The mechanisms of speciation, applied in a cultural context, might not only help better understand how cultures form, but also provide means for designing and constructing culture. In this essay we focus on reproductive isolation, a subtopic of speciation, and try to broadly look at similarities and differences in isolation between species and cultures, making some conjectures about what kinds of conditions are favorable for cultures to “speciate”.

The application of biological ideas, especially evolutionary biology, to sociology (sociobiology) and the concept of dual inheritance theory (1,2) have aimed in this direction, but for reasons we speculate on later, neither has looked much at how biological models of speciation can inform sociological models of cultural genesis. Instead the focus has been on other evolutionary mechanisms (but see (3) for an exception). One way of imagining species and cultures is as groups of individuals who are more similar to each other than to other individuals in other cultures or species. The similarities of the individuals can be genetic, phenotypic and behavioral in biology; and demographic, behavioral, and memetic in sociology.

In evolutionary biology (and we suggest culturally), reproductive isolation is when a group of individuals exchange extreme amounts of information amongst themselves relative to how much they exchange outside their group (4). Analogous to the exchange of DNA among individuals in sexually reproducing species, the spread or flow of ideas/attitudes and behaviors among individuals in cultures represents how cultures change and evolve. One might expect to see reproductive isolation or its analog in terms of information exchange in cultures as in species. What we seem to find across cultures throughout history is a lot of variation in the degree of informational isolation, with higher levels of isolation associated with successful cultural maintenance(2). It is less obvious whether the same holds for cultural genesis.

For example, the origins of Christian culture have varying degrees of isolation, from the Western Christians who merged with Roman culture quickly, to the Eastern Christians who maintained isolation for quite awhile (5). The Protestant Reformation was aided by the Ottoman Empire, which provided some indirect isolation from the Catholic powers of the time (6). Buddhist cultures emerged from monasteries that had some isolation from the rest of the world, but also much interaction with it (7). Anabaptists (ex:the Amish) have emerged with a large amount of isolation that they kept even after their genesis (8). Further, many examples exist of hybridization of cultures to form new (heritably stable) cultures, but that is not what we will focus on here. Historians have already considered this mechanism of cultural genesis (9).

What can evolutionary sociologists learn from evolutionary biologists through understanding reproductive isolation, especially in the genesis of new cultures? In biology, isolation stands as a precondition for generating a new species. In cultures, it seems to be much less important. The question is why? Cultures are more analogous to breeds, or varieties in biology (2), which might explain why cultures are able to exchange information and hybridize. It still does not explain why information isolation might be less important for cultures to speciate, because the creation of new breeds in the biological world still requires reproductive isolation.

To address this, we’ll need to look deeper into information isolation in cultural genesis as a consequence of more fundamental mechanisms, taking our cues from genetic speciation. There are three detailed mathematical models of genetic speciation, all of which have some empirical support (10,11,12). To understand these theories and how they might apply to cultural genesis, it is necessary to introduce the concept of adaptive landscapes, invented by the evolutionary biologist S. Wright almost 100 years ago.

The adaptive landscape is a function from either gene space, phenotype space, or to make a connection with evolutionary game theory, heritable strategy space, to the real numbers, with its value representing fitness (or equivalently the rate of growth of populations of individuals or groups (11)). If we model two strategies, this has a nice visual representation as a topographic landscape, with the height representing fitness, and movement in the other two dimensions representing the proportion of individuals adopting one or both strategies. This concept has been generalized so that the adaptive landscape is dependent not only on the strategy of an individual (or mean strategy of a species), but on the strategies of other individuals within a species and other interacting species in an ecosystem. It is not static but can change with time, as new species come into being, grow and sometimes become extinct, and as resources change (11).

