concept binary function in category c++

This is an excerpt from Manning's book Functional Programming in C++.
In general, folding doesn’t require that the arguments and the result of the binary function passed to it have the same type. Folding takes a collection that contains items of type
T, an initial value of typeR(which doesn’t need to be the same asT), and a functionf: (R, T) → R. It calls the specified function on the initial value and the first item in the collection. The result is then passed to the functionfalong with the next item in the collection. This is repeated until all items from the collection are processed. The algorithm returns the value of typeR: the value that the last invocation of functionfreturned (see figure 2.6). In the previous code example, the initial value was 0 (or 1 in the case of multiplication), and the binary functionfwas addition (or multiplication).
You don’t do anything overly complicated here. You take a binary function—
operator> : (int, int) → bool—and create a new unary function object from it that does the same thing as the greater-than operator, but with the second argument fixed to a specific value (see figure 4.2).