Complexity and the "Learning Organization"

October 28, 1996

(published in Complexity, Vol. 2 No. 5, 1997)


High performing, creative teams which thrive in complex environments represent the most exciting yet least understood phenomenon in business today.  Improving management's ability to cultivate and sustain such teams is both a business and a cultural imperative.  Managers must get in touch with, and re-examine their presuppositions on how organizations create and deliver value as our "fabricative" work approach is subsumed by "creative synthesis" in the knowledge economy.  


Nonlinear models emerging from Dynamical Systems and Complexity research are opening the door to more insightful ways of perceiving and managing business and work. This note makes "rough and ready" use of these models in exploring the nature of learning organizations, from team to economy.   While the "fit" may be less than perfect, it might best be assessed relative to the goodness of fit of currently employed linear models.   The intent of this disclaimer on the manner of use of these nonlinear models is best expressed via a quote from Wittgenstein [1], "My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them, --- as steps --- to climb up beyond them.  (He must, so to speak, throw away the ladder after he has climbed up it.)"


* * *


The Dipolar Value of Teams


The overall value (V) of the team can no longer be measured and rewarded solely in terms of "fabricative" tangible results (F).  The latent value of group-creativity (C), which may emerge in tangible results, must also be measured and managed.  A problem presents itself here in the tendency of management processes to consider only the "real" component of value and to ignore "emergent value".  That is, the value of creative teams may thought of in terms of a complex variable; V = F + iC.  Rewards and reinvestment schema which do not account for the complex value in "learning organizations" may retard or inhibit their evolution.


The "learning organization" is the conceptual plasma anointed to deliver solutions to rising complexity in business.  One way to envision the "learning organization" is with respect to chance and chaos, ... in terms of "creative interference" within a network of collaborating people through the interplay of purpose, knowledge and random business perturbations.  Complexity models emanating from research into the behavior of "simple" physico-chemical systems, appear ripe ground for the gathering of insights relevant to the "learning organization".


An important bridge between business teams and complexity theory can be found in their common systems structure.  That is, business teams employ dual "polar" modes of thinking and working which seem to equate well to the concepts of equilibrium (stable periodic) systems and nonequilibrium (chaotic) systems.   Teams, like most natural processes, are capable of both "fabricating" results through local rule-based assembly processes and/or "creating" results through ad hoc "spatial" collaboration.  These "dipolar" qualities can be seen as being both complementary and "antagonistic".


Probabilistically, the former pole associates with the adding up of "incompatible" (non-interfering) events, while the latter pole associates with the spatial interference of "independent" events.   While the traditional emphasis has been on "fabricating", increasingly complex environments are demanding a higher emphasis on "creative" work modes.  This shift in emphasis is well depicted in the film Apollo 13 when the team, confronted with a life-threatening technical disaster, subordinates their linear, equilibrium "fabricative" mode to their nonlinear "far-from-equilibrium" "creative" mode.


The fundamentally differing process modes of "fabricating" via linear, non-overlapping assembly and "creating" via nonlinear interference appear to be archetypes in productive processes, and will provide the "backbone thread" in this discussion.  The term "interference" is used in preference to "interaction", to imply the degree of participative commitment observed in high performing teams.  Interference and interaction express the relative degree of commitment to teamwork, as does the pig or the chicken, respectively, to a bacon and egg breakfast. 


A convenient way for introducing these productive archetypes in the context of chance, may be to review the classic quantum behavior experiment in which electrons pass through one or the other of two slits on their way to a target screen. 


 Examining the probabilities in the case (a) where "we look" at which of two alternative slits the electron passes through, and the case (b) where "we don't look" at which slit it goes through, we get the following result; [3]  (where big P is probability of the event and little p is probability amplitude).


(a) If we look .....   P_L = p1**2 + p2**2


(b) If we don't look ..... P_NL = (p1 + p2)**2   = P_L + 2*p1*p2


From probability definitions, it appears that if "we look", we concern ourselves only with "incompatible" events (i.e. probability sums), or events which can have no overlap.  In other words, if "we look", we get the mechanical sum of the parts.  Whereas if we "don't look", we open the door to consideration of both "incompatible" events and "independent" events (i.e. probability products), so that we get a creative lagniappe in the form of an emergent spatial correlation.


It seems that even with electron-teams, we have the choice of looking at their process as being mechanical or creative.


While porting model concepts from the macro level of teams to the micro level of electrons may seem a bit radical and scientifically questionable, the focus here is on four "generic" points-of-interest or "patterns", relevant to "production"  as follows;


1. The polar archetypes of non-overlapping assembly (i.e. piecewise fabrication) and creative interference (i.e. joint creativity) seem pervasive in natural processes


2. There is a difficulty at all levels in attempting to model processes as being either purely "fabricative" or purely "creative"


3.  At all levels, probabilistic spatial correlations, suggestive of large-scale organizational coherency, appear to be present only in creative mode.


4.  While the "fabricative" archetype is tangible-causal, the "creative" archetype is latent-emergent, suggesting a relationship between overall value (V) of the process, and its "fabricative" and "creative" value poles, "F" and "C" of the form  V = F + iC.


Business appears to be shifting emphasis from "fabrication" to "creativity" to the point where a polarity flip is occurring in which the former is subordinated to the latter.  For example, in the ideal "learning organization", it would seem that the "real" or "fabrication" component of value is the emergent offspring of "creative" value.  This is in contrast to the industrial age model where tangible, quantitatively measured "fabrication" value was "all she wrote".


As we transition from "fabricating organizations" to "learning organizations"  a number of questions arise, including; how and when to shift emphasis or reverse polarity, how to recognize and manage the more elusive creative-emergent mode, and how cultural factors influence the embracing of one polarity rather than the other.  Each of these questions will be discussed in the context of macro experience, against a backdrop of physics and complexity models.


Polarity Reversals in Teams


As mentioned above, it would be hard to find a better example of a polar reversal from a dominant linear, non-overlapping assembly orientation to a dominant nonlinear interfering orientation, than the transformative experience which is so graphically illustrated in the film "Apollo 13".  Challenged by a technical disaster, the team does a polarity flip from operating as a hierarchy of local, specialized equilibrium centers, into a strongly spatially correlated, creatively interfering web enterprise.  Both team performance and the actualization of team-member potentials make an enormous forward leap.


In Apollo 13 as in other such team experiences, this "polarity reversal" seems to involve an associated "flip" in team members' modes of perception or consciousness.  Interviews with members of high performing business teams in the upstream petroleum industry suggest that a similar phenomenon occurs as is commonly encountered in sports or ballet; i.e. a transitioning from a piecewise mechanical assembly of thoughts into a seamless creative mental synthesis.  As in the quantum behavior experiment, it seems that if one persists in "looking" at the process in a mechanical context, one cannot "get into" the natural interfering mode at the same time.


It is hard not to compare this transitioning between the two modes to physico-chemical systems transitions as are described by Prigogine and Nicolis [2], in both a probabilistic as well as behavioral sense.  The probabilistic aspects become particularly important relative to appropriate rewards and reinvestment feedback, as will be discussed later.


While emphasis in business is on tangible results which are delivered through "fabrication", both modes of value creation are commonly represented on business teams.  In the past, where there was a dominant focus on "fabrication", it appears as if the "creative" mode was seen in a background support role and taken very much for granted.  In the "learning" organization, these perceptions appear to be reversed.


There are at least two problems associated with this mixed mode of working. First, non-overlapping assembly may have worked fine in an environment which emphasized manual skills, but in a knowledge-work environment, it tends to "strip-mine" knowledge workers, extracting contributions which are narrow with respect to both the overall business process and the natural potentials of the employee.  Second, having been taking for granted for so long, the "group-creative" mode, and "organizational learning" are currently neither well understood nor well nurtured, and tend to be inadvertently suppressed by deeply ingrained cultural beliefs.


