Cognition – Magrathea TLC

Authenticity

Some define innovation as a lack of self-restraint. But one has to admit that lingering in a painting of what once was (Mark Fisher) generates stagnation and that is not the way of the world. El Bulli’s celebrated chef Ferran Adria instead defines tom innovate as to not copy. Montroll (1982) introduces MAXENT as a method for predicting human behaviour, because people tend to employ every position of the applicable Gauss distribution.

You have to agree that it is hard to do something that hasn’t been done before and that is why some avoid it. Seth Godin writes in The Practice that you can choose the pattern to live by. You may opt to follow an industrial one that guides your everyday activities by their expected outcomes. However, this offers different activities compared to those that establish and reinforce your identity from your own perspective (your self), and others’.

Your identity is the resultant of continual interactions with what you encounter in life. Following the industrial pattern, Godin continues, your identity instead gets to be made up of outside elements not of your choice. This affects the way you perceive and the way you are perceived. It generates a difference of perspectives on your identity and leads to friction in interactions. And, Godin argues, your products contribute to the world.

According to Godin in order to contribute to your self and to the world you must innovate by making something new in a creative process. Activities that follow an industrial pattern tend to point in the direction of desired (final, systemic) outcomes. The pattern inspires confidence in the outcomes because everything in its organisational realm is directed there according to a plan.

Or you follow a pattern that develops your identity in both of the senses. Here the approach is to do stuff and then find out what we like, instead of defining what we like and then acting in accordance. Inspiration has no practical meaning to the professional, but the need to act continually has: love what you do replaces do what you love. By following what you (turn out to) like while doing it, you find the pattern that contributes most to our identity and to the world.

Start today doing stuff without a preset plan in order to find out what you and others love about the process. In this way you develop an individual pattern that contributes to your identity guiding your future actions. The awareness that, although innovative, it may not lead to anything requires generosity without aiming for a final outcome, but for a beautiful process.

This requires the faith that this approach will add value to our selves and to the world. It can not provide confidence that we act so as to benefit the performance of an industrial system. This faith originates from the conviction that that radical pursuit of creativity lasts because it corresponds with your self and with the world.

You are more authentic if regardless of outside influence you determine your identity through actions that accord with what you favour. Heidegger qualifies the identity as authentic that results from disallowing external influence to shape the mind without conscious decision. Outside ideas as a consequence rarely remain unscathed leaving a trace on your mind. And in that way you contribute to innovation.

The Order of Time

Carlo Rovelli, Penguin Books, ISBN 978-0-141-98496-4

In the universe there is the possibility of time and space. Masses of bodies modify the structure of the possible space and time between them: closer to a mass there is less time. dpb: if this concerns how much time passes, then does this mean that the grain of the structure of time is finer closer to a mass than further removed? And does that mean that closer to a mass more steps are needed compared to farther away? Time is different at every locus: it is relative and it has no unity.

In the laws of physics there is no inherent difference between past and future. Why is the past so different from the future to us? Based on Carnots proposition, Clausius posits that, if everything else remains equal, heat cannot pass from a cold body to a hot one. Rovelli writes: ‘This is the only basic law of physics that distinguishes the past from the future‘. None of the others do so’ Not Newton, Maxwell, Einstein, Heisenberg, Dirac, Schrodinger or elementary particle scientists. But dpb: is this indeed the only asymmetry known to physics? Or the only time related one?

The link between time and heat is fundamental: ‘In every sequence of events that becomes absurd if projected backwards, there is something that is heating up‘ (p 23). This means it is the only irreversible one, all the others are reversible. dpb: Is it inconceivable that irreversible change occurs without something heating up? Chaos theory teaches that deterministic systems can produce irreversible randomness. If such a system reverses its behavior at some point it is likely to deviate from its trajectory back to where the observation started in unpredictable ways, and hence to be irreversible (disorder from order). Another source of irreversibility is complexity, especially life, because what higher level of organization emerges will resist reversal to a previous, less organized state (order from disorder). Take e.g an autopoietic system seeking to maintain its operational closure and its present organization (configuration). It will last longer than expected (Schrodinger).

Clausius formulated the 2nd law, namely dS=>0, implying that heat never passes from cold to hot bodies. But e.g. Popper (1965) argues that heat gets passed from colder to warmer bodies all the time and that entropy is not homogeneous over different locations and is a stochastic parameter. Boltzmann: entropy production is growing disorder into less particular less special situations. dpb: to what is the (decreasing) order attributed when entropy increases? Or in other words: what are these situations and what are their changes? For example: the emergence of an organism increases order of the system in focus but also increases entropy. The point is that entropy by definition defines and scopes the system, because it is a statistical notion extending the system until the energy balance is nil.

Take sunlight for example, a source of low entropic energy, cast on a stone. The atoms that compose the stone are agitated by the high-grade energy of the sunlight. Caused by the agitation the atoms heat up and the temperature of the entire stone rises. The amount of energy carried by the stone increases with what was carried by the sunlight. However, that energy dissipates and becomes less special, aka less capable of work: entropy is produced. At night, when the air cools, the heat carried by the stone (or rather by its agitated atoms) is transferred to the atoms of the air surrounding it. They become agitated and the energy of the atoms in the stone is transferred to the atoms in the air. Entropy is produced and the energy carried in the agitated air atoms is capable of still less work, becoming even less special. From an entropic perspective the system includes the stone, the sun and the air molecules.

dpb: Take the example when the atoms composing a cat are agitated when sunlight is cast on the cat. The cat is a self-referencing (autopoietic) system which is organized such that its body temperature is kept in bounds such that it keeps functioning. The temperature of the entire cat rises, until the cat reacts by taking physiological measures to cool its body sweating through its soles. The molecules in the surrounding soil and air are agitated, heating it. Entropy is produced and the same amount of energy as initially input by the sunlight into the atoms of the cat is capable of ever less work, entering a less special state. The system is constituted of the cat, the sun, the soil and the air molecules. The level of organization increased by the participation of the cat, but the entropy production is roughly the same.

dpb: The state of a system becomes less special when entropy is produced: the system becomes less ordered, aka less organized, and the ignorance of the observer regarding the likelihood of future states increases. The stone and the cat are defined by us as we observe the processes and their relations taking place within their physical boundaries. But they are determined by their organization as a stone and a cat respectively.

dpb: Thus, entropy is produced regardless of the level of organization of the discussed systems. While they produce entropy their organizations don’t change: the organization of the stone and the cat remain the same. What does it mean to say that the level of organization decreases when entropy is produced but the organization of the stone and the cat remain the same? The incoming sun(light) and the atoms composing the air are external to the systems in our focus, but not regarding the production of entropy.

However, the amount of uncertainty regarding the future states – in phase space – of these systems increases. This explains the production of entropy in an existing system. If an organism emerges – appears for the first time – the order of what are (post-hoc) its components increases into their organization as the observed system, but the entropy increases nonetheless. The appearance of organisms does not reverse entropy production!

Increasing (dis)order is on account of the observer. When entropy is maximum all the possible next states are equally special, it is observed, and when minimum, say say a crystal at 0 K, the next state is the same as the current one. The notion of particularity only comes to be if we see things in a blurred and approximate way. dpb: the states become less particular from the perspective of someone taking a blurred view. Does blurred mean that e.g. a person cannot distinguish atoms and the relations between them, or the different kinds of fields such as energy and material? Can there also be a relation between blurring and emergence, because observing an emerged system we are incapable to reduce its behavior at the larger scale to that of its components at the smaller scale? And if so, is this knowable, or in other words: can an observer keep the behavior of the components of any system in focus until the point where they become part of the unity they have emerged into? And is the observer in this case aware of the emerged unity? And if so then is the observer one observer or has he become two observers?

According to Rovelli, blurring means that we are incapable of seeing the microscopic level where there is no difference between past and present, and between cause and effect. An observer is required for that! There is a loss of direction, because there is no intrinsic difference between past and future on the smallest scales.

Time passes more slowly for what moves. dpb: I understand this effect in the same way as (potential) time slows down closer to a mass. The dimensions of the squares (graining) of the grid onto which time is canvassed become smaller on the topology closer to a mass and at high relative speed.

There is no single or absolute time and every point in space has a proper time. ‘Now’ is meaningless, because there is no present anywhere that corresponds to any other. The question of a present in/of the universe is therefore improper. The temporal structure of the universe can be said to be made of partially ordered cones. Each approximate generation roughly followed by another. The structure of space-time is not stratified but scattered, and without common direction (wobbly). There is no universal present.

According to Aristotle time is the measure of change, including thinking of change. dpb: this implies that change is a series of events at which configurations transpose to other configurations, whereby whereby these configurations generate corresponding behavior, and events are observations of these transpositions. According to Newton even if nothing happens, time (of some kind) still passes. Leibniz defends the former idea that time results from counting of events, but the latter has caught on and is the more widely accepted. Until the end of the 19th century every place had its proper time, only later local time was replaced with global time schedules. This interpretation is closer to the notion of relativity of time than the wide spread absolute Newtonian interpretation.

dpb: events are grained or quanted into (or: on) the space-time grid of potential space and time, whereby they occupy space and time. When closer to a mass or moving, the time grid particular to it is finer grained and the thing is therefore observed to move slower. In this view there is no empty space or time, aka space and time only come to be when occupied.

Aristotle maintains that the position of a thing in space and time is identified by what surrounds it. There is no ‘empty’ space or time. dpb: this connects with the idea of a rhizome spanning up its own dimensions plus the ones required to describe it by the observer, but not its behavior at the ‘emerged’ scale (Deleuze and Guattari 1987).

What does it mean when space and time can be void i.e nothing is there as Newton suggests? Not even electromagnetic fields or other things imperceptible to people. Einstein suggest that time and space are like canvasses on which reality (substances) comes to exist as fields. Gravity, one of these fields, through mass determines the topology of space-time a gravitational field. dpb: Time and space are not absolute but relative, suggesting they interpenetratingly depend on neighboring elements. Does it mean that space and time are potential until something occupies them? Do they substantiate from a potential (virtual) to the real under the conditions of the actual, or in other words that they become?

Moving on to consider quantum properties of space and of time, then in addition to situationalized, individualized, localized and independent as previously described, time is also distributed. This leads to the aspects of granularity, indeterminacy, and relativity which further break down our understanding of time. dpb: the entropic direction of time seems problematic, because time would take the direction of every whole system ‘scoped’ by the entropy production, covering many different loci in space time. But every point in space time has develops its particular space and time.

