In the second chapter of her dissertation, Preiser tackles the difficulties of knowing complex systems. She starts with four quotes from four modern philosophers of science who all capture the issue nicely, but Bruno Latour is the most succinct: "We have taken science for realist painting, imagining that it made an exact copy of the world" (Pandora's Hope 78). Latour's implication, of course, is that science is not an exact copy of the world. What his statement overlooks is that realist painting isn't an exact copy either.
The very term copy implies not exact, not the same, but something that affords some utility: more portable and handy and reproducible. We make copies because the original is too unwieldy — too complex — to deal with handily. The models of the world — the knowledge — that we carry around in our heads, books, and computers to use for our various purposes are all inexact copies that greatly reduce the complexity of reality to make the models easy to use. A realist landscape will hang on our walls — the landscape itself will not. We reduce the landscape to a two by four foot stretch of canvas in order to make it fit and to make it easy to transport when we move or sell it, but that reduction always leaves things out that are quite likely important to the actual landscape itself. For instance, the painting may not capture the increasingly dry weather conditions that make the forest susceptible to fire and blight. The painter can overlook those details to satisfy his own purposes, the forest cannot.
Likewise, a scientist's model of the forest will omit some details and dynamics to make the forest intelligible — to paint an intelligent picture — and usually for some purpose — perhaps to convince Congress to act on global warming, but the reductions in her model always leave out details that may very well prove to be critical later on. Only the forest is exactly itself. Copies are not. As George Box has told us so very well: "All models are wrong, but some are useful." If complete correspondence between reality and your model (poem, painting, or formula) is your objective, sorry. Everything we know is wrong, but some of it is useful.
Of course, I overstate my case. Day to day, we find it useful to say that we know things, and we can usually rely on this knowledge in our proximate zones of influence, but we must always be aware that our knowledge like our influence extends only so far. We can always reach a limit where our knowledge breaks down and becomes error. I reach that limit every time I write. That's where all the insight waits.
I find it rather humorous that in its reaction against reductionist thinking, complexity reveals that knowledge always reduces reality. Yet, in a strange way and quite unexpectedly, this tension between a desire for holism and the necessity of reductionism is the zone of best complexity thinking. Complexity must operate in that hot, volatile zone between the certainty that our knowledge models reveal something reliable and testable about reality while at the same time leaving out something that is important about reality. In other words, we can be confident that we know something wrong. We can have reasonable hopes that it may be useful to us in certain situations, and we can be certain that it will prove incorrect in other situations. Strangely enough, we are both enabled by what we know and limited by what we know, equally. This is the hot zone within which human knowledge must work.
So what does Preiser say about the problem of knowledge?
Preiser addresses two core problems with modern knowledge:
- the failure of Newtonian/Cartesian reductionism to cope with the complex issues raised by the discovery of the quantum particle, evolution, and relativity, and
- the restructuring of knowledge itself in the face of these complex issues.
Preiser notes that our current dominant scientific epistemology follows from classical Cartesian/Newtonian, which in short, asserts that the descriptions of reality produced through isolation, observation, and the establishment of regularities describes reality as it is. Preiser claims that this reductionist epistemology works well enough for mechanical, closed systems, but is inadequate for dealing with complex systems with emergent properties, which leads to problems for our knowledge generating practices. She intends to correct this issue with a post-reductionist epistemology that incorporates a more holistic complexity view while coping with the necessary reductionism inherent in any epistemology.
She then gives an overview of Cartesian/Newtonian reductionism, which posits five key features of Newtonian natural systems: they are deterministic, closed, reversible, atomistic, and universal. This model became the basis of the modern scientific method and epistemology, and it was so successful that its assumptions of unchangeable, timeless properties and laws that govern the universe soon spread throughout Western thought. This view has a number of implications. First, natural systems can be known by analyzing and isolating their parts into elementary matter and interactions that follow universal and uniform laws. Science, and by extension true knowledge, is thus the process of classification, measurement, and rational organization. Newtonian reductionism was expressed in the universal languages of mathematics and logic, which precisely represented the real world as it is.
But, Preiser cautions, the fault lines within the Newtonian scientific model finally cracked with the discovery of the quantum particle, that Gordian knot of interactions and exchanges rather than a single, unified thing. The Newtonian model could not formalize the behavior and fundamental nature of quantum particles. Moreover, relativity and evolution revealed that the concepts of space and time, absolutes in Newton's model, must be changed to account for new experience and insights. These ruptures in the Newtonian model allowed complexity theory to emerge as a new view of reality.
Many complexity theorists have recognized what Preiser calls the first problem of knowledge: that a gap has emerged between our knowledge of the world and the world itself because of the empirical difficulties of describing the physical and phenomenal characteristics of complex phenomena. The logic of classical science cannot keep up with the generative, flexible, and pluralist nature of knowledge needed to describe complex systems. Complex phenomena challenge the five Newtonian postulates mentioned above: they are non-linear rather than linearly deterministic, open rather than closed, contingent in time rather than reversible, neither compressible nor universal, but always unfolding in a local, complex ecosystem. Preiser insists that we need new methods and vocabularies to usefully describe complexity.
Preiser insists that developing these new conceptual frameworks for knowledge requires recognizing four different kinds of reductionism inherent in any knowledge system:
- Ontological reductionism claims that all physical and non-physical phenomena can be explained in terms of matter, particles in motion.
- Epistemological reductionism claims knowledge in one discipline can be reduced to another discipline, ultimately to physics.
- Methodological reductionism claims that all systems are best investigated at the lowest, simplest possible level.
- Causal reductionism claims that all emergent properties of a system can be explained by their causal relations to the basic elements of the system, thus denying any downward causation in emergent phenomena.
