- The Idea: Our Way of Thinking Often Does Not Match Our World
An added perspective on thinking.
For the past century, Western thinking has focused on the idea that the world is organized around discrete objects which aggregate and have simple relationships. The paradigm is that of physics. All things should be explainable through rules, laws, and algorithms. The observer is not a part of the observation since the "things" themselves constitute reality. Despite the successes which this frame for thinking has produced, it has a serious deficiency: How is it that the actions and behaviors of reflexive anticipatory creatures are best described by rules for non-thinking non-reflexive non-anticipatory objects? How is it that context is deemed to not matter? And what about complexity? Those relationships which cannot be described by the simple? The physics based frame has no answer and instead discards these issues with the magic words: Ceteris paribus. But we do not live in a Ceteris paribus world. Thus arises the need for some other frame of thought. Without access to these other frames our tools for understanding are inadequate for the world around us.
- The Problem: The World is Often Complex & Context Dependent
We Do Not Live in a Simple World
But Our Method of Thinking Assumes We Do
We give priority to physics and physics like ways of thinking -- where objects are discrete, come together in aggregates, and form simple relationships.
Such thinking assumes that the goal is "reliable prediction." We enquire so that we can reliably predict and we then act based upon the models we create to represent such predictions. But, in parts of the world (especially those parts involving the items listed at left) reliable prediction is a logical impossibility. Moreover it is often impossible to disregard the role of the observer. This part of our world is complex -- not simple -- and the simple rules do not apply.
Complicated & Complex are DIFFERENT
Complicated is about aggregates,probabilities, and unfoldings.
Complex is about interweavings -- where relationships and organization matter.
If we lack the basisc impulse to think about complex relationships and their implications, our thinking tools are insufficient to deal with the complex portion of our world.
- Cognitive Phenomenology: The Core of Epi-Thinking
Arise in the Process of Sorting the Undifferentiated into the Complex and Attuned or the Simple and Ordered
"Atoms are not things, they are only tendencies. -- Werner Heisenberg"
"While most people understand first-order effects, few deal well with second- and third-order effects. Unfortunately, virtually everything interesting lies in fourth-order effects and beyond." -- Jay Forrester
This Narrated Presentation describes the Critical Element of Epi-Thinking
- Science I: A Method for the Simple and Complicated
Traditional "Western" Science
The goal of traditional western science is explanation via reliable prediction.
Scientists examine and model the world so as to enable us to make reliable predictions. These reliable prediction models are the product of either deduction or induction via probabilistic inference.
A key assumption in this realm is that the predictions based on such models are a suitable basis for next actions. Should the predictions prove to be unreliable then the basis for next actions becomes confounded by uncertainty. It seems to be the rare scientist who fully states the basis on which these claims of reliablity are made.<
Another key assumption in the Science I world view is that the "laws" of science are independent of context. This assumption allows the scientist to assert Ceteris Paribus (all other things being equal) with impunity. What is seldom asked is, "Does the assumption of Ceteris Paribus matter?"
In Physics, Ceteris Paribus rarely matters. By contrast, in sociology or economics Ceteris Paribus matters greatly. The methods, models, and emphasis on reliable prediction in traditional Western Science I work best in those realms where either Ceteris Paribus does not matter or where it can be claimed as an assumption.
The models and world view of traditional science are based on restrospective looks back and on the assumption that such retrospective views will continue to apply (at least in the near future). Traditional science is about what "was" and then predicts the future based upon the past.
Both deduction and induction via probabilistic inference are restrospective analytic techniques. The make use of past data, attempt to nullify any observer dependent effects, and then model a prediction. The concept of feedback and feed forward loops which may change conditions or interpretations and thus further actions is usually specifically excluded from the modeling.
A key technique of modern science is to find abstract representations which can be manipulated in models to stand for the "real thing." Often these representations take the form of a mathematical expression. Formalisms then are allowed to dominate the discussion of theory and the very notion of embodiment goes into remission.
The consequence of such a representation centered view is that the labels used to designate categories are seen as "average examples" of the category and not as exemplars. The notion of differentiation in embodiment is often overlooked. Labels and categories are treated as "codes" with defined (non-contingent) meanings.read more
- Science II: A Method for the Complex and Context Dependent
When reliable prediction is not possible and the focus shifts to the future and to future actions, the goal shifts to "better" attunement to the affordances present in the situation. "Better" is a relative subjective judgment left to the observer/actor.
Affordances are possibilities offered to an actor by the present environment/situation. They are a joint product of actor, environment, awareness and attention. If an actor is unaware of a possibility, that possibility is NOT afforded by the situation. If an actor pays no attention to a theoretic affordance, that affordance is NOT embodied. Categories and labels are insufficient to describe affordances.read more<
With a focus on environment, situation, affordances, and possible actions, the possibility of emergence grows in importance. Emergence is the embodiment of possibilities which differ from those available to components and assemblages when components are organized into a new interwoven whole (the family is emergent from its members, matter has properties which differ as states/phases change).
The possibility of emergence and the importance of situation combine to place an increased emphasis on resilience. Resilience is the capacity of a system or actor to maintain its integrity in the face of dramatically changed circumstances. The quest for optimization tends to presume ceteris paribus. Resilience reverses that presumption.Emergence read more Resilience read more
Many actors in the world are endowed with abilities which go beyond action and programatic reaction. These actors anticipate the consequences of potential actions before they take them and then reflexively consider potential outcomes of those actions before undertaking them.
Reflexive anticipation by definition includes feed forward and feedback loops which act as both enablers and constraints in reaction to the affordances presently available. Reflexive anticipation cannot be modelled in an algorithmic manner and, in the absence of ceteris paribus, cannot be predicted. Simulations are not substitutes for prediction, but instead serve to "prime" models -- and such priming then enters back into the reflexive anticipation loop. We often refer to reflexive anticipation as "learning."read more
In situations of reflexive anticipation abstract representations seldom provide sufficient information to answer the key questions of "what if?" or "how?". Linear representations of cause and effect fail to take into account the effects of feedback/feedforward loops, learning, attention, and situation. Narratives do.
But, it is not just any kind of narrative which works -- instead it is narratives which purport to tell astory of explanatory mechanism, a "how come" or a follow-on to a "what if". These narratives allow the observer to retell the story from his/her own perspective and to make substitute "what if" conditions. Explanatory narratives are spun from cues observed when a situation is examined for affordances. They are not abstract nor categorical.read more
- Implications of Epi-Thinking
Implications of Epi-Thinking
Action by sentient beings is most often the result of reflexive anticipation. This implies that when we are projectingf the actions of others we need to take into account the context, history, attention and forecasting abilities which those others will bring to the situation at hand.
To attribute to others a non-thinking non reflexive response -- as we do when we make use of a probability distribution to assert what others will "likely" do -- is to ignore the power of reflexive anticipation to not only affect the choices otehrs make but also the situations in which they allow themselves to be in.
Thus it is critical to ask what reflexive anticipation is occurring.<
All too often category membership or a label is used to both define a situation (synecdoche) and then to "explain" likely behaviors and outcomes. The error of synecdoche is to rely on the assumption that the chosen "part" (the label or category membership) adequately reflects the "whole" given the likelihood of context dependence.
The assumption that a given model (and/or its results) are context invariate needs to be challenged before one makes a decision based upon that implicit assumption. Unless the question is asked and the assumption challenged, actual situation dependency can be overlooked or ignored.
It is critical to ask if one is giving full consideration to the individuality of the situation at hand or is instead relying on labels or category membership.
Fundierung is the relationship which we take for granted in the exercise of function. Often the explicit recognition of that relationship is capable of altering the exercise of the function itself.
To focus on the makeup of the pen with which we write is to stop focusing on both the act of writing and the writing itself.
The action of shifting attention to either reflexive anticipation or assumed synedoche is to question the cognitive fundierung which is taken for granted in the decision making which precedes action. It involves asking questions which are seldom asked and pausing to stop what might otehrwise seem to be "automatic" modes of thought.
There is great power in this, perhaps momentary, shift in attention.
Homologies are models (using the exacting definition of Rosen) which can be used to describe two or more situations (which then "share" that underlying homology). Homologies are applied by asking questions of the model and retelling the narratives which result.Indexicals, by contrast, are representations which are used to stand in (like an index) for the individual items they supposedly represent.
The error which often implicitly occurs is to confuse the functions of homology and indexical. Asserting a homological relationship for an indexical is the assertion of a weak analogical relationship which ignores context dependency. Asserting indexicality to a given homology is to give the model more status than its embodiment and again to ignore context dependency. By asking about reflexive anticipation, synecdoche and fundierung the applicability of homology or indexicality can be better assessed.
- What is Next: Applying Epi-Thinking
Asking whether context matters is epi-thinking. Assuming that labels and categories apply and/or that context can be disregarded will only work in situations which call for Science I. Demanding a role for context in Science I scenarios will likely only produce noise -- much like ignoring the role of context in Science II scenarios.
When actions are the focus, context is critical. Basing actions and reactions on a model which fails to explicitly take context into account will lead to error unless context does not matter.
When categorization is the focus, context is far less critical. Category membership is often invariate to context, and when context does matter it often can be used to distinguish sub-categories.
Actions have a fundierung relationship with context. Cognition
has a fundierung relationship with semiotic affordances.<
Asking "Does the assumption of Ceteris Paribus matter?" goes a long way to determining if one is operating in the Science I world or the Science II world. So does asking what happens if the ceteris paribus assumption was discarded.
Epi-thinking makes explicit recognition that often the Ceteris Paribus assumption is asserted but not mentioned. Instead there is an assumption that a given model (and/or its results) are context invariate. Unless the question is asked and the assumption challenged, actual situation dependency can be overlooked or ignored.
Epi-thinking asks "Is the goal prediction (where the focus is often NEXT action) or attunement to affordances?".
Prediction is based upon category membership and models of prior actions and behaviors. Attunement is tied to current perception and anticipation of future situations. Prediction looks back; attunement looks forward.
Science I is about making predictions based upon retrospective observation (via deduction and probabilistic inference). Science II is about making abductive hyoptheses regarding prospective futures. The techniques involved in drawing conclusions while looking backward and looking forward are seldom the same and are unlikely to work well when applied in the incorrect realm. The act of selecting perspective (forward or backward) and heuristic for conclusions is one of epi-thinking. To "allow the situation to choose" is not.
There are situations when explanation by means of category membership is more than suffcient. There are other situations which demand the generation of an explanatory narrative. Epi-thinking involves recognizing both that these situations differ and making the appropriate inquiries to determine which situational type applies.
When explanation by category membership is sufficient, then the quest is to generate the "best" label. Once that label is ascribed, the rules and algorithms of Science 1 will allow reliable predictions to be generated, along with their resulting implications for next actions. But when explanatory narratives are called for, and reliable prediction logically implausible, then the quest is to generate "good questions" which can further the process of abductive inquiry. Epi-thinking is about knowing when to generate labels and when to ask probing questions.
- Key Concepts & Vocabulary
The Vocabulary of Epi-Thinking
Critical Concepts of Epi-Thinking
- Reference Works
Books which influenced Epi-Thinking
Journal Articles and Web Sites which influenced Epi-Thinking
Consider attending the Modes of Explanation Conference in Paris May 22-24, 2013