EVOLUTION, ENTROPY AND WORK
James Fournier
Ph.D. Candidate
California Institute of Integral Studies
1453 Mission Street, San Francisco CA 94103
jim@geoman.com
Research Paper
The discussion of the thermodynamics of evolution dates back at least to Schrodinger’s (1945) seminal
work, What is Life? which appeared long after an obscure but groundbreaking article by Lotka (1922).
There has been a great deal of discussion of the thermodynamics of evolution over recent decades, and yet many
issues remain unresolved, especially surrounding the question of entropy. The topic of this paper, involves the basic
thermodynamic calculation of entropy, and specifically the relationship between energy, heat flow, entropy and
work. It would appear that in their haste to develop a comprehensive theory explaining all of evolution, recent
theorists have lost sight of the thermodynamic concept of work in their treatment of entropy (Swenson, 1989/1997).
This is particularly ironic, as thermodynamics was originally developed in an effort to better understand the
interrelationship among energy, heat flow, and work.
Keywords: Thermodynamics, Evolution, Energy, Entropy, Work
1. A Review of Basic Thermodynamics
Thermodynamic entropy has units of energy over absolute temperature. If the term is to have any meaning this is
where we must begin. There may be metaphors for entropy, which may be useful in information theory or in other
approaches to understanding complex systems, but if we label these with the word
“entropy,” which has a very specific physical meaning, we lose sight of any real value of the term. Similarly,
if we wish to engage in a thermodynamic analysis of a problem we must be careful to use words which have very
specific technical meaning appropriately. A specific example of this is the word
“efficiency.” Efficiency has a very specific technical meaning in thermodynamics. It is intimately related
to the concept of work. A brief review of basic thermodynamic concepts is in order.
From the First Law, the basic formula describing a simple thermodynamic system, or heat engine, is:
Q(h) – Q(c) – W = 0 [eq. 1]
where Q(h) is the heat flowing in, Q(C) is the heat flowing out, and W is the work done by the system. or, in terms of Q
(h):
Q(h) = Q(c) + W [eq. 2]
This says that the heat flowing in from the high temperature reservoir is equal to the heat flowing out to the low
temperature reservoir, plus the work done by the heat engine. The key insight which led to the
development of the Second Law is that W, the work done by the system, must always be less than Q(h), the heat
flowing in from the high temperature reservoir. In other words, all of the heat flowing in cannot be converted to work,
some heat must be discharged and flow out to the cold reservoir. It is the ratio of the work done by the system vs.
the heat flowing in, which determines the efficiency of the system:
Evolution, Entropy and Work
W
Efficiency = -------- [eq. 3] Q(h)
The maximum theoretical efficiency for any system is: T(h) – T(c)
Efficiency max = ---------------- [eq. 4]
T(h)
For any designer of heat engines it is the amount of work, and therefore the efficiency, which are of
greatest interest. It soon became obvious that some portion of the thermal energy, or any type of energy input, must
be discharged in the form of heat to a colder reservoir, or sink, in order to extract any work at
all from a high temperature reservoir or energy source. It was this observation which led to the articulation
of the Second Law of Thermodynamics: For any irreversible process, entropy always increases.
In our system, a heat engine and two heat reservoirs, the change in entropy, delta S, in a complete cycle is: Q(c)
Q(h)
delta S = ------- – -------- [eq. 5] T(c) T(h)
or substituting for Q(c):
Q(h) – W Q(h)
delta S = -------------- – -------- [eq. 6] T(c) T(h)
For a theoretical reversible cycle delta S = 0. In this case, the decrease of Q(h) by W in the numerator of the first
term is exactly proportional to the smaller denominator represented by the lower temperature reservoir, T(c),
and the terms cancel in Equation [6]. In a theoretical reversible cycle the work done, W, is
a maximum, and the change in entropy is zero.
For the irreversible, and therefore real, case W is always less than the maximum reversible value.
Therefore, the numerator of the first term is larger, and delta S > 0, as required by the Second Law.
Reducing W increases the magnitude of the first term, thereby increasing the value of delta S. When W is zero, Q(c)
is equal to Q(h), all of the heat is flowing out to the cold reservoir, and delta S is at a maximum,
as shown in Equation [5].
From the First Law, the heat flowing in, Q(h), must be equal to the heat flowing out, Q(c), plus any work, W, done by
the system; as the energy flowing in as heat must be equal to the energy flowing out as some combination of heat
and work. In addition, the Second Law says that some portion of the energy flowing out must be heat; the heat
flowing in can't all be converted to work.
It becomes apparent from Equation [6] that the magnitude of the increase in entropy, delta S, is directly linked to the
amount of work done by the heat engine on its environment. If it does no work, then Q(h) is equal to Q(c) and the
entropy increase in maximal. However, if the system is as efficient as possible (in the thermodynamic sense of the
word), and therefore does as much work as possible, then the increase in entropy is minimized. If the process
could be truly reversible, then the increase in entropy could, in theory,
be zero. In reality it is impossible for any actual macroscopic processes to be fully reversible. What is
most important to focus on here, however, is the direct negative relationship between the magnitude of the increase
in entropy, and work done by the system. Work most often represents that portion of the process which is
reversible. Work can be used to de-disperse things, to put things back into a state of higher potential
energy. Whether by lifting something with respect to gravity, or by making chemical bonds of a higher potential
energy, it is work which can reverse the normal tendency of things to dissipate toward thermal equilibrium.
2
Evolution, Entropy and Work
2. Thermodynamic Models of Evolution
Much of the recent theorizing about the thermodynamics of evolution attempts to build a model up from non-living
systems.(Goerner, 1994) In the end this may prove to be fruitful, but it may be easier to first start from the other
direction. When humans build machines we are mimicking what nature does, at least in
so far as we are faced with the similar constraints as those facing any other organisms attempting to further their own
survival. Therefore, we need only look at the most basic thermodynamics of the situation to see that the problem,
both for humans and for other organisms, has always been how to extract as much useful work from a potential
energy gradient as possible. There are two ways to do this. One is to increase the magnitude of energy flow across
the gradient. The other is to increase the (thermodynamic) efficiency of
its use, i.e. maximize the work extracted and therefore, by definition, minimize the entropy production. The
first approach can lead to increased entropy production, if the thermodynamic efficiency of the process remains
constant. But the second approach, which incidentally seems to be highly optimized in natural systems, may
actually lead to a reduction in the entropy increase. In any case, the increase in entropy is an indicator after the fact,
not a cause of the behavior of the system, which may be much better understood in terms of energy flows, work,
and the overall thermodynamic efficiency.(Corning, 1997a) Indeed, it becomes plausible to hypothesize that
in strictly thermodynamic terms living systems are the work extracted from the energy flux across the
potential. However, it might be more correct to say that their structures are composed of the energy
embodied by that work in matter. The following thought experiments may make this more understandable.
