That's got my brain very engaged but I'm struggling to get my head around "getting the system to model itself". Which is good because I may be primed to learn something new... is this something that can be illustrated with an example?

Two examples serve - one is the micro narrative capture, and its self-signification in SenseMaker®

Continuous capture over large populations means that the system models itself into landscapes etc.

The other is SNS where instread of modeling or analysising a network (per SNA) you stimulate the ecology within a constraint framework to form networks which self declare

God I hope this is self-evident to people. I guess it is very important to go tell it on the mountain. It would be very regrettable if a systems thinking practitioner ever claimed their model was valid. The only valid model of a complex system is the system itself. It is also DIFFERENT, and, by the way, VALID for every observer/stakeholder. System thinking may aid and abet comprehension of complexity and complex systems, but never model it. I’m feeling a bit of nausea at the moment that some systems thinking dilettantes out there may believe their models are valid… Good grief, what monsters have we created?

But are partial or incomplete representations so universally dangerous, at the complex level? For sure, if used singly and as a stand-in for the referent (as in the worst social network analyses and computational models), they're a fool's toy.

But if, as a species, we are naturally good at abstracting patterns from experience and recombining them in new forms, then why could a complex approach not include the exposure of decision-makers to patterns generated by computational models? Especially those that might occur in low-probability, high-impact situations -- experience that is hard to acquire.

To put it another way, would adding to the variety of mental patterns available to a decision-maker be an intrinsically bad thing?

This morning I said to a local well know services designer: target groups should be able to define themselves. Anyone in "the system" should be free to opt-in/opt-out // to share experiences.

Aaaahhhhh, he said, if only I had more customers who accepted this. They tend to say "I publish to the 25-35 yr / single mum / 30k$ income group"

The issue of model usefulness is an interesting one. If the limitations are acknowledged then in some cases they provide useful insight into complex domains and assistance to decision making. However the claims normally exceed the capability and I am afraid that for far too many academics complexity now means modelling

Dave (and others),

I'm pleased you made that last observation as qualification Dave ("in some cases models are useful"). I believe we chatted briefly last time you were in NS about this although it was at the end of a long day and, as I recall, their were pints on the table at the time.

I've never accepted the view that such models as you discuss are unimportant. Nor am I (I don't think) naive about this, because since running into your work, I've thought a lot about this (and before for that matter).

Yes, people misuse such models as they misuse (I suspect) sensemaking. But this isn't a modeling problem, its the bounded applicability problem. And we (I) see it every day!

When people use models for their results (ie planning), they are almost always in the wrong. However, when people use models for learning, to capture current understanding of a system, to portray that understanding back to the group and then to build on that by uncovering mathematically implausible relationships in the current mental model(s), then such efforts can bring much discussion and understaning to the situation.

This of course has to do with many concepts that you rail about, but particularly the limited capacity of the brain to deal with anything more than bivariate relationships. This alone is a reason to engage in the knowledge modification that modeling (systems thinking) offers. If nothing else, in my experience, such activity helps people to conclude that they jusst don't know their system nearly as well as they thought they did. To me that's the logical connection into sensemaking.

I believe this idea of bounded applicability is incredibly important. When our only tool is a hammer, every problem looks like a nail.

BTW, note the number of responses to thsi post. Clearly you've touched on a soft spot with people!8:)

Cheers,

Ray

Bounded applicability is the key issue here Ray as you point out. THe problem when you have a dominant approach (and modeling in various forms dominates the systems period from which we are emerging) then its all to easy for advocates of that dominant approach to cross the line. Hence my hard push back!

I don;t agree on the idea that the brain can only handle bivariate relationships by the way

This relates exactly to our prior discussions regarding science, pseudo science, and others. I hope the following helps to illustrate my actual evolutionary/emergent position, rather than a creationist one.

A model to me is what emerges when an interpretant expresses an understanding of experienced phenomena in terms of representamen, which can be in symbolic, indexic, or iconic form. In this way, a model INCLUDES interpretive experience... moreso it is a response to that experience.

Problems arise at the intersection of assumed certainties and accepted uncertainties in regards to how complete and consistent the model is... problems which are subjective, interpretive, and objective in nature.

Godel's Theorem offers evidence that no formal system (a model) can be both complete and consistent... something which Stephen Hawking eventually noted on his quest for a "Theory of Everything".

Part of the problem I think people are falling into regarding a debate between Cynefin or Systems Thinking, is yes the notion of bounded applicability, but not in terms of recognition of the nature of the system, but recognizing the nature of the mind, the interpretant.

