Conclusions

Superquestions for Supercomputers (Part #10)

[Parts: First | Prev | All] [Links: To-K | From-K | From-Kx | Refs ]

It would appear that there is a form of "disconnect" between the applications for which supercomputers are currently envisaged and the "superordinate questions" which might ensure their relevance to "global societal problems" and the "important issues facing society" -- to the issues with which human beings are ultimately concerned. The capacity to consider such superquestions has in effect been carefully designed out as with the challenges faced by individuals and collectivities in engaging with complexity..

"Excessive complexity": It is typically argued that so many of the challenges of governance inhibiting effective response to global challenges are "too complex" to be amenable to the quantitative power of supercomputers -- despite the decades of research on complex systems dynamics. As with the exclusion of the population factor from the response to climate change framed by the above-mentioned Kaya Identity, "simpler" problems are then selected for attention in what amounts to an exercise in conceptual gerrymandering. In terms of the larger systemic chellanges, might this be described as "keyhole science"? The problems and questions that are "too complex" for mathematicians are then left to the capacities of those "of more limited intelligence" in positions of authority -- with minimal computer support. In the case of climate change, this will in all likelihood result in the recommendation and implementation of geo-engineering solutions of the most questionable technical simplicity -- since these lend themselves to exploration with the kinds of models amenable to supercomputer calculation (Geo-engineering Oversight Agency for Thermal Stabilization (GOATS), 2008).

Surprise: Despite the widespread use of computers for economic and financial modelling, clearly the methodology is fundamentally defective in relation to the case for superquestions. The point may be succinctly made through the now famous question of the Queen of England to the assembled faculty of the London School of Economics with regard to the financial crisis: Why did nobody notice it? (The Queen asks why no one saw the credit crunch coming, The Telegraph, 5 November 2008). The LSE director of research responded: At every stage, someone was relying on somebody else and everyone thought they were doing the right thing.

However, the point to be made is that neither the superordinate question nor that response are appropriately integrated into the models for which supercomputers are currently designed. The pattern of the response can be usefully elaborated further (Responsibility for Global Governance Who? Where? When? How? Why? Which? What? 2008). Given the proven strategic expertise of computers in chess, a strategic distinction might be fruitfully made, using chess metaphors, between:

• the Queen's superordinate "move": To what other challenges of governance, such as overpopulation, might such a superquestion be fruitfully addressed, as variously highlighted (Karen A. Cerulo, Never Saw It Coming: cultural challenges to envisioning the worst, 2006; Nassim Nicholas Taleb, The Black Swan: the impact of the highly improbable, 2007)
• the Knight's "move": Traditionally held to symbolize power in multidirectional movement and the use of intelligence information in the formulation of strategy and countermeasures. For that reason it figured on the crest of the US Army 312th Military Intelligence Battalion, known as the "Silent Warriors". In the context of the financial and other crises of governance, as illustrated by the response above of the LSE director, it effectively symbolizes fast footwork in advantageously reframing strategic situations. Perhaps appropriately, "knight's move thinking" (asyndesis or entgleisen) also refers pejoratively to a pathological pattern of discourse characterized by only remotely related ideas.

It might be asked how many teraFLOPS (and misguided dependency on their results) are required to increase the risk of a "terra flop" -- a global crisis of the form increasingly anticipated -- in contrast with the number required to avoid the kind of "flop" which characterizes so many current global plans.

Reformulation: Given the case made for supercomputer capacity, it is useful to explore the argument of physicist Peter Rowlands in a domain where such capacity is currently required (Removing redundancy in relativistic quantum mechanics, 2005):

Essentially, the problem arises from the use of an intrinsically 2-dimensional mathematical structure to represent a 3-dimensional reality; we can, for example, show that the problem is immediately solved and the singularity removed when the intrinsically 3-dimensional quaternions are introduced. Exactly, the same kind of reasoning can be applied to relativistic quantum mechanics, where problems emerge from the imposition of a matrix representation. Relativistic quantum mechanics as represented by the Dirac equation and quantum field theory produces at least one type of singularity that appears to be an artefact of the system - the infrared divergence. It also leads to infinities that have to be removed by renormalization, even in the ideal case of free particles where there is apparently no real source for the divergent terms. In addition, there appears to be a great deal of redundancy. For example, QCD calculations using Feynman diagrams derived from the standard gamma matrix representation require ten million calculations for a six gluon interaction, whereas the alternative algebraic approach using twistor space... reduces the calculations required to only six. Even with this method, it is clear that redundancies are still visible; so the question we should ask is whether it is possible to find a coordinate system for the fermionic state which removes redundancy entirely. [emphasis added]

It might then be asked whether the governance problems that are alleged to be "too complex" are a consequence of their formulation using the 2-dimensional matrices (characteristic of the input/output charts of spreadsheets), as is typically the case -- and is notably the case in the 4-fold figure above. Formulated otherwise, the impossible "millions" of calculations might be reduced "to only six" -- a theme explored separately (Geometry of Thinking for Sustainable Global Governance, 2009; Metaphorical Geometry in Quest of Globality, 2009), notably in terms of the implications for identity and engagement (Geometry, Topology and Dynamics of Identity, 2009; Topology of Valuing: dynamics of collective engagement with polyhedral value configurations, 2008). How might the tentative list of superquestions be better formulated, perhaps as "superordinate questions", to render them amenable to sophisticated analysis -- with output suitably constrained to be widely comprehensible?

