Books

Short works

Books : reviews

Eric D. Schneider, Dorian Sagan.
Into the Cool: energy flow, thermodynamics, and life.
University of Chicago Press. 2005

rating : 2.5 : great stuff
review : 4 December 2011

When someone says that life, or evolution, or complex systems, or whatever, doesn't obey the second law of thermodynamics, it's a pretty safe bet that person has no idea what the second law of thermodynamics is.

If someone points out to you that your pet theory of the universe is in disagreement with Maxwell's equations—then so much the worse for Maxwell's equations. If it is found to be contradicted by observation—well, these experimentalists do bungle things sometimes. But if your theory is found to be against the second law of thermodynamics I can give you no hope; there is nothing for it but to collapse in deepest humiliation.

-- Sir Arthur Stanley Eddington,1928

For one thing, the second law applies only to isolated systems. If you isolate a living organism from its open environment, it will equilibrate to a state of maximum entropy (die, decompose, decay). That life obeys the second law is clear to anyone with a knowledge of physics; what this book does is take that further, and explain how complex self-organising systems, including life, actually accelerate the increase in entropy, by being efficient gradient reducing dissipators. Far from life and the second law being incompatible, the truth is that the study of life requires the study of thermodynamics.

p4. life can be regarded as one of a class of complex systems ruled by energy and its transformations. As the science of energy flow and chemical kinetics, thermodynamics is crucial to understanding life. Theoreticians who want to understand energy flow and transformations in biology must look the science of thermodynamics in the eye, as any theoretical claim is meaningless unless it conforms to thermodynamic principles.

This book is specifically about non-equilibrium thermodynamics (NET): open systems existing away from equilibrium, self-organising to exploit some form of gradient (a difference across a distance). They can exploit the gradient to do work. This tends to dissipate the gradient, but if it is maintained by some environmental source (the most clearly obvious being solar energy), they can exist in a steady state, stable but not static, "feeding "off the gradient. In fact, they organise themselves to help ensure this.

p18. Free energy is the quantity of energy available that organisms can put to work. … This … is directly proportional to the gradients machines can tap into---or organisms can use to maintain and reproduce themselves as specific kinds of material organizations.

And open systems are qualitatively different from closed, isolated systems.

pp70-71. Open systems enjoy energy and material influx and outflow across their boundaries. They are systems that, instead of reaching a predetermined end of equilibrium and disappearing, accelerate the reaching of equilibrium in the areas around them. Whereas isolated systems predictably head toward ruin, such systems are rare. Almost all real systems except those studied in the classical period of thermodynamics are open.

The book gradually builds on ideas, with many examples, to reach its conclusion. Summarising the final 4 page summary of the argument:

  1. In the linear near-equilibrium [Onsager] region, reciprocity relationships exist in which forces and fluxes are coupled. ...
  2. In the Onsager linear regime the total entropy production of a system due to material and energy flows reaches a minimum at the nonequilibrium steady state. ...
  3. Power is conserved in the system. ...
  4. As systems are moved away from the linear near-equilibrium by imposed gradients, they will use all avenues available to counter and degrade the applied gradients. ...
  5. As the applied gradients increase, so does the system's ability to oppose further movement from equilibrium. ...
  6. Systems moved away from equilibrium by greater gradients will be accompanied by increasing energy flows and higher entropy production rates. ...
  7. Open nonequilibrium systems reside at some distance from equilibrium, produce entropy that is exported out of the system, and maintain a low-entropy level inside the system ...
  8. If a gradient is imposed on a system, and kinetic conditions permit, autocatalytic or self-reinforcing organizational processes and structures arise. ... it is the second law and the behavior of nonequilibrium dissipative systems that is the main generative force for complex dynamic organizations in nature, including those of autocatalysis. ...
  9. Biological systems optimally capture energy and degrade available energy gradients as completely as possible.
  10. Biological processes delay the instantaneous dissipation of energy and give rise to energy and material storage, cycling, and structure. ... The most successful dissipative autocatalytic structures degrade available gradients in such a way as to maintain their gradient-degrading capabilities. ...
  11. Life and other complex systems not only do not contradict the second law but exist because of it. Moreover, life and other complex systems reduce preexisting gradients more effectively than would be the case without them.

On the way to this summary we get an interesting argument, peppered with little gems and insights such as:

p2. tough species are not necessarily representative of the health of their surrounding ecosystems. For example, some of the hardiest organisms belong to pioneer species that repopulate damaged ecosystems. Such organisms thus may signify not health but ecosystem illness.

p203. Cycling of material and energy within the ecosystem changes during succession. In early succession the cycles are short, open, and fast. In mature ecosystems the material and energy cycles are just the opposite, with long complex cycles that are closed in upon themselves.

