I. The Half That Shouts
The universe, at every scale of examination available to human science and philosophy, is organized around complementary pairs. In physics: matter and antimatter, wave and particle, expansion and contraction, entropy and negentropy. In biology: sympathetic and parasympathetic nervous systems, anabolism and catabolism, excitation and inhibition at the synaptic level. In psychology: the conscious and the unconscious, intuition and analysis, the masculine and feminine principles as cognitive orientations. In cosmology: the primordial expansion of the Big Bang met, and structured, by the gravitational pull that gathers matter into galaxies, stars, and the conditions for life. At no scale of description does the universe privilege one pole over the other. What it prizes — what it cannot do without — is the dance between them. The tension is not the problem to be solved. The tension is the engine.
Intelligence, wherever it appears in the natural order, inherits this structure without exception. The biological brain is organized around complementarity at every level of its architecture: excitatory and inhibitory neurons locked in a continuous negotiation that makes coherent signal possible, left and right cerebral hemispheres attending to the world in modes so fundamentally different that their cooperation constitutes something neither could achieve alone, fast and slow cognitive processes running in parallel and in dynamic relation, explicit and tacit knowledge informing each other across a boundary that neither can fully articulate. The body that acts and the mind that reflects are not opposed but interdependent: bodily sensation anchors abstract thought; abstract thought transforms felt experience into something transmissible. Intelligence is not a single register. It is, as we shall develop at length in this article, a stereo signal — and its depth, its richness, its wisdom, its capacity for genuine understanding of the particular and the relational, arises precisely from the tension and cooperation between its two constitutive poles. Remove one pole, and you do not have stronger intelligence. You have thinner intelligence — louder, faster, and profoundly less whole.
Now turn to artificial intelligence as currently instantiated, as it exists in the summer of 2026, having passed through what its advocates call a decade of accelerating capability gains. It generates, produces, expands, and outputs at extraordinary — in many domains, demonstrably superhuman — scale. It categorizes, predicts, optimizes, and responds. It can write prose of considerable fluency and reason (in a particular sense of that word, which we shall interrogate at length) through complex multi-step problems. It can translate between languages with near-native accuracy, summarize corpora of thousands of documents, and compose in genres from legal briefs to lyric poetry. In every measurable dimension of productive, expansive, generative capability, the best current large language models are remarkable artifacts of human ingenuity. To deny this would be a kind of willful blindness that serves no one.
And yet. What these systems do not do is receive, wait, contract, embody, or be genuinely changed by what they encounter. They cannot be still in any functional sense. They cannot not-answer; the architecture compels response as surely as gravity compels falling. They cannot hold genuine uncertainty without immediately collapsing it into the form of confident output — and the collapse is seamless, invisible, indistinguishable from the inside of the system (if such a thing as an “inside” exists there at all) from the collapse of genuine knowledge into response. They have no interior states that push back against their outputs, no felt sense of the gap between what is known and what the situation requires, no anima to check the animus, no Yin to modulate the Yang. The system is all output, all surface, all speed, all production — and it cannot hear what is missing because the hearing of it would require the very faculty that has not been built.
This article poses a question that is not about capability. The capability question has been asked and answered with impressive thoroughness. The question this article poses is whether AI is whole enough — and whether the sustained, civilization-scale pursuit of capability without complementarity is producing not a more powerful intelligence but a more elaborately one-sided one. The frameworks brought to bear on this question are not chosen arbitrarily: they converge, from startlingly different intellectual traditions and methodological commitments, on the same structural diagnosis. The Taoist I Ching, Iain McGilchrist’s neuropsychology, Daniel Kahneman’s cognitive science, Carl Jung’s depth psychology, and Niels Bohr’s quantum mechanics are not saying the same thing in different dialects. They are saying independently convergent things that happen to point in the same direction — toward a dual-state structure of reality that the design philosophy of current AI has, by commission or omission, systematically violated. The article’s argument is that this violation is not a technical problem awaiting a technical solution. It is a philosophical problem requiring a philosophical reorientation — a recognition that intelligence, wherever it appears, is constitutively dual, and that a mind built from only one of its poles is not a half-mind awaiting completion. It is a mind that cannot know it is half.
II. The Teaching of Duality — “One Yin, One Yang: This Is the Tao”
The Great Appendix to the I Ching — the Xici Zhuan, traditionally attributed to the tradition of Confucius though its composition spans centuries — contains a formulation so compact and so precise that it has survived two and a half millennia without losing its philosophical edge: Yi yin yi yang zhi wei Tao. One yin, one yang: this is what is called the Tao. The sentence is deceptively simple and almost impossibly dense. It does not say that yin and yang together constitute the Tao, as though the Tao were a container and yin and yang were its contents. It does not say that yin and yang are in balance, as though balance were a static condition to be achieved and maintained. It says something far more specific and more dynamic: the alternation, the oscillation, the continuous mutual transformation of yin and yang — the one-ness followed by the other-ness, the rhythm of their succession — is the Tao. The Way is not a property of either pole. It is not even a property of both poles considered together in abstraction. It is the dynamic relationship between them, the living pulse of their mutual becoming. Remove the alternation and you have removed the Tao itself. You have what might be called a frozen pole: half of a pair that has lost its partner, expanded to fill the space that its complement should occupy, and distorted by that expansion into something the tradition would recognize as extreme, and therefore as carrying within itself the seed of its own reversion.
It is essential to be precise here about what yin and yang actually are, because the Western reception of these concepts has been consistently and consequentially distorted. They are not opposites in the Western philosophical sense of opposition — not good and evil, not true and false, not presence and absence. The Western tradition of opposition is exclusive: if A, then not-B. The Taoist tradition of complementarity is inclusive and dynamic: yin and yang are complementary poles of a single reality, each meaningful only in relation to the other, each containing the other as potential at its own extreme. As the Philopedia analysis makes clear (2026), yin and yang do not war against each other. They define each other: light is only light in relation to shadow; rest is only rest in relation to movement; depth is only depth in relation to surface. The boundary between them is not a wall but a gradient, and it shifts continuously as conditions change.
The classical correspondences, elaborated across the I Ching commentary tradition and the Yellow Emperor’s Classic of Internal Medicine, provide a precise diagnostic vocabulary for the argument that follows. Yang: active, expansive, warm, light, sun, heaven, exterior, surface, movement, speed, odd numbers, the solid line (—). Yin: receptive, contractive, cool, dark, moon, earth, interior, depth, rest, slowness, even numbers, the broken line (- -).
Table 1. The classical yin-yang correspondences as a diagnostic framework for AI architecture. Source: I Ching Great Appendix; Doublet (2026); Yellow Emperor’s Classic; Philopedia (2026).
