The evolution of language and biology

Sandman has a post up, Can There Be A Synthesis Between Cultural And Biological Evolution?, taking off on the PLoS Biology article, Across the Curious Parallel of Language and Species Evolution. Read both. I would add one important point though: linguistic and biological evolution are simply subsets of evolutionary dynamics. That is why Martin Nowak's book of that name, Evolutionary Dynamics, naturally has a section on the evolution of language. Several evolutionarily oriented thinkers have attempted to translate models originally developed for biology into the domain of culture. Cultural Transmission and Evolution and Culture and the Evolutionary Process are two works which I think are good introductions to the field.

I think it is important to emphasize that various evolutionary genetic parameters such as selection, drift and migration are operative in cultural evolution. This means that the evolutionary dynamics of language extend beyond the debate as to the relevance of functionalism to anthropology, just as evolution on the molecular level is more than simply adaptation. Additionally, the adaptation of cultural traits must be measured as both intrinsic and extrinsic to the mind. That is, particular cultural adaptations may be evaluated against their fitness implications against the environment (e.g., the extent of clothing) or social milieu (e.g., structures which mediate rank order and enable group cohesion to operate smoothly). But others are almost certainly canalized by the constraints and incentives which are inherent within the structural parameters of the human mind. Any given cultural trait may span all these dimensions, or limit itself to one dimension. For example, it seems rather clear that religious ideas are strongly constrained/adapted to the nature of the human mind. On the other hand, the question as to whether religious ideas are constrained/adapted exogenously to the mind is an open question, they may well be, or they may not.

Then there is the matter of substrate. Biological evolution is gifted with a discrete and concrete unit of replication, the physical gene. The structure of the gene and the process of meiosis result in the emergence of endogenous parameters which have an affect on biological evolution; e.g., the parameter of recombination which is conditioned by the physical nature of DNA. In contrast, cultural evolution has no atomic unit to which we can ascribe bookkeeping functions. Cultural evolution may be considered to operate by principles of blending of traits, but this misses some important nuances.

For example, the transmission process in cultural evolution is extremely malleable and exhibits far less invariance and regularity than that of biological evolution mediated by genes. A child may adopt the values of one parent to the total exclusion of others. Except in the case of mitochondria one can generally state that the offspring receives half of its genetic material from each parent in the case of diploid sexually reproducing organisms. Many social insects are characterized haplodiploidy. Asexual organisms reproduce in a clonal manner. And so forth. The set of various modes of transmission is finite and not so numerous as to be intractable (ergo, the reasonableness of a few categories such as haploid, diploid, haplodiploid, sexual, asexual, hermaphroditic, etc.).

Cultural transmission is a much more statistical affair. Though one can posit models of total transmission of cultural values via one parent (e.g., Judaism transmitted through the mother), these are often exceptional cases, and generally the rate of deviation from the norm is high. Blending inheritance in evolutionary biology was always problematic because of the rapid diminishing of variation, but in cultural process this is not an issue. Additionally, the nature of variation across populations differs a great deal in culture from that of genetics. Genetic variation usually exhibits a clinal pattern so that demes only change gradually in character; this is due to the power of gene flow in eliminating between group variance. This is one reason that many evolutionary biologists are skeptical of the power of higher than individual level selection, within group genetic variation is usually much greater than between group variance, and the power of selection is directly proportional to the extent of said variance.

No so for culture. Going back to language, consider the rapid changes in dialect that can occur across group boundaries. Because it seems plausible that we have a "language organ" the trait is atypical in some ways, but it still exhibits differences from biological variation that illustrate the general pattern. In small tribes one can posit that there is almost no within group variance in accent. If an individual from the outside marries into the group their offspring will carry half their genes, but, it is likely that very little of their cultural traits will be passed down because of the strong incentives for the children of a foreigner to adhere closely to the norms of their tribe. While genetic variation can be thought of as the gentle rolling sweeps of hills gradually rising and fall across the landscape, cultural variation many manifest as a set of sharply demarcated plateaus and ravines cutting a very rugged and fragmented landscape. This of course is a better context for higher-than-individual level selection to occur!

