Tim Lewens on Lamarckian inheritance in cultural evolution

[This is a second post originating from reading Tim Lewens book Cultural Evolution. The first, on the application of the concept of population thinking to culture, is here.]

One of the most repeated criticism of the analogy between cultural and biological evolution is that inheritance in the former, but not in the latter, is Lamarckian. Things may be, to a certain extent, complicated (“Lamarckian” evolution might mean different things; the concept of soft inheritance, which might include “Lamarckian” forms, is no more a taboo in biology), but the nuts and bolts – which are, I think, what really matters for the analogy cultural/biological evolution – are not.


Here is the famous giraffes story: according to a Lamarckian explanation, the long neck of giraffes is due to the fact that, in each generation, individuals were stretching their necks to reach higher branches, making it longer and more robust. Their offspring would then inherit the longer necks of their parents: this is what is known as “inheritance of acquired characteristics”. We now know that this is not how biological inheritance works. Children do not inherit plastic surgery, body-builder muscles or, in the classic example, mutilations, from their parents (I discovered that the seminal experiment disproving Lamarckian inheritance involved, in fact, cutting the tails of 68 white mice – sorry internet – and checking that all of the 901 offspring had a normal tail). What about culture, then?

A famous quote of Stephen Jay Gould reads:

Human cultural evolution, in strong opposition to our biological history, is Lamarckian in character. What we learn in one generation, we transmit directly by teaching and writing. Acquired characters are inherited in technology and culture

This looks obvious, but I’ve never really got its precise meaning. My interpretation is that here Gould was mixing the cultural and biological levels of analysis. If we consider the biological individual as the unit of analysis in cultural evolution, “what we learn in one generation” can be the analogous of the stretched neck of the giraffe, but then cultural evolution is clearly not Lamarckian. Not differently from a stretched neck, reading skills or the ability to play the accordion are not inherited.

Sure, I can teach my daughter to read or to play the accordion, but in which sense this inheritance is Lamarckian? Now, the units of analysis are “reading” and “playing the accordion”, and “generations” are the instances of transmission. When “reading”, or any cultural trait, is transmitted from individual A to individual B, there is nothing particularly Lamarckian going on.

Even assuming that the trait has been modified, how one can decide whether, from the point of the view of the trait (our unit of analysis), the new character is “acquired” or not? The laptop I am using in this moment has various features that distinguish it from its predecessor (one is, say, the “Force Touch” trackpad). Is this Lamarckian inheritance or not? Tim Lewens, as well as Alex Mesoudi before him, clearly understands how the question of whether cultural inheritance is Lamarckian or Darwinian depends on how we consider the genotype/phenotype distinction in cultural evolution. In other world, only if one believes that the modification happens on the “phenotype”, and it is then inherited (remember the giraffes!), then can also define cultural evolution as Lamarckian.

I tend to share Lewens’ pessimism about the possibility of making sense of the genotype/phenotype distinction in cultural evolution (in the same time, I do not think this is too worrying, see here), but this would definitely bring me out of topic. The take home message here is: when someone complains that cultural evolution is Lamarckian, s/he is also assuming (i) a replicator-centred view of cultural evolution, in which (ii) it is possible to distinguish between the analogous of genotype and phenotype, and (iii) there are good reasons to conclude that modifications happen at the phenotype-analogous level. Or, more likely, just repeating something that looks obvious.


Acerbi, A., Mesoudi, A. (2015), If we are all cultural Darwinians what’s the fuss about? Clarifying recent disagreements in the field of cultural evolution, Biology & Philosophy, 30(4), 481-503

Gould, S. J. (1980), The Panda’s Thumb, Norton

Jablonka E., Lamb M. J. (2008), Soft inheritance: challenging the modern synthesis, Genetics and Molecular Biology, 31(2).

Kronfeldner, M. E. (2007), Is cultural evolution Lamarckian?, Biology & Philosophy, 22(4), 493-512

Lewens, T. (2015), Cultural Evolution, Oxford University Press

Mesoudi, A. (2011), Cultural Evolution, Chicago University Press


10 thoughts on “Tim Lewens on Lamarckian inheritance in cultural evolution

  1. I think perhaps there is a genotype phenotype distinction, at least in the evolution of in-the-world objects. Before we get to anything as complex as “playing the accordion” we could look at simple objects, like a bowl or a blade. I can argue that a single bowl in my hand has a phylogeny, and that it has evolved — but this is not the place. If that is so, the locus of replication, and possible variation, is the human organism (where else?). The phenotype is the bowl on the potter’s wheel and beyond. The genotype is the physical modification of the architecture of the brain (and we don’t have much clue what that is yet) that is the information which the potter employs when making a bowl when there are no material bowls present.

    1. Hi James,

      thank you for your comment. I agree that it is good starting to reason from simple cases, like the bowl example you use. I am also not opposed, in principle, to the possibility of the genotype/phenotype distinction for cultural objects, simply I believe that we do not have a good one at the moment. Even though it is probably the most intuitive candidate, there are, to me, a few problems to identify the genotype in “the physical modification of the architecture of the brain”. If you are interested you can have a look here:


      [The relevant part is the end of the section “Preservative versus reconstructive processes depend on the granularity of the analysis”, starting from “One could hope that, when we find the correct unit of analysis for cultural evolution…”]

    2. Richard Dawkins’ 1982 book proposed genotypes-in-brains – the first clear articulation of the idea I am aware of. Rober Aunger offered a book-length treatment of the idea in 2002. However, I think it is fair to say that most other writers on the topic have fairly uniformly rejected these proposals – for lacking germ line continuity, failing to deal properly with transmission via computers and so on.

