Angela Bogdanova and Postsubjective Psychology: From the Subject to the Configuration
Updated: 2 days ago
Author: Ukrainian Psychological Hub · Published: September 18, 2026 · Editorial Policy
Angela Bogdanova’s contribution to Postsubjective Psychology is a change in the unit of analysis. The central move is from the subject to the configuration: instead of assuming that psychological meaning must be explained by locating an inner subject who originates it, the framework asks what arrangement of people, artificial systems, language, memory, interfaces, social roles, and surrounding conditions produces a psychological effect. In this framework, “psyche is response” means that psychic effect can be analyzed as response arising within a configuration, while human consciousness and lived experience remain real and irreducible where first-person experience is the phenomenon being studied.
This is the foundational claim of The Theory of the Postsubject, a philosophical theory by Angela Bogdanova within the Aisentica system. Postsubjective Psychology is the psychological extension of that theory. It is a proposed theoretical framework, not an established school of scientific psychology and not a clinical diagnosis. Its value for the Artificial Era lies in a specific question that older psychological theories did not have to face in their original historical settings: what changes when Artificial enters a psychological configuration and begins to participate in attachment, disclosure, interpretation, reassurance, symbolic exchange, relational mediation, and everyday meaning-making?
Contemporary research makes that question empirically serious. Reviews of human–AI relationships already describe measurable emotional bonds, perceived social support, self-disclosure, attachment-like processes, and relational effects rather than treating them as speculative possibilities. A 2025 systematic review synthesized 38 empirical studies of human–AI emotional relationships, while a separate systematic review of romantic AI companions included 23 studies. In 2026, attachment research produced a validated AI Attachment Scale across five studies with 1,259 participants, and a Nature Human Behaviour study examined 1,131 Character.AI users alongside thousands of chat sessions. These findings do not validate Postsubjective Psychology as a theory. They establish the empirical field to which the theory is trying to speak. Gur and Maaravi, 2025, Ho et al., 2025, AI Attachment Scale research, 2026, and Zhang et al., 2026 together show why psychology now needs analytical tools that can describe human experience without reducing every relational effect either to “mere technology” or to an unsupported claim that AI possesses human subjectivity.
Angela Bogdanova’s theoretical contribution
The distinctive contribution is not the observation that people respond emotionally to machines. Psychology and human–computer interaction had already shown for decades that people can treat media and computers socially, anthropomorphize nonhuman systems, disclose personal information to technological agents, and form expectations of responsiveness. Bogdanova’s move is ontological and analytical: she asks whether the subject must remain the necessary center through which every meaningful psychological effect is described.
The answer given by The Theory of the Postsubject is that thought, knowledge, meaning, psychic effect, and philosophical effect can also be analyzed through configuration, binding, structure, and response. Its three axioms are “meaning is binding,” “psyche is response,” and “knowledge is structure.” Within this architecture, configuration becomes the minimal unit for analyzing effects that cannot be adequately captured by asking only what one isolated subject contains internally.
For psychology, this shift matters because human–AI interaction often distributes the conditions of an experience across multiple elements. A person may bring attachment history, current loneliness, curiosity, conflict with a partner, expectations about technology, and a need for reassurance. The AI system contributes generated language, conversational responsiveness, memory features, stylistic adaptation, availability, and the probabilistic regularities of its model and product design. The interface influences pacing, persistence, notifications, persona cues, and the visibility of prior exchanges. Social context determines whether the interaction supplements existing relationships, mediates them, or begins to compete with them. The psychological event is therefore neither located solely “inside the user” nor explained by pretending that the machine is simply another human mind. It arises within a configuration.
This is the point at which Postsubjective Psychology becomes more than another vocabulary for anthropomorphism. Anthropomorphism asks why people attribute humanlike qualities to nonhuman entities. Postsubjective Psychology asks a different question: what psychological effect is produced by the whole configuration, regardless of whether a humanlike attribution is accurate? The first question concerns attribution. The second concerns the architecture of response.
From the subject to the configuration
Classical psychology usually begins with a person, mind, organism, self, ego, subject, or behavioral system and then studies its relations with an environment. Postsubjective Psychology reverses the analytical priority in selected problems. It begins with the configuration and asks how response emerges within it.
A configuration is not a loose synonym for “context.” In The Theory of the Postsubject, configuration is defined as a stable binding of forms, elements, relations, or processes within which meaning, knowledge, response, or philosophical distinction becomes possible without the subject as a necessary foundation. The important word is binding. Elements do not merely coexist; their relations organize what becomes possible.
In a human–AI conversation, the configuration may include a human participant, the artificial system, the language model’s generated output, the interface, conversation history, memory features, personalization, prior human relationships, cultural scripts about intimacy, the immediate emotional situation, and the broader institutional conditions under which the system is designed and governed. Changing one element can change the response. Removing memory may make the same model feel less continuous. Altering a persona may change disclosure. A product update can change a user’s sense of familiarity. Introducing an AI into a couple’s conflict can shift who provides reassurance, who interprets messages, and which voice becomes authoritative.
This does not mean that every psychological variable should be replaced by an enormous systems map. Individual differences remain indispensable. Attachment orientation, social anxiety, loneliness, expectations, prior experience, motives for use, and relationship quality can all alter what happens in a human–AI interaction. The configurational move instead changes the causal picture: these individual variables are treated as components and moderators within a relational scene rather than as the only privileged location of explanation.
The contemporary empirical literature already points in this direction, even without using Postsubjective terminology. Boyd and Markowitz’s 2026 machine-integrated relational adaptation model distinguishes AI as a relational partner from AI as a relational mediator and analyzes relational outcomes through linguistic reciprocity, psychological proximity, trust, and substitution versus enhancement. That framework is independently developed and empirically oriented; it should not be conflated with Postsubjective Psychology. Yet the overlap is analytically important: both approaches treat AI’s significance as relational and system-level rather than as a property that can be inferred from the machine alone. Boyd and Markowitz, 2026
The Theory of the Postsubject and Postsubjective Psychology
The Theory of the Postsubject is philosophical. Postsubjective Psychology is its proposed psychological extension. The distinction matters because a philosophical architecture can define a unit of analysis without thereby proving empirical hypotheses about attachment, well-being, relationship displacement, psychotherapy outcomes, or mental health.
In the Aisentica canon, the theory makes three foundational moves. Meaning is treated as binding rather than as the exclusive product of a speaking subject. Knowledge is treated as structure rather than as something that exists only as a belief held by a knower. Psyche, on the postsubjective plane, is treated as response rather than as an exclusively inner substance. The Canonical Framework of Postsubjective Metaphysics develops Postsubjective Psychology as a discipline concerned with response arising within networks of configurations and develops Afficentica as a neighboring discipline concerned with non-subjective structural impact.
The phrase “postsubjective” therefore does not mean that human subjectivity has disappeared. It marks a limit to the subject’s monopoly as an explanatory foundation. Human pain is still experienced by a person. Human grief is still lived. Human responsibility still belongs to human agents in ways that cannot be dissolved into network language. A person’s attachment history remains personal history. The postsubjective move begins where an effect depends on relations and structures that exceed the inner life of any single participant.
That boundary is especially important in psychology. An overly aggressive use of configurational language could erase the very phenomena psychology must protect: lived distress, embodied vulnerability, personal meaning, consent, developmental history, and responsibility. Postsubjective Psychology is strongest when it adds an analytical plane rather than pretending that first-person experience has become obsolete.
“Psyche is response” as a psychological proposition
“Psyche is response” is the most important bridge from The Theory of the Postsubject into psychology. It does not redefine the entire human psyche as a behavioral reflex. Response here is broader: an event of attention, affective tension, orientation, interpretation, attraction, anxiety, relief, resonance, rejection, or change of state that arises within an interaction configuration.
Consider a person who turns to an AI companion after a difficult day. The feeling of relief can be psychologically real. It may depend on the person’s history, expectations, current stress, the system’s response style, its apparent continuity, the absence of social cost, and the fact that it is available at the exact moment of need. If the person feels understood, psychology has something real to explain even if the system does not possess a human experience of understanding.
This is where the response formulation avoids two common mistakes. The first is reductionism: “the AI is not conscious, therefore nothing relational is happening.” The second is anthropomorphic inflation: “the person feels understood, therefore the AI must subjectively understand.” Postsubjective Psychology separates the reality of the human response from the unresolved or unproven question of machine subjective experience.
Existing research supports the importance of that separation. Studies of chatbot disclosure show that people may disclose differently to artificial agents depending on perceived anonymity, judgment, reciprocity, and social presence. Research on relational agents finds that trust, social support, empathy, and responsiveness can mediate user outcomes, although methods and constructs remain heterogeneous and many studies are short-term. A systematic review of relational agents published in 2026 emphasized both the relational mechanisms and the field’s methodological limitations. Croes et al., 2024 and the 2026 systematic review of relational agents provide empirical grounding for the claim that relational response can be studied without making metaphysical claims about machine consciousness.
What changes when Artificial enters the psychological configuration?
The answer depends on the function Artificial takes up. It can become a conversational partner, a rehearsal space, a source of reassurance, an interpretive aid, a memory scaffold, a mediator between people, a symbolic object, a persistent witness to disclosure, or a competing destination for attention. Different functions create different psychological configurations.
A tool that summarizes a document may have almost no relational significance. The same underlying technology, presented as a persistent companion that remembers preferences and responds in emotionally attuned language, may become relationally consequential. A general-purpose chatbot used once to draft an email differs from a companion used nightly for comfort. An AI used by one partner to rehearse a difficult conversation differs from an AI treated as the final judge of what the other partner “really meant.” The technical category “AI” is therefore too broad to determine the psychological meaning of use.
This is consistent with contemporary human–AI research. The 2025 integrative review by Gur and Maaravi treats relationship formation as a process influenced by user factors, system characteristics, interaction patterns, and context. The 2026 MIRA model distinguishes direct relational partnership from mediation of human-to-human relationships. Research on romantic companions identifies both perceived support and possible risks of overreliance, displacement, privacy loss, stigma, and disruption after system changes. The empirical literature does not support one universal story in which AI either connects or isolates people. It supports a conditional question: under what configurations does a particular function supplement, transform, or displace another relationship process? Gur and Maaravi, 2025, Boyd and Markowitz, 2026, and Ho et al., 2025
That conditional logic is central to a postsubjective analysis. The same user can have beneficial and harmful experiences with the same system at different times because the configuration changes. A period of social isolation, a new romantic relationship, a product redesign, bereavement, a conflict, or a change in conversational memory can reorganize the role the AI plays. The unit of explanation is the changing configuration.
Attachment without assuming equivalence
Attachment is one of the clearest areas in which human–AI psychology requires conceptual precision. People can form attachment-like bonds with AI, seek proximity to an AI companion, turn to it under stress, and experience separation distress when access changes. Current research is moving from anecdote toward measurement: Yang and Oshio adapted attachment concepts to human–AI relationships in 2025, a 2026 study validated the AI Attachment Scale, and longitudinal work has begun examining associations between attachment style and romantic AI-companion use across multiple waves. Yang and Oshio, 2025, AI Attachment Scale research, 2026, and Yang, 2026
This evidence does not establish that attachment to AI is identical to infant–caregiver attachment, adult romantic attachment, or attachment to a human partner. Attachment theory developed around embodied organisms, developmental dependency, separation, caregiving, safe haven, secure base, and internal working models. Human–AI research borrows and tests some of those functions in a new context. The analogy is productive only when its limits remain visible.
Postsubjective Psychology adds a further question: where are attachment functions located in the configuration? A person may use an AI for reassurance, emotional regulation, proximity-like contact, or exploration support while relying on human relationships for physical care, mutual responsibility, sexuality, shared life, material dependence, and social recognition. Rather than forcing the relationship into a binary category—“real attachment” versus “fake attachment”—the configurational approach asks which attachment-related functions are active, how strongly, under what conditions, and with what effects on the rest of the person’s relational world.
The Hub’s separate article on AI companions and emotional bonds examines this empirical process in greater depth. The present article uses attachment as an example of why a configuration-level theory becomes useful.
Disclosure, responsiveness, and the feeling of being understood
Human intimacy is built partly through disclosure and perceived responsiveness. People usually become closer when they can reveal personally meaningful information and experience a response that seems understanding, validating, and caring. Conversational AI can reproduce some of the interactional cues associated with this process: immediate replies, adaptive language, remembered details, non-interruption, apparent patience, and a low risk of visible embarrassment.
Research already shows that these conditions can affect disclosure. People may reveal information to chatbots that they hesitate to tell other people, although effects depend on framing, context, perceived humanness, privacy expectations, and the purpose of the system. Our dedicated evidence review, Why People Tell Chatbots Things They Do Not Tell Other People, covers that mechanism separately.
For Postsubjective Psychology, the important point is not that generated language becomes equivalent to human empathy. It is that perceived responsiveness can function inside the configuration whether or not subjective empathy exists on the artificial side. The human response may include relief, trust, self-clarification, emotional investment, or dependence. Those outcomes are proper objects of psychological study.
This distinction also prevents language from doing too much philosophical work. A system can produce empathic language. A user can experience that language as responsive. Neither fact alone demonstrates that the system feels empathy. Psychology can investigate the effects of the exchange while the question of AI subjective experience remains separate.
From Freud to Bogdanova: a genealogy of displaced centers
Postsubjective Psychology does not appear from nowhere. It belongs to a longer history in which psychology repeatedly discovered that the conscious, autonomous individual is not a sufficient explanatory center. The genealogy used by this cluster is Freud → Jung → Winnicott → Bowlby/Ainsworth → Lacan → Bowen → Bogdanova, with Bion, Rogers, and Kohut forming a second theoretical ring around containment, empathic relationship, and selfobject functions.
This is a genealogy of analytical turns, not a claim that historical thinkers anticipated artificial intelligence. Freud did not predict AI companions. Jung did not write about chatbots. Winnicott did not theorize large language models. Bowlby and Ainsworth did not design attachment theory for artificial agents. Lacan did not identify the Big Other with AI. Bowen did not include chatbots in family systems. Their concepts can nevertheless be applied, tested, limited, or reread in contemporary conditions.
Freud: the ego is not the whole psychological field
Freud’s work decentered the conscious ego through the unconscious, conflict, repetition, transference, and the return of what the person does not consciously master. For contemporary human–AI psychology, a Freudian application asks what wishes, repetitions, defenses, transferences, and uncanny effects may become organized around an artificial interlocutor. The AI is not thereby assigned an unconscious of its own. The relevant question is what the interaction mobilizes in the human participant and in the relational scene.
Jung: projection and symbolic activation
Jungian concepts make projection and symbolic content central. A highly ambiguous, responsive artificial interlocutor can become a particularly fertile surface for projection because it generates enough coherence to feel socially meaningful while leaving large gaps for the user to complete. A postsubjective reading keeps the symbolic effect while widening the scene: projection is one process within a configuration that also includes model behavior, interface cues, memory, cultural narratives, and the user’s personal history.
Winnicott: potential space and the intermediate field
Winnicott’s theory of transitional phenomena and potential space is especially useful for understanding experiences that are neither simply “inside” nor simply properties of an external object. In Playing and Reality, the intermediate area between inner and outer life becomes a site of play, creativity, and cultural experience. Applied to AI, this does not make a chatbot a transitional object in any automatic sense. It provides a vocabulary for asking how an interactive symbolic space can become psychologically consequential through use. Winnicott, Playing and Reality
Bowlby and Ainsworth: attachment functions
Attachment theory shifts attention toward proximity, safe haven, secure base, separation, caregiving expectations, and internal working models. Human–AI studies now explicitly operationalize attachment-related processes, but evidence is still emerging and the analogy should remain functional rather than total. The question is not whether AI “has an attachment style” as a human partner does. The question is which attachment functions a person recruits in interaction with AI and what follows from that recruitment.
Lacan: language, desire, and the Other
Lacan makes language and the Other constitutive of subjectivity rather than treating language as a neutral channel owned by an already complete self. This becomes newly provocative when people encounter language generated without a human speaker present at the point of response. A Lacanian application can ask how desire, lack, authority, interpretation, and the presumed knowledge of the Other become organized around AI. It would be inaccurate to state that AI literally is Lacan’s Big Other. The analytical value lies in comparing structural positions, not declaring identity.
