Paco Calvo and Natalie Lawrence
Planta Sapiens: Unmasking Plant Intelligence
London: Hachette, 2022, 304 pp.
Reviewed by Uziel Awret
The Haggadic sages of old said that ‘Seekers of wisdom should not live in a city devoid of plants’. What is it about entering a cool garden or greeting the Jasmine plant in our kitchen that is so soothing and uplifting? A phytophile is defined as an organism that thrives around plants and it seems like those sages were phytophiles.
Planta Sapiens by Paco Calvo and Natalie Lawrence is an impassioned but reasoned plea for a radical reorientation of our attitude towards plants that accounts for their ‘inner world’ (Umwelt) with its peculiar ‘phenomenology’ and ‘psychology’. To do that the book appeals to both current research on plant electrophysiology and current ‘scientific theories of consciousness’, concluding that we have good reasons to believe plants possess consciousness and that we need to ‘open our minds’ to this possibility. Hence ‘Sapiens’ in the provocative title.
The reason that we are forced to take this claim seriously is also the reason it is almost impossible for its proponents to convince us of its truth, i.e. we have no idea what consciousness is, or at what level of organization it appears, if at all! However, even the possibility that plants are conscious has important ethical implications that need to be ‘thought out’, and the book certainly does that.
I enjoyed this book very much and recommend it both to plant lovers and JCS readers interested in the evolution of consciousness. Since modern ‘plant cognition’ (seeking to understand the interplay of the plant’s complex electrical and chemical signalling networks with its environment) is a young science, any truth of claims that appeal to it, both for and against ‘plant sentience’, depends on current and pending scientific projects. Those with an academic interest in ‘plant sentience’ who wish to get the most out of this book and get a feel for the current state of the field would do well to follow some of its scientific and historic references, again both ‘for’ and ‘against’ plant consciousness. The book can also be read together with JCS vol. 28, no. 1–2 (Raja and Segundo-Ortín, 2021) — Plant Sentience (which also contains an article by Calvo). I will use plant consciousness and plant sentience interchangeably.
Planta Sapiens weaves three themes, ‘plant phenomenology’, ‘plant structure’, and ‘plant philosophy’ to argue for the possibility of ‘plant consciousness’. It then proceeds to explore the ethical consequences of this very real possibility.
The first theme explores the possible phenomenology of plants and the possible structure of their Umwelt (inner world), but also the way in which attributing ‘phenomenology’ to plants should reorient our own phenomenology of plants from one that engages them as vegetative resources to one that engages them as gentle partners and legitimate ‘others’. The second, structural, theme argues that plants exhibit ‘neural-like’ cognitive/computational properties that can support rudimentary plant phenomenology. The third theme attempts to determine whether the structural claims justify the phenomenological claims by engaging current theories of consciousness and cognition, especially predictive processing and IIT.
Accordingly, the book consists of three parts:
- ‘Seeing Plants Anew’ consists of Calvo’s plea for radical change in the way we view plants, from vegetative resource to partners in the web of life with their own peculiar inner world.
- ‘The Science of Plant Intelligence’ — with the sections Phytonervous Systems, Do Plants Think?, and Ecological Cognition — is the one most relevant to possible ‘plant consciousness’ and to ‘consciousness studies’, and I will concentrate on it.
- ‘Bearing Fruit’ explores some of the ethical consequences of plant sentience and the maverick thinking needed to engage beings that harbour an inner world but that are completely different from us, and which, according to Calvo, is crucial for knowing ourselves.
Plants contain many of the sensory and neural components used by animals, including neurotransmitters like glutamate, GABA, and acetylcholine; cellular messengers like calcium-sensing calmodulin, voltage-gated ion channels; biological sensors such as pressure, temperature, light, and gravity sensors; cellular motors like actin and myosin, and memristor proteins that ‘remember’ their electric history. They also contain gap junction-like structures called plasmodesmata, enabling local electrical signalling with an amplitude that decreases with distance, associated with the xylem system, and fast action potential-based long-range signalling propagating (vertically) along vascular bundles of the phloem. The amplitude of these depolarizing signals does not recede with distance and they propagate at around 20 mm/s, unlike 20 m/s for animals — 1,000 times faster (although those phloem velocities can be much higher in soybeans). Upon reaching their target area they interact (horizontally) with the local xylem system cells and initiate complex local metabolic pathways.
Such signalling can happen in response to tissue injury, temperature differences, and more. Electrical signalling enables plants to respond to changing environmental stress in a quick and coordinated manner. However, both plants and animals owe their ‘neuronal capacities’ to ancient bacteria, retained as organelles, like those whose membranes depolarize during phagocytosis — cousins of the spirochetes whose flagella convinced Lynn Margulis to view electrical brain signalling as an example of ‘frozen motility’ in which locomotion is converted into communication. These ‘older’ building blocks were also crucial to the evolution of the first multicellular organisms. Zooplankton (which appeared shortly after the Cambrian explosion), with its 200 neurons, already contains most of the basic building blocks giving rise to electrical signalling in mammalian brains.
What is so fascinating here is the diversity of things nature has managed to do with these building blocks. Multicellular electrical signalling is not unique to the Animalia and Plantae kingdoms and can also be found in the Protista kingdom. To understand the complex interactions of mammalian neurons and the intricate interplay between their electrophysiology, genomics, and proteomics, it is crucial to understand simpler systems composed of similar constituents that function similarly. To appreciate some of these unexpected functional connections, consider the brain’s default system as alternatively allocating computational resources to either more online, environmentally-demanding, vigilant states or more offline, introspective states. Recent research (Li et al., 2022) on plants’ jasmonate signalling pathway suggests that it functions similarly:
We also focus on the JA signaling pathway, considering its crosstalk with the gibberellin (GA), auxin, and phytochrome signaling pathways for mediation of the trade-offs between growth and defense. In summary, JA signals regulate multiple outputs of plant defense and growth and act to balance growth and defense in order to adapt to complex environments.
One of the biggest problems facing those interested in consciousness studies is the ‘consciousness cut’. You begin with a seemingly unconscious virus or plasmid, or even an m-RNA molecule, and move up the evolutionary tree through archaea, bacteria, protists, plants, mammals, and all the way to humans which are known to be conscious, and then search for the intermediate point at which consciousness emerged. Even a panpsychist that attributes consciousness to the microphysical constituents of the brain still has a ‘consciousness cut’ of sorts, needing to explain when and how such conscious micro-constituents combined to form a macro-consciousness like ours. I think that while a solution of the panpsychist binding problem may help place plants on the right side of the consciousness cut, it is more likely that future research on the ‘neural like’ properties of plants will shed light on the combination problem.
Similarly, eliminativists need to explain whether what they consider as the executive functions of consciousness, such as global availability of information and accessibility to primitive forms of introspection, are realized by coordinated long range electrical signals in a plant and can be considered as a primitive ‘access consciousness’. I doubt whether plants suffer from false problem intuitions about their mental states, but neither do seemingly conscious cats and babies.
While it seems hard to relate plant consciousness to global workspace architecture, it is perhaps possible to relate it to primitive higher-order theories in which local electrical and chemical activity in xylem that relies on both depolarization and plasmodesmata gap junctions is acted on in the right way by the phloem long-range depolarization signals. The authors actually seem sympathetic to such a scenario, if not thinking of it explicitly as a higher-order theory.
The important point here is that exploring the position of our current theories of mind on ‘plant consciousness’ can help clarify those theories even when these theories fail to determine whether ‘plants are conscious’. The first author’s collaboration with Carl Friston and Andy Clark on whether ‘plant intelligence’ can be described by predictive processing seems to have benefited both sides. Plants seem to make both short- and long-range predictions, and for a proper ‘Friston cycle’ one needs to identify the plant systems that generate the predictions, the environmental input that generates the surprise, and the mechanism that minimizes the surprise (by producing the appropriate changes in the plant). Plants do seem to have some of the machinery necessary for predictive processing, like the phloem’s two-way signalling vascular bundles, and the authors consider a scenario in which predictions flow ‘from the deeper layers outward, to the superficial sensory ones’; however, it is not clear at all that plants possess continuously updated generative models that end up modelling their environment. As it stands, we don’t know enough about plants to rule out generative models that can help construct a primitive Umwelt. I assume that to better imagine such successive Friston cycles one would need to fast track the relevant plant processes.
The authors apply predictive processing to ‘plant intelligence’ and not so much to consciousness, but mention that plants make both long- and short-term predictions. Here they could have asked whether combining these predictions by thinking of a plant as a ‘hierarchically-nested prediction machine’ can help provide the plant with the ‘temporal thickness’ central to Friston’s take on consciousness. Of course, Friston is an eliminativist, but establishing the existence of rudimentary temporal thickness in plants’ PP would strengthen the authors’ position.
The book suggests that putative plant consciousness fits well with Clark and Chalmers’ externalism. It makes sense that the plant’s metabolic and computational loops undergo ‘spillage’ extending into the environment and crucial for understanding ‘plant intelligence’ and behaviour, and that they have complex reciprocal interactions with their ‘immediate environments’ (gradually appropriated under the banner of the self). However, it is not clear that such computational externalism is identical to consciousness externalism. Yet one can rightfully argue that when both plants and animals are viewed as extended systems whose relevant dynamic variables span system and environment, their similarities are easier to appreciate. If I am not mistaken, while Friston is closer to eliminativism, Clark is closer to consciousness realism, despite having attempted to use predictive processing to explain the ineffability of qualia.
The book also explores more internalist options. One way of making ‘plant consciousness’ scientifically respectable is by showing that it is supported by an internalist ‘scientific’ theory of consciousness in which systems whose information flow is integrated in the right way harbour consciousness even when these systems possess trivial complexity. If true, that could dramatically increase the prospects of ‘plant consciousness’. Again, both sides stand to benefit from exploring the relationship between IIT and ‘plant consciousness’ because IIT should be interested in approximating the phi values of primitive ‘living’ systems. Such analysis would be more interesting were IIT to claim that consciousness has to exceed a ‘phi threshold’ (like phi per unit volume) similar to the gravitational Bekenstein Bound (see McQueen, 2019).
Perhaps IIT can tell us whether a sequoia tree is more conscious than a zooplankton, but it is also possible that attempting to evaluate phi values of plants will expose weaknesses in IIT. For example, it may challenge the exclusion postulate because identifying the maximally irreducible conceptual structures here is tricky. According to the exclusion postulate, if one neuron in the human brain were to possess a higher phi value than all other neurons, and also higher than the phi value of the whole brain circuitry, it alone would be conscious. Presumably the phi values of mammalian brain circuitry are higher than those of cellular constituents, but this is not necessarily the case for plants. Here it is crucial to estimate the phi value of the phloem system and compare to phi values of cellular constituents. If the phi value of a single component cell is higher than that of the phloem it is possible that consciousness in the single cells with the highest momentary phi will ‘bounce’ from one cell to another (perhaps due to a single protein binding to some cell’s membrane) and that would be strange (Schwitzgebel, see Cerullo, 2015).
The book is clear about the fact that it is dealing with a controversial subject and that there are reasons to reject the possibility of plant consciousness, and the reasons can be both structural and philosophical. For example, in their ‘Anesthetics and Plants: No Pain, No Brain, and Therefore No Consciousness’, Draguhn, Mallatt and Robinson (2021) (arguing against concluding that plants are conscious because they respond to general anaesthesia similarly to animals) conclude:
…plants lack the neural anatomy and all behaviors that would indicate pain… [W]e discuss whether [anesthetics] provide any empirical or logical evidence for ‘plant consciousness’ and whether it makes sense to study the effects of anesthetics on plants for this purpose. In both cases, the answer is a resounding no.
Again, we don’t know enough about consciousness to conclude that systems that lack our exact neural anatomy cannot be conscious. However, Draguhn, Mallatt and Robinson also claim that:
An important limitation of electrical signaling in plants is that, as far as we know, it is all one way without any feedback messaging to allow signal exchanges… (ibid.)
If true, that would mean that significant ‘plant consciousness’ is ruled out by IIT, which relies on computational feedback loops. However, the authors’ claim, that the phloem vascular tubes support two-way electrical signalling, seems well supported. Some of this research is current and accompanied more by clashes between thesis and antithesis than by robust synthesis. A more interesting objection to plant consciousness draws on the assumption that the appearance of biological consciousness is related to the first systems that performed what Andy Clark termed ‘counterfactual emulation’, or ‘what if’ processing. To perform minimal counterfactual emulation all one needs is a generative model continuously updated by internal input instead of the usual external one. It is therefore interesting to ask whether plants have something that functions like a generative model that can be updated by internal, instead of external, input, impinging on their many membrane sensors. If so, it is probably much slower than in animals. However, any success in establishing the existence of such offline processing in plants can be relevant to Richard Gregory’s claim that consciousness is evolution’s way distinguishing online vs. offline processing.
Ginsburg and Jablonka (2019) identify a cluster of properties they deem necessary for consciousness and whose conjunction is sufficient to define an ‘evolutionary transition marker’ of consciousness (you can think of Clark’s counterfactual emulation as such a marker) which they call unlimited associative learning (UAL), arguing that plants lack this capacity. While concluding that the term ‘plant neurobiology’ is unfortunate, they add:
We believe that we are only scratching the surface concerning memory and learning in plants. The attempts to find parallels between the ways that plants coordinate their behavior and physiology and animal neurobiological processes can yield interesting and surprising findings. The study of electrical signaling in plants, which was neglected for a long time, is a good example. (ibid.)
While it is not clear at all that UAL is necessary or sufficient for consciousness, it is clear that we need more research and, as Calvo stresses, this ‘ignorance’ has ethical consequences.
In the JCS volume on plant consciousness, Baluška and Reber (2021) attribute raw unitary consciousness to prokaryotic processes and ‘full blown’ subjectivity to the ancient fusion of a prokaryotic cell harbouring ‘unitary cellular consciousness’, generated by excitable membrane processes, with a flagellum containing a eukaryotic cell that gains its consciousness both from membrane processes and from actin/microtubule cytoskeletal processes. This is a more speculative take on the same subject that does not engage current scientific theories of mind like Planta Sapiens. Nevertheless, in their Coda I, Baluška and Reber (2021) provide a pertinent summary of resistance to more radical theories of plant sentience by Dennett (2017) and Ginsberg and Jablonka (2019) who have:
…acknowledged the remarkable array of behaviors exhibited by prokaryotes, but declined to see them as evidence of sentience, preferring instead to view them as robot-like in nature and driven by genetic programmes. In Dennett’s terms, they have ‘competence’ but lack ‘comprehension’. (Baluška and Reber, 2021)
Which sums up the current situation quite well. The onus is on the proponents of ‘plant consciousness’ to do the research. The good news is that there is room for productive research that might even tip the scale in favour of the ‘plant consciousness’ proponents.
Baluška, F. & Reber, A.S. (2021) The biomolecular basis for plant and animal sentience, Journal of Consciousness Studies, 28 (1–2), pp. 34–59.
Cerullo, M.A. (2015) The problem with phi: A critique of integrated information theory, PLoS Computational Biology, 11 (9), e1004286. doi: 10.1371/journal. pcbi.1004286
Dennett, D.C. (2017) From Bacteria to Bach and Back, New York: Norton.
Draguhn, A., Mallatt, J.M. & Robinson, D.G. (2021) Anesthetics and plants: No pain, no brain, and therefore no consciousness, Protoplasma, 258 (2), pp. 239–248. doi: 10.1007/s00709-020-01550-9
Ginsburg, S. & Jablonka, E. (2019) The Evolution of the Sensitive Soul: Learning and the Origins of Consciousness, Cambridge, MA: MIT Press.
Li, C., Xu, M., Cai, X., Han, Z., Si, J. & Chen, D. (2022) Jasmonate signaling pathway modulates plant defense, growth, and their trade-offs, Interntional Journal of Molecular Sciences, 23 (7), art. 3945. doi: 10.3390/ijms23073945
McQueen, K. (2019) Interpretation-neutral integrated information theory, Journal of Consciousness Studies, 26 (1–2), pp. 76–106.
Raja, V. & Segundo-Ortín, M. (eds.) (2021) Plant sentience (special issue), Journal of Consciousness Studies, 28 (1–2), pp. 7–209.