# Chapter 21: Biological and comparative applications
## 21.1 What an application must establish
The finite constitution gives a rule on a realized operational domain. Applying it to an organism is not accomplished by substituting the word “brain” for a vertex of a graph. An application must identify the components, the time scale, their physically admitted interventions, the internal and external interfaces, the relevant preparations, and the information available to the process itself. It must then determine whether the operational reduction is adequate for the proposed experience-facing targets.
Three questions should remain separate. Does a chosen physical model predict the experimental records? Does its proposed core decomposition remain stable under justified changes of resolution? Does the independently elicited phenomenal organization agree with the structure that A2 assigns? Success on the first question is indispensable but is not automatically success on the other two. Failure on the first cannot be repaired by invoking primitive awareness.
It is therefore useful to distinguish a *candidate application* from a *validated realization*. A candidate application specifies an organism or artificial process and proposes an $R^\ast$ model. A validated realization would require evidence that the supplied model and its interfaces capture the relevant physical organization, together with a justified interpretation of the experience-facing measurements. No such completed validation for a human, animal, plant, or deployed language model is asserted here.
## 21.2 Development and the absence of adult prerequisites
The developmental case tests the temptation to identify consciousness with the mature capacities of its investigators. The review by Bayne and colleagues argues for taking early infant experience seriously while examining the uncertainty about its onset and form [[2](/consciousness/monograph/bibliography#bib-Bayne2023)]. This is not a license to assign an exact developmental threshold from a spectral gap or a recursion count. Its relevance is that adult language, explicit philosophical reflection, and an elaborate autobiographical self cannot simply be built into an allegedly universal admission rule.
Within SPC-2, an immature vessel can qualify through an endogenous recurrent core without implementing adult counterfactual deliberation. Learning may change the physical kernels, the repertoire of available actions, the predictive quotient, and the content of an episode. A0 does not have to increase as a child learns a word; A2 permits the organization of a perspective to change. This captures the intended distinction between an underlying awareness commitment and acquired scaffolding without pretending to have measured a quantity of primitive awareness.
The model also distinguishes a repertoire from its current point. A system can have a rich capacity for discrimination while occupying a simple current state; conversely, an apparently vivid report does not establish a large or unified predictive structure. A developmental interpretation should analyze both the available organization and the actual state within it. The words “less developed” refer here to specified functional capacities, not lesser moral worth or a smaller entitlement to protection.
## 21.3 Sleep, anesthesia, and the lived scene
Clinical descriptions often bring three questions together: does the vessel continue, is there experience now, and can the person subsequently describe it? Separating them does not make assessment simple, but prevents a failure of memory or motor output from deciding a different question by default. The distinction in Section [2.5](/consciousness/monograph/awareness-contents-and-the-personal-model#sec:vessel-scene-memory) applies throughout this discussion.
### 21.3.1 Dream formation and reduced external constraint
In this framework, dreaming is interpreted as a scene generated when external constraint is reduced while the vessel's generative and scene-forming organization remains sufficiently coupled. The experienced setting can be rich even when it is weakly constrained by the current environment. Reduced constraint does not mean complete sensory isolation, and internally generated does not mean detached from physical realization.
This account is an interpretive model rather than a claim that all dreams arise through one mechanism. Siclari and colleagues compared reports of dreaming and their absence across REM and non-REM sleep [[50](/consciousness/monograph/bibliography#bib-Siclari2017)]. Their work supports distinguishing sleep stage, reported experience and recall; it does not turn the proposed coupling language into an established SPC-2 mechanism. The theory neither identifies all sleep with dreaming nor treats dream recall as the criterion of dream occurrence.
### 21.3.2 Dreamless and unremembered intervals
An unremembered dream and a genuinely scene-less interval are different possibilities. In the former, experience occurred but was not retained or recovered for a later report. In the latter there was no current experience. It would be misleading to describe that absence as a subject observing darkness, or to preserve experience by stipulating that everyone always dreams and merely forgets.
On the biological interpretation, coupling sufficient to form a scene can fail while the organism and much of its regulation continue. During a genuinely scene-less interval there is no currently instantiated localized experiential subject. This is compatible with the organism's survival and ordinary biographical continuity. Within SPC-2, a genuine qualification gap ends the experiential episode under A3; when qualification returns, a new episode begins. An unremembered dream does not establish such a gap.
### 21.3.3 General anesthesia
General anesthesia can disrupt the organization required for an integrated lived scene while substantial biological and local neural activity remains. Neuroscientific accounts distinguish changes in large-scale interaction from a simple cessation of brain activity [[29](/consciousness/monograph/bibliography#bib-Mashour2024)]. The effects of particular drugs, doses and clinical contexts need not be identical. Unresponsiveness, disconnected experience, amnesia and absence of experience are also distinct possibilities; retrospective reports concerning experimentally induced unresponsive states under propofol and dexmedetomidine illustrate why they must be distinguished [[45](/consciousness/monograph/bibliography#bib-Scheinin2021)]. Those findings do not establish how often experience occurs under deep surgical anesthesia, and recollections may depend on the transition toward arousal.
Within the Shadow interpretive model, sufficient disruption of coupling between explicit/attended and unconscious/generative organization prevents a lived scene; restored coupling permits manifestation to resume. This does not assert that “one side is switched off” as a universal neurophysiological fact. A biological person can remain physically continuous, with background processes active, throughout a genuinely scene-less interval. Awareness, as the knowing aspect of ${\mathsf{U}}$, is not said to be created or destroyed by the anesthetic. During that interval there is no current lived scene and no currently instantiated localized experiential subject. There is no observer experiencing darkness: darkness would itself be a scene.
A genuine qualification gap ends the experiential episode under A3. When qualification returns, a new episode begins, although the biological/person-level vessel has remained physically continuous. For a genuinely scene-less interval, the first-person sequence can be represented schematically as
$$
\text{scene}_1\;\longrightarrow\;\varnothing\;\longrightarrow\;\text{scene}_2.
$$
The middle term denotes no experience, not an experienced empty interval. This makes intelligible why deep general anesthesia can feel, on return, as though no time elapsed: there is no intervening scene presenting that duration. Physical time still passes. An impression of immediacy can also follow failed recall, so it does not by itself establish that experience was absent.
A3 can continue an episode across a change of anesthetic regime when qualification and nonbranching provenance persist, including a change to disconnected experience. Neither the clinical label nor a report of remembering nothing establishes loss of qualification. The coupling account interprets possible biological realizations of A1–A3; it adds no separate scene gate and supplies no clinical test for a particular patient.
### 21.3.4 Local anesthesia and partial sensory loss
Local anesthesia provides a different case: a peripheral intervention can block or alter signals from a restricted region while the person's larger lived scene continues. Additional sedation or other clinical factors must be considered separately. The relevant contrast is between a changed incoming channel and a disruption of the organization sustaining the whole scene. Local sensory silence is not, by itself, destruction of a subject.
In the framework's terms, a sensory boundary changes what can enter or constrain experience. Other sensory, bodily, mnemonic and affective organization may remain. This illustrates why a missing modality, a failed report route and a loss of experiential admission cannot be treated as the same event.
## 21.4 Candidate cortical correlates of scene-forming coupling
In mature mammalian cortex, large-scale recurrent interaction and effective connectivity are strong candidates for the physical realization of scene-forming coupling; selected fast-band coordination may contribute to that organization. This is a biological interpretation of organized interaction among distributed processes, not an identification of consciousness with tissue activity. Effective connectivity concerns how activity in one population influences another; observational estimates of directional dependence and responses to an imposed perturbation provide different, method-dependent evidence for it.
The anesthesia literature supports the distinction. Human intracranial recordings during propofol induction showed that local neuronal relationships could persist while activity became restricted to brief windows occurring asynchronously across cortical regions [[25](/consciousness/monograph/bibliography#bib-Lewis2012)]. Cortex need not be globally silent for communication to fragment. Lee and colleagues found reduced estimated frontal-to-parietal feedback around loss of responsiveness under ketamine, propofol and sevoflurane, despite different spectral effects [[24](/consciousness/monograph/bibliography#bib-Lee2013)]. Their directional statistic and behavioral endpoint do not establish absence of every kind of experience. In rats, Imas and colleagues found preferential impairment of frontal–posterior feedback at 50 Hz under volatile anesthetics, while effects at 30 Hz differed [[21](/consciousness/monograph/bibliography#bib-Imas2005)]. These findings motivate studying recurrent organization and its disruption, rather than assigning one meaning to all gamma activity.
Agent-specific perturbational findings sharpen the point. Sarasso and colleagues used transcranial magnetic stimulation with EEG: propofol produced relatively local responses, xenon produced widespread but stereotyped responses, and both had low perturbational complexity. Ketamine preserved more differentiated responses and high complexity, alongside subsequent vivid dream reports [[44](/consciousness/monograph/bibliography#bib-Sarasso2015)]. Broad propagation alone therefore does not establish differentiated conscious integration, and reduced feedback in a particular estimate cannot be a universal criterion for scene absence. Neither retrospective dream reports nor their absence escape the memory limitations discussed above.
Hameroff emphasized gamma-range synchrony and anesthesia-related disruption of conscious integration [[19](/consciousness/monograph/bibliography#bib-Hameroff2010)]. That emphasis is relevant here; his proposed dendritic mechanism, microtubule computation and Orch-OR are not adopted. Gamma power, phase coordination and effective interaction are distinct measurements. No frequency band is awareness itself, either of the conscious/generative modes, or an admission condition. REM dreaming, ketamine-related disconnected experience and seizure states must be assessed without treating fast activity, synchronization or responsiveness as sufficient for conscious integration. There is no “gamma on = conscious” rule.
*The relevant invariant is organizational integration versus its failure, not a single number on a monitor.* Here integration means differentiated, effective interaction capable of supporting the proposed scene organization, not maximal synchrony or merely extensive connections. This is the interpretive target across implementations, not an additional mathematical invariant already proved for cortical measurements. No universal EEG threshold, necessary cortical band or unique anatomical locus is established.
These network findings motivate candidate biological realizations of scene-forming coupling in a rich mammalian $R^\ast$. They are not a new admission law, not A1, and not a proof of A0 or A2. During a genuinely scene-less anesthetic interval, no localized experiential subject is currently instantiated under SPC-2, while the biological vessel remains physically continuous and the person biographically continuous. Awareness as an aspect of ${\mathsf{U}}$, formal admission, the biological coupling account, its measured correlates and empirical validation remain separate levels of claim. The combined SPC-2 interpretation places dreams and anesthesia within a source-aspect constitution; it claims no novelty for recurrent processing or network integration individually. Existing anesthesia and predictive-processing accounts support or motivate parts of this pattern without establishing the ontology or the constitution [[24](/consciousness/monograph/bibliography#bib-Lee2013), [44](/consciousness/monograph/bibliography#bib-Sarasso2015), [9](/consciousness/monograph/bibliography#bib-CarhartHarrisFriston2019)].
## 21.5 Disorders of consciousness and incomplete access
Bodien and colleagues reported task-related fMRI or EEG responses in 60 of 241 participants who lacked observable responses to commands in their examined sample [[6](/consciousness/monograph/bibliography#bib-Bodien2024)]. This finding constrains the use of overt motor behavior as a sole measure of relevant cognitive activity. It does not show that every neural response measures all experience, nor that every negative neural test demonstrates absence of experience.
In an $R^\ast$ model, overt report is one output channel, not the definition of the process. A motor pathway can fail while another internal discrimination or response route remains. The predictive profile can therefore retain distinctions that a particular report channel loses. The aperture-adequacy theorem gives a formal way to express the issue: a descriptor that collapses conditions differing in an independently justified target cannot support an exact bridge for that target.
This is a methodological implication, not a diagnostic instrument. No clinical decision should be based on the two-bit example or an unvalidated assignment of perspectival cores. An application in patients would require appropriate clinical expertise, consent or surrogate procedures, validated measurements, and uncertainty reporting. The framework's insistence on separating presence, access, memory, and output is intended to prevent premature classification, not to replace established care.
## 21.6 Blindsight and divided processing
The classical blindsight literature describes residual visual discrimination in a field defect without the corresponding ordinary report of seeing [[43](/consciousness/monograph/bibliography#bib-Sanders1974)]. Its role here is to make content-specific discrimination, confidence, access, and phenomenal attribution distinct questions. The condition of an entire person cannot be inferred from a single missing visual report.