Quantum effects occur at Planck time scales of 10E-44. On smaller scales there is no time (it is meaningless). Also time assumes only special discrete values: most times do not exist. Hence time is not continuous but granular, leapfrogging from one instance to another separated by intermittencies at the Planck scale.

Time is indeterminate: it is not absolute but in a superposition, its present state is unknown. Something may take place before and after something else does. What takes place is resolved at an event of interaction aka observation. The only point where it becomes concrete is in the relation with the ones it interacts with (or is observed by), instead of uniformly for the rest of the universe. In this view, the world is made up of events taking place in a disorderly way, not things. Events are short-lived and processual, and things persist and are object-like.

The notion of the event as the basic unity of reality fits well with how we experience the world, because it is spatially and temporally delimited. Objects in this view are long events. dpb: Can we say that objects are series of events that are long compared to the duration of human life, and perhaps to other things? In this sense time equals happening.

Thus, the idea of presentism, meaning that reality moves from one present to the next, no longer works. The world is not a single linear succession of presents. The opposite idea of eternalism (block universe), that past, present and future are equally real means that change is an illusion realistically can’t be true. The point is that the present develops in a disorderly distributed and indeterminate way. But our grammar isn’t suitable: ‘In the world there is change, there is a temporal structure of relations between events that is anything but illusory’ (p 100).

We do not need the time variable to describe the world, but the variables that describe it, and how they change relative to each other. If a number of these relations are established then we might say when an event takes place: ‘.. what the relations are between these variables’ (p 103). dpb: or in other words these are differences (relations) between two series of differences (variables), aka assemblages. The differences should have the adjective comparable to indicate the potential of correlations between correlations.

Fields become manifest in granular form. The grains are not immersed in space but they form it, the structure of space is granular. Spatiality is the web of interactions between these grains. They exist only during those interactions. The latter are the happening of the world, the development of reality.

‘Time emerges from a world without time, ..’ (p 117). dpb: will this be the link to blurring and approximation Rovelli mentions early on in the text as the insight of Boltzmann?

The conventional view is that time determines energy determines the macro state. An opposing view is that the macro state determines energy determines time, whereby the macro state equals the blurred view of systems and their behavior. Or in other words: time becomes determined because / as an effect of the blurring, namely the incompleteness of the description of the corresponding micro-states. dpb: Rovelli refers to this notion as thermal time, because it corresponds to the probabilistic notion of entropy at the macro scale. In processual terms the macro-state can mean the actual, namely the organization of a system restricting the behavior of its constituent elements (components). Possible names for time caused by blurring cased not by heat: emergence time, organization / order time, identity time?

Measurement of speed and position of a particle are non-commutative. The order of the interactions (measuring) matter causing a primitive form of time to develop. Time developing in changing macro states (thermal, others?) and time developing in changing quantum states (quantum) are very similar. Rovelli suggests that this is the time we know, whereby quantum indeterminacy and the large quantities of particles cause the blurring. Time is ignorance. dpb: Emergence takes place in physical systems at various scales and I hoped that Rovelli would express a stronger link to emergence &c.

dpb: so, again: what is blurring? We, and every physical system, interact with a limited number of other systems. We interact with them through only a limited number of variables (aka correlations, assemblages). The configurations we do not notice seem to be equivalent to us. Thus our – and other physical systems’ – vision of the world is blurred.

Thus, blurring is the pivot of the theory of Boltzmann: entropy is the number of configurations of a system we are not aware of. This can of course vary per observed system as well as per observing system, because what is not visible for one can be visible for another. Blurring is not only mental, because interactions with the micro-systems exist. Entropy is relative to the observer aka the one interacting with the system in hand, like speed a property of the object relative to others.

The entropy of the world depends on its configuration as well as on our (in general physical systems’) blurring of them. This depends on which variables of the world – our part of the world – we interact with.

Thus, the seemingly low entropy in the far past may be a result of the limited set of variables we interacted with and our blurring of them. Our interactions are limited to a small number of macroscopic variables. The microscopic configurations are blurred. Since we are responsible for that, with the supposed low entropy as a result. In this view the arrow of time is not universal but it depends on the physical system doing the viewing, us in this case, from our ‘special corner’ of the universe.

Indexicality refers to words which assume a different meaning depending on their use. dpb: like a variable? Indexing makes the perspective of its use explicit. It is not possible to say anything from outside the world. Existence takes place from within (so as the map is useful if we know our position versus it). Thus, time is not external to us (or any physical system) but we are situated within it, seeing time from the inside.

Any energy goes to thermal energy. Low entropy goes to high: production. Without entropy there is no past and no future. Entropy production is increasing disorder. But evolution of organisms increases order by way of organization while entropy is still produced. dpb: Again: What is the subject of entropy? The relevant one is the change of order in phase space (not real space). Even though there are temporary pockets of order in the grand scheme (e.g of evolution), the very grand scheme is of increasing disorder.

The past leaves traces in the future. CR suggests that this is because of low entropy into the past, because there is nothing else to separate the past from the future. dpb: blind spot. Common causes in the past of current phenomena point at low entropy in the past, because of the increasing improbability of their occurring together? ‘In a state of thermal equilibrium, or in a purely mechanical system, there isn’t a direction to time identified by causality‘ (p 146). dpb: but it isn’t just causality but irreversibility.

Time is related to us. What are we? We are not individual entities, but relational identities. What produces this:1) point of view 2) in connection with the working of our brain, we categorize vis-a-vis other humans (the reflection of us we get back from our kind) 3) memory links our histories through narratives.

