Order of parameters in quantified predicates

112 Views Asked by At

I'm studying up for my midterm in Discrete Math and I've been looking at sample questions and their solutions. There is one I don't really understand and I was hoping someone could help me out.

2. Let the domains of x and y be the set of all integers.
Compute the Boolean values of the following quantified predicates:

  All x, Exist y, (x^2 < y)
  Exist y, All x, (x^2 < y)
  Exist x, All y, (x^2 >= y)
  All y, Exist x, (x^2 >= y)

Solution:
  All x, Exist y, (x^2 < y) = T
  Exist y, All x, (x^2 < y) = F
  Exist x, All y, (x^2 >= y) = F 
  All y, Exist x, (x^2 >= y) = T

I'm not really sure if I'm understanding this or not. The first solution appears to say there exists one integer that is greater than every integer squared? I guess that makes sense on a per-integer basis, but the second solution appears to say the same thing, in a different order, but is false.

I know there's more here that I'm just not seeing. Would somebody mind explaining the nature of the problem and solutions? It would mean a lot, thanks!

2

There are 2 best solutions below

0
On BEST ANSWER

$\forall x.P$ means that every possible value of $x$ will make $P$ true.

$\exists x.P$ means that there is a value of $x$ that will make $P$ true. One such value is enough, but there has to be at least one.

Usually each variable is restricted to some domain. For example, since this is a discrete math course, can we stipulate that $x$ and $y$ represent integers?

Now consider, the sentence, "For $x = 4,$ there is a value of $y$ that satisfies $x^2 < y.$" Is that true?

How about, "For $x = 17,$ there is a value of $y$ that satisfies $x^2 < y$"?

In fact, we could set $x$ equal to any integer, and there would still be a value of $y$ that satisfies $x^2 < y.$ So the statement, "There is a value of $y$ that satisfies $x^2 < y,$" which we can write as $\exists y.(x^2<y),$ is true for all $x.$ And that is what $\forall x.\exists y.(x^2<y)$ says.

Now consider the sentence, "If $y = 4,$ every possible integer $x$ satisfies $x^2 < y.$" Is it true?

How about, "If $y = 17,$ every possible integer $x$ satisfies $x^2 < y$" ?

In fact, is there any value to which we can set $y$ so that $\forall x.(x^2<y)$ (that is, so that every possible integer $x$ will satisfy $x^2 < y$)? No, there is not. The statement $\exists y.\forall x.(x^2<y)$ asserts that there is such a value of $y,$ so that statement is false.

2
On

The first quantified predicate $\forall x \, \exists y \, (x^2 < y)$ means: "If you pick any real number $x$ and square it, then I can find a real number $y$ that is bigger than it". Indeed, this is true. Given any $x \in \mathbb R$, just take $y = x^2 + 1$.

The second quantified predicate $\exists y \, \forall x \, (x^2 < y)$ means: "I can think of a real number $y$ that is bigger than the square of any real number $x$ that you pick.". Indeed, this is false. Given any $y \in \mathbb R$, if you pick $x = \sqrt{|y|}$, then $x^2 = ||y|| = |y| \geq y$.


In general, the order of different types of quantified variables matters because the second variable is allowed to depend on the first. Indeed, the second variable is often a function of the first variable. This concept is very important to understand, particularly when doing $\epsilon$-$\delta$ proofs in real analysis.