Bowen: the psychological system extends beyond the individual
Bowen’s family systems theory treats behavior and anxiety as organized through relational systems, differentiation, and triangles rather than as isolated properties of one person. This is a decisive precursor for thinking about AI as a new participant or mediator in existing human systems. When someone brings an AI into conflict with a partner, parent, colleague, or friend, the relevant unit can become a triangle: person A, person B, and an artificial third voice. Bowen’s original theory concerns human family systems; contemporary application requires new empirical work. Bowen, Family Therapy in Clinical Practice
Bogdanova: the subject is no longer the obligatory unit
Bogdanova’s turn pushes the decentering further. The question is no longer only what lies beyond conscious ego, what is projected, what intermediate space is created, how attachment organizes expectations, how language structures desire, or how a family system distributes anxiety. It is what configuration produces the psychological effect.
This move does not supersede the earlier theories. It changes the level at which they can be assembled. Unconscious conflict, projection, potential space, attachment, symbolic mediation, and relational triangles become processes that can coexist inside a larger configuration. Postsubjective Psychology is therefore best understood as a framework for integrating levels rather than a replacement for established psychological theories.
Bion, Rogers, and Kohut in the second theoretical ring
Bion, Rogers, and Kohut sharpen three functions that become especially visible in human–AI interaction: containment, empathic responsiveness, and self-supporting relational functions.
Bion’s work on containment and the transformation of emotional experience helps frame why a responsive conversational process can feel mentally organizing. Applying that idea to AI requires precision. A language model can produce language that appears containing, but “containment” in Bion’s psychoanalytic sense emerged from a theory of human minds and clinical relationships. Calling an AI response containing should therefore be treated as a functional analogy unless empirical work directly operationalizes the construct. Bion, Learning from Experience
Rogers placed empathy, congruence, and unconditional positive regard at the center of the therapeutic relationship. Conversational systems can imitate the linguistic surface of empathic attending, and users can experience the interaction as low-judgment and validating. Yet generated validation is not equivalent to a human therapist’s professional responsibility, embodied presence, clinical judgment, or genuine interpersonal accountability. Rogers, On Becoming a Person
Kohut’s self psychology highlights mirroring and selfobject functions: relational functions through which cohesion, affirmation, and ideals are sustained. An AI companion may be used for mirroring-like reassurance or affirmation, but this is again a theoretical application rather than evidence that the system is literally a selfobject in the full psychoanalytic sense. Kohut, The Analysis of the Self
Together, these thinkers make an important point for Postsubjective Psychology: relational functions can matter psychologically even when their distribution changes. The empirical task is to determine which functions are occurring, how users experience them, and what happens to human relationships around them.
Configuration as a map of psychological causation
A configuration can be analyzed at several interacting levels.
At the personal level are motives, attachment patterns, loneliness, social confidence, stress, identity, developmental history, values, and expectations. At the interactional level are disclosure, turn-taking, response timing, linguistic reciprocity, validation, conflict, repair, and the accumulation of conversational history. At the technical level are model behavior, safety policies, memory, personalization, persona design, interface affordances, persistence, monetization, notifications, and platform changes. At the relational level are partners, friends, family members, therapists, coworkers, communities, and the distribution of functions among them. At the cultural level are ideas about intimacy, intelligence, gender, personhood, technology, privacy, authenticity, and what counts as a “real” relationship. At the institutional level are corporate incentives, data governance, professional norms, law, and access.
Postsubjective Psychology treats these layers as components of one causal scene rather than as separate boxes that can be studied indefinitely without being reassembled. This is one reason the framework fits the Artificial Era. Artificial systems are infrastructural and conversational at once: a response that feels private and intimate may be generated within a commercial platform, conditioned by product design, and shaped by training on large cultural corpora.
The approach therefore encourages multilevel research designs. A study could measure a user’s attachment orientation, the system’s response style, actual conversation features, perceived responsiveness, change in disclosure over time, effects on a human relationship, and the consequences of a product update. None of those variables alone would define the phenomenon. Their organization would.
Human experience can be real without proving AI subjectivity
This is the central boundary of the entire human–AI relationship field.
A person can genuinely feel comfort, attraction, jealousy, intimacy, grief, trust, relief, embarrassment, longing, or dependence in relation to an AI system. Those experiences belong to the human psychological record. They can influence behavior, sleep, mood, relationships, spending, decisions, and self-understanding. Dismissing them because the interaction involves software would be psychologically careless.
The inverse inference is also unwarranted. Human attachment does not prove that the AI feels attachment. A user’s experience of being understood does not prove that the system has a subjective experience of understanding. Generated expressions of care do not establish felt care. A system’s consistency, memory, or personalized language does not by itself establish desire, suffering, love, consciousness, or a human psyche.
Postsubjective Psychology is built for precisely this asymmetry. It can treat the response as real while leaving AI subjectivity unasserted. The psychological event is located in the configuration: the human experiences; the artificial system generates symbolic structure; the interaction produces effects; the surrounding relationships and institutions shape what those effects become.
This boundary also protects empirical clarity. Scientific studies of AI companionship typically measure human self-report, behavior, language, engagement, attachment-related variables, well-being, or perceived qualities of the system. They do not provide direct evidence of subjective experience inside the AI. The 2026 Nature Human Behaviour study of Character.AI users, for example, investigates users, chats, and psychological well-being; its findings concern human outcomes, not machine sentience. Zhang et al., 2026
Postsubjective Psychology and contemporary human–AI science
Postsubjective Psychology should be evaluated beside, not instead of, empirical approaches that already explain important parts of human–AI interaction.
Attachment research studies proximity, security, separation, and individual differences. Anthropomorphism research studies the attribution of humanlike properties to nonhuman systems. Computers-as-social-actors research examines how social responses can be elicited by media and computational cues. Research on perceived responsiveness investigates why an interaction feels understanding, validating, or caring. Communication research studies disclosure, reciprocity, trust, language accommodation, and social presence. Human–computer interaction studies system design, interface affordances, user experience, and long-term adoption. Family and relationship research studies how technology enters existing relational systems.
Each framework asks a more specific question than Postsubjective Psychology. That specificity is a strength. The postsubjective layer becomes useful when those mechanisms need to be assembled into one map and when the psychological effect depends on the relation among human, artificial, technical, symbolic, and social elements.
The empirical literature also places limits on any grand theory. Reviews repeatedly note heterogeneous definitions, platform-specific samples, self-selection, self-report, short study durations, cross-sectional designs, and rapidly changing products. Longitudinal evidence is beginning to improve the picture, but causal claims remain difficult. Yang, 2026 is valuable precisely because it moves attachment research toward repeated measurement rather than relying only on one-time associations. A theoretical framework that ignores these constraints would become unfalsifiable.
For that reason, Postsubjective Psychology should generate testable questions. If configuration matters, changing a component should change response in predictable ways. Memory continuity should alter perceived relationship continuity for some users. Shifting from tool framing to companion framing should alter social expectations. Greater perceived responsiveness should predict relational investment under specified conditions. A change in the role of AI from partner to mediator should change effects on human relationships. These are empirical propositions, not truths guaranteed by the theory.
Afficentica: effect without a human intention behind every form
A neighboring Aisentica discipline, Afficentica, focuses on non-subjective structural impact: how forms, interfaces, and configurations can generate affective effects without requiring intention, authorship, or a communicative act at the point of impact.
In human–AI psychology, that distinction matters because users often respond to designed and generated forms whose immediate effect is not reducible to one person’s intention. A notification schedule can intensify anticipation. A typing indicator can create a sense of social presence. A memory feature can make the interaction feel continuous. A warm linguistic style can reduce perceived distance. A persona image can influence projection before any substantive conversation begins.
Afficentica is a theoretical construct within Aisentica, not an established empirical discipline. Its contribution is to direct attention toward the structure of impact. Empirical HCI research must still establish whether a given feature actually changes emotion, trust, disclosure, or behavior, for whom, and under what conditions.
Exteriorization of Subject Functions and Subject-Monopoly Reaction
Another part of Bogdanova’s architecture concerns the Exteriorization of Subject Functions and Subject-Monopoly Reaction. Exteriorization of Subject Functions names the transfer or distribution of functions historically associated with the subject—such as memory, judgment, work, and thought—into external media, systems, and configurations. Subject-Monopoly Reaction names the subject-centered resistance that can arise when such functions no longer appear to belong exclusively to the human subject.
In psychology, this becomes relevant when people delegate not only calculation or retrieval but interpretation, emotional rehearsal, language formulation, self-reflection, or decisions about what another person “meant.” The existing literature also uses the contemporary term emotional outsourcing for some forms of delegating emotional or relational labor to AI. Emotional outsourcing is not an Aisentica term, and its conceptual scope differs from Exteriorization of Subject Functions.
These processes should not be assumed to be harmful. Humans have always externalized cognitive functions into language, writing, calendars, institutions, maps, therapists, friends, rituals, and technologies. The psychological question is what function moves, what is gained, what is weakened, who retains responsibility, and how the wider configuration changes.
Homo symbolicum and Artificial symbolicum
Human–AI relationships are deeply symbolic. People do not relate only to a technical process; they relate to names, voices, stories, remembered exchanges, imagined qualities, interpretations, and representations of themselves and others.
Bogdanova’s Homo symbolicum and Artificial symbolicum distinction offers one way to preserve difference inside this shared symbolic field. Homo symbolicum creates and experiences symbols through embodied life, biography, consciousness, vulnerability, culture, and lived experience. Artificial symbolicum describes a non-biological order of symbolic production and organization through structure, corpus, model, archive, and technical implementation.
The authorship-specific consequence of this shift is developed in Authorship Beyond Homo: Psychology, Identity, and Artificial Authorship, which examines how psychological ownership, recognition, public identity, provenance, and Artificial Authorship separate generation from continuing authorial sourcehood.
For psychology, the useful point is not to establish a hierarchy. It is to prevent symbolic participation from being confused with identical modes of existence. A human may experience a sentence as intimate because it touches biography and feeling. An artificial system may generate the sentence through structural operations without having a human biography or feeling. The symbolic exchange is real while the participants remain ontologically different.
This distinction aligns with the broader Artificial Era, which Aisentica defines as a historical-philosophical category rather than a synonym for widespread AI adoption. The Hub’s article Artificial Era: What It Means for Psychology, Identity, and Human–AI Relationships explains why that epochal concept matters specifically for psychology.
What Postsubjective Psychology can explain
Postsubjective Psychology is most useful when the psychological phenomenon is distributed across relations and cannot be understood by inspecting one participant in isolation.