Finally, there is the issue of biology-culture evolution. This is perhaps the most important point of synthesis and cross-fertilization between these two domains of evolutionary dynamics. Lactase persistence for example is probably the classic case. But there are others, such as the CNV work on amylase, or the affect of Toxoplasma gondii on the central tendency of various societies in regards to personality. Exciting times....


More like this

I have often said that I tend to see "group selection" as a lesser evolutionary force when set against lower levels of evolutionary processes, e.g., "individual" or "gene" level selection. By group selection I do not mean pro-social tendencies, or the success of individuals who band together as a…
In reviewing a paper which sketches out the boundary conditions under which group-level natural selection would result in the emergence of altruism as a genetically encoded trait, I stated: ... I would look to cultural group selection, because there are many cases of women being assimilated into a…
There have been a spate of articles about E. O. Wilson'sdrive to put group selection back into mainstream conversation among evolutionary biologists over the past year. Wilson has kept the torch alive for this particular paradigm since the 1970s, when it was prominently featured in his famous book…
I recently finished reading The Faith Instinct: How Religion Evolved and Why It Endures, a new book by Nicholas Wade, a science writer for The New York Times. Before giving it the "full treatment" I thought it behooved me to revisit some of the scientific literature which Wade relies upon to give…

Linguistics is discrete too -- you have the /f/ phoneme in your brain or not, you've stored the word "book" in your lexicon or not, you've flipped the "objects after or before the verb" switch in your brain, you combine the meaning of "black" and "dog" to understand what "black dog" means (discrete intersection of discrete sets).

Cultural transmission isn't like inheritance; it's like infection, which is why it can change so quickly and show such clear-cut spatial heterogeneity. That's also why we don't have to worry about getting something from the father or mother -- we get infected by reservoirs of infection (like "the English language in this place and time").

That's also why you see the gene-culture patterns that you do. Take the case of higher-frequency irregulars remaining in a language, while the low-frequency irregulars exponentially die off. If you use an irregular frequently (like "went"), that's the same as coughing a pathogen into the air frequently -- it infects more people than an irregular that you only cough up once every 10 years (like "clove").

I think the reason most people don't pursue the clearly better infectious model is that they don't teach epidemic models in pop gen or evo bio -- you'd need a broad math bio course for that. Now, the big wigs do -- Feldman and Cavalli-Sforza's book treats S-I-R models, Bartholomew's book on stochastic models in social processes from the '60s does too, and the more recent ones use it or a formal equivalent (I think called the "random copy" model).

But that hasn't filtered to everyone else, even though it's already there in the "meme" concept from Dawkins. Basically, evo bio people need to learn a broader class of models, and things become much clearer.

Just to be explicit, you're still using an evolutionary framework, but you're investigating how infectious agents and their hosts evolve, against and with each other.

Take a look at this article about language, culture, and brain evolution:
Greenberg, G., Partridge, T., Weiss, E. & Haraway, M. M. (1999). Integrative levels, the brain, and the emergence of complex behavior. Review of General Psychology, 3,168-187.

By Gary Greenberg (not verified) on 24 Jul 2008 #permalink

If I recall correctly, Robert Pennock used the evolution of language analogy in his book: "Tower of Babel: The Evidence against the New Creationism"


At the higher levels, cultural evolution is not just Lamarckian but intentional. If you look, for example, at the rise of republican, secular, liberal governments starting in the 17th centuries, there was an enormous intellectual effort to understand history and society and to invent and construct an improved sociopolitical form. Many participants thought that they were merely reviving classical forms of some kind, but they weren't, and many of them knew that they weren't.

You can still call it evolution (variation and selective retention), but intentional variation is added to blind variation.

It's common enough for social scientists, historians, satirists, and ideologues of various kinds to claim that history is entirely blind and that results of social action are never what was intended, but I find it hard to believe that when I look at the sweep of history. When they go beyond that and also claim to know what the hidden, secret forces of history really were, I tend to think that it's just self-promoting propaganda.

By John Emerson (not verified) on 24 Jul 2008 #permalink

Just to clarify: linguistic evolution is pretty much blind (I was thinking of political and social evolution.)

