To be more precise, Ocaml being a functional language, any function, 
when applied
to one argument, may return a new function which can in turn be applied 
to another
argument.

Thus if you write :

(a (b)) (c),

Ocaml (and most other functional programming languages), understand that

a(b) is supposed to be a function, that returns another function, then 
applied to c.

In practice, this trick is used extensively throughout functional 
programs, so
that placing parentheses around arguments would result in a humongous number
of parentheses.  For this reason, a new convention for parentheses was 
enforced:
you don't put any parentheses around function arguments, unless it is 
necessary
for disambiguation (for instance, if you want b to be applied to c, and 
you don't
place parentheses around the function part, so that (a(b))(c) is simply 
written

a b c

This means : a applied to b, and then to c,

Now, if you want "a applied to the result of applying b to c", you write

a (b c)

Please note there are no parentheses around c.

In practice, most functions taking several arguments are described in this
manner, instead of being described as function taking a pair as argument.
This is known as "currification" because Curry was one of the early 
mathematicians
to advocate the idea that mathematics (and programming) could be described
with only one-argument functions.

In your case, both print_int and fac are one argument functions, as can 
be seen from
their type.

Yves