Content uploaded by John Thomas
Author content
All content in this area was uploaded by John Thomas on Oct 29, 2018
Content may be subject to copyright.
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution
License 4.0 (http://creativecommons.org/licenses/by/4.0/).
* Corresponding author: johntom@cogtools.com
Complex Adaptive Blockchain Governance
Thomas John1*, Mantri Pam1
1Cognitive Tools Ltd. LLC, P.O. Box 695; 255 North Ave; New Rochelle, NY 10801, USA
Abstract. The blockchain revolution upholds the decentralizing ideal of “control nothing.” It is natural that such a
pursuit would face issues of governance that demand reasonable control; control that is both operational as well as
adaptive in nature. Eliminating middlemen and handing over controls to a trusted system of trustless agents does not
thereby bestow trust across time. This is especially true when relentless change is the order of the day. Issues of
governance rise up when blockchain systems (especially those that have embedded smart contracts) are forced to
operate increasingly away from their original intent. Smart contracts need governance when beset with the problem
of the unknown-unknowns. Guided by the axiomatic approach, this paper looks at the paradoxical issue of blockchain
governance from a Complex Adaptive Systems (CAS) perspective that helps frame the fundamental problem of
decentralization. The objective is to solve the Blockchain Governance Kernel Design. Real-life examples are used to
illustrate the findings.
Keywords: Blockchain; Stigmergy; Governance; Complex Adaptive System; Heterarchical Hierarchy; Unknown-
Unknowns; Smart Contract
1 Introduction
Consider C.P. Snow’s proposition of the growing
chasm between “The Two Cultures” [1]; i.e., between the
sciences (which includes the social sciences) and the
humanities, but now from a design perspective. Design as
a discipline that deals with human artifacts (be they social,
technical or socio-technical), has to bridge Snow’s chasm
in every single instance of design. This is because
meaning and purpose, i.e., the root Functional
Requirements (FR’s) that mandate any given design,
ultimately reside in the human-centric humanities which
includes disciplines such as languages, history,
philosophy, arts and the law [2-4]. Thus, for example, the
design of an equitable governance system is ultimately
rooted in the realm of law and justice. The research
reported herein integrates across both the above cultures
in order to make explicit the kernel (or germline, as
opposed to somatic or operational) governance design in
the context of the ongoing blockchain revolution.
There is a fundamental difference in the requisite
bridging-over that is necessary when considering
technical versus social systems. Technical system
designers have well-developed disciplines such as
cognitive engineering, business-analysis, ergonomics,
and others to help establish the preamble and move the
design activity into the technical realm. In contrast, social
system design is barely a discipline. There is no similar
preamble body of knowledge that helps translate the
social system FR’s and DP’s (Design Parameters) into the
language of the social sciences. There are no rich
traditions, no well-accepted bodies-of-knowledge as to
how design operates in the social realm. For example, it is
only recently that the nascent concept of stigmergy is
helping disambiguate Adam Smith’s economy-wide,
organizing principle of the “invisible hand” [5]. A
disciplined approach to design, therefore, is more evident
in the technical as compared to the social/organizational
realms. The design of technological artifacts is more
amenable to principled structuring as compared to the
design of social artifacts. As Prof. Suh has noted, the ad-
hoc approach is the accepted norm in the case of social
artifacts [2]:
“In many organizations well-defined FRs are often
lacking or not completely understood by everyone in
the organization, and the organizational structure
does not have specific DPs to satisfy FRs. The job of
the management is to define FRs and establish DPs,
but this has been done ad hoc, very much as in other
fields of design.”
While technical system designs have come a long way
since the 1990s when the above critique was first made,
social system designs remain as is. Heretofore, the social
and the technical have existed side-by-side, content to drift
apart in their separately evolving cultures. However, now
we are entering the realm of massively fine-grained socio-
technical systems such as the world of IoT (Internet of
Things). The above divide across these two cultures is
therefore not sustainable. The odd marriage between the
two cultures does not scale; instead, it has the potential to
result in large-scale, out-of-balance socio-technical
system failures. Every system has a certain capacity for
change beyond which it starts to show pathologies.
Current social systems are ill-prepared to receive dramatic
2
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
influxes of technology such as the promise of IoT [6].
Design of the governance kernel of socio-technical
systems is a challenge of the first order. This challenge is
two-fold. Governance has to operate both at the agent as
well as at the institutional level. Firstly, the design of
social systems is more nuanced than the design of a purely
technical system. This is because natural laws are easier to
discern compared to the social. For example, the ideal of
justice enshrined in the rule of law that governs a given
social order is not as apparent as, the force of gravity
enshrined in the Newtons law of gravity. Natural laws
govern the physical sciences; there is, therefore, no other
governance needed. In contrast, design (be it technical,
social, or socio-technical) requires governance--since it
needs to account for the human element. The bounded
rationality problem conditions human decision-making.
While natural laws are universal in scope, human decision
making is constrained to operate within the limited
knowledge scope of individuals and groups. Humans have
different knowledge bases and perceptions. These
differences inevitably lead to misunderstandings and
conflicts that then needs governance at the agent level.
A second issue that social system designers face is that
of institutionalized injustice. Human agents are endowed
with free-will. As we just discussed, when bounded
rationality meets free-will, human agents may commit
erroneous decisions. However, free-will is also a
significant reservoir for corrective governance forces that
unleash when agents face institutional-level injustices.
Indeed, the exercise of free will (both individually as well
as collectively) has the power to change the course of
nations and organizations. In other words, organizational
units do not have the power to dictate commands that
violate the moral code unilaterally. Given the painful
history of how humans have socially engineered their way
into political/criminal dominance over others, and then
institutionalized this dominance, it is crucial to safeguard
against such abuses via proper governance. Moreover,
with the advent of massive socio-technical systems, the
prospective danger of such institutionalized dominance
via the exploit of cognitive biases is abundantly clear and
present. Cultures may exhibit differences regarding their
respective tolerance for injustice; but eventually, social
elements rebel and overthrow nodes of injustice. This, of
course, is a costly process of redress that a proper
governance model can help address upfront. Thus, if
properly designed, elements of governance can help
safeguard against acts of injustice at the institutional level.
Governance operates at two levels: the agent and the
institutional. Governance at the agent level primarily
needs to safeguard against bounded rationality;
governance at the institutional level needs to safeguard
against institutionalized injustice. As discussed in Section
7 on CAS, these two levels of governance bifurcate into
the α and β levels of governance respectively. Given the
apparent differences in size and operational scale, making
course-corrections at the institutional level is far more
demanding than the same at the agent level. It is akin to
maneuvering and changing course of a massive oil tanker
versus an agile sports-car. Even so, and as we shall see,
the kernel design for both of these governance systems
may be obtained via a surprisingly similar lower-triangle
arrangement of heterarchical hierarchies. This is because
in both cases, governance ultimately does reduce to the
problem of unknown-unknowns in the context of
knowledge architectures that operate either at the agent or
the institutional level. In either case, the addition of
heterarchic controls may, therefore, help solve the
governance issue.
In essence, this is the promise of the blockchain
revolution (i.e., the addition of decentralized heterarchic
controls). The blockchain technology initially was created
as a support system for bitcoin transactions. However, it is
now turning out to have far-reaching economy-wide
implications for conducting peer-to-peer transactions
absent any gatekeeper middlemen. Governance in a
knowledge-economy implies the establishment of
appropriate regulatory nodes that help keep open the flow
of information and knowledge with proper regard for
privacy and property-right concerns.
Note however that the design of a CAS system (that is
emergent, self-organizing and adaptive) is a step removed
from the traditional design of systems that are
predominantly non-self-organizing. Here the design is left
incomplete at a meta-level; the final stages are
orchestrated in a self-actualizing boot-strap from its
inchoate embryonic state to its fully actualized adult form.
The closest exemplars of self-organization may be found
in the realm of biological entities that have brought forth
their ever adaptive, ever-evolving designs via genetic trial
and errors spanning immense temporal expanses. Trapped
within the sparse coils of the DNA (which consists of
about 1.5 GB of DVD-sized data), one may witness the
essence of the Information Axiom operating in a self-
organizing context. Herein, the genetic code orchestrates
the embryonic self-articulation and development of a
complex living entity (consisting of about 150 zettabytes
of data and requiring about 30 Manhattan-size datacenters
to merely store) that can struggle, adapt and thrive in
heretofore novel and unknown environments with ever-
changing risks and opportunities. Biological as well as
blockchain-based socio-technical systems may be studied
under the rubric of CAS (Section 7) to help elicit issues of
decentralization, self-organization, emergence, etc.
Section 2 surveys the relevant literature on governance
as it relates to the blockchain technology. Section 3
highlights the issue of trust amongst trustless agents--
which fundamentally undergirds the blockchain way.
Section 4 studies the rise of complexity as a driving force
for creating new organizational structures. Section 5
summarizes research on organizational design. Section 6
presents the phenomenon of stigmergy and stigmergic
gearings that help structure a CAS. Section 7 describes the
CAS system; i.e., the basic as well as iterative. Section 8
establishes the concept of decentralization (that underpins
much of the blockchain technology) in terms of CAS.
Section 9 defines the role of governance. Section 10 looks
at the phenomenon of Emergence and Self-Organization
in the context of governance. Section 11 explores
3
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
heterarchies and hierarchies in the context of governance.
Section 12 helps reduce the problem of governance (both
agent level as well as institutional) to the heterarchical-
hierarchy of knowledge architectures. Section 13
examines governance regarding the unknown-unknowns.
Section 14 frames the Axiomatic Design framework for
CAS. Section 15 structures the basic blockchain design
from an axiomatic perspective. Section 16 extends this
design to include smart contracts. Section 17 discusses the
kernel blockchain governance design. Section 18
concludes and wraps up the current work.
2 Literature Review
Governance of digital assets takes on a whole new
meaning when almost anything and everything can be
tokenized and traded by quasi-anonymous agents which
include machines and IoT’s.
In [7], Campbell-Verduyn provides a broad-brush
introduction to the issue of global governance in the
context of the blockchain. Blockchain-based trading
platforms started showing the cracks in the operative
governance framework (or the lack of it) when enterprises
such as Mt. Gox and SilkRoad started surfacing the
underlying “deep and dark-web” side of the new
technology. The key issue raised is whether the
blockchain technology “gives rise to new governance
problems and pathologies? [7]” Every shift in the
technological front leaves many who are vulnerable to
new information asymmetries. If so, what is the role of
governance in mitigating these asymmetries while
allowing innovation to proceed forward? Hastily drawn
governance rules could very well kill the golden goose
that may just turn out to be immensely transformative and
liberating. Also, as the governance researchers clearly
understand, blindly subscribing to the Lawrence Lessig
formulation of “Code is the Law [8]” is an invitation to
enter a Hobbesian Leviathan monolith. It is therefore
critical to understand the foundational basis of human
trust; and how much of it could be replicated via a trusted
system of trustless agents?
The fundamental problem of interdisciplinary
complexity inherent in the blockchain ecosystem is
echoed in [9] wherein Lopp asserts that: "One challenge
to understanding bitcoin is that it is a multifaceted cross-
disciplinary system that is constantly evolving."
In [10], Ehrsam (co-founder of Coinbase) asserts the
strategic significance of blockchain governance in that it
could be “the largest determinant of our future trajectory
as a species” when it potentially gets used to bootstrap
powerful AI across the distributed landscape. Crossing
over to the humanities end, Ehrsam highlights that it is
governance that “keeps communities together and, in turn,
gives a token value.” He further argues for on-chain
governance (i.e., code is law) in terms of its consistency,
fairness and speed of decision making. He does, however,
caution that the Leviathan metasystem could easily get
exploited if flaws were to be discovered; also, that it
becomes “harder to change once instituted.”
As a direct rebuttal to Ehrsam’s stand on the costs and
benefits of on-chain governance, Zamfir (lead developer
at Ethereum’s Casper protocol) asserts in [11] that
blockchain governance (as well as governance in general)
cannot “be understood as a design problem.” The reason
suggested as to why governance falls outside the purview
of design is because governance is a process, and
processes presumably fall outside the scope of design on
account of the dynamics involved. This, of course, is an
untenable position in favor of adhocracy; processes can
and should be subject to design. Nevertheless, Zamfir’s
highlighting of the need for adaptiveness when designing
governance structures is on target. Furthermore, it
dovetails well with the adaptiveness embedded in a CAS
architecture. Zamfir also takes issue with Ehrsam’s stand
on on-chain governance as being “incredibly risky” in
inviting automatic upgrades of the governance processes
without adequate human due-process and oversight.
Again, a proper design of the governance kernel ought to
make clear the appropriate contexts where one may resort
to on-chain versus off-chain governance.
In [12], DuPont forensically analyses the governance
failure in the Ethereum based DAO (Decentralized
Autonomous Organization). The DAO promised
transparency, efficiency, fairness and a democratic
decision-making process. In just a month, it managed to
raise USD 250 Million; yet within days of its launch, it
suffered a massive “attack” (draining it off USD 35
million) from which it never recovered. The study
exposed the “inherent complexity of bringing to life an
algorithmic and experimental organizational model.”
After the attack there were three post-attack options
presented:
• Code is Law (on-chain): Let the attack stand.
