Path: csiph.com!eternal-september.org!feeder.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Ben Bacarisse Newsgroups: comp.theory Subject: Re: Infinitely Recursive input on HP Proofs Date: Thu, 16 Mar 2017 10:19:06 +0000 Organization: A noiseless patient Spider Lines: 66 Message-ID: <87d1dhttgl.fsf@bsb.me.uk> References: <918df253-d4f0-4370-8f73-88e6690380a1@googlegroups.com> <87o9x6wxby.fsf@bsb.me.uk> <8737ehx3zz.fsf@bsb.me.uk> <8cb5d54a-10a7-48c4-8cb1-fcc81ba7e4d3@googlegroups.com> <878to6v4za.fsf@bsb.me.uk> <7b6313c0-3f5a-4bea-8249-67bdfe61d5f3@googlegroups.com> <87o9x1ucpi.fsf@bsb.me.uk> Mime-Version: 1.0 Content-Type: text/plain Injection-Info: mx02.eternal-september.org; posting-host="368b5322df020f7350229613c200b4e1"; logging-data="29039"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX18RJ/NJy79YHeKxypk09KzgY0J6SpWFaeQ=" Cancel-Lock: sha1:pNVtBzapQKzqIDtRkPLwXATH3XY= sha1:UlVqHNdIyVt+V9Slra4saU1sSqE= X-BSB-Auth: 1.98d66d7a2ddef9b9392f.20170316101906GMT.87d1dhttgl.fsf@bsb.me.uk Xref: csiph.com comp.theory:11932 peteolcott writes: > On Wednesday, March 15, 2017 at 10:23:24 PM UTC-5, Ben Bacarisse wrote: >> peteolcott writes: >> >> > On Wednesday, March 15, 2017 at 12:12:45 PM UTC-5, Ben Bacarisse wrote: >> >> I'll cut to the chase... >> >> peteolcott writes: >> >> >> >> > So when H^ evaluates Wm Wm (for clarity we will call Wm_1 and Wm_2) >> >> > the execution trace of Wm_1 indicates that it makes a copy of Wm_2 >> >> > which for clarity we will call Wm_3. >> >> > >> >> > The next step of the execution trace of Wm_1 indicates that it begins >> >> > its own evaluation execution trace of Wm_2 Wm_3. >> >> > >> >> > The execution trace of Wm_2 indicates that it makes a copy of Wm_3 >> >> > which for clarity we will call Wm_4. >> >> >> >> None of the Wm have an execution trace. They are stings in the alphabet >> >> of the TM H (and H^). The only thing that can have has any "execution >> >> trace" is H^ and we have no idea what it does after copying the tape >> >> other than it behaves like some other machine called H. >> > >> > To provide cognitive leverage and thus eliminate (Fred Brooks) >> > inessential complexity we will assume that the Turing Machine >> > Description of H^ is executed inside of (an interpreter) >> > Universal Turing Machine. >> >> That complicates matters -- so much so that you've lost track of the >> argument altogether. You find it hard enough to separate machine from >> representations without adding another layer of description. > > It reduces the names from three(H^, w^, Wm) to one: H^ That hides a distinction that you need to have made as explicit as possible. Other people might cope with it being hidden but I don't think you can. Furthermore, you now introduce a machine whose execution you remove from the explanation. This will confuse you even more. >> > The purpose for making this simplifying assumption is so >> > that we can assume H^, w^ and Wm are all names of the same >> > literal string. >> > >> > To provide a little more insight into what a potential halt >> > decider PHD might do to decide halting we could hypothesize >> > that it might simulate the execution of its input TMD in its >> > own UTM (interpreter). Since it could simulate this execution >> > step-by-step there is no risk that this PHD would get stuck >> > in an infinite loop. >> >> You *could* make that hypothesis but it would render any conclusion you >> draw from it contingent on it. The simpler argument (the one in Linz) >> assumes only that H is a halting decider. > > When ^H1(^H2, ^H3) is invoked ^H1 is evaluating what ^H2 would > do on its input ^H3. This requires ^H2 to know what ^H3 would > do on its input, yet it has no input. The simplest course of action (for me) is for you to continue to ignore what I say. My main purpose is to keep the record straight and your failure to address these points will be a clear signal to future readers that you don't know what's going on. -- Ben.