Tuesday, January 26, 2016

Do plants challenge our notion of cognition?

Welcome back everyone. This week's post is by guest blogger Saray Ayala.

A recent book by plant scientist Stefano Mancuso and science writer
Alessandra Viola, Brilliant Green, tells us that plants are intelligent. To many people, “plants + intelligence” is not the most intuitive combination. They will be surprised to find out that the notions of plant intelligence and plant cognition are no news, and no joke either, for a number of plant scientists (in the image, a New York Times article reporting Francis Darwin’s defense of plant intelligence in 1908 at the British Association for the Advancement of Science), and for a group of philosophers who work on minimal cognition (e.g. Paco Calvo, Fred Keijzer). However, these proposals are typically met with intense skepticism and thus remain largely underground. It’s worth unpacking this resistance to the notion of plant cognition. Let’s see how our notion of cognition fares under the green challenge. So, why can’t plants be intelligent/cognitive?

One tempting way of rejecting the possibility of plant cognition is to say, “they are too different from humans.” An anthropocentric notion of cognition makes it easy to declare plants incapable of what by definition is the sole territory of (a select group of) animals. That is an uninteresting, and also unfair, strategy. If we are to offer a serious resistance to plant cognition, let’s do it relying on a more substantive notion of cognition than “whatever humans characteristically do”.

Another class of common responses to the plant cognition challenge is along the lines of “they are too simple”, “they don’t engage in higher-level cognitive activities such as problem-solving”. The idea of plants as simple organisms is common, but studies in plant science show that plants are very complex systems, in terms of both behavior and physiology (examples abound, see reviews here and here). Plants also have proven to be outstanding at solving adaptation problems faced by living organisms on this planet – at the very least, good enough to become the class of organisms constituting the vast majority of the biomass on Earth. Plants may be looked down upon in the hierarchy of living things, but let’s face it: plants are the true dominant species on this planet!

Mancuso & Viola would say we can stop here: their main argument is that complexity and adaptability are all we need to establish plant intelligence. Indeed, if any form of successful complex-problem-solving ability counts as cognition, then it is hard to deny plants are cognitive. But is that an interesting notion of cognition? According to it, any living species has cognition, for it has successfully adapted. Unless we’re content with cognition meaning simply “being alive”, there are more questions to answer: how do plants solve those problems? What are the mechanisms they use? Solving a survival problem by merely reacting to stimuli is quite different from solving it through a set of computations over representational states. Figuring out whether and, if so, how plants compute information is what some philosophers are trying to articulate (see here, here, and here). Now, how do we evaluate plants’ abilities? How do we decide whether plants’ adaptive strategies are complex in the right way? Is there a standard measure of cognition that we can use? In spite of huge progress in figuring out how particular cognitive skills work (e.g. memory, attention, decision-making), Cognitive Science is still lacking consensus about what exactly cognition is. But “we know it when we see it” is not a good working definition, especially when we start disagreeing about what we see. Plants’ reputation as simple reactive organisms may blur our ability to “know cognition” when we “see” it in plants. After all, intellectual abilities have been denied, in spite of blatant evidence to the contrary, to many groups in the past, from women to different racial groups. We should at least try to avoid falling into the same pit again.

A reasonable strategy is to compare plants’ strategies to paradigmatic cases of cognition. This raises new questions, e.g., at which level should the comparison take place? At the functional (computational) level? Algorithmic level? At the physiological implementation level instead? Do we need to find a central controller, á la mammalian brain? But even research in human cognition (e.g. distributed and extended cognition) is abandoning its obsession with the brain. And if cognitive processes (and mental events in general) are multiply realizable in different structures (as Putnam argued long ago), searching for animal-like structures in plants is misguided. To escape the ghost of anthropocentrism, we have to acknowledge that a good and unbiased test for plant cognition can’t set similarity to humans as one of its major criteria. The search for human-like cognition in a clump of stems and roots might be doomed from the start; we should perhaps search, instead, for plant-like cognition. But again, what would it be like?

