[Fis] What is ³Agent²?

Bruno Marchal marchal at ulb.ac.be
Mon Oct 23 18:39:54 CEST 2017 

Dear Gordana,




On 20 Oct 2017, at 11:02, Gordana Dodig-Crnkovic wrote:

>
> Dear Terry, Bob, Loet
>
> Thank you for sharing those important thoughts about possible  
> choices for the definition of agency.
>
> I would like to add one more perspective that I find in Pedro’s  
> article which makes a distinction between matter-energy aspects and  
> informational aspects of the same physical reality. I believe that  
> on the fundamental level of information physics we have a good ND  
> simplest example how those two entangled aspects can be formally  
> framed.
> As far as I can tell, Terrys definition covers chemical and  
> biological agency.
> Do we want to include apart from fundamental physics also full  
> cognitive and social agency which are very much dominated by  
> informational aspects (symbols and language)?
> Obviously there is no information without physical implementation,


Hmm... I am not sure. Elementary arithmetic determines all semi- 
computable relative information state (with Oracles). So, with the  
numbers, once you accept the addition laws and the multiplication  
laws, information "grows" from inside, and consciousness differentiates.
When the information get deeper and deeper, in Bennett sense of debth,  
dreams can stabilize and physical reality are "correctly" inferred,  
and eventually derived from arithmetic.

That might not make your point below invalid.

It is yet an important metaphysical point. The incompleteness theorem  
entails the existence of a sort of canonical information flux, or  
consciousness differentiation internal to elementary arithmetic, or  
elementary combinators, or to any universal machinery (universal in  
the mathematical Church-Turing-Post-Kleene sense).

We can decide to consider the arithmetical beings being zombies, but  
this would entails a very special definition of matter to make it  
differ from the testable "arithmetical distribution".

We can't have weak mechanism and weak materialism, and the evidences  
might side on a mathematical (somehow theological or psychological)  
origin of the physical reality.

Incompleteness entails that all (platonist, classical reasoner  
machine) are confronted with many different, and conflicting, views  
about itself. Indeed it enforces the Theaetetus' distinctions, between  
true, provable, knowable, observable, sensible:

p,
[]p,
[]p&p,
[]p&~[]f,
[]p&~[]f&p

With p sigma_1 arithmetical (equivalently: partial computable) this  
gives a proposition account of a theology, testable as it explains how  
the physical laws emerges from some "dream percolation" in arithmetic.

The physical is very important, but like in Plato, it could be, and  
seemed to be, the border of another non physical, more mathematical,  
plausibly arithmetical, reality.



> but when we think about epistemology and the ways we know the world,  
> for us and other biological agents there is no physical interaction  
> without informational aspects.
> Can we somehow think in terms those two faces of agency?
> Without matter/energy nothing will happen, nothing can act in the  
> world but that which happens and anyone registers it, has  
> informational side to it.

Without matter/energy nothing physical will happen. But if we assume a  
very weak form of digital mechanism, arithmetic justfies limiting  
dreams, with rich indexical, relative amount of information, from  
"inside arithmetic". And what we take as the physical might be what  
emerges from a first person statistics on those dreams.

The logic of which is testable, and up to now, it matches the data  
(thanks to QM without coilapse of the wave).

It is just premature to conclude that information (in the 1p and 3p  
sense) needs the physical. The physical might be an invariant in a  
notion of normal sharable number dream. (A dream can be defined by a  
computation containing the emulation of a Löbian machine (they know  
they are universal) with respect to different or not universal numbers.

In arithmetic, the universal numbers infers that below their  
substitution level, if it exists, they are confronted to a statistics  
on infinity of universal numbers, and above, locally, only with a  
finite (but huge) number of universal machine/number.

I am aware I ask a huge spiritual or theological effort, coming back  
to Plotinus, and Parmenides, and Plato, if not Pythagoras.

But in epistemology, computable can be defined in very elementary  
theories and languages. The deep reason is the closure of the partial  
computable functions for cantor diagonalization (Gödel's called that a  
Miracle), and its price: the non computability of most predicate on  
most machines behaviors (like halting), and the loss of control and  
the art of letting go the things which go without saying.

The universal (Löbian) machine can already defeat all normative or  
reductionist theory about their first person. They know that their  
soul is not a machine!




> For human agency (given that matter/energy side is functioning)  
> information is what to a high degree drives agency.

OK.


>
> Do you think this would be a fruitful path to pursue, with “agency”  
> of elementary particles and agency of social institutions as two  
> limit cases?

Agency of elementary particles? I am not sure this would not make all  
the number relations into an agent. Social institution are closer,  
perhaps even more the corporations, but none are really autonomous. I  
don't know.


