*2025-03-05* ## Introduction What does it mean to understand something? What does it mean to think? Do we even have adequate language to describe these fundamental cognitive processes? Despite centuries of philosophical inquiry and decades of scientific research, we still struggle to articulate the mechanics of our own thinking. The mind observes everything except itself with clarity. Like the eye that cannot see itself without a mirror, our consciousness finds it challenging to examine its own operations. In this essay, I present a descriptive model of cognition designed to serve as a medium through which we can discuss thinking and understanding. This model does not claim to describe the actual neurological processes occurring within our brains. It does not even claim to describe anything that objectively happens in our reality. Rather, it translates whatever actually is happening into something we can talk about, something we can comprehend. It provides a framework and vocabulary that allows us to examine our own thinking processes with greater precision. The methodology employed here is both observational and functional. This model has been developed through connecting and synthesizing introspective observations and existing theories about cognition. It approaches the mind as a system of black boxes whose internal mechanisms remain unknown, but whose inputs, outputs, and functional behaviors can be observed and described. Consider how we retrieve memories. We know we have the ability to recall information from the past. We can think of this as a function that exists in our mind—a "remembering" function that we can call upon. We don't know precisely how this happens at a neurological level, but we know that it happens. By treating "remembering" as a black box function, we can examine its behavior: what kinds of inputs it accepts, what outputs it generates, how it performs under different conditions, and how it interacts with other cognitive functions. This model does not aim for scientific precision or neurological accuracy. Rather, it provides a set of linguistic tools for communication about thinking processes. It offers a level of granularity beyond common terms like "processing," "remembering," or "critical thinking," without venturing into the territory of neural mechanisms or brain structures. A note on memory: Memory is extraordinarily complex, and this model does not attempt to explain it fully. For our purposes, we simply acknowledge that we can store memories in different ways, we can access memories through different pathways, and we sometimes lose access to certain memories. We might theorize about different types of memory representation, but that's as far as we will go in this examination. ## Foundations of Cognitive Processing At the most fundamental level, we exist in a world of stimuli. Our senses continuously capture information from our environment—light waves strike our retinas, sound waves vibrate our eardrums, molecules activate our olfactory receptors, surfaces make contact with our skin. For simplicity in our model, we treat these stimuli as discrete units, though in reality they form a continuous stream of information. When these discrete stimuli enter our mind, they become information units, or simply "units." A unit is any discrete piece of information in our mind—anything we can perceive, think about, or recall. Everything in our mind can be conceptualized as a unit. This abstraction allows us to talk about mental content in a consistent way, regardless of whether we're discussing visual perceptions, abstract concepts, emotional states, or memories. Our mind provides a workspace to hold these units in short-term memory. This workspace has limited capacity, which explains why we can only consciously think about a certain number of things at once. The workspace serves as a stage where units can be manipulated, combined, compared, and transformed. It is where our conscious cognitive operations take place. Beyond processing sensory input, we have the ability to generate units independently, without external stimulation. We can imagine things that don't exist, recall things that aren't present, and create entirely new mental constructs. This generative capability is crucial for planning, creativity, and abstract thinking. Among the units we can generate, one type stands out as particularly important: the relation unit. A relation unit, as the name suggests, establishes a connection between two or more other units. These relations can represent any type of connection—similarity, causality, sequence, hierarchy, or any other way in which information can be linked. The simplest and most universal relation might be the similarity relation: "this unit is similar to that unit." Relations are arbitrary in the sense that they can represent any type of connection the mind needs to establish. This flexibility allows us to create complex networks of meaning from simple elements. Consider how we understand a simple object like a cup. We relate its physical appearance to its function, its material composition, our memories of using similar cups, the concept of containers generally, and countless other aspects of our knowledge and experience. All of these connections are formed through relation units. When we relate multiple units together, we begin to perceive this connected set of units as a cohesive whole—as a single higher-order unit. For example, when we see the parts of a face (eyes, nose, mouth, etc.), we don't just perceive them as separate elements but as components of a unified face. This process of relating units and perceiving them as a new whole can be understood in two ways: either as a unique cognitive operation, or as the generation of a new unit that relates to all its component units through hierarchical relations. This ability to create hierarchical structures of units is fundamental to how we organize information. It allows us to manage complexity by chunking information into meaningful wholes. When we relate two units, we create a connection between them, forming a simple structure. As we add more connections and more units, these structures become increasingly complex and meaningful. Now we arrive at a central concept in this model: the model itself. When we have a connected set of units, we call this a model. Models are the bedrock of this cognitive framework—they allow us to represent, manipulate, and understand information in our minds. A model is essentially a structure of interconnected units that represents something in our world or in our imagination. Models are important because they provide a way to visualize and manipulate information mentally. Consider how we think about problems. A problem is itself a model—a set of connected units that includes units describing why this is a problem, units defining the boundaries of the problem, and units specifying what a solution must satisfy. By conceiving of a problem as a model, we can examine its components, restructure it, and work toward a solution. This framework gives us our first requirement for understanding: Understanding occurs when every unit in our mental workspace is related to at least one other unit. When all units are connected in meaningful ways, we experience the sensation of comprehension. Isolated, unrelated units create cognitive dissonance and confusion. The mind naturally seeks to establish connections between units, to integrate them into existing models or to form new models that accommodate them. Yet we clearly don't process each individual unit of information every time we think or perceive. That would be cognitively overwhelming and inefficient. Instead, we rely on higher-level processes of abstraction and recognition, which we'll explore next. ## The Concept of Models in Cognition Models are the cornerstone of this cognitive framework. A model is a set of related units that forms a cohesive representation of something—whether that's a physical object, an abstract concept, a process, or a situation. Models have boundaries, which define what is included in the model and what remains outside it. These boundaries are often fluid and context-dependent, expanding or contracting based on our current focus and needs. Models differ fundamentally from isolated facts. A fact might be represented as a single unit or a simple relation between units, but a model encompasses numerous interrelated units that collectively represent something more complex. Models provide context, meaning, and structure that isolated facts cannot convey. We are, essentially, model thinkers. Our cognition doesn't operate primarily on isolated bits of information but on structured models that represent aspects of our experience and understanding. When we think about an apple, we don't just activate a single unit labeled "apple"; we engage a complex model that includes units for its typical appearance, taste, texture, uses, categorical relationships to other fruits, and personal memories associated with apples. Physical objects provide some of the most straightforward examples of how we construct models. When we perceive or think about an object like an apple, our mind creates a model that includes not just its sensory properties but also predictions about how it will behave. We know that if dropped, it will fall; if cut, it will reveal flesh of a certain color and texture; if left too long, it will decay. These behavioral predictions are integral parts of our models of physical objects. Our models of physical reality extend beyond specific objects to encompass abstract properties like space, time, and causality. Space isn't just an empty container but a complex model that includes concepts like distance, position, orientation, and movement. Time isn't just a linear progression but a model that includes duration, sequence, simultaneity, and change. Forces aren't directly observable but are modeled as invisible influences that explain and predict physical behaviors. These fundamental models underpin our understanding of physical reality and allow us to make predictions about how the world works. Beyond physical reality, we construct models of abstract concepts and mental constructs. Concepts like "justice," "democracy," or "happiness" are not directly perceivable through our senses, yet we build complex models that allow us to think about and discuss them. These conceptual models might include examples, counter-examples, defining features, related concepts, emotional associations, and various instantiations we've encountered. The richness of these models explains why concepts can be so nuanced and context-dependent—they're not simple definitions but complex structures of interrelated units. Not all models are visual or consciously accessible. We also construct non-visual and intuitive models that operate largely beneath our awareness. Feelings, intuitions, aesthetic judgments, and creative insights often involve models that we cannot easily articulate or visualize. The "spark" of creativity or the intuitive sense that something is right or wrong may emerge from complex modeling processes that operate outside our conscious awareness but nevertheless influence our thinking and behavior. Even mental operations themselves can be understood as models. The process of comparing two ideas involves a comparative model; the act of making a decision engages a decision-making model; the experience of remembering activates a memory retrieval model. These operational models guide how we perform various cognitive tasks without requiring conscious attention to their mechanics, much as we don't think about the specific muscle movements involved in walking or riding a bicycle. The power of models lies in their ability to organize information in ways that facilitate understanding, prediction, and interaction with our world. By structuring information into coherent models rather than processing isolated bits of data, our cognitive system achieves efficiency and depth of understanding that would otherwise be impossible. Models allow us to see patterns, make inferences, and navigate complexity with relative ease. ## Cognitive Operations with Models Having explored what models are, we now turn to what we do with them—how we create, store, retrieve, recognize, and generate models in our cognitive processes. When we encounter new information or experiences, we create models to represent them. These models, once formed, are stored in our memory. The storage process is not like filing away a document in a cabinet; it's more like integrating new information into an existing network. When we store a model, we're storing both the specific instance of that model and its connections to other models and units in our cognitive landscape. Storage is selective and influenced by numerous factors, including emotional significance, relevance to our goals and interests, repetition, and distinctive features. Not everything we perceive or think about gets stored with equal fidelity or permanence. Some models fade quickly from memory, while others become deeply entrenched and readily accessible. As we accumulate similar models in our memory, a crucial cognitive function called abstraction comes into play. Abstraction is the process of extracting common elements from a set of similar models to create a more general, abstract model. This abstract model retains the essential features shared by the individual instances while omitting specific details that vary across instances. For example, after encountering many different apples, our mind abstracts the common features to form a general concept of "apple-ness." This abstract model includes the typical characteristics of apples but leaves many details underspecified. The abstract model serves as a template or prototype that represents the category as a whole rather than any specific instance. Abstract models are still composed of units and relations, just like instance models. However, they're more flexible and less detailed. Instead of specifying exact colors, shapes, or sizes, abstract models often include ranges or typical values. They might leave certain aspects entirely unspecified, creating "slots" that can be filled in different ways in different instances. This flexibility allows abstract models to represent categories, concepts, and patterns that encompass many individual instances. Once we've stored models, both instances and abstractions, we need to be able to retrieve them. Retrieval is the process of accessing stored models and bringing them into our conscious workspace. This process can be triggered by external cues (seeing something that reminds us of a past experience), internal cues (thoughts or feelings that activate associated models), or deliberate search efforts (trying to remember someone's name). Retrieval is not always successful or accurate. We may retrieve partial models, combine elements from different models, or fail to retrieve relevant models altogether. The accessibility of stored models depends on factors such as recency, frequency of use, emotional significance, and contextual cues. Models that are frequently accessed or strongly connected to our current context are more readily retrieved than those that are rarely used or contextually distant. When we encounter new stimuli, our cognitive system doesn't start from scratch to build understanding. Instead, it attempts to match the incoming information against our existing abstract models. This pattern-matching process, which we can call recognition, allows us to quickly identify and categorize new experiences based on their similarity to familiar patterns. Recognition operates largely automatically and often unconsciously. When you see a new type of fruit that resembles an apple, your mind automatically activates your abstract apple model and checks for matches. If the match is close enough, you recognize it as a type of apple. If the match is partial, you might categorize it as "apple-like" or search for a better matching category. If no good match is found, you may need to generate a new model. This pattern-matching process is extraordinarily efficient. Instead of having to relearn what each new object is, how it behaves, and how to interact with it, we can leverage our existing models to make rapid inferences. When you recognize something as an apple, you immediately have access to a wealth of knowledge about its likely properties and behaviors without having to discover them anew. Sometimes, however, we encounter situations for which we don't have adequate existing models, or we need to combine or modify existing models to address new challenges. This is where model generation comes into play. Generating a model involves creating new units and relations, often by combining elements from existing models in novel ways. Model generation can be deliberate, as when we consciously try to solve a problem or understand a new concept, or it can occur spontaneously, as in creative insights or dreams. Many aspects of model generation happen unconsciously—we often aren't aware of how our mind constructs new models until they emerge into consciousness as "aha" moments or creative ideas. To illustrate these processes in action, consider what happens when you encounter an unfamiliar object. Each visual feature of the object registers as a stimulus and enters your mind as a unit. Your brain quickly begins to relate these units, identifying edges, shapes, textures, and other patterns. It then attempts to match this emerging model against your existing abstract models. If it finds a match—recognizing the object as a type of chair, for instance—it can immediately apply all the associated knowledge from your chair model. If no good match is found, it will work to generate a new model, perhaps by combining elements of existing models ("it's like a chair but with wheels and a lever"). This interplay of model storage, abstraction, retrieval, recognition, and generation forms the core of how we process and understand information. These operations allow us to build on past experience, recognize patterns, make predictions, and adapt to new situations with remarkable efficiency and flexibility. ## Using Models Having examined how we create, store, and retrieve models, we now turn to how we apply them to understand reality, direct our attention, and navigate the complexities of information processing. Our perception of reality itself can be understood through the lens of modeling. We don't experience reality directly or completely; we experience a model of reality constructed by our cognitive system. This model is necessarily simplified, selective, and shaped by our existing knowledge, expectations, and limitations. The units and connections that form our model of reality represent only a fraction of what might be perceived or known, filtered through the lens of what our cognitive system deems relevant or meaningful. Given this perspective, we can see that reality itself might be thought of as consisting of units and connections. After all, if this is how we understand things, then our very conception of reality is shaped by this framework. This doesn't mean reality objectively consists of discrete units and relations, but that our understanding of reality is structured this way. The map is not the territory, but the map is all we have direct access to. Attention plays a crucial role in how we use models. Our cognitive workspace has limited capacity, so we cannot process all available information simultaneously. Attention is the mechanism that selects which information enters our workspace and becomes available for conscious processing. Attention is directed by multiple factors. We naturally pay attention to what seems important or relevant to our current goals and concerns. When we don't know what's important in a new situation, we pay attention to what interests us or what stands out perceptually. Attention can be captured automatically by sudden changes, unusual patterns, or emotionally significant stimuli. It can also be directed voluntarily, as when we deliberately focus on a specific aspect of our experience or a particular problem we're trying to solve. The allocation of attention profoundly influences which models we activate and how we use them. By directing attention to certain aspects of a situation and away from others, we shape which connections are formed, which patterns are recognized, and ultimately, what we understand and remember. Two people with different attentional focuses may construct entirely different models of the same situation, leading to different understandings, decisions, and memories. This selective nature of attention leads us to consider bias. In this model, bias can be understood as a systematic failure to account for all relevant units of information. When we focus exclusively on certain aspects of a situation while ignoring others that are equally relevant, we create a distorted model that may lead to flawed understanding or poor decisions. Bias is essentially a misuse of attention—directing it consistently toward some types of information and away from others, often without awareness that we're doing so. Our models become skewed, emphasizing certain connections and patterns while minimizing or ignoring others. These biases can be reinforced over time as we repeatedly attend to the same kinds of information and activate the same patterns of connection. Understanding, within this framework, can be defined as the state that occurs when we have successfully connected all units of information in our current workspace into a coherent model. When every unit is meaningfully related to the whole, we experience the "click" of comprehension. Isolated, unconnected units create cognitive discomfort and drive us to seek connections until everything fits together. The quality of understanding depends on the quality of the model we construct. A sophisticated, nuanced model with many levels of connection will yield a deeper understanding than a simplistic model with few connections. A model that accurately represents the relevant aspects of what it models will yield truer understanding than one that distorts or omits key elements. A model that connects to many other models in our knowledge base will be more meaningful and useful than one that remains isolated. Learning, in this framework, is fundamentally associative. We learn by creating new models and by connecting new information to existing models. The more extensively a model is connected—both internally among its component units and externally to other models—the better it is learned and the more accessible it becomes. This view explains why prerequisites are so important in learning. Without the necessary foundation models, we cannot form meaningful connections for new information. Trying to learn advanced concepts without understanding the basics is difficult precisely because we lack the models to which new information can be connected. Learning proceeds most effectively when new information can be integrated into a rich network of existing knowledge. Thinking itself can be understood as the process of creating, retrieving, modifying, and connecting models. When we think about a problem, we're manipulating models in our workspace—breaking them apart, recombining them, comparing them to other models, and generating new ones. The quality of our thinking depends largely on the quality and variety of models we can bring to bear, and on our skill in manipulating and connecting them. Creative thinking often involves combining models from different domains to generate novel insights or solutions. Analytical thinking involves breaking down complex models into their component units and examining the relations between them. Critical thinking involves evaluating models for consistency, coherence, and correspondence with other trusted models. Each type of thinking represents a different way of working with models in our cognitive workspace. ## Applications and Implications This descriptive model of cognition offers valuable insights into numerous aspects of human experience and has practical applications in various domains. Let's explore some of these applications and their implications. In education, this model suggests that effective teaching involves helping students build robust, well-connected models. Rather than focusing solely on transmitting information, educators should emphasize how new knowledge connects to existing knowledge and how different concepts relate to each other. Learning activities that require students to actively construct models—such as concept mapping, problem-solving, project-based learning, and peer teaching—are likely to be more effective than passive information reception. The model also explains why learning often follows a non-linear trajectory. Initial learning in a new domain can be slow and effortful because students must build new models from scratch. As these foundational models develop, learning can accelerate dramatically because new information can be connected to existing structures. Eventually, learning may plateau as the basic models are well-established and only refinements or specialized extensions remain to be added. In problem-solving and creativity, this model highlights the importance of having diverse models available and being able to connect them in novel ways. Creative insights often come from recognizing patterns or relationships between seemingly unrelated domains—essentially, connecting models that aren't typically connected. Problem-solving frequently involves restructuring our model of the problem, either by adding new units and relations or by activating different abstract models that suggest alternative approaches. This perspective suggests that creativity can be enhanced by deliberately expanding one's repertoire of models through diverse experiences, interdisciplinary learning, and exposure to different perspectives. It also suggests that effective problem-solving might involve consciously activating different models and exploring how they might apply to the current situation, rather than fixating on the first approach that comes to mind. In communication, this model helps explain why the same message can be understood so differently by different people. Each person interprets information through the lens of their existing models, making connections that align with their knowledge, beliefs, and experiences. Effective communication requires considering the models the audience likely possesses and presenting information in ways that connect to these models appropriately. This insight has implications for everything from education and marketing to conflict resolution and cross-cultural communication. When communication fails, it's often because the sender has assumed the receiver has models that they don't actually possess, or because the message activates different models than intended. Clarifying shared models and explicitly building bridges between different perspectives can significantly improve understanding. In artificial intelligence and cognitive computing, this model offers a framework for designing systems that process information more like humans do. Rather than treating knowledge as a collection of discrete facts or procedures, AI systems might be designed to construct, store, retrieve, and manipulate models with interconnected units and relations. Such systems might better capture the contextual, associative nature of human knowledge and the flexibility of human thinking. This approach aligns with recent developments in AI, such as neural networks and deep learning, which build distributed representations rather than symbolic rules. However, it suggests that truly human-like AI would need mechanisms not just for pattern recognition but for active model construction, abstraction, and creative recombination—capabilities that remain challenging for current systems. In personal development and metacognition, this model provides a vocabulary for reflecting on our own thinking processes. By recognizing when we're struggling to form adequate models, when we're applying inappropriate models, or when our attentional biases are distorting our understanding, we can take steps to improve our cognitive processes. Practices like mindfulness, reflective journaling, and deliberate exposure to diverse perspectives can help us become more aware of our modeling processes and more intentional in how we direct them. The model also has implications for understanding cognitive differences and neurodiversity. People may differ in their modeling processes in various ways—in the types of units they most readily process, in how they form and use relations, in their abstraction tendencies, in their attentional patterns, and so on. What might appear as deficits from one perspective might represent different, but equally valid, ways of constructing and using models. This view encourages appreciation for cognitive diversity and adaptation of educational and work environments to accommodate different modeling styles. In mental health, this model offers a perspective on how cognitive distortions and emotional difficulties might arise from problematic modeling patterns. Anxiety might involve overactivation of models that predict threat or harm; depression might involve pervasive activation of negative self-models and pessimistic world-models; obsessive-compulsive patterns might reflect difficulty in determining when a model is complete or accurate enough to stop checking. Therapeutic approaches might aim to help individuals recognize these patterns and develop more balanced, flexible modeling processes. While this is not a clinical model and should not replace established therapeutic frameworks, it provides a way of thinking about cognitive aspects of mental health that complements other approaches. It aligns particularly well with cognitive-behavioral perspectives that emphasize how our interpretations shape our experiences and how changing our thought patterns can lead to emotional and behavioral changes. ## Limitations and Future Directions Like any model, this descriptive framework of cognition has limitations and leaves many questions unanswered. Recognizing these boundaries helps us use the model appropriately and identifies opportunities for future development. First, the model operates at a descriptive, functional level rather than explaining underlying mechanisms. It doesn't address how units and relations are physically implemented in the brain, how models are encoded in neural structures, or how cognitive operations map to neural processes. This limitation is deliberate—the model aims to provide a useful vocabulary for discussing cognitive phenomena without making claims about their neurological basis. Nevertheless, connections to neuroscience could enrich the model and provide testable predictions. Second, the model doesn't fully account for the dynamic, temporal aspects of cognition. While it acknowledges that models change over time and that processing unfolds sequentially, it doesn't provide detailed descriptions of how these temporal dynamics work. How quickly do models form and change? How do sequential dependencies in processing affect model construction? How do different timescales of cognitive operation interact? These questions remain largely unexplored within the current framework. Third, the role of emotion in this model needs further development. Emotions clearly influence which models we activate, how we direct attention, what we remember, and how we evaluate information. While emotions can be represented as units and incorporated into models, the current framework doesn't fully capture their pervasive influence on cognitive processing. A more developed version might integrate emotional and cognitive aspects more thoroughly. Fourth, the social dimensions of cognition are not extensively addressed. Human cognition doesn't occur in isolation but is profoundly shaped by social interaction, cultural context, and shared knowledge. Models are often socially constructed and transmitted. A fuller account would explore how individual cognitive models relate to collective knowledge structures and how social processes influence model formation and use. Fifth, the current model focuses primarily on explicit, conscious cognition. While it acknowledges unconscious processes, it doesn't fully explore the relationships between conscious and unconscious modeling or how implicit knowledge structures influence explicit reasoning. These relationships are complex and not fully understood even in more specialized cognitive theories, but they represent an important area for future development. Despite these limitations, the model offers a useful framework for thinking about cognition and for discussing mental processes in more precise terms than everyday language typically allows. Its value lies not in its completeness or neurological accuracy, but in its ability to provide a structured vocabulary for exploring how we think, understand, and learn. Future development of this model might proceed in several directions. Connections to empirical research in cognitive psychology and neuroscience could provide validation for certain aspects of the model and suggest refinements to others. Applications in specific domains like education, communication, or artificial intelligence might lead to more detailed elaborations of how the model applies in particular contexts. Integration with other theoretical frameworks, from embodied cognition to predictive processing, could enrich the model and situate it within broader discussions of how minds work. While the model is presented here as a descriptive framework rather than a scientific theory, it does suggest potential avenues for empirical investigation. How do people actually form and use models in different contexts? What factors influence which models are activated and how they're applied? How do individual differences in modeling processes relate to differences in learning styles, problem-solving approaches, or creative tendencies? Such questions could be explored through experimental studies, observational research, or computational modeling. ## Conclusion The descriptive model of cognition presented here offers a framework for understanding how we think, learn, and understand. By conceptualizing cognition in terms of units, relations, and models—and the operations we perform with them—it provides a vocabulary for discussing mental processes that bridges the gap between everyday language and specialized scientific terminology. At its core, this model suggests that we are model builders and model users. We navigate our world by constructing mental representations that capture the patterns, relationships, and structures we perceive. These models allow us to organize information, make predictions, solve problems, and communicate with others. The quality of our thinking depends largely on the quality of our models and our skill in manipulating them. The model emphasizes several key aspects of cognition. First, it highlights the constructive nature of understanding—we don't just receive information passively but actively build models to make sense of it. Second, it underscores the associative character of knowledge—understanding involves connecting new information to existing models and relating different models to each other. Third, it acknowledges the selective nature of attention and the potential for bias that this selectivity creates. Fourth, it recognizes the crucial role of abstraction in learning—our ability to extract patterns from specific instances and to apply these patterns to new situations. This framework has practical implications for how we approach education, problem-solving, communication, and personal development. It suggests strategies for enhancing learning, creativity, and understanding based on how our cognitive processes actually work. By becoming more aware of our modeling processes, we can become more intentional in how we direct them and more effective in how we use them. While the model has limitations and leaves many questions unanswered, it provides a starting point for thinking more clearly about thinking itself. In a world of increasing information complexity and accelerating change, understanding our own cognitive processes becomes increasingly valuable. This model offers one approach to gaining that understanding—not by making claims about neural mechanisms or computational algorithms, but by providing a descriptive framework that aligns with our lived experience of thinking, learning, and knowing. Ultimately, this model is itself a model—a simplified representation that captures some aspects of cognition while necessarily omitting others. Its value lies not in its perfect correspondence to cognitive reality but in its utility for helping us think about thinking. Like all good models, it should be used where it's helpful, modified where it's inadequate, and supplemented with other perspectives where appropriate. In the spirit of model thinking itself, it's a tool for understanding, not a definitive description of what understanding is.