Phenomenalism
A later, secularised version of idealism: phenomenalism argues that statements about physical objects are really disguised statements about actual and possible sense experiences. "There is a table in the next room" means "if I were to go into the next room, I would have table-like sense experiences." The table isn't a mind- independent object; it's a stable pattern in possible experience. J.S. Mill called physical objects "permanent possibilities of sensation." The problem: Phenomenalism struggles with counterfactuals. "If I looked through the window, I'd see a tree" — but what makes this counterfactual true? Not the tree (phenomenalism says there's no mind-independent tree). Not the possible experience (which doesn't exist yet). The theory seems to require something that grounds the counterfactual — something that makes it the case that certain experiences would occur — and it's hard to see what that something could be other than a mind-independent reality. Which is what phenomenalism was trying to eliminate.
The Adverbial Theory
There's a fifth option that tries to avoid both sense-data and naïve realism. The adverbial theory says: when you see a red apple, you don't perceive a red sense- datum, nor do you directly perceive a red apple. Instead, you sense redly and sense applishly . The redness is not a property of a mental object (the sense-datum) or a physical object (the apple). It's a way of sensing — an adverbial modification of the act of perception itself. This avoids postulating sense-data as mental objects (which creates the veil-of- perception problem) while allowing that perception has qualitative character. But the theory struggles with complexity. When you see a red square and a blue circle, you sense redly-and-squarely-and-bluely-and-circularly. The adverbial theory can't easily distinguish "red square plus blue circle" from "red circle plus blue square" — it groups the properties but can't structure them. This is the many-property problem , and it remains unresolved.
The positions on perception map cleanly onto the three modalities. Direct realism operates in the immanent mode: you are in direct contact with the world. Indirect realism operates in the omniscient mode: you access the world through structural representations (sense-data) that model it from outside. Idealism operates in the transcendent mode: the mind generates the world of experience from within itself.
Each captures something real. You do have direct perceptual contact with reality (direct realism is right about the immanent-modal aspect of perception). You do process the world through representational structures (indirect realism is right about the omniscient-modal aspect). And the mind does actively participate in constructing perceptual experience (idealism is right that perception is not passive reception). The IDM says: perception is an interaction between subject and world. It has all three modal aspects simultaneously. The debate persists because each position absolutises one aspect and denies the others. Direct realism says perception is only interaction. Indirect realism says it's only representation.
Idealism says it's only mental construction. In reality, it's all three — distinct, inseparable, and non-interchangeable — and understanding how they relate dissolves the apparent contradictions. This is not just a theoretical resolution. It has practical implications. If perception is interaction (not just representation), then the quality of your perceptual experience depends on the quality of your engagement with the world. A birdwatcher sees things in a forest that a casual walker misses — not because they have better eyes but because their participatory involvement (immanent mode) is richer. A sommelier tastes things in a wine that a casual drinker doesn't — not because they have a superior tongue but because their structural knowledge (omniscient mode) and their trained sensitivity (immanent mode) work together. Perception is a skill, not just a passive reception — and like all skills, it can be developed.
You're reading this page through a screen. The colours are produced by LED pixels, the text is rendered by software, and the entire visual experience is mediated by technology and neurology. Are you perceiving the page directly? Through a representation? Or is the page nothing beyond your perception of it? Now extend this to social media, deepfakes, AI-generated images, and virtual reality. The philosophical question about perception isn't academic — it's the fundamental question of the information age: how do you know what you're perceiving is real? And the IDM's answer — perception is interaction, and the quality of perception depends on the quality of engagement — suggests that the antidote to digital illusion is not more information but deeper participation in reality.
Discussion Questions
1. You see a sunset. A direct realist says you see the sun itself. An indirect realist says you see sense-data caused by the sun. An idealist says the sun is a collection of ideas in minds. Which account best matches your experience? Which best matches what science tells us about the physics of light and the neuroscience of vision? 2. Locke's primary/secondary distinction says the apple's shape is "real" but its colour isn't. Does this seem right? If secondary qualities are "in the mind," does that make them less real? 3. Virtual reality headsets create convincing visual experiences of objects that don't exist. Does VR support indirect realism (we always perceive representations, and VR just feeds us different ones) or does it support direct realism (in VR, we directly perceive the virtual environment, which is a real thing)? 4. Berkeley said "to be is to be perceived." Social media culture sometimes operates as if this is true — things only "exist" (matter, count, are real) if they're posted, liked, and perceived by others. Is this a version of idealism?
What does it miss?
Chapter Six
Rationalism and Empiricism
Where does knowledge come from? Two great traditions give opposing answers.
Rationalism says: the most fundamental knowledge comes from reason alone, independent of sensory experience. Some truths can be known a priori — before and apart from experience — through the power of the mind itself.
Empiricism says: all knowledge ultimately derives from sensory experience. The mind at birth is a blank slate ( tabula rasa ), and everything we know is built up from what we see, hear, touch, taste, and smell. This debate runs through the entire history of Western philosophy and directly shapes how you think about truth, evidence, and certainty. It also connects to live questions about AI, data science, and whether algorithmic pattern-recognition counts as "knowledge."
The Rationalist Case
Plato argued that genuine knowledge — knowledge of the Forms, the eternal, unchanging structures behind the shifting appearances of the sensory world — cannot come from the senses. The senses give you opinions about particular, changing things. Reason gives you knowledge of universal, permanent truths. Descartes argued that the senses demonstrably deceive (Chapter 7), and that the only certain knowledge — the cogito — comes from pure reason. He further claimed that certain ideas are innate : the idea of God, the idea of mathematical truths, the idea of substance. These ideas are not derived from experience but are built into the mind's structure. Leibniz argued that necessary truths — mathematical and logical truths that could not possibly be false — cannot be grounded in experience, because experience only shows us what is , not what must be . The fact that 2 + 2 = 4 is not something you learn by counting objects. It's a truth of reason that would hold even if no objects existed. Leibniz distinguished between truths of reason (necessary, a
priori) and truths of fact (contingent, a posteriori), and argued that even truths of fact must have rational explanations — the
Principle of Sufficient Reason holds that nothing exists without a reason why it exists rather than not. This will reappear in Chapter 13 as the foundation for a version of the cosmological argument. The rationalist case is strongest for mathematics, logic, and conceptual truths. "All bachelors are unmarried" is true by definition — you don't need to survey bachelors to confirm it. "The interior angles of a triangle sum to 180 degrees" is provable by pure reasoning. These are a priori truths: known independently of experience. The modern case for innate ideas has been strengthened by Noam Chomsky's poverty of the stimulus argument in linguistics. Children learn language far more quickly and accurately than the data they receive could explain. They hear fragmentary, error-filled speech and yet produce grammatically complex sentences they've never heard before. Chomsky argued that this is only possible if the mind comes equipped with an innate "universal grammar" — a built-in linguistic structure that isn't learned from experience but is a precondition for language learning. If something as complex as grammar is innate, perhaps other fundamental cognitive structures are too. Cognitive science has added further evidence. Infants as young as a few months display surprise when objects behave in physically impossible ways — disappearing, passing through solid barriers, moving without being touched. This suggests that basic expectations about physical causation and object permanence are not learned from experience (the infant hasn't had enough experience to learn them) but are innate features of the cognitive architecture. The rationalist would say: the mind comes equipped with certain expectations about how the world works, and experience fills in the details. The empiricist would respond: even these early expectations could be learned in the womb or in the first weeks of life, and calling them "innate" overstates the case. The debate, in this form, has become partly an empirical question — and partly a question about what we mean by "innate" in the first place.
The Empiricist Case
John Locke (1632–1704) argued that the mind at birth is a tabula rasa — a blank slate. There are no innate ideas. Everything in the mind got there through experience: either through sensation (input from the external world via the senses) or through reflection (the mind's observation of its own operations). Even
complex abstract ideas — justice, infinity, causation — are built up from simpler ideas that ultimately derive from sensory experience.
Locke distinguished between simple ideas (the colour red, the feeling of warmth, the taste of sweetness — which the mind receives passively and cannot create from nothing) and complex ideas (friendship, government, unicorn — which the mind constructs by combining, comparing, and abstracting from simple ideas). The mind is active in constructing complex ideas, but it cannot invent the raw materials. All raw materials come from experience. This is a powerful framework. It explains why people raised in different environments have different beliefs — they've had different experiences. It explains why science progresses through observation rather than pure thought — because knowledge of the world comes from the world, not from the mind. And it grounds knowledge in something publicly accessible (experience) rather than something private and uncheckable (innate ideas). David Hume pushed empiricism further. He divided all knowledge into two categories —
Hume's fork
: relations of ideas (mathematical and logical truths, which are certain but tell you nothing about the world) and matters of fact (claims about the world, which are informative but never certain). Anything that doesn't fall into one of these categories — including most of metaphysics and theology — is, in Hume's famous phrase, "nothing but sophistry and illusion" and should be "committed to the flames."
Hume's sharpest argument concerned causation . We believe that fire causes heat, that dropped objects fall, that the sun will rise tomorrow. But what do we actually observe ? We observe constant conjunction — fire is always accompanied by heat. We don't observe any necessary connection between them. The belief that fire must produce heat goes beyond anything experience can justify. It's a habit of the mind — what Hume called "custom" — not a rational inference.
This leads directly to the problem of induction . Inductive reasoning — inferring general laws from particular observations — depends on the assumption that the future will resemble the past. But this assumption cannot itself be justified inductively (that would be circular: "the future will resemble the past because in the past the future has always resembled the past") or deductively (there's no logical contradiction in imagining that the laws of nature change tomorrow). The entire edifice of scientific knowledge rests on an assumption that cannot be rationally justified. This is one of the most important unsolved problems in philosophy.
Karl Popper tried to solve the problem by redefining science. Instead of inductively confirming theories (which Hume shows is impossible), science falsifies them. A theory is scientific not because it can be proven true but because it can be proven false . "All swans are white" is scientific because a single black swan would falsify it. "God works in mysterious ways" is not scientific because nothing could falsify it — any evidence can be accommodated. Popper's falsificationism sidesteps induction by saying we never confirm theories, we only eliminate false ones. But this solution has its own problems: no single observation can cleanly falsify a theory (the Duhem-Quine thesis), and scientists in practice do use positive evidence to support theories, not just negative evidence to eliminate them. Thomas Kuhn took a more radical approach: science doesn't progress by steadily accumulating knowledge (whether through induction or falsification) but through paradigm shifts — revolutionary changes in the basic framework through which scientists understand the world. Normal science works within a paradigm (Newtonian physics, for example), solving puzzles according to its rules. But anomalies accumulate, and eventually the paradigm breaks down and is replaced by a new one (Einsteinian physics). The shift is not a logical progression but something closer to a gestalt switch — a new way of seeing the same data. This raised uncomfortable questions about whether science tracks objective truth or merely shifts between equally "valid" perspectives. The empiricist case is strongest for knowledge of the physical world. You can't know the boiling point of water by pure reason — you have to observe it. You can't know what a pineapple tastes like without tasting one. These are a posteriori truths: known through experience. The distinction between a priori and a posteriori maps onto another important distinction: analytic and synthetic truths. An analytic truth is true by virtue of the meanings of its terms ("all bachelors are unmarried"). A synthetic truth adds new information not contained in the concepts ("the cat is on the mat"). Most a priori truths seem to be analytic, and most a posteriori truths seem to be synthetic. But is this mapping complete? That's where Kant enters.
Kant's Synthesis
Immanuel Kant (1724–1804) claimed to have been "awakened from dogmatic slumber" by Hume. His response was the most ambitious synthesis in the history of epistemology.
Kant's
Copernican revolution in philosophy
: previous thinkers assumed that knowledge must conform to objects — that the mind passively receives information from the world. Kant reversed this: objects must conform to our knowledge. The mind doesn't passively receive reality; it actively structures it. Just as Copernicus solved astronomical problems by making the Earth revolve around the sun rather than the reverse, Kant solved epistemological problems by recognising that the mind shapes experience rather than merely receiving it. Specifically, Kant argued that the mind comes equipped with categories of understanding — innate conceptual structures (causation, substance, unity, plurality, possibility, necessity) that organise raw sensory input into coherent experience. You don't learn that every event has a cause from experience — you bring the concept of causation to experience, and it's what makes experience intelligible in the first place.
The mind also structures experience through the forms of intuition : space and time. These are not features of the world-in-itself but frameworks the mind imposes on sensory data. You don't perceive space and time — you perceive through space and time. They are the conditions under which experience is possible, not part of what is experienced. This generates Kant's crucial innovation: synthetic a priori knowledge — truths that are informative about the world (synthetic) but known independently of experience (a priori). "Every event has a cause" is Kant's example: it's not true by definition (it's synthetic), but you don't learn it from experience (it's a priori — it's a precondition for experience to be coherent at all). "7 + 5 = 12" is another: it genuinely adds new information (it's synthetic) but you know it through reason, not observation (it's a priori). But Kant's synthesis comes with a cost. If the mind structures everything we experience, then we can only know the world as it appears to us — the phenomenal world . The world as it is in itself — the noumenal world , or the thing-in-itself — remains permanently inaccessible. We can never step outside our own categories to see reality "raw." This is a kind of permanent epistemic limitation — we are, in a sense, trapped inside our own cognitive architecture.
"AI systems trained on massive datasets seem to 'know' things they were never explicitly taught. Is that empiricist knowledge? Rationalist knowledge?
Something else?"
A large language model processes billions of text samples and extracts patterns — grammatical rules, factual associations, reasoning templates — that were never explicitly programmed. Is this empiricism (knowledge derived from data, which is a form of experience) or rationalism (the system has extracted structural truths that go beyond any individual data point)? The IDM's answer: it's knowing without understanding. The AI has extracted structural patterns (omniscient-modal knowledge) from data. But it has no first-person experience of what the data means — no understanding in the immanent-modal sense. It can tell you that "grief" is associated with "loss," "tears," and "mourning" — but it has never grieved. The rationalism- empiricism debate, mapped onto AI, reveals the Incommensuration Theorem in action: the AI has immense knowledge (whether you call it empiricist or rationalist) and zero understanding. Both the rationalist and the empiricist would claim the AI as support for their position. The IDM says both are right — and both are missing the modality that makes knowledge meaningful .
The rationalism-empiricism debate maps directly onto the knowing/understanding distinction. Rationalism privileges knowing — third- person, structural, representational. The truths of reason are precisely the truths you can access from outside the domain, through formal analysis.
Empiricism privileges understanding — first-person, participatory, experiential. The truths of experience are precisely the truths you can only access from inside the interaction. The debate persists because both sides are right about their own mode and wrong to claim it's the only one. Rationalism correctly identifies that structural truths (logic, mathematics) are not derived from sensory experience. Empiricism correctly identifies that experiential truths (what pineapple tastes like, what grief feels like) are not derived from reason alone. The Incommensuration Theorem explains why: knowing and understanding are categorically different, and neither can produce the other. Kant's synthesis is the closest the canon gets to the IDM position. He saw that both modes are needed and that the mind actively participates in structuring experience. But he stopped short: he treated the "noumenal" world as permanently inaccessible, whereas the IDM says it's accessible — through understanding, through the immanent-modal contact that doesn't pass through the representational bottleneck.
Discussion Questions
1. Hume's problem of induction: you can't logically prove that the future will resemble the past. Does this bother you? Do you trust that the sun will rise tomorrow? If so, on what basis — reason, experience, habit, or something else? 2. AI systems trained on large datasets are extremely good at pattern recognition. Is this empiricist knowledge (patterns derived from data/experience) or something else? Does an AI that has processed every piece of music ever written "understand" music? 3. Kant says the mind structures experience using categories like causation. If a different species had a different mental architecture, would they experience a different reality? Is there one reality or many? 4. Is mathematical knowledge discovered or invented? Rationalists say discovered (maths is real, and we access it through reason). Empiricists tend toward invented (maths is a human construction derived from patterns in experience). What do you think, and why?
Chapter Seven
The Challenge of Scepticism
This chapter asks the question that keeps philosophy students up at night: how do you know anything is real? The question seems extreme — even silly. Of course the world is real. But the sceptical challenge has never been adequately answered within the Western canon, and its echoes are everywhere in your generation's experience: simulation theory, deepfakes, algorithmic manipulation, the feeling that nothing online can be trusted. The ancient sceptical worry that appearances might not correspond to reality has become a daily lived experience. Understanding why scepticism is so hard to refute — and where it actually breaks — is more practically important now than at any point since Descartes.
Descartes' Method of Doubt
René Descartes, writing in 1641, decided to doubt everything he could possibly doubt — systematically, rigorously, without mercy — to see if anything survived. If something could survive total doubt, it would be a foundation for certain knowledge. Stage 1: The senses deceive. I've been fooled before — optical illusions, mirages, dreams that seemed real. If my senses have deceived me even once, how can I trust them at all? Maybe everything I perceive is an illusion. (This doesn't mean it is — just that I can't be certain it isn't.) Stage 2: The dream argument.
There's no reliable way to distinguish being awake from dreaming. Right now, you believe you're reading a book. But you've had dreams in which you believed you were reading a book. How do you know this isn't one of those dreams? You don't. So maybe your entire experience right now is a dream, and none of the objects you perceive actually exist.