Ricard SolŽ & Brian Goodwin (2000), Signs of Life: How Complexity Pervades Biology, New York: Perseus Books.

[Add to notes on chaos, complexity and emergence]

The examination of life itself appears a fundamental example of complexity theory. The origins of life continues to confound modern scientists aiming to create life out of 'primordial soup': even Darwin recognised that there must be a beginning for evolution (SolŽ & Goodwin, 2000: 240; this is where artificial life begins to speculate on life-forms using computer simulations). Building upon the work of mathematicians and physicists, biologists have developed a keen interest in the emergent order out of the complexities of life's material foundations. That a local part is dynamically linked to the global whole is demonstrated in the case of the human organism where various part can indicate information about the condition of the system as a whole and imbalance in any part therein (take Chinese medicine as a good example of this).

SolŽ & Goodwin see this as an opportunity to move beyond the scientific impulse to explain, predict and control nature that operates in a linear fashion of orderly cause and effect, to an understanding of human participation in nonlinear natural processes, not least in the way the observer is embedded in this (2000: 1 & 28). They point out that an understanding of nonlinearity is nothing new, the stable cycle of the clock pendulum has served as a paradigm for control in this respect where adjustment is simply made by altering the length and mass of the pendulum and - the dualistic 'tick-tock' logic of the mechanical age. That the world is not clock-work but more chaotic was recognised by Henri PoincarŽ (in 1913) observing that small differences in initial conditions produce enormous errors the final phenomena. This is not simply a 'stochastic or random process, where the irregularity arises from the cumulative effects of a multitude of many extraneous influences', but is a 'generic dynamic state' containing its own intrinsic logic (SolŽ & Goodwin, 2000: 3). Unlike the clock pendulum, chaos is both deterministic and unpredictable. 

[leads to Lorenz 50 years later - see other notes]

Complexity is neither complete order nor disorder, but 'displays nontrivial correlations that are not reducible to smaller, more fundamental units' (SolŽ & Goodwin, 2000: 34). In this sense, it is a fractal object.  

The principles of the living things, for SolŽ & Goodwin, lie in the complex operations and dynamic organisation of the gene and the cell. Genes do not simply generate coherent behaviour but a pattern of order involves emergent properties, 'changes of state can occur spontaneously, without any defined internal or external cause. By definition these changes are epigenetic phenomena: dynamic processes that arise from the complex interplay of all the factors involved in cellular activities including the genes.'  (2000: 63) They explain how 'epigenesis' is a term from Aristotle to describe how embryos develop from the interaction of emergent parts in contrast to 'preformatism' where the embryo simply grows (this is essentially 'genetic determinism' and a gross oversimplification). The study of bacteria has been fruitful in this connection, and in particular how mutations occur.

Researchers have discovered that cells can operate as closed networks (what Kaneko and Yomo call 'autocataytic'), demonstrating all the properties of a system to maintain itself and grow. Also 'Cell division occurs when a particular constituent, defined as a division factor, exceeds a threshold. Cells can also die of starvation.' (SolŽ & Goodwin, 2000: 64) Taken together, cells demonstrate nonlinear dynamics wherein genes do not determine the development of the organism, but stabilise 'generic patterns of emergent complexity in these multicellular systems' (2000: 66). 
[link to the cell in the commodity - Marx/Mandel?]
Stuart Kauffman approximated the regulatory dynamics of genes in terms of binary switches, a gene state of 'on' or 'off' - Boolean logic applied to gene dynamics. He set up a network where genes are selected at random from a set, and allowed to interact with random couplings, and examined the state transitions that the network underwent. The network demonstrated a unexpected high degree of dynamic order (in SolŽ & Goodwin, 2000: 72-5). In this, very simple rules of interaction with a complex system can be seen to generate 'emergent order or what Kauffman has called "order for free" in evolution: unexpected constraints on the large-scale dynamic patterns of gene activities that provide the living state with the properties needed for generating flexible, adaptive and robust behaviour.' (in SolŽ & Goodwin, 2000: 78) 
[this is what Katherine Hayles was talking about as a radical critique of the way we understand nature - reducible to binary logic]
This is not a deterministic theory; the gene is reducible to code as a way of understanding these properties and dynamic relations. 