The first model of speciation (which we will necessarily simplify for space constraints), the shifting balance theory, is very similar to theories of phase transitions in physics. In this theory, reproductive isolation of a small group (the incipient species) with one or several mutations from a mother species is necessary for speciation. Isolation and small group size (Mayr coined the term “founder effect”, which has other consequences we will not get into) are needed in order to produce a large enough genetic drift away from the local fitness maximum of the mother species. This genetic drift is akin to Brownian motion, or phase-space diffusion in phase transitions (unfortunately, genetic selection is akin to what is known in physics as drift, a source of confusion). Genetic drift allows the incipient species to first move downhill in the adaptive landscape (representing reduced fitness relative to the mother species), against selective forces, and then through a “mountainpass” (saddle-point), where selection can take over again and lead the incipient species to another maximum where it becomes a bona-fide species. In practice, reaching a saddle point through drift is far from guaranteed, drift more often than not does not result in speciation. The smallness of the initial group is necessary because the drift is inversely proportional to group size (13). Too big of a group and the mutation leading away from the mother species may be lost. On the other hand, smallness of the group is no guarantee of fixing the mutation, as drift could just as likely go towards losing the mutation. Similarly, the reproductive isolation from the mother species is helpful (and becomes necessary the bigger the size of the mother species population) because without it, the mutation(s) may become lost, swamped (through the mechanisms of genetic drift and gene flow from mother culture) by the non-mutated allele in the mother species. The shifting balance mechanism is consistent with what is called sympatric speciation.

In the second and third models, the adaptive landscape also changes while the genetics, phenotypes, or heritable strategies change. Similar to the shifting balance theory above, both these theories involve some form of reproductive isolation for a new species to arise. In the first, competitive speciation, the adaptive landscape changes in such a way so a species finds itself no longer at a maximum, but at a fitness saddle point. Selection for assortative mating (like choosing to mate with like and avoiding individuals who are very different from oneself), or Koinophilia (wanting to mate with individuals close to the average group appearance) then drives two increasingly isolated populations away from the saddle point towards two distinct fitness maxima, as hybrids would have lower fitness. Assortative mating/Koinophilia thus functions as a reproductive isolation mechanism, but for competitive speciation to occur, the mutations, epigenetic mechanisms, or preexisting variation must find directions leading to higher fitness. This is not necessarily going to happen if the variation or rate of mutation is too small, because there are so many possible directions to go in strategy space and only one or a few of them lead to higher fitness. Like finding a needle in a haystack, an effective entropic barrier has to be overcome (14). The faster the mutation rate, or epigenetic change, the faster the entropic barrier can be overcome.

In the third model (consistent with allopatric speciation), the reproductive isolation happens due to a geographic barrier between sufficiently different environments (such as a river, an ocean or a mountain) that allows a divergent evolution that brings the two groups to far enough places in the adaptive landscape, so that they find new maxima in fitness that are still separated even if the geographic barrier is removed, and the adaptive landscape changes (though not to its original form). In this latter model of speciation, the reproductive isolation seems unnecessary and coincidental, but if the two species were somehow able to mix they would have lower fitness than if they stayed separate, because what produces high fitness in one environment is probably not optimal in the other.

Since we are ultimately interested in explaining why some cultures need more or less isolation than others when they form, we will not look at the third model of geographic/allopatric isolation. It is an interesting question why cultures that have been geographically isolated and diverged sufficiently and are brought back together are able to hybridize more than species, but we will not pursue it here, we are interested in how the genesis of cultures is possible even without geographical isolation. In the following, we focus on the shifting balance and competitive speciation models (without geographical barriers) and try to formulate a hypothesis that allows cultures to form with varying levels of isolation.

The shifting balance theory (with Mayr’s extensions) might be more applicable to cultural phenomena than biological ones, it hasn’t been observed as much in biological speciation. Apparently “blind” drift and the founder effect are weak forces compared to selection. But we see this scenario (shifting balance) all the time in cultural phenomena such as new companies having to spend much money before they make money with a new venture. This suggests that another force is operative in cultural speciation.

Regardless of the mechanism, some reproductive isolation happens in cultural genesis, but the reproductive isolation is usually not as severe as in species. Why might this be so? Two candidate hypotheses can be eliminated quickly as fundamental explanations:

  • A. Assume that cultural genesis happens through competitive speciation and hypothesize that assortative mating is selected for less in cultures than in species, allowing for less isolation. The reproductive units of culture (memes), whether they are considered to be brain patterns, ideas, values, behavioral patterns, or strategies, can change more easily and quickly than DNA, and the difficulty of getting through entropic barriers should therefore be diminished in cultures. Memes may be at one extreme of an information continuum starting with genes (hardware), through epigenetic changes (firmware) to ideas (software). So if going through entropic barriers can happen faster, there should also be a faster selection for assortative mating and hence more isolation. Despite this, there seems to be less isolation with cultural genesis, again suggesting that another mechanism allowing for less isolation is operative in cultures but not in species.
  • B. Or else, if the shifting balance model is applicable to cultural genesis, fitness barriers still exist, but for cultures, the barriers between peaks of the adaptive landscape may be smaller than for species, hence drift is not as important and hence isolation is not as important. Perhaps it is true that fitness barriers are smaller, but why should this be? A more fundamental explanation is needed.

