Groups | Search | Server Info | Login | Register


Groups > comp.ai.philosophy > #29735

Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question

From olcott <polcott2@gmail.com>
Newsgroups comp.theory, sci.logic, comp.ai.philosophy
Subject Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question
Date 2023-06-18 09:32 -0500
Organization A noiseless patient Spider
Message-ID <u6n4ho$1m6pt$1@dont-email.me> (permalink)
References (6 earlier) <i8tjM.5978$Zq81.1390@fx15.iad> <u6lq6v$1i475$1@dont-email.me> <joujM.1824$VKY6.722@fx13.iad> <u6lsjq$1id16$1@dont-email.me> <knCjM.62$_%y4.58@fx48.iad>

Cross-posted to 3 groups.

Show all headers | View raw


On 6/18/2023 7:02 AM, Richard Damon wrote:
> On 6/17/23 11:10 PM, olcott wrote:
>> On 6/17/2023 9:57 PM, Richard Damon wrote:
>>> On 6/17/23 10:29 PM, olcott wrote:
>>>> On 6/17/2023 8:31 PM, Richard Damon wrote:
>>>>> On 6/17/23 7:58 PM, olcott wrote:
>>>>>> On 6/17/2023 6:13 PM, Richard Damon wrote:
>>>>>>> On 6/17/23 5:46 PM, olcott wrote:
>>>>>>>> On 6/17/2023 4:09 PM, Ben Bacarisse wrote:
>>>>>>>>> Richard Damon <Richard@Damon-Family.org> writes:
>>>>>>>>>
>>>>>>>>>> Except that the Halting Problem isn't a "Self-Contradictory" 
>>>>>>>>>> Quesiton, so
>>>>>>>>>> the answer doesn't apply.
>>>>>>>>>
>>>>>>>>> That's an interesting point that would often catch students 
>>>>>>>>> out. And
>>>>>>>>> the reason /why/ it catches so many out eventually led me to 
>>>>>>>>> stop using
>>>>>>>>> the proof-by-contradiction argument in my classes.
>>>>>>>>>
>>>>>>>>> The thing is, it looks so very much like a self-contradicting 
>>>>>>>>> question
>>>>>>>>> is being asked.  The students think they can see it right there 
>>>>>>>>> in the
>>>>>>>>> constructed code: "if H says I halt, I don't halt!".
>>>>>>>>>
>>>>>>>>> Of course, they are wrong.  The code is /not/ there.  The code 
>>>>>>>>> calls a
>>>>>>>>> function that does not exist, so "it" (the constructed code, 
>>>>>>>>> the whole
>>>>>>>>> program) does not exist either.
>>>>>>>>>
>>>>>>>>> The fact that it's code, and the students are almost all 
>>>>>>>>> programmers and
>>>>>>>>> not mathematicians, makes it worse.  A mathematician seeing 
>>>>>>>>> "let p be
>>>>>>>>> the largest prime" does not assume that such a p exists.  So 
>>>>>>>>> when a
>>>>>>>>> prime number p' > p is constructed from p, this is not seen as a
>>>>>>>>> "self-contradictory number" because neither p nor p' exist.  
>>>>>>>>> But the
>>>>>>>>> halting theorem is even more deceptive for programmers, because 
>>>>>>>>> the
>>>>>>>>> desired function, H (or whatever), appears to be so well 
>>>>>>>>> defined -- much
>>>>>>>>> more well-defined than "the largest prime".  We have an exact
>>>>>>>>> specification for it, mapping arguments to returned values. 
>>>>>>>>> It's just
>>>>>>>>> software engineering to write such things (they erroneously 
>>>>>>>>> assume).
>>>>>>>>>
>>>>>>>>> These sorts of proof can always be re-worded so as to avoid the 
>>>>>>>>> initial
>>>>>>>>> assumption.  For example, we can start "let p be any prime", 
>>>>>>>>> and from p
>>>>>>>>> we construct a prime p' > p.  And for halting, we can start 
>>>>>>>>> "let H be
>>>>>>>>> any subroutine of two arguments always returning true or 
>>>>>>>>> false". Now,
>>>>>>>>> all the objects /do/ exist.  In the first case, the 
>>>>>>>>> construction shows
>>>>>>>>> that no prime is the largest, and in the second it shows that no
>>>>>>>>> subroutine computes the halting function.
>>>>>>>>>
>>>>>>>>> This issue led to another change.  In the last couple of years, 
>>>>>>>>> I would
>>>>>>>>> start the course by setting Post's correspondence problem as if 
>>>>>>>>> it were
>>>>>>>>> just a fun programming challenge.  As the days passed (and the 
>>>>>>>>> course
>>>>>>>>> got into more and more serious material) it would start to 
>>>>>>>>> become clear
>>>>>>>>> that this was no ordinary programming challenge.  Many students 
>>>>>>>>> started
>>>>>>>>> to suspect that, despite the trivial sounding specification, no 
