Darwin's Doubt

by Stephen C. Meyer

  Self-Organization and Epigenetic Information

  To see why let’s look first at how Kauffman attempts to explain the epigenetic “positional” information that directs the organization of cells in the second phase of animal development. Kauffman attempts to explain this “positional” information by offering an entirely hypothetical and, ultimately, question-begging proposal. He invokes an idea sketched out in the 1940s by the famous English mathematician Alan Turing.15 Turing proposed that specific arrangements of cells in animal development might ultimately derive from the diffusion and specific arrangement of crucial molecules—presumably something like the morphogen proteins present in embryonic cells. (Recall that morphogens, or morphogen proteins, influence cell differentiation and organization during animal development.) Rather than attributing the distribution of these morphogenic proteins to preexisting genetic and epigenetic information in cells, as occurs during development in modern animals, Turing postulated that the distribution of these molecules might have originated in the first place independently of such information as the result of simple chemical reactions. He imagined one molecule producing both a copy of itself (“autocatalyzing”) and, in addition, producing a different molecule as well. Then he envisioned one of these molecules inhibiting the production of the other, thus allowing, through repeated cycles, the production of more and more of one molecule and less and less of the other. Turing thought the resulting nonuniformities in the patterns of distribution of those molecules would eventually result in nonuniform patterns in the distributions of different cells, possibly resulting in different animal forms.

  Kauffman expanded upon this proposal as a way of understanding how crucial positional information might organize as the result of chemical interactions of different molecules. Nevertheless, his proposal suffers from an obvious difficulty: it lacks any chemical or biological specificity. In explaining the proposal, Kauffman does not mention any specific chemicals or proteins that would behave in the way he envisions. Instead, he describes the behavior of hypothetical molecules that he labels with the indistinct monikers “X” and “Y.” More important, Kauffman offers no evidence that chemicals interacting in the way he envisions could create specific biologically relevant configurations or distributions of morphogen proteins—apart, that is, from the processes that generate specifically arranged distributions of these proteins in preexisting information-rich embryonic cells today.

  Instead, it is inherently implausible to think that the specificity necessary to coordinate the movements and arrangements of the billions or trillions of cells present in adult animal forms could be established by the interactions of one or two simple chemicals, even if they formed autocatalytic cycles. Kauffman himself seems tacitly to acknowledge the difficulty of generating biological specificity from the reactions of chemicals alone. He notes, in critique of his own model, that patterns of molecular diffusion produced by chemical autocatalysis would depend crucially upon “the initial conditions.”16 In other words, getting a biologically relevant information-rich arrangement of morphogenic proteins would require starting with a very specific (presumably information-rich) arrangement of autocatalyzing molecules.

  Kauffman encounters this same problem in attempting to explain the origin of the first life as the result of autocatalytic reactions starting from a prebiotic soup. In The Origins of Order, he acknowledges that generating an autocatalytic, or self-reproducing, set of molecules—a crucial step in his origin-of-life scenario—would require “high molecular specificity”17 in the initial set of peptides or RNA molecules. In other words, it would require specificity of arrangement and structure, that is to say, functional information.

  Self-Organization and Genetic Information

  And what about the specifically genetic information necessary to the earlier phase of animal development? Does Kauffman’s self-organizational theory explain the origin of the “genetic regulatory networks” necessary to cell differentiation? Again, it does not. Instead, in an even more obvious way, it begs the question of the origin of these regulatory networks. Indeed, though Kauffman discusses cell differentiation as a kind of “self-ordering” or self-organizational process, he acknowledges that the predictable pathways of differentiation that characterize this process derive from preexisting gene regulatory networks. As Kauffman notes, the spontaneous ordering tendencies in cell differentiation are “inherent in a wide class of genomic regulatory networks.”18 Indeed, the genetic information in the gene regulatory networks does not come from self-ordering processes of cell differentiation. Instead, cell differentiation, to the extent that it can be properly described as “self-ordering,” results from preexisting genetic sources of information. Thus, the self-organizational process that Kauffman cites cannot explain the origin of genetic information, because it derives from it, as Kauffman’s own description reveals.

  In a later book, At Home in the Universe: The Search for the Laws of Self-Organization and Complexity, Kauffman does offer computer simulations of two “model systems”19 that seek to explain, at least in principle, how genetic information might have self-organized. In one example, he describes a system of buttons connected by strings.20 The buttons represent novel genes or proteins and the strings represent self-organizational forces of attraction between the proteins. Kauffman suggests that when the complexity of this system reaches a critical threshold (as represented by the number of buttons and strings), new modes of organization might arise in the system “for free”21—without intelligent guidance—similar to the way that water spontaneously changes to ice or vapor under specific conditions.

  Kauffman asks his readers to imagine a system of many interconnected lights. Each light can flash in a variety of states—on, off, twinkling, and so on. Since each light can adopt more than one possible state, the system may adopt a vast number of possible states. Further, in his system, rules determine how past states influence future states. Kauffman asserts that, as a result of these rules, the system would, if properly tuned, eventually produce a kind of order in which a few basic patterns of light activity recur with greater than random frequency. Since these patterns represent a small portion of the total number of possible states in which the system can reside, Kauffman suggests that self-organizational laws might similarly find highly improbable biological outcomes—perhaps even functional sequences of bases or amino acids within a much larger sequence space of possibilities.22

  It’s not hard to see why these simulations also would fail to account for the origin of the new genes and proteins needed to produce the Cambrian animals. In both of his examples Kauffman presupposes significant sources of preexisting information. In his buttons and strings simulation, he intends the buttons to represent proteins, themselves the result of preexisting genetic information. Where did that information come from? Kauffman doesn’t say, but it is an essential part of what needs explanation in the history of life. Similarly, in his light system, the order that allegedly arises “for free”—that is, apart from an intelligent input of information—only does so if, as Kauffman acknowledges, the programmer “tunes” the system to keep it from either (a) generating an excessively rigid order or (b) devolving into chaos.23 This tuning presumably involves an intelligent programmer selecting certain parameters and excluding others—that is, inputting information. In fact, in summarizing the import of this illustration, Kauffman insists that it shows how the “orderliness of the cell, long attributed to the honing of Darwinian evolution, seems instead likely to arise from the dynamics of the genome network,”24 that is, from preexisting—unexplained—sources of genetic information.

  In addition, Kauffman’s model systems are not analogous to biological systems because they are not constrained by functional considerations. A system of interconnected lights governed by preprogrammed rules may well settle into a small number of patterns within a much larger space of possibilities. But since these patterns have no function and need not meet any functional requirements, they have no specificity ana
logous to that in the genes of actual organisms. Kauffman’s model systems do not produce sequences or systems characterized by specified complexity or functional information. They produce modules of repetitive order distributed in an aperiodic manner, yielding mere complexity (i.e., information only in the Shannon sense).25 Getting a law-governed system to generate repetitive patterns of flashing lights, even with a certain amount of variation, is interesting, but not biologically relevant. A system of lights flashing “Vote for Jones,” on the other hand, would model a biologically relevant outcome, at least, if such a functional sequence of letters arose without intelligent agents programming the system with equivalent amounts of functionally specified information.

  Kauffman on the Cambrian

  Kauffman also proposes a specific self-organizational mechanism to explain some aspects of the Cambrian explosion. According to Kauffman, new Cambrian animals emerged through “long-jump” mutations that established new body plans in a discrete rather than gradual fashion.26 He recognizes that mutations affecting early development are almost inevitably harmful.27 Thus he concludes that body plans, once established, will not change, whatever subsequent evolution may occur. This keeps his proposal consistent with a top-down pattern in the fossil record in which higher taxa (and the body plans they represent) appear first, only later to be followed by the multiplication of lower taxa representing variations within those original body designs.

  Even so, Kauffman’s proposal begs the most important question: What produces the new Cambrian body plans in the first place? By invoking “long-jump mutations,” he identifies no specific self-organizational process that can produce such changes. Moreover, he concedes a principle that undermines his own proposal. As noted above, Kauffman acknowledges that mutations early in development are almost inevitably deleterious. Yet developmental biologists know that these are the only kind of mutations that have a realistic chance of producing large-scale evolutionary change—the big jumps that Kauffman invokes. Though Kauffman repudiates the neo-Darwinian reliance upon random mutations, he must invoke the most implausible kind of random mutation to provide a self-organizational account of the new Cambrian body plans.

  Developmental Toolkits and Self-Organizational Processes

  More recently, another advocate of self-organization, Stuart Newman, a cell biologist at the New York Medical College, has published several papers suggesting that self-organizational processes can help explain the origin of body plans. In a paper in the volume produced from the Altenberg 16 conference, Newman develops a model that resembles Kauffman’s, but one that offers more biological specificity.28

  Newman, like Kauffman, invokes self-organizational processes. But Newman sees these processes acting dynamically and in coordination with a genetic “toolkit.” His model emphasizes the importance of a highly conserved (i.e., similar) set of regulatory genes in all the major Cambrian taxa. In his view, this common “developmental genetic toolkit”29 has been used “to generate animal body plans and organ forms for more than half a billion years”30 since the inception of the animal kingdom.

  But if all the animal taxa have the same toolkit, why are the various forms of animals and higher metazoan taxa so different from one another? For Newman, the answer to this question requires understanding how self-organizing processes influence the interaction of cells during development and how they cause genes to acquire different functions affecting the interactions of cells.

  For example, he attributes the emergence of multicellularity to cells acquiring the capacity “to remain attached to one another after dividing.”31 This capacity in turn derives not from generating new genes and proteins (as neo-Darwinism would assume). Instead, it derives from the repurposing of old genes and proteins in response to specific self-organizational (and epigenetic) processes such as the “physical force of adhesion.”32 Newman proposes, further, that once the first multicellular organisms had arisen, they would have “set the stage for additional physical processes to come into play”33—processes that could alter the expression and function of still other genes in the developmental genetic toolkit, resulting in wholly new and different body plans. As Newman explains, “The phenomenon of multicellularity opened up possibilities for these molecules to become involved in the molding of bodies and organs.”34

  Newman envisions new animal body plans resulting from different cells sticking to each other in different configurations because of different forces of attraction between molecules on the surface of cells and because of different patterns of distribution of crucial molecules within cells. He calls these self-organizational forces and factors “dynamical patterning modules” (or DPMs).35 Figure 15.2 shows some typical ways that cells cluster together or arrange themselves as the result of these self-organizational forces. Newman lists many “dynamical patterning modules,” or self-organizational forces, responsible for the spontaneous emergence of these different cell clusters, including “adhesion, shape and surface polarization, switching between alternative biochemical states, biochemical oscillation, and the secretion of diffusible and nondiffusible factors.”36

  FIGURE 15.2

  Dynamical Patterning Modules (DPMs), showing the different ways that, according to biologist Stuart Newman, cells can stick to one another (“aggregate”) and form structures during animal development.

  To get a handle on what Newman has in mind, think of cells as Lego blocks. There are many different ways of connecting Lego blocks, depending on the shape of the blocks and the pattern of raised bumps and indentations on the blocks. These patterns allow small groups of Legos to be arranged into different modular structures: cubes, walls, circular rings, and so forth. Each of these smaller modular structures can then be combined to make many different larger structures, from airplanes and skyscrapers to submarines and castles. In a similar way, Newman suggests that different forces of adhesion between cells and different patterns of molecular diffusion within and between cells will generate many different patterns or motifs of multicellular organization, which in turn function as modular elements that can be combined in various ways to make diverse animal forms.37

  Do these self-organizational processes account for the origin of animal body plans in the Cambrian explosion or the information necessary to produce new animal forms? Again, they do not. Instead, Newman, like Kauffman, either fails to offer an adequate mechanism for generating crucial sources of biological information, or he begs the question by presupposing the existence of various sources of information.

  Assume a Toolkit

  In the first place, Newman obviously presupposes the existence of a “developmental genetic toolkit,” that is, a whole set of genes, including regulatory genes, that help to direct the development of animal body plans. Where does this genetic information come from? He doesn’t specify, though presumably he might be assuming the neo-Darwinian mechanism somehow produced the genetic information in the toolkit. If so, he leaves his model vulnerable to the criticisms outlined in Chapters 9 through 12. He certainly does not cite any specifically self-organizational process to explain the origin of the genetic toolkit. He also incorrectly seems to presuppose that the genes present in this common toolkit provide all the genetic information necessary to specify individual body plans. But this overlooks a host of recent findings showing that individual species within specific taxa often require genes for development that are specific to those species and taxa.38 Thus, these genes would not have been present in a common metazoan toolkit of the kind Newman postulates.

  Second, Newman does not account for the origin of the information necessary to organize modular arrangements or groups of cells into whole animal body plans. The forces at work in his dynamical patterning modules explain, at best, only the arrangements of small groups of cells, not the arrangements of those modular cell clusters into tissues, organs, and whole body plans.

  Think, again, of arranging Lego blocks. There are many ways of arranging small numbers of Lego blocks. These various arrangements form common
structural motifs such as: two blocks stuck together at right angles; several curved blocks forming circular rings, stacked blocks forming hollow squares or walls or cube-like shapes; blocks arranged as prisms or cylinders; flat layers of blocks stacked two bumps thick or three bumps thick or more. Though these structural elements stick together because of interactions between the bumps and indentations on each block, those bumps and indentations themselves do not specify any particular larger structure—a castle or an airplane, for example—because each motif may be combined or recombined with many other structural motifs in numerous different ways. The shape and properties of the modular elements do not dictate the type of larger structure that must be built from them. Instead, to build a particular structure, the modular elements must be arranged in particular ways. And since there are many possible ways to arrange these modular elements, only one or a few of which will result in a desired structure, every Lego set includes a blueprint with step-by-step instructions—in other words, additional information.

  In a similar way, producing a body plan from the different types of cell clusters generated by Newman’s dynamical patterning modules (DPMs), would also require additional information. Newman does not account for this information. He correctly highlights the way certain recurrent motifs for organizing groups of cells seem to form spontaneously as the result of physical interactions between individual cells (his DPMs). He does not, however, establish that these groups of cells must arrange themselves into specific tissues, organs, or body plans in response to any known physical process or law. Instead, it seems entirely possible that these modular elements (cell clusters) have many “degrees of freedom” and can be arranged in innumerable ways. If so, then some additional information—an overall organismal blueprint or set of assembly instructions—would need to direct the arrangement of these modular elements. Newman does not consider this possibility. Nor does he cite any law-like self-organizational process that would eliminate the need for such information to direct animal development.