Attacker gains USD 35 million
• Soft Fork (off-chain): Let USD 35 million vanish.
• Hard Fork (off-chain): Return the funds to the
“investors” who willingly participated in the
trading platform but felt taken advantage of. The
attacker lost USD 35 million.
Ultimately the DAO that was supposed to be hands-off
and accepting of the “Code is Law” dictum, violated its
own governance and did a hard-fork by reverting to
human governance.
Based on the DAO failure, Voshmgir highlights the
problem of the Unknown-Unknowns in [13]: While
machine consensus can radically reduce bureaucracy, the
question of how to deal with unknown unknowns that
manifest over time has not yet been resolved.
3. Trust: Sed Quis Custodiet Ipsos
Custodes?
At its root, all socio-technical systems either rely on
trust or make costly allowances for a potential breach of
trust. With social trust in place, one can tentatively
foresee, plan and design our socio-technical engagements.
Trust is the underlying basis for designing socio-technical
artifacts. Society pays an enormous overhead for securing
4
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
trust. Traditionally, trust-systems (both internally as well
as externally) are hierarchically orchestrated. Internal
hierarchies preside over breaches of trust within an
organization. External hierarchies preside over breaches
that cross organizational boundaries. Hierarchic
governance structures are inherently flawed in the sense
that the top nodes can be compromised over time. Hence
Lord Acton's dictum in [14] that “power tends to corrupt,
and absolute power corrupts absolutely.” This same
sentiment was expressed two millenniums ago by the
Roman poet Juvenal when he coined the phrase [15]: Sed
quis custodiet ipsos custodes? (i.e., but who will guard the
guardians?). Going further back into antiquity, similar
suggestions may be discerned in Plato’s The Republic.
Having established an authoritarian top-down hierarchical
social order, Socrates is here first shown to raise the
problem of corruption at the top nodes; he then appeals to
self-governance in order to help resolve this conflict; i.e.,
the self-indulgent notion that the “best man has within
himself the divine governing principle” [16]. However,
within The Republic itself, there are passages [17] that
indicate that Plato looked down upon such a self-
deceptive, self-referential design construct as a
contemptable “noble lie,” (γενναῖον ψεῦδος: i.e., a lie or
wrong opinion about the true origins).
Fundamentally, there is no escaping the fact that all
hierarchical systems lack sufficient design parameters to
help resolve the problem of the top-nodes going rogue. It
was founding father, James Madison who first recognized
in [18] the fatal flaw in a purely hierarchical design
(symbolized as |H henceforth and as in [19]). Instead he
proposed the now institutionalized heterarchic governance
(symbolized as |h henceforth and as in [19]) within the US
federal government; i.e., a model of checks and balances
along with a clear separation of powers across three co-
equal hierarchies in order to help make sure that the top
nodes always have external oversight. It, therefore, may
be argued that the ongoing US experiment in self-
governance is indeed a lower-triangle decoupled design as
the coupling via self-referentiality so evident in the
Platonic logic has now been eliminated.
The blockchain model is likewise potentially
revolutionary in scope in that it has within it, the ability to
democratize and make ubiquitously available such
heterarchic controls across all levels of any given socio-
technical system; not just at the highest echelons of the US
Federal government.
4. Rising Socio-Technical Complexity
In [20], Prof. Bar-Yam suggests that society at large is
shifting away from deeply hierarchic models, and in favor
of more and more decentralized, heterarchic or mixed-
mode |h-|H control. This is because distributed
governance/control has a larger capacity for dealing with
increasing socio-technical complexity. One may witness
this in the academic realm, where interdisciplinarity is on
the rise; and (traditionally) hierarchical disciplinary group
boundaries are becoming porous.
Heterarchic linkages add extra burden in the realm of
governance for the simple reason that heterarchies do not
play nice; they instead jostle for dominance. Here,
governance involves sense-making across domains and
disciplines. And in order to be coherent and make sense,
one of the erstwhile co-equals eventually starts to
dominate the heterarchic complex. For example, in the
case of the US Federal government, historically we do
have three co-equal branches. However, over time, the
judiciary (given its ability to explicate the governance
narrative and create consistency across the total political
span) has carved out a dominant long-term role; the
executive (given its protagonist role in the near term)
similarly dominates the contemporary stage; and much of
the legislative branch stands reduced in stature The
executive and judiciary have usurped much of the law-
making ability of the legislative at both the consequential
coarse as well as the fine-grain. The judiciary is not per se
at fault; instead the default is in the purview of the other
two as they lack explicit grasp of a missing FR, namely
the need for historical consistency similar to judicial
review. This is indeed a flaw in the founding design; for
there is a hidden sense-making functional requirement that
has been left unaddressed. Each of the co-equal branches
ought to have had an ongoing sense-making role. Contrary
to Emerson [21], consistency is not “the hobgoblin of little
minds”; it in fact is what provides the directive stigmergic
thrust (Section 6). Court precedents are not easily
overturned; established case law sets the stage for what
follows. Such binding sense-making is missing in the
other two co-equal branches.
Sense-making is intimately related to the nature and
shape of human knowledge. As discussed in Section 12,
human knowledge is heterarchically hierarchical. Human
knowledge dynamics, therefore, have a key role in the
evolution of governance. Governance ultimately refers to
the over-arching body of knowledge that provides
guidance wherever conflicts arise. While heterarchical
contributions enrich the growing corpus, it is the
hierarchical aspect of human knowledge that is
responsible for sense-making. Sense-making is
fundamentally hierarchical in nature. Absent the distilling
of such knowledge hierarchies, information fails to make
sense. Larger the number of agents engaged in abstract,
high-order sense-making, greater the chance that the
engagement will devolve into heterarchic nonsense. Such
is the fundamental weakness that the legislative faces.
While it is heterarchically able to bring multiple points of
views to the governance table, it is unable to integrate this
into coherent hierarchically-sound agreements. Here, for
example, is a report [22] on the attempt to merely
determine the number of federal criminal laws at hand:
“In 1982, while at the Justice Department, Mr.
Gainer oversaw what still stands as the most
comprehensive attempt to tote up a number. The
effort came as part of a long and ultimately failed
campaign to persuade Congress to revise the
criminal code, which by the 1980s was scattered
among 50 titles and 23,000 pages of federal law.
5
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Justice Department lawyers undertook "the
laborious counting" of the scattered statutes "for the
express purpose of exposing the idiocy" of the
system, said Mr. Gainer.”
Consequences of such accountability failures in the
governance corpus are dire. For example, the above report
summarizes the legislative failure on sense-making with
the following comment [22] by law professor Prof. John
Baker: “There is no one in the United States over the age
of 18 who cannot be indicted for some federal crime. That
is not an exaggeration.”
Sense-making is especially relevant in the modern
context of vastly expanded and heterarchically-rich, socio-
technical systems such as the coming world of IoT's where
machines directly transact with other machines at an
unprecedented scale. The speed and scale of modern
socio-technical operations make it abundantly clear that
humans are increasingly left out of the decision loop as it
is beyond our human comprehension. Fundamentally,
humans are conceptual entities. The new era of
human/machine symbiotics [23] we are now entering,
ultimately has to make sense. Governance is
fundamentally rooted in sense-making. It is this that is at
stake when dealing with technologies of trust across the
human/machine divide. And unless we are careful, just as
the case was with the judiciary dominating the sense-
making role and by default, taking up the “first amongst
equals” position, machines may likewise be deliberately
or accidentally programmed to serenade us with
convincing but deceptive “noble lies” that exploit our
individual and collective cognitive biases & weaknesses.
When cast in this sense, the design of an equitable
governance structure for the coming blockchain way of
organizing our socio-technical systems may prove to be of
existential import.
It is worth noting here that Cynefin [24-25] is also
about sense-making in an interdisciplinary setting. Indeed,
as the Welsh word Cynefin suggests, the emphasis is on
multiple-belongings, i.e., multiple domains that mash-up
to create the unwieldiness of modern complexity.
5. Complex Socio-Technical
Organizational Design
Prof. Banathy was a pioneer in advocating for a
disciplined design of social systems. He was of the
opinion that we are now squarely in the "postindustrial
information/knowledge era [3]." Consequently,
organizational designs that arose in the industrial machine
age does not scale given the exponential rise in the "speed,
intensity, and complexity of change." Compared to design
initiatives in other disciplines, he was painfully aware of
the lack of attention regarding social-systems design. Prof.
Banathy held that when considering social systems design,
there is a tangible "shift from product thinking to process
thinking [3]." In direct contrast to this view was
Ethereum's lead developer, Mr. Zamfir who dismissed
processes as being outside the purview of design.
Social-systems have an overarching “concern for
justice [3].” Ethics, therefore, plays a pre-eminent role
when designing social systems. The system ought to be
equitable and just to all parties concerned, be they central
or peripheral in the activities subsumed. In other words,
governance plays a central role in all social system
designs. The proper design of the governance unit for a
blockchain-based social system ought to help establish
clear boundaries that when breached may trigger
appropriate smart contracts and/or legal recourse.
Blockchain governance is therefore not independent and
free-floating outside the appropriate societal moral code.
What is, however, a departure from the norm is the fact
that the blockchain based governance structures can
adjudicate many of the conflict scenarios with machine-
like precision and efficiency. It is as if vast parts of the
social order may now be governed by benevolent
arbitration judges that reside in the form of smart contracts
coded up within the machine.
Social system designers are often faced with problems
that are "anything but well defined [3]." Such problems
may have inconsistent FR's, inconsistent constraint-sets,
upstream designs that are coupled, problem-space that is
dynamic across time, etc. This is to be expected since we
are dealing with inconsistent ontologies across far-flung
inter-disciplinary fronts. While keeping the holistic view,
design may, therefore, have to proceed in small iterative
steps across the FR-DP divide. When inconsistencies are
detected, this may lead to repeat back-tracks up the design
hierarchy.
Social systems are primarily designed towards the
benevolent nurture, growth, and development of the
human potential. A properly designed governance unit
helps adjudicate equity throughout the system--both in its
homeostatic phase as well when the system adapts and
evolves beyond its stable stage; i.e., “self-organization
incorporates self-transcendence, the creative reaching out
of a system beyond its boundaries [3].” CAS is capable of
modeling such shape-shifting behaviors. Díaz and Olaya
highlight the role of emergence in [4]:
“Human beings co-design the social systems that
they form, this is why those designs might be
intentional up to some point but they are also
emergent, dynamic, incomplete, unpredictable, self-
organizing, evolutionary and always ‘in the
making’.”
In 2014, Prof. Norman and others put forth the
DesignX framework [26] for tackling the design of
complex socio-technical systems. It had nine problem
categories within its scope. These categories include some
of the problem areas mentioned above such as cognitive-
biases, bounded-rationality, interdisciplinarity,
requirements & constraints that do not always cohere (but
can periodically or chaotically change), precedent designs
that are inherently coupled, non-linearity in the element-
to-element interactions, causality that operates across
multiple scales and long/unpredictable latencies. Armed
with these complexities, Prof. Norman critiqued the
Axiomatic Framework along the following lines [26]:
6
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
“With sociotechnical systems, it is seldom
possible to follow the Independence Axiom: two-way
or even n-way interdependencies are common.
Moreover, these interdependencies are often
unknown, discovered only after the fact.”
In other words, the design-matrix (that tracks FR-DP
couplings two-by-two), is inadequate when dealing with
FR-DP clusters that may not compose between them into
a static, well-integrated, 2-dimensional matrix. The
example referred to in [26] that illustrates this
phenomenon pertains to the design of the treatment
schedule in an elderly healthcare service where patients
often present multiple ailment complexes and severe side-
effects from earlier treatments. What usually starts out as
a single-organ failure quickly devolves into a chaotic
complex of treatment and care that spans multiple
specialties [26]:
“When patients have multiple chronic conditions,
a common occurrence in the elderly, there are
numerous different professionals involved in the
treatment, with complex interconnections among
them (including, in some cases, a lack of
communication). These problems defy easy
analysis.”
The above set of nine problem categories are some of
the fundamental design challenges of the modern world--
and there are no easy answers. However, easy answers
include the "muddling through" approach advocated in
the DesignX framework. This approach advocates small,
incremental steps that in principle, refuses to consider the
problem as a whole [26]:
“This approach requires a different design
philosophy than might be used when considering the
project as a whole. Now, the design must be
modular, with multiple small, relatively independent
parts, incremental changes that can be
implemented, and linkages that are designed for
flexibility.”
Indeed, if such a "muddling through" approach were to
be institutionalized in medical-care, it would be cause for
alarm. Furthermore, such an approach fails to take
advantage of some of the modern tools at our disposal,
such as stigmergy (Section 6), Axiomatic Design [27],
Cynefin [24,28], Agent Base Modeling (ABM) [29],
Data-Sciences [30] and others. Each of these approaches
attempts to learn workable heuristics that are holistic in
scope while also attempting to meet the current
expediency. These are a more responsible approach than
merely "muddling through." Indeed, the muddling-
through approach may be considered as being
unnecessarily defeatist in embracing of the adhocracy
philosophy of yesteryears--even as complexities abound.
