Fwd: Re: Understanding you-folks

[glen ep ropella]

Seems my other email address is jammed up ... made it to the archives, though.

-------- Forwarded Message --------
Subject: Re: [FRIAM] Understanding you-folks
Date: Wed, 6 Jul 2016 13:51:13 -0700
From: glen ☣ <[hidden email]>
To: The Friday Morning Applied Complexity Coffee Group <[hidden email]>


Well, we're in some rarefied air, now ... but since most of you are in Santa Fe, you're used to it.  I'll give you 3 ways (I find interesting) that meaning might be non-computational (or at least not 1st order):

    1) Rosen: loopiness/closure,
    2) situatedness, embeddedness (Penrose: quantum embeddedness), and
    3) coherence (who?).

(1) has to do with higher order operations.  A variable takes on meaning when (partially) convolved into an anticipatory agent ... some process that expects/anticipates the future.  (2) A variable takes on meaning when it interacts with the milieu (probably bound by a light cone). And (3) a variable takes on meaning when/if it perfectly integrates with every sentence (again probably bound by some inferential proximity) in the system.

These are types of binding that are distinct from, say, plugging in a constant or yet another schema.

On 07/06/2016 01:34 PM, Marcus Daniels wrote:
> What does it mean to _mean_ something?   Just grounding in some real world phenomenology?   Or does it require sensors and actuators -- robotics?

--
☣ glen

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

 "(1) has to do with higher order operations.  A variable takes on meaning when (partially) convolved into an anticipatory agent ... some process that expects/anticipates the future.  (2) A variable takes on meaning when it interacts with the milieu (probably bound by a light cone). And (3) a variable takes on meaning when/if it perfectly integrates with every sentence (again probably bound by some inferential proximity) in the system.

These are types of binding that are distinct from, say, plugging in a constant or yet another schema."

I don't see how this is unexpected or suggests anything that can't be computed.   The agent takes a higher order function as an argument and then uses it.  That is exactly like plugging in a constant.   In fact, any decent compiler that inlines using interprocedural optimization will actually treat it that way.  The generated object code will be the caller function with the argument incorporated into it, e.g. g(f,m,s) turns into just h(m,s).   And users of languages like Haskell do this sort of partial application all the time [without particular concern for how it works].   If g is the anticipatory agent that is responding to other stimuli, s, then f may take on a different meaning than if it were incorporated into some other function like g'.   Likewise if there are m and m' for different milieu environments different behaviors could occur for h(m,s) or h(m',s).    That being realized, it doesn't mean that f has any magical meaning.  It can still be understood completely without th
 at coupling.  

Marcus
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[glen ep ropella]

I don't disagree with you.  But the question is less about whether any part of "an answer" is definable as computation and more about a value judgement on the results (or inputs) of any particular computation.  If there is such a thing in the universe as a non-computational process (oracle) that sits inside the agent (or lurking out in the milieu), then when the computational cascade hits that non-computational process, that's when the binding/grounding/meaning obtains.

It's obvious that higher order processes (like quantification over quantifications) can be at least simulated in any modern computer programming language.  (assuming the parallelism theorem)  And languages like Coq can even help express intuitionistic logics.  Whether such things are exactly equivalent to the higher order math we use in things like analysis is an open question, I think(!).  I could easily be wrong, since I don't really grok things like HoTT.  But even if unification were demonstrable everywhere, it's still a reasonable question for the lay person to wonder about.  Where does vernacular "computation" stop and this high-falutin fancy-pants "computation" begin?  The same sort of question occurs in questions about the neural correlates of consciousness.



