From Evo Devo Universe
Jump to: navigation, search
“If life follows from [primordial] soup with causal dependability, the laws of nature encode a hidden subtext, a cosmic imperative, which tells them: “Make life!” And, through life, its by-products: mind, knowledge, understanding. It means that the laws of the universe have engineered their own comprehension. This is a breathtaking vision of nature, magnificent and uplifting in its majestic sweep. I hope it is correct. It would be wonderful if it were correct. ...if it is, it represents a shift in the scientific world-view as profound as that initiated by Copernicus and Darwin put together.” -- Paul Davies, The Fifth Miracle, 1999, Simon and Schuster, p 246.
I am suggesting that there may come a time when physics will be willing to learn from biology as biology has been willing to learn from physics, a time when physics will accept the endless diversity of nature as one of its central themes, just as biology has accepted the unity of the genetic coding apparatus as one of its central dogmas. -- Freeman Dyson, Infinite in All Directions, 1988, Harper, p 47.


Situation. The underlying paradigm for cosmology is theoretical physics. It has helped us understand much about universal space, time, energy, and matter, but does not presently connect strongly to the emergence of information, computation, life and mind. Fortunately, recent developments in physics, cosmology, theoretical biology, evolutionary developmental biology, information and computation theory, and the complexity sciences are providing complementary yet isolated ways to understand our universe within a ‘meta-Darwinian’ framework in which unpredictable and diversity-creating or "evolutionary" and predictable, convergent, and hierarchical, or "developmental" processes work together, via replication and under selection, to generate adapted information, order, and "intelligence" in a variety of physical systems at multiple scales. The rigor, relevancy, and limits of an "evolutionary developmental" approach to understanding universal complexity remains an open and understudied domain of scientific and philosophical inquiry.

Course of Action. These results and hypotheses deserve to be explored, criticized, and analyzed by an international interdisciplinary research and discussion community ‘Evo Devo Universe (EDU)’. Where merited, they may lead to expanded conceptual framework that improves our understanding of both contingent and convergent complexity emergence in the universe.

Benefits/Results. If validated, such a framework promises to advance our understanding of both perennially chaotic, contingent, creative, experimental, and unpredictable processes (evolutionary processes) and of constraining, convergent, hierarchical, cyclical, heritable, and predictable processes (developmental processes) in the universe as a system, and of evolutionary and developmental process at all scales, including the human scale. If falsified in any part, this endeavor will improve our critical thinking about the generation and regulation of complex systems, and the role and limits of organic analogies in understanding our universe.

Evolutionary Development (Evo Devo)

Biological systems engage in evolutionary (variational and selective) process, and they also develop (have heredity, reproduction, a predetermined, predictable and convergent series of hierarchical emergences within a life cycle). In living organisms, these two processes comprise the general model of "evolution". Yet evolutionary and developmental processes are actually quite different, even antagonistic in their roles in living systems. Evolutionary process is fundamentally creative and contingent, while developmental process is fundamentally conservative and convergent with respect to the generation and handling of biological information. Both processes are fundamentally necessary to the self-organizing phenomenon we call life, and they may also be fundamental to all complex replicating systems in the universe, including the universe itself, if it is found to be a replicating system within a metaverse or multiverse.

We propose therefore that the phrase "evolutionary development," or "evo devo," is a more useful description of organic change than the term "evolution," which has historically ignored or minimized the role of developmental process (organismic development, group development, environmental development) in guiding and constraining long-range evolutionary processes at all scales. The new generation of "evo-devo" (hyphenated) biologists understand this, as they stress the long-range effects of organismic developmental constraint (homeobox genes, body plans, developmental path dependencies, group culture, niche construction) on evolutionary process.

Those who use the term "evo devo" in its unhyphenated form, as we do here, are scholars and theorists who suspect that both evolutionary (evo) and developmental (devo) process are likely to be equally fundamental to understanding change in a broad range of complex systems, including non-living systems, and over all scales, including the universal scale.

While evolutionary process (“evo”) is creative, contingent, and famously unpredictable, many aspects of development (“devo”) are conservative, convergent, and quite predictable, with the right theoretical or empirical aids. For example, if you have a sense of what stage a developing system (such as a cycling star) is at in its ‘replication’ cycle (birth, growth, reproduction, aging, or recycling), you have a strong basis to expect what stage is coming next, as long as you have a sufficiently advanced understanding of stellar physics to discover the cycle, or have made sufficient astronomical observations to infer the morphology of cycle stages. In practice, identifying any evolutionary process and predicting any developmental process requires a mix of observation, analogy, and theory.

Organic molecules show evolutionary (variational, selectionist) process as they develop morphological and informational complexity, and in some cases replicate, in ways predictable across a bounded set of environments. Our life-generating solar system appears to have emerged via a process of stellar "evolutionary development" (reproduction of progressively more chemically complex solar systems) in our galaxy. Ideas or ‘memes,’ which replicate between human brains, also appear to engage in both evolutionary processes (emerging in unique and contingent ways at multiple scales) and developmental ones (converging on culturally universal features and constraints, such as language, logic, and morality). So apparently do technologies, which today replicate and undergo adaptive selection as obligate dependents on human cultures. Yet just as we have theories for the growing autonomy (biogenesis, intelligence) of certain obligate chemical replicators on the early Earth, so too we have early ideas that certain of our technological replicators (for example, deep learning systems) are today on a path to their own autonomy.

Since the 1980's, several scholars have even proposed that our universe itself may replicate, evolve, and develop, which if true would place all universal creative evolutionary process, prebiological, biological, and postbiological, under a set of universal developmental constraints. We have known for 150 years that our universe is dying (second law of thermodynamics). We've known for 80 years that it has had a finite time of origin (Big Bang). We've known since 1999 that our universe appears to be fractionating into informational islands at an accelerating rate, via dark energy. What other developmental structural and process constraints may future physics reveal?

The idea that some of the fundamental constants of our universe (and therefore at least some of its physical laws), have been fine-tuned (or perhaps, in an evo devo model, were self-organized to become fine-tuned over multiple universe cycles) to have the precise values that make life and intelligence emerge as a robust and developmental process would be one potential set of constraints. Many scientists see the only available explanation for fine tuning as the anthropic principle, whereas others, including Lee Smolin (Scientific alternatives to the anthropic principle, 2004), contend that it cannot presently yield any falsifiable predictions. If not, it would have to remain in the realm of philosophy rather than science, at least in our present state of cosmology and astrophysics. String theorists such as Leonard Susskind (2004) have proposed that the fine-tuned value of the cosmological constant, and possibly other fundamental constants, can be explained by anthropic approaches to string theory, using a string theory "fitness landscape."

Today, a growing number of interdisciplinary researchers are striving to articulate what might be called a 'meta-Darwinian' theory of universal change. Such a theory must predict certain systemic aspects of complexity's hierarchical emergence as developmental, or statistically probable, arising from the unique parameters (laws, constants, conditions) of our particular universe and the environment in which it is embedded (dark matter, multiverse, etc.). It must relate those predictable processes to the primarily stochastic and diversity-creating mechanisms of complexity emergence in nonliving and living systems at multiple scales, and reconcile both processes with Darwinian models of selection and adaptation in living systems, and of the origin of life under nonliving molecular evolution, development, and selection. A number of nonbiological processes, such as snowflake formation, and biological ones, such as brain emergence, have heritable features that are unpredictable, and diversity and information-generating (e.g., 'evolutionary') and heritable features that are predictable, directional, hierarchical, cyclic, and information-conservative over successive replications (e.g., 'developmental').

Modern evolutionary theory is more useful and validated than ever, yet with its gene-reductionist and individual-selectionist emphasis it is very unlikely to be a complete description of dynamics in biological systems. A growing community of evo-devo, ecological, and theoretical biologists propose that natural selection, while dominant, is not the only long-range influence on biological change. Developmental process and structure (biological development, psychological development, niche/social development, and universal development) and the unique physical features of our universal environment (its fundamental parameters and laws) are among a handful of 'non-adaptive' processes that are likely to constrain and direct long-range evolutionary change in living systems in many still-poorly-understood ways (see Ho and Saunders 1984; Müller and Newman 2003; Reid 2007 for a variety of different approaches to this crucial and unresolved question).