Even so, Prof. Norman's critique about the inadequacy
of the design-matrix in tracking complex coupling
clusters is well taken and needs to be addressed. Indeed,
biological systems are highly coupled. In fact, an
uncoupled biological system may legitimately be
considered to be dead. Thus, while the design axioms
continue to inspire (i.e., the FR-DP mapping could be
considered as "form follows function" in the biological
realm), the tools used to implement the axiomatic design
framework may need to be extended.
For example, the AD/CT extension [31] recognizes the
time-dependence of a given design. In other words, the
design matrix (along with its couplings) are not static and
unchanging across all the operational phases that a given
design is faced with. It is in fact, time-dependent. In this
expanded sense, the overall design is an ensemble of
appropriately governed designs that are either pre-set or
just-in-time improvisations composed of known
elements. The underlying time-dependence may be
periodic or aperiodic. AD/CT can help streamline and
resolve many of the objections raised by Prof. Norman.
6. Governance
Governance has taken center-stage on account of at
least two major global trends, and often working at cross
purposes [32]:
1. Globalization
2. Democratization
Globalization as the top-down legacy framework is
being challenged by the bottom-up democratic fervor that
has swept across the world stage ever since ubiquitous
smart-phones and blogs brought about an overthrow of the
highly scripted and staged “noble lie.” Now added into
the mix is the promise of the blockchain technology that
threatens to flatten and shorten the existing value-chains,
end-to-end. Consequently, a traditional firm now faces
competitive and regulatory challenges from multiple
dimensions. A governance misstep in any of the exposed
fronts may have serious consequences.
Governance is about collective decision-making, and
may be defined as in [32] as follows:
“Governance is about the rules of collective
decision-making in settings where there are a plurality
of actors or organisations and where no formal
control system can dictate the terms of the relationship
between these actors and organisations.”
Note the emphasis on:
1. Rules
2. The collective scope
3. The decision-making process, and the
4. Lack of formal control systems
While formal rules may easily get coded in smart
contracts, informal, on-the-fly, negotiated rules are much
harder to codify. Also, the collective scope squarely places
modern-day governance in the unwieldy heterarchic side
of the ledger as opposed to the well-behaved hierarchic
side that may be safely encoded in smart contracts. Also,
the decision-making process itself has to have its own
slowly-changing meta-rules as to “who can decide what,
and how decision-makers are to be made accountable
[32].” But the most challenging aspect of modern-day
governance is the realization that really “no one is in
charge”; i.e., “no formal control system can dictate the
relationships and outcomes”. And as we shall see (in
sections 7-8), this aspect of “no formal control” is what
7
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
makes modern governance a CAS problem. The
provenance of decentralized control may be traced at least
as far back as the writing of the Old Testament [33]:
“Go to the ant, you sluggard; consider its ways
and be wise! It has no commander, no overseer or
ruler, yet it stores its provisions in summer and
gathers its food at harvest.”
The lack of formal control mechanisms is a distinct
departure from traditional models of top-down
governance.
7. Stigmergy
So how is it that non-conceptual entities like ants, that
even though lacking a central controlling agent, are still
able to coordinate and collaborate in vast numbers (i.e., in
billions [34])? The answer to this is stigmergy.
Etymologically it is of Greek origin (stigm-oi meaning
pricking, signing, marking; and erg-on meaning work),
while entomologically it is from a study in 1959 by P.P
Grasse on termites [35]:
“The stimulation of the workers by the very
performances they have achieved is a significant one
inducing accurate and adaptable response, and has
been named stigmergy.”
Stigmergy denotes call to work based on local signs or
markings left by self or other agents at some time in the
past and during the course of their work (either as a side-
effect of the said work or as something in addition to the
work). These markings aggregate to provide
organizational directives available at various levels, both
within the environment as well as within and between
agents, thus leading to the visual of stigmergic gearings
(Fig. 1). Thus, even though there is no one controlling,
there is nevertheless system-wide control. These gear-
trains may or may not all engage simultaneously; instead,
they may be asynchronously meshed in different
groupings as per some meta-level (αi-βj) logic.
Examples of stigmergy abound in nature. For example,
the pheromone markings that an agent ant (αi) leaves
behind as it navigates an unknown terrain helps it to
navigate back home instead of being lost. Moreover, if
perchance, it does chance upon a choice food item, these
same trails then help recruit other ant compatriots in
jointly squirreling away the find back to the nest (Fig. 2).
The pheromone trails that aggregate across the
environment is the emergent pattern (βj). Gearing
upwards, the pattern-making potential (i.e., the chance
ability to create, sense and communicate asynchronously
via pheromones), must have had to be evolutionarily
written into the genetic constitution of the ant or its
predecessors (at some remote point in the past).
Stigmergy is thus a two-way street; it is not just that the
agent is leaving tell-tale markings in the environment; the
environment is also signing back but at a much more
glacial gearing pace. This captures the A and the Ω, but
there could be many more engagements across the span
that could veer off laterally.
Fig. 1. Stigmergic gearings
Fig. 2. Stigmergic trails (using NetLogo [36])
Thus, these gear-trains are not just linearly laid-out;
instead, they constitute fractal-networks, and networks-of-
networks that branch off and engage other such gearings.
Governance is the two-way aggregate effect of the multi-
faceted β-level gear-train network on the evolving α-level
agent body, and vice-versa.
Stigmergic gear trains may work to enhance or inhibit a
given agent/group-level activity, thus leading to non-
linear effects [37]:
“In more complex self-organising systems, there
will be several interlocking positive and negative
feedback loops, so that changes in some directions
are amplified while changes in other directions are
suppressed.”
Also, these stigmergic gear-trains are a direct analog to
the modern blockchain; except that nature displays
abundant varieties of these offerings, each of which has
painstakingly been forged in its exacting survival-of-the-
fittest workshop. One such example is the immune system
which we discuss in Section 15. Blockchain designers
would be wise to study similar nature-inspired chains.
The two-way nature of stigmergic-governance (as
mentioned above) does not thereby imply any additional
agency embedded in the environment; but that the
environment is also signing back (in an “action/reaction”
Newtonian sense) and thus shaping the evolution of the
protagonist agent. Longer the stigmergic chain, greater the
need to unitize and embed the ricochet as second nature
within the agent. Emergence is the global precipitation of
these stigmergic patterns into the environment; while
submergence is the local embedding of the unitized write-
back (i.e., the ricochet) into the constitution of the agent
for facilitating future emergence.
As illustrated above, stigmergy helps in the social
organization of lower level life forms such as ants and
termites. However, stigmergic ordering is not just limited
to the lower forms; indeed, much of human organization
(or the lack of it) may be attributed to stigmergic successes
8
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
and failures. For example, the organizing market power
captured in Adam Smith’s “invisible hand” metaphor may
be attributed to stigmergy [5]:
“Adam Smith’s ‘‘invisible hand’’ metaphor used to
denote the unintended emergent consequences of a
multiplicity of individuals’ actions, is stigmergic in all
but name…”
Indeed, the price of goods are the pheromone markings
that helps organize the vast reaches of our global economy
without explicit direction (i.e., the invisible in the
“invisible hand” metaphor). Stigmergy operates as a
problem-solving coordination mechanism wherever living
entities are faced with problems that are beyond their
limited individual ken. Billions of ant’s and other insects
wouldn’t be able to coordinate and thrive but for their
stigmergic know-how. It is therefore not surprising that we
humans have also been engaging in stigmergic rituals
without explicitly knowing it. Parunak analyses a whole
slew of such human-level stigmergic processes in [38],
including forest trail-formation, highway traffic-flows,
democratic elections, document editing, social-media
groupings, viral-marketing, Google page-ranks, peer-to-
peer computing, Amazon-style recommender-systems,
etc. The blockchain is yet another stigmergic innovation
to help coordinate human (as well as human-machine)
activities. In each of these systems, the patterns that
emerge have significant potential to help organize and
scale the human potential.
The study of emergence spans many domains including
economics [39], biology [40], sociology [41],
mechatronics [42], nanotechnology [43], spatial
computing [44], philosophy [45] etc.
From a designer’s viewpoint, stigmergy is critical in
learning to read nature without stumbling on “intelligent
design [46].” It is crucial for understanding and
deciphering designs in nature towards creating and
validating an integrated perspective that spans across the
natural as well as the artificial. It is the causal thread that
connects function and form; i.e., the Functional
Requirement (FR) to its Design Parameter (DP) in the
natural world. Given the scale, scope, immense time
frames as well as the vast combinatorial sweep across
which nature operates, it behooves us as keen students of
design, to perk up and listen. Deciphering the submerged
building blocks would render much of the biological order
transparent and seamless across the artificial/natural
divide. This is of significance given that while throughout
20th century, physics was the dominant science, the 21st
century is the century of biology. And stigmergy can help
bridge the conceptual bridge across these vastly different
sciences. But that requires carefully mapping the
underlying gear-train. Or in the words of Francis Bacon:
“nature, to be commanded, must be obeyed [47].” In this
context, the axiomatic framework could be fruitfully
employed in tracking the myriad gear-trains that nature
employs to keep its machinery fine-tuned and humming.
For example, from a design-matrix perspective, the
aforementioned +/- feedback effects across the gear-train
complex could be qualitatively captured in a matrix as
shown in Fig. 3 below. Here the design matrix captures the
delivery of the FR along the diagonal (denoted X), as well
as its off diagonal +/- control-set gearings to help keep the
main FR-DP (along the diagonal) on its track. And moving
across the A-Ω gear-train spectrum, the aforementioned
unitization would result in the familiar design hierarchy
that helps drill down from the macro-view and into the
micro.
Fig. 3. Design Matrix of Stigmergic Gearings/Hierarchy
Unlike direct and imperial command and control in the
human realm, stigmergic command is far subtler, indirect,
distributed and democratic. This is because there is no
master gear; no single point of control; instead there are
very many stigmergic gears distributed across the
unwieldly control landscape. Indirection implies
asynchrony as well as a lack of specificity as to who the
recipient of the message is. Needless to say, stigmergy is
thus intimately connected with governance as it is via
stigmergic gearings that the individuals in a group as well
as the group as a whole is able to orchestrate a global order
despite having “no commander, no overseer or ruler
[33].”
Note that the stigmergic signal is not a clear explicit
imperative command for the next agent to do something
or the other. Instead it is like a mark left on a shared
common blackboard for the next agent that comes by, to
use it as it pleases. For example, the openly displayed
stigmergic signal may very well be read by an adversary
and then put to nefarious uses against the original agent
and/or its group. Stigmergic signals are therefore not an
imperative command of what should be done next; instead
it is a recording of what has transpired up to the point when
the mark was made. The imperative element instead
resides in the collective emergences as well as in the agent
that picks up the baton. Thus, from an agent modeling
perspective, the "what next" is probabilistic.
In an abstract sense, markets are primarily an
expression of the human quest for freedom (via division
of labor); i.e., freedom to make better use of our scant
resources, especially time. The blockchain technology
helps escalate this timeless quest as it unblocks and frees
up the agents towards engaging in many more degrees of
freedom and ad-hoc mashups than previously imaginable.
If prior to the advent of the blockchain, the markets were
operating along highly choreographed pathways; after the
advent of the blockchain, the markets have now entered a
world of jazz-like improvisations that have the potential to
topple many of the strait-jacketed middle-men controlled
pathways. With the elimination of superfluous
middlemen, not just organizations; whole industries may
be flattened.
9
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Note that lock-in is a problem that stigmergic systems
(such as cryptocurrencies) face. As suggested in [48]:
“…path-based idiosyncrasies may become
locked in as material artifacts, institutions,
notations, measuring tools, and cultural practices.”
Under lock-in, stigmergic systems are unable to fork
away from the dominant strand on account of lack of
followership. The first movers therefore have a strategic
advantage which is not easily overcome. In the natural
world diseases, viruses and bacteria are the heterarchical
pathways that nature exploits in order to constantly stress-
test the resilience/viability of the total evolving biomass
away from stagnant lock-in. Likewise, in the blockchain
context one may expect similar heterarchic thrusts and
parries across the unguarded/evolving attack surfaces.
While stigmergic-coordination is remarkable in its
scale and scope, it is not the same as cognitive thought and
reasoning. Attempts to anthropomorphize stigmergy and
posit the existence of an "extended mind" is in error. There
is no "extended mind" agency, no stigmergic cognition,
and therefore no basis for stigmergic epistemology as in
[5]. Hypothetically speaking, if the extended mind exists
in a distributed, asynchronous sense, it must incarnate
whenever an agent partakes or contributes to the growing
stigmergic corpus. It is then like the luminance of the
lightning bug--it comes and goes out of existence. Even
so, the fundamental problem is that of will. No executive
center animates the extended mind figment. All will,
action, and responsibility remain vested in the underlying
agents. This restriction has jurisprudential implications for
the blockchain enterprise. Legally speaking, one cannot
litigate against the emergent βj-pattern; one may only sue
the αi-agents either individually or collectively [49]. The
extended-mind concept is a flawed concept; it serves no
rational purpose. Thinking along this line could place the
blockchain founders in legal jeopardy.
Similar restrictions also apply to humans operating
stigmergically as a group. There is no organ that can be
posited as a repository of group cognition. Stigmergic
epistemology [5] is, therefore, an oxymoron.
