Consciousness: objective definition

 

Concept: consciousness

1. Reduction to perceptual roots and basic axioms/principles

A. Perceptual-level roots (publicly accessible phenomena and direct observations)

• Contrasts in state within a single organism: awake vs deep sleep, anesthesia, coma; dreaming vs dreamless sleep; sedation levels in medicine.
• Direct first-person data (as a special case of perception): seeing colors/shapes, hearing sounds, feeling pain/pleasure, having a mental image, attending to one thing rather than another, a continuous “field-like” presence of contents.
• Behavioral indicators in others (humans and nonhumans): flexible, goal-directed action; orienting and exploration; learning from novel situations; deliberate choice; report or report-like communication; context-sensitive pain avoidance; attention shifts; conflict monitoring and error correction.
• Neural/physiological contrasts (purely as observable correlates): patterns that track the above state changes (e.g., changes under anesthesia vs wakefulness), long-range integration vs fragmentation, REM vs NREM distinctions.
• Pathologies and dissociations: blindsight (accurate guessing without reported seeing), neglect, vegetative state vs minimally conscious state, locked-in syndrome—showing that some functions can proceed without conscious access, and that conscious access has distinctive profiles.

B. Basic fundamental axioms and general principles

• Existence and identity: things are what they are; a system’s states have determinate properties.
• Causality and non-contradiction: the same causes under the same conditions yield the same effects; a state cannot both have and lack a property in the same respect at the same time.
• Consciousness is always consciousness-of-content: when present, it includes determinate contents (self/world, thoughts/feelings) rather than being a contentless nothing.
• Organisms are agents maintaining homeostasis and goals; their nervous systems (or functionally analogous architectures) process information to guide action.
• Information processing can be integrated (many parts jointly constrain/shape content) or fragmented; degrees of arousal modulate whether such processing is globally effective.
• Distinction between unconscious processing and conscious access: much processing is unconscious; conscious states are those whose contents become globally available to multiple subsystems (perception, evaluation, decision, memory, verbal/motor output), enabling flexible control.
• Unity at a time: conscious contents co-appear in a single, coherent field for the subject when present.

(Working lemmas tying observation to concept)

• If a system’s internal representations are globally available and integrated in a way that guides flexible, context-sensitive behavior (and, where feasible, report), then the system is in a conscious state.
• If representations are isolated, strictly modular/reflexive, or unavailable to global control, the processing is unconscious.

2. Intermediate steps and principles to reconstitute the concept

• Start from concrete cases of being awake, feeling pain, seeing, hearing, deliberating; note their contrasts with sleep/anesthesia/coma.
• Abstract the common pattern: a unified, present, contentful “field” in which multiple modalities and self-related signals are co-available and can flexibly influence action and memory.
• Differentiate from mere reactivity: reflexes and habit-loops can proceed without that unified, globally available content.
• Generalize across species and systems: where we cannot obtain verbal report, use functionally equivalent indicators (flexible generalization, context-sensitive valuation, cross-modal binding, counterfactual learning, goal reprioritization after new information).
• Establish boundary conditions: minimal consciousness does not require language or self-reflection; metacognition and narrative self are higher-level add-ons, not necessary for baseline consciousness.

Intermediate principles (bridge rules)

• Global availability/broadcast: contents accessible to multiple subsystems at once.
• Integration: contents are co-constrained into a single, coherent scene/stream.
• Valuation: contents are appraised for significance and can modify goals/behavior.
• Flexible control: access to contents alters behavior beyond fixed stimulus-response mappings.
• Subject-level organization: the integrated contents belong to one agent maintaining its own boundaries and goals.

3. Measurable essential and distinguishing characteristics (omitting specific measurements)

Essential characteristics (necessary and jointly sufficient in typical cases)

• Integration and unity: internally represented contents are bound into a single, coherent field for one agent at a time.
• Global availability/access: those contents are available to multiple internal processes (decision-making, memory updating, action selection, communication/output).
• Present-centered guidance: the globally available contents can guide flexible, context-sensitive control here-and-now (including planning into the near future).
• Valence/appraisal: contents are evaluated for relevance (good/bad, important/irrelevant) and this evaluation can influence priorities and actions.

Distinguishing characteristics (vs purely unconscious processing)

• Reportability or report-equivalent behavior when channels exist (verbal in humans; communicative/choice-based proxies in nonverbal agents).
• Cross-modal binding and selective attention within the unified field.
• Capacity to learn from one-off episodes and to flexibly apply that learning in novel contexts.
• Optional but common: metacognitive access (awareness of knowing), not required for minimal consciousness.

Conceptual common denominators (shared, scalable dimensions)

• Degree of integration, degree of global access, degree of flexibility, and strength of valuation coupling—these vary by level without changing the kind of state.

4. Objective definition (genus, differentia, essentials)

Genus

• A functional-organizational state of a living or artificial agent.

Differentia

• In which internally represented contents about self and/or environment are (i) integrated into a single, unified field for that agent, and (ii) globally available to multiple internal subsystems at once, such that (iii) these contents, together with their valenced appraisal, can flexibly guide goal-directed behavior and, where possible, be expressed in report-like outputs.

Essential characteristics in the definition

• Unity/integration of content.
• Global availability/access.
• Valenced appraisal coupled to flexible control.

Distinguishing characteristics in the definition

• Flexible, context-sensitive guidance of behavior beyond fixed reflexes.
• Report or report-equivalent expression when output channels permit.

Conceptual common denominators in the definition

• Integration, accessibility, flexibility, valuation (scalable across kinds and degrees).

Criteria for an objective definition and assessment

• Ties to reality: Refers to directly observable contrasts (wake vs anesthesia, flexible control vs reflex), known dissociations, and measurable correlates. Yes.
• Non-contradiction and clarity of boundaries: Provides positive markers and exclusion criteria (mere reactivity, isolated processing). Yes, with acknowledged gray zones.
• Non-circularity: Avoids defining consciousness by synonyms like “awareness” alone; instead uses organizational/functional terms (integration, global availability, flexible control, valuation). Yes.
• Essential vs accidental: Identifies characteristics that are necessary for the kind (unity, access, valuation, flexible control) and omits accidents (specific sensory modality, language). Yes.
• Explanatory and predictive power: Entails testable implications (loss under deep anesthesia/coma; presence with global broadcasting/integration; functional changes with attention/valuation). Yes.
• Generality and parsimony: Applies across species and potential artificial systems without invoking mysticism or anthropocentrism; uses a minimal set of core properties. Yes.
• Inter-subjective checkability: Supports public assessment via behavior and physiology, not only private report. Yes.

Validity and universality

• Valid as a naturalistic, objective definition anchored to observable phenomena and general principles. It is designed to be general across humans, nonhuman animals, and engineered agents that instantiate the same organization.
• Universality caveat: As science advances, boundary cases (e.g., simple animals, plants, distributed collectives, or novel AI architectures) may prompt refinement. The definition is broad and current, but open to empirical revision at the margins.

Practical implication for the “planet full of other conscious beings” question

• Apply the criteria functionally: look for unified, globally available, valenced representations that flexibly guide behavior (with appropriate proxies for report). Many mammals and birds, and likely some cephalopods, satisfy these conditions; simpler organisms and plants show sophisticated reactivity and signaling but may lack the unified, globally available field needed for minimal consciousness. The definition provides a non-anthropocentric, testable basis for assessment