Frans de Waal on language and cognition

BonobosFrans de Waal has published an excellent essay on the relationship between language and cognition in Aeon magazine. Both de Waal and Aeon are very much worth paying attention to, which is why I’m devoting this post to the essay, entitled “The link between language and cognition is a red herring.” Though, as it turns out, that link isn’t really a red herring.

de Waal begins by pointing out that the desire of a number of scientists working on animal cognition to “talk” to their objects of study is rather weird, since animals clearly don’t talk and even if they could — as Wittgenstein argued, and de Waal reminds his readers — we probably wouldn’t understand their meanings.

Moreover, he continues, it is by now well known that when humans themselves verbally explain their own feelings and inner thought processes we cannot trust such explanations. All too often they are the result of after-the-fact rationalizations rather than clear insights into one’s own mental processes. (I do agree with that, but as I pointed out in the past, let’s not take that line of reasoning too far, or all human communication, including scientific one, goes out the window.)

In fact, says de Waal, his “distrust” of language runs deeper, because he is not convinced of the allegedly fundamental role that language plays in the thinking process. Here is where things become interesting, and where I diverge at the least in part from the author.

If language is foundational to thought, de Waal writes, then “speaking two languages at home and a third one at work, my thinking must be awfully muddled. Yet I have never noticed any effect, despite the widespread assumption that language is at the root of human thought.”

Let’s set aside the contradiction inherent in the fact that de Waal is now asking to trust his privileged insight into his own thought processes, something he said a few lines above we shouldn’t do for people in general.

I too speak more than one language and haven’t noticed any particular muddling up of my thought processes, but that seems irrelevant to the issue of the relationship between language and thought. For one thing, I also sense very clearly that I alternatively think in one language or the other, but never both. Even when my Italian sentences come out mixed with English words (or, more rarely, vice versa), that’s because my brain tends to think in one mode or the other, and occasionally has troubles switching back and forth, attempting instead to “translate” a given sentence on the fly from, say, English in my mind to spoken Italian, or vice versa.

A better point brought up by de Waal is our more than occasional inability to express our feelings and emotions using language. Though, again, this seems far less of a conclusive argument that it appears at first. I would argue that feelings and emotive states are not examples of thinking, and that it is therefore not at all surprising that we have a hard time articulating them to others. Indeed, cognitive therapies are in part about thinking through our feelings, so to speak, in the attempt to understand and alter them, if they do not lead to constructive behaviors.

de Waal mentions Jean Piaget, who was not about denying thought to pre-verbal children, and concluded therefore that cognition is independent of language. Setting aside that Piaget’s theories have come under a significant amount of fire in the intervening decades, I think it is a bit of a fals dichotomy (if we want to use the language of informal fallacies) to cast the relationship between language and thought as all-or-nothing. Surely some thinking goes on in both pre-verbal children and a number of other animal species (most likely other higher primates), but that doesn’t mean that language isn’t thereby a precondition for more sophisticated thinking, especially of the analytic type.

Rather surprisingly, de Waal admits that “I consider humans [to be] the only linguistic species. We honestly have no evidence for symbolic communication, equally rich and multifunctional as ours, outside our species,” a refreshing admission of human exceptionalism by a leading biologist.

Of course he also immediately adds that “critical pieces [of human-like behavior] such as power alliances (politics) and the spreading of habits (culture), as well as empathy and fairness (morality), are detectable outside our species. The same holds for capacities underlying language.” Again, this is true, not surprising (given that humans evolved from other animals), and insufficient to deny a major role of language in the thinking process.

It is therefore a bit puzzling when de Waal concludes his essay by stating that “the manifest reality of thinking by nonlinguistic creatures argues against the importance of language.” It most certainly doesn’t. Rather, it points toward the idea that thinking, just like morality — another major interest both of de Waal and mine — evolved gradually on the basis of instinct, and yet made a huge leap forward when Homo sapiens (and, possibly, some of our closest, now extinct, relatives) came on the evolutionary stage. Language is not necessary for low-level thinking, but it is certainly required for the sort of complex, captivating communication — like de Waal’s own books and articles — that human beings are the only species on earth capable of producing.

166 thoughts on “Frans de Waal on language and cognition

  1. Disagreeable Me (@Disagreeable_I)

    Hi Socratic,

    > No, Coel, I’d call that an idea, not a concept.

    I would call it a concept.

