Consider a scalar-valued function of arity n [$]f = f(x_1,... ,x_n)[$] From this function we want to fix [$]n-1[$] arbitrary arguments leaving one argument free as it were, resulting in a (higher-order) function of arity 1.. The idea is to cycle over each argument from 1 to n, thereby creating a function of arity 1.
g1(x) = f(x, 2.0, 1.0)
g2(y) = f(2.0, y, 1.0)
g3(z) = f(2.0, 2.0, z)
Can it be done?
I bet Mathematica or Haskell yes?