Cultural attraction, “standard” cultural evolution, and language

Speaking Our Minds (SOM) was a great pleasure to read. This slim book provides even a non expert like myself with an accessible but, at the same time, in-depth treatment of language evolution. Scott-Phillips proposes us a coherent and, according to him, exhaustive, picture of the origins and evolution of language. The big questions are answered: we can proceed to the next topic.

I wonder how the community of linguists will feel in regard to this bold attitude (by the way, I am all for bold attitudes). As for myself, I can comment on a particular aspect of the book, that is, the role assigned to cultural attraction in explaining some of the features of language.

The basic idea behind the concept of cultural attraction is spelt out with remarkable clarity in SOM. In short, cultural transmission, differently from biological transmission, is mainly a reconstructive process. Each time we “copy” a cultural trait we are in fact reconstructing it, starting from some piece of information that we gather from others. Individual modifications are not rare, and they are not errors. They are the crux of the cultural transmission process, and, importantly, they tend to be oriented in non-random ways (hence the notion of attractor).

One example – which I discovered reading this book – is tonal languages. In tonal languages (like Mandarin Chinese) the pitch that one uses to pronounce a word makes a difference to its meaning. It has been discovered that the distribution of tonal languages is associated with the distribution of two genes that regulate neural development. These are not genes for tonal languages, as individuals without the genes can learn them (and vice versa), but they may represent a factor of cultural attraction if these genes make it easier (for example) to detect or produce pitch differences. Imagine a population in which few individuals have the variants. Language changes that give more importance to pitch will be, in this population, rare and generally not re-constructed. The population will converge on a non-tonal language. The opposite will happen in a population in which the majority of individuals have the genetic variants in question.

Scott-Phillips gives a few other examples of factors of attraction that may shape language attributes, some related to biological or cognitive features (like the example above), and others related to communicative needs, drawing mainly on the research from the Edinburgh Language Evolution Group. Overall, The case is convincing: cultural attraction is likely to have an important role in determining the features of the languages we speak today, and the details of their evolution. But is that all? What about all the researches that use a more “standard” evolutionary framework to study language, that is, that consider it like a culturally transmitted replicator?

The idea that languages evolve like biological species, with a process of descent with modification, has a long and successful history. Darwin’s famous endorsement of language evolution testifies to that. Phylogenetic analyses are today used routinely in cultural evolution, and while their application to different domains is far from being uncontroversial, their success is at least partly due to the fact that they have been productively applied to language evolution, providing stimulating results. If phylogenetic analysis works for languages, what does this tell us about the feasibility of using standard evolutionary tools to understand their historical dynamics?

Recent researches showed that the rate of changes of words is correlated with their frequency of use. Words that are similar in related languages (a classic example is terms for numbers: think about one in English, un in French, uno in Italian, etc.) are also words that evolve at very slow rate, and, interestingly, are the words that are used with high frequency in daily life. This suggests a classic evolutionary pattern, one of generally faithful transmission with random modifications. Frequency of use would indeed affect rates of replacement by reducing the “mutation rate”, as words used frequently would be, for example, remembered more easily than words only rarely used.

My general perspective is that various domains of human culture are characterised by various degrees of reconstruction and preservation in the transmission of their traits, and when domains are close to the “preservative extreme”, it is useful, for pragmatic reasons, to consider them as standard evolutionary systems. Moreover, in the same cultural macro-domain, like language in this case, different aspects may be situated in different regions of the preservation/reconstruction continuum. More than asking which aspects are in general more important, it may be more productive to ask when and why transmission is preservative or reconstructive, and what the consequences are for the resulting cultural dynamics. For example, one may wonder whether the contemporary widespread use of media favouring strongly preservative transmission (such as “sharing” something on Facebook, or “re-tweeting” it) may play a role in contemporary language evolution.

In sum, I strongly believe that the cluster of ideas surrounding the notion of cultural attraction (the importance of individual reconstruction in cultural transmission, the fact that modifications to cultural items are generally not random, the importance of universal, or at least relatively stable, factors of attraction), developed in the past years by anthropologists like Dan Sperber and others, is one of the most important contribution to the contemporary study of cultural evolution. I am also open to considering whether cultural attraction forces are responsible for the most interesting attributes of languages, as one could infer from Scott-Phillips’ book. A further step would be to identify which features of languages are due to cultural attraction forces and which features are due to processes included in “standard” cultural evolution models, such as random modification of words, simple contextual learning biases, and similar, and how the various processes interact. The material presented in Speaking Our Minds may be an excellent starting point for this endeavour.

