People get turned off mathematics in various ways. If you teach it as sort of just stuff you need to know to balance your checkbook — which is nonsense because none of us balance our checkbooks; computers do that for us. On the other hand, because language is so important to us as living creatures, everyone is interested in language one way or another, be we language mavens or just interested in listening to the radio or reading or novels. You know, language is a fundamental part of what we ask.
In fact, in a book I wrote in 2000, called The Math Gene, I actually made a case based on sort of rational reconstruction of human evolutionary development, that mathematics and language are actually two sides of the same coin in terms of evolutionary development. Human beings, when we developed the capacity for language — and nobody knows when that was; it might be as recent as 50,000 years ago — but when our ancestors developed language capacity, at that moment they developed the capacity for mathematics. It's the same capacity. It just plays out in different ways.
A lot of the problem in mathematics is that an awful lot of what goes on in the school system is basically trying to train the mind to do what a $10 calculator can do: follow rules and algorithms and procedures. And one thing that we do know is, that the human brain does not find that natural. The human brain is analogical, not logical. And so, when we try to force it to be procedural and exact, the brain simply doesn't like it.
It was important for many thousands of years to be able to do computation and calculation because that was the basis of commerce and trade and buying and selling. And you had to do it in your head or with an abacus board or something. So for hundreds of years, it was actually important to train the mind to follow rules to do computations and get the right answer. Well, now we've automated that. And we carry around devices in our pockets that can do that. Which means that we can spend more time letting the brain do things that the brain is really well suited for that computers can't do very well: making value judgments, making analogical leaps.