This was my second (and last) post of the week.

All the best and best to All,

Bruno





>
>
> Gordana
>
>
>
> ______________________________________________
> Gordana Dodig Crnkovic, Professor of Computer Science
> Department of Computer Science and Engineering
> Chalmers University of Technology
> School of Innovation, Design and Engineering, Mälardalen University
> http://www.mrtc.mdh.se/~gdc/
> General Chair of is4si summit 2017
> http://is4si-2017.org
>
>
> From: Fis <fis-bounces at listas.unizar.es> on behalf of Loet  
> Leydesdorff <loet at leydesdorff.net>
> Organization: University of Amsterdam
> Reply-To: "loet at leydesdorff.net" <loet at leydesdorff.net>
> Date: Friday, 20 October 2017 at 08:40
> To: 'Bob Logan' <logan at physics.utoronto.ca>, 'fis' <fis at listas.unizar.es 
> >
> Subject: Re: [Fis] What is “Agent”?
>
> Dear Bob and colleagues,
>
> I agree with the choice element. From a sociological perspective,  
> agency is usually defined in relation to structure. For example, in  
> terms of structure/actor contingencies. The structures provide the  
> background that bind us. Remarkably, Mark, we no longer define these  
> communalities philosophically, but sociologically (e.g., Merton,  
> 1942, about the institutional norms of science). An interesting  
> extension is that we nowadays not only perceive communality is our  
> biological origins (as species), but also in terms of communicative  
> layers that we construct and reproduce as inter-agency (interactions).
>
> The relation with the information issue is not obvious. I worked on  
> this a bit in the first half of the 90s:
> "Structure"/"Action" Contingencies and the Model of Parallel  
> Distributed Processing, Journal for the Theory of Social Behaviour  
> 23 (1993) 47-77.
> The Production of Probabilistic Entropy in Structure/Action  
> Contingency Relations, Journal of Social and Evolutionary Systems 18  
> (1995) 339-56.
> Best,
> Loet
>
> Loet Leydesdorff
> Professor emeritus, University of Amsterdam
> Amsterdam School of Communication Research (ASCoR)
> loet at leydesdorff.net ; http://www.leydesdorff.net/
> Associate Faculty, SPRU, University of Sussex;
> Guest Professor Zhejiang Univ., Hangzhou; Visiting Professor, ISTIC,  
> Beijing;
> Visiting Fellow, Birkbeck, University of London;
> http://scholar.google.com/citations?user=ych9gNYAAAAJ&hl=en
>
>
> From: Fis [mailto:fis-bounces at listas.unizar.es] On Behalf Of Bob Logan
> Sent: Friday, October 20, 2017 6:11 AM
> To: Terrence W. DEACON <deacon at berkeley.edu>
> Cc: fis <Fis at listas.unizar.es>
> Subject: Re: [Fis] What is “Agent”?
>
> Dear Terry and FIS friends - I agree with all that Terry has said  
> about agency. I do wish to however to point out that an agent has  
> choice and a non-agent has no choice. I would suggest that the  
> defining characteristic of an agent is choice and therefore an agent  
> must be a living organism and all living organisms are agents.  
> Agents/living organisms have choice or are capable of choice or  
> agency and they are the only things that have choice or can  
> interpret information. Abiotic non-agents do not have information  
> because they have no choice. We humans can have information about  
> abiotic objects but those objects themselves do not have that  
> information as they have no mind to be informed. That includes this  
> email post, it is abiotic an has no agency. It has information by  
> virtue of you reading it because you are able to interpret the  
> visual signs with which I have recorded my thoughts. Marshall  
> McLuhan would add to my comments that “the user is the content” as  
> well as saying that Shannon’s work was not a theory of information  
> but a "theory of transportation”. I think of Shannon’s work in a  
> similar light. I also do not regard Shannon’s work as a theory of  
> information but it is a theory of signals. Shannon himself said his  
> theory was not about meaning and I say what is information without  
> meaning and that therefore Shannon only had a theory of signals.
>
> Another insight of McLuhan’s that of figure and ground is useful to  
> understand why we have so many different definitions of information.  
> McLuhan maintained that one could not understand a figure unless one  
> understood the ground in which it operates in. (McLuhan might have  
> gotten this idea from his professor at Cambridge, I. A. Richards,  
> who said that in order to communicate one needs to feedforward [he  
> coined the term btw] the context of what one is communicating.) The  
> different definitions of information we have considered are a result  
> of the different contexts in which the term information is used. We  
> should also keep in mind that all words are metaphors and metaphor  
> literally means to carry across, derived from the Greek meta  
> (literally ‘across') and phorein (literally 'to carry'). So the word  
> information has been carried across from one domain or area of  
> interest to another. It entered the English language as the noun  
> associated with the verb 'to inform', i.e. to form the mind. Here is  
> an excerpt from my book What Is Information? (available for free at  
> demopublishing.com):
> "Origins of the Concept of Information - We begin our historic  
> survey of the development of the concept of information with its  
> etymology. The English word information according to the Oxford  
> English Dictionary (OED) first appears in the written record in 1386  
> by Chaucer: 'Whanne Melibee hadde herd the grete skiles and resons  
> of Dame Prudence, and hire wise informacions and techynges.' The  
> word is derived from Latin through French by combining the word  
> inform meaning giving a form to the mind with the ending “ation”  
> denoting a noun of action. This earliest definition refers to an  
> item of training or molding of the mind.” This is why abiotic  
> objects have no information as I claimed above because they have no  
> mind that can be informed.
> I hope that by informing you of the origin of the word information I  
> have shed some light on our confusion about what is information and  
> why we have so many definitions of it. It might even shed some light  
> for that matter as to what is an agent. Got the ticket? If so that  
> makes me a ticket agent. I hope you get the joke. all the best - Bob
>
>
>
> ______________________
>
> Robert K. Logan
> Prof. Emeritus - Physics - U. of Toronto
> Fellow University of St. Michael's College
> Chief Scientist - sLab at OCAD
> http://utoronto.academia.edu/RobertKLogan
> www.researchgate.net/profile/Robert_Logan5/publications
> https://www.physics.utoronto.ca/people/homepages/logan/
>
>
> On Oct 19, 2017, at 7:11 PM, Terrence W. DEACON  
> <deacon at berkeley.edu> wrote:
>
> AUTONOMOUS AGENCY: The definition I propose for autonomous agency It  
> is open to challenge. Of course, there are many ways that we use the  
> term 'agent' in more general and metaphoric ways. I am, however,  
> interested in the more fundamental conception that these derived  
> uses stem from. I do not claim that this definition is original, but  
> rather that it is what we implicitly understand by the concept. So  
> if this is not your understanding I am open to suggestions for  
> modification.
>
> I should add that it has been a recent goal of my work to describe  
> an empirically testable simplest model system that satisfies this  
> definition. Those of you who are familiar with my work will  
> recognize that this is what I call an autogenic or teleodynamic  
> system. In this context, however, it is only the adequacy of the  
> definition that I am interested in exploring. As in many of the  
> remarks of others on this topic it is characterized by strange-loop  
> recursivity, self-reference, and physicality. And it may be worth  
> while describing how this concept is defined by e.g. Hofstadter, von  
> Foerster, Luhmann, Moreno, Kauffman, Barad, and others, to be sure  
> that we have covered the critical features and haven't snuck in any  
> "demons". In my definition, I have attempted to avoid any cryptic  
> appeal to observers or unexamined teleological properties, because  
> my purpose is instead to provide a constructive definition of what  
> these properties entail and why they are essential to a full  
> conception of information.
>
> CENTRALITY OF NORMATIVE PROPERTIES: A critical factor when  
> discussing agency is that it is typically defined with respect to  
> "satisfaction conditions" or "functions" or "goals" or other  
> NORMATIVE properties. Normative properties are all implicitly  
> teleological. They are irrelevant to chemistry and physics. The  
> concept of an "artificial agent" may not require intrinsic teleology  
> (e.g. consider thermostats or guidance systems - often described as  
> teleonomic systems) but the agentive properties of such artifacts  
> are then implicitly parasitic on imposed teleology provided by some  
> extrinsic agency. This is of course implicit also in the concepts of  
> 'signal' and 'noise' which are central to most information concepts.  
> These are not intrinsic properties of information, but are  
