Roughly, a tuple of arguments could be thought of as a vector, whereas a tuple of parameters could be thought of as a covector (i.e., linear functional). When a function is called, a parameter tuple becomes "bound" to an argument tuple, allowing the function instance itself to be computed to yield a return value. This would be roughly analogous to applying a covector to a vector (by taking their dot product (or, rather, matrix-product of row vector and column vector)) to obtain a scalar.