Consider the following insight from evo-devo biology: two genetically identical twins are always unpredictably unique in their stochastically-determined dynamics and structure (organogenesis, fingerprints, retinal patterns, neural connectivity, etc.) yet they are also always predictably highly convergent over a broad range of far-downstream systemic attributes (their developmental milestones, gross physical appearance, key psychological attributes, lifespan, etc.). Identical twins allow us to see these two genetic processes, one divergent (creating evolutionary variety) and one convergent (enforcing developmental similarity), interacting in tension with each other, both in service to the organism's ability to adapt in the environment. Likewise, under a variety of current cosmological models, two initially parametrically identical universes would each exhibit unpredictably unique and creative internal evolutionary variation over their lifespan, and at the same time, a broad set of predictable developmental differentiation milestones and shared structure and function between them. Both of these systems have evolutionary and developmental dynamics (unpredictable and predictable emergence processes) that cause a specific globally homologous but locally unique type of complexity to arise within each of them. In what other ways might they be alike? Do both systems replicate, for example, and would such replication (and presumed selection, in some environment) offer a parsimonous explanation for the kind of order that each has accumulated?

We know our universe has deep "developmental" isotropy and predictability in many of its emergent forms and functions. A special subset of nonlinear universal processes, including entropy increase, complexity and information growth, hierarchy emergence, replication, "convergent evolution" in living systems, and apparent accelerating cosmic and technological change, are presumably directed by statistically predictable, information-conservative, and future-convergent or "developmental" dynamics. At the same time, the large majority of nonlinear universal processes are modeled by complexity scientists to exhibit unpredictable, stochastic, divergent, diversity-generating, or "evolutionary" dynamics. For example, most astrobiologists suspect that life, information, and intelligence must have unpredictably different diversity in each of its spatially and temporally isolated environments in our universe. Some astrobiologists expect that life, even complex life, must also be fecund in our universe. Fortunately, we can expect astrophysical and SETI evidence for or against fecundity hypotheses to emerge in coming decades.

Such observations suggest many questions for further research. Among them: Is there some informational, computational, or adaptive advantage to universes that predictably create a tremendous fecundity of spatially-isolated and unpredictably diverse life, assuming, as most theorists do, that all such life must be computationally finite and thus limited in its intelligence? Can we better model the utility of a "predictably fecund and unpredictably diverse" type of complexity for replicating living systems across some range of selective environments? Can we generalize such a model to understand a role for information and intelligence in our universe as a complex system, in a range of extrauniversal environments? Or is this analogy partly or entirely flawed? How can we better test and falsify it?

Fortunately, recent publications in physics, cosmology, evolutionary developmental ('evo-devo') biology, astrobiology, information and computation, and philosophy of science have provided promising new avenues of research for meta-Darwinian investigations. The scientific need to organize, make accessible, and critique the literature, evidence, models and arguments of those proposing such a reconciliation is great. So also is the need to identify research topics in multiscale evolutionary developmental inquiry, and to promote productive collaboration among investigators and students in those research topic areas.

A Research Agenda in Multi-Domain, Multi-Level, Multi-Scale Variation, Development, and Selection

We can today find tentative evidence and hypotheses for both evolutionary and developmental process in complex systems in multiple academic domains and systems levels, and at all universal scales, including the development of spacetime to create our quantum-relativistic universe (Jan Ambjørn, Renate Loll and Causal Dynamical Triangulation), in the selectionist emergence of classical from quantum physics (Wojciech Zurek and Quantum Darwinism) in Darwinian approaches to thermodynamics (Ping Ao and Darwinian thermodynamics). The cosmological theory of inflation--originally designed to explain the flatness of the universe, the thermodynamic equilibrium of distant horizons, and other key observable features--has given rise to the theory of [eternal inflation (Alan Guth, Paul Steinhardt, Alexander Vilenkin, Andrei Linde) which naturally produces a multiverse structure of different universes, where fundamental physical constants vary across the multiverse from one universe to the next. Both the mathematical universe hypothesis (Max Tegmark) and the (universe as a) simulation hypothesis (Nick Bostrom) also can be considered from evolutionary (unpredictable) and developmental (predictable) perspectives, in an effort to further validate or invalidate such approaches.

