I agree that visible consistency can bring social benefits. I also agree that sticking with a single plan (which is a type of consistency) is in some cases preferable to constantly re-examining the plan, given the time costs and other costs involved in such examination.
That said, it is fairly clear that if you’re aiming to have accurate beliefs, your past claims and actions aren’t any more relevant than are the claims and actions of someone else with comparable rationality and domain knowledge. And it is fairly clear that most of us stick with our beliefs (vs. the beliefs of comparable others that we know) even in cases where our aim is truth, and that our tendency to hold onto our own past beliefs (vs. those of comparable others) impairs our access to truth. So I’d like to become, not “more self-inconsistent” in a generic sense, but more able to update my beliefs in useful directions—and updating requires changing beliefs.
As to actions… you make a good point there that the benefits of consistency and inconsistency are not obvious and that I glossed over the question above. It’s a good question to highlight. To give some rough reasoning now: I and others seem, in my observation, to waste inordinate amounts of resource sticking with the particular action-patterns we’ve used in the past, instead of experimenting with the patterns others are using. We drive with whatever driving habits we picked up in our first two years, not withstanding any risk to our lives; we use whatever habits of social interaction we happen to have picked up, whether or not they help us learn from others and form good relationships; we go about our work according to particular patterns; we come up with self-justifying stories for why our past mistakes must have been optimal, and must the the right way to respond to our present choices. Identity, and self-consistency, are part of this story. It’s a part I’d like to reduce in myself, and in anyone whose epistemic accuracy and practical effectiveness I care about. But this is admittedly not a full analysis of the merits of self-consistency heuristics.
I agree that visible consistency can bring social benefits. I also agree that sticking with a single plan (which is a type of consistency) is in some cases preferable to constantly re-examining the plan, given the time costs and other costs involved in such examination.
That said, it is fairly clear that if you’re aiming to have accurate beliefs, your past claims and actions aren’t any more relevant than are the claims and actions of someone else with comparable rationality and domain knowledge. And it is fairly clear that most of us stick with our beliefs (vs. the beliefs of comparable others that we know) even in cases where our aim is truth, and that our tendency to hold onto our own past beliefs (vs. those of comparable others) impairs our access to truth. So I’d like to become, not “more self-inconsistent” in a generic sense, but more able to update my beliefs in useful directions—and updating requires changing beliefs.
As to actions… you make a good point there that the benefits of consistency and inconsistency are not obvious and that I glossed over the question above. It’s a good question to highlight. To give some rough reasoning now: I and others seem, in my observation, to waste inordinate amounts of resource sticking with the particular action-patterns we’ve used in the past, instead of experimenting with the patterns others are using. We drive with whatever driving habits we picked up in our first two years, not withstanding any risk to our lives; we use whatever habits of social interaction we happen to have picked up, whether or not they help us learn from others and form good relationships; we go about our work according to particular patterns; we come up with self-justifying stories for why our past mistakes must have been optimal, and must the the right way to respond to our present choices. Identity, and self-consistency, are part of this story. It’s a part I’d like to reduce in myself, and in anyone whose epistemic accuracy and practical effectiveness I care about. But this is admittedly not a full analysis of the merits of self-consistency heuristics.