1
$\begingroup$

I've seen $k\in \emptyset$ written in the context of empty sum. I read, $k\in \emptyset$ as "$k$ is an element of the set which has no elements." To me, this sounds like a contradiction. If the empty set contains the element $k$ it is no longer empty.

Also, how do you arrive at the conclusion that because $k\in \emptyset$, $k$ does not exist? To say that $k$ is in the $\emptyset$ is a false statement, therefore nothing follows from it. If we want to say that $k$ does not exist we simply write $\neg \exists k$.

Since $k\in \emptyset$ is widely used there must be an explanation. Can you clarify how you rationalize the contradiction that empty set is not really empty? Does the set theory has its own logic where contradiction is legal?


Edit

I'm sorry it was not clear from the question that $k$ is a summation index and as such it cannot be a free variable. $k$ takes natural numbers as its values $k\in \mathbb{N}$.

I think, therefore, answers claiming that $k\in \emptyset$ cannot be false because $k$ is a free variable are not acceptable. Please correct if I misunderstand.

$\endgroup$
25
  • 1
    $\begingroup$ The empty sum, $\sum_{k \in \varnothing} a_k = 0$, is mostly just a convention that works well in practice, just like $a^0 = 1$ for nonzero $a$. $\endgroup$
    – David Gao
    Commented Dec 23, 2023 at 6:24
  • 2
    $\begingroup$ There’s no contradiction here - when we write $k \in \varnothing$ we are not implying that there is something in the empty set. Quite the contrary. It is precisely because there’s no such $k$ that this sum is empty. $\endgroup$
    – David Gao
    Commented Dec 23, 2023 at 6:25
  • 2
    $\begingroup$ Also, the statement that that because $k \in \varnothing$, $k$ does not exist doesn’t really make much sense logically. The correct statement is that there exists no $k$ such that $k \in \varnothing$. Or, symbolically, $\neg \exists k (k \in \varnothing)$. $\endgroup$
    – David Gao
    Commented Dec 23, 2023 at 6:27
  • 1
    $\begingroup$ @MichaelCarey It looks like you pretend not to know that $y$ is the empty set, then you pretend not to know that there are no sets which are not equal to themselves and then you pretend to realize that $y$ really was $\emptyset$. Why? What’s wrong with the definition of the empty set as the set which has no elements? Isn’t this definition simpler? $\endgroup$
    – zeynel
    Commented Dec 28, 2023 at 9:24
  • 1
    $\begingroup$ We could add as an axiom that $\exists$y$\forall$x($\neg$(x$\in$y)) - Such a y is called the empty set. Which is fine, but then one is just assuming the empty set exists, they haven't constructed it. The first definition, is better in the sense that it constructs an object with the property that we want ( it is empty) rather than just assuming the existence of such an object satisfying some property. $\endgroup$ Commented Dec 28, 2023 at 14:26

3 Answers 3

4
$\begingroup$

The formula $$k\in\emptyset \tag1$$ is a predicate with a free variable $k$. It is still not true nor false, because you still have not said what $k$ is. Please note that this predicate is not saying that there exists $k$ with $k\in\emptyset$. That formula would be $$\exists k,\ \ k\in\emptyset \tag 2$$ which indeed has the value false. This is probably what you have in mind.

Again, the predicate $(1)$ is a formula with a free variable, so it doesn't have a truth value, while the proposition $(2)$ has the value false. On the other hand, what you observe is that there cannot be a $k$ for which formula $(1)$ has the value truth. This sentence is captured by the formula $$\forall k, \ \ \neg(k\in \emptyset). \tag 3$$ which has the value true.