Some references…
On cultural attraction: see Claidière et al. 2014, How Darwinian is cultural evolution? and Dan Sperber’s book Explaining culture Tonal languages and genes: Dediu & Ladd 2007, Linguistic tone is related to the population frequency of the adaptive haplogroups of two brain size genes, ASPM and Microcephalin The Edinburgh Language Evolution Group: see e.g. Kirby et al. 2008, Cumulative cultural evolution in the laboratory: An experimental approach to the origins of structure in human language Language as a “culturally transmitted replicator”: Pagel 2009, Human language as a culturally transmitted replicator Rate of changes of words is correlated with their frequency of usage: Pagel et al. 2007, Frequency of word-use predicts rates of lexical evolution throughout Indo-European history “My general perspective…”: Acerbi & Mesoudi 2015, If we are all cultural Darwinians what’s the fuss about? Clarifying recent disagreements in the field of cultural evolution


  • Thom Scott-Phillips
    Thom Scott-Phillips 24 June 2015 (00:40)

    20 years ago evolution was a “dirty word” in linguistics (see McMahon, 1994, Understanding Language Change, p.314). Much has changed: the cultural evolution of languages is now a thriving area of research. There are now many findings about the factors that influence how and why languages change in the ways that they do, and in particular how and why systems of conventional codes begin to acquire the sort of features that identify them as languages. In chapter 5 of SOM, I suggest that these findings are best framed within the context of Cultural Attraction Theory (CAT). In the comments above Alberto continues a productive discussion on this topic, which he has had with myself and others over the past year or so (see previous entires on his blog, and his recent paper on the topic). At the centre of this discussion is the relationship between CAT and other frameworks that have been used to study cultural evolution. Let me in this response sketch one of the reasons why believe that CAT is the right framework for explaining culture. This will allow me to then respond directly to some of Alberto’s queries.

    For many disciplines, the key intellectual question is why the world is the way it is, and not some other way. In physics, for instance, this leads to the search for physical laws. In linguistics, this question sometimes goes by the name of Greenberg’s problem: why are languages the way they are? Biology asks why organismic form is the way it is, and the theory of natural selection is so important exactly because it provides an answer to this question (see below).

    By analogy, a theory of culture should aim at explaining why cultural items take the forms that they do. The items in question are mental and public representations, each of which can be cultural to different degrees. So, the goal is to find a way to explain the relative distributions of tokens of different types of cultural representation, both public and mental. This is the agenda of what has been called an epidemiology of representations.

    Cultural Attraction Theory (CAT) is a proposed answer to the challenge of an epidemiology of representations. As Alberto summarises, the central claim is that in the process of cultural propagation, factors of attraction tendentially cause cultural items to be modified in non-random ways, such that some cultural representations become common in a population, and some do not. If this is correct, it provides a genuinely causal framework with which to explain culture. More specifically, it implies that to explain why a given representation (or set of representations) is common (or not) in a population, we must identify the corresponding factors of attraction.

    Let me elaborate on the importance of this point, through a comparison with the theory of natural selection. The idea that there is often a good fit between organism and environment predates Darwin. Darwin did of course contribute a great deal of empirical data and rigour to this finding, but the contribution for which he is most celebrated is that he identified a specific process that links the two sides of this observation. Specifically, he proposed a process by which one side can, in a largely systematic way, cause the other to take certain forms, and not others.

    This finding motivates a number of research agendas, of which I here highlight just one: adaptationism. Once Darwin described the causal link between form and environment, he justified explaining organismic form by reference to the corresponding selective environment. This means that for any given trait of interest, the explanatory challenge is to identify the corresponding selective environment(s) in which the trait evolved and is maintained (including the possibility that there are no such environments). Over the past 40 or so years, behavioural ecology, among other disciplines, has achieved vast empirical success pursuing this agenda.

    With this background in mind, let me state simply why I believe that CAT is the right framework for language evolution, and indeed for explaining culture in general: it fulfils the same epistemic function for culture that natural selection does for biology. That is to say, it provides a causal, mechanistic explanation of why cultural items should be expected to gravitate in particular directions, and not others. Just as with natural selection, this finding, assuming it is true, motivates a number of research agendas, including one that roughly corresponds to adaptationism. We could, somewhat clumsily, call this agenda “attractionism”. The basic idea is that the prevalence and stability (or absence) of some cultural item can be explained by identification of the corresponding factors of attraction.

    Language evolution has pursued attractionism with some success, albeit without any reference to CAT. I listed some examples in a preview of SOM that I wrote for the language evolution blog Replicated Typo. However, these findings are not presently unified within one general framework, and it is this problem that CAT can address. Moreover, by framing language evolution in this way, we can, as I said in the précis to this book club, link explanations of structure in language with explanations of structure in other cultural domains.

    Let me finish with some quick replies to three specific and important suggestions that Alberto makes. My hope is that the above background will help to make transparent my perspective on each of these.

    “when domains are close to the “preservative extreme”, it is useful, for pragmatic reasons, to consider them as standard evolutionary systems”
    I agree that this may be a useful idealisation in some cases. But we should nevertheless recognise that cases of cultural propagation that are largely or exclusively preservative still lie on a continuum with other cases. Any truly general framework should address the whole continuum, and this is what CAT aims to do.

    “More than asking which aspects are in general more important, it may be more productive to ask when and why transmission is preservative or reconstructive, and what the consequences are for the resulting cultural dynamics”
    I agree entirely. This is part of the agenda of attractionism.

    “A further step would be to identify which features of languages are due to cultural attraction forces and which features are due to processes included in “standard” cultural evolution models”
    This is a false dichotomy. The processes described in “standard” cultural evolution models are a subset of cultural attraction in general.

  • Mathieu Charbonneau
    Mathieu Charbonneau 25 June 2015 (10:47)

    As a man with no horse in this race, I have no problem in accepting both Alberto’s replicator view of cultural stability (which should really be relabelled as something in the like of the “conservative” approach, since it is clear since Henrich & Boyd’s (2002) paper that the conservative approach needs not rely on cultural replicators) and the cultural attractor view of cultural stability (I found the Claidière & Sperber (2007, 2010) papers right on point as a clarifying a viable alternative to the conservative view). As long as both approaches are used to study the stability of some tradition, acknowledge one another as a relevant alternative causal-mechanistic explanation, and pay due respect to the real psychological processes underlying the stability of the studied tradition (and do not revel themselves into general mathematical models with little empirical grounding), I see no deep opposition between these views. Which one set of mechanisms is supposed to be the most important – faithful transmission plus discriminatory selection vs. transformation circling around a nearby cognitive gravitational point –appears to me as a problem for a case-by-case empirical investigation of specific traditional stability, but not one for a more general, united theory of cultural evolution. I do not want to partake into this debate, and candidly (and quite frankly), I am still a bit puzzled why there is such a debate in the first place.

    But this is not my point. Rather, I would like to suggest a third possible way that traditions might stabilize, and, as I hope to show, whereas it seems to fail to explain actual language stability, the suggested “mechanism” might have some consequences for early (proto-)language evolution, both in phylogenetic and ontogenetic processes.

    As Thom rightly claims in his response to Alberto’s comment: “For many disciplines, the key intellectual question is why the world is the way it is, and not some other way. In physics, for instance, this leads to the search for physical laws. In linguistics, this question sometimes goes by the name of Greenberg’s problem: why are languages the way they are? Biology asks why organismic form is the way it is, and the theory of natural selection is so important exactly because it provides an answer to this question (see below).” (my emphasis)

    I am not sure I agree with how Thom deals with the question, however. It is one thing to explain why things are the way they are given other possible ways they could be, and another to examine the range of way things can be in contrast to ways it is not possible for them to be. In physics, it is a puzzle just how to explain why the laws of physics landed on exactly the right kind of specific values they obey to in contrast to other values (as any slightly different values would not have led to any recognizable universe).