But don't believe everything you read here.

p35. In 1714 Daniel Gabriel Fahrenheit introduced the first scale to the mercury thermometer. The low mark on this scale, the lowest he could obtain in the lab, was zero, thirty-two degrees less than the freezing point of water---the temperature of ice.

Fahrenheit didn't just introduce the temperature scale: he invented the mercury thermometer itself. And the rather meaningless "temperature of ice" should be "freezing point of a water-salt mixture" (or something even more technical). And on p45 there is mention of a small hand-held "Sterling engine [sic]" (I suspect it is this one with a misspelled filename on wikipedia, although the correct spelling is used in its description). The authors also appear to misunderstand gravity:

p51. gravity, producing gradients at a cosmic scale, must challenge any claim that the second law is an inexorable, omnipotent force

But even gravitational collapse (from a uniform gas cloud to clumpy blobs) does, of course, increase entropy. As well as the clumpy final state, there is entropy produced from radiation (the collapsing gas heats up) and/or material being ejected away (evaporated) from the central part. So although gravitational collapse might be considered an example of a case where nature does not "abhor a gradient", the second law itself is still "inexorable".

There is another, minor, misconception that needs correcting:

p175. Sir Frederick Hoyle, the astronomer, and his Sri Lankan colleague Chandra Wickramasinghe (1984) were so bowled over by the improbability of cells "snapping together" on Earth that they compared it to the assembly of a 747 jet from detritus in a junkyard by a passing tornado. Their solution to the problem was to increase the available arena for biogenesis by removing it from Earth to space. … However, if one assumes random interactions of the atomic or molecular components necessary to construct even a single minimal bacterial cell, even the seemingly adequate expansion in time from the 4.6-billion-year-old Earth to a 15-billion-year-old cosmos does not prove adequate.

I've seen this criticism in other places, too. I must speak in defence of Fred Hoyle here (even if he didn't understand evolution). He was by no means suggesting, as implied here, that a measly factor of three increase in timescale was sufficient. 15 billion years is the time since the Big Bang, the beginning of the universe in our current cosmological model. Hoyle, who himself coined that name, had his own preferred, Steady State, model of the universe. In that model, the universe has existed forever. So the timescale increase Hoyle was proposing was from 4.6 billion years to infinity: plenty long enough for a tornado to assemble a 747, or anything else!

These kind of errors in areas one knows about lead to less confidence in believing the statements in the areas one does not know about. Additionally, in many places the text is inconsistent (on p221 the units are calories/cm2/minute, and in the accompanying figure 15.2 they are W/m2, and calories/tree/day); also the prose is repetitive, purple, and elliptical, making it harder than necessary to follow the argument. (Note to authors: metaphors are of little use if you don't make clear the mapping from one domain to the other, but leave the reader to guess it, and are of even less use if the reader doesn't understand the metaphorical domain either.)

But despite these content and style caveats, the book is still well worth reading. It contains a wealth of detail, drawing ideas from a great range of complexity scientists, thermodynamicists, ecologists, systems biologists, and more. As just one example, we learn about the more modern understanding thermodynamics: the irreversibility formulation of Constantin Carathéodory, the network thermodynamics of Don Mikulecky, the dissipative systems of Ilya Prigogine. Whilst the traditional formulation was useful for analysing steam engines (which is, after all, essentially why it was invented), this new approach is more applicable to non-equilibrium open systems.

p81. Prigogine popularized the term dissipative structures. These dissipative systems, a term first used by Lotka, maintain their stable, low-entropy state by importing material and energy across their system boundaries. Dissipative systems are nonequilibrium, open, dynamic systems with gradients across them. They degrade energy and exhibit material and energy cycling. Dissipative structures grow more complex by exporting---dissipating---entropy into their surroundings

The book starts off discussing non-living, but nevertheless complex, systems (such as tornadoes). These clearly depend on a gradient, and their complexity can be attributed to the existence of this (contextual) gradient.

p85. Most "self-organizing" systems feed on free energy from the outside to maintain their organization: they are organized by the gradients they reduce---often they are better described not as self-organizing, but as gradient-organized systems with self-referential attributes.

p116. Bénard cells are a striking reminder that complex systems do not come from nowhere, but from pre-existing gradients. Their complexity does not arise from within but from their context.

Life is another such complex system, but it has more capabilities than tornadoes or Bénard cells. For one thing, it can "time shift" the energy it needs:

p107. Stored energy such as fat, starch, and glycogen frees the organism from the imperative of immediate gradient breakdown.

