms and I mistakenly think that the arms are being operated by my action, my statement "I am tying the knot" would not be incorrect unless the "I" referred to myself. If I discovered my mistake I might say "Oh, it is not I who am tying the knot." Both of these statements have their proper meaning only if the "I" refers to myself.

Second, both statements pertain to a sense of agency. My first statement, that I am tying the knot, is based on an introspective awareness of what I am trying to do. If, for example, I was just standing around watching the robotic arms move, but not trying to make them move, then I would not normally make this statement. I make the statement based on an introspective sense of my own agency, and although I can be wrong about the agency itself (about who is moving the robotic arms) I cannot be wrong about that sense of agency. Cole and his colleagues are surely right to say that this sense of agency is reinforced by peripheral feedback (proprioception of one's own arms and/or vision of the robotic arms). Even without that feedback, however, I could still say "I am trying to tie the knot," and no one would be in a position to tell me that I made a mistake about that -- "No, sorry, it's not you who are trying."

I have suggested elsewhere (Gallagher, 2000) that senses of agency and ownership for action depend upon temporal aspects of experience. Cole, Sacks and Waterman note a short delay between motor command and robotic movement. How long could that delay become before the senses of agency and ownership break down?