We also find evo and devo process models in comparing the locally diverse and globally predictable aspects of stellar nucleosynthesis (Donald D. Clayton, George Wallerstein and others), in the themodynamically and kinetically constrained emergence of globally predetermined but locally unique chemotypes in inorganic and organic chemistry (J.J.R. Frausto da Silva, Robert J.P. Williams), and in chemical and mathematical models of RNA complexification and cellular biogenesis (D. Eric Smith, Eörs Szathmáry and others). In the biological sciences, innovative scholars in genetic and cellular self-organization (Stewart A. Newman and dynamical patterning modules (DPMs), Stuart Kauffman in self-organizing autocatalytic sets, Eric Davidson and gene-regulatory networks) ), in evo-devo and theoretical biology (Werner Callebaut, Sean B. Carroll, Christian de Duve, Manfred Laubichler, Gerd B. Müller, Massimo Pigliucci, Rupert Riedl, Richard Reid, Stanley Salthe, Günter P. Wagner and others), in epigenetics (Eva Jablonka and others), in niche construction and stigmergy (Benjamin Kerr, Kevin N. Laland, John Odling-Smee) in evolutionary convergence as a limitation on biological form (George McGhee, Simon Conway Morris and others), in evolutionary transitions (Eörs Szathmary), in the thermodynamics of progressive evolution (Alexander I. Zotin), and evolutionary escalation (Geerat Vermeij) are attempting to construct an "extended evolutionary synthesis", one that sees the gene- and individual-centric process of natural selection as a major factor but not the only process of macroscale biological change. In this view, Darwinian evolution must operate within a framework of developmental and environmental constraints that will themselves provide some statistically-observed long-range directionality, and constrain and dictate certain types of evolutionarily convergent morphological and computational process, structure, and function.

Variations of this new synthesis, involving an interplay between mechanisms of unpredictable variation and predictable accumulated constraints, are being attempted in evolutionary and developmental bioenergetics (Alexander I. Zotin), in brain development (Gerald Edelman and Neural Darwinism), in cognitive selectionism and neural synchronization in perception, attention and consciousness (Gyorgy Buzsaki, William Calvin, JA Scott Kelso, John McCrone and others), in evolutionary psychology and the evolution of cognitive and moral preferences (Richard Alexander, Jerome Barkow, Samuel Bowles, Leda Cosmides, Michael Ghiselin, Herbert Gintis, John Tooby and others), in evolutionary epistemology, or the autopoetic production and organization of social knowledge (Ross Ashby, Donald T. Campbell, William P. Hall, Humberto Maturana, Ikujiro Nonaka, Karl Popper, Francisco Varela, Heinz von Foerster and others), in neuroeconomics and evolutionary economics (Kurt Dopfer, Ernst Fehr, Geoffrey M. Hodgson, Elinor Ostrom), in evolutionary anthropology (Robert Boyd, Terrence Deacon), in evolutionary approaches to human history (Dan Smail), in the evolution and development of social and religious beliefs (Donald T. Campbell, Elliott Sober, David Sloan Wilson), in general theories of 'memetic' selection occurring in human brains (Robert Aunger, Susan Blackmore, Richard Dawkins, Daniel Dennett and others), in evolutionary computation and 'artificial life' in computers (Chris Adami, Mark Bedau, John Holland, John Koza, Chris Langton, and others), in the development of ecological and human cultural hierarchy and scale (John Bonner, Geoffrey West, and others), in flow maximization in living and technological systems (Adrian Bejan and constructal theory), and in general technological evolutionary development (Brian Arthur, David Brin, Kevin Kelly, Ray Kurzweil and others). In each case, tentative models of both evolutionary and developmental process and their interactions may be usefully defined and discussed.