2. The Benard Cell Experiment
First, let us examine the now paradigmatic Benard cell. The Benard cell experiment is carried out in a flat circular
dish filled with a thin layer of water, which is uniformly heated from below in a controlled manner.
At a certain critical rate of thermal input, if the water layer is precisely the right thickness, the heat will be dissipated
more rapidly through the formation of a coherent pattern of hexagonally arrayed convection cells than through
turbulent boiling.(Benard, 1900) In this case we have a system in which energy moving across a potential
gradient spontaneously gives rise to an ordered pattern or structure. While this may be a
very useful critique of the 19th century idea that all physical systems tend toward a state of greater disorder,
it really has very little to do with the thermodynamic behavior of living systems. Its repeated invocation
seems to be an artifact of early thinking about dissipative structures.(Prigogine, 1984; Swenson,
1989/1997) The Benard cell does, however, also inadvertently illustrate a key distinction which should be
recognized when comparing various types of systems. The question becomes one of whether or not the
system does work which contributes to structures that endure in matter. This distinction is of pivotal
importance to living systems. Living systems create structures that do endure. While perhaps so obvious
as to be taken for granted, this distinction allows us to recognize why the Benard cell is not an appropriate example
for describing the thermodynamic behavior of living systems. In order to create enduring structures in matter
living systems must do work. Thus, if we are looking for thermodynamic examples of living system, we must look for
systems that do work and create enduring structures. This appears to be a necessary, but not sufficient, condition
for describing living systems. Thus, perhaps it might prove to be an interesting criterion to employ in pursuing an
investigation into how non-living systems could have evolved into living systems. That question, however, is not the
focus of this paper.
In the case of the Benard cell, one might be able to arrange a frame of reference in which it appears that work is
being done to create, and temporarily maintain, the coherent pattern among the convective cells. This coherent
pattern is, as has been pointed out repeatedly over the last few decades, what allows the system to
dissipate the available thermal energy most rapidly (Swenson, 1989/1997). However, unlike a living system, this
coherent pattern does not contribute to the establishment of any enduring structure in matter. As soon as the
source of high energy potential is interrupted, the pattern immediately collapses, and any energy which had been
contributing to maintaining the structure, is immediately converted to thermal dispersion and dissipated. Thus,
the claim that the Benard cell creates increased entropy as rapidly
as possibleoften erroneously referred to as “efficiently”appears to be a correct (Swenson, 1989/1997). But, while
this is an interesting physical phenomenon, showing that macro scale ordered patterns can
3
Evolution, Entropy and Work
spontaneously emerge out of the dynamics of energy flow across a potential gradient, it has nothing to do
with the thermodynamic behavior of living systems. Living systems, by contrast, use part of the energy flowing
through them to do work. To the extent that this work is used to create structures in matter, that portion of the
energy which does the work is sequestered in matter. That portion of the energy does not flow into a low
temperature reservoir, and thus does not immediately contribute to entropy production. Thus, in many cases
living systems actually create less entropy, than comparable non-living systems operating across the same
potential gradient over a given period of time. It seems that the second law cannot be violated, but it can be
stalled. In that sense life could be regarded as an energy “kiting” scheme.
3. Photosynthesis
This becomes particularly apparent when one examines the thermodynamic behavior of organisms engaging
in photosynthesis. Here energy from photons coming from the high temperature source, namely the Sun, is
converted to potential energy in the form of excited electrons and used to do work, namely creating higher
energy molecules in the form of carbohydrates and molecular oxygen in the atmosphere. The energy represented by
photons which are actually captured and used for photosynthesis, about twelve percent of those actually striking a
leaf, are converted into chemical potential energy.(Corning, 1997b, p.5) This energy is bound up in the carbon
compounds which make up the structure of the organism, i.e. organic material, and in the higher chemical potential
energy state of the atmosphere. This portion of the incident solar energy is not immediately reradiated in the form of
thermal dispersion, and thus does not immediately contribute to increased entropy. Therefore, it would appear that
throughout the time when there have been photosynthesizing organisms on Earth, the overall rate of entropy
production on Earth has in fact been less than it would otherwise have been without them. We can see the
difference in the energy sequestered in the vast accumulation of biomass and fossil fuels on Earth. Energy
which was chemically bound up on Earth rather than being disbursed into space as heat did not contribute to
increased entropy over that period.
5. Respiration
It is only because of the long term accumulation of energy by photosynthesizing organisms, and the
resultant slowing of entropy production, that a large energy reserve has been built up on Earth. It is the existence of
this reserve that has made it possible for other organisms to consume and transform that energy more rapidly. It is
only with the advent of aerobic organisms engaged in respiration that it becomes possible to burn up these
reserves more rapidly, and therefore to increase the rate of entropy production. If one only examines the behavior of
organisms engaged in respiration over the period following the onset of the increase in atmospheric oxygen
levels, one might conclude that there is a pervasive trend toward increased entropy production in all of
evolution (Swenson, 1989/1997). But, this is apparently a characteristic, not the cause, of a trend which
has only existed over some portions of the evolutionary history of life on this planet. Moreover, in spite of this
recent trend, wherein the oxidation rate associated with the respiration of ever more complex organisms has
been steadily increasing, the overall rate of entropy production of the planet as a whole has still remained less
than that of a comparable planet without life. The evidence for this observation is found in the overall
tendency of the Earth to continue to accumulate carbon-based biomass, while maintaining a stable concentration
of atmospheric oxygen. This represents the sequestering of energy which would otherwise have been reradiated
into Space contributing
to increased entropy, but which instead did not do so. A more interesting observation in this area concerns the
steady increase in the oxidation rate per unit mass by ever more complex aerobic organisms. This trend toward
increasing energy flux density in increasingly complex organisms is most interesting, but does not
by itself lend any real support to the claim that entropy increases as rapidly as possible. There are two
reasons why this is the case. In the first place, one would have to know how much work was performed by those
organisms before one could make any meaningful calculation of entropy. Secondly, the trend itself is only valid
over a portion of the evolutionary history, and as has previously been pointed out, was only made possible
by a prior prolonged period of energy accumulation and collateral entropy rate reduction.