    According to a few dictionaries I looked at and Wikipedia and so on, a concept is an abstraction that enables higher order thinking, whereas an idea is more of a particular case.

    Have a look at the opening paragraphs in and you’ll see a description of the usage that Coel and I would support.

    So a chimp may have the idea a human might render as “I think I’ll go fishing for termites now”, but fishing for termites itself is an abstraction of the process of finding a stick and a good termite mound and poking the stick in, just as a termite is an abstraction of tasty wriggly white thing A and tasty wriggly white thing B and tasty wriggly white thing C and ….

    Concepts such as “stick” and “fishing for termites” and “termite” enable the chimp to think thoughts such as the above rather than thoughts such as “I think I’ll find something long and thin and poke it into a big hard thing which contains white wriggly things”, but even that thought requires abstracted concepts such as “long”, “thin”, etc. Without concepts and abstractions of some kind, I think an animal is limited pretty much to simplistic stimulus/response action and cannot learn anything, because learning is more or less the acquisition of concepts (in my language and Coel’s at least).

    > since chimps simply don’t have teaching-explaining communicative learning.

    You keep bringing up how animals are not as good at communicating as we are.

    This is a trivial point we can all agree on. We all agree that the communication that exists between animals is pathetically impoverished compared to what humans can achieve. Language is extremely useful. Humans are exceptional. But the question is whether language is required not only for complex communication but also for complex thought. So your comments along this line are redundant and miss the point in my view.

    Liked by 4 people

  2. Disagreeable Me (@Disagreeable_I)

    Hi Dan,

    The case you have made mostly refers to arguments elsewhere. That’s fine. I would be interested to try to read Philosophical Investigations some time to see if I can make any headway on it, or alternatively to try Kripke. But I’m not going to be able to do so any time soon.


  3. Thomas Jones

    Massimo: “yet [I] find the underlying denial of human exceptionalism simply bizarre.”

    But I don’t think that’s de Waal’s central point when, for example, he writes:

    “But although [ape] gestures are more context-dependent than other signals and greatly enrich communication, comparisons with human language remain a stretch.”

    What is bizarre is my telling one of our dogs to “go left” and my wife with a laugh pointing this out by asking, “Are you aware of what you just said?” 🙂


  4. Philosopher Eric

    I think Robin and others here are making a great point. Sure we can limit the definition of “concept” to language, though that severely limits the “concept” (if you will) of concept. I can’t tell you how often I fail to communicate with someone in the construction trades with mere words, but then I am able to get my meaning across with a quick sketch. I couldn’t say what I meant right, and the person that I was communicating with couldn’t say it right either, but with a picture the person is now able to understand. Yes language is an amazingly useful tool, though sometimes there are better ways to convey concepts/ideas than with language. In fact often you can’t even understand something yourself here until you draw it out.


  5. Robin Herbert

    The claim seems to be that concepts require words. I have given counter examples – specifically a child who has no words at all but appears to be using concepts very successfully.

    Either that child has no concepts, or else the claim that concepts require words is wrong. There is no other possibility.

    I don’t see why it is beyond the scope of this forum to be explicit and say that you don’t think this child is using concepts at all, if that is what you mean.

    But if you do say that, that a child can ask for help about the particular kind of swipe required to avoid a particular kind of object in a computer game, or ask for a video to be transferred from one digital device to another, without having any concept of these things – then we probably just both using the word “concept” in a different way.


  6. brodix


    While the history isn’t exactly clear, it would seem likely the ability to actually start fire arose from napping flint, likely with something with a high metal content, given that would make it heavy and hard.

    I think there is bit of a horizon line issue here, that what we perceive as a fairly clear distinction between non-verbal and verbal communication would have been a feedback process that built on itself, much as language continues to evolve today.

    It seems likely that the linear, rational thought process would have evolved from having to physically navigate and then storing memory of this process, i.e., narration. To then as this process evolved, symbols for the variety of experiences encountered would have become necessary, in a bottom up process.


  7. Daniel Kaufman


    No, that’s incorrect. The arguments apply directly to the points made here about the possession of concepts by non-linguistic creatures. That you don’t accept it is fine, but don’t pretend it doesn’t apply. That a non-linguistic creature can possess concepts, the possession of which involves the capacity to make inferences of various kinds and use terms correctly, is straightforwardly contradictory. This is how the performative account of concepts that one finds in Wittgenstein works. And there are arguments *for* it and against competing views that need to be confronted, before one simply dismisses it.

    It is worth noting that on other views, like the view that concepts are mental representations, it makes even *less* sense to ascribe concepts to non-linguistic creatures, as representations have semantic — i.e. referential — content.

    You can read whatever you like. Or not. I am not trying to convince you. Just speaking to the issues at hand, from the perspective of my knowledge and expertise.

    And now, really, it’s enough. I’m getting flashbacks of the endless, interminable, round and round and round arguments in the days of Scientia.


  8. dbholmes

    Hi Dan,

    “As for your gap-type argument, “Could you explain how…?”, they are notoriously invalid.”

    Just to be clear that was not meant as an argument. Perhaps I should have used “describe” instead of “explain”. I wanted to hear how you would talk about these behaviors, to get an idea if I would agree and so if this was mainly us talking past each other due to different word usage.

    “Wittgenstein treats concepts not as entities to be discovered, but as techniques of using words. To have mastered a certain concept is to have mastered the technique of the use of a certain word in some language or other… Concepts are human creations, made not found. They are comparable to instruments made for human purposes, and their acquisition is comparable to the mastery of the technique of using an instrument. They are rule-governed techniques of word use.”

    From this I take it we are largely talking past each other. You use a very different definition of “concept” (and so likely “thought” and “action”) than I do. And this doesn’t seem to be a thing one can argue about, unlike muddled mind/brain language confusions. It is not wrong for you (and Wittgenstein) to have that usage, I can see how it is a consistent/cogent idea, but neither is it wrong for others to have a broader account… though our different usages will likely lead to confusion when discussing things.

    Toward this end I would still like to hear how you would describe those activities, so I have an idea how I can pitch my wording to create less confusion. If it’s agreeable/catchy enough I may take it on board for general usage.

    I think a solid point of disagreement would be that language and words are wound up with everything we do. Much yes, all no.

    Also, thanks for the detailed replies and links. They are useful.

    Liked by 1 person

  9. Daniel Kaufman


    I don’t know how to be clearer with you than I have been. For the last time.

    “To have mastered a certain concept is to have mastered the technique of the use of a certain word in some language or other. To possess a concept is to be able to use a word or phrase correctly, to explain what one means by it in a given context, and to respond with understanding to its use.”

    At some point in his development, a child develops this technique. He doesn’t have it as a fetus, and he does have it at some later point in time. To ask at what precise point in the development he gets it is to misunderstand how techniques are acquired. Aristotle gives a similar account of how moral virtue is acquired. It too is a kind of technique, the acquisition of he compares to the acquisition of virtues of craft. (techne) And it would be equally mistaken to badger him about the precise point at which the child finally has it.


    I think one of the real problems with the discussions here — and back in the days of Scientia — is the inability to allow things to stand where they are. When people have expressed their arguments and positions as completely as they can, then everyone needs to step back and let it go. The record is there for readers to see and they can judge the arguments for themselves.

    But this endless, manic, borderline compulsive need to keep going on and on and on and on is what creates aggravation, hostility, and ultimately fights.


  10. Daniel Kaufman

    dbholmes: Yes, i suspect that’s right. That there is fundamental disagreement as to what a concept is, which means there will be disagreement to how concepts are acquired. But this is often the case with theoretical terms, where the various disputants cannot agree to allow ordinary language usage to be the ultimate arbiter.


  11. Daniel Kaufman

    dbholmes: It seems to me that the notion of a concept that many are working with here is a somewhat inchoate notion of a kind of mental picture. The problem is that if it is a picture *of* something, then it is representational and that requires a public language (in light of the rule following and private language arguments), and if it is *not* a picture of something, then it doesn’t provide us with any substantive account of what a concept is.


  12. SocraticGadfly

    I/m with Dan. I’ve talked plenty about both the semi-uniqueness of human language, and details of what that means and how it works, and even given it a dash of Hume.

    Per Thomas, it’s nice, or rather, “nice,” that de Waal says that, but Massimo’s bullet point, or nut graf, is that he then undermines himself. And he does.

    Liked by 1 person

  13. dbholmes

    Hi Massimo (& all), I want to be clear that I think language is an extraordinarily important capacity (and tool). It distinguishes humans from all other species (as far as we can tell) and it allows for a much greater range of activity and evaluation.

