Talk:Self-organization/Archive 1

From Wikipedia, the free encyclopedia

Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Contents

Bibliography

Linking bibliographies

To 129.26.132.3 and Lexor:

I agree that there should be no links inside the titles. The proper way to do is to link the entire title; the linked pages are supposed to be book reviews. It might be appropriate to link the authors, though. -- Miguel

Links to authors are OK (I have already done this myself in the biblio), but keywords in titles are already listed in the article text and especially in the "See also" sections. Linking to the book title is appropriate if the review exists (or will likely be created at some stage), especially for important works (e.g. Darwin's Origin of Species, or in this context, perhaps Origins of Order by Kauffman or Design for a Brain by Ashby). I don't think every book title should be wikilinked as a matter of course, because it's unlikely that every title will have have a review written for it. -- Lexor 01:50, 26 Aug 2003 (UTC)
On the other hand, empty links encourage people to fill in the content. Someone who has read Origins of Order and sees the empty link may feel inclined to write a page about the book, and the idea might not have crossed their mind otherwise. -- Miguel
True, but not for every book in the bibliography. It's unlikely that there will ever be descriptions for every book in the biblio list (especially for articles with long biblio lists). I agree for "important" books (which is suitably vague, I know). I think links should be provided for books which we would most likely want a description, or think we might plan (or suggest to somebody else) to write later, but I don't create links for every single book title, I think it gets distracting, not to mention the problems with disambiguation for common titles. I think that there should be some editorial discretion as to which book titles we want to have pages for. I agree that Origins of Order would be a good one. There, I just created a wikilink for it... ;-) -- Lexor 01:15, 27 Aug 2003 (UTC)

Old vs. new books

Is there any justification for the division between "older" and "newer" technical books? If a single list was sorted chronologically, the reader of the article could decide for him/herself. On growth and form is a timeless classic, and the book by Eigen and Schuster is to a large extent still current because 30 year is not a long time for a revolutionary contribution. -- Miguel

The division was somewhat taken from the source I used. Feel free to merge them and sort alphabetically. Nice work on the article, by the way! It's looking pretty substantial now. -- Lexor 21:00, 16 Aug 2003 (UTC)

Copyright issues

I rewrote stub from scratch as the existing paragraph was take wholesale from the external link cited.

To the original submitter: please don't copy text from external sites without having obtained permission to release under the GFDL, see Wikipedia:Copyrights. -- Lexor 09:19, 3 Aug 2003 (UTC)

I have obtained permission to adapt some Notebooks under the GFDL from Cosma Shalizi. -- Lexor 09:13, 6 Aug 2003 (UTC)

Subject: wikipedia self-organization article
To: Cosma Shalizi
Date: 05 Aug 2003 06:57:08 -0700

I remembered reading your notebooks about self-organization 
(http://www.cscs.umich.edu/~crshalizi/notebooks/self-organization.html)
which has good snappy definitions and references and biblio entries,
and I was wondering whether you'd be willing to let me adapt it and
contribute it to the project and of course you are cited for your
contribution.

You could edit them again yourself, if you wished, even anonymously.
Basically you retain complete copyright (see:
http://wikipedia.org/wiki/Wikipedia:Copyrights), but license it to
wikipedia under the GNU Free Documentation License (GFDL:
http://www.gnu.org/licenses/fdl.html), which is like the GNU GPL, but
for documentation, which you're probably familar with, so it has the
same "copyleft" nature.

From: Cosma Shalizi
Subject: Re: wikipedia self-organization article
Date: Tue, 5 Aug 2003 16:02:29 -0400

Which brings me to the Wikipedia: by all means, adapt away!

Wikis as an example of self-organisation

Does anyone else see a wikiwiki as an example of a self-organising system? I gues it would make a great example on this page. --Anthony Liekens 22:30, 14 Sep 2004 (UTC)

No, a wiki-system is (mostly) not self-organizing. A system that consists of a wiki-system plus its users is a self-organizing system, where a significant fraction of the system (e.g., the wiki contents) is organized by elements of the system. -- Peter, 24 Jan 2005

However, a wiki is a great example of stigmergy. Should we cross-link "self-organization" with "stigmergy" and with other forms of organization? --DavidCary 23:29, 11 Feb 2005 (UTC)

Wikis may not be Self-organizing systems, but Wikipeida is one! I think that using Wikipedia as an example of self-organising organisation could make a very good example for this page.--GarOgar--

I think the concept of wiki (software + contents + participating humans) is a self-organising system, because it increases its complexity and internal order without external intervention (the system is considered as a whole). Still, can you find any emergent property? Ste nohype 16:25, 14 September 2007 (UTC)

Featured article?

With just a little more work this could be a good Featured Article candidate. --Erauch 04:16, 5 Feb 2005 (UTC)

File a request for peer review. — Miguel 06:27, 2005 May 20 (UTC)

Conflated?

Might the word conflated be out of place in this article? There is a Wiktionary entry for it, but nevertheless my more conservative nature must wonder if perhaps its usage here, at least without reference to definition, might not overstep a bit.

