Short Answer Questions (submit online)
- Explain Popper's anti-inductivist view of confirmation.
- What is the Bayesian approach to inductive confirmation?
- How does Norton's material theory purport to solve the problem of induction?
For Further Discussion
- Denying induction. Popper denies that inductive confirmation is a legitimate form of reasoning.
- Do you think that inductive confirmation in science is any more trustworthy than the predictions of Paul the Octopus?
- Ellis and Silk claim that many theories used in modern cosmology can be adjusted so as to fit almost any data that may appear in cosmology. Is there any real problem with a theory that can fit any data? Why or why not?
- Modern theories of cosmology probably can't be adjusted to fit any data whatsoever. However, some can be adjusted to fit a large amount of different kinds of data. Is this a problem? Why or why not?
- Popper argues that the only way science produces knowledge is through falsification. Thus, we know that Aristotle's theory of freefall is false, but not that Einstein's theory is (even approximately) true. Instead, Popper says that positive tests "corroborate" a theory, in that they render the theory more "worthy" of our consideration, but not of our belief. What exactly does this mean? Why might it be preferable to mere verification?
- Norton, following Wes Salmon, argues that corroboration is not enough to explain the scientific practice of using theories to make predictions. What does Norton mean by this?
- A further well-known complaint about Popper's anti-inductivism is that, without induction, the choice of which hypothesis to test becomes a matter of mere psychology. Is this a reasonable suggestion? Why or why not?
- The Bayesian approach. The Bayesian account of inductive confirmation holds that a piece of evidence E provides a degree of confirmation to a hypothesis H given by the conditional probability P(H|E).
- How does Bayesianism confirm a hypothesis in practice? Give an example other than the London rain example.
- Bayesianism claims to solve the paradox of the raven. How can this be done? Write it down explicitly.
- What is the Bayesian problem of the priors? What options are there to avoid it?
- What is the Bayesian problem of ignorance? Is there any addition to Bayesianism that can avoid it?
- Norton's material theory of induction. Norton says that material facts, not an inductive schema, are what "power" or justify inductive inferences.
- How does Norton justify the inference that "All bismuth melts at 271°C" after a few observations?
- How is this distinct from the inference that "All wax melts at 91°C" after a few observations? Which one of the two inferences is justified according to Norton's view?
- What exactly does Norton mean by a "material fact"? What do you think should count as a such?
- According to Norton, material facts are typically justified by further material facts, leading to a regress. But he claims this is different than the regress that one encounters in Hume's Problem as it is normally stated. Why?
- Why does Norton say that the material theory makes induction just a matter of doing more science?
- Do you think the regress problem is fatal for Norton's account?