2

What is the negation of this statement?

Let the sequences $\{x_{n}\}$ and $\{y_{n}\}$ be given. There exist a positive rational $a$ and a positive integer $N$ such that $x_{n} - y_{n} \geq a$ for all positive integer $n$ with $n \geq N$.

My answer is,

Let the sequences $\{x_{n}\}$ and $\{y_{n}\}$ be given. For every positive rational $a$ and every positive integer $N$, there is a positive integer $n$ with $n \geq N$ such that $x_{n} - y_{n} < a$.

Is this correct?

fiverules
  • 777
  • 2
    Yes, it is correct. +1 – DonAntonio May 30 '14 at 18:12
  • Hmm, but I'm stuck on some problem using this.. – fiverules May 30 '14 at 18:15
  • My deepest condolences, @fiverules...but what has that to do with your question? If you've problems post a new question. – DonAntonio May 30 '14 at 18:33
  • @DonANtonio - what about the quantifiers? – Hans Engler May 30 '14 at 18:51
  • What about them, @HansEngler ? This is precisely what's wrong with your answer: the quantifiers in this case are after the given two sequences, so these ones aren't affected by them – DonAntonio May 30 '14 at 18:54
  • 1
    @HansEngler - I think that in support of the reading of "given" as meaning an (implicitly) assumed couple of specific sequences, is the more "natural" reading of the mathematical sentence. It sound "strange" to assert that every couple of sequnce are "converging" to the same value... – Mauro ALLEGRANZA May 30 '14 at 18:58
  • @MauroALLEGRANZA - I see your point. However, this may very well be an exercise in predicate logic, not a statement from real analysis. Unfortunately the OP hasn't told us. – Hans Engler May 30 '14 at 19:31
  • @HansEngler - yes, sure. My firts "interpretation" of it was with the "given" as specific, but it can be read as well as quantified... – Mauro ALLEGRANZA May 30 '14 at 19:45

2 Answers2

2

What problem ? Your translation is fine.

Assuming that the statement of the problem regards the convergence of a couple of "given" sequences, we have that the sentence :

There exist a positive rational $a$ and a positive integer $N$ such that for all positive integer $n$ with ...

has the "form :

$\exists a \exists N \forall n \varphi$,

where $\varphi$ is : $x_n−y_n ≥ a$

Negating it we get :

$\lnot \exists a \exists N \forall n \varphi$, i.e. $\forall a \forall N \exists n \lnot \varphi$

and $\lnot \varphi$ is $\lnot (x_n−y_n ≥ a)$, i.e. $x_n−y_n < a$, which fits exactly with your translation.

0

Your negated statement does not start out correctly. It should be "there are sequences ...", not "let the seqeunces ... be given".

Write the statement using quantifiers, then negate it. For notational brevity, set $\mathcal{S}$ be the set of all sequences.

I take the phrase " let the sequences be given ..." to mean "given any sequences ...". The original statement is $$ \forall (x_n), (y_n) \in \mathcal{S} \, \exists a \in \mathbb{Q} \cap (0, \infty) \, \exists N \in \mathbb{Z}^+ \, \forall n \in \mathbb{Z} \; n \ge N \Longrightarrow x_n - y_n \ge a $$ Negating this results in $$ \exists (x_n), (y_n) \in \mathcal{S} \, \forall a \in \mathbb{Q} \cap (0, \infty) \, \forall N \in \mathbb{Z}^+ \, \exists n \in \mathbb{Z} \, \ni \, n \ge N \wedge x_n - y_n < a $$ Now translate this into words.

Hans Engler
  • 15,439
  • The first lines are wrong here: the two sequences are given, it is not said "for any two sequencers", or "there are two sequences s.t....". – DonAntonio May 30 '14 at 18:54
  • @DonAntonio - now we have two versions. I think the OP should decide which works in his situation. If the sequences appear in some previous context (e.g. if they have be constructed or otherwise specified), then no quantifiers are needed indeed. – Hans Engler May 30 '14 at 19:13
  • What "two versions" do we have ,@Hans ? I can see only one. And of course the quantifiers are needed no matter how the sequences appear or appeared before. – DonAntonio May 30 '14 at 19:16
  • @DonAntonio - Please explain which variables need to be bound by quantifiers in your opinion. In particular, what quantifiers are associated with the two sequences, if any? – Hans Engler May 30 '14 at 19:25
  • None, @Hans. Wasn't I clear enough before? The sequences are not quantified: they are given and thus fixed. Again, what "two version"? – DonAntonio May 30 '14 at 19:33
  • @DonAntonio - I disagree. Here is another version of the kind of statement that the OP wrote. Consider the sentence "Let the function $f:[0,2] \to \mathbb{R}$ be given. Then $f'(1) = 0$". Is this a statement? Is it true or false? There are no quantifiers, so you ought to be able to decide. – Hans Engler May 30 '14 at 20:29
  • That's a statement, which can be proved false easily by given a counterexample. As it stands, anyone in mathematics would understand that there is a hidden quantifier "for all functions..." . Not at all the case in the OP post. I really don't get it: this should be a piece of cake for any professional mathematician. And even if in some very special context one could understand what you did, why not better go for the simplest, easiest interpretation? If, after this, the OP says something else, then we can fix easily what needs to be fixed... – DonAntonio May 30 '14 at 20:39
  • @DonAntonio - just to conclude this: Hidden quantifier indeed, that's what going on in both statements. Please respond in chat. – Hans Engler May 30 '14 at 20:53