For example, the practice of seeking non-overlapping contributions from organizational subcomponents, by clearly defining local roles, responsibilities and quantitative measures, runs directly counter to the need for creative interference (spatial coherence), which derives from an ad hoc knowledge sharing environment.  In addition, latent value potentials embodied in creative interference are lost or abandoned in the single-minded focus on the "tangibles" of "fabrication".


Recognizing and Managing Latent-Emergent Value


Nonlinear, creative interference in the macroworld of teams is difficult to get a handle on for three reasons; we can't speak directly about it, we can't remember it, and we can't quantitatively measure it.


We can't speak about it directly or remember it because it is higher dimensional and intrinsically geometrical.  For example, we can't directly describe how we ride a bicycle, although riding a bicycle is a sustained learning process involving only ten interfering variables [4], (i.e. business phenomena can involve far more).  It is impossible to convey in words the information contained in ten-dimensional or higher, phase space trajectories since business language is inherently constrained to three or four dimensions (i.e. poetry is typically banned).


Neither can we "explicitly remember" a learning experience such as a bicycle ride, since we store such pattern-based experience in "implicit memory" [5], a type of memory which does not allow voluntary recall but only a "knowing" when we once again see or experience the thing or episode.  For example, though we cannot "explicitly remember" and voluntarily recall all 50 faces we saw at yesterday's meeting, we will likely be able to "implicitly remember" most of them, or "know them when we see them" if we view them in a "lineup" in which they are interspersed with strangers.


Thus, we can neither linguistically describe, nor voluntarily recall nonlinear interference effects such as organizational learning experiences, ... a frustrating circumstance for employers (and machines) vying for increased control over knowledge.


While measurement in the fabricative environment is straight forward due to the non-overlapping constraint, measurement in creatively interfering environments is complicated by the notion of latent value associated with spatial correlation.  While this is hard to visualize in complex business situations, a billiards example may serve to illustrate the effect.


A manager of a billiards team (which inherently deals with creative interference) would look beyond "scoring statistics" to the latent value implicit in the spatial configurations; for example he/she might ask the questions; ...  is the team moving towards snookering themselves? ... are they over-emphasizing short term scoring opportunities to the point that they can no longer extract value from the residual configuration? ... are the players going for best opportunities first (re-configuring versus sinking balls) to optimize full cycle team results, even though their mates may harvest the opportunites they set up?


Since billiards is a classic case of "deterministic chaos", there can only be probabilistic measures for the "value" of a billiard ball configuration.  Such interference configurations apparently fall into the higher dimensional phase space category, similar to the bicycle-riding system, which cannot be "explicitly remembered" nor described in words nor quantitatively measured.   The dominant role of latent, interference value in team-creative situations implies that honesty, trust and humility will be needed for system optimization.  That is, no-one knows for sure if you choose to sink a ball and forego a better (with respect to full cycle team results) opportunity to improve the configuration; you must trust that if you take no ownership of the scoring opportunities you set up, your teammates will do the same; and if your ego drives you to go for high personal scores to the detriment of creative interference, full cycle team performance will necessarily be compromised.


Rewarding only quantitative results in a creative environment where latent value is tied up in spatial interference patterns appears likely to drive the system back towards the "fabrication" model and suppress creativity.   Sporting games such as ice hockey try to  mitigate this exposure by rewarding creative interference in the form of "play-making" (i.e. assists) in the same measure as scoring (both a goal and an "assist" merit one point).  Thus it is quite possible to be recognized as a "star" for creative play rather than for high scoring (e.g. Wayne Gretzky).   In business, however, "play-making" and "assists" are far less visible to management, and an insistence on honesty, trust and humility becomes the joint responsibility of the team.   Since behavioral information is only locally accessible, the maintaining of large spans of control by management depends on quantitative measures which discount creative interference contributions.