4
Evolution, Entropy and Work
6. The Planet as a Whole
The question of the overall thermodynamic behavior of the planet as a whole is also worth considering in greater
detail. The maximum possible increase in entropy for a given system is found when all of the thermal
energy flowing in as heat from the high temperature source, is also flowing out as heat to the low temperature sink.
This is essentially the case for a dead planet such as the Moon, or as far as we can tell Venus or Mars. In this case
all of the solar radiation which strikes the planet is reradiated out into Space within a relatively brief period. The only
other possibility is that some of the energy flowing in could go into heating the planet. But the dead planet
appears to be pretty much at equilibrium, or perhaps more accurately at a stable steady state, as there may
also be some residual thermal energy being radiated as heat from volcanic activity by the planet as it cools very
slowly. The source of this heat is largely nuclear decay within the core of the planet, although it is possible that there
may also be some leftover thermal energy from mechanical collisions when it was formed, or even conversion
of a small amount of gravitational energy to heat in movements of its crust.(Cloud, 1988) But the rate of heat flow
from all of these sources is stable enough that as a practical matter the dead planet is at a stable equilibrium
temperature averaged over any representative period.
It is also important to understand that changing the relative position of the temperature of the planet within the
temperature gradient between the Sun and Space has no effect on the ongoing change in entropy. Any increase on
one side is exactly offset by a decrease on the other, i.e. if the planet were colder and there was therefore more
entropy increase associated with photons from the Sun being absorbed by it, then that increase would be
offset by the reduced increase in entropy when photons were radiated from the colder planet back out into
Space. The overall rate of entropy production is entirely determined by the temperature difference between
the source, the Sun, and the sink, cold dark Space, and has nothing to do with the relative intermediate temperature
of the planet. The rate of entropy production is therefore already pegged at a maximum for any dead planet. There
is simply nothing for any incident energy to do other than
to be reradiated out into space, and therefore all of the energy leaving the Sun and striking the planet is already
contributing to the maximum possible rate of entropy production.
7. The Transition to Life
Once a planet has evolved photosynthesizing organisms, it is clearly operating at an overall rate of entropy
production which is less than that of a dead planet. This is the case because, as has already been outlined,
photosynthesizing organisms use some of the energy to do work to create higher energy compounds, which
sequester a portion of the energy rather than allowing it to be reradiated out into space. If a portion of the energy
striking the planet is not immediately contributing to the increase in entropy then the overall rate of entropy
production must be less than the theoretical maximum for any given period of time. What is more interesting is the
question of what happened during the earlier period before photosynthesis starts. More interesting still is the
question of what happened before there was really life per se, but after autocatalytic chemical reactions start
occurring. It would seem that as soon as such reactions begin to sequester significant amounts of incident
energy in the form of chemical compounds at higher than equilibrium potential energy, the process has
started. Indeed, most hypotheses for the spontaneous emergence of life out of some mixture of multiple
autocatalytic reaction cycles depend upon the existence of some sort of high-energy soup. The ultimate energy
source for the creation of such a soup might be some combination
of thermal radiation from the Earth and/or photons from the Sun. Even if this energy were converted into a hydrologic
cycle giving rise to lightening, or some other form, the key transition point in this model would
be when some portion of that energy began to be sequestered as complex chemical compounds, proteins or other
chemical precursors to life. From a thermodynamic standpoint it is impossible to say where in this sequence life
actually begins, even though from a biochemical, or at least from a biological standpoint it might be clear that such a
system could not yet be said to contain life. This begs the question of whether a living system or pre-living system
must contain life, or rather living entities. Indeed, as Gaia theory begins
to suggest the whole system might in some sense be said to be living, even before individual entities could
be said to be alive.(Lovelock, 1987; 1995) The key observation appears to be that living, and perhaps even pre-
living, systems tend to drive the planet away from its previous condition of maximum entropy.
5
Evolution, Entropy and Work
8. Human Intervention
It is only with the advent of human use of fire that the overall rate of entropy actually begins to increase on Earth.
The first indications of this might have been when hunting and gathering peoples began to manage large areas of
forest by regularly burning the undergrowth more extensively and frequently than had been the case due to
lightening strikes. From there humans eventually moved on to slash and burn agriculture, and eventually to larger
and larger scale deforestation. The desire for wood, both as fuel and building material, combined with the
need for more grazing area for cattle and sheep has driven a steady pattern of deforestation wherever human
populations have spread. Especially in Europe and the Middle East, the spread of human civilization has been
basically synonymous with the spread of deforestation for at least four thousand years. Yet, throughout this
period, while the rate of entropy production due to human activity was steadily increasing, total combustion
still did not even come close to approaching the point where the biosphere could no longer sequester far more
energy than was being dissipated as heat due to combustion. It has only been since the industrial revolution that the
overall rate of entropy increase on the Earth may have for the first time approached, or even exceeded, that of the
planet before life began. For it was only when humans began to tap, and burn, the fossil hydrocarbon reserves of
the planet, disbursing a large portion of that energy as heat, that the rate of entropy production on Earth spiked
upward. In this act
of essentially hyper-respiration human technology at once mimicked, and surpassed, the behavior of all
other aerobic organisms including ourselves. Prior to this point the biosphere, in combination with the
hydrologic and geologic systems of the planet, has apparently always been capable of maintaining a net
positive balance in the current account of energy flow, always sequestering a net surplus.(Cloud, 1988) Now,
in a relative flash on the scale of geologic time we are literally burning through those reserves, massively
accelerating the rate of entropy production, and indeed the rate of increase of the rate of entropy production. This
resembles more a rip, or tear, in the fabric of the energetic net life has so carefully woven, than an extension of life’
s natural functioning. This situation will most likely only make sense in retrospect, as a necessary and
even inevitable stage in our evolution, if we are able to recognize the implications of the situation now and
very rapidly adapt our technology and behavior.
