When someone puts forward ideas in public, and persuades people, but then changes his mind, he ought to tell people. They shouldn't go by his old ideas with no chance for an update and to maybe change their minds too.
For example, if Thomas Szasz had decided he was mistaken that mental illness is a myth, then he would have been responsible for publishing a retraction and correction, and explaining why he changed his mind. Not doing so would have been immoral and irresponsible.
I do not know exactly what I have communicated about David Deutsch over the last decade, in public. This is partly an issue of memory, partly an issue of some things being communicated inexplicitly (without directly saying them, but they still come across), and partly an issue of trying to remember what was said in private or in public.
Let me clarify my relationship with David. I have known David for over a decade and had many, many discussions with him. For David's book The Beginning of Infinity, I provided over 200 pages of especially appreciated comments and edits. I made and own the website and discussion group for the book. David is a founder of Taking Children Seriously (TCS) and Autonomy Respecting Relationships (ARR). I own the dicussion groups for both of those, too.
We no longer associate closely. Things changed. I have learned a lot from David and I used to think we agreed more than I now think. I now regard David as rejecting some important good ideas. For some of these ideas, I had thought I learned them from David, but I've changed my mind about that.
Here are some things I have changed my mind about.
I believe I have communicated that David is a world class expert on TCS, ARR, and some other parts of philosophy. I thought he was. However, he has stopped talking about a lot of that stuff and has said things exposing misconceptions. So I've changed my mind.
I think have communicated that I consider David a better philosopher than myself with higher status and more knowledge. I have changed my mind.
In the past I think I basically said David is always right. I did not mean it literally but I did mean something, and I have changed my mind.
I believe I have communicated that David is super rational. That I endorse him and his ideas pretty much without exception. That I'm a big fan. I've changed my mind.
I've said that David is a fan of Ayn Rand. He made this claim to me and I accepted it. I've changed my mind.
The list of issues I now know that I disagree with David about includes qualia, mirror neurons, Edmund Burke, Thomas Szasz, Ayn Rand, Ludwig von Mises, William Godwin's economics, deduction, hard to vary, meta discussion, justificationism, the value of school and academia, and the right approach to email discussions. Note that I have left some out to respect David's privacy.
Despite David's TCS reputation, and arguments against school, he actually has a a much more favorable opinion of university and academia than I do. His position on school is incompatible with TCS.
I believe I have communicated that David has the utmost intellectual integrity and responsibility. He does not. I thought he did; I was surprised when he acted otherwise; I've changed my mind.
People can seem more rational than they are as long as they are right frequently. This can happen when they already know a lot of things, but are not learning new things. When there are serious criticisms of their thinking then they are put to a harder test. Critical challenges can be particularly revealing about someone's character. David has done poorly on several.
I still consider The Fabric of Reality and The Beginning of Infinity to be very good books. Despite some flaws, they are world class. And there are other things David has written that are good.
I made every effort to avoid this outcome. For example, I tried to help David by explaining his misconceptions and offering him new ideas.
I have learned from this. In the future I will hold people to a higher standard. Many of my comments about David were years past, and I have improved my judgment.
Update, Feb 2020: Related to this post, I've written: