1. The Question

What makes us human? The standard answers are three: sociality, tools, language. None is sufficient.

Sociality is not uniquely human. Honeybees maintain division of labor more elaborate than early human societies. Wolf packs coordinate flanking maneuvers, relay pursues, and role rotations. Termites construct structures a million times their body size. If sociality defined humanness, honeybees would be more human than we are.

Tools are not sufficient either. Chimpanzees fish for termites with twigs, New Caledonian crows manufacture hooked tools, sea otters crack shellfish with stones. More telling, Oldowan stone tools remained essentially unchanged for approximately 700,000 years in the archaeological record (Morgan et al. 2015, Nature Communications), demonstrating that early tool production and transmission required none of the capacities we associate with being human.

Language is not the fundamental definition but a co-emergent product of self-completion. Communication (signals, calls, pheromones) belongs to 11DD+12DD, the biological cognitive substrate shared by many animals. Language, however, requires both 13DD—a self that says "I"—and 14DD—institutionalized symbolic rules (grammar and syntax as social contract). A person who loses the capacity for language remains human, because self-completion does not depend on language, though language depends on the emergence of self.

What, then, is the essence of being human?

This paper argues: what makes us human is not mere self-recognition, but the reflexive problematization of one's own annihilation (personal mortality reflexivity) and the structural response it compels. Self-completion means a subject not only possesses self (self-recognition) but that this self has grown strong enough to confront its own extinction as a problem demanding structural response. The formal marker of self-completion is the reflexive problematization of personal mortality—fear of death is its most common phenomenological manifestation, but the marker is not limited to fear (Stoic equanimity, martyrdom, religiously peaceful death are all post-completion response forms, not refutations of self-completion). The structural response to self-completion is myth-ritual closure: a self-consistent symbolic system addressing death. Religion is its most prevalent historical form, but the formal marker is the system's self-consistent closure, not any particular religious tradition.

In the terminology of SAE (Self-as-an-End; foundation paper: Qin, DOI: 10.5281/zenodo.18528813), this is a phase transition within 13DD (the individual subjectivity dimension). The construct layer is 11DD (perception) plus 12DD (cognition)—the biological cognitive substrate. The emergent layer is 13DD self-completion.

This paper applies the four-stage phase-transition model from SAE Methodology Paper VI (DOI: 10.5281/zenodo.19464506), cross-validates a single phase-transition structure using three independent evidence lines (animal behavior, archaeology/paleoanthropology, developmental psychology), and predicts r>>1 asymmetry: the distance from germination to flip-point far exceeds that from flip-point to establishment.

2. What Is Not 13DD

Before constructing the positive argument, three commonly confused concepts must be excluded.

2.1 Cooperation is not self

The cognitive demands of cooperation are lower than those of self-recognition. Wolf pack hunting is highly complex—flanking, relay pursuit, role switching—yet wolves fail mirror self-recognition tests. Great apes pass mirror tests but cooperate less effectively than wolves. Cooperation and self are independent developmental lines; cooperation precedes self-recognition in the construct-layer (11DD+12DD) accumulation sequence.

Cooperation can be entirely driven by 12DD predictive capacity without requiring self. Species with long prediction windows sustain temporally extended behavioral coordination that resembles organized social behavior, yet this remains continuous accumulation within 12DD and does not alter DD level. Domesticated animals demonstrate this most clearly.

2.2 12DD prediction depth is not self: dogs and cats

Dogs. Domestication locked dog 12DD prediction onto a single channel—human intention—pushing the prediction window deep enough to sustain delayed obedience, yet dogs fail mirror self-recognition. Domestication replaced self with deep 12DD prediction: humans make every decision that would require self, permanently closing the dog's self-emergence space. Dog "loyalty" is pseudo-13DD: 12DD colonized to simulate 13DD output (structurally identical to anorexia as pseudo-14DD in SAE Biology Note 3, DOI: 10.5281/zenodo.19501120). Wolves are the pre-domestication reference: broad-spectrum, independent 12DD prediction. Domestication collapsed predictive degrees of freedom onto one channel—deep but narrow. This is regression: not shallower 12DD, but collapsed predictive freedom.

Cats. Cat 12DD prediction windows are extremely short—responding only to immediate stimuli, unable to sustain cross-temporal behavioral commitments. Human misread cat "non-compliance" as self, but it is merely that the prediction window is too short for delayed reinforcement to restructure—cats cannot be colonized, not because self resists, but because 12DD is too short to grasp.

Dogs simulate self through deep (colonized) prediction; cats are projected as self through short (uncolonizable) prediction. Neither possesses 13DD, but humans project 13DD onto them in opposite ways. Both cases demonstrate the same point: 12DD, however deep or shallow, is not 13DD.

2.3 Deep 12DD does not automatically trigger 13DD: Neanderthals

Neanderthal brain volume exceeded that of Homo sapiens. Their tool manufacture was elaborate. Evidence exists for symbolizing and mortuary behavior: intentional burial at Shanidar Cave (Pomeroy et al. 2020), deep-cave speleothem constructions at Bruniquel Cave (~176.5 ka), and U-Th dated cave art on the Iberian Peninsula potentially attributable to Neanderthals. This implies extremely deep 12DD penetration, possibly reaching the germination zone or beyond.

However, current evidence is insufficient to support the claim that Neanderthals possessed a stable, self-consistently closed myth-ritual system—no confirmed narrative scene art (the earliest narrative cave art is attributed to Homo sapiens, Sulawesi, ≥51.2 ka, Aubert et al. 2024, Nature), no large-scale ritual spaces independent of sapiens influence. The exact position of Neanderthals within the four-stage model cannot currently be determined with precision, but the insufficiency of closure evidence suggests they may not have crossed the flip-point, or approached it without reaching establishment.

This provides an important structural insight: deep 12DD does not automatically trigger 13DD emergence. Le Chatelier buffering can be sustained at high cognitive levels for extended periods. This may also explain the outcome of population competition: Constraint Five from the Prequel (DOI: 10.5281/zenodo.19503158) implies that once Homo sapiens achieved 13DD completion, its diffusion rate far exceeded the rate at which Neanderthals could independently complete the phase transition.

3. Definitions

Definition 1: Construct layer (11DD+12DD)

11DD (perception) plus 12DD (cognition) constitute the biological cognitive substrate. This includes but is not limited to: sensory integration, pattern recognition, causal reasoning, spatial navigation, temporal memory, social signal recognition, inhibitory control. The construct layer is a continuous accumulation process with no internal phase transition.

Definition 2: 13DD (emergent layer)

13DD is the individual subjectivity dimension. The emergence of 13DD is not a sharp boundary but a phase-transition window with width: germination marks the onset of emergence (self recognizes self); establishment marks its completion (myth-ritual closure). Self-completion means self has grown strong enough to (1) confront its own annihilation as an inescapable problem, and (2) produce a structural response. Great apes occupy the germination state of 13DD—already emerging but not yet complete.

Definition 3: Language (13DD+14DD)

Language = 13DD + 14DD. Language is not a component of the construct layer (11DD+12DD) but a co-emergent product of 13DD (self) and 14DD (institutionalized symbolic rules). Communication (signals, calls, pheromones) belongs to 11DD+12DD; any organism with perception and cognition communicates. Language requires "I" as speaker (13DD) plus grammar and syntax as social contract (14DD).

Critical clarification: the two legs of 13DD+14DD do not emerge simultaneously. The germination state of 13DD is sufficient to support basic linguistically mediated self-expression—a 2-year-old saying "no" requires only germination-state 13DD (self speaking), not fully emerged 14DD (grammatical contract not yet in place). This is why "terrible twos" language is so simple: two- or three-word combinations without complex syntax. The complete language system (recursive structure, grammatical contract) is not achieved until after 14DD emerges.

Note: Language (13DD+14DD) has two projections in the SAE framework. In the emergence direction, it projects as language (the symbolic interaction protocol between 13DD and 14DD). In the base direction, it projects as law (the negation-suppression defense between 13DD and 14DD). Both are structures for processing the collision between individual (13DD) and institution (14DD), differing in direction. This paper addresses the emergence projection; the SAE Jurisprudence Series addresses the base projection.

Definition 4: Bridge

A bridge is a channel connecting two subjects or a subject and its environment. Bridge thickness is a consequence of 12DD prediction depth and direction, not an independent variable. The bridge between dog and human is thick because dog 12DD has been pushed deep in the direction of predicting human intention; the bridge between cat and human is thin because cat 12DD prediction windows are too short. Bridge thickness does not alter DD level at either end.

Definition 5: Pseudo-13DD

A state in which a lower layer (12DD or below) occupies the 13DD position. Two forms: (a) deep prediction from colonized 12DD simulates self (dogs), (b) short prediction window producing non-compliance projected as self (cats). Diagnostic criterion: ability to say no—not stimulus rejection, but linguistically mediated self-negation.

Definition 6: Four-stage phase-transition structure

Following the framework of SAE Methodology Paper VI, the emergence of 13DD proceeds through four stages:

Stage 1: Germination. Self-recognition. Self recognizes self, but cannot perform any linguistically mediated operation on self.

Stage 2: Spectral flip. Linguistically mediated self-expression. Say no—self actively negates the environment through language. The emphasis in "say no" falls not only on "no" (negation) but equally on "say" (linguistic mediation). Language and self co-emerge at this stage: without self there is no "I" speaking; without language, self cannot self-refer. The spectral flip is not "language enabling self-expression" but self-expression and language appearing together in a single phase transition. This is empirically the most difficult stage to observe directly: in the archaeological line, spoken language leaves no material trace (structural blind spot); in the animal line, it is directly absent; in the developmental line, the evidence (terrible twos / negation development) is the strongest but remains indirect proxy. The spectral flip is the strongest structural inference and the weakest direct evidence.

Stage 3: Flip-point. Personal mortality reflexivity. Self recognizes that self will cease to exist and confronts this as an inescapable problem. Fear of death is the most common phenomenological manifestation, but the formal marker is reflexive problematization itself, not limited to fear as an emotional form. This is the critical phase-transition point.

Stage 4: Establishment. Myth-ritual closure. The structural response of self to death. No longer a scattered fear response, but a self-consistent symbolic system—narrative, ritual, and worldview forming a closed structure addressing death. Religion is its most prevalent historical form.

Definition 7: r>>1 (asymmetry ratio)

The distance from germination to flip-point far exceeds the distance from flip-point to establishment. This is the instantiation in this domain of the core prediction from SAE Methodology Paper VI. Structural intuition for the asymmetry: Le Chatelier buffering operates at full capacity during the germination period, absorbing construct-layer (11DD+12DD) accumulation over extended durations. Once the buffer is breached (self becomes strong enough to self-refer its own annihilation), establishment follows rapidly, because the buffering mechanism itself has fallen below the minimum threshold required to maintain the old pattern.

4. Three Evidence Lines

4.1 Evidence line 1: Animal behavior

4.1.1 Germination: mirror test

Mirror self-recognition (MSR) is the standard operational definition of self-recognition. Strong evidence species include great apes (Gallup 1970, Science) and human infants (~18–24 months). Strong non-primate candidates include bottlenose dolphins (Reiss & Marino 2001, PNAS) and Asian elephants (Plotnik et al. 2006, PNAS).

Gray-zone cases deserve particular attention. Reports of cleaner wrasse (Kohda et al. 2019, 2022) passing the mark test have reignited fundamental debates about what MSR measures. Magpie MSR evidence was not replicated in Soler et al. 2020. Horse MSR has been reported at group level but with methodological disputes (Baragli et al. 2021). Manta rays and octopuses show preliminary contingency-checking behavior far short of mark-test standards.

The evidential structure most favorable to this paper's argument is this: even in MSR-passing species, no further leap from self-recognition to higher-order self-behavior has been observed. Cleaner fish post-MSR can "recall body size" (Kobayashi et al. 2024), but this remains body-as-object representation, not a symbolic self-model. MSR-passing species remain in the germination zone—precisely supporting the judgment that the buffer has not been breached.

4.1.2 Spectral flip and beyond: absent

No non-human animal exhibits linguistically mediated self-expression. Great apes do not say no (in this paper's technical sense). Animals cannot reach the spectral flip, not because they lack a tool (language), but because 13DD has never emerged sufficiently to support language—without self there is no language, and without language self cannot self-refer.

The animal evidence line is permanently stuck at germination. This is not an absence of evidence but a structural absence—living evidence that the Le Chatelier buffer has never been breached.

4.1.3 Animal death responses: not death awareness

The animal thanatology literature requires careful parsing. Chimpanzee mothers carrying dead infants (Lonsdorf et al. 2020) show behavioral evidence that they may recognize death has occurred, but this is recognition of another's death, not personal mortality. Crow responses to dead conspecifics (Swift et al. 2020) are best explained as information gathering and danger learning, not death concepts. Elephant responses to the dead are reviewed in dedicated literature (Goldenberg & Wittemyer 2020), but inference from behavior to concept remains highly contested.

Monsó's minimal death-understanding criteria require evidence that an animal (1) first expects the individual to be alive, then (2) grasps non-functionality and irreversibility. Cross-species, current evidence supports this: many species show systematic, prolonged responses to the dead; some behaviors are consistent with partial death recognition; but strong claims about human-like death concepts—especially personal mortality—remain evidentially underconstrained.

4.2 Evidence line 2: Archaeology and paleoanthropology

4.2.1 Germination: cave handprints

Cave handprints are the archaeological mirror test—"this is my hand." What is depicted is not prey or environment but a trace of one's own body. The earliest handprints (Sulawesi, ~40+ ka) require more precise dating to establish temporal relations with other cave art, but as self-referential material traces, their symbolic function differs from animal-scene depiction.

4.2.2 Spectral flip: teaching traces (structural blind spot)

The archaeological line has a structural blind spot at the spectral flip: spoken language leaves no material trace. Language as a co-emergent product of 13DD+14DD is invisible in the archaeological record. It almost certainly preceded grave goods but left no direct evidence.

Indirect evidence comes from tool standardization. Morgan et al. 2015 demonstrated experimentally that Oldowan tools could be transmitted through imitation, but high-fidelity transmission of Acheulean handaxes required teaching, especially linguistically mediated teaching. Lombao et al. 2017 (Scientific Reports) further reported that verbal teaching outperformed gestural teaching in efficiency and information retention.

However, this inference has limits. Ferar 2026's "puppet method" experiment showed that knapping-naive individuals can produce handaxe shapes without cultural transmission, meaning Acheulean standardization cannot strictly entail language. The Morgan and Lombao conclusions concern transmission efficiency and fidelity, not strict necessity.

This paper's position: the jump in tool standardization during the Oldowan-to-Acheulean transition (~2.6 Ma to ~1.76 Ma) is consistent with teaching and proto-language as facilitating conditions, but does not constitute strict entailment. The archaeological line's evidential strength at the spectral flip is weaker than at other stages—this honest acknowledgment is itself a finding.

4.2.3 Flip-point: grave goods

Burial itself is not evidence of personal mortality reflexivity. Burial can be hygienic behavior—disposing of decomposing remains. Evidence for reflexive problematization consists of grave goods and symbolically marked mortuary practices: orientation choices, ochre use, placement of non-utilitarian objects with the dead. These do not occur without self confronting death as a problem.

The Operationalization is important: 12DD deep prediction can process "irreversible physical damage (non-functionality)," so burial for sanitary or predator-prevention reasons is the 12DD optimum. Only 13DD can experience "ontological void (personal mortality)"—hence non-utilitarian grave goods. Burial is 12DD's best solution; grave goods are 13DD's patch against the void.

The Neanderthal burial controversy falls precisely on this distinction. The Shanidar "flower burial" (Solecki's original interpretation) has been significantly weakened: Hunt et al. 2023 concluded that pollen clumps most likely originated from nesting solitary bees, not intentionally placed flowers. But Pomeroy et al. 2020 supports deliberate burial at Shanidar. La Ferrassie and other sites also provide evidence for intentional burial.

The key question is not whether Neanderthals buried their dead (they very likely did) but whether burial was accompanied by symbolic behavior. Current evidence tends toward: some intentional body placement, but weak support for ritual offerings. This is consistent with the framework—Neanderthals may have been near the flip-point without clearly crossing it.

For Homo sapiens, evidence of symbolic mortuary practice is stronger. Ochre use and beadwork appear around ~100 ka, typically used as anchors for the emergence of symbolic systems.

4.2.4 Establishment: myth-ritual closure

The marker of establishment is not individual symbolic objects but the self-consistent closure of a symbolic system—from scattered symbolic use to a coherent worldview.

The strongest early ritual-space evidence comes from Manot Cave (Barzilai et al. 2024, PNAS): during the Early Upper Paleolithic (~37–35 ka BP), a ritual compound segregated from living areas, featuring an engraved boulder with geometric signs (possibly representing a tortoise), wood ash for illumination, and acoustics suitable for communal gathering. Multiple independent evidence lines converge on "non-mundane collective practice."

An earlier candidate includes the Neanderthal speleothem constructions at Bruniquel Cave (~176.5 ka), sometimes interpreted as evidence for deep-cave communal activity. But the inference from ritual behavior to "myth-ritual closure" requires additional independent support.

Sulawesi narrative cave art (≥51.2 ka, Aubert et al. 2024, Nature) displays scene-based depiction and possible therianthrope figures, implying narrative imagination. This marks the jump from symbolic objects to symbolic narrative, but does not yet equal complete myth-ritual closure (which is almost unrecoverable from the archaeological record).

The main limitation of the archaeological line at the establishment stage: we can observe ritual spaces and narrative art, but cannot directly observe mythic texts. Inference from "ritual behavior" to "worldview closure" is always indirect.

4.3 Evidence line 3: Developmental psychology

4.3.1 Germination: ~18 months

Mirror self-recognition in standard developmental paradigms emerges around 18–24 months. Cross-cultural research indicates onset variation (related to mirror exposure and social practices), but this is temporal variation, not capability absence.

A directly relevant neuroscience finding: Bulgarelli et al. 2019 (Developmental Cognitive Neuroscience) reported that 18-month-olds who passed mirror self-recognition ("Recognizers") had stronger fronto-temporoparietal resting-state connectivity than non-passers. This connectivity pattern substantially overlaps with adult DMN regions implicated in self-reflection.

4.3.2 Spectral flip: ~2 years

The "terrible twos" in developmental psychology are typically framed through autonomy and self-regulation. Hughes et al. 2020 (Developmental Science) linked toddler noncompliance to early executive function and parent-child interaction quality.

Negation research in the language-acquisition literature directly supports the operationalization of "say no": toddlers' earliest "no" is rejective (rejecting offers/commands), later developing toward more adult-like denial, typically in the 2–2.5 year range. Cross-linguistic research (2025) shows that the distribution of negative functions varies with language environment, but the basic timeline is cross-culturally consistent.

A critical operational distinction: 12DD stimulus rejection and 13DD self-assertion are not the same. An animal refusing disliked food is a 12DD optimal-prediction response. 13DD "no" rejects not just an object but the other's framework—toddlers frequently say "no" to candy they actually want, purely to test the boundaries of their own subjectivity. This negation against one's own 12DD biological interest, purely to assert self-boundaries, is the true marker of the spectral flip.

The emphasis in "say no" falls not only on "no" but equally on "say." A 2-year-old saying "no" requires only germination-state 13DD—self speaking—without fully emerged 14DD (grammatical contract not yet in place). This is why terrible-twos language is so simple. Animals cannot reach the spectral flip because 13DD has never emerged to a level sufficient to support linguistically mediated self-expression.

Neurodevelopmental data: Fiske et al. 2024 (Imaging Neuroscience) used longitudinal fNIRS to track inhibitory control development from 10 to 16 months, finding a shift from right-lateralized PFC/parietal recruitment toward broader bilateral recruitment including right inferior frontal gyrus and bilateral dorsolateral/orbitofrontal PFC. The researchers interpreted this as a possible reorganization period—supporting "phase transition" framing over "smooth maturation."

4.3.3 Flip-point: ~4–5 years

The modern literature on children's death cognition decomposes "understanding death" into sub-components: universality, irreversibility/finality, cessation/nonfunctionality, and causality.

Menendez et al. 2020 reported that before age 5, children begin developing universality and irreversibility understanding; by age 5, most understand cessation of bodily processes; by age 6, many understand that death can have multiple causes. A 2026 narrative review confirmed that children grasp universality, irreversibility, and biological cessation before age 12.

The flip-point is around 4–5 years: children begin asking "Will people die?", "Will I die?", "Where do you go after death?" This marks self as strong enough to make its own disappearance a problem.

This time-point notably overlaps with the classic theory-of-mind milestone—passing the false belief task (~4 years). TPJ and mPFC are repeatedly confirmed in adult belief reasoning and perspective-taking (neuroimaging meta-analyses), and functional specialization of these regions strengthens significantly around 4–5 years. A defensible position: ToM network consolidation is a necessary (but not sufficient) scaffold for explicit death reasoning.

4.3.4 Establishment: nearly instantaneous

Richert & Corriveau 2022 (Annual Review of Developmental Psychology) reported that by ~5 years, children tend to claim that unobservable natural and supernatural phenomena exist, with confidence tracking cultural consensus and home/community endorsement. Menendez et al. 2020 applied this directly to death: children construct death understanding through conversation, media, and ritual participation, with biological and religious explanations often coexisting rather than being treated as contradictions.

Cross-cultural experiments (Vanuatu vs. U.S.) showed that brief narrative primes with natural vs. supernatural cues modulate afterlife endorsement patterns across age groups, consistent with rapid uptake of culturally scaffolded frameworks under testimony.

Qualification on "instantaneous": the literature supports early and efficient uptake of testimony-based frameworks, but "nearly instantaneous" depends on what counts as "acceptance" (explicit assent, behavioral compliance, emotional comfort, or stable belief). The most defensible claim: young children can coordinate multiple explanatory systems early—including religious accounts of what happens after death—without needing complete biological death mastery first.

4.4 Textual verification: Gilgamesh

Beyond the three evidence lines, humanity's oldest surviving long narrative directly narrativizes the 13DD phase transition.

The Epic of Gilgamesh (c. 2100 BCE, Sumer) has a core structure: Gilgamesh's companion Enkidu dies. Enkidu's death is not a plot event but the flip-point—Gilgamesh recognizes that he too will die. "My friend, whom I loved, has turned to clay. Shall I not, like him, lie down and never rise again?" This is the prototypical expression of self confronting its own annihilation as a problem.

Gilgamesh then crosses the world seeking immortality (attempting structural response), ultimately fails—the herb of eternal life is swallowed by a serpent. He returns to Uruk, gazes at its walls, and accepts death.

Humanity's earliest written long story has personal mortality reflexivity as its subject. This is not coincidence. After 13DD completion, the first thing requiring narrativization is the flip-point itself. The Epic of Gilgamesh is a literary fossil of the 13DD phase transition. Additional cross-civilizational narratives appear in Appendix A.

5. The r>>1 Asymmetry

5.1 Developmental line

From germination (~18 months) to flip-point (~54 months, i.e. ~4.5 years): approximately 36 months.

From flip-point to establishment: nearly simultaneous.

r ≈ 36 / ε, where ε approaches zero (establishment distance is nearly unmeasurable in months). r>>1 is cleanest on the developmental line.

5.2 Evolutionary line

From germination (great-ape-level MSR, at least several million years ago) to flip-point (appearance of symbolic mortuary practice, ~100 ka order of magnitude): millions of years.

From flip-point to establishment (ritual spaces and narrative art, ~50–35 ka): tens of thousands of years.

Rough estimate: r on the order of 100 (millions of years vs. tens of thousands of years). This estimate is very coarse, depending on dating uncertainties and operationalization of stage markers, but r>>1 is robust at the order-of-magnitude level.

5.3 Animal line

The animal line provides the extreme case of r = ∞: stuck in germination for millions of years, flip-point never reached. Le Chatelier buffering never breached.

5.4 Resolution problem in the flip-to-establishment interval on the archaeological line

A limitation must be acknowledged: on the archaeological line, the signal for "mortality awareness" (burial, repeated body placement, grave goods) and the signal for "myth-ritual closure" (ritual spaces, complex narratives, institutionalized myth) do not always cluster tightly in time. Depending on proxy choices and evidential conservatism, they may be separated by tens of thousands of years. This makes the archaeological r>>1 argument less clean than the developmental line.

But this is precisely the value of cross-validating three evidence lines: the developmental line provides the cleanest r>>1 evidence (high resolution, controllable variables), the archaeological line provides evolutionary-scale validation (low resolution but enormous time span), and the animal line provides an extreme control where the buffer was never breached. Each line has limitations, but their limitations do not overlap.

6. Remainder Rate and Species Orders of Magnitude

The strength of Le Chatelier buffering can be quantified another way: across Earth's history, how many species arose, and how many reached each DD level?

6.1 Data

The total number of multicellular eukaryotic species across the Phanerozoic (~540 million years) is standardly estimated from current species count (~8.7 million, Mora et al. 2011, PLoS Biology) and extinction rate (99%–99.9%, Jablonski standard estimate). Taking the upper bound: 8.7 million / 0.001 ≈ 8.7 billion species.

The vast majority remained at 9DD+10DD (basic multicellular organizational level). Species with 12DD cognitive capacity—possessing central nervous systems, capable of learning and prediction—roughly encompass the mammals, birds, and cephalopods, totaling on the order of tens of thousands of species historically.

Species reaching 13DD (self-completion): 1. Homo sapiens.

6.2 Two-step structure of the remainder rate

The remainder rate used in the SAE Physics Series Finale (DOI: 10.5281/zenodo.19464378) and the Prequel (DOI: 10.5281/zenodo.19503158) is approximately 10^{-5} per step. From 9DD+10DD to 13DD requires two steps:

9DD+10DD (multicellular organisms) → 11DD+12DD (cognitive species): 8.7 billion × 10^{-5} ≈ 87,000

11DD+12DD (cognitive species) → 13DD (self-completion): 87,000 × 10^{-5} ≈ 0.87

Two steps, each ~10^{-5}, total remainder rate ~10^{-10}. From 8.7 billion multicellular species to 0.87 13DD species.

6.3 Interpretation

0.87 ≈ 1. At the order-of-magnitude level, Earth across its Phanerozoic history (540 million years) is at precisely the scale to produce approximately one 13DD species. This should not be read as "nearly didn't happen"—Earth has roughly 1 billion years of habitable window remaining; had Homo sapiens not emerged at this point, subsequent species would eventually have crossed the flip-point. The correct reading of 0.87 is: by the current stage of the Phanerozoic, the remainder rate predicts approximately one 13DD species emergence.

The critical point is Constraint Five (Prequel): once the first 13DD species emerges, its diffusion rate far exceeds the rate of independent emergence of a second 13DD species. The slot is permanently occupied. There will not be a second. This is consistent with the Lonely Star Theorem at the planetary level: neither more nor less—exactly one.

The specific value of ~10^{-5} per step has not yet been rigorously proven within the SAE framework (the Physics Series Finale provides a derivation pathway but has not concluded); this paper uses it as established series data. If the precise remainder rate is revised in future, the order-of-magnitude argument in this section will need corresponding update, but the two-step structure is invariant.

7. The Position of Neuroscience

This paper does not claim that any single neural structure "equals" self. Self is an emergent layer and cannot be reduced to the construct layer. What neuroscience can provide is a measurement window into construct-layer states before and after the phase transition.

7.1 Cross-species DMN comparison as construct-layer accumulation indicator

The Default Mode Network (DMN) in humans is tightly coupled with self-referential thought and temporal projection (recalling past / imagining future). Core DMN hubs include medial prefrontal cortex (mPFC), posterior cingulate cortex/precuneus (PCC/Prec), and angular gyrus.

Key cross-species comparative data come from Garin et al. 2022 (Cell Reports): in non-hominid primates, DMN homologous regions fractionate into two networks, with weak mPFC-PCC connectivity—precisely the strong connectivity that is a core feature of the human DMN. Chimpanzee DMN equivalence is closer to human (Barks et al. 2015, Cerebral Cortex), with PET evidence showing task deactivation of cortical midline areas during social discrimination, consistent with the human DMN task-negative effect.

Significance for this argument: if self-completion requires a tightly integrated self-model—linking value/identity (mPFC) with autobiographical scene construction/context (PCC/Prec and medial temporal contributions)—then connectivity architecture, rather than presence of individual regions, becomes the plausible indicator of whether construct-layer accumulation has reached phase-transition conditions. Non-hominid primates have DMN components but insufficient connectivity—the construct layer is accumulating but has not reached the flip-point.

7.2 Developmental timeline alignment with four stages

Bulgarelli et al. 2019: 18-month Recognizers have stronger fronto-temporoparietal connectivity—corresponding to germination.

Fiske et al. 2024: 10–16 month PFC inhibitory control reorganization—possibly corresponding to neural preparation for the spectral flip.

TPJ/mPFC functional specialization around age 4 for false-belief-task performance—corresponding to the scaffold for the flip-point.

These nodes do not map one-to-one onto the four stages (a single neural transition can support multiple behavioral changes simultaneously), but they support a picture of "stage-like jumps" rather than "smooth maturation."

7.3 Neural markers of death awareness

Hirano et al. 2021 (Cerebral Cortex Communications): when humans think about their own death, PCC (core DMN hub) is specifically activated in the self condition, and fear-of-death modulates right supramarginal gyrus response (negative linear relationship) and PCC response (inverted-U relationship). PCC involvement is interpreted as processing death-related thought as a self-relevant future agenda—directly connecting death cognition to core DMN function.

A 2026 Neuropsychologia systematic review found that compared with generally unpleasant stimuli, death-related stimuli tend to reduce insular activity. This suggests death-related processing differs from generic threat processing, possibly involving a distinctive mode of self-referential processing rather than amplified fear circuitry.

The claim these data support is epistemological: personal mortality reflexivity is not stronger fear—it is a distinctive form of self-referential processing.

8. Relation to Existing Theories

8.1 Terror Management Theory (TMT)

TMT (Greenberg, Solomon, & Pyszczynski) proposes that objective self-awareness makes death salient and that cultural worldviews (including religion) buffer existential anxiety. This is closest to this paper's "mortality reflexivity → myth-ritual closure" chain.

But TMT lacks a four-stage structure, lacks r>>1 quantitative prediction, and lacks a three-line cross-validation architecture. More importantly, TMT treats death anxiety → cultural worldview as a single-step causal relationship, whereas this paper's framework decomposes it into a multi-step phase transition from germination to establishment, each step with distinct markers and evidential standards.

TMT's empirical foundations are contested: large-scale replication efforts have challenged some classic mortality-salience effects, and meta-analytic debates highlight potential publication bias and heterogeneity. This paper's use of TMT is limited to: TMT's theoretical framework supports the overall direction of "death awareness → cultural response," but this paper does not depend on TMT's specific experimental paradigms.

8.2 Varki's MORT theory

Ajit Varki's Mind Over Reality Transition (MORT) theory is this paper's closest competitor. MORT treats the emergence of mortality salience as the key evolutionary event of human uniqueness and links it to cultural consequences.

This paper's incremental contributions: (1) four-stage structure rather than single-step crossing, (2) r>>1 quantitative asymmetry prediction, (3) three independent evidence lines cross-validating rather than a single evolutionary narrative, (4) integration with the SAE framework providing cross-domain theoretical unity (the same phase-transition structure applies to metabolic oncology, experimental design methodology, economics, etc.).

8.3 MSR critique literature

The modern MSR literature is increasingly cautious about MSR as a unified "self-awareness" indicator. The cleaner-fish controversy, magpie replication failure, and trained-macaque MSR studies are all eroding MSR's reliability as a clean species-level cognitive boundary.

This is actually favorable to this paper. If MSR is merely a narrow "self-as-body-object" competence, then MSR can persist over long evolutionary durations without yielding a mortality-reflexive or symbolically closed self—making the argument easier. This paper places MSR at the germination position rather than the flip-point, precisely consistent with the MSR literature's self-critique.

9. Nontrivial Predictions

Prediction 1: Animals will never reach the spectral flip

Without the emergence of 13DD, no non-human species can exhibit linguistically mediated self-negation (say no in the technical sense). If a future species is found capable of this, it must simultaneously possess (a) MSR capacity and (b) some form of symbolic negation expression.

Falsification conditions: Discovery of a species that cannot pass MSR but can say no, or discovery of a species that passes MSR and can say no but does not develop death awareness. The former falsifies the germination → spectral-flip ordering; the latter falsifies the spectral-flip → flip-point ordering.

Prediction 2: Archaeological evidence for the spectral flip will always be indirect

Spoken language is invisible in the archaeological record. Future discoveries may provide more indirect evidence of teaching traces (e.g., sudden jumps in tool standardization), but cannot provide direct evidence of speech itself.

