Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Math Bite: Irrationality of √m (1999)

fermatslibrary.com

Post reply on HN
Print view View original

11–20 of 27 posts

Re: Math Bite: Irrationality of √m (1999)

#11

Post by vog »

jamesfisher wrote: post #10

I don't understand. The end of the proof says r/p is a fraction, presumably because r must be an integer. But why must r be an integer? There seems to be an implicit assumption that m is an integer, but the explicit assumptions only give the much weaker statement that " m is not a perfect square".

> why must r be an integer?

r is a shorthand for (m-n^2)q-2np, which is an integer because m,n,p,q are all integers.

> There seems to be an implicit assumption that m is an integer, but the explicit assumptions only give the much weaker statement that "m is not a perfect square".

Yes, it would have been more explicit to say "m is an integer that is not a perfect square."

On the other hand, this is pretty clear from the context. This is like looking at a computer program and saying "foo has no side-effect" without stating that "foo is a function".

Re: Math Bite: Irrationality of √m (1999)

#12

Post by iand »

tempodox wrote: post #4

The print is too small and the document cannot be scaled, only pixellated. This makes it quite unusable for me. Too bad, math deserves better.

It's pixelated because the paper is a scan of a 17 year old paper. The Fermat's Library site is designed to allow people to annotate and comment on old classic papers, most of them from before digitisation of journals.

Re: Math Bite: Irrationality of √m (1999)

#13

Post by contravariant »

vog wrote: post #2

That proof is really great! It's always nice to see alternatives to well-known proofs. This demonstrates that you can always tackle problems from different angles. However, I slightly disagree with the introduction of the proof: | The really interesting thing about this proof is that it doesn't use divisibility, just mathematical induction in its "Z is well-ordered" form. That's quite a bold statement. It would mean…

>> The really interesting thing about this proof is that it doesn't use divisibility, just mathematical induction in its "Z is well-ordered" form.

>That's quite a bold statement. It would mean that this proof should generalize to other well-ordered sets that do not provide any notion of divisibility.

It doesn't, for the somewhat obvious reason that divisibility makes sense in any set that has a notion of 'multiplication'. The proof itself doesn't use any statements of the form 'x is divisible by y' though, not even implicitly as far as I can tell.

Even fraction laws don't really count as 'using' divisibility. Since fractions can be defined in a way that doesn't directly use divisibility (although the concepts are obviously related).

Being well ordered is a far more restrictive property than 'having a notion of divisibility'.

> Moreover, these fractions are not even fractions in Z, but in R, or at least in Z[sqrt(m)]. So there's still a lot of specialized structure involved in this proof.

According to the proof the fractions are in Q not R.

Re: Math Bite: Irrationality of √m (1999)

#15

Post by tribe »

JerTheRipper wrote: post #14

I'm trying to work this out in my head, but I believe that this holds if m is a rational such that its numerator and denominator are not both perfect squares (e.g., √(2/3) is irrational but √(4/9) is rational). Am I wrong?

When this proof has been presented to me in the past, I believe we assume that m is a positive integer. This is mentioned in one of the comments, but I agree that it should be more clear in statement of the theorem.

Re: Math Bite: Irrationality of √m (1999)

#16

Post by plookn »

JerTheRipper wrote: post #14

I'm trying to work this out in my head, but I believe that this holds if m is a rational such that its numerator and denominator are not both perfect squares (e.g., √(2/3) is irrational but √(4/9) is rational). Am I wrong?

Good point! The only part we need to check is the base case. Since the result holds for n=0, I'd say you're right.

Re: Math Bite: Irrationality of √m (1999)

#18

Post by jameshart »

vog wrote: post #9
jfarmer wrote: post #8

Earlier quoted context omitted.

By "doesn't use divisibility" I assume the author means that the point of contradiction doesn't rest on the "divisibility properties" of the integer, not that division is never used. In particular, this proof doesn't rely on the fundamental theorem of arithmetic.

> I assume the author means that the point of contradiction doesn't rest on the "divisibility properties" of the integer I see what the author meant by that. I just think it was stated in a slightly exaggerated way. The "divisibility properties" of the integers are still used. That part has just has been moved to another corner of the proof, by transforming fractional equations. One of the comments in the article is…

Indeed - I think the claim that alpha can be expressed as p/q with some 'lowest q' seems to rest on divisibility properties.

Re: Math Bite: Irrationality of √m (1999)

#19

Post by fluxion »

brlewis wrote: post #17

> We may assume that q is as small as possible (Estermann's key idea) Making the denominator as small as possible uses divisibility.

Not necessarily. Strictly mathematically a (non-negative) fraction q/p is just the equivalence class of all non-negative integer pairs (s, t), t != 0, such that

qt = sp

As such we may make q as small as possible by finding a representative in the equivalence class where the numerator is smallest (using the well ordering principle of the of the natural numbers). You can identify this pair uniquely (when the fraction is not 0) by using divisibility but that isn't required here.

Re: Math Bite: Irrationality of √m (1999)

#20

Post by brlewis »

fluxion wrote: post #19
brlewis wrote: post #17

> We may assume that q is as small as possible (Estermann's key idea) Making the denominator as small as possible uses divisibility.

Not necessarily. Strictly mathematically a (non-negative) fraction q/p is just the equivalence class of all non-negative integer pairs (s, t), t != 0, such that qt = sp As such we may make q as small as possible by finding a representative in the equivalence class where the numerator is smallest (using the well ordering principle of the of the natural numbers). You can identify this pair uniquely (when the fraction i…

By the same reasoning, divisibility does not use divisibility. Instead of asking if x is divisible by y you create a set of all products of the form ny where n is an integer, then check to see if x is a member of that set.

I don't buy it. qt = sp looks too much like x = ny to me.

Post reply on HN