Abstract
M.A. (Philosophy)
The purpose of my project is to provide necessary and sufficient conditions for a prediction to be considered good. Alex Broadbent (2013) claims that a good prediction is a stable prediction, thereby providing an internalist account of judging predictions. In contrast, this project demonstrates that an externalist approach to identifying good predictions is not only possible but, on the proposed view, more reliable too. Robert Nozick’s notion of sensitivity provides a means of understanding what makes a good prediction. It is argued that a good prediction is a sensitive prediction; one where a prediction activity tracks the truth of the claims and assumptions used to produce prediction claims. To gauge whether a prediction activity tracks the truth it is suggested that we ask the following question of the prediction: if the claims and assumptions appealed to in the prediction activity were false, would the same prediction claims have been made if at all? If the prediction claims would have been made in spite of this, then the prediction is not sensitive. Otherwise, the prediction satisfies the following tracking condition for good prediction: in the closest possible world to our own where one or more of the claims and assumptions appealed to in the prediction activity are false the prediction claims would be different or not made at all.