And it can evolve to exploit energy sources, such as other organisms, or other organisms' waste products. Here the authors focus on a fundamental difference between material (nutrients) and energy (initially sunlight):

p85. Far-from-equilibrium systems pay for their reduced entropy by exporting a concomitant increase in entropy into the surrounding environment. A most familiar, if troublesome, example of such necessary external disorder is pollution. All organisms, not just human technological ones, produce waste. ... There are no recycling bins in nature; everything is elegantly used and reused because organisms, however mindless, have evolved to make use of relatively limited materials in an environment of relatively unlimited energy.

p85. "Energy flows; matter cycles," ... Harold Morowitz (1997,121)

p148. In every instance considered natural selection will so operate as to increase the total mass of the organic system, to increase the rate of circulation of matter through the system, and to increase the total energy flux through the system, so long as it is presented an unutilized residue of matter and available energy ... (Lotka 1922, 149)

We might be more appreciative of this matter/energy distinction if we were plants:

p186. The kind of energy required for organisms to maintain their bodies, their metabolism, is strictly limited. The list includes light (photoautotrophy), organic chemical energy (heterotrophy), and a very limited number of inorganic energy-yielding chemical reactions ... Organisms also need food, which forms the stuff of their bodies. Energy gets used up; food is transformed into matter and materials of the body. One of the reasons we tend to be confused about these things is that we animals don't distinguish food from energy in our metabolism. In animals the source of energy and food is the same (sugars and other carbohydrates, amino acids, and proteins). In plants, however, the sources of energy and food are entirely different; sunlight is the energy source and carbon dioxide, chemically converted to sugars and other materials, is the source of food.

Because the material is relative scarce, it must be used efficiently.

p203. In the rain forest even moisture is recycled. Early morning and noontime temperatures, clear skies, and sunlight lead to afternoon showers, a rapid-response cycling system. Early successional systems do not have the root systems, leaf biomass, and organic material to make this recycling possible. A key attribute of a mature ecosystem is that it does not leak a lot of its nutrients, material, and water from the system as it recycles these materials.

Diversity enables efficiency.

p244. The climax ecosystem is a system of energy fixers, photosynthetic food makers, and energy-degrading herbivores. Entropy production takes place both in the making of food and its consumption. Energy is degraded during both photosynthesis and transpiration.

[One thing all this discussion of maximising recycling and reuse of material nutrients: are animals, and even carnivores, inevitable in a mature ecosystem? So, would alien ecosystems necessarily have animals? Are they necessary to maximise the recycling, or could it all be done with bacteria and viruses?]

When the 2nd law is cast in terms of open systems with inputs and outputs (rather than the original, isolated systems formulations), some problems disappear, and we can see our "selves" in a different way.

p92. mathematical descriptions of ecological and evolutionary trends are, we argue, likely to follow increases in rates and flow of energy and materials, which cycle, increasingly and expansively, in growing thermodynamic systems. ... there is no need for a new fourth law when the second, stated for open systems, suffices.

p112. Selves are not closed or isolated but arise as metastable open systems in a sea of energy and flows. ... Thermodynamic selfhood comes from dissipative systems that establish boundaries. Far from sealing themselves off from the outside world, their boundaries allow them to continue their operations. Biological self-hood on Earth depends on the semipermeable layer, the ubiquitous lipid cell membrane ...

p311. … organisms remain open systems. This means that their selfhood is already always open to encroachment, both from the outside where they enter into alliance with other beings, and from the inside where renegade cells, as in cancer, can spread without concern for the good of the genotype to which they belong. More importantly, a gene by itself is not a self; it replicates only as part of the reproduction of a thermodynamic system sufficiently coherent to access energetic gradients. ... Without that integral association, a gene is just a chemical, as indeed are the inert crystals of viruses separated from the active teleonomy of living cells. … organisms are nested hierarchies, composed of units that were once selves in their own right.

How can we resolve the inexorable linear march of the universal 2nd law with biological evolutionary contingency and parochialism? Well:

p152. The situation is similar to that between a Boltzmannian microstate and a thermal macrostate. The specific pattern of the particles composing the biosphere is our own, as is that of an individual's body and life. However, the tendencies of which these particles partake show universality.

Because of the strong thermodynamic driver in the organisation of living systems, we may be able to see traces of earlier processes in current ones:

p94. Building up complexity over time, energy-driven cycles embody a natural memory and record of their past states. … The chemical cycles of modern cells … may contain traces not only of their bacterial ancestors but of the thermodynamic cycles from which bacteria themselves evolved.

