> On 25 Jan 2019, at 14:53, Philip Thrift <cloudver...@gmail.com> wrote:
> 
> 
> 
> On Friday, January 25, 2019 at 6:27:44 AM UTC-6, Bruno Marchal wrote:
> 
>> On 24 Jan 2019, at 15:19, Philip Thrift <cloud...@gmail.com <javascript:>> 
>> wrote:
>> 
>> 
>> 
>> On Thursday, January 24, 2019 at 7:14:15 AM UTC-6, Bruno Marchal wrote:
>> 
>>> On 23 Jan 2019, at 19:01, Philip Thrift <cloud...@gmail.com <>> wrote:
>>> 
>>> 
>>> 
>>> On Wednesday, January 23, 2019 at 5:52:01 AM UTC-6, Bruno Marchal wrote:
>>> 
>>>> On 22 Jan 2019, at 01:49, Philip Thrift <cloud...@gmail.com <>> wrote:
>>>> 
>>>> One of the oddest of things is when physicists use the language of 
>>>> (various) theories of physics to express what can or cannot be the case. 
>>>> It's just a language, which is probably wrong.
>>>> 
>>>> There is a sense in which the Church/Turing thesis is true: All out 
>>>> languages are Turing in their syntax and grammar. What they refer to is 
>>>> another matter (pun intended).
>>> 
>>> They refer to the set of computable functions, or to the universal machine 
>>> which understand that language. But not all language are Turing universal. 
>>> Only the context sensitive automata (in Chomski hierarchy) are Turing 
>>> universal. Simple languages, like the “regular” one are typically not 
>>> Turing universal. Bounded loops formalism cannot be either.
>>> 
>>> But the notion of language is ambiguous with respect to computability, and 
>>> that is why I prefer to avoid that expression and always talk about 
>>> theories (set of beliefs) or machine (recursively enumerable set of 
>>> beliefs), which avoids ambiguity. 
>>> For example, is “predicate calculus” Turing universal? We can say yes, 
>>> given that the programming language PROLOG (obviously Turing universal) is 
>>> a tiny subset of predicate logic. But we can say know, if we look at 
>>> predicate logic as a theory. A prolog program is then an extension of that 
>>> theory, not something proved in predicate calculus.
>>> Thus, I can make sense of your remark. Even the language with only one 
>>> symbol {I}, and the rules that “I” is a wff, and if x is wwf, then Ix is 
>>> too, can be said Turing universal, as each program can be coded by a 
>>> number, which can be coded by a finite sequence of I. But of course, that 
>>> makes the notion of “universality” empty, as far as language are concerned. 
>>> Seen as a theory, predicate calculus is notoriously not universal. Even 
>>> predicate calculus + the natural numbers, and the law of addition, 
>>> (Pressburger arithmetic) is not universal. Or take RA with its seven 
>>> axioms. Taking any axiom out of it, and you get a complete-able theory, and 
>>> thus it cannot be Turing complete.
>>> 
>>> Bruno
>>> 
>>> 
>>> 
>>> Here's an example of a kind of "non-digital" language:
>>> 
>>> More Analog Computing Is on the Way
>>> https://dzone.com/articles/more-analog-computing-is-on-the-way 
>>> <https://dzone.com/articles/more-analog-computing-is-on-the-way>
>>> 
>>> 
>>> The door on this new generation of analog computer programming is 
>>> definitely open. Last month, at the Association for Computing Machinery’s 
>>> (ACM) conference on Programming Language Design and Implementation, a paper 
>>>  <https://people.csail.mit.edu/sachour/res/pldi16_arco.pdf>was presented 
>>> that described a compiler that uses a text based, high-level, abstraction 
>>> language to generate the necessary low-level circuit wiring that defines 
>>> the physical analog computing implementation. This research was done at 
>>> MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and 
>>> Dartmouth College. The main focus of their investigation was to improve the 
>>> simulation of biological systems. 
>>> 
>>> 
>>> Configuration Synthesis for ProgrammableAnalog Devices with Arco
>>> https://people.csail.mit.edu/sachour/res/pldi16_arco.pdf 
>>> <https://people.csail.mit.edu/sachour/res/pldi16_arco.pdf>
>>> 
>>> Programmable analog devices have emerged as a powerful
>>> computing substrate for performing complex neuromorphic
>>> and cytomorphic computations. We present Arco, a new
>>> solver that, given a dynamical system specification in the
>>> form of a set of differential equations, generates physically
>>> realizable configurations for programmable analog devices
