5

Let us consider the statement $\exists x P(x)$ - translated into English, "there exists an $x$ in our universe of discourse such that $P(x)$ is true." In writing the negation of this, we are taught to switch quantifiers ($\exists \leftrightarrow \forall$) and to negate that statement $P(x)$.

Thus,

$$\neg (\exists x P(x)) = \forall x (\neg P(x))$$

However, humor me for a second. Let's consider what negation is - it is the "logical complement." The negation of a statement is always false when the statement is true, and vice versa. In that light, why would we not say the following is also the negation?

$$\neg (\exists x P(x)) = \not \exists x P(x)$$

Or, taking this a bit further, why would we not write this as well?

$$\forall x (\neg P(x)) = \not \exists x P(x)$$

Both seem to imply the same thing: there does not exist an $x$ such that $P(x)$ is true (and thus for all $x$, $P(x)$ is false, or, rather, $\neg P(x)$ is true).

So is there some underlying reason why we don't do negations in this way? As far as I can tell, they mean the same thing, yet I always have seen the $\forall$ version as above. Looking around MSE, I've only seen some posts which have $\neg \exists$ (basically the same as $\not \exists$), but they're only in the context of simplifying a logical expression.

So I guess, if indeed these are logically equivalent, my follow-up question would be - why is $\forall$ considered a simplification of $\neg \exists$ or $\not \exists$?

My only guess is that "$\not \exists$" isn't a standard notation, or so I recall from some notes my complex analysis professor gave us last semester. Or perhaps to say "for all $x$, this is false" more immediately is understood (or a more "direct" way of saying it) than "there does not exist $x$ such this is true?"

(Footnote of note: I haven't had much education in predicate logic and such. We went over it for a little while in one of my classes so I understand some basics like the above but we never went into much detail. So I apologize if this question is poorly framed or worded.)

PrincessEev
  • 43,815
  • 2
    The idea of "there exists no $x$ such that $P(x)$ is true" is equivalent to "for all $x$, $P(x)$ is not true". I think it is just easier to work with something like "for all $x$, $P(x)$ is not true" rather than saying "there is no $x$ such that $P(x)$ is true". – Dave Jan 01 '19 at 00:11
  • Okay, that resolves one doubt of mine regarding their equivalence. Though it then raises my follow up question - why is the latter phrasing favored? – PrincessEev Jan 01 '19 at 00:12
  • 1
    I edited my comment to say that I think the "for all" statement is easier to work with than "there exists no..". I'm not entirely sure why one is used over the other, but in fact I know some people who often phrase it as "there exists no...". – Dave Jan 01 '19 at 00:14
  • Just a small point: you mentioned "logical complement"; this is not wrong, but could lead to confusion? – ryang Aug 29 '22 at 09:26

2 Answers2

7

The two statements are equivalent, the reason why we write it that way it is because it is easier to deal/prove it. It is irrelevant which way it is easier to say it, the important thing is being able to use.

Just consider the statement $\exists x \in \mathbb Z, x^2+(x+1)^2 \mbox{ is even}$. This is not true, now try to show that this is false.

Try to prove separately each of the following two statements:

  • $\not\exists x \in \mathbb Z, x^2+(x+1)^2 \mbox{ is even}$
  • $\forall x \in \mathbb Z, x^2+(x+1)^2 \mbox{ is odd}$

It is easier to deal with the second one.

N. S.
  • 132,525
  • 1
    Ahhhh, I see, I can't believe I looked over something so basic. Thanks for your insight, I appreciate it. ^_^ – PrincessEev Jan 01 '19 at 00:20
  • 2
    Good answer, +1. This is basically what I was trying to say with the "for all" statements being easier to work with. – Dave Jan 01 '19 at 00:26
3

Which form to consider simpler is basically a matter of taste and convention.

There are some accounts of predicate logic that consider $\exists$ the only primitive quantifier and treat $\forall x\,\varphi$ as an abbreviation for $\neg\exists x\neg\varphi$.

My impression is that this is something of a minority option these days, but it is not wrong as such.