One of the most mind-expanding books that you’ll ever read is Evolution in Four Dimensions by Eva Jablonka and Marion Lamb. They remind us that evolution is about variation, selection, and heredity, not genes. Genes provide one mechanism of heredity but there are others, including epigenetic mechanisms, forms of social learning found in many species, and forms of symbolic thought that are distinctively human. They provide a concise history of why evolutionary theory became so gene-centric during the 20th Century and how it needs to be expanded to include the other three dimensions.
Eva Jablonka is a Professor at the Cohn Institute for the History of Philosophy of Science and Ideas at Tel Aviv University in Israel. I talked with her by Skype on November 6 2014. Our conversation provides a panoramic tour of evolutionary theory based on heredity, not just genes.DSW: Welcome, Eva. I’m so pleased to be talking with you.EJ: Hello, David.DSW: I want to talk to you about the definition of evolution and the need for it to go beyond genetic evolution. This is the topic of your great book, Evolution in Four Dimensions, which I have adopted as the first text for almost all of my courses. That’s how much I think of it. Let’s begin by discussing your background. What is your training that enables you to write such a book?EJ: I am a geneticist. I did a PhD in genetics and molecular biology; in fact, on DNA methylation and chromatin structure. Before that, I did a Masters thesis in microbiology. At the same time, I was deeply interested in philosophy of biology. While I was doing a PhD in genetics, I was also writing papers for philosophy of biology journals. I thought that I should combine the two because theoretical biology and evolutionary biology need a very strong conceptual basis. I ended up being in some kind of twilight zone between the two things. For me it was a productive combination.DSW: Great! Everyone knows that Darwin knew nothing about genes. For him, evolution was about variation, selection and heredity, a resemblance between parents and offspring. Nevertheless, nowadays, whenever you say the word “evolution,” most people hear “genes.” That’s true for a professional evolutionist, as much as for the lay public. How is it that the study of evolution became gene-centric?EJ: It is related to the strong focus on heredity that is apparent already in the second 19th century, when many theories of heredity were developed. Once evolution became an accepted theory it was clear that one has to think very seriously about heredity. In order to have cumulative evolution, heredity is necessary. Darwin himself had a theory of heredity, which was, in fact, one of the most Lamarckian theories of heredity around at the time! The point is, however, that he needed a theory of heredity to consolidate his theory of evolution, and he did develop one.The other reason heredity became focal was because of the Industrial Revolution. The population was growing and there was an urgent need to feed people so improvements in agriculture became pertinent. It was clear that breeding and selection were of great importance, and selection must be based on heritable variation. The study of heritable variation was therefore important from a practical point of view.So there was a huge focus on heredity at that period, and once Mendel’s laws were re-discovered and accepted – it took some time – people equated biological heredity with his hereditary element, the genes. The genes were, at first, just computational entities. Nobody knew what they were. The person who invented the term “gene,” [Wilhelm] Johannsen, did not think about genes as little particles. He thought about them as processes. He was an Aristotelian, and the distinction between potential and actualization was important for him. The genotype was the (hereditary) potential and the phenotype was the actualization. That’s why when he wrote his famous paper about the gene concept, heredity and the genotype-phenotype distinction, he said that his theory about the gene is ahistorical. Development doesn’t matter to heredity. That was a view that was very quickly accepted. It was compatible with the views of [August] Weismann in the 19th century, who was very much against the idea that acquired, developmental somatic variations can be inherited. Weismann argued that only variations that impinge on the germ-line directly can be inherited.DSW: Let’s fast forward to the new synthesis. You said that the new synthesis is as interesting for what it left out as for what it kept in. Let’s speak on that theme.EJ: Many things were excluded, but let me talk here about just two. First, the modern synthesis assumed that the only inherited variations are genetic. Everything else is not really inherited in a biological sense. Therefore, for heredity, all we need to be interested in are genes. But although it is true that genes are very important for heredity, the architects of the synthesis excluded other hereditary factors.The second thing they excluded is the possibility that development can affect heredity. They said that the hereditary variation, whatever its origin, cannot be influenced by the developmental history of the organism or