REPRESENTATIONAL REDESCRIPTION

Introduction and Core Definition of Representational Redescription

Representational Redescription (RR) is a seminal concept within cognitive developmental psychology, describing the fundamental mental process through which a child transforms existing internal representations of knowledge into new, more explicit, and flexible formats. This sophisticated mechanism allows the child to move beyond merely successful, implicit behavior toward conscious understanding and the versatile application of knowledge across diverse contexts. It constitutes an iterative and cyclical process of internal reorganization, where knowledge that is initially procedural and inaccessible becomes progressively formalized and available for verbalization, reflection, and integration with other cognitive domains. The core thesis of RR addresses the profound challenge of explaining how competence, achieved through practice and interaction, evolves into conscious, explicit knowledge, making way for new and more sophisticated use of that knowledge structure.

The concept posits that cognitive development is not solely about the acquisition of novel information or the refinement of input data, but rather the systematic restructuring of the format in which information is stored internally. This internal computational process serves as a crucial bridge between highly specialized, modular knowledge and general, accessible understanding. A key feature of RR is its emphasis on the endogenous nature of this restructuring; the child’s mind actively works on its own products, redescribing existing, successful representations to make them computationally more robust and communicable. This internal drive for reorganization is essential for the transition from performance mastery to conceptual awareness, fundamentally altering how the child interacts with and reasons about the world around them.

Theoretical Foundation: The Contribution of Annette Karmiloff-Smith

Representational Redescription was first systematically proposed and elaborated by the British psychologist Annette Karmiloff-Smith during the 1980s and 1990s. Her work sought to reconcile the insights of Piagetian constructivism, which emphasized developmental stages based on interaction with the environment, with emerging nativist perspectives, which highlighted domain-specific innate predispositions. Karmiloff-Smith utilized RR as a framework to explain how initial domain-specific biases could lead, through self-generated internal reorganization, to generalized, explicit knowledge systems. Her approach provided a robust mechanism for developmental change that accounted for both the initial specialized nature of certain cognitive functions and the eventual broad flexibility observed in adult cognition.

Karmiloff-Smith argued that traditional models often overlooked the continuous internal work performed by the cognitive system after behavioral mastery has been achieved. She posited that developmental change occurs in cycles within specific domains (such as language, number, or physics), where a period of implicit mastery is followed by a period of internal redescription. This process generates new forms of representation that are increasingly abstract and detached from the original sensory or motor input. The theoretical foundation rests on the belief that the drive for redescription is fueled by the system’s inherent need for compression, efficiency, and the ability to compare representations across different contexts, thus reducing cognitive load and enhancing flexibility. The theory distinguishes itself by focusing on the format change of representations, rather than merely the content change, offering a powerful explanation for the qualitative shifts observed during childhood development.

The J-E-I Levels of Representation

The Representational Redescription model defines development as proceeding through a sequence of four distinct representational levels, often summarized using the acronyms I (Implicit), E1 (Explicit 1), E2 (Explicit 2), and E3/J (Explicit 3/Justified Knowledge). These levels describe the accessibility and flexibility of the knowledge stored within a specific cognitive domain. The initial level, the Implicit (I) level, is characterized by procedural knowledge. At this stage, the child can successfully perform a task, but the knowledge driving the behavior is encapsulated, tied directly to the input, and inaccessible to conscious reflection or verbal report. Behavior is competent, yet the child cannot articulate the rules governing their actions.

The second level, Explicit 1 (E1), marks the first significant step in redescription. The implicit representation is re-coded into a more compressed, stable mental format. Crucially, this representation becomes consciously available to the child, forming an internal model. However, E1 representations remain largely tied to the original domain context and are not yet available for sharing or generalization outside that specific domain. This initial move toward explicitness often results in the temporary disruption of behavior, leading to the well-documented U-shaped developmental curve, as the child attempts to impose a simple, conscious rule onto the complex, successful implicit behavior.

The subsequent levels involve progressive abstraction. At Explicit 2 (E2), the representation is further redescribed, becoming decoupled from the initial perceptual and motor inputs. E2 representations are highly abstract and are now available for inter-domain comparison and communication with others. They are stored in a common, generalized format, making them flexible tools for reasoning. Finally, the Explicit 3 (E3 or J – Justified) level represents the highest stage of flexibility. The knowledge is fully verbalizable, consciously controlled, and integrated into a coherent theoretical framework. At this level, the child can not only state the rule but also justify its existence and apply it universally, demonstrating complete mastery and conceptual awareness.

Mechanisms of Redescription: The Process Itself

The actual mechanism driving Representational Redescription is an internal computational process triggered by the cognitive system itself, rather than solely by external environmental failure. Unlike Piagetian accommodation, where errors necessitate restructuring, RR often occurs after a period of behavioral success (I-level mastery). The system, having accumulated a large amount of effective, but computationally heavy, implicit data, seeks to compress and simplify this information. This compression involves taking the procedural steps and outcomes stored implicitly and recoding them into a compact, symbolic, and declarative format.

This process of recoding is iterative and results in a change of format—from a procedural representation to an explicit, data-driven representation. The new format is inherently more flexible because it is less bound by the constraints of the original input stimuli. For instance, an E1 representation, while conscious, is still constrained by its origin, whereas an E2 representation is sufficiently abstracted to be manipulated purely internally, without reference to the external world. This continuous internal recycling of successful information ensures that development is driven both by environmental interaction (leading to I-level success) and by the system’s own intrinsic need for cognitive optimization and theoretical coherence.

The computational effort required for redescription explains the temporary dip in performance often observed when transitioning from the implicit to the first explicit level. As the child attempts to consciously access and formalize their implicit knowledge, they often oversimplify the rules, temporarily replacing complex but effective procedures with overly restrictive, conscious rules. This temporary regression is a predictable consequence of the mind actively restructuring its own knowledge base, demonstrating that cognitive development is rarely a smooth, linear progression but rather a series of cyclical restructurings within each domain.

Domain Specificity and Generalization

A critical feature of the Representational Redescription framework is its nuanced position regarding domain specificity versus domain generality. Karmiloff-Smith strongly asserted that the initial stages of development, particularly the I-level, are highly domain-specific. Children appear to possess specialized learning mechanisms or predispositions (modules) for processing information related to areas like language acquisition, intuitive physics, or face recognition. These domains operate largely independently, allowing for specialized efficiency in specific tasks.

However, the power of the RR mechanism lies in its ability to transcend these initial boundaries. While the *content* being redescribed is domain-specific (e.g., knowledge about grammatical rules, or knowledge about balance), the *process* of redescription itself is posited to be domain-general. The transformation from implicit to explicit knowledge uses a common computational mechanism regardless of the cognitive domain involved. The entire developmental trajectory, moving from I to E3, can be viewed as the gradual dissolution of strict modularity. As representations reach the E2 and E3 levels, they are stored in a format that allows them to be shared and integrated across modules, thereby facilitating true conceptual generalization and the development of high-level reasoning abilities that rely on cross-domain knowledge.

Therefore, RR provides a dynamic model where initial specialized competence gradually evolves into generalized conceptual understanding. The redescription process serves the function of taking knowledge that is “trapped” within specialized modules and elevating it to a central, shared cognitive workspace. This explains how children eventually develop the capacity to build coherent, overarching theories about the world, rather than maintaining isolated pockets of procedural knowledge. The ability to abstract core principles and apply them analogously across different types of problems is a direct result of successful representational redescription.

Empirical Support and Developmental Manifestations

Empirical evidence for Representational Redescription comes from various developmental tasks that demonstrate the characteristic cyclical patterns of performance mastery, temporary regression, and subsequent conceptual advancement. The most frequently cited example is Karmiloff-Smith’s work on the Balance Beam Task. In this experiment, children must predict which side of a pivot beam will drop based on the placement of weights.

At the I-level, very young children often solve the task successfully by implicitly considering both weight and distance, demonstrating behavioral mastery. However, when asked to explain their reasoning (the attempt to access E1), they cannot articulate the rule. As the child begins the redescription process (E1/E2 transition), they often impose a simple, explicit, but incorrect rule, such as “always focus only on the weight.” This conscious rule-following leads to errors in situations where distance is the determining factor—a clear example of the U-shaped curve where performance temporarily degrades as explicit knowledge takes precedence over implicit success. Only at the E3/J level do children successfully redescribe their knowledge to integrate both weight and distance into a fully verbalizable and flexible principle (the torque rule), demonstrating conceptual understanding.

Similar manifestations of RR have been observed in other domains. In language acquisition, children often move from implicit mastery of irregular past tense verbs (e.g., “went”) to a temporary stage of over-regularization (e.g., “goed”) when they explicitly realize the existence of a rule-based system, only to resolve this conflict by redescribing both the regular and irregular forms into a coherent explicit grammatical system. In drawing development, children initially draw successfully but implicitly (I-level). They then move to a stage where they consciously impose schematic rules (E1/E2), sometimes making their drawings less naturalistic, before achieving the final stage of flexible, conscious control over artistic representation (E3). These examples confirm the cyclical, non-linear nature of developmental progress predicted by the RR theory.

Significance and Conceptual Distinctions

Representational Redescription holds profound significance for developmental theory because it provides a detailed mechanism for endogenous cognitive change. It shifts the focus from purely external inputs driving learning to the internal, self-generated restructuring of knowledge. This framework offers a compelling explanation for how children become consciously aware of the abilities they already possess implicitly, a phenomenon that other theories often struggled to address adequately. By focusing on changes in the format of representations, RR illuminates the transition from “knowing how” (procedural knowledge) to “knowing that” (declarative knowledge).

Conceptually, RR distinguishes itself sharply from purely maturation-based models, as it emphasizes active computational work, and from traditional Piagetian models. While Piaget emphasized equilibration driven by external failure (disequilibrium), Karmiloff-Smith argued that redescription is driven by internal pressures for computational efficiency and coherence, often occurring *after* successful behavior is achieved. This difference is key: the child is driven to redescribe successful implicit knowledge because the implicit format is computationally opaque and inflexible, even if behaviorally effective. Therefore, RR offers a powerful synthesis, acknowledging the role of innate biases while detailing the constructive processes that lead to flexible, human-level intelligence. The theory remains foundational for understanding the trajectory of complex learning and the nature of conscious awareness in development.