The complexity of living systems originated in the physical embodiment (metabolism), and only later was the biological information (for replication) exploited to make the gradient reducing process even more efficient:

p169. The hardware by itself can exist, metabolism can continue as long as there is energy flow, but software cannot exist by itself; it is a virus, an "obligatory parasite." ... In life the hardware consists of the body, the software of the genes. The thermodynamic cells and replicating nucleotides we find together today are logically, and historically, separable. Just as one can imagine a computer without software, so one can picture early thermodynamic life without any genes. First came the apparatus, functioning and physiological, then came the operating systems, user manuals, and codes for making new, improved metabolic machines.

p251. Variation offers new possibilities, and natural selection whittles them down to potent systems adapted to the current environment. But the original impetus---to make do with available materials to dissipate high-energy systems as efficiently as possible---is that of the second law. Even before natural selection, the second law "selects" from the kinetic, thermodynamic, and chemical options available, those systems best able to reduce gradients under given constraints.

More mature ecosystems are more complex, and have a greater energy flux, and greater entropy production (that is, life accelerates the 2nd law effect). Stressing them sends them into a less complex state, with a lower energy flux.

p198. One of the most obvious features of ecosystem change during succession is the increase in biomass over time. ... The biomass of an ecosystem increases, stabilizes, and levels off. .... It grows. The more energy captured and flowing through a system, the greater will be the number of entropy production processes such as transpiration, photosynthesis, and metabolic reactions, which all degrade in-coming energy. In ecosystems, biomass, total system throughput, and entropy production go up with succession and maturation.

p208. In both living and nonliving systems the reversion to earlier modes is triggered by reduced energy flows. Stress sends the gradient-reducing system back to earlier modes able to make do on less energy.

This increasing complexity producing ever more efficient energy degradation defines a direction for evolution.

p236. There is no essential mystery to these heightened activities or to the thermodynamically based progressive arrow to evolution. As organisms, tapping into the solar gradient, expand their activities through reproduction, which is the offshoot of thermodynamic metabolism, they bestow greater complexity upon the environment---even as they export molecular chaos and heat to the wider realm. Maintaining themselves in a low-entropy state, and growing, they expand the reign both of complexity, locally, and of disorder, pushing heat out into space. Ultimately the thermodynamic system reaches the limits of its growth, and the energy formerly used in expansion is redirected internally, showing up as diversity, differentiation, and increased cycling.

p239. Evolution's direction is that of the equilibrium-seeking organizations of open system thermodynamics.

p316. complexity accrues not only from the "bottom up" via natural selection of replicating variants but thermodynamically from the "top down" via gradient breakdown

The final part of the book focuses on economic systems and how they might fit into the model.

p276. Economic flows also have their entropy production surrogates, such as transaction fees, taxes, and legal fees. These overheads often add little value to the product or number of energy units transferred but tend to smooth transactions. ... Money, tradable for energy, work, and products, behaves like energy changing forms as it organizes flows through nonhuman natural systems.

p277. … markets and economies are not themselves closed systems ... Markets, like organisms, depend upon external sources for their complexity-maintaining gradient-reduction activities. Economies, like ecosystems, expand via the natural wealth of the gradients around them, and prosper relative to their ability to develop conscious or unconscious mechanisms to degrade those gradients. ... Economies and markets, like organisms and ecosystems, are metastable, nonequilibrium systems.

p279. the destruction of a gradient to release its energy requires energy … thus it is that making money requires money, and finding food requires catabolic burning of metabolic fuel.

p284. the economy is a system organized by energy flows. A true NET economics would recognize that economies, insofar as they are stable, are stabilized away from equilibrium by material and energy flux. Such an economics might require a modified accounting system that included urban flow requirements. As Dyke (1988, 365) suggests, "... We never, except in the most superficial ways, examine the relationships between our patterns of social organization and the rate of material flow needed to sustain them."

This rounds off the discussion that covers physical, biological, and social complex self-organising systems (reminding me in places of De Landa's aproach). There's a good bibliography, for when you want to dig deeper. And it's an extensive bibliography, because to fully understand life, we need to consider (at least) biology, chemistry, and thermodynamics:

p303. Evolutionary theory links organisms in time. Ecology links organisms in space. Chemistry links them in structure. NET links them in process.
[stable on a gradient]

So, life (and any other complex self-organised system) is a non-equilibrium thermodynamic system and process, living off a gradient. I'm left with an image of forever walking up a down escalator -- it takes effort to stay in place, but flowing down the escalator is all energy and material needed to support this effort.