5

How do I symbolize the statement "there does not exist a $x$ for all $y$, $B(x,y)$".

Using $\neg(\exists x) (\forall y) B(x,y)$ would mean it is not the case that there exist a $x$ for all $y$, $B(x,y)$ or that the negation would be for both the quantifiers.
If we write $(\neg (\exists x))(\forall y) B(x,y)$ then what would negation of quantifier mean ? As far as I know negation is for statements and not quantifiers. Interchanging of quantifiers is also I think not allowed.

The basic question was if it is possible to negate one quantifier and not the other. Like for example how to symbolize not for all $x$ $B(x,y)$ is true for all $y$ ?

ryang
  • 44,428
  • 1
    "there does not exist an $x$ [such that] for all $y$, $B(x,y)$" precisely means "it is not the case that there exist an $x$ [such that] for all $y$, $B(x,y)$", i.e. $\neg(\exists x) (\forall y) B(x,y)$ or equivalently $(\forall x)(\exists y)\neg B(x,y)$. – Anne Bauval Sep 23 '24 at 05:07
  • 2
    Also noteworthy: The statement $(\neg (\exists x))(\forall y) B(x,y)$ does not carry a lot of meaning, since $\neg$ should be applied to a formula, which the expression $\exists x$ is not. – ClemensB Sep 23 '24 at 05:09
  • Thanks, but what about not for all $x$ and $y$ versus not for all $x$ and for all $y$ ? They are not the same. In this case what would $\neg (\forall x) (\forall y) B(x,y)$ mean ? – user221985 Sep 23 '24 at 05:17
  • 1
    The sentence $\neg (\forall x)(\forall y) B(x,y)$ tells you there is a formula $B(x,y)$ and this formula does not hold for all $x,y$, so it is equivalent to $(\exists x) (\exists y) \neg B(x,y)$. Your other statement "not for all $x$ and for all $y$ is difficult for me to translate. Is your "and" logical? Then you would maybe have something like $\neg (\forall x) B(x,y) \wedge (\forall y) B(x,y)$. – ClemensB Sep 23 '24 at 07:47
  • By not for all $x$ and for all $y$ I meant: not for all $x$ $B(x,y)$ is true for all $y$. – user221985 Sep 23 '24 at 08:29
  • Do you mean $\forall{} y\neg{} \exists x B(x,y)$? Seems to me the reversed order of the quantifiers might be what you want. – Fictional Sep 23 '24 at 16:01
  • This is a binding problem. To what does the symbol $\neg$ bind? It must bind to a complete statement, and the only complete statement it can bind to in the expression $\neg(\exists x) (\forall y) B(x,y)$ is $(\exists x) (\forall y) B(x,y)$. In other words, $\neg(\exists x) (\forall y) B(x,y)$ is the same as $$\neg\bigl((\exists x) (\forall y) B(x,y)\bigr)$$ – Lee Mosher Sep 23 '24 at 18:31

3 Answers3

8

How do I symbolize the statement "there does not exist a $x$ for all $y$, $B(x,y)$"?

Using $\neg(\exists x) (\forall y) B(x,y)$ would mean it is not the case that there exist a $x$ for all $y$, $B(x,y),$ or that the negation would be for both the quantifiers.

If we write $(\neg (\exists x))(\forall y) B(x,y)$ then what would negation of quantifier mean ?

You need to first try to make sense of the original statement. In this case, it isn't actually meaningful unless we change it to something like

  • there does not exist an $x$ for which this holds: for each $y$, $B(x,y)$

or

  • there is no $x$ such that for each $y$, $B(x,y).$

Symbolically:

  • $\lnot\bigg(\exists x\Big(\forall y \big(B(x,y)\big)\Big)\bigg).$

Notice that we are actually negating the entire sentence; quantifiers and variables don't attain any truth value, so, indeed, cannot be negated.

More human-friendly:

  • $\lnot\,\exists x\,\forall y\, B(x,y)$
  • $\lnot\,\exists x\,\forall y\, Bxy.$

P.S. Equivalently:

  • $\forall x\, (\lnot\,\forall y\; Bxy)$
  • $\forall x\;\exists y\,( \lnot\, Bxy).$

Addendum to include comment under another answer

The sentence There is a lid for every pot is not translated as ∃l ∀p or as ∀p ∃l, which are not sentences, let alone meaningful. Because of its hanging quantifier, this sentence is technically ambiguous, and can actually be read as either of the following:

  • ∀p ∃l (p∈P → (l∈L ∧ F(l,p))

    or, equivalently, ∀p∈P ∃l∈L F(l,p)

    Each pot has a lid

    For each pot, there is some lid such that the lid fits the pot

  • ∃l ∀p (l∈L ∧ (p∈P → F(l,p))

    or, equivalently, ∃l∈L ∀p∈P F(l,p)

    Some lid fits all pots

    There is some lid such that for every pot, the lid fits the pot.

ryang
  • 44,428
3

Let $B(x,y) : xy=0$
Let our Domain be Integers.

We can make the Claim : $$\{ \exists x : [\forall y : B(x,y)] \}\tag{1}$$ In other words ,
$$\{ \exists x : [\forall y : xy=0] \}$$

It is indeed true , we can take $x=0$ , hence there does exists some $x$ such that $xy=0$ for all $y$ which are in our Domain Integers.

Hence the Negation must be false.

Now when we change the Domain to be Irrational numbers , there is no such number $x$ (Irrational number) which makes $xy=0$ for all $y$ which are in our Domain Irrational numbers.

Hence the Negation must be true now.