Falsification condition: Discovery of direct material traces of spoken language (physically near-impossible, but logically necessary to state).

Prediction 3: r>>1 can be quantitatively verified in child development

The temporal distance from germination to flip-point (~18 months to ~4–5 years) should be systematically greater than that from flip-point to establishment (~4–5 years to acceptance of myth-ritual framework). This can be directly quantified in longitudinal developmental studies.

Falsification condition: Longitudinal data showing flip-to-establishment distance comparable to or greater than germination-to-flip distance (i.e. r ≈ 1 or r < 1).

Prediction 4: Neanderthal archaeological record should lack myth-ritual closure

The exact position of Neanderthals within the four stages cannot currently be precisely determined, but if this paper's framework is correct, their archaeological record should contain symbolizing and mortuary behavior evidence (consistent with germination zone and partial flip) but lack a stable, self-consistently closed myth-ritual system—specifically (a) large-scale ritual spaces independent of sapiens influence, (b) narrative scene art, (c) recoverable myth-ritual closure structure.

Falsification condition: Discovery of Neanderthal-made narrative scene art, or a large-scale ritual space independent of sapiens influence supported by multiple converging independent evidence lines.

Prediction 5: No continuous relationship between 12DD depth and 13DD emergence

Cross-species comparison should show that increases in cognitive capacity (12DD penetration depth) do not lead linearly toward self-completion (13DD emergence). Some species will have very deep 12DD with no 13DD indications (e.g. Neanderthals, if future evidence supports this judgment), while Homo sapiens 13DD emergence appears to have occurred "suddenly" once cognitive capacity reached a certain threshold.

Falsification condition: Cross-species data showing a smooth linear relationship between cognitive capacity indicators and self-completion indicators, with no threshold effect.

10. Open Questions

• What is the archaeological proxy for the spectral flip? Teaching traces (tool standardization) are the best current candidate, but Ferar 2026 weakens "standardization = language" as strict reasoning. Are there better proxies?

• Where exactly are Neanderthals in the four stages? Current evidence is insufficient for precise positioning. If future evidence reveals Neanderthal symbolic-system evidence independent of sapiens influence, this framework would need to revise its Neanderthal staging judgment, but not the four-stage structure itself.

• How to quantify r across evidence lines? On the developmental line, r can be directly calculated in months. On the evolutionary line, r depends on archaeological dating and proxy choices. On the animal line, r = ∞. Can the r values from all three lines be unified into a single dimensionless comparison framework?

• Diversity of myth-ritual systems. This paper uses myth-ritual closure as the formal marker of establishment, but different civilizations' structural responses to death vary enormously. Are these differences variation within the establishment stage, or do they point to different establishment pathways?

• After self-completion, then what? How do 13DD-complete subjects form institutions? Individual 13DD relationships are constrained by Dunbar's number (~150 people)—beyond this scale, self-to-self direct relationships cannot be maintained, and 14DD (institution) must emerge to organize larger groups. A noteworthy material signature: within 150 people, spoken language (supportable by germination-state 13DD alone) suffices to maintain all social information—who owes whom, who feuds with whom, where prey can be found—without externalizing information to any medium. The appearance of writing (Sumer, ~3200 BCE, when settlements had reached thousands to tens of thousands) was not a technological advance but an inevitable product of 14DD emergence: once groups exceeded face-to-face management range, contracts, laws, and accounts had to be externalized beyond individual memory. Spoken language → writing may be the material signature of 13DD → 14DD. Political organization from band to tribe to chiefdom to state (Service's sequence) corresponds in the SAE framework to 14DD emerging from a 13DD construct layer. Does this emergence also exhibit a four-stage phase-transition structure? Does r>>1 similarly hold? This is the subject of Paper 2.

11. Conclusion

What makes us human is not sociality (honeybees are more social), not tools (crows use tools too), not cooperation (wolf cooperation is more effective), not 12DD prediction depth (dogs predict human intention more accurately than any great ape yet have no self), nor even cognitive capacity (Neanderthal brain volume exceeded that of Homo sapiens).

What makes us human is self-completion: self grown strong enough to reflexively problematize its own annihilation and produce a structural response through myth-ritual closure. This is the emergence of 13DD from the 11DD+12DD construct layer, proceeding through four stages (germination, spectral flip, flip-point, establishment), with r>>1.

Three independent evidence lines—animal behavior, archaeology/paleoanthropology, developmental psychology—cross-validate the same phase-transition structure. The animal line is stuck at germination; the archaeological line has a structural blind spot at the spectral flip (spoken language leaves no trace); the developmental line provides the most complete observable sequence. Each line has limitations, but their limitations do not overlap. Humanity's oldest literary text (the Epic of Gilgamesh) directly narrativizes the flip-point itself, providing independent corroboration from the textual tradition.

This paper's core contribution: placing mirror self-recognition, personal mortality reflexivity, and myth-ritual closure within a single phase-transition framework—without precedent in the existing literature. Existing theoretical connections are all pairwise (MSR ↔ self, death cognition ↔ thanatology, mortality ↔ religion via TMT). The four-stage structure and the r>>1 quantitative prediction are new.

---

Relation to the SAE framework: This paper is the first in the SAE Anthropology Series, addressing the emergence of 13DD (individual subjectivity) from 11DD+12DD (cognitive substrate). The Series Prequel (DOI: 10.5281/zenodo.19503158) provides cosmic-scale background structure: the Lonely Star Theorem and Constraint Five's species-competition extension. Paper 2 will address the emergence of 14DD (institution) from 13DD—how self-complete subjects form political and social organization. Construct-layer/emergent-layer relations, Le Chatelier buffering, r>>1 asymmetry, and bridge independence are all concepts formalized in SAE Methodology Paper VI, here instantiated in the anthropological domain. Cross-domain consistency of phase-transition structure (metabolic oncology, experimental design, economics, now anthropology) provides cumulative evidence for the generality of the SAE framework.

---

Appendix A: Cross-Civilizational 13DD Narratives

Beyond Gilgamesh, at least three independent civilizational traditions preserve the narrative structure of the 13DD phase transition.

Genesis (Hebrew tradition). Adam and Eve eat from the tree of knowledge of good and evil—"you will be like God, knowing good and evil." Knowing good and evil is self-completion: self's judgment leaps from passive acceptance of environment to active evaluation. The punishment? "You will surely die." The narrative structure precisely encodes this paper's thesis: the cost of self-completion is knowing you will die. Eden is the germination zone; eating the fruit is the flip-point; expulsion is the irreversibility of establishment.

The Buddha's departure (Indian tradition). Siddhartha within the palace walls knows nothing of old age, sickness, or death (germination zone—self exists but is buffered; Le Chatelier buffering provided by the king's protection). Leaving the palace, he encounters an old man, a sick man, a dead man (flip-point—self recognizes self will cease to exist). Renunciation and seeking the path (establishment—structural response). The narrative even preserves the r>>1 asymmetry: twenty-nine years within the palace (germination period); from departure to renunciation, nearly instantaneous.

Achilles in the underworld (Greek tradition). Homer's Odyssey, Book XI: Odysseus encounters Achilles' shade in the underworld. Achilles says: "I would rather be a living slave than king among the dead." This is 13DD at its extreme—self-completion pushed to its limit, where even heroic honor (proto-14DD institutionalized value) cannot compensate for acceptance of death. Among all Greek heroes, Achilles is the only one who knew in advance he would die and chose to go to war anyway—his choice itself is a structural response to self-completion.

---

References

(In order of first appearance)

Morgan, T. J. H. et al. (2015). Experimental evidence for the co-evolution of hominin tool-making teaching and language. Nature Communications, 6, 6029.

Qin, H. (2024). SAE Foundation Paper 1. DOI: 10.5281/zenodo.18528813.

Qin, H. (2026). SAE Methodology Paper VI: Phase Transition Window and Experimental Design. DOI: 10.5281/zenodo.19464506.

Qin, H. (2026). SAE Biology Note 3. DOI: 10.5281/zenodo.19501120.

Qin, H. (2026). SAE Learning Series Paper 4. DOI: 10.5281/zenodo.19491926.

Qin, H. (2026). SAE Anthropology Series Prequel: The Cosmic Background of Anthropology. DOI: 10.5281/zenodo.19503158.

Gallup, G. G. Jr. (1970). Chimpanzees: Self-recognition. Science, 167(3914), 86-87.

Reiss, D. & Marino, L. (2001). Mirror self-recognition in the bottlenose dolphin. PNAS, 98(10), 5937-5942.

Plotnik, J. M. et al. (2006). Self-recognition in an Asian elephant. PNAS, 103(45), 17053-17057.

Kohda, M. et al. (2019). If a fish can pass the mark test, what are the implications for consciousness and self-awareness testing in animals? PLoS Biology, 17(2), e3000021. doi: 10.1371/journal.pbio.3000021. Update: Kohda, M. et al. (2022). PLoS Biology, 20(2), e3001529. doi: 10.1371/journal.pbio.3001529.

Soler, M. et al. (2020). Replication of the mirror mark test experiment in the magpie (Pica pica) does not provide evidence of self-recognition. Journal of Comparative Psychology, 134(4), 363-371. doi: 10.1037/com0000223.

Baragli, P. et al. (2021). If horses had toes: demonstrating mirror self recognition at group level in Equus caballus. Animal Cognition, 24(5), 1099-1108. doi: 10.1007/s10071-021-01502-7.

Kobayashi, T. et al. (2024). Cleaner fish body-size recall after MSR.

Lonsdorf, E. V. et al. (2020). Chimpanzee mothers carrying dead infants.

Swift, K. N. et al. (2020). Crow neuroimaging study: responses to dead conspecifics.

Goldenberg, S. & Wittemyer, G. (2020). Elephant thanatology review.

Monsó, S. (2021). How to tell if animals can understand death. Erkenntnis, 86, 1411-1431. doi: 10.1007/s10670-019-00175-4.

Hunt, C. O. et al. (2023). Shanidar pollen reassessment.

Pomeroy, E. et al. (2020). New Neanderthal remains at Shanidar.

Aubert, M. et al. (2024). Sulawesi narrative cave art. Nature.

Barzilai, O. et al. (2024). Manot Cave ritual compound. PNAS.

Lombao, D. et al. (2017). Verbal vs gestural teaching in knapping. Scientific Reports.

Ferar, N. (2026). Puppet method: handaxe shaping without cultural transmission.

Bulgarelli, C. et al. (2019). fNIRS connectivity and mirror self-recognition at 18 months. Developmental Cognitive Neuroscience.

Hughes, C. et al. (2020). Toddler noncompliance and executive function. Developmental Science.

Fiske, A. et al. (2024). Longitudinal fNIRS of infant inhibitory control. Imaging Neuroscience.

Menendez, D. et al. (2020). Children's death understanding via subcomponents.

Richert, R. A. & Corriveau, K. H. (2022). Religious cognition in children. Annual Review of Developmental Psychology.