One possibility is that cultures have horizontal (in the same generation) information transmission, as contrasted with sexually reproducing species, which have mostly vertical (from one generation to the next) information transmission, making recruitment from the current population more important. Nevertheless, some bacteria also have horizontal transmission of DNA and short reproductive times, and they still maintain reproductive isolation from other species of bacteria. A tradeoff exists between recruitment and the energy it takes to invest in the process, especially filtering the right from the wrong allele. This mechanism could just be natural selection, but in the shifting balance model this would act to suppress the incipient species/culture and would be counterproductive, as gene flow from mother species/culture plus drift would act just as with vertical transmission. If the scenario of the competitive speciation model is operative, this mechanism would be equivalent to assortative mating with vertical transmission. Perhaps bacteria have not evolved a good filtering mechanism, and different cultures differ in their isolation by how good a filtering mechanism they have, or other ways of making the tradeoffs involved in keeping out invading strategies. This can likely be tested empirically from the historical record.

Other tradeoffs may exist in cultures to mitigate the need for complete isolation, such as needing resources from the mother culture. This is an example of a tradeoff of immediate benefit for long term cost (being absorbed by the mother culture and never making it to a higher fitness peak).

We propose that the most important feature that may relax the isolation requirement for cultural genesis relative to speciation is the increased ability of cultures to provide an internal selection for a putative fitness maximum even before reaching it externally. Gene networks and even more so species, create an internal environment that has its own selection criteria. In cultures that internal environment is easier to change than in species, using mechanisms such as foresight and memory. However, there is the added risk that relying too much upon internal selection without returning to a consideration of external fitness landscapes can result in cultural extinction.

If we think of the shifting balance model of speciation, in addition to a random walk (genetic drift) in gene (or strategy) space, plus the force of selection (proportional to the gradient of the adaptive landscape(12)), an incipient culture can “see” (through reason and/or inspiration), albeit imperfectly, a portion of the adaptive landscape and keep going in the direction of a mountain pass, against the force of external memetic selection away from the current peak. If we think of the competitive speciation model, foresight allows cultures to more quickly find the direction that goes uphill among the millions of other directions. In both these cases, a culture possessing foresight will be able to speciate more quickly than with drift and random mutations alone. Possibilities that lead to worse places could be tried in thought (and more recently computer simulation) and eliminated before trying them “in vivo”.

In addition, once a new fitness peak is able to be perceived, the trip through the valley need not be as difficult as without that perception. In other words, the adaptive landscape may become less steep and even never have to go downhill if internal selection is stronger than external selection, which provides a more fundamental explanation than hypothesis B above. Another way to say this is that foresight allows for delayed “gratification” (or more accurately delayed external fitness maximization) and the ability to endure reduced external fitness and invest in material (hardware) or spiritual (software) infrastructure that would give an external selective advantage later on, while giving an immediate internal selective advantage. This may very well be a common theme in the birth stories of many religions and other memetic software programs.

The other component that differentiates the birth of new cultures from the birth of species is the ability to learn from the past, or the evolution of memory. Indeed, foresight is dependent on this. With the evolution of memory and foresight, genesis of cultures has acquired a Lamarckian component, though the Darwinian mechanisms of mutation, selection and drift are still operative and hence some degree of isolation from the mother culture might still probably important for forming a new culture.

The trade-off between immediate external and delayed external fitness optimization, and between internal and external fitness optimization made possible by foresight, memory and possibly other mechanisms is akin to other evolutionary trade-offs, such as between investing in the germline (next generation) and soma (current generation), or group and individual selection mechanisms. It allows for finding a “sweet spot” between the two extremes of no new culture (no isolation and optimizing for current fitness), and a new culture that is too hard to form because of hardship/low current fitness (complete isolation and optimizing for only delayed fitness).

We hope scholars will take an interest in helping uncover more about cultural speciation. Dynamic game theoretical and optimal control models of foresight, internal selection and delayed gratification (in addition to the old mechanisms of mutation, external selection and drift) could be developed and tested on the historical record and with incipient vs mother cultures such as the ones found in intentional communities or newly formed religious groups. Factor analysis could be performed to rank foresight along with drift and mutation in the formation of new cultures. Conditions such as trust (and the conditions engendering trust, such as cultures adhering to Ostrom principles or utilizing Greenbeard mechanisms) could be studied for their interactive effects on foresight and delayed gratification.

Practical applications might include a mother culture making the effort to allow and support groups who want to form new cultures which it might otherwise suppress, in order to deal with the environmental, social and spiritual problems of our time. A mother culture might also choose to support groups who want to maintain their old cultures, with as much isolation as they need instead of stigmatizing or attacking them, or forcing them to trade with the global economy. It may be vital to generate and conserve memetic diversity in this way, similar to the efforts of conservation biologists to maintain genetic diversity (though the biologists do not have to fight homogenization or forced gene flow). Another similar application is for intentional communities to intentionally isolate themselves, except for the minimum needed for recruitment,  building infrastructure, and paying taxes. As we mentioned, thanks to foresight and the need for recruitment, this isolation need not be extreme, and may change in its degree as the community evolves.

Last, as promised in the introduction, we would like to speculate on why cultural isolation has not made it into the vocabulary and understanding of how cultures form in the social sciences. Cultural isolation has been proposed by historians in a roundabout way (a new culture arising once the old one is almost dead) (15,16,17), so this is not a new idea, but the evolutionary implications have not been worked out, especially in cases where the old culture is still alive. We propose that this blind spot has occurred because:

  1. Isolation has gotten stigmatized through cults that use reproductive isolation to attempt new cultures, but often not in a direction of higher fitness.
  2. Isolation has gotten stigmatized through xenophobic tribal behaviors of certain cultures. But isolation does not have to involve xenophobia and the demonizing of the other.
  3. A certain breed of humanist culture (akin to an invasive species) has homogenized many cultures, through the carrot of trade, but military and police sticks are used as well. One of the mechanisms it uses to maintain itself might be stigmatizing isolation.
  4. The invasive breed of humanism is popular among economists and politicians and a few influential evolutionary sociologists.

Luckily, there are other breeds of Humanist culture that also have the strategy of “live and let live” and appreciate diversity. We hope they prevail.

[1] Richerson, P. J. and Boyd, R. (1978). A Dual Inheritance Model of the Human Evolutionary Process I: Basic Postulates and a Simple Model. Journal of Social and Biological Structures 1(2): 127–154.

[2] Mulder, M.B. and Durham, W.H. (2005). Are Cultural Phylogenies Possible? In R. Boyd & P. J. Richerson (Eds), The Origin and Evolution of Cultures, p. 319.

[3] Pál, C., & Miklós, I. (1999). Epigenetic inheritance, genetic assimilation and speciation. Journal of Theoretical Biology, 200(1), 19–37.

[4] Reproductive Isolation [link] (accessed 8/5/16).

[5] Schonfield, H. J. (1968, 2012). Those incredible Christians. New York: B. Geis Associates; Distributed by Grove Press.

[6] Iyigun, M. (2008). Luther and Suleyman*. Quarterly Journal of Economics, 123(4), 1465-1494. [link] [7] Buddhist monasticism: history and development [link] (accessed 8/5/16).

[8] Anabaptism [link] (accessed 8/5/16).

[9] Quigley, C. (1961, 1979). The evolution of civilizations: An introduction to historical analysis. New York: Macmillan.

[10] Bolnick, D. I., & Fitzpatrick, B. M. (2007). Sympatric Speciation: Models and Empirical Evidence. Annu. Rev. Ecol. Evol. Syst. Annual Review of Ecology, Evolution, and Systematics, 38(1), 459-487. [link] [11] Vincent, T. L., & Brown, J. S. (2005). Evolutionary game theory, natural selection, and Darwinian dynamics. Cambridge: Cambridge University Press.

[12] Traulsen, A. & Hauert, C. (2009). Stochastic evolutionary game dynamics. In H. G. Schuster (Ed.), Reviews of nonlinear dynamics and complexity. Weinheim: Wiley-VCH.

[13] Chouteau, M., & Angers, B. (2012). Wright’s Shifting Balance Theory and the Diversification of Aposematic Signals. PLOS ONE, 7(3), e34028.

[14] Van Nimwergen, E. & Crutchfield, J.P. (2000). Metastable evolutionary dynamics. Crossing fitness barriers or crossing via neutral paths? Bulletin of Mathematical Biology, 62(6) 799-848.

[15] Vico, G. (1725). Scienza Nuova.

[16] Spengler, O., & Atkinson, C. F. (1926). The decline of the West. New York: A.A. Knopf.

[17] Toynbee, A.J. (1934–1961). A Study of History. New York: Oxford University Press.


Published On: August 24, 2016

Iuval Clejan

Iuval Clejan

Iuval Clejan has a PhD in Physics from Boston University. He worked as a process simulation engineer at Motorola making memory devices, as a molecular biologist studying aging and DNA repair in nematodes at UNC, and as an engineer and farmer implementing alternative technology and agriculture at several intentional communities. His other current interests are culture genesis, local economies, the connection between continuous game theory and control theory and the physics of signature flipping surfaces.

Christopher Congleton

Christopher Congleton

Christopher Congleton is a social scientist who applies the science of sharing to help solve ecological and social problems. He integrates human ecology and technological development with natural systems, grounded in critical theory and practice. He has extensive international experience, including facilitating public open space design charrette for the City of Stavanger in Norway and a two year stint as an English Instructor in Korea. His recent work includes a project for the Rainforest Action Network in which he interviewed stakeholders, including staff, partner organizations, and constituents and analyzed data generated by hundreds of participants across the U.S.A. in weekend-long visioning processes to produce a summary of the resulting collective vision. He is also cofounder of the Davis Bike Church. He is interested in connecting with others using evolutionary theory to remove barriers to a just and sustainable world. 

Gil Benmoshe

Gil Benmoshe

Gil Benmoshe is an independent scholar studying intentional communities in the US. His graduate education is in developmental psychology but his current research employs both behavioral economic and game theoretic approaches to understanding success and failure of intentional communities. Gil is currently working in collaboration with a few academic institutions as well as the federation of egalitarian communities. He started several intentional communities and currently lives at Cambia community in Louisa, VA.


  • Ted Howard says:

    Some very useful ideas in this article, and some very dangerous ones.

    Bullying and dominance are possible social strategy sets present and dominant in many cultures. I see no ethical basis for conserving such diversity.

    Similarly I see no ethical basis for conserving any cultural modality (in so far as such cultures impinge on the freedom of individuals acting responsibly).

    Cultures are a very poor mimetic analogue of a species.

    Individual human minds are much closer analogues to mimetic ecosystems, within which exist many mimetic species in complex associations.

    Cultures, in as much as they have form and structure seem to me to be mere shadows of the complexity within the individuals with in a culture, and say more about the major modalities of sets of ecosystems present in the individuals with those cultures.

    Once individuals pass second level abstraction, then communication with outside agents becomes improbable, with each successive level of abstraction attained by such individuals reducing the probability of communication by further orders of magnitude. Even third level abstractions are almost impossible to communicate, by the one gets to 10th level abstractions, it is unlikely that anyone on the planet is likely to get the message transmitted.

    Our world is already post cultural, in many very significant senses. The proportion of individuals within any particular cultural that are post cultural can vary substantially, both between cultures generally and geographically or socially within cultures and subcultures (in as much as those terms have any meaning in such contexts).

    There are many conventions in reality that are arbitrary, and it only matters that there is agreement, like which side of the road to drive on. Not having such conventions impose a lot of risks and costs. Adopting such conventions (protocols) can significantly reduce risk and increase efficiency. In that sort of context, there can be some value in culture, and there can also be significant costs as technologies evolve but interfaces remain static (I have one computer board where most of the board is taken up with interfaces, the chips that do the work are tiny). Rapidly changing technology can demand rapidly changing standards. Some can find that difficult.

    And when one is exploring new territory, such protocols are not normally required. Anyone doing such explorations accepts that risk is a part of the process, and traffic or interfaces are unlikely to be significant issues.