Reductionism creates a kind of blindness when knowledge seekers ignore the complex systems at hand to investigate the simpler elements and then to explain the complex system only in terms of the simpler elements or systems. This reductionism ignores its own blind spots in order to claim universal truth. Many with a more holistic sense of reality have argued against this reductionism, but holism itself cannot escape reductionism. Indeed, Preiser argues that it is impossible to avoid the four kinds of reductionism, which are all implicated with one another, and that most complexity theorists fall into one of two traps: those who see no distinction between the system and its environment in some holistic approach and those who insist that all complexity can indeed be measured and simulated by computational models to reveal universal laws.
Following Cilliers, Preiser insists that a rigorous understanding of complexity must be aware that any description of complexity involves some reduction of reality. This understanding leads to a performative tension that destabilizes the dichotomy between either holism or reductionism. It's always both. Thus, any engagement with complex systems is always a dynamic interaction among the nature of phenomena (ontology), our knowledge of it (epistemology), and our methods for studying it (methodology) in a dialectical (Cilliers) or a dialogical (Morin) process that Preiser calls general complexity, after Morin.
General complexity is at once coherent and open with the result that our understanding is never absolute but always contingent and skeptical of itself, allowing the researcher to reflect critically on her knowledge generating practices. She is no longer certain that her models fit reality like its mirror image as she shifts her focus from the properties of entities in classical science to the relations among entities and the echoing relations among relations in complexity science, which of necessity leads her to an entirely new epistemology as new knowledge requires new ways of modelling reality, new ways of framing reality to gather knowledge from it through observation and interpretation. But, as Preiser warns repeatedly, no model can capture the full complexity of any complex system as such systems are radically contextual and radically open. In some ways, a system's degree of complexity can be measured by the degree of difficulty in modelling the system. Any modelling system (a particular science or novel, I think) must decide what observables of a given real system to include and which to exclude in order to function as a model and to generate knowledge about that system. Knowledge, then, always limits a contextual and open system in order to understand it and use it, but it never knows when the parts of the system excluded by our models — which in the real system are still interacting non-linearly with the parts included in our models — will become relevant. Given that we cannot avoid the reductionism of any model, of any knowledge, then we must embrace up front and constantly the limitations of our models. An irreducible gap exists between complex reality and our knowledge of that reality; thus, to create knowledge, we must use reductionist strategies to be able to say anything meaningful about complex systems at all, but our models too seldom acknowledge what's left out, and thus they all have blind spots. This is the nature of knowledge as revealed by complexity: that knowledge is limited, but as Preiser argues, this limitation is not a disaster but a condition for knowledge. Limits enable knowledge. As Dutch philosopher Cornelis Anthonie van Peursen explains, we need a horizon that limits our field of vision for the act of seeing to take place. This horizon is formed by the interaction of the observer and the environment, and is situated in both at once. It is both inside (subjective) and outside (objective) the observer.
Having explored the first problem of knowledge, or the epistemological rupture that occurs when moving from the reductionist Newtonian paradigm to the complexity paradigm, Preiser frames the second problem of knowledge, arguing that knowledge generating practices and the notion of knowledge itself changes in the face of complexity.
Preiser complains that most current complexity science is still reductionist: concerning itself with measurement and uncovering regular laws — an approach that, according to Morin, recognizes complexity by decomplexifying it. The heart of the error of decomplexifying lies in the assertion that what is left out of the measurements and calculations are not of importance, but as Cilliers insists, they are of utmost importance as they are still a vital, perturbing part of the real system being measure and calculated in the model, and in complex systems, even small parts can have large effects (the butterfly effect of chaos theory). Preiser proposes Morin's concept of general complexity that replaces the concept of disjunction between emergent features of a system and its underlying structures with the concept of distinction between emergent and underlying structures that recognizes both their independence and dependence in the system. This is a post-reductionism that is self-aware of the blind spots of its own practices and disarms the animosities of opposing paradigms without uniting them into a grand monist truth. Post-reductionist denies neither reductionism nor holism, but holds them in dialectical tension and assumes that the most useful knowledge lies in the interplay of both. This new approach to generating knowledge requires a new language and vocabulary.
Preiser claims that complex knowledge is hybrid and difficult: because complex knowledge acknowledges dynamic relationships as well as entities, it is not static or fixed, but dynamical and provisional, not limited to a stable entity, a fact, but branching out to other knowledge regimes so that there is always a surplus of signification in which meaning is open, infinitely disseminated, and ultimately uncontainable (rhizomatic, in Deleuzional terms). The process of generating, storing, and using knowledge becomes a dynamic complex system itself. While, complex knowledge rejects both the absolute totality of knowledge and the possibility of representing something fully, it does not reject knowledge, truth, and representation in some anything-goes relativism. Rather, it challenges us to know and engage the limits of our knowledge, and to re-invent if necessary. In short, complex knowledge is the ghost of reality, and haunts those liminal spaces where knowing meets non-knowing.
So does Preiser clarify (reduce to a working model) this complex knowledge? I hope so. That's why I'll read the rest of her dissertation. But in this chapter she reinforces for me issues in writing fiction that almost all fiction writers and readers struggle with: where to put the frame of beginning and ending? what to put in the middle and what to leave out? and to understand the implications of all those forces that are perturbing the narrative but could find no space or time for expression. No one can tell the whole story, so how do you tell an engaging story?
I think the best writers have always understood intuitively the complexity of the world. Of course, formula fiction is a closed little system with neat actors interacting in highly regular and predictable ways (stock characters with fixed plots), but the best fiction is open to the world, mapping new terrains to see what happens, following ghosts in that liminal space between knowing and unknowing. That's the good stuff.