its ancestors. In this sense variation is functionally random.DSW: They also, in some ways, excluded the whole developmental process, as if you could map genetic variation onto phenotypic variation and ignore the developmental process that maps one onto the other. Could you elaborate on how that happened?EJ: There were many reasons for that, conceptual, social-political and experimental. There were experiments that showed that you cannot have selection in pure lines. Johannsen’s experiments, for example. He showed that in pure lines there are phenotypic variations, but selection is ineffective. You cannot select the extremes, for example, and make the tall plants taller over generations through selection. There were, however, at the same time, some experiments that did show selection in pure lines, but they were not very consistent, and people didn’t know how to interpret them.Other experiment that showed the effect of the environment on organisms, were also ignored. Many of these experiments were published in very good journals but there was no interpretative framework. Unfortunately, some of these experiments were done by charlatans, we always have frauds, sadly, in all domains of science. When the charlatans were exposed, not just the specific experiments, but the whole area became tainted. In the Soviet Union, before and especially during the Cold War, there was this horrible man, Trofim Lysenko, who was a charlatan and very much against Mendelian genetics. He regarded Mendelism a bourgeois aberration. He worked within the Stalinist framework and he caused a lot of harm to Russian genetics. Because of him, the whole tradition that developed in the U.S.S.R. was dismissed as nonsense in the West. But it is wrong to assume that this was indeed the case, that every Russian scientist who discovered that development impinged on heredity was a charlatan.Some Russian scientists thought that stress, for example, induces heritable variation. They did good experiments but their work was ignored because they became associated with Lysenko, the anti-Mendelian tradition, and the Cold War with its strong anti-Soviet prejudices. They had great tradition of good work in the U.S.S.R., and there were a lot of honest people who worked within a very different paradigm from that then current in the West. Their paradigm was that the hereditary processes are part of the physiology of the cell, and that they are responsive to developmental changes. They did ingenious experiments in aphids, for example, and in silver foxes that showed phenomena that are best interpreted in terms of epigenetic inheritance.The hereditary variation they found was difficult to classify. It was not completely random, and it was not completely directed. People didn’t have concepts for those kinds of things, and certainly, they didn’t have mechanisms. This was important because once we had molecular biology and an understanding of how DNA is transcribed into RNA, how RNA translated into proteins and especially once we had the central dogma, it seems that developmental changes, which were assumed to affect proteins, cannot be inherited.There was no molecular mechanism to describe the developmental effects that the Russians and other scientists (in the West) described. Biologists didn’t have a clue how this worked. Maybe, many thought, it’s all an artifact or can be explained in terms of genetic variations. It is not difficult to interpret every complex pattern of heredity in terms of genes. [Thomas Hunt] Morgan once said, “Give me six genes, with a few epistatic interactions, and I’ll explain everything to you.” Every result, can be so explained if you have enough genetic factors and enough interactions among them.DSW: This is a good opportunity to move to the second dimension of evolution, epigenetics. Now we do have mechanisms, and we can begin to understand how epigenetics counts as an inheritance system. Maybe you could talk about that a little.EJ: Yes. It’s difficult to decide when it started, but I think modern epigenetics started in the 1980s. It began with the intensification of the study of DNA methylation and the understanding that DNA methylation has a dual role. It is part of the regulatory system and it is a cellular heredity system. DNA methylation is commonly involved in the repression of genes: when there is a lot of methylation in promoter regions there is often transcriptional silencing. On the other hand, it is part of cell memory, so somehow, the patterns of methylation on DNA are reconstructed during cell division. Methylation has these two aspects, developmental and hereditary.Once you realize that there are these two aspects in development, within the organism, you can think, “Well…, maybe sometimes such variations are also inherited between generations?” This was the insight that Marion and I had. We said, “Well, if you have these mechanisms, is it really true that everything is deleted, completely, during gametogenesis? Biology doesn’t work this way. It doesn’t make sense.” If germ-line cells somehow change their methylation patterns because of certain developmental signals that occurred during development and if these epigenetic marks can be inherited, then we can