    This kind of question needs not be set in a high-level, quasi-metaphysical framework as in the case of why the physical laws are calibrated the way they are. Something of a similar puzzle can be found in evolutionary biology. For instance, in the last 30 years or so, an important question in evolutionary developmental biology (Evo-Devo) has focused not on why one possible, usually observed variant, strove whereas another observed variant failed to be stabilized (natural selection is usually enough to settle these issues). Rather, a different, key question is asked. Simply: “Why were either known forms possible in the first place, and moreover, in contrast to other, non-observed forms?” In a more localized, biological scenario, the question translates as: “Why are some variants of a given trait (e.g., mutations) possible whereas others are not?” As far as I know, neither the conservative view (defend by Alberto, among others) nor the transformative view (defended by Thom, among others) of cultural stability address this question, neither in fact nor in principle. I would like to suggest that the “internal view” (see below) I suggest here offers an alternative explanation to cultural stability to both conservative and transformative mechanisms, and that it is a an important part of any process of cultural evolution (language evolution included), as it is the only one equipped to explain why some cultural (language) forms are not possible.

    For Evo-devo, the question of why observed forms are possible in the first place is a central one. Major advances in evolutionary biology have shown that what might look like natural selection (conservation or attraction in the cultural cases) is merely due to constraints in the sorts of phenotypes that can be produced at all, i.e., even before any external stability process can act on these variants (a phenomenon referred to as developmental convergence; see McGhee 2007, 2011). A biological form might be stable enough only because it cannot evolve any more. Natural selection or any cultural analog process have nothing to do with such stability. It is simply the space of possible producible variants that restricts what can be stably produced and reproduced again, and which other forms cannot (see Charbonneau (2015) for a cultural analog of this process). This means that a biological form might be very stable only because of the set of physical, developmental, and material constraints they evolve in forces them to stop evolving, and thus stagnate, without any adaptive mechanism being required to explain their stability. This latter form of stability is studied in Evo-devo as an impediment to the evolvability of species (Altenberg 1994, 1995; Wagner & Altenberg, 1996). Such “internal” constraints can have major phylogenetic effects (I say they are internal only because they are not the effect of the immediate selective environment of the trait but merely the effect of the space of produciblevariants (for selection to act, the phenotype needs first to be produced)).

    To get back to the evolution of languages, because this is what we are discussing here, I would be very curious to read any participant’s thought on what sorts of constraints might limit languages to evolve in a limited space of variation. I am no language expert but, for instance, I have always found that the claim that Modern Human languages are infinite in scope fundamentally absurd (Thom suggests, with others, that the infinite scope of languages is a qualitative difference, whereas coding systems can only account for quantitative change, and that this serves as a genuine phenotypic difference-maker between what the code model in contrast to the ostensive model can explain, SOM, p. 47). Of course, assuming some sophisticated generative mechanisms as those of UG, you can produce –in principle- an infinite scope of propositions. But constraints on working memory (not to mention any human being’s limited life span) will limit the complexity of the sentences anyone can actuallyproduce. I dare anyone to demonstrate we can produce an infinite set of sentences, and thus what such a phenotype would actually look like.

    I agree that, when considering Modern Human languages, there might not be any obvious constraints on just what propositions we can produce, but to me this seems more the effect that we cannot imagine a more complex form of language that we are in fact capable of. This observation might seem trivial at first, but I would argue that when this limitation is set in the context of the evolution of languages (for instance what earlier forms of languages might have looked like), it might have some important evolutionary role to play.

    Consider the following. Newport (1990) argues that the differences in language acquisition proficiency at different maturational stages might be an effect of the cognitive constraints in the learner’s working memory. In a nutshell, the older you are, the less proficient you are in learning a language for the first time because your other cognitive capacities are too sophisticated to learn the language in the “normal” way. This is supposed to be a consequence of the fact that ordinary language acquisition depends on a bottleneck-effect imposed by the limitations of the early cognitive capacities that infant of the right age find themselves constrained by. In other words, it is good to have limited cognitive capabilities early on in your development because those constraints simplify what you need to learn, and as your cognitive capabilities increase throughout your development, you can incrementally build a more sophisticated understanding of the language you are nurtured to learn.

