The evolution (or diachrony) of “language evolution”

Oxford University Press just published the first issue of its new Journal of Language Evolution, seemingly a logical consequence of the increased popularity of evolution-oriented studies at least since the first Evolang conference in 1996. But what is “language evolution”?

One would think that the opening editorial of a new journal would say something about its scope, i.e. that it would tell us a bit more about what falls under “language evolution” in the sense of the new journal. But Dediu & de Boer (2016) are letting us down; it seems that they simply presuppose that everyone knows what “language evolution” is (though their prominent mention of the Société de Linguistique de Paris suggests that they are primarily thinking of the origin of language; on the other hand, they also use the expression “language origins and evolution”, as if these were different things).

But it is not difficult to see that the term language evolution has undergone a significant semantic change over the last twenty years. Until the mid-1990s, it referred almost exclusively the evolution of language as a human-specific trait, i.e. primarily the origin and development of its biological prerequisites. This was something that linguists were hardly interested in for a long time, and Pinker & Bloom’s (1990) paper on “the evolution of the language faculty” was the first to attract widespread attention of linguists with its attempt to link linguists’ hypotheses with a Darwinian perspective.

Around the same time, some linguists engaged in diachronic studies started making more ambitious claims about general principles of language change, especially grammaticalization, and in 1994, Bybee et al.’s book The evolution of grammar appeared, the first book dealing with diachrony that talked about “evolution” instead of “diachrony”. Bybee and her colleagues had been using the old terms “grammaticization” and “language change” (cf. also Bybee 2015) for quite some time in the 1980s, and I had been following their work closely, so it came as a real surprise to me that the book title contained the sexy term “evolution” when it came out in 1994. Maybe this was due to the publisher’s recommendation (when publishers get their income from selling books, they are tempted to use misleading titles, in a way that is apt to confuse readers and thus harms rather than advances knowledge)?

Be that as it may, around a decade later, three ambitious book-length studies were published that drew detailed analogies between biological (Darwinian) evolution and linguistic diachrony: Croft (2000), Blevins (2004), and Ritt (2004). All three contained the word evolution in their titles, but for good reasons: They were inspired by the biological theories of selection and adaptation. To my mind, at the time, the change in the meaning of evolution, compared to Pinker & Bloom, was fairly mild: We were talking about developments governed by a pattern of replication and selection, whether of genes (creating the biological foundations of language) or of linguistic memes (linguemes, in Croft’s terms). It seemed very clear to me that these two kinds of “evolutionary” approaches to language (the biological and the diachronic) were asking completely different questions, even though their proposed answers shared some intriguing aspects.

Now in 2016, we have a Journal of Language Evolution, and confusingly, the first issue is primarily about none of the three uses of evolution that have been familiar to linguists; instead, it is about what linguists call (world-wide) areal typology, as well as the question of whether there is a causal effect of the natural environment on the form of languages, specifically on the existence and the complexity of tonal systems. (There is a target article by Everett et al. (2016) which claims that complex tone systems tend to occur in humid climates, because ambient desiccation reduces the performance of the vocal cords; plus a wide range of commentaries, cf. Collins’s (2016) blog post.)

To be sure, the question whether languages are adapted to climate is related to the endeavors of Croft, Blevins and Ritt, in that it takes one key concept from biology (adaptation), but it does not talk about replication, selection or change at all. Even though these are of course present in the background, the paper is really about a functional relationship, of the sort that has long been widely studied in functionally oriented typology and universals research (e.g. Givón 2010). The issue addressed thus combines observations of world-wide areal typology (as first seriously studied by Dryer 1989 and Nichols 1992) with a proposed functional explanation of the observed patterns.

In addition to Dediu & de Boer’s introduction to the journal, there is also an introduction by Greenhill (2016) to the theme of an “effect of the environment on language”. This paper actually does mention variation, replication and selection, so it seems to put the discussion in a true evolutionary context, but sadly, without mentioning the important work by Croft, Blevins and Ritt. And when Greenhill says that “languages evolve”, one gets the impression that it is not much different from saying that “languages change”.