On 07/06/2016 04:00 PM, Marcus Daniels wrote:
> I don't see how this is unexpected or suggests anything that can't be computed.   The agent takes a higher order function as an argument and then uses it.  That is exactly like plugging in a constant.   In fact, any decent compiler that inlines using interprocedural optimization will actually treat it that way.  The generated object code will be the caller function with the argument incorporated into it, e.g. g(f,m,s) turns into just h(m,s).   And users of languages like Haskell do this sort of partial application all the time [without particular concern for how it works].   If g is the anticipatory agent that is responding to other stimuli, s, then f may take on a different meaning than if it were incorporated into some other function like g'.   Likewise if there are m and m' for different milieu environments different behaviors could occur for h(m,s) or h(m',s).    That being realized, it doesn't mean that f has any magical meaning.  It can still be understood completely without th
>  at coupling.


--
glen ep ropella ⊥ 971-280-5699
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

"Where does vernacular "computation" stop and this high-falutin fancy-pants "computation" begin?  The same sort of question occurs in questions about the neural correlates of consciousness."

I don't buy there is a meaningful distinction -- I mean one that should be preserved -- between those who do analysis and those who do calculations.    (There's history and prejudice about this of course.)   There's just differences in domain knowledge, and a formalism gap where misunderstandings occur.    Ideally there would be automated proofs (and context) all the way down.  If a code can be shown to be correct and as intended, and tolerances of the machine are known as are numerical sensitivities in the algorithms, then there is no reason to trust analytical results more or less than calculations.   The latter just takes longer.  As for neural correlates of consciousness, same thing:  If it turns out there are neat, modular ways to model cognitive function, great, but decoupling different kinds of evidence into stove-piped areas of expertise is just throwing away information.

Marcus    
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[gepr]

Hm.  I can't shake the feeling you're relying on some ambiguity in "meaningful distinction".  If you admit distinctions in things like domain knowledge, correctness, verified code, tolerances, sensitivities, etc., then why not admit there are meaningful distinctions in _types_ of computation?

Sure, it all may boil down to distinctions of degree rather than kind.  But (going back to definiteness) any 2 points in a continuum are still distinct points, even if they are arbitrarily close.  And discretization of continuous data has proven a valuable technique.  So, we're not going to stop doing it.  Hence, we'll end up with at least 2 types of computation, anyway, the one called "living systems" versus the purely mechanical ... even if, in full reduction, they are fundamentally the same kind.  So, we may as well allow the distinction now and see where it takes us.


On 07/06/2016 06:30 PM, Marcus Daniels wrote:
> I don't buy there is a meaningful distinction -- I mean one that should be preserved -- between those who do analysis and those who do calculations.    (There's history and prejudice about this of course.)   There's just differences in domain knowledge, and a formalism gap where misunderstandings occur.    Ideally there would be automated proofs (and context) all the way down.  If a code can be shown to be correct and as intended, and tolerances of the machine are known as are numerical sensitivities in the algorithms, then there is no reason to trust analytical results more or less than calculations.   The latter just takes longer.  As for neural correlates of consciousness, same thing:  If it turns out there are neat, modular ways to model cognitive function, great, but decoupling different kinds of evidence into stove-piped areas of expertise is just throwing away information.


--
☢ glen

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Owen Densmore]

Just to calibrate: has the OP been satisfied? 

I *think* so, we discussed FSM's discussing their input string and their final state and whether that was the designated accept state.

And tho a Turing Machine is more than a FSM, the vocabulary of states, input strings and so on should answer the OP.

I'm not sure the additional ideas on computation were coherent enough to add to his interest, but then, knowing Nick, I could be wrong!

Hope the book reading is progressing with success, given our help.

   -- Owen

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

 ``Hence, we'll end up with at least 2 types of computation, anyway, the one called "living systems" versus the purely   mechanical ... even if, in full reduction, they are fundamentally the same kind.  So, we may as well allow the  distinction now and see where it takes us. ''

When the result of a calculation is surprising, it is typical that people dig up analytical results to test against.   Usually analytical results are very narrowly scoped and can't test the full capabilities of a calculating device.   If they could there would be no need for the calculating device.   In this situation it would be better if the theoreticians could participate in more of the implementation of the device or at least review a list of properties of the code that can be shown to be true.   (And have confidence the properties were true.)  The vernacular of "computation" puts the device implementation outside of the scope of the theoreticians.   It's another set of people that do that, usually from an under-specified set of requirements.   The division between the vernacular and the fancy-pants version of computation encourages underspecification.   Conversely, having a higher level conception of computation can facilitate the engineers to move toward the science too.

