Sometimes it's not the dumb question or the right question but "the appropriate question" to address the situation. It's good to remember that we all bring our previous experiences, preconceptions and (if you like) prejudices with us. Testing those preconceptions can be a good thing and sometimes you can do this in a neutral non-confrontational manner.
Here's an example which I witnessed many years ago. The situation was a meeting between an experienced consultant and a customer and the topic revolved around "orders". Only, something seemed to be going wrong and they were getting at cross-purposes.
The "dumb question" to the customer was: "How many orders do you process a year?"
The answer was (do you want to pause here and guess): "One to three (if we are lucky)!"
The customer was an engineering firm which received, at most, a dozen proper enquiries a year and converted one (or 3) of those to an order. Imagine the difference in mindset between that and someone (the consultant) who was thinking in retail terms. In this case, the quantifying question completely changed the dynamics of the meeting because it created the opportunity for the customer to explain in detail just exactly what they did. BTW an answer of "100 Million" would have been equally effective.
There's a whole class of "dumb questions" which revolve around quantifying the problem in even the crudest way. Nowadays I tend to view the answers to these crude quantities in terms of "powers of ten".