Anthropomorphically positing otherwise is an error. Same
is true for the rest of the philosophical train (i.e., there is
no validity to stigmergic metaphysics, ethics, politics,
aesthetics, etc.). However, what can be studied is the
validation of stigmergically arrived conclusions via
cognitive means resident in independent individual
entities. Thus, the philosophic underpinnings of
stigmergically arrived truths revert to normal philosophy.
There can, therefore, be no stigmergic validation of design
or governance. Thus, no matter how good a stigmergically
arrived design may be, it still needs independent analysis
and validation using the normal tools of human reasoning.
Conceptual knowledge therefore has dominance.
8. Complex Adaptive Systems (CAS)
Professor John H. Holland (1929-2015) is rightly
considered the father of genetic algorithms. He also laid
the foundational work in the study of CAS. As he has
described it [50], CAS’s “are systems that have a large
number of components, often called agents that interact
and adapt or learn.” Holland proposed a two-tiered
system as shown in Fig. 4a below. The lower α-tier
follows a fast-dynamic and is engaged in the flow of
resources between diverse agents (αi grouped in level i)
that are also leaving behind stigmergic markings; while
the upper β-tier follows a slower-dynamic that captures
and aggregates the stigmergic markings into emergent
patterns (βj grouped in level j), which is then emitted
system-wide as stigmergic signals that help the governed
agents to self-organize and scale.
Fig. 4. Complex Adaptive System: Basic vs. Iterative
Note that for the sake of simplicity, all the agents in
Fig. 4(a) have been placed homogeneously in the lower
tier, while all the emergent artifacts have been placed
homogenously in the upper tier. Such a simplification does
not quite capture the aforementioned gear-train logic.
Instead, what may be happening is that each follow-on
feedback-loop/iteration is bifurcating the target
population into higher levels of organizational
complexity. In each subsequent iteration, the population is
now composed of bifurcated ensembles of agent-nodes
and artifacts (as indicated by the dotted-ovals in Fig. 4(b)).
Gearing, therefore, iteratively creates self-organization
and structure in both of these interacting entity spaces. In
each follow-on iteration, the respective number of nodes
in each of these dotted-ovals is asymptotically decreasing
(with allowance for population dynamics) while the
dotted-ovals proliferate. Agents may, of course, migrate
across these boundaries.
While it is true that the β-tier is entrusted with the
governance mandate (i.e., the role of control and
governance) of a vast network of unwieldy, decentralized
agents that populate the α-tier, it is likewise, not immune
from further restructuring (i.e., higher order gearings).
In his post "Notes on Blockchain Governance [51],"
Buterin (who founded Ethereum) gives partial evidence of
the α-β CAS structuring in the blockchain architecture
when he writes:
“Generally speaking, there are two informal
models of governance, that I will call the “decision
function” view of governance and the “coordination”
view of governance.”
The coordination view is the β-tier view, while the
decision-function view pertains to the α-tier agents that act
on the coordination signals. Also, as mentioned earlier in
Section 7, it is the decision-function view that has legal
liability.
10
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Buterin also gives evidence about the layering (at least
in the β-tier) when he writes that "the coordination model
of governance…exists in layers [51]."
9. Centralization/Decentralization & CAS
Here we explore centralization vs. decentralization in
the context of CAS. In [52], Paul Baran was the first to
outline the distinction (see Fig. 5) between the two:
“Although one can draw a wide variety of
networks, they all factor into two components:
centralized (or star) and distributed (or grid or mesh).
The centralized network is obviously vulnerable as
destruction of a single central node destroys
communication between the end stations. In practice,
a mixture of star and mesh components is used to form
communication networks.”
Fig. 5. Distributed vs. Decentralized from a CAS Perspective
When cast in the CAS-framework, it is clear that
centralized vs. decentralized is relative. Absent the β-tier,
α-tier agents are distributed (Fig. 5a). It is only in the
governance context of the β-tier that the α-agents may be
considered as centralized (Fig. 5b) or decentralized (Fig.
5c). Also, there is a natural progression across these three
self-organizing architectures. Extending this natural
progression, one could easily see that there are far more
architectures beyond the above mentioned three. Indeed,
the full rigor of network sciences may be put to use in
extending the categories. For example, the mammalian
nervous system is a hybrid architecture that incorporates
both centralized as well as distributed control. The
vocabulary, therefore, needs to be enriched with
mathematical rigor. Also, note the loss-factor approach
being used in defining these critical concepts; i.e., the
“centralized network is obviously vulnerable as
destruction of a single central node destroys
communication between the end stations.” While such
vulnerability duly needs to be noted, it, unfortunately, fails
to capture the positive functions that the β-tier provides. It
is akin to saying that the stoppage of the heart muscle will
result in death which doesn’t quite describe the positive
function the heart muscle performs. This same loss-
function approach may be witnessed in the works of other
researchers who followed Baran’s approach.
Nevertheless, the above four preliminary concepts may be
summarized as follows:
• Distributed/Non-Distributed: Pertains to the
agent-spread in the α-tier, and across its various
bifurcations/groupings.
• Centralized/De-Centralized: Pertains to the
governance/control-logic in the β-tier (as
arranged across its various
bifurcations/groupings).
With the β-tier providing the controlling logic, all of
the α-agents would appear as centralized to the β-tier if
indeed there was agency embedded in β. But there is no
agency in the β-tier; hence usage of “control” as it pertains
to the β-tier is purely anthropomorphic.
Buterin categorizes centralization vs. decentralization
along the following three axes [53]:
• “Architectural (de)centralization: how many
agents is a system made up of? How many of these
agents may be lost, without loss of function?
• Political (de)centralization: how many agents
ultimately control all other agents/infrastructure
the system is made up of?
• Logical (de)centralization: how monolithic or
dispersed are the underlying data
structures/interfaces? How much of this
infrastructure may be lost, without loss of
function?”
In parsing the above classification, there is a level of
ambiguity that is clarified using the CAS-framework:
• Architectural (de)centralization:
o “How many agents is a system made up of?”
Agent population size pertains to the α-tier; this
does not directly pertain to the issue of
centralization vs. decentralization.
o “How many of these agents may be lost, without
loss of function?” This refers to system resiliency
and therefore does pertain to the governing β-tier.
However, Buterin does not explicate how merely
the counting of α-level agent’s, as well as the
fraction that may be lost, is thereby sufficient in
helping establish the system as being centralized
vs. decentralized. Indeed, until we look at the
controlling patterns that have been established in
the β-tier, it is impractical to say whether a given
system is architecturally centralized or
decentralized. The example that Buterin provides
is equally ambiguous: “traditional corporations
are architecturally centralized as it has just one
head office.” While this may sound plausible,
note that this has nothing to do with the number of
agents as well as the loss function. It is, therefore,
a non-sequitur.
• Political (de)centralization:
o “How many agents ultimately control all other
agents/infrastructure the system is made up of?”
While the element of control is salient in the
context of centralization/decentralization, it is an
error to think that the controlling logic of a system
11
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
(that often outlives the agent life-expectancy) is to
be found in such short-lived agent entities.
Consider the suggested example: “traditional
corporations are politically centralized (one
CEO).” This again is erroneous; the political
power is indeed being exercised by the transient
CEO agent; but the pattern of power assimilates in
the office of the CEO, which is a β-tier artifact that
outlives any given CEO. So, the question is not
regarding how many agents control the rest of the
organization; instead, it is whether the rules and
protocols vested in the office of the CEO help
establish a centralized or a decentralized
organization (which in the case of the traditional
corporation, is indeed centralized—but it doesn’t
have to be).
• Logical (de)centralization:
o “How monolithic or dispersed are the underlying
data structures/interfaces? How much of this
infrastructure may be lost without loss of
function?” Now, this is indeed part of the β-tier,
as it pertains to information flows and the patterns
around it. However, consider the example
suggested: “traditional corporations are logically
centralized (can’t really split them in half).” A
centralized corporate database, in and of itself
does not guarantee centralized control; indeed, a
decentralized organization could very well
harness a centralized corporate database. Instead,
the focus ought to be on the β-tier rules and
protocols that help establish centralized vs.
decentralized control.
Using the above ambiguous framework, Buterin
classifies the blockchain technology along the three axes
[53]:
“Blockchains are politically decentralized (no
one controls them) and architecturally
decentralized (no infrastructural central point of
failure) but they are logically centralized (there is
one commonly agreed state and the system behaves
like a single computer).”
The problem with Buterin’s classification scheme is
that it doesn’t address the heart of the decentralization
issue. Also, the three suggested axes are ad-hoc and could
be easily augmented; for example, they could very well
include economic (e.g., microfinance), aesthetic (e.g.,
decentralized control among jazz musicians), scientific
(e.g., the citizen science movement), etc.
Consider the legal implications of the above
misclassifications. If (as Buterin asserts) no one controls
the blockchain, then no one may be litigated against.
However, that is not what is happening in the real world.
For example, SilkRoad had six of its decentralized
servers tracked down, and its founder Ross Ulbricht,
arrested for money-laundering [54]. Ripple is currently
facing multiple class-action lawsuits (with CEO Bradley
Garlinghouse named as a defendant) claiming securities
law violation [55]. Similarly, Tezos and its founders face
multiple class-action lawsuits [56] for securities law
violation. The case against Tezos is significant as it was
designed as a meta-level operator that would smoothen
all future governance issues. However, because of poor
corporate governance structuring and the resultant fallout
between the founders, their ICO (Initial Coin Offering)
got stalled, resulting in the lawsuits [56]:
“One thing is clear though: there is a certain
irony in how Tezos, the cryptocurrency aiming to
solve governance issues on the blockchain, crashed
due to governance issues.”
From a legal as well as business point of view, it is
critical to understand the relative nature of centralization
versus decentralization. What appears decentralized for
agents at the αi-level (and below) is indeed centralized for
agents operating at the level of αi+1 (and above) and under
the direction of βi. That being the case, agents at the αi +1
(and above) are legally liable. Buterin is therefore in error
when he claims that “blockchains are politically
decentralized (no one controls them) [53].” In fact,
wherever two or more human agents engage, legal
disputes are certainly possible. Furthermore, litigation is
more than likely when necessary boundaries are left
unstated. In the blockchain context, disputes can occur
between agents at the same level, or across levels. It
serves no one any favors when the leadership tries to hide
the agency issue behind the decentralization veil
whenever disputes cross levels. True blockchain
leadership would be in setting up timely and appropriate
responsibilities, limitations and boundaries as the system
scales. Or as Robert Frost would wisely but reluctantly
suggest, “good fences make good neighbors [57].”
Srinivasan and Lee [58] have proposed a Lorenz
Curve/Gini Coefficient based framework to help quantify
the degree of decentralization in a given system. The Gini
coefficient spans the range of 0.0-1.0. The closer the
coefficient is to 1.0, more centralized the system. A
related concept, the Nakamoto Coefficient is also
reported in [58] that tracks the agent level thresholds that
tip the cumulative area under the Lorentz curve into 51%
control. From the CAS-perspective, the key insight is in
the fact that this framework suggests studying the
blockchain system as being composed of 6 essential
subsystems, namely:
1. Mining: by reward
2. Client: by codebase
3. Developers: by commits
4. Exchanges: by volume
5. Nodes: by country
6. Ownership: by addresses
This approach comes closer to the CAS-ideal as it
acknowledges the existence of a variety of controlling
gear-trains that need to be independently & jointly
tracked. Also, note that the metrics are focused on
stigmergic outputs (such as measuring the developer-
focused commit distribution). But lacking an integrated
framework, this approach is unable to combine the
Gini/Nakamoto subsystem measurements into a coherent
system-level measure; it is therefore forced to treat the
subsystems as stand-alone. Furthermore, the concept of
12
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
decentralization is far more generic than just the
blockchain context; for example, the above six
subsystems play no role when considering the degree of
decentralization in a corporate organization. Here, the
underlying bipartite α-β CAS machinery is what is
missing. By highlighting and referring to the generic CAS
machinery, we may be able to liberate the decentralization
concept to its rightful stature. Also, from a principled
design perspective, it is important to articulate the driving
functional requirements in the above endeavor; i.e., the
“why” we are looking for decentralization here vs.
centralization there. For example, taking a page from the
biological realm, there is a reason why parts of our
nervous system are under central control; while other
parts are under decentralized control. Blindly optimizing
along the decentralization ideal would miss out on these
hybrid architectures.
10. Emergence, Self-Organization and
Governance
In [37], Wolf and Holvoet differentiate between
emergence and self-organization. Referring to Fig. 4,
emergence is the upward moving arrow from α to β; while
self-organization is the downward pointing feedback loop
from β to α. These two flows can and often do occur
asynchronously. Long-winded asynchronous loops easily
confound the tracing of the causal structures. Governance
is predominantly associated with the downward β→α
command orchestrations, but it is equally important to
underscore the formative α→β pattern-captures. In this
sense, emergence ought to precede self-organization.
However, it is possible to graft foreign patterns and
artifacts onto an immature blockchain offering, resulting
in lack of coherence. All the key elements (including the
governance units) selected from the overall blockchain
ecosystem [59] needs to cohere within the evolving
context of a given blockchain community setting to make
a unique blockchain offering. Once the base structures
have materialized, every new emergence and its
corresponding self-organizational restructurings ought to
be appropriately governed. Given the speed, anonymity,
and heterarchically-hierarchical reach of the blockchain
based markets, traditional governance structures do not
seamlessly carry over. Governance artifacts ought to have
the same level of speed, anonymity-piercing and
heterarchically-hierarchical reach (on an as-needed basis)
that closely parallels any of the breaches across these
dimensions. Or to quote Callimachus of Cyrene, we have
got to "set a thief to catch a thief [60]," but now in real-
time.
11. Hierarchy, Heterarchy & Governance
Closely related to the conceptual pair of centralization-
decentralization is the conceptual pair of hierarchy-
heterarchy. We briefly considered this earlier when
discussing the issue of trust (Section 3) as well as the rise
of complexity (Section 4). For example, in Section 3 we
asserted that “it, therefore, may be argued that the ongoing
US experiment in self-governance is indeed a lower-
triangle decoupled design.” Such a design may be
depicted as shown in Fig. 6 below:
Fig. 6. Heterarchically Hierarchical (|h-|H) Governance
Heterarchic control of hierarchical systems (i.e., |h-|H
as depicted above) is an expensive proposition as it
demands a delicate balancing act of power-sharing
between competing hierarchies. It is therefore rarely used,
except in providing governance of the very top-most
nodes; and in the cases of national import. However, the
problem of “who will guard the guards themselves [15]”
occurs throughout the system; not just at the top nodes.
Indeed, Lord Acton’s insight that “power tends to corrupt,
and absolute power corrupts absolutely [14],” is
applicable across all nodes, (except maybe the lower-
most) in every socio-technical hierarchy. This is because
as a hierarchical system scales, it provides sufficient
latency for information flow, sufficient nooks and
crannies to bury the proverbial skeletons of misconduct. It
is similar to the distinction between local vs. global
maximum in the field of mathematical programming (Fig.
7). Thus, while there may be just one global optimum,
there may indeed be many local optima based on the local
settings. Likewise, in hierarchic (as well as in heterarchic)
organizations, there may be local as well as global top
nodes that may be compromised. Indeed, it may even be
asserted that it is the local top nodes that jostle to take on
the mantle of the global top node (Shakespeare’s Othello
vs. Iago being a case in point [61]). Hence it may be
crucial to nip the bud of corrosive power at the early local
stages before it scales and migrates over to the global slot.
Fig. 7. Problem of the Local vs. Global Top-Node
If properly designed, the blockchain approach could
effectively democratize and make available the above |h-
|H governance architecture across the board. This is why
establishing a proper governance model for the blockchain
approach has significant beneficial implications society-
wide. As discussed in Section 2, the fundamental problem
in democratizing and scaling up the governance kernel is
in clearly understanding when to use the machine vs. when
13
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
to apply human intervention; in other words, when is it
appropriate to use on-chain vs. off-chain governance vs. a
mixed setup?
How should one go about adding heterarchic controls
(via the blockchain technology) into traditional hierarchic
governance models, and across the board? Here, an
analogy may help in grasping the auxiliary evidence
scheme that the blockchain technology offers. Consider
the task that a particular lawyer is faced with, i.e., of
ascertaining the veracity of a given client or witness. The
task is to check if the client is telling the truth. One way is
to painstakingly check each statement, to check for
internal consistency and to independently validate it
against other bodies of evidence. However, there is yet
another way to check if the client is lying; and that is to
bring in a micro-expressions expert. Micro-expressions
[62] are fleeting (i.e., lasting less than half a second)
betrayals of inner conflict that the subject is incapable of
hiding or suppressing. The act of concealment is being
orchestrated by the pre-frontal cortex, which the amygdala
effectively short-circuits by leaking the subterfuge in an
involuntary micro-expression. In this sense, micro-
expressions are auxiliary evidence streams that may help
catch a lie. Using the blockchain technology is akin to
using micro-expressions to help adjudicate a conflicted
situation; it provides easily verifiable auxiliary streams of
evidence that may be available to anyone in public.
However, note that on its own merit, micro-expressions do
not reveal the factual basis of the conflicted case; only that
whatever the client is asserting has an element of
concealment in it. In this sense, it is preserving the client-
attorney privilege as far as the micro-expressions expert is
concerned. Likewise, the blockchain technology
cryptologically conceals the factual basis of a given
transaction; but it has the potential to reveal if that
transaction is conflicted with something prior that
happened and as recorded within the ledger the blockchain
controls. Such self-on-self is what makes current
blockchain governance hierarchical in nature. It, however,
does not have to be so. Similar to the US Federal
Government (Fig. 6), co-equal blockchain governance
structures may be designed to mesh heterarchically (at
both local as well as global top-nodes).
The problem of hierarchic governance was painfully
evident on June 17th, 2016, when the Ethereum based
DAO (Decentralized Autonomous Organization) suffered
the infamous DAO attack that legally exploited
weaknesses in its code-base [3]. Section 6 briefly
described Ethereum. It is a programmable, Turing-
complete blockchain infrastructure that can authenticate
and run code (in the form of smart contracts), not just keep
track of the underlying transactions. The DAO was built
on top of Ethereum as a decentralized, cryptocurrency-
based, crowd-funded platform where investors could
directly fund and manage new enterprises that would, in
turn, run on Ethereum. In a period of just one month, the
DAO was able to raise the equivalent of 250 million USD,
the largest crowdfunding success as of May 2016.
However, the DAO attack fundamentally crippled the
visionary zeal. Faced with dire losses (in the order of 35
million USD), the principals banded together in an ad-hoc
manner to perform a hard-fork; i.e., to violate their own
pre-established rules of conduct, to revert back to the
genesis state while simultaneously changing the rules of
operation to make it favorable to the majority. This is
indeed the ancient problem of governance at the top nodes
of an organizational hierarchy, be it human or technology-
based at the top-nodes. Merely handing the administration
of agreements between willing agents over to smart
contracts (i.e., a rule-based digital logic used to verify and
enforce an agreed upon contract between two or more
agents) does not obviate the top-node problem.
Hierarchically governed socio-technical designs are
fundamentally coupled on account of too few DP’s. To
understand how one may go about introducing elements
of heterarchic gear-train governors into a predominantly
(smart contract based) hierarchic mix, one has to delve
into the architecture of human knowledge alongside the
issue of the unknown-unknowns.
12. Heterarchically-Hierarchical
Knowledge and Governance
Earlier, in Section 4 we had asserted that governance
is intimately related to sense-making, which in turn is
related to the nature, shape, and dynamics of human
knowledge. It is by understanding the epistemological
roots of human knowledge that one may formulate the
proper division of labor between the human and the
machine (i.e., between off-chain and on-chain
governance). In other words, what is it that the human is
good at; likewise, what is it that the machine is good at?
Smart contracts are smart only to the extent that the
human ingenuity has embedded the smarts within them,
including the necessary smarts for knowledge dynamics
originating both within as well as outside one’s ken.
Given the abstract nature and spread of human
knowledge, it may be observed that knowledge has a
dynamical and heterarchically-hierarchical (|h-|H)
structure as shown in Figs 8.a-f below. This figure is
adapted from [19]. Concretes are far more numerous than
abstractions; this implies that domain-specific human
knowledge (Fig. 8b) has a conical/hierarchical shape.
Induction flows along an upward arch, while deduction
flows along a downward arch. Abductive cascades utilize
both inductive as well as deductive streams in problem-
solving (including designerly) situations [63]. These
distinctions ought to inform the on-going debate as to the
proper division-of-labour between humans and machines:
induction (that favours human faculties) versus deduction
(that favours the machine) ought to be the proper role
demarcation between the two sets of entities in any socio-
technical system. Call this demarcation the Inducto-
Deductive Front (IDF) shown as the dotted line in Fig. 8.a-
d. For abductive cascades (with the IDF at cascade apex),
both human and machine agents would need to work in
close symbiotic coordination [23,64].
14
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Fig. 8. Human knowledge as heterarchically hierarchical (|h-|H)
The rate of change in the knowledge corpus is more
pronounced along the lower rungs as compared to the
higher, abstract levels (Fig. 8a). Reverse-salients (Fig. 8c)
are lagging knowledge fronts (the known-unknowns) that
occur because of differentials in growth spurts across
domains that are close enough to make sense if conceptual
barriers didn’t exist. When they do gap-close, it ripples
across the knowledge fabric radially (i.e., hierarchically-
|H) as well as tangentially (i.e., heterarchically-|h).
Another source of knowledge dynamic is the archstand
[63]—an integrated external perspective such as the Non-
Euclidean framework that led to the Theory of Relativity.
When stand-alone domains are organized using domain
kinship metrics, one may expect these conics to exhibit a
self-similar fish-scale (hierarchically-heterarchic) fractal
structure (Fig. 8e). Humans are at the mesoscale.
Unknowns from the macro-world dominate the outer
realms; unknowns from the micro dominate the inner
regions. Knowledge is sandwiched between these two
outer and inner circles-of-ignorance that are expanding
and contracting respectively. Regions beyond are the
ultimate terra incognita; the vast unknown-unknowns.
Between hierarchies and heterarchies, hierarchies
exhibit relatively stable vertical linkages; whereas
heterarchies exhibit dynamic ties that are conceptual
mashups in the making. At finer grains, hierarchies may
contain heterarchies and vice-versa, and switch
dominance across time (Fig. 8f). Knowledge flux
involves the constant jostling between heterarchies and
hierarchies. Without hierarchies, higher-level heterarchies
do not form nor engage; without heterarchies, hierarchies
tend to become stale, iconoclastic and insular. The
emergence/flourishing of a discipline arises from
heterarchic assaults and hierarchic defenses; both forces
are necessary. Heterarchies encourage falsifiability while
hierarchies encourage verifiability; both are essential.
Therefore, in the context of governance, both of these
forces ought to be judiciously engaged. When
heterarchical assaults reach above the IDF, inductive
human ingenuity ought to be marshaled; in contrast, when
heterarchic assaults land below the IDF, the machines may
well be capable of handling the issue. Likewise, when
issues of verifiability range above the IDF, it would again
demand human ingenuity to overcome the default. But if
it occurs below the IDF line, the smart contract
infrastructure may be sufficient to handle it.
In Section 4 (on rising complexity) we had indicated
that there is an on-going phase shift away from deep-
hierarchies and into hybrid |h-|H systems with many ad-
hoc laterals. Interdisciplinarity is on the rise; and
traditional disciplines are heterarchically being cross-
pollinated. This has been discussed at length in [19]. One
of the progressive schemes (i.e., the Jantschian) is as
shown in Fig. 9 below. In CAS-terms, such a progression
is to be expected, given the upward gearing across α↔β.
Fig. 9. Terms of Interdisciplinarity (Jantschian) [19]
While the overall envelope of the unknown-unknowns
is as shown in the knowledge sandwich of Fig.8.e, there
are many nuances (such as the case of reverse-salients, i.e.,
known-unknowns) that need to be addressed. We turn to
the issue of unknown-unknowns next.
13. The Unknown-Unknowns and
Governance
Defense Secretary, Donald Rumsfeld popularized the
issue of the unknown unknowns [65]:
“Subject. What you know. There are known
knowns. There are known unknowns. There are
unknown unknowns. But there are also unknown
knowns. That is to say, things that you think you know
that turns out you did not.”
The problem is how to parse this with logical
consistency in mind.
15
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Fig. 10. The Unknown-Unknown Knowledge Asymmetry Exploit
For the sake of brevity, let us notate these options as KK
(Known-Known), KU (Known-Unknown), UU
(Unknown-Unknown) and UK (Unknown-Known).
Elsewhere [66] Rumsfeld also opined that:
“There are known knowns. These are things we
know that we know. There are known unknowns.
That is to say, there are things that we know we don’t
know. But there are also unknown unknowns. These
are things we don’t know we don’t know.”
This quote considers just three of the options: KK, KU,
and UU, with UK missing. The way Rumsfeld has parsed
the second option KU (i.e., things that we know we don’t
know) suggests that the first lettering is about our state of
confidence in our state of knowledge; and the second
lettering is about the base state of our knowledge. Thus,
the missing variant UK in the above formulation indicates
poor confidence in a given assertion that we accept.
Once we parse this base structure, it then becomes clear
that there are many more shades of the Unknown-
Unknowns that lurk in the shadows, especially when we
start considering issues of stigmergic knowledge as well
as what our adversaries likewise know. Understanding the
problem of the Unknown-Unknowns is central to
understanding how blockchain governance is likely to
evolve. For example, in analyzing the DAO debacle,
Voshmgir highlights the problem of the unknown-
unknowns faced by on-chain “codified governance
rulesets” (CGR’s) [13]:
“In reality, formalised and codified governance
rulesets can only depict known knowns and known
unknowns, but have very limited capabilities to
properly deal with unknown unknowns.”
The earliest formulation of one of the combinations may
have been by the poet John Keats in Endymion wherein
love-struck demi-god Endymion ponders the mysteries
that wrap the object of his affections, the moon: “O known
Unknown! from whom my being sips [67].” The realm of
the unknown-unknowns can inspire as well as frustrate
inquiry. Here Keats puts forth the idea that things of
beauty have both familiar as well as unknown facets; and
that we come to grasp the unknown by systematically
working our way to the edge of the known realms.
Indeed, there are many pathways into the realm of the
unknowns, not just the four that Rumsfeld put forth. One’s
true state of knowledge about some pertinent issue may be
cross-mapped against what the society-at-large (or your
adversary in a game-theoretic sense) is aware of. Also
relevant to the problem is how the knowledge being
claimed was arrived at; i.e., whether conceptually or
stigmergically? In the human context, stigmergic
knowledge gets coded in mores, heuristics and, habits of
individual thought and action (both at the individual as
well as at the societal level).
If it was conceptually arrived at, then it has a greater
chance of error; but if true, it has far-reaching potential to
scale. In contrast, if it was stigmergically arrived at, then
its basis may be stronger (provided it avoids the problem
of the aforementioned stigmergic lock-in); but being pre-
conceptual, it does not scale easily. It is therefore of
strategic value to convert stigmergic knowledge into the
conceptual realm.
So, what are the combinatorial possibilities that
populate the realm of the unknown-unknowns? Denote
you (or your teams) state of knowledge in small-caps.
Denote the state of knowledge of society-at-large (or
perhaps your adversary) in large-caps. The resultant
combinatorics may then be bifurcated along the following
dimensions:
• Process by which knowledge is gained: C/S
(Conceptual/Stigmergic) for {Society, Adversary}
vs. c/s (conceptual/stigmergic) for {you, team}
• Confidence in knowledge possessed: H/L
(High/Low) for {Society, Adversary} vs. h/l
(high/low) for {you, team}
• True status of knowledge possessed: T/F
(True/False) for {Society, Adversary} vs. t/f (true,
false) for {you, team}
Mapping the resultant combinations into the
known/unknown characterization is fairly straightforward
with stigmergically derived true and false states in small-
caps {k, u}; and conceptually derived true and false states
in large-caps {K, U}. Thus CHTcht(KK) would denote
both you as well as your adversary possessing
conceptually-derived knowledge that is of high-
16
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
confidence and happens to be true, leading to a situation
of conceptually derived known-knowns. When the
adversary’s knowledge is mapped against one’s own,
there are 64 combinations as shown in Fig. 10 above. Note
that the matrix assumes a two-player game structure,
though one or both players could represent coordinated
groups. Also note that the UK style coding (in parenthesis)
is different from the Rumsfeldian coding as it denotes two
opposing agents. Each new such player expands the
combinations by a multiple of 8. These combinations
create knowledge asymmetries (with comparative
advantage to the {K, k} team, if paired against a {U, u}
adversary) that are ripe for exploitation, thus triggering
governance. These asymmetries are, therefore, at the
foundation of the governance conundrum, that in its
essence, checks to see if agreed-upon knowledge flows
have been thwarted to result in the given asymmetry.
Consider for example the cell CHTshf (Ku) highlighted in
green in the top-right quadrant of Fig.10. Here the
adversary’s knowledge about some matter (say the true
worth of a smart contract) is conceptual, of high
confidence and true; in contrast, your knowledge about
that same matter is mere stigmergic hearsay, but of high
confidence and happens to be wrong. Diagonally across
from CHTshf (Ku) is the diametrically opposite case of
SHFcht (uK) (highlighted in red) where the asymmetry
now favors the individual actor as opposed to society at
large. Here the socially networked group is operating
stigmergically, has fatally high confidence in its findings
which in fact is wrong; in contrast, the lone operator is
operating conceptually, has high confidence in its findings
and is in fact right. In the world of finance, hedge-funds
try to exploit SHFcht (uK) types of knowledge
asymmetries. And when a smart contract is executed based
upon such asymmetries, there are bound to be outcries of
failures in governance. This indeed is what transpired in
the case of the DAO-attack.
Playing the role of the adversary, if indeed one wishes
to widen the {K, k}- {U, u} gap even further strategically,
it may be worth introducing Axiomatic
Design/Complexity Theory based complexing red-
herrings as suggested in [68]. This may be even more
potent when dealing with a {k}-{u} type knowledge gaps
(which happenstance is much of the operative human
knowledge); the reason being that it is challenging to
debug stigmergic linkages that have been deliberately
sabotaged for the explicit design purpose of throwing off
one’s adversaries.
14. Axiomatic Design for Complex
Adaptive Systems
The creative mashup between two diametrically
opposed design methodologies (i.e., the top-down
Axiomatic approach vs. the bottom-up Design Patterns
approach) was discussed in [69-70]. Thus, the top-down
V-approach was juxtaposed with the bottom-up Λ-
approach to create the N-model. As it turns out, the N-
model comports well with the Complex Adaptive Systems
framework. The design-patterns approach that leads with
the upward-stroke of Λ is akin to the α→β emergent stroke
in a CAS system; likewise, the axiomatic approach that
leads with the downward stroke of V is akin to the β→α
self-organization stroke in a CAS-system. Together they
compose to make the N-model which indeed is the overall
gearing dynamic behind the α↔β CAS system. However,
as mentioned earlier (in Section 1), the design of a CAS
System is a step removed from the traditional design of
systems that are predominantly non-emergent/non-self-
organizing. There is an inherent embryology of the CAS
system that the designer has to yield to; i.e., the CAS
designer needs to think more like a farmer rather than an
engineer and adjust to the vagaries of emergence, such as
that between pests and pollinators [71].
To come up with a design that is holistic and emergent
requires the designer to be steeped in the practice of
design; i.e., it is combinatorically challenging. Also,
emergence requires beneficial interaction between the
design elements, thus favoring lower-diagonal decoupled
vs. uncoupled designs, which is not the usual norm. In the
case of the diagonal design, the whole is equal to the sum
of the parts. Emergence, however, requires beneficial
interaction, which is feasible only if non-diagonal
elements are present. Thus, in most cases, the uncoupled
has dominance over the decoupled given the lower
informational complexity. Emergence is the rare
occurrence that could be flipping this dominance to
combinatorically win the race with lower information
content. This, however, is merely a hypothesis that needs
to be validated. Many of the biological systems (given the
enormous temporal-combinatorial space that they have
been stigmergically operating over and finessing the
information axiom) have strong elements of emergent
qualities (such as life, consciousness, everything that
pertains to the emotional faculties, etc.). Biological
designs present rich opportunities to test this hypothesis.
By adopting the axiomatic approach, the β→α design
is decomposed both
• laterally and non-hierarchically across the various
realms such as customer, functional, physical,
process (CR, FR, DP, PV, etc.) as shown in Fig. 11
below, and
• vertically and hierarchically within each of the
above realms.
In a rapidly evolving design context, it is impractical to
approach design in staged, linear waterfall fashion as in
CR↔FR↔DP↔PV. Instead, it is better modeled (as
shown in Fig. 11 below) as a fully linked network of
information nodes. The linear structuring still dominates,
but it is now augmented with auxiliary flows. Each of
these realms has its own α↔β CAS structure that forms
over time. Furthermore, since domain knowledge is
hierarchical, the design trace that leverages this
knowledge is likewise hierarchical.
By visualizing the design in the context of knowledge
hierarchies, one may begin to appreciate the historical
import of Prof. Suh's work [2]. In fact, something similar
17
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
(see Fig. 12) happened in Renaissance Italy around 1420,
with the invention of linear perspective [72] by the Italian
architect/artist Filippo Brunelleschi. Ancient Rome indeed
did have something close to linear perspective; however,
the ancients used multiple vanishing points in its
paintings, thus leaving a sense of lack of coherence in the
presentation. Brunelleschi did study the ancients. He then
came back to Florence to revolutionize the world of
representational art as we now know it. With a single
vanishing point, all the objects in the field of vision
compose in a realistic, coherent, eye-pleasing fashion.
Indeed, juxtaposing any of the art-works prior to
Brunelleschi's approach, one immediately senses the
flatness and lack of proportions in the former vs. the three-
dimensionality and compositionality in the later.
Fig. 11. Design Information Flow Network
Fig. 12. Perceptual vs. Conceptual Mashups
Likewise, Prof. Suh's work on axiomatic design is at
least of equal stature (if not more); for what Brunelleschi
stipulated in the realm of perceptuals, Prof. Suh has
stipulated in the realm of conceptuals. The ability to bring
unity and coherence in the realm of the conceptual artifact
space is monumental in scope, especially in the field of
education in general; and not just design education. Here,
the teaching of anatomy and physiology from a “Form
follows Function” perspective [73] is worth noting as it
offers significant insights into the potential scope. As was
the case for paintings prior to Brunelleschi’s perspective
drawing, much of education today is a sprawl that lacks
conceptual unity and coherence. This same sprawl is
evident in the blockchain realm [9]. Again, there is
untapped potential in modeling instructional sciences
along the biological template.
15. Basic Blockchain Design
The web has evolved from a sprawling network of
hyperlinks in the 1990s (Web1), to being programmable
(Web2) in the 2000's--thus enabling social media, e-
commerce, and other similar restructurings. These
restructurings allowed a few to scale upwards and enjoy
global reach.
So now we are on to the third phase; i.e., Web3. The
problem with Web2 was that (as was the case with the
Napster-model with its centralized set of index files), at
the center of many of the Web2 business models, there
exists a centralized database that amasses immense power
to structure and shepherd the flow of thought and
commerce. These models implicitly took advantage of
power-laws that favor the highly connected central nodes.
It is true that on the one hand, these Leviathans have
enabled tremendous productivity gains compared to what
existed prior; but on the other hand, they have de facto
established governance-in-stealth for all the peripheral
nodes. It is not to say that there is any malevolent element
in these designs; it is merely that anything so centralized
(as per Lord Acton’s dictum) will perforce be restrictive
towards the free-flow of thought and association.
Fig. 13. Basic Blockchain Design
To put things in perspective, every google page-rank
discriminates against those who create and search for the
road-less-traveled; for the average person rarely goes
beyond the first five search results [74]. In contrast, Web3
portends to be genuinely democratic by eliminating all
central nodes and associated intermediaries who currently
enjoy a high degree of betweenness-centrality [75] by
inserting themselves in between nodes that otherwise
would be P2P. Human/machine agents now genuinely can
18
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
engage P2P without the need for gatekeepers and
connectors. To put this in context, the ongoing revolution
portends the ability "to build ridesharing without Uber,
apartment sharing without Airbnb, and social media
without Facebook and Twitter [76]."
The top-level CR for the blockchain may be stated as
follows:
Need consensually-trusted, immutable,
distributed and decentrally-managed, verifiable,
publicly and efficiently searchable record of all
transactions (since genesis) that are private and
discreet, but stigmergically-marked for public-
viewing, that pertains to a given economic activity
and that may be made by adversarial/trust-less
agents.
When restated in the FR-DP framing, the above CR
translates into the basic blockchain design as shown in
Fig. 13 above. The design-matrix indicates a decoupled,
lower-triangle design. The couplings systematically build-
up across the FR list, top to bottom. For example, the FR:
Trustless-Trust (highlighted in red) is being delivered
using all the previous DP’s, along with the Consensus
Model DP.
In the BitCoin case, just as soon as consensus is
achieved, new tokens are released as a reward for the
successful miner who expended computational resources
to help bring about consensus. Thus, in this phase of token
creation, a small part of the overall design has one of the
design matrices in a different form which leads off with
the consensus model. This agrees with the extended
AD/CT (i.e., TDPC: Time-Dependent Periodic
Complexity). In the discussion below, we have chosen to
focus on just the broad design (as shown in Fig. 13 above)
and ignore all such finer variations.
The blockchain is a CAS system that fundamentally
operates on stigmergy. Discreet stigmergy requires agents
be allowed to mark their environment (here, the
blockchain ledger) discretely (i.e., without having to
reveal their respective true identities). This is similar to
ants leaving pheromone droppings—except that in the
blockchain context, the agents enjoy a certain degree of
anonymity. Cryptographic tokens (bitcoin, ether, gas, etc.)
are the cash-like pheromones that various stakeholders use
to engage in economic activities. They are cash-like in the
sense that they shield the privacy of the agents; but they
are stigmergic in the sense that the transaction now has a
permanent, publicly-viewable/traceable record. Thus,
when spent, the tokens leave their stigmergic markings
that if properly aggregated, could help evolve the system
forward. Cryptographic stigmergy [77] looks at the
stigmergic design of the overall system to help precipitate
direction-providing emergences and the corresponding
self-organization around it.
Transaction security is obtained via standard,
cryptologic hash functions such as SHA-256.
The Merkle tree data-structure that encodes all the
transactions in a given block is designed to help verify the
existence and validity of the growing chain of transactions
in a computationally efficient fashion (costing less than
O(Log2(N)) in space and time).
The paradoxical property of trustless-trust is obtained
via various consensus models (such as PoW: Proof-Of-
Work; PoS: Proof-Of-Stake, etc.). Cryptoeconomics [78]
studies the creation of economic incentives (such as
tokens allocated to miners for performing computationally
intensive work such as PoW) to bring about consensuses
in a distributed and potentially adversarial setup. For
example, the PoW model embedded in the BitCoin system
allows decision-making via consensus (via the Byzantine
Fault Tolerant algorithm [79]) despite approximately 1/3rd
of all agents going rogue. Blockchain offerings may be
differentiated using the consensus model differentiator
that power their respective answers to the problem of
securing Trustless Trust.
As a biological analog, the immune system [80] is a
perfect example of the blockchain along with its own
Proof-of-Work (called the fever) which is the
computational struggle that various cells from the immune
system go thru in order to recognize invading antigens as
a friend or a foe. Once identified, the body never forgets.
