Review - From Complexity to Life
On the Emergence of Life and Meaning
by Niels Henrik Gregersen (Editor)
Oxford University Press, 2002
Review by Rainer Kamber, Lic.Phil.
Jan 21st 2004 (Volume 8, Issue 4)

This volume is advertised on the back cover as an "excellent introduction to the sciences of complexity". In fact, only two of the ten contributions are concerned with the concept of complexity as such, and both the cursory paper by Charles H. Bennet (on empirical and abstract correlates of complexity and their formalization) and the more technical one by Gregory Chaitin (that relates the concept of complexity to his algorithmic theory of information) have been published elsewhere much earlier. What about "the sciences of complexity"? If we take the ten authors and their respective subject matters as a measure for the scope of such a field then it includes at least physical cosmology, mathematics, information theory, biochemistry and biophysics, and theology. Apart from theology this seems to be a reasonable list to start from. But then this book is not introductory in character but represents a collection of selected topics in these "sciences of complexity" and it is clearly aimed at an academic audience. Most of the contributions are rather technical and the selection of topics and authors has likely been based less on didactic and more on doctrinal concerns. The volume basically incorporates the attempt to link recent physical cosmology and speculations in theoretical biology with an argument for "creative divinity" in the history of the universe. After reading the final contribution by theologian and editor N. H. Gregersen it occurred to this reviewer that a more accurate title for the book could have been "From Design Theism to Complexity Deism" since at least five of the authors seem to embrace such positions. This point is furthered by the fact that the symposium that spawned the volume in 1999 was one of a series within a program called "The Humble Approach Initiative", organized and financed by the John Templeton Foundation. It is part of the mission of the Templeton Foundation (that sponsors the one-million-dollar "Templeton Prize For Progress Toward Research or Discoveries about Spiritual Realities") "to pursue new insights at the boundary between theology and science".

The volume (243 p.) is divided into three sections, the first containing the two papers on the notion of complexity. The five articles in section two are concerned with matters in theoretical biology and physical cosmology, and three articles in the final section deal with theological and philosophical questions such as the problem of reductionism, the notion of emergence, and recent shifts in theological cosmology. Below is an attempt to overview the main subject matters tackled in the book.

The main topics that are to ground this particular quest for the discovery of spiritual realities originate in physical cosmology and theoretical biology. One of the implications of the second law of thermodynamics is that the "order" of all physical systems cannot increase over time. Order is a relative notion and a given state will generally be considered more ordered if its description requires more information than that of a state considered less ordered. In its statistical definition (Boltzmann 1872 and 1877), the "entropy" of the macrostate of a physical system is proportional to the logarithm of the number of all its compatible microstates. As a consequence, it is a statistical fact that, for any physical system, its state at time t will be more ordered than its state at t+n. Even before the notion of entropy was introduced (Clausius 1865) reflection on thermodynamics had led Thomson (Lord Kelvin 1852) to cosmological discussions, preparing the way for Helmholtz (1854) who came to the conclusion that the universe is determined to end in "heat death", that is, in a state of thermodynamic equilibrium after the total diffusion of thermal energy it contains. This view created stirs not only among ecclesiastic circles but also among philosophers and scientists regarding an accurate interpretation of thermodynamics. But, as Paul Davies puts it in his contribution, the "significance [of the second law] for both theology and human destiny has been overstated." (p. 76) Why, the big bang and the following expansion of the universe actually brought it further from its possible maximum entropy state, mainly because of the newly generated energy gradient between matter and radiation and because of the forces of gravity starting to develop their pull on everything, but also because of the helium-hydrogen (that is, the neutron-proton) gap created in the early universe. In short, the history of the universe "is one of entropy rising but chasing a moving target, because the expanding universe is raising its maximum entropy at the same time." (p. 82)