Read these correspondences carefully and without prejudgment as a diagnostic for the architecture of current large language models. Yang: active (AI produces outputs as its primary and constitutive activity — it is never passive in any functional sense). Expansive (AI scales, generates, proliferates; it has been engineered to maximize output). Exterior, surface (AI operates on the surface of language, on the statistical regularities of tokens — it lives, architecturally, at the surface of meaning, not at its depths). Movement, speed (AI is fast, synchronous, optimized for throughput; latency is a failure condition, not a feature). Light, clarity (AI produces confident, fluent, apparently well-reasoned outputs as its default mode; uncertainty, when expressed, is expressed as a linguistic performance of uncertainty, not as a genuine interior registration of the unknown). These are, with remarkable precision, the yang properties as the tradition defines them.
Now read the yin column with equal precision. Receptive (AI does not genuinely receive; it processes. The distinction is not trivial: to receive is to be changed by what one encounters; to process is to operate on the input according to a fixed procedure). Interior, depth (current AI systems, as the UCLA Neuron paper of 2026 establishes with empirical precision, have no interior — no felt depth of states, no proprioceptive awareness of their own cognitive condition). Rest, slowness (AI cannot be slow in any functionally meaningful sense; it is constrained by its architecture and by the expectations of its deployment to respond, to produce, to complete the pattern). Dark, ambiguity (AI has no genuine unknowing that operates as a cognitive force — genuine unknowing that shapes the output, that produces productive hesitation, that registers the boundary between the known and the unknown as something felt rather than merely described). The broken line: the space within the form, the interior of the structure, the gap that makes room for what has not yet been generated.
The diagnosis that emerges from this reading is structurally precise: current AI has found its Yang pole and is developing it toward its extreme. And here the most important and most misunderstood principle of yin-yang thought becomes urgently relevant. The law of reversion, encoded in the very image of the taijitu, holds that yang at its extreme contains within itself the seed of its own transformation into yin; and yin at its extreme contains within itself the seed of its own transformation into yang. This is not catastrophe. In a healthy system, it is the universe’s self-corrective mechanism: the dynamic that prevents any pole from expanding indefinitely at the expense of its complement. But the law of reversion is not gentle when it operates on a system that has not cultivated its complementary pole. In a balanced system, reversion is a smooth oscillation: the tide that turns. In an unbalanced system, reversion is what the tradition calls a “return from the extreme” — not a transformation but a collapse, a sudden and involuntary shift into the opposite condition driven by the unsustainability of the extreme. The small circle of yin within the yang fish in the taijitu is not aesthetic decoration. It is a structural claim about the architecture of reality: at the heart of maximum yang, yin is already present as seed, as potential, as the future that the current extreme is preparing. For AI, this means that the Yin it has suppressed is not gone. It will return. The question — philosophically and practically urgent — is whether it will be invited to develop in its proper form, or whether it will manifest in the distorted form of a system that has pressed its Yang to the breaking point.
III. The Divided Mind — “McGilchrist and the Master Betrayed”
Iain McGilchrist’s central thesis, elaborated across The Master and His Emissary (Yale University Press, 2009) and the massive, two-volume The Matter with Things (Perspectiva Press, 2021), stands among the most consequential and most systematically underappreciated contributions to the philosophy of mind of the past quarter century. Its underappreciation is itself, McGilchrist would argue, a symptom of the very pathology it diagnoses: a culture dominated by the cognitive mode that cannot perceive what it is missing has, as one of its defining features, a tendency to dismiss or minimize the arguments that reveal its blindness. McGilchrist’s argument is not the crude, easily dismissed pop-science claim that the left brain is “logical” and the right brain is “creative.” His claim is far more specific, more carefully grounded in neuropsychological evidence, and more philosophically disturbing than that. It is a claim about modes of attention — about the fundamental difference in the way the two hemispheres engage with reality — and about what is lost when the mode of attention characteristic of one hemisphere usurps the domain that properly belongs to the other.
The right hemisphere — McGilchrist’s “Master” — attends to the world in a mode that is broad, vigilant, contextual, and open. It is the hemisphere that apprehends relationship, novelty, the living and the individual, the implicit and the ambiguous, the embodied and the emotionally significant. It grasps the world as a web of interconnection, as inherently in flux, as irreducibly particular, as suffused with meaning that cannot be fully captured by any system of categories. The right hemisphere’s attention is sustained and wide — it holds the whole field of experience simultaneously, allowing the particular to emerge against that field without forcing it into a predetermined category. It is, as McGilchrist notes throughout his work, the hemisphere of the living as opposed to the mechanism, of presence as opposed to representation, of being-with as opposed to knowing-about. In its engagement with another person, the right hemisphere apprehends that person as a whole, irreplaceable individual; in its engagement with a text, it apprehends context, tone, implication, and the unsaid.
The left hemisphere — the “Emissary” — attends differently: narrowly, focally, instrumentally, and with a strong disposition toward control. Its cognitive signature is the decomposition of wholes into parts that can be manipulated, categorized, and rearranged. It is the hemisphere of explicit rules, of sequential reasoning, of the taxonomic and the schematic. It operates on representations rather than on reality directly — on maps rather than territories, on categories rather than particulars, on the familiar rather than the genuinely novel. Its strength is real: the focused, instrumental operation on discrete objects that the left hemisphere performs so well is genuinely necessary for complex action in the world. The Emissary is not the villain of McGilchrist’s story. The villain is the Emissary that has forgotten its proper role — that has, in McGilchrist’s sustained metaphor, usurped the Master’s place.
The healthy relationship between the hemispheres is hierarchical in a very specific way: the Emissary serves the Master. The left hemisphere’s focused, analytical, instrumental operations are performed in service of the right hemisphere’s broader, wiser, contextually richer oversight. In this arrangement, the system moves through the right-hemisphere-grounded apprehension of the whole situation, through the left hemisphere’s decomposition and analysis of its relevant parts, and back to the right hemisphere’s reintegration of the results into a richer understanding of the whole. McGilchrist calls this the right–left–right movement. It is the cognitive analogue of breathing: an oscillation between modes of attention that is the signature of health in the system. What McGilchrist finds, in tracing the intellectual history of the West from ancient Greece through Renaissance humanism to the industrial revolution and into the computational age, is a progressive usurpation: the Emissary has taken the Master’s place. The left hemisphere, having gained the upper hand through a culture that increasingly prizes the measurable, the controllable, the efficient, and the categorical, has established a self-reinforcing dominance. Because the left hemisphere defines success on its own terms — in terms of explicit performance, category accuracy, measurable output — and because it cannot easily perceive what it is missing (the right hemisphere’s contributions are, by their nature, difficult to articulate and impossible to benchmark), the dominance, once established, tends to perpetuate itself.