have the inheritance of developmental variation. This kind of inheritance, we argued, can be adaptive in some types of environment, so it is likely to have evolved.We reasoned that if this is true we must find evidence for it in the literature. So we went to the old literature of genetics, and we started digging and looking at results that were not explained. We found a lot of those, discovered by very good geneticists, who published in Nature, Science, in the best journals. But they were forgotten sooner or later, at best were mentioned as interesting aberrations. We showed that we can reinterpret these results, not all of them, but many of them, in terms of epigenetic inheritance. We focused on DNA methylation that was then the best understood epigenetic system, although we were aware that other mechanisms are possible.I want to say a few words about Robin Holliday, who passed away some time ago. He was a great pioneer, a great inspiration for us. He wrote a paper, with John Pugh, about DNA methylation. He and Pugh, and independently Art Riggs, realized that DNA methylation is an inheritance system—a within-organism inheritance system. That was a great insight, and Holliday started working on it, trying to understand a phenomenon such as cancer and aging. Maybe aging is an epigenetic inheritance phenomenon, with epigenetic mistakes inherited within cells, accumulating, and leading to the catastrophe of aging and, eventually, death. People picked this up and started to work on these aspects of within-organism epigenetic inheritance. This didn’t threaten the modern synthesis view of evolution because they thought about within organism cell heredity, not about trans-generational heredity.Marion and I started to take it outside ontogeny into heredity and evolution. We were basing our initial thoughts on the biological logic of the system, on what we knew about development and gametogenesis. Soon, our conjecture was supported by experiments especially on transgenic organisms, in plants and mice that showed the inheritance of DNA methylation patterns. There were initially very few of those. When we were developing the idea, it was sometime in 1987, there was very little evidence at the molecular level.DSW: If I understand it correctly, DNA methylation is not common in insects and other non-mammals. Is that right? As important as it is as a mechanism of epigenetics….EJ: It’s not the only epigenetic inheritance system. We focused on methylation because this is how the study of modern epigenetics started. The fruit fly Drosophila has very little methylation. There is methylation in other insects – in bees, for example, you have much more methylation than in Drosophila. It was clear from the outset that there are several important epigenetic inheritance mechanisms, not just one . It’s not like DNA replication, which is universal. With epigenetic inheritance, you have a whole family of mechanisms. DNA methylation is one of them, but there are many others, and they may play different roles in different groups of organisms, and some may even be lost in some groups. An important epigenetic inheritance mechanisms, that was fully characterize only after 1998 and is now being investigated very intensely, is the one that is mediated by small RNAs. It is responsible for extensive epigenetic inheritance in the nematode worm C. elegnas, for exmaple.What is interesting is that if you are looking at DNA methylation, modifications of histones (which also can be reconstructed within cell lineages) and the small RNA regulatory system, they are functionally linked. They form, together, a very complex system that we don’t understand fully, as yet. We’re getting there.DSW: Let’s quickly go to the third and fourth dimensions of evolution, to get the whole package on the table. That would be forms of social learning found in many species, and forms of symbolic thought that are distinctively human. I won’t say “uniquely” human. I don’t care about uniqueness. We can say distinctively human. Tell us about those two mechanisms of inheritance, and then we can go on from there.EJ: Once you start thinking that there are several ways of transmitting variations from one generation to the next, to think about heritable variation in different ways, then you start thinking about additional routes, for example, about information that is transmitted through social learning. As we know, in many animal species, there are such traditions. A lot of people say, “Yes, that’s true, but it’s not quite biology, is it?" But if it is not biology what is it? Many people are very surprised to hear that rats can learn through social learning, for example. But even with rats' social learning they still are not sure that this transfer of information is, really, biological.This gets worse when you’re talking about human culture or symbolic culture, that for many people belongs to a totally different sphere. In some ways, symbolic cultural transmission is, indeed, very different from other forms of information transfer. In fact, each of the systems of inheritance has its own distinct features. For example, with behavior and symbolic evolution, you don’t have a big material overlap between generations. When