Could a better word be chosen?

Could the word be a link to http://en.wiktionary.org/wiki/Conflation ?

Could there be one of those odd little parenthetical numbers with a question mark or whatever it is after the word that off-site-links to the definition on wikitionary?

And why the fuck is Wiktionary missing so many words. Seems like that's the kind of fundraiser that would be easy to accomplish.

A Prominent announcement somewhere on Wikipedia that Some Month/Week will be "Fill in the Holes Week" on Wiktionary, or some such. Post a "Missing Word List" and just get 'er done. Or at least get her leaped forward.

Anyhoo,, I'd humbly suggest that conflated, while clearly defendable, is a bit too masturbatory for this article.

-:)Ozzyslovechild 03:36, 20 May 2005 (UTC)

Examples of self-organization without emergence and vice versa

Sometimes the notion of self-organization is conflated with that of the related concept of emergence. Properly defined, however, there may be instances of self-organization without emergence and emergence without self-organization, and it is clear from the literature that the phenomena are not the same. The link between emergence and self-organization remains an active research question.

Could someone please provide either an example of instances of self-organization without emergence and emergence without self-organization or a specific citation to literature that the phenomena are not the same?

The Self-Organizing Systems FAQ for Usenet newsgroup comp.theory.self-org-sys makes no mention of such a distinction. -- Nick 15:41, 15 Jun 2005 (UTC)

Please have a look at "T. De Wolf, and T. Holvoet, Emergence Versus Self-Organisation: Different Concepts but Promising When Combined, Engineering Self Organising Systems: Methodologies and Applications, Lecture Notes in Computer Science, 2005, Volume 3464, May 2005, pages 1 - 15" which can be found at http://www.cs.kuleuven.be/~tomdw/publications/

Indeed the description as given is vague and unhelpful; it sounds like someone has just (mis)copied from somewhere. If "properly defined" — what is the proper definition? Who has given it? There certainly seems to be little consensus among scientists, and usage tends to differ between disciplines.

Emergence is certainly possible without self-organization. To take an example, everyone accepts Conway's Game of Life as an example of emergent complexity. But it is not self-organized: change the rules even slightly and you no longer get the same collective behaviour (see, e.g. Bak, How Nature Works). Ditto at the critical point of a phase transition you get emergent scale-free structures (e.g. fractal cluster shapes in percolation) but this is not self-organized because it relies on very precise tuning of a control parameter. I suspect that the Wolfram CA rule pictured in the article here is also an example of emergence but not self-organization, since I doubt very much it is robust to alteration of variable parameters.

By contrast is self-organization possible without emergence? I think that's dubious although I'm open to alternative suggestions on that. IMO self-organization is probably best thought of as emergence that is robust with respect to variable parameters in the dynamics.

The De Wolf and Holvoet paper seems to me to be quite a confused discussion of the issues and I don't think its propositions can or should be just taken and put in the article without careful discussion. I don't have time right now but I will try to post a detailed analysis of its arguments on this talk page sometime soon. WebDrake 23:24, 7 October 2005 (UTC)

The main idea of the De Wolf and Holvoet paper is as follows: 'Emergence' denotes the fact that there are global or macroscopic properties that emerge solely from local interactions between the parts without the parts having any notion of the global property. 'Self-Organisation' on the other hand denotes the fact that there is no external control and that the system autonomously achieves its required behaviour, it organises by itself. As such Emergence can exist without self-organisation if a system is considered where multiple interacting parts in combination with external influences are needed to achieve the global behaviour. It all depends on where you define the boundary of the system. For example, if you take the temperature evolution of a heated metal plate. If you consider the heating source as external to the system then this is an example of emergence without self-organisation because the external heating control is needed to achieve the temperature evolution. Also, Self-organisation can exist without emergence if we have a system that organises itself but the global behaviour is known as a plan in each part. For example, in computer science, consider a software system consisting out of multiple autonomous entities which have to organise themselves as a group in a certain shape on a 2D plane. If such a system is engineered by supplying each entity with a global map of the wanted shape then this is not an emergence behaviour but still the system organises itself. In addition to this difference between emergence and self-organisation, the paper then argues that often a system exhibits both phenomena in combination which could explain why a lot of confusion exists in defining them. --Tomdw 16:19, 14 November 2005 (UTC)

Sorry, Tom, but I think the second point about self-organization without emergence is nonsense. How can it be "self-" organized if each component is given a map telling it what to do? That's instruction, not self-organization. WebDrake 15:42, 14 December 2005 (UTC)
It is not instruction because the map is just information each component can use to still autonomously achieve the required behaviour. It is not a control plan that states the full sequence of actions for each component. It jus states what all the components as a whole have to achieve, a global goal. As such the system is still "self-" organized because there is no external control at runtime but there is no emergence because each component is aware of the required global end result. --Tomdw 15:18, 16 December 2005 (UTC)
OK, I'll buy that. I think it's an interesting distinction and I'm not surprised that it's engineers, rather than, say, physicists or mathematicians, who came up with it. Still, while these ideas should certainly be discussed, I'm uncomfortable with the idea of putting them across in a way that suggests they are widely accepted. Definitions of "self-organization" and "emergence" are still too vaguely agreed on for that. WebDrake 10:21, 19 December 2005 (UTC)