In moving beyond questions of relative optimization to the threat of collapse into total chaos, it may be worthwhile to revisit the bicycle example.  However, in this case it's more insightful to consider a thought experiment in which the bicycle is ridden by a team of five tiny but powerful people, situated so as to manage the dual variables of position and velocity associated with; handlebars, left pedal, right pedal, drive wheel and front (braking) wheel.  In the absence of the inevitable potholes in the pathway, this is a ten dimensional interference problem.


As previously suggested, sustained creative interference is needed to keep the bike upright and on course, which is not describable in words, not quantitatively measurable and not explicitly rememberable.  Thus a successful bicycle ride by the "tiny-team" would seem to involve the sharing of a mental view of a family of phase space trajectories associated with a successful ride.  In this example, after a period of learning, a false movement of the pedal by one team member can be "covered" by a compensatory move of the handlebars by another, introducing a degree of resilience into the system.  Thus it was in the Apollo 13 project where diversity and overlapping capability inherent in the interfering variables delivered a resilience which averted a collapse into total chaos.  And thus it was that the training of a new team member was a nonlinear, experiential task, in which words and documents played a support role.


In view of the above-described properties of dipolar value creation systems, there are a number of problems associated with making rewards and reinvestment programs overly dependent on quantitative measures (as opposed to behavioral measures).  One is that ANY differential in rewards based on quantitative subsystem measurements (pedaling proficiency, ball-sinking etc.) is likely to introduce confusion and weaken spatial coherency necessary for the creative mode.  In the bicycle "tiny-team" case, and as affirmed in interviews with high performance team members, one must shift out of a piecemeal mechanical focus on how things work, into an interference (harmonizing etc.) orientation.  If the pedaler is "looking" at "pedaling performance goals", he/she cannot at the same time be tuning into high dimensional phase space trajectories.


Another problem is that since value builds in creative environments, through creative interference, at any given stage the value may be more or less manifest in real results or stored as latent potential (recalling that V = F + iC).  If the focus is exclusively on quantitative results, this equates to looking only at the "real" component of value and ignoring the "complex" component.   If assessments are made on a full life-cycle basis (e.g. on the final outcome of a billiards game), this is not a problem since latent value has at that point emerged as measurable value.  But if performance is assessed on the basis of the "real" component at arbitrary intervals, for example as Wall Street assesses companies through their quarterly reports, this practice may fail to account for latent value resident in creative interference potentials.  And since rewards and reinvestment are often based on such assessments, this tends not only to "starve out" creative resources, but also to encourage a focus on "speed of fabrication", dissipating the spatial coherency necessary for organizational learning.


One of the saving graces of the results-oriented rewards system in the past has been the physical collocation of "fabricative" and "creative" contributions.   As organizations become "virtual" through the availability of network technologies, there is likely to be a progressive diffusion and decoupling of both "F" and "C", which will make this compensatory spillover in rewards and reinvestment much less likely, and the need for behavior-oriented rewards and reinvestment an imperative.


Such a shift is becoming increasingly evident.  Examples include Southwest Airlines, where CEO Herb Kelleher hires and rewards "great attitudes" [6] and Motorola, where Bob Galvin, head of Motorola's Excom leads an initiative to re-invest through "behavioral leaders" [7].  In the former case, hiring programs were developed which cost several times the norm, to recruit staff who not only possessed the "pedaling proficiencies" but also the right "attitudes" relative to promoting synergies in the workplace.  In the latter case, a multi-billion dollar investment has been made, through the vehicle of Motorola University, to develop "behavioral leadership", or "role model" leaders who are differentially rewarded and empowered as investors of company funds, on the basis of their ability to catalyze learning synergies in the workplace.


The Influence of Culture on Work Approach


The focus on non-overlapping mechanical assembly processes and rewarding and re-investing on the basis of quantitative measurements is linear theory based practice which appears to be deeply ingrained in the culture of many organizations.