Marcus
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Stephen Guerin-5]

Nick,

Owen asks:

> has the OP (original post) been satisfied?  

Has the this email thread answered your original question what an Accept state is? And why it is called an Accept state?

Are we in an accept or reject state. Or like many threads is this non-halting?

-S

On Thu, Jul 7, 2016 at 8:10 AM, Owen Densmore <[hidden email]> wrote:

Just to calibrate: has the OP been satisfied? 

I *think* so, we discussed FSM's discussing their input string and their final state and whether that was the designated accept state.

And tho a Turing Machine is more than a FSM, the vocabulary of states, input strings and so on should answer the OP.

I'm not sure the additional ideas on computation were coherent enough to add to his interest, but then, knowing Nick, I could be wrong!

Hope the book reading is progressing with success, given our help.

   -- Owen

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[glen ep ropella]

Heh, to be as clear as possible, there were 4 questions in the OP and several follow-up questions, summarized below.  I think the additional ideas on computation were (mostly) addressing the follow-up questions, particularly the _exploration_ of the idea that not all inference is computational.  But those additional ideas also address the OP question #2 to some extent.  We have 1 answer to OP #1 from Dave.

On 07/02/2016 08:30 AM, Nick Thompson wrote:> Dear Friammers,
> 1.       Has anybody read this book?
> 2.       Do you understand it?
> 3.       WTF is an Accept State?
> 4.       And why is it called an “Accept State?”

On 07/05/2016 06:25 PM, Nick Thompson wrote:
> If one has to use an “artificial” stop rule such as “quit when you get to the tenth decimal point”, is such a problem deemed “computable” or “non-computable”?  Can one “compute” the square root of two?

On 07/06/2016 11:33 AM, Nick Thompson wrote:> Thanks, Glen,
> I assume that the following is NOT a program in your sense.
> ;;Compute the sum of 2 and 2;;.
> Begin
> Ask Dad, "Dad, what is the sum of 2 and 2?
> Dad says, "Four"
> Four
> End.  
> It is, however, an algorithm, right?

On 07/06/2016 12:05 PM, Nick Thompson wrote:
> I guess what I was fishing for is some sort of exploration of the idea that not all procedures for arriving at answers are computations.  

On 07/07/2016 11:32 AM, Stephen Guerin wrote:
> Nick,
>
> Owen asks:
>> has the OP (original post) been satisfied?  
>
> Has the this email thread answered your original question what an Accept state is? And why it is called an Accept state?
>
> Are we in an accept or reject state. Or like many threads is this non-halting?

> On Thu, Jul 7, 2016 at 8:10 AM, Owen Densmore <[hidden email] <mailto:[hidden email]>> wrote:
>>
>>     Just to calibrate: has the OP been satisfied?
>>
>>     I *think* so, we discussed FSM's discussing their input string and their final state and whether that was the designated accept state.
>>
>>     And tho a Turing Machine is more than a FSM, the vocabulary of states, input strings and so on should answer the OP.
>>
>>     I'm not sure the additional ideas on computation were coherent enough to add to his interest, but then, knowing Nick, I could be wrong!
>>
>>     Hope the book reading is progressing with success, given our help.

--
glen ep ropella ⊥ 971-280-5699

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[gepr]

But what you're arguing for is essentially the idea that all special-purpose devices (should not can) be replaced by universal computers.  That's unreasonable.  It makes good engineering and scientific sense to divvy up types of computation.  The distinction in the question of whether the kind of computation we do on paper is the same as what IBM Watson does is a meaningful distinction.  It's also meaningful to try to dissolve the distinction.  But that doesn't devalue the tasks and thinking that rely on the distinction.