>>> that are algebraically equivalent to the specified system.
>>> On a set of benchmarks from the biological domain, Arco
>>> generates configurations with 35 to 534 connections and 28
>>> to 326 components in 1 to 54 minutes.
>>> 
>>> 
>>> - pt
>> 
>> Intersting.
>> 
>> Yet, that does not violate the Church-Thesis, even if very useful FAPP. But 
>> such computations arise in arithmetic, either directly, or through a 
>> infinite sequence of approximations. If all decimals of the analog 
>> phenomenon needs to be taken into account, then we are out of my working 
>> hypothesis, and even evolution theory becomes wrong, as evolution and life 
>> becomes sequences of miracles. But the goal of the authors here is not 
>> learning anything in metaphysics, just doing efficacious machine. In that 
>> case mechanism explains the plausible necessity of such moves, including 
>> quantum computations (which also do not violate Church’s thesis).
>> 
>> Bruno
>> 
>> 
>> 
>> 
>> 
>> 
>> I don't believe in (or know what are) miracles (although a real 
>> hypercomputer - one you could give any statement of arithmetic to - e.g. 
>> Goldbach's conjecture  -  and it could check through all - infinite number 
>> of -  integers and tell you "true" or "false" within the hour - would be 
>> basically a miracle), but I do think that substrate matters.
>> 
>> Hence in the PLTOS view (program, language, transformer/compiler, object, 
>> substrate), substrate can't be eliminated in the semantics of program. In 
>> other words, in real programming, there are no such things as 
>> substrate-independent programs.
>> 
> 
> Because you assume some primary substrate. And then you need, coherently, to 
> assume no-mechanism. No problem, but the current evidence favours Mechanism, 
> and there has never been any evidence for substrate. Adding substrate in the 
> picture makes the mind-body problem almost non soluble, at least without 
> invoking some precise non computationalist theory of mind. I start from the 
> computationalist of mind, shows that we have to derive a phenomenology of 
> matter in a special (self-referentially based) manner, and nature seems to 
> confirm this. The illusion of matter is easier to explain once we have a 
> theory of consciousness, than to derive a theory of consciousness from some 
> notion substrate (which are conceived usually as being inert).
> We are working in different theories. You might think about a way to motivate 
> your ontological commitment in some primitive substance. The books in physics 
> does not provide such motivation, as they do not aboard the mind-body problem 
> (even if Everett Quantum Mechanics already look like a solution to the 
> mechanist mind-body problem).
> 
> Bruno
> 
> 
> 
> 
> 
> Just to note that the "substrate" terminology is used in computing (as above):
> 
>   Programmable analog devices have emerged as a powerful computing substrate
>   for performing complex neuromorphic and cytomorphic computations.  
> 
> It's a word combined with "computing" like love and marriage.

But those analog computations, although they could be very useful in practice, 
do not change the consequence of the theory, unless you claim that they provide 
us with a method to compute new functions, which would violate the 
Church-Turing thesis (which I doubt very much). Keep in mind that with 
mechanism, the physical reality has analog part, which might or not be used by 
our bodies, although there are no evidences (that I know) for this. I follow 
the idea of not adding any hypothesis in a theory, unless there are strong 
evidences for them.

Bruno




> 
> - pt
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Everything List" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to everything-list+unsubscr...@googlegroups.com 
> <mailto:everything-list+unsubscr...@googlegroups.com>.
> To post to this group, send email to everything-list@googlegroups.com 
> <mailto:everything-list@googlegroups.com>.
> Visit this group at https://groups.google.com/group/everything-list 
> <https://groups.google.com/group/everything-list>.
> For more options, visit https://groups.google.com/d/optout 
> <https://groups.google.com/d/optout>.

-- 
You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at https://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.

Reply via email to