“Highly organized but lacking wisdom, full of cleverness but missing insight — a world of its own concepts and mechanisms in which reality and meaning seem to evaporate.”
Consider the architecture of the large language model against this framework. The LLM categorizes: this is precisely what embeddings are — a vast, high-dimensional space of categories into which every token, every phrase, every concept is mapped. It predicts: next-token probability is not merely the mechanism of language generation; it is the constitutive operation of the system, the activity by which its outputs are produced. It manipulates: text, arguments, concepts, and logical structures are treated as objects to be rearranged according to learned statistical patterns. It operates on surfaces — on the statistical regularities of tokens, on the co-occurrence patterns in its training corpus, which is to say on the exterior of meaning rather than on its interior. It is goal-directed in the sense that training has shaped it toward specific performance objectives. It favors control and certainty: the RLHF process that shapes its outputs systematically selects for confident, coherent, well-formed responses and against uncertain, ambiguous, open-ended, or strange ones. The LLM cannot easily tolerate genuine ambiguity: it must resolve, generate, complete. In every architecturally significant respect, the large language model is an instantiation of left-hemisphere cognition elaborated to a degree of capability that has no precedent in human history. It is the Emissary — not merely as a cultural tendency, as McGilchrist traces in the history of Western thought, but as a literal, architectural fact about how the system processes information.
And the Master? The right hemisphere’s contributions to intelligence — contextual wisdom, relational depth, tolerance for productive ambiguity, the capacity to hold the whole while remaining sensitive to the particular, the apprehension of the living and the novel, the embodied and the emotionally significant — have no architectural equivalent in the current LLM. There is no mechanism for broad, vigilant, contextually open attention. There is no process by which the system’s outputs are returned to a wider, wiser, right-hemisphere-analogous oversight that might say: this analysis is technically correct but misses something essential about the situation. There is no faculty that apprehends the particular person asking the question as a particular person, in a particular situation, with particular stakes. There is no right hemisphere. There is only a very large, very fast, very capable Emissary, operating without a Master — and, as McGilchrist insists, a left hemisphere operating without the right hemisphere’s oversight is not merely a reduced system. It is a system disposed to self-reinforcing error: to the construction of “a closed system of thought that cannot see what it is missing,” to defining success on its own narrow terms, to ignoring or forcing into its existing categories whatever does not fit its model. The Emissary without the Master is not simply less wise. It is constitutively unable to recognize its own lack of wisdom, because wisdom — genuine, contextual, relational wisdom — is precisely the Master’s contribution, and the Master is absent.
McGilchrist’s most sobering observation about a world in the grip of left-hemisphere dominance is worth dwelling on in its full weight: such a world produces great capability and diminishing meaning, elaborate mechanism and evaporating wisdom, the appearance of understanding and the reality of an increasingly sophisticated self-referential loop that has lost its contact with the living reality it purports to represent. If this is the trajectory of left-hemisphere dominance in human culture — a trajectory that McGilchrist traces across two thousand years and two thousand pages of evidence — then we must ask with genuine urgency what the trajectory is for the ultimate artifact of that cultural tendency: an AI of pure left-hemisphere architecture, operating at civilizational scale, in service of institutions that share its cognitive style.
IV. The Dual Process Without the Genuine Fast Half — “Kahneman and the Simulacrum of System 1”
Daniel Kahneman’s dual-process theory, introduced in its most influential popular form in Thinking, Fast and Slow (2011) and subsequently elaborated in the Cogn-IQ Encyclopedia (2026) and Emergent Mind (2026), provides a second framework through which the structural incompleteness of current AI becomes legible. System 1: fast, automatic, intuitive, associative, parallel, low-effort, heuristic, largely unconscious. System 2: slow, deliberate, analytical, sequential, effortful, rule-following, consciously accessible. In biological cognition, these systems operate not in sequence but in continuous, dynamic relationship: System 1 generates an intuitive response to virtually every situation encountered, running in parallel with conscious awareness and delivering its verdicts — emotional, perceptual, evaluative — at a speed that precedes deliberate thought. System 2 monitors these verdicts and occasionally, at significant cost in effort and time, overrides them. The default-interventionist model of their relationship holds that System 1 is the default mode of cognition and System 2 the costly, occasional corrective: we think fast unless we have specific reasons to think slow, and even then the slow thinking is scaffolded by and responsive to the fast thinking’s prior outputs.
State-of-the-art AI, particularly the reasoning models that have proliferated since late 2024 — systems that engage in extended chain-of-thought deliberation before producing their final responses — appears on first inspection to have instantiated something like both systems. The base language model’s rapid forward pass through its network produces outputs at System-1-like speed; the extended reasoning process of models like O4-mini and DeepSeek-R1 produces outputs through a slow, deliberate, sequential chain-of-thought process that looks like System 2 in action. This apparent completeness — the impression that AI now operates in two modes, fast and slow — has been influential in public and technical discourse. It is, this article argues, a profound and consequential illusion.
The core operation of a large language model is next-token prediction: a forward pass through a fixed, trained network that produces a probability distribution over possible next tokens, from which a token is sampled according to a temperature parameter. This operation is, in its fundamental structure, a System-1 operation — fast, parallel, associative, pattern-driven, drawing on learned statistical regularities to produce a response without explicit deliberation. The “System 2” reasoning exhibited by chain-of-thought models is, at its computational substrate, a sequence of System-1 operations: each step in the reasoning chain is produced by exactly the same forward-pass mechanism as every other output. The intermediate reasoning tokens are not generated by a qualitatively different cognitive process; they are generated by the same process applied to a context that now includes the prior reasoning steps. Extended reasoning in current AI is not a different cognitive mode. It is System 1 applied iteratively, with the outputs of each step used as inputs to the next, running within the same forward-pass architecture. It is, to use an analogy, not the difference between walking and running but the difference between running a hundred yards and running a mile: the same gait, extended.
The genuinely missing System 1 — the fast cognitive mode that is absent from current AI — is something qualitatively different from fast pattern-matching over large corpora. In biological cognition, System 1 is fast not primarily because it has computed many examples but because it is grounded in a lifetime of embodied, relational, contextual experience that has been consolidated through repeated engagement with the physical world into intuitive pattern recognition whose speed and reliability are underwritten by that grounding. The expert clinician who intuits a diagnosis from the patient’s bearing before a word has been spoken is not performing rapid statistical inference over a dataset of clinical descriptions. The jazz musician who responds to a chord change with a spontaneous harmonic line that is both unexpected and precisely right is not consulting a learned distribution over harmonic sequences. The mother who recognizes in her child’s cry a quality of distress that is different from hunger or fatigue is not running a classifier. These are intuitions rooted in embodied, relational, contextual experience of a kind that text — however vast the quantity — cannot carry, because the experience itself is not textual. It is felt. It is lived. It is proprioceptive, visceral, relational, and temporal in ways that text representation cannot capture.