9. Technological Analogs of Organisms
From the point of view of the energy flow and entropy analysis, which we have so far been applying
primarily to biological organisms, current fossil-fuel energy technology functions like a hyper-animal. Human-
made machines oxidize hydrocarbons to extract the embodied energy, while remaining dependent upon
photosynthesizing organisms to maintain atmospheric equilibrium. Biological respiration is essentially
combustion, but carried out in a controlled manner to yield the maximum thermodynamic efficiency. It seems
clear, given the preceding analysis, that we as a species are now facing a fundamental point of inflection in a very
long-term oscillation in the thermodynamics of the biosphere. This point is no less profound than when organisms
capable of aerobic respiration first emerged in response to the build up
of oxygen in the atmosphere. We now face the symmetrical point. This time, hyper-animals, i.e. our
machines, threaten to exceed the capacity of the atmosphere to absorb the CO2 they give off without
adversely destabilizing the narrow range of climatic temperature we require to support our vast population. The
current situation is in many ways exactly the opposite of the one which gave rise to our predecessors, and launched
the whole lineage of ever more complex organisms extracting ever greater amounts of energy with oxygen from the
atmosphere. Now, having made technological extensions of ourselves, which have pushed that process further, we
must implement technology which will bring the system back into balance.
The key insight may be to discern the fundamental difference between plants, which engage in
photosynthesis, and animals, which engage in respiration, and to recognize that there is potentially a
parallel classification system between animal-like technologies which engage in combustion, or more precisely
oxidation, and plant-like technologies which represent a technological analog of photosynthesis. The obvious
technological parallel will be with photovoltaic cells. Yet, photovoltaics by themselves cannot really be described as
meta-plants. Like plants, they do convert sunlight into available energy, but they do not directly affect the
atmospheric balance between CO2 and O2. If, however, their electrical output is used
6
Evolution, Entropy and Work
to perform electrolysis, splitting water into hydrogen and oxygen, then photovoltaics might represent plants
on one side of a potentially balanced energetic cycle, analogous to the carbon cycle in nature. A human- made
closed-loop analog of nature might even work without carbon, instead using only H2, O2 and water. Energy
technology may actually be headed in this direction already. The chemical composition of fuels have moved steadily
from wood, to charcoal, to peat, to coal and then through the sequence of successively smaller and lighter fossil
hydrocarbons from crude oil, to natural gas. This sequence inevitably converges
on pure hydrogen, H2, as the smallest lightest energy storage and transport medium. Hydrogen can be
highly problematic as a fuel for combustion, but the optimal technology for the use of H2 does not involve
atmospheric combustion. Oxidation is instead carried out in a controlled manner in a fuel cell. This is the
technological analog of respiration. Fuel cells combine H2 and O2 to yield an electric current and water. These
technological analogs of photosynthesis and respiration are based on silicon instead of carbon. Silicon, the
basis of rock, is essentially the higher octave of carbon in the periodic table. Thus, it might be appropriate to
describe such a new closed-loop cycle as an octave of nature, or even as “Meta-Nature.”
Baring the sudden advent of nuclear fusion or zero-point energy, the rapid deployment of such technology will be
necessary for human civilization to continue. While this is already obvious to many, it may shed new light on the
problem if we are able to understand it within the framework of the very long-term behavior of energy flows in
nature. Such analysis may illustrate, more clearly than previous approaches, that what we must now do is
essentially create meta-plants, systems which are capable of an analog of photosynthesis, in order to initiate a
new sustainable balance with the vast amount of human technology which already behaves like essentially meta-
animals. It may also be helpful to recognize that in general, living systems tend to maximize their access to
available energy, while striving to be as efficient as possible, in the true thermodynamic sense of the word. We
would do well to follow their example.
References
Benard, H. 1900. “Les Tourbillons Cellulaires dans une Nappe Liquide,” Rev. General Science Pur. Appl.
Vol. 11:1261-1271.
Cloud, P. 1988. Oasis in Space, W.W. Norton & Co.
Corning, P. A. & Kline, S.J. 1997a. “Thermodynamics, Information and Life Revisited, Part I To Be or
Entropy” Syst. Res. Behav. Sci. Vol. 14:0:1-22 John Wiley & Sons, Ltd.
Corning, P. A. & Kline, S.J. 1997b. “Thermodynamics, Information and Life Revisited, Part II
‘Thermoeconomics’ and ‘Control Information’” Syst. Res. Behav. Sci. Vol. 14:0:1-29 John Wiley & Sons. Corning, P.
A. in press. “On the Concept of Synergy and its Role in the Evolution of Complex Systems,” Journal of Social and
Evolutionary Systems
Goerner, S. J. 1994. Chaos and the Evolving Ecological Universe Gordon & Breach.
Lotka, A. J. 1922. “Contribution to the Energetics of Evolution,” Proceedings of the National Academy of
Sciences Vol. 8:147-151.
Lovelock, J. E. 1987. Gaia : A New Look at Life on Earth Oxford University Press
Lovelock, J. E. 1995. The Ages of Gaia : A Biography of Our Living Earth W.W. Norton & Company
Prigogine, I. 1984. Order Out of Chaos Bantam. Schrodinger, E. 1945. What is life? New York: Macmillan.
Swenson, R. 1989a. “Emergent Attractors and the Law of Maximum Entropy Production: Foundations to a
Theory of General Evolution,” International Federation for Systems Research Pergamon Press.
Swenson, R. 1989b. “Emergent Evolution and the Global Attractor: The Evolutionary Epistemology of Entropy
Production Maximization,” Proc. 33rd Annual Meeting, ISSS, 2-7 July, Edinburgh Vol. 3:46-53. Swenson, R. 1989c.
“The Earth as an Incommensurate Field at the Geo-Cosmic Interface: Fundamentals to
a Theory of Emergent Evolution,” Proc. 1st Int. Congress on Geo-Cosmic Relations Amsterdam: PUDOC Science
Publishers.
Swenson, R. & Turvey, M.T. 1991. “Thermodynamic reasons for Perception-Action Cycles,” Ecological
Psychology Vol.3(4):317-348 Lawrence Erlbam Associates, Inc.
Swenson, R. 1997a. “Autocatakinetics, Evolution, and the Law of Maximum Entropy Production: A Principled
Foundation Toward the Study of Human Ecology” Advances in Human Ecology Vol. 6:1-47
Swenson, R. 1997b. “Thermodynamics and Evolution,” The Encyclopedia of Comparative Psychology
New York: Garland Publishers, Inc.
7
James Fournier
Ph.D. Candidate
California Institute of Integral Studies
1453 Mission Street, San Francisco CA 94103
jim@geoman.com
Research Paper
The discussion of the thermodynamics of evolution dates back at least to Schrodinger’s (1945) seminal
work, What is Life? which appeared long after an obscure but groundbreaking article by Lotka (1922).
There has been a great deal of discussion of the thermodynamics of evolution over recent decades, and yet many
issues remain unresolved, especially surrounding the question of entropy. The topic of this paper, involves the basic
thermodynamic calculation of entropy, and specifically the relationship between energy, heat flow, entropy and
work. It would appear that in their haste to develop a comprehensive theory explaining all of evolution, recent
theorists have lost sight of the thermodynamic concept of work in their treatment of entropy (Swenson, 1989/1997).
This is particularly ironic, as thermodynamics was originally developed in an effort to better understand the
interrelationship among energy, heat flow, and work.
Keywords: Thermodynamics, Evolution, Energy, Entropy, Work
1. A Review of Basic Thermodynamics
Thermodynamic entropy has units of energy over absolute temperature. If the term is to have any meaning this is
where we must begin. There may be metaphors for entropy, which may be useful in information theory or in other
approaches to understanding complex systems, but if we label these with the word
“entropy,” which has a very specific physical meaning, we lose sight of any real value of the term. Similarly,
if we wish to engage in a thermodynamic analysis of a problem we must be careful to use words which have very
specific technical meaning appropriately. A specific example of this is the word
“efficiency.” Efficiency has a very specific technical meaning in thermodynamics. It is intimately related
to the concept of work. A brief review of basic thermodynamic concepts is in order.
From the First Law, the basic formula describing a simple thermodynamic system, or heat engine, is:
Q(h) – Q(c) – W = 0 [eq. 1]
where Q(h) is the heat flowing in, Q(C) is the heat flowing out, and W is the work done by the system. or, in terms of Q
(h):
Q(h) = Q(c) + W [eq. 2]
This says that the heat flowing in from the high temperature reservoir is equal to the heat flowing out to the low
temperature reservoir, plus the work done by the heat engine. The key insight which led to the
development of the Second Law is that W, the work done by the system, must always be less than Q(h), the heat
flowing in from the high temperature reservoir. In other words, all of the heat flowing in cannot be converted to work,
some heat must be discharged and flow out to the cold reservoir. It is the ratio of the work done by the system vs.
the heat flowing in, which determines the efficiency of the system:
Evolution, Entropy and Work
W
Efficiency = -------- [eq. 3] Q(h)
The maximum theoretical efficiency for any system is: T(h) – T(c)
Efficiency max = ---------------- [eq. 4]
T(h)
For any designer of heat engines it is the amount of work, and therefore the efficiency, which are of
greatest interest. It soon became obvious that some portion of the thermal energy, or any type of energy input, must
be discharged in the form of heat to a colder reservoir, or sink, in order to extract any work at
all from a high temperature reservoir or energy source. It was this observation which led to the articulation
of the Second Law of Thermodynamics: For any irreversible process, entropy always increases.
In our system, a heat engine and two heat reservoirs, the change in entropy, delta S, in a complete cycle is: Q(c)
Q(h)
delta S = ------- – -------- [eq. 5] T(c) T(h)
or substituting for Q(c):
Q(h) – W Q(h)
delta S = -------------- – -------- [eq. 6] T(c) T(h)
For a theoretical reversible cycle delta S = 0. In this case, the decrease of Q(h) by W in the numerator of the first
term is exactly proportional to the smaller denominator represented by the lower temperature reservoir, T(c),
and the terms cancel in Equation [6]. In a theoretical reversible cycle the work done, W, is
a maximum, and the change in entropy is zero.
For the irreversible, and therefore real, case W is always less than the maximum reversible value.
Therefore, the numerator of the first term is larger, and delta S > 0, as required by the Second Law.
Reducing W increases the magnitude of the first term, thereby increasing the value of delta S. When W is zero, Q(c)
is equal to Q(h), all of the heat is flowing out to the cold reservoir, and delta S is at a maximum,
as shown in Equation [5].
From the First Law, the heat flowing in, Q(h), must be equal to the heat flowing out, Q(c), plus any work, W, done by
the system; as the energy flowing in as heat must be equal to the energy flowing out as some combination of heat
and work. In addition, the Second Law says that some portion of the energy flowing out must be heat; the heat
flowing in can't all be converted to work.
It becomes apparent from Equation [6] that the magnitude of the increase in entropy, delta S, is directly linked to the
amount of work done by the heat engine on its environment. If it does no work, then Q(h) is equal to Q(c) and the
entropy increase in maximal. However, if the system is as efficient as possible (in the thermodynamic sense of the
word), and therefore does as much work as possible, then the increase in entropy is minimized. If the process
could be truly reversible, then the increase in entropy could, in theory,
be zero. In reality it is impossible for any actual macroscopic processes to be fully reversible. What is
most important to focus on here, however, is the direct negative relationship between the magnitude of the increase
in entropy, and work done by the system. Work most often represents that portion of the process which is
reversible. Work can be used to de-disperse things, to put things back into a state of higher potential
energy. Whether by lifting something with respect to gravity, or by making chemical bonds of a higher potential
energy, it is work which can reverse the normal tendency of things to dissipate toward thermal equilibrium.
2
Evolution, Entropy and Work
2. Thermodynamic Models of Evolution
Much of the recent theorizing about the thermodynamics of evolution attempts to build a model up from non-living
systems.(Goerner, 1994) In the end this may prove to be fruitful, but it may be easier to first start from the other
direction. When humans build machines we are mimicking what nature does, at least in
so far as we are faced with the similar constraints as those facing any other organisms attempting to further their own
survival. Therefore, we need only look at the most basic thermodynamics of the situation to see that the problem,
both for humans and for other organisms, has always been how to extract as much useful work from a potential
energy gradient as possible. There are two ways to do this. One is to increase the magnitude of energy flow across
the gradient. The other is to increase the (thermodynamic) efficiency of
its use, i.e. maximize the work extracted and therefore, by definition, minimize the entropy production. The
first approach can lead to increased entropy production, if the thermodynamic efficiency of the process remains
constant. But the second approach, which incidentally seems to be highly optimized in natural systems, may
actually lead to a reduction in the entropy increase. In any case, the increase in entropy is an indicator after the fact,
not a cause of the behavior of the system, which may be much better understood in terms of energy flows, work,
and the overall thermodynamic efficiency.(Corning, 1997a) Indeed, it becomes plausible to hypothesize that
in strictly thermodynamic terms living systems are the work extracted from the energy flux across the
potential. However, it might be more correct to say that their structures are composed of the energy
embodied by that work in matter. The following thought experiments may make this more understandable.
2. The Benard Cell Experiment
First, let us examine the now paradigmatic Benard cell. The Benard cell experiment is carried out in a flat circular
dish filled with a thin layer of water, which is uniformly heated from below in a controlled manner.
At a certain critical rate of thermal input, if the water layer is precisely the right thickness, the heat will be dissipated
more rapidly through the formation of a coherent pattern of hexagonally arrayed convection cells than through
turbulent boiling.(Benard, 1900) In this case we have a system in which energy moving across a potential
gradient spontaneously gives rise to an ordered pattern or structure. While this may be a
very useful critique of the 19th century idea that all physical systems tend toward a state of greater disorder,
it really has very little to do with the thermodynamic behavior of living systems. Its repeated invocation
seems to be an artifact of early thinking about dissipative structures.(Prigogine, 1984; Swenson,
1989/1997) The Benard cell does, however, also inadvertently illustrate a key distinction which should be
recognized when comparing various types of systems. The question becomes one of whether or not the
system does work which contributes to structures that endure in matter. This distinction is of pivotal
importance to living systems. Living systems create structures that do endure. While perhaps so obvious
as to be taken for granted, this distinction allows us to recognize why the Benard cell is not an appropriate example
for describing the thermodynamic behavior of living systems. In order to create enduring structures in matter
living systems must do work. Thus, if we are looking for thermodynamic examples of living system, we must look for
systems that do work and create enduring structures. This appears to be a necessary, but not sufficient, condition
for describing living systems. Thus, perhaps it might prove to be an interesting criterion to employ in pursuing an
investigation into how non-living systems could have evolved into living systems. That question, however, is not the
focus of this paper.
In the case of the Benard cell, one might be able to arrange a frame of reference in which it appears that work is
being done to create, and temporarily maintain, the coherent pattern among the convective cells. This coherent
pattern is, as has been pointed out repeatedly over the last few decades, what allows the system to
dissipate the available thermal energy most rapidly (Swenson, 1989/1997). However, unlike a living system, this
coherent pattern does not contribute to the establishment of any enduring structure in matter. As soon as the
source of high energy potential is interrupted, the pattern immediately collapses, and any energy which had been
contributing to maintaining the structure, is immediately converted to thermal dispersion and dissipated. Thus,
the claim that the Benard cell creates increased entropy as rapidly
as possibleoften erroneously referred to as “efficiently”appears to be a correct (Swenson, 1989/1997). But, while
this is an interesting physical phenomenon, showing that macro scale ordered patterns can
3
Evolution, Entropy and Work
spontaneously emerge out of the dynamics of energy flow across a potential gradient, it has nothing to do
with the thermodynamic behavior of living systems. Living systems, by contrast, use part of the energy flowing
through them to do work. To the extent that this work is used to create structures in matter, that portion of the
energy which does the work is sequestered in matter. That portion of the energy does not flow into a low
temperature reservoir, and thus does not immediately contribute to entropy production. Thus, in many cases
living systems actually create less entropy, than comparable non-living systems operating across the same
potential gradient over a given period of time. It seems that the second law cannot be violated, but it can be
stalled. In that sense life could be regarded as an energy “kiting” scheme.
3. Photosynthesis
This becomes particularly apparent when one examines the thermodynamic behavior of organisms engaging
in photosynthesis. Here energy from photons coming from the high temperature source, namely the Sun, is
converted to potential energy in the form of excited electrons and used to do work, namely creating higher
energy molecules in the form of carbohydrates and molecular oxygen in the atmosphere. The energy represented by
photons which are actually captured and used for photosynthesis, about twelve percent of those actually striking a
leaf, are converted into chemical potential energy.(Corning, 1997b, p.5) This energy is bound up in the carbon
compounds which make up the structure of the organism, i.e. organic material, and in the higher chemical potential
energy state of the atmosphere. This portion of the incident solar energy is not immediately reradiated in the form of
thermal dispersion, and thus does not immediately contribute to increased entropy. Therefore, it would appear that
throughout the time when there have been photosynthesizing organisms on Earth, the overall rate of entropy
production on Earth has in fact been less than it would otherwise have been without them. We can see the
difference in the energy sequestered in the vast accumulation of biomass and fossil fuels on Earth. Energy
which was chemically bound up on Earth rather than being disbursed into space as heat did not contribute to
increased entropy over that period.
5. Respiration
It is only because of the long term accumulation of energy by photosynthesizing organisms, and the
resultant slowing of entropy production, that a large energy reserve has been built up on Earth. It is the existence of
this reserve that has made it possible for other organisms to consume and transform that energy more rapidly. It is
only with the advent of aerobic organisms engaged in respiration that it becomes possible to burn up these
reserves more rapidly, and therefore to increase the rate of entropy production. If one only examines the behavior of
organisms engaged in respiration over the period following the onset of the increase in atmospheric oxygen
levels, one might conclude that there is a pervasive trend toward increased entropy production in all of
evolution (Swenson, 1989/1997). But, this is apparently a characteristic, not the cause, of a trend which
has only existed over some portions of the evolutionary history of life on this planet. Moreover, in spite of this
recent trend, wherein the oxidation rate associated with the respiration of ever more complex organisms has
been steadily increasing, the overall rate of entropy production of the planet as a whole has still remained less
than that of a comparable planet without life. The evidence for this observation is found in the overall
tendency of the Earth to continue to accumulate carbon-based biomass, while maintaining a stable concentration
of atmospheric oxygen. This represents the sequestering of energy which would otherwise have been reradiated
into Space contributing
to increased entropy, but which instead did not do so. A more interesting observation in this area concerns the
steady increase in the oxidation rate per unit mass by ever more complex aerobic organisms. This trend toward
increasing energy flux density in increasingly complex organisms is most interesting, but does not
by itself lend any real support to the claim that entropy increases as rapidly as possible. There are two
reasons why this is the case. In the first place, one would have to know how much work was performed by those
organisms before one could make any meaningful calculation of entropy. Secondly, the trend itself is only valid
over a portion of the evolutionary history, and as has previously been pointed out, was only made possible
by a prior prolonged period of energy accumulation and collateral entropy rate reduction.
4
Evolution, Entropy and Work
6. The Planet as a Whole
The question of the overall thermodynamic behavior of the planet as a whole is also worth considering in greater
detail. The maximum possible increase in entropy for a given system is found when all of the thermal
energy flowing in as heat from the high temperature source, is also flowing out as heat to the low temperature sink.
This is essentially the case for a dead planet such as the Moon, or as far as we can tell Venus or Mars. In this case
all of the solar radiation which strikes the planet is reradiated out into Space within a relatively brief period. The only
other possibility is that some of the energy flowing in could go into heating the planet. But the dead planet
appears to be pretty much at equilibrium, or perhaps more accurately at a stable steady state, as there may
also be some residual thermal energy being radiated as heat from volcanic activity by the planet as it cools very
slowly. The source of this heat is largely nuclear decay within the core of the planet, although it is possible that there
may also be some leftover thermal energy from mechanical collisions when it was formed, or even conversion
of a small amount of gravitational energy to heat in movements of its crust.(Cloud, 1988) But the rate of heat flow
from all of these sources is stable enough that as a practical matter the dead planet is at a stable equilibrium
temperature averaged over any representative period.
It is also important to understand that changing the relative position of the temperature of the planet within the
temperature gradient between the Sun and Space has no effect on the ongoing change in entropy. Any increase on
one side is exactly offset by a decrease on the other, i.e. if the planet were colder and there was therefore more
entropy increase associated with photons from the Sun being absorbed by it, then that increase would be
offset by the reduced increase in entropy when photons were radiated from the colder planet back out into
Space. The overall rate of entropy production is entirely determined by the temperature difference between
the source, the Sun, and the sink, cold dark Space, and has nothing to do with the relative intermediate temperature
of the planet. The rate of entropy production is therefore already pegged at a maximum for any dead planet. There
is simply nothing for any incident energy to do other than
to be reradiated out into space, and therefore all of the energy leaving the Sun and striking the planet is already
contributing to the maximum possible rate of entropy production.
7. The Transition to Life
Once a planet has evolved photosynthesizing organisms, it is clearly operating at an overall rate of entropy
production which is less than that of a dead planet. This is the case because, as has already been outlined,
photosynthesizing organisms use some of the energy to do work to create higher energy compounds, which
sequester a portion of the energy rather than allowing it to be reradiated out into space. If a portion of the energy
striking the planet is not immediately contributing to the increase in entropy then the overall rate of entropy
production must be less than the theoretical maximum for any given period of time. What is more interesting is the
question of what happened during the earlier period before photosynthesis starts. More interesting still is the
question of what happened before there was really life per se, but after autocatalytic chemical reactions start
occurring. It would seem that as soon as such reactions begin to sequester significant amounts of incident
energy in the form of chemical compounds at higher than equilibrium potential energy, the process has
started. Indeed, most hypotheses for the spontaneous emergence of life out of some mixture of multiple
autocatalytic reaction cycles depend upon the existence of some sort of high-energy soup. The ultimate energy
source for the creation of such a soup might be some combination
of thermal radiation from the Earth and/or photons from the Sun. Even if this energy were converted into a hydrologic
cycle giving rise to lightening, or some other form, the key transition point in this model would
be when some portion of that energy began to be sequestered as complex chemical compounds, proteins or other
chemical precursors to life. From a thermodynamic standpoint it is impossible to say where in this sequence life
actually begins, even though from a biochemical, or at least from a biological standpoint it might be clear that such a
system could not yet be said to contain life. This begs the question of whether a living system or pre-living system
must contain life, or rather living entities. Indeed, as Gaia theory begins
to suggest the whole system might in some sense be said to be living, even before individual entities could
be said to be alive.(Lovelock, 1987; 1995) The key observation appears to be that living, and perhaps even pre-
living, systems tend to drive the planet away from its previous condition of maximum entropy.
5
Evolution, Entropy and Work
8. Human Intervention
It is only with the advent of human use of fire that the overall rate of entropy actually begins to increase on Earth.
The first indications of this might have been when hunting and gathering peoples began to manage large areas of
forest by regularly burning the undergrowth more extensively and frequently than had been the case due to
lightening strikes. From there humans eventually moved on to slash and burn agriculture, and eventually to larger
and larger scale deforestation. The desire for wood, both as fuel and building material, combined with the
need for more grazing area for cattle and sheep has driven a steady pattern of deforestation wherever human
populations have spread. Especially in Europe and the Middle East, the spread of human civilization has been
basically synonymous with the spread of deforestation for at least four thousand years. Yet, throughout this
period, while the rate of entropy production due to human activity was steadily increasing, total combustion
still did not even come close to approaching the point where the biosphere could no longer sequester far more
energy than was being dissipated as heat due to combustion. It has only been since the industrial revolution that the
overall rate of entropy increase on the Earth may have for the first time approached, or even exceeded, that of the
planet before life began. For it was only when humans began to tap, and burn, the fossil hydrocarbon reserves of
the planet, disbursing a large portion of that energy as heat, that the rate of entropy production on Earth spiked
upward. In this act
of essentially hyper-respiration human technology at once mimicked, and surpassed, the behavior of all
other aerobic organisms including ourselves. Prior to this point the biosphere, in combination with the
hydrologic and geologic systems of the planet, has apparently always been capable of maintaining a net
positive balance in the current account of energy flow, always sequestering a net surplus.(Cloud, 1988) Now,
in a relative flash on the scale of geologic time we are literally burning through those reserves, massively
accelerating the rate of entropy production, and indeed the rate of increase of the rate of entropy production. This
resembles more a rip, or tear, in the fabric of the energetic net life has so carefully woven, than an extension of life’
s natural functioning. This situation will most likely only make sense in retrospect, as a necessary and
even inevitable stage in our evolution, if we are able to recognize the implications of the situation now and
very rapidly adapt our technology and behavior.