>>>>>>>>> program
>>>>>>>>> could do the job.  I always felt a bit uneasy doing this, as if 
>>>>>>>>> I was
>>>>>>>>> not being 100% honest, but it was a very useful learning 
>>>>>>>>> experience for
>>>>>>>>> most.
>>>>>>>>>
>>>>>>>>
>>>>>>>> sci.logic Daryl McCullough Jun 25, 2004, 6:30:39 PM
>>>>>>>>     You ask someone (we'll call him "Jack") to give a truthful
>>>>>>>>     yes/no answer to the following question:
>>>>>>>>
>>>>>>>>     Will Jack's answer to this question be no?
>>>>>>>>
>>>>>>>>     Jack can't possibly give a correct yes/no answer to the 
>>>>>>>> question.
>>>>>>>>
>>>>>>>> It is an easily verified fact that when Jack's question is posed 
>>>>>>>> to Jack
>>>>>>>> that this question is self-contradictory for Jack or anyone else 
>>>>>>>> having
>>>>>>>> a pathological relationship to the question.
>>>>>>>
>>>>>>> But the problem is "Jack" here is assumed to be a volitional being.
>>>>>>>
>>>>>>> H is not, it is a program, so before we even ask H what will 
>>>>>>> happen, the answer has been fixed by the definition of the codr 
>>>>>>> of H.
>>>>>>>
>>>>>>>>
>>>>>>>> It is also clear that when a question has no yes or no answer 
>>>>>>>> because
>>>>>>>> it is self-contradictory that this question is aptly classified as
>>>>>>>> incorrect.
>>>>>>>
>>>>>>> And the actual question DOES have a yes or no answer, in this 
>>>>>>> case, since H(D,D) says 0 (non-Halting) the actual answer to the 
>>>>>>> question does D(D) Halt is YES.
>>>>>>>
>>>>>>> You just confuse yourself by trying to imagine a program that can 
>>>>>>> somehow change itself "at will".
>>>>>>>
>>>>>>>>
>>>>>>>> It is incorrect to say that a question is not self-contradictory 
>>>>>>>> on the
>>>>>>>> basis that it is not self-contradictory in some contexts. If a 
>>>>>>>> question
>>>>>>>> is self-contradictory in some contexts then in these contexts it 
>>>>>>>> is an
>>>>>>>> incorrect question.
>>>>>>>
>>>>>>> In what context is "Does the Machine D(D) Halt When run" become 
>>>>>>> self-contradictory?
>>>>>> When this question is posed to machine H.
>>>>>>
>>>>>> Jack could be asked the question:
>>>>>> Will Jack answer "no" to this question?
>>>>>>
>>>>>> For Jack it is self-contradictory for others that are not
>>>>>> Jack it is not self-contradictory. Context changes the semantics.
>>>>>>
>>>>>
>>>>> But you are missing the difference. A Decider is a fixed piece of 
>>>>> code, so its answer has always been fixed to this question since it 
>>>>> has been designed. Thus what it will say isn't a varialbe that can 
>>>>> lead to the self-contradiction cycle, but a fixed result that will 
>>>>> either be correct or incorrect.
>>>>>
>>>>
>>>> Every input to a Turing machine decider such that both Boolean return
>>>> values are incorrect is an incorrect input.
>>>>
>>>
>>> Except it isn't. The problem is you are looking at two different 
>>> machines and two different inputs.
>>>
>> If no one can possibly correctly answer what the correct return value 
>> that any H<n> having a pathological relationship to its input D<n> 
>> could possibly provide then that is proof that D<n> is an invalid 
>> input for H<n> in the same way that any self-contradictory question is 
>> an incorrect question.
>>
> 
> But you have the wrong Question. The Question is Does D(D) Halt, and 
> that HAS a correct answer, since your H(D,D) returns 0, the answer is 
> that D(D) does Halt, and thus H was wrong.
> 
sci.logic Daryl McCullough Jun 25, 2004, 6:30:39 PM
    You ask someone (we'll call him "Jack") to give a truthful
    yes/no answer to the following question:

    Will Jack's answer to this question be no?

For Jack the question is self-contradictory for others that
are not Jack it is not self-contradictory.

The context (of who is asked) changes the semantics.

Every question that lacks a correct yes/no answer because
the question is self-contradictory is an incorrect question.

If you are not a mere Troll you will agree with this.