The mind can see traces of the past left in the brain. But threading them as an interpretation of the flow of time is an internal process of the mind. It is integral to the mind. The mind constitutes (a representation of) time through the retention of traces on the brain of past events (Husserl).

‘Time opens up our limited access to the world. Time, then, is the form in which we beings whose brains are made up essentially of memory and foresight interact with the world: it is the source of our identity‘ (p 164).

Chemical Organization Theory and Autopoiesis

E-mail communication of Francis Heylighen on 29 May 2018:

Inspired by the notion of autopoiesis (“self-production”) that Maturana and Varela developed as a definition of life, I wanted to generalize the underlying idea of cyclic processes to other ill-understood phenomena, such as mind, consciousness, social systems and ecosystems. The difference between these phenomena and the living organisms analysed by Maturana and Varela is that the former don’t have a clear boundary or closure that gives them a stable identity. Yet, they still exhibit this mechanism of “self-production” in which the components of the system are transformed into other components in such a way that the main components are eventually reconstituted.

This mechanism is neatly formalized in COT’s notion of “self-maintenance” of a network of reactions. I am not going to repeat this here but refer to my paper cited below. Instead, I’ll give a very simple example of such a circular, self-reproducing process:

A -> B,

B -> C,

C -> A

The components A, B, C are here continuously broken down but then reconstituted, so that the system rebuilds itself, and thus maintains an invariant identity within a flux of endless change.

A slightly more complex example:

A + X -> B + U

B + Y -> C + V

C + Z -> A + W

Here A, B, and C need the resources (inputs, or “food”) X, Y and Z to be reconstituted, while producing the waste products U, V, and W. This is more typical of an actual organism that needs inputs and outputs while still being “operationally” closed in its network of processes.

In more complex processes, several components are being simultaneously consumed and produced, but so that the overall mixture of components remains relatively invariant. In this case, the concentration of the components can vary the one relative to the other, so that the system never really returns to the same state, only to a state that is qualitatively equivalent (having the same components but in different amounts).

One more generalization is to allow the state of the system to also vary qualitatively: some components may (temporarily) disappear, while others are newly added. In this case, we no longer have strict autopoiesis or [closure + self-maintenance], i.e. the criterion for being an “organization” in COT. However, we still have a form of continuity of the organization based on the circulation or recycling of the components.

An illustration would be the circulation of traffic in a city. Most vehicles move to different destinations within the city, but eventually come back to destinations they have visited before. However, occasionally vehicles leave the city that may or may not come back, while new vehicles enter the city that may or may not stay within. Thus, the distribution of individual vehicles in the city changes quantitatively and qualitatively while remaining relatively continuous, as most vehicle-position pairs are “recycled” or reconstituted eventually. This is what I call circulation.

Most generally, what circulates are not physical things but what I have earlier called challenges. Challenges are phenomena or situations that incite some action. This action transforms the situation into a different situation. Alternative names for such phenomena could be stimuli (phenomena that stimulate an action or process), activations (phenomena that are are active, i.e. ready to incite action) or selections (phenomena singled out as being important, valuable or meaningful enough to deserve further processing). The term “selections” is the one used by Luhmann in his autopoietic model of social systems as circulating communications.

I have previously analysed distributed intelligence (and more generally any process of self-organization or evolution) as the propagation of challenges: one challenge produces one or more other challenges, which in turn produce further challenges, and so on. Circulation is a special form of propagation in which the initial challenges are recurrently reactivated, i.e. where the propagation path is circular, coming back to its origins.

This to me seems a better model of society than Luhmann’s autopoietic social systems. The reason is that proper autopoiesis does not really allow the system to evolve, as it needs to exactly rebuild all its components, without producing any new ones. With circulating challenges, the main structure of society is continuously rebuilt, thus ensuring the continuity of its organization, however while allowing gradual changes in which old challenges (distinctions, norms, values…) dissipate and new ones are introduced.

Another application of circulating challenges are ecosystems. Different species and their products (such as CO2, water, organic material, minerals, etc.) are constantly recycled, as the one is consumed in order to produce the other, but most are eventually reconstituted. Yet, not everything is reproduced: some species may become extinct, while new species invade the ecosystem. Thus the ecosystem undergoes constant evolution, while being relatively stable and resilient against perturbations.

Perhaps the most interesting application of this concept of circulation is consciousness. The “hard problem” of consciousness asks why information processing in the brain does not just function automatically or unconsciously, the way we automatically pull back our hand from a hot surface, before we even have become conscious of the pain of burning. The “global workspace” theory of consciousness says that various subconscious stimuli enter the global workspace in the brain (a crossroad of neural connections in the prefrontal cortext), but that only a few are sufficiently amplified to win the competition for workspace domination. The winners are characterized by much stronger activation and their ability to be “broadcasted” to all brain modules (instead of remaining restricted to specialized modules functioning subconsciously). These brain modules can then each add their own specific interpretation to the “conscious” thought.

In my interpretation, reaching the level of activation necessary to “flood” the global workspace means that activation does not just propagate from neuron to neuron, but starts to circulate so that a large array of neurons in the workspace are constantly reactivated. This circulation keeps the signal alive long enough for the different specialized brain modules to process it, and add their own inferences to it. Normally, activation cannot stay in place, because of neuronal fatigue: an excited neuron must pass on its “action potential” to connected neurons, it cannot maintain activation. To maintain an activation pattern (representing a challenge) long enough so that it can be examined and processed by disparate modules that pattern must be stabilized by circulation.

