# What is wrong with this paper on arXiv which claims that the set of real numbers is countable?

This is a follow up to: Cantor's diagonal argument and alternate representations of numbers

What is wrong in this paper on arXiv: The Continuum is Countable: Infinity is Unique, Laurent Germain?

The author builds a tree and shows that he can represent, through this tree, any real number. Thus, real numbers are countable?! Here is another related research Counterexamples to Cantorian Set Theory E. Coiras

• Note that $\pi$ does not appear anywhere in the tree... – Jair Taylor Jun 12 '19 at 21:43
• why not? there is a path in the tree characterizing this particular number, 𝜋, or 𝜋 - 3, or whatever number you wish – Daniel S. Jun 12 '19 at 21:44
• Is there no oversight to which articles can be published in arxiv? – Mariah Jun 12 '19 at 21:46
• What could be the reason that this paper did not win all the prizes? A conspiracy, or maybe it's just that some people only pretend to know math? – Asaf Karagila Jun 12 '19 at 22:12
• If there is one representation of $\Bbb R$ that is uncountable, and you claim there is another representation of $\Bbb R$ that is countable, then clearly that new representation is missing a whole lot of the numbers. If it contained all of them, then there would be a bijection between the representations and they'd have the same cardinality. The notion of an "alternative representation" producing a different cardinality is flawed. – lurker Jun 13 '19 at 0:22

In this tree only numbers with finite decimal expansion appear. The numbers with infinite decimal expansion never actualy appear in the tree. You can imagine them as being the limits of specific paths in the tree, but they nevertheless are never counted when you list all the elements in the tree.

Note that to prove the countability of real numbers, you'd need to list all infinte paths in the tree; listing all elements of the tree is not enough.

The author of the paper in the proof of Proposition 2.2. claims

"When N is large, counting the number of nodes is the same as counting thenumber of paths (to infinity) in the tree."

but this is not true.

• Adam, aren't you strongly relying on the representation of the numbers? What if there is an alternative representation of the numbers, for which we associate to each infinite path a new "symbol"? – Daniel S. Jun 12 '19 at 21:53
• The author claims (Proposition 3.2) that every real number can be defined by a node of the tree, but that's not true. Real numbers (and actualy many of rational numbers as well) are defined by infinte paths in this graph, and there's no way to know the whole infinte path knowing only one node. – Adam Latosiński Jun 12 '19 at 21:58
• Here is another related research arxiv.org/abs/1404.6447 --- any comments on that one? – Daniel S. Jun 13 '19 at 18:15

It is true that the tree can "represent" any real number, for some meaning of "represent", but it doesn't follow that the real numbers are countable. While the paper presents a tree in which every natural number is represented by a node, the tree representing real number represents real numbers as paths. No irrational, or indeed non-repeating, real number appears as a node. The paper's argument seems to be saying that after a finite number of steps, the number of nodes equals the number of paths, and thus the number of paths in the entire tree is equal to the number of nodes, but this logic fails. While each finite path corresponds to a node, infinite paths correspond to no node. The set of infinite paths is larger than the set of finite paths.

Some examples of the paper's "logic":

When N is large, counting the number of nodes is the same as counting the number of paths (to infinity) in the tree.

Here, the letter "N" is not bolded, so apparently they are using it to refer to an unknown number, rather than the set of natural numbers. However, I don't see anywhere that they define it, and relying on bold versus non-bold to distinguish between two concepts is poor choice. Also, if N is "large", that implies that it's finite, but then they switch to saying "to infinity".

The cardinality of the set of the nodes of this infinite tree is $$10^{ℵ_0}$$ counted 9 times.

Bare statement with no support.

Indeed, we see in Figures 1 and 2 that the set N is of cardinality $$(0, 1, 2, 3, 4, 5, 6, 7, 8, 9)^N$$ counted 9 times.

This is simply handwaving. Just drawing a diagram and then declaring that "we see" something is not an argument. $$(0, 1, 2, 3, 4, 5, 6, 7, 8, 9)^N$$ would be the number of infinitely long sequences of digits. Natural numbers consist of finitely long sequences of digits.

Suppose that $$x := 0.4444444.......$$ appears in this tree. Then by his construction $$x$$ is reachable after a finite number of steps from $$0$$ and moreover the node preceding $$x$$ differs from $$x$$ in one decimal spot. If it takes $$n < \infty$$ steps to reach $$x$$ then $$x$$ differs from $$0$$ in only a finite number of decimal places.

I don't know, I didn't read the article but that's what this picture seems to say, no?

The Axiom of Infinity says that there exists a set that contains the natural number 1 and, if it contains the natural number n, it also contains the natural number n+1. That is, the set |N={1,2,3,4,...}. But while such a set |N itself is infinite, every member of |N is still finite.

The point of the Axiom of Infinity is to say that we don't have to reach the end of the defining algorithm to accept a set as complete. All we have to do is reach any possible member.

The algorithm you found can't reach the end of a simple repeating fraction like 1/3=0.333... . The author of the paper is confusing the fact that the Axiom of Infinity says the decimal string "0.333..." exists, with actually defining it with a finite number of steps.