    When cast into an evolutionary scenario (assuming ontogeny recapitulates phylogeny), this succession of developmental constraints suggests that early humans were only capable of learning a proto-form of language in a very restricted space of language variation. This is because their cognitive abilities could not allow for much linguistic variation in the first place. If there were any cultural evolution at such early stages, “cultural” language evolution might appear very stable if only because there were not so much possible variants to evolve into in the first place. In such populations, we would observe stable proto-languages if only because of the cognitive constraints. However, as cognitive capacities increased throughout our lineage (e.g., as a larger working memory evolved), the range of possible languages and their range of possible complexity might have broadened. What started with a uniform “proto-language”, uniform not because of conservative nor transformative mechanisms, but because of internal constraints in the space of possible language variants, became more and more subject to cultural evolution as the space of possible language variants broadened.

    Agreed, the “internal” constraints I have pointed at here might not apply to modern languages. Again, I am no language expert. I am only suggesting a different way to look at cultural stability. I have argued elsewhere that such constraints might explain stability in other areas of cultural evolution, especially technological evolution (Charbonneau 2015). Nevertheless, I am very curious to know if any such constraints might apply to modern language capabilities, as a cause of cultural stability that neither falls under the conservative paradigm suggested by Alberto, nor the transformative scenario of Thom.

    Altenberg, Lee. “The Evolution of Evolvability in Genetic Programming.” In Advances in Genetic Programming, edited by K. E. Kinnear, Jr., 47-74. Cambridge, MA: MIT Press, 1994.

    ———. “Genome Growth and the Evolution of the Genotype-Phenotype Map.” In Biocomputation: Comptuational Models of Evolution, edited by W. Banzhaf and Frank H. Eeckman, 205-59. New York: Springer-Verlag, 1995.

    Charbonneau, Mathieu. “Mapping Complex Social Transmission: Technical Constraints on the Evolution Cultures.” Biology & Philosophy 30 (2015): 527-46.

    Claidière, Nicolas, and Dan Sperber. “Imitation Explains the Propagation, Not the Stability of Animal Culture.” Proceedings of the Royal Society B 277 (2010): 651-59.

    ———. “The Role of Attraction in Cultural Evolution.” Journal of Cognition and Culture 7 (2007): 89-111.

    Henrich, Joseph, and Robert Boyd. “On Modeling Cognition and Culture: Why Cultural Evolution Does Not Require Replication of Representations.” Journal of Cognition and Culture 2, no. 2 (2002): 87-112.

    McGhee, George R., Jr. Convergent Evolution: Limited Forms Most Beautiful. Cambridge, MA: MIT Press, 2011.

    ———. The Geometry of Evolution: Adaptive Landscapes and Theoretical Morphospaces. Cambridge: Cambridge University Press, 2007.

    Newport, Elissa L. “Maturational Constraints on Language Learning.” Cognitive Science 14 (1990): 11-28.

    Schank, Jeffrey C., and William C. Wimsatt. “Generative Entrenchment and Evolution.” PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association 2 (1986): 33-60.

    Wagner, Günther P., and Lee Altenberg. “Complex Adaptations and the Evolution of Evolvability.” Evolution 50, no. 3 (1996): 967-76.

    • Alberto Acerbi
      Alberto Acerbi 25 June 2015 (11:29)

      Hi Mathieu, thank you for your comment.

      I will read it with the time it deserves (at a first reading though, it seems to me that the “internal constraints” you talk about are good examples of attractors, by orienting changes/mutations in non-random directions).

      I just wanted, for now, to clarify my position – even though I thought it was pretty clear to be honest. I am absolutely not defending a “replicator” view of cultural stability against an “attractor” view. In my comment (and elsewhere) I have in fact defended exactly what you say in the first paragraph, i.e. that the two views are compatible, and especially, as you say, that this is “a problem for a case-by-case empirical investigation of specific traditional stability” (I almost literally wrote this in So we agree completely, perhaps with the difference I think it is in fact an interesting debate, as long as it is used to clarify specific empirical cases, as I am trying, as others, to do.

      Regarding the replicator/conservative terminology: I agree again. I used the term “replicator” (in italic) to refer explicitly to a paper from Mark Pagel (see the references at the bottom of my comment) and, more generally, to an extremist view of cultural evolution to juxtapose to the attractor view.

      Hope to continue this discussion!