Thinking a bit about infinite regresses in justification. My argument in a paper I published recently in this blog (otherwise unpublished and five years old...!) was that when I say I have a good reason (a good justification, a reliable process etc) I am somehow deferring and coupling, that is, deferring to a reason that I endorse and coupling my belief to an existing and accepted chain of reasons. The issue, of course, is whether I can buy the justification of my belief by coupling it to an infinite chain of reasons or by deferring to an infinite process. I don't know. But it is interesting to bite the bullett and claim that there is nothing else to justification than good deferral and good coupling. That is, deferral to a commonly accepted reason and coupling to a commonly accepted (infinite) chain. After all, justification always makes appeal to accepted reasons. These reasons can be out there in chains and processes and to justify could be no more than to accommodate a belief to them. If I am a good detector of red - because the chain that starts with me being a good detector of detectors of red and goes on to me being a good detector of detectors of detectors of red etc is out there to be used -, I can say I'm justified in spotting something red.