On 07/07/2016 11:18 AM, Marcus Daniels wrote:
> When the result of a calculation is surprising, it is typical that people dig up analytical results to test against.   Usually analytical results are very narrowly scoped and can't test the full capabilities of a calculating device.   If they could there would be no need for the calculating device.   In this situation it would be better if the theoreticians could participate in more of the implementation of the device or at least review a list of properties of the code that can be shown to be true.   (And have confidence the properties were true.)  The vernacular of "computation" puts the device implementation outside of the scope of the theoreticians.   It's another set of people that do that, usually from an under-specified set of requirements.   The division between the vernacular and the fancy-pants version of computation encourages underspecification.   Conversely, having a higher level conception of computation can facilitate the engineers to move toward the science too.

--
☢ glen
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

I don't think special purpose devices should be replaced by universal computers.   Universal computers are slow for some things.   However, universal computers should have as high of fidelity models of those devices as possible.     It should be possible "in the future" to understand, with the precision of a program on a universal computer, the justification of every decision leading to a device and thus the outputs of the device.  Technical papers should basically be literate computer programs.  

-----Original Message-----
From: Friam [mailto:[hidden email]] On Behalf Of glen ?
Sent: Thursday, July 07, 2016 1:48 PM
To: The Friday Morning Applied Complexity Coffee Group <[hidden email]>
Subject: Re: [FRIAM] Fwd: Re: Understanding you-folks


But what you're arguing for is essentially the idea that all special-purpose devices (should not can) be replaced by universal computers.  That's unreasonable.  It makes good engineering and scientific sense to divvy up types of computation.  The distinction in the question of whether the kind of computation we do on paper is the same as what IBM Watson does is a meaningful distinction.  It's also meaningful to try to dissolve the distinction.  But that doesn't devalue the tasks and thinking that rely on the distinction.

On 07/07/2016 11:18 AM, Marcus Daniels wrote:
> When the result of a calculation is surprising, it is typical that people dig up analytical results to test against.   Usually analytical results are very narrowly scoped and can't test the full capabilities of a calculating device.   If they could there would be no need for the calculating device.   In this situation it would be better if the theoreticians could participate in more of the implementation of the device or at least review a list of properties of the code that can be shown to be true.   (And have confidence the properties were true.)  The vernacular of "computation" puts the device implementation outside of the scope of the theoreticians.   It's another set of people that do that, usually from an under-specified set of requirements.   The division between the vernacular and the fancy-pants version of computation encourages underspecification.   Conversely, having a higher level conception of computation can facilitate the engineers to move toward the science too.

--
☢ glen
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[gepr]

OK.  But you did express that you thought the distinction (between paper math and computation) isn't meaningful (at least not in perpetuity).  Yet you admit that (in perpetuity) we should preserve the distinction at least for the sake of efficiency/performance.  You have to admit that can seem paradoxical.

Re: _technical_ papers being literate computer programs ... I agree.  But a recurring theme in this forum is the poor job journalists do communicating scientific efforts.  Analogously, we can predict that when/if all technical papers are literate programs, we'll have a similar problem.  This same conversation will continue to occur when the Nicks of the world ask the you-folks of the world what some program means.  So the distinction will persist as long as there are general intelligences (Nicks) attempting to parse domain-specific artifacts.

The solution you imply (a plethora of easily instantiable models on universal computers) is a good one, I think.  That way the generally intelligent, specifically ignorant can at least understand special-purpose devices through analogical reasoning.  But you could have said that outright without the above paradox! 8^)


On 07/07/2016 01:00 PM, Marcus Daniels wrote:
> I don't think special purpose devices should be replaced by universal computers.   Universal computers are slow for some things.   However, universal computers should have as high of fidelity models of those devices as possible.     It should be possible "in the future" to understand, with the precision of a program on a universal computer, the justification of every decision leading to a device and thus the outputs of the device.  Technical papers should basically be literate computer programs.