It keeps the evidence of all its successful struggles in its
growing ledger, i.e., its collection of antibodies.
The blockchain ledger is a growing linked list of
transaction records that have been bundled into timed
blocks. The chained ledger has been accumulating these
blocks ever since the genesis of the blockchain under
study. Each such timed block of transactions is bundled in
an efficient, easily verifiable data-structure such as the
Merkle tree. Each new addition contains in its header the
hash of the previous block. This makes both the nodes as
well as the overall chain highly resistant to modification.
Longer the chain grows, harder it is to break.
Finally, Network Resilience is obtained via the P2P
distributed protocols.
Satoshi Nakamoto designed the PoW based blockchain
for the P2P bitcoin cryptocurrency [81]. The contractual
logic that is embedded in the BitCoin blockchain may be
abstracted out and generalized to help secure trusted
transactions across the whole gamut of global economic
activity. The Ethereum project was the first to recognize
the value of such decoupling’s. While the Blockchain
provides the underlying infrastructure, it is what gets built
on top of it that defines the business offering. Each such
offering provides unique affordances targeting specific
business eco-systems. The rules and boundaries of these
eco-systems (along with their governance protocols) are
established via the logic of the smart contracts.
16. Smart Contract and Governance
Smart contracts are contracts written in code that will
execute when matching conditions that make up the
agreement are met. In other words, it is “cocked, locked
and ready to fire”; there are no off-ramps. Smart contracts
have been envisioned across multiple domains, including
crowdfunding, financials (buying and selling of
tangibles/intangibles, insurance, derivatives), legal, etc.
19
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Smart contracts are point-to-point with all middlemen
having been dis-intermediated. It, therefore, has
substantial potential to inflict losses in the hands of the
naïve, careless or uninitiated (i.e., the KU type
asymmetry).
The blockchain-based smart contract technology has the
potential to transform society as a whole for the better;
better in the sense of faster, cheaper and fairer
transactions. Thus, given the enormous potential to
smoothen the flow of commerce while bringing down
costs, it is incumbent on the designers of blockchain based
smart contracts to get the governance aspect done right. If
it is designed for scaling, it ought to cover for the variety
of knowledge asymmetries that exist across the spectrum
of participants as well as the dynamics along a fast-
moving front. Voshmgir emphasizes the pace with which
the context for a smart contract design could rapidly move
away from its original intent [82]:
“First use cases show that as circumstances
change, protocols can become inappropriate for the
new environment and require modification.”
In other words, it is not just that the design space is
rapidly evolving; here the FR’s themselves are rapidly
evolving; i.e., the half-life of any given FR is also rapidly
being cut short. This in itself is highly unprecedented in
the world of design. In other words, there is a high
premium for designing systems based on first principles as
compared to short-sighted pragmatics.
Since the smart contract offering sits on top of the
blockchain infrastructure, the CR for the design may be
stated as follows:
Need the ability to structure and verify auto-
executing contracts that incorporate arbitrarily
complex business rules and trade on a given
blockchain offering.
Fig. 14. Blockchain-Based Smart Contract Design
When restated in the FR-DP framing, the above CR
translates into the blockchain-based, smart contracting
infrastructure as shown in Fig. 14 above. The smart
contract is the structured, auto-executing contract. It is
neither a legal contract nor necessarily smart. Just as any
other trading instrument, the legality of the contract
needs to be ironed out in the appropriate legal setting.
The smart in the smart contract depends on how well the
coding reflects the underlying economic incentives. For
example, the DAO as a smart contract [12] was anything
but smart.
The blockchain infrastructure wouldn’t necessarily
have the necessary data/logic to verify if and when all the
preconditions specified in a given smart contract have
been sufficiently met. An oracle is a 3rd party service that
exists outside the blockchain to heterarchically reach out
and help verify the preconditions encoded within the
smart contract. These artifacts provide ways and means
to interface with the real world. Being outside the
underlying blockchain setting, they may have unique
governance issues that need to be addressed
independently. Also, they could provide heterarchic
governance monitoring services.
17. Blockchain Governance Kernel
Design
As was discussed in Section 13, knowledge
asymmetries create governance issues. However,
knowledge asymmetries are the basis for initiating any
successful trade; and is therefore not wrong per se. Indeed,
wealth creation requires such knowledge asymmetries.
Also, the half-life of knowledge is short and getting
shorter. In other words, there are no guarantees that a
given vantage point will remain forever. However, at any
given time, if the asymmetries are severe or resulted from
a prior information-agreement breach, it then opens up
problems of poor governance. Good governance,
therefore, involves creating adequate channels of
information flow for timely decision-making for all parties
to participate in sufficient amount of openness while also
allowing specific strategic/proprietary information to
remain hidden and off the grid (either permanently, or at
least for a while). Insights from Cryptographic Stigmergy
[77] as well as Cryptoeconomics [78] would be needed to
design the appropriate information signaling mechanisms
and economic incentives that help streamline the required
information flows. Here we delimit the context to the
kernel governance design to help decide between the off-
chain/on-chain approaches.
When we place the matrix of the Unknown-Unknown
asymmetries (Fig. 10) alongside the Inducto-Deductive
Front (IDF), it raises the issue of how the design-matrix
gets transformed when one of the parties on either side is
a machine working off a highly specified Codified
Governance Ruleset (CGR) as opposed to a human
working off a more abstract Principled Governance
Ruleset (PGR)? While the CGR could be codified into the
on-chain governance modules, the PGR would be
administered in pre-agreed, human-centered, arbitration-
like off-chain governance setups. Either may cross
hierarchical boundaries; i.e., PGR/CGR may be either |h
or |H in scope. Fig. 15 shows the kernel governance design
indicating when one or the other ought to be used.
When knowledge context is conceptual and below the
IDF, CGR-coded machines can adjudicate governance
modules coded as on-chain, smart contracts (i.e., via X2
CGR, Fig. 15a). In contrast, when the knowledge context
is conceptual but above the IDF, governance remains off-
chain and human adjudicated (i.e., via X1 PGR, Fig. 15a).
20
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
When the knowledge context is stigmergic and above
the IDF, governance remains firmly off-chain and human
adjudicated (i.e., via X3 PGR in Fig. 15b). The case where
the knowledge context is stigmergic and below the IDF is
a bit more nuanced. For even though the knowledge
context is safely below the IDF (and therefore could use
CGR), given the stigmergic uncertainties, it always needs
human oversight. It is a case of the decoupled design (as
shown in the lower half of Fig. 15b). It is, therefore, a
mixed case that uses both off-chain as well as on-chain
logic. This is the fundamental answer to the on-chain vs.
off-chain governance debate between Ehrsam [10] and
Zamfir [11] that we discussed in Section 2.
Fig. 15. Governance Kernel Design
From a CAS-perspective, the various configurations (in
Fig.15 a-b) are in dynamic flux; protagonists are
continually re-positioning for strategic advantage.
However, each of these configurations has β-level patterns
that frame the governance issue. Most dynamic (and
therefore requiring the highest amount of governance
effort) are those patterns that have intrinsic knowledge
asymmetries (i.e., the UK type); and the ones that need the
least amount of governance have fundamental knowledge
symmetries (i.e., UU, KK type). However, here too, if the
β-level patterns indicate that one or both parties suffer
from high-confidence coupled with poor-grasp (or
alternatively, good grasp, but low confidence), then
problems of governance may surface. This adds greater
onus to the “Know Your Customer” (KYC) type
guidelines.
Fig. 16 shows the overall composition of the
Blockchain design along with the governance sub-
modules (as discussed above) factored in. DAO was
missing these governance sub-modules. Dominance of the
governance sub-modules is evident given its top-row
position. Within the governance sub-modules (and in
agreement with Section 7), the conceptual dominates the
stigmergic on account of its logical consistency.
However, the conceptual does need to factor in the
broader inductive base that the stigmergic provides.
Controlling knowledge-architectures dictate the
requisite governance logic. |H-structures demand |H-
governance; likewise, |h-structures demand heterarchic,
Madisonian-like governance (Fig. 6). In either case, PGR
will need to be invoked whenever dealing with pre-
conceptual, stigmergic facts as well as concepts that reach
beyond the IDF. Consider the |h-case. The blockchain
architecture allows for |h-controls both in the
somatic/operational (i.e., via the consensus model) as well
as germline/kernel (i.e., via stigmergic/conceptual PGR's
& CGR's) cases (Fig. 16). Same is true for the less
demanding |H-case. Furthermore, there is no fundamental
difference between the governance of agents at the α1
level versus those engaged at a more evolved αn, n>1 level.
Either the agent conforms to the governing β-level; or,
alternatively, the β-level itself needs to evolve towards
greater consistency. The former is the case of operational
governance; the latter is the more demanding, shape-
shifting, kernel/germline governance that in dealing with
the various shades of unknown-unknowns, helps evolve
the CAS system towards greater consistency and reach.
Fig. 16. Governance Kernel for the Blockchain Design
19. Conclusions
Given the leveling of the playing field, the dis-
intermediation of the middle-men, and the transparency of
blockchain-based transactions, it is highly likely that the
information flows are on the verge of scaling
exponentially. Stigmergy steps in when information flows
scale beyond aided/unaided human cognitive limits. In
other words, the α↔β gearing (as discussed in Section 7)
will most likely ramp up as the technology gains
mainstream support. This paper has provided CAS based
governance guideposts as to what may be expected.
Salient points include:
• Review of blockchain governance to highlight novel
pathologies and the problem of unknown-unknowns.
• Issue of trust as it relates to the top nodes in a
hierarchical organization going rogue; and its solution
via heterarchical control (courtesy James Madison)
that conforms to a decoupled lower-diagonal design.
• How organizational structures as well as sense-
21
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
making changes with rising complexity.
• The unique challenges faced in the context of large
scale socio-technological organizational designs.
• The challenge of governance where “formal controls”
are missing.
• The role of stigmergic gearing in both biological as
well as human decision-making context.
• The original formulation of an iterative CAS.
• The original formulation of how the iterative CAS
framework may be utilized to help understand the
crux of the centralization/decentralization issue.
• The framing of governance from a heterarchically-
hierarchic human knowledge architecture
perspective. Then using this approach to
fundamentally frame the issue of human vs. machine
dominance in decision making (i.e., induction vs.
deduction).
• The framing of the Unknown-Unknown nuances and
the way these show up in the context of governance.
• The distinction between emergence vs. self-
organization; and how the disruption of the natural
flow between these two processes can lead to
governance pathologies.
• Highlighting the historical significance of Axiomatic
Design as providing a unifying conceptual vanishing
point (similar to the perceptual vanishing point in the
case of perspective drawings); except being in the
realm of conceptuals, it has far greater import,
especially in education.
• The blockchain technology when viewed from an
axiomatic perspective is seen to be a lower-triangle
decoupled design.
• The promise as well as the governance problem of
smart contracts.
• The design of the blockchain governance kernel as
helping decide between on-chain vs. off-chain
governance.
References
1. C. P. Snow. The Two Cultures: And a second Look
Hardcover. Cambridge University Press; First Thus
edition (1965)
2. N. P. Suh. The Principles of Design. 1st ed. New
York: Oxford University Press; 1990
3. B. H. Banathy. Designing Social Systems in a
Changing World. Springer. 1997.
4. C. Olaya, C. García-Díaz. Social Systems
Engineering: The Design of Complexity. Wiley.
December 2017
5. L. Marsh, C. Onof. Stigmergic epistemology,
stigmergic cognition, Cognitive Systems Research
(2007), doi:10.1016/j.cogsys.2007.06.009
6. J. Compton. How Blockchain Could Revolutionize
The Internet Of Things. Forbes. Jun 27, 2017.
https://www.forbes.com/sites/delltechnologies/2017
06/27/how-blockchain-could-revolutionize-the-
internet-of-things/#a969acf6eab4
7. M. Campbell-Verduyn. Introduction: What are
blockchains and how are they relevant to
governance in the contemporary global political
economy? In Bitcoin and Beyond. Cryptocurrencies,
Blockchains, and Global Governance. Edited by
Campbell-Verduyn M. Routledge Press. 2018
8. L. Lessig. Code: And Other Laws of Cyberspace,
Version 2.0 2nd Edition. Basic Books. 2006
9. J. Lopp. Nobody Understands Bitcoin (And That's
OK). Mar 11, 2017.
https://www.coindesk.com/nobody-understands-
bitcoin-thats-ok/
10. F. Ehrsam. Blockchain Governance: Programming
Our Future. Nov 27, 2017.
https://medium.com/@FEhrsam/blockchain-
governance-programming-our-future-c3bfe30f2d74
11. V. Zamfir. Against on-chain governance. Dec 1,
2017. https://medium.com/@Vlad_Zamfir/against-
on-chain-governance-a4ceacd040ca
12. Q. DuPont. Experiments in Algorithmic
Governance: A history and ethnography of “The
DAO,” a failed Decentralized Autonomous
Organization. In Bitcoin and Beyond.
Cryptocurrencies, Blockchains, and Global
Governance. Edited by Campbell-Verduyn M.