Plant cognition presents a great and timely challenge for our concept of cognition (and related concepts, such as mind and intelligence). Will our concept of cognition be able to survive the challenge without changing? Conceptual change is a common move, and a necessary one, in scientific progress. Our notions of cognition and mind have changed dramatically over the last century, and the idea of cognition in non-human animals is no longer laughable. Could plants be next in line?

The consequences of acknowledging cognition in plants will go far beyond revolutionizing Cognitive Science and, possibly, restructuring the traditional hierarchy of living things. Just as cognition in non-human animals has been used in arguments for animal rights, the idea of plant rights would gain traction (Switzerland, by the way, is already there). It could reinforce and expand the arguments in favor of deep ecology; it would introduce new parameters into arguments for environmental justice (e.g., justifications of agricultural biotechnology for the benefit of humans would have to meet an extra challenge); finally, it would definitely complicate food ethics. Whatever happens, we can expect the questions about plant cognition to sprout into an exciting and fruitful debate.

Saray Ayala-López
Department of Philosophy
San Francisco State University


  1. Saray, you want to defend the book Brilliant Green which claims that plants are intelligent and have cognitive states. You never mention whether the book defends the claim that plants are conscious.

    Unfortunately you do not mention whether the authors of the book Brilliant Green offer any definition for what counts either as being an intelligent plant or as being a plant with cognitive states, although you suggest that there are a lot of bad definitions out there, ones that are too human-centric. But there is also a tone of optimism in your post that suggests there IS a good definition to be found, and it will imply that plants are intelligent and capable of cognitive states.

    However, the burden of proof is clearly on your shoulders to provide such a definition since the field of botany doesn’t seem to be producing the definitions you are hoping for. The leading contributors in that field treat plants as exotic mechanisms, don’t they? And couldn’t this be for a very good reason? Yes, plants have adapted over the centuries as their environment has changed, and yes plants bend their stems toward the sun instead of away from the sun, but these adaptions and changes have good explanations within botanical science that do not need of the concept of cognitive states. You need to produce an example of plant behavior whose explanation requires us to apply cognitive states to the plants.

    1. Bradley, thank you for your comments.

      While I share Manusco & Viola’s interest in rethinking our opinions about plants, I do not defend what they say in Brilliant Green. I rather take their book as an excuse to raise questions about cognition by criticizing their (unsatisfying) argument for plant intelligence which capitalizes on plants’ complexity and remarkable adaptive abilities. (So, to address your second point: they define intelligence in terms of complex problem-solving abilities.) My criticism of their claims targets mainly over-inclusiveness, rather than anthropocentrism.

      With regards to consciousness: you raise an interesting question about the relationship between consciousness and cognition – a question that hasn’t been quite resolved even in the mainstream accounts of cognition. There is, however, consensus that many cognitive processes unfold beyond consciousness, which suggests that the debate about plant cognition doesn’t hinge on whether plants are conscious.

      With regards to your last comment about the burden of proof: if my main objective were to prove that plants are indeed cognitive (among other things, by tailoring the definition of cognition to fit plants), then yes. But that is not my objective. Finding a good, non-anthropocentric definition of cognition is everyone’s job. Plants offer a good test case. Whether or not they end up having cognitive states will ultimately be an empirical question. What I aim to do in my post is to raise a possibility that we lack the right tools to determine whether cognition is actually there.

      PS. The links contained in the post point to some reviews of different behaviors that are interesting to discuss. One example originally studied by Schwartz A, Koller D. (1986) is off-line nocturnal reorientation by plant leaves in Lavatera cretica (which contrasts with a clearly reactive on-line tracking, see Calvo Garzón (2007) for a brief philosophical analysis).

  2. Saray,
    Thanks for this challenging post. While the concept of plant cognition is not entirely new to me (my professor for neurobiology was engaged in early research on pain perception in plants back in the 80s), it still shakes me out of my dogmatic animal-centric slumber. A shake up I (we) need every once in a short while. Reading your account of the recent efforts of scientists and philosophers reminded me of the arrogance of those in the past, for example Decartes, who believed non-human animals to be mere machines, and their screams at being vivisected while alive as the mere sound of the machine winding down. I'd like to think the progress we, as a people, have made with regard to the recognition of non-human animal cognition won't stop at similar prejudices with regard to plants.

    If we can get ourselves out of the limited thinking that "if it ain't got eyes, or if it don't get up and move, or if it ain't got a brain" then it can't be a cognitive being, we might actually see much more of the value in our world -- value which plants have, not merely for us on our terms, but for themselves on their terms. One nice implication of some of this research is that it might actually offer a more solid conceptual grounding for deep ecological claims of value without having to rely on a pseudo-spiritual or mystical view of non-human nature. It might also demand that we give pause, before we clear-cut yet more landscapes for lumber, or for yet another housing development or quarry or mall.

    1. Christina,
      Thank you for your comment. I totally agree, we do need to regularly shake our “dogmatic animal-centric slumber”. Exploring different kinds of injustice (especially the products of culpable ignorance or implicit bias) made me more concerned about our epistemic responsibilities. Of course, moral concerns are good enough of a reason to attend to what we could be missing about plants (as you suggest with the reminder of Descartes’ attitude towards non-human animals). But even setting moral concerns apart, there’s value in being epistemically responsible. And that is a good enough reason to be, even if minimally, open to considering the possibility of plant cognition.

      Changing gears a bit: I agree with your comment about possible consequences for deep ecological claims of value in nature. An appeal to cognitive capacities of plants will definitely pave the way to arguments different from the ones we’re seeing now.

    2. Chris,
      Should we care about plants? Yes, but not for the wrong reasons. We should care about the success of plants in the human food chain, or else we risk humans losing their food supply since animals depend upon eating plants. We should care about a forest fire destroying all the plants in the forest because of future erosion problems and the loss of animal habitats and the aesthetic blight. But we shouldn’t care about plants because they have a degree of consciousness. Consciousness probably does come in degrees, but we already have a fine term for the exotic behavior of plants. It is called being “alive.” Botanical researchers recognize this when they refuse to call plants “conscious” and instead treat plants as exotic, living mechanisms whose behavior is far more complex than that of a sandstorm or a lightning strike.

    3. Brad, thanks for your reflections. I think we should care about all living things not only for their value to us on our terms, but also for their value on their own terms. That it might be difficult for us to assess what those terms might be, shouldn't justify our neglecting them. If it turns out that some plant life is also conscious life, I would want us to be concerned about that, specially about our use of that organism. That our existence depends on them should make us pause to be more considerate of them and of their value on their own terms, not less so.

    4. Saray, thank you for your response. Yes, I agree that it is important to consider non-moral responsibility, especially epistemically responsibility and consistency, as these serve to ground or anchor our other moral and political claims. I believe the more we know about our fellow beings and their capacities, the more we situate ourselves to be better decision makers in other regards, including what we eat, where we build our living and productive spaces, and what ultimately deserves our protection (or being left alone). That we might not like the implications of what we learn should not be underestimated, though.

  3. Saray, thanks for this interesting post.

    If we were willing to privilege cognitive science, the field that defines itself as the study of cognition, it seems like we should allow that any system that performs computations and stores the results in memory for later use is engaging in at least a rudimentary form of cognition. I'm not sure how much value this would have for deep ecology or proponents of environmental justice, though, since it's an ability that we associate with pocket calculators.

    My students get tired of hearing me say this, but I think to make progress thinking clearly about cognitive/epistemic abilities we have to be willing to treat them as matters of degree. To me the important question here might not be whether plants are capable of cognition at all, but what grade of cognition they can perform. I know next to nothing about this, but I wonder if the salient behavioral difference between plants and animals is locomotion, and the corresponding cognitive difference here might be the ability to use stored information to make predictions about the effects of its own behavior.

    1. Randolph,

      Thanks you for your comments. I am glad you raised the question of treating cognitive abilities as a matter of degree. I agree this is the right way to go. In fact, the philosophers working on unusual forms of cognition are trying to articulate a notion of minimal cognition, emphasizing, as you say, not whether or not plants or bacteria are cognitive, but rather the kind and degree of cognition they have, and how interesting that degree is. Some offer a (debatable) definition of minimal cognition in terms of sensorimotor coordination, based on a functional comparison with animal nervous system (for example Duijin, Keijzer, and Franken 2006, who take Escherichia coli as a case study, here
      http://staffwww.dcs.shef.ac.uk/people/A.Sharkey/minimal.pdf )

      About locomotion: locomotion may not be necessary “to use stored information to make predictions about the effects of its own behavior”; how about predicting outcomes of a changed salinity balance or release of a chemical into the soil or air? Even if locomotion were found necessary to possess the cognitive property of using stored information to make predictions about the effects of its own behavior, plants actually do move, but at a different time-scale. In that sense, this cognitive property is not impossible in plants. We need to determine whether cognition requires a particular time-scale. Again, we find ourselves wondering whether a strict comparison with the time scale characteristic of animals is fair.

      Regarding your first comment on how much value plant cognition could have for arguments in deep ecology: you raise a very interesting point. I agree that acknowledging certain skills in artifacts does not make us attribute any special value to them. It may turn out that the key point is the combination “living organism + minimal cognitive skills”.

    2. I like your point about time scale a lot, especially since so many other great insights in science have depended on our ability to reboot our sense of what is possible in accord with discoveries about how much time is really available (geology, cosmology, biological evolution) as well as how rapidly incredibly complex interactions (cellular metabolism, computation) occur.

    3. Sorry I'm jumping into this a bit late. I had a question about what makes an account of cognition anthropocentric. Two answers come to mind.

      1. Cook the account to make it such that all and only people have cognitive abilities.
      2. Build the account up from paradigmatic examples of things that are clearly cognitive (i.e., people).

      I agree that 1. is objectionable and anthropocentric. 2. is also anthropocentric (in an important sense), but it doesn't strike me as objectionable. It's still an open, empirical question whether or not things in addition to people have cognitive abilities. This is how I'm assuming the cognitive scientists Randy mentioned did it: any system that performs computations and stores the results in memory for later use. I'm assuming they came to this account by thinking of what's essential to the cognition that people do. It turns out, of course, that not only people do this (pocket calculators, etc). But still anthropocentric, right?

  4. Kyle, I'll be interested to hear what Saray has to say regarding your question about the charge of anthropocentrism.

    As far as how cognitive scientists actually proceeded, I think it might be more accurate to say that cognition as computation has its roots in thinking about the possibility of machine intelligence. I think the tradition that approached cognition anthropocentrically (your 2) tended to characterize it more abstractly as the ability to use concepts and/or natural language.

    1. But that doesn't undermine your basic point that people studying the possibility of machine intelligence were still using human intelligence as the paradigmatic case.

    2. That's right, Randy. I'm wondering if there's any good alternative to doing that -- being anthropocentric in that sense. I suppose there's this third possibility:

      3. Deliberately devise an account of cognition so that plants count as doing it, too.

      But 3. looks problematic for the same kind of reason that 1. is.

  5. Cognition exists in extreme variety, a variety rooted in evolutionary continuity. Tests for cognition must fit the species being tested; each species (or even group of interconnected-interdependent species) has its/their Umwelt/en (see Jakob von Uexkull) or niche-related point-of-view. Testing outside of that creates bias. Understanding that niche is time consuming, difficult, and according to some never fully possible (constrained by human perspective). Animals and plants are thought to have evolved multi-cellularity separately but still retain aspects of our single-celled ancestor. Plants and animals have converged on similar traits as well. What would be a quality argument for pre-emptively denying the existence of cognitive processes in plants? Plant cognition may vary as much as animal cognition does, and be inseparable from interspecific and ecological relationships. A neurochauvinistic denial of plant cognition (so much more to deny: volition/reason/etc.) ignores our ancestral and convergent commonalities, and ignores the qualitative importance of our differences. It also ignores ecological context. I suggest Plants as Persons by Matthew Hall (2011) - a great and all too short book on ethics, cultural shifts in human perspectives of plants.