2

I'm trying to do the following problem in my book, but I don't understand how the book got their answer.

The problem: Determine whether the following relations are equivalence relations:$\newcommand{\relR}{\mathrel{R}}$

The relation $\relR$ on $\mathbb{R}$ given by $x\relR y$ if and only if $|x-y|\leq1$.

The answer only says it isn't transitive and gives this example: $(1\relR2)\wedge(2\relR3)$, but $1\not\relR 3$. Where did they get those numbers from?

As for the problem being reflexive and symmetric, please correct me if I'm wrong but here is what I assume it to be:

Reflexive: For any $x$ such that $x\relR x\Rightarrow x\leq1$

Symm: For any $x$, $y$ such that $x\relR y\Rightarrow|x-y|\leq1\text{ and }1\geq|x-y|$

Travis Willse
  • 99,363

4 Answers4

1
  • Reflexivity means $xRx$ which is $|x-x|=0\le 1$ verified.
  • Symmetric means $xRy\implies yRx$ which is $|x-y|\le 1\implies |y-x|\le 1$ verified too.

Now transitivity is not verified.

  • transitivity means $xRy\text{ AND } yRz\implies xRz$

So can you find $x,y$ at distance $1$ apart, $y,z$ at distance $1$ apart, but $x,z$ are further apart ?

A simple example is $x=1,y=2,z=3$ since $|x-y|=|1-2|=1\le 1$ and $|y-z|=|3-2|=1\le 1$ but $|x-z|=|1-3|=2>1\quad$ so $\require{cancel}x\cancel{R}z$

Of course you could choose any other numbers, for instance $7,8,9$ or $-0.5,0,0.6$, the book selected $1,2,3$ because these are "easy" numbers to plug in.

zwim
  • 28,563
  • Does picking numbers always help or do you have to be careful? – SirChryse Feb 19 '19 at 00:23
  • 1
    I just want to be sure about this now, if it was |x-y| >= 1 then that would not be reflexive, because 0 is not >= 1, correct? – SirChryse Feb 19 '19 at 00:24
  • One counterexample is enough to prove the relation is not transitive. So as soon as you picked some numbers that invalidate it, you are OK. For your second question, yes, it would not be reflexive, you are correct. – zwim Feb 19 '19 at 00:25
  • @ChrisD93 Yes, that is a perfect counterexample. :) – Nico Feb 19 '19 at 00:26
1

Let $xRy$ if and only if $|x-y|\le 1$.

Reflexivity: Notice $|x-x|=0\le 1$. Thus $xRx$.

Symmetry: Let $xRy$. Then $|x-y|=|y-x|=1$, which implies $yRx$.

The numbers for the proof that $R$ is not transitive are cooked up (there are many such examples), but you can see why there is a problem:

$|1-2|=1\le 1$, which proves that $1R2$. Similarly, $|2-3|\le 1$, so $2R3$. But then we can compute $|1-3|=2\not\le 1$, so $1\not R 3$.

This is contradicts the requirement of transitivity, which would imply (in particular) that $$1R2\;\wedge\; 2R3\quad\Rightarrow\quad 1R3.$$

Nico
  • 410
1

Reflexive: For any x such that xRx ---> x <= 1

No, reflexivity requires that $\def\R{\operatorname R}\forall x{\in}\Bbb R~(x\R x)$, which is clearly true (given the definitions for absolute value, substraction, and the $\leqslant$ comparator).$$\forall x{\in}\Bbb R~(\lvert x-x\rvert\leqslant 1)$$

Symm: For any x,y such that xRy ---> |x-y| <= 1 and 1>= |x-y|

No, symmetry requires that $\forall x{\in}\Bbb R\,\forall y{\in}\Bbb R~(x\R y\to y\R x)$, which is clearly true. $$\forall x{\in}\Bbb R\,\forall y{\in}\Bbb R~(\lvert x-y\rvert\leqslant 1\to\lvert y-x\rvert\leqslant 1)$$


Transivity requires that $\forall x{\in}\Bbb R\,\forall y{\in}\Bbb R\,\forall z{\in}\Bbb R\;((x\R y\land y\R z)\to x\R z)$.   The truth value for this universal statement not so obvious, so we shall look into the possibility of counterexamples.   We just need to demonstrate one counterexample to disprove a universal quantified statement.

Our relation, $\R$ is not transitive if $\exists x{\in}\Bbb R\,\exists y{\in}\Bbb R\,\exists z{\in}\Bbb R\;(x\R y\land y\R z\land x\require{cancel}\cancel{\R}z)$ .   That is to say, should there exist some $x,y,z$ where $y$ is at most a distance of one from each of $x$ and $z$, but $x$ is more than one from $z$.

$$\exists x{\in}\Bbb R\,\exists y{\in}\Bbb R\,\exists z{\in}\Bbb R\;\big(\lvert x-y\rvert \leqslant 1\,\land\, \lvert y-z\rvert\leqslant 1\,\land\,\lvert x-z\rvert \gt 1\big)$$

So what real numbers could possible make that so?

Well, $1,2,3$ easily fit that bill. $$\lvert \mathbf 1-\mathbf 2\rvert \leqslant 1\,\land\, \lvert \mathbf 2-\mathbf 3\rvert\leqslant 1\,\land\,\lvert \mathbf 1-\mathbf 3\rvert \gt 1$$

Graham Kemp
  • 129,094
0

An equivalence relation $\sim$ satisfies three axioms.

  1. Reflexivity. $x \sim x$ for all $x$.
  2. Symmetry. If $x \sim y$ then $y \sim x$ for all $x, y$.
  3. Transitivity. If $x \sim y$ and $y \sim z$ then $x \sim z$ for all $x, y, z$.

If all three hold, then $\sim$ is an equivalence relation. If any one of them fails to hold, then $\sim$ is not an equivalence relation. Any equivalence relation induces a partition of a set, and any partition of a set induces an equivalence relation.

To prove that your relation breaks Axiom 3, recall that for any $x, y, z $ $|x - y| \leq |x - z| + |z - y|$. This property is called the triangle inequality.

Fomalhaut
  • 2,106
  • I know that, but for this case how would I know if it's reflexive or not? Would I plug in numbers? – SirChryse Feb 19 '19 at 00:15
  • To determine whether it is reflexive, you have to use the property of absolute values. – Fomalhaut Feb 19 '19 at 00:17
  • @ChrisD93 look at my answer again. To prove that it is not transitive, you need to use the triangle inequality property of absolute values. – Fomalhaut Feb 19 '19 at 00:20
  • We did not learn that – SirChryse Feb 19 '19 at 00:22
  • @ChrisD93, well, now you did. :) – Fomalhaut Feb 19 '19 at 00:22
  • Actually, you don't need the triangle inequality to prove transitivity fails. You need only do the computation with the explicit example given above. You can see either of the other answers for the computation. – Nico Feb 19 '19 at 00:25
  • @Nico the reason the specific example fails is because of the triangle inequality. You can prove abstractly that the triangle inequality means that transitivity cannot hold for any x, y, z. – Fomalhaut Feb 19 '19 at 00:27
  • @TomislavOstojich Yes, except for silly cases. I suppose my comment was more about your didactic style than mathematical fact. The student mentioned that they hadn't learned the triangle identity, so why make them toil with abstract reasoning when they are holding a perfectly good counterexample right in their hands? – Nico Feb 19 '19 at 00:31