Routledge Press. 2018
13. S. Voshmgir. Blockchain’s Problem with Unknown
Unknowns.2017. https://medium.com/blockchain-
hub/blockchains-problem-with-unknown-unknowns-
6837e09ec495
14. Lord Acton (John Emerich Edward Dalberg). Letter
to Archbishop Mandell Creighton. Apr. 5, 1887.
https://history.hanover.edu/courses/excerpts/165acto
n.html
15. Juvenal (Decimus Iunius Iuvenalis). Satire VI, lines
347–348.
http://www.thelatinlibrary.com/juvenal/6.shtml
16. Plato, Republic.
http://www.perseus.tufts.edu/hopper/text?doc=Perse
us%3Atext%3A1999.01.0168%3Abook%3D9%3A
page%3D590
17. Plato, Republic.
http://www.perseus.tufts.edu/hopper/text?doc=Perse
us%3Atext%3A1999.01.0168%3Abook%3D3%3As
ection%3D414c
18. J Madison, A. Hamilton. The Federalist Papers : No.
51. Friday, February 8, 1788.
http://avalon.law.yale.edu/18th_century/fed51.asp
19. J. Thomas, A. Zaytseva. Mapping
Complexity/Human Knowledge as a Complex
Adaptive System. 2016 Wiley Periodicals, Inc., Vol.
21 No. S2. DOI 10.1002/cplx.21799. 24 June 2016.
https://onlinelibrary.wiley.com/doi/abs/10.1002/cplx
.21799
20. Y. Bar-Yam, Complexity rising: From human
beings to human civilization, a complexity profile,
Encyclopedia of Life Support Systems (EOLSS
UNESCO Publishers, Oxford, UK, 2002)
22
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
21. R. W. Emerson. Self-Reliance. 1841.
https://emersoncentral.com/texts/essays-first-
series/self-reliance/
22. G. Fields, J. R. Emshwiller. Many Failed Efforts to
Count Nation's Federal Criminal Laws. New York
Times. July 23, 2011
23. J. R. Licklider. Man-Computer Symbiosis. IRE
Transactions on Human. Factors in Electronics,
HFE-1. 4-11. 1960.
24. D. E. Snowden, M. E. Boone. A Leaders
Framework for Decision Making. Harvard Business
Review. Nov 2007
25. E. Puika. D. Ceglarek. The quality of a design will
not exceed the knowledge of its designer; an
analysis based on Axiomatic Information and the
Cynefin Framework. Procedia CIRP 34 (2015)
26. D. A. Norman, P. J. Stappers. DesignX: Complex
Sociotechnical Systems. She Ji. The Journal of
Design, Economics, and Innovation Vol. 1, No. 2,
Winter 2015
27. A. Scott-Wright A, A. A. Boxwala, Y. Denekamp,
R. A. Greenes, D. Tate. Applying Axiomatic Design
Methodology for Guideline Revision. AMIA
Annual Symposium Proceedings. 2003;2003:1000.
28. T. Shannon. Cynefin and Healthcare.
https://frectal.com/book/chaos-complex-
complicated-simple-and-cynefin/cynefin-
and%C2%A0healthcare/
29. S. S. Jones, R. S. Evans. An Agent Based
Simulation Tool for Scheduling Emergency
Department Physicians. AMIA Annual Symposium
Proceedings. 2008;2008:338-342.
30. E. Schadt. The role of big data in medicine.
November 2015
https://www.mckinsey.com/industries/pharmaceutic
als-and-medical-products/our-insights/the-role-of-
big-data-in-medicine
31. N. P. Suh. Complexity: Theory and Applications.
1st ed. New York: Oxford University Press; 2005
32. V. Chhotray, G. Stoker. Governance Theory and
Practice A Cross-Disciplinary Approach. Palgrave
Macmillan. 2009
33. Bible. New International Version. Proverbs 6:6-8.
https://www.biblegateway.com/passage/?search=Pro
verbs+6%3A6-8&version=NIV
34. M. Walker. Ant mega-colony takes over world.
BBC. July 1, 2009.
http://news.bbc.co.uk/earth/hi/earth_news/newsid_8
127000/8127519.stm
35. P. P. Grassé. Insectes Sociaux VI. 79. 1959
36. Ants. NetLogo Models Library
http://ccl.northwestern.edu/netlogo/models/Ants
37. T.D. Tom De Wolf, T. Holvoet. Emergence Versus
Self-Organisation: Different Concepts but Promising
When Combined. In Engineering Self-Organising
Systems: Methodologies and Applications (Lecture
Notes in Computer Science). Edited by Brueckner S.
A, Serugendo, G. D. M, Karageorgos A. Springer.
2005
38. H. V. D. Parunak. A survey of environments and
mechanisms for human-human stigmergy. 2006.
Environments for Multi-Agent Systems II: 163-186.
39. J. F. Padgett, W. W. Powell. The Emergence of
Organizations and Markets. Princeton University
Press (October 14, 2012)
40. S. A. Kauffman. The Origins of Order: Self-
Organization and Selection in Evolution. Oxford
University Press; 1 edition (June 10, 1993)
41. R. K. Sawyer. Social Emergence: Societies As
Complex Systems. Cambridge University Press
(October 27, 2005)
42. F. Klein, H. Giese. Separation of Concerns for
Mechatronic Multi-agent Systems Through
Dynamic Communities. In: R. Choren, A. Garcia, C.
Lucena, A. Romanovsky (eds) Software
Engineering for Multi-Agent Systems III. SELMAS
2004. Lecture Notes in Computer Science, vol 3390.
Springer, Berlin, Heidelberg
43. S. E. H. Murph, E. Simona. Nanoscale Materials
Fundamentals and Emergent Properties in
Anisotropic and Shape-Selective Nanomaterials:
Structure-Property Relationships. In: S. E. H.
Murph, Dr. G. K. Larsen, K. J. Coopersmith (eds)
Anisotropic and Shape-Selective Nanomaterials -
2017 Structure-Property Relationships. Springer; 1st
ed. 2017 edition (July 15, 2017)
44. M. Viroli, J. Beal, K. Usbeck. Operational
Semantics of Proto. Science of Computer
Programming. October 22, 2012
45. M. A. Bedau, P. Humphreys. Emergence:
Contemporary Readings in Philosophy and Science.
A Bradford Book; 1 edition (March 21, 2008)
46. F. Heylighen. Stigmergy as a generic mechanism for
coordination: definition, varieties and aspects.
ECCO Working paper 2011-12
http://pespmc1.vub.ac.be/Papers/Stigmergy-
Springer.pdf
47. F. Bacon. Novum Organum (1620)
https://en.wikisource.org/wiki/Novum_Organum/Bo
ok_I_(Spedding)
48. M. J. Doyle., L. Marsh. Stigmergy 3.0: From Ants to
Economies, Cognitive Systems Research (2012),
49. S. D Palley. How to Sue A Decentralized
Autonomous Organization. Mar 21, 2016.
https://www.coindesk.com/how-to-sue-a-
decentralized-autonomous-organization/
50. J. H. Holland. Studying complex adaptive systems.
Journal of Systems Science and Complexity 2006,
19(1), 1-8.
51. V. Buterin. Notes on Blockchain Governance. Dec
17, 2017.
https://vitalik.ca/general/2017/12/17/voting.html
52. P. Baran. On Distributed Communication Networks.
IEEE Transactions of the Professional Technical
23
MATEC Web of Conferences 223, 01010 (2018) https://doi.org/10.1051/matecconf/201822301010
ICAD 2018
Group on Communications Systems. Volume CS-
12, Number 1, March 1964.
53. V. Buterin. The Meaning of Decentralization. Feb 6,
2017. https://medium.com/@VitalikButerin/the-
meaning-of-decentralization-a0c92b76a274
54. D. L. Leger. How FBI brought down cyber
underworld site Silk Road. USA TODAY. May 15,
2014
https://www.usatoday.com/story/news/nation/2013/
10/21/fbi-cracks-silk-road/2984921/
55. C. Harper. Another Class Action Filed Against
Ripple, Claims XRP Has “Hallmarks of a Security”.
Jul 5, 2018
https://bitcoinmagazine.com/articles/another-class-
action-filed-against-ripple-claims-xrp-has-
hallmarks-security/
56. T. Maas. The Curious Tale of Tezos —from a $232
MILLION ICO to 4 class action lawsuits. Apr 6.
2018 https://hackernoon.com/the-curious-tale-of-
tezos-from-a-232-million-ico-to-4-class-action-
lawsuits-6f411b7aad7e
57. R. Frost. Mending Wall. 1914
https://www.bartleby.com/104/64.html
58. B. S. Srinivasan, L. Lee. Quantifying
Decentralization: We must be able to measure
blockchain decentralization before we can improve
it. July 27, 2017. https://news.earn.com/quantifying-
decentralization-e39db233c28e
59. S. Ivanov. Op Ed: What Do We Mean When We
Talk About the "Blockchain Ecosystem”? Feb 26,
2018. https://bitcoinmagazine.com/articles/op-ed-
what-do-we-mean-when-we-talk-about-blockchain-
ecosystem/
60. M. Yeroulanos, O. Taplin. A Dictionary of Classical
Greek Quotations. I.B. Tauris & Co Ltd. 2017
61. W. Shakespeare. Othello. Arden Shakespeare; 3rd
edition. http://shakespeare.mit.edu/othello/full.html
62. Paul Ekman Group. Micro Expressions.
https://www.paulekman.com/micro-expressions-2/
63. J. Thomas. Archstand Theory of Design for
Innovation. PhD Thesis, 1995. MIT. Available at:
http://dspace.mit.edu/handle/1721.1/11722
64. G. Kasparov, M. Greengard. Deep Thinking: Where
Machine Intelligence Ends and Human Creativity
Begins. PublicAffairs; 1 edition (May 2, 2017)
65. D. Rumsfeld. The Unknown Knowns
https://www.youtube.com/watch?v=nAnKdq5Yty8
66. E. Morris. The Certainty of Donald Rumsfeld (Part
1). New York Times. March 25, 2014.
https://opinionator.blogs.nytimes.com/2014/03/25/th
e-certainty-of-donald-rumsfeld-part-1
67. J. Keats. Endymion. Book-1. The Poetical Works of
John Keats. (1795–1821).
http://www.bartleby.com/126/32.html
68. J.T. Foley, E. Puik, D. S. Cochran. Desirable
complexity. The 10th International Conference on
Axiomatic Design, ICAD 2016. Procedia CIRP 53.
69. J. Thomas, P. Mantri. Axiomatic Design/Design
Patterns Mashup: Part 1. In: 9th International
Conference on Axiomatic Design (ICAD), Procedia
CIRP, 2015, 34, 269-275. Available at:
http://www.sciencedirect.com/science/article/pii/S22
12827115008501
70. J. Thomas, P. Mantri. Axiomatic Design/Design
Patterns Mashup: Part 2. In: 9th International
Conference on Axiomatic Design (ICAD), Procedia
CIRP, 2015, 34, 276-283. Available at:
https://www.sciencedirect.com/science/article/pii/S2
212827115008513
71. M.E. Saunders, R.K. Peisley, R. Rader, G.W. Luck.
Pollinators, pests, and predators: Recognizing
ecological trade-offs in agroecosystems. Ambio. A
Journal of the Human Environment. February 2016,
Volume 45, Issue 1, pp 4–14.
72. S. Y. Edgerton. The Mirror, the Window, and the
Telescope: How Renaissance Linear Perspective
Changed Our Vision of the Universe. Cornell
University Press (January 29, 2009)
73. T. H. McConnell, K. L. Hull. Human Form, Human
Function: Essentials of Anatomy & Physiology.
Lippincott Williams & Wilkins. 2010
74. M. Jacobson. How Far Down the Search Engine
Results Page Will Most People Go? SEO. 2017.
https://www.theleverageway.com/blog/tag/seo/
75. D. Prountzos, K. Pingali. Betweenness Centrality:
Algorithms and Implementations. PPoPP '13
Proceedings of the 18th ACM SIGPLAN
symposium on Principles and practice of parallel
programming. Volume 48 Issue 8, August 2013.
76. S. Voshmgir, V. Kalinov. Blockchain-A Beginners
Guide. blockchainhub.
https://blockchainhub.net/blockchain-technology
77. P. Goorha. The Return of ‘The Nature of the Firm’:
The Role of the Blockchain . The JBBA. The
Journal of The British Blockchain Association. 2018
Vol 1, Issue 1.
78. Awesome Cryptoeconomics. An awesome curated
list of Cryptoeconomic research and learning
materials. https://github.com/jpantunes/awesome-
cryptoeconomics
79. M. Castro, B. Liskov. Practical Byzantine fault
tolerance. OSDI '99 Proceedings of the third
symposium on Operating systems design and
implementation. Pages 173-186
80. A. K. Abbas, A. H. H. Lichtman, S. Pillai. Cellular
and Molecular Immunology, Elsevier; 9 edition
(May 25, 2017)
81. S. Nakamoto. Bitcoin: A Peer-to-Peer Electronic
Cash System. https://bitcoin.org/bitcoin.pdf
82. S. Voshmgir. Disrupting governance with
blockchains and smart contracts. In Strategic
Change: Briefings in Entrepreneurial Finance. The
Future of Money and Further Applications of the
Blockchain. Volume 26, Issue 5. September 2017.