Avoidance of innovation: Of relevance to the general theme of this argument is that an earlier paper of Peter Rowlands, submitted to the arXiv open archive for physics, was suppressed by the administrators as being "inappropriate" and "of no interest" to users -- and without further explanation (see paper trail at: The Suppression of Dr. Rowlands' Quantum Physics Paper). Rowlands' comment on the process is relevant to the control of innovation in a global knowledge society -- but in this case raises questions relevant to the mindset governing investment in supercomputers:

However, it was novel and original in its approach, which, of course, is the whole reason for doing research in the first place. The arXiv is not a journal with specific stated policies for inclusion.  It claims to represent the whole of physics, and it does not say anywhere that it will refuse to publish papers that fall outside the narrow interests of its moderators.  This covert censorship is even more insidious in the light of arXiv's pretended policy of being open.

Physics would seem to remain insensitive to a problem of knowledge management recognizable in governance. This insensitivity notoriously engendered assertions such as the following:

• Heavier-than-air flying machines are impossible. (Lord Kelvin, president, Royal Society, 1895).
• There is nothing new to be discovered in physics now. All that remains is more and more precise measurement (Lord Kelvin).
• Flight by machines heavier than air is unpractical and insignificant, if not utterly impossible. (Simon Newcomb, 1902).
• Space travel is bunk (Sir Harold Spencer Jones, Astronomer Royal of Britain, 1957, two weeks before the launch of Sputnik)

This is curious in the light of the widely-cited assertion by Niels Bohr that innovative theories in fundamental physics need to be "crazy enough". Worse still is the possibility that the future will judge the mathematics and physics of today -- convinced as it is of the need for supercomputers -- as being "not even wrong". More curiously, to an entry on the Status of Superstring and M-theory (in a blog entitled Not Even Wrong, 13 December 2008), Samuel Prime responds:

You wonder if the enormous advancements in physics (and science) are prompting researchers to delve into, speculate, and think about superquestions and more adventerous ideas that, by nature, go beyond the currently accepted methods of science. It's hard to imagine science not evolving in its methods.... We are hitting the boundaries of experiment. (More specifically, experiment that can be humanly done.) These many superquestions and supercuriosities show that we are more curious by questions that now seem untestable by our current state of abilities, at least to a high degree of certainty.

What questions might the future consider it appropriate to have been formulated  now? The point to be stressed is that such "interesting questions" are of a kind that needs to be used to refine "superquestions" -- if the same mindset is not to be used to define the (above-mentioned) "global societal problems", and the "important issues facing society", for which supercomputers are assumed to enable solutions. It is unfortunate that in practice the greater the sophistication of analysis, as enabled by supercomputers, the less the relevance to the practical challenges of governance and sustaining its credibility.

Unconventional pattern exploration: Following his widely-cited work on a pattern language and his subsequent work on order in nature, Christopher Alexander has suggested a program of research of great potential relevance to governance (New Concepts in Complexity Theory: an overview of the four books of the Nature of Order with emphasis on the scientific problems which are raised. 2003; Harmony-Seeking Computations: a science of non-classical dynamics based on the progressive evolution of the larger whole, International Journal for Unconventional Computing (IJUC), 5, 2009). Its implications, notably in the light of "superquestions", are discussed separately (Harmony-Comprehension and Wholeness-Engendering: eliciting psychosocial transformational principles from design, 2010).

Through enhancing human pattern recognition capacity, whether through geometry or dynamic rhythm, is it the case that superquestions do not need supercomputers -- provided the capacity of the human brain is appropriately enabled to formulate and engage with superquestions? Is the issue rather how to reframe the "coordinate system" which currently introduces redundant singularities? In this respect it is appropriate to look for a mnemonic reframing of FLOPS:

• FLOPS: currently an abbreviation for the computing term: floating point operations per second... perhaps to...
• FLOPS: potentially an abbreviation for a governance term: floating perspective operacy per (sub)system

The mnemonic highlights the need to encompass multiple perspectives (effectively "floating" between them), enabling appropriate action (operacy) in their associated (sub)systems.

Enabling collective intelligence: It is striking, at the time of writing, that individual tax payers (via dubious governmental intermediaries) are being called upon to bail out various "developed" countries (Ireland, following Greece, potentially to be followed by Portugal, Spain and Belgium) -- after previously bailing out banks and corporations "too big to fail".  Most curiously no one has been effectively held responsible for such extreme mismanagement and there is every sign that "business as usual" is expected to continue -- appropriately rewarded by bonuses acknowleged to be obscene, awarded to those taking the most extreme risks (Extreme Financial Risk-taking as Extremism -- subject to anti-terrorism legislation?, 2009). Mismanagement with impunity has become the primary characteristic of governance at the highest level (Emergence of a Global Misleadership Council: misleading as vital to governance of the future? 2007).

The academic community has been significantly complicit in this process, whilst disclaiming any responsibility whatsoever. It is therefore appropriate to ask whether that mindset will ensure appropriate use of supercomputers -- to remedy, rather than exacerbate, the challenge for the population at large in continuing to bail out systemic stupidity, the antithesis of collective intelligence.

The set of tentative superquestions might be usefully understood as derivative of a more fundamental, superordinate question for a global knowledge-based society. This might be framed as how does the rapidly developing computing power, distributed or otherwise, enable integrative dialogue capable of enabling the emergence of collective intelligence.

A danger is that techno-enthusiasm considers this as already to be an inherent characteristic of current web interactivity, social networking and the like, when these have as yet to demonstrate a quantum leap in capacity to respond coherently to emerging crises -- as demonstrated on a minor scale by the Gulf oil spill (Enabling Collective Intelligence in Response to Emergencies, 2010) and, at the time of writing by the situation in Haiti.

"Purpose, culture, process, and people replace strategy, structure, and systems as our superordinate questions."
James Champy, Reengineering Management: mandate for new leadership. 1995

[Parts: First | Prev | All] [Links: To-K | From-K | From-Kx | Refs ]