    • Christopher Congleton says:

      Hi Ted,

      Thank you for your thoughtful commentary, we provide responses to some of your specific points

      1. We don’t like bullying or dominance either! Yet, we can imagine there might be an ethical
      basis for conserving bullying and dominance (and diversity of traits in general). That doesn’t
      have to mean we let them rule our affairs, just that we don’t have to stamp them out.

      2. You note that cultures are a very poor memetic analogue of a species.
      We agree – we hypothesized that cultures might be analagous to breeds and this looks a ripe
      place for inquiry. The starting point is to model and see – it’s an open empirical question how
      to best model such things. Certain phenomena like group level selection, the nesting/nursery
      effects of cultures and breeds or species (creating a selective advantage internally though
      there is still a selective disadvantage externally), foresight, or immune response, might be
      hard to explain at the lower level without taking into account the emergent higher level. To
      understand such phenomena at whatever level (gene, gene network, cells, organs, individuals,
      breeds, memes, meme networks, brains, cultures, etc.), we might need to acknowledge that complex
      networks are more than the sum of their nodes, and even more than the sum of the interactions
      around each node. Here is a table listing all the known similarities and differences between
      cultures, breeds and viruses that can prime our empirical research:

      We agree that part of culture resides in human minds, just like part of what we mean by a species or breed resides in the genome of individual organisms comprising said species or breed. But not all.

      There are 3 possibilities with the meme network within individual minds theory (as there are with the meme network within the group of minds, or the gene network within an individual, or the gene network within the group of individuals):

      A. The individual finds themselves isolated geographically from other individuals in another
      part of the fitness landscape, without having their memes subject to selection on the geographic
      path that brought them there (because it happens quickly on evolutionary scales). We only
      touched on this briefly in the paper (geographic, or allopatric isolation/speciation)

      B. The individual is at a current fitness maximum and starts going downhill in meme space in
      search of a higher maximum. Without foresight and isolation from the mainstream memes, the memes
      he is changing are getting selected against and also swamped by the unchanged version of those
      mainstream memes. In this case a combination of foresight and isolation needs to happen (a
      priori) in order to withstand the trip through the valley. The less foresight, the more
      isolation she needs to make it to another peak. Even when she finds another uphill direction
      after reaching the bottom of a valley and selection starts acting in her favor, she still needs
      some isolation (and foresight) to not be swamped by memes from the mainstream memes. If she
      makes it to another peak, she may not need isolation anymore if this peak is higher, but she
      would still need it if the peak is lower. This scenario (without the foresight) is what we
      referred to in the paper as shifting balance. Examples are likely to include religious leaders
      (Jesus, Mohammed, Buddha) and a few pioneers of science (Galileo, Copernicus).

      C. The fitness landscape changes (due to other individuals or some experience). What used to be
      a maximum is no longer so. It has developed one uphill direction (so technically it’s a saddle
      point) among a multitude of downhill directions (which haven’t changed their downhill nature).
      In this case, isolation is selected for (a posteriori) as the individual starts following the
      uphill direction, because hybrids would have lower fitness and energy would be better spent in
      mating with one’s own memotype. This is what we called competitive speciation in the paper, with
      the mechanism of selective mating or koinophilia providing the isolation. But these mechanisms
      operate at the level of groups, not individuals, though we can envision them also operating on
      groups of memes within individual minds. Two examples of the former are Newton and Einstein. Two
      examples of the latter are religious people who can compartmentalize their exotic strange
      beliefs from “normal” life beliefs and attitudes or doctors who can regularly perform surgeries
      but keep empathy and compassion alive.

      3.Can you provide an example or series of examples to elaborate what you mean by different
      levels of abstraction and why higher levels of organization imply more isolation from the same
      levels? Are you suggesting a connection between reproductive or cultural isolation of species,
      breeds, and cultures and computer science isolation of different levels of abstraction on the

      4. What do you mean by post-cultural? Can you point to single individuals who are not dependent
      on other individuals for their very survival?

      5. You invoked conventions like which side of the road to drive on, which we’d call
      “coordination”. Coordination seems likely to start as an arbitrary agreement or happenstance,
      with path dependency cementing patterns of coordination within cultures once there are enough
      adopters. Coordination can happen without much communication, which makes it very efficient, and appealing from the standpoint of designing cultures.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.