--
☣ glen

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

"OK.  But you did express that you thought the distinction (between paper math and computation) isn't meaningful (at least not in perpetuity).  Yet you admit that (in perpetuity) we should preserve the distinction at least for the sake of efficiency/performance.  You have to admit that can seem paradoxical."

I don't understand why you connect special purpose devices with paper math vs. computation.   I claim the problem with paper math is that 1) the former does not carry or enforce correctness checks, 2) it is not put in context -- things are pulled out of thin air as "the reader should know this", and 3) there isn't a formal mapping or harness to a universal computer.  So, going back to the "Ask Dad" approach to computing things, one could imagine a detailed model of the sorts of calculations Dad could do very well, but nonetheless leave the actual calculation to him.  I could share this model with other people and we could agree it was a "Certified Dad compliant" interface.
Regarding 2), ideally a paper's citations and bibliography will provide nodes on the semantic graph to start pulling, but it isn't required or consistently enforced by publishers and it certainly isn't machine readable.   The audience of today's technical literature is assumed to be other human domain experts, not, say, a Watson.

"Re: _technical_ papers being literate computer programs ... I agree.  But a recurring theme in this forum is the poor job journalists do communicating scientific efforts.  Analogously, we can predict that when/if all technical papers are literate programs, we'll have a similar problem.  This same conversation will continue to occur when the Nicks of the world ask the you-folks of the world what some program means.  So the distinction will persist as long as there are general intelligences (Nicks) attempting to parse domain-specific artifacts."

If all domain-specific artifacts were built up with machine readable ontologies, then the general intelligent agents will have threads to pull to start putting the artifacts in context.   Perhaps some kinds of agents, like humans, would benefit from additional `analogy modules' to assist with mapping large semantic graphs into similar pre-existing ones.  That would be an accelerator for learning, not a question of having a sufficient semantic representation.

Marcus
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[glen ep ropella]

On 07/07/2016 02:11 PM, Marcus Daniels wrote:
> I don't understand why you connect special purpose devices with paper math vs. computation.   I claim the problem with paper math is that 1) the former does not carry or enforce correctness checks, 2) it is not put in context -- things are pulled out of thin air as "the reader should know this", and 3) there isn't a formal mapping or harness to a universal computer.

Well, I disagree with all 3 of those assertions.  But it's a soft disagreement.  I rely on softening the definitions of "correctness checks", "put in context", "harness to universal computer".  Paper math is a social enterprise and that sociality is the correctness check.  Similarly, it is put in the context of its application and/or the larger body of math.  And the "universal computer" it is harnessed to is (proximally) the human brain/CNS and (distally) logic/reasoning as a whole.

Paper math is a semi-semantic computation.  This is nothing more than a restatement of Hilbert's program.  It is a (canonical) use case of a special purpose device: the human brain.  It's interesting and meaningful to ask whether or not computers can do the math humans do.  I think the answer keeps coming up "yes" ... but people smarter than me are not convinced.  So, we shouldn't be stubbornly reductionist.  It hurts nobody to let them have the distinction ... at least for now and possibly forever.

> If all domain-specific artifacts were built up with machine readable ontologies, then the general intelligent agents will have threads to pull to start putting the artifacts in context.   Perhaps some kinds of agents, like humans, would benefit from additional `analogy modules' to assist with mapping large semantic graphs into similar pre-existing ones.  That would be an accelerator for learning, not a question of having a sufficient semantic representation.

Well, OK.  But there's still an assumption that the infrastructure will be complete, high quality, and credible.  Is there room for gaming and misinformation in such systems?  Can our ontological mesh lie to people?  ... create idiot savants? ... be used to rig elections?  If so, then it most assuredly _is_ a question of sufficient semantic grounding.