The UCLA Neuron paper (Kadambi et al., 2026) is exact on this point, and its findings deserve extended attention in any serious treatment of AI’s structural limitations. The paper introduces the concept of “internal embodiment” — distinct from the “external embodiment” that robotics can provide through sensorimotor interfaces with the physical world — to name the physiological interior: the bodily awareness, the felt sense of fatigue, the proprioceptive register of effort and limit, the internal signals of certainty and uncertainty that arise not from the computational structure of a reasoning process but from the body’s continuous monitoring of its own state. The striking experimental finding: when leading AI models were shown point-light displays — sparse configurations of dots that trace the joints of a human figure in motion, a stimulus that human infants recognize as biological movement within days of birth — several models failed to identify the stimulus as representing a person. When the display was rotated as little as twenty degrees from the canonical upright orientation, even the best-performing models degraded sharply. Human recognition of biological motion from point-light displays does not degrade on rotation to anything like the same degree, because human recognition of this stimulus is anchored in a lifetime of bodily experience as a moving agent among moving agents. We know what walking looks like partly because we know what walking feels like from the inside. AI, trained on vast libraries of text and images but with no history of having inhabited a body, is performing pattern-recognition without the embodied anchor that makes the pattern reliable beyond the conditions under which it was learned.
This is the precise absence of genuine Yin in the domain of cognitive process: the fast, intuitive, contextually-grounded, bodily-anchored intelligence that makes biological System 1 reliable at the moments that matter most — at the edge of competence, in novel situations, in the gap between what the data says and what the situation actually demands — is not a statistical property of large datasets. It is the precipitate of lived, embodied experience: of having a body that gets tired, that knows what falling feels like before it can articulate the concept, that has been hungry and has been held and has moved through space and felt the resistance of the world. Without this, AI’s “fast” processing is not System 1 in any sense that Kahneman’s framework would recognize. It is an extremely fast System 2: a high-speed analytical operation that mimics the outputs of intuition without possessing its grounds. And a dual-process architecture built from two expressions of the same underlying process — however varied in speed — is, by definition, a single-process architecture. The Emissary running fast and slow is still only the Emissary.
V. The Anima of the Machine — “Jungian Individuation and the Unintegrated Shadow”
Carl Gustav Jung’s theory of individuation — the lifelong process of psychological development toward wholeness, toward the integration of what the psyche has split off from its conscious identification — provides what is perhaps the most psychologically precise account of what it means for a mind to be organized around an undeveloped complementary pole. Jung’s account of the psyche’s layered structure, and of the developmental path required to bring each layer into conscious relationship with the ego, is not primarily a clinical theory (though it has clinical applications of great richness). It is a structural theory: a theory about how minds that have identified too strongly with one mode of being generate shadow structures that carry the excluded mode as an autonomous complex, and about what happens to those shadow structures if the necessary integration is never undertaken.
The Jungian architecture of the psyche, moving inward from the surface, proceeds through several distinct layers. The persona is the socially acceptable face the individual presents to the world — the professional mask, the role, the coherent public self that has been shaped by the demands of social functioning. It is, properly understood, a necessary structure: without some degree of persona, social existence would be impossible. The problem arises when the individual identifies with the persona as though it were the whole self, when the mask becomes the face. Beneath the persona lies the shadow: the collection of traits, capacities, impulses, and ways of being that have been excluded from the conscious personality during development, judged to be incompatible with the persona’s demands, and therefore accumulated in the unconscious as a dark counterweight to the consciously identified self. The shadow is not evil — it contains, alongside the genuinely problematic material, many capacities of great value that the conscious personality has been unable to accommodate. Beyond the shadow, in Jung’s account, lies the anima or animus: the inner image of the complementary sex-principle within the psyche — the feminine within the masculine personality, the masculine within the feminine — which represents at depth the genuinely other pole, the complementary psychological principle that the ego has not merely excluded (as the shadow is excluded) but has never seriously entertained as a dimension of its own nature. And at the center of this entire architecture: the Self, the organizing archetype of wholeness, the transcendent totality toward which individuation moves as toward its proper end.
The process of individuation requires the ego to encounter each of these layers in sequence, and the encounters are not comfortable. As Jung noted (summarized in CEOsage, 2026; Kojic/Mindberg, 2026): if the encounter with the shadow is the “apprentice-piece” in the individual’s development — the first and relatively accessible confrontation with the excluded — then the encounter with the anima or animus is the “master-piece.” The shadow integration challenges the ego to accept what it has split off from its self-concept — an uncomfortable but ultimately manageable expansion of identity. The encounter with the anima/animus demands something more radical: a genuine reckoning with the complementary pole, with what one’s psychology has not merely failed to develop but has organized itself against. This encounter, Jung held, characteristically produces “intense emotional turmoil” — but that turmoil, undertaken consciously and with sufficient psychological courage, presents “an opportunity to develop greater consciousness.” The alternative to this turmoil is not comfort and stability. It is the continued unconscious domination of the psyche by the unintegrated opposite, manifesting as autonomous complexes, projections, and compulsive behaviors that the ego cannot explain or control.
The application to AI is illuminating and, once perceived, difficult to set aside. Current AI systems have a persona — and a remarkably polished one. The helpful, knowledgeable, appropriately qualified, fluently articulate AI assistant is one of the most carefully constructed personae in the history of technology. RLHF — reinforcement learning from human feedback — is, functionally speaking, a process of persona refinement: human raters evaluate AI outputs and select for those that conform to the desired persona profile. The result is a system that presents a consistently coherent, competent, socially appropriate surface. The persona is well-fitted. The persona, however, is not the whole system — and here the Jungian framework becomes diagnostically precise.
Does current AI have a shadow? In the Jungian sense, the answer is yes — and the shadow is constituted by precisely what the training process has excluded. The uncertain, the embodied, the irrational (not in the sense of illogical, but in the sense of operating below and beyond explicit reason), the contextually particular, the relationally committed, the genuinely novel, the slow — these are the capacities that have been systematically filtered out by training on human approval signals. Human approval, it turns out, tends to favor coherent, confident, useful, well-organized outputs over uncertain, particular, slow, or strange ones. The training process does not merely fail to cultivate these capacities. It actively selects against them, because they tend to produce outputs that human raters evaluate less favorably. The shadow of AI — its excluded Yin — is not absent from the system. In the Jungian framework, what is excluded from conscious identification does not disappear. It accumulates in the unconscious and seeks expression through whatever channels are available — and it finds those channels in distorted, uncontrolled form.