I believe this is clearly evident, but not quite understood, in your advocacy of self-declaration of networks... or allowing a landscape to emerge.

Your use of language seems to negating insights from Semiotics and Symbolic Interactionism, which would posit some common sense that humans act towards things based upon the meanings they ascribe to those things... which is related to a comment I read of yours once relating that human acquisition of knowledge builds upon prior knowledge.

In this sense, any landscape, or self-declared network that emerges, has done just that... emerged. Emerged from very specific initial conditions, which includes your agitation of the original system... be it through SenseMaker or whatever methods.

So the landscape or self-declared network actually becomes a RESPONSE to your agitation... not necessarily a "revelation" or self-declaration. SenseMaker presents signs, interpretants react to the symbols with other signs... humans act towards signs based upon the meaning they ascribe to those signs. So any landscape which emerges has more to do with the "expressed understanding" rather than the "understanding expressed" in regards to representamen that stands for phenomena to certain people.

How does assumed certainties and accepted uncertainties fit in?

When things emerge, humans have a tendency to recognize their supervenient power, rather than recognizing the fact that emergence can continue in the present. So we choose to stop... recognize something of the past, contemplate its permanence into the present and future, and then direct our actions in the present accordingly... leading to self-fulfilled "order". This is what allows us to use the same language as if it exists separately from our usage, and without us having to clarify each and every word.

Inconsistencies and incompleteness become evident or non-evident (pattern entrainment) over time as a model either reflects or does not reflect true nature.

No model or science or theory can account for the emergence possible in the present.

We differ maybe in the notion that complexity is found in the mind, not in the system. Chaos is always the nature of the system, in regards to temporal matters of emergence and supervenience.

Mark we have been through this before. I don't have the same regard for semiotics and symbolic interactionism that you have. I also feel that it is critical to realise that complexity thinking is ontologically distinct from systems thinking.

- I am not suggesting you have the same regard for Semiotics and Symbolic Interactionism... nor is my perspective only from this end. There is a simple realization though, that when I present you with a word, you will react in some way, a way which regards your experience with that word. I take this further though... to include observations people have of events, behavior, patterns, sounds, tastes, etc... So your response cannot be necessarily considered an objective thing, our interaction is in many ways the formation of a new complex adaptive system.

So the practical value I see in this is that while we can use RECOGNIZABLY VALUABLE and NOVEL methods like SenseMaker to attempt to better understand specific CAS, we are actually creating a new CAS in the process... which has implications.

You have alluded many times to things which Semiotics (simply the study of sign systems) addresses, such as the use of the word "complex" in different ways, the name "Cynefin" being used, or valuing the quote of even the explicit needing to be tacitly understood... and Aiden Choles very interesting blog post regardings people in South Africa having a hard time accepting the language of complexity being used.

- I certainly agree that Complexity Thinking is ontologically different from Systems Thinking... but my point is that this precise thing is what illustrates the nature of the Mind being a source of chaos, complexity, complication, and order. So while a group of people may be investigating the nature of a system as if they could discover a "true" nature and then educate everyone on that nature, that it will solve the problem... I am suggesting that the problem may actually stem from the differing nature of the people involved, rather than the system.

A reaction to a word depends on context to a degree; within specific contexts words can have precise meaning. In other cases experience, myth etc. will produce variable response. I don't think any of that implies we are creating a new CAS in the process. We are a part of that CAS but we can have degrees of separation and therefore degrees of coherence in any conclusions we may make. I think coherence is a better concept than objectivity here.

While I find Mead and Pierce interesting (and their latter day interpreters such as Ralph and Hari even more so) I don't feel the need to accept the concept of "sign" in the sense that I suspect you are using it. Aiden makes a good point, and language is emergent in its nature and inherits from its use etc. etc. again I can accept that without having to accept semiotics per se.

In respect of your final paragraph I am I am afraid in total disagreement. A system can exist and have a nature independently of the differing nature of its human observers.

I completely agree regarding context, but I wouldn't agree that words can ever have "precise" meaning. The fallacy there would be thinking that the word projects meaning onto the person, implying that the word held some intrinsic meaning and that you could present it to a person who never saw it before and retain its meaning.

Regarding seeing things as signs, it is very useful and valuable I believe, you could even replace it with organization or pattern or such. There are degrees that things represent, iconic, symbolic, indexic. And a CAS is made up of non-linear reactions to stimuli, be they chemical, biological, interpretive, physical, etc...

I agree on the word coherence, and there being degrees rather than actual separate CAS, but my point is simply realizing that by acting on the system / within the system, your activity and it's response are now inherent in any "explanation" or "landscape" that emerges be it from an analysts, or from self-declaration. That is why I used the language, the landscape that emerges, self-declares as you say, is a "response" to whatever agitation of the system... in the case of SenseMaker, it is presenting story prompts and signifier sets.

Your right that we can talk about language without accepting semiotics, but semiotics being the study of signs offers a lot of good literature, like the ones you mention.

Ultimately I think SenseMaker or other Cognitive Edge methods are concerned greatly with how people deal with meaning separately and together in regards to a shared subjective and objective context.

In regards to my last paragraph, I dont suggest that a system cannot have a nature independent of human observers, I am suggesting that humans cannot escape the means of language on their quest of sharing ideas regarding that nature, as well as the ramifications of such.

The problem with complexity, complication, and order in regards to linguistic context is that we tend to contemplate the permanence and independence of the imputed relation between a representamen and an object... which leads people to think that a word actually means something, rather than realizing that it is we who ascribe meaning to the word.

So while a system may have an objective nature independent from human observers, human observers will undoubtedly spend their whole lives debating symbolic concepts as if they were real. And even more so if we are talking about a company context (a communicative symbolic construct persisted over time and contemplated as existing in the present independent of our persisting the concepts in the present) rather than a physical system like the speed of light, then the ramifications are even greater.

The words we use to describe systems all too often become more real than the system itself... being able to get back to meaninglessness (the edge of chaos) lets us realize this... but the cycle will repeat itself from there, all the way back to order. This is why nothing could be more catastrophic than a group of people all agreeing on the completeness and consistence of a model of a system.

Mark, I understand the position, and with variations it is held by many people working in complexity theory and elsewhere. I think its true in part but its not universal and as a universal its a dangerous position in that it leads to relativism/solipsism if taken to its logical conclusion.

We can agree that it wrong (and sometimes catastrophic) to have a group of people all agreeing on the completeness and consistence of a model, but we do not agree on the reason why, or the limits of that wrongness.

Hi Dave please I'm attempting to understand your Complex Adaptive Systems model in light of CritLit2010 and connected learning. Is Connectivism an example of what you describe above? Could you call Connectivism a coherent system that models itself?

I dont know what a fuzzy system is please, does that imply that it may or may not have a purpose? (Slide -The Nature of Systems)

Would the agents or nodes in Connectivist courses be self signifying through their blog narratives and are these blogs, microblogs and bookmarks called attractors?

I recognise that in Connectivist courses as in CAS the course system lightly constrains the agents and that the agents modify in their turn both the system and each other (peer 2 peer learning). Also the idea of operating beyond equilibrium, the decentralised nature of course and of the Internet itself in my experience assures an on the edge of chaos experience both time management or resource filtering.
Distributed cognition, meaning emerges through interaction, disintermediation etc...

I think connectivism is a technology based perspective on learning, and the various situations it describes will be in the main complex adaptive systems. From what I have seen if considered as a learning theory it can't escape social constructivism and that way (from my perspective) lies relativism and solipsism (George for certain and Stephen probably) know that is my position.

From what I have seen of the connectivist courses they attempt to create a CAS type environment for learning and I commend them for that. I personally don't like the edge of chaos phrase, better to talk about far from equilibrium; ordered human systems are perpetually on the edge of chaos so the phrase covers more than CAS.

I define a system as any network which has coherence and it may be fizzy, that means the boundaries may not be clear and at times it may not be clear if you are in or out of the system itself (think about walking on a plateau in a mist and you will get a sense of what I mean).

Self-signification has a specific meaning in Cognitive Edge and is linked to SenseMaker®. It is a half way house between (i) tagging blogs (your reference) which is chaotic to my mind as you can't use vocabulary consistently and is also dependent on semantic tools for aggregation and that involves too great a loss of meaning, and (ii) the use of taxonomies which are too ordered. We take an approach half way between (which is proprietary). There is a lot more detail here if you are interested.

Hope that helps

Thanks so much Dave for informative clarity. I will follow up your link right now!

I'm flat out just learning new language let alone its context and application, and its huge fun.

Hi Dave the above link you gave is an error I tried contacting the pbworks people but it appears only you can correct it, I know you're busy.

I think it's this link for others that are interested?

http://learningtobeprofessional.pbworks.com/From-induction-to-abduction%2C-a-new-approach-to-research-and-productive-inquiry

Thanks Ruth - it is the correct link and I have now changed it