Garin, C. M. et al. (2022). Cross-species DMN comparison. Cell Reports.

Barks, S. K. et al. (2015). Chimpanzee DMN via PET. Cerebral Cortex.

Hirano, K. et al. (2021). Neural correlates of self-relevant death processing. Cerebral Cortex Communications.

Varki, A. (2009). Human uniqueness and the denial of death. Nature, 460, 684. doi: 10.1038/460684c. See also: Varki, A. & Brower, D. (2013). Denial: Self-Deception, False Beliefs, and the Origins of the Human Mind. Twelve Books.

Greenberg, J., Solomon, S., & Pyszczynski, T. (1997). Terror management theory of self-esteem and cultural worldviews. Advances in Experimental Social Psychology, 29, 61-139. doi: 10.1016/S0065-2601(08)60016-7.

Mora, C. et al. (2011). How many species are there on Earth and in the ocean? PLoS Biology, 9(8), e1001127. doi: 10.1371/journal.pbio.1001127.

一、问题的提出

人之所以为人，标准答案通常有三个：群居，工具，语言。这三个答案都不够。

群居不是人独有的。蜜蜂的社会分工比早期人类复杂得多。狼群的协作猎杀需要角色分配和实时信号协调。白蚁能建造比自身大一百万倍的结构。如果群居是人之所以为人的标志，那蜜蜂比人更"人"。

工具也不是。黑猩猩用树枝钓白蚁，新喀里多尼亚乌鸦制造钩状工具，水獭用石头敲开贝壳。更重要的是，Oldowan石器在考古记录中保持了大约七十万年的基本不变（Morgan et al. 2015, Nature Communications），这说明早期工具的制作和传播可以完全不依赖我们认为"人"应该有的那些能力。

语言也不是。语言不是人之所以为人的根本定义，而是self完备过程的共生产物。交流（信号、叫声、信息素）属于11DD+12DD的生物认知层，许多动物都有。但语言是13DD+14DD的——它同时需要13DD的self（"我"在说）和14DD的制度化符号规则（语法和句法是社会契约）。一个失去语言能力的人仍然是人，因为self完备不依赖语言，但语言的出现依赖self的涌现。

那么人之所以为人的本质是什么？

本文的论点是：人之所以为人，不在一般自我识别，而在self对自身消亡的反身问题化（personal mortality reflexivity）及其符号-仪式回应。self完备的意思是：一个主体不仅拥有self（自我识别），而且这个self强到能够把自身的消亡当作问题来面对，并且产生结构性的回应。self完备的formal marker是self将自身消亡反身化/问题化——其最常见的现象学表现是怕死，但不限于恐惧（斯多葛式的平静面对、烈士的主动赴死、宗教传统中的安宁死亡，都是self完备后的不同回应形式，而非对self完备的否证）。self完备的结构性回应是关于死亡的符号-仪式系统闭合（myth-ritual closure）——宗教是其最普遍的历史形式，但formal marker是符号系统的自洽闭合，不限于任何特定宗教传统。

在SAE（Self-as-an-End，自我目的论框架；基础论文见Qin, DOI: 10.5281/zenodo.18528813）的术语中，这是13DD（个体主体性维度）内部的相变。构层是11DD（感知）加12DD（认知）的生物认知基底。涌现层是13DD的self完备。

本文使用SAE Methodology Paper VI（DOI: 10.5281/zenodo.19464506）的四阶段相变模型，用三条独立证据线（动物行为，考古/古人类学，儿童发展心理学）交叉验证同一个相变结构，并预测r>>1的不对称性：从萌芽到翻转的距离远大于从翻转到确立的距离。

二、排除项：什么不是13DD

在建立正面论证之前，先排除三个容易混淆的概念。

2.1 协作不是self

协作的认知要求低于自我识别。狼群的协作猎杀高度复杂：侧翼包抄，接力追逐，角色轮换。但狼不通过镜子测试。大猿通过镜子测试，但协作能力不如狼。这说明协作和self是两条独立的发展线，协作在构层（11DD+12DD）的积累序列中甚至排在自我识别之前。

更具体地说，协作可以完全由12DD的预测能力驱动，不需要self。预测窗口长的物种能维持跨时间的行为协调，看起来像高度组织化的社会行为，但这仍然是12DD内部的连续积累，不改变DD层级。这一点在驯化动物身上看得最清楚。

2.2 12DD预测深度不等于self：狗与猫

狗。 驯化把狗的12DD预测方向锁死到人类一个通道上——预测窗口长到足以维持延迟服从，但狗不通过镜子测试。驯化用12DD的深预测替代了self：人类替狗做所有需要self才能做的决定，狗的self涌现空间被永久关闭。狗的"忠诚"是12DD被殖民后模拟出的伪13DD（pseudo-13DD，与SAE Biology Note 3中anorexia作为pseudo-14DD是同一结构）。狼是驯化前的参照：12DD预测广谱且独立，驯化把预测自由度塌缩到人类一个通道——深但窄，这就是退化。

猫。 猫的12DD预测窗口极短，只对即时刺激反应，无法维持跨时间的行为承诺。猫的"不配合"被人类误读为self，但实际上只是预测窗口太短，无法被延迟奖惩重构——猫无法被殖民，不是因为有self在抵抗，而是因为12DD太短抓不住。

狗用深预测伪装self，猫用短预测被投射为self。两者都没有13DD，但人类投射13DD的方式恰好相反。这两个例子说明的是同一件事：12DD无论多深、多浅，都不是13DD。

2.3 12DD穿透深度再深也不自动触发13DD：尼安德特人

尼安德特人的大脑容量甚至大于智人。工具制作复杂精细。存在若干symbolizing和mortuary行为的证据：Shanidar洞穴的intentional burial（Pomeroy et al. 2020），Bruniquel Cave的深洞石笋构筑物（约176.5 ka），以及伊比利亚半岛部分洞穴艺术的U-Th断代可能指向尼安德特人。这意味着他们的12DD穿透深度极深，且可能在相变的萌芽区甚至更远。

但目前尚无足够证据支持尼安德特人拥有稳定的、可自洽闭合的符号-仪式系统——没有确认的叙事性场景艺术（最早的叙事性洞穴艺术来自智人，苏拉威西岛，≥51.2 ka，Aubert et al. 2024, Nature），没有独立于智人影响的大规模仪式空间。尼安德特人在四阶段中的确切位置目前无法精确判定，但closure证据的不足暗示他们可能尚未越过翻转点，或虽接近翻转但未达到确立。

这提供了一个重要的结构性启示：12DD穿透深度再深，也不自动触发13DD的涌现。Le Chatelier缓冲可以在高认知水平上长期维持。这也可能解释了种群竞争的结果：约束五（前置篇，DOI: 10.5281/zenodo.19503158）的逻辑意味着，一旦智人的13DD完备，其扩散速度远快于尼安德特人独立完成相变的速度。

三、定义

定义1：构层（11DD+12DD）

11DD（感知）加12DD（认知）构成生物认知基底。包括但不限于：感觉整合，模式识别，因果推理，空间导航，时间记忆，社会信号识别，抑制控制。构层是一个连续的积累过程，不存在内部相变。

定义2：13DD（涌现层）

13DD是个体主体性维度。13DD的涌现不是一个sharp boundary，而是一个有宽度的相变窗口：萌芽是涌现的开始（self认出self），确立是涌现的完成（符号-仪式系统闭合）。13DD的完备（self-completion）意味着self强到能够（1）把自身的消亡当作问题来面对，（2）产生结构性的回应。大猿处于13DD的萌芽态——已经涌现但未完备。

定义3：语言（13DD+14DD）

语言 = 13DD + 14DD。语言不是构层（11DD+12DD）的组件，而是13DD（self）和14DD（制度化符号规则）共同涌现的产物。交流（信号、叫声、信息素）属于11DD+12DD，任何有感知和认知的生物都有交流。但语言需要"我"在说（13DD）加上语法和句法作为社会契约（14DD）。

关键澄清：13DD+14DD的两条腿不同时涌现。13DD的萌芽态足以支撑基础的语言化self表达——2岁儿童说"no"只需要13DD萌芽态（self在说），不需要完整的14DD（语法契约尚未涌现）。这就是为什么terrible twos的语言如此简单：两三个词的组合，没有复杂句法。完整的语言系统（递归结构、语法契约）要等14DD涌现后才完成。

注：语言（13DD+14DD）在SAE框架中有两种投影。在涌现（emergence）方向上投影为语言（13DD与14DD之间的符号互动协议），在基底（base）方向上投影为法律（13DD与14DD之间的否定压制防御）。两者都是处理个体（13DD）与制度（14DD）碰撞的结构，但方向不同。本篇处理涌现投影；法学系列处理base投影。

定义4：桥

桥是连接两个主体或一个主体与环境之间的通道。桥的粗细是12DD预测深度和方向的结果，不是独立变量。狗与人之间的桥粗，是因为狗的12DD在预测人类意图方向上被推得很深；猫与人之间的桥细，是因为猫的12DD预测窗口太短。桥的粗细不改变两端的DD层级。

定义5：伪13DD

低层（12DD或更低）占据了13DD位置的状态。两种形式：（a）12DD预测窗口被殖民后的深预测模拟self（狗），（b）12DD预测窗口短导致的不配合被投射为self（猫）。辨别标准：是否能say no——不是拒绝刺激，而是语言化的自我否定。

定义6：四阶段相变结构

沿用SAE Methodology Paper VI的框架，13DD的涌现经历四个阶段：

阶段一：萌芽。 self识别。self认出self，但不能对self做任何语言化操作。

阶段二：谱翻转。 语言化的self表达。say no——self通过语言对环境进行主动否定。Say no的重点不仅在no（否定），更在say（语言化）。语言和self在这个阶段共生涌现：没有self就没有"我"在说，没有语言self就无法自指。谱翻转不是"语言使能了self表达"，而是self表达和语言在同一个相变中一起出现。这是四阶段中经验上最难直接观测的一段：考古线上口语不留痕（结构性盲区），动物线上直接缺席，儿童线上的证据（terrible twos / 否定语言发展）是最强的但仍是间接代理。谱翻转是最强的结构推断，最弱的直接证据。

阶段三：翻转。 Self将自身消亡反身问题化（personal mortality reflexivity）。self认出self会消失，并将此作为不能不面对的问题。怕死是最常见的现象学表现，但formal marker是反身问题化本身，不限于恐惧这一种情感形式。这是最关键的相变点。

阶段四：确立。 符号-仪式系统闭合（myth-ritual closure）。self对死亡的结构性回应。不再是散点式的恐惧反应，而是自洽的符号系统——关于死亡的叙事、仪式、世界观形成闭合结构。宗教是其最普遍的历史形式。

定义7：r>>1（不对称比）

萌芽到翻转的距离远大于翻转到确立的距离。这是SAE Methodology Paper VI的核心预测在本领域的实例化。不对称性的结构性直觉：Le Chatelier缓冲在萌芽期全力运转，构层（11DD+12DD）的积累长期被吸收。一旦缓冲被突破（self强到能够自指其消亡），确立很快，因为缓冲机制本身已经跌破了维持旧pattern的最低要求。