On the one hand it is still highly likely that the universe will obey natural laws and either be obliterated in the "big crunch" or gradually reach its maximum entropy state in its expansion and the burnout of black holes. On the other hand it will go on to differentiate for a very long time. Davies finds it conceivable (with Dyson 1979) that its further development widens the gap between it and its theoretical state of maximum entropy and that the history of the universe may thus even be infinite (p. 84). This image of an indefinitely developing universe intuitively seems to give rise to the question of a "fourth law of thermodynamics" to account for this kind of development. Notably, such a law may still lack an explanandum because physical theory has so far been able enough to explain the actual cosmological development that Davies discusses. This seems to be one of the conclusions of the article by Ian Stewart who deepens the account of the "gravitic" universe. Stewart prefers to give the speculative fourth law the form of a second law of gravitation, a law that could account for the kind of rising complexity incorporated in the increasing clumpiness of the material universe. He suggests that there are quite different entropy "gaps" between the early and the contemporary universe and that the notion of such gaps, and thus the explanandum for a fourth law, may be an epistemic artifact stemming from the "coarse graining" underlying the modeling assumptions that statistically distribute the states of bodies in a thermodynamical system: "The increase of entropy in classical thermodynamics [...] is a statistical law. [The] negentropic nature of gravity is also statistical, for similar reasons. There are clear physical differences, however: elastic collisions are short range and repulsive, whereas gravity is long-range and attractive. With such different force-laws, it is hardly surprising that the generic behavior should be so different" (p. 138). A "fourth" law would seem to be, then, a "context-dependent" law since its scope could only cover the macrostructure of the universe and its development (p. 140). It need not even be a physical law in any orthodox sense since the existing laws of physics seem to suffice for explaining clumpiness in the universe.

In contradistinction, there is a real scientific explanandum regarding the initial state of the universe. Cosmology simply postulates an "initial" state of the universe such as it is observable today. The explanandum stems from the proposition that the initial state of the universe was in a low, but not in a maximum entropy state. Davies specifies the nature of this improbable state: The initial state "was one on which the matter and radiation were close to thermodynamic equilibrium but the gravitational field was very far from equilibrium. The universe started, so to speak, with its gravitational clock wound up but the rest in an unwound state" (p. 81-2). Quantum cosmology is able to explain the history of the universe from such an initial state but it obviously leaves the improbable initial state as such unexplained. Davies and Stewart, however, discuss no explanatory connection between a possible fourth law and an explanation of the initial state of the universe.

The search for a fourth law of thermodynamics has not only been motivated by the apparent (but not physically inexplicable) negentropy of the macroscopic universe but also by theoretical biology. To Stuart Kauffman, the existence of what he has termed "autonomous agents" similarly constitutes a phenomenon that seems at odds with physicalism, that is, the doctrine that all natural phenomena can be explained with nothing but the laws of a completed physics. Kauffman has been concerned with the fact that "living entities--bacteria, plants, and animals - manipulate the world on their own behalf: the bacterium swimming upstream in a glucose gradient [...]; the paramecium [a unicellular microorganism], cilia beating like a Roman warship's oars, hot after the bacterium" (p. 48). All this in seemingly flagrant violation of the second law. This general view has driven much of the systems biology movement from about 1900 to the 1950's (cf. e.g. Hirth 1900 and Auerbach 1910). What has differed is the exact formulation of what it means to be an autonomous agent, escaping the second law. According to Kauffman, "an autonomous agent is a self-reproducing system able to perform at least one thermodynamic work cycle" (p. 50). What this could mean Kauffman explicates more thoroughly in his "Investigations" (2000). It can be illustrated with cell metabolism. The steady state (that is, the flow equilibrium) of such metabolisms in open systems such as cells incorporated by the continuous intake and outflow crucially depends on the cells ability to process matter and energy. In order to do this, cells need the ability to put energy (e.g., in the form of enzymes) into the process. The organelles of the cells provide this ability. Such processes are examples of autonomous agency, although they are, of course, purely physico-chemical in character. Nevertheless, such entities clearly violate the second law (as long as they persist). This brings us to a further topic tackled in this volume, the problem of reductionism.

If a cell is an autonomous agent and thus, according to Kauffman, incorporates life then life is a physico-chemical process. One could express this fact by saying that biological explanations of life are reducible to explanations of the physico-chemical processes that constitute it. But this way of putting it seems to be contested by some of the authors in this volume. In his introduction to the volume, Davies rejects any "hope of reducing the operation of the human brain to that of subatomic particles" (p. 5). Nevertheless, neurophysiological models of the brain are fully based upon the physical descriptions of their constituent parts. Their reducibility to quantum mechanics is thus surely possible in principle (whether it is feasible or desirable at this point is a different question). But maybe Davies meant to challenge the possibility of reducing the operations of the mind (not the brain) to particle physics. Put like this, such reduction would involve not neurophysiological but psychological theories. All that contemporary neuroscience is able to tell are correlative relations between mental (that is, psychological) and neurophysiological (that is, physico-chemical) events. Accordingly, the challenge of reduction in this case fully depends on whether or not it will be possible to state neurological explanations laws that allow for mental explanation, or even prediction. This illustrates well that we cannot sensibly talk of a reduction relation regarding brains and subatomic particles. Reduction should be viewed as a relation between theories instead of a relation between minds and brains, or brains and subatomic particles (cf. Nagel 1961, chs. 11-12). Furthermore, if reduction thus understood is to be successfully carried out we need full-blown theories where the empirically meaningful content of a theory to be reduced can be fully expressed in the terms and relations that constitute the reducing theory. Discussion about the possibility of reduction between specific theory domains is thus utterly dependent on the availability of the according theories that would allow for the theoretical relation of reduction in the first place. Whether these will ever be available is an epistemic and not an ontological question. Nevertheless, many authors of the "emergentist" tradition in theoretical biology contend that such reduction will be impossible in principle.

In this vein, Harold J. Morowitz and Arthur Peacocke adress the subject matter of reduction and emergence in their respective contributions. Morowitz at first wants to view "emergence and reductionism not as antagonistic approaches but as complementary aspects of understanding" (p. 177) and Peacocke submits the view that emergence is "the entirely neutral name for that general feature of natural processes wherein complex structures [...] develop distinctively new capabilities and functions at levels of greater complexity" (p. 188). But then the issues get confused. "Emergence is the attempt to predict the structure and activities at one hierarchical level from an understanding of the properties of the agents at the next lower-level." Here, one can substitute "describe" for "predict" to get semantically identical concepts of emergence and reduction. Later, Morowitz suggests "unpredictable novelty" to distinctly characterize emergent phenomena (p.177). This different use of the term conforms well to that of the British Emergentists who coined it (cf. e.g. Stephan 1999, 85f. and McLaughlin in Beckermann et al. 1992). But a further characteristic of emergent phenomena that they suggested was the irreducibility (in principle) of emergent phenomena to phenomena on "lower" levels – a conceptual feature that Morowitz clearly does not share in his understanding ot emergence. Thus the reader is left with the mere postulate that the "whole is more than the sum of its parts" within the emergent orders of the natural world since Morowitz basically goes on to state that what is emergent is also "resultant" from properties and dispositions at "lower" levels. So why would there be ontological instead of only epistemic reasons for the impossibility of prediction of "resultant" phenomena? Morowitz' brief "emergentist" account of natural history and the layered structure of nature that follows is further burdened by the lack of explication of the notions of "lower" and "higher". For what is it that would make something belong to a "higher" level if it is unpredictable from, or even irreducible to theories in other domains if it could stand in no theoretical relation whatsoever to those domains? The notion of levels in nature thus already presupposes the possibility of theoretical relations between domains of phenomena (also cf. Kim 2002).

Arthur Peacocke acknowledges the distinction between an epistemic and an ontological notion of levels and emergence by suggesting that for any emergent entity to claim a genuine ontical (and, consequently, theoretical) status it would need to be intrinsically causally efficacious (p. 189-90). Put differently, it cannot be the constituents of a whole that incorporate its causal efficacy but  it need be the whole as an emergent and presumably nonreducible entity. An example is consciousness in higher animals. Consciousness could be viewed as being composite and united at the same time - which makes it a challenging subject matter for ontological inquiry because in what sense can an entity be a "whole" other than by just being a composite entity? Unfortunately, although Peacocke promisingly titles the second section of his contribution "The Relation of Wholes and Parts in Complex Systems", he really only discusses one specific kind of relation that a whole could conceivably have to its part, that is, the causal relation: the whole as causally constituted by its parts. More specifically, Peacocke takes on the subject matter of downward causation. Among other positions, Peacocke discusses the concepts of "strucuring causes" (Dretske 1993) and "boundary conditions" (Polanyi 1967, 1968). These are attempts to conceive of downward causation as not so much causal forces acting top down but as structures in wholes that limit the range of behavior of their constituent parts. (One further important way of conceptualizing wholes and their parts not discussed by Peacocke is represented by mereology and the complementary discussion on non-mereological wholes. Cf. e.g. Simons 1987, Van Inwagen 1990, Scaltsas 1990.)

An accurate explanation or interpretation of life and its conceptual and theoretical representations in the empirical sciences needs to be complemented with the history of life in the form of the theory of evolution. According to William Dembski, an "evolutionary algorithm" is "any well-defined mathematical procedure that generates contingency via some chance process and then sifts it via some lawlike process". Dembski proposes that evolutionary algorithms are to account for "specified complexity", a further tentatively characteristic feature of living things, apart from Kauffman's autonomous agency. Dembski shares Davies' definition (in his "Fifth Miracle", 1999) of 'specificity' as 'conformity to an independently given pattern' and 'complexity' as 'requiring a complicated instruction-set to be characterized' (p. 93). Dembski tries to show that evolutionary algorithms cannot account for specified complexity thus understood and he takes the well-known example of Dawkins (1986) as his target. To generate a sentence such as METHINKS IT IS LIKE A WEASEL (from Hamlet, act iii., sc. 2; presumably an instance of specified complexity and thus an analogy for, say, a cell) out of a randomly sequenced process of letter selection, a powerful algorithm will dramatically ease the task. Dawkins suggested to alter exclusively those letters not agreeing with the target sequence in each selection step and he demonstrated that such an algorithm actually produces the target sequence within a number of steps that lies many orders of magnitude below a completely random selection process in each step. Dembski contends that this explanatory strategy simply reshuffles specified complexity instead of explaining it (p. 95). Crucial for Dembski's attack is the notion of a "fitness function" since it is the latter that is to be maximized by such an algorithm. In Dawkins example, the fitness function assigns higher fitness to those sequences that have more letters in common with the target sequence METHINKS IT IS LIKE A WEASEL. The problem now seems to be where fitness functions originate. Dembski argues that a fitness function represents itself an instance of specified complexity and thus generates a vicious explanatory regress. "An evolutionary algorithm is supposed to find a target within phase space. To do this successfully, however, it needs more information than is available to a blind search." (p. 104-5) A potential solution according to the author is "for the phase space itself to constrain the informational context." (p. 106) But whence the constraints? Dembski submits that "even if Darwinian evolution is the means by which the panoply of life on Earth came to be, the underlying fitness function that constrains biological evolution would not be a free lunch and not a brute given but a finely crafted assemblage of smooth gradients that presupposes much prior specified complexity." (p. 113) This may be a premature conclusion. First, the fitness function would need to have some kind of empirical correlate in order to be explanatory within this mathematical model in the first place. Dembski tacitly presupposes the empirical correlates of variation and selection, but he is quiet on selective factors external to the letter sequence (or the genome, or the organism) that could provide for such a correlate. Remember that the explanandum was to be specified complexity, and that specificity was characterized as conformity to an independently given pattern: The explanandum thus explicitly involves an element external to its proper systemic boundary. Secondly, it is one thing to point to a possible explanatory gap in specific theorems about evolution but quite something else to suggest craftsmanship in the development of life.

This reviewer feels that it is a weakness of this volume that several of its contributors fall back on the strategy of filling presumed explanatory gaps in the empirical sciences with kinds of explanations whose further ontological and methodological implications are in almost all cases not thought through well. This starts in Davies' introduction where he suggests that the "multiverse" explanation regarding the fine-tuning of the universe may be on a par with designer theories (p. 12). Davies seems unaware of the methodological difference between the purely semantical character of multiverse theories (since there can in principle be no empirical evidence for multiple universes in their sense) and the ontological implications of invoking a designer god and his or her causal efficacy within this universe. The latter obviously contradicts physicalism while the former does not. Further, Davies is unclear about the exact position of theological explanation in the scientific edifice. Although he notes that the "emergence of meaning" is a "postscientific quest" (an unexplicated notion) he still seems to ask for the kind of explanations regarding the origin of meaning that one cannot help but regard as scientific explanations, that is, statements that nomically or causally link objects or events in the universe and that are empirically grounded.

The tactic reappears in the third section with Morowitz asserting without further ado that emergence "maps onto the theological constructs" that he identifies with the Judaic God of History, the Greek mythology, and the "immanent God" of "the religion of Spinoza and Einstein" and the book of Job (p. 180-5). This mapping remains unargued for, although Morowitz' stance on it is clear. Natural science is to be interpreted as a tale about the "unfolding of the universe of the immanent God into all its novelties and surprises [that] we designate as emergence." In a similar vein, Arthur Peacocke concludes his discussion about emergence and downward causation with "a world in which over the course of space-time new realities have emerged" (p. 197). But if, in his final analysis, one of these realities is God, as he asserts on p198, then this conclusion is composed of claims whose epistemic warrants are surely incommensurable. After seeing many of the central questions regarding the concept of emergence and downward causation raised and several interesting strategies to tackle them cursorily sketched the reader is left, again, with doctrine and lack of argument: "[The] very existence of all-that-is, with that inherent creativity to bring persons out of quarks [...], is for me and all theists only explicable by postulating an Ultimate Reality which is [God]."

In the final contribution, N. H. Gregersen seems at first straightforward about the strictly theological nature of his inquiry, that is, the exploration of "the religious significance of complexity studies." To his mind, the presumed fact of the "robustness" of self-organizational processes "might seem to make any reference to an external creator obsolete." Gregersen counters this threat by arguing for an understanding of self-organization "as the apex of divine purpose. [...] God's design of the world as whole favors the emergence of autonomous processes in the particular course of evolution" (p. 207). We should thus "see God as continuously creating the world by constituting and supporting self-organizing processes." (p. 208) Gregersen contrasts this view to the deterministic arguments from design that, to his mind, are unable to properly account for the explanatory power of scientific theories of self-organization and the according role of chance in the development of the natural world. It is these elements that he intends to relate sensibly to the "anthropic principle". The strong version of this principle states that the universe must have those properties which allow life to develop within it at some stage of its history (Denton 1998, cit. in Gregersen) whereas its weak version states that the existence of life may be a contingent fact in the history of the universe but that the latter is still not only adapted to the development of life but that life itself and its evolution is restricted by the parameters of an anthropic universe (Barrow and Tipler 1986, cit. in Gregersen). Gregersen favors the weak version since it seems better suited to an integration of a contemporary scientific account of the history of the universe and theological cosmology (Ellis 1993, cit. in Gregersen). According to him, the "causal approach" and the "noncausal qualitative approach" are two ways in which this could be cashed out. The causal approach is an answer to the question of what "effective difference" God can make in the history of the universe, the problem of "natural theology". According to this approach, God can be conceived of as a causal agent in "quantum processes", or by "constraining the possibilities on the world as a whole", or by managing "probability patterns throughout evolution". The explanandum for which God is the explanans here is thus, once again, the presumed explanatory gap left by contemporary science (a strategy that Gregersen reconstructs and criticizes in design theories on p. 210-11): "How can we account for the preferred pathways of evolution (and their changing probability rates) if these are not fully explainable from the underlying laws of basic physics." According to Gregersen, God incorporates the "preferential principle that explains why nature follows this particular route rather than all other energetically possible routes" (p. 226). God is here seen both as the Creator of the basic laws of nature (in his transcendent mode) and as a causal agent that fills the gaps of natural chance in the development of life (in his immanent mode). But if God is really to be assigned a material causal role in the universe then she will be the subject matter of the empirical sciences. This surely seems to be neither a reasonable nor a desirable perspective.

Auerbach, Felix (1910); Ektropismus oder die physikalische Theorie des Lebens. Leipzig: Wilhelm Engelmann.

Beckermann, Ansgar/Flohr, Hans/Kim, Jaegwon (eds. 1992); Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism. Berlin etc.:Walter de Gruyter.

Boltzmann, Ludwig (1872); "Weitere Studien über Wärmegleichgewicht unter Gasmolekülen", in: Wiener Berichte, 66:275-370. Repr. in Boltzmann's Wissenschaftliche Abhandlungen (Leipzig, 1909; New York, 1968), I, 316-402; trans. "Further Studies on the Thermal Equilibrium of Gas Molecules", in: Kinetic Theory, ed. S. G. Brush, 2 vols. (Oxford, 1966), II, 88-175.

---(1877); "Über die Beziehung zwischen dem zweiten Hauptsatze der mechanischen Wärmetheorie und der Wahrscheinlichkeitsrechnung, respektive den Sätzen über das Wärmegleichgewicht", in: Wiener Berichte, 76:373-435; repr. in Abhandlungen, II, 164-223.

Clausius, Rudolf (1865); "Über verschiedene für die Anwendung bequeme Formen der Hauptgleichungen der mechanischen Wärmetheorie", in: Poggendorffs Annalen, 125:353-400; trans. "On Several Convenient Forms of the Fundamental Equations of the Mechanical Theory of Heat", in: The Mechanical Theory of Heat. London, 1867, pp327-65.

Dawkins, Richard (1986); The Blind Watchmaker. New York: Norton.

Dyson, Richard (1979); "Time Without End: Physics and Biology in an Open Universe", in: Reviews of Modern Physics 51:447-60.

Helmholtz, Hermann von (1854); "Über die Wechselwirkung der Naturkräfte und die darauf bezüglichen neuesten Ermittelungen der Physik", in: Populäre wissenschaftliche Vorträge, 2nd. ed. (Braunschweig, 1876), pp91-136; trans. in Popular Scientific Lectures, ed. M. Klein (New York, 1961)

Hirth, Georg (1900); Die Ektropie der Keimsysteme. Munich.

Kim, Jaegwon (2002); "The Layered Model: Metaphysical Considerations", in: Philosophical Explorations 5:2-20.

Nagel, Ernest (1961); The Structure of Science. Problems in the Logic of Scientific Explanation. New York etc.: Harcourt, Brace & World.

Scaltsas, Theodore (1990); "Is a Whole Identical to its Parts?" in: Mind 396(99):583-98.

Simons, Peter (1987); Parts. Oxford University Press.

Stephan, Achim (1999); Emergenz. Von der Unvorhersagbarkeit zur Selbstorganisation. Dresden: Dresen University Press.

Thomson, William [Lord Kelvin] (1852); "On a Universal Tendency in Nature to the Dissipation of Mechanical Energy", in: Proceedings of the Royal Society of Edinburgh, 3:139-42. Also in Philosophical Magazine 4:304-06; repr. in Mathematical and Physical Papers of Lord Kelvin (London 1911), 1, 511-14.

Van Inwagen, Peter (1990); Material Beings. Cornell University Press. (Thoroughly discussed in Philosophy and Phenomenological Research, vol. 53, issue 3)

© 2004 Rainer Kamber

Rainer Kamber, Lic.Phil., Basel University, Program Mensch Gesellschaft Umwelt MGU, Basel, Switzerland