It is true that there is no widely accepted definition for both concepts and maybe there never will be. However, this distinction is certainly useful and becomes more accepted from an engineering point of view. I of course cannot speak for physicists or others. --Tomdw 14:08, 2 January 2006 (UTC)

Self-organized criticality and related pages

In case anyone watching Self-organization has missed it, note that after Kku's clever separation of Self-organized criticality into its own article, WebDrake has rewritten and dramatically expanded that page into a very good description of SOC. He's also made some stub examples of SOC which people might like to contribute to, and he's thinking of doing further work on Power law and Self-organization itself over the next month or two. It's worth having a look at the other pages involved. Maybe people have thoughts on possible further improvements. -- JimR 06:07, 16 October 2005 (UTC)

Categories

I've added to and re-ordered the categories of this page. The logic is as follows:

  • Cybernetics, though historically important, isn't thought of much in terms of modern concepts of self-org, so I sent it to the back of the list.
  • Dynamical systems and systems theory come first, though I think maybe the latter should be relegated to the back along with cybernetics, since it seems more historical than present.
  • Then applied & interdisciplinary physics, stat mech and supramolecular chemistry.

All good? — WebDrake 19:52, 17 October 2005 (UTC)

Hi Drake. I created Category:Self-organization, because there quite a few pages that can fit in there, I think. Karol 07:53, 18 October 2005 (UTC)

Definition of self-organization/sources used

A good part of the article, particularly as regards definition of self-organization, distinction from emergence, etc., seems to be taken pretty directly from the Cozma Shalizi notebooks.

I don't object to CS's ideas on their own terms but I think his notebooks are a personal viewpoint and do not represent a broad enough view of what "self-organization" has come to mean in different parts of the scientific community. For example as CS himself notes, the concept of "self-organization" used in self-organized criticality is different.

First, I think it is not clear that self-organization exclusively refers to systems where there is a non-directed increase in order or complexity. Talk of the "organization" of a system need not necessarily refer to the degree of order present but simply how it is arranged: e.g. "the molecular organization is completely random". CS's PhD thesis suggests that the laws of thermodynamics are an emergent phenomenon of molecular interaction but that it is not self-organizing. In the instinctive sense of "self-organization", as opposed to CS's definition, this is bizarre. If the behaviour of the molecules, and the collective result, is not self-organized, what on earth is it? It is not being directed by any outside entity, so it must be arranging (=organizing) itself. Even if we were to throw out the term "organizing", the system is self-something, and I suggest it is the self part of the definition that is more important since it is the more general. (Self-organization in the sense I have outlined here clearly includes self-organization as the article currently defines it, but is broader in its scope.)

Secondly, both CS's work and this page cite certain systems as "self-organizing" where I think that claim is dubious, the most obvious example being certain cellular automata. Certainly these systems may display an increase in structural order or complexity, but I would argue that insufficient attention is paid to the self- part. For example, many CA displaying emergent complexity rely on a hefty input of effort from the computer in order to maintain the integrity of their rules, information transmission, etc., via the use of error-correcting codes and so on. The behaviour of many CA relies on the rules being "perfect" all the time, not subject to noise, etc. For example the Game of Life only "works" with the given update rule: change it even slightly and the system will no longer behave the same. It may be "organizing" according to CS's definition, but it is not self-organizing because the parameters of its dynamics must be so specific.

I don't feel it would be appropriate to just charge on in to the article with these objections so I'd like there to be a debate here before any changes are made. Ideas, anyone?

— WebDrake 20:20, 17 October 2005 (UTC)

Self-organization & emergence

I wonder is the articles Self-organization and emergence should be related more tightly? Also, should the categories Category:Self-organization and Category:Emergence (just created it) be directly related somehow? (the difference is not very clear to me) Karol 09:40, 13 November 2005 (UTC)

Self-organization vs. entropy

The second law of thermodynamics as defined on its page only affects the entropy of an isolated system, whereas a closed system can exchange heat and work, but not matter, with its surroundings. The statement in this section that "However, a closed system can gain macroscopic order while increasing its overall entropy. Specifically, a few of the system's macroscopic degrees of freedom can become more ordered at the expense of microscopic disorder." therefore seems odd, as it can gain entropy from its surroundings. Is this what's meant? ..dave souza, talk 21:36, 20 May 2006 (UTC)

Having come closer to clarifying this issue at Evolution#Self-organization and entropy I've added the statistical thermodynamics connection here, and have changed the "closed system" reference above to avoid a contradiction, hope someone can clarify this point. ...dave souza, talk 09:05, 24 May 2006 (UTC)