> extrinsically imposed distinctions (e.g. noise as signal to the  
> repair person). So I consider the analysis of agency and its  
> implicit normativity to be a fundamental issue to be resolved in our  
> analysis of information. Though we can still bracket any  
> consideration of agency from many analyses my simply assuming it  
> (e.g. assumed users, interpreters, organisms and their functions,  
> etc.), but this explicitly leaves a critical defining criterion  
> outside the analysis. In these cases, we should just be clear that  
> in doing so we have imported unexplained boundary conditions into  
> the analysis by fiat. Depending on the goal of the analysis (also a  
> teleological factor) this may be unimportant. But the nature and  
> origin of agency and normativity remain foundational questions for  
> any full theory of information.
>
> On Thu, Oct 19, 2017 at 12:47 PM, Stanley N Salthe <ssalthe at binghamton.edu 
> > wrote:
>> Here is an interesting recent treatment of autonomy.
>>
>> Alvaro Moreno and Matteo Mossio: Biological Autonomy: A Philosophical
>>
>> and Theoretical Enquiry (History, Philosophy and Theory of the Life  
>> Sciences 12);
>>
>> Springer, Dordrecht, 2015, xxxiv + 221 pp., $129 hbk, ISBN  
>> 978-94-017-9836-5
>>
>>
>>
>> STAN
>>
>>
>> On Thu, Oct 19, 2017 at 11:44 AM, Terrence W. DEACON <deacon at berkeley.edu 
>> > wrote:
>>> AN AUTONOMOUS AGENT IS A DYNAMICAL SYSTEM ORGANIZED TO BE CAPABLE  
>>> OF INITIATING PHYSICAL WORK TO FURTHER PRESERVE THIS SAME CAPACITY  
>>> IN THE CONTEXT OF  INCESSANT EXTRINSIC AND/OR INTRINSIC TENDENCIES  
>>> FOR THIS SYSTEM CAPACITY TO DEGRADE.
>>>
>>> THIS ENTAILS A CAPACITY TO ORGANIZE WORK THAT IS SPECIFICALLY  
>>> CONTRAGRADE TO THE FORM OF THIS DEGRADATIONAL INFLUENCE, AND THUS  
>>> ENTAILS A CAPACITY TO BE INFORMED BY THE EFFECTS OF THAT INFLUENCE  
>>> WITH RESPECT TO THE AGENT’S CRITICAL ORGANIZATIONAL CONSTRAINTS.
>>>
>>> On Wed, Oct 18, 2017 at 6:00 PM, Koichiro Matsuno <CXQ02365 at nifty.com 
>>> > wrote:
>>>> On 19 Oct 2017 at 6:42 AM, Alex Hankey wrote:
>>>>
>>>> the actual subject has to be non-reducible and fundamental to our  
>>>> universe.
>>>>
>>>>    This view is also supported by Conway-Kochen’s free will  
>>>> theorem (2006). If (a big IF, surely) we admit that our fellows  
>>>> can freely exercise their free will, it must be impossible to  
>>>> imagine that the atoms and molecules lack their share of the  
>>>> similar capacity. For our bodies eventually consist of those  
>>>> atoms and molecules.
>>>>
>>>>    Moreover, the exercise of free will on the part of the  
>>>> constituent atoms and molecules could come to implement the  
>>>> centripetality of Bob Ulanowicz at long last under the guise of  
>>>> chemical affinity unless the case would have to forcibly be  
>>>> dismissed.
>>>>
>>>>    This has been my second post this week.
>>>>
>>>>    Koichiro Matsuno
>>>>
>>>>
>>>>
>>>> From: Fis [mailto:fis-bounces at listas.unizar.es] On Behalf Of Alex  
>>>> Hankey
>>>> Sent: Thursday, October 19, 2017 6:42 AM
>>>> To: Arthur Wist <arthur.wist at gmail.com>; FIS Webinar <Fis at listas.unizar.es 
>>>> >
>>>> Subject: Re: [Fis] What is “Agent”?
>>>>
>>>> David Chalmers's analysis made it clear that if agents exist,  
>>>> then they are as fundamental to the universe as electrons or  
>>>> gravitational mass.
>>>>
>>>> Certain kinds of physiological structure support 'agents' - those  
>>>> emphasized by complexity biology. But the actual subject has to  
>>>> be non-reducible and fundamental to our universe.
>>>>
>>>> Alex
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Fis mailing list
>>>> Fis at listas.unizar.es
>>>> http://listas.unizar.es/cgi-bin/mailman/listinfo/fis
>>>>
>>>
>>>
>>>
>>>
>>> -- 
>>> Professor Terrence W. Deacon
>>> University of California, Berkeley
>>>
>>> _______________________________________________
>>> Fis mailing list
>>> Fis at listas.unizar.es
>>> http://listas.unizar.es/cgi-bin/mailman/listinfo/fis
>>>
>>
>>
>
>
>
>
> -- 
> Professor Terrence W. Deacon
> University of California, Berkeley
> _______________________________________________
> Fis mailing list
> Fis at listas.unizar.es
> http://listas.unizar.es/cgi-bin/mailman/listinfo/fis
>
> _______________________________________________
> Fis mailing list
> Fis at listas.unizar.es
> http://listas.unizar.es/cgi-bin/mailman/listinfo/fis

http://iridia.ulb.ac.be/~marchal/



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listas.unizar.es/pipermail/fis/attachments/20171023/4d01ae3c/attachment.html>

More information about the Fis mailing list