--
glen ep ropella ⊥ 971-280-5699

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

"It's interesting and meaningful to ask whether or not computers can do the math humans do.  I think the answer keeps coming up "yes" ... but people smarter than me are not convinced.  So, we shouldn't be stubbornly reductionist.  It hurts nobody to let them have the distinction ... at least for now and possibly forever."

I'm claiming that a universal computer is a good way to normalize the forms and to check that the manipulations between the forms are sound.   The point is to track what the special purpose machines are doing, not to do it.  The theoreticians would still do the creative side.
More than just LaTeX but less than AI.  

Marcus

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[gepr]

I agree completely!

On 07/07/2016 03:11 PM, Marcus Daniels wrote:
> I'm claiming that a universal computer is a good way to normalize the forms and to check that the manipulations between the forms are sound.   The point is to track what the special purpose machines are doing, not to do it.  The theoreticians would still do the creative side.
> More than just LaTeX but less than AI.


--
☣ glen

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Nick Thompson]

Sorry,  

It took me a bit to realize that I was the OP.  

This has been tremendously useful for me, because it has given me a sense of what you all agree on and what is controversial.  Author of the book, of course, writes as if everything he says would be agreed upon by everybody in theworld, including.  

I will restart the book with all of this discussion behind me.  

Thanks to you all,

OP

Nicholas S. Thompson
Emeritus Professor of Psychology and Biology
Clark University
http://home.earthlink.net/~nickthompson/naturaldesigns/


-----Original Message-----
From: Friam [mailto:[hidden email]] On Behalf Of glen ep ropella
Sent: Thursday, July 07, 2016 3:38 PM
To: The Friday Morning Applied Complexity Coffee Group <[hidden email]>
Subject: Re: [FRIAM] Fwd: Re: Understanding you-folks


Heh, to be as clear as possible, there were 4 questions in the OP and several follow-up questions, summarized below.  I think the additional ideas on computation were (mostly) addressing the follow-up questions, particularly the _exploration_ of the idea that not all inference is computational.  But those additional ideas also address the OP question #2 to some extent.  We have 1 answer to OP #1 from Dave.

On 07/02/2016 08:30 AM, Nick Thompson wrote:> Dear Friammers,
> 1.       Has anybody read this book?
> 2.       Do you understand it?
> 3.       WTF is an Accept State?
> 4.       And why is it called an “Accept State?”

On 07/05/2016 06:25 PM, Nick Thompson wrote:
> If one has to use an “artificial” stop rule such as “quit when you get to the tenth decimal point”, is such a problem deemed “computable” or “non-computable”?  Can one “compute” the square root of two?

On 07/06/2016 11:33 AM, Nick Thompson wrote:> Thanks, Glen,
> I assume that the following is NOT a program in your sense.
> ;;Compute the sum of 2 and 2;;.
> Begin
> Ask Dad, "Dad, what is the sum of 2 and 2?
> Dad says, "Four"
> Four
> End.  
> It is, however, an algorithm, right?

On 07/06/2016 12:05 PM, Nick Thompson wrote:
> I guess what I was fishing for is some sort of exploration of the idea that not all procedures for arriving at answers are computations.  

On 07/07/2016 11:32 AM, Stephen Guerin wrote:
> Nick,
>
> Owen asks:
>> has the OP (original post) been satisfied?  
>
> Has the this email thread answered your original question what an Accept state is? And why it is called an Accept state?
>
> Are we in an accept or reject state. Or like many threads is this non-halting?

> On Thu, Jul 7, 2016 at 8:10 AM, Owen Densmore <[hidden email] <mailto:[hidden email]>> wrote:
>>
>>     Just to calibrate: has the OP been satisfied?
>>
>>     I *think* so, we discussed FSM's discussing their input string and their final state and whether that was the designated accept state.
>>
>>     And tho a Turing Machine is more than a FSM, the vocabulary of states, input strings and so on should answer the OP.
>>
>>     I'm not sure the additional ideas on computation were coherent enough to add to his interest, but then, knowing Nick, I could be wrong!
>>
>>     Hope the book reading is progressing with success, given our help.

