irrational number in the middle

shahar

Full Member
Joined
Jul 19, 2018
Messages
526
There is a way to prove that between 2 irrational numbers there is another irrational number?
 
There is a way to prove that between 2 irrational numbers there is another irrational number?
Sure. Given any two real numbers a and b the number a+b2\dfrac{a + b}{2} will be between them.

If a + b is irrational, then we are done because half of that is also irrational.

If a + b is rational then the irrational parts of a and b are negatives of each other. So then take a+ba4a + \dfrac{b -a}{4}.

-Dan
 
Sure. Given any two real numbers a and b the number a+b2\dfrac{a + b}{2} will be between them.

If a + b is irrational, then we are done because half of that is also irrational.

If a + b is rational then the irrational parts of a and b are negatives of each other. So then take a+ba4a + \dfrac{b -a}{4}.

-Dan
How do you define "the irrational part of a number"?
 
There is a way to prove that between 2 irrational numbers there is another irrational number?

Start with irrational numbers a and b (a<b). It's possible to find integer values ai,bi,na_i \, ,\, b_i \,,\,n such that...

a<ain<bin<b\displaystyle a \lt \frac{a_i}{n} \lt \frac{b_i}{n} \lt b

See spoiler if you can't work out how to do this. Next, find a new irrational number "c" such that...

ain<c<bin\displaystyle \frac{a_i}{n} \lt c \lt \frac{b_i}{n}

OP, could you do this?

a<ain<bin<ba \lt \frac{a_i}{n} \lt \frac{b_i}{n} \lt b
na<ai<bi<nbna \lt a_i \lt b_i \lt nb
The following values provide an answer (there are many more possible)...
n=3baai=na+1bi=nb1n = \left\lceil \frac{3}{b-a} \right\rceil \\ a_i = \lfloor na + 1 \rfloor \\ b_i = \lceil nb - 1 \rceil
 
How do you define "the irrational part of a number"?
I did say that wrong, didn't I? I suppose a better concept would be to use the fractional part. The concept I'm reaching for is that the non-integer part of the irrational numbers add to a rational number (between 0 and 1, but that isn't important.) So if we have two rational numbers a and b such that a + b is rational then we know that {a} + {b} = rational. But I guess we really don't need to comment on the possibility. All we need to do is find a rational number around b - a and add it to a. (As Cubit is saying above.)

-Dan
 
I did say that wrong, didn't I? I suppose a better concept would be to use the fractional part. The concept I'm reaching for is that the non-integer part of the irrational numbers add to a rational number (between 0 and 1, but that isn't important.) So if we have two rational numbers a and b such that a + b is rational then we know that {a} + {b} = rational. But I guess we really don't need to comment on the possibility. All we need to do is find a rational number around b - a and add it to a. (As Cubit is saying above.)

-Dan
I loved your method/ post, and I immediately knew the kind of thing you intended by "the irrational part of a number" but I couldn't tie it down more formally myself :unsure: :LOL:

I've thought of an alternative (and simpler) method than post#4... Start with two different irrational numbers a and b (a<b, b>0). It's always possible to find an integer value n,n>1n, n>1, such that...

a<b×n1n<b\displaystyle a \lt b \times \frac{n-1}{n} \lt b

For example, this works if a=sqrt(2), b=pi, and n=2

2<π×212<π\displaystyle \sqrt{2} \lt \pi \times \frac{2-1}{2} \lt \pi

1.414... < 3.141... * (1/2) < 3.141...
1.414... < 1.570... < 3.141...

I'm pretty sure that a rational times an irrational is always irrational. I'll leave it as a problem for OP to find a minimum acceptable value of n in terms of a and b
 
Last edited:
We can prove it considering this proposition, For any real numbers x,yx,y with x<yx < y there exists a rational tt such that x<t<yx < t < y.

Now we can prove your statement : Let x,yx,y two irrationals, regarding the proposition above there exists a rational tt such that x2<t<y2    x<t2<y\dfrac{x}{\sqrt{2}} < t < \dfrac{y}{\sqrt{2}}\iff x < t\sqrt{2} <y. And t2t\sqrt{2} is an irrational.

It remains to prove the first statement : (x,yR)(x<y    tQ:x<t<y)(\forall x,y\in \R)(x < y\implies \exists t\in \mathbb{Q} :x < t < y)

The idea is the following one: Imagine you are walking on the real number line in a direction of xx and yy, what length must be your step to make sure that one of your steps will be between xx and yy. (You will need to use the Archimedean property of R\R).
 
Last edited:
Top