NOTE: this is a working draft, not a finished product. feel free to contact me with constructive feedback, but please do not quote or distribute this draft, although you may link to this page, which will be updated.

Which animals possess consciousness? When and in which lineages did consciousness evolve? These questions are being actively researched, and it is fair to say the current state of the science is characterized by uncertainty and debate rather than a neat consensus. In fact, I think it is accurate to say that a full-blown scientific revolution has been taking place over the last 25 years or so, in the sense that the dominant way of thinking about and relating to the topic in the scientific community has been replaced.

Over this time, attitudes toward animal consciousness have shifted outside of academic science as well, with legal protections being extended to new groups of animals in many countries, and even with the fairly dramatic occurrence of a change in official position by the Catholic Church to extend consciousness and spiritual agency to non-human animals, which is controversial and seen by many as a reversal of orthodoxy.

I recently wrote, together with Dr. Jonathan Birch, a revision of the Stanford Encyclopedia of Philosophy entry on Animal Consciousness, which I first worked on 10 years ago when I did the last revision. This revision gave me an opportunity to comprehensively review the literature and reflect on what has changed over the past 10 years, and while that work informed the revisions to the entry, I decided to reflect on the topics at more length in this article.

In those 10 years, the field of animal consciousness research has changed dramatically, going from a fringe inter-disciplinarity to a proper subject with its own journals, a bunch of books, and a recognizable landscape of several more or less mature (or at least coherent) theoretical views. Those years have seen a proliferation of new scientific research on consciousness in non-human animals. This article will review that research in an unabashedly opinionated way, rooted in my own theoretical perspective, focusing on the findings that I think are most salient and informative.

Intro

At the turn of the century, it was an entrenched orthodoxy that subjective consciousness could not be studied in nonhuman animals. In 2012, a group of scientists signed a publicized declaration accepting the likelihood of consciousness in “non-human animals, including all mammals and birds, and many other creatures, including octopuses.” (Cambridge Declaration on Consciousness, 2012). In 2024, a similar declaration gave a slightly more detailed assessment with graded evidence: “First, there is strong scientific support for attributions of conscious experience to other mammals and to birds. Second, the empirical evidence indicates at least a realistic possibility of conscious experience in all vertebrates (including reptiles, amphibians, and fishes) and many invertebrates (including, at minimum, cephalopod mollusks, decapod crustaceans, and insects).” (NYU Declaration) Both declarations indicate uncertainty about the precise distribution of consciousness outside of birds and mammals, reflecting that researchers have advanced several possible scenarios for which groups have and lack it, and when it emerged in evolutionary history. However, the uncertainty represents active empirical and theoretical research, rather than the attitude of helpless mystery that was common early in the century (examples)

In the late 90s and early 2000s, most research on animal consciousness was largely dedicated toward defending the very possibility of studying it as a topic against skeptical attacks. Academic philosophy of mind was largely occupied with the debate over whether or not consciousness could be considered ‘physical’ (Mary room thing). The most influential theoretical work on consciousness had a marked tone of skeptical rigor, as analytic philosophers mostly struggled to deal with David Chalmers’ framing of the ‘hard problem of consciousness’(1995), and Ned Blocks’ framing of the problem of phenomenal vs access consciousness (1995). These methodological quandaries seemed to cast a pall over empirically minded philosophy, as if no work could proceed that did not solve them on their own terms.

The emerging interdisciplinary field of consciousness studies was largely focused on the problem of isolating the ‘neural correlates of consciousness’ in humans (and primates insofar as they could serve as model organisms standing in for humans), but in a way that was unfortunately not placed in evolutionary context or necessarily grounded with good phenomenology.

So workers in animal consciousness were still basically occupied with Griffin’s project of convincing biologists that animal behavior and theory of consciousness were in any way compatible, much less synergistic.

Arguments were generally less directed at evaluating or even considering alternate theories of animal consciousness, and more focused on defending basic concepts and fending off skeptical arguments that the entire project of studying animal consciousness was futile or incoherent. In other words, we used to argue about whether we could study animal consciousness instead of actually studying it much. In the 2010s, more work has focused on investigation of specific components or levels of consciousness, looking at how animals subjective experience of their world plays a role in their behavior and ecology.

One of the most interesting trends of the last decade has been researchers beginning to go beyond the binary question of whether a given group of animals possesses consciousness, distinguishing specific states, forms, gradations, or dimensions of conscious experience, in order to map them phylogenetically:

  • Sentience, or the ability to feel affective valence (to suffer and enjoy)
  • Noetic consciousness, or the ability to know, understand, semantic grasping
  • Mental maps (Tolman) or images (Feinberg and Mallatt 2016)
  • Mental time travel, the ability to imagine or think about the past or future. - Self-awareness (in various senses: embodied/sensorimotor, social, temporal, narrative, etc.)
  • Intersubjectivity, collective intentionality, ‘theory of mind’
  • Language-structured thought (reference, syntax, recursion, or other properties)
  • Goal-directedness, teleology, purposiveness; Tolman,Trestman, Ledoux 2019, Jablonka and Ginsburg 2019
  • Personhood (Singer et al 1994, Davis , Marino 2013)

Multi-trait and dimensional approaches can create space to compare and contrast forms of consciousness in different forms of animal life, as required for an animal-centric (rather than anthropocentric) study of animal consciousness.

These shifts in the field indicate a true scientific revolution has been taking place, i.e. a new way of doing science…(paradigm)

This theoretical reorientation, toward exploring the diversity of conscious experience in animals rather than agonizing endlessly over species-solipsistic skepticism, exemplifies what Lars Chittka has called a ‘Copernican Revolution’ (2024, NYU conference) currently underway in animal consciousness studies.

Copernicus’ advancement of the heliocentric theory of the solar system shook the European worldview to its core, by revealing that humans do not occupy the spatial center of the universe.

As the scientific community accepts that consciousness is not uniquely or specially human, this urges a change in thinking about consciousness itself: human consciousness is no longer the core or default model of consciousness–the center of the universe of consciousness. Just as Earth is just one world among uncountably many wildly diverse planets in the observable universe, human consciousness is just one way of experiencing life among the many, wildly diverse forms of life we can observe right here on earth. As we regard animals more and more different from ourselves in terms of their bodies and their ways of relating to their environment and one another, we regard the possibility of conscious experiences more and more different from our own.

But how can we know the truth about what other animals experience? Human experience itself varies tremendously, and we struggle famously to accurately describe even our own experience of life, much less other people we know. Cultural gaps can make the actions and expressions of others seem incomprehensible. How can we hope to understanding others across the biological gulf between distantly related species?

Evolutionary origin scenarios for consciousness

# Scenario label Phylogenetic starting point Main idea
1 Human-unique / Higher-Order-Thought (HOT) < 1 M.Y.A. — language-using Homo Consciousness requires language-based self-reflection; only humans qualify.
2 Dual cortex-dependent origins (mammals + birds) Carboniferous period, _ ~320 M.Y.A._ — separate events in early mammals & early birds Laminated (or pallial) cortex deemed necessary; reptiles, amphibians, fishes excluded.
3 Amniote common-ancestor origin Carboniferous Unified thalamo-cortical / limbic “world model” emerged once in early amniotes.
4 Vertebrate-wide midbrain origin Early Cambrian, ~ 520 M.Y.A. — base of Vertebrata Subcortical midbrain circuitry sufficient; cortex only enriches content.
5 Triple convergent origins Cambrian–Ordovician window Consciousness evolved independently in vertebrates, cephalopods, arthropods via large brains + distal senses.
6 Pan-metazoan nerve-net origin Pre-Cambrian, >600 M.Y.A. Diffuse nerve nets already generated “overall sensation” in earliest eumetazoans.
7 Cellular / pan-biotic origin Origin of Eukaryotes or origin of cellular life, >1 B.Y.A. Every living cell has proto-experience; consciousness co-extensive with life itself.

Human-unique / Higher-Order-Thought (HOT)

  • Proponents: Carruthers 1998–2023; early Dennett; Davidsonian language theories

  • Mechanism / evidence hinge: Phenomenal consciousness is argued to require syntactic language.

  • Predictions: No non-linguistic animal (apes, cetaceans, corvids) can possess genuine subjective experience; metacognitive error-monitoring in animals must be sub-personal heuristics.

