There is yet a further problem with Newman’s proposal. Even the capacity for cells to self-organize into dynamical patterning modules probably derives from prior unexplained sources of information. Newman’s DPMs undoubtedly form as the result of interactions between molecules on the surface of cells, and as the result of chemical gradients between cells—with the specific configuration and properties of those molecules determining the exact structure of the individual DPMs. In that sense, the DPMs do, of course, self-organize, but clearly the specific ways that cells typically cluster together will depend upon highly specific and complex forces of interaction between the molecules and groups of molecules on the surface of these cells. Many of the molecules that contribute to these interactions are no doubt proteins, obvious products of genetic information. But, in addition, cell-to-cell interactions are affected by the arrangement of proteins and other molecules on the surface of cells (such as the sugars in the sugar code, see Chapter 14) as well as by the arrangement of structures made of proteins. But these molecular arrangements are, in turn, specified by either preexisting genetic or, more probably, epigenetic sources of information and structure. Thus, Newman’s analysis shows that the self-ordering tendencies (or biological laws of form), to the extent they exist, depend upon preexisting sources of biological information. Newman, again, does not explain where that information comes from.
Newman does emphasize how epigenetic sources of information affect the expression and the function of gene products during the process of animal development. He notes that different gene products may perform different functions depending upon the organismal context in which they find themselves. But Newman does not explain with any specificity how the genes in the common toolkit acquire different functions in response to self-organizational processes or where the epigenetic information comes from that determines those functions. Yet, clearly context-dependent gene expression depends upon a host of other preexisting epigenetic sources of information and structure.
Recall, for example, the discussion of cell membrane targets from Chapter 14. These targets provide an important source of epigenetic information by influencing the positioning of crucial morphogenic proteins. Yet, the arrangements of the targets on the cell membrane do not self-organize as the result of simple chemical interactions between the proteins out of which they are made39—that is, the proteins do not determine the location of the membrane targets on the interior of the cell. Instead, the location and the structure of membrane targets are transmitted from parent membrane to daughter membrane, a process that transmits preexisting epigenetic structural information from the parent membrane. Newman does not attempt to explain the origin of this information or structure by any known self-organizational process. Instead, the evidence indicates that interior and exterior membrane targets, cytoskeletal arrays, the sugar code, and many other sources of epigenetic structural information do not self-organize as the result of physical interactions between their respective molecular subunits.40
Order vs. Information
Self-organizational theorists face, in addition, a conceptual distinction that has cast doubt on the relevance of their theories to biological systems. Self-organizational theorists seek to explain the origin of “order” in living systems by reference to purely physical or chemical processes (or laws describing those processes). But what needs to be explained in living systems is not mainly order in the sense of simple repetitive or geometric patterns. Instead, what requires explanation is the adaptive complexity and the information, genetic and epigenetic, necessary to build it.
Yet advocates of self-organization fail to offer examples of either biological information or complex anatomical structures arising from physics and chemistry alone. They either point, as Newman and Kauffman do, to embryological development unfolding predictably as the result of preexisting information-rich gene products, cell membranes, and other cell structures. Or they offer examples of purely physical and chemical processes generating a kind of order that has little relevance to the features of living systems that most need explanation.
In the latter case, self-organizational theorists often point to simple geometric shapes or repetitive forms of order arising from or being modified by purely physical or chemical processes. They suggest that such order provides a model for understanding the origin of biological information or bodyplan morphogenesis.41 Self-organizational theorists have cited crystals, vortices, and convection currents (or stable patterns of flashing lights) to illustrate the supposed power of physical processes to generate “order for free.” Crystals of salt do form as the result of forces of attraction between sodium and chloride ions; vortices can result from gravitational and other forces acting on water in a draining bathtub; convection currents do emerge from warm air (or molten rock) rising in enclosed spaces. And some molecules found in living systems do adopt highly ordered structures and recognizable geometric shapes as the result of the physical interactions of their constituent parts alone. Nevertheless, the type of order evident in these molecules or physical systems has nothing to do with the specific “order” of arrangement—the information or specified complexity—that characterizes the digital code in DNA and other higher-level information-rich biological structures.
This is easiest to see in the case of the information encoded in DNA and RNA. Some of what follows may be familiar from my discussion in Chapter 8, but it bears repeating. The bases in the coding region of a section of DNA or in an RNA transcript are typically arranged in a nonrepetitive or aperiodic way. These sections of genetic text display what information scientists call “complexity,” not simple “order” or “redundancy.”
To see the difference between order and complexity consider the difference between the following sequences:
Na-Cl-Na-Cl-Na-Cl-Na-Cl
AZFRT
The first sequence, describing the chemical structure of salt crystals, displays what information scientists call “redundancy” or simple “order.” That’s because the two constituents, Na and Cl (sodium and chloride), are highly ordered in the sense of being arranged in a simple, rigidly repetitive way. The sequence on the bottom, by contrast, exhibits complexity. In this randomly generated string of characters, there is no simple repetitive pattern. Whereas the sequence on the top could be generated by a simple rule or computer algorithm, such as “Every time Na arises, attach a Cl to it, and vice versa,” no rule shorter than the sequence itself could generate the second sequence.
The information-rich sequences in DNA, RNA, and proteins, by contrast, are characterized not by either simple order or mere complexity, but instead by “specified complexity.” In such sequences, the irregular and unpredictable arrangement of the characters (or constituents) is critical to the function that the sequence performs. The three sequences below illustrate these distinctions:
Na-Cl-Na-Cl-Na-Cl-Na-Cl (Order)
AZFRT
Time and tide wait for no man (Specified complexity)
What does all this have to do with self-organization? Simply this: the law-like, self-organizing processes that generate the kind of order present in a crystal or a vortex do not also generate complex sequences or structures; still less do they generate specified complexity, the kind of “order” present in a gene or functionally complex organ.
Laws of nature by definition describe repetitive phenomena—order in that sense—that can be described with differential equations or universal “if-then” statements. Consider, for example, these informal expressions of the law of gravity: “All unsupported bodies fall” or “If an elevated body is left unsuspended, then it will fall.” These statements represent reasonably accurate law-like descriptions of natural gravitational phenomena precisely because we have repeated experience of unsupported bodies falling to the earth. In nature, repetition provides grist for lawful description.
The information-bearing sequences in protein-coding DNA and RNA molecules do not exhibit such repetitive “order,” how
ever. As such, these sequences can be neither described nor explained by reference to a natural law or law-like “self-organizational” process. The kind of nonrepetitive “order” on display in DNA and RNA—a precise sequential “order” necessary to ensure function—is not the kind that laws of nature or law-like self-organizational processes can—in principle—generate or explain.
Otherwise, the nucleotide bases would repeat rigidly—such as ACACACACACACACACAC—in a way that would not allow DNA to store or convey specified information. A curious feature of the chemistry of DNA allows any one of the four nucleotide bases to attach to any site on the interior backbone of the DNA molecule. This chemical indeterminacy makes it possible for DNA and RNA to store any one of a virtually unlimited number of different arrangements of nucleotide bases—in effect, to encode any genetic message. But this indeterminacy also categorically defies explanation by deterministic law-like forces of chemical attraction. And because forces of attraction do not determine the sequence of nucleotide bases in DNA or RNA, the origin of the specific arrangement of the bases—the information in DNA and RNA—cannot be attributed to self-organizing forces of attraction either.
Hubert Yockey, a leading innovator in the application of information theory to molecular biology, first recognized the problems associated with invoking self-organization to explain the origin of biological information. These theories fail, he argued, for two reasons. First, they do not distinguish order from information. And, second, the information in the DNA molecule does not derive from law-like forces of attraction.42 As he explained in 1977: “Attempts to relate the idea of order … with biological organization or specificity must be regarded as a play on words that cannot stand careful scrutiny. Informational macromolecules can code genetic messages and therefore can carry information because the sequence of bases or residues is affected very little, if at all, by [self-organizing] physicochemical factors.”43
Much the same thing is true of many vital sources of epigenetic information. The forces of attraction between constituent proteins in membrane targets or cytoskeletal arrays, for example, do not determine the structure or location of these epigenetic structures and the positional information they provide. The origin of these structures cannot be attributed to self-organizing forces of attraction either. Instead, in each case, information-rich epigenetic structures are generated from preexisting sources of epigenetic information.
Thus, self-organizational theories explain well what does not mainly need explaining in biology, namely, repetitive or simple geometric forms of order. Self-organizational theorists do cite structures that might have self-organized. But these examples are typically extremely modest in scope. They include repeating patterns of atoms in crystals; simple geometric figures; patterns of lines, triangles, and streaks; vortices; spiral wave currents; and simple shapes that glide across computer screens.44 None exhibit the specified complexity that characterizes the digital information in DNA and RNA or the complex arrangements of proteins, cells, tissues, and organs necessary to build a functioning form of animal life.
Natural Magic or True Cause?
In 2007, I participated in a private meeting of evolutionary biologists and other scientists who shared the conviction that a new theory of biological origins is now needed. In attendance were several prominent advocates of the self-organization approach. During the meeting, these scientists presented intriguing analogies from physics and chemistry to show how order might have arisen “for free”—that is, without intelligent guidance—in the biological realm. Yet the order they described in these analogies seemed to have no direct relevance to the complexity—indeed the specified complexity—of genes or cell membranes or animal body plans. Other scientists at the conference challenged the advocates of self-organization to cite known processes that could produce biologically relevant form and information.
Near the end of the meeting one advocate of self-organization privately acknowledged to me the validity of these critiques, admitting that, for now, “self-organization is really more of a slogan than a theory.” Stuart Kauffman, perhaps attempting to make a virtue of the necessity of accepting this explanatory deficit, has recently celebrated the self-organizational perspective for embracing what he calls “natural magic.” In a lecture at MIT, he concluded: “Life bubbles forth in a natural magic beyond the confines of entailing law, beyond mathematization.”45 He went on to explain that one benefit of the self-organizational perspective is that it allows us to be “reenchanted” with nature and to “find a way beyond modernity.”46
Since the beginning of modern science, scientists have championed a commonsense principle of scientific reasoning known as the vera causa principle. This principle holds that explaining a particular phenomenon or event requires identifying a “true cause” that is known from experience to have the power to produce the event or phenomenon in question. The early modern scientists affirmed this principle as one of the key aspects of a scientific approach to understanding nature. This stood in opposition to the magical thinking that had gone before in which people attributed powers to nature that they had never observed it manifesting. As the scientific revolution matured, endeavors such as alchemy, for example, were eventually rejected precisely because the alchemists could not identify a cause that could effect the transformation they were seeking to demonstrate.
Self-organizational theories have clearly failed to provide a vera causa for the origin of biologically relevant forms of “order”—the functional complexity and specified information present in living systems. Instead, they either beg the question as to the ultimate origin of biological information or point to physical and chemical processes that do not produce the specified complexity that characterizes actual animals.
Viewed in this light, Kauffman’s recent discussion of natural magic and calls for a “reenchantment” with nature sound less like a bold new initiative to reconcile science and spirituality (which is what he intended) than a tacit admission that self-organizational theories have failed to identify known physical and chemical processes capable of generating the form and information present in actual living systems. Indeed, after years of attempting to solve the problem of the origin of form, Kauffman’s recent ruminations about “natural magic” sound a lot like an admission that a profound mystery remains.
16
Other Post-Neo-Darwinian Models
When Stephen Jay Gould was first wrestling with the question of how new forms of animal life could have arisen so quickly in the fossil record, he considered many possible mechanisms of change. In the famed 1980 paper in which he declared neo-Darwinism “effectively dead,”1 he didn’t just propose allopatric speciation and species selections as new evolutionary mechanisms. He also granted a rehearing to a long discredited idea. Specifically, he argued that large-scale “macromutations” might generate significant innovations in form relatively quickly.2
In the 1930s and 1940s, this idea had been associated with University of California at Berkeley geneticist Richard Goldschmidt. Aware of the many discontinuities in the fossil record, Goldschmidt envisioned radical transformations in the form of animals arising in even one generation as the result of such large-scale mutations. He endorsed, for instance, the view of the German paleontologist Otto Schindewolf (1896–1971) that “the first bird hatched from a reptilian egg” and, thus, in Goldschmidt’s words, “that the many missing links in the paleontological record are sought for in vain because they have never existed.”3 If a bird hatched directly from a reptilian egg as the result of heritable, large-scale mutations, then such a sudden leap or “saltation” would obviously leave no fossil intermediates behind.
Neo-Darwinists rejected this idea as biologically implausible in the extreme. They argued that changing so many functionally integrated anatomical and physiological systems so quickly would inevitably result in deformed mutants, not different integrated systems of organs constituting a whole new animal.4 Goldschmidt’s macromutations, they contended, would produce not w
hat Goldschmidt called “hopeful monsters,” but “hopeless monsters”—that is, nonviable organisms.5
Though Gould wanted to reconsider a role for large-scale mutations, he carefully disassociated his proposal from Goldschmidt’s much ridiculed idea. Instead, he suggested that the mutations affecting genes in animal development might generate larger increments of morphological innovation than the mutations that affected other genes. These “developmental mutations,” he thought, might generate modular parts of biological systems in a short time—without needing to generate whole new forms of life in a single generation. He offered, as an example, the possibility that the gill arch bones of ancient jawless fish, though not the whole fish, might have arisen in one step as the result of a developmental macromutation. Gould explained: “I do not refer to the saltational origin of entire new designs, complete in all their complex and integrated features… . Instead, I envisage a potential saltational origin for the essential features of key adaptations.”6
In response to heavy criticism from neo-Darwinists, Gould later downplayed the role of such larger-scale developmental mutations in the theory of punctuated equilibrium. Other evolutionary biologists, however, took his idea as an inspiration and developed theories that emphasize such developmental mutations as a driving force in macroevolution. Evolutionary theorists and developmental biologists such as Rudolf Raff, Sean B. Carroll, and Wallace Arthur have developed a subdiscipline of biology known as evolutionary developmental biology, or “evo-devo” for short. The evolutionary developmental biologists have since formulated alternative models that challenge a key aspect of the neo-Darwinian triad. Whereas neo-Darwinism envisions new form arising as the result of slow, incremental accumulations of minor mutations, evolutionary developmental biologists argue that mutations affecting genes involved in animal development can cause large-scale morphological change and even whole new body plans.
Darwin's Doubt Page 34