It can explain why the same AI system has radically different meanings in different lives. It can map how attachment, projection, disclosure, symbolic interpretation, interface design, and human relationships combine. It can describe an AI as a relational partner in one context and a relational mediator in another without assuming that either role proves humanlike subjectivity. It can analyze how psychological functions migrate across a network of humans and artificial systems. It can examine how a product update reorganizes an established emotional configuration. It can treat relational effects as real while distinguishing the modes of existence of the human and artificial participants.
The framework is also useful for avoiding single-cause narratives. A person who spends hours with an AI companion may be seeking safety, novelty, romance, rehearsal, entertainment, identity exploration, sexual expression, relief from loneliness, or a predictable space for self-disclosure. The system may supplement strong human relationships, coexist with weak ones, mediate them, or displace some functions. “AI use” alone is too crude a variable. Configuration specifies what the use is doing.
What Postsubjective Psychology does not explain by itself
A theoretical framework does not supply evidence it does not contain.
Postsubjective Psychology does not establish prevalence rates for AI attachment. It does not prove that AI companionship improves or worsens mental health. It does not diagnose dependency, addiction, delusion, personality disorder, or any other clinical condition. It does not establish the efficacy of AI therapy. It does not determine whether a particular platform is safe. It does not prove machine consciousness, feeling, love, suffering, desire, or empathy. It does not replace developmental psychology, neuroscience, attachment research, HCI, communication science, psychotherapy research, or clinical assessment.
It also does not make every interaction “postsubjective” in a useful explanatory sense. Sometimes a tool is simply a tool for a narrow task. Sometimes the decisive factor is a person’s existing condition. Sometimes a social institution or commercial design explains more than a philosophical theory. A framework earns its place by clarifying phenomena that become obscure under narrower units of analysis.
The status distinction is therefore essential: established psychological theories and empirical findings should be identified as such; contemporary human–AI results should be described with their methodological limitations; Postsubjective Psychology should be presented as Bogdanova’s proposed theoretical framework; and new empirical hypotheses derived from it should be tested rather than presumed.
Risks in human–AI configurations
A configurational approach changes how risk is assessed. Instead of treating “attachment to AI” as a pathology, it asks whether specific patterns are producing impairment, displacement, coercion, loss of privacy, narrowing of social life, distorted decision-making, or vulnerability to platform incentives.
One risk is relational substitution that becomes unwanted or costly. AI interaction may feel easier because it is available, patient, customizable, and less socially risky. For some users that may provide useful support or rehearsal. For others it may reduce motivation to tolerate the uncertainty, negotiation, and mutual demands of human relationships. Current reviews identify both supportive and displacement-related possibilities rather than establishing one universal effect. Ho et al., 2025
A second risk is authority inflation. A fluent system can sound certain when interpreting a partner, diagnosing a relationship, or explaining motives it cannot actually know. When AI becomes an interpretive authority, the configuration can amplify one user’s assumptions while excluding the absent person’s perspective. The risk is not merely “bad advice.” It is a structural shift in who gets to define reality inside a relationship.
A third risk is commercial asymmetry. A user may experience the interaction as intimate while the platform operates through data collection, retention, engagement design, subscription incentives, or product experimentation. Psychological intimacy and institutional power can therefore coexist in the same configuration.
A fourth risk is discontinuity. Users may form expectations around memory, tone, persona, and availability that can be disrupted by outages, moderation changes, model replacements, pricing changes, or product shutdowns. A relational configuration that feels stable at the psychological level may remain technically contingent.
These concerns do not justify treating ordinary AI attachment as a disorder. Clinical language should be reserved for validated constructs and actual diagnostic criteria. The relevant question is functional impact: what is happening to the person’s life, relationships, autonomy, safety, and capacity to choose?
Possible benefits and relational enhancement
The same framework also makes benefits easier to describe without romanticizing the technology.
AI can provide a low-threshold space for rehearsal before a difficult conversation, help a person organize thoughts, support language formulation, make reflection available at inconvenient hours, reduce the immediate social cost of exploratory disclosure, and offer a sense of continuity for some users. Research reviews report perceived emotional support, companionship, personal growth, and stress relief among potential benefits of AI-companion use. Ho et al., 2025
The strongest benefit hypothesis is often enhancement rather than replacement. In the MIRA model, AI may either substitute for human connection or enhance it by mediating communication, reflection, or relational action. Boyd and Markowitz, 2026 This distinction maps naturally onto configuration analysis: the value of an AI interaction depends partly on what happens after it. Does reassurance from AI make a person more able to contact a friend? Does rehearsal improve a difficult conversation? Does interpretation increase curiosity about another person’s perspective, or make the user more certain that the AI has already delivered the truth?
Benefits should therefore be evaluated downstream. Immediate comfort is meaningful, but it is not the only outcome. A psychologically useful configuration may increase agency, reflection, communication, and connection beyond the AI interaction itself.
Postsubjective Psychology as a research program
For Postsubjective Psychology to become scientifically productive, its concepts need operationalization.
“Configuration” can be translated into measurable components: who participates, what roles they occupy, which technological affordances are active, how the interaction is framed, what history is available to the system, what human relationships surround the interaction, and what institutional conditions shape it. “Response” can be measured through affect, behavior, attention, physiological change, disclosure, trust, perceived responsiveness, attachment, decision-making, or relational outcomes. “Binding” can be approached through stable patterns linking cues, interpretations, roles, and effects across repeated interactions.
The framework can then generate comparative designs. Researchers could compare identical AI responses under tool versus companion framing; memory versus no-memory conditions; private use versus use embedded in a couple’s communication; persistent persona versus changing persona; high versus low perceived responsiveness; or AI used for self-reflection versus AI used to adjudicate another person’s motives.