-- 
Copyright 2023 Olcott "Talent hits a target no one else can hit; Genius
hits a target no one else can see." Arthur Schopenhauer

Back to comp.ai.philosophy | Previous | NextPrevious in thread | Next in thread | Find similar


Thread

ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 00:54 -0500
  Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 08:09 -0400
    Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 11:59 -0500
      Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 13:43 -0400
        Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 13:23 -0500
          Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 16:27 -0400
    Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Ben Bacarisse <ben.usenet@bsb.me.uk> - 2023-06-17 22:09 +0100
      Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 16:46 -0500
        Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Jeff Barnett <jbb@notatt.com> - 2023-06-17 16:03 -0600
          Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 19:18 -0400
            Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 18:44 -0500
              Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 21:46 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 21:35 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 23:03 -0400
        Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 19:13 -0400
          Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 18:58 -0500
            Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 21:31 -0400
              Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 21:29 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-17 22:57 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-17 22:10 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 08:02 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 09:32 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 12:31 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 11:41 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 12:54 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 12:09 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 13:46 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 13:05 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 14:20 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 13:30 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 14:43 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 13:47 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 15:19 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 14:26 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 16:10 -0400
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-18 18:43 -0500
                Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-18 19:59 -0400
                Does input D have semantic property S or is input D [BAD INPUT]? olcott <polcott2@gmail.com> - 2023-06-18 22:31 -0500
                Re: Does input D have semantic property S or is input D [BAD INPUT]? Richard Damon <Richard@Damon-Family.org> - 2023-06-19 07:38 -0400
                Re: Does input D have semantic property S or is input D [BAD INPUT]? olcott <polcott2@gmail.com> - 2023-06-19 09:30 -0500
                Re: Does input D have semantic property S or is input D [BAD INPUT]? Richard Damon <Richard@Damon-Family.org> - 2023-06-19 20:45 -0400
                Re: Does input D have semantic property S or is input D [BAD INPUT]? olcott <polcott2@gmail.com> - 2023-06-19 22:57 -0500
                Re: Does input D have semantic property S or is input D [BAD INPUT]? Don Stockbauer <donstockbauer@hotmail.com> - 2023-06-20 00:33 -0700
                ChatGPT discussion (was: Re: Does input D have semantic property S or is input D [BAD INPUT]? vallor <vallor@vallor.earth> - 2023-06-20 11:16 +0000
                Re: Does input D have semantic property S or is input D [BAD INPUT]? Richard Damon <Richard@Damon-Family.org> - 2023-06-20 07:19 -0400
                Re: Does input D have semantic property S or is input D [BAD INPUT]? olcott <polcott2@gmail.com> - 2023-06-20 10:09 -0500
                Re: Does input D have semantic property S or is input D [BAD INPUT]? Richard Damon <Richard@Damon-Family.org> - 2023-06-20 11:48 -0400
  Ben Bacarisse specifically targets my posts to discourage honest dialogue olcott <polcott2@gmail.com> - 2023-06-20 10:06 -0500
    Re: Ben Bacarisse specifically targets my posts to discourage honest dialogue Richard Damon <Richard@Damon-Family.org> - 2023-06-20 11:48 -0400
      Re: dishonest subject lines Ben Bacarisse <ben.usenet@bsb.me.uk> - 2023-06-20 17:02 +0100
        Ben Bacarisse specifically targets my posts to discourage honest dialogue olcott <polcott2@gmail.com> - 2023-06-20 12:25 -0500
  Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] olcott <polcott2@gmail.com> - 2023-06-20 14:57 -0500
    Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] Richard Damon <Richard@Damon-Family.org> - 2023-06-20 16:34 -0400
      Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] olcott <polcott2@gmail.com> - 2023-06-20 15:42 -0500
        Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] Richard Damon <Richard@Damon-Family.org> - 2023-06-20 16:52 -0400
          Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] olcott <polcott2@gmail.com> - 2023-06-20 16:39 -0500
            Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] Richard Damon <Richard@Damon-Family.org> - 2023-06-20 17:53 -0400
              Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] olcott <polcott2@gmail.com> - 2023-06-20 17:07 -0500
                Re: Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts to discourage honest dialogue] Richard Damon <Richard@Damon-Family.org> - 2023-06-20 18:52 -0400
  Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts] olcott <polcott2@gmail.com> - 2023-06-20 14:59 -0500
  Refutation of the Ben Bacarisse Rebuttal [Ben targets my posts] olcott <polcott2@gmail.com> - 2023-06-20 15:00 -0500
  ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question vallor <vallor@cultnix.org> - 2023-06-21 19:10 +0000
    Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question vallor <vallor@vallor.earth> - 2023-06-21 19:23 +0000
    Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-21 14:59 -0500
      Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-21 19:01 -0400
        Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-21 19:40 -0500
          Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-21 22:47 -0400
            Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-21 21:58 -0500
              Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-22 07:26 -0400
                Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question olcott <polcott2@gmail.com> - 2023-06-22 09:18 -0500
                Re: ChatGPT and stack limits (was: Re: ChatGPT agrees that the halting problem input can be construed as an incorrect question Richard Damon <Richard@Damon-Family.org> - 2023-06-22 21:06 -0400

csiph-web