Indeed, it is my impression that “language evolution”, in addition to being used in the sense of “origin and evolution of language as a human-specific trait”, has simply come mean “language variation and change at a global (or large-scale) level”. Linguists studying change in Old Russian or in lower-class Newcastle English are not likely to use the term “evolution”, but not because their questions are essentially different.

Thus, we now have three senses of the term “evolution” in the context of the study of languages:

(1) origin of language as a human-specific trait (Pinker & Bloom, and also most of McMahon & McMahon 2012)

­(2) language change studied from the perspective of variation and selection (Croft 2000, etc.)

(3) language patterns and language change at a global level (Everett et al. and related work)

And this does not even include grammaticalization (Bybee et al. 1994).

As usual, this semantic change of “evolution” has happened because different communities are interested in different concepts, and people (including scientists) are reluctant to coin new terms for new concepts, preferring to adopt old terms from neighbouring communities. Moreover, even linguists tend to be unaware of semantic changes and thus sloppy about terminological use.

It seems quite possible that the new Journal of Language Evolution is actually interested in studies in all the senses of “evolution”; after all, they are not totally untrelated to each other. But one strong desideratum at least from my perspective would be a paper (maybe in that very journal) that discusses how the various senses of “evolution” relate to each other, and the editors should say what the scope of their journal is (can I submit a paper on grammaticalization in Old Russian?). And for those who use the term “evolutionary linguistics” (or even “Evolutionary Linguistics”, as in Seán Roberts’s recent blog post about the paper by Everett, Blasi and Roberts), it would be good to say what it is (the Wikipedia article seems to include everything, rather naively.)


Blevins, Juliette. 2004. Evolutionary phonology: The emergence of sound patterns. Cambridge: Cambridge University Press.

Bybee, Joan L. 2015. Language change. Cambridge: Cambridge University Press.

Bybee, Joan L., Revere Perkins & William Pagliuca. 1994. The evolution of grammar: Tense, aspect and modality in the languages of the world. Chicago: The University of Chicago Press.

Collins, Chris. 2016. Climate does not affect the way you speak. Post on Humans Who Read Grammars blog.

Croft, William. 2000. Explaining language change: An evolutionary approach. London: Longman.

Dediu, Dan & Bart de Boer. 2016. Language evolution needs its own journal. Journal of Language Evolution 1(1). 1–6. doi:10.1093/jole/lzv001.

Everett, Caleb, Damián E. Blasí & Seán G. Roberts. 2016. Language evolution and climate: The case of desiccation and tone. Journal of Language Evolution 1(1). 33–46. doi:10.1093/jole/lzv004.

Givón, T. 2010. The adaptive approach to grammar. In Bernd Heine & Heiko Narrog (eds.), The Oxford Handbook of Linguistic Analysis, 27–49. Oxford: Oxford University Press.

Greenhill, Simon J. 2016. Overview: Debating the effect of environment on language. Journal of Language Evolution 1(1). 30–32. doi:10.1093/jole/lzv007.

McMahon, April & Robert McMahon. 2012. Evolutionary linguistics. Cambridge: Cambridge University Press.

Pinker, Steven & Paul Bloom. 1990. Natural language and natural selection. Behavioral and Brain Sciences 13(04). 707–727.

Ritt, Nikolaus. 2004. Selfish sounds and linguistic evolution: A Darwinian approach to language change. Cambridge: Cambridge University Press.

Roberts, Seán. 2016. 1st issue of The Journal of Language Evolution: discussion on tone and humidity. Post on Replicatedtypo blog

14 thoughts on “The evolution (or diachrony) of “language evolution”