And when contrasted to biological evolution, "cultural evolution" includes everything not genetically transmitted: language, family, religion, social forms, technology, etc. Often "culture" is used to name the so-called superstructural elements such as the arts, religion, ritual, entertainment, as opposed to the infrastructure of the economy, technology, and government, but that's all culture from the biological point of view.

By John Emerson (not verified) on 24 Jul 2008 #permalink

Agnostic: You make a good case. And with lateral transmission of DNA heritable genomic elements can have infectious properties. Also, some symbionts are transmitted intergenerationally with enough reliability that they can be viewed as as a kind of inheritance.

Perhaps infection and inheritance are on a continuum of lateral and vertical transmission (leaving aside another axis of probability of transmission and another of fidelity).

And are constructed niches (modified ecology, artificially selected species, infrastructure, architecture etc) infectious, truly heritable, or neither? Their transmission through time and space may be due to replication, active maintainence, or simple persistence. (See J. Scott Turner.)

Genes and language get a lot of attention due to their particulate, high fidelity transmission, but all of these transmitted entities - epigenes, memes, constructed niches - are important features of evolutionary dynamics.

John Emerson: How much of that cultural intentionality has been shaped by a larger cultural structure created by past selection, instruction (Lamarckian), and chance? I'm not arguing against your point, just noting that the intentional aspect itself be the result of underlying selection and other processes.

I believe that meme thinking can be valid in the same way as genes. But if you compare the effort that is put into identifying genes and mapping them with the 'bar room coversation' level of most memetic discussion there is a stark contrast!

The other problem is that it is systems which are selected, not really genes or memes as such - not specific items of information but the self-reproducing and potentially growing systems within which information becomes information (and without which there is no information).

Using genes as a shorthand for systems (the rest of the system being 'held constant' as it were) is usually OK - but there seem to be bigger problems when people talk about memes without conisdering the systems which give them meaning.

I once presented a lunchtime seminar analyzing the meme for celibacy, which Dawkins uses as an example in the Selfish Gene - and people (psychologists and biologists) just tuned-out and glazed-over long before I reached the end of my argument!

The point was the 'celibacy' as a behaviour had different meanings within different systems. Nowadays it means voluntarily refraining from sex and thereby having no genetic offspring while retaining control over the disposal of one's estate - but in Medieval Roman Catholic Europe celibacy meant not having legal heirs. RC priests often had genetic offspring, but they did not have legal heirs to whom they could transmit their property. (In other words, Dawkins in 1976 was confused about the meaning of medieval celibacy!)

So 'celibacy' is _not_ a meme - it is part of a system, and celibacy is part of a different system now than it was in medieval times; and the modern system of secular celibacy did not evolve from the medieval system of legal celibacy.

(Interestingly, celibacy also evolved within the RC church, and has a totally different justification in modern doctrine - something to do with altruism and dedicating one's life to the RC Church - than it did in medieval times.)

Going further - (I got this from Ernest Gellner) Medieval celibacy was just one mechanism by which the warrior ruling class controlled the intellectual ruling class by preventing them from building rival dynasties. Other methods were having the intellectuals as slaves (Romans), non-citizens (Jews), or Eunuchs (Arabs).

My main reservation about memetics is therefore that people are not really *interested* enough to put-in the time and money needed to find-out stuff.

After all, memetics is really a new version of 'history of ideas' and that has never been a very influential or popular branch of history (I love history of ideas, and apparently so does Razib, but we are not really typical of evolutionary scientists).

How much of that cultural intentionality has been shaped by a larger cultural structure created by past selection, instruction (Lamarckian), and chance? I'm not arguing against your point, just noting that the intentional aspect itself be the result of underlying selection and other processes.

Any intentionality at all makes a contrast with blind variation, and also with unconscious, unintentional Lamarckian change. (By the latter I mean, for example, phonetic change, such as the way English systematically lost the "gh" sound in words like "thought" -- it wasn't always a silent letter. No one decided to do that, it just happened).

How much of that cultural intentionality has been shaped by a larger cultural structure created by past selection, instruction (Lamarckian), and chance?