Dual cortex-dependent origins in mammals and birds

  • Proponents: Rose 2002; Key 2015; narrow interpretations of Edelman’s dynamic-core model.

  • Mechanism: Laminated neocortex (mammals) or convergently laminated pallium (birds) is the minimal neural substrate; thalamo-cortical re-entry indispensable.

  • Predictions: Sharp functional cliff at mammal/bird boundary—e.g., fish pain responses are nociceptive reflexes, not felt; reptile motivational trade-offs should be absent.

Amniote common-ancestor origin

  • Proponents: Edelman 2003; Cabanac 2009.

  • Mechanism: Reentrant thalamo-cortical / limbic loops unify affect and perception into a “world model” once in early amniotes, inherited by mammals, reptiles, and birds.

  • Predictions: Reptiles, but not amphibians or fish, exhibit motivational trade-offs, emotional fever, play.

Vertebrate-wide midbrain origin

  • Proponents: Merker 2005, 2007; Panksepp 2017; Feinberg & Mallatt 2016.

  • Mechanism: Tectal–isthmic–basal diencephalon circuitry forms an integrative, value-laden egocentric world model; cortex elaborates but does not create consciousness.

  • Predictions: Primary (anoetic) consciousness in all vertebrate classes—including cartilaginous & bony fishes; behavioural gradations track neural elaboration, not presence/absence.

Triple convergent origins

  • Proponents: Feinberg & Mallatt 2016; Trestman 2013 / 2018 (Complex Active Body); Ginsburg & Jablonka 2019 (UAL).

  • Mechanism: Consciousness has a computational role in complex cognition that is implemented differently in the three taxa. Authors point to multisensory integration, predictive processing, spatial and object-oriented cognition, and complex learning, as functions that increase the likelihood of consciousness.

  • Predictions: Octopuses, bees, and crows converge on similar cognitive feats; sessile or simple-bodied descendants (clams, barnacles) lack consciousness despite neurons.

Pan-metazoan nerve-net origin

  • Proponents: Ginsburg & Jablonka 2007; Bronfman et al. 2016.

  • Mechanism: Diffuse nerve nets already support global sensory states coupled to associative learning (“overall sensation”)—a minimal phenomenal field.

  • Predictions: Cnidarians and ctenophores display rudimentary valenced behaviour; only non-neuronal animals (sponges, placozoans) are unconscious.

Cellular / pan-biotic origin

  • Proponents: Margulis 2001 The Conscious Cell; Reber & Baluška 2019, 2023 (“Cellular Basis of Consciousness”).

  • Mechanism: Autopoietic membrane dynamics, ion flux, and cytoskeletal signalling endow every living cell with primitive subjective valence.

  • Predictions: Proto-experience in bacteria, plants, protists; consciousness ceases only at death. Requires novel micro-behavioural methods to test.

Human solipsism and exceptionalism

Humans are clearly conscious, in that the concept of ‘consciousness’ was invented by humans to describe their own condition. If you are having an experience of reading this article, presumably you are both human and conscious.

There is a long history of the view that consciousness is unique to humans, although it has probably never been the dominant view in the world—it is in fact rather peculiar Europe and the Middle East, beginning in Rome and continuing with medieval Christian and Islamic religious philosophy. The Hindu, Buddhist, and Jain intellectual traditions which are the cultural background for much of the world’s population hold animal consciousness as an assumption. Similarly, belief in animal consciousness is also important in many indigenous knowledge systems from around the world (Trestman and Birch 2025).

20th and 21st century American and European science unfolded against a cultural background of 1) religious orthodoxy about humans’ unique role in creation, 2) Descartes’ influential identification of consciousness with a capacity for rational cognition and subsequent rejection of sentience in nonhuman animals, and 3) 20th century Skinnerian anti-mentalism.

Renee Descartes argued, if a hypothetical, all powerful demon overwhelmed your senses and even memories with deceptive illusions, you could at least still be certain that you existed—not ‘you’ as the particular human person you take yourself to be, with your body and identity and possessions and social network intact, but at least ‘you’ as a subject of consciousness—someone who can think and feel. This is the meaning of the famous dictum “I think, therefore I am”. If I’m experiencing anything at all, I can at least know that I exist in order to experience. Because this does not apply to anyone else’s thoughts and feelings, one’s self-knowledge can be known in a way that other-knowledge cannot. While my own feelings cannot possibly be an illusion, the feelings of the person who appears to be crying in front of me could, as far as I know, be a phantasm created by a demon, or a dream, or a ‘the Matrix’-style virtual reality technology.

‘Solipsism’ refers to this difference in certainty in our grasp of our own minds and the minds of others, and also to the corollary that if we only treat what is certain as real, than the minds of others are not. If anything that could be an illusion must be distrusted as such, than the existence of other minds, as well as the particular thoughts and feelings they may appear to convey, all must be distrusted, regarded as unreal.

While it is largely agreed that there is no way to disprove solipsism in a logical sense, philosophers seem to agree that is simply okay as an expedient to move past it and accept the reality of the minds of others. What this result shows is that the most important things that we ‘know’ aren’t really certain, rather than showing that we should ignore the feelings of others because we can’t prove those feelings exist with logical certainty. However, the same standard has generally not been used for nonhuman animals. While only a few philosophers have actually ever put forward positive arguments to the effect that consciousness is human unique, unequal application of skepticism about mind in humans and nonhuman is much more common, with many philosophers being willing to dismiss nonhuman consciousness as an unknowable mystery. This air of mystery allows forms of denialism–I’ll discuss two, ‘pragmatic’ and ‘scientistic’, to operate

Carruthers, following Davidson, has argued that to consciously experience anything, you need the concept of experience, which requires language. Dennett argued something similar, although with the twist of illusionism: consciousness is not real, it’s an illusion or imaginary conundrum due to inaccurate language; so having it requires language.

I’ve previously argued at length against both Dennett’s anti-realist view (1991, 1995), and Carruthers’ Higher-Order Thought position (1998a,b, 2000) (See Animal Concsiousness and Does the stream of consciousness have a rate of flow?). Dennett’s views, controversial but never very committal, evolved over time to admit some animal consciousness: “Human consciousness is to animal consciousness roughly as language is to bird song. Birdsong communicates, but not much.” (Dennett 2020 ). Carruthers’ views do not seem to have changed much (2018).

No new philosophical argument that phenomenal consciousness is unique to humans, i.e. that there is nothing it is like to be anything in the world other than a human, has (to our knowledge) been published.

To my knowledge, there is no other argument that concludes that only humans have it. However, in the absence of positive arguments or claims for the human-uniqueness of consciousness

However, far more widespread than any explicit argumentation based on evidence and theory, have been what we may call the pragmatic and scientistic modes of denial. Pragmatic denial amounts to simply ignoring the experience (and welfare) of animals because it is inconvenient, and silencing or avoiding arguments and evidence for animal consciousness, rather than directly countering them based on an evidential assessment. This stance is neither very ‘philosophical’ nor very ‘scientific’, although both philosophers and scientists have done a lot of it, and it has likely more causally influential on human belief and behavior than more intellectual forms of denial. Example include denying that fish or lobsters are sentient because that would imply that catch and release angling and live-boiling, respectively, are cruel practices (See Great Fish Pain Debate, Wallace 2004), or holding that chickens, even if sentient, are so stupid that their own condition doesn’t matter to them (Davis 2014). Researchers may be subject to this bias as well, whether their subjects are dolphins or bees (Marino ; Chittka 2021 bit on Von Fritsch and nociception).

Scientistic denial refers to treating animal consciousness as, if not probably unreal, ‘outside the scope of science’ and therefore subject to ontological erasure. The usual justification for this scoping is the assumption that verbal report has a special evidential relationship to subjective experience that other forms of behavior lack; Only humans can give verbal report, so this gives human consciousness a unique epistemic status. (Dennett 1990s, Key …, Ledoux …, MS Dawkins …, ). However, as increasingly powerful large language model (LLM) artificial intelligence technology has shown, it is not language-like string production per se that offers evidence of consciousness, but the adaptively goal-directed, affectively motivated, use of language to communicate, which is an inherently social behavior. Indeed, it is probably futile to fully understand human language outside of the context of goal-directed, affectively motivated, social behavior, since this is how humans learn language (Trestman 2015). We interpret human language behavior as evidence of human experience because we already take humans to be conscious; our understanding of the experience of the speaker guides both our interpretation of sounds as language, and of the meaning of what is said in any instance. Further, as much of the research cited in the remainder of the article demonstrates, a new generation of researchers has expanded the scope of science to include animal consciousness by finding new experimental ways to probe aspects of subjective experience.

As animal consciousness research has gained momentum in the scientific and philosophical communities and most researchers have come to accept that at least some nonhuman animals (e.g., bats) have at least some form of basic phenomenal consciousness, the focus has shifted to two types of debate: (1) relative evidence for consciousness in specific contested taxa (e.g. fish), and evolutionary origin scenarios based on the inferred distribution of consciousness; and (2) attempts to delineate specific forms of consciousness with narrower distributions. The latter are often paired with attempts to specify forms of consciousness which are human unique, particularly if they can putatively explain other features of distinctively human behavior. A number of authors have put forward different hypotheses about human-unique forms of consciousness, none of which are uncontroversial:

  • Logical recursion and the resulting transformation to language-based thought (Pinker, 1994; Hauser, Fitch, and Fitcsh 2002)
  • Self-consciousness or ‘autonoesis’, mental time travel (Tulving 2005; Suddendorf and Corballis 1997)
  • Intersubjectivity, ‘theory of mind’, collective intentionality (Tomasello et al 2005)
  • Concept-mediated emotional experience (Barret 2017, 2022)

Mammals and other Vertebrates

Humans have much in common biologically with all mammals, and several arguments have been put forward to the effect that consciousness must be shared across mammals, because the neurophysiological traits and functions identified as critical correlates of consciousness in humans are ancestral traits shared by all mammals. The close evolutionary relationship and biological similarities between humans and other mammals tend to support not just a high confidence that other mammals have consciousness per se, but confidence that their experience shares many common features and similarities with human experience. The 2024 NYU declaration emphasizes uncertainty about the status of other vertebrates (reptiles, amphibians and fishes) relative to mammals (and birds), granting a ‘realistic possibility’ of consciousness to the former and ‘strong scientific support’ for the latter. This uncertainty represents active research and ongoing scientific debate; what specific differences between vertebrate lineages can help us to determine the distribution of forms of consciousness?

Edelman proposed that consciousness is grounded in the brain in a ‘dynamic core’ of continuously looping thalamocortical activity, which unifies self-defining interoceptive value information and exteroceptive perceptual information (Edelman and Tononi 2000). On this basis, he argued that the reentrant core – and hence, consciousness – emerged in its most simple form “some time around the divergence of reptiles into mammals and then into birds”, before eventually further elaboration of the reentrant core “linked semantic and linguistic performance to categorical and conceptual memory systems… enabling the emergence of higher-order consciousness.” (Edelman 2003). On the other hand, Merker (2005) hypothesized that consciousness originated much earlier, is grounded in sub-cortical structures and only gains complexity with the elaboration of the upper brain, and is shared among all vertebrates. A parallel debate, related specifically to emotion experience, has unfolded between Ledoux,

who has argued that emotion experience is constructed in the cortex (Ledoux 2000, 2019) and Panksepp, who has argued that emotion experience is grounded in “emotional operating systems that are concentrated in sub-neocortical, limbic regions of the brain” (Panksepp 2004, 2017). Although the theories above disagree about both the mechanisms of consciousness and the distribution of consciousness in other vertebrates, all agree that mammals possess consciousness, since mammals share all of the cortical and sub-cortical anatomy in question.

Tulving (199x, 2005) and Suddendorf and Corballis (1997, 2007) argued that only humans can imagine, remember, or anticipate themselves in situations other than the present, what they call mental time travel (MTT). Many mammals, birds and fish exhibit behavior such as food caching, nest building, tool use, or migration that seems to suggest foresight. For example, tayras — a members of the weasel family found in Central and South America — hide unripe sapote and plantain fruits and recover them when ripe, fine-tuning their behavior to the maturation and ripening of each fruit (Soley et al. 2011). Similar cache-and-recovery have been found from a variety of mammal and bird species (reviewed in Roberts 2012). This type of ‘what-where-when’ memory, connected to flexible behavior, suggests but does not prove, that animals can recall previous experiences in order to plan and make decisions. Panoz-Brown et al. (2018) demonstrated that rats can reliably choose the symbol that marked the nth-to-last (e.g. 2nd to last) arm of a maze they freely explored. They argue that this implies that the rats replay the sequence of sensory impressions experienced while exploring the maze in order to recover the order in which they encountered the symbols. Neural recording studies show that activity in hippocampal place cells ‘replays’ routes a rat has traveled through a maze, but also traces other possible routes revealed by exploration, including ones the rat has not traveled (Gupta et al 2012). Such replay may also offer clues to dreaming (a famously paradoxical form of unconscious consciousness) in other mammals (). Tolman argued that when faced with especially difficult or ambiguous choices, animals such as rats engage in a process he called ‘vicarious trial and error’ (VTE), which amounts to simulating possible forward runs cognitively in order to evaluate them and choose the best. (1938). The discovery of predictive hippocampal ‘theta sequences’ involved in goal-directed navigation apparently supports this interpretation (Wikenheiser and Redish 2015, Redish 2016). Thus, some degree of episodic recollection and prospection may be more widespread among mammals, albeit perhaps over qualitatively shorter time-scales compared to the human examples of MTT typically discussed (e.g. Tulving 2005).

Sibling to the living mammals are the reptiles, which includes birds as an offshoot of the otherwise extinct dinosaur lineage, familiar reptiles such as lizards, snakes, and turtles, little-researched animals such as amphisbaenians, and extinct animals such as dinosaurs, pterosaurs and pleseiosaurs (see http://tolweb.org/Amniota). These animals and amphibians such as frogs and salamanders comprise the tetrapods, who are all of the remaining descendants of a single lineage of ancient fish that made the transition to land. Hence, tetrapods (including mammals and ourselves) are one lineage of vertebrate, along with all living and extinct fishes. Which of these animals experience consciousness, and of what sort?

Birds

Outside of mammals, birds are the group of animal to whom researchers have most readily attributed consciousness, with few researchers explicitly denying consciousness to birds. As noted, the 2012 Cambridge Declaration claimed that all birds possess consciousness, and even the researchers who most prominently deny consciousness in fishes, invertebrates, and reptiles, such as Key (2015), do accept bird consciousness, claiming it has as an independent evolutionary origin from mammalian consciousness. In many ways, birds tend to seem more similar to us than other reptiles, perhaps largely because they are endothermic, constantly using stored energy to maintain a steady, high blood temperature even in cold conditions, and simply because their bodies are soft rather than scaly. However, they are no more closely related to us, and in general people’s intuitions about similarity in reptiles do not seem to accurately track biological relatedness. For example, lizards and crocodilians might seem superficially similar, but in fact, crocodiles are more closely related to birds than they are to other reptiles, and even turtles are more closely to birds than either is to snakes and lizards (ref).

Many humans have direct experience with being ‘outsmarted’ by birds (especially farmers and gardeners), as reflected in many world-wide stories about birds as tricksters, particularly corvids. This impression of problem-solving intelligence has been amply confirmed in experimental work… Puzzle boxes, tools, construction, physical intelligence. Birds also appear capable of all of the types of learning associated used as benchmarks for nonhuman animal consciousness.

Many birds display high degrees of sociality with long term affiliations of many sorts, including high degree of mate-pair bonding and extended parental care, as well as extended kin and non-kin affiliative networks. all of which provide suggestive evidence for an active, subjectively engaged, emotional life.

Birds are charismatic and conspicuous for their communication, song, and vocal mimicry abilities. Like mammals, they display communicative and emotional behavior that is conspicuous and, if not easy to understand, easy to recognize as communication or as emotion of some sort. Other than apes, the mammals most closely related to us, and perhaps dogs, which have been extensively domesticated, the animals that are most demonstrably capable of partial but still functionally communicative use of human language are birds, especially parrots (Pepperberg and other refs).

The complexity and nuance of bird communication systems offer clues to their experience of the world, including of each other. Suzuki (2018, 2020) showed that in song birds (japanese tits), alarm signals evoke a ‘search image’ of a predator. After hearing snake-specific alarm calls, the birds search along the ground and actively investigate ambiguous looking model snakes, whereas general alarm calls evoke different search behavior and don’t attract attention to the model snakes. The authors argue that this provides evidence for referential communication, and is a strong example of cross-modal integration, as the interpretation of the call evoked a specific visual expectation or ‘mental image’ in the receiver. Birds even appear to use referential communication across species boundaries within tightly knit multi-species foraging groups (Suzuki 2015, 2020).

Referential calls are known from mammals such as ground squirrels (Slobodchikoff …, and dolphins) and bird species, but have not been clearly documented for other vertebrates – and the honeybee waggle dance system likely qualifies as spatially referential in a quite different way. The flexible, adaptive social usage of referential communication seen in birds and mammals hints at higher-order forms of consciousness sometimes proposed to be human-unique, such as intersubjectivity, the capacity of one individual to perceive the subjectivity of another, and collective intentionality, the ability of multiple subjects to knowingly coordinate their attention on a common object (SEP collective intentionality).

As in mammals, many bird species have demonstrated impressive feats of ‘episodic-like memory’, perhaps indicating that they can recollect even quite distant experiences, at least in selectively tuned respects (reviewed in Roberts 2012). For example, if scrub jays are prevented from recovering their caches for long enough, they will recover only nonperishable items (peanuts, in the study), ignoring their caches of otherwise preferred but perishable food (Clayton et al. 2003). Eurasian jays have been show also remember information unrelated to a previous task — color as well as spatial location of a cup where food is cached, suggesting they have a sensory memory of the event, not just declarative what-where-when memory https://journals.plos.org/plosone/article? id=10.1371/journal.pone.0301298#pone.0301298.ref015() As with nonhuman mammals, the occurrence of impressive time and social cognition feats call into question the claim the human-uniqueness of autonoetic consciousness, and current research is currently probing related questions.

Non-avian reptiles

Most researchers now accept that non-avian reptiles likely have some degree of consciousness and sentience. They are motivated to seek pleasure and avoid pain, are capable of Pavlovian learning, and show behavioral and physiological signs of emotion, including anxiety at being handled (reviewed in Cabanac et al. 2009). Like mammals and birds, they benefit from environmental and social enrichment (Burghardt 2013), and exhibit a variety of play (Burghardt 2015, Dinets 2015). Despite traditionally being classified as solitary, many reptile species exhibit positive social affiliations to particular individuals, including conspecifics and sometimes humans (Doody, Dinets and Burghardt 2021). A number of species of lizard and snake guard their eggs, and many crocodilians are well known to exhibit extended parental care of young after hatching (e.g. Carl and Darlington 2017).

There has long been a perception or bias by brain and behavior researchers that non-avian reptiles are qualitatively more cognitively primitive than mammals and birds, and that debate continues in the form of scenarios about the evolution of noetic consciousness. The idea that non-avian reptiles are primitive creatures compared to the ‘higher vertebrates’, mammals and birds, became entrenched in neuroscience, psychology, and the popular awareness with the idea of ‘triune brain’ concept (MacLean 1985???), in which the ’reptilian brain’ is related to unthinking, reflexive survival functions, whereas the mammalian cortex contains emotion and more flexible thought, and the specifically human prefrontal cortex contains language, symbolism and ‘higher

thought’. Despite the lack of evolutionary or neurobiological support (see Ledoux 2019 ch 39), this erroneous concept that reptile brains are similar to partial or incomplete mammal brains, is still widely propagated in psychology education and may continue to influence research (reviewed in Cesario et al, 2020;).

Ledoux has argued that while a form of primary anoetic consciousness exists across vertebrates, only mammals (and birds, through convergent evolution of cortex-like brain structures) can engage in noetic consciousness and instrumental, means-end goal-directed action (2019).
This yields a hypothesis that recapitulates MacLeans’, despite Ledoux’s criticisms of Maclean’s empirical basis: “in early vertebrates, non-cognitive anoetic consciousness may have provided important advantages in navigation and foraging, including novel processing capacities for sensing the world, learning relations between sensory processing and biologically significant stimuli, and selection of actions based on past consequences. With cortical expansion in mammals, and then primates, additional survival advantages related to behavioural control may have come with more complex anoetic and novel cognitive (noetic) consciousness, with the latter involving the use of mental models. And in humans, with still further cortical expansions and the addition of language, verbal-based cognitive noesis and autonoesis have allowed us to know ourselves as entities with a past and present, and also to anticipate possible futures.” (Ledoux 2022)

Ledoux’s claim that other vertebrates are incapable of noesis (knowing, understanding, having ‘mental models’), echoes Maclean’s in attempting to reduce all reptile behavior to unthinking reflex: “There is no convincing evidence that goal-directed instrumental behavior occurs in early vertebrates (reptiles, amphibians, fish). Some claim to have demonstrated goal-directed instrumental learning in invertebrates, but the evidence mostly involves the modification of innate behaviors rather than the learning of a novel behavior on the basis of the value of its consequences. Compelling evidence for goal-directed behavior has really only been found in mammals and birds.” (Ledoux 2019, p )

For Ledoux, unlike birds and mammals, living reptiles seem to be stuck in the evolutionary past, able to learn simple associations between stimuli and contingencies of reward or punishment, but incapable of understanding or reasoning flexibly about these relationships or innovating new actions. Their consciousness is anoetic (‘unknowing’), consisting of feelings and perception of stimuli without integration of these into anything like a coherent model of reality.

However, Cabanac and colleagues (2009) argued that all amniotes (reptiles, birds, and mammals, but not amphibians) are capable of conscious emotional experience and a degree of noesis. This view is based on a theory of consciousness as a unified representational space, “an abstract private model of reality”, which allows animals to simulate possible courses of action, using affective valence (pleasure or pain) as a ‘common currency’ to evaluate and choose between actions based on expected consequences (which are based on prior experience). Consciousness is inferred from: observations of motivational tradeoffs behavior (“pleasure maximization”), play, navigational detouring (which requires an animal to pursue a series of nonrewarding intermediate goals in order to obtain an ultimate reward), expression of emotion and sensory pleasure, emotional fever (an increase in body temperature in response to, e.g. the stress of gentle handling), and taste aversion learning (Cabanac et al. 2009). Cabanac’s world model theory suggests ‘noetic’ and perhaps even aspects of ‘autonoetic’ consciousness.

Panksepp and colleagues (Fabbro et al 2015) have argued that a core self implying body ownership and agency, a common repertoire of basic emotions, noetic consciousness and some autonoetic consciousness are likely widespread among vertebrates, although ‘self-consciousness’ is unique to mammals and birds. On this view, consciousness is grounded in the vertebrate brain’s core systems for representing the bodily self in relation to the world and objects in it, including object perception and semantic memory. Their line of argument complements Merker’s (2005) argument that consciousness is grounded in subcortical brain structures common to all vertebrates, and functions to: a) stabilize representations of body and world against the sensory fluctuations of self-generated motion, b) integrate interoceptive information about the body’s internal needs and states with exteroception of the world, and c) afford high-level decision-making abstracted from sensorimotor details. These are essentially noetic functions; they are about extracting meaning from the world, so if Merker is correct, a capacity for noesis is built into the vertebrate brain, including all reptiles. .

Recent experimental work on observational learning bolsters the attribution of these aspects of consciousness to diverse non-avian reptiles, and even suggests higher-order states of consciousness such as ‘intersubjectivity’ (awareness of other individuals as

subjects/agents analogous to the the observer). Observational learning has been observed in turtles and lizards (reviewed in Doody, Dinets and Burghardt 2021, ch 11). Gutnick, Weissenbacher, and Kuba (2019) showed that giant tortoises could more quickly learn an operant discrimination task (to approach and bite one of two colored toys to earn a reward) if they first watched conspecifics. Gaze following and behavioral response to perceived gaze of conspecifics and humans has been observed in tortoises (Wilkinson et al. 2010) and lizards (). Recently discovered social hunting in boas (Dinets 2017) and sea kraits (Somaweera et al 2023) potentially offer promising case studies of naturalistic behavior for social cognition and the possibility of intersubjectivity in snakes.

Amphibians

Cabanac argued that consciousness is shared by all amniotes — the clade that includes all descendants of the common ancestor of living mammals and reptiles. He also acknowledged it was possible in other groups, such as teleosts and mollusks, but specifically excluded amphibians, based on negative results in experiments looking for taste aversion learning, emotional fever, absence of observations of play, and differences in the role of dopamine in motivation systems. (Cabanac and Cabanac 2009; Cabanac and Bernieri 2000; Cabanac and Cabanac 2000; Cabanac and Cabanac 2004) and taste aversion (Paradis and Cabanac 2004)

However, new research has shown that amphibians do display indicators of sentience, although our knowledge is currently limited by how little their behavior has been studied overall (Wiklinson 2024 NYU talk). Nociception, anesthesia, and neural pathways for pain integration are now well described in some frogs (Guenette, Giroux, and Vachon 2013). Amphibians can learn to avoid noxious stimuli, although their repertoire of responses may be tightly constrained by innate operant biases. For example, shocks by suppressing their righting reflex, although not by (Harvey). Example of negative results through bad methods (aversion learning electricity, which causes freezing…)

Amphibians benefit from environmental enrichment (Burghardt 2013), and display reliable behavioral preferences for naturalistic environments (Ramos and Ortiz-Diez 20201), social contact () and even for affiliating with specific individuals (Lamb et al., 2022).

While much amphibian behavior may seem simple and constrained in terms of means-end adaptation, Poison dart frogs stand out among amphibians that have been studied for both the social and spatial complexity of their parental care behavior. Most species lay eggs in tiny pools in fallen leaves. Both parents care extensively for eggs and young, traveling back and forth between a water source and the egg cache to keep it hydrated, and then eventually transfer the tadpoles to larger pools. Field experiments, as well as experiments with frogs in water mazes show that they use ‘mental map’-like spatial memory (Liu et al. 2019), which seems to be a clear example of noesis. Poison dart frogs, uniquely among amphibians, have also been observed in play-like behavior (Burghardt 2015).

Fishes

Most vertebrates are fish. Humans are tetrapods, descended from a single lineage of fish that left the sea. Our closest living relatives that we recognize as ‘fish’ are lobe finned fishes like lungfish and coelacanths. Together with ray-finned fish — a spectacularly diverse group include groups like salmon, tuna, bass, minnows, catfish, cichlids, seahorses, goldfish and zebra fish— we make up the bony fish, sibling to the ‘cartilaginous fishes’ (Chondrichthyes) which includes sharks, rays and their relatives.

Despite the tremendous diversity of these animals, the scientific and public debate around sentience in fishes has dominated by the question over whether fish in a small number of fish can suffer pain from the kind of injury they receive when affected by being caught on a hook. This was largely a result of science being mobilized in service of the political and public-intellectual contention around the legality of recreational catch-and-release angling, first in (formerly West) Germany and subsequently in the United States, despite the far greater human impact on fish welfare through aquaculture and environmental destruction (Jaqcuet, Franks, and Vettesse 2020)

A small number of fish-sentience skeptics have persistently argued that “fish lack a cerebral cortex or its homologue and hence cannot experience pain or fear”, in the words of Key (2015), paraphrasing the most highly-cited paper along these lines (Rose, 2002), in order to summarize his own view. However, as many people point out when first hearing this argument, it would also

seem to prove that fish cannot see or hear, which is demonstrably false. As discussed above, prominent general hypotheses about consciousness in the brain can accommodate the possibility of fish consciousness (e.g. Merker 2005, Fabbro et al 2015). Feinberg and Mallatt (2016) argue that even lampreys, which diverged from other vertebrates over 400 million years ago, possess the neuroanatomical hallmarks of consciousness, which therefore are likely shared among all vertebrates:

sufficient overall neural complexity,
hierarchical organization,
isomorphism between neural fields and exteroceptive fields,
reciprocal interactions between top-down and bottom-up influence,
multi-sensory convergence, and neuroanatomical mechanisms for memory and attention.

Woodruff (2017) details a theory of how fishes’ brains may satisfy the requirements of consciousness, focusing on isomorphic sensory maps, mechanisms of selective attention, perceptual binding, declarative memory and avoidance learning,

Fish respond to pain perception in similar ways to mammals: as if they experience and remember discomfort. Sneddon et al (2014) proposed 15 evidential criteria for discovering pain experience in animals, arguing they jointly make a strong case for for pain- sentience in fish:

  1. Integration of nociception involving brain areas that regulate motivated behaviour (including learning and fear) 2. Endogenous modulation of nociception (e.g. by opioids)

  2. Nociception causes physiological stress responses (change in e.g. respiration and heart rate or hormonal levels )

  3. Responses are flexible, not just reflexive

  4. Long term behavior change that reduces encounters with the stimulus

  5. Protective behaviour, e.g. limping or extra grooming
  6. Analgesia reduces pain behaviors and stress response
  7. Individuals in pain will self-administer analgesia

  8. Individuals in pain will pay a cost to self-administer analgesia

  9. Pain interferes with task performance, learning and memory, indicating high attention salience 11. Conditioned place avoidance and avoidance learning to painful stimuli

  10. Learning based on relief

  11. Long-lasting change in a suite of responses especially those relating to avoidance of repeat noxious stimulation

  12. Avoidance of the noxious stimulus can be traded off against other motivational requirements 15. Individuals will pay a cost to avoid noxious stimulus

Fish become stressed and actively seek to escape a painful stimulus, and subsequently remember and avoid it in a predictive way (Sneddon et. al 2003, Sneddon 2015; Brown 2015, 2016; Braithwaite 2010). They move slowly and rub or groom affected areas after injury. Anasthesia reduces these effects, and fish can learn to self-administer anesthesia, but only actually do so if in pain (Sneddon 2015). Fish can learn also learn to avoid artificial lures for up to a year after a single instance of being hooked (Beukema 1970), suggesting that they experience this evolutionarily novel event as harmful, learn to associate it with the lure, come to anticipate harm rather than food from the potential action of biting the lure and act accordingly. Fish display optimistic or pessimistic judgment bias in exploring a maze after a surprising reward or aversive event (being briefly caught in a net) (Espigares, Martins and Oliveira 2018), which is interpreted to indicate subjective mood.

Fish also display evidence of a range of positive emotions, including exploratory curiosity (Franks et al 2023), preferences for environmental enrichment, and social affiliation with specific individuals, i.e. ‘friendship’ (reviewed in Franks, Sebo and Horowitz 2018). Examples of play described in fish include: “needlefishes leaping over floating sticks and even a turtle… balancing of twigs and batting of balls in mormyrid fish species, stingrays batting around balls and competing for the opportunity to do so, and cichlid fish that repeatedly strike a self-righting thermometer.” (Burghardt 2015) Franks, Graham and von Keyserlingk (2018)

discovered ‘heightened shoaling’ in zebrafish; a form of spontaneous and apparently self-reinforcing social behavior involving closely packed, high synchrony swimming and reduced aggression, which has many of the characteristics of play.

Fish exhibit nuanced social behavior and at least some can recognize other individuals visually () and acoustically (), which suggest they may have various degrees and forms of self-awareness. Cleaner fish (Kohda et al 2023) and manta rays (Ari and D’Agostino 2016) exhibit curious, variable, exploratory mirror-oriented behavior potentially indicating aspects of sensorimotor or social self- awareness, similar to that found in many mammal and bird species, and in fact to that of humans viewing a mirror for the first time as an adult (Rochat and Zahavi 2011). Guppies, like many fish, engage in dyadic predator inspection (DPI)—a cooperative antipredator behavior performed as a pair in which close synchrony is critical—and preferentially spend time near their preferred DPI partners even when perceived predator risk is low and fish aggregate in larger, relaxed shoals. When perceived predation risk is high, they avoid large groups and maintain small groups with preferred partners, indicating that they trade-off a social-affiliative preference for larger shoals against the perceived relative safety of more tight-knit shoals (Heathcote et al 2017).

Many species of cleaner fish engage in complex interactions with their clients, often much larger predatory fish, who allow them to groom them for parasites, also often biting away bits of mucus and flesh, which make up much of the cleaner’s diet. Clients often show ‘jolts’ of discomfort, which sometimes proceed to leaving or chasing the cleaner. Cleaner wrasse engage in a unique behavior of ‘tactile stimulation’ toward their clients, which Bshary and Wurth (2001) compare to massage, and classify as a form of manipulation by which the cleaners pacify the clients, as cleaners directed more massage and more massage per bite toward more dangerous client species, and often timed massage close to bites, indicating its tactical usage. Wild surgeonfish (which had likely experienced wrasse cleaning) when temporarily held in tanks and given the opportunity to position themselves near model cleaner wrasse with brush bristles attached, which could be moved by the experimenters by pulling a string, often chose to receive the stimulation, which measurably reduced hormonal stress in the fish that received it (Soares et al 2011). This would appear to indicate that the cleaners are aware of the state of stress of their clients and use the stimulation to regulate it against the perturbation of their bites, arguably a form of intersubjectivity.

It should be kept in mind that, to a far greater extent than with reptiles and mammals, since the divides between major groups are so much more ancient, fish are very diverse, and only a tiny proportion of described species have been studied.

Other Phyla (Invertebrates)

Most animals are invertebrates. The vertebrate lineage represents just one of approximately 34 known phyla — ancient lineages of animals characterized by differences in fundamental anatomical organization and the developmental processes that generate it. Each of these lineages is derived from a relatively simple state (i.e. relatively few cells and cell-types, a minimal central nervous system, limited sensory capacities). Hence, the invertebrates that are complex enough to attract attention as candidates in consciousness debates, especially cephalopod mollusks (e.g., octopi and squids), and arthropods (e.g. crustaceans, insects and spiders), each evolved their complexity largely independently from vertebrates, and, at least in the case of cephalopods and arthropods, from each other (Coombs and Trestman forthcoming). The biological differences between vertebrates and invertebrates have posed daunting challenges to scientists trying to understand invertebrate animals behavior and sensory world, and recognition of consciousness in these groups has lagged far behind that of vertebrates.

While the diversity of all animals is staggering, three lineages of living animals (vertebrates, arthropods, and cephalopod mollusks) stand out as having individuals with large, fast, complex, bodies, and large brains capable of spatial cognition (Trestman 2013a, Coombs and Trestman 2024 in review). Several authors have argued that there is strong evidence for consciousness in animals in each of these three groups (Feinberg and Mallatt 2016, Trestman 2018, Prinz 2018, Godfrey-Smith 2019, Jablonka and Ginsburg 2019, Birch 2024 forthcoming). Feinberg and Mallat (2016, xxx) argued in detail that precisely these three lineages satisfy their neuro-architectural and behavioral criteria for consciousness (discussed above), indicating that consciousness originated independently at the root of each.

Trestman (2013a, 2018, forthcoming) proposed a collection of traits that jointly define a ‘complex active body’ (CAB) and evolution driven by selection on related features as markers of consciousness, in particular, bodily. These traits include: flexible, jointed appendages with multiple degrees of freedom and adaptations for high-speed locomotion or object manipulation, and true eyes and other distal senses. Adaptive control of a CAB is hypothesized to require basic cognitive embodiment (BCE), a set of cognitive abilities related to extracting stable spatial information from unstable, dynamic sensorimotor information; BCE constrains the evolution of CABs. Consciousness is hypothesized to satisfy the core requirements specified in BCE; Specifically, the solution lies in consciousness’ invariant temporal structure: a short-term subjective past recursively constituted by a sequence of retained moment-phases, and a branching short-term subjective future composed of a tree-structure of predicted moment- phases, with branch-points representing probabilistic uncertainty and choice (Husserl (refs); Trestman 2013b, 2018, 2023; Yoshimi 2016); compare also to Friston and Solms (). Therefore, consciousness is likely ancestral in vertebrates, cephalopods, and arthropods, as each lineage is characterized by macroevolution of form driven by selection on active features per se (Trestman 2018).

Jablonka and Ginsburg (2019) have argued at length that a suite of complex learning abilities, which they call unlimited associative learning (UAL) is a strong marker of phenomenal consciousness. UAL indicates both sentient and noetic or cognitive aspects of consciousness, as it requires the ability to make flexible perceptual judgments combining felt needs and perception of the external world. UAL combines compound learning, time-spread learning.

In addition to the more general theoretical approaches reviewed above, which make hypotheses at large evolutionary scales and apply to many species of animal, focused experimental work has allowing researchers to build up a more detailed understanding of specific ‘model conscious organisms’ such as honeybees, and hermit crabs, reflecting the higher degree of scientific confidence expressed in, for example, the NYU 2024 declaration.

Molluscs (Including Octopuses, Squid and Cuttlefish)

Mollusks are a diverse phylum of animals, including bivalves (e.g. oysters and clams), generally assumed to lack consciousness, as well as gastropods (snails and slugs), representing an acute gray area for many researchers, as discussed in Schwitzgebel’s focused study (2020). However, most cognition and consciousness research has focused on cephalopods such as squids, cuttlefish and especially octopuses, which have often been singled out as the invertebrates most likely to be conscious (2012 Declaration, Godfrey-Smith ?, ???). Cephalopods include the largest invertebrate animals (squids), largest invertebrate brains (octopus) and eyes (squids), and are notoriously clever and sensitive animals, remarkable for their stealth and predatory abilities, flexible learning and problem solving, and engaging behavior directed towards humans and other creatures (e.g. Godfrey-Smith, YouTube: Elora and Egbert).

Cephalopods show a number of markers for sentience. They exhibit flexible learning in relation to sensory pleasure, such as preferred food items, and aversive stimuli such as the stings of anemones or electric shocks. Octopuses use trial and error when attempting to prey on hermit crabs adorned with anemones, alternately approaching from different angles and varying tactics in order to probe for opportunities to attack with minimal stings (Mather 2020). This exemplifies matching a means to an end based on sequential experimentation and selection of the easiest, least energetically taxing, or most successful option, which has been proposed as a hallmark of purposive, or goal-driven, behavior and cognition (Tolman, 1932; Trestman, 2010; Ginsburg and Jablonka 2023). Many species of octopus are central place foragers, roaming an extensive hunting ground and returning to a particular refuge. Octopus are capable of using a ‘cognitive map’ of their territory, as they can head directly home from the new location after being displaced by an experimenter, and also appear to remember where in their territory they have recently hunted, and avoid those areas (reviewed in Mather 2020a)

Cephalopods show a general drive to reduce uncertainty about their environment and exhibit a variety of exploratory and uncertainty-reducting behavior towards objects and their surroundings (Mather 2020b). Reef squid observed over time by Mather (2010, 2018) adapted their responses to approaches by commonly encountered species of potentially dangerous fish. Adult squid appeared to make fine-grained judgements of the danger posed by fish according to species, swimming speed, and size, with speed mattering more than size for some species, and the reverse for others. Juveniles didn’t make these distinctions, and the overall average distance from predators was much higher for adults despite adults moving more efficiently, so it seems squid learned from experience to improve their predictions of how fish moved in order to reduce overall risk, i.e. proximity to danger. Squid also used their dynamic skin coloration differently for different species; their ‘zebra display’, normally used to warn other squid to keep their distance, was sometimes used toward nonpredatory parrotfish but never towards predatory jack or snapper, which elicited a distinct ‘pale’ escape response. This level of categorization can arguably be considered a form of interspecies intersubjectivity or ‘theory of mind’ (Mather 2018). While many species of cephalopods live their whole lives without affiliative relationships to any other creature (Mather 2018), many species do engage in various forms of aggregation which are highly variable and little understood (Godfrey-Smith 2017)

Cephalopods also appear able to anticipate the future to some extent. Cuttlefish eat less crab (a less preferred food) if shrimp are predictably available at night, but not if supply is unpredictable (Billard et al 2020). After being trained that a cue indicates they’ll get a live shrimp (favorite food) after a delay, cuttlefish can learn not to take a piece of dead shrimp (a less preferred but still readily taken reward), but only if the delay isn’t too long (Schnell et al 2021). This seems to involve not just thinking about how to satisfy present desires in the future, but to anticipate the impact of present decisions on future motivational state (i.e. if they eat the crab they’ll be too full to enjoy the shrimp) and so seems to indicate the possibility of autonoetic consciousness.

Arthropods (Including Crustaceans, Insects and Spiders)

Arthropods includes insects, crustaceans, spiders, scorpions, centipedes and many other less familiar animals. This is an ancient and tremendously diverse group of animals, perhaps earliest animals to evolve complex active bodies and brains, and so if the function of consciousness is to solve problems raised by the control of complex active bodies, it may have evolved first in the arthropod lineage, in a common ancestor of all living arthropods (Feinberg and Mallatt 2016).

Crustaceans

Crustaceans such as crabs, lobsters, and crayfish are among the largest and longest lived invertebrates, and those that are most familiar to people, since they are eaten in large quantities. Unlike most human prey, crustaceans are often handled while alive and killed during cooking (boiled alive) by consumers, rather than in a factory prior to purchase. As a result, people are confronted directly with behavioral evidence of their suffering. This often results in moral conflict, since “the lobster… behaves very much as you or I would if we were plunged into boiling water (with the obvious exception of screaming)… it takes a lot of intellectual gymnastics and behaviorist hairsplitting not to see struggling, thrashing, and lid-clattering as… pain-behavior” (Wallace 2004). Seafood vendors often distribute pseudoscientific misinformation about invertebrates, for example, that “Neurophysiologists tell us that lobsters, like insects, do not process pain. Neither insects nor lobsters have brains.” (Capeporpoiselobster.com, sampled May 2024)

Crustaceans do in fact have brains, which appear to be sufficiently complex to support integrated pain perception and aversion learning, indicating pain experience rather than mere nociceptive escape reflexes (Feinberg and Mallat 2016; Crump et al 2022; Jablonka and Ginsburg 2019; Elwood 2019), and although live boiling is still widely practiced in the United States and elsewhere, it has been banned in New Zealand, Switzerland, Norway, and the UK.

Crump and colleagues (2022) reviewed neurological and behavioral evidence for sentience in crustaceans, using criteria similar to Sneddon’s (2014) listed above. They concluded that substantial evidence exists for sentience in most major crustacean lineages, and strong evidence in true crabs (Brachyura), the best studied group neuroanatomically. The available evidence is limited mostly by lack of studies, rather than negative results, indicating that more study is likely to uncover more positive results.

Hermit crabs, and their behavior orientated to the discarded gastropod shells they typically inhabit, have served as a model system for experiments probing crustacean sentience and cognition (Elwood and Appel 2009; Elwood 2003, 2019, 2022). Shells are life- critical for wild hermit crabs; they must offer protection from predators, but not be too heavy or awkwardly shaped to inhibit mobility, and crabs apparently judge shell adequacy quite generally, as they can inhabit shells of a variety of species of gastropod, and even artificial shells. They must find new shells as they grow, and are constantly under threat of being un-shelled by competitors, so their cognition—and if it exists, their subjective experience of the world—has been under intense evolutionary pressure for evaluating shell quality and making decisions accordingly. When experimentally given a small shock inside a shell,

crabs will sometimes abandon the shell, although their likeliness of abandoning it depends on the quality of the shell, and crabs that smelled the scent of a predator were less likely to abandon it, indicating the crabs could consider and balance the perceived risk of another puzzling shock against the other factors. Some crabs that abandoned the shell directed aggressive behavior against it, potentially indicating confusion and motivational conflict at the strange stimulus; even crabs that did not immediately abandon the shells in which they were shocked later showed an increased willingness to switch shells, indicating they remembered the shock as aversive and continued to associate it with the shell. Conflicts over shells are a major feature of hermit crab life, and during conflicts over shells, crabs appear to make integrative judgements about their own shell quality and the availability of other shells, as well as the current relative fighting ability of themselves and their opponents, although exactly how these judgments are made is an area of active research. Crabs also appear to be able to include information about the size and shape of their when predicting their ability to pass through a gap, which indicates a sophisticated form of bodily-spatial self-awareness .

Many crustaceans are surprisingly social, such as spiny lobsters, which engage in unique single-file mass migrations (Herrnkind et al 1973) and exhibit tightly coordinated antipredator behavior (Lavalli and Herrnkind 2009), and the little studied eusocial shrimps of genus Synalpheus, which show coordinated colony defense and behavioral differentiation similar that of hymenoptera and termites (Duffy, Morrison, and McDonald 2002). At least three groups of crustaceans, crayfish, stomatopods (mantis-shrimp) and hermit crabs, are capable of recognizing other individuals and recalling previous interactions such as fights (Aquiloni and Tricarico 2015). Some crayfish are known to exhibit parenting behavior and kin recognition (Aquiloni and Gherardi 2008).

Insects

Insects are the invertebrates that are most familiar to most people. They are incredibly numerous and diverse, and it therefore follows that whether or not insects are conscious implies a tremendous difference in the world around. Does every fly and ant have its own subjectivity, perspective and values?

One large hurdle to accepting insect consciousness has been that insects are supposedly too small, or that their brains contain too few total neurons or neuron types (Feinberg and Mallat 2016; Ledoux, Key, tons of others ???), although the aforementioned analogy with fish vision should quickly demonstrates the fallacy of this argument as well, since insects can see despite their entire bodies being smaller than the human visual cortex. However, Feinberg and Mallatt (2016) concluded that the insect brains display all other hallmarks of consciousness, and Klein and Barron (2016) also argued the insect brain performs al the core functions required for consciousness.

Researchers disagree in their assessment of insect behavior. Tye (1997) argued, radically for the time, that there was “ample evidence” that “honey bees, like fish, are phenomenally conscious: there is something it is like for them… Honey bees make decisions about how to behave in response to how things look, taste, and smell. They use the information their senses give them to identify things, to find their way around, to survive. They learn what to do in many cases as the situation demands. Their behavior is sometimes flexible and goal-driven.” West-Eberhard (2003) argued that anti-anthropomorphic bias was a “curse” that had caused researchers to neglect internal mechanisms of insect and other animal behavior and left a conceptual void. She argued that motivational conflict between fear and competitive aggression was evident in the evolution of social signals stingless bees (Melaponinae), and that the direct foraging flights of many social hymoptera exemplify prospective goal directed behavior, as individuals can fly directly to patches of different resource (food, water, pulp, nectar) determined prior to departure by perception of colony needs (West-Eberhard 2003, p 314). Chittka (2021) argued that “if we apply the same behavioral and cognitive criteria as we do to much larger-brained vertebrates, then bees qualify as conscious agents with no less certainty than dogs or cats.” In contrast, Allen-Hermanson (2008) argued that honeybees and other insects are ‘natural zombies’ without subjective consciousness, their appearance of consciousness being a collection of well-adapted tropistic reflexes.

Studies on bees, revealing pattern recognition, concepts of ‘same’ and ‘different’, navigation, communication, and visual working memory (reviewed in Srinivasan 2010), and on mood and cognitive bias (Mendl et al. 2011).

Recent studies with bumblebees “have consistently shown that bees improve their foraging performance as they accumulate experience, by approximating the shortest possible route to visit all flowers, prioritizing visits to flowers offering greater rewards, and trading off accuracy of route repeatability against flight speed. These behaviors are incompatible with hardwired movement

rules, indicating that bees acquire a spatial memory of flower locations and use this information to minimize travel costs.”(Lihoreau et al. 2013)

Domesticated European honeybees have been studied behaviorally more than any other insect.
Honeybees’ long term, flexible ‘landscape memory’, established during non-foraging orientation flights and subsequently refined throughout their foraging career, allows them to return home after far-ranging exploratory flights or after displacements, or even to short-cut between a familiar foraging location and one indicated by a waggle-dance (Menzel and Greggers 2015).

Chittka (2017) in reviewing findings on honeybee and bumblee bee cognition and behavior concludes that the time has come to consider, as a broad strategic shift, that bees “have a basic understanding of the outcome of their own actions, and those of other bees.”

cross-modal spatial inference (molyneux’s problem Chittka 2021)

Bees also alter their foraging behavior after being attacked by predators, either avoiding a location or flower-type associated with the attack, or approaching more cautiously (Chittka, 2017; Wang et al., 2018).

Bees visit the flowers they do as part of a route they have optimized through experience. Constancy emerges as part of an adaptive solution. A bee’s habitual routes, or ‘trap-lines’, are a part of its extended cognitive phenotype that emerges through complex forms of embodied spatial learning that are consistent with the phenomenological theory of consciousness:

A recently discovered aspect of the honeybee waggle dance offers striking behavioral evidence of prospective cognition: if a waggle-dance follower observes a dance indicating a location where they have been attacked by predators or other bees, or where they have encountered poor forage, they often respond with a distinctive ‘stop signal’, which inhibits performance of and recruitment to the dance, and hence also inhibits foraging at the inhibited location (Nieh 2010). Further, the signal appears to encode ‘referential’ information about degree of threat—in the Asian honeybee Apis Cerana, stop signals recorded from bees who had been attacked by the large and extremely dangerous hornet Vespa mandarina had a higher frequency and more strongly suppressed dancing and dance- following by others, compared to stop signals following encounters with the smaller Vespa velutina (Nieh et al,
2019

Individual bumblebees make slower, more accurate decisions when they must integrate information about risk (of a simulated predator attack or a tasting aversive quinine) with learned expectations about relative reward at artificial feeders. However, when it becomes more difficult to distinguish between safe high-reward and dangerous feeders, they will switch to safe, low-reward feeders in order to avoid having to make the risky, time-consuming judgment (Ings et al., 2008; Wang et al. 2018). This suggests a flexible metacognitive awareness of strategy (i.e. make vs avoid the difficult judgment), as well as showing that bees spend longer on difficult or higher-stakes judgments.

Bees seem to show an awareness of their own bodies and of the bodies of their fellow bees as analogous to their own (embodied intersubjectivity). Bumblebees, for example, can learn by imitation, as demonstrated for artificial tasks such as pulling a string to bring a reward into reach, or rolling a ball into a target area, even when the reward is out of sight (reviewed in Chittka 2017, 2021). The honeybee waggle dance offers another example: dance observers are more likely to follow the waggle-dances of performers who are similar to them in body size. (Waddington 1989). This is adaptive because similar sized bees will have similar success rates at extracting pollen or nectar from given types of flower. How bees judge the size of the dancer relative to themselves, including whether it involves learning, is unknown.

Other fiindings suggest cognitive self-awareness/autonoesis. Individual bumblebees make slower, more accurate decisions when they must integrate information about risk (of a simulated predator attack or a tasting aversive quinine) with learned expectations about relative reward at artificial feeders. However, when it becomes more difficult to distinguish between safe high-reward and dangerous feeders, they switch to safe, low-reward feeders to avoid the risky, time-consuming judgment (Ings et al., 2008; Wang

et al. 2018). This suggests a flexible metacognitive awareness of strategy (i.e. make vs avoid the difficult judgment), as well as showing that bees spend longer on difficult or higher-stakes judgments.

Other than bees…
Individual recognition and transitive inference in Polistes (Tibbets 2019) Termites learn to avoid nasty smells in a Y-maze (Ding and Li 2024)

Carabid beetles of several different species learned over time to adjust their behavior when interacting with dangerous ants in artificial environments, depending on whether the ants were high or low aggression and whether they were tethered and unable to pursue the beetle. Beetles showed ‘catalog learning’, improving in their use of different combinations of their innate behavioral patterns in order to more effectively avoid danger without being overly cautious (Reznikova and Dorosheva 2013).

Arachnids

Spiders and other arachnids such as scorpions, pseudoscorpions, whip-spiders, solifuges and vinegaroons, occupy more of a gray area in the current scientific literature. They were excluded from direct mention in the 2024 NYU Declaration, due to a relative paucity of studies having looked for the signs of sentience in any of these groups, as defined in the criteria sets proposed by e.g. Sneddon, Crump and Elwood. However, arachnids do also show a number promising markers of sentience. Nociceptors have not been described in arachnids, but they use of a variety of escape, defensive, and wound-tending behaviors in response noxious stimuli such as heat, shock, and toxins such as acetylcholine and bradykinin, and also display signs of an arousal and stress modulation system (reviewed in Kralj-Fišer and Matjaž 2019). When encountering ants for the first time, Edwards and Jackson (1994) found spiderlings first approached and then stalked ants as prey, before switching to anti-predator behavior upon encountering dangerous resistance, and subsequently avoiding ants, implying that they reclassified ants based on their aversive experience.

It should also be noted that macroevolutionary arguments such as Feinberg and Mallat (2016, p 185), Trestman (2018), and Ginsburg and Jablonka (2019) imply that all arthropods are likely conscious, including arachnids, unless consciousness has been secondarily lost. Arachnids have diversified into an impressive array of body plans based on novel configurations of weaponry and sensors exemplifies, including variations on the opposed, crab-like lateral claws of scorpions, pseudoscorpions, whip-spiders and vinegaroons, the scorpions’ unique forward facing tail, and the specialized antenniform legs of whip-spiders and vinegaroons, which arguably constitute a distinct sense modality, as well as the incredible versatility with which spiders use their silk spinnerets and legs to construct webs. This exemplifies to an extreme degree the concept of CAB-driven macroevolution argued by Trestman (2018) to be a marker of consciousness. In keeping with this hypothesis’ emphasis on spatiality, diverse arachnids show impressive abilities to guide behavior with integrated spatial representations. Jumping spiders are know to take detours (Jackson and Wilcox 2003). They are also able to circle around to the back of their prey and remain motionless when an enemy is facing them, which are forms of gaze-tracking. Diverse arachnids exhibit central place foraging (Ortega-Escobar et al 2023), a spatially demanding lifestyle that requires an advanced toolkit of navigational abilities, and which Chittka (2021) hypothesized to be a driver of hymenopteran cognitive evolution. Web building spiders appear to hold representations of prey and of their webs in memory. Rodriguez et al (2011) found that when spiders search for experimentally pilfered prey, their memory of the prey influences the searching behavior; spiders searched longer for larger lost prey items, and continued searching until the lost prey was found despite finding other distracting items, which the authors interpreted to show that the search behavior was guided by a mental representation of the target, based on the memory of when the prey was cached.

Black widows show path integration, an egocentric spatial awareness of their distance and direction to home (Sergi 2022). In what is proposed as a novel assay for consciousness, Sergi examined the behavior of black widow spiders when they were placed back into a web the spider had created but not occupied in a week, with a cricket already in the web, in order to test whether a mismatch between the spider’s expectation or subjective mental representation of the web’s layout could compete for its attention with the prey item. In the control group, the previously occupied web had the same spatial structure as the one the spider had been in since; no spiders searched, and almost all approached the cricket directly. In the experimental conditions, the spatial

layout of the two chambers, and hence webs, differed either slightly or greatly; roughly half of these spiders engaged in a distinctive searching behavior, walking about the web and probing it, for a time before proceeding to devour the cricket. Sergi interpreted these to show that these spiders were distracted from the cricket and the drive to feed on it by the mismatch between expectation and actuality, indicating subjective experience (2022).

In terms of possible autonoesis and intersubjectivity, many species of arachnid have complex social lives and even many those classified as ‘solitary’, such as jumping spiders, display elaborate courtship rituals () and some degree of social affiliative behavior (), signs which are often interpreted as indicating preferences or aesthetic appreciation in vertebrates (Prum 2017). Example species of spiders, amblypygids (whip spiders), uropygids (vinegaroons), and schizomida (short-tailed whip scorpions) are all known to exhibit extended parental care, including guarding and carrying of eggs and young (reviewed in Rayor and Taylor 2006). Amblypygids, large arachnids with pronounced mushroom bodies (Sinakevitch, Long and Gronenberg 2021), show sibling-sibling and mother-offspring affiliative behavior, with offspring orienting toward the mother, who will “sit in the middle of a group of her offspring and stroke their bodies with her whips”. In response to an experimental disturbance (the cage being lightly rattled), sibling interactions ceased and young hide under their mother (Rayor and Taylor 2006).

Other invertebrates and other organisms

Invertebrates other than arthropods have scarcely been examined closely in consciousness studies…

Another possibility is that consciousness evolved even earlier in animal history and is even more widely distributed among animals, and hence has a function that is even more fundamental to animal life. Ginsburg & Jablonka (2007a,b) attribute a primitive form of “overall sensation” as a by-product of even the simplest nerve nets in animals. They argue that as these states became harnessed to learning and motivation that they acquired the functional properties of “basic consciousness”. If this is right, than consciousness may have arisen not independently in arthropods, mollusks and vertebrates, but only once in the common ancestor of these ancient groups, very early in animal evolution.

Other potential candidates out of scope: plants, fungi, single cells, software, robots, aliens…