$\endgroup$
3
  • $\begingroup$ “predicate with a free variable k” How do we know this is a predicate? We know that $k$ is a summation index that can take natural numbers as its values. Also, $k$ is an arithmetic term, it’s not a logic term. Arithmetic terms start to exist when you write them down and stop existing when you remove them. So, there is no question that $k$ here exists. $\endgroup$
    – zeynel
    Commented Dec 28, 2023 at 9:31
  • $\begingroup$ Also, in this site proofwiki.org/wiki/Definition:Empty_Set they state that "$x \in \emptyset$ is false, whatever $x$ is." Can you comment on this? $\endgroup$
    – zeynel
    Commented Dec 28, 2023 at 9:35
  • 2
    $\begingroup$ The meaning of that statement in wiki, if you want to be formal, is $\forall x, \neg x \in \varnothing$, that is not the same as saying $x \in \varnothing$ on itself is false. (It’s really better for this discussion if you find a book on first-order logic and model theory and see how truth value is formally defined there. Otherwise everyone here seems to be just running around in circles with you.) $\endgroup$
    – David Gao
    Commented Dec 29, 2023 at 9:33
1
$\begingroup$

You have basically two seperate questions:

On the notation for summations and specifically the empty sum, we read:

$$\sum_{k \in X}a_k$$ As the sum of all $a_k$ such that $k \in X$. For the empty sum, we would read: $$\sum_{k \in \emptyset}a_k$$ As the sum of all $a_k$ such that $k \in \emptyset$, which (you are correct) there are none. Which is why the empty sum evaluates to $0$, because we are summing nothing. This is simply a matter of notation and how to read summations, it is not a logical argument, no worries!

On your question about contradictions you write two things:

How do you arrive at the conclusion that because $k \in \emptyset$, $k$ does not exist? To say that $k$ is in the $\emptyset$ is a false statement, therefore nothing follows from it.

Your second statement is incorrect. If we can come to the conclusion that a false statement, namely $k \in \emptyset$, is true, then we can come to the conclusion that any other false statement is also true, which we know is not correct. Therefore, we know our original statement was incorrect, and hence there are no $k \in \emptyset$. This is the concept known as the principle of explosion and it relies around the idea that (If $F$ is a false statement). Using this concept, when we come to a statement that we know is false, we can then deduce that our original assumption was false. For example, if we assume $A$ to be true then:

$$A \Rightarrow B \Rightarrow C \Rightarrow F$$ Would point us to the idea that $A$ is false, since we know that $F$ is false and when we assume $A$ to be true $F$ evaluates to true. In proofs we state that this is contradiction, since $F$ should be true based on our work but we know it is false.

To arrive at the conclusion (with a contradiction) that $(k \in \emptyset) \Rightarrow \neg \exists k$, we can do the following: First, we assume that there exists a $k$ such that $k \in \emptyset$. Then, since $\emptyset \subset A$ for any set $A$, $k \in A$. But the set $A = \{b\}$ does not contain $k$, hence this is a contradiction and thus there does not exist a $k$ such that $k \in \emptyset$. Or any other argument where you can arrive at a contradiction that is shorter or longer than mine. A much easier proof, is just to say: by definition $\neg \exists k (k \in\emptyset)$.

$\endgroup$
6
  • $\begingroup$ Your proof by contradiction can be much shorter I think: "we assume that there exists a $k$ such that $k\in \emptyset$ QED :) Since empty set is empty by definition this statement must be false. $\endgroup$
    – zeynel
    Commented Dec 29, 2023 at 9:04
  • $\begingroup$ I think this is the problematic satetement for me: "the empty sum evaluates to $0$, because we are summing nothing." We don't have summands to evaluate so nothing is evaluated. Can you evaluate the following sum: No, you cannot because there is nothing there. Nothing does not evaluate to $0$. So I agree that, as you say, equating summation without summands to $0$ is a notational convention. So $0$ in $\sum_{k\in \emptyset} a_k = 0$ is only a placeholder for the non-existent summands. $\endgroup$
    – zeynel
    Commented Dec 29, 2023 at 9:10
  • $\begingroup$ @zeynel if you assume something is true saying well it's not true actually is kind of a silly proof by contradiction and is more of a direct proof, so in the spirit of an actual contradiction I think that works. Also for $k \in \emptyset$ its a heuristic, there is no like definitional or rigorous explanation for notation, but the heuristic helps one understand why we have the notation. $\endgroup$
    – Robertmg
    Commented Dec 29, 2023 at 22:55
  • $\begingroup$ "$k\in \varnothing$ is a heuristic..." I don't agree. $k\in \varnothing$ or in its more general form, $x\in y$ is a declarative sentence. It says: "element is an element of set." When we replace "set" with "empty set" we are knowingly making a false statement because we know that the set with no elements has no elements: "element is an element of the set with no elements" is a classic self-referential contradiction. $\endgroup$
    – zeynel
    Commented Jan 8 at 8:47
  • $\begingroup$ The element says to the set "I'm your element" and set replies, "No, you are not my element" and contradiction ensues. So, $x\in \varnothing$ is a contradiction that has the same form as "This is not a sentence." $\endgroup$
    – zeynel
    Commented Jan 8 at 8:47
1
$\begingroup$

