ArticlePDF Available

Self-Organization vs. Self-Ordering events in life-origin models

Authors:
  • The Gene/Epigene Emergence Project

Abstract and Figures

Self-ordering phenomena should not be confused with self-organization. Self-ordering events occur spontaneously according to natural “law” propensities and are purely physicodynamic. Crystallization and the spontaneously forming dissipative structures of Prigogine are examples of self-ordering. Self-ordering phenomena involve no decision nodes, no dynamically-inert configurable switches, no logic gates, no steering toward algorithmic success or “computational halting”. Hypercycles, genetic and evolutionary algorithms, neural nets, and cellular automata have not been shown to self-organize spontaneously into nontrivial functions. Laws and fractals are both compression algorithms containing minimal complexity and information. Organization typically contains large quantities of prescriptive information. Prescriptive information either instructs or directly produces nontrivial optimized algorithmic function at its destination. Prescription requires choice contingency rather than chance contingency or necessity. Organization requires prescription, and is abstract, conceptual, formal, and algorithmic. Organization utilizes a sign/symbol/token system to represent many configurable switch settings. Physical switch settings allow instantiation of nonphysical selections for function into physicality. Switch settings represent choices at successive decision nodes that integrate circuits and instantiate cooperative management into conceptual physical systems. Switch positions must be freely selectable to function as logic gates. Switches must be set according to rules, not laws. Inanimacy cannot “organize” itself. Inanimacy can only self-order. “Self-organization” is without empirical and prediction-fulfilling support. No falsifiable theory of self-organization exists. “Self-organization” provides no mechanism and offers no detailed verifiable explanatory power. Care should be taken not to use the term “self-organization” erroneously to refer to low-informational, natural-process, self-ordering events, especially when discussing genetic information.
Content may be subject to copyright.
This article was originally published in a journal published by
Elsevier, and the attached copy is provided by Elsevier for the
author’s benefit and for the benefit of the author’s institution, for
non-commercial research and educational use including without
limitation use in instruction at your institution, sending it to specific
colleagues that you know, and providing a copy to your institution’s
administrator.
All other uses, reproduction and distribution, including without
limitation commercial reprints, selling or licensing copies or access,
or posting on open internet sites, your personal or institution’s
website or repository, are prohibited. For exceptions, permission
may be sought for such use through Elsevier’s permissions site at:
http://www.elsevier.com/locate/permissionusematerial
Author's personal copy
Physics of Life Reviews 3 (2006) 211–228
www.elsevier.com/locate/plrev
Review
Self-organization vs. self-ordering events in life-origin models
David L. Abel a,, Jack T. Trevors b,1
aThe Gene Emergence Project, The Origin-of-Life Foundation, Inc., 113 Hedgewood Dr. Greenbelt, MD 20770-1610, USA
bDepartment of Environmental Biology, University of Guelph, Rm 3220 Bovey Building, Guelph, Ontario, Canada N1G 2W1
Accepted 5 July 2006
Available online 12 October 2006
Communicated by L. Perlovsky
Abstract
Self-ordering phenomena should not be confused with self-organization. Self-ordering events occur spontaneously according to
natural “law” propensities and are purely physicodynamic. Crystallization and the spontaneously forming dissipative structures of
Prigogine are examples of self-ordering. Self-ordering phenomena involve no decision nodes, no dynamically-inert configurable
switches, no logic gates, no steering toward algorithmic success or “computational halting”. Hypercycles, genetic and evolutionary
algorithms, neural nets, and cellular automata have not been shown to self-organize spontaneously into nontrivial functions. Laws
and fractals are both compression algorithms containing minimal complexity and information. Organization typically contains large
quantities of prescriptive information. Prescriptive information either instructs or directly produces nontrivial optimized algorithmic
function at its destination. Prescription requires choice contingency rather than chance contingency or necessity. Organization
requires prescription, and is abstract, conceptual, formal, and algorithmic. Organization utilizes a sign/symbol/token system to
represent many configurable switch settings. Physical switch settings allow instantiation of nonphysical selections for function
into physicality. Switch settings represent choices at successive decision nodes that integrate circuits and instantiate cooperative
management into conceptual physical systems. Switch positions must be freely selectable to function as logic gates. Switches
must be set according to rules, not laws. Inanimacy cannot “organize” itself. Inanimacy can only self-order. “Self-organization” is
without empirical and prediction-fulfilling support. No falsifiable theory of self-organization exists. “Self-organization” provides
no mechanism and offers no detailed verifiable explanatory power. Care should be taken not to use the term “self-organization”
erroneously to refer to low-informational, natural-process, self-ordering events, especially when discussing genetic information.
©2006 Elsevier B.V. All rights reserved.
Keywords: Chaos theory; Crystallization; Decision nodes; Dissipative structures; Dynamically-inert configurable switches; Fractals; Hypercycles;
Genetic information; Programming; Origins; Molecular; Evolution; Prebiotic; Primordial soup; Codons; Proteins; Life; Nucleic acids; Cells;
Genes; Computable cells; Environment
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .212
*Corresponding author. Tel.: +301 441 2923; fax: +301 441 8135.
E-mail addresses: life@us.net (D.L. Abel), jtrevors@uoguelph.ca (J.T. Trevors).
1Tel.: +519 824 4120 x 53367; fax: +519 837 0442.
1571-0645/$ – see front matter ©2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.plrev.2006.07.003
Author's personal copy
212 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
2. Self-orderingevents ....................................................................212
2.1. Crystallization...................................................................213
2.2. Self-ordered dissipative structures . . . . . . . ...............................................213
2.3. Spontaneously ordered nanotubules and other self-assembled structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
2.4. Spontaneous biopolymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
2.5. Theroleofvariablefree-energybindingofaminoacids .......................................216
2.6. Self-orderedstructurescansometimesmimictrueorganization ..................................216
3. Organization .........................................................................216
4. Laws and fractals are both compression algorithms ...............................................218
4.1. Lawsarethemselvescompressionalgorithmsforreamsofempiricaldata ...........................218
4.2. Fractals create the illusion of complexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218
5. Open systems far from equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
6. “Self-ordering” and “self-organizing” cannot be used interchangeably . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
7. Self-organizationhasnotbeendemonstratedtoexist ............................................221
7.1. Noempiricalevidence..............................................................221
7.2. No rational plausibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
7.3. Nopredictionfulllment............................................................222
7.4. No falsifiability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
8. The problem of prescriptive sequencing . . . . . . . . ...............................................223
9. Discussion ..........................................................................224
10. Conclusions .........................................................................225
Acknowledgements.........................................................................225
References...............................................................................225
1. Introduction
Self-ordering should not be confused with self-organization. Typically, self-ordering and self-organization are
merged as a result of unrefined definitions. Ill-defined distinctions lead to category errors of fallacious inference.
Systems science, biosemiotics, biocybernetics, life-origin research, neural network and cellular automata studies, and
many fields of natural science research all suffer from this confusion.
In addition to failing to perceive the necessary and sufficient conditions of bona fide organization, confusion exists
about the nature of intuitive information. Shannon knew the limitations of his uncertainty measurements, and stated
them in his first paper [119,143]. His measurements would not be applicable to meaning or aboutness. As the von
Weizsäckers have pointed out [140], Weaver also understood what the negative logarithm of an event’s probability
could not provide: “Two messages, one heavily loaded with meaning, and the other pure nonsense, can be equivalent
as regards information” [138]. Weaver’s only mistake was to call the quantifiable uncertainty to which he referred, “in-
formation”. Reduced uncertainty (mutual entropy) does not provide what Abel has termed, “prescriptive information”
[2,3,128]. Prescriptive information either instructs or directly produces nontrivial optimized algorithmic function.
To begin to understand both organization and the prescriptive information upon which organization depends, the
limitations of the self-ordering phenomena of physical nature are examined.
2. Self-ordering events
Self-ordering phenomena occur according to natural “law” propensities. Self-ordering events are physicodynamic.
Crystallizations and the spontaneously forming dissipative structures of Prigogine [99–101] are examples of self-
ordering. They do not algorithmically “self-organize”. Rather, they self-order out of seeming chaos and the complex
interacting force relationships of inanimate nature. They are constrained by environmental starting conditions and the
pre-existing orderliness described and predicted by the “laws” of nature. No other factors such as instructions are re-
quired for self-ordering phenomena to occur. Except within the formal mathematical expression of “laws” themselves,
self-ordering phenomena involve no decision nodes, no dynamically-inert configurable switches, no logic gates, and
no steering toward algorithmic optimization. “Dynamically inert” (dynamically incoherent) means that physicody-
namics plays no role in the setting of a configurable switch [106,107]. Gravity, for example, plays no role in which
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 213
way a switch knob on a horizontal switch board is flipped. Despite the switch knob’s being physical, the agent oper-
ator’s choice alone determines which way the knob is pushed. The essential element of configurable switches of all
types is dynamic incoherence (inertness). The setting of the switch is not coherent with physicodynamic causation.
Function is never an issue with the self-ordering phenomena of nature. Self-ordering is oblivious to whether com-
putational halting, cellular metabolism, or function of any kind are achieved. Just as objective being in philosophy is
oblivious to shoulds and oughts,being is also oblivious to formal algorithmic choices made with the intent of achiev-
ing utility. Self-ordering events do not pursue design and engineering success. Computational halting and potential
protometabolism are not desires or goals of self-ordering phenomena.
Anti-informationists are well-represented in Refs. [7,13,24,29,51,55,56,69–72,78,79,113,114,139]. The school of
thought known as infodynamics treats information as nothing more than a subset of physicodynamic interactions. Yet
empirical evidence and prediction fulfillment are sorely lacking for unaided physicodynamics producing nontrivial
computational or metabolic function. Schrödinger’s What Is Life? started the negentropy concept [115]. Brillouin
contributed to a physical concept of information and organization [16–18].
The Russian E. Liiv has of late helped popularize infodynamics as it relates to a defined negentropy: “The negen-
tropy of systems as a special form of reality” [77]. Not all, however, agree that negentropy is real or even possible.
Yockey points out that Boltzmann specifically excluded a negative constant from his mathematical definition of physi-
cal entropy [143]. Shannon employs a negative constant in his Hequation [119]. Yockey argued that every probability
distribution is unique, and that the Sprobability distribution phase space of Boltzmann’s physical entropy (S) cannot
be equated or synthesized with Shannon’s “informational” uncertainty (H) probability distribution despite seemingly
identical Sand Hequations. Section 5below addresses such issues in “open systems far from equilibrium”.
When attempts are made to reduce semantic and prescriptive information to physicodynamics, the patterns created
by self-ordering become the object of much attention. An examination of the informational capabilities of self-ordering
physical events follows.
2.1. Crystallization is a self-ordering phenomenon typically of the same molecular species. Crystallization pro-
ceeds spontaneously under conditions of dehydration, cooling, and supersaturation of solutions. No decision-node
choices are required for crystallization to occur. Physicodynamic thresholds are reached, not purposefully chosen or
controlled. No configurable switches must be set. Apart from agent manipulation and selection of initial conditions,
precipitation is not controlled to achieve some purpose.
Almost no Shannon uncertainty is associated with crystallization. Aside from impurities, the probability of a
particular crystalline structure forming for each molecular species under the same conditions approaches 1.0. The
information retaining potential, therefore, of a crystal is extremely low. Only crystal impurities provide opportunity
for information retention in such a monotonous physical matrix. But crystal impurities are rarely controllable in na-
ture. Impurities that are controllable tend to be too regular for programming and significant information retention. The
limited amount of prescriptive information that can be retained in crystal layers is one reason Cairns–Smiths model
of initial “clay life” did not progress [25,26].
2.2. Self -ordered dissipative structures arise spontaneously out of seeming chaos. The nonlinear nature of com-
plex, multi-variable phenomena make precise prediction extremely difficult. Yet the underlying law-like self-ordering
tendencies of nature make spontaneously forming vortices at bathtub drains, hurricanes, and tornadoes unsurprising.
Computerization has made possible large numbers of complex interacting computations in minimal time. Bennett
used such computation time as a measure of complexity [9,10]. But Bennett’s “logical depth” presupposes cybernetic
concepts that are foreign to spontaneous self-ordering events.