--
glen ep ropella ⊥ 971-280-5699

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com


============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Robert Wall]

Yikes! Sorry Stephen for misspelling your name, but at least I did it consistently.  😕

- Rebort

On Thu, Jul 7, 2016 at 6:30 PM, Nick Thompson <[hidden email]> wrote:

Sorry,

It took me a bit to realize that I was the OP.

This has been tremendously useful for me, because it has given me a sense of what you all agree on and what is controversial.  Author of the book, of course, writes as if everything he says would be agreed upon by everybody in theworld, including.

I will restart the book with all of this discussion behind me.

Thanks to you all,

OP

Nicholas S. Thompson
Emeritus Professor of Psychology and Biology
Clark University
http://home.earthlink.net/~nickthompson/naturaldesigns/


-----Original Message-----
From: Friam [mailto:[hidden email]] On Behalf Of glen ep ropella
Sent: Thursday, July 07, 2016 3:38 PM
To: The Friday Morning Applied Complexity Coffee Group <[hidden email]>
Subject: Re: [FRIAM] Fwd: Re: Understanding you-folks

Heh, to be as clear as possible, there were 4 questions in the OP and several follow-up questions, summarized below.  I think the additional ideas on computation were (mostly) addressing the follow-up questions, particularly the _exploration_ of the idea that not all inference is computational.  But those additional ideas also address the OP question #2 to some extent.  We have 1 answer to OP #1 from Dave.

On 07/02/2016 08:30 AM, Nick Thompson wrote:> Dear Friammers,
> 1.       Has anybody read this book?
> 2.       Do you understand it?
> 3.       WTF is an Accept State?
> 4.       And why is it called an “Accept State?”

On 07/05/2016 06:25 PM, Nick Thompson wrote:
> If one has to use an “artificial” stop rule such as “quit when you get to the tenth decimal point”, is such a problem deemed “computable” or “non-computable”?  Can one “compute” the square root of two?

On 07/06/2016 11:33 AM, Nick Thompson wrote:> Thanks, Glen,
> I assume that the following is NOT a program in your sense.
> ;;Compute the sum of 2 and 2;;.
> Begin
> Ask Dad, "Dad, what is the sum of 2 and 2?
> Dad says, "Four"
> Four
> End.
> It is, however, an algorithm, right?

On 07/06/2016 12:05 PM, Nick Thompson wrote:
> I guess what I was fishing for is some sort of exploration of the idea that not all procedures for arriving at answers are computations.

On 07/07/2016 11:32 AM, Stephen Guerin wrote:
> Nick,
>
> Owen asks:
>> has the OP (original post) been satisfied?
>
> Has the this email thread answered your original question what an Accept state is? And why it is called an Accept state?
>
> Are we in an accept or reject state. Or like many threads is this non-halting?

> On Thu, Jul 7, 2016 at 8:10 AM, Owen Densmore <[hidden email] <mailto:[hidden email]>> wrote:
>>
>>     Just to calibrate: has the OP been satisfied?
>>
>>     I *think* so, we discussed FSM's discussing their input string and their final state and whether that was the designated accept state.
>>
>>     And tho a Turing Machine is more than a FSM, the vocabulary of states, input strings and so on should answer the OP.
>>
>>     I'm not sure the additional ideas on computation were coherent enough to add to his interest, but then, knowing Nick, I could be wrong!
>>
>>     Hope the book reading is progressing with success, given our help.


--
glen ep ropella ⊥ <a href="tel:[phone]" value="[phone]">971-280-5699

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com


============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[Marcus G. Daniels]

Glen wrote:
    1) Rosen: loopiness/closure,
     [..]

  (1) has to do with higher order operations.  A variable takes on meaning when (partially) convolved into an anticipatory agent ... some process that expects/anticipates the future.  

