Epistemology

Learning Objectives

By completing this book, you will be able to:

1. Explain what epistemology is and why the quality of problem formulation depends on the epistemic assumptions of the person formulating it.

2. Distinguish between propositional, procedural, and tacit knowledge, and recognize which one dominates when a user articulates an intention.

3. Apply Popper’s criterion of falsifiability to evaluate whether a hypothesis or problem formulation is genuinely informative or irrefutably empty.

4. Analyze how paradigms, according to Kuhn, and research programs, according to Lakatos, determine which problems a person can even perceive and formulate.

5. Recognize the implicit epistemic assumptions in any problem formulation and make them explicit for critical evaluation.

6. Apply epistemic humility as a design tool in interaction with intelligent systems, formulating problems that recognize the limits of one’s own knowledge.

1. Introduction: What Is Epistemology?

Epistemology is the branch of philosophy that studies knowledge: what it is, how it is obtained, what its limits are, and how we distinguish genuine knowledge from mere opinion, unfounded belief, or the illusion of understanding. Its name comes from the Greek episteme, meaning knowledge, and logos, meaning study. It is, literally, the study of how we know what we believe we know.

“Imagine that a manager says with complete confidence: ‘Our customers want lower prices.’ It sounds like knowledge. But how does he know? Did he conduct a rigorous survey? Did he infer it from three complaints he received last week? Does he assume it because he himself would want lower prices? Or is he simply repeating something he heard at a conference? The statement is the same in all four cases, but the epistemic quality, meaning the strength of the knowledge supporting it, is radically different. And that difference will determine whether the problem he formulates from that belief is well framed or a castle built on sand.”

Epistemology is fundamental to the Architecture of Intention because every formulation of a problem rests on assumptions of knowledge. When a person says, “I need X,” they are implicitly claiming: “I know what my situation is, I know what causes it, I know what would solve it, and I know that X is the solution.” Each of those “I know” statements is an epistemic claim that may be true, partially true, or completely wrong.

“An entrepreneur tells an AI: ‘I need a social media strategy to increase my sales.’ Inside this request there are at least four epistemic assumptions: (1) I know that my sales need to increase, although perhaps I need to reduce costs; (2) I know that social media is the right channel, although perhaps it is not; (3) I know that a strategy will solve the problem, although it may be a product problem, not a marketing problem; and (4) I know enough about my situation to formulate the right request, even though the WYSIATI effect from Module I tells us that I probably do not. Epistemology teaches us to examine each of those ‘I know’ statements before acting on them.”

In Module I, you learned how the mind thinks, through cognitive psychology; how it is wired, through neuroscience; how it takes shortcuts, through biases; and what drives it, through motivation. All of that describes the mechanism of intention. Now you need a discipline that evaluates the quality of that intention. That discipline is epistemology.

1.1 Historical Context: From Plato to Modern Science

The question “What is knowledge?” is at least 2,500 years old. Plato, in his dialogue Theaetetus, proposed the classic definition: knowledge is justified true belief. In other words, for something to count as knowledge, you need three things: you must believe it, it must be true, and you must have good reasons to believe it.

“If you believe there is traffic on the highway because you saw it on a real-time navigation app, that is justified true belief: you believe it, it is true, and you have evidence. If you believe there is traffic because ‘there is always traffic at this time’ and it turns out there is, that is true belief but weakly justified: you were right, but by heuristic, not by current evidence. If you believe there is no traffic because you do not want there to be traffic, that is desire disguised as knowledge, neither justified nor necessarily true.”

This definition dominated Western philosophy for centuries. But in 1963, the philosopher Edmund Gettier published a three-page article that destroyed it. Gettier showed, with elegant counterexamples, that you can have a belief, that belief can be true, and it can be justified, and yet it may still fail to constitute knowledge because the justification and the truth coincide by pure chance.

“Imagine that you look at the wall clock and it says 3:15. You believe it is 3:15. It turns out that it really is 3:15. And your justification, looking at a clock, seems reasonable. But the clock stopped yesterday at 3:15. Your belief is true and justified, but it is not knowledge: you were right by coincidence. Gettier problems show that knowledge is more fragile than it appears.”

Gettier problems are not academic curiosities. They are the formal version of something that happens constantly in the formulation of intentions: people who have the right answer for the wrong reasons and therefore cannot replicate their success or diagnose their failure. For the Architecture of Intention, the lesson is clear: it is not enough for the user to be right. Why they are right matters, because only then can they correct course when circumstances change.

2. Types of Knowledge: Knowing That, Knowing How, Knowing Without Knowing

Not all knowledge is of the same kind. When we talk about “what the user knows” when formulating an intention, we need to distinguish between at least three forms of knowledge because each one is articulated differently and creates different challenges for problem formulation.

2.1 Propositional Knowledge: Knowing That

Propositional knowledge is knowledge of facts: knowing that the Earth revolves around the Sun, that the interest rate is 5%, that your company had losses last quarter. It is knowledge that can be expressed in declarative propositions, statements that are true or false.

“A chief financial officer who says ‘Our operating costs increased by 12% this year’ is expressing propositional knowledge. It can be verified, quantified, and communicated without ambiguity. It is the easiest type of knowledge to include in a problem formulation because it can be articulated explicitly.”

Propositional knowledge travels best between people and systems. An AI system can process “our costs increased by 12%” without difficulty. The problem is that most human intentions are not composed exclusively of propositional knowledge. They are contaminated, or rather enriched, by the other two types.

2.2 Procedural Knowledge: Knowing How

Procedural knowledge is knowledge of how to do something: knowing how to ride a bicycle, negotiate a contract, write code, or read a customer’s body language. It is practical knowledge, often acquired through experience, and crucially, it cannot always be articulated in propositions.

“An experienced salesperson ‘knows’ when a customer is about to close a deal. If you ask how they know, they will probably say something vague: ‘You can see it in their face,’ ‘their tone changes,’ or ‘they start asking questions about delivery times.’ But they cannot give you a precise algorithm. They know how to do it without being able to fully explain what they know. This creates a serious problem when they try to formulate a request to an AI system: ‘Help me identify customers ready to buy’ omits exactly the knowledge that would make that request useful.”

Gilbert Ryle, the British philosopher, made this distinction famous in his book The Concept of Mind (1949). Ryle argued that reducing all knowledge to propositions is a fundamental error, what he called “the intellectualist legend.” Knowing how to ride a bicycle is not knowing a proposition about balance; it is an irreducibly different type of knowledge.

2.3 Tacit Knowledge: Knowing Without Knowing That You Know

Michael Polanyi, a Hungarian-British physical chemist and philosopher, introduced the concept of tacit knowledge with a phrase that became famous: “We know more than we can tell.” Tacit knowledge is knowledge we possess but cannot fully articulate, and we are not always even conscious of possessing it.

“An experienced doctor who looks at a patient and ‘feels’ that something is wrong before the tests confirm it is using tacit knowledge. Thousands of previous patients have trained pattern recognition processes that operate below the threshold of consciousness. The doctor cannot articulate exactly what they saw, perhaps a subtle color in the skin, a breathing pattern, or a microexpression of pain. This knowledge is real, valuable, and often accurate. But it is invisible to any system that depends on the user explaining what they know.”

For the Architecture of Intention, tacit knowledge is simultaneously the most valuable and the most problematic. It is the most valuable because it contains years of compressed experience. It is the most problematic because, by definition, the user cannot include it explicitly in their formulation. When an expert says, “I do not know why, but this does not convince me,” their tacit knowledge is speaking. Ignoring it because it does not come articulated as a proposition would be a serious epistemic mistake.

3. Falsifiability: Popper’s Revolution

Karl Popper (1902-1994) is perhaps the most influential philosopher of science of the twentieth century. His central contribution, the criterion of falsifiability, fundamentally changed how we distinguish scientific knowledge from pseudoscience, and it has direct implications for how we should formulate problems and intentions.

3.1 The Problem of Demarcation

Popper began with a simple but profound question: what separates science from non-science? Before Popper, the dominant answer was verifiability: a theory is scientific if it can be verified through observations. But Popper noticed a devastating logical problem with this criterion.

“If I say ‘all swans are white,’ how many white swans do I need to observe in order to verify the claim? A million white swans do not prove it, because there could always be a black one I have not seen. But a single black swan is enough to refute it. The asymmetry is total: no number of confirmations can prove a universal theory, but a single refutation can destroy it. Popper concluded that verification is impossible; what is possible is falsification.”

Popper proposed that a theory is scientific not because it can be proven true, but because it can be proven false. The criterion is not “Can I confirm this?” but “What observation would make me abandon this?” A theory that nothing can refute says nothing about the world. It is irrefutably empty.

“Popper used an eloquent contrast. Einstein’s theory of relativity made specific and risky predictions, such as that the light from a star would bend when passing near the Sun, which could be measured and, if they failed, would destroy the theory. In contrast, Popper observed that certain psychological theories could explain any behavior after the fact: if the patient improved, the theory worked; if the patient worsened, it also worked. If a man saved a child, it was sublimation; if he harmed the child, it was repression. The theory could never be wrong, and therefore it could never teach us anything new.”

3.2 Falsifiability as a Formulation Tool

The criterion of falsifiability is not only for scientists. It is an extraordinarily useful tool for anyone who formulates problems, hypotheses, or intentions. The Popperian question par excellence is: “What evidence would make me change my mind?” If you cannot answer that question, you do not have a hypothesis. You have a dogma.

“A manager claims: ‘Our sales team needs more training.’ Let us apply Popper: what observation would refute this claim? If the answer is ‘nothing, because there is always room for improvement,’ then the claim is not really a diagnosis; it is an irrefutable commonplace. A falsifiable formulation would be: ‘I believe that if we train the team in closing techniques, the conversion rate will increase by at least 10% in three months.’ Now there is something to measure, a deadline, and a clear criterion for failure.”

In interaction with AI systems, lack of falsifiability manifests in requests that have no defined criterion of success. If the user cannot describe what an unsatisfactory response would look like, they also cannot really describe what they want. They are delegating to the system not only execution but the definition of the problem itself.

“Compare these two requests: (1) ‘Make me a good logo for my company’ versus (2) ‘I need a minimalist, monochrome logo that works at small size and evokes professional trust; it must not look like a medical logo or use script typography.’ The second is falsifiable: each criterion can be evaluated and declared met or unmet. The first is irrefutable: what does ‘good’ mean? If you do not like the result, is it because the system failed or because you never defined what you wanted?”

3.3 Corroboration, Not Confirmation

Popper introduced a subtle but crucial distinction between confirmation and corroboration. A theory is not “confirmed,” since that would be verification, which we have already seen is impossible for universal theories. Instead, it is “corroborated” when it survives serious attempts to refute it. The difference is more than semantic: a confirmed theory is treated as definitive; a corroborated theory is treated as provisionally accepted but always open to revision.

“A marketing team launches a campaign and sales increase. Does this confirm that the campaign worked? Not necessarily. Perhaps sales rose because of seasonality, because of a competitor’s mistake, or because of a factor no one measured. What the team can say is this: the hypothesis that the campaign works has not yet been refuted. It is corroborated, not confirmed. That epistemic humility changes how you interpret the result and what you decide to do next.”

The distinction between confirmation and corroboration connects directly with the confirmation bias studied in Module I. You will remember that people seek evidence that confirms their beliefs and ignore evidence that contradicts them. Popper offers us the antidote: instead of seeking confirmation, seek refutation. If your hypothesis survives genuine attempts to destroy it, you have something valuable. If you have only searched for supporting evidence, you do not have knowledge. You have self-complacency.

4. Paradigms: The Structure of Scientific Revolutions

If Popper taught us how science should ideally work, Thomas Kuhn (1922-1996) showed us how it actually works. His book The Structure of Scientific Revolutions (1962) is one of the most cited academic texts in history, and for good reason: it radically changed how we understand the relationship between knowledge, community, and change.

4.1 What Is a Paradigm?

A paradigm, in Kuhn’s sense, is much more than a theory. It is the complete set of beliefs, values, techniques, vocabulary, and assumptions shared by a community. It is the lens through which that community sees the world, defines what counts as a legitimate problem, what methods are acceptable, and what answers are satisfactory.

“Before Copernicus, the astronomical paradigm placed the Earth at the center of the universe. This was not just a theory: it determined which observations were considered important, which problems were worth solving, such as adjusting epicycles to explain the retrograde motion of planets, and which questions were unthinkable, such as ‘What if the Earth moves?’ The paradigm did not only answer questions. It determined which questions could be formulated.”

For the Architecture of Intention, the concept of paradigm is extraordinarily useful because it explains something that the mental models from Module I already suggested but did not fully resolve: why equally intelligent and well-intentioned people may be unable to see certain problems or conceive certain solutions. It is not that they lack information; it is that their paradigm makes certain problems literally invisible.

“In the business world, think about how the paradigm of ‘operational efficiency’ dominated management for decades. If your paradigm says success comes from optimizing internal processes, the problems you see are problems of costs, time, and waste. A problem such as ‘our customers need an emotional experience, not a cheaper product’ is invisible within that paradigm, not because it does not exist, but because the paradigm has no vocabulary to formulate it.”

4.2 Normal Science vs. Revolutionary Science

Kuhn distinguishes two phases in scientific activity. Normal science is routine work within an accepted paradigm: solving puzzles using the tools and assumptions of the current paradigm. Most scientific work, and by analogy most problem solving in any domain, is normal science: operating within the established framework.

“A development team optimizing an existing product is doing ‘business normal science’: it works within the paradigm that the current product is basically correct and only needs incremental improvements. The problems it formulates are of the type ‘How do we reduce load time?’ and ‘How do we improve the conversion rate?’ All valid questions within the paradigm. But the question ‘What if the entire product is solving the wrong problem?’ cannot be formulated from normal science.”

Anomalies are observations that the current paradigm cannot explain satisfactorily. At first, they are ignored or rationalized. But when enough anomalies accumulate, the paradigm enters crisis. And when the crisis is severe enough, a scientific revolution can occur: the old paradigm is replaced by a new one that explains the anomalies and reorganizes the entire field.

“Kodak dominated photography for decades. Its paradigm was ‘photography is chemistry: film, development, printing.’ When digital photography began to emerge, it was not an anomaly within its paradigm. It was a completely different paradigm. Kodak could not formulate the problem correctly, ‘How do we transition to digital?’ because its paradigm told it that digital was an inferior niche. The anomalies accumulated until it was too late.”

4.3 Incommensurability: When Paradigms Do Not Speak to Each Other

Kuhn introduced a disturbing concept: incommensurability. Two paradigms cannot be directly compared because they do not share the same criteria, the same vocabulary, or even the same problems. It is like trying to decide whether a symphony is “better” than a mathematical equation. The question does not make sense because they belong to different frameworks of evaluation.

“In a company, the finance department and the design department may operate under incommensurable paradigms. Finance measures success in ROI, margins, and cash flow. Design measures success in user experience, aesthetic coherence, and emotional satisfaction. When both formulate ‘what our product needs,’ they ask questions that sound the same but mean radically different things. An AI system that does not detect this incommensurability will produce answers that satisfy one paradigm and frustrate the other.”

5. Research Programs: Lakatos’s Middle Ground

Imre Lakatos (1922-1974), a Hungarian mathematician and philosopher, offered a synthesis between Popper’s critical rationalism and Kuhn’s historical contextualism. Where Popper said “a refutation should destroy the theory” and Kuhn said “refutations are ignored until there is a better alternative,” Lakatos proposed something more nuanced: the concept of a research program.

5.1 Hard Core and Protective Belt

Lakatos observed that real scientific theories are not isolated propositions that are refuted one by one, as Popper’s simplest version suggests. They are complex systems with a layered structure. At the center there is a hard core: the fundamental assumptions that the community decides not to abandon. Around it there is a protective belt of auxiliary hypotheses that can be modified to absorb anomalies without touching the core.

“Think of a startup. Its hard core could be: ‘people want to manage their personal finances from their mobile phone.’ This assumption is not touched. Around it there are auxiliary hypotheses: ‘the interface must be minimalist,’ ‘the business model is freemium,’ and ‘our typical user is 25-35 years old.’ When something does not work, downloads are low or retention drops, the startup modifies the protective belt: it changes the interface, adjusts the pricing model, or redefines the target audience. All of this happens without questioning the hard core. It is rational to do this, until it stops being rational.”

5.2 Progressive vs. Degenerative Programs

Lakatos’s most practical contribution is his criterion for distinguishing when it is rational to persist and when it is time to change. A research program is progressive when its modifications to the protective belt generate new predictions that are confirmed: the program does not merely absorb anomalies but discovers unexpected phenomena. It is degenerative when the modifications only serve to explain, after the fact, what already happened, without predicting anything new.

“A company that adjusts its marketing strategy and, as a result, discovers a customer segment it did not know it had is in a progressive program: the adjustments are generating new knowledge. A company that constantly changes its strategy only to justify why the last quarter’s results were not as expected, without those changes producing better future predictions, is in a degenerative program: it is spending intellectual energy keeping alive a framework that no longer works.”

For the Architecture of Intention, Lakatos offers an invaluable guide for when to insist and when to pivot. The question is not “Was my hypothesis refuted?” as in Popper, or “Does my community still support it?” as in Kuhn, but “Are my adjustments generating new knowledge, or only post-hoc rationalizations?”

“A user who has asked an AI three times to reformulate their business plan, each time adding exceptions and conditions to explain why the previous numbers did not add up, may be in a degenerative program. They are not discovering anything new; they are protecting their hard core, ‘my business idea is viable,’ with an increasingly thick and less convincing protective belt. A good system would detect the pattern and ask: ‘Is it possible that the central premise needs revision?’”

6. Epistemic Assumptions and Problem Formulation

Everything we have explored in this topic, the types of knowledge, falsifiability, paradigms, and research programs, converges on one central idea: every problem formulation contains implicit epistemic assumptions. These assumptions determine the quality of the formulation as deeply as the data or logic that accompany it.

6.1 Invisible Assumptions

An epistemic assumption is an implicit belief about what is known, how it is known, and what degree of certainty that knowledge has. Most people are not aware of their epistemic assumptions. They operate like the air they breathe: invisible, omnipresent, and determining.

“When someone says, ‘I need a CRM system to improve my relationship with my customers,’ they are assuming: (1) that the customer relationship is deficient, a diagnostic assumption; (2) that a technological system can improve a human relationship, a causal assumption; (3) that a CRM is the right technology, a solution assumption; and (4) that they know enough about CRMs for the request to be useful, a competence assumption. Four assumptions, all invisible, all questionable.”

Alfred Korzybski, the father of general semantics, coined the phrase “the map is not the territory.” This metaphor is perfect for epistemic assumptions: every problem formulation is a map of reality, not reality itself. And like every map, it simplifies, omits, and distorts. The problem arises when we confuse the map with the territory, when our assumptions seem so obvious that we forget they are assumptions.

6.2 Taxonomy of Epistemic Assumptions

Epistemic assumptions in problem formulation can be classified into four categories:

Situational assumptions: “This is how things are.” Beliefs about the current state of reality. They may be outdated, partial, or simply wrong.

“A director who assumes that ‘our employees are not engaged’ may be confusing silence with disinterest. They may be highly engaged but lack channels to express it, or they may be engaged with aspects of the work that the director does not value.”

Causal assumptions: “X causes Y.” Beliefs about cause-and-effect relationships. They often confuse correlation with causation, ignore intermediate variables, or assume linearity where there is complexity.

“If a manager observes that teams using a specific tool perform better, they may assume that the tool causes the performance. But perhaps high-performing teams choose better tools, so the causality is reversed. Or perhaps both are caused by a third factor: a competent team leader who chooses good tools and also manages the team well.”

Solution assumptions: “The answer is Z.” Beliefs about what would solve the problem. They often confuse means with ends, replicate solutions that worked in other contexts, or reflect the user’s familiarity more than the suitability of the solution.

Competence assumptions: “I know enough.” Beliefs about one’s own ability to formulate the problem correctly. This connects directly with the Dunning-Kruger effect from Module I: those who know the least are often the most confident in their formulation.

6.3 Epistemic Humility as a Design Tool

Epistemic humility is not insecurity or false modesty. It is the honest recognition of the limits of one’s own knowledge: knowing what you do not know and, even more importantly, suspecting that there are things you do not know that you do not know. It is perhaps the most difficult and valuable intellectual virtue for problem formulation.

“Donald Rumsfeld, regardless of one’s opinion of him, articulated a useful epistemic taxonomy: there are things we know we know, known knowns; things we know we do not know, known unknowns; and things we do not know we do not know, unknown unknowns. The first are data; the second are identified questions; the third are blind spots that no analysis can cover a priori. Epistemic humility is the ability to keep the existence of the third category present while operating with the first and second.”

For the Architecture of Intention, epistemic humility translates into concrete design principles: systems that ask before executing, that make explicit the implicit assumptions in the user’s request, that present alternatives the user did not consider, and that admit the limits of their own responses. A system that projects certainty where there is uncertainty is not being useful; it is being dangerous.

7. Practical Application: Epistemology in Action

Let us bring together the concepts of this topic in an integrated scenario that shows how epistemology operates at every moment of problem formulation.

7.1 Case Study: International Expansion

Gabriela, CEO of a successful accounting software company in Mexico, asks an AI system: “I need a strategy to expand into Colombia. Our product already works well here, so we want to replicate the model.”

Let us analyze the epistemic assumptions present:

Situational assumptions: Gabriela assumes that she knows enough about the Colombian market to formulate the request. But does she know Colombian accounting regulations? Does she know how the needs of Colombian SMEs differ from those of Mexican SMEs? Does she know the local competitive landscape? Her propositional knowledge about Colombia may be limited, and her tacit knowledge, the knowledge she accumulated from understanding the Mexican market for years, is not automatically transferable.

Causal assumptions: “Our product works well here” contains an implicit causal attribution: “success is due to the product.” But it could be due to Gabriela’s network of contacts in Mexico, to a timing advantage, to a weak competitor that no longer exists, or to regulatory conditions specific to Mexico. The representativeness heuristic from Module I leads her to assume that Mexico and Colombia are “similar,” and therefore what worked there will work here.

Solution assumptions: “Replicate the model” is a solution assumption that has not been questioned. Why replicate instead of adapt? Why Colombia and not another market? Why geographic expansion and not product expansion? The request arrives with the solution predetermined. The real problem, “how to grow as a company,” has been prematurely reduced to a single path.

Operating paradigm: Gabriela operates within the paradigm of “scale by replicating,” a common mental model in the startup world. This paradigm makes alternatives invisible, such as licensing, local partnerships, or a completely different approach to the Colombian market.

Falsifiability: The request is not falsifiable. What result would make it fail? What is the criterion? Within what time frame? Without these parameters, neither Gabriela nor the system can evaluate whether the strategy works.

“A system with good intention architecture would respond: ‘I can help you with that. Before designing the strategy, I want to verify some assumptions. First: what specific evidence do you have that there is demand in Colombia for your type of product? Second: when you say your product works well in Mexico, to what do you attribute that success, the product, the team, the market, or a combination? Third: have you considered alternatives to geographic expansion, such as deepening your presence in the Mexican market or expanding your product line? And fourth: what minimum result would justify the investment, and within what time frame?’ Each question attacks a different epistemic assumption.”

7.2 Design Principles Derived from Epistemology

1. Make assumptions explicit before executing. Before responding to a request, identify the epistemic assumptions that support it and present them to the user for verification.

2. Ask for evidence, not conviction. “Are you sure?” is not an epistemically useful question. “What concrete evidence supports this claim?” is.

3. Apply the falsifiability test. If the user’s request does not have clear criteria for success and failure, help them formulate those criteria before proceeding.

4. Identify the operating paradigm. Detect the frame of reference from which the user is formulating and offer perspectives from alternative frameworks.

5. Distinguish between progressive and degenerative programs. If the user is iterating without progressing, point it out respectfully: perhaps the central premise needs revision.

6. Externalize tacit knowledge. Use specific questions and hypothetical scenarios to help the user articulate what they know but cannot express directly.

8. Glossary of Key Terms

9. Final Reflection and Self-Assessment

Before moving on to the next topic, Ends vs. Means and Framing Errors, take a moment to evaluate your understanding. The point is not to memorize definitions but to verify that you can apply the concepts.

10. Bibliography and Recommended Readings

Essential Readings

• Popper, K. (1959). The Logic of Scientific Discovery. Routledge. [La lógica de la investigación científica. Editorial Tecnos]

• Kuhn, T. (1962). The Structure of Scientific Revolutions. University of Chicago Press. [La estructura de las revoluciones científicas. Fondo de Cultura Económica]

• Polanyi, M. (1966). The Tacit Dimension. University of Chicago Press. [La dimensión tácita. Paidós]

Complementary Readings

• Lakatos, I. (1978). The Methodology of Scientific Research Programmes. Cambridge University Press.

• Gettier, E. (1963). “Is Justified True Belief Knowledge?” Analysis, 23(6), 121-123.

• Ryle, G. (1949). The Concept of Mind. Hutchinson. [El concepto de lo mental. Paidós]

• Popper, K. (1963). Conjectures and Refutations. Routledge. [Conjeturas y refutaciones. Paidós]

Next Topic

The next book in this series addresses Ends vs. Means and Framing Errors: how to distinguish between what you really want to achieve and the instruments you chose to achieve it, a confusion behind most poorly formulated problems. You will connect the epistemology you have just studied with George Pólya’s work on problem solving, Donald Schön’s work on reflective practice, and Rittel and Webber’s work on the “wicked problems” that defy all conventional formulation.