The fundamental nature of Prigogine’s dissipative structures within chaos theory [101] remains self-ordering rather
than self-organizing. The title of Prigogine and Stengers book was correct: Order Out of Chaos, not Organization
Out of Chaos. The title of Nicolis and Prigogine’s earlier book [86] was incorrect in the title, “Self-Organization in
Nonequilibrium Systems”: and correct in the subtitle, “From Dissipative Structures to Order Through Fluctuations”.
Unfortunately, others since have continued to blur the distinction between order and organization [32,33,36–38,54,63–
68,82,95,123,137]. The illegitimate merging of the two concepts now seems almost universal. The “category error” of
logic theory leads to countless faulty inferences.
No programming is necessary for the spontaneous occurrence of a water vortex at a bathtub drain, the behavior of a
sand pile, the shape of a candle flame, the formation of a tornado or hurricane. Such dissipative structures form out of
Author's personal copy
214 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
a continuous stream of instantaneously self-ordered states and phase changes that have nothing to do with cybernetic
organization. No reason or empirical data exist in chaos theory to suggest that a physical environment is capable
of marshalling cybernetic organizational skills. The environment cannot program or compute. Chance and necessity
cannot set configurable switches to optimize genetic algorithms in genomes [2,3]. No evidence exists for attributing
algorithmic organization to chaos theory.
2.3. Spontaneously ordered nanotubules and other self-assembled structures
Spontaneous self-assembly of various physically ordered structures has been the topic of much research. Pohl et
al. [98], for example, described a spontaneous process that takes place on a single-atom-thick film of silver sprinkled
with sulfur. A lacework nanopattern emerges with precision as the sulfur atoms pierce the silver with great regularity.
Each hole is over 20 times further apart than the force field generated by each atom. The authors envisioned the
generation of nanostructures that might exceed the best currently known patterning techniques.
Papaseit et al. observed microtubule self-assembly to be gravity-dependent [89]. The list of publications and types
of spontaneous self-assemblies in physics alone seems endless. But under careful analysis, every instance of inanimate
self-assembly can be traced to complex interactions of self-ordering force relationships. No algorithmic integration or
organization is required to explain the spontaneously formed structure.
From a biological perspective, one must distinguish between physically self-ordered structures and those that are
essentially “designed, engineered, and manufactured” by the linear digital programming instructions of codon strings.
Clark et al. studied the bio-nanoscience of self-assembling supramolecular protein-nucleic nanostructures [31].Re-
blova et al. [102] found that the core region of RNA “kissing complexes” form cation-binding pockets with highly
negative electrostatic potentials. These pockets show nanosecond-scale breathing motions coupled with oscillations
of the entire molecule. Amos used electron crystallography to study the structure of tubulin to predict structural in-
teractions with nucleotides, drugs, motor proteins and microtubule-associated proteins [6]. Surrey et al. identified
parameter combinations that determine the generation of asters, vortices, and a network of interconnected poles. They
found that microtubules and their associated motor proteins in eukaryotic cells can be organized into various large-
scale patterns [125]. Rothemund et al. made nanotubes from DNA double-crossover molecules (DAE-E tiles) [111].
But the structure of these nanotubes can be explained by a simple model based on the geometry and energetics of
β-form DNA.
In no instances of self-ordering are functional selections required at true decision nodes. Bifurcation points may
exist, but choice with intent is not needed to decide which course to take. Formal computation and control are not
involved in self-ordered self-assembly. Optimization of algorithmic programming is always missing from natural-
process, physicochemical, self-ordering events.
If the assembly of RecA on single-stranded DNA is examined, this process is measured and interpreted as a finite-
state machine [8]. The machine is able to discriminate fine differences between sequences. Such discrimination is a
formal computational operation, not a physicodynamic self-ordering phenomenon. An iterative cascade of multistage
kinetic proofreading amplifies minute differences, including single base changes. Such a Turing-like machine seems
to be able to compute integral transforms. We do not see such processes arising independent of digitally programmed
life. Prescriptive information is required to organize such a system. Decision nodes and efficacious configurable switch
settings are involved. Such digital programming is unique to life. It should not be confused with the low-informational
self-ordering events of inanimate nature.
2.4. Spontaneous biopolymerization
Leman et al. have showed that carbonyl sulfide (COS), a simple volcanic gas, can bring about the formation of
peptides from amino acids under mild conditions in an aqueous solution. But this reaction, like others that have been
published (e.g., Stanley Miller type experiments), produce only short stochastic-ensemble peptides rather than long
functional sequences.
The leading group of experimenters with clay adsorption polymerizations is that of James Ferris [43,62,84] Ferris’
group has concentrated mostly on the polymerization of oligoribonucleotides [62,84]. Minimal gene length in any
life-origin model is generally thought to be 30–60 monomers (mers). 10 mers is the maximum length that forms in
solution. Ferris’ group has been able to produce 55 mer lengths on clay: montmorillonite for activated nucleotides;
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 215
illite and hydroxylapatite for joining activated amino acids directly [44]. The longer lengths are formed by successive
’feedings’ with monomers. Polymerization is similar to solid-phase synthesis of biopolymers.
The problem with polymerization on clay is the strong tendency to produce highly ordered sequences rather than
informational sequences. Montmorillonite aligned monomeric sequences offer no help in explaining the origin of
genetic instructions.
Bernd M. Rode’s group has also been a leader in pursuing spontaneous peptide formation, including homochiral
preferences. Homochirality refers to a pure population of the same optical isomer (all right-handed gloves, or all left
handed gloves). Most of Rode’s group papers have been first-authored by either Plankensteiner [96,97] or Bujdak [21,
23]. Bujdak has also worked with clay adsorption [20,22]. Certain amino acids can catalyze peptide bond formation
in “Salt-Induced Peptide Formation (SIPF)” reactions. Rode argues that SIPF reactions in connection with adsorp-
tion processes on clay minerals is the most likely universal mechanism for initial biopolymerizations in a Peptide
Worl d [109].
Certainly Rode and many others are correct that physicodynamic biases exist, especially with clay adsorption.
However, the reasoning strikes us as being a nonsequitur in the statement, “Reaction-inherent preferences of certain
peptide linkages make the argument of “statistical impossibility” of the evolutionary formation of the “right” peptides
and proteins rather insignificant” [109]. In our opinion, physicodynamic bias only reduces through self-ordering ten-
dencies the vast sequence spaces needed for prebiotic molecular evolution. Even if vast sequence spaces had been
available in a theoretical primordial soup, no known mechanism exists for the prebiotic selection of prescriptive
sequences. Prescriptive sequences require freedom of selection at each successive decision node. This freedom is
anti-thetical to the self-ordering “necessity” of physicodynamic bias.
Fontana et al. studied phenotypic equivalence classes of genotypes in view of neutral genetic drift [46]. They then
computed a statistical topology organizing the set of RNA shapes. Neighborhood relations among phenotypes were
correlated with the statistics of neighborhood relations among equivalence classes of genotypes in genotype space.
Reza Ghadiri’s group was able to replicate peptides using hydrophobic interactions between side chains of short
alpha-helical peptides [112]. Some view this as a nonlinear molecular information transfer processes. But, as Mellersh
points out [81], starting with RNA polymerase presupposes and incorporates considerable prior programming into
Ghadiri’s supposedly materialistic origin of protein self-replication. The initial information input is monumental (two
32-mer peptides). Mellersh argues that the Ghadiri group is merely joining together two large pre-activated chains.
This is not really self-replication, but bioengineering. Continuity with present day biochemistry is also lacking.
How did the RNA polymerase achieve its functional sequencing? Even if there were some natural process explana-
tion for this, how would all of the other needed substrates and enzymes acquire their critical monomeric sequencing at
the same time and place? The self- or mutual replication of RNA polymerase would also use up most of the resources
needed for other essential biopolymers to explore sequence space.
Homochirality theoretically could have been established through some mechanism similar to the self-replicative
peptide model suggested by M. Reza Ghadiri’s group [112] or the nonenzymatic templating of Bernd Rode’s group.
But no known physicochemical processes can produce anything close to a pure homochiral population of levo amino
acids or dextro ribonucleotides. The question always remains, “How would the template acquire its homochirality
in a prebiotic environment?” Ghadiri’s group artificially amplifies an initially minimal enantiomeric difference in a
racemic mixture of peptide fragments.
Despite having noteworthy success in producing self-replicating peptides, Ghadiri’s research does not demonstrate
self-organized autocatalytic networks capable of homeostatic metabolism. In addition, they stated quite accurately,
“These results augur well for the rational design of functional peptides”. Like the ribozyme engineers using SELEX,
the Ghadiri group is very honest and up-front about the fact that the only way they accomplished this feat with
polypeptides was through artificial selection, not natural selection. This rational design was formal, not physical. Of
course the objects being manipulated were physical. But the rational design that alone produced the desired homochiral
result was formal.
Take away rational symbol selection, and noise immediately begins to corrupt the functionality of linear digital
sequences. This is true of any other kind of “meaningful” message that depends upon a particular syntax abiding by
semantic and pragmatic rules. In the case of proteins, the “letters” of the sequence comprising the long “word” are
monomers with specific Rgroups. The correct “letters” must be chosen at the individual covalently-bound primary-
structure (sequence) level. This must be done before protein folding ever begins. Spontaneous natural-process peptide
Author's personal copy
216 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
formations have no ability to select and covalently bind “correct” amino acids. A prebiotic environment cannot antic-
ipate future folding and binding needs for complex life.
2.5. The role of variable free-energy binding of amino acids
Amino acid (AA) free-energy binding creates a sequencing bias. Such a physicodynamic bias creates highly or-
dered (patterned) sequences. Not only do the sequences have no foresight into how to generate a folding program, their
highly patterned sequencing precludes programming. Spontaneous peptide formation cannot anticipate the tertiary-
structure binding needs of complex integrated metabolism. Rigid covalent bonds firmly establish sequencing before
hydrogen bonding, hydrophobicities, polarities, and van der Waals forces secondarily fold polyamino acid chains
into needed shapes. Once in a long while the sequencing arising from free-energy preferences could just happen to
correspond to the sequencing needed for functional tertiary structure and metabolic binding function (e.g., a mem-
brane protein of archaebacteria or prokaryote). But such an association would be extremely rare. It certainly would
not establish a cause-and-effect relationship between physicochemical bias and algorithmic linear digital program-
ming of needed tertiary function. The latter requires freely configurable, dynamically inert (dynamically incoherent)
switches [106]. Physicochemical bias precludes such freedom, greatly reducing the Shannon information-retaining
potential of any biopolymer.
Freely configurable switches are observed in all known current life. When proteins are formed by living organisms,
their sequences are much closer to random than highly ordered or patterned by physical bias. Kok et al. [73] and Thax-
ton et al. [127] analyzed twenty-five ancient proteins and observed the peptide bond frequencies to have a distribution
predicted by random statistics rather than physicodynamic bias. The coding of proteins is unrelated to the free-energy
AA bonding tendencies. This is not surprising. Given the mechanism of translation, one would expect freedom from
such bias. Amino acids are located on the opposite ends of each t-RNA from the anticodon. If AA free-energy binding
tendencies determined sequencing, the high information content needed to instruct and control life, its development,
growth, and reproduction would be severely compromised. Other investigators have affirmed the conceptually ideal
nature of genetic coding and control [15,48].
2.6. Self-ordered structures can sometimes mimic true organization
Anti-informationists often appeal to epigenetic factors, prions, self-replicating peptides, regulatory proteins and
small RNAs to argue life is merely physical without need of formal genetic algorithmic control mechanisms. “The
appearance of design” [38] is all that is granted to the most highly organized phenomenon known to science, that of
life. A corollary of this perspective is to often view cases of merely self-ordered objects and events as being evidence
of self-organization.
Garcia-Ruiz et al. were able to synthesize inorganic micron-sized filaments with curved, helical morphologies
suggestive of biological forms [52]. Self-ordering abiotic and inorganic hydrocarbons readily condensed onto these
filaments. The filaments were similar to supposed cyanobacterial microfossils from the Precambrian Warrawoona
chert formation in Western Australia. These structures were once thought to be the oldest terrestrial microfossils.