All those factors are always in play. I'm very suspicious of attempts to link historical change and evolutionary change over time spans of less then a millenium or two, though.

It's one thing to say that voluntary social action always takes place against a background of inherited dispositions, inherited social and cultural patterns into which people are indoctrinated, and various external influences such as climate change, changes in economic productivity, etc. It's just self-aggrandizement when a historian, social scientist, or ideologue claims that they know the real causes of everything, and that all of the players at the time were deluded and never really did anything at all, because blind forces control history.

By John Emerson (not verified) on 25 Jul 2008 #permalink

GBC As you say, medieval celibacy was about not leaving legal heirs. However, Gellner seems to be putting a strange twist on the story. The warrior caste had no objection to high church officials leaving heirs, or even to having high church offices becoming hereditary in certain families. It was, after all, their sons who held such offices. It was always the Papacy which held out for the power of Rome to appoint the Bishops. Medieval celibacy represents the triumph of the Church over the landed aristocracy not the other way round.

By freethinker (not verified) on 25 Jul 2008 #permalink


I probably didn't express this very well: Gellner argued that both military and prieathood were needed to rule a society of any size and complexity - and he was discussing how they managed the 'division of powers' so as to minimize conflict and make the system sustainable.

The Evolution of Language and Biology
Khanâs discussion presents a good case for the difference between cultural and biological evolution. There now exists a published scientific theory of Cultural Evolution (Gehlsen, 2009 - find this book on Amazon). I will use the Complex-Systems Theory of Culture to clarify a few points.

For agnostic:
The cultural transmission of information is not a discrete process. You seem to be confusing the process with the products of the process. The word âdogâ may be a discrete structure but only when it is written. People say the word many different ways with many distinct accents, and which dog you think about can vary dramatically.

Models donât clarify anything. Models are fundamentally descriptive, and descriptions encompass a virtually infinite amount of variation, which tends to expand discussions rather than clarify them. Scientific theories âexplainâ things in ways that reduce arguments by pointing the way to relevant issues. Memes donât exist (the concept is vacant), and cultures are not infections.

For John Emerson:
Cultural Evolution is not Lamarckian at any level, and it isnât âintentionalâ because we canât control it. Humans may have âintentionsâ, but the unintended consequences of human actions always out weigh our âintendedâ purposes. When you look at the âsweep of historyâ with 20-20 hindsight you can reconstruct it to be anything you want. Try predicting cultures and future events and you will discover that the unexpected far out weighs the intended or even the predictable.

For Colugo:
You are moving closer to the point. Cultural structures are reliably âreplicatedâ with some degree of similarity into the next temporal frame. All units in cultural evolution are difficult to quantify because cultural transmission is not discrete (even an âinfectionâ model would provide a basis for discrete units if it were valid). Cultural transmission is a continuous and non-deterministic process, which means that it canât be broken down into discrete units. Most people have trouble with evolutionary concepts because they are non-deterministic, which frustrates our linear deterministic scientific heritage. Cultural evolution is even more difficult because it is pure process. There are no structural replicating entities. The only structures that are produced are the material residues of our knowledge and behavior.

For BGC:
Your discussion of âcelibacyâ as a meme demonstrates my point that âmemeâ is a vacant concept. There is no way for evaluating whether two putative memes are analogous or homologous.

Cultural replication works like this - a human (an intentional autonomous agent) is born into a rich environment of information (a culture). The vast amount of information that radiates at us each day is mostly perceived at a non-conscious level (in fact learning takes place at a non-conscious or subconscious level). In other words, no body has conscious control over what they learn (culture is not a Lamarckian process).

Cultures are faithfully replicated when individuals are thoroughly socialized within similar cultural environments. On the other hand, humans are a complex organism, which means that some will behave in unexpected and unpredictable ways (cultures are non-deterministic at many levels).

Cultures are open systems, which means they are maintained by a flow of information. Any aspect of culture can be explained by answering these four questions:
1. What information is transmitted?
2. How is the information transmitted?
3. Who is transmitting the information?
4. Who is receiving the information?

This is the power of theory; it tells you what questions are relevant.

The Complex-Systems Theory of Culture is employed in the examination of current cultural issues at: