The fantastic "Overcoming Bias" blog has a piece on how we really shouldn't try to solve a problem until we've done as much as we can to understand the problem.
It is astonishing how difficult it is to follow this advice, even when one tries to.
I consciously try to spend most of my initial meetings with clients asking questions, developing my model of their problem and establishing what properties any solution should have. However, it's really hard... one cannot help mentally considering solutions (or solution fragments) and using them to drive the questioning.
Part of this is healthy: many clients have multiple problems or multiple priorities and so driving questioning in a direction where you have a chance of being helpful is clearly useful. However, considering potential solutions gets in the way of developing a deeper understanding of the client's problems because, as the article suggests, the human brain has a very strong tendency to hold onto a solution once it has been conceived.
I found a language to articulate this on a course a few months ago. It was based on Huthwaite's "SPIN Selling" methodology. The course was one of the most useful (and challenging) ones I've been on and was focussed pretty much exclusively on addressing the problem identified in the Overcoming Bias piece. Specifically, a structured model for questioning that allowed one to arrive naturally at the fullest possible statement of the client's problem and needs prior to moving on to solutioning.
This article was timely; it has reminded me I need to make a renewed effort to implement it!