My special spin on conceptual metaphor is based on my gut reaction that Lakoff’s description is too tidy. You’re brain is a big ol’ kludged together association engine with zillions of overlapping neural networks. Mappings between source and target domains are going to be messy. Let’s say we both have a network for the metaphorical phrase “balls to the wall.” They’ll be different. When yours lights up, it will tend to activate other associated networks that are different than mine.
That may not matter for communication that allows the two of us to coordinate our actions, but I suspect it may matter more for problem-solving. A way to put it is that I like Lakoff’s explanation for how metaphorical communication *works*, but I’m less pleased that it doesn’t have much to say about how communication *doesn’t* work. As an old-timey software tester, thinking about how and why things don’t work interests me more.
There's more toward the end of this podcast transcript 
.
See also the Logjam Metaphor for Concepts, Distributed Logjam, and Kinds of Metaphor for more.