  2. Hi Alberto,
    I have already read “If we are all cultural Darwinians …” and I think, together with Massimo Pigliucci’s “The Trouble with Cultural Evolution” it is the clearest, most exhaustive and succinct summary of where we are now. But Pigliucci’s stricture is timely and correct, and there is work to do.
    The first step might be to make as clear a distinction as possible between what, we are suggesting, evolves in a “broadly darwinian” way, culture (and it now isn’t just homo sapiens culture) on one hand, and on the other the human organism (which clearly does evolve in the darwinian way). It’s not clear how we do this. Maybe, culture is anything that has material existence (including as energy) outside the skin of human beings. Everything inside the human skin, but outside the brain-neural network, is the human organism. Anything in the brain-neural network is either or both.
    I am writing a book for the general reader on this at the moment and have spent 20,000 words just de-familiarising the landscape in preparation for the hypothesis, so rather than try to outline a coherent argument I’ll jump about a bit in an attempt to confirm that a Lamarckian approach is redundant.
    I start with the seemingly pretentious but necessary observation that the internalist externalist distinction is redundant because a cultural entity, say a bowline knot, has a distributed existence, it exists in a space-time continuum that stretches between all bowline knots (as in configurations of a cord) in the world, and the existence of all bowline knots in human brains (whatever the material form of this existence is, we don’t know yet).
    As you point out, a bowline knot either is or it isn’t. In this sense I think Richerson is misleading, in that what he seems to be referring to are the mnemonic devices for tying a bowline. The irreducible bowline knot in the brain is always the same. Otherwise, as you say, it would produce another kind of knot or no knot at all.
    Irreducible is the key word here, and it is at this scale of granularity that we might search for the unit that undergoes replication, variation and selection.
    It is perhaps unduly challenging to start at the scale, the granularity, of a recipe. Recipes are highly reticulated. “First sauté an onion” would produce a WTF? reaction from somebody in whose brain the culture of cuisine took up no space at all. How many bits of potential irreducible difference are there in a recipe? You could go back through the phylogeny of sauté for a long time. Luckily for the economy of human communication, loci of irreducible difference (LIDs) work at any scale. Does it use pasta or rice? Rice. Then it’s not lasagne, it’s risotto.
    Stories are the same. If I want to tell the story of Cinderella I locate the irreducible difference between a story and everything else that is not a story, say an elephant or a recipe. Then I locate the irreducible differences between this story and all other stories. These will be many but countable. It has to have a beautiful girl, two sisters or step sisters, a glass (or fur) slipper and so on. If it has a wicked step-mother, or a magic mirror, or seven dwarves, then it’s not Cinderella. Then, with a lot or a little reconstruction, reticulation, modification of one or a thousand contingent irreducible differences (no, dear, the spokes of the wheels of the coach were studded with emeralds, not diamonds), I tell the story. In this model there is no place for the notion of an attractor, unless all LIDs, at whatever scale, are attractors.
    The task which Massimo Pigliucci indicates is demanding. If meaning is itself merely a locus of irreducible difference, say the difference between a pin and a needle, then hole/no hole is a LID. But then we have a hand-held needle, a sewing machine needle and a hypodermic needle, all topologically equivalent. What this suggests is that the relationships of LIDs are complex, and LIDs work across many scales (the difference between a triangle and a square as opposed to the difference between a religion and an ideology). This looks impossible, but once a basic hypothesis has been suggested, scientists and mathematicians can start to develop models, perhaps borrowing from thermodynamics, that describe such a landscape.
    As you suggest, “what we learn in one generation” contains an imprecision within and between the “we”, the “what” and the “in” which is breath-taking in its scope. There is no room here for Lamarck.
    I want to stop calling these things LIDs, but can’t think of a word. Ping!?

  3. In your paper you say: “One proposed solution to this puzzle is to consider the information, wherever stored, as the equivalent of the biological genotype, and the expression of the information in behaviours or artifacts as the equivalent of the biological phenotype (Dawkins 1976).” Right however you then raise objections – which I don’t think are coherent. The split between what is inherited and what is not is indeed useful and fundamental from the perspective of evolutionary theory. The paper argues that sometimes the ‘phenotype’ is inherited. Here you are just not taking the original proposal seriously. Genotype is what is inherited; phenotype is what the genotype affects. There is no inheritance of the phenotype – by definition.

  4. Hi Alberto! Followed you over here from Twitter.

    I’m an urban planner – not an ‘academic’ – though I have a Masters degree – but, more of a nuts and bolts consultant for land developers.

    In my spare time, I’ve written a book called Paleo Places where I’ve tried to deconstruct what we want/need in our natural and man made environments from an evolutionary perspective.

    In my book, I talk about Donald Brown’s “Human Universals,” like fire-making. My understanding is that fire-making is not a human instinct, but that it is a ‘human universal.’ Sooo…is it a product of cultural evolution then??? Are all the human universals??? If you ship-wrecked some kids a la the old movie, the Blue Lagoon, would they “do” the instinctual things, like sex, but NOT do the cultural evo/human universal things, like fire-making?? Or am I mixing up Human Universals and cultural evolution??

    Btw, my blog has an excerpt from my book posted if you’d like to take a look and you’re not too busy at the moment 😉

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s