Maria Konnikova takes a psychological experiment on memory into an excursion on literature: "On writing, memory, and forgetting: Socrates and Hemingway take on Zeigarnik".
In this view, talking something through—completing it, so to speak, off the page—impedes the ability to actually create it to its fullest potential. Somehow, that act of closure, of having talked through a piece of work, takes away the motivation to finish. It’s like the order has already been delivered to the waiting customer. Once done, it escapes from the mind to make way for the next client. And the best of both worlds may or may not exist.
This for me is one of the perils of blogging. Once I've written something up, it has a sense of completion, so I'm not typically in a hurry to publish it elsewhere. This weakness of course is counterbalanced by a great strength: having more eyes on something makes the idea stronger. But it does take discipline to carry a research agenda through of the many strains of writing. I think it's also a fundamental element of interdisciplinary research, carrying an idea through the excursions into the different lines of evidence needed to examine it.