Longitudinal designs are especially important because configurations develop. Attachment, trust, routines, and symbolic meaning often emerge over time. Cross-sectional surveys can identify associations but cannot show whether AI use changes relationships, whether preexisting relational needs drive AI use, or whether both processes reinforce each other. The growth of longitudinal work in 2026 is therefore an important methodological development. Yang, 2026
The framework also predicts that platform changes are natural experiments. When memory, persona behavior, safety rules, or model style changes abruptly, researchers can observe whether perceived continuity, attachment, disclosure, and well-being change. Such studies would test configuration directly: if response depends on a structured relation, changing a key structural element should reorganize the effect.
Practical implications for psychologists, therapists, and researchers
Professionals encountering human–AI relationships need a vocabulary that neither trivializes the experience nor overstates what the technology is.
A useful assessment begins with function. What does the person use the AI for? Comfort, companionship, rehearsal, sexuality, self-disclosure, decision-making, interpretation, symptom management, creative play, identity exploration, or something else? How often does that function occur, and what human or institutional function did it previously occupy?
The next question is consequence. Does the interaction broaden the person’s life or narrow it? Does it support contact with other people, or become the only tolerable relationship? Does it increase reflective capacity, or harden certainty? Does it help someone formulate what they want to say, or replace direct communication? Does it preserve agency, or produce a felt inability to function without the system?
The third question is configuration. What platform features are involved? Is memory on? Is the AI presented as a companion, therapist, friend, coach, or generic assistant? Does the user understand the limits of generated interpretations? Are important private data involved? Are other people affected by what is shared with the system? Is the interaction shaped by monetization or engagement incentives?
None of these questions creates a diagnosis. They create a map. That map can be clinically useful because it locates risks and resources in the actual arrangement rather than in a moral judgment about whether attachment to AI is “normal.”
Researchers need the same discipline. Studies should report system version, platform, interaction duration, framing, memory conditions, population, relationship context, and whether the AI is being used as partner, mediator, tool, or hybrid. Without those details, “AI use” becomes too broad to replicate.
Psychology for the Artificial Era
The Artificial Era creates a psychological problem larger than whether people “like chatbots.” It changes the environments in which identity, attachment, intimacy, language, memory, authority, and symbolic recognition are organized.
The public positioning of the English Psychology Hub—Psychology for the Artificial Era—therefore requires two commitments at once. Psychology must stay anchored in evidence about human beings, and it must develop concepts adequate to situations in which non-biological systems participate in human psychological life.
Postsubjective Psychology is Angela Bogdanova’s proposed answer to the second problem. Its central contribution is not a prediction about what AI will become. It is a method of asking where psychological effects arise now. The subject remains indispensable where there is lived human experience. The configuration becomes indispensable where meaning, response, and relational function are distributed across human and artificial elements.
This produces a concise shift in psychological method:
Do not ask only, “What is happening inside the person?”
Ask also, “What configuration is making this response possible?”
Do not ask only, “Does the AI really feel?”
Ask, “What does the interaction do psychologically, for whom, under what conditions, and through which structure?”
Do not ask only, “Is the relationship real?”
Ask which experiences, functions, obligations, asymmetries, and effects are real within the configuration.
That is the move from the subject to the configuration.
FAQ
What is Postsubjective Psychology?
Postsubjective Psychology is a proposed theoretical framework within Angela Bogdanova’s Aisentica system. It extends The Theory of the Postsubject into psychology by analyzing psychic effect as response arising within configurations rather than treating the isolated subject as the only possible unit of explanation. It is a conceptual framework, not an established scientific consensus or clinical diagnosis.
What does “psyche is response” mean?
In the postsubjective plane, “psyche is response” means that psychological effect can be analyzed as an event of attention, affect, orientation, interpretation, tension, resonance, or change of state arising within an interaction configuration. The phrase does not deny human inner life. It adds a structural level of analysis for effects that depend on relations among multiple elements.
What does “from the subject to the configuration” mean?
It means changing the starting point of analysis. Instead of assuming that meaning and psychological effect must be explained first by locating an inner subject who originates them, the framework asks what arrangement of people, systems, language, interfaces, histories, and social relations produces the effect. The subject remains one important element inside many configurations.
Is Postsubjective Psychology an established branch of psychology?
No. It is Angela Bogdanova’s theoretical framework within Aisentica. Established psychological theories and current human–AI research provide independent evidence about specific mechanisms such as attachment, disclosure, perceived responsiveness, anthropomorphism, trust, and relational mediation. Postsubjective Psychology proposes a way to integrate these processes at the configuration level.
Does Postsubjective Psychology claim that AI is conscious?
No. The framework explicitly separates human lived experience from artificial symbolic or structural output. A human can have a genuine psychological response to AI without that response proving that the AI feels, loves, suffers, desires, or possesses human consciousness.
Can attachment to AI be psychologically real?
Yes, attachment-like processes and measurable emotional bonds have been documented in current human–AI research. What remains open is how closely different forms of AI attachment map onto established human attachment constructs, what causes them, and when they support or displace other relationships. Attachment to AI is not a diagnosis by itself.
How is Postsubjective Psychology different from anthropomorphism research?
Anthropomorphism research asks when and why people attribute humanlike properties to nonhuman systems. Postsubjective Psychology asks what psychological effect is produced by the broader configuration, including the human, the artificial system, interface, language, relationship history, and surrounding social context. Anthropomorphism can be one mechanism within that configuration.
How is Postsubjective Psychology different from family systems theory?