A2 is deliberately demanding in this setting. If two conditions share the complete nominated predictive structure yet differ in an independently established phenomenal relation, A2 fails for that realization or the physical descriptor is inadequate. It is not enough to point to a successful forced-choice response and call that response experience. A2 is a psychophysical identification requiring justification, not a linguistic decision to redefine all discrimination as phenomenology.
Split-brain research is particularly relevant to subject individuation. Pinto and colleagues argued for divided perception without two independent perceivers in the cases they studied [[30](/consciousness/monograph/bibliography#bib-Pinto2017)]. Volz and colleagues dispute that inference, emphasizing cross-cueing, ipsilateral motor control and possible subcortical transfer [[54](/consciousness/monograph/bibliography#bib-Volz2018)]. Whatever position one takes on that interpretation, graph partition and phenomenal subject count cannot simply be read off from the phrase “split brain.” The physically operative routes, remaining common control, temporal grain, and independently assessed behavior all matter.
The exact SCC rule makes a definite conditional assignment once those data are supplied. Its sharp weak-coupling transition is a vulnerability rather than a fact established by the clinical literature. A biologically adequate realization might preserve important internal routes after a particular anatomical intervention; an overly coarse model might erase them. Conversely, retaining every negligible feedback route might over-unify the system. This makes the physical boundary and resolution problem central to testing A1.
## 21.7 Animals and nonhuman presentation
A conscious subject need not possess human-style language, autobiography, explicit metacognition, or a reflective self-concept. Animals can in principle be conscious under this theory when an appropriate $R^\ast$ realizes A1; this is not a universal species list or a declaration that every proposed animal model is certified. Comparative accounts can describe several dimensions of an animal's experiential capacities rather than place all species on one scale [[5](/consciousness/monograph/bibliography#bib-Birch2020)]. This does not identify those dimensions with the present predictive object. The dolphin, whale, and insect comparisons should begin with a modest logical point: different sensory and action repertoires can support different ways of organizing a world. There is no requirement that every potential perspective contain adult human visual categories, spoken language, or a human bodily model. A nonhuman comparison must use its own physically meaningful input and action spaces rather than treating human reportability as the universal standard.
Under SPC-2, two species could have different predictive structures even when a particular task score agrees. Conversely, a particular operational structure might be reproduced across different substrates without establishing that their entire phenomenal lives are identical. Cross-system phenomenal comparison is meaningful only after specifying which physical structures and protocols the comparison preserves. A2 supplies a conditional structural comparison, not unrestricted imaginative access to another organism's life.
No numerical consciousness threshold for insects or marine mammals follows from the O1 spectra. Those finite complexes are source-response examples, not measured nervous systems. Their value is methodological: provenance and sector geometry can matter even when spectra agree. The lesson transfers as a requirement to preserve relevant structure, not as a numerical species classifier.
## 21.8 Plants, fungi, and minimal systems
Plant consciousness remains contested in the literature. Taiz and colleagues argue that plants neither possess nor require consciousness; Segundo-Ortín and Calvo explore a more permissive analysis of plant cognition and consciousness [[52](/consciousness/monograph/bibliography#bib-Taiz2019), [46](/consciousness/monograph/bibliography#bib-Segundo2022)]. These positions should not be collapsed into agreement that signaling proves experience. Communication, adaptation, and distributed regulation establish physical capacities; the psychophysical conclusion remains a further claim.
The present constitution is potentially permissive in a different way: a simple recurrent device can qualify under A1. The requirement of nontrivial predictive structure does not impose a high intelligence threshold. That implication should not be concealed because it is counterintuitive. It is one of the costs and possible tests of the proposed law. An application to a plant, fungal system, or controller still requires the same $R^\ast$ specification rather than an inference from the material category alone.
A theory that intends to exclude minimal recurrent systems needs an additional principle. It could introduce a physically motivated scale, an integration condition stronger than SCC membership, or a different admission rule. None follows simply from preferring a familiar list of conscious organisms. Such an extension would be a new version of the psychophysical constitution and should be compared explicitly with A1, not presented as an unnoticed consequence of it.