四、三条证据线

4.1 证据线一：动物行为

4.1.1 萌芽：镜子测试

镜子自我识别（MSR）是self识别的标准操作化定义。强证据物种包括大猿（Gallup 1970, Science）和人类婴儿（约18-24个月）。非灵长类的强候选包括瓶鼻海豚（Reiss & Marino 2001, PNAS）和亚洲象（Plotnik et al. 2006, PNAS）。

灰色案例值得特别注意。清洁鱼（Kohda et al.）通过标记测试的报告引发了MSR到底测量什么的根本性争论。喜鹊的MSR证据在Soler et al. 2020的重复实验中未被复现。马的MSR在群体层面有报告但存在方法学争议（Baragli et al. 2021）。蝠鲼和章鱼有初步的contingency checking行为但远未达到标记测试的标准。

对本文论点最有利的证据结构是：即使在通过MSR的物种中，没有发现从自我识别到更高层级self行为的进一步跃迁。清洁鱼在通过MSR后能"回忆体型"（Kobayashi et al. 2024），但这仍然是body-as-object层面的表征，不是符号化的self模型。MSR-pass物种停在萌芽区——这恰好支撑了"缓冲没有被突破"的判断。

4.1.2 谱翻转及以后：缺席

没有任何非人类动物展现出语言化的self表达。大猿不say no（在本文的技术意义上）。动物到不了谱翻转，不是因为缺少一个工具（语言），而是因为13DD从未涌现到足以支撑语言——没有self就没有语言，没有语言self也无法自指。

动物那条证据线永远卡在萌芽区。这不是证据的缺失，而是结构性的缺席——Le Chatelier缓冲从未被突破的活证据。

4.1.3 动物的死亡反应：不是死亡意识

动物thanatology的文献需要仔细区分。黑猩猩母亲carrying死婴（Lonsdorf et al. 2020）的行为证据表明她们可能识别到死亡已经发生，但这是识别他者的死，不是personal mortality。乌鸦对死去同伴的反应（Swift et al. 2020）的最佳功能解释是信息收集和危险学习，不是死亡概念。大象对死者的复杂社会反应有专门的综述文献（Goldenberg & Wittemyer 2020），但从行为到概念的推理仍然高度争议。

Monsó的最小死亡理解标准要求证据表明动物（1）先预期个体是活的，然后（2）理解non-functionality和irreversibility。跨物种看，当前证据支持的是：许多物种对死者有系统性的持续性反应，部分行为与局部的死亡识别一致，但关于人类式的死亡概念——特别是personal mortality——的强主张仍然缺乏证据。

4.2 证据线二：考古/古人类学

4.2.1 萌芽：洞穴手印

洞穴手印是考古版的镜子测试——"这是我的手"。手印画的不是猎物，不是环境，而是自己身体的痕迹。最早的洞穴手印（苏拉威西岛，约40 ka以上）与其他洞穴艺术的时间关系需要更精确的断代数据，但手印作为自我指涉的物质痕迹，其符号功能不同于动物场景描绘。

4.2.2 谱翻转：教学痕迹（结构性盲区）

考古线在谱翻转处有一个结构性盲区：口语无法留下物质痕迹。语言作为13DD+14DD的共生涌现产物，在考古记录中是invisible的。它几乎肯定发生在随葬品之前，但留不下直接证据。

间接证据来自工具标准化。Morgan et al. 2015的实验表明，Oldowan工具可以通过模仿传播，但Acheulean手斧的高保真传播需要教学，特别是语言化的教学。Lombao et al. 2017（Scientific Reports）进一步报告语言教学在效率和信息保留上优于纯手势教学。

但必须承认这个推理的局限。Ferar 2026的"puppet method"实验表明，knapping-naive的人可以在没有文化传播的情况下制作出手斧形状，这意味着Acheulean标准化不能严格推出语言的存在。Morgan和Lombao的结论是关于传播的效率和保真度，不是严格的必要条件。

本文采取的立场是：Oldowan到Acheulean的转换（约2.6 Ma到约1.76 Ma）中工具标准化程度的跳跃，与教学和原始语言作为前提条件的假说一致（consistent），但不构成严格论证（entailment）。考古线在谱翻转处的证据力度弱于其他阶段——这个诚实的承认本身就是一个发现。

4.2.3 翻转：随葬品

丧葬本身不是personal mortality reflexivity的证据。丧葬可以是卫生行为——处理腐烂的尸体。反身问题化的证据是随葬品和符号化的丧葬实践：朝向的选择，赭石的使用，与死者一同放置的没有实用功能的物品。这些在没有self把死亡当作问题来面对的情况下不会出现。

尼安德特人丧葬的争议恰好落在这个区分上。Shanidar洞穴的"花葬"（Solecki原始解释）在近年被显著削弱：Hunt et al. 2023的重新评估认为花粉团块的最可能来源是筑巢的独居蜜蜂，而不是有意放置的花束。但Pomeroy et al. 2020的发现支持Shanidar存在deliberate burial。La Ferrassie等遗址也有支持intentional burial的证据。

关键问题不是尼安德特人是否掩埋死者（很可能是），而是掩埋是否伴随符号化行为。目前的证据中心倾向于：有一些intentional body placement，但支持ritual offerings的证据薄弱。这与本文的框架一致——尼安德特人可能在翻转点附近，但没有明确越过。

对于智人，符号性丧葬实践的证据更强。赭石使用和珠饰出现在约100 ka前后，通常被用作符号系统涌现的锚点。

4.2.4 确立：符号系统闭合

确立的标志不是单个符号物品，而是符号系统的自洽闭合——从散点式的符号使用到完整的世界观。

最强的早期仪式空间证据来自Manot Cave（Barzilai et al. 2024, PNAS）：在早期旧石器时代晚期（约37-35 ka BP），一个与生活区分隔的仪式空间，有刻有几何符号的巨石（可能是龟的表征），有用于照明的木灰，有适合集体聚集的声学特性。多条独立证据线汇聚在"非日常的集体实践"上。

更早的候选包括Bruniquel Cave的尼安德特人石笋构建（约176.5 ka），有时被解释为深洞集体活动的证据。但从仪式行为到"符号系统闭合"的推理需要更多独立代理支撑。

苏拉威西的叙事性洞穴艺术（≥51.2 ka，Aubert et al. 2024, Nature）展现了场景化的描绘和可能的兽人（therianthrope）形象，暗示了叙事性的想象能力。这是从符号物品到符号叙事的跳跃，但仍然不等于完整的符号-仪式系统闭合（后者在考古记录中几乎无法直接恢复）。

考古线在确立阶段的主要局限是：我们能看到仪式空间和叙事艺术，但不能直接看到神话文本。从"仪式行为"到"世界观闭合"的推理总是间接的。

4.3 证据线三：儿童发展心理学

4.3.1 萌芽：约18个月

镜子自我识别在标准发展范式中约在18-24个月出现。跨文化研究表明存在onset时间的变异（与接触镜子的经验和社会实践有关），但这是时间变异，不是能力缺席。

一项直接相关的神经科学发现：Bulgarelli et al. 2019（Developmental Cognitive Neuroscience）报告，18个月时通过镜子测试的婴儿（"Recognizers"）比未通过的婴儿有更强的fronto-temporoparietal静息态连接。这个连接模式与成人DMN中自我反思的区域高度重叠。

4.3.2 谱翻转：约2岁

"Terrible twos"在发展心理学中通常用自主性（autonomy）和自我调节（self-regulation）来解释。Hughes et al. 2020（Developmental Science）将toddler的不合作行为与早期执行功能和亲子互动质量联系起来。

语言习得文献中的否定（negation）研究直接支撑"say no"的操作化：toddler最早产生的"no"是拒绝性的（reject offers/commands），之后才发展出更接近成人的否认用法（denial），通常在2-2.5岁区间。跨语言研究（2025）显示否定功能的分布因语言环境而异，但基本时间线跨文化一致。

操作化区分至关重要：12DD的拒绝（stimulus rejection）和13DD的否定（self-assertion）不是一回事。动物拒绝不喜欢的食物是12DD的最优预测反应。13DD的"no"不仅拒绝客体，更是拒绝他者的框架——toddler经常在自己想要那块糖的时候，纯粹为了验证自己的主体性边界而说"no"。这种违背12DD生物学利益、纯粹为了彰显self边界而做出的否定，才是谱翻转的真正标志。

Say no的重点不仅在no，更在say。2岁儿童说"no"只需要13DD的萌芽态——self在说，但完整的语法契约（14DD）尚未涌现。这就是为什么terrible twos的语言如此简单。动物到不了谱翻转，不是因为缺一个工具，是因为13DD从未涌现到足以支撑语言化的self表达。

神经发育数据：Fiske et al. 2024（Imaging Neuroscience）用纵向fNIRS追踪10到16个月的抑制控制发育，发现从右侧偏侧化的PFC/顶叶招募转向更广泛的双侧招募，包括右侧下额叶回和双侧背外侧/眶额PFC。研究者将此解释为可能的reorganization period——支持"相变"修辞多于"平滑成熟"叙事。

4.3.3 翻转：约4-5岁

儿童死亡认知的现代文献将"理解死亡"分解为子组分：普遍性（universality），不可逆性（irreversibility/finality），功能停止（cessation/nonfunctionality），和因果性（causality）。

Menendez et al. 2020的综述报告：5岁前儿童开始发展普遍性和不可逆性理解；5岁左右大多数理解身体功能停止；6岁左右理解死亡可以有多种原因。2026的叙事综述确认12岁前儿童理解普遍性，不可逆性和生物停止。

翻转点约在4-5岁：儿童开始问"人会不会死"，"我会不会死"，"死了以后去哪里"。这标志着self已经强到能把自身的消失当作问题。

值得注意的是，这个时间点与theory of mind（心智理论）的经典里程碑——通过false belief task（约4岁）——高度重叠。TPJ和mPFC在成人的信念推理和视角转换中被反复证实参与（neuroimaging meta-analyses），而这些区域的功能专门化在约4-5岁显著增强。一个可辩护的立场是：ToM网络的巩固是显式死亡推理的必要（但非充分）脚手架。

4.3.4 确立：几乎即时

Richert & Corriveau 2022（Annual Review of Developmental Psychology）的综述报告：到约5岁，儿童倾向于声称不可观察的自然和超自然现象存在，其信心追踪文化共识和家庭/社区的认可。Menendez et al. 2020将此直接应用于死亡：儿童通过对话，媒体和仪式参与来建构死亡理解，生物学解释和宗教解释常常共存而非被视为矛盾。

跨文化实验（瓦努阿图 vs 美国）显示，简短的叙事提示（自然 vs 超自然线索）就能调节不同年龄段对来世的认同模式。这与"在testimony下快速接受文化脚手架框架"一致。