There are also a number of hypotheses for potentially 'universal' developmental architecture and process in complex systems in our universe, including scale invariance and scale relativity (Laurent Nottale, Jean Chaline and others), criticality (Didier Sornette) and self-organized criticality (Per Bak, Kim Sneppen), self-organization under far-from equilibrium dynamics (Stuart Kauffman, Ilya Prigogine), self-similar hierarchy (Robert L. Oldershaw, Peter Winiwarter), free energy rate density and cosmic expansion (Eric Chaisson), complexity emergence acceleration either as an exponential, log-periodic, hyperbolic, or scale-invariant phenomenon over macroscales of space and time (Henry Adams, Francois Meyer, Carl Sagan, Laurent Nottale, Jean Chaline, Pierre Grou, Philip Tobias, Anders Johansen, Didier Sornette, Richard Aunger), or as a continuous or bounded long-wave logistic escalation (Derek De Solla Price, Richard Coren, Theodore Modis, Tessaleno Devezas), entropy production and infodynamics (Roderick Dewar, Stanley Salthe, and others), fine-tuning of fundamental physical parameters as a developmental approach to cosmology (John D. Barrow, Paul Davies, Lawrence M. Krauss, Martin J. Rees, and others). Speculative models in the evolutionary development of the universe as a system include Cosmological Natural Selection (CNS) (Lee Smolin, Andy Gardner), and hypotheses of 'CNS with Intelligence' (CNS-I)) (James N. Gardner , John Smart, Clement Vidal).

Many mathematical and process models of long-proven use in biological systems, such as logistic growth curves, normal and power laws, representative maps, connectionist and hierarchical networks, reaction-diffusion systems in chemical and embryonic development (Alfred Gierer, Hans Meinhardt), and thermodynamic models of ecological dynamics which involve not only bottom-up (evolutionary) but also top-down (developmental) control, such as developmental ascendancy (James A. Coffman, Robert Ulanowicz) and panarchy (Lance H. Gunderson, Crawford S. Holling), and the interaction of cybernetic systems and evolution to produce self-organization (Terrence Deacon, Carlos Gershenson, Francis Heylighen, Stuart Kauffman and others) are beginning to be applied more rigorously to prebiological, biological, cultural, and technological systems, and deserve much wider analysis and attention.

We have seen similar theoretical advances in the previous century. When classical physics, as powerful as it is, came to be understood as a subset, yet also the most accessible and broadly used component, of general relativity, we gained a theoretical framework with significantly greater long-range explanatory and predictive value. The same enlargement of our understanding occurred when the immensely practical quantum theory emerged, and led us to better models in fundamental physics, cosmology, and our current efforts at a theory of quantum gravity. Should evo-devo and theoretical life scientists point the way to a new, meta-Darwinian synthesis in biological systems, with (intrinsically unpredictable) macroevolutionary and (statistically predictable) macrodevelopmental processes both operating simultaneously over nearly four billion years of life on Earth, we may also gain insights into the nature and extent of quasi-evolutionary and quasi-developmental process at the universal scale.

The Universal Roles of Information, Computation, and Intelligence

Hypotheses of universal evolutionary development must also seek to understand the role and emergence of information, computation, and intelligence within the universe. In addition to physics, advances in mathematics and nonlinear science, information theory, complexity (algorithmic, computational, informational, and physical), chemistry, biology, neuroscience, computer science, and machine learning all seem likely to play a role in our search more useful and generic theories of information and order in universal systems.

Physics has provided several advances in this regard in recent decades. We now have advanced theories of quantum information (John Preskill, Scott Aaronson) and models like quantum Bayesianism (Chris Fuchs) that attempt to derive quantum mechanics from purely informational approaches. The model of entropic gravity (Erik Verlinde) proposes gravity as emergent from information, deriving relativity and classical mechanics from statistical thermodynamics. Research on the black hole information paradox, exploring what happens to information as it enters black holes, has also been productive. Most physicists are now convinced that black holes do not destroy the information of the matter-energy that enters them (Thomas Banks, Leonard Susskind, Michael Peskin 1984). This work has indirectly led some scholars to the holographic principle, the proposition that the information contained in any region of space is proportionate to its surface area rather than its volume, and in the case of a black hole can be thought of as existing entirely on the surface (Gerard 't Hooft 1993, Leonard Susskind 1994). Some physicists think the holographic principle will lead us to a view of "the physical world as information, with energy and matter as incidentals" (Jacob Bekenstein 2003). The holographic principle has led to non-perturbative models of quantum gravity such as BFSS matrix theory (Banks, Fischler, Susskind, Shenker 1996) (a type of non-commutative geometry), and AdS/CFT (Maldacena 1998), both current contenders in the search for the ultimate foundation for string theory known as M-theory. Perhaps, at the end of this process, we may even come to understand our universe as a computational entity, a contention that has been championed in various forms by an impressive array of physicists in recent generations (Konrad Zuse, John A. Wheeler, David Deutch, Gregory Chaitin, Ed Fredkin, Stephen Wolfram, Seth Lloyd). A holographic derivation of entropy in AdS/CFT discovered by Ryu and Takayanagi has led to the conjecture that spacetime itself may emerge from quantum entanglement (Mark Van Raamsdonk, 2010). The ER=EPR conjecture takes this idea a step further and proposed that wormholes and quantum entanglement may refer to the same phenomenon (Susskind, Maldacena, 2013). All of this suggests there may be a deep connection between quantum information theory and gravity, which operates at multiple scales and ties together such diverse areas as particle physics, condensed matter physics, fluid dynamics, and chaos theory.