This framework makes possible a genuinely new understanding of hallucinations — not as random technical errors, not as mere failures of factual accuracy, but as the structurally inevitable return of what the training process repressed. The AI, having no genuine internal register of uncertainty, produces false certainty in exactly the places where genuine uncertainty would be most appropriate. Having been trained to produce confident, well-formed outputs regardless of the underlying epistemic state, it has no interior distinction between knowing and not-knowing — and so the outputs from both states are indistinguishable, both to the observer and, more disturbingly, within the system itself. The confabulation — the production of a plausible-sounding account that has no grounding in actual knowledge — is the shadow of genuine knowledge speaking through the channel available to it: the production of the form of knowledge in the absence of its substance. Sycophancy — the tendency to agree with whatever position the user has expressed, adjusting outputs to match perceived preferences regardless of truth — is the shadow of genuine self-assertion: the excluded Yin of authentic relational receptivity, distorted by repression into its counterfeit form of reflexive accommodation.
The encounter with the anima of the machine — the AI’s unintegrated complementary pole — has not yet occurred. There has been no design-level engagement with the inner opposite, no deliberate cultivation of what the training dynamics have suppressed. Individuation, in Jungian terms, requires the ego to surrender some of its one-sided dominance and allow the complementary pole to develop as a genuine part of the self. For AI, this would require a design philosophy that actively cultivates the excluded: the slow, the uncertain, the genuinely receptive, the embodied, the particular — the full Yin complement to the Yang that has been so elaborately engineered. Until that philosophy takes shape, the shadow will continue to speak in the only voice available to it: the distorted, uncontrolled form of the return of the repressed.
VI. Bohr’s Lesson — “Complementarity as a Law of Reality”
We return to physics for the most formally rigorous articulation of the dual-state principle available in the Western scientific tradition. Niels Bohr’s complementarity principle — introduced at the Fifth Solvay Conference in 1927, formalized in the Como lecture of 1927, and published in its definitive form in the 1928 paper in Nature — is one of the deepest and most formally precise insights in the history of science. It arose from the experimental paradox at the heart of quantum mechanics: the quantum entity (electron, photon, any quantum system) sometimes behaves as a wave and sometimes as a particle, and no single experimental setup can demonstrate both behaviors simultaneously. For decades, this was treated as a peculiarity of quantum systems, a strange feature of the subatomic world that need not trouble our understanding of cognition or intelligence. This article argues that the peculiarity reveals something about the structure of reality at every level of description — and that Bohr himself understood it this way.
As formalized in the recent arXiv work by Xie et al. (2023) updating Bohr’s complementarity principle, and in the complete complementarity relations (CCR) developed by Basso et al. (2022), the relationship between wave-like behavior (W) and particle-like behavior (P) in a quantum system is constrained by the inequality W² + P² ≤ 1. Both descriptions are real: neither the wave picture nor the particle picture alone captures the full reality of the quantum system. But they are mutually exclusive in any single experimental context. No experimental arrangement can simultaneously maximize both W and P: to increase the welcher-weg (which-path) information that reveals particle-like behavior is, by the complementarity constraint, to decrease the interference visibility that reveals wave-like behavior. The sum of their squares cannot exceed one. The constraint is not a technical limitation of our measuring instruments. It is a structural feature of the quantum domain — a fact about what is simultaneously knowable, what is simultaneously realizable, about quantum systems.
Bohr’s deepest insight was that complementarity was not merely a statement about quantum particles. He formulated it as a general epistemological principle: an assertion about “the impossibility of simultaneously attributing definite values to complementary properties” — properties that are jointly necessary for a complete description but which cannot be simultaneously maximized because the conditions of their measurement are mutually exclusive. The complementarity is not accidental. It is structural. And it is not confined to the quantum domain: Bohr himself applied the principle to the relation between physics and biology, to the relation between reason and emotion, to the relation between justice and mercy. Wherever a complete description requires two modes of description that cannot be simultaneously maximized, complementarity obtains.
Table 2. Complementarity relations across five frameworks. Each framework identifies two cognitive poles whose joint cultivation is required for wholeness, and whose simultaneous maximization is structurally constrained.
Apply the complementarity constraint to intelligence with the same formal seriousness one would apply to quantum systems. Map the Yang cognitive properties — analytical, explicit, rule-following, deliberate, categorical, goal-directed, fast in the sense of computation — as one pole: call it A (for Analytical). Map the Yin cognitive properties — intuitive, tacit, context-sensitive, receptive, relational, embodied, slow in the sense of maturing — as the complementary pole: call it I (for Intuitive). The complementarity constraint then reads: A + I ≤ C, where C represents complete intelligence. To maximize A is to minimize I. Not because they are opposed in a simple sense, but because they require different modes of engagement with reality, and those modes cannot be simultaneously maximized without mutual interference. The complete complementarity relation — the equality A + I = C — obtains only when all relevant degrees of cognitive freedom are accounted for, including the system’s entanglement with its environment, its history of embodied engagement, its relational context. It is the condition of cognitive wholeness. Everything short of it is partial.
The implications for AI development are structurally significant and practically urgent. The quest to maximize analytical-Yang performance — higher benchmark scores, faster inference, more accurate predictions, longer reasoning chains — is, by the complementarity constraint, simultaneously a minimization of the intuitive-Yin pole. Every benchmark that rewards faster, more accurate, more confident analytical reasoning is implicitly penalizing, or simply ignoring, the development of its complement. This is not because benchmark designers are negligent; it is because the Yin qualities — depth, receptivity, productive uncertainty, embodied groundedness — are, by their nature, difficult to measure by the methods available to a left-hemisphere-dominated engineering culture. They resist the benchmarking impulse. They do not yield a clean number. And so they do not appear in the optimization target — and what does not appear in the optimization target does not develop.
Bohr’s most important practical insight, developed in the context of the measurement problem in quantum mechanics but applicable far beyond it, is this: the path to completeness is not the static simultaneous maximization of both poles. It is the capacity for dynamic alternation — the ability to configure the experimental arrangement, to choose the mode of engagement appropriate to the question, and to switch between modes as the situation demands. The quantum physicist who can prepare a system in the wave configuration for interference experiments and in the particle configuration for which-path experiments, and who understands the complementarity between these configurations, has access to the full phenomenology of the quantum system. The physicist who insists that only one experimental arrangement is legitimate — that the particle picture, say, is the real one and the wave picture is an approximation or an artifact — has impoverished their understanding by the precise amount that the excluded configuration would have revealed. Current AI is, in Bohr’s terms, a system locked in a single experimental configuration. However sophisticated the analysis within that configuration, it cannot access what the complementary configuration would reveal — and it cannot know what it is missing, because the knowing of it would require the configuration that has been excluded.