对"即时"措辞的限定：文献支持早期和高效的testimony-based框架接受，但"几乎即时"取决于你把什么算作"接受"（显式同意，行为遵从，情感安慰，还是稳定信念）。最可辩护的主张是：幼儿能够在早期协调多个解释系统——包括关于死后发生什么的宗教说明——而不需要先完全掌握生物学死亡知识。

4.4 文本验证：吉尔伽美什

三条证据线之外，人类现存最古老的长篇叙事直接就是13DD相变的叙事化。

《吉尔伽美什史诗》（约公元前2100年，苏美尔）的核心结构是：吉尔伽美什的挚友恩奇都死了。恩奇都的死不是一个情节事件，而是翻转点——吉尔伽美什从此意识到自己也会死。"我的朋友，我所爱的，已经变成了泥土。我难道不也会像他一样躺下，永远不再起来吗？"这是self把自身的消亡当作问题的原型表达。

此后吉尔伽美什跨越世界去寻找永生（试图结构性回应），最终失败——永生草被蛇吞食。他回到乌鲁克城，看着城墙，接受了死亡。

人类最早写下的长故事，主题就是self对自身消亡的反身问题化。这不是巧合。13DD完备后，第一件需要被叙事化的事情就是翻转点本身。吉尔伽美什史诗是13DD相变的文学化石。其他跨文明的类似叙事见附录A。

五、r>>1的不对称性

5.1 儿童线

从萌芽（约18个月）到翻转（约54个月，即4.5岁）：约36个月。

从翻转到确立：几乎同步。

r ≈ 36 / ε，其中ε接近零（确立距离以月计几乎不可测）。r>>1在儿童线上最干净。

5.2 演化线

从萌芽（大猿级别的MSR，至少几百万年前）到翻转（符号化丧葬的出现，约100 ka量级）：几百万年。

从翻转到确立（仪式空间和叙事艺术，约50-35 ka）：几万年。

不对称比r粗略估计在100左右（几百万年 vs 几万年）。这个估计非常粗糙，依赖于考古断代的不确定性和各阶段标志的操作化定义，但r>>1在数量级上是稳健的。

5.3 动物线

动物线提供的是r = ∞的极端案例：卡在萌芽区几百万年，翻转从未发生。Le Chatelier缓冲从未被突破。

5.4 考古线在翻转-确立区间的分辨率问题

需要承认的一个限制是：考古线上，"死亡意识"的信号（burial，repeated body placement，grave goods）和"符号系统闭合"的信号（仪式空间，复杂叙事，制度化神话）在时间上不总是紧密聚集。根据你选择的代理指标和证据标准的保守程度，两者之间可能相隔数万年。这使得考古线上r>>1的论证不如儿童线那样干净。

但这正是三条证据线交叉验证的价值所在：儿童线上r>>1的证据最干净（分辨率高，变量可控），考古线提供演化尺度的验证（分辨率低但时间跨度大），动物线提供缓冲从未被突破的极端对照。三条线各有局限，但它们各自的局限不重叠。

六、余项率与物种数量级

Le Chatelier缓冲的强度可以用另一种方式量化：地球历史上产生了多少物种，其中多少到达了各个DD层级？

6.1 数据

地球历史上（Phanerozoic，约5.4亿年以来）总共出现过的多细胞真核生物物种数，标准估计基于现存物种数（约870万，Mora et al. 2011, PLoS Biology）和灭绝率（99%至99.9%，Jablonski标准估计）。取上限：870万 ÷ 0.001 ≈ 87亿个物种。

这些物种绝大多数停留在9DD+10DD（多细胞生物的基本组织层级）。具有12DD认知能力的物种——有中枢神经系统、能进行学习和预测的——大约在哺乳类、鸟类、头足类这个范围，历史上总计约数万个物种量级。

到达13DD（self完备）的物种：1个。智人。

6.2 余项率的两步结构

SAE物理系列收束篇（DOI: 10.5281/zenodo.19464378）和前置篇（DOI: 10.5281/zenodo.19503158）使用的余项率为每步约10^{-5}。从9DD+10DD到13DD需要两步跃迁：

9DD+10DD（多细胞生物）→ 11DD+12DD（认知物种）：87亿 × 10^{-5} ≈ 8.7万

11DD+12DD（认知物种）→ 13DD（self完备）：8.7万 × 10^{-5} ≈ 0.87

两步余项率，每步10^{-5}，总余项率10^{-10}。从87亿个多细胞物种到0.87个13DD物种。

6.3 解释

0.87 ≈ 1。在数量级上，地球在其Phanerozoic历史（5.4亿年）中恰好处于产生约一个13DD物种的量级。这不应解读为"差一点就不会产生"——地球还有约10亿年的宜居窗口，如果智人未在此时涌现，后续物种迟早会越过翻转点。0.87的正确含义是：到Phanerozoic的当前阶段，余项率预测恰好约一个13DD物种涌现。

关键在于约束五（前置篇）：一旦第一个13DD物种涌现，其扩散速度远快于第二个13DD物种独立涌现的速度。槽位被永久占据。不会有第二个。这与孤独恒星定理在行星层面的推论一致：不多不少，就一个。

余项率10^{-5}的具体值尚未在SAE框架内严格证明（物理系列收束篇提供了推导路径但未完结），本文将其作为系列已有数据使用。如果未来余项率的精确值有修正，本节的数量级论证需要相应更新，但两步跃迁的结构不变。

七、神经科学的位置

本文不主张任何单一神经结构"等于"self。self是涌现层，不能还原为构层。但神经科学能提供的是相变前后构层状态的测量窗口。

6.1 DMN的跨物种比较作为构层积累的指标

Default Mode Network（DMN）在人类中与自我参照思维和时间投射（回忆过去/想象未来）高度耦合。DMN的核心hub包括medial prefrontal cortex（mPFC），posterior cingulate cortex/precuneus（PCC/Prec），和angular gyrus。

关键的跨物种比较数据来自Garin et al. 2022（Cell Reports）：在非人类灵长类中，DMN同源区域分裂为两个网络，且mPFC与PCC之间的连接薄弱，而这种强连接恰好是人类DMN的核心特征。黑猩猩的DMN等价性更接近人类（Barks et al. 2015, Cerebral Cortex），有PET证据显示社会判别任务中皮质中线区域的去激活，与人类DMN的任务负效应一致。

对本文论点的意义：如果self完备需要一个紧密整合的自我模型——将价值/身份（mPFC）与自传性场景建构/语境（PCC/Prec和内侧颞叶贡献）联接——那么连接架构（connectivity architecture）而非单个区域的有无，就是构层积累是否达到相变条件的可能指标。非人类灵长类有DMN的组件，但连接不够——构层在积累，还没到翻转点。

6.2 发育时间线与四阶段的对齐

Bulgarelli et al. 2019：18个月的Recognizers有更强的fronto-temporoparietal连接——对应萌芽。

Fiske et al. 2024：10-16个月的PFC抑制控制reorganization——可能对应谱翻转的神经准备。

TPJ/mPFC在约4岁的false belief task相关功能专门化——对应翻转的脚手架。

这些节点与本文的四阶段不是一对一映射（一个神经转变可以同时支撑多种行为变化），但它们支撑的是"阶段性跳跃"而非"平滑成熟"的总体图景。

6.3 死亡意识的神经标记

Hirano et al. 2021（Cerebral Cortex Communications）：人类思考自身死亡时，PCC（DMN核心hub）在self条件下特异性激活，且恐惧-死亡（fear-of-death）调节了右缘上回的反应（负线性关系）和PCC的反应（倒U型关系）。PCC的参与被解释为将死亡相关思维作为self-relevant future agenda来处理——直接将死亡认知连接到DMN的核心功能。

2026年Neuropsychologia的系统综述发现：与一般不愉快刺激相比，死亡相关刺激更倾向于降低insula活动。这暗示死亡相关加工不同于一般威胁加工，可能涉及self-referential processing的独特模式而非恐惧回路的放大。

这些数据支撑的主张是认识论性的：personal mortality reflexivity不是更强的恐惧，是self-referential processing的一种特殊形态。

八、与已有理论的关系

7.1 Terror Management Theory（TMT）

TMT（Greenberg, Solomon, & Pyszczynski）提出客观的自我意识使死亡变得突出，文化世界观（包括宗教）缓冲存在焦虑。这与本文的"mortality reflexivity → myth-ritual closure"链最为接近。

但TMT没有四阶段结构，没有r>>1的定量预测，没有三条证据线的交叉验证架构。更重要的是，TMT把死亡焦虑→文化世界观当作一个单步因果关系来处理，而本文的框架把它拆解为从萌芽到确立的多步相变，每步有不同的标志和证据标准。

TMT的实证基础存在争议：大规模重复实验挑战了一些经典的mortality salience效应的稳健性，meta-analytic的争论指出了可能的publication bias和异质性。本文对TMT的使用限于：TMT的理论构架支撑"死亡意识→文化回应"的总体方向，但不依赖其特定实验范式。

7.2 Varki的MORT理论

Ajit Varki的Mind Over Reality Transition（MORT）理论是本文最近的竞争对手。MORT把mortality salience的涌现当作人类独特性的关键进化事件，并把它与文化后果联系起来。

本文的增量贡献在于：（1）四阶段结构而非单步跨越，（2）r>>1的定量化不对称预测，（3）三条独立证据线的交叉验证而非单一进化叙事，（4）与SAE框架的整合提供了跨领域的理论统一性（同一个相变结构适用于代谢肿瘤学，实验设计方法学，经济学等）。

7.3 MSR批评文献

现代MSR文献越来越谨慎地对待MSR作为统一"自我意识"指标的地位。清洁鱼争议，喜鹊重复失败，训练猕猴通过MSR的研究——这些都在瓦解MSR作为clean species-level cognitive boundary的可靠性。

这对本文实际上是有利的。如果MSR只是"self-as-body-object"的狭窄能力，那么MSR可以在进化上长期持续而不产生mortality-reflexive或symbolically closed的self，就更容易论证了。本文把MSR放在萌芽位置而非翻转位置，恰好与MSR文献的自我批评一致。

九、非平凡预测

预测1：动物永远到不了谱翻转

在13DD未涌现的情况下，没有任何非人类物种能展现语言化的self否定（say no的技术意义）。如果未来发现某个物种能做到这一点，该物种必须同时拥有（a）通过MSR的能力和（b）某种形式的符号化否定表达。

否证条件： 发现一个不通过MSR但能say no的物种，或者发现一个通过MSR并能say no但不发展出死亡意识的物种。前者否证萌芽→谱翻转的顺序，后者否证谱翻转→翻转的顺序。

