The latest Circuit Breaker podcast with Bob and Greg goes into incredible depth on "Prototyping to Learn". When Bob first told me about his approach to prototyping, I balked. The idea of building out multiple versions of something seemed wasteful, like going backwards or sideways instead of forwards. I didn't want to hear it. After a long time I realized he wasn't telling me to do double or triple the work, he was talking about the methods I use to make decisions.
There are a lots of decision-making steps along the product-development path. Those of us in the software industry who were influenced by Kent Beck et al talk about "spiking." It means trying to build just enough of something to expose the unknowns that we would have missed otherwise. It's not unusual to spike two different approaches to the same problem to see the difference in their characteristics.
Another example is in the shaping room. We've got a good idea on the wall, everyone is nodding their head and ready to commit, but not exactly buzzing with excitement. Then one of us says: "Is there a very different way we can approach this?" Sometimes it's to highlight a doubt, but it can also work the opposite way. Briefly entertaining a very different idea can strengthen everyone's resolve by showing, through contrast, how strong the existing path actually is.
And of course you only have to look over a graphic designer's shoulder to see what prototyping is about. You never see one sketch in their notebook, and one corresponding vector image on their screen. You see variation after variation, sometimes wildly different ideas, both in the sketchbook and in screenful after screenful of Illustrator canvas.
These examples only scratch the surface. The big idea is to think of prototyping not as a single costly effort to build and verify a single guess, but as a way to learn, to uncover what we don't know, to find the best way forward for the unanswered question at hand.