paper
active
2023
ⓒ 4
paper:multiple-ways-to-implement-and-infer-sentience

Multiple ways to implement and infer sentience

ByNicolas Rouleauⓘ·Michael LevinⓘAllen Discovery Center At Tufts University, Allen Discovery Center, Tufts University + 11 more
DOI 10.51291/2377-7478.1817OpenAlex W4384024093
Substrate-independent consciousness theoriesSubstrate-independent sentience frameworksMultiple Realizability SentienceSubstrate IndependenceAnimal SentienceSegundo-Ortin & Calvo on Plant SentienceanticipationBasal CognitionBiological DegeneracyBlindsightClassical ConditioningCognitioncognitive functionscontext-dependent habituationCooperationdouble standard in inferring sentienceembodied behavioursgoal-directed behaviours+12 more

TL;DR

Rouleau and Levin argue that the evidential standard applied to inferring sentience is inconsistently applied: behavioural signatures used to attribute subjective experience to non-human animals — goal-directedness, anticipatory reorientation, classical conditioning, game-theoretic risk evaluation, and mimicry — are present in plants yet routinely dismissed without principled justification. The core theoretical move is invoking multiple realizability (Bickle, 2006) and substrate independence (Bostrom, 2003) to argue that felt states, like computation, need not depend on any single physical medium, including the neural circuitry of vertebrates. Plants synthesize glutamate — the most abundant excitatory neurotransmitter in the human central nervous system — and propagate action-potential-like depolarizations along distributed vascular networks, providing at least partial electrochemical homology with animal nervous systems despite radically different anatomical organization. Rouleau and Levin extend the argument beyond plants to metaplastic nanowire networks (Loeffler et al., 2023), robotic systems (Clawson and Levin, 2022), and artificial intelligences, proposing a system-agnostic inferential framework they call a behaviour-based, substrate-independent approach to sentience attribution. Drawing on biological degeneracy — the principle that structurally dissimilar neural regions can achieve identical functional outcomes, as in blindsight responses mediated by subcortical nuclei — the paper argues that if diverse brain architectures can instantiate sentience within animals, non-neural tissues and non-biological substrates cannot be excluded a priori. This implies that current taxonomies of sentience are anthropocentric artefacts rather than principled scientific boundaries, and that a unified, system-agnostic framework spanning cybernetics, bioengineering, materials science, and biomedicine is both possible and necessary.

What to take away

  1. 1. Plants display classical conditioning, context-dependent habituation, and game-theoretic risk evaluation — the same behavioural criteria used to attribute sentience to non-human animals — yet are systematically excluded from sentience attributions without independent empirical justification.
  2. 2. Glutamate, the most abundant excitatory neurotransmitter in the human central nervous system, is synthesized and used for signalling by plants, indicating a degree of electrochemical overlap between plant and animal physiologies that is rarely acknowledged in sentience debates.
  3. 3. Plants propagate action-potential-like depolarizations along distributed vascular networks of aqueous conduits rather than along neurons, making them a concrete biological test case for substrate independence of electrophysiological signalling.
  4. 4. The paper introduces a behaviour-based, substrate-independent inferential framework for sentience attribution, in which evaluation criteria focus exclusively on observable response patterns — anticipation, flexibility, goal-directedness, adaptability — without reference to the physical medium producing them.
  5. 5. Biological degeneracy (Mason, 2015) — the principle that structurally dissimilar neural regions achieve the same functional outcomes, exemplified by blindsight responses mediated by subcortical and brainstem nuclei with independent duplicate maps of retinal inputs (Cowey, 2010) — is used to argue that diverse tissue organizations within animal brains can already instantiate the same cognitive functions, weakening the case for neural exclusivity.
  6. 6. The substrate-independent framework is explicitly extended beyond plants to metaplastic nanowire networks (Loeffler et al., 2023, Science Advances 9(16), eadg3289), robotic systems (Clawson and Levin, 2022), and artificial intelligences, making sentience attribution a question of behavioural pattern rather than substrate identity.
  7. 7. A replication-ready methodological commitment in this framework is to evaluate candidate sentient systems using the same behavioural checklist — goal-directedness, anticipation, flexibility, adaptability, context-dependent learning — currently applied to non-human animals, applied uniformly across biological and non-biological substrates.
  8. 8. The double-standard critique establishes that attributing sentience to non-human animals is itself an inferential leap from observable behaviour, never a direct measurement, meaning the evidential gap between human sentience and plant sentience differs in degree but not in kind from the gap between human sentience and animal sentience.
  9. 9. An open hypothesis raised is that plant sentience is likely the tip of the iceberg of sentient systems yet unknown (citing Ramstead et al., 2019, Physics of Life Reviews 31, 188–205), predicting that the universe contains a far wider distribution of sentient substrates than current neurocentric frameworks can accommodate.
  10. 10. A critical unresolved question the paper raises is whether any purely behavioural criterion can distinguish genuine sentience from functional mimicry of sentience, given that even verbal self-report in humans is classified as an embodied behaviour subject to the same inferential limitations.

Peer brief — for seminar discussion

Rouleau and Levin wrote this piece as a commentary in Animal Sentience (33:30, 2023, DOI 10.51291/2377-7478.1817) responding to Segundo-Ortin and Calvo's review of plant neurobiology, but it uses that occasion to advance a broader theoretical claim about how sentience should be inferred and what substrates can implement it. The paper does three things: it endorses and sharpens the double-standard critique, introduces a behaviour-based substrate-independent inferential framework, and extends the argument from plants to a much wider class of systems including artificial intelligences, metaplastic nanowire networks (Loeffler et al., 2023, Science Advances 9(16), eadg3289), and robotic platforms (Clawson and Levin, 2022). The load-bearing finding is that the evidential standard currently used to attribute sentience to non-human animals — behavioural signatures such as anticipation, classical conditioning, flexible goal pursuit, game-theoretic risk evaluation, and mimicry — is fully satisfied by plants, which also synthesize glutamate (the dominant excitatory neurotransmitter in the human CNS) and propagate action-potential-like depolarizations along distributed vascular networks. Rather than arguing plants definitely have sentience, the paper argues the burden of proof is symmetrical: since all sentience attribution, including that applied to humans, is inferential and behaviour-based, the denial of plant sentience requires positive evidence of exclusion, not merely the absence of neurons. The theoretical apparatus mobilised is multiple realizability (Bickle, 2006, Stanford Encyclopedia) and substrate independence (Bostrom, 2003, Philosophical Quarterly 53(211)), supplemented by biological degeneracy (Mason, 2015, Complexity 20) — illustrated by blindsight responses mediated by subcortical nuclei with independent duplicate retinal maps (Cowey, 2010) — to show that even within vertebrate brains, functionally identical outcomes arise from structurally dissimilar tissue. The implication is a prediction: a system-agnostic, behaviourally grounded framework would unify sentience research across cybernetics, bioengineering, materials science, and biomedicine, and plant sentience would represent merely the nearest instance of a broad distribution of non-neural sentient systems. An alternative method the framework could have employed is integrated information theory (IIT), which provides a substrate-neutral mathematical measure of consciousness (Φ) and would have allowed quantitative comparison across plant, animal, and artificial systems rather than relying on qualitative behavioural checklists. The most obvious point a critical reader would push back on is that the behaviour-based framework risks collapsing the distinction between functional mimicry and genuine sentience: thermostats display goal-directed, anticipatory, and adaptive behaviour under certain descriptions, and the paper provides no principled threshold at which behavioural complexity becomes sufficient evidence for subjective experience rather than sophisticated input-output mapping. The substrate-independence move, while philosophically defensible, effectively defers rather than resolves the hard problem — it broadens the candidate set without supplying a detection criterion that would falsify any given attribution. The paper's scope is also limited to theoretical argumentation and literature synthesis; no new empirical data are presented, which means the framework's empirical fruitfulness remains a promissory note rather than a demonstrated result.

Frameworks (1)

Findings (3)

Claims (15)

Hypotheses (2)

Questions (2)

Original abstract (expand)

Segundo-Ortin & Calvo’s (S&C’s) thorough review of “plant neurobiology” presents evidence supporting the possibility of plant sentience. They make a compelling case that plants anticipate, assess risk, cooperate, mimic, and pursue goals, as do their animal counterparts. S&C point out that there is a double standard: behavioural patterns associated with subjective experiences in humans are considered valid for inferring cognition in non-human animals but not in diverse other systems including plants. We argue that cognitive functions, including sentience, can potentially be achieved by very different systems and their disparate substrates. We offer some context from the basal cognition literature and suggest that the deep insights of neurobiology have relevance far beyond neurons.

Related work— refs + corpus + external arXiv

Cited / in-corpus / arXiv badges show which signals surfaced each row. Multi-source rows weighted higher.

+13 more

Similar preprints — Semantic Scholar

Cited by (1)

Cross-corpus bridges (12)

same_concept_as · Nomic cosine

External markdown files that talk about the same concept as this entity.

  • aboutblank_kb
    Rouleau, Nicolas and Levin, Michael (2023) Multiple ways to implement and inferpapers/edited/Multiple ways to implement and infer sentience_edited.md0.863
  • aboutblank_kb
    System-Agnostic Approach To Sentienceframeworks/system-agnostic-approach-to-sentience.md0.858
  • aboutblank_kb
    Is it possible that sentient systems exist using mechanisms fundamentally different from human neural consciousness?questions/is-it-possible-that-sentient-systems-exist-using.md0.853
  • aboutblank_kb
    Rouleau, Nicolas and Levin, Michael (2023) Multiple ways to implement and inferpapers/linkified/rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md0.853
  • aboutblank_kb
    Rouleau, Nicolas and Levin, Michael (2023) Multiple ways to implement and inferpapers/cleaned/Multiple ways to implement and infer sentience.md0.851
  • aboutblank_kb
    Why do cognitive scientists attribute sentience to animals based on behavioral patterns but not to plants displaying identical cognitive behaviors?questions/why-do-cognitive-scientists-attribute-sentience-to-animals.md0.845
  • aboutblank_kb
    System-Agnostic Approach To Sentienceframeworks/system-agnostic-approach-to-cognition.md0.838
  • aboutblank_kb
    Levin, Michael (2022) Generalizing frameworks for sentience beyond naturalpapers/edited/Generalizing frameworks for sentience beyond natural species_edited.md0.836
  • aboutblank_kb
    Levin, Michael (2022) Generalizing frameworks for sentience beyond naturalpapers/linkified/levin-michael-2022-generalizing-frameworks-for-sentience-beyond-natural.md0.835
  • aboutblank_kb
    Levin, Michael (2022) Generalizing frameworks for sentience beyond naturalpapers/cleaned/Generalizing frameworks for sentience beyond natural species.md0.832
  • aboutblank_kb
    Why should plants be denied the possibility of sentience when they display cognitive behaviors analogous to animals?questions/why-should-plants-be-denied-the-possibility-of.md0.832
  • aboutblank_kb
    How can we determine if non-biological entities like Large Language Models or bioengineered systems possess genuine sentience versus sophisticated information processing?questions/how-can-we-determine-if-nonbiological-entities-like.md0.828