  1. First, we want to thank Martin for this very timely and important post that made us realize that we assumed more than we should have in what concerns the coverage of the journal. As the executive editors of the Journal of Language Evolution (JoLE), we certainly do not want our potential readership and contributors to get the impression that we limit the scope of the journal to environmental influences on language! JoLE aims to be about how language came to be and how it came to be the way it is today. Biological evolution plays a very important role, but we have learned over the last two decades or so that to really understand this question we must take cultural evolution (and its interaction with biological evolution) into account. Both types of evolution operate in complex environments where functional, anatomical, social and other constraints certainly play a role. The contents of our first issue, especially the target article and associated comments address the idea that even environmental factors may play a role in shaping the present-day distribution of linguistic diversity, but also the Introduction to linguistic diversity, might give the wrong impression that JoLE is concerned by modern (and historically recent) language only. Far from it, but we share the conviction – in essence a type of Uniformitarianism – that one cannot really think about deep time and the processes that made language be the way it is without seriously considering the shape and dynamics of the present. So, to answer Martin’s rhetorical question “can I submit a paper on grammaticalization in Old Russian?”, we’d say certainly, as long as the processes and patterns uncovered are general enough so as to successfully apply to deep time (for example, in the near future, JoLE will feature a discussion of the concept of homology in historical linguistics and an analysis of the evolution of the numeral system in Indo-European, both topics, we believe, of quite general relevance to the journal)!
    It is true that we have perhaps said too little about the scope of the journal in our introduction. One reason was that we did indeed take for granted that our readers would understand what we mean by language evolution. Another reason was however that we feared to propose a scope that would be considered too narrow by some of our target audience.
    In order to prevent future misunderstandings, we will update our call for papers on the journal website:
    Bart de Boer & Dan Dediu (Executive editors of the Journal of Language Evolution)

    • Many thanks, Bart and Dan, this makes some things clearer. However, it is not obvious to me that in order to understand the origin of the capacity for language, one needs to take cultural evolution into account. I’m generally sympathetic to the idea, but I could also imagine that the genesis of the capacity for language preceded any cultural variation, or was at least independent of it (this is of course the nativist view). Moreover, for my ears, “language” (as a mass noun) has to refer to the capacity for language, so the expression “how language came to be and how it came to be the way it is today” cannot really refer to origin of language and the diachronic development of languages, respectively (as you seem to intend) – it can only refer to the origin of the capacity for language and the development of that capacity in recent times (e.g. development of subordination only over the last few thousand years, as was once commonly thought). – But I can certainly see that there are many similarities between biological evolution and diachronic change/development/evolution, and considering these two different things together can certainly be insightful (I just wouldn’t want to give up the distinction).

  2. Diachronic linguists study in detail the dynamics of changes occurring within particular lects – full-fledged modern language resources which already are highly systematic. However, their high-level design features seem to be very robust, and perhaps they could provide observable, empirical evidence for the dynamics revealed by simulation models.

    The double-layered structure of language resources rests on relationships between units of meaning (conceptual categorisations) and conventionalised signals grounded by bodily states and sensorimotor modalities. As such, the structure of any particular lect results from the same long-term glossogenetic processes based on the mechanism of cultural transmission.

    Studies like those cited by Thom suggest that cultural transmission involves a kind of pressure dynamics towards structural generalisation of (any kind of) behavioural patterns. With regard to the communicative behaviour of early humans some researchers hypothesise that it might even be powerful enough to kick-start the glossogeny within an already ecologically functional communication system.

    When concentrating on this replicative aspect of language resources, we perforce leave aside questions about the very capacity to acquire and use them (1), and this is perhaps most evident in historical studies on ‘language change’ (3).

    From a more broad theoretical perspective, there is a need for a conceptual leap to catch the relations between (3) and (2). I would say that language resources with their modern systematicity arose from, or could be “placed between” the pragmatic domain (inferring meanings) on the one hand and the pure ostensive domain of “meaningful” signals (expressed in an available/convenient modality).

  3. Simon Kirby and I addressed these issues in our 2010 review paper, ‘Language evolution in the laboratory’ (2010). I think what we wrote there is still a cogent description of what language evolution is about: “The term evolution can be understood in a wide sense as simply change over time. If so, then the evolution of language might refer both to the biological process whereby the capacity for language arose in our species and the ongoing historical process of language change. However, a narrower concept characterises the field more accurately. Language evolution researchers are interested in the processes that led to a qualitative change from a non-linguistic state to a linguistic one”. The paper is here: (see specifically Box 1).

    To elaborate, let’s distinguish three things:
    1. The biological evolution of the capacity to acquire and use languages.
    2. The change from very simple systems, with few if any of the characteristic features of languages (e.g. no or very little syntax), into a system that is characteristically linguistic.
    3. The change from one characteristically linguistic system to another.
    Historical and diachronic linguistics is classically concerned with (3). This is not to deny that it has something to say about (2) as well – clearly it does – but (2) is not a/the central concern of these fields. Language evolution, on the other hand, is concerned with (1) and (2), but not (3). I would expect that most people in the field would accept at least some version of this characterisation.