    I didn’t see de Waal denying this. Just it’s necessity for any and all activity, including thought (though here he would seem to be using a different account of “thought” than Wittgensteinians use).

    Liked by 1 person

  14. garthdaisy

    I always find it strange when when people use the term “animals” to mean “non-humans.” Humans are animals. Yes there is human exceptionalism. But there is also dolphin exceptionalism and bat exceptionalism.

    As for the rest, it seems as usual it all depends on your definitions of words like “concept” “thinking” etc. One can define “concept” as being something only language can create, or one can define it as something that one can have without language. And then those two people talk past one another for a very long time.

    Liked by 2 people

  15. dbholmes

    Hi Dan, I thought my own usage (which appears to include DM, Coel, and Robin… and de Waal?) is the more “common” usage. But that is really more of an empirical question I guess and I don’t know that to be true.

    I am going to read through all the material you provided (as I have time… still have one to go) and I’m going to think over your argument about representations (mental pictures) requiring language. This seems flawed to me, but I get that it requires an argument from my side and not a feeling.

    Liked by 1 person

  16. Daniel Kaufman

    Sorry, I meant dbholmes, not Garth.

    The common definition of “concept”:

    “Something conceived in the mind; a thought or notion.

    An abstract or generic idea generalized from particular instances.

    What exists in the mind as a representation.”

    Representation is at the heart of our common notion of a concept, and representation is referential and therefore, semantic, in nature.

    Wittgensteinians reject mentalistic accounts of representation, which is why the accounts they give are essentially performative.


  17. Massimo Post author


    “I always find it strange when when people use the term “animals” to mean “non-humans.” Humans are animals. Yes there is human exceptionalism. But there is also dolphin exceptionalism and bat exceptionalism.”

    That exactly the sort of commentary that makes me go “groan.”

    Yes, *of course* humans are animals. Do you seriously doubt anyone here doesn’t realize it? Especially your host, who is a biologist?

    That’s most obviously *not* what we are talking about. We are talking about humans as possibly exceptional animals, in one sense or another — in this case, in the sense of language and thinking.

    Second groan for dolphin exceptionalism. Please get back to me when dolphins develop nuclear weapons. Or blog’s.


  18. dbholmes

    Hi Dan,

    “Sorry, I meant dbholmes, not Garth.”


    Anyway, thanks for the definition. I don’t necessarily disagree with what you gave for the common definition… except at the end when representations require references requires semantics. That seems to be the point of disagreement, but I’ll work on arguments for my side (for later).


  19. Daniel Kaufman

    dbholmes: How can something be a representation and not referential? It’s what they are. It’s their referential characteristics that given them content.

    I’m afraid that the referential character of representations is not controversial in philosophy. What’s controversial is how to make sense of it.


  20. brodix


    Given all our perceptions of the larger reality are representations, as we don’t think at the speed of light and so have to distill out these sense impressions and then further distill them through the filters of prior thoughts, beliefs, concepts, etc, it seems doubtful other creatures possess the ability to sense their situations without following a similar process, if less evolved.

    So what is “language?” It would seem to be combinations of sounds and symbols used to organize, edit and arrange our understanding of reality.

    I’m just not seeing anything here really external to normal evolutionary processes.

    Nature certainly developed levels of biological complexity, as well all varieties of signaling, millions, if not hundreds of millions of years ago, that match, if not exceed, our current cultural, civil and economic systems.

    I look around the world today and I see a great deal of grandiosity and ego, from buildings, to militaries, to technological and theoretical modeling, but is that really exceptional? Are our computers really more advanced than biology? Does language really exceed chemistry?

    I still see language as part and parcel of our bootstrapping ourselves upward.


  21. couvent2104

    > Second groan for dolphin exceptionalism. Please get back to me when dolphins develop nuclear weapons. Or blog’s.

    Or when they can play Bach like Glenn Gould, or play football like Maradona, I would like to add.


  22. brodix


    Wouldn’t Pavlov’s bell be representative of food?

    If someone says to you, “lobster tails, saluted in butter,” that would be representative, no? Yet you might have sensations somewhat similar to those dogs.

    If they were to say, “Apollo moon mission,” or “Trump Tower,” or “hunger in Bangladesh,” would the concepts elicit some degree of emotional sense on your part? If they elicited absolutely zero emotion on your part, would there be any content for you, in these concepts? Simply another ringing bell, signifying nothing.


Comments are closed.