Life without a Replicator

A creation myth can be a powerful window into a culture’s most important belief systems. In Chapter Three I described the creation myth of biology, the conventionally accepted story of how life originated with an initial replicating molecule, which eventually mutated and evolved into the diversity of forms we see today. Lost amid all the speculation about possible RNA or peptide candidates for the “first replicating molecule” is the fact that no such molecule exists today. There is no gene, no sequence of DNA or RNA, that can replicate itself—not without an awful lot of help from other genes. In fact, what genetic reproduction entails is many, many genes cooperating in their mutual replication. Each gene plays a very limited—though sometimes indispensable—role in this collective replication. Perhaps it provides instructions for the production of a protein important in sex hormone production; perhaps it turns on another gene that initiates cell differentiation at a crucial stage of development. Each of these and other functions certainly are necessary for a gene’s replication, but they are not sufficient. Even in the simplest organism, even in a virus, no gene can autonomously replicate itself.

In mitotic cell division, a diversity of enzymes and signaling proteins contribute indispensably to the whole process, each coded for by a different gene and each useless by itself. Helicase, polymerases, cyclin, and CDK are each produced from a distinct gene, and if even one is missing, mitosis fails utterly. Genes do not replicate themselves. At its most fundamental level, reproduction (and survival) is a cooperative effort.

Why, then, is biogenetic theory overwhelmingly focused on explaining the origin of the “first replicator”? Why do we look for this mythical molecule capable of self-replication, when life does not work that way? In Chapter Three I described the cultural biases that motivate and reinforce the selfish gene conception of life. A world of clearly demarcated, competing selves, in which cooperation is incidental and not a necessary or fundamental feature of life, is consistent with the philosophies of the Scientific Revolution and our self-imposed separation from nature and each other. We find it as well in our systems of money, education, medicine, law, and religion. Projecting our own culture onto life, we see it as red in tooth and claw, driven from the outset by a ruthlessly selfish competition to survive. And the subject of this “self”ishness is the gene, the replicator.

Because it is an elegant and parsimonious theory, and because it fits in so well with our culture’s conception of self and world, the neo-Darwinian synthesis has persisted as the dominant paradigm for a long time, despite several formidable problems that have never been resolved. Chief among these is the survivability of intermediates, and the closely related problem of irreducible complexity (IC). At each level of organization, standard evolutionary theory runs up against enormous leaps in complexity that are difficult to reconcile with undirected, random mutation. And the solution points toward not only a different conception of the self that is relationally defined, but also a different understanding of the relative importance of cooperation and competition, a different conception of the nature of life that is not nasty, brutish, and short, a different attitude toward the program of control, and a vastly different understanding of progress, human society, and human relationship.

The first such leap of complexity that evolutionary theory must contend with is, of course, the origin of life. The original version of the problem was the chicken-and-the-egg situation in which DNA requires proteins to catalyze its replication, while proteins can only be produced by DNA. Which came first? A possible solution came with the discovery in 1982 of “ribozymes”: RNA molecules that can perform both catalytic and information-storage functions, implying the possibility of an “RNA world” without the need for proteins. However, this solution only pushes the chicken-and-egg problem into another realm, as the next paragraph describes (you can skip it if you want). Despite these enormous difficulties, research on the “RNA world” remains, in the words of one commentator, a “medium-sized industry” unto itself.[27] Motivating this effort is the conviction that there must have been a “first replicator” somehow, because our entire conception of what life is, what the self is, and how the universe works is dependent on it. It is the original discrete and separate self. RNA world theories are full of all kinds of highly contrived, ad hoc pre-conditions, based on the assumption that “something like this had to have happened.”

First there is the problem of getting an initial prebiotic soup full of β-D-ribonucleotides, about which two leading researchers, Joyce & Orgel, write, “We conclude that the direct synthesis of the nucleosides or nucleotides from prebiotic precursors in reasonable yields and unaccompanied by larger amounts of related molecules could not be achieved by presently known chemical reactions.”[28] Worse yet, even if such synthesis were achieved, there is no plausible mechanism to resolve the left-handed L-isomers from the right-handed D-isomers, which is necessary because the presence of the L version (which does not exist in biology) inhibits the polymerization of the D version.[29] A further problem is the synthesis of polynucleotides with the correct 3′,5′ linkages, which would be in the minority in the absence of some specific catalyst.[30] Thirdly, even if some pre-replication polymerization mechanism existed, there is the difficulty of traversing the enormous search space of possible polynucleotides for one that engages in self-replication, yet without the evolutionary search mechanism of error-prone replication—another chicken-and-egg problem.[31] To give you some idea of this difficulty, an RNA polymerase, which carries out only one of the functions an RNA replicase would need, has been created in the lab; however, it can only perform polymerization of strands three nucleotides long and itself contains a hundred nucleotides, implying a search space on the order of 4100.[32]