But circulation, as noted, does not imply invariance or permanence, merely a relative stability or continuity that undergoes transformations by incoming stimuli or on-going processing. This seems to be the essence of consciousness: on the one hand, the content of our consciousness is constantly changing (the “stream of consciousness”), on the other hand that content must endure sufficiently long for specialized brain processes to consider and process it, putting part of it in episodic memory, evaluating part of it in terms of its importance, deciding to turn part of it into action, or dismissing or vetoing part of it as inappropriate.

This relative stability enables reflection, i.e. considering different options implied by the conscious content, and deciding which ones to follow up, and which ones to ignore. This ability to choose is the essence of “free will“. Subconscious processes, on the other hand, just flow automatically and linearly from beginning to end, so that there is no occasion to interrupt the flow and decide to go somewhere else. It is because the flow circulates and returns that the occasion is created to interrupt it after some aspects of that flow have been processed and found to be misdirected.

To make this idea of repetition with changes more concrete, I wish to present a kind of “delayed echo” technique used in music. One of the best implementation is Frippertronics, invented by avant-garde rock guitarist Robert Fripp (of King Crimson): https://en.wikipedia.org/wiki/Frippertronics

The basic implementation consist of an analogue magnetic tape on which the sounds produced by a musician are recorded. However, after having passed the recording head of the tape recorder, the tape continues moving until it is read by another head that reads and plays the recorded sound. Thus, the sound recorded at time t is played back at time t + T, where the interval T depends on the distance between the recording and playback heads. But while the recorded sound in played back, the recording head continues recording all the sound, played by either the musician(s) or the playback head, on the same tape. Thus, the sound propagates from musician to recording head, from where is is transported by tape to the playback head, from where it is propagated in the form of a sound wave back to the recording head, thus forming a feedback loop.

If T is short, the effect is like an echo, where the initial sound is repeated a number of times until it fades away (under the assumption that the playback is slightly less loud than the original sound). For a longer T, the repeated sound may not be immediately recognized as a copy of what was recorded before given that many other sounds have been produced in the meantime. What makes the technique interesting is that while the recorded sounds are repeated, the musician each time adds another layer of sound to the layers already on the recording. This allows the musician to build up a complex, multilayered, “symphonic” sound, where s/he is being accompanied by her/his previous performance. The resulting music is repetitive, but not strictly so, since each newly added sound creates a new element, and these elements accumulate so that they can steer the composition in a wholly different direction.

This “tape loop” can be seen as a simplified (linear or one-dimensional) version of what I called circulation, where the looping or recycling maintains a continuity, while the gradual fading of earlier recordings and the addition of new sounds creates an endlessly evolving “stream” of sound. My hypothesis is that consciousness corresponds to a similar circulation of neural activation, with the different brain modules playing the role of the musicians that add new input to the circulating signal. A differences is probably that the removal of outdated input does not just happen by slow “fading” but by active inhibition, given that the workspace can only sustain a certain amount of circulating activation, so that strong new input tends to suppress weaker existing signals. This and the complexity of circulating in several directions of a network may explain why conscious content appears much more dynamic than repetitive music.

Chemical Organization Theory as a Modeling Tool

Heylighen, F., Beigi, S. and Veloz, T. . Chemical Organization Theory as a modeling framework for self-organization, autopoiesis and resilience . Paper to be submitted based on working paper 2015-01.

Introduction

Complex systems consist of many interacting elements that self-organize: coherent patterns of organization or form emerge from their interactions. There is a need of theoretical understanding of self-organization and adaptation: our mathematical and conceptual tools are limited for the description of emergence and interaction. The reductionist approach analyzes a system into its constituent static parts and their variable properties; the state of the system is determined by the values of these variable properties and processes are transitions between states; the different possible states determine an a priori predefined state-space; only after introducing all these static elements and setting up a set of conditions for the state-space can we study the evolution of the system in that state-space. This approach makes it difficult to understand a system property such as emergent behavior. Process metaphysics and action ontology assume that reality is not constituted from things but from processes or actions; the difficulty is to represent these processes in a precise, simple, and concrete way. This paper aims to formalize these processes as reaction networks of chemical organization theory; here the reactions are the fundamental elements, the processes are primary; states take the second place as the changing of the ingredients as the processes go on; the molecules are not static objects but raw materials that are produced and consumed by the reactions. COT is a process ontology; it can describe processes in any sphere and hence in scientific discipline; ‘.. method to define and construct organizations, i.e. self-sustaining networks of interactions within a larger network of potential interactions. .. suited to describe self-organization, autopoiesis, individuation, sustainability, resilience, and the emergence of complex, adaptive systems out of simpler components’ [p 2]. DPB: this reminds me of the landscape of Jobs; all the relevant aspects are there. It is hoped that this approach helps to answer the question: How does a system self-organize; how are complex wholes constructed out of simpler elements?