a cell divides, the daughter cell is part of the mother cell. The transmitted information has material continuity. However, when you are thinking about transmission through behavioral means such as imitation, or through symbolic means, you don’t have this material overlap. This has many important consequences.Because there are genuine and important differences between different types of information transmission, people tend to treat them as if they belonged to completely different worlds and completely different types of discourse. Cultural evolution in animals was not considered real evolution, and cultural evolution in humans, symbolic evolution in humans, wasn’t evolution, it was history, and history is assumed to be something else altogether.But in a cultural symbolic system the distinctions between development, learning, heredity and evolution becomes very difficult. History and evolution are continuous. It is a very complex system, the most complex because one has to think about all four dimensions of heredity, but this doesn’t mean that culture is not part of the biological realm.DSW: No. B. F. Skinner had a classic paper, titled Selection by Consequences in Science. He put his finger on why it is that we need to think of all these things in the same way. It’s the same way that Darwin did; that once you combine variation, selection and heredity, then there’s this plasticity to these entities, which cause them to be molded by their consequences. This is basically what adaptationism is, for genetic adaptationism. The same thing holds, Skinner said, for the plasticity of individual learning. We can say the same for culture. It’s at that level that these things all are a common inheritance system. Of course, in their mechanisms, they become very different, and those differences make a difference. These inheritance systems, as you are careful to point out, are very different from each other, in some respects. Nevertheless, they allow organisms to adapt to their environments, and you get that core part of evolution – adaptationism, basically – that’s open-ended adaptation to change. Let’s talk about what distinguishes symbolic thought from forms of learning found in many species.EJ: Things can be learned in very different ways. When you have symbolic representation and transmission, you have the ability to combine things that do not belong just to the here and now. This is one of the great things that symbols allow you to do. They allow you to think about futures and about imaginary or non-imaginary pasts. You can engineer the future, if you want, through this kind of symbolic communication and representation. You can’t do it with social learning that is not symbol-based. Regular animal social learning is still largely in the here and now. The variation in the symbolic variation is very stable and allows multiple combinations, and the formation of multiple scenarios. Moreover, there are many different ways of symbolizing, and you can combine these ways of symbolizing, too.DSW: Let me play it back in my own words. It is that symbolic thought has the kind of combinatorial diversity that genetically combination has?EJ: Absolutely, it does.DSW: That is different than both epigenetics and forms of social learning. Yes, they are inheritance mechanisms, but they don’t have the combinatorial diversity that genetically combination has and that the recombination of elements of symbolic thought have. Is that a good representation of your views?EJ: I think there are far less constraints on both genetic and symbolic variations, but I think that there are many possibilities for epigenetic and behavioral heritable variations too, more than we usually think. I do agree with you, however, that combinatorial variation is inherent in DNA and in the digital structure of this polymer, and it is also inherent in the symbolic system, which is made up of representations that can be decontextualized. If I say “dog,” I can relate to a dog that I saw, a dog that I will see, a dog I love, a dog I hate; a dog I never saw, a pink little dog.DSW: I can also say, “You’re a dog,” and that will be transferred, basically.EJ: Exactly, and this decontextualization of the single word, this metaphoric usages, allows these enormous combinatorial possibilities. The number of meaningful expressions that you can generate is unlimited.DSW: I want to talk about aspects of symbolic thought that are exceptionally guided. That would include such things as flat-out, intentional planning. We want to do something, and then we choose among alternatives to do that. Also, the scenario building. Human thought is often very future-oriented. For some people, I think this disqualifies them from being evolutionary. I disagree with that. I think that intentional decision-making, for example, is explicitly evolutionary. It’s an explicit variation and selection process. There is also a chance element that goes into what you consider for what you’re trying to do. These forward-directed, intentional elements of human thought, I think, fit within the evolutionary paradigm. They count as evolutionary, for example, but I think other people tend to disagree and want to make some kind of dividing line between randomness and guidedness. I wonder if you could talk about that. This goes back to Lamarckian forms of inheritance and genetic evolution. Let’s first talk about it in the realm of symbolic thought. What is the role of guided processes, and does that make them non-evolutionary? Or do they remain evolutionary?EJ: If you think that evolution is a change in the constitution of heritable types in a population, why not? Types do not have to be genotypes, they can be epigenotypes, behavioral types, symbolic types. If we agree on that broad definition, I don’t see why change over time with respect to them is not evolution. Symbolic evolution is a very different type of evolution, among other things it is a developmental system that can direct its own evolution. But this does not make it non-evolutionary.This is not confined to cultural evolution. We need to recall that we have learned that cells are very clever. They have a genetic-engineering kit within each cell, which it can use in all kinds of circumstances. Evolution has led to very clever and sophisticated systems, including systems that can direct their own evolution. There’s nothing wrong with it, and we shouldn’t be surprised at it. In fact, we would be surprised if it didn’t happen, in the long term. I don’t see why it is not evolution. I think that the notion of evolution most people have is very gene-centric. This version may be simple, but it far too limiting. It cannot account for actual evolutionary changes satisfactorily.DSW: That’s how I want to finish up our conversation. There is a lot of shyness in embracing this larger view of evolution. There are different forms of shyness, different forms of reticence. As my experience, I think it’s also yours, that some of our biological colleagues who are doing state-of-the-art work on biological evolution are very reluctant to expand the definition beyond genetic evolution. Even if they can expand it to include epigenetics, then they’re reluctant to include these two other forms, social learning and symbolic thought.On the other hand, our colleagues in the human-related disciplines, both the humanities and the human social sciences, often, historically, have defined their fields to be apart from biology. Sociology is often defined in a way to be apart from even psychology, not to speak of biology. We have these historical trends, and then we also have the problem that, often, the big complaint about evolution is that it’s biologically deterministic and so on. There’s a reticence on that end to think about such things as symbolic thought as within the orbit of evolutionary science. What are your own feelings about those sorts of things?EJ: I agree with you. I belong to a generation that had to fight very hard to make people accept that the notion of heredity must be extended, and if so, this must also apply to the notion of evolution. Once you have heritable variation of different types, you need think about how these variations and the mechanisms of transmission underlying them work, and how they interact in the real world. You have to think in a very complex way. I am still engaged in arguing with people about that, trying to make them go beyond the gene-centered paradigm on which, I, too, was brought up.But things are changing. A young colleague of mine told me that maybe I’m a dinosaur, and there is no need for these conceptual arguments. She’s a young and very bright epigeneticist. She said, “Oh, it’s not a problem. It’s just an empirical question. The question is: what exactly is the mechanism? How are things inherited through epigenetic mechanisms, to what extent and conditions, and in what organisms? How do different mechanisms of epigenetic inheritance, methylation, small RNA’s, etc. interact? This is what we have to figure out. It’s not a conceptual problem anymore.” Maybe she is right.DSW: We hope she’s right. We’d love to be dinosaurs.EJ: Yes, I too hope she’s right.DSW: How about on the human side, people that are steeped in culture? They appreciate the cultural diversity. They appreciate the fact that who we are is very much a matter of the symbols that we carry, but are unaccustomed to thinking about this as an evolutionary process and, in fact, tend to think about evolution as just genes. Could you have that kind of conversation?EJ: There was a very strong reactions against the genetic-based view of evolution by people from the social sciences, who feel that the geneticists' view is irrelevant to their own work, to their concerns. If it’s irrelevant then all one can say is: “You do your job, we’ll do our job. We don’t have any common ground.”But if you accept that there is this complexity, that there is reconstruction of phenotypes between generations and this reconstruction uses many developmental resources, the genetic resources, the epigenetic resources and the cultural resources, and all can make a difference to heredity both directly and indirectly, the picture may change. Importantly, since all these resources interact, and they form a system. This system has its own dynamics – its own stability and its own opportunities for change. Once we understand how the system is stabilized, understand the feedback interactions, then we can also understand how one can change it. If you present things in this way to social scientists – at least, that was my experience – they can accept it, and we can work together. If you tell social scientists, as some biologists do, that culture is just a very thin veneer on what really matters for human nature, they will say, “Well, thank you, very much. We go our way. You go your way.”DSW: Yes, really taking the concept of cultural evolution seriously is part of what this all about. That’s no veneer.EJ: No, it’s a very deep, complicated and beautiful process. As you say, it has this future-oriented aspect to it, which is very unique to human, as far as we know, to our symbolic representation and communication system. It makes cultural evolution and cultural history very different from animal culture.DSW: Here’s my final question, is there any meaningful distinction between biology and non-biology?EJ: It’s a very difficult question. We all make distinctions. We made a distinction between physics and biology. We make distinctions between biology and psychology. We all make distinctions between different levels of description of phenomena. That’s alright. From this point of view, you have to make distinctions because you describe things at very different levels. If you are somebody who describes things at the molecular level, then it is good to be clear about your domain of research. This is a legitimate kind of distinction.When we study something like culture, it, too, has its own level of description. But in order to understand this level of description, we have to understand how it is constructed. Biological resources are part of what matters here, so you cannot exclude biology. There is something that I call the “biology of culture,” but it doesn’t reduce culture to biology. It does, however, incorporate biological factors and biological interactions into the fabric of the cultural system. It mean that biological resources are all important. Sometimes, genetic differences are not important for a particular cultural phenomena, while for other cases, they may be quite important.DSW: You said, earlier, that epigenetics had its breakthrough when it became mechanistically well-understood. Everything has a mechanism. Everything has a physical mechanism. When learning and symbolic thought, as the physical mechanism, is better understood – that would neurobiologically and so on, developmentally – then maybe there will be a breakthrough there, in terms of seeing it, more clearly, as an inheritance system.EJ: I think it will help, but I worry about trying to reduce cultural phenomena just to that, just to the neurobiological mechanisms. This is a danger because there is a cultural level of description. Culture is indeed a system of interactions at many, many different levels, that crucially involve neurobiology, but the system is more than sum of its parts and requires its own level of analysis.DSW: Now, that’s where the distinction between ultimate and proximate causation comes in very handy. Not just a proximate description, that’s never adequate for any evolutionary explanation. There will always be a functional explanation, in addition to a mechanistic explanation.EJ: Yes. But I think the distinction between ultimate and proximate, when you’re thinking hereditary variation that are developmentally or culturally constructed is breaking down a bit. Cultural evolution is both proximate and ultimate.DSW: Also, there are levels of description that don’t go all the way down to the neurobiological level. On some of our own work, we’re thinking about, for example, of sacred texts as like a cultural genome. Why is that? There’s high-fidelity replication. Sacred texts are taken seriously. They tend to translate into action. They have many components to them that are deferentially invoked in response to the environment, so that different parts of the Bible, for example, are expressed under different circumstances. You have an expression and a phenotypic consequence kind of thing. That’s a mechanistic description, which is still not all the way down, but we’re finding it very useful to be thinking about culture that explicitly, as an inheritance system, as an evolutionary process.EJ: It’s very interesting to think about the evolution of these texts in this way. We also want to understand how is it that these texts become sacred in the first place. What made them sacred, and how is sacredness maintained? Going back to our DNA analogy, DNA is not just a very stable molecule. In fact it is not very stable at all. It requires all the machinery of the cell to make it stable, to edit away and to get rid of mistakes. It’s the same about the sacred text, there are social-cultural processes that maintain the sacredness of the text or allow changes in it. That’s a very important part of what we need to understand. A good example, are the Mormons. They are very good at modifying things, according to circumstances, and maintaining the core of their religion.DSW: That is an excellent example of this evolutionary thinking applied to our culture. To that, I say, “Amen.” That can be the end of our wonderful conversation.EJ: Thank you, very much.DSW: Thank you!