The authors showed that abiotic and morphologically complex microstructures that are identical to currently accepted
biogenic materials can be synthesized inorganically. Thus self-ordering phenomena can mimic life’s algorithmic or-
ganization. Upon careful examination, such structures were found to be inanimate.
Epigenetic factors notwithstanding, empirical life has invariably been found to be fundamentally cybernetic. Linear,
digital, genetic control of metabolism has been suggested as the most consistent singular criterion for differentiating
life from nonlife [143,146]. No nonliving material in nature manifests this property.
3. Organization
Organization involves choice contingency rather than chance contingency or law-like necessity [2,3]. Organiza-
tion requires purposeful selection from among real options. Organization is fundamentally algorithmic, goal-oriented,
and formal. Organization requires dynamically-inert (dynamically incoherent) configurable switch settings [106] to
instantiate formal choices into a physical matrix. Organization utilizes a sign/symbol/token system to represent those
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 217
choices. Symbols can be alphanumeric or other kinds of physical symbol vehicles. But their function is representa-
tional, not just physicodynamic. Triplet codons function in molecular biology as a Hamming “block code” of physical
symbol vehicles in a sign system. They do not function physicochemically in their coding role. They represent a spec-
ified amino acid which is located on the opposite end of each t-RNA from the anticodon that binds with the codon.
Representationalism using signs/symbols/tokens involves the matter-symbol problem, or philosophy’s “problem of
reference”: How do symbols come to stand for physical structures? [27,57],[92, p. 11],[141].
Every example of true organization can be traced to cybernetic steering [11,34,110,136]. Algorithmic decision-
node programming choices cannot be actualized by physicodynamics. Dynamics cannot generate “choice with intent”
or formal computation. To instantiate free selections at decision nodes into a physical system requires the program-
ming of dynamically-inert configurable switches [106]. These switches are not set by physicodynamic forces and
laws. “Dynamically inert (dynamically incoherent)” was defined in the first paragraph of Section 2.
Instruction sets can be transmitted, received, and understood at the destination using a common coding-decoding
system [119]. Coding systems, like choice with intent, is formal, not physical. Switch settings represent pragmatic
choices at successive decision nodes in order to integrate circuits and instantiate cooperative management into phys-
ical systems [134–136]. Switch positions must be freely selectable to function as logic gates. Switches must be set
expediently according to rules, not laws. Rules can be broken at will, often at the expense of achieving successful
function [92, pp. 15–16],[93, p. 69].
Unlike digital systems, Hoffmeyer [61] considers analog systems to be dynamically coherent processes that spec-
ify a different dynamic system. But formality is still required to specify selections for function when designing an
analog device. Switch settings never violate physical law. But physical law cannot set the switches so as to achieve
formal computational success [2,3,128]. Chance, necessity, and extended periods of time cannot generate algorithmic
programming. No empirical or rational support exists for believing that the environment can generate a sophisticated
configurable system such as a computable cell. In addition, the code appears to be not only formal, but formally ideal
for life [15,47].
A physicodynamically induced sequence could not be trusted to have any algorithmic function. The sequence
would be “ordered” by redundant law and chance variation rather than by formal programming. The sequence would
tend to be either random or highly patterned by law, such as a polyadenosine adsorbed onto a clay surface. Natural
law produces only monotonous, redundant, low informational order. Natural law governance of configurable switches
would set all switches the same, thereby precluding the freedom needed to program those switches. The program
would contain all 0’s, or all 1’s. Nothing imaginative and complex could be computed from such an operating system
or program. Yet biological computation in superior to the finest computer systems in existence [58].
We reluctantly use “self-organization” in our title for the benefit of those searching the all-too-widely-accepted
term in the literature. But we question the rational validity and physical reality of the notion of self-organization.
The only self that can organize its own activities is a living cognitive agent. In models of inanimate nature, including
theorized primordial soup models, we maintain that “self-organization” is a nonsense term that should be critically
analyzed for lack of both meaning and scientific content. The literature is replete with publications using the term
self-organization erroneously to refer to self-ordering events.
Kauffman is one of many who continues his pursuit of “self-organization” out of mere self-ordering and statisti-
cal mechanics phenomena [67]. Purely physicodynamic cause-and-effect events are low-informational. They do not
require engineering choices at successive decision nodes. No configurable switches have to be set in a particular way
at each bifurcation point. Logic gates are irrelevant to the process. In self-ordered structures, the matrix or latticework
is regular in a low-informational repeating pattern. No formal categorizations or teleonomic selective compartmen-
talization is required. The event just spontaneously self-orders by natural process with no selectable goal or desired
function. The phenomenon is not algorithmically integrated. It is not organized by choice contingency. The process
does not require programming to steer it toward formal computational halting. It is not contingent (could have oc-
curred differently). Events occur by “necessity”. It self-orders “by law”, though our understanding of the interaction
of combinations of the particular force fields involved may be poorly understood. The unimaginative matrices will
never generate computations from buttons and strings, or from islands of quantum dots, because they contain no
dynamically-inert, formally set, configurable switches.
Deamer [39],Luisi[126], and many others envision amphiphilic compounds spontaneously self-assembling into
bimolecular layers. These in turn form closed membranous vesicles that supposedly incorporate self-organizing proto-
Author's personal copy
218 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
life. The formation of a bilipid micelle is a classic example of a self-ordering event. No design or engineering decisions
are required. Like a soap bubble, the structure just spontaneously forms through natural process. It self-orders.
Sowerby et al. looked for clues in life-origin science as to how self-reproducing molecular machines might have
originated. They looked to the spontaneous adsorption of purine and pyrimidine bases that become “self-organized”
into monolayers and adsorbed onto the surfaces of crystalline solids [121]. The formation of these layers has nothing
to do with organization. Their alignment and adsorption onto a surface is an example of natural process self-ordering.
Whitesides and Boncheva correctly use the term “self-assembly”: “Self-assembly is a process in which compo-
nents, either separate or linked, spontaneously form ordered aggregates” [142]. Fletcher et al. have developed an
artificial protein scaffold that allows linking of an in-vitro synthesized protein directly to the nucleic acid which
encodes it through a self-assembly process [45]. They too correctly avoid use of the term “self-organization”.
Base-pairing is an example of self-assembly of a complementary strand. It proceeds by a self-ordering natural
process. Only the sequencing of the template’s switch-settings, if that template contains prescriptive information
worthy of being copied, is organizational. But the replication of the template proceeds by natural process except
where catalysis is required by formally instructed enzyme-sequences.
4. Laws and fractals are both compression algorithms
Laws and fractals contain minimal uncertainty, complexity, and information. They represent a unique synthesis of
highly-ordered natural-process patterning with formal algorithmic compression. The formal component of a compres-
sion algorithm is always simple.
4.1. Laws are themselves compression algorithms for reams of empirical data
High order content hidden in experimental data allows compression down to an extremely low-informational for-
mal description. Ocham’s razor states the simplest explanation for any natural phenomenon should be preferred. The
simplest and most pristine formulae are considered “elegant”. Ocham’s razor is prized as a general principle of sci-
ence precisely because law-like behavior is so highly ordered that endless data can be compressed to one simple
equation (e.g., F=ma;E=mc2). Such parsimonious equations are not complex. This does not deny that they have
tremendous pragmatic value. But they nonetheless contain very little Shannon uncertainty or prescriptive information.
No newly discovered law or theory of everything will explain the vast information content found in genomic
instructions. Highly informational algorithmic organization cannot arise from such simple, highly ordered, low in-
formational formulae. Thus quests such as “the search for laws of self-organization and complexity” [64] make no
sense.
4.2. Fractals create the illusion of complexity
Fractals correspond to redundant, monotonous, highly-ordered states. Fractals select members of geometric shapes
for endless reuse. Parts of fractals resemble the whole in a self-similar pattern. Like laws, the instructions producing
fractals are compressible to parsimonious descriptions. The only significant algorithmic aspect of fractals is found in
the compression algorithm that can reduce their high order content to an extremely low informational Kolmogorov
reduction [76]. They have little or no potential to organize or instruct anything sophisticated. Nearly all of the switches
of fractals minimal program are set the same way. Nothing imaginative can result. Almost no instructions are provided,
certainly not the cybernetic prowess required for a living cell or complex eukaryotic organism. The Cantor Dust Fractal
(see Fig. 1) is a prime example: “Take a line segment, remove the center third. Repeat Ntimes”. Such a fractal may
seem “complicated” compared to the same fractal with many fewer iterations, but it certainly is not “complex”.
Maximum complexity is located at the opposite end from maximum order on a uni-dimensional vector that flows
from high order to high complexity [3,143,144]. Maximum complexity is a random sequence with no compressibility.
Maximum order is purely redundant pattern (e.g., a polyadenosine: pick adenosine, repeat Ntimes). Fractals are
anything but random. Fractals are highly ordered and among the least complex of all structures. Hence fractals are of
minimal interest in explaining highly informational organization.
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 219
Fig. 1. The Cantor Dust Fractal representative of the high order content of all fractals. Fractals create the illusion of high complexity. Their low
complexity is demonstrated by the simple Kolmogorov compression algorithm: “Take a line segment, remove the center third. Repeat Ntimes”.
5. Open systems far from equilibrium
Nicolis and Prigogine [86] argued that for any time interval (T ),
S =Si+Se.
Where Siis the change in entropy internal to a system and Seis the change in entropy resulting from an external
flow of mass or energy through that open system. Thaxton et al. [127] summarized some of the thermodynamic
problems associated with life-origin scenarios. In a closed system, Si0 according to the Second Law. Also in a
closed system, Se=0, which is why in a closed system total S is always positive.
When Se>0, the system is open to an influx of mass or energy. But the mere influx of mass/energy by itself does
not necessarily mean the system will become metabolically homeostatic. The influx of raw energy is far more likely
to add to the unusable, wasted energy generated internally by the system. An influx of energy usually only speeds
degradation of the system being intruded.
What generates the possibility of metabolic homeostasis and growth in living organisms is the prescriptive in-
formation [2,3,128]. Prescriptive information either instructs or directly produces sophisticated algorithmic utility. It
comes in the form of dynamically-inert configurable switch-settings [106] that have been programmed in a certain
way so as to successfully compute. Computation has never been observed to arise independent of choice contingency.
Neither chance contingency nor the necessity described by law has ever been observed to produce algorithmic op-
timization. Prescriptive information alone produces organization [2,3,128], not physicodynamics. The engineering
mechanisms generated by this prescriptive information are alone what makes homeostatic metabolism possible in
the cell possible. Such metabolism never violates the second law of thermodynamics. The mechanisms merely make
energy trapping, transduction, and utilization for cellular work possible. Metabolic function is instructed by preset
configurable switches.
One of the sources of so much confusion between self-ordering events and true organization is the failure to
realize that the definition of “entropy” is formal, not physical. “Energy not available for work” presupposes that
we know the definition of “work”. What is “work”? “Work”, like “function”, is a formal engineering concept that
extends far beyond mere physical ISness. Like “shoulds” and “oughts”, work and function lie in a different category
and dimension entirely from mere order and complexity [3]. Even Shannon, Kolmogorov, and Chaitin concepts are
formal. Both high order and high complexity can exist in systems that have no function. Neither work nor function
is reducible to physicality. They both have formal components. When entropy is thought of in purely physical terms,
a serious category error of logic theory has been committed which leads to fallacious inferences.
The result of failing to realize the formal component of entropy is confusing entropy with order/disorder. Both
highly ordered and disordered states can manifest variable amounts of entropy. A highly ordered crystalline structure
can trap and store considerable usable energy, or a crystal can be a near thermodynamic dead end. Highly complex
and disordered states can possess considerable usable energy given the right algorithmic mechanism to harness it.
Entropy is not synonymous with “disorder”. Published scientific literature is replete with nonsense emanating from
this source of confusion.
Another catastrophic result of failing to appreciate the formal component of entropy is trying to define work solely
in terms of physical states. Work is a directed process. Work accomplishes something useful in a larger engineering
context. Mere physicality is blind to concepts of “usefulness” and “usability”. Work is almost always algorithmically
directed. Mere phase changes do not constitute usable work unless those phase changes have been deliberately incor-
porated into an algorithmic scheme. Even analog systems must be designed with choices. Agents value usefulness.
