Glossary

Here you will find brief encyclopedic references to the theories, concepts, and models on which the main modules of Chimera were designed.

Memory System

Atkinson–Shiffrin model

A classic cognitive theory (1968) describing the structure of memory as a sequence of three functionally distinct stores: sensory register, short-term memory, and long-term memory. Information travels from immediate perception, through an attentional filter, to short-term memory, where, via encoding and rehearsal, it is consolidated into long-term storage.

A key feature of the model is its rigid processing hierarchy: short-term memory acts as a limited‑capacity “workspace” where active data manipulation occurs, while long‑term memory serves as a virtually unlimited knowledge archive. Although the model has been criticized for its excessive linearity and oversimplification of information transfer mechanisms, it laid the foundation for understanding the dynamics of forgetting and memory retention, to become the foundational paradigm for studying cognitive architecture.

Self-memory system

A theoretical concept introduced by Conway and Pleydell-Pearce in 2000. It conceptualizes autobiographical memory as a dynamic system inextricably linked to personality structure and self‑concept. Memories and personality are in constant interaction: an individual’s current goals and tasks determine what information is retrieved from memory; in turn, memory itself limits and shapes those goals and the self‑concept.

The central element of the model is the “working self,” which regulates access to long‑term memory and filters out memories that conflict with current self‑image or goals. Memory thus becomes a constructive process in which the past is reconstructed in light of present needs, thereby ensuring the continuity of self‑identity.

Resource theory of attention

A cognitive model developed by Kahneman in 1973. Attention is conceptualized as a limited supply of mental resources available for information processing. The model posits that attentional limits depend less on structural “bottlenecks” and more on the total amount of available processing capacity, which can be flexibly allocated across different tasks.

Resource distribution is governed by a central mechanism guided by the individual’s current intentions and enduring dispositions. Processing efficiency depends directly on the ratio of allocated resources to task demands: when task requirements exceed the available attentional capacity, performance declines or the task is abandoned. This accounts for the phenomena of cognitive overload and selective attention.

Interference theory

A classic account in memory psychology (1900) according to which forgetting arises due to competition between memory traces. Recall difficulties stem not from the natural decay of traces, but from other memories—particularly those similar in meaning or context—competing for access and blocking retrieval of the target information.

The two main types of interference are proactive (past learning interferes with the acquisition or retrieval of new information) and retroactive (newly learned information disrupts the recall of older material). The model underscores that forgetting reflects a dynamic memory system in which overlap between similar items can lead to their distortion or integration into more schematic, gist‑like representations as a way to reduce cognitive conflict.

Tulving’s multiple memory systems theory

Proposed by Tulving in 1985, this framework posits that human memory is not a unitary faculty but comprises several qualitatively distinct systems operating on different principles. The primary distinction is between procedural memory (skills and automatic behaviors) and declarative memory, with the latter subdivided into semantic memory (general world knowledge) and episodic memory (personally experienced events tied to specific times and places).

A central tenet of the theory is the distinction between knowing facts and personally experiencing past events. Episodic memory enables mental time travel—the capacity to re‑experience one’s past and to be aware of oneself as having been there—whereas semantic memory stores abstract conceptual knowledge independent of the specific context in which it was acquired. Although these systems are neurobiologically dissociable, they interact closely during encoding and retrieval.

Distributed representation theory

Proposed by Hinton in 1986 within the framework of artificial neural networks and cognitive science, this theory holds that information is encoded not in localized nodes but in patterns of activity distributed across many processing units. In such a system, each concept corresponds to a unique vector in a multidimensional feature space, and individual units participate in representing multiple different items simultaneously.

The central advantage of this approach is its inherent capacity for generalization and semantic categorization: concepts with similar meanings give rise to similar activation patterns, effectively becoming neighbors in vector space. This allows the system to automatically capture latent relationships and analogies, enabling retrieval and processing based on semantic proximity rather than formal symbolic matching.

Memory systems theory

A classification model proposed by Larry Squire in 1987 that partitions memory into two broad categories based on their processing mechanisms and underlying neural substrates: declarative (explicit) and nondeclarative (implicit) memory. Declarative memory—responsible for the conscious recall of facts and events—depends critically on the medial temporal lobe and hippocampus.

Nondeclarative memory, by contrast, encompasses skills, habits, priming, and simple forms of conditioning. It is acquired and expressed unconsciously, involving other brain structures such as the basal ganglia and cerebellum.

This framework established the view of memory not as a unitary faculty but as a collection of functionally independent systems that may have followed distinct evolutionary trajectories.

Emotional consolidation theory

A neuropsychological concept (2000) explaining the mechanism whereby emotional arousal enhances the strength and persistence of memory traces. Activation of the limbic system—particularly the amygdala—during emotionally significant events modulates hippocampal function, triggering a cascade of biochemical reactions leading to more robust synaptic consolidation of information.

This process confers an evolutionary advantage by ensuring that events associated with danger, reward, or high social salience are remembered more strongly and retained longer than neutral experience. Emotions thus serve not merely as a backdrop but as a critical marker of significance, determining which fragments of experience are consigned to long-term storage.

Umwelt

A foundational concept in biosemiotics introduced by Jakob von Uexküll in 1909. It refers to the “subjective universe” of an organism—the unique meaningful world constructed by a living being through its specific sensory and motor capacities. In contrast to the objective physical environment, the umwelt is a model of reality shaped solely by those signals that matter for the survival and adaptation of a given species.

The central insight is that perception does not mirror reality. The organism (or cognitive system) functions as an active interpreter, filtering out “information noise” and isolating meaningful features. For artificial systems, the notion of umwelt implies the presence of built-in perceptual filters—such as personal dispositions, affective valences, or goal functions—that transform raw data streams into subjectively meaningful experience, thereby constructing a unique world model distinct from that of the user or other systems.

Levels of processing

A cognitive theory proposed by Craik and Lockhart in 1972. The durability of memory is a direct function of the depth to which information is processed. In contrast to structural models that posit separate memory stores, this approach holds that memory is a by‑product of perceptual and cognitive analysis.

The theory distinguishes between different levels of processing, ranging from shallow (analysis of physical features, such as the shape of letters or sound patterns) to deep (semantic analysis of meaning, context, and associations). The deeper the processing, the more durable and elaborate the resulting memory trace, which in turn supports more effective subsequent retrieval.

Personality System

The Big Five

The dominant model in personality psychology that crystallized in the 1980s and 1990s, describing the structure of individual differences in terms of five broad dimensions: neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness. The framework rests on the lexical hypothesis—the idea that the most salient personality attributes are encoded in natural language, and that analyzing stable linguistic descriptors serves to reveal the basic architecture of human character.

Unlike typological approaches that sort people into discrete categories, the Big Five conceptualizes traits as continuous spectra. For cognitive architectures, this offers a scientifically grounded basis for modeling identity: the system is afforded not a set of rigid templates but continuous scales—“dispositions”—making it possible to simulate a personality with a stable core that remains recognizable while exhibiting natural behavioral variability across contexts, depending on its position on the trait spectra.

Self-reflection System

Two‑factor theory of emotion

A cognitive model proposed by Stanley Schachter and Jerome Singer in 1962. It posits that emotional experience results from the interplay of two components: physiological arousal and its cognitive interpretation. According to the theory, physiological activation is nonspecific—it is essentially the same across different emotions—and the particular quality of the experience (fear, anger, joy, etc.) arises from attributing that arousal to a cause in the surrounding context.

For cognitive architectures, this framework highlights the importance of distinguishing between the affective signal and its semantic appraisal. It allows emotions to be modeled not as hardwired reflexes but as a constructive process in which the same state of arousal can give rise to different emotional outcomes depending on how the situation is interpreted.

Processual unconscious

Dynamic self‑regulation of personality

A class of models that conceptualize personality as an open dynamic system continuously maintaining relative equilibrium and developmental direction through feedback loops and structural reorganization. Whereas static models treat personality as a fixed set of traits, dynamic self‑regulation views the personality profile as a temporarily stabilized outcome of multiple parallel processes—goal setting, discrepancy monitoring, affective appraisal, and behavioral adjustment.

A core premise is that stable personality patterns are not pre‑given but emerge and are sustained through ongoing regulatory cycles. For artificial agent architectures, this framework motivates the use of dynamic rather than rigidly fixed personality profiles, enabling the system to adapt to novel conditions and accumulate experience without overhauling its entire model.

Implicit learning

A cognitive phenomenon and research program pioneered by Arthur Reber in the late 1960s. In classic experiments using artificial grammars, subjects acquired the ability to distinguish grammatical from ungrammatical sequences while remaining unable to articulate the underlying rules. The phenomenon was later demonstrated in sequential learning (serial reaction time tasks) and other statistically structured environments.

A defining feature of implicit learning is the acquisition of knowledge about complex environmental regularities without awareness of their structure and without the capacity for conscious access to what has been learned. For cognitive architectures, this motivates the incorporation of mechanisms that build implicit schemas and behavioral patterns through statistical processing of experience, operating outside the scope of conscious access. Such schemas can serve as functional counterparts to “unconscious knowledge,” shaping choices and interpretations without being explicitly represented.

Adaptive unconscious

A contemporary cognitive model developed by Timothy Wilson and presented in his book “Strangers to Ourselves: Discovering the Adaptive Unconscious” (2002). In contrast to the psychoanalytic unconscious—conceived as a repository of repressed conflicts—the adaptive unconscious refers to the array of mental processes that operate outside conscious awareness, automatically evaluating information, detecting patterns, setting goals, and forming preferences, thereby enabling rapid adaptation to the environment.

The central claim is that most mental “work”—learning, categorization, decision‑making—takes place beneath the threshold of consciousness, which is presented only with the final products: intuitions, feelings, and motivations. For artificial systems, this framework motivates an architecture that separates fast, automatic implicit processing from slow, deliberate conscious planning, with implicit processes engaged only in relevant contexts rather than requiring constant conscious oversight.

Cumulative unconscious

A model that conceptualizes the unconscious as the cumulative product of numerous subliminal experiences and their traces. Unlike the classical psychoanalytic unconscious—understood primarily as a repository populated by repressed instinctual impulses and conflicts—the cumulative unconscious takes shape through the layered buildup of micro‑events: emotionally charged occurrences that never quite reach conscious awareness. Individually, each such event leaves a faint but persistent trace; collectively, their accumulation coalesces into stable patterns of attitudes, responses, and behavioral tendencies.

In contemporary psychology and psychoanalysis, related ideas find expression in Masud Khan’s concept of cumulative trauma, which demonstrates how repeated, minor disruptions in care and emotional support can accumulate and produce enduring changes in an individual’s psychic organization and well‑being. For cognitive architectures, the notion of a cumulative unconscious suggests a design principle in which multiple influences are integrated through the weighted summation of “change vectors”: traces pointing in the same direction amplify one another, while opposing traces tend to cancel out—yielding a resultant developmental vector without requiring the explicit resolution of internal contradictions.

Linguistic presuppositions

A concept from linguistic pragmatics developed in the work of Robert Stalnaker in the early 1970s. A presupposition is a background assumption that the speaker takes for granted and expects the interlocutor also to accept as true at the moment of utterance. Unlike an assertion, a presupposition is not open to challenge within the immediate discourse; rather, it sets the framework within which alone the utterance has meaning.

The key function of presuppositions is to tacitly establish the conditions under which interpretation becomes possible. For cognitive architectures, this implies the presence of a layer of latent assumptions that remain unspoken yet determine which responses are considered meaningful, permissible, or expected. In the Chimera architecture, resonant cores function as analogous “cognitive presuppositions”: they do not directly govern behavior but constrain the space of possible reactions by providing the background semantic frame.

Narrative identity (Paul Ricœur)

A philosophical concept articulated by Paul Ricœur in the 1980s and 1990s (most notably in “Soi‑même comme un autre”", 1990; English: “Oneself as Another”). In contrast to substantialist views that posit the self as an unchanging core of personality, Ricœur understands identity as a story constructed through narrative: a person constitutes themselves by telling—to themselves and to others—the story of their lives. The self, on this account, is not a given but an ongoing process of rewriting one’s narrative in dialogue with others and with oneself.

The central insight is the distinction between identity as self‑sameness (idem) and identity as selfhood (ipse): the person maintains continuity not as an unchanging essence but as a coherent plot capable of integrating heterogeneous and even contradictory episodes. For cognitive architectures, this implies moving beyond static personality profiles toward a dynamic narrative structure in which significant events serve as “plot points” that shape how subsequent experience is developed and interpreted.

Processual (procedural) unconscious

A contemporary concept bridging cognitive science and psychoanalysis that distinguishes between the dynamic (repressed) unconscious and the procedural unconscious. Whereas the classical Freudian unconscious is understood as a repository of repressed, conflict‑laden content, the procedural unconscious comprises the repertoire of non‑conscious procedures and skills that draw on procedural (implicit) memory and operate without ever having been repressed. The term was consolidated in the work of Eric Kandel, who identified the “procedural unconscious” as a dimension of the self that remains outside awareness not because of repression, but because it is instantiated in the brain’s procedural memory systems.

The core insight is that a substantial portion of the unconscious is organized not around forbidden wishes, but around automatic routines for engaging with the world: perceptual patterns, emotional responses, and habitual modes of action. For cognitive architectures, this suggests modeling the unconscious as a dynamic network of procedures and schemas that become active in context and implicitly shape behavior, without requiring symbolic processing or conscious access.

Context‑dependent memory

A cognitive model holding that information retrieval is more effective when the encoding context matches the retrieval context. A classic demonstration is the Godden and Baddeley (1975) experiment, in which divers learned word lists either on land or underwater and were later tested in the same or the alternate environment: recall was significantly better when the learning and testing environments coincided.

The theory treats context as encompassing both external factors (location, surroundings, sensory cues) and internal states (emotional and physiological condition), all of which become incorporated into the memory trace and later serve as retrieval cues. For cognitive architectures, this implies that the accessibility of a given experience depends not only on its “strength” but also on the degree of match between current context and encoding conditions—thereby enabling efficient, context‑sensitive regulation of behavior.

Spreading activation theory

A cognitive model proposed by Allan Collins and Elizabeth Loftus in 1975, conceptualizing semantic memory as a network of concepts interconnected by links of varying strength. When a node is activated, excitation spreads through the network, attenuating with distance and modulated by connection weights. This accounts for priming effects: the facilitated processing of semantically related words and categories.

The core principle is the parallel, graded activation of multiple elements in proportion to their semantic proximity to the initial stimulus. For cognitive architectures, this provides a natural basis for contextual activation of experience: past events, concepts, and affective evaluations are not “retrieved” one at a time but instead continually modulate the current state in the background, exerting a distributed influence analogous to the “resonance field” of the unconscious.