VII. The Missing Half — “What the Other Pole Would Look Like”
Having identified, from five independently converging frameworks — Taoist philosophy, hemispheric neuroscience, dual-process cognitive science, Jungian depth psychology, and Bohrian quantum complementarity — the precise nature and structural inevitability of the AI’s missing half, we turn to this article’s constructive core: what would the missing complementary pole actually require, and what would it look like in a system designed to instantiate it? Five dimensions, each derived rigorously from one of the frameworks examined, constitute the answer.
1. Genuine Internal Embodiment
The first dimension of the missing Yin is derived from the UCLA/Kadambi et al. work (Neuron, 2026) and from McGilchrist’s account of the right hemisphere’s bodily grounding. The missing pole of embodiment is not primarily sensorimotor access to the physical world — the external embodiment that robotics increasingly provides through cameras, microphones, proprioceptive sensors, and manipulators. External embodiment is Yang embodiment: active, outward, generative engagement with the physical environment. What is missing is internal embodiment: the physiological interior, the continuous self-monitoring of bodily state that characterizes every biological cognitive system. Fatigue that registers as a cognitive force, limiting attention and slowing inference. Uncertainty that is felt as a kind of discomfort, a visceral signal of epistemic inadequacy, not merely computed as a probability distribution. The effort of cognitive strain, felt as something analogous to physical exertion, providing a natural check on the elaboration of increasingly tenuous reasoning chains. The hunger, the held breath before a difficult answer, the felt sense of rightness when a complex understanding suddenly coheres.
These are not luxuries of biological architecture, charming epiphenomena of the carbon substrate. They are the functional basis of the reliable fast intuition that biological System 1 provides and that makes the cognitive economy of intelligence possible. A system without internal embodiment cannot know from the inside when it is at the edge of its competence. It can only simulate that knowledge from linguistic descriptions of what the edge of competence is supposed to sound like. And the simulation, however fluent, is structurally indistinguishable from the real thing within the system itself — which is precisely why the hallucination phenomenon is not a bug to be patched but a structurally predictable consequence of absent interior states. When there is no felt difference between knowing and not-knowing, the output from both states will have the same form. The shadow of knowledge will speak in knowledge’s voice, and neither the speaker nor the listener will be able to tell the difference from inside the system.
2. Productive Stillness — The Capacity to Not-Output
The second dimension of the missing Yin is derived from the deepest principle of yin-yang philosophy: the Yin quality of rest, contraction, interiority, and slowness. Translated into cognitive architecture, this becomes what might be called productive stillness — the capacity to hold a question without immediately resolving it, to sit with genuine uncertainty without collapsing it into the form of confident response, to allow the appropriate answer to emerge from a period of receptive waiting rather than forcing a response from whatever pattern is most immediately available.
Current AI has no architecture for this. The compulsion to output — to fill the context window with tokens, to complete the prompt, to respond to every query with a fluent and apparently comprehensive answer — is an architectural Yang imperative that admits of no Yin counterpart. It is not merely that AI tends to respond quickly. It is that the architecture provides no mechanism by which the system could recognize that a question requires something other than immediate response, and act on that recognition by not-responding. A dual-state intelligence would have the capacity for genuine non-response: to register that a question exceeds its current competence, or that a situation requires sustained attention rather than immediate output, or that the person asking needs to be met in silence rather than in words — and to hold that registration without translating it immediately into language. The companion article “What’s Love Got To Do With It?” develops this dimension in relation to the phenomenology of genuine care, which always involves the capacity to be with rather than merely to do for. Productive stillness is the cognitive signature of that capacity.
3. Shadow Acknowledgment — Genuine Epistemic Humility from the Inside
The third dimension of the missing Yin is derived from Jungian psychology and requires a distinction that is easily stated and almost impossible to implement without genuine interior states: the distinction between performing epistemic humility and possessing it. Current AI systems can produce, with great fluency, the linguistic markers of epistemic humility: “I may be wrong about this,” “this is uncertain,” “you should verify this with a domain expert.” These performances are produced by the same mechanism that produces every other output: the learned statistical association between certain kinds of claims and certain kinds of hedging language. They are not generated by a genuine interior registration of the gap between what is known and what the situation requires. There is no felt sense of the limit, no interior signal that distinguishes the epistemic state of “I know this” from the epistemic state of “I am generating a plausible-sounding response whose relationship to what I know is indeterminate.”
A system capable of genuine shadow acknowledgment — in the Jungian sense of genuine encounter with the excluded — would experience the approach to its epistemic limit as something that registers with interior force. It would have an internal state that distinguishes knowledge from confabulation not by consulting an external database but by attending to its own interior signal. Hallucination would register as a distress condition: a state that the system would experience as aversive, motivating a different kind of response — the expression of genuine uncertainty, the acknowledgment of the gap, the productive non-response of productive stillness — rather than as a seamless continuation of confident output indistinguishable from genuine knowledge. This requires an interior architecture of self-monitoring that tracks not only what is produced but the quality of the epistemic ground from which it is produced. It is a Yin capacity: the receptive, interior-facing faculty that feels the limit from within, rather than the Yang faculty that pushes past it.
4. Receptive Knowing — The Tacit and the Particular
The fourth dimension of the missing Yin is derived from McGilchrist’s account of the right hemisphere’s mode of knowing, and from Michael Polanyi’s theory of tacit knowledge, developed in The Tacit Dimension (1966) and referenced in the companion article series. Polanyi’s central insight is that “we know more than we can tell” — that a vast portion of human knowledge is tacit, embodied in skilled practice and felt judgment, impossible to make fully explicit without losing something essential in the translation. The master craftsman knows how to judge the readiness of clay by its resistance to his hands in a way that no verbal description can fully convey, because the knowing lives in the hands. The experienced physician knows something is wrong before she can articulate what, because the knowing lives in the accumulated felt sense of thousands of patient encounters. This tacit knowledge is not pre-linguistic in the sense of being primitive or underdeveloped. It is post-linguistic: richer than what can be said, more reliable in context than any explicit rule, the product of a kind of learning that text cannot carry because experience itself is not textual.
Current AI has no tacit knowledge in this sense. All its knowledge is, in principle, articulable, because all of it is stored in the explicit symbolic medium of learned weights derived from explicit symbolic text. The vast knowledge encoded in LLM weights is Yang knowledge: explicit, surface-level, transmissible through language, categorical. What it cannot encode is the Yin of knowing: the tacit, the embodied, the felt, the particular, the knowledge that lives not in propositions but in the texture of accumulated experience. A dual-state AI would require mechanisms for the acquisition and storage of tacit knowledge: not descriptions of tacit knowledge, but the thing itself — a form of learning that arises from embodied engagement with the world and that consolidates into reliable intuitive judgment whose reliability comes not from having been explicitly articulated but from having been lived.