预测2：考古线的谱翻转证据将始终是间接的

口语在考古记录中是不可见的。未来的考古发现可能提供更多教学痕迹的间接证据（如工具制作的标准化程度的突变），但不可能提供口语本身的直接证据。

否证条件： 发现了直接的口语物质痕迹（这在物理上几乎不可能，但逻辑上需要说明）。

预测3：r>>1在儿童发育中可被量化验证

萌芽到翻转（约18个月到约4-5岁）的时间距离应该系统性地大于翻转到确立（约4-5岁到符号-仪式框架接受）的时间距离。这可以在纵向发育研究中直接量化。

否证条件： 纵向数据显示翻转到确立的时间距离与萌芽到翻转的时间距离相当或更大（即r≈1或r<1）。

预测4：尼安德特人的考古记录应该缺乏符号-仪式系统闭合

尼安德特人在四阶段中的确切位置目前无法精确判定，但如果本文框架正确，他们的考古记录中应该存在symbolizing和mortuary行为的证据（与萌芽区和部分翻转一致），但缺乏稳定的、可自洽闭合的符号-仪式系统——即（a）独立于智人影响的大规模仪式空间，（b）叙事性场景艺术，（c）可恢复的myth-ritual closure结构。

否证条件： 发现尼安德特人独立制作的叙事性场景艺术，或独立于智人影响且具有多条独立证据线汇聚的大规模仪式空间。

预测5：12DD穿透深度与13DD涌现之间没有连续关系

跨物种比较应该显示：认知能力（12DD穿透深度）的增加不线性导向self完备（13DD涌现）。会有一些物种12DD很深但完全没有13DD迹象（如尼安德特人，如果未来证据支持这个判断），而智人的13DD涌现在认知能力达到某个阈值后似乎"突然"出现。

否证条件： 跨物种数据显示认知能力指标与self完备指标之间存在平滑的线性关系，没有阈值效应。

十、开放问题

• 谱翻转的考古代理指标是什么？教学痕迹（工具标准化）是目前最好的候选，但Ferar 2026的实验削弱了"标准化=语言"的严格推理。是否存在更好的代理？

• 尼安德特人到底在哪个阶段？当前证据不足以精确定位。如果未来发现尼安德特人独立于智人影响的符号系统证据，本文的框架需要修正尼安德特人的阶段判断，但不需要修正四阶段结构本身。

• r的跨证据线比较如何量化？儿童线上r可以用月数直接计算。演化线上r需要考古断代和代理指标选择。动物线上r = ∞。三条线的r值能否统一到一个无量纲的比较框架中？

• 符号-仪式系统的多样性。本文把myth-ritual closure作为确立的formal marker，但不同文明对死亡的结构性回应差异极大。这些差异是确立阶段内部的变异，还是指向不同的确立路径？

• self完备之后，然后呢？ 13DD完备的主体们如何形成制度？13DD的个体关系受Dunbar数（约150人）约束——超过这个规模，self-to-self的直接关系无法维持，必须涌现14DD（制度）来组织更大群体。一个值得注意的物质签名是：150人以内，口语（13DD萌芽态即可支撑）足以维持所有社会信息——谁欠谁的、谁和谁有仇、哪里有猎物——不需要把信息外化到任何介质上。文字的出现（苏美尔，约3200 BCE，群体已达数千至数万人）不是技术进步，是14DD涌现的必然产物：群体超过面对面管理范围，契约、法律、账目必须外化到个体记忆之外。口语→文字可能就是13DD→14DD的物质签名。从band到tribe到chiefdom到state（Service序列）的政治组织演化，在SAE框架中对应的是14DD从13DD构层的涌现。这个涌现是否也有四阶段相变结构？r>>1是否同样成立？这是本系列Paper 2的主题。

十一、结论

人之所以为人，不在群居（蜜蜂更群居），不在工具（乌鸦也用工具），不在协作（狼的协作更高效），不在12DD预测深度（狗预测人类意图比任何大猿都准但没有self），甚至不在认知能力（尼安德特人的脑容量比智人大）。

人之所以为人，在于self的完备：self强到能够将自身的消亡反身问题化，并产生符号-仪式系统的结构性回应。这是13DD从11DD+12DD构层的涌现，经历四个阶段（萌芽，谱翻转，翻转，确立），r>>1。

三条独立证据线——动物行为，考古/古人类学，儿童发展心理学——交叉验证了同一个相变结构。动物线卡在萌芽区，考古线在谱翻转处有结构性盲区（口语不留痕），儿童线提供最完整的可观测序列。三条线各有局限，但局限不重叠。人类最古老的文学文本（吉尔伽美什史诗）直接叙事化了翻转点本身，提供了来自文本传统的独立旁证。

本文的核心贡献是：将镜子测试，personal mortality reflexivity和符号-仪式系统闭合放在同一个相变框架内，这在现有文献中没有先例。已有的理论连接都是两两之间的（MSR↔self，death cognition↔thanatology，mortality↔religion via TMT）。四阶段结构和r>>1的定量预测是新的。

---

与SAE框架的关系：本文是SAE Anthropology Series的第一篇，处理13DD（个体主体性）从11DD+12DD（认知基底）的涌现。系列前置篇（DOI: 10.5281/zenodo.19503158）提供了宇宙尺度的背景结构：孤独恒星定理和约束五的物种竞争推广。系列第二篇将处理14DD（制度）从13DD的涌现——即self完备的主体们如何形成政治和社会组织。构层-涌现层的关系，Le Chatelier缓冲，r>>1的不对称比，桥的独立性——这些都是SAE Methodology Paper VI中形式化的概念在本领域的实例化。相变结构的跨领域一致性（代谢肿瘤学，实验设计，经济学，现在是人类学）为SAE框架的普遍性提供了累积证据。

---

附录A：跨文明的13DD叙事

吉尔伽美什之外，至少三个独立文明传统保存了13DD相变的叙事结构。

创世记（希伯来传统）。 亚当和夏娃吃了知善恶树的果子——"你们便如神知道善恶"。知善恶就是self完备：self的判断力从被动接受环境跃迁到主动评估。惩罚是什么？"你必定死。"叙事结构精确地编码了本文的论点：self完备的代价就是知道自己会死。伊甸园是萌芽区，吃果子是翻转，被逐出是确立后的不可逆。

佛陀出城（印度传统）。 悉达多在宫墙内不知老病死（萌芽区——self存在但被屏蔽，Le Chatelier缓冲由父王的保护提供）。出城看到老人、病人、死人（翻转——self认出self会消失）。出家寻道（确立——结构性回应）。叙事甚至保留了r>>1的不对称结构：在宫中二十九年（萌芽期），出城到出家几乎即时。

阿喀琉斯在冥府（希腊传统）。 荷马《奥德赛》第十一卷：奥德修斯在冥府见到阿喀琉斯的亡魂。阿喀琉斯说："我宁愿活着做奴隶，也不愿做死人的王。"这是13DD的极致表达——self完备到极点，连英雄荣誉（proto-14DD的制度化价值）都换不了对死亡的接受。在所有希腊英雄中，阿喀琉斯是唯一事先知道自己会死还选择出征的——他的选择本身就是self完备后的结构性回应。

---

参考文献

（按首次出现顺序）

Morgan, T. J. H. et al. (2015). Experimental evidence for the co-evolution of hominin tool-making teaching and language. Nature Communications, 6, 6029.

Qin, H. (2024). SAE Foundation Paper 1. DOI: 10.5281/zenodo.18528813.

Qin, H. (2026). SAE Methodology Paper VI: Phase Transition Window and Experimental Design. DOI: 10.5281/zenodo.19464506.

Qin, H. (2026). SAE Biology Note 3. DOI: 10.5281/zenodo.19501120.

Qin, H. (2026). SAE Learning Series Paper 4. DOI: 10.5281/zenodo.19491926.

Qin, H. (2026). SAE Anthropology Series Prequel: The Cosmic Background of Anthropology. DOI: 10.5281/zenodo.19503158.

Gallup, G. G. Jr. (1970). Chimpanzees: Self-recognition. Science, 167(3914), 86-87.

Reiss, D. & Marino, L. (2001). Mirror self-recognition in the bottlenose dolphin. PNAS, 98(10), 5937-5942.

Plotnik, J. M. et al. (2006). Self-recognition in an Asian elephant. PNAS, 103(45), 17053-17057.

Kohda, M. et al. (2019). If a fish can pass the mark test, what are the implications for consciousness and self-awareness testing in animals? PLoS Biology, 17(2), e3000021. doi: 10.1371/journal.pbio.3000021. Update: Kohda, M. et al. (2022). PLoS Biology, 20(2), e3001529. doi: 10.1371/journal.pbio.3001529.

Soler, M. et al. (2020). Replication of the mirror mark test experiment in the magpie (Pica pica) does not provide evidence of self-recognition. Journal of Comparative Psychology, 134(4), 363-371. doi: 10.1037/com0000223.

Baragli, P. et al. (2021). If horses had toes: demonstrating mirror self recognition at group level in Equus caballus. Animal Cognition, 24(5), 1099-1108. doi: 10.1007/s10071-021-01502-7.

Kobayashi, T. et al. (2024). Cleaner fish body-size recall after MSR.

Lonsdorf, E. V. et al. (2020). Chimpanzee mothers carrying dead infants.

Swift, K. N. et al. (2020). Crow neuroimaging study: responses to dead conspecifics.

Goldenberg, S. & Wittemyer, G. (2020). Elephant thanatology review.

Monsó, S. (2021). How to tell if animals can understand death. Erkenntnis, 86, 1411-1431. doi: 10.1007/s10670-019-00175-4.

Hunt, C. O. et al. (2023). Shanidar pollen reassessment.

Pomeroy, E. et al. (2020). New Neanderthal remains at Shanidar.

Aubert, M. et al. (2024). Sulawesi narrative cave art. Nature.

Barzilai, O. et al. (2024). Manot Cave ritual compound. PNAS.

Lombao, D. et al. (2017). Verbal vs gestural teaching in knapping. Scientific Reports.

Ferar, N. (2026). Puppet method: handaxe shaping without cultural transmission.

Bulgarelli, C. et al. (2019). fNIRS connectivity and mirror self-recognition at 18 months. Developmental Cognitive Neuroscience.

Hughes, C. et al. (2020). Toddler noncompliance and executive function. Developmental Science.

Fiske, A. et al. (2024). Longitudinal fNIRS of infant inhibitory control. Imaging Neuroscience.

Menendez, D. et al. (2020). Children's death understanding via subcomponents.

Richert, R. A. & Corriveau, K. H. (2022). Religious cognition in children. Annual Review of Developmental Psychology.

Garin, C. M. et al. (2022). Cross-species DMN comparison. Cell Reports.

Barks, S. K. et al. (2015). Chimpanzee DMN via PET. Cerebral Cortex.

Hirano, K. et al. (2021). Neural correlates of self-relevant death processing. Cerebral Cortex Communications.

Varki, A. (2009). Human uniqueness and the denial of death. Nature, 460, 684. doi: 10.1038/460684c. See also: Varki, A. & Brower, D. (2013). Denial: Self-Deception, False Beliefs, and the Origins of the Human Mind. Twelve Books.

Greenberg, J., Solomon, S., & Pyszczynski, T. (1997). Terror management theory of self-esteem and cultural worldviews. Advances in Experimental Social Psychology, 29, 61-139. doi: 10.1016/S0065-2601(08)60016-7.

Mora, C. et al. (2011). How many species are there on Earth and in the ocean? PLoS Biology, 9(8), e1001127. doi: 10.1371/journal.pbio.1001127.