    Scott-Phillips, T. C. & Kirby, S. (2010). Language evolution in the laboratory. Trends in Cognitive Sciences, 14(9), 411-417.

    • I think Thom’s formulation, specifically the three things he distinguishes, is the right way of looking at things.

      Personally, what I find very exciting is the transition between items (2) and (3), and the possibility that the change from “very simple systems” into systems that are “characteristically linguistic” (to use Thom’s words characterizing item (2)) is still in progress. Or, in other words, that the uniformitarianist hypothesis needs to be qualified, ie. that languages even as recently as 10,000 years ago weren’t quite like today’s languages.

      Two lines of thought that illustrate this are (a) Talmy Givón’s suggestion that earlier human language was predominantly SOV, and that SVO is a more modern word order; and (b) some of my own work in progress suggesting that simplicity in compositional semantics correlates positively with simplicity in social and political structure, which entails that ongoing increase in socio-political complexity is driving an ongoing increase in linguistic complexity (at least in one important domain).

      To tie this back to the discussion of terminology, I would thus say that actual languages aren’t just changing but also, in some cases, evolving, ie. still on a trajectory whose starting point was no language.

  4. To my mind this problem has much to do with the very word language and its casual, underspecified use elsewhere, i.e. outside meta-level discussions like the one here. And exactly because EoL is such a multidisciplinary endeavour, we would need technical terms for the notions referring to the specific aspects of the phenomenon under study.

    For a start I would plead for making a clear distinction between 1) the notion of a (bio-cognitive) capacity for language, 2) the notion of language resources and 3) the notion of verbal behaviour (without any Skinnerian connotations).

    As a type of communicative behaviour, verbal behaviour would rest on the exploitation of language resources (LR), understood as a particular type of semiotic resources available to animal species. We could conceptualise LR as socially maintained behavioural patterns (see Tamariz, 2014) for expression of concepts, anchored in a communication system that only we, humans, have arrived at over time. Being population-specific, every single instance of LR rests on symbolic conventions which reflect how the community of users of a given lect see the world.

    As for the capacity, researchers in EoL are more and more inclined to speak of a mosaic-like diversity of adaptations. Many scholars by now would agree that the evolutionary dynamics which gave rise to the origins of a shared LR in relative small communities of early humans cannot be explained solely with reference to a biological capacity for language (cf. Hauser et al., 2014). Rather, the earliest LR might have emerged under selection pressures which reflected fairly weak cognitive biases in favour of processing and memorising structural analogies perceived in the signalling behaviour of others and ones own performances. Amplified by repeated episodes of transmission, such early germs of lectal structure could have given rise to the systematic, rule-like organisation of a shared, highly conventionalised communicative behaviour (verbal behaviour).

    With terminological distinctions like this, we can state that researchers interested in the origin of the human capacity for language ask very different questions and use completely other methods than scholars interested in the origin of (the structure of) language resources. The later is currently being investigated using (computational, mathematical) models and simulations. However, as David mentioned, there is also a role to play for the research on ‘language change’. Although it deals with the full-fledged modern lects, it certainly can contribute to the understanding of the evolutionary development of language resources, by adopting the notion of dynamic uniformitarianism, as discussed in Hurford, 2011 (p. 645).

    Anyway, I agree with Sean that there is a need for this kind of clarifying, terminological work. One interesting attempt is the forthcoming Evolang contribution by Jérôme Michaud:
    He proposes a nicely arranged taxonomy, with subcategories of language resources framed at the relevant social structures and time scales.

    Tamariz, M. (2014). Experiments and Simulations Can Inform Evolutionary Theories of the Cultural Evolution of Language. In: Pina, M. & Gontier, N. (ed.), The Evolution of Social Communication in Primates. A Multidisciplinary Approach. New York: Springer, 249–288.

    Hauser, M.D., Yang, C., Berwick, R.C., Tattersall, I., Ryan, M., Watumull, J., Chomsky, N. & Lewontin, R. (2014). The mystery of language evolution. Frontiers in Psychology 5(401), DOI: 10.3389/fpsyg.2014.00401.