The brain is perhaps easier to compute, requiring a model of networked simple elements or interconnected neurons. For instance, in 1982, John Hopfield introduced a very simple model of a neural network, comprising of interconnected binary units, that would 'learn' by association to describe a basic structure and dynamics (in SolŽ & Goodwin, 2000: 124-5; although clearly this is not pedagogy as such). Clearly there are massive limitations in terms of any analogy to neural activity in the brain, but this is where learning really has taken place in the development of more sophisticated models that ironically confirm that the functioning of brain is chaotic. The brain demonstrates many of the same properties of other nonlinear dynamic systems, involving multiple attractors requiring order and disorder to operate (SolŽ & Goodwin, 2000: 145).  

In a social context, the study of insects reveals some of the emergent patterns more overtly:
'It is very interesting to see how simple is the answer to our initial problem: individuals dealing only with local, noisy information are able to generate ordered, large-scale structures through the amplification of initial perturbations. These individuals are unaware of the progressive emergence of higher-order structures, although it is they who create them. There is an underlying dialogue between the individuals and the structures they create.' (SolŽ & Goodwin, 2000: 155)
This communicative aspect is crucial to the stability of the system. Ilya Prigogine and Isabelle Stengers indicate that there is an antagonism between stable and instable forces; as a system's complexity increases, fluctuations also increase in parallel (SolŽ & Goodwin, 2000: 157). The delicate balance of an ecosystem reflects these principles, small changes (environmental or as a result of human action) can produce radical shifts in the food chain or in species survival. 


On Darwin:

Darwin's theories of evolution are much cited but also misunderstood. SolŽ & Goodwin describes its key contributions as the recognition that there is an intrinsic (chemical) source of variability in living organisms, and that population size is limited by the finite nature of the resources available (2000: 248). The resource limitations clearly relates back to the issue of the variability of the organisms, itself generating through reproduction. Darwin's The Origin of Species articulates this emergence as evolution by natural selection - as a 'historic process of successive accumulation of changes' (2000: 249) or increased divergence as a result of being able to adapt to environmental changes. It is Stephen Jay Gould who advocates contingence as part of the process, taking into account nonlinear dynamical aspects of life as 'network dependent' (in SolŽ & Goodwin, 2000: 269). It is important not to conflate this critique with the reactionary creationism of fundamentalist Christians - whose sophistication stopped with the first chapter of Genesis (Bateson, 2000: 434). Cybernetics first introduced the concept of self-corrective systems that somewhat complicated matters. It was Russel Wallace who in effect proposed the first cybernetic model (at the time of Darwin) by confirming 'that natural selection acts primarily to keep the species unvarying', but that at higher levels, it may act to keep 'constant that complex variable which we call "survival"' (Bateson, 2000: 435). 

Artificial life:   
 
Some of this can be demonstrated in 'artificial life' scenarios and the study of life-like phenomena. The term is derived from John Von Neumann's experiments in constructing a self-replicating automata, and applied to biology and evolution by Thomas Ray in his studies of complex ecologies. His model, Tierra, is a set of computer programs that compete for processing time not food. SolŽ & Goodwin explain the process: that a short ancestral program evolve into diverse forms of increasing length and behavioural complexity (2000: 272). In time, social behaviour emerged as did extinctions as a result of parasites that relied on others to reproduce. What researchers in this area have found is that the more robust organisms were the more complex ones, able to withstand the effects of mutations. Steve Grand made a hugely successful game called Creatures in 1992, in which synthetic life-forms were propagated over online networks by a community of non-experts in the field (2001: 10). Like non-artificial life, interactions at all levels are fundamental to the health of the system.

This is no simple Frankenstein scenario, although Steve Grand would point to the parallel problem of the missing 'Žlan vital' or soul within artificial life (2001: 2; and he is making a distinction from 'A-life' which is the scientific discipline rather than the quest to make artificial creatures). It is not perhaps surprising that at the time of Mary Shelley, electricity, as the new technology of the time, held the imagination in terms of possibilities of generating life from seemingly lifeless matter. After all, in 1780, Luigi Galvini had demonstrated that electricity passed through amputated frog's legs could re-animate them. Grand sees this as part of the quest for the understanding the nature of life. He describes how in the middles ages, it was thought that a special chemical substance held the answer. Grand sees this 'vitalist' perspective associated with a materialist one, in seeking to find an embodiment of the life force or a physical form for something as spiritual as the life essence (2001: 4). As a computer programmer, he sees the logic of regarding mechanical processes as capable of human endeavour and recognising this can be traced historically back to Charles Babbage's Analytical Engine, itself interpreted by Ada Lovelace as an early computer that might 'think' (herself the wife of the poet Shelley, whose friend Lord Byron was the father of Ada Lovelace). For Grand, materialism is truth of the matter but not the whole truth - it can provide answers that are too simple rather than complex (2001: 5). 