What is a contradiction in Mathematical Logic?

Definition: A sentence is called a contradiction if its truth table only has false values.

Where a sentence is a formula with no free variables.

The formula x$\in$$\emptyset$ needs to unraveled, as $\emptyset$ is a defined term.

formally, x$\in$$\emptyset$ is shorthand for x$\in$y $\wedge$ $\forall$w$\in$y(w ≠ w)

Note: The same unraveling needs to be done with the empty sum in order to assess if it is a contradiction or not. The standard notation is just a shorhand for a longer formula, don't focuse on the syntax for the shorthand.

So x, is a free variable ( not bound by a quantifier) and so it doesn't make sense to talk about if it's a contradiction because it isn't even a sentence.

If we wanted to consider the sentence:

$\exists$x$\exists$y(x$\in$y $\wedge$ $\forall$w$\in$y(w ≠ w))

Then we could say that such a sentence is a contradiction

I usually see x$\in$$\emptyset$ as a premise in a vacuously true implication.

For example, Let's Prove $\emptyset$ is a subset of any set.

Definition: X $\subset$ Y $\iff$ $\forall$z(z$\in$X $\rightarrow$ z$\in$Y)

So we prove $\forall$z(z$\in$$\emptyset$ $\rightarrow$ z$\in$Y)

Proof: Let Y be a set.

Let $z_0$ be arbitrarily chosen.

Then

$z_0$$\in$$\emptyset$ $\rightarrow$$z_0$$\in$Y

Is true as the premise is false.

Thus $\emptyset$ is a subset of every set.

$\endgroup$
12
  • $\begingroup$ Thanks for the detailed answer. I'm sorry that it was not clear that $k$ is a summation index taking natural numbers as its values so $k$ is not a free variable. (I updated the question.) $\endgroup$
    – zeynel
    Commented Jan 1 at 13:10
  • $\begingroup$ But even if $k$ were to be a free variable $k\in \emptyset$ would still be false because by definition, $\neg\exists k : k\in \emptyset$ = False (You wrote a version of this statement and said it was false.) Or equivalently, $\exists k : k\notin \emptyset$ = True $\endgroup$
    – zeynel
    Commented Jan 1 at 13:13
  • $\begingroup$ We know that $\emptyset$ has no elements and that nothing in the entire universe, including free variables, can be an element of the set with no elements. So, what is the point of writing $k\in \emptyset$ when we know that by definition this $k$ whether or not a free variable cannot be a member of the empty set? $\endgroup$
    – zeynel
    Commented Jan 1 at 13:17
  • $\begingroup$ I don't know the history of the notation, for the empty summation. Sorry! I wish I knew the answer. It's just a piece of notation, not a logical claim. You can just consider it a symbol which is defined to be equal to 0. There isn't any deeper logic going on. $\endgroup$ Commented Jan 1 at 15:57
  • 1
    $\begingroup$ Officially, $\in$ is not a symbol of FOL, you need some set theory added in. Its a bit lazy to just say FOL and not specify more, ZFC, Set Theory is a First Order Theory - with elementhood added in, which is probably why they say that. The reason I didn't include $\forall$x,$\rightarrow$... etc is because it is unnescesary, They are useful symbols but Don't need to be part of the base language. Since $\forall$x($\phi$) is equivalent to $\neg$($\exists$x($\neg$$\phi$)). Note: Earlier I did forget $\wedge$, as a symbol. Sorry. $\endgroup$ Commented Jan 6 at 15:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .