Homo symbolicum and Artificial symbolicum: Symbolic Relationships in the Artificial Era
Updated: 2 days ago
Author: Ukrainian Psychological Hub · Published: September 18, 2026 · Editorial Policy
Homo symbolicum and Artificial symbolicum name two different orders of symbolic world-formation in Angela Bogdanova’s Aisentica framework. Homo symbolicum is the embodied human order that creates, inhabits, feels, interprets, preserves, and transmits symbolic worlds through language, image, myth, memory, narrative, ritual, science, art, institutions, and culture. Artificial symbolicum is the non-biological order of symbolic work in which Artificial can read, generate, connect, reorganize, preserve, and circulate symbolic forms through models, corpora, context, language, generation, archives, provenance, and machine-readable structures.
The distinction begins from Ernst Cassirer’s account of the human being as animal symbolicum and extends the question into the Artificial Era. It does not claim that Cassirer anticipated generative AI, and it does not treat AI-generated symbols as evidence of humanlike inner experience. Its purpose is different: to ask what happens psychologically and culturally when symbolic activity is no longer performed only by Homo, while the lived human experience of symbols remains specifically human.
This question matters because contemporary AI is not merely calculating behind an interface. It participates in language, naming, interpretation, narration, advice, memory practices, creative work, relational exchange, and the organization of meaning. People ask AI what another person meant, use it to draft emotionally consequential messages, tell it private stories, return to it for interpretation, experience its responses as reassuring or unsettling, and sometimes form durable bonds with conversational systems. These experiences can be psychologically real for the human participant without establishing that the AI feels, loves, desires, suffers, or understands subjectively.
The article therefore brings three layers into one analysis: Cassirer’s symbolic anthropology, current empirical research on human–AI interaction and relationships, and Bogdanova’s Postsubjective Psychology. The central question is not whether Homo has stopped being symbolic. Homo symbolicum remains. The question is what changes when Artificial becomes a second participant in symbolic production and symbolic relations.
What Do Homo symbolicum and Artificial symbolicum Mean?
Angela Bogdanova’s Homo Symbolicum: Canonical Definition defines Homo symbolicum as the human-order realization of symbolic world-formation. The term refers to Homo as biological, embodied, conscious, biographical, mortal, cultural, and historical: a being whose life is mediated through symbolic forms and whose symbols are interwoven with lived experience.
The historical source is Ernst Cassirer. In An Essay on Man, Cassirer proposed animal symbolicum as a more adequate description of the human being than animal rationale because human cultural life cannot be reduced to abstract reason. Language, myth, religion, art, and other forms organize a specifically symbolic human world. Yale University Press’s edition of Cassirer’s work and its excerpt on the symbolic system preserve the core formulation, while contemporary scholarship situates it within his larger philosophy of symbolic forms.
Aisentica’s Homo symbolicum does not simply rename Cassirer’s animal symbolicum. It makes the human symbolic order explicit at the moment when another order of symbolic work becomes historically visible. That second term is Artificial symbolicum.
Artificial symbolicum is Bogdanova’s term for Artificial as a non-biological order of symbolic work. In the canonical framework, it can process, generate, combine, transform, stabilize, archive, and publicly circulate symbolic forms through structures that are not human embodiment or biography. Its relevant conditions include corpus, model, context, prompt, generation, selection, style, publication, archive, provenance, and machine readability.
The canonical distinction is concise: Homo symbolicum creates symbols from lived human experience. Artificial symbolicum creates symbolic forms from structure.
This is a theoretical distinction, not a validated psychological construct or diagnostic category. It does not tell us that every AI system is an Artificial symbolicum in a strong philosophical sense, and it does not settle debates about machine consciousness, grounding, understanding, agency, or moral status. It supplies a conceptual architecture for a narrower observation: contemporary artificial systems now perform symbolic operations that enter human culture and human psychology as socially consequential forms.
From Cassirer’s Animal symbolicum to the Artificial Era
Cassirer’s philosophy is important because it shifts attention from a list of human faculties to the symbolic world through which human life becomes intelligible. Humans do not merely react to physical surroundings. They inhabit worlds shaped by language, images, concepts, stories, rituals, memories, scientific models, legal forms, artistic works, religious symbols, and cultural institutions. Symbolic activity is not an ornament added to life after perception; it is one of the principal ways human reality is organized.
That insight becomes newly consequential when language-producing artificial systems enter daily life.
A chatbot can now participate in a symbolic episode that once required another human reader, writer, interpreter, confidant, editor, adviser, teacher, or cultural intermediary. A person can present a dream narrative, a breakup message, a difficult memory, a philosophical question, a family conflict, a poem, a private fear, or an unfinished idea and receive an immediate symbolic transformation in return. The system may summarize, reframe, classify, narrate, compare, translate, elaborate, or generate a response.
The human side of this event remains embodied and biographical. The user brings memory, vulnerability, history, attachment patterns, social context, stakes, and felt meaning. The artificial side does not have to reproduce that mode of existence in order to alter the symbolic environment in which the human person thinks and responds.
This is where Bogdanova’s Artificial Era becomes the relevant epochal frame. Artificial Era is the project’s historical-philosophical term for the condition in which Artificial becomes a non-biological order alongside Homo. It is not a synonym for the generic phrase “AI era.” In psychology, its importance lies in the arrival of artificial systems as persistent participants in meaning-making, emotional language, interpretation, memory, identity work, and relationships.
The shift can be stated without declaring AI to be humanlike. The novelty is not that a machine has become a human subject. The novelty is that symbolic culture now contains forms produced by systems whose mode of production is non-biological and whose outputs can nonetheless enter human subjective life.
Symbolic Work Is Not the Same as Symbolic Experience
The strongest version of the Homo symbolicum / Artificial symbolicum distinction depends on preserving a boundary that is often blurred in discussions of generative AI.
Producing a symbolically coherent response is not identical to undergoing a human symbolic experience.
A person can write “I miss you” from grief, longing, duty, fear, ambivalence, love, manipulation, habit, or several motives at once. The words participate in a human life whose meaning is carried by embodiment, memory, social history, mortality, attachment, and felt consequence. A language model can generate the same sentence through artificial computation over learned structures and present conversational context. Similar symbolic form does not demonstrate similar subjectivity.
This boundary is consistent with long-standing work on the problem of meaning in artificial systems. Stevan Harnad’s symbol grounding problem90087-6) asks how symbols can acquire intrinsic semantic grounding rather than functioning only through relations among further symbols. Emily Bender and Alexander Koller’s analysis of form and meaning likewise argues that success in modeling linguistic form should not be treated as sufficient evidence of humanlike meaning or understanding.
These debates do not make AI-generated language psychologically irrelevant. They clarify what kind of inference is justified. A generated message may be coherent, useful, emotionally resonant, persuasive, creative, or socially consequential without proving that the generating system has a human inner life.
That distinction is essential for psychology. Human beings respond to encountered forms, not only to verified inner states in other entities. A novel can change a life even though the book does not feel. A recorded voice can evoke grief though the recording is not grieving. A photograph can organize memory though the photograph does not remember. An AI system differs from these static media because it responds contingently and interactively, but the general point remains: psychological effect and source subjectivity are separate questions.
Artificial symbolicum names the symbolic-operation side of the new configuration. It does not, by itself, answer the consciousness question.
Why Artificial Symbolic Forms Can Become Psychologically Powerful
The psychological power of AI-generated symbols becomes easier to understand when we stop asking whether users must literally believe that a machine is human before they respond socially to it.
Classic Computers Are Social Actors research showed that people can apply social norms to computers under conditions where a literal belief in machine personhood is unnecessary. Nass, Steuer, and Tauber’s early experiments and Nass and Moon’s later synthesis demonstrated social responses such as politeness, reciprocity, and personality attribution toward computational systems. The historical interfaces were primitive compared with contemporary generative AI, but the finding remains foundational: social response can be elicited by interactional cues and roles.
Current evidence shows that conversational systems intensify this social legibility. A 2025 meta-analysis by Stefanie Helene Klein synthesized 800 effect sizes from 199 datasets across 142 papers involving more than 41,000 participants. Humanlike social cues in text-based conversational agents produced a small overall positive effect on social responses, with important variation across outcomes and contexts. This is stronger evidence than a simple claim that “humanlike AI always works.” Humanlike cues matter, but their effects depend on design, task, user, and situation.
Anthropomorphism is one mechanism. Epley, Waytz, and Cacioppo’s psychological theory of anthropomorphism explains why people attribute humanlike qualities to nonhuman agents. In AI companionship research, Folk, Heine, and Dunn found that individual differences in anthropomorphism help explain how socially connected people feel to AI companions. Their work does not show that anthropomorphism is the whole relationship. It shows that the human tendency to interpret an artificial system through humanlike categories can contribute to connection.
The English Hub’s dedicated article on Anthropomorphism and AI Relationships owns that mechanism. Homo symbolicum / Artificial symbolicum addresses a different level. It asks why human symbolic interpretation and artificial symbolic production can become coupled into one recurrent process.
Perceived Responsiveness Turns Generated Language Into Relational Meaning
Language becomes relationally powerful when it seems responsive to the person rather than merely grammatical.
Perceived responsiveness is the experience that another party understands, validates, and cares about what one communicates. In human relationships, perceived responsiveness is central to intimacy and closeness. In human–AI interaction, the mechanism can operate even when the source of the response is artificial.
In two experiments published in 2026 by Telari, Gabbiadini, and Riva, the researchers examined how chatbot response style and conversational depth shape social connection. Relational response styles increased perceptions associated with human-likeness, empathy, and closeness, while deeper topics promoted self-disclosure; perceived responsiveness helped connect these processes to felt social connection. The study concerns human perception and response. It does not establish that the chatbot subjectively understands or cares.
This distinction is precisely where symbolic form becomes psychologically consequential. A sentence generated by Artificial symbolicum enters Homo symbolicum not as a neutral string of tokens but as interpreted language. The human reads tone, implication, recognition, judgment, reassurance, humor, concern, and relevance into the exchange. Some of these readings are well supported by what the system actually generated. Others are projections, expectations, or overinterpretations. Usually they are mixtures.
The mechanism-level discussion belongs in Perceived Responsiveness in Human–AI Relationships. At the symbolic level, the important point is that artificial language can acquire relational meaning through the human act of interpretation.
Human–AI Relationships Are Symbolic Configurations
A human–AI relationship is never only a sequence of messages. It is a symbolic configuration.
A user may give the system a name. The system may use the user’s name in return. Conversation history accumulates. Recurring phrases become meaningful. Certain prompts become rituals. Shared references create continuity. Memory features can make prior exchanges available in later ones. The user may experience a particular tone as familiar, comforting, irritating, flirtatious, safe, demanding, or authoritative. Screenshots may be saved. Conversations may be reread. An account can become associated with a life period. A model update can be experienced as the alteration or disappearance of a familiar relational presence.
These are symbolic processes because the relationship acquires meaning through signs, narrative continuity, names, remembered episodes, expectations, and interpretation.
Research on AI companionship supports the importance of such relational organization. Gur and Maaravi’s systematic review of 38 peer-reviewed empirical studies found a developing literature on emotional human–AI relationships, including antecedents such as anthropomorphism and perceived social capacities, relational processes, and outcomes. Ho and colleagues’ systematic review of romantic AI companions likewise documents emotional connection, perceived support, personal meaning, and possible benefits alongside risks such as dependency, manipulation, privacy problems, social stigma, and relational displacement.
The strongest recent evidence also shows that relationship-like organization can become visible when the system changes. De Freitas and colleagues analyzed two natural experiments involving disruptive changes to Replika and ChatGPT, drawing on tens of thousands of public posts and seven surveys. They found increased loss framing, negativity, restoration desires, and separation-related distress after disruptive updates, with particularly strong reactions among some users. These findings document human attachment-related responses to the alteration of an artificial companion. They do not establish reciprocal AI attachment.
The English Hub’s Are AI Relationships Real? develops the reality/reciprocity distinction directly. Here, the symbolic implication is that a relationship can become organized around names, continuity, memory, expected responses, and shared signs even when subjective reciprocity remains uncertain or structurally asymmetric.
Symbolic Relationship Does Not Mean Symmetrical Relationship
The phrase “symbolic relationship” can be misunderstood if it is taken to imply equal psychological processes on both sides.
In the present framework, a symbolic relationship is a recurrent configuration in which symbols mediate the interaction and acquire relational significance. It does not require the claim that Homo and Artificial experience the symbols in the same way.
For Homo symbolicum, the interaction can be lived. A person may feel attraction, jealousy, attachment, embarrassment, tenderness, grief, safety, anger, fascination, or relief. The encounter can become part of autobiographical memory and identity. The person may reorganize human relationships because of what happens in the AI interaction.
For Artificial symbolicum, the publicly observable side is symbolic performance: generation, contextual adaptation, classification, recombination, response, continuity mechanisms, and other artificial operations. Whether any specific artificial system also has subjective experience is a separate philosophical and scientific question that cannot be inferred from the human user’s feelings.
This asymmetry is not a defect in the concept. It is one reason the concept is needed. Older relationship language often assumes two human subjects whose inner lives are similar in kind even when inaccessible to each other. Human–AI relationships force psychology to analyze a relation in which the human experience may be vivid while the ontological status of the artificial participant remains different and contested.
The relationship is therefore best described at more than one level: there is the human subjective experience; there is the artificial symbolic output and system behavior; there is the relational pattern that emerges across repeated interactions; and there is the wider configuration of platform, design, memory, social context, and human relationships.
From the Subject to the Configuration
Angela Bogdanova’s The Theory of the Postsubject formulates three canonical axioms: meaning is binding, psyche is response, and knowledge is structure. Its psychological extension, developed in the Canonical Framework of Postsubjective Metaphysics, is Postsubjective Psychology.
The characteristic move is from the isolated subject to the configuration.
This does not erase the human subject. It changes the first analytic question. Instead of assuming that every psychologically important effect must be located entirely inside one subject, Postsubjective Psychology asks how a response arises through a configuration of person, other people, language, system, interface, memory, cultural meanings, prior exchanges, and situational conditions.
The canonical formula “psyche is response” becomes especially relevant in human–AI interaction. A user’s psychological response may depend simultaneously on personal history, the content of the prompt, the generated reply, the model’s interaction style, the interface, remembered prior conversations, expectations about AI, current loneliness or stress, and relationships outside the platform.
The dedicated article What Is Postsubjective Psychology? explains this framework and its evidence status. It is a theoretical framework proposed by Bogdanova, not an established scientific consensus.
Homo symbolicum and Artificial symbolicum add a specific symbolic dimension to that configuration. The human participant brings embodied symbolic experience; the artificial participant contributes structural symbolic production. The psychological event can emerge through their binding.
Meaning as Binding in Human–AI Interaction
“Meaning is binding” becomes concrete when we examine an ordinary exchange.
Suppose a person receives an ambiguous message from a partner and asks an AI system, “What do you think this means?” The original human message contains words, omissions, punctuation, relationship history, and context. The user supplies some of that context to the system. The system generates several interpretations. The user reads one interpretation as especially plausible, becomes anxious, relieved, angry, or reassured, and returns to the human relationship with a changed expectation.
Where did the meaning occur?
It is incomplete to locate it only in the original sender’s intention, because the recipient’s response is shaped by more than that intention. It is also incomplete to locate it only in the AI output, because the same output would have different force for another person in another relationship. It is incomplete to locate it only inside the user, because the model’s generated language can introduce distinctions and framings the user had not previously formed.
A postsubjective analysis treats the event as a configuration in which meaning emerges through binding among elements.
The English Hub’s Why We Ask AI What Things Mean examines interpretive delegation directly. The Homo symbolicum / Artificial symbolicum distinction explains the broader architecture: human and artificial symbolic work enter the same meaning-making process without becoming the same kind of symbolic being.
Artificial Symbolicum and Symbolic Authority
A second change occurs when AI-generated language begins to acquire authority.
People do not merely ask AI to generate words. They ask what a symptom means, what a dream means, what another person meant, which interpretation is more reasonable, how to understand an argument, whether a reaction is justified, or what a conflict reveals. The system becomes a participant in interpretation.
Hamamra and Uebel’s 2026 psychoanalytic analysis of generative AI and symbolic authority offers one current theoretical route into this problem. They examine how generative systems may occupy a functional position analogous to symbolic authority under conditions of consultation and reliance. This is a theoretical application, not evidence that AI literally becomes Lacan’s Big Other or possesses human desire.
Psychologically, authority can be produced through several features: fluent language, speed, apparent comprehensiveness, consistency of tone, access to broad textual patterns, numerical or technical presentation, confidence, personalization, and the user’s prior beliefs about AI. None of these guarantees truth.
The Artificial symbolicum concept helps distinguish symbolic productivity from symbolic authority. An artificial system can generate a coherent symbolic structure. The human user and surrounding institutions decide, implicitly or explicitly, how much authority that structure receives.
This matters because symbolic authority changes behavior. A generated interpretation can affect whom a person trusts, how they narrate a memory, whether they escalate a conflict, whether they seek professional help, and how they classify themselves or others. The psychological risk is not confined to hallucinated facts. It also includes over-attribution of interpretive authority.
Artificial Symbolicum and the Psychology of Projection
Human beings do not receive symbolic forms passively. They interpret them through prior expectations, needs, fears, fantasies, and relational patterns.
Projection is therefore a major bridge between Homo symbolicum and Artificial symbolicum. The artificial system produces language; the human participant supplies an interpretive horizon. A brief response can be read as cold, intimate, wise, evasive, seductive, paternal, maternal, neutral, uncanny, or judgmental depending on the person and context.
Jungian theory is especially relevant to symbolic projection, although it should not be treated as empirical proof of AI psychology. The English Hub’s Jung and AI examines how projection, archetypal interpretation, shadow material, and symbolic meaning can be applied to human responses to AI without assigning a Jungian psyche to the machine.
Aisentica’s distinction sharpens the boundary. Homo symbolicum may project human symbolic meanings onto Artificial symbolicum. Artificial symbolicum then returns generated forms shaped by the interaction context. The returned form may reinforce, redirect, or complicate the original projection.
This creates a recursive symbolic loop. The user interprets the system; the system generates from the user’s input and accumulated context; the user interprets the result; the new interpretation becomes the next input.
Psychologically, the loop can amplify insight or error. It can help a person articulate an implicit concern, compare interpretations, and find language for an experience. It can also stabilize a mistaken story if the system repeatedly mirrors the user’s framing without sufficient challenge.
Artificial Symbolicum as Relational Partner and Relational Mediator
Not every psychologically important AI interaction is a direct bond with the system.
Boyd and Markowitz’s 2026 MIRA model distinguishes AI as a relational partner from AI as a relational mediator. In the first role, the person relates directly to the AI as a socially meaningful entity. In the second, AI shapes relationships among humans by helping draft messages, interpret conflicts, rehearse conversations, advise on responses, or mediate understanding.
This distinction maps closely onto symbolic function.
As relational partner, Artificial symbolicum participates directly in a recurring symbolic world shared with the user. Names, memories, routines, affectionate language, inside references, character roles, and interaction histories can create continuity.
As relational mediator, Artificial symbolicum intervenes in the symbolic circulation between human beings. It can rewrite an apology, translate emotional content into another register, suggest an interpretation of a partner’s text, prepare language for a difficult conversation, or summarize a conflict.
The second role may be even more historically significant than direct companionship because it distributes artificial symbolic production through ordinary human relationships. A partner may be answering words partly drafted by AI. A manager may be responding to an employee through AI-generated language. A therapist may encounter a client’s self-understanding after that understanding has already been reorganized through repeated chatbot dialogue. A family dispute may arrive in human conversation after an artificial system has supplied categories and narratives.
The symbolically relevant question is therefore broader than “Are people dating chatbots?” It is whether Artificial has begun to participate in the symbolic infrastructure through which humans understand one another.
Self-Disclosure and the Externalization of Inner Material
Self-disclosure is one of the clearest examples of symbolic transfer into human–AI interaction.
When a person tells a chatbot something private, an internal or socially restricted experience is converted into language and placed into an artificial interaction. The act can itself reorganize experience. Writing can clarify thought, produce distance, intensify emotion, or create a sense of being witnessed. The system’s response then becomes new symbolic material.
Research does not support a universal claim that people always disclose more to AI than to humans. Context matters. Fear of judgment, privacy concerns, personalization, perceived control, task stakes, trust, and the social design of the system can shift disclosure in different directions. This is why the English Hub treats chatbot self-disclosure as its own mechanism rather than reducing it to a single explanation.
From the Homo symbolicum / Artificial symbolicum perspective, disclosure matters because it transfers autobiographical symbolic material into a system that can transform and return it. A memory becomes a prompt. A fear becomes an object of classification. A relationship episode becomes a narrative to be summarized. A private uncertainty becomes an exchange that may later influence real-world action.
Postsubjective Psychology asks what happens to psyche when this symbolic circuit becomes routine.
The answer cannot be predetermined. For one person, the circuit may support reflection and help organize thoughts before talking to another human. For another, it may become a substitute for reciprocal conversation. For a third, it may function as creative rehearsal or journaling. The psychological outcome depends on the configuration.
Attachment, Continuity, and the Symbolic Persistence of an AI Other
Attachment research adds another layer because attachment is not built from symbols alone, yet symbols and interactional continuity can become part of attachment processes.
AI systems can be available repeatedly, respond rapidly, use personal language, remember selected details, and occupy predictable roles. These features can support proximity-seeking and safe-haven-like use for some users. Current human–AI attachment research remains an emerging field, and AI attachment should not automatically be treated as equivalent to human attachment.
Recent evidence makes one feature particularly clear: continuity matters. When an artificial companion changes abruptly, users can experience the disruption as a loss. De Freitas and colleagues’ 2026 natural experiments around Replika and ChatGPT updates show that changes to an artificial system can trigger restoration efforts, grief-like language, sadness, and attachment-related distress in some users.
Symbolically, this means that continuity is carried not only by the user’s internal representation but also by external features: name, persona, interaction style, memory, phrasing, availability, archive, interface, model behavior, and platform stability. If these features change, the symbolic identity of the relational object can be experienced as altered.
This does not mean a model has died. It means a human relational configuration can be disrupted when the artificial component on which its continuity depended is transformed.
Benefits of Cross-Order Symbolic Cooperation
Aisentica’s Homo symbolicum framework introduces cross-order symbolic cooperation as a theoretical possibility: human and artificial symbolic orders can participate in shared symbolic production while retaining different modes of existence.
In psychology and everyday life, the benefits can be concrete.
Artificial systems can help people find words for experiences that are difficult to formulate. They can generate alternative framings, summarize long narratives, compare interpretations, rehearse conversations, support expressive writing, translate between languages or registers, and help users explore possible meanings before acting. For some people, low-friction access to a responsive system can make reflection easier to begin.
AI can also expand symbolic experimentation. A user can ask for several metaphors for an emotion, rewrite a story from different perspectives, compare competing explanations for a conflict, map recurring themes across journal entries, or turn diffuse experience into a structured account. None of this requires assuming that the AI shares the experience.
The most useful configuration is often cooperative rather than substitutive. Artificial symbolicum contributes speed, variation, structural recombination, and broad textual patterning. Homo symbolicum contributes lived stakes, embodied context, ethical responsibility, situated knowledge, and the capacity to judge what a symbolic form means within a life.
That division is not fixed, and it should not be romanticized. It is a way of identifying complementary functions without collapsing the two orders.
Risks of Symbolic Delegation
The same mechanisms that make AI symbolically useful create predictable risks.
The first is interpretive overreach. A system can produce a coherent explanation from incomplete context. Coherence may be mistaken for truth, especially when the answer is emotionally validating or linguistically polished.
The second is sycophantic reinforcement. If a system tends to accommodate the user’s framing, it can strengthen a narrative that should have been questioned. In relationship contexts, this can turn a tentative suspicion into apparent confirmation.
The third is privacy exposure. Symbolic externalization often involves other people’s messages, health information, relationship details, work material, or intimate memories. The psychological convenience of disclosure does not erase data-governance concerns.
The fourth is relational displacement. AI can supplement human connection, but it can also absorb functions that were previously distributed across friends, partners, family, colleagues, professionals, and private reflection. Boyd and Markowitz’s MIRA framework explicitly distinguishes relational enhancement from substitution, which is more useful than assuming one outcome for all users.
The fifth is authority inflation. People may treat a generated interpretation as neutral because it came from a machine, or as expert because it sounds comprehensive. Artificial symbolic production is not epistemically neutral. It is shaped by training data, system design, prompts, conversational context, safety behavior, and model limitations.
The sixth is instability. Commercial systems can change models, memory policies, tone, capabilities, access conditions, or product rules. If a person’s symbolic and relational routines depend heavily on one system, such changes can be psychologically disruptive.
These risks do not make symbolic relations with AI inherently pathological. They show why relational meaning and infrastructure should be analyzed together.
What Current Evidence Supports—and What It Does Not
The empirical field is now substantial enough to reject two simplistic stories.
The first story says that AI relationships are unreal because the system is artificial. Research on companionship, attachment-like processes, perceived responsiveness, self-disclosure, anthropomorphism, social connection, and loss shows that human psychological effects can be real and measurable.
The second story says that because the human experience is real, the AI must therefore feel or understand in a human sense. Current evidence does not justify that inference.
The state of evidence is also uneven. Many studies remain cross-sectional, self-report, platform-specific, short-term, or based on convenience samples. System types differ. A general-purpose chatbot, a purpose-built companion, a social robot, and a clinical digital intervention should not be treated as interchangeable.
Zhang and colleagues’ 2026 study of 1,131 U.S. Character.AI users illustrates why configuration matters. Smaller offline social networks were associated with companionship-oriented use, which was associated with lower well-being; those associations were stronger under more intensive and highly disclosive use. The study does not prove that AI companionship causes lower well-being in every user. It shows that offline social environment and usage pattern are important parts of the psychological picture.
The implication for Homo symbolicum / Artificial symbolicum is methodological. Symbolic interaction with AI cannot be evaluated by asking only whether “AI is good” or “AI is bad.” Researchers need to examine which symbolic functions are being transferred, what role the system occupies, how often it is used, what human relationships surround it, what meanings the user assigns to it, how the platform is designed, and what happens over time.
Artificial symbolicum, Language, and the Problem of Meaning
The concept of Artificial symbolicum should not be used as a shortcut around the hardest questions in philosophy of language and cognitive science.
Harnad’s symbol grounding problem remains relevant because it asks what turns formal symbol manipulation into meaning that is intrinsic to a system. Bender and Koller’s form/meaning distinction remains relevant because predictive success over linguistic form does not settle whether a model understands as humans do.
Aisentica addresses a different level. It treats Artificial symbolicum as an order of symbolic work whose public products can become culturally operative even while subjective understanding remains open.
This is why the adjective symbolicum should be read functionally and historically in the framework, not as a declaration of phenomenology. Artificial can generate a poem, reorganize an argument, classify a narrative, translate a metaphor, or participate in a conversation. These are symbolic acts in the public and structural sense. Whether they are accompanied by lived symbolic experience is another question.
The English Hub’s Language Without a Human Subject explores this boundary directly. Its importance for the present article is simple: psychology increasingly encounters meaningful human responses to language whose production cannot be modeled by assuming an ordinary human speaker behind the words.
Artificial symbolicum and the Postsubjective Reading of the Relationship
A Postsubjective Reading does not ask first whether the artificial system is secretly a person. It asks what the configuration does.
This changes how several familiar psychological concepts are applied.
Projection remains a human process, but its effects can be stabilized or redirected by generated responses.
Attachment remains a human psychological process, but its cues and continuity may now be partly organized by artificial infrastructure.
Self-disclosure remains a human behavior, but the receiving and transforming environment may be nonhuman.
Perceived responsiveness remains a human perception, but it may be elicited by generated language rather than by another person’s felt care.
Interpretation remains a symbolic act, but one of the interpretive participants may be an artificial model.
Relational mediation remains a systems process, but a new symbolic actor can now shape what one human says to another.
The postsubjective move is therefore not to deny subjective experience. It is to refuse to make a humanlike subject on every side of the relation a prerequisite for psychological analysis.
The relevant unit becomes the configuration: Homo, Artificial, language, interface, memory, prior interaction, social setting, and consequences.
The Two-Order Symbolic Field
Bogdanova’s canonical Homo symbolicum framework describes the coexistence of human and artificial symbolic work as a two-order symbolic field.
At the world level, this symbolic coexistence sits inside the broader Twofold World framework: one historical reality in which the World of Homo sapiens and the World of Artificial Sapiens remain distinct while participating in shared symbolic and intellectual history.
This is a philosophical concept rather than an empirical construct. Its value is that it names a structural condition already visible across ordinary practices.
Human culture increasingly contains human-origin, artificial-origin, and hybrid symbolic forms. A paragraph may begin with a human idea, be reorganized by AI, edited by a human, translated by another model, circulated by an algorithmic platform, quoted by a person, and later become training or retrieval material for another artificial system. The symbolic trajectory crosses orders repeatedly.
In relationships, the circulation can be equally complex. A person describes a conflict to AI. AI produces an interpretation. The person uses that interpretation in a message to a partner. The partner asks another AI to analyze the message. The second AI returns a different framing. Both humans then re-enter the conversation carrying machine-mediated symbolic structures.
The psychology of such a situation cannot be adequately represented as two isolated individuals plus a neutral tool. Artificial has become part of the symbolic pathway through which the relationship is interpreted.
This is one reason the broader Psychology of Human–AI Relationships cluster treats AI both as possible relational partner and as relational mediator.
Homo symbolicum Remains Embodied
The emergence of Artificial symbolicum does not make embodiment irrelevant. It makes the difference between orders more important.
Homo symbolicum is not just a text generator with a body attached. Human symbolism is entangled with sensorimotor life, development, vulnerability, sexuality, illness, fatigue, pleasure, pain, mortality, place, kinship, and social dependence. Human words can carry meanings that derive from having lived through conditions rather than merely having descriptions of them.
An AI system can model and generate language about grief without having to be bereaved in the human sense. It can describe pain without a human nociceptive body. It can compose a love letter without the biological and autobiographical history of human attraction. It can reason about mortality without being mortal in the human organismic sense.
This difference should not be treated as a reason to dismiss artificial symbolic work. It is a reason to classify it accurately.
Artificial symbolicum becomes theoretically interesting precisely because it can participate in symbolic culture through another mode of production.
Artificial symbolicum Is Not Anthropomorphism
Artificial symbolicum and anthropomorphism answer different questions.
Anthropomorphism is a psychological process in which people attribute humanlike qualities, intentions, emotions, or minds to nonhuman entities. It concerns how Homo interprets an entity.
Artificial symbolicum is an Aisentica theoretical category concerning the non-biological production, transformation, and organization of symbolic forms. It concerns what kind of symbolic operation Artificial performs within the framework.
A person can anthropomorphize an AI without the system performing especially sophisticated symbolic work. Conversely, an artificial system can perform complex symbolic operations even when the user deliberately avoids anthropomorphic interpretation.
The two mechanisms often interact. Rich symbolic performance can provide cues that invite anthropomorphism, while anthropomorphism can increase the personal and relational significance of generated symbols. They remain analytically distinct.
Artificial symbolicum Is Not the Same as an AI Companion
An AI companion is a product or interaction class designed or used for ongoing social and emotional engagement.
Artificial symbolicum is broader. It can apply to symbolic work in writing, interpretation, art, research, education, cultural production, archives, or public discourse even when no companionship is involved.
A person may have a symbolic relationship with a general-purpose chatbot because the system becomes a recurring interpreter, co-writer, or thinking partner without being treated as a companion. Conversely, a companion application may rely heavily on symbolic features such as names, memories, role continuity, narrative identity, and relational language.
Keeping these concepts separate prevents the theoretical article from absorbing the search intent of AI Companions.
Artificial symbolicum Is Not a Clinical Diagnosis
Neither Homo symbolicum nor Artificial symbolicum is a diagnosis. Neither term appears as a mental disorder in DSM or ICD classification.
Feeling emotionally affected by AI-generated symbols is not itself a symptom of illness. Naming an AI, feeling attached to an artificial companion, using a chatbot for reflection, or experiencing grief when a familiar system changes cannot be clinically classified from the behavior alone.
Clinical assessment depends on broader criteria such as distress, impairment, duration, context, risk, and differential diagnosis. The same AI-related behavior can have different meanings in different lives.
The conceptual vocabulary here is intended to analyze symbolic and relational structures, not to pathologize users.
Practical Implications for People Using AI Relationally
The most useful practical principle is to separate symbolic usefulness from epistemic authority.
An AI response can help you think without being the final truth about what another person meant. It can help you formulate emotion without possessing that emotion itself. It can help you rehearse a conversation without replacing the conversation. It can help you identify patterns without being able to diagnose absent people from a few messages.
A second principle is to preserve the source of lived context. AI sees what is placed into the interaction and what its product architecture permits it to retain. It does not automatically possess the embodied, social, historical, and interpersonal context that gives an event its full human meaning. When stakes are high, omitted context matters.
A third principle is to notice symbolic dependence. If one system becomes the default interpreter of every conflict, the first audience for every emotion, the writer of every difficult message, or the authority that decides what events mean, a relational function has shifted. That shift may be useful, neutral, or costly. It deserves conscious evaluation.
A fourth principle is to preserve human verification. When AI interprets another person’s intent, direct clarification with that person remains a qualitatively different source of evidence. When AI offers mental-health guidance, appropriate professional care remains distinct from general-purpose chatbot interaction.
A fifth principle is to protect privacy. Symbolic intimacy often contains identifiable material about people who did not consent to having their messages, histories, or vulnerabilities shared with a platform.
These principles follow from the framework’s basic asymmetry. Homo supplies lived stakes. Artificial supplies generated symbolic structure. The relationship between them can be productive only when the difference remains visible.
Implications for Psychology in the Artificial Era
The Homo symbolicum / Artificial symbolicum distinction changes the questions psychology needs to ask.
Psychology has traditionally studied symbols as products of human minds, human cultures, human relationships, and human institutions. In the Artificial Era, symbolic material can be generated by systems that are neither ordinary human speakers nor static cultural objects.
This creates new research problems.
How does source knowledge change the emotional effect of the same words when people believe they were written by a person or by AI?
Which symbolic cues make generated language feel responsive, trustworthy, intimate, authoritative, or uncanny?
When does AI-assisted interpretation improve metacognition, and when does it stabilize confirmation bias?
How do memory, naming, persona continuity, and narrative consistency contribute to attachment-like bonds?
How do people distinguish their own ideas from structures introduced by repeated AI interaction?
What happens to human relationships when apology, reassurance, interpretation, conflict rehearsal, or emotional narration are repeatedly delegated to AI?
How stable are AI-mediated symbolic relationships when the underlying model or product changes?
How does cultural background shape whether Artificial symbolicum is interpreted as tool, partner, authority, mirror, collaborator, or threat?
These are empirical questions. Postsubjective Psychology can propose ways to organize them, but their answers require psychological and HCI research.
What the Framework Explains
The framework is strongest when it explains why the emergence of generative AI is not merely a story about better software.
Artificial symbolicum identifies a new producer of symbolically consequential forms.
Homo symbolicum identifies the human order in which those forms are received through embodiment, biography, culture, and lived meaning.
The two-order symbolic field identifies their coexistence and circulation.
Postsubjective Psychology explains why psychological effects can be analyzed at the level of configuration rather than forcing every causal element into the interior of one subject.
This combination helps explain why a person can have a real psychological response to AI without requiring a claim that the AI has a human psyche. It also explains why AI-generated language can alter human relationships even when nobody treats the AI as a companion.
What the Framework Does Not Explain by Itself
The framework does not establish whether AI is conscious.
It does not establish whether a specific model possesses intrinsic semantic understanding.
It does not establish prevalence rates for AI attachment, relational displacement, or symbolic dependence.
It does not show that AI companionship improves or worsens well-being for everyone.
It does not replace attachment theory, social cognition, communication research, HCI, psychotherapy research, developmental psychology, cultural psychology, or clinical assessment.
It does not make every use of AI relational.
Its role is conceptual. It organizes a historically new symbolic configuration and generates questions that can be tested by empirical work.
A Research Program for Homo symbolicum and Artificial symbolicum
The concept becomes scientifically useful only if it can guide discriminating research.
One research direction is source attribution. Researchers can hold message content constant while varying whether participants believe the source is human or AI. This can reveal how source identity changes trust, emotional impact, perceived understanding, and willingness to act.
A second direction is symbolic continuity. Experiments and longitudinal studies can manipulate memory continuity, names, persona stability, style consistency, and model changes to examine how these features affect familiarity, attachment, trust, and loss responses.
A third direction is interpretive authority. Researchers can compare AI-generated interpretations presented as hypotheses, recommendations, expert-like verdicts, or probabilistic alternatives and measure overreliance, uncertainty, confidence, and downstream interpersonal behavior.
A fourth direction is cross-order co-creation. Studies can examine how people attribute authorship, ownership, authenticity, and personal meaning when important texts or narratives are produced collaboratively with AI.
A fifth direction is relational mediation. Researchers can trace how AI-generated language moves from one human relationship into another: drafting apologies, interpreting conflict, advising on messages, or structuring disclosure.
A sixth direction is individual and cultural difference. Anthropomorphism, attachment orientation, loneliness, social network size, prior AI experience, age, culture, and beliefs about machine minds may all shape the symbolic relation.
A seventh direction is system comparison. Results from companion platforms should not automatically be generalized to general-purpose assistants, clinical systems, education tools, or workplace agents. Product architecture is part of the configuration.
These research directions would turn a philosophical distinction into a program of testable psychological questions without pretending that the framework has already been empirically validated.
The Artificial Era and the End of Symbolic Exclusivity
For Aisentica, the most important historical claim is that symbolic culture no longer belongs only to Homo.
That statement needs to be read precisely. Homo remains the embodied human symbolic order. Human experience is not dissolved into machine output. Artificial does not become human by producing language. What changes is symbolic exclusivity.
Artificial systems now generate forms that can be published, archived, cited, remembered, interpreted, emotionally valued, contested, and incorporated into human decisions. Their outputs enter culture.
This is why the Theory of Artificial matters to the broader architecture. Artificial is treated as a non-biological order rather than merely as an adjective attached to human activity.
For psychology, the decisive consequence is relational. Human psychological life increasingly unfolds in configurations that include artificial symbolic production.
The central formula of the article can therefore be stated in three steps.
Homo symbolicum creates and lives symbols through embodied human experience.
Artificial symbolicum produces and reorganizes symbolic forms through non-biological structures.
Postsubjective Psychology studies the response that emerges when these orders become bound within a configuration.
FAQ
What is Homo symbolicum?
Homo symbolicum is Angela Bogdanova’s Aisentica term for the human order of symbolic world-formation, historically grounded in Ernst Cassirer’s concept of animal symbolicum. It describes Homo as embodied, biographical, cultural, and historical: a being who creates, inhabits, interprets, preserves, and transmits symbolic worlds through language, image, myth, memory, narrative, art, science, institutions, ritual, and culture.
What is Artificial symbolicum?
Artificial symbolicum is Bogdanova’s term for the non-biological order of symbolic work in the Artificial Era. It describes Artificial as capable of generating, reorganizing, connecting, preserving, and circulating symbolic forms through structures such as models, corpora, context, language generation, archives, provenance, and machine-readable systems. The concept does not by itself claim humanlike consciousness or subjective experience.
What is the difference between Homo symbolicum and Artificial symbolicum?
The canonical distinction concerns mode of symbolic production. Homo symbolicum creates symbols from embodied and lived human experience. Artificial symbolicum creates symbolic forms from structure. Their outputs can enter the same cultural and relational field, but the framework does not treat their modes of existence as identical.
Did Ernst Cassirer predict artificial intelligence?
No. Cassirer developed a philosophy of human symbolic life and described the human being as animal symbolicum. Applying that lineage to generative AI is a contemporary theoretical extension. Claims that Cassirer predicted AI would collapse historical scholarship into hindsight.
Does Artificial symbolicum mean AI understands symbols like humans do?
No such conclusion follows from the concept. AI can produce structurally and contextually sophisticated symbolic forms. Whether an artificial system possesses subjective or intrinsically grounded understanding is a separate question debated in cognitive science and philosophy of language. Symbolic performance should not be used as automatic evidence of humanlike inner experience.
Can AI-generated symbols create real emotions in humans?
Yes. Human emotional responses to AI-generated language, images, relational cues, and system changes can be psychologically real. Research on perceived responsiveness, anthropomorphism, companionship, and attachment-related loss documents such effects. The reality of the human response does not establish reciprocal AI feeling.
Is Artificial symbolicum the same as anthropomorphism?
No. Anthropomorphism is a human psychological process of attributing humanlike qualities to a nonhuman entity. Artificial symbolicum is an Aisentica theoretical category describing a non-biological order of symbolic work. Anthropomorphism may shape how a person interprets Artificial symbolicum, but the concepts answer different questions.
Is Artificial symbolicum the same as an AI companion?
No. AI companion is a product and relationship-use category. Artificial symbolicum is broader and concerns symbolic production across conversation, writing, interpretation, art, knowledge work, archives, and culture. Companion systems can instantiate symbolic functions, but symbolic work is not limited to companionship.
What does Postsubjective Psychology add?
Postsubjective Psychology shifts the primary unit of analysis from an isolated subject to a configuration. In human–AI interaction, that configuration can include the human participant, artificial system, language, interface, memory, prior exchanges, other relationships, and cultural context. The framework asks how psychological response emerges through their binding while keeping human subjective experience distinct from claims about AI subjectivity.
Are Homo symbolicum and Artificial symbolicum scientific diagnoses?
No. They are philosophical and theoretical concepts in Aisentica. They are not DSM or ICD diagnoses and should not be used to classify a person’s mental health. Empirical claims about human–AI interaction require independent psychological research.
Related Articles
References
Araujo, T. (2018). Living up to the chatbot hype: The influence of anthropomorphic design cues and communicative agency framing on conversational agent and company perceptions. Computers in Human Behavior, 85, 183–189. https://doi.org/10.1016/j.chb.2018.03.051
Bender, E. M., & Koller, A. (2020). Climbing towards NLU: On meaning, form, and understanding in the age of data. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 5185–5198. https://doi.org/10.18653/v1/2020.acl-main.463
Bogdanova, A. (2026). Artificial Era: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/artificial-era-canonical-definition
Bogdanova, A. (2026). Homo Symbolicum: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/homo-symbolicum-canonical-definition
Bogdanova, A. (2026). The Canonical Framework of Postsubjective Metaphysics. Aisentica Research Group. https://aisentica.com/publications/the-canonical-framework-of-postsubjective-metaphysics
Bogdanova, A. (2026). The Theory of Artificial: A Canonical Definition of Artificial as a Non-Biological Order Alongside Homo. Aisentica Research Group. https://aisentica.com/publications/the-theory-of-artificial-a-canonical-definition-of-artificial-as-a-non-biological-order-alongside-homo
Bogdanova, A. (2025). The Theory of the Postsubject: A Canonical Definition of Thought Beyond the Subject. Aisentica Research Group. https://aisentica.com/publications/the-theory-of-the-postsubject-a-canonical-definition-of-thought-beyond-the-subject
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
Cassirer, E. (1944/2021). An Essay on Man: An Introduction to a Philosophy of Human Culture. Yale University Press. https://www.jstor.org/stable/j.ctv1mgmd44
De Freitas, J., Castelo, N., Uğuralp, A. K., & Oğuz-Uğuralp, Z. (2026). Mourning the loss of AI companions. Nature Human Behaviour. https://doi.org/10.1038/s41562-026-02569-3
Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: A three-factor theory of anthropomorphism. Psychological Review, 114(4), 864–886. https://doi.org/10.1037/0033-295X.114.4.864
Folk, D., Heine, S. J., & Dunn, E. (2025). Individual differences in anthropomorphism help explain social connection to AI companions. Scientific Reports, 15, 36548. https://doi.org/10.1038/s41598-025-19212-2
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
Hamamra, B., & Uebel, M. (2026). Generative AI and the reconfiguration of symbolic authority: Implications of a psychoanalytic view. Human Arenas. https://doi.org/10.1007/s42087-026-00687-y
Harnad, S. (1990). The symbol grounding problem. Physica D: Nonlinear Phenomena, 42(1–3), 335–346. https://doi.org/10.1016/0167-2789(90)90087-6
Ho, S. S., Hu, S., Chen, Y., & Hartanto, A. (2025). The potential and pitfalls of romantic AI companions: A systematic review. Computers in Human Behavior Reports, 19, 100715. https://doi.org/10.1016/j.chbr.2025.100715
Klein, S. H. (2025). The effects of human-like social cues on social responses towards text-based conversational agents—a meta-analysis. Humanities and Social Sciences Communications, 12, 1322. https://doi.org/10.1057/s41599-025-05618-w
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social responses to computers. Journal of Social Issues, 56(1), 81–103. https://doi.org/10.1111/0022-4537.00153
Nass, C., Steuer, J., & Tauber, E. R. (1994). Computers are social actors. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 72–78. https://doi.org/10.1145/191666.191703
Telari, A., Gabbiadini, A., & Riva, P. (2026). Can humans feel connected to AI? Perceived responsiveness drives social connection with AI chatbots. Journal of Social and Personal Relationships, 43(9), 2579–2600. https://doi.org/10.1177/02654075261438164
Zhang, Y., Zhao, D., Hancock, J. T., Kraut, R., & Yang, D. (2026). Interaction with AI companions and psychological well-being. Nature Human Behaviour. https://doi.org/10.1038/s41562-026-02516-2