9. Technological Analogs of Organisms
From the point of view of the energy flow and entropy analysis, which we have so far been applying
primarily to biological organisms, current fossil-fuel energy technology functions like a hyper-animal. Human-
made machines oxidize hydrocarbons to extract the embodied energy, while remaining dependent upon
photosynthesizing organisms to maintain atmospheric equilibrium. Biological respiration is essentially
combustion, but carried out in a controlled manner to yield the maximum thermodynamic efficiency. It seems
clear, given the preceding analysis, that we as a species are now facing a fundamental point of inflection in a very
long-term oscillation in the thermodynamics of the biosphere. This point is no less profound than when organisms
capable of aerobic respiration first emerged in response to the build up
of oxygen in the atmosphere. We now face the symmetrical point. This time, hyper-animals, i.e. our
machines, threaten to exceed the capacity of the atmosphere to absorb the CO2 they give off without
adversely destabilizing the narrow range of climatic temperature we require to support our vast population. The
current situation is in many ways exactly the opposite of the one which gave rise to our predecessors, and launched
the whole lineage of ever more complex organisms extracting ever greater amounts of energy with oxygen from the
atmosphere. Now, having made technological extensions of ourselves, which have pushed that process further, we
must implement technology which will bring the system back into balance.
The key insight may be to discern the fundamental difference between plants, which engage in
photosynthesis, and animals, which engage in respiration, and to recognize that there is potentially a
parallel classification system between animal-like technologies which engage in combustion, or more precisely
oxidation, and plant-like technologies which represent a technological analog of photosynthesis. The obvious
technological parallel will be with photovoltaic cells. Yet, photovoltaics by themselves cannot really be described as
meta-plants. Like plants, they do convert sunlight into available energy, but they do not directly affect the
atmospheric balance between CO2 and O2. If, however, their electrical output is used
6
Evolution, Entropy and Work
to perform electrolysis, splitting water into hydrogen and oxygen, then photovoltaics might represent plants
on one side of a potentially balanced energetic cycle, analogous to the carbon cycle in nature. A human- made
closed-loop analog of nature might even work without carbon, instead using only H2, O2 and water. Energy
technology may actually be headed in this direction already. The chemical composition of fuels have moved steadily
from wood, to charcoal, to peat, to coal and then through the sequence of successively smaller and lighter fossil
hydrocarbons from crude oil, to natural gas. This sequence inevitably converges
on pure hydrogen, H2, as the smallest lightest energy storage and transport medium. Hydrogen can be
highly problematic as a fuel for combustion, but the optimal technology for the use of H2 does not involve
atmospheric combustion. Oxidation is instead carried out in a controlled manner in a fuel cell. This is the
technological analog of respiration. Fuel cells combine H2 and O2 to yield an electric current and water. These
technological analogs of photosynthesis and respiration are based on silicon instead of carbon. Silicon, the
basis of rock, is essentially the higher octave of carbon in the periodic table. Thus, it might be appropriate to
describe such a new closed-loop cycle as an octave of nature, or even as “Meta-Nature.”
Baring the sudden advent of nuclear fusion or zero-point energy, the rapid deployment of such technology will be
necessary for human civilization to continue. While this is already obvious to many, it may shed new light on the
problem if we are able to understand it within the framework of the very long-term behavior of energy flows in
nature. Such analysis may illustrate, more clearly than previous approaches, that what we must now do is
essentially create meta-plants, systems which are capable of an analog of photosynthesis, in order to initiate a
new sustainable balance with the vast amount of human technology which already behaves like essentially meta-
animals. It may also be helpful to recognize that in general, living systems tend to maximize their access to
available energy, while striving to be as efficient as possible, in the true thermodynamic sense of the word. We
would do well to follow their example.
References
Benard, H. 1900. “Les Tourbillons Cellulaires dans une Nappe Liquide,” Rev. General Science Pur. Appl.
Vol. 11:1261-1271.
Cloud, P. 1988. Oasis in Space, W.W. Norton & Co.
Corning, P. A. & Kline, S.J. 1997a. “Thermodynamics, Information and Life Revisited, Part I To Be or
Entropy” Syst. Res. Behav. Sci. Vol. 14:0:1-22 John Wiley & Sons, Ltd.
Corning, P. A. & Kline, S.J. 1997b. “Thermodynamics, Information and Life Revisited, Part II
‘Thermoeconomics’ and ‘Control Information’” Syst. Res. Behav. Sci. Vol. 14:0:1-29 John Wiley & Sons. Corning, P.
A. in press. “On the Concept of Synergy and its Role in the Evolution of Complex Systems,” Journal of Social and
Evolutionary Systems
Goerner, S. J. 1994. Chaos and the Evolving Ecological Universe Gordon & Breach.
Lotka, A. J. 1922. “Contribution to the Energetics of Evolution,” Proceedings of the National Academy of
Sciences Vol. 8:147-151.
Lovelock, J. E. 1987. Gaia : A New Look at Life on Earth Oxford University Press
Lovelock, J. E. 1995. The Ages of Gaia : A Biography of Our Living Earth W.W. Norton & Company
Prigogine, I. 1984. Order Out of Chaos Bantam. Schrodinger, E. 1945. What is life? New York: Macmillan.
Swenson, R. 1989a. “Emergent Attractors and the Law of Maximum Entropy Production: Foundations to a
Theory of General Evolution,” International Federation for Systems Research Pergamon Press.
Swenson, R. 1989b. “Emergent Evolution and the Global Attractor: The Evolutionary Epistemology of Entropy
Production Maximization,” Proc. 33rd Annual Meeting, ISSS, 2-7 July, Edinburgh Vol. 3:46-53. Swenson, R. 1989c.
“The Earth as an Incommensurate Field at the Geo-Cosmic Interface: Fundamentals to
a Theory of Emergent Evolution,” Proc. 1st Int. Congress on Geo-Cosmic Relations Amsterdam: PUDOC Science
Publishers.
Swenson, R. & Turvey, M.T. 1991. “Thermodynamic reasons for Perception-Action Cycles,” Ecological
Psychology Vol.3(4):317-348 Lawrence Erlbam Associates, Inc.
Swenson, R. 1997a. “Autocatakinetics, Evolution, and the Law of Maximum Entropy Production: A Principled
Foundation Toward the Study of Human Ecology” Advances in Human Ecology Vol. 6:1-47
Swenson, R. 1997b. “Thermodynamics and Evolution,” The Encyclopedia of Comparative Psychology
New York: Garland Publishers, Inc.
7