In other words, the emergence of even the simplest imaginable replicating molecule is only plausible within a highly complex chemical system—a system that includes molecules generally only produced by living systems. It would seem that life is a prerequisite for life. Since the disproof of “spontaneous generation” in the 19th century, such a maxim has indeed corresponded to our every observation. On the other hand, at all stages of evolution life has undergone sudden leaps in complexity, each of which admits to the same chicken-and-egg explanatory dilemma. The question of the origin of life is a special case of a more general question: what is the origin of complexity, order, and organization?

One theory that tries to explain the origin of life without a “first replicator” is known as complexity theory. Elucidated by Kauffman in his book Origins of Order,[33] this theory is an important step toward a biological paradigm no longer based on our culture’s present conception of self. In describing the origin of life, Kauffman takes advantage of the “order for free” concept described above in the “Order without Design” section. Complex evolving structures appear in many mathematical, physical, and chemical systems in the presence of certain basic conditions such as feedback. Could life be one of them?

The key to Kauffman’s account of biogenesis is the idea of an autocatalytic set. Many of the problems with the standard selfish gene theory come from the absurd unlikelihood of obtaining such a molecule from the building blocks likely to be found in the prebiotic soup. In Kauffman’s theory it is not necessary for a molecule to appear that can catalyze its own formation. All that is necessary is to have a set of molecules each of which catalyze a step in the formation of one or more other molecules in the set. The final step in the formation of each molecule in the set must be catalyzed by another member of the set, a condition called catalytic closure. Based on combinatoric reasoning, Kauffman argues that the emergence of autocatalytic sets is highly probable if not inevitable when molecular diversity crosses a certain concentration threshold.

Let’s look at a simplified example, an autocatalytic loop in which A catalyzes the formation of B, B catalyzes C, C catalyzes D, D catalyzes E, and E catalyzes A. One way to look at it is to say that A is the replicator and uses B, C, D, and E as tools to achieve replication. However, that is an arbitrary designation, because we could say the same thing about any other member of the set. Each is dependent on the others.

Figure 3: An autocatalytic loop

An autocatalytic loop like this lies at the heart of Kauffman’s autocatalytic sets, but it gets far more complicated than that. Imagine a system of multiple loops and chains, loops within loops, mutual cross-feed relationships connecting them, inhibitory connections, preferential reactions given different substrate concentrations . . . very soon the picture starts looking very much like a metabolism or an ecosystem. There is still no unequivocally identifiable unit that might be said to be the replicator, but we may impose somewhat arbitrary boundaries on various subsystems within the system and call these parts alive, recognizing that while they might be dependent for certain of their reagents on other subsystems, they are able to maintain a constant internal environment.

Figure 4: A very simple autocatalytic set. Each node represents a ligation/cleavage reaction comprising three elements. The dotted lines represent the catalytic action of a fourth element.

Figure 5: Map of protein-protein interactions. Left-hand diagram by Hawoong Jeong, right-hand diagram by Erzsebet Ravasz.[34]

The resemblance of autocatalytic sets to metabolisms and ecosystems is surely not coincidental. When ecologists draw a nodes-and-arrows graph of the interdependencies among species, it looks like a vast elaboration of the autocatalytic set pictured above. Moreover, as Kauffman points out, the modern cell is an autocatalytic set in which DNA, RNA, proteins, and various intermediate products all contribute to each others’ synthesis. Just as in our autocatalytic loop above, it is arbitrary to say that the DNA is the replicator and the RNA, proteins, etc. just its tools. Instead of saying that our genes act collectively, by coding for proteins, to replicate themselves, why not say that the proteins act collectively, through catalyzing the chemical steps in DNA transcription and RNA translation, to replicate themselves?

Stuart Kauffman suggests a self that is cooperative at its very origin, yet his model still carries a subtle projection of separation. Is a cell really autocatalytic? What about a human being? No. At best we can say that each contains autocatalytic systems and systems-within-systems. Each requires a “food set” of molecules that it cannot produce itself. A human being cannot produce sugar from sunlight, nor the free molecular oxygen our metabolism requires, nor a number of essential amino acids, fatty acids, and vitamins. These substances and their sources are therefore excluded from self.

When we wall off part of an autocatalytic set to define an organism, we recognized a certain arbitrariness to that definition. We could just as easily have named the whole as the unit of life, or a smaller part, just as we could say in a human that our organs are alive, our cells are alive, our mitochondria are alive. True, none are viable on their own, but neither are human beings. Nor is any life form, at any level.

The biological definition of the organism, and therefore of the self, usually draws on the concept of the phenotype: the “expression” of the nuclear DNA. (Already this definition is problematic because DNA is expressed very differently under different environmental conditions.) However, the resulting organism is often no more viable—that is, no more capable of survival and replication—than an isolated human organ or cell. Most life forms are so utterly dependent on symbiotic relationships with other life forms as to call into question the validity of the phenotypic definition. Without the bacteria in their rumens, for instance, cows would be unable to digest cellulose and would quickly starve. Is the bacteria part of the cow, or a separate organism? Or is our identification of discrete organisms as the atomic elements of an ecosystem a source of confusion?

It may very well be that contrary to the selfish gene theory, in which the organism came first and eventually mutated, evolved, and generated the ecosystem, it is in fact the ecosystem that is primary, while organisms are merely semi-autonomous off-buddings that arose after the ecosystem developed to a certain level of complexity. If so, the quest to find the first “replicator” is a diversion, an artifact of a cultural prejudice about the nature of the self as a discrete, independently existing entity. If our boundaries of self were more fluid, as they were in other cultures, more open, less rigid in their demarcation of the universe into self and other, then perhaps we would not be so fixated on finding the replicator and better able to understand a different genesis story.

Kauffman’s work suggests a story of life based much more on cooperation than on competition. In the simple example of the autocatalytic loop, each element is indispensable to the viability of the whole. Remove one and the whole system disintegrates. A more complex system such as an actual ecosystem is more robust—remove C, and there will be other pathways to get from B to D, or if not, from B to E—but the loss of one element, or one species, will typically still initiate a cascade of other losses causing the entire system to collapse into a simplified, but still viable, subset of the original. The system is cooperative in a another way, too: the whole is greater than the sum of the parts. Split an autocatalytic set into two pieces, and both may be non-autocatalytic and hence incapable of maintaining themselves away from chemical equilibrium. Autocatalysis, and therefore life, is an emergent property arising out of complexity. Life on all levels is a collective.

This is an example—quite literally a living example—of the limits of reductionistic logic. As in any system with emergent order, there is something about the whole that eludes analysis, that cannot be understood by taking it apart. Poets have long known that something of a flower is lost when it is reduced to just so many botanical properties of stamen, sepal, anther, and petal, just as life is more than the collection of enzymes, fatty acids, DNA, proteins and so forth that make it up. The ideology of science has long been that higher-order properties, such as beauty, meaning, and love, are either human projections that are not really there, or simply collective terms for a number of lower-order “real” phenomena. What is happiness, really? Merely a set of biochemically determinate conditions: hormone levels, neurotransmitter levels, activation of brain regions, etc. Today, however, we can demonstrate mathematically that irreducible higher-order properties exist, and we can see them as well in living systems.

But my dear reader, when I say “life is more than a collection of enzymes, fatty acids. . .” please don’t think I am proposing to add yet one more ingredient, an immaterial spirit to inhabit and animate the body. The truth is far more marvelous. I do not believe in an immaterial spirit, but I do believe in a material spirit! Spirit is not separate from matter, it is an emergent property of matter. On the one hand, yes, there is nothing more than the masses and forces of physics and chemistry,[35] but that does not mean emergent higher-order phenomena are unreal. Spirit is just as real as Langton’s ant highways. It has real effects and real explanatory power, but when we take things apart to look for it, it isn’t there. The soul is not just another ingredient in a living being, which is why experiments purporting to prove the soul’s existence by weighing the body before and after death are misguided.[36] That doesn’t mean it isn’t real though! The same goes for other emergent phenomena like happiness, consciousness, beauty, and selfhood, which puts them forever outside traditional paradigms of engineering and control. As in any complex, non-linear, feedback-ridden system, monkeying with the parts can have unexpected or even contrary effects on the whole, which is why attempts to “manage” ecosystems are so problematic.

A technology of wholes rather than parts will look very different from what we have today, running contrary to the scientific intuitions of the last several centuries. If posed with an herb that has a healing effect, science would take it apart chemically in search of the “active ingredient”. Happiness was to be explained not as a holistic state of the entire person, but isolated as a limited set of neurotransmitter levels, hormone levels, and related responses. Consciousness was similarly to be identified in a certain part of the brain, the “seat of consciousness”. Soil fertility was to be explained reductionistically as well, in terms of percentages of a finite list of minerals and other ingredients. Science was based on distilling, purifying, extracting the active principal and separating out the dross.

Closely related to this goal was the program of control. For once we had grasped the active ingredient in pure form, we could wield it to manipulate reality to our liking. We could make the soil fertile, we could make the patient happy, we could make a medicine in pure form more precisely effective than the original herb containing it. Future technology will emerge from our understanding of all the things that vanish when you take them apart, and that are real nonetheless. It will be a technology of the emergent.

When René Descartes pondered the nature of the self, he took the very same approach of isolating the essential principle. In fact it is our concept of self, enunciated so clearly by Descartes, which has set the template for our investigations into the world as well. Separating out the dross, Descartes arrived at what he believed to be the essential, purified kernel of selfhood, a mote of am-ness separate from the body, the emotions, the sense-impressions, and the thoughts, but gazing down upon them, experiencing them but apart from them. Descartes’ self is the audience for what Daniel Dennett calls the “Cartesian theater”, viewing the play of thoughts, experiences, sense data, and emotions on the stage of the brain.

But like the original replicator, like the vital principle of life, when we take the self apart to find its essence, we discover that it too is not there. This has certainly been the case in neuroscience, which has found that properties such as consciousness and memory are not localized anywhere in the brain. Buddhism has arrived at an identical finding, that the self has no objective, discrete reality, but emerges from relationships spanning the entire cosmos; there are no separate individuals, all are interconnected, interdependent, interdefined.

The vision of life, the organism, and therefore the self as an arbitrarily-bounded open subsystem itself composed of numerous interdependent sub-subsystems, and therefore without a discrete objective reality, conflicts with many of the founding assumptions of modern philosophy, medicine, economics, religion, law, and psychology. The new conception of self will give birth to momentous changes in all these areas. The Cartesian self is fundamental to the dualism that informs so much of modern thought: I think, therefore I am. Descartes believed in an irreducible kernel of selfhood or am-ness that is the true “I”, discrete and separate. That conception of self is wholly consistent with the intuitions of the Age of Separation. Now we are discovering that at the very basis of life, no such self exists, and we shall observe that fact again and again, level after level, through the cell, the organism, the ecosystem, and the whole planet. The progressive alienation of ourselves from the community of life, our progressive distancing from nature, is based ultimately on an illusion. The Scientific Revolution gave voice to this illusion; today science is undermining its very foundations.


[27] I did the research for this section several years ago. Now it seems that the RNA world is falling rapidly out of favor.


[28] Joyce, Gerald F. and Leslie E. Orgel, “Prospects for Understanding the Origin of the RNA World,” from The RNA World, Second Edition, Cold Spring Harbor Laboratory Press, 1999. p. 68


[29] Joyce, G.F., Visser G.M., van Boeckel C.A.A., van Boom J.H., Orgel L.E., and van Westrenen J. “Chiral selection in poly(C)-directed synthesis of oligo(G)”. Nature , vol. 310, 1984, pp. 602-604.


[30] Joyce and Orgel, p. 51


[31] Joyce and Orgel, p. 62


[32] Bartel, David P. “Recreating an RNA Replicase.” The RNA World, p. 143-159


[33] Kauffman’s more recent book, At Home in the Universe, will be more accessible to the lay reader. It applies the same concepts far beyond biogenesis and evolution, and is an excellent resource to help develop non-dualistic intuitions about the origin of order and beauty in the universe.


[34] Displayed at


[35] Also keep in mind that even in the world of masses and forces, there is much we are ignorant of.


[36] The emergent property of soul might be associated with a bound energy that contributes to the mass of a living being. Death brings a decoherence of innumerable processes of life, an enormous loss of embodied information and energy, and thus an equivalent loss of mass. I realize that standard physics admits no way for the mass—measured at several ounces—to just disappear without being converted to energy on the order of 9×10^16 joules. Where does it go? I’m not going to try to answer that question right now.