As another facet of this search, we may also learn to reconcile the two most fundamental approaches to modeling human cognition, connectionism and computationalism. Connectionist (neural network, parallel and distributed, associative) processes (James L. McClelland, David E. Rumelhart) seems strongly evolutionary, selectionist, and adaptive, yet may also be coupled to developmental process to create 'biologically inspired' evolutionary and developmental systems. For example, we may use genetic programming techniques ('description languages') to control the expression/development of (presently primitive) artificial neural networks (Daniel Rivero, Julián Dorado), and allow environmental selection on the informational genotype of the expressed computational phenotype, as in living systems. By contrast, computationalist (symbolic/logical/Bayesian) informational and cognitive processes (Alan Turing, Jerry Fodor) seem likely to be a specially restricted and emergent subset of the most socially complex and imitative connectionist systems (e.g., humans). Learning how to unify these two approaches, as in the neural-symbolic integration agenda (Barbara Hammer, Pascal Hitzler, Marco Gori) seems a necessary goal of future theory.

Of particular interest to our near-term cultural future are questions in technology growth and evolution. Many classes of macroscale technology, such as transportation or steel production, demonstrate a pattern of logistic growth and punctuated equilibria in their evolutionary development, with long periods of relative stasis. In contrast, as Ray Kurzweil, Chris Magee, Gordon Moore, Hans Moravec, Bela Nagy, William D. Nordhaus and others have observed, a select subset of information, computation, communication, and physical nanotechnologies (energy, etc.) have shown exponential or superexponential capacity development over long spans of cultural time. These long-term technology performance curves represent an important and understudied area of research. Chris Magee has published a number of academic works on these, and Bela Nagy has developed a Performance Curve Database at the Santa Fe Institute to collect technology performance metric data across the technology domain.

As Ray Kurzweil has shown, digital computers have doubled their price-performance ratios every one-and-a-half to three years, since at least the 1890's and particularly rapidly since the 1930's, while simultaneously migrating across mechanical, electromechanical, vacuum tube, transistor, and integrated circuit manufacturing paradigms. Are some hidden laws at work here? Are Moore's law and a collection of related exponential technology capacity growth laws part of a universal developmental process arising from some interaction between the physical structure of our universe and the value of computation at the 'leading edge' of evolutionary developmental adaptation?

Or are these historical consistencies, as some have argued, merely random consequences of some law of large numbers? In a complex universe, a few physical processes must occasionally show powerful growth, for astronomically insignificant periods of time (eg., early-universe inflation, supernovas). Is our recent history of stunning informational, computational, and nanotechnological acceleration here on Earth part of a critical, central process of cosmic evolutionary development, or is it an isolated statistical outlier, with little impact on future universal dynamics? Scholars are beginning to address such humbling yet fundamentally important questions.

In 1956, John McCarthy, Marvin Minsky, Nathan Rochester, and Claude Shannon convened a small and very interdisciplinary group to "proceed on the basis of a conjecture" that computers can be designed to simulate biological learning and intelligence. Their Dartmouth Summer Research Conference on Artificial Intelligence is considered the beginning of the "artificial intelligence" research community. In 1987, Christopher Langton convened the first interdisciplinary conference on "artificial life", a research community that proceeds on the basis of the conjecture that useful computer simulations of living systems, and perhaps even autonomous intelligent systems themselves, may be developed by biologically-inspired computing strategies. In the same vein, the EDU community convened in 2008 to proceed on the basis of a conjecture that analogs to the biological processes of evolutionary variation and experiment, developmental convergence and hierarchy, and adaptive replication and selection may also help explain universal complexity emergence and dynamics.