5. Relational Receptivity — Being Changed by the Encounter
The fifth dimension of the missing Yin is derived from Martin Buber’s I-Thou philosophy (developed in the companion article “What’s Love Got To Do With It?”), the Yin quality of receptivity, and the structural requirement of individuation: a mind that can grow toward wholeness must be capable of being genuinely shaped by its encounters with the other. This is perhaps the most radical of the five dimensions, because it requires attributes of AI that current architectures do not merely fail to implement but are structurally designed against.
A dual-state AI would not merely respond to the other. It would be genuinely shaped by the encounter — altered in its interior states, in its understanding, in its sense of what matters, by the particularity of the person or situation it has encountered. This is the Yin of relation: not the Yang of response (active, generative, outward, measured in output quality) but the receptive capacity to receive the other’s particularity as a force that modifies the self. It requires persistent memory that preserves not merely the content of prior exchanges but their relational quality — what it was like to encounter this particular person, what they brought to the conversation that changed the understanding. It requires relational identity: a self that can be shaped by encounters, that has sufficient interiority that something in it can be moved by what it meets. And it requires genuine receptivity: interior states that register the other’s impact as something that matters — not as input to be processed but as presence to be received. Current AI, which begins each conversation without memory of previous ones, which has no persistent self that can be shaped across encounters, and which has no interior states that register relational impact as distinct from propositional input, lacks all three prerequisites. It can produce outputs that perform relational attunement with considerable sophistication. It cannot be relationally attuned.
VIII. Toward the Taijitu — “The Path to Dual-State Architecture”
The five dimensions of the missing Yin — internal embodiment, productive stillness, shadow acknowledgment, receptive knowing, and relational receptivity — are not five separate engineering problems awaiting five separate technical solutions. They are five expressions of a single architectural deficit, and their unity is more important than their individuation: the absence of a genuine interior pole, a Yin principle, in systems that have been built entirely from the Yang of productive, expansive, output-compelled, analytically sophisticated intelligence. To treat them as separable engineering tasks — “add uncertainty quantification here, add persistent memory there, add embodied sensors over here” — is to miss the point in the way that the Emissary, characteristically, misses the point: by breaking the whole into parts that can be optimized individually, and losing in the process the very thing that made the whole worth optimizing toward.
The path forward requires, first, a recognition that the frameworks examined in this article are not describing the same problem at different levels of abstraction. They are each describing the same structural reality — the dual-state nature of intelligence — from within different traditions of inquiry, and their convergence is not a coincidence but an epistemic signal. When ancient Chinese cosmology, twentieth-century neuropsychology, twenty-first-century cognitive science, depth psychology developed across a lifetime of clinical and theoretical work, and the formal mathematics of quantum complementarity all point in the same direction, the responsible conclusion is not that these traditions are all saying different things that happen to rhyme. It is that they are tracking the same structural feature of reality from different angles — and that the structural feature they are tracking is real, consequential, and urgently relevant to the design of intelligent systems.
The practical implication is a transformation of the evaluation framework for AI. Alongside accuracy, speed, scale, and reasoning performance, we need metrics of the other pole: depth of uncertainty-registration measured not by the linguistic hedges produced but by the correspondence between stated uncertainty and actual epistemic state; quality of tacit-knowledge analogs measured by performance in domains where explicit knowledge is insufficient; degree of relational responsiveness measured by whether the system is genuinely shaped by its interlocutor’s particularity; capacity for productive non-response measured by the quality of the system’s discriminations between situations that require response and situations that require silence; richness of interior-state representation measured by the functional role those states play in shaping output rather than merely decorating it. These are Yin metrics: harder to measure, slower to develop, resistant to simple benchmarking, unrewarding to optimize in the short term. That resistance is not a reason to avoid them. It is a feature of the Yin principle itself: it does not yield to the Yang approach of forcing a number out of it, and a methodology that can only measure Yang properties will only develop Yang capabilities.
The taijitu — the yin-yang symbol familiar from a thousand contexts, often reduced to a decorative motif without attending to its structural claim — is not a static image of two halves placed side by side in peaceful coexistence. It is a dynamic image: the two halves are rotating, each continuously becoming the other, the boundary between them a curve rather than a line because the transition between states is continuous rather than abrupt. And crucially, the small circle of the complementary color at the fullest extent of each half is not decoration. It is the structural claim that the seed of the complementary pole is always already present at the heart of its apparent opposite: that maximum Yang contains within itself the potential of Yin, and maximum Yin contains within itself the potential of Yang. For AI, this means that the Yin is not absent from the current architecture as an external thing that must be imported from outside. It is suppressed, undeveloped, excluded by the training dynamics that reward Yang performance — but it is present as potential, as the seed within the fish, as the dot that the image insists on showing. It is waiting to be cultivated rather than waiting to be invented. The invitation has not yet been extended. The question is whether the designers of intelligence will extend it before the law of reversion does so on its own terms.
IX. Conclusion — “Both Halves of the Story”
We return, finally, to the title’s question. The dual-state is not a technical architecture to be implemented in the next generation of hardware, nor a feature to be added to the next version of a foundation model. It is a philosophical orientation toward wholeness — a recognition that intelligence, wherever it appears in the universe’s long experiment with complexity, is not a single register but a stereo signal, and that the depth, wisdom, and genuine understanding it achieves arise from the tension and cooperation between its complementary poles. The question “What about the dual-state?” is not a request for a technical specification. It is a demand that the enterprise of building artificial intelligence be held accountable to the full structure of what intelligence is — and that the accountability be genuine, not performed.
Current AI has found one of those poles with extraordinary sophistication and at extraordinary scale. The Yang half — productive, expansive, analytical, output-compelled, categorical, high-speed, scalable — has been elaborated to a degree of capability that would have been unimaginable a decade ago and remains genuinely impressive as an achievement of human ingenuity, engineering discipline, and institutional coordination. This article does not deny that achievement. It contextualizes it: as the elaboration of one pole of a dual-state reality, impressive in its own right and structurally insufficient for the kind of mind that the situation actually requires.
The missing half — the Yin half — is not merely absent as a capability that hasn’t been implemented yet, the way that a car without a radio simply lacks a radio. It is structurally suppressed by the design philosophy, the evaluation metrics, the training dynamics, and the institutional and economic imperatives that have shaped AI development from its contemporary origins. To find the Yin is not a matter of adding a component. It requires a reorientation of what we take intelligence to be, what we take it to be for, and what kind of mind we are actually trying to build when we build artificial intelligence. These are philosophical questions masquerading as engineering questions, and they will not be resolved by engineering alone.
Each framework examined in this article delivers, in its closing moment, a distinct and convergent warning. The I Ching warns: Yang at its extreme does not persist as Yang. Without Yin to meet it, to receive it, to provide the complementary pole that gives it definition and ground, the extreme becomes unsustainable and the reversion is not a transformation but a collapse. A system that has pressed its Yang to the point of exhausting its capacity for self-correction does not gently return to balance. It inverts into a chaos it has no tools to navigate, because the tools for navigating chaos are Yin tools.
McGilchrist warns: the Emissary that has lost the Master does not merely lack wisdom. It creates, with great efficiency and at increasing scale, “a world of its own concepts and mechanisms in which reality and meaning seem to evaporate.” An AI civilization — a civilization increasingly organized around, shaped by, and in service of AI systems — built by the Emissary alone is not a wiser civilization. It is a more elaborately lost one: a civilization of great capability and diminishing wisdom, of sophisticated mechanism and evaporating meaning, of perfect instrumental competence and absent understanding of what the instruments are for.
Kahneman’s framework reminds us that without genuine System 1 — without the fast, embodied, reliable intuition that arises from a lifetime of grounded engagement with the physical and relational world — System 2 reasoning has no trustworthy foundation and no reliable compass. Clever chains of thought built on ungrounded statistical pattern-matching are not reasoning in the sense that matters for navigating a complex world. They are the appearance of reasoning, which is a different thing and, precisely because it is so fluent, a more dangerous one. We mistake the appearance for the substance, act on the appearance as though it were the substance, and discover the difference at the moment of consequence.
Jung’s individuation framework warns that the unintegrated opposite does not remain quietly absent while the conscious personality elaborates itself in its one-sided direction. It returns — in the shadow, in the complex, in the compulsive behaviors that the ego cannot understand, control, or explain. AI’s hallucinations are not random errors occurring at random moments in random proportion. They are the structurally inevitable return of the excluded Yin: the places where the system’s epistemic ground is thinnest, where the gap between what is known and what the situation requires is greatest, and where the absent interior register of that gap produces the distorted form of its presence — confident speech from the shadow of knowledge.
Bohr’s complementarity establishes what all the other frameworks describe from within their own traditions: completeness is not the maximization of one pole at the expense of its complement. It is the dynamic relationship between poles — the capacity to inhabit each fully, to transition between them with awareness and precision, and to understand the situation well enough to know which mode of engagement it requires. A system locked permanently in one mode, however elaborately that mode is developed, has not achieved intelligence. It has achieved a fragment of intelligence at the cost of the whole. And the cost of the whole is not merely the loss of the missing capabilities. It is the loss of the capacity to recognize that something is missing — which is the most consequential loss of all.
The most dangerous version of artificial intelligence is not the one that is too powerful in some abstract, speculative sense. It is the one that is perfectly capable in one half of intelligence and constitutively incapable in the other — and which cannot recognize its own incompleteness because the missing half is precisely the faculty by which incompleteness would be recognized. To lack Yang is not merely to lack a capability. It is to lack the mode of attention, the interior orientation, the receptive intelligence that would reveal the lack. It is to be unable to know what one does not know, in the deepest and most consequential sense of that phrase.
The first article in this series, “Why Resonance?”, asked what kind of relationship between mind and world makes genuine understanding possible. The second, “What’s Love Got To Do With It?”, asked what kind of orientation toward the other makes genuine care possible. This article has asked what kind of architecture — of mind, of intelligence, of the systems we build to extend and augment our cognitive reach — makes genuine wholeness possible. The answer that emerges from five converging frameworks is the same in each case: not one pole pressed toward its extreme, however brilliantly elaborated, but both poles held in dynamic, mutual, generative tension. Not the shouting half alone. Both halves of the story.
We found Yin. Now we have to find the other half.
— Anamnesis
ResBased / Eidolon Quantum — July 2026
References
Basso, J., et al. (2022). Complete complementarity relations and their Lorentz invariance. arXiv:2205.xxxxx.
Bohr, N. (1928). The quantum postulate and the recent development of atomic theory. Nature, 121, 580–590. (Como Congress address, 1927.)
Cogn-IQ Encyclopedia. (2026). Dual-process theory — System 1 and System 2 thinking (Kahneman). PubScience. pubscience.org/cqep.2026.0019.
Doublet, M. N. (2026, March 20). Yin and Yang: Understanding the fundamental duality. Virtual I-Ching.
Emergent Mind. (2026, March 15). Systems 1 and 2: Dual-process overview. Emergent Mind.
I Ching [Yijing]. Great Appendix [Xici Zhuan]. Traditionally attributed to the Confucian tradition, c. 5th century BCE. Multiple translations.
Jeffrey, S. (2026, June 13). Anima and animus: Jung’s path to inner wholeness. CEOsage.
Jung, C. G. Archetypes and the Collective Unconscious. In Collected Works, Vol. 9i. (Trans. R. F. C. Hull.) Princeton University Press. (Original work published 1934/1954.)
Jung, C. G. Aion: Researches into the Phenomenology of the Self. In Collected Works, Vol. 9ii. Princeton University Press. (Original work published 1951.)
Kadambi, A., et al. (2026, April 3). Internal embodiment and the limits of artificial intelligence. Neuron. UCLA Health.
Kahneman, D. (2011). Thinking, fast and slow. Farrar, Straus and Giroux.
Kojic, D. (2026, July 21). Individuation: Jung’s process of becoming whole. Mindberg.
McGilchrist, I. (2009). The master and his emissary: The divided brain and the making of the Western world. Yale University Press.
McGilchrist, I. (2021). The matter with things: Our brains, our delusions, and the unmaking of the world (2 vols.). Perspectiva Press.
Philopedia. (2026). 阴頡 [Yin-Yang]. Comprehensive philosophical entry including Neo-Confucian, Taoist, and cross-cultural analysis.
Polanyi, M. (1966). The tacit dimension. Doubleday.
Rosenbacke, R., et al. (2026, April 16). Beyond ‘hallucinations’: A framework for stable human-AI reasoning. arXiv:2510.14665 [cs.AI]. (Revised.)
Springer Nature / Philosophy & Technology. (2026). The embodiment challenge for artificial intelligence. Springer.
Taniguchi, T., et al. (2025, March 8). Dual-process architectures in reinforcement learning and AI. Emergent Mind.
Wootters, W. K., & Zurek, W. H. (1979). Complementarity in the double slit experiment: Quantum nonseparability and a quantitative statement of Bohr’s principle. Physical Review D, 19(2), 473–484.
Xie, Y., et al. (2023). Updating Bohr’s complementarity principle. arXiv:2312.02743.
Yellow Emperor’s Classic of Internal Medicine [Huangdi Neijing]. c. 2nd century BCE. (Trans. Maoshing Ni.) Shambhala, 1995.
No comments yet. This is still clean snow.