In general , we do not know whether the Claim in true or not.
We just take the written Claim and try to make the Negation.

$$\lnot \{ \exists x : [\forall y : B(x,y)] \}$$ We can use the "Standard" laws : $$\{ \forall x : \lnot [\forall y : B(x,y)] \}\tag{2A}$$ $$\{ \forall x : [\exists y : \lnot B(x,y)] \}\tag{2B}$$

Using the given Example :
$$\{ \forall x : [\exists y : xy\not=0] \}\tag{2C}$$

Domain Integers will make that false : for all integers $x$ , we can not get some $y$ which makes $xy\not=0$ , because we might have $x=0$ then there is no $y$ to ensure $0y\not=0$

Domain Irrational numbers will make that true : for all Irrational numbers $x$ , we can get some $y=1/x$ which makes $xy=1$ , thus we can ensure $xy\not=0$

SUMMARY :

The given Statement is (1)
The Negation is (2A) when we consider only the outer quantifier
The Negation is (2B) when we consider the outer quantifier with the inner quantifier too
The Negation is (2C) when we consider the outer quantifier with the inner quantified too , while considering the Example $B$ to change $=$ to $\not =$, though that is not very general.

ADDENDUM :

The "Standard" laws are well known. Check here :
https://math.libretexts.org/Courses/SUNY_Schenectady_County_Community_College/Discrete_Structures/02%3A_Logical_Reasoning/2.04%3A_Quantifiers_and_Negations

It is just :

$$\lnot \forall x : X(x) \equiv \exists x : \lnot X(x)$$ $$\lnot \exists x : X(x) \equiv \forall x : \lnot X(x)$$

Prem
  • 14,696
  • Thanks, but my doubt was is it possible to negate one of the quantifiers and not the other. If so, how to symbolize it. For example how to write not for all $x$, $B(x,y)$ is true for all $y$? – user221985 Sep 23 '24 at 08:31
  • I updated my answer to cover that Doubt , @user221985 , Consider (2A) then move to (2B) there. – Prem Sep 23 '24 at 08:40
  • Yes, but what if only the outer quantifier has to be negated ? – user221985 Sep 23 '24 at 08:43
  • That is (2A) where only outer quantifier gets negated. – Prem Sep 23 '24 at 08:46
  • Or is it the case that $\neg (x)(y) B(x,y)$ equivalent to $(y)\neg(x) B(x,y)$ ? – user221985 Sep 23 '24 at 08:55
  • No , that is invalid : We can not generally change the order , whether involving Negation or not. – Prem Sep 23 '24 at 10:17
0

In my opinion the statement

there does not exist a $x$ for all $y$, $B(x,y)$

is a bit ambiguous, and I tend to interpret it differently to the other answers. I tend to read the following statement, without the negation,

there exists an $x$ for all $y$ [such that] $B(x,y)$

as $$\forall y\ \exists x\ B(x,y),$$ similarly to how you say "there is a lid for every pot", which in my understanding means that for every pot there exists a lid (which fits it) and not the other way around. I read the original statement as its negation, i.e. $$\lnot(\forall y(\exists x(B(x,y))))=\exists y(\forall x(\lnot B(x,y))).$$

On the other hand, if this is an exercise question, the choice of variables $x,y$ and the order in which they appear in $B(x,y)$ suggest that $x$ comes first in the formula and then $y$ ;) So @ryang's and @Prems' answers are probably the desired solutions to the question in this case.

julian
  • 1
  • 1
    The unambiguous way to read ∀t ∃d P(d,t) is "every pot has a lid". "There is a lid for every pot" is just ambiguous: notice that it could mean either ∀t ∃d P(d,t) or ∃d ∀t P(d,t). The two natural-language corrections suggested in my answer are genuine, and not just artificial sentence constructions designed to be compliant with textbook exercises... – ryang Sep 23 '24 at 15:02
  • My sincere apologies if my last sentence came across flippant, this was not intended at all, I certainly agree that your answer is genuine. I changed the wording a bit now. – julian Sep 23 '24 at 15:33
  • The lid-pot sentence was meant to give an example where in natural language the quantifiers come in the opposite order of how the quantifiers come in formal language, as a justification of why I tend to read the OPs statement the way I wrote in my answer. I agree that formally the lid-pot sentence is ambiguous, but I assume most people understand it as $\forall t\ \exists d\ P(d,t)$, like in the saying https://en.wiktionary.org/wiki/there%27s_a_lid_for_every_pot#English. – julian Sep 23 '24 at 15:49
  • No no, don't worry about that, I took no offence at all. $\quad$ 2. "in natural language the quantifiers come in the opposite order of how the quantifiers come in formal language" I think you are referring to the issue of hanging quantifiers (ambiguity due to mixed quantifiers not all specified at the beginning of the sentence).
  • – ryang Sep 23 '24 at 15:56
  • Hm, it would be an issue of hanging quantifiers if you are saying "there is a lid that fits on every pot". "There is a lid for every pot" is only quantifiers and does not technically contain the $P(d,t)$, so not sure the quantifiers are hanging in this sentence. – julian Sep 23 '24 at 16:08
  • The sentence "There is a lid for every pot" is ambiguous precisely due to its hanging quantifier. It isn't symbolised as ∃d ∀t or ∀t ∃d, which are not sentences and certainly not meaningful! As mentioned in my first comment, one of its possible readings is ∀t ∃d (t∈T → (d∈D ∧ F(d,t)) (F(d,t) means "d fits onto t") or, equivalently, just ∀t∈T ∃d∈D F(d,t). – ryang Sep 24 '24 at 05:30