Reaction Networks

A reaction network consists of resources and reactions. The resources are distinguishable phenomena in some shared space, a reaction vessel, called the medium. The reactions are elementary processes that create or destroy resources. RN = <R,M>, where RM is a reaction network, R is a reaction, M is a resource: M = {a,b,c,…} and R is a subset of P(M) x P(M), where P is the power set (set of all subsets) of M and each reaction transforms a subset Input of M into a subset Output of M; the resources in I are the reactants and the resources in O are the products; I and O are multisets meaning that resources can occur more than once. R:x1+x2+x3+..→y1+y2+… The + in the left term means a conjunction of necessary resources x: if all are simultaneously present in I(r) then the reaction takes place and produces the products y.

Reaction Networks vs. Traditional Networks

The system <M,R> forms a network because the resources in M are linked by the reactions in R transforming one resource into another. What is specific for COT is that a reaction represents the transform from a multiplicity of resources into another multiplicity of them: a set I transforms to a set O. DPB: this reminds me of category theory. My principal question at this point is whether the problem of where organization is produced is not relocated: first the question was how to tweak static object into self-organization, now it is which molecules in which quantities and combination to conjuncture to get them to produce other resources and showing patterns at it. In RN theory the transform of resources can occur through a disjunction or a conjunction: the disjunction is represented by the juxtaposed reaction formulae, the conjunction by the + within a reaction formula.

Reaction Networks and Propositional Logic

Conjunction: AND: &; Disjunction: OR: new reaction line; Implication: FOLLOWS: →; Negation: NOT: -. For instance: a&b&c&..→x. But the resources at the I side are not destroyed by the process then formally a&b&..→a&b&x&… Logic is static because no propositions are destroyed: new implications can be found, but nothing new is created. Negation can be thought of as the production of the absence of a resource: a+b→c+ d = a→c+ d – b. I and O can be empty and a resource can be created from nothing (affirmation, →a) or a resource can create nothing (elimination, a→ or →-a). Another example is a→a and hence a+(-a) = a-a→ and →a-a: the idea is that a particle and its anti-particle annihilate one another, but they can be created together from nothing.

Competition and cooperation

The concept of negative resources allow the expression of conflict, contradiction or inhibition: a→-b what is the same as a+b→0 (empty set): the more of a produced, the less of b is present: the causal relation is negative. The relation “a inhibits b” holds if: : a is required to consume but not produce b. The opposite “a promotes b” means that a is required to produce but not to consume b. When the inhibiting and promoting relations are symmetrical, a and b inhibit (a and b competitors) or promote (a and b cooperators) each other, but they do not need to be. Inhibition is a negative causality and promotion is a positive influence. If only positive influences or an even number of negative influences are included in a cycle then negative feedback occurs. When the number of negative influences is uneven then a positive feedback occurs. Negative feedback leads to stabilization or oscillation, positive feedback leads to exponential growth. In a social network a particular message can be promoted, suppressed or inhibited by another. Interaction sin the network occur through their shared resources.

Organizations

In COT and organization is defined as a self-sustaining reaction system: produced and consumed resources are the same: ‘This means that although the system is intrinsically dynamic or process-based, constantly creating or destroying its own components, the complete set of its components (resources) remains invariant, because what disappears in one reaction is recreated by another on, while no qualitatively new components are added’ [p 8]. DPB: I find this an appealing idea. But I find it also hard to think of the basic components that would make up a particular memeplex, even using the connotations. What in other words would the resources have to be and what the reactions to construct a memeplex from them? If the resource is an idea then one idea leads to another, which matches my theory. But this method would have to cater for reinforcement: and the idea itself does not much change, it does get reinforced as it is repeated. And in addition how would the connotation be attached to them: or must it be seen as an ‘envelope’ that contains the address &c, and that ‘arms’ the connoted idea (meme) to react (compare) with others such that the ranking order in the mind of the person is established? And such that stable network of memes is established such that they form a memeplex. The property of organization above, is central to the theory of autopoiesis, but, as stated in the text, without the boundary of a living system. But I don’t agree with this: the RC church has a very strong boundary that separates it from everything that is not the RC church. And so the RN model should cater for more complexity than only the forming of molecules (‘prior to the first cell’). The organization of a subRN <M’,R> of a larger RN <M,R> is defined by these characteristics: 1. closure: when I(r) is a part of M’ then O(r) is a part of M’ for all resources 2. semi-self-maintenance: no existing resource is removed, each resource consumed by some reaction is produced again by some other reaction working on the same starting set and 3. self-maintenance: each consumed resource x element of M’ is produced by some reaction in <M’,R> in at least the same amount as the amount consumed (this is a difficult one, because a ledger is required over the existence of the system to account for the quantities of each resource). ‘We are now able to define the crucial concept of organization: a subset of resources and reactions <M’,R> is an organization when it is closed and self-maintaining. This basically means that while the reactions in R are processing the resources in set M’, they leave the set M’ invariant: no resources are added (closure) and no resources are removed (Self-maintenance)’( emphasis of the author) [p 9]. The difference with other models is that the basic assumption is that everything changes, but this concept of organization means that stability can arise while everything changes continually, in fact this is the definition of autopoiesis.

Some examples

If a resource appears in both the I and the O then it is a catalyst.

Extending the model

A quantitative shortcoming, a possible extension, is the absence of relative proportions and of the relative speeds of the reactions. To extend quantitatively the model can be detailed to encompass all the processes that make up some particular ecology of reactions.