    Hurford, J.R. (2011). The Origins of Grammar: Language in the Light of Evolution. Oxford, UK: Oxford University Press.

    • Yes, these distinctions are necessary, and linguists make them routinely. For “capacity for language”, they often say “language faculty”; for “language resources”, they generally say “language structure”, and for “verbal behaviour”, they generally say “language use”. When linguists hear “language evolution”, they generally think about the origin of the language faculty, and when they talk about developments of language structure, they say “language change”. If you talk about language, please don’t ignore the huge literature of linguistics.

      • Thank you for this quick reply. I’m afraid this is exactly what I’m trying to avoid by preferring the terms capacity and resources (verbal behaviour being admittedly less telling than the other two).

        Linguistics is one from many disciplines engaged in the study of language origins and evolution. Having primary linguistic background myself, I personally doubt that it will ever become the leading discipline in this area. That’s why I really prefer to talk about capacity for language, because it seems to be a fairly neutral term for referring to the phylogenetic aspects of language origin. It certainly is more neutral than, say, the Chomskyan „language faculty”. The latter is loaded with theoretical connotations such as syntactocentrism, commitment to a very strong form of nativism, making direct analogy between observed structural complexity of language resources and biologically determined components of the mind/brain(!) responsible for acquiring and processing them, etc.

        As for the glossogenetic aspects, there is a difference between „resources” and „structure”. Talking about language resources allows us to refer generally to the replicative nature of linguistic structure, leaving aside the particular morphosyntactic features of modern lects.

        So, whenever we talk about the high-level design features of linguistic structure (duality of pattering, combinatoriality, compositionality etc.), the term language resources seems to be very handy.

        The distinction between phylogeny and glossogeny was originally made by Tecumseh W. Fitch in the context of cultural evolution (another broad area of research, including a linguistic part). In discussions with peer linguists, I very often feel that there is a kind of reluctance to talk about language in such a broad terms. It has even motivated me to translate into Polish a paper by J.R. Hurford, entitled „Linguistics from an evolutionary point of view”. The proposed terminological distinctions (capacity, resources & behaviour) are inspired by this work, and I coined them in Polish to avoid using the word „language” throughout the text without differentiating those meanings.

        Hurford, J.R. (2012). Linguistics from an evolutionary point of view. In: R. Kempson, T. Fernando i N. Asher, (ed.), Handbook of the Philosophy of Science. Volume 14: Philosophy of Linguistics. Amsterdam: Elsevier BV, 473–498. DOI: 10.1016/B978-0-444-51747-0.50014-7

  5. Simon Greenhill explains his understanding as follows (in a personal comment to me):

    “Possibly the issue is that for people coming from an evolutionary biology background, using “language evolution” to mean “the origin of language” is decidedly odd.

    Why? Because evolution is a dynamic process of change. And it’s not just “change”, it’s a specific type of change – many things change over time: food (raw -> decay), people (infant -> adult -> death), etc. But this is not evolution – evolving systems generate variation, pass this variation to others, and filter out some of these variants based on some criteria. Languages evolve, they don’t change.”

    My reaction to this is that it may well be useful to make a distinction between two kinds of developmental processes (maybe “evolutionary” and “mutational” developments), but insisting that only the latter be called “change” will not work, because people who have the habit of using a certain term for a certain meaning will not change their habits just because some other people think that they should use another term. This practice will only lead to confusion. The only reasonable way to make people’s usage more precise is to use novel terms for making the desired distinction.

  6. Yes, it is an interesting and non-obvious question. No doubt everyone’s definition of ‘language evolution’ (as a research field) is a bit different, but I’m sure there is enough overlap between them for the concept to have a recognisable meaning. But this problem comes in all sorts of guises, e.g. as I’m working on my Evolang talk about the teaching of language evolution at the university level, the trickiest question is what ‘language evolution content’ is in this context (e.g. does language vs animal communication count?). I’m tempted to look to sociology of science and say it is whatever the privileged denkkollektiv (here: the Evolang community) recognises under this term…

  7. This is an issue I think about a lot! And still more needs to be done – a paper on this would be very interesting. But I’d just like to defend our climate paper a bit.