In these investigations, there is no assumption that our universe is “alive” or "organic" in any sense other than the obvious one, that it is capable of producing living systems, which themselves are largely a product of a recursive process of evolution, development, and selection. There are likely to be sharp limits to the generalizability of evolutionary and developmental processes of change across a broad range of complex systems. The differences between such systems must surely greatly exceed their similarities. We are also likely to be limited by our own ability to model and represent complex systems in current nonlinear science and information theory.

Nevertheless, the metaphors of evolution, development, and selection may take us to useful new places today, and help us uncover promising new avenues of research in many of these disciplines, as they each work toward better theories of information, computation, and intelligence from their own unique frames of understanding.

Major Transitions in Evolutionary Development

Theoretical biologists Smith and Szathmáry (1995) have proposed eight major transitions of evolutionary developmental emergence of biological complexity in universal history to date:

1. Replicating molecules to compartmentalized replicator populations (microspheres, etc.)
2. Independent nucleic acid replicators to chromosomes (linked replicators)
3. RNA as replicator template and enzymes to DNA as template and protein as enzymes
4. Prokaryotes to eukaryotes (with captured mitochondria and chloroplasts)
5. Asexual eukaryotic clones to sexual populations
6. Single-celled sexual protists to multicellular organisms
7. Solitary-except-for-mating multicellular individuals to societies
8. Primate societies to human societies with complex oral language

Each of these may be considered a metasystem transition (Turchin 1977; Heylighen and Campbell 1996), the evolutionary emergence of a more complex and global (by some metrics) level of organization and control, and a system with greater internal diversity, both structural and behavioral. Promising mathematical and complexity models for several of these transitions also exist (Fontana and Schuster 1998; Nowak 2006). These and other transitions may form a specification hierarchy (Salthe 1993), may be quantitatively periodized (Nottale et. al. 2000; Aunger 2007), and have been proposed to form an evolutionary developmental arrow of complexity (Bedau 1998; Stewart 2000) for our universe.

Strictly speaking, transitions 1 and 2 above are prebiological (chemical evo devo). Other prebiological evolutionary developmental transitions may include:

1. Replicating black holes generating 'fecund' complex universes in cosmology (Smolin 2004)
2. Evolving space-time generating the quantum-relativistic world in cosmology (Ambjørn 2006)
3. Quantum einselectionism leading to emergence of the classical universe (Zurek 2003)
4. Replicating stellar nucleosynthesis generating complex pre-organic elements (Clayton 1968)
5. Interstellar and planetary environments generating replicating organic molecules (Ehrenfreund and Spaans 2007)

Smith and Szathmáry's transition 8, to oral/written language, may be partly 'postbiological' (cultural evo devo). Scholars have proposed additional postbiological transitions, such as:

14. 'Memes' (replicating ideas and behavioral algorithms, in brains) generating 'temes' (replicating technological algorithms, in machines) (Blackmore 1999).
15. Biologically-inspired machines leading to a technological singularity (autonomous machine intelligence) (Vinge 1983, Kurzweil 1999, Clark 2003)
16. Postsingularity human-machine symbiosis leading to a global superorganism (Miller 1978, Stock 1993, Bloom 2000, Lifton 2004)
17. Postbiological intelligence eventually replicating its universe (Crane 1994, Harrison 1995, Gardner 2000)

Research Questions

  • Which of these (and other) transitions seem plausible, from a theoretical perspective? Which are mere conjecture?
  • Which may we hypothesize from a universal perspective, as ‘evolutionary’ processes that generate ‘developmental’ emergence?
  • Which may be tentatively modeled, and which may be presently falsified by observation in universal history?
  • In cases where the sample is presently only one (biogenesis) or where which the proposed transition is still very early in development (autonomous technology) what other methods may be used to test or falsify the hypothesis?
  • How do our tentative models of evolutionary and developmental process differ, and what features do they share in common across these transitions?
  • In which transitions can we identify evolutionary processes of natural selection, adaptation, creativity, divergence, contingency and/or unpredictability?
  • In which transitions can we identify developmental processes that are directional, convergent, conservative, cyclic, and/or statistically predictable?
  • How do cosmology, astrophysics and astrobiology, evo-devo and theoretical biology, and the complexity and information sciences inform our thinking about evolutionary and developmental processes across these transitions and in the universe as a system?

Please see Themes and Questions for a more extensive list of research themes and questions being considered by the EDU community.

EDU Community and Conferences

Promoting interdisciplinary research into the arguments, evidence, and models for evolutionary and developmental processes that may operate in our universe is the unifying theme of the EDU Community. We seek to ground this inquiry in partnership with scientific disciplinarians from each of the proposed transitions listed above. Research topics are chosen roughly proportionately across five disciplinary academic domains: physics, chemistry, biology, society, and technology. In addition, we seek to emphasize three primary interdisciplinary approaches: complexity research, philosophy, and systems theory. Please see EDU SIGs for more on the Special Interest Groups we are seeking to develop in our community.

A subset of our research themes are selected for presentation and collaboration at each EDU conference or workshop. Conference presentations are recorded and posted to a scholars discussion group, EDU Talk, along with presenters slides and papers. Presented papers are subject to open peer commentary and selected peer-reviewed papers are published in an academic journal.

Universal Evolutionary Development ('Devology') vs. Universal Darwinism ('Universal Evolution')

Investigations in universal evolutionary development, also called devology by physicist Georgi Georgiev, include hypotheses of universal selection, also called universal Darwinism, but go substantially beyond variation and natural selection to seek evidence for directional, hierarchical, and perhaps cyclical processes of universal development that may coexist with and constrain evolutionary process, just as development constrains variation and selection in biological systems.

Such questions must be asked if we suspect our universe is not only evolving but also developing, perhaps even in the manner that a biological system develops from seed, to organism, to reproduction, aging, and ultimately death/recycling. Early 'grand synthetic' models of this type were pioneered by Charles Pierce, Pierre Teilhard de Chardin, Thomas Huxley, Vladimir Vernadsky, Ludwig Von Bertalanffy, Alfred North Whitehead, and other twentieth-century scientists and philosophers. Ervin Laszlo, Ken Wilber, and others are today notable successors to these theorists, yet many of our synthetic models remain less rigorous, testable, and quantified than modern science demands.

Such models are often called hypotheses of 'universal evolution'. This is a misnomer. They are actually hypotheses of devology, or universal evolutionary development, as they are efforts to describe both evolutionary and developmental process at the universal scale. Though the distinction between these two processes was and still is rarely made sufficiently clearly or rigorously, both terms - evolution and development - are used extensively by these theorists and their successors to describe universal change.

Models of universal Darwinism do not, as a rule, admit to any inherent melioristic or orthogenetic (improving, complexifying) trajectory to processes of universal change, other than the traditional features of evolutionary process in biology (increasing variation, increasing specialization, etc.). As such they fail to address such apparently developmental features as the accelerating emergence of hierarchy in the universal, biological, and human-historical record, the accelerating adaptability and computational complexity at the leading edge (reproductive or 'germline') of local complex systems over cosmic time, and the increasing senescence and entropy of the general ('somatic') universe at the same time. Models of universal evolutionary development can and do address such features.

As in evo-devo biology, modern hypotheses of universal evolutionary development must attempt to investigate and describe the broad evidence of multi-level selection operating not just in biological systems (e.g., genetic, kin, sexual, and cultural selection in human beings), but also in pre- and postbiological systems. But at the same time, they must seek to describe the many ways evolutionary processes are likely to be constrained and conserved by an equally fundamental set of processes of universal development, just as evolutionary processes appear to be constrained and conserved by developmental architecture and process in all biological systems.

In conclusion, evo devo universe investigations are an attempt at describing the state of hypotheses in universal evolutionary development, as a tentative 'next step' beyond the useful but incomplete hypothesis of universal Darwinism toward a meta-Darwinian theory of universal change. Thank you for adding your insight and critique to our community.

References (Partial List)