Grand sees the task of artificial life of putting the life back into technology (2001: 9). Incidentally, in Marxist terms, one might equate this with putting life back into, or re-animating, dead labour. Clearly societies can be seen to regulate themselves through laws and social mores in a similar way to the ways an organism regulates itself through catalysis and adaptation (Grand, 2001: 71; 'catalysis' is the process by which something facilitates or speeds up a chemical reaction without destroying itself in the process - hence 'autocatalysis' where the parts themselves further perpetuate the process as feedback - just like adaptation to change is feedback too). Persistent phenomena such as this, through feedback use a form of predication that can be seen to be like learning. 

This is not simply a 'sausage-machine' as Grand calls it to undermine some of the top-down models of developing computer programs (implying a commercial imperative too unintentionally). His preference is for emergence stressing that things emerge from data not from code. This is counter to the accepted logic of computer science in which it is assumed that code drives data. Although computers deal with the language of processing, they also deal with structures. He says: 'Instead of telling the computer what to do, we are telling it what to be' (2001: 103). This is a different sense of programming in which behaviours emerge in a less deterministic manner. Here he is making the distinction between algorithms where instructions are followed and second-order structures where groups of objects act in parallel (by 'parallel' he simply refers to interactions with a population - this makes them potentially 'intelligent' for Grand, or behave as if it were alive, 2001: 106) as if there are many computers running simultaneously. 

This is what AI research has been attempting, running serial algorithms to demonstrate the ability to think. In most cases, these programs or 'expert systems' demonstrate the stored intelligence of their programmers, but fail as demonstrations of intelligence. To show intelligence, to Grand (and evoking Castells formulation unwittingly), the program must not simply follow rules but make them (2001: 110). To do this, it is important to recognise that the human brain, that gives the appearance of intelligence, is only ever a network of billions of very stupid machines. A computer program can work in serial and parallel - both synchronically and diachronically - and in this way operates as a network of behaviours from the 'inside-out' with no central 'outside-in' controller. It is the interaction between components at all scales that make a living organism less and less stupid. Grand sees this a critical issue in artificial intelligence in that it is fruitless to try to develop one function at a time, they all need to be implemented simultaneously in order to build an integrated whole machine otherwise it will not be integrated (2001: 169). He emphasises the point:

'Life is not the stuff of which it is made - it is an emergent property of the aggregate of that stuff. Even the stuff itself is no more than an emergent property of a still smaller whirlpool of interactions. Living beings are high-order persistent phenomena, which endure through intelligent interaction with their environment. This intelligence is a product of multiple layers of feedback. An organism is therefore a localized network of feedback loops that ensures its own continuation.' (Grand, 2001: 175)
   
randomness (add to section on this as distinct from emergent phenomena):

Randomness is a notoriously poorly defined term. Grand says 'just because something is indeterminate (by which I mean it cannot be known), we must not conclude that it is undetermined (has no prior cause).' (2001: 248). Randomness is an effect that has causes that simply remain unknowable or irrelevant. So even random events have definite causes. In the case of Grand's game Creatures, the user is part of the system as a form of external random noise but the complexity of the rest of the behaviours are entirely determined and yet as a whole unpredictable. 

On the question of order emerging from apparent random events, Bateson looks to Genesis for a clear articulation of the separation of land and water and species as a problem of description. This leads to the question: 'If random events lead to things getting mixed up, by what non-random events did things come to be sorted?' (2000: 343; in an essay about creationism). Do we call it evolution or something else like creation? And if so what kind - from the many creation myths around the world to contemporary theories of evolution.
  

Complexity and Markets:

The same logic can be applied to the marketplace, where products and companies compete for dominance. SolŽ & Goodwin describe the mathematical model of George P—lya (2000: 280). The formulation considers an urn with one black ball and one white ball. The rules are simple: if you pick one ball you return it with another ball of the same colour. When this is repeated, the system reaches well-defined steady states but their number is infinite with any possible ratio of white to black balls possible. This has been used to explain how the economy follows dynamical patterns, and self-organises in space and time. Speculators follow this logic in trying to predict the behaviour of markets as nonlinear systems - displaying predictable and unpredictable elements - as well as trying to understand collapses and extinctions (the dot.com crash for example).