Self-organization

If we apply the rules for closure and maintenance we can know how organization emerges. If a reaction is added, a source for some resource is added which interrupts closure, or a sink is added which interrupts the self-maintenance. In general a starting set of resources will not be closed; their reactions will lead to new resources and so on; but the production of new ones will stop if no new resources are possible given the resources in the system; at that point closure is reached: ‘Thus, closure can be seen as an attractor of the dynamics defined by resource addition: it is the end point of the evolution, where further evolution stops’ [p 12]. In regards to self-maintenance, starting at the closed set, some of the resources will be consumed but not produced in sufficient amounts to replace the used amounts; these will disappear from the set; this does not affect closure because loss of resources cannot add new resources; resources now start to disappear one by one from the set; this process stops when the remaining resources only depend on the remaining ones (and not the disappeared ones): ‘Thus, self-maintenance too can be seen as an attractor of the dynamics defined by resource removal. The combination of resource addition ending in closure followed by resource removal ending in self-maintenance produces an invariant set of resources and reactions. This unchanging reaction network is by definition an organization’ [p 12]. Every dynamic system will end up in a attractor, namely a stationary regime that the system cannot leave: ‘In the attractor regime the different components of the system have mutually adapted, in the sense that the one no longer threatens to extinguish the other they have co-evolved to a “symbiotic”state, where they either peacefully live next to each other, or actively help one another to be produced, thus sustaining their overall interaction’ [p 12]. DPB: from the push and pull of these different attractors emerges (or is selected) an attractor that manages the behavior of the system.

Sustainability and resilience

An organization in the above sense is by definition self-maintaining and therefore sustainable. Many organizations grow because they produce more resources than they consume (e.g. positive feedback of resources: overproduced). Sustainability means the ability of an organization to grow without outside interference. Resilience means the ability to maintain the essential organization in the face of outside disturbances; a disturbance can be represented by the injection or the removal of a resource that reacts with others in the system. Processes of control are: buffering, negative feedback, feedforward (neutralizing the disturbance before it has taken effect). The larger the variety of controls the systems sports, the more disturbances it can handle, an implementation of Asby’s law of requisite variety. Arbitrary networks of reactions will self-organize to produce sustainable organizations, because an organization is an attractor of their dynamics. DPB: this attractor issue and bearing in mind the difficulties with change management, this reminds me of the text about the limited room an attracted system takes up in state-space (containment) explains why a system once it is ‘attracted’ it will not change to another state without an effort of galactic proportions. ‘However, evolutionary reasoning shows that resilient outcomes are more likely in the long run than fragile ones. First, any evolutionary process starts from some arbitrary point in the state space of the system, while eventually reaching some attractor region within that space. Attractors are surrounded by basins, from which all states lead into the attractor (Heylighen, 2001). The larger the basin of an attractor, the larger the probability that the starting point is in that basin. Therefore, the system is more likely to end up in an attractor with a large basin than in one with a small basin. The larger the basin, the smaller the probability that a disturbance pushing the system out of its attractor would also push it out of the basin, and therefore the more resilient the organization corresponding to the attractor. Large basins normally represent stable systems characterized by negative feedback, since the deviation from the attractor is automatically counteracted by the descent back into the attractor. .. However, these unstable attractors will normally not survive long, as nearly any perturbation will push the system out of that attractor’s basin into the basin of a different attractor. . This very general, abstract reasoning makes it plausible that systems that are regularly perturbed will eventually settle down in a stable, resilient organization’ [p 15].

Metasystem transitions and topological structures

A metasystem transition = a major evolutionary transition = the emergence of a higher order organization from lower order organizations. COT can be understood in this way if an organization S (itself a system of elements, albeit organized) behaves like a resource of the catalyst type: invariant under reactions but it has an input of resources it consumes I(S) and an output of resources it produces O(S), resulting in this higher order reaction: I(S) + S → S + O(S), assume that I(S) = {a,b} and O(S) = {c,d,e}, then this can be rewritten as a+b+S → S+c+d+e. S itself constitutes of organized elements and it behaves like a black box processing some input to an output. If S is resilient it can even respond to changes in its input with a changed output. Now the design space of meta-systems can be widened to include catalyst resources of the type S, organizations that are self-maintaining and closed.

Concrete applications

It is possible to mix different kinds of resources; this enables the modeling of complex environments; this is likely to make the ensuing systems’ organizations more stable. ‘Like all living systems, the goal or intention of an organizatrion is to maintain and grow. To achieve this, it needs to produce the right actions for the right conditions (e.g. produce the right resource to neutralize a particular disturbance). This means that it implicitly contains a series of “condition-action rules” that play the role of the organization’s “knowledge”on how to act in its environment. The capability of selecting the right (sequence of) action(s) to solve a given problem constitutes the organization’s “intelligence”. To do this, it needs to perceive what is going on in its environment, i.e. to sense particular conditions (the presence or absence of certain resources) that are relevant to its goals. Thus, an organization can be seen as a rudimentary “intelligence” or “mind”’ [p 20]. DPB: I find this interesting because of the explanation of how such a model would work: the resources are the rules that the organization needs to sort out and to put in place at the right occasion.