I'm not sure if this is what you are getting at, but would the following scenario also be an instance of this:   An organization has a set of goals and let's say they are underspecified or the goals compete with one another.  Now, whenever someone  is upset about whatever they might be upset about (or sees some opportunity to schmooze to their superiors), they make an appeal, and, in some circumstances, their superiors make reference to the policies to bring order.    In doing so, they produce something of the form of an inverse problem.   "You all should see this is in an instance of a class of policy X."   Let's just say, for the sake of argument, that is not at all clear or is debatable.   Where does that leave the reader?   One probably does not question the superior court, as it were, because it is stated as fact.   Instead I think what happens is that the underlings search for generating functions to fit the set of constraints and then socialize the solutions to reduce future r
 isk.   That is, they seek (effective) unification that is deterministic.   1 answer, not 0, not more than 1.    But it isn't unification in the logic programming sense, it just looks that way.

This sort of system leads to each agent making a sort of master equation of the environment and using it to predict risk (and reward) from above.   Meaning does not exist until the underlings scurry around filling in the free variables with a self-consistent (and consensus) set of values.   I would claim that in the real world there is usually no shared typing system except in exceptional cases like the U.S. Judicial system.   Mostly it is just the evolution of anticipatory behaviors from high or low fitness.  Mommy and the kids just try to keep Daddy the tyrant from losing his cool, and in doing so evolve an effective control system.

Marcus

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com

[gepr]

Yes, I think so; except the goals need not be underspecified or contradictory.  The condition (or action or assertion) made by one of the anticipatory agents within the system can be an unambiguous member of the set defined by the policy.  The loopiness comes in because that condition is defined in reference to the policy that defines it.  Perhaps an example would be "A staff member is an employee (not a contractor) iff that staff member has all the properties generally thought to belong to employees (as opposed to contractors)."  This is a common ploy used by state tax agencies to extract extra taxes from corporations.  A book keeper in such a company might assert that some staffer should not be paid as an employee because they're the only employee who, say, telecommutes.

The binding/grounding/meaning in such a circular system can be volatile.  But it's not (necessarily) due to being unspecified/vague.  I suppose it's technically ambiguity, multi-valued.  The underlings, including _any_ member of the corporation like owners, board members, *EOs, etc., can use their influence to knead the corporation-level grounding of the policy in whatever way they can.  And it stays this way until the corporation officially files a request with the state agency to get a ruling on that particular condition (telecommuting).  Once the state makes the ruling, then that property (telecommuting) is "hard" bound/grounded one way or the other.


On 07/13/2016 06:58 AM, Marcus Daniels wrote:
> I'm not sure if this is what you are getting at, but would the following scenario also be an instance of this:   An organization has a set of goals and let's say they are underspecified or the goals compete with one another.  Now, whenever someone  is upset about whatever they might be upset about (or sees some opportunity to schmooze to their superiors), they make an appeal, and, in some circumstances, their superiors make reference to the policies to bring order.    In doing so, they produce something of the form of an inverse problem.   "You all should see this is in an instance of a class of policy X."   Let's just say, for the sake of argument, that is not at all clear or is debatable.   Where does that leave the reader?   One probably does not question the superior court, as it were, because it is stated as fact.   Instead I think what happens is that the underlings search for generating functions to fit the set of constraints and then socialize the solutions to reduce future r
>  isk.   That is, they seek (effective) unification that is deterministic.   1 answer, not 0, not more than 1.    But it isn't unification in the logic programming sense, it just looks that way.
>
> This sort of system leads to each agent making a sort of master equation of the environment and using it to predict risk (and reward) from above.   Meaning does not exist until the underlings scurry around filling in the free variables with a self-consistent (and consensus) set of values.   I would claim that in the real world there is usually no shared typing system except in exceptional cases like the U.S. Judicial system.   Mostly it is just the evolution of anticipatory behaviors from high or low fitness.  Mommy and the kids just try to keep Daddy the tyrant from losing his cool, and in doing so evolve an effective control system.


--
☢ glen
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com