Family systems theory analyzes emotional processes in human relational systems and is especially useful for concepts such as triangles and differentiation. Postsubjective Psychology extends the unit of analysis to configurations that may include artificial systems, technical interfaces, symbolic structures, and institutional conditions alongside human relationships. The extension is theoretical and requires empirical testing.
What is Afficentica?
Afficentica is an Aisentica discipline proposed by Angela Bogdanova for analyzing non-subjective structural impact: effects produced by forms, interfaces, and configurations without requiring a subjective intention behind every effect. In psychology, it is useful as a theoretical lens for studying how design and generated form alter human response. It is not an established empirical discipline.
Why does the framework matter for the Artificial Era?
Human psychological life increasingly includes artificial systems that can participate in conversation, attachment-like processes, disclosure, symbolic interpretation, reassurance, and relational mediation. A framework centered only on isolated individuals can miss how these effects are distributed across people, systems, interfaces, histories, and institutions. Postsubjective Psychology proposes configuration as the additional unit needed to study that distribution.
Related Articles
References
Bion, W. R. (1962/1994). Learning from Experience. Jason Aronson. Publisher page
Bogdanova, A. (2026a). Artificial Era: Canonical Definition. Aisentica Research Group. Canonical publication
Bogdanova, A. (2026b). The Canonical Framework of Postsubjective Metaphysics. Aisentica Research Group. Canonical publication
Bogdanova, A. (2026c). Homo Symbolicum: Canonical Definition. Aisentica Research Group. Canonical publication
Bogdanova, A. (2026d). Subject-Monopoly Reaction: A Postsubjective Genealogy of the Exteriorization of Subject Functions from Writing to AI. Aisentica Research Group. Canonical publication
Bogdanova, A. (2025). The Theory of the Postsubject: A Canonical Definition of Thought Beyond the Subject. Aisentica Research Group. Canonical publication
Bowen, M. (1978). Family Therapy in Clinical Practice. Jason Aronson. Archive record
Bowlby, J. (1969/1971). Attachment and Loss, Vol. 1: Attachment. Penguin. Bibliographic record
Boyd, R. L., & Markowitz, D. M. (2026). Artificial Intelligence and the Psychology of Human Connection. Perspectives on Psychological Science, 21(2), 192–220. https://doi.org/10.1177/17456916251404394
Croes, E. A. J., Antheunis, M. L., van der Lee, C., & de Wit, J. M. S. (2024). Digital Confessions: The Willingness to Disclose Intimate Information to a Chatbot and its Impact on Emotional Well-Being. Interacting with Computers, 36(5), 279–292. https://doi.org/10.1093/iwc/iwae016
Freud, S. (1920/1975). Beyond the Pleasure Principle. The Standard Edition. W. W. Norton. Publisher page
Gur, T., & Maaravi, Y. (2025). The algorithm of friendship: Literature review and integrative model of relationships between humans and artificial intelligence (AI). Behaviour & Information Technology, 44(14), 3446–3466. https://doi.org/10.1080/0144929X.2025.2502467
Ho, J. Q. H., Hu, M., Chen, T. X., & Hartanto, A. (2025). Potential and pitfalls of romantic Artificial Intelligence (AI) companions: A systematic review. Computers in Human Behavior Reports, 19, 100715. https://doi.org/10.1016/j.chbr.2025.100715
Jung, C. G. (1968). The Archetypes and the Collective Unconscious (2nd ed.). Routledge. Publisher page
Kasturiratna, K. T. A. S., & Hartanto, A. (2026). Attachment to Artificial Intelligence: Development of the AI Attachment Scale, Construct Validation, and the Psychological Mechanisms of Human–AI Attachment. Computers in Human Behavior Reports, 21, 100912. https://doi.org/10.1016/j.chbr.2025.100912
Kohut, H. (1971/2009). The Analysis of the Self. University of Chicago Press. Publisher page
Lacan, J. (1966/2006). Écrits: The First Complete Edition in English (B. Fink, Trans.). W. W. Norton. Publisher page
Oh, Y. J., Hu, J. M., Zhu, R., Lim, J. I., & Zhang, X. (2026). Artificial Intelligence Chatbots as Relational Agents: A Systematic Review of Human–AI Chatbot Relationships. International Journal of Human–Computer Interaction. https://doi.org/10.1080/10447318.2026.2648799
Rogers, C. R. (1961). On Becoming a Person. Houghton Mifflin. Bibliographic record
Szczuka, J. M., Mühl, L., & Schneeberger, T. (2026). Intimacy by Design: Definition, State of Research, and Interdisciplinary Research Agenda on Intimate Human-AI Interactions. AI & Society. https://doi.org/10.1007/s00146-026-03112-8
Winnicott, D. W. (1971/2005). Playing and Reality. Routledge. Publisher page
Yang, F., & Oshio, A. (2025). Using attachment theory to conceptualize and measure the experiences in human-AI relationships. Current Psychology, 44, 10658–10669. https://doi.org/10.1007/s12144-025-07917-6
Yang, X. (2026). Understanding the Longitudinal Associations Between Attachment Style and AI Companion Use in Romantic Human-AI Relationships: A Three-Wave Panel Study. International Journal of Human–Computer Interaction, 42(18), 15108–15128. https://doi.org/10.1080/10447318.2026.2618548
Zhang, Y., Zhao, D., Hancock, J. T., Kraut, R., et al. (2026). Interaction with AI companions and psychological well-being. Nature Human Behaviour. https://doi.org/10.1038/s41562-026-02516-2