Stigmergy as a universal Coordination Mechanism (II)

Heylighen, F. . Stigmergy as a universal coordination mechanism II: Varieties and Evolution . Cognitive Systems Research (Elsevier) 38 . pp. 50-59. 2016

Abstract

‘One application is cognition, which can be viewed as an interiorization of the individual stigmergy that helps an agent to a complex project by registering the state of the work in the trace, thus providing an external memory’[p 50]. DPB: I understand this as: according to this hypothesis, stigmergy exists prior to cognition; this means that natural but non-living processes use stigmergy on an external medium; once they are alive they are (in addition) capable of internalizing stigmergy, namely by internalizing the medium. The process of internalization of individual stigmergy is the same as (the development of?) cognition. This is another way of saying that the scope of a system changes so as to encompass the (previously external) medium on which the stigmergy takes place. The self-organization is now internalized. Cognition is now internalized. How does this view on the concept of cognition relate to the concept of individuation as a view on cognition?

1. Introduction

‘To bring some order to these phenomena, the present paper will develop a classification scheme for the different varieties of stigmergy. We will do this by defining fundamental dimensions or aspects, i.e. independent parameters along which stigmergic systems can vary. The fact that these aspects are continuous (“more or less”) rather than dichotomous (“present or absent”) may serve to remind us that the domain of stigmergic mechanisms is essentially connected: however different its instances may appear, it is not a collection of distinct classes, but a space of continuous variations on a single theme – the stimulation of actions by their prior results’ [p 50]. DPB: this reminds me of the landscape of Jobs: at the connection of the memes and the minds, there is a trace of the meme left on the brain and a trace of the brain is added to the meme, leaving the meme and the brain damaged. This means that from the viewpoint of the brain the memeplex is the medium and from the viewpoint of the meme the brain is the medium. The latter is more obvious to see: traces can be left in individuals’ brains. The former implies that changes are imposed on the memeplex; but the memeplex is represented by the expression of ideas in the real and in the mind; the real is an external medium, accessible through first order observations; the expression of the memeplex existing in the mind is an external medium, because it exists in other persons’ minds and in versions of the Self, both accessible through second-order observations. Back to the landscape: it is there anyhow, the difference in states is how the Jobs are connected and as a consequence how they are bounded and how they individuate.

2. Individual vs. collective stigmergy

Ants do not require a memory, because the present stage of the work is directly discernible by the same ant, and also by a different ant. Because they have no memory, the work can be continued by the same ant, but by another just as well.

3. Sematectonic vs. marker-based stigmergy

Sematectonic means that the results of the work itself are the traces that signify the input for the next ant and the next state (Wilson Sociobiology, 1975). Marker-based means that the stigmergic stimulation occurs through traces in the shape of markers such as pheromones left by other individuals (ants, termites!) before them, and not by traces of the work itself indicating a particular stage (Parunak, H.V.D., A survey of environments and mechanisms for human-human stigmergy, In Environments for multi-agent systems II (Weyns, Parunak, Michel (Eds.), 2006). Marker signals represent symbols, while sematectonic signals the concrete thing. But this is not straightforward: the territory boundary indicated with urine markers are an indication of the fact that there is an animal claiming this territory, while the urine contains additional information specific for that animal. To spread urine evenly around the claimed area and to interpret the information contained by it is useful for both the defender and the visitor in order to manage a potential conflict. And hence to reduce the uncertainties from the environment for both. The point is that the urine represents both information about the object and about the context.

4. Transient vs. persistent traces

‘We have conceptualized the medium as the passive component of the stigmergic system, which undergoes shaping by the actions, but does not participate in the activity itself’ (emphasis of the author) [p 52]. But a medium is bound to dissipate and decay, unless the information is actively maintained and reconstructed; without ongoing updates it will become obsolete, especially as the situation changes. No sharp distinction can be made between transient and persistent traces, they are the extremes of a continuum. A persistent trace does not require the simultaneous presence of the agent, while a purely transient trace does require their simultaneous presence. Synchronous stigmergy means to broadcast some signal, releasing information not directed at any one in particular. ‘A human example would be the self-organization of traffic, where drivers continuously react to the traffic conditions they perceive’ [p 53]. DPB: the gist of this example is that the behavior of the drivers is the signal: it is synchronous, not directed at anyone in particular, and it is sematectonic, because it represents the state of the system.

5. Quantitative vs. qualitative stigmergy

Quantitative stigmergy means that stronger conditions imply more forceful action to follow, or, in practical terms: the stronger conditions imply a higher probability of action. Qualitative stigmergy refers to conditions and actions that differ in kind rather than in degree: thís trace leads to thát action. There is no clear distinction of these two categories.

6. Extending the mind

‘Traditionally, cognition has been viewed as the processing of information inside the brain. More recent approaches, however, not that both the information and the processing often reside in the outside world (Clark, 1998; Dror & Harnad, 2008; Hollan, Hutchins & Kirsh 2000) – or what we have called the medium. .. Thus the human mind extends into the environment (Clark & Chalmers, 1998), “outsourcing” some of its functions to external support systems. .. In fact, our mental capabilities can be seen as an interiorization of what were initially stigmergic interactions with the environment’ (emphasis of the author) [p 54]. DPB: beetje brakke quote. This reminds me of the idea that a brain would not have been required if the environment was purely random. Just because it is not, and hence patterns can be cognized, it is relevant to avail of the instrument for just that: a brain embodying a set of condition-action rules to generate an action from the state of the environment sensed by it. Stigmergic activity lacks a memory: its state represents its memory as it reflects its every experience. But now the system is dependent on the contingencies of the part of the environment that is the medium: in order to detach itself from the uncertainties of the environment it internalizes memory and information processing.

7. The evolution of cooperation

In a stigmergic situation the defector does not weaken the cooperator: the cost of a trace is sunk.