1

According to wikipedia, every finite set is computable. Definition: set $S \subset N$ is computable if there exists an algorithm which defines in finite time if a given number $n$ is in Set.

Question: what is wrong with this counter-example:

  • given some $TM$
  • $S \subset N$
  • Lets assume $S$ could contain only $0$, i.e., either $S = \{0\}$ or $S = \emptyset$
  • if a given $TM$ halts then $S=\{0\}$ otherwise $S=\emptyset$

So set $S$ is finite, but not computable, since we cannot "compute" if a given $TM$ halts.

What is wrong above?

Ayrat
  • 1,065
  • 1
  • 8
  • 23
  • 1
    You know that - according to your definition - $S = {0}$ OR $S = \emptyset$ (exclusive or); you don't know which of the two sets you are referring, but you know that both ${0}$ and $\emptyset$ are computable. – Vor Jul 09 '13 at 14:29
  • so you are saying that my definition of the set is invalid? That is, i defined a function that maps TM to a set, but not the set itself? Hm, this raises the question what is a valid definition of the set... – Ayrat Jul 09 '13 at 14:49
  • 1
    ... in other words you are able to (formally) prove that for the set $S$ there exists an algorithm which defines in finite time if a given number n is in the set; it is one of the two algorithms: 1) given n, return n==0 2) given n, return false – Vor Jul 09 '13 at 15:26
  • thanks, @Vor, but why cannot i extend it to halting problem: there is infinite # of TM, there is inf # of possible answers for all these TM, but there is an algorithm which answers correctly, because we "can" just enumerate all such algorithms, and our algorithm is one of them. I think this proof fails because we cannot enumerate all of them, right? (due to "diagonalization" argument) – Ayrat Jul 09 '13 at 19:01
  • No, for the halting problem there is no algorithm which answers correctly, it can be proven that such an algorithm cannot exist. – svinja Jul 10 '13 at 06:33

3 Answers3

6

A similar question was answered here:

How can it be decidable whether $\pi$ has some sequence of digits?

We can ignore the question "if a given TM halts" as it is irrelevant what the question actually is, let's just name the condition C.

If C is true, then the correct algorithm is if n == 0 return true else return false. If C is false, the correct algorithm is return false. Whether C is true or false, one of these two algorithms is correct for every n, so such an algorithm exists.

Additionally, "does a given TM halt" is computable for the same reason - the correct algorithm is either return true or return false. What is not computable is a function that answers "does this TM halt" for any TM.

Community
  • 1
svinja
  • 472
  • 2
  • 6
2

The problem is, you haven't defined a language. You have defined a function that returns a language.

What you call $S$ is really $S(M)$, for some Turing Machine $M$. What is undecidable is the function problem: given a Turing Machine $M$, determine $S(M)$.

Once you have a fixed $S$, deciding which numbers are in this $S$ is always decidable if $S$ is finite. This is because all finite languages are regular. If $L = \{w_1, w_2, \cdots, w_n\}$, then the regular expression $R = w_1 + w_2 + \cdots + w_n$ accurately describes $L$.

Joey Eremondi
  • 29,754
  • 5
  • 64
  • 121
0

The simplest answer I can think of: because we can make a finite look-up table for answers. In other words, if you have a finite language $L = \{w_0, \ldots, w_{k-1}\}$, consider the following algorithm:

Input: $x$
$L \leftarrow [w_0, \ldots, w_{k_1}]$
for $i$ from 0 to k-1 do
$~~$ if $x = L[i]$
$~~$$~~$ accept
reject

Now, the problem with your argument is that you are thinking of $M$ as input in one place and not thinking of it as input in another place. If $M$ is not part of the input then the machine does not decide the halting problem but only a fixed instance of it, so the algorithm for deciding it doesn't imply that the halting problem is decidable.

Note that we don't need to construct (uniformly in $M$) the algorithm deciding the halting of a given machine $M$, we only need to show that it exists, i.e. the process of finding the machine that decide if $M$ halts or not does not need to be a computable process (and it can't, if it could then your argument would go through).

On the other hand, if we assume that $M$ is a part of the input, then the set is not finite anymore. It is essentially $A = \{\langle M,b\rangle \mid \text{$b=0$ and $M$ halts}\}$. What you are doing is taking a slice of this set $B_M = \{ w \in A \mid \mathsf{fst}(w)=M \}$ which is finite and therefore decidable. But the decidability of all slices of a set does not imply the set itself is decidable. We need to be able to decide membership in $A$ uniformly in $M$, i.e. we need a single algorithm that works for all $M$, not a different algorithms for each $M$.

Kaveh
  • 22,231
  • 4
  • 51
  • 111