Last time, I ended with the question, "What goals should we have?". Before, I said that "should" only makes sense in reference to goals, so how can this question be answered?
Obviously, using a goal to justify itself is circular reasoning. But you could justify a goal using other goals. If a goal helps you achieve your other goals, you should have it, in the same way you should do anything else that furthers your goals. Conversely, if a goal hinders your other goals, you shouldn't have it, in the same way you shouldn't do anything else that hinders your goals.
But then, what about the other goals? How do you determine whether or not you should have them? In the same way, referring to each other. This will form an infinite regress of self-reference, but that's not necessarily insurmountable. It could probably be represented similarly to Google's PageRank algorithm, which determines the importance of a website based on the number and importance of websites that link to it.
But doesn't that end up being just as circular as before? Well... Yes. And given two or more sets of goals which support each other equally well, and the unlimited ability to modify your goals, I don't know how you could determine which set of goals you "should" adopt. For that matter, I don't know what "should" means in that context.
But as it happens, I don't think we do have the unlimited ability to modify our goals. I think our goals are at least partially constrained by our biological nature. We need to eat. You can deliberately refrain from doing so, but I think that's more acting on a conflicting goal rather than not having the goal to eat.
And if that is the case, then the goals you can't change give you a starting point to base the goals you can change around.