    I don’t think it’s fair to say that our article does not discuss replication and selection a la Croft. In fact, we cite him directly and talk about both concepts. From section 3.1:
    “In order to link this effect to wider change, we take the perspective that the locus of language change is the production and perception of individual utterances in conversation (e.g., Croft 2000; Enfield 2014). In this view, linguistic constructions are the units on which selection applies. Units ‘replicate’ by being used in utterances in conversation, making them available for further replication by other speakers. These constructions vary in form and function, and are in competition given limited time resources, a pressure for efficient communication and the potential for roughly the same meaning to be expressed in many ways. These factors set the scene for Darwinian cultural evolution. In the case of pitch, linguistic elements which require careful control of pitch should be less likely to be replicated in dry climates.
    … an alternative (and not mutually exclusive) pressure may come from the potential for miscommunication. … items should be less likely to be ‘replicated’ from utterance to utterance in dry conditions.”

    So I’m not sure what the distinction is between ‘language change’ and ‘language change at a global level’. They may use different methodologies, but are really asking the same kind of questions. For example, in my blog post, I give two ideas for future research on the question: A ‘macro’ approach which uses phylogenetics and global data (which has an explicit model of evolution); and a ‘micro’ approach where a case study of a handful of particular languages can help confirm or disconfirm the hypothesis (

    But the definition of language evolution is a constantly recurring issue. The evolution of language conference also has a tension between studies on the origin of language and studies on language change. But I think it’s possible to talk about the former by studying the latter. The work of Simon Kirby and colleagues on the emergence of compositionally by cultural evolution (iterated learning experiments) demonstrates that that kind of structure does not require a direct biological explanation. Similarly, work on patterns in global typology helps see *what evolved* more clearly. That is, work on cultural evolution can bridge the gap between language origins and global patterns. So I would define ‘evolutionary linguistics’ as any approach to language which uses the theory of evolution as a guiding principle. To point out a big contribution with this approach (I think): Jim Hurford’s ‘Language in the Light of Evolution’ series.

    But I have one more alternative meaning of language evolution: the study of the necessary ingredients for and constraints on the emergence of complex communication systems. I think this is what I study: I’m interested in potential communication systems and in animals who *don’t* have the kinds of systems we do, just as much as in real languages. I’m going to have to think more about this, though.

    (Regarding the capitalisation of Evolutionary Linguistics, I’m afraid that I do tend to hold the shift button when I’m feeling important. I think it’s a hangover from programming in Java. 🙂 )

    • Sean: your alternative meaning, “the study of the necessary ingredients for and constraints on the emergence of complex communication systems… and in animals who *don’t* have the kinds of systems we do”, is very close to the focus of Protolang ( ), which is “…the requirements for evolving (proto)language… through which hominins have gained a degree of specificity that is not found in other forms of animal communication and cognition”.
      Perhaps we should be clear about how we define protolanguage, as well as language evolution.

  8. Martin’s raising of this terminological issue is welcome and to the point. I would just like to add that what Martin has to say about the new journal would also seem to be equally relevant to the name of the department (of the Max Planck Institute for Human History) that he and I are members of, namely, “Linguistic and Cultural Evolution”. My impression is that in the linguistic part of our department, more work is being done on Martin’s senses (2), language change studied from the perspective or variation and selection, and (3), language change at a global level, than on sense (1), origin of language — even though I share Martin’s feeling that sense (1) is somehow the original sense of the term “language evolution”.

    This also makes me wonder to what extent these three senses are even distinguishable with respect to the “other half” of our department, which studies “culture”. For language, in order to distinguish between evolution (in the traditional sense of origins of language) and language change (as in classical historical / diachronic linguistics), it is necessary to make reference to some version of uniformitarianism, whereby reconstructed proto-languages may not fall outside the range of contemporary languages with regard to their typological features. Thus, while uniformitarianism is assumed in diachronic linguistics, it is, by definition, denied when engaging in the investigation of language origins. But has anybody ever proposed something like the uniformitarian hypothesis for cultural evolution? Off the top of my head I would doubt it, but perhaps someone more knowledgeable could chip in here.

Comments are closed.