Physicality values nothing. “Energy available for work” entails more than spontaneous phase transitions.
Author's personal copy
220 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
A third problem with failing to realize the formal component of entropy is the confusion of physical entropy with
the Shannon’s probabilistic uncertainty in communication engineering. Physical phase spaces often have no cross-
over with numerical and alphabetical symbol string combinatorial probability spaces [143,145]. As mentioned earlier,
the merging of thermodynamic entropy with Shannon entropy has been called into question [143, p. 84].Thetwo
probability spaces are nonisomorphic even though the formulae are the same. Information theory lacks the integral of
motion present in thermodynamics and statistical mechanics. In thermodynamics and statistical mechanics, there is no
code linking the two “alphabets” of linear stochastic ensembles that we see in the information theory of biomessages.
Kolmogorov–Solomonoff–Chaitin complexity does not reside in the domain of stochastic ensembles of statistical
mechanics. In addition, Boltzmann’s definition of entropy specifically excluded a negative constant from the mathe-
matical definition of entropy [12]. All of the probabilities involved in S are of necessity nonnegative numbers from 0
to 1. This makes the notion of “negentropy” an illegitimate extension from the thermodynamic entropy equation. No
physical system, life included, escapes the domain of the 2nd Law of thermodynamics.
In an open system far from equilibrium, prescriptive information in the form of pre-programmed configurable
switch settings provide the formal mechanisms required to capture, store and use otherwise unusable energy. This
qualifies as “work”. It is accomplished through organization. Solar energy, for example, is wasted energy apart
from algorithmic capabilities instantiated into the control mechanisms of such structures as chloroplasts. Sponta-
neous electron transfers may occur through natural process. But it is the algorithmic harnessing of such energy
transfers into a utilitarian schema that defines and produces “useful work”. Cybernetic steering is required to trans-
duce, store, and call up that energy in a usable form when needed. Without such steering toward eventual function,
organization is impossible. No purely physical system devoid of algorithmic control from a formal source can “self-
organize”.
6. “Self-ordering” and “self-organizing” cannot be used interchangeably
Self-ordering events and algorithmic organization lie in completely different categories with no cross-over and
no set intercepts [3]. Only after instantiation of formal choices into a physical medium of dynamically-inert config-
urable switch-settings are physicodynamic factors secondarily introduced. The selection of desired switch-settings is
not reducible to physicodynamics except in the instructed physical act of throwing the switch. This is the point of
instantiation of cognitive intent into physicality. But the setting of the switch on a horizontal circuit board panel in
one direction rather than another is formal, not physicodynamic.
Stegmann shows that at least one class of genetic information exists wherein molecular processes exhibit the
semantic properties of aboutness, error, and information storage [122]. Aboutness [19,59] is not an attribute of
chance and necessity. No self-ordering event conveys representational meaning, except what might be constructed
within the mind of an interpreting agent. Adami is correct when he insists that “information is always about some-
thing” [4]. But aboutness will not be found in comparing two combinatorial syntaxes alone. As Adami argues
[5], aboutness exists in an environmental context. But significance must be read into, or applied to, sequences by
agents using arbitrary rule-based assignments. The inanimate environment does not make value judgments regarding
meaning, significance, or the worth of potential function. Such assignments of significance are formal, not physi-
cal.
Science, mathematics, and logic demand exact definitions to avoid the invalid logical inferences known as “category
errors”. Linguistic equivocation results in faulty models and conclusions. Once considerable publication capitol is
invested in models and theories, linguistic equivocation often grows into obfuscation by proponents to support the
mirage of their pseudo explanatory mechanism. This approach is not scientific.
Not even a satisfactory hypothetical model has been provided for “self-organization”. Yet the literature liberally
appeals to this imaginary process as proven fact. “Self-organization” is granted causal status for “complex” phenomena
we do not understand. A healthy scientific skepticism is necessary. Organization is not a physical cause. It is the
product of incorporating formal choices into physicality using dynamically-inert, deliberately chosen configurable
switch settings.
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 221
7. “Self-organization” has not been demonstrated to exist
7.1. No empirical evidence
We are aware of no publications demonstrating bona fide “self-organization” in nature. Self-ordering has been
observed. But self-ordering shows no dependence upon, and plays no role in, cybernetic programming. Evidence of
spontaneous generation of algorithmic optimization is simply absent from the literature. Physical objects have never
been observed to spontaneously compute. They self-order. But they do not spontaneously program. Neither random
number generators nor stochastic ensembles have ever been observed to produce computational halting. There is no
reason to ascribe formal creativity to physical ISness. Computational halting is no more derivable from the ISness of
physicality than are “shoulds” and “oughts”. Algorithmic optimization requires steering toward “success” [28,35,83].
Physicality possesses no such motives. It does not and cannot make wise formal programming choices.
Genetic programming is the most sophisticated cybernetic phenomenon known to humans. No human-made system
is as complex as biological computation, organization, and regulation. Life is programmed with degrees of freedom
to respond appropriately to diverse environmental eventualities. For example, in prokaryotic organisms genes are
expressed under certain environmental conditions. No natural process model has come close to explaining the reality
of the formal component to biological cybernetics.
7.2. No rational plausibility
Leslie Orgel [87] examined the plausibility of theories that postulate the development of spontaneous com-
plex chemical organization. He concluded that “theories that involve the organization of complex, small-molecule
metabolic cycles such as the reductive citric acid cycle on mineral surfaces make unreasonable assumptions about the
catalytic properties of minerals and the ability of minerals to organize sequences of disparate reactions”. Orgel argues
that “data in the Beilstein Handbook of Organic Chemistry that have been claimed to support the hypothesis that the re-
ductive citric acid cycle originated as a self-organized cycle can more plausibly be interpreted in a different way” [87].
Even a theorized protometabolism involves far more than just self-replication. Life cannot be reduced to mere self-
replication or miceller GARD’s. GARD’s are “graded autocatalysis replication domains”—mutually catalytic sets of
simple organic molecules envisioned to be capable of self-replication and rudimentary chemical evolution [116–118].
Even a protometabolism requires true organization. Stochastic ensembles possess no organizational skills.
It is common for theorists to limit the discussion to only the first steps toward life. But why would inanimate na-
ture make any steps toward life? Evolution has no predetermined goal or end point. The environment could care less
whether anything functions, let alone whether anything comes to life. In a prebiotic molecular-evolution environment,
no differential survival or differential reproduction exists yet. Natural selection does not exist yet in a prebiotic envi-
ronment. Organization requires selection at the decision-node rigid covalent-bond level of sequence formation. Each
switch setting must be made so as to contribute to eventual computational success. Organization is a formal process,
not a physicodynamic necessity. No rational scientific basis exists for blindly believing in a relentless uphill push by
mere physicality toward formal algorithmic optimization and organizational sophistication.
A minimal degree of integration is required of many cellular biochemical pathways and cycles to produce home-
ostatic metabolism, differential survivability, and differential reproduction. Only then does selection pressure on
phenotypes come into play. It is not sufficient to keep stating “That is too complex to have been there from the begin-
ning”. Simpler Ganti like scenarios [49–51] are too accidental and momentary to be sustained without linear digital
memory and heritability programmatically organizing and maintaining that protometabolism. The wheel would have
to be reinvented with each new generation. Homeostatic metabolism is statistically prohibitive enough as it is as a
one-time event. No Metabolism-First model can be sustained without rapid incorporation of any minimal successes
into a recorded, integrated, heritable, cybernetic scheme.
“Self-organization” is not rationally sound. Cybernetics requires not only “anticipation” of pragmatic needs, but
prescription in the form of a programmable fulfillment strategy. Inanimacy and physical nature cannot anticipate
the future needs of living organisms. Yet anticipation is needed to program any kind of step-wise procedure for
sophisticated utility/function. Random number generators do not produce computational halting. Neither does fixed
law-like behavior.
Author's personal copy
222 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
Physicality is not logically capable of generating the abstract representationalism inherent in sign system use.
Chance and necessity cannot generate choice contingency required for symbol selection and arbitrary assignment of
meaning to those symbols [2,3,128]. The physical symbol vehicles themselves are not capable of generating rules and
engineering cooperative integration between parts. It is irrational to attribute algorithmic organization to “self” when
“self” consists of inanimate objects that cannot choose. If there is any universal criterion of agency, it is the ability to
choose with intent. Programming is impossible without this attribute. Organization requires programmed integration.
Components must be formally controlled into a cooperative effort to produce a functional cell.
Physicality is divided from symbolization by an “epistemic cut” [60],[90, p. 36],[94]. A description of an event is
separated from the objective event itself. “Semantic closure” must occur [91,92,130,132]. This is similar to the mea-
surement problem in quantum theory. The genotype-phenotype dichotomy in biology also portrays a similar epistemic
cut. Options cannot be selected stochastically if any engineering function is expected. Symbols must be selected for
their representational value. The meaning of these symbols is defined by their function. Codons, like the letters of
words, function as symbols, not as chemical reactants. In the absence of repeated observation, falsifiability, prediction
fulfillment, and rational respectability, the notion of self-organization is wishful thinking and speculation [88].
We question Eigen’s recent definition of information “meaning” [41]. “The conclusion is reached that information
content is generated via selection, which can be described as a phase transition in information space”. We agree with
Eigen that “information content [what we would call “prescriptive information” content, at least] is generated via
selection”. But we challenge the ability of phase transitions to reflect any of the attributes of selection pressure or
of biological instructions (not all of which are genetic). A more constructive definition of “meaning” in biological
information is, “The meaning of bioinformation is the biofunction that information instructs or produces at its des-
tination” [1]. Phase transitions do not instruct anything, nor do they contribute to biological algorithms. They fail to
explain selections at the bifurcation points of networks and neural net equivalents. They cannot explain efficacious
settings of configurable switches. We need a better paradigm than phase changes to explain selection of functional
monomeric sequencing.
Biopolymers are essentially strings of decision nodes. Monomers function as alphabetical symbols in a sign sys-
tem [2]. The decision at each node is more than a mere coin flip [128]. Self-replication of “gibberish” strings offers
far less to any life origin model than we suppose. Each nucleotide in an oligonucleotide represents the equivalent
of a four-way fork in the road. The right road must be taken at many successive four-way forks prior to the realiza-
tion of any after-the-fact phenotypic biofunction. We need a mechanism that explains the correct selection of each
of these road options. After-the-fact phenotypic selection does not explain how the computed phenotype came into
existence. We cannot continue pointing to clay or lipid templates without explaining where/how the template acquired
its instructive sequencing.
7.3. No prediction fulfillment
A key tool of verification in science is the ability to predict future physicodynamic events using nonphysical, formal
models. Not one realization of prediction of self-organization of a nontrivial cybernetic system has been experienced
and published, to our knowledge. Given the millions of supposedly self-organized systems that exist in the empirical
world, there should be no problem observing such a prediction fulfillment of spontaneous self-organization. Little
reason or excuse remains for clinging to the metaphysical presupposition of self-organization.
The Stigmergic Systems website [120] defines “Self-organization” as the “unplanned organization that emerges
from an open system of interacting components. The system can be thought of as lifting itself up by its boot straps”.
By what scientific mechanism does this self-help system arise in nature? The website responds, “It is impossible to
even begin to understand how all these different influences have an effect on the organization that emerges”. It is
impossible for these authors to distinguish such self-organization dreams from “wish fulfillment”. Any biophysical
explanation is absent.
When predictions of infrequent events fail, it is common to appeal to long periods of time as a caveat. But the
longest possible period of time based on current estimates is 1017 seconds (14 billion years). This maximum time
can be factored into probabilities to adjust for infrequent events. Even after multiplying probabilities by 1017,self-
organization of life models yield probabilities far less than every published probability bound. The most widely
referenced probability bound in scientific literature is also the most conservative: 1050 [14, p. 28]. The most lib-
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 223
eral and permissive is 10150 [40]. The latter probability bound grants chance enormous potential. Yet most models
of biological self-organization are far less probable than 10150.
7.4. No falsifiability
The notion of self-organization is not falsifiable by hypothesis-driven scientific method. Hypotheses must be
testable, especially to merit respectability as a scientific theory. No experimental design has been offered that can
test and potentially falsify such an open-ended postulate as hoped-for self-organization. Its source is not scientific,
but metaphysical. In the conclusions of this paper we offer two testable null hypotheses that are potentially falsifiable.
Demonstration of a single exception would falsify either null hypothesis.
Kurakin [75] provides an example of many investigators who are progressively abandoning mechanistic concepts
of molecular motors and protein translocations in favor of stochastic models. The latter are seen as being a more
fruitful conceptual framework for understanding biological organization at the molecular level. Once stochastic self-
organization has been metaphysically presupposed as the central integrating theme of biology, the paradigm takes
precedence over empirical, predictive, and rational inconsistencies. The lack of falsifiability of self-organization leaves
the notion with widespread acceptance.
8. The problem of prescriptive sequencing
The problem of prescriptive and functional sequencing exists whether we begin our life-origin model in a Pep-
tide/Protein World [109],orinaPreRNA/RNAWorld[53]. The first question that must be resolved is the degree to
which polymerization is dynamically inert. This is the degree to which sequencing is dynamically incoherent rather
than coherent. Incoherence is freedom from physicochemical cause-and-effect determinism. Physicodynamic inco-
herence in turn subdivides into two subcategories of interest to life-origin science: how dynamically inert is:
(1) peptide bond formation between the 20 main biological amino acids, and
(2) 35-phosphodiester bond formation between the five heterocyclic nucleotides.
In both categories of untemplated polymerization in a prebiotic environment, physicochemical self-ordering ten-
dencies would severely preclude instantiation of prescriptive information into the sequence. The reason is that freedom
would be restricted in coding selections if physicochemical “necessity” forced certain selections. Law-like determin-
ism would tend to code every sequence the same. Physicochemical bias would reduce each linear digital programming
“choice” of monomer to a noncontingent, fixed, cause-and-effect result with only mild random variation of alterna-
tives.
Let us begin with the spontaneous formation of polyamino acids. As mentioned briefly above, Kok, Taylor and
Bradley [73,127] disproved the notion that differences in chemical binding forces between amino acids would lead to
nonrandom dipeptide distributions. If amino acids Aand Bhave the strongest bonding affinities for each other in a
mixture of A,B,C,D,E,F, and G, then we would expect many more AB dipeptides to form than AG’s o r BEs. Kok,
Taylor and Bradley analyzed 25 different proteins, finding a nearly random distribution of peptide bond frequencies.
Bonding preferences appeared to play almost no role in protein coding [73]. Sequences were dynamically incoherent
with cause-and-effect physical law. They were dynamically inert.
In current life, polyamino acid sequencing is prescribed by codon sequence, not by amino acid free-energy bonding
preferences. If in a Peptide World bonding forces determined sequencing, the high degree of self-ordering would have
produced low-informational, monotonous, redundant sequences of the same few amino acids. The peptides would
have lacked the many specific functionalities needed even for the simplest of protometabolisms. All spontaneously
formed peptides would have tended to be the same. This would have restricted the sample space of three-dimensional
shapes that is invariably cited as mechanism for molecular evolution [46,133]. Given the short period of time between
earth’s cooling (4.0 to 3.9 Ga) and the 3.8 billion-year-age of life [85,131], life could not have evolved from a peptide
world out of such a highly self-ordered minimal sample space. The nearly ideal nature of genetic coding could only
have arisen out of freedom of programming selection for function [47].
Aside from templating and base-pairing, ribonucleotides show little physicochemical bias in polymerization of in
aqueous solution. Any differential availability of nucleotides in a primordial soup would have produced weighted
Author's personal copy
224 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
means in the formation of stochastic ensembles, however. But this would only reduce the number of bits of Shan-
non uncertainty and complexity. Information retention in the RNA matrix of a forming RNA World would have been
severely restricted by both the limited availability of certain nucleotides, and/or by varying nucleotide binding ener-
gies. The key to instantiation of prescriptive information into nucleic acid sequences is the fact that polymerization of
nucleotides corresponds to freely configurable switch settings. Physical laws do not determine sequencing in aqueous
solution. Only templating and the replication of existing informational sequences by base-pairing is physicodynami-
cally constrained. Source code, on the other hand, is truly controlled by symbol selection and by encryption/decryption
rules, not laws and constraints. The fixed laws of nature cannot explain formal instructions or coding function.
9. Discussion
Cybernetics is not achieved through physicodynamic constraints and cause-and-effect laws. It is achieved using
arbitrary rules governing a theory of logic and “language convention”. Expedient programming choices must be made
at each critical decision node. Not all decision nodes in biopolymers are critical. Some sections of primary structure
become buried in the tertiary structure of the protein where the Rgroups play little or no role.
The selection of physical symbol vehicles and messages are inert with respect to dynamics [106]. Semiotic function
is not achieved through law-like behavior. It is achieved through formal cybernetic prescription [74]. Prescription
can be achieved only through the sequencing of particular alphabetic symbol selections and/or configurable switch-
settings. Switches can be freely set because no physicodynamic preference exists for any of the four nucleotides to
bind in a nontemplated single-stranded positive strand.
Only the replication of an existing template is physicodynamic. But the programming of the template itself must be
formal, not physicodynamic, if the template is to convey nontrivial prescriptive information and functional meaning to
its replicant. No physical theory can explain this dynamically-inert, digital coding function [143,146]. Self-ordering
could never generate the freely programmable configurable-switch-settings required for organization. Functional
choices at successive decision nodes cannot be made by chance or by fixed laws. Self-ordered phenomena can-
not program the organization of algorithmic systems. Genetic control is algorithmic and cybernetic. It employs a
representational sign system, dynamically-inert configurable switches, coding encryption/decryption, formal compu-
tational halting, and meaningful messages understood in terms of eventual metabolic success. Cellular and organismic
processes require all of the above. These lead to differential survival and reproduction required for evolution and
diversification of species. But natural selection cannot generate the genetic programs that compute the fittest already
living phenotypes. Selection pressure favors only existing, already-computed, phenotypes.
“Selection for function” has escaped quantification efforts in information science. Yet it is this selection for function
that is the essence of organization. It occurs at successive decision nodes. Both functional and nonfunctional selections
are possible at each decision node within the constraints of the laws of physics. Physicodynamics cannot make effi-
cacious choices. If law made the choice, all switches would be set the same. No computational function would result.
If the choice were random, no sophisticated utility would result. Neither chance nor necessity can explain molecular
machines and computable cells. Protein and ribozyme conformation is prescribed by the sequencing of primary struc-
ture. Primary structure in turn determines minimum-free-energy folding constraints into a three-dimensional structure.
Sequencing is formal, not physical. In biology, the sequencing of nucleotide selections into digital prescriptions and
the use of a formal encryption/decryption, rule-based system is what organizes life. No such cybernetic system can
self-order or self-assemble.
The emergence of agents is not possible from a connectionist state-determined system [80]. Neural networks and
connectionist models are dynamically coupled to, or coherent with, their environment [30,103]. The aim of research
into connectionist systems is to be able to explain emergent classifications (Eigen-behavior). This classification is
considered emergent because “it is the global result of the local, state-determined, interaction of the basic components
of the self-organizing system with its environment” [105]. But such a system precludes the most fundamental aspect
of agency: choice contingency. Choice contingency in turn requires freedom from cause-and-effect determinism and
random noise at configurable switches. Agency is able to choose with intent. Connectionist systems cannot properly
be called “embodied agents” because connectionist systems are dynamically coherent with their environments. If
embodied agents were connectionist systems, no mind-body problem would exist to ponder. Thoughts would be
determined by cause-and-effect physicodynamics [108, p. 7]. Thoughts would be either random or self-ordered into a
fixed regularity. Logic gates would produce either noise, or be locked into one fixed position by necessity. Dynamically
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 225
coherent agent-environment couplings cannot, therefore, give rise to “embodied cognition”. The connectionist system
cannot be isolated from its environmental determinism to achieve choice contingency.
Molecular biology preceded temporally and causally all forms of brain-secreted cognition. Each amino acid pre-
scription is the “letter” of a protein “word”. Biomessages have real meaning. No direct physicochemical binding or
reaction occurs between the codon symbol and the prescribed amino acid. The genetic code table is formal, not phys-
ical. No amount of physicochemical self-ordering events can explain its abstract concept and freedom from dynamic
constraints. Infodynamic reductionism cannot explain the literal genetic algorithms of living organisms. Systems bi-
ology must confront the need for selection at the genetic level [124]. Although integrated algorithmic systems can
be instantiated into physicality, they are fundamentally formal, not physical. Physicodynamics is incapable of self-
organizing such a formal system of genetic control.
Separate manuscripts will address why neither hypercycles nor evolutionary algorithms can give rise to bona fide
self-organization.
10. Conclusions
The numerous published papers referring to inanimate “self-organization” usually fall into one of three categories:
(1) The authors are mistakenly referring to natural-process self-ordering phenomena that have nothing to do with
algorithmic organization,
(2) The process described involves hidden experimenter interference (artificial selection and investigator involvement
in the experimental design). The most common form of behind-the-scenes investigator involvement consists of
choosing which iterations to pursue in pseudo evolutionary algorithms (e.g., ribozyme engineering experiments
using SELEX [42,104,129]).
(3) The term is used to refer to an imagined ingenuity of physicality for which no scientifically verifiable reality
exists.
We propose the following fully falsifiable null hypotheses to test the ingenuity of spontaneous physicodynamics:
(A) Self-ordering phenomena cannot generate cybernetic organization.
(B) Randomness cannot generate cybernetic organization.
We invite falsification of either or both of these null hypotheses. Demonstration of a single exception would falsify
either null hypothesis.
Our contention is that all organizational systems are fundamentally formal. Dissipative structures (e.g., hurricanes,
tornadoes, candle flames) are not true organizational systems. They are a string of moment-by-moment self-ordered
states incapable of organizing physicality into nontrivial computational utility. If anything, dissipative structures de-
stroy physically instantiated, algorithmically programmed material sign systems (MSS) [106] and the functionality
they engineer.
We maintain that “self-organization” is a nonsense term when applied to inanimate nature and pre-biotic molecular
evolution. The notion is without empirical and prediction-fulfillment support. “Self-organization” is not falsifiable,
provides no mechanism, and offers no plausible explanatory power. The concept fails every test of scientific re-
spectability, and should be skeptically viewed until further research provides more verifiable substance to the notion.
Acknowledgements
Dr. Abel is supported by grants from The Origin of Life Foundation, Inc., a US 501-(c)-3 science foundation. JTT
acknowledges the support of the Canadian Foundation for Innovation and the Ontario Innovation Trust.
References
[1] Abel DL. Is life reducible to complexity? In: Workshop on life: a satellite meeting before the Millennial World Meeting of University
Professors. Modena, Italy: University of Modena and Reggio Emilia; 2000. p. 3–4.
[2] Abel DL, Trevors JT. More than metaphor: Genomes are objective sign systems. J Biosemiotics 2006;1(2):253–67.
Author's personal copy
226 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
[3] Abel DL, Trevors JT. Three subsets of sequence complexity and their relevance to biopolymeric information. Theoretical Biology and
Medical Modeling 2005;2(29), open access at http://www.tbiomed.com/content/2/1/29.
[4] Adami C. Introduction to artificial life. New York: Springer/Telos; 1998.
[5] Adami C. Information theory in molecular biology. Phys Life Rev 2004;1:3–22.
[6] Amos LA. Focusing-in on microtubules. Curr Opin Struct Biol 2000;10:236–41.
[7] Balanovski E, Beaconsfield P. Order and disorder in biophysical systems: A study of the correlation between structure and function of DNA.
J Theor Biol 1985;114:21–33.
[8] Bar-Ziv R, Tlusty T, Libchaber A. Protein-DNA computation by stochastic assembly cascade. Proc Nat Acad Sci 2002;99:11589–92.
[9] Bennett CH. On the logical “depth” of sequences and their reducibilities to incompressible sequences. 1989.
[10] Bennett DH. Logical depth and physical complexity. In: Herken R, editor. The universal turing machine: A half-century survey. Oxford:
Oxford University Pres; 1988.
[11] Bertalanffy von L. The history and status of general system theory. In: Klir GJ, editor. Trends in general systems theory. New York: John
Wiley; 1972. p. 53–4.
[12] Boltzmann L. Weitere Studien über das Wärmegleichgewicht unter Gasmolekulen. In: Sitzungsberichte II Abteilung, vol. 66. Wien:
Königliche Academie der Wisschenshaft; 1877. p. 275.
[13] Boniolo G. Biology without information. History Philosophy Life Sci 2003;25:255–73.
[14] Borel E. Probabilities and life. New York: Dover; 1962.
[15] Bradley D. Informatics. The genome chose its alphabet with care. Science 2002;297:1789–91.
[16] Brillouin L. The negentropy principle of information. J Appl Phys 1953;24:1153.
[17] Brillouin L. Science and information theory. New York: Academic Press; 1962.
[18] Brillouin L. Life, thermodynamics, and cybernetics. In: Maxwell’s demon, entropy, information, and computing. Princeton, NJ: Princeton
University Press; 1990.
[19] Bruza PD, Song DW, Wong KF. Aboutness from a common sense perspective. JASIS 2000;51:1090–105.
[20] Bujdak J, Eder A, Yongyai Y, Faybikova K, Rode BM. Investigation on the mechanism of peptide chain prolongation on montmorillonite.
J Inorganic Biochem 1996;61:69–78.
[21] Bujdak J, Remko M, Rode BM. Selective adsorption and reactivity of dipeptide stereoisomers in clay mineral suspension. J Colloid Interface
Sci 2006;294:304–8.
[22] Bujdak J, Rode BM. The effect of smectite composition on the catalysis of peptide bond formation. J Molecular Evolution 1996;43:326–33.
[23] Bujdak J, Rode BM. On the mechanisms of oligopeptide reactions in solution and clay dispersion. J Peptide Sci 2004;10:731–7.
[24] Bunge M. Realism and antirealism in social sciences. Theor Decis 1993;35:207–35.
[25] Cairns-Smith AG. The origin of life and the nature of the primitive gene. J Theor Biol 1966;10:53–88.
[26] Cairns-Smith AG. Seven clues to the origin of life. Cambridge: Cambridge University Press; 1990.
[27] Cassirer E. The philosophy of symbolic forms, vol 3: The phenomena of knowledge. New Haven, CT: Yale University Press; 1957.
[28] Chambers L. The practical handbook of genetic algorithms applications. 2nd ed. Washington, DC: Chapman & Hall/CRC; 2001.
[29] Chargaff E. Essays on nucleic acids. Amsterdam: Elsevier; 1963.
[30] Chen H. Machine learning for information retrieval: Neural networks, symbolic learning, and genetic algorithms. JASIST 1995;46:194–216.
[31] Clark J, Shevchuk T, Swiderski P, Dabur R, Crocitto L, Buryanov Y, Smith S. Mobility-shift analysis with microfluidics chips. Biotechniques
2003;35:548–54.
[32] Comazine S, Deneubourg J-L, Franks N, Sneyd J, Theraulaz G, Bonabeau E. Self-organization is biological systems. Princeton, NJ: Princeton
University Press; 2001.
[33] Csilling A, Janosi IM, Pasztor G, Scheuring II. Absence of chaos in a self-organized critical coupled map lattice. Physical Review E:
Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics 1994;50:1083–92.
[34] Danesi M. Modeling systems theory. J Biosemiotics 2005;1:213–25.
[35] Davis L. Handbook of genetic algorithms. New York: Van Nostrand Reinhold; 1991.
[36] Dawkins R. The selfish gene. 2nd ed. Oxford: Oxford Univerisy Press; 1989; 1976.
[37] Dawkins R. The blind watchmaker. New York: WW Norton and Co; 1986.
[38] Dawkins R. Climbing mount impossible. 1996.
[39] Deamer DW, Akeson M. Nanopores and nucleic acids: Prospects for ultrarapid sequencing. Trends Biotechnol 2000;18:147–51.
[40] Dembski WA. No free lunch. New York: Rowman and Littlefield; 2002.
[41] Eigen M. Natural selection: a phase transition? Biophys Chem 2000;85:101–23.
[42] Ellington AD, Szostak JW. In vitro selection of rna molecules that bind specific ligands. Nature 1990;346:818–22.
[43] Ferris JP. Prebiotic synthesis on minerals: Bridging the prebiotic and rna worlds. Biol Bull 1999;196:311–4.
[44] Ferris JP, Hill Jr AR, Liu R, Orgel LE. Synthesis of long prebiotic oligomers on mineral surfaces. Nature 1996;381:59–61.
[45] Fletcher G, Mason S, Terrett J, Soloviev M. Self-assembly of proteins and their nucleic acids. J Nanobiotechnol 2003;1:1.
[46] Fontana W, Schuster P. Shaping space: The possible and the attainable in rna genotype-phenotype mapping. J Theor Biol 1998;194:491–515.
[47] Freeland SJ, Hurst LD. The genetic code is one in a million. J Mol Evol 1998;47:238–48.
[48] Freeland SJ, Knight RD, Landweber LF, Hurst LD. Early fixation of an optimal genetic code. Mol Biol Evol 2000;17:511–8.
[49] Gánti T. Biogenesis itself. J Theor Biol 1997;187:583–93.
[50] Gánti T. On the early evolutionary origin of biological periodicity. Cell Biol Int 2002;26:729–35.
[51] Gánti T. The principles of life. Oxford, UK: Oxford University Press; 2003.
[52] Garcia-Ruiz JM, Hyde ST, Carnerup AM, Christy AG, Van Kranendonk MJ, Welham NJ. Self-assembled silica-carbonate structures and
detection of ancient microfossils. Science 2003;302:1194–7.
Author's personal copy
D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228 227
[53] Gesteland RF, Cech TR, Atkins JF. The rna world. Cold Spring Harbor: Cold Spring Harbor Laboratory Press; 1999.
[54] Gleick J. Chaos: Making a new science. New York: Penguin Books; 1987.
[55] Goodwin BC. Evolution and the generative order. In: Goodwin BC, Saunders P, editors. Theoretical biology: Epigenetic and evolutionary
order from complex systems. Edinburgh: Edinburgh University Press; 1989.
[56] Gordon KHJ. Were rna replication and translation directly coupled in the rna (+protein?) world? J Theor Biol 1995;173:179–93.
[57] Harnad S. The symbol grounding problem. Physica D 1990;42:335–46.
[58] Higgins CM. Sensory architectures for biologically inspired autonomous robotics. Biol Bull 2001;200:235–42.
[59] Hjorland B. Towards a theory of aboutness, subject, topicallity, theme, domain, field, content, and relevance. JASIST 2001;52:774–8.
[60] Hoffmeyer J. Code-duality and the epistemic cut. Ann NY Acad Sci 2000;901:175–86.
[61] Hoffmeyer J. Life and reference. Biosystems 2001:60.
[62] Huang W, Ferris JP. Synthesis of 35–40 mers of rna oligomers from unblocked monomers. A simple approach to the rna world. Chem
Commun (Camb) 2003;12:1458–9.
[63] Janosi IM, Scheuring II. Reply to “comment on ‘absence of chaos in a self-organized critical coupled map lattice’ ”. Physical Review E:
Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics 1995;52:2116–7.
[64] Kauffman S. At home in the universe: The search for the laws of self-organization and complexity. New York: Oxford University Press;
1995.
[65] Kauffman SA. Antichaos and adaptation. Sci Am 1991;265:78–84.
[66] Kauffman SA. The origins of order: Self-organization and selection in evolution. Oxford: Oxford University Press; 1993.
[67] Kauffman SA. Investigations. New York: Oxford University Press; 2000.
[68] Kauffman SA, Johnsen S. Coevolution to the edge of chaos: Coupled fitness landscapes, poised states, and coevolutionary avalanches. J Theor
Biol 1991;149:467–505.
[69] Kay L. Who wrote the book of life? A history of the genetic code. Stanford, CA: Stanford University Press; 2000.
[70] Keller EF. The century of the gene. Cambridge, MA: Harvard University Press; 2000.
[71] Keller EF. Decoding the genetic program. In: Beurton P, Falk R, Rheinberger H-J, editors. The concept of the gene in development and
evolution. Cambridge: Cambridge University Press; 2000. p. 159–77.
[72] Kitcher P. Battling the undead; how (and how not) to resist genetic determinism. In: Singh RS, et al., editors. Thinking about evolution:
Historical philosophical and political perspectives. Cambridge: Cambridge University Press; 2001. p. 396–414.
[73] Kok RA, Taylor JA, Bradley WL. A statistical examination of self-ordering of amino acids in proteins. Origins Life Evol Biosph
1988;18:135–42.
[74] Korzeniewski B. Confrontation of the cybernetic definition of a living individual with the real world. Acta Biotheor 2005;53:1–28.
[75] Kurakin A. Self-organization versus watchmaker: Molecular motors and protein translocation. Biosystems 2006;84:15–23.
[76] Li M, Vitanyi P. An introduction to Kolmogorov complexity and its applications. New York: Springer; 1997.
[77] Liiv E. Infodynamics: Generalized entropy and negentropy. http://staff.ttu.ee/~elmliiv/publications.htm (Last accessed June, 2006).
[78] Lwoff A. Biological order. Cambridge, MA: MIT Press; 1962.
[79] Mahner M, Bunge MA. Foundations of biophilosophy. Berlin: Springer; 1997.
[80] McCllelland JL, Hinton GE. The appeal of parallel distributed processing. In: Rumelhart DE, McClellans JL, editors. Parallel distributed
processing: Explorations in the microstructure of cognition, vol 1: Foundations. Cambridge, MA: MIT Press; 1986. p. 3–44.
[81] Mellersh A, Wilkinson AS. RNA bound to a solid phase can select an amino acid and facilitate subsequent amide bond formation. Orig Life
Evol Biosph 2000;30:3–7.
[82] Mikulecky DC. Complexity, communication between cells, and identifying the functional components of living systems: Some observations.
Acta Biotheor 1996;44:179–208.
[83] Mitchell M. An introduction to genetic algorithms. Bradford Books; 1998.
[84] Miyakawa S, Ferris JP. Sequence- and regioselectivity in the montmorillonite-catalyzed synthesis of rna. J Am Chem Soc 2003;125:8202–8.
[85] Mojzsis SJ, Arrhenius G, McKeegan KD, Harrison TM, Nutman AP, Friend GRL. Evidence for life on earth before 3,800 million years ago.
Nature 1996;384:55–9.
[86] Nicolis G, Prigogine I. Self-organization in nonequilibrium systems: From dissipative structures to order through fluctuations. New York:
Wiley–Interscience; 1977.
[87] Orgel LE. Self-organizing biochemical cycles. Proc Natl Acad Sci USA 2000;97:12503–7.
[88] Overman DL. A case against accident and self-organization. New York: Rowman and Littlefield Publishers; 1997.
[89] Papaseit C, Pochon N, Tabony J. Microtubule self-organization is gravity-dependent. Proc Natl Acad Sci USA 2000;97:8364–8.
[90] Pattee HH. The evolution of self-simplifying systems. In: Laszlo E, editor. The relevance of general systems theory. New York: George
Braziller; 1972. p. 32–41; p. 193–195.
[91] Pattee HH. Cell psychology: An evolutionary approach to the symbol-matter problem. Cognition Brain Theory 1982;5:325–41.
[92] Pattee HH. Evolving self-reference: Matter, symbols, and semantic closure. Commun Cognition – Artificial Intelligence 1995;12:9–28.
[93] Pattee HH. Causation, control, and the evolution of complexity. In: Andersen PB, Emmeche C, Finnemann NO, Christiansen PV, editors.
Downward causation: Minds, bodies, and matter. Aarhus, DK: Aarhus University Press; 2000. p. 63–77.
[94] Pattee HH. The physics of symbols: Bridging the epistemic cut. Biosystems 2001;60:5–21.
[95] Pincus SM. Irregularity and asynchrony in biologic network signals. Methods Enzymol 2000;321:149–82.
[96] Plankensteiner K, Reiner H, Rode BM. Catalytically increased prebiotic peptide formation: Ditryptophan, dilysine, and diserine. Origins
Life Evol Biosph 2005;35:411–9.
