The integrand is positive, so no integral could be negative. The theorem requires \(f\) continuous on all of \([-1,1]\); here \(1/x^2\) is unbounded near \(0\) and the Riemann integral does not exist.
It links derivatives and integrals: a definite integral equals an antiderivative evaluated at the endpoints, and differentiating an integral with variable upper limit returns the integrand.