# About a universal (functionally complete) function producing a constant

I read in a note the following:

Suppose we have a boolean function $f(x,y,z,w)$, if $f(a,a,a,a)=1$ then $f$ can't be functionally complete.

Why is that? how does it imply that $f$ can't produce nand or nor?

The claim is false if you're allowed to use constants. Take $f(x,y,0,0) = f(x,y,1,1) = 1$ and $f(x,y,0,1) = f(x,y,1,0) = x \text{ NAND } y$. If you're not allowed to use constants, then every function that you can get by composing $f$ with itself will have the property that if all inputs are 1, then so is the output. NAND doesn't have that property.
• Oh comparing properties directly. Is it possible to also deduce which one (or both) of $not$ or $and$ can't be produced from that function? – shinzou Mar 6 '16 at 18:53
• Just after I wrote this I figured that $and$ does have that property but $not$ doesn't.. – shinzou Mar 6 '16 at 19:00