[97] Plankensteiner K, Reiner H, Rode BM. Stereoselective differentiation in the salt-induced peptide formation reaction and its relevance for the
origin of life. Peptides 2005;26:535–41.
Author's personal copy
228 D.L. Abel, J.T. Trevors / Physics of Life Reviews 3 (2006) 211–228
[98] Pohl K, Bartelt MC, de la Figuera J, Bartelt NC, Hrbek J, Hwang RQ. Identifying the forces responsible for self-organization of nanostructures
at crystal surfaces. Nature 1998;1999:238–41.
[99] Prigogine I. From being to becoming. San Francisco: WH Freeman and Co; 1980.
[100] Prigogine I, Mayne F, George C, de Haan M. Microscopic theory of irreversible processes. Proc Nat Acad Sci USA 1977;74:4152–6.
[101] Prigogine I, Stengers I. Order out of chaos. London: Heinemann; 1984.
[102] Reblova K, Spackova Na, Sponer JE, Koca J, Sponer J. Molecular dynamics simulations of rna kissing-loop motifs reveal structural dynamics
and formation of cation-binding pockets. Nucl Acids Res 2003;31:6942–52.
[103] Reidys C, Forst CV, Schuster P. Replication and mutation on neutral networks. Bull Math Biol 2001;63:57–94.
[104] Robertson DL, Joyce GF. Selection in virtro of an rna enzyme that specifically cleaves single-stranded DNA. Nature 1990;344:467–8.
[105] Rocha LM. Eigenbehavior and symbols. Sys Res 1996;13:371–84.
[106] Rocha LM. Evolution with material symbol systems. Biosystems 2001;60:95–121.
[107] Rocha LM, Hordijk W. Representations and emergent symbol systems. Cognitive Sci 2000.
[108] Rocha LM, Hordijk W. Material representations: From the genetic code to the evolution of cellular automata. Artif Life 2005;11:189–214.
[109] Rode BM. Peptides and the origin of life. Peptides 1999;20:773–86.
[110] Rosen R. Drawing the boundary between subject and object: comments on the mind-brain problem. Theor Med 1993;14:89–100.
[111] Rothemund PW, Ekani-Nkodo A, Papadakis N, Kumar A, Fygenson DK, Winfree E. Design and characterization of programmable DNA
nanotubes. J Am Chem Soc 2004;126:16344–52.
[112] Saghatelian A, Yokobayashi Y, Soltani K, Ghadiri MR. A chiroselective peptide replicator. Nature 2001;409:797–801.
[113] Salthe SN. Meaning in nature: Placing biosemitotics within pansemiotics. J Biosemiotics 2005;1:287–301.
[114] Sarkar S. Information in genetics and developmental biology: Comments on Maynard Smith. Philos Sci 2000;67:208–13.
[115] Schrödinger E. What is life: The physical aspect of the living cell. Cambridge: Cambridge University Press; 1944.
[116] Segre D, Ben-Eli D, Deamer DW, Lancet D. The lipid world. Orig Life Evol Biosph 2001;31:119–45.
[117] Segre D, Ben-Eli D, Lancet D. Compositional genomes: Prebiotic information transfer in mutually catalytic noncovalent assemblies. Proc
Natl Acad Sci USA 2000;97:4112–7.
[118] Segre D, Lancet D, Kedem O, Pilpel Y. Graded autocatalysis replication domain (gard): Kinetic analysis of self-replication in mutually
catalytic sets. Orig Life Evol Biosph 1998;28:501–14.
[119] Shannon C. Part i and ii: A mathematical theory of communication. Bell Syst Tech J 1948;XXVII:379–423.
[120] Small P. Stigmergic systems. http://www.stigmergicsystems.com/5/1/2006.
[121] Sowerby SJ, Holm NG, Petersen GB. Origins of life: A route to nanotechnology. Biosystems 2001;61:69–78.
[122] Stegmann UE. Genetic information as instructional content. Philos Sci 2005;72:425–43.
[123] Stomp AM. Genetic information and ecosystem health: Arguments for the application of chaos theory to identify boundary conditions for
ecosystem management. Environ Health Perspect 1994;102(Suppl 12):71–4.
[124] Strange K. The end of “naive reductionism”: Rise of systems biology or renaissance of physiology? Am J Physiol Cell Physiol
2005;288:C968-74.
[125] Surrey T, Nedelec F, Leibler S, Karsenti E. Physical properties determining self-organization of motors and microtubules. Science
2001;292:1167–71.
[126] Szostak JW, Bartel DP, Luisi PL. Synthesizing life. Nature 2001;409:387–90.
[127] Thaxton CB, Bradley WL, Olsen RL. The mystery of life’s origin: Reassessing current theories. Dallas, TX: Lewis and Stanley; 1984.
[128] Trevors JT, Abel DL. Chance and necessity do not explain the origin of life. Cell Biol Int 2004;28:729–39.
[129] Tuerk C, Gold L. Systematic evolution of ligands by exponential enrichment—rna ligands to bacteriophage—t4 DNA-polymerase. Science
1990;249:505–10.
[130] Umerez J. Semantic closure: A guiding notion to ground artificial life. In: Moran F, Moreno JJ, Chacon P, editors. Advances in artificial life.
Berlin: Springer; 1995. p. 77–94.
[131] Van Zuilen MA, Lepland A, Arrhenius G. Reassessing the evidence for the earliest traces of life. Nature 2002;418:627–30.
[132] Vaneechoutte M. The scientific origin of life. Considerations on the evolution of information, leading to an alternative proposal for explaining
the origin of the cell, a semantically closed system. Ann NY Acad Sci 2000;901:139–47.
[133] Viedma C. Formation of peptide bonds from metastable versus crystalline phase: Implications for the origin of life. Orig Life Evol Biosph
2000;30:549–56.
[134] von Neumann J. The general and logical theory of automata. In: Newman JR, editor. The world of mathematics, vol 4. New York: Simon
and Schuster; 1956.
[135] von Neumann J, Aspray W, Burks AW. Papers of John von Neumann on computing and computer theory. Cambridge, MA: MIT Press,
Tomash Publishers; 1987.
[136] von Neumann J, Burks AW. Theory of self-reproducing automata. Urbana: University of Illinois Press; 1966.
[137] Waldrop MM. Complexity. New York: Simon and Schuster; 1992.
[138] Weaver W. The mathematics of communication. Sci Am 1949.
[139] Weber BH, Depew DJ. Natural selection and self-organization. Biol Philos 1995.
[140] Weizsäcker von E, Weizsäcker von C. Information, evolution and “error-friendliness”. Biol Cyber 1998;79:501–6.
[141] Whitehead AN. Symbolism: Its meaning and effect. New York: Macmillan; 1927.
[142] Whitesides GM, Boncheva M. Supramolecular chemistry and self-assembly special feature: Beyond molecules: Self-assembly of mesoscopic
and macroscopic components. Proc Nat Acad Sci 2002;99:4769–74.
[143] Yockey HP. Information theory and molecular biology. Cambridge: Cambridge University Press; 1992.
[144] Yockey HP. Origin of life on earth and Shannon’s theory of communication. Comput Chem 2000;24:105–23.
[145] Yockey HP. Information theory, evolution and the origin of life. Inform Sciences 2002;141:219–25.
[146] Yockey HP. Information theory, evolution, and the origin of life. Cambridge: Cambridge University Press; 2005.
... The most common cases of structuration at this level are self-ordering and self-assembly. Self-ordering refers to the capacity of molecules to form stereoregular shapes, such as spherically formed micelles (Abel & Trevors, 2006), whereas self-assembly refers to the more basic capacity of molecules to accumulate in water (Vendruscolo et al., 2003). On the other hand, conceptual demarcations between self-ordering and self-organization are disputed. ...
... Selfordering seems to be a common way of referring to molecular regularity. In some studies, the emergence of molecular order is associated with self-organization (Luisi, 2006), whereas this kind of usage is criticized by other authors (Abel & Trevors, 2006). Restricting the application of the concept of self-organization to the individual organization of a system appeals to the dynamics that I associated with regulative self-organization. ...
Thesis
Full-text available
In this thesis, I discuss the organism's self-organization from the perspective of relational ontology. I critically examine scientific and philosophical sources that appeal to the concept of self-organization. By doing this, I aim to carry out a thorough investigation into the underlying reasons of emergent order within the ontogeny of the organism. Moreover, I focus on the relation between universal dynamics of organization and the organization of living systems. I provide a historical review of the development of modern ideas related to self-organization. These ideas have been developed in relation to various research areas including thermodynamics, molecular biology, developmental biology, systems theory, and so on. In order to develop a systematic understanding of the concept, I propose a conceptual distinction between transitional self-organization and regulative self-organization. The former refers to the spontaneous emergence of order, whereas the latter refers to the self-maintaining characteristic of the living systems. I show the relation between these two types of organization within biological processes. I offer a critical analysis of various theories within the organizational approach. Several ideas and notions in these theories originate from the early studies in cybernetics. More recently, autopoiesis and the theory of biological autonomy asserted certain claims that were critical toward the ideas related to self-organization. I advocate a general theory of self-organization against these criticisms. I also examine the hierarchical nature of the organism's organization, as this is essential to understand regulative self-organization. I consider the reciprocal relation between bottom-up and top-down dynamics of organization as the basis of the organism's individuation. To prove this idea, I appeal to biological research on molecular self-assembly, pattern formation (including reaction-diffusion systems), and the self-organized characteristic of the immune system. Finally, I promote the idea of diachronic emergence by drawing support from biological self-organization. I discuss the ideas related to constraints, potentiality, and dynamic form in an attempt to reveal the emergent nature of the organism. To demonstrate the dynamicity of form, I examine research into biological oscillators. I draw the following conclusions: synchronic condition of the organism is irreducibly processual and relational, and this is the basis of the organism's potentiality for various organizational states.
... The presence of irreversible processes contributing to the decrease of system entropy would be balanced by a larger increase in the rest of the universe, i.e., ∆S U /∆t > 0. Therefore, from a global perspective, one would be forced to disagree with the definition of self-organization, unless one considers the system to perpetually interact with an environment able to supply energy and order, as pointed out by Heinz von Forster in the '60s [51]. A complementary perspective is that dissipative structures in non-living systems, such as flames or hurricanes, are not true organizational systems since inanimate units cannot organize, but just self-order, themselves [52]. ...
Preprint
Full-text available
When a large number of similar entities interact among each other and with their environment at a low scale, unexpected outcomes at higher spatio-temporal scales might spontaneously arise. This nontrivial phenomenon, known as emergence, characterizes a broad range of distinct complex systems -- from physical to biological and social ones -- and is often related to collective behavior. It is ubiquitous, from non-living entities such as oscillators that under specific conditions synchronize, to living ones, such as birds flocking or fish schooling. Despite the ample phenomenological evidence of the existence of systems' emergent properties, central theoretical questions to the study of emergence remain still unanswered, such as the lack of a widely accepted, rigorous definition of the phenomenon or the identification of the essential physical conditions that favour emergence. We offer here a general overview of the phenomenon of emergence and sketch current and future challenges on the topic. Our short review also serves as an introduction to the Theme Issue "Emergent phenomena in complex physical and socio-technical systems: from cells to societies", where we provide a synthesis of the contents tackled in the Issue and outline how they relate to these challenges, spanning from current advances in our understanding on the origin of life to the large-scale propagation of infectious diseases.
... The presence of irreversible processes contributing to the decrease of system entropy would be balanced by a larger increase in the rest of the universe, i.e. S U / t > 0. Therefore, from a global perspective, one would be forced to disagree with the definition of self-organization, unless one considers the system to perpetually interact with an environment able to supply energy and order, as pointed out by Heinz von Forster in the 1960s [51]. A complementary perspective is that dissipative structures in non-living systems, such as flames or hurricanes, are not true organizational systems since inanimate units cannot organize, but just self-order, themselves [52]. ...
Article
When a large number of similar entities interact among each other and with their environment at a low scale, unexpected outcomes at higher spatio-temporal scales might spontaneously arise. This non-trivial phenomenon, known as emergence, characterizes a broad range of distinct complex systems—from physical to biological and social—and is often related to collective behaviour. It is ubiquitous, from non-living entities such as oscillators that under specific conditions synchronize, to living ones, such as birds flocking or fish schooling. Despite the ample phenomenological evidence of the existence of systems’ emergent properties, central theoretical questions to the study of emergence remain unanswered, such as the lack of a widely accepted, rigorous definition of the phenomenon or the identification of the essential physical conditions that favour emergence. We offer here a general overview of the phenomenon of emergence and sketch current and future challenges on the topic. Our short review also serves as an introduction to the theme issue Emergent phenomena in complex physical and socio-technical systems: from cells to societies , where we provide a synthesis of the contents tackled in the issue and outline how they relate to these challenges, spanning from current advances in our understanding on the origin of life to the large-scale propagation of infectious diseases. This article is part of the theme issue ‘Emergent phenomena in complex physical and socio-technical systems: from cells to societies’.
... For growing DS, these mass-energy exchanges build organization by dissipating potentials in its immediate surroundings [19]. Due to these exchange processes, spontaneous self-organization [20] at relatively reducing disorder level can emerge in DS, thus producing higher ordered spatial macro DS [21]. ...
Article
Full-text available
Dissipative structures (DS) exist at all scales, systems, and at different levels of complexity. A thermodynamic theory integrating simple and complex DS is introduced, which addresses existence of growing/decaying DS based on their entropy analysis. Two entropy-based dimensionless ratios are introduced, which explain negentropy-debt payment and existence of DS with growth or decay. It is shown that excess negentropy debt payment is needed and beneficial for growing DS; but for decaying DS, it hastens its approach to perish and is counter-productive. Growing complex DS tend to pay lower negentropy debt to their surroundings, due to involvement in other activities enabled by complexity; e.g. mediation for survival that is linked to their mortality. Hence, disorder of complex DS increases, due to which, their growth can be un-sustained, leading to entry in decay-phase in spite of availability of adequate mass-energy in-flows. Proper handling or reduction of complexity enables growth in the direction of ideal growth (without increase in disorder of DS), which is limited only by availability of adequate mass-energy in-flows.
... How big or how old that system needs to be is still a matter for debate. Arguments by Abel and Trevors (2006) and Abel (2009) suggest that within the framework of Big-Bang type cosmologies naturalistic protogene formation still faces almost insuperable difficulties. However, by whatever process life has emerged, this event of origination must be reckoned as unique, and the subsequent spread of life throughout the universe ensured by the processes of "panspermia." ...
Chapter
Full-text available
A range of astronomical observations are shown to be in accord with the theory of cometary panspermia. This theory posits that comets harbor a viable biological component in the form of bacteria and viruses that led to origin and evolution of life on Earth. The data includes (1) infrared, visual and ultraviolet spectra of interstellar dust, (2) infrared spectra of the dust released from comet Halley in 1986, (3) infrared spectra of comet Hale-Bopp in 1997, (4) near and mid-infrared spectra of comet Tempel I in 2005, (5) the discovery of an amino acid and degradation products attributable to biology in the material recovered from the Stardust Mission in 2009, (6) jets from comet Lovejoy showing both a sugar and Ethyl alcohol and finally, (7) a diverse set of data that has emerged from the Rosetta mission. The conjunction of all the available data points to cometary biology and interstellar panspermia as being inevitable.
... Their order vanishes with the constraints, and they do not show operating life. 46 An operating organized system may involve intelligent agents, but not necessarily-as sometimes required for complex adaptive systems (CASs). For instance, expressions like the following are used for CASs: the systems are characterized "by nodes (agents) that make decisions based on past actions and on new input" 47 or in the cases of social science and ecosystems, "partly independent operators work together or in a network and give rise to an emergent complex behavior." ...
Article
Full-text available
In the framework of thermodynamics of irreversible processes, patterns of macroscopic evolution of operating organized systems from various fields (engineering, biology, cosmology) are coupled to the increase in their entropy. An extension of Boltzmann’s equation is proposed to characterize the entropy evolution. It is shown that such a “top-down” approach allows us to merge empirical data in a single inclusive model. A method is proposed to quantitatively assess the minimum semantic information gained during the life of the systems. This allows us to compare systems with different types of organization and lifespans. An example of calculation is given for the universe. The method also offers a challenging view to “bottom-up” approaches in progress. The article can be freely accessed following the link: https://doi.org/10.1063/1.5128315
... In particular, this section proposes some perspectives reasoning on the role of the subcellular scale namely the dynamics of genes. 1,33,72 In addition, it is shown that the application of the multiscale approach can be addressed to the study of various classes of systems in life sciences. ...
Article
Full-text available
This paper presents a review on the mathematical tools for the derivation of macroscopic models in biology from the underlying description at the scale of cells as it is delivered by a kinetic theory model. The survey is followed by an overview of research perspectives. The derivation is inspired by the Hilbert's method, known in classic kinetic theory, which is here applied to a broad class of kinetic equations modeling multicellular dynamics. The main difference between this class of equations with respect to the classical kinetic theory consists in the modeling of cell interactions which is developed by theoretical tools of stochastic game theory rather than by laws of classical mechanics. The survey is focused on the study of nonlinear diffusion and source terms.
Article
This essay aims to define the origin, expansion, and evolution of living matter. The first formations, identified as remains, fossils, traces etc. of life are almost as old as the Earth itself. During four billion years, life on the Earth has continuously existed and been implemented in the range of conditions, ensuring the dropping-liquid state of water. During the entire period of life existence, its evolution was going with the tendency of multidirectionality, after each catastrophe tending to the diversity and vastness of distribution, and all the currently living species, regardless of their complexity, have the same evolutionary age. The property of reproductive surplus (multiplication) is inherent in all the living matter. The reproduction of all the living matter is implemented via the “development” – a process of continuous occurrence of something new that did not exist in the previous moment in the reproduced individual at each specific moment of time with the tendency towards the reproduction of a “copy”. In its fundamental basis, Life is based on a programme, its material support is implemented and exists not in the field of causative-consecutive events, but in the field of programmed-causative-consecutive events. This predetermines the “biology laws”, the behaviour of the material constituent of Life at each time period, and the future of the material constituent of life.
Conference Paper
The damage of natural equilibrium nowadays requires urgent action. All global changes we face are the forms of self-reorganization process in nature. The long history of chemistry creates the means for overpowering nature for natural resources exploitation, for the sake of humankind. Chemistry is one of the key sciences to understand natural changes. Some probes in analytical chemistry can be the key information about the whole phenomena. Method development and validations in analytical chemistry must be in line with the need to preserve nature. This way technology must have good intension to assess its process and its risk management. Moreover, the power of big data can aid the need for chemists to examine earth condition as well as global common problems. In this paper, a rough survey about ethical issues regarding chemistry and its impact on the environment is reported. Ethics development in chemistry education needs to be put into action, from the very beginning of the knowledge transfer. An eco-reflexive chemistry education must be based also on chemist’s responsibility towards nature as a whole.
Chapter
In diesem Kapitel wird für die breite Leserschaft erörtert, ob und wie die Thermodynamik dazu beitragen kann, das Leben grundlegend zu beschreiben. Statt einzelne Merkmale des Lebens aufzuzählen, werden hier zunächst die Begriffe der Entropie, Emergenz und Komplexität allgemein verständlich herangezogen. Wesentliche Bestandteile dieses Kapitels sind auch die Unterscheidung zwischen der Selbstordnung und Selbstorganisation und die davon ausgehende Frage, ob Leben synthetisch nachempfunden werden kann. Die Leserschaft wird zuletzt mit einem naturalistischen Konzept konfrontiert, demzufolge es zwischen dem Leben und Nicht-Leben keine qualitative Unterscheidung geben muss, sondern in dem das Leben im Sinne einer Vereinheitlichung lediglich eine Teilklasse vieler chemisch-physikalischer Prozesse repräsentiert, zu der auch wir selbst gehören.
Article
Full-text available
Electrophoretic mobility shift analysis (EMSA) is a well-characterized and widely used technique for the analysis of protein-DNA interaction and the analysis of transcription factor combinatorics. Currently implemented EMSA generally involves the time-consuming use of radiolabeled DNA and polyacrylamide gel electrophoresis. We are studying the bionanoscience of self-assembling supramolecular protein-nucleic nanostructures. We have undertaken these studies because they promise to enhance our understanding of assemblies formed during prebiotic evolution, provide tools for analysis of biological processes like DNA recombination, and may lead to the development of nanoscale biosensors designed for site-specific molecular targeting. During the course of that work, we noted that EMSA of these complex structures could be effectively implemented with microfluidics chips designed for the separation of DNA fragments. In this report we compare the two techniques and demonstrate that the microfluidics system is also capable of resolving complex mixtures produced by decorating DNA recombination intermediates with mixtures of DNA binding proteins. Moreover, the microfluidics chip system improves EMSA by permitting analysis with smaller samples, avoiding the use of radiolabeling, and reducing the time involved to a matter of minutes.
Chapter
Nach der mechanischen Wärmetheorie gehorchen die thermischen Eigenschaften von Gasen und anderen Stoffen trotz der Tatsache, daß diese Stoffe aus einer großen Anzahl von Molekülen zusammengesetzt sind, die sich in schneller ungleichmäßiger Bewegung befinden, wohlbestimmten Gesetzen. Die Erklärung dieser Eigenschaften muß auf die Wahrscheinlichkeitsrechnung gegründet werden, und dazu muß man die Verteilungsfunktion kennen, die zu jedem Zeitpunkt die Anzahl von Molekülen in jedem Zustand bestimmt. Um diese Verteilungsfunktion f(x,t) = Anzahl der Moleküle, die zur Zeit t die Energie x haben, zu ermitteln, wird für f eine partielle Differentialgleichung hergeleitet, indem untersucht wird, wie sich f während eines kleinen Zeitintervalls infolge von Stößen zwischen den Molekülen ändert.
Book
As reports of genocide, terrorism, and political violence fill today's newscasts, more attention has been given to issues of human rights-but all too often the sound bites seem overly simplistic. Many Westerners presume that non-Western peoples yearn for democratic rights, while liberal values of toleration give way to xenophobia. This book shows that the identification of rights with contemporary liberal democracy is inaccurate and questions the assumptions of many politicians and scholars that rights are self-evident in all circumstances and will overcome any conflicts of thought or interest. Rethinking Rights offers a radical reconsideration of the origins, nature, and role of rights in public life, interweaving perspectives of leading scholars in history, political science, philosophy, and law to emphasize rights as a natural outgrowth of a social understanding of human nature and dignity. The authors argue that every person comes to consciousness in a historical and cultural milieu that must be taken into account in understanding human rights, and they describe the omnipresence of concrete, practical rights in their historical, political, and philosophical contexts. By rooting our understanding of rights in both history and the order of existence, they show that it is possible to understand rights as essential to our lives as social beings but also open to refinement within communities. An initial group of essays retraces the origins and historical development of rights in the West, assessing the influence of such thinkers as Locke, Burke, and the authors of the Declaration of Independence to clarify the experience of rights within the Western tradition. A second group addresses the need to rethink our understanding of the nature of existence if we are to understand rights and their place in any decent life, examining the ontological basis of rights, the influence of custom on rights, the social nature of the human person, and the importance of institutional rights. Steering a middle course between radical individualist and extreme egalitarian views, Rethinking Rights proposes a new philosophy of rights appropriate to today's world, showing that rights need to be rethought in a manner that brings them back into accord with human nature and experience so that they may again truly serve the human good. By engaging both the history of rights in the West and the multicultural challenge of rights in an international context, Rethinking Rights offers a provocative and coherent new argument to advance the field of rights studies. © 2009 by The Curators of the University of Missouri. All rights reserved.
Chapter
This book asserts that the Civil War marks the end of one era of American legal history, and the beginning of another. Abraham Lincoln's famous Gettysberg Address is viewed as the beginning of a new kind of “covert” constitutional law – one with a stronger emphasis on equality in the wake of the abolition of slavery – which was legally established in the Amendments made to the U.S. Constitution between 1865 and 1870. The author asserts that the influence of this “secret constitution”, which has varied in degree from Reconstruction to the present day, is visible in the rulings of the Supreme Court on issues hinging on personal freedom, equality, and discrimination.