That is, linear theory comes into play in the form of presuppositions which are at a level lower than the questions we normally ask.   Getting at the root source of these presuppositions is an exercise in what Argyris [8] calls "double loop learning".


In this case the question is not; "how can we devise more equitable ways to measure and reward results in creative team environments", but rather; "are traditional results-based measurement and management approaches appropriate for creative team environments?"


Several new questions arise from this level of inquiry, including;


1. Are traditional statistical estimates of causality appropriate for correlating business resources/practices with business success?


2.  Is the application of rewards and reinvestment at the local portals of emergence of results appropriate for nurturing both "fabricative" and "creative" organizational modes?


In America, there seems to be a strong cultural presupposition that, for example; if cancer is statistically correlated with smoking, then smoking is a CAUSE of cancer, or; if business success is statistically correlated with Fred's participation, then Fred is a CAUSE of success.   Business corollaries also appear to emerge as follows;


(a) Those who CAUSE success are the best stewards of re-investment


(b) Investment applied at the local portals of emergence of results will "trickle-down" through causal channels to equitably nurture contributory resources and practices.


This causal, trickle-down view of business-economy appears to be the son of culture and the mother of behavior in the traditional business world, and tends to boost both the ego and pocketbook of those who would "fabricate" results most rapidly.


There is little evidence of concern amongst senior management (or investors for that matter) that rising complexity in the form of multi-variable interference may be eroding the presuppositions which underpin statistics-based rewards and reinvestment schema.  Yet researchers in nonlinear phenomena suggest that the spatial correlation which emerges in "transition" or "learning" phases is at the expense of equilibrium statistics in a system.   Such findings seem to jive with the bicycle and billiards examples above.  Questions like; "who is 'most' responsible for winning the billiards game, ... the highest scorer?" seem out of place in situations involving creative interference.


Unfortunately, the billiards player who performs best in an "organizational learning" context, by sacrificing scoring opportunities when there are better overall opportunities to improve the ball configuration, will necessarily have an inferior scoring record relative to an equally competent teammate who exploits personal scoring opportunities on any and all occasions.  Thus the practice of "rewarding results" would seem to foster internal competition and distrust, creating an environment inherently disadvantageous to the evolution of "learning organizations".  


Both regional and local culture appear to influence the degree to which this type of thinking is impacting the evolution of "learning organizations".  For example, in Japan where there is a great respect for power, one might expect to see an even larger "spread" in compensation between senior management and the average wage.  In fact, there is almost an order of magnitude less of a spread than in the US.  It appears that there is much more "humility" in terms of causality in Japan.  Cultural "interference" archetypes such as the Zen "sound of one hand clapping" or "Karma" which "selects" a particular result out of a diverse soup of interfering potentialities, appear to moderate the notion of a "statistical causality".   A similar "humility" with respect to the "causality" of success appears also to be present in high performing "learning" teams (e.g. Apollo 13).  In this case, "humility" appears to equate to a more balanced view of the origins of value contributions which incorporates both tangible-causal and latent-emergent contributions.  Thus an effective "Ego-nomics" comes into play, born from cultural presuppositions rooted in mathematics, which appears to be modulating the evolution of "learning organizations".


In America, the influence of "ego-nomics" is particularly strong, with senior management compensation being more than 40 times as great as the average wage.  According to economist Edward Wolff of New York University [9], wealth follows earnings and inequality of wealth now stands at a 63-year high.  "The top 1 percent controls 42 percent of the nation's household assets and 50 percent of its financial assets, such as stocks and bonds."


There are, thus, some very strongly established feedback patterns which suggest that a review of manager/investor assumptions on "causality" and "trickle-down" in rewards and reinvestment schema,  and their mathematical underpinnings, may be in order.


Logical Incompleteness and Cognitive States


As suggested earlier, high performing, "learning teams" appear to make greater than normal use of pattern-oriented knowledge synthesis.   This is articulated by team members in the context of shifting from "factory" to "creative" cognitive modes.


The intention here is not to get into cognitive theory but simply to try to reconcile experiential data coming from highly successful teams with the "mathematics" which underpins cultural presuppositions.


In examining why the latent value in "organizational phase space" tends to be ignored by business managers, Gregory Chaitin's suggestion [10] comes to mind, that the incompleteness expressed by Goedel's theorem may be more pervasive and significant than previously assumed.  The proposition that "all Cretans are liars", which provided an avenue of attack for Goedel's original proof, seems relevant to the topics of "learning organizations" and "creative interference", since the proposition presupposes both consciousness (i.e. one must first "know" truth in order to lie), and interference (i.e. "lying" inherently involves interference between two or more people).


In order to prove that "all Cretans are liars", a proposition that implies both consciousness and interference, it appears that we need more than unconscious, non-interfering axioms.  That is, the Cretan proposition appears to be "complex" or "dipolar" while syllogistic logic is "real" and "mechanical".   Perhaps the "incompleteness" alluded to in Goedel's theorem is also playing havoc in the domain of management.  The counter-intuitive shockwave that hit mathematicians in 1931 may be only just now knocking at management's door.


As in the case of quantum behavior and mixed mode teams, if we "look" (or "manage") in such a way as to guarantee non-interfering subsystems, then we constrain ourselves to a linear summing of non-overlapping parts.  However, if we "do not look" (i.e. we empower self-managing teams), we open the door to creative (and/or destructive) synergies.  How "tangential" we can get (i.e. to what extent we should constrain C/F) would appear to depend on questions of diversity and resilience to business perturbations.  A time-based plot of C/F for the Apollo 13 team would probably show a very large peak following the technical disaster, implying a closer encounter with total chaos than most teams would be capable of recovering from.


To summarize the "logical completeness" issue; for managers to understand "organizational learning", a process which implies both consciousness and interference, the manager will need more than mechanical-causal axioms.   As emergent business trends seem already to be saying, "organizational learning" is best measured and managed in terms of "behaviors" which promote synergistic interference (large-scale spatial correlation) and through which "cycles of learning" can be progressively improved upon.   In other words, the move to "behavior" as the prime management parameter, by definition, equates to the employing of conscious, interfering axioms.


The shifting of rewards and reinvestment away from a pure "results" base seems unlikely in companies where the egotism of senior management is high, since the notion of "responsibility for success" (i.e. "statistical causality") and therefore the degree of "spread" in rewards, is likely to be severely diluted.  Executive compensation may have to go through the same "stochastic analogue of bifurcation" [1] as is described in complexity theory, as the system properties which business most values and rewards transition from a unipolar manual skills orientation, to a dipolar knowledge management orientation.


Teams and companies who lead the way in transforming themselves into "learning organizations" seem graced by leaders who sense the "emergent value" of empowered teams and who possess more than average measures of honesty, trust and humility.


[1] Wittgenstein, Ludwig, "Tractatus Logico-Philosophicus", 1921


[2] Prigogine and Nicolis, "Exploring Complexity", 1989


[3] Feynmann, Richard P., "Six Easy Pieces", 1963 (published 1994)


[4] Stewart, Ian, "Does God Play Dice?", 1989


[5] Schacter, Daniel, "In Search of Memory", 1996


[6] Kelleher, Herb, "Quality Awareness", Dallas, Oct. 9, 1995


[7] Galvin, Robert W., "The Learning Organization", Keynote Address, "Knowledge Advantage II Conference", Chicago, Nov. 16, 1995


[8] Argyris, Chris, "Education for Leading-Learning", 1993


[9] Bryant Quinn, Jane, "A Paycheck Revolution in '96", Newsweek, Feb. 19, 1996


[10] Chaitin, Gregory, Web site: