Metaphysics and epistemology

In short

Metaphysics asks what there ultimately is — substance, cause, mind, self, God, universals, time — and epistemology asks what knowledge is, how it is justified and how far it can reach. Together they form the core of "general philosophy" or "problems of philosophy" papers in BA Philosophy under UGC CBCS/NEP four-year frameworks, and they supply the vocabulary used in every other Western philosophy paper. They are examined in university theory papers, in UGC-NET Paper II (Philosophy), in entrance tests such as CUET-PG and JNU/DU/Jadavpur MA admissions, and in UPSC Philosophy Optional Paper I Section A. For jobs, the skills they train — defining terms precisely, spotting hidden premises, constructing and attacking arguments — are what law firms, policy and research organisations, UPSC interviews, analytics and content/AI-annotation employers actually test.

Explained

What the two branches are, and how they divide

• Metaphysics: the study of what is real and what kinds of things exist. Name comes from the ordering of Aristotle's treatise *after* (*meta*) the *Physics*; Aristotle himself called it "first philosophy" and the "science of being qua being".
• Divisions: ontology (what exists, categories of being), cosmology (origin, order and causation of the universe), philosophical theology/rational theology (God, arguments for and against), philosophical psychology (self, mind, immortality).
• Epistemology (Greek *episteme* = knowledge, *logos* = account): the nature, sources, structure, limits and validity of knowledge.
• Typical epistemology questions: What is knowledge? Is it possible? What are its sources? What is truth? What justifies a belief?
• Why they are interdependent: a claim about what can be known constrains what can be said to exist (Kant), and a claim about what exists constrains how it could be known (realism).
• Descriptive vs revisionary metaphysics (P. F. Strawson, *Individuals*, 1959): descriptive metaphysics describes the structure of our actual conceptual scheme; revisionary metaphysics proposes a better one.
• Critiques of metaphysics: Hume (no impression, no idea), Kant (reason overreaches when it leaves experience), logical positivism (A. J. Ayer: metaphysical statements are unverifiable, hence meaningless), later Wittgenstein (confusions of language), Quine (rehabilitates ontology via "ontological commitment" — what the bound variables of our best theory must range over).

Core metaphysical problems

• Appearance and reality: the distinction between how things seem and how they are; Parmenides (being is one, change is illusion) vs Heraclitus (all is flux); Bradley's *Appearance and Reality* (1893) treats relations as self-contradictory and reality as a single Absolute experience.
• Substance: that which exists in its own right and bears properties. Monism — one substance (Spinoza: God/Nature, with thought and extension as attributes; also Advaita Vedānta's Brahman). Dualism — two (Descartes: thinking and extended substance). Pluralism — many (Leibniz's monads, atomism, Nyāya-Vaiśeṣika's nine substances).
• Critique of substance: Locke calls it "something, I know not what"; Berkeley rejects material substance (*esse est percipi*); Hume reduces mind to a "bundle of perceptions".
• Universals and particulars: do redness, triangularity, justice exist? Platonic realism (Forms exist independently, *ante res*); Aristotelian moderate realism (universals exist only in particulars, *in rebus*); conceptualism (they exist as concepts in the mind); nominalism (only names are general — Ockham; resemblance nominalism; Quine's class nominalism).
• Causation: Aristotle's four causes — material, formal, efficient, final. Hume's analysis: a cause is a constant conjunction of contiguous, temporally prior events; the "necessary connection" is a habit of the mind, not an observed feature — hence the problem of induction. Kant replies that causality is an a priori category the understanding imposes, making objective experience possible. Mill's methods (agreement, difference, joint method, residues, concomitant variations) treat causation empirically.
• Mind-body problem: how mental and physical relate.
  1. Interactionist dualism (Descartes) — two-way causal traffic; objection: how does unextended mind push extended matter, and what of conservation of energy?
  2. Parallelism (Leibniz) — no interaction; pre-established harmony, like two synchronised clocks.
  3. Occasionalism (Malebranche) — God causes the correlation on each occasion.
  4. Epiphenomenalism — body affects mind only; mental events are causally idle by-products.
  5. Double-aspect / neutral monism (Spinoza, Russell) — one reality, two aspects.
  6. Idealism — only minds and ideas (Berkeley); materialism/physicalism — only the physical.
  7. Logical behaviourism (Ryle's attack on the "ghost in the machine" as a *category mistake*), identity theory (mental states are brain states — Place, Smart), functionalism (mental states are functional roles, realisable in different substrates — Putnam, Fodor), eliminativism (Churchland).
  8. The "hard problem" (Chalmers) and Nagel's "What is it like to be a bat?": why is there subjective experience (qualia) at all?
• Free will and determinism: hard determinism (every event, including choice, is necessitated; so no moral responsibility); libertarianism/indeterminism (some choices are genuinely free and self-caused); soft determinism/compatibilism (Hobbes, Hume, Mill, Frankfurt — freedom is absence of constraint plus acting from one's own desires, which is compatible with causal determination); fatalism is distinct — the outcome is fixed whatever one does. Related: Kant's resolution — determined as phenomenon, free as noumenon.
• Personal identity: what makes a person at t₂ the same as at t₁? Soul/substance view (Descartes, Butler); bodily/brain continuity; memory/psychological-continuity view (Locke's "consciousness makes personal identity"; Reid's brave-officer objection shows memory is not transitive); bundle/no-self view (Hume; and Buddhist *anattā*, the chariot/Nāgasena argument in the *Milindapañha*).
• Space, time and God: absolute vs relational space-time (Newton vs Leibniz); A-series/B-series (McTaggart's argument for the unreality of time). Arguments for God — ontological (Anselm, Descartes; Kant's objection that existence is not a predicate), cosmological/first cause (Aquinas; Hume and Russell: why must the series have a first term?), teleological/design (Paley's watch; Hume's and Darwin's objections), moral (Kant's God as a postulate of practical reason). The problem of evil (Epicurus, Hume) and the free-will defence (Plantinga).

What knowledge is: the standard analysis and its collapse

• Three senses of "know": propositional/factual (*knowing that* Delhi is the capital), ability (*knowing how* to swim), acquaintance (*knowing* a person or a pain) — Russell's knowledge by acquaintance vs by description; Ryle's knowing-how vs knowing-that.
• Belief, truth, knowledge: belief is a mental attitude that may be false; knowledge cannot be false. Knowledge entails belief and truth but is not reducible to them.
• The tripartite / JTB analysis (traced to Plato's *Theaetetus* and *Meno*): S knows that p iff (i) p is true, (ii) S believes that p, (iii) S is justified in believing that p. Each condition is claimed necessary, the three jointly sufficient.
• Gettier's refutation (Edmund Gettier, "Is Justified True Belief Knowledge?", 1963): cases where all three conditions hold but we deny knowledge, because the belief is true by luck. Standard examples — the man who will get the job and has ten coins in his pocket; Smith's "Jones owns a Ford or Brown is in Barcelona"; the fake-barn and stopped-clock cases.
• Post-Gettier repairs: (a) no-false-lemmas — the justification must not pass through a falsehood; (b) causal theory (Goldman) — the fact must cause the belief; (c) reliabilism — the belief must be produced by a reliable process; (d) defeasibility — no truth that would defeat the justification; (e) virtue epistemology (Sosa, Zagzebski) — the truth must be *credited to* the knower's intellectual virtue; (f) tracking/sensitivity and safety (Nozick: if p were false, S would not believe p); (g) Williamson's knowledge-first — knowledge is primitive and unanalysable.

Sources of knowledge, and the rationalism-empiricism axis as used here

• A priori vs a posteriori: justification independent of experience vs dependent on it. Analytic vs synthetic: predicate contained in the subject vs not. Necessary vs contingent: could not have been otherwise vs could. Kant's four combinations and his key claim: there are synthetic a priori judgements (mathematics, "every event has a cause"). Kripke later separates the metaphysical from the epistemic: "water is H₂O" is necessary but a posteriori; a standard metre stick case gives the contingent a priori.
• Sources recognised in the Western tradition: perception, introspection, memory, reason/intuition, inference, testimony (now a major research area — Coady), and in religious epistemology, revelation and mystical experience.
• Perception, theories of: naive/common-sense realism (we perceive objects directly as they are); representative realism/causal theory (Locke: we perceive ideas caused by objects; primary qualities resemble, secondary qualities do not); phenomenalism/idealism (Berkeley, Mill: objects are actual and possible sense-data); sense-datum theory (Moore, Russell, Price) and the argument from illusion; critical realism; direct realism restated by Austin (*Sense and Sensibilia*) and by disjunctivists.
• Rationalism (reason as the source — innate ideas, deduction, certainty) vs empiricism (experience as the source — *tabula rasa*, induction, probability) — the detailed arguments of Descartes, Locke, Berkeley, Hume and Kant belong to the modern-philosophy paper; what matters here is the *structural* problem: reason alone risks emptiness, experience alone risks scepticism.
• Comparative note (Indian parallel, frequently asked): Indian epistemology frames the same questions as pramāṇa-śāstra — accepted means of knowledge are *pratyakṣa* (perception), *anumāna* (inference), *upamāna* (comparison), *śabda/āgama* (verbal testimony), *anupalabdhi/anupalabdhi-abhāva* (non-apprehension, for absence) and *arthāpatti* (postulation); Cārvāka accepts only perception, Buddhism perception and inference, Nyāya four, Advaita six. Theories of error (*khyāti-vāda*) and the self-validity of knowledge (*svataḥ-prāmāṇya*) parallel Western debates on justification and defeasibility.

Structure of justification, scepticism and truth

• The regress problem: if every belief needs justification by another belief, we face infinite regress, circularity, or arbitrary stopping.
  1. Foundationalism — some basic beliefs are self-justifying or immediately justified (Descartes' *cogito*, incorrigible sense-reports); non-basic beliefs rest on them. Problem: are basic beliefs rich enough, and are they really indubitable? (Sellars' "myth of the given".)
  2. Coherentism — a belief is justified by fitting a mutually supporting system (Bradley, Blanshard, BonJour, Davidson). Problem: the isolation/plurality objection — coherent fiction.
  3. Infinitism (Klein) and contextualism (justification standards shift with context — DeRose, Lewis).
• Internalism vs externalism: must the justifier be accessible to the subject's reflection (internalism — evidentialism), or can it be an external fact about reliability (externalism — reliabilism)? Illustrated by the clairvoyant and the chicken-sexer cases.
• Scepticism: *global* (Pyrrhonian suspension of judgement) vs *local* (about the external world, other minds, induction, the past). Classic engines: Descartes' dream and evil-demon arguments, the modern brain-in-a-vat, Agrippa's trilemma, the problem of the criterion. Replies — Cartesian certainty, Moore's "Here is one hand" (common sense), Kant's transcendental argument, Wittgenstein's hinge propositions in *On Certainty*, Putnam's semantic externalism against the vat hypothesis, and contextualist denials of closure.
• Theories of truth (epistemology's companion question): correspondence (truth is agreement with fact — Aristotle, Russell, Tarski's semantic formalisation); coherence (truth is systematic consistency — Bradley); pragmatic (truth is what works, what is verified in inquiry — Peirce, James, Dewey); semantic/deflationary (Tarski; Ramsey's redundancy theory — "p is true" says no more than p); identity and consensus theories. Criteria of truth vs definitions of truth is a standard distinction.
• Realism and its rivals about the external world: realism (objects exist mind-independently), subjective idealism (Berkeley), absolute/objective idealism (Hegel, Bradley), transcendental idealism with empirical realism (Kant), phenomenalism, instrumentalism and internal realism (Putnam). Related dispute: scientific realism vs anti-realism, with Popper's falsifiability, Kuhn's paradigms and the demarcation problem.

Doing the exam: method and argument form

• Deductive vs inductive: a valid deductive argument's conclusion cannot be false if the premises are true; an inductive argument is at best probable. Validity ≠ truth; soundness = validity + true premises.
• Necessary and sufficient conditions: "p only if q" gives q as necessary; "p if q" gives q as sufficient. Most definitions in both branches are offered as *necessary and jointly sufficient* conditions — which is exactly what counterexamples like Gettier's attack.
• Counterexample method: to refute "all and only A are B", produce one A that is not B. This single move is behind Gettier cases, Reid on Locke, Hume on causation, and most exam answers that earn high marks.
• Thought experiments as evidence: Ship of Theseus, Theseus-style teletransportation (Parfit), Mary the colour scientist, Chinese Room (Searle), Twin Earth (Putnam), Trolley cases. State the case, extract the intuition, say what thesis it refutes.
• Answer structure that scores: define the terms → state the thesis in numbered premises → give the strongest argument for → give the standard objections with named philosophers → give the reply → state and defend your own position in one or two sentences.

Key points

• Metaphysics - inquiry into the ultimate nature of reality; Aristotle's "first philosophy", the science of being qua being.
• Ontology - the branch of metaphysics listing what kinds of thing exist and their categories.
• Epistemology - inquiry into the nature, sources, structure and limits of knowledge, and into truth and justification.
• Substance - what exists independently and bears properties; monism (Spinoza), dualism (Descartes), pluralism (Leibniz).
• Universals - repeatable properties; realism (*ante res*/*in rebus*), conceptualism, nominalism.
• Hume on causation - only constant conjunction, priority and contiguity are observed; necessary connection is a mental habit, which generates the problem of induction.
• Kant's reply - causality is an a priori category of the understanding; synthetic a priori judgements are possible, so metaphysics beyond experience is not.
• Mind-body positions - interactionism, parallelism, occasionalism, epiphenomenalism, double aspect, identity theory, functionalism, eliminativism.
• Category mistake - Ryle's charge that Cartesian dualism treats the mind as one more thing of the same logical type as the body.
• Compatibilism - freedom is acting from one's own unimpeded desires, which determinism does not exclude; distinct from fatalism.
• Personal identity - soul, body/brain, memory (Locke) and bundle/no-self (Hume, Buddhism) views; Reid's brave-officer case breaks memory transitivity.
• JTB analysis - knowledge = justified true belief; all three conditions necessary, claimed jointly sufficient.
• Gettier problem - JTB can be satisfied by a belief true only by luck, so JTB is not sufficient; repairs include no-false-lemmas, causal, reliabilist, defeasibility and virtue accounts.
• A priori / analytic / necessary - three distinct contrasts; Kant's synthetic a priori and Kripke's necessary a posteriori show they do not coincide.
• Theories of perception - naive realism, representative realism (primary vs secondary qualities), phenomenalism, sense-datum theory, direct realism.
• Regress problem - foundationalism, coherentism, infinitism and contextualism are the four standard answers; internalism vs externalism cuts across them.
• Scepticism - dream, evil demon, brain-in-a-vat and Agrippa's trilemma; answered by Descartes, Moore, Kant, Wittgenstein's hinges and Putnam.
• Theories of truth - correspondence, coherence, pragmatic, semantic/deflationary; keep the *definition* of truth distinct from a *criterion* of truth.
• Verifiability critique - Ayer and the logical positivists held metaphysical claims meaningless; Popper replaced verifiability with falsifiability, and Quine restored ontology through ontological commitment.

Examples

Metaphysics and epistemology use argument forms, not formulas; the cases below are to be reproduced and analysed the way derivations are in a science paper.

Case 1 — Testing the JTB definition with a Gettier case (full exam working).

1. State the definition: S knows p iff p is true, S believes p, S is justified in believing p.
2. Construct the case: Smith has strong evidence that Jones will get the job (the employer said so) and has counted ten coins in Jones's pocket. He infers "the man who will get the job has ten coins in his pocket."
3. Add the twist: Smith himself gets the job, and unknown to him he happens to have ten coins in his pocket.
4. Check the three conditions: the proposition is true; Smith believes it; he is justified (good evidence, valid inference).
5. Verdict: we deny he *knows* it, because its truth has nothing to do with his evidence. So the three conditions are not jointly sufficient — a counterexample refutes a sufficiency claim with a single instance.
6. Repair and test the repair: no-false-lemmas blocks this case (the lemma "Jones will get the job" is false) but not the fake-barn case (no false lemma there) — so the repair is incomplete, which is why reliabilist and virtue accounts were developed.

Case 2 — Hume's argument against necessary causal connection, laid out as premises.

1. Every idea is derived from a prior impression (Hume's copy principle).
2. In any causal sequence — the cue ball striking the second ball — we observe contiguity in space, priority in time and constant conjunction.
3. We do not observe any impression of *necessary connection* or of a power in the cause.
4. Therefore the idea of necessity has no external impression as its source.
5. Its real source is the felt determination of the mind to pass from the impression of the cause to the idea of the effect after repeated experience.
6. Consequence: causal inference rests on a habit, not on reason, and induction cannot be rationally justified without circularity (any justification would itself be inductive).
7. Kant's counter-move: deny premise 1's scope — some concepts are not derived from experience but are conditions of experience; causality is such a category, so "every event has a cause" is synthetic a priori.

Case 3 — Testing personal identity against two scenarios.

• *Ship of Theseus*: every plank is replaced one by one, and the old planks are reassembled elsewhere. Which is the original? Brings out that identity over time may be a matter of continuity of form, of matter, or of convention — and that "same" may mean different things for artefacts and for persons.
• *Reid's brave officer*: a boy is flogged for robbing an orchard; as a young officer he takes a standard and remembers the flogging; as an old general he remembers taking the standard but not the flogging. On Locke's memory criterion the general is the officer, the officer is the boy, but the general is not the boy — identity would fail to be transitive, which no identity relation can do. Repair: psychological *continuity* through overlapping chains of memory, not direct memory.

Case 4 — The regress of justification, worked through.

• Claim: "The Mumbai local is delayed." Why? "The app says so." Why trust the app? "It is usually accurate." Why believe that? …
• Three exits: stop at a basic belief that needs no further support (foundationalism); let the beliefs support one another in a system (coherentism); accept an endless chain of distinct reasons (infinitism).
• Each has its cost: foundationalism must show basic beliefs are both secure and informative; coherentism must rule out a coherent fiction; infinitism must explain how an unfinished chain confers justification.
• Internalist vs externalist test on the same case: if an unreliable app made the belief true by accident, the internalist may still call it justified (the evidence seemed good), the externalist will not (the process was not reliable).

Case 5 — Comparison tables for revision.

Where it is used

• Civil services and state PSC Philosophy optional: UPSC Paper I covers Plato to contemporary analytic philosophy — substance, causality, mind-body, truth, scepticism are standing questions; the same material carries the Ethics paper's foundations.
• Law and the judiciary: standards of proof ("beyond reasonable doubt" is applied epistemology), burden of proof, expert and eyewitness testimony, and the admissibility rules now contained in the Bharatiya Sakshya Adhiniyam, 2023; legal personhood and identity arguments appear in company-law and constitutional reasoning.
• AI, data and tech roles in India (TCS Research, Wipro, Infosys, Fractal, AI-annotation and RLHF teams): Searle's Chinese Room, functionalism and the qualia debate frame work on machine understanding; epistemology frames model evaluation, hallucination, evidence and bias audits — roles in AI ethics, policy and red-teaming hire humanities graduates for exactly this.
• Teaching and research: eligibility through UGC-NET/JRF Philosophy, assistant professorships, ICPR (Indian Council of Philosophical Research, New Delhi) fellowships, doctoral work at IITs' HSS departments (IIT Bombay, Kanpur, Madras) where philosophy of mind and of science are active areas.
• Journalism, fact-checking and public policy: evaluating sources and testimony, distinguishing claim from evidence, writing defensible arguments — Alt News/Boom-style verification work, think-tank research (Takshashila, CPR, ORF).
• Medical and research ethics committees: personal identity and criteria of death, consent capacity, and the Transplantation of Human Organs and Tissues Act, 1994 framework turn on metaphysical criteria of the person; institutional ethics committees apply them under ICMR's National Ethical Guidelines.
• Competitive-exam reasoning and corporate aptitude rounds: necessary/sufficient conditions, validity, assumption and inference questions in CAT/CLAT/GRE critical reasoning are the logic half of this topic applied.

Interview questions

• Q: What is the difference between metaphysics and epistemology? A: Metaphysics asks what ultimately exists and what reality is like; epistemology asks what knowledge is, how beliefs are justified and how far knowledge extends. Each constrains the other: what can be known limits what one may claim to exist.
• Q: State the tripartite definition of knowledge and say what is wrong with it. A: Knowledge is justified true belief. Gettier showed in 1963 that all three conditions can be met while the belief is true only by luck, so they are not jointly sufficient, though each is still widely held to be necessary.
• Q: Why does Hume deny that we perceive causal necessity? A: All ideas come from impressions, and we have impressions only of constant conjunction, temporal priority and contiguity — never of a power or necessity; so the idea of necessary connection comes from the mind's habit, not from the world.
• Q: How does Kant answer Hume? A: By making causality an a priori category the understanding imposes on appearances, so causal judgements are synthetic a priori and objectively valid of experience — but useless beyond it, which rules out speculative metaphysics.
• Q: What is a category mistake, and how did Ryle use it? A: Ascribing to something a predicate from a logical type it cannot belong to — like asking where the university is after being shown all its buildings. Ryle argued Cartesian dualism makes just this error by treating mind as another substance alongside body.
• Q: Is compatibilism a real solution to the free-will problem? A: It solves the problem it sets itself — freedom as absence of external constraint and action from one's own desires is clearly compatible with determinism — but critics say it redefines freedom and leaves untouched the question whether one could have willed otherwise.
• Q: Difference between a priori and analytic? A: A priori is an epistemic notion (justified without experience); analytic is semantic (the predicate is contained in the subject). They come apart: Kant's synthetic a priori arithmetic, and Kripke's necessary a posteriori "water is H₂O".
• Q: What is the regress of justification and what are the ways out? A: If each belief needs another belief to justify it, justification threatens to regress infinitely. Foundationalism stops at basic beliefs, coherentism makes support mutual, infinitism accepts the endless chain, and contextualism makes the stopping point context-relative.
• Q: Internalism or externalism about justification — which would you defend? A: Externalist reliabilism better explains how ordinary perceivers and children know things without reflective access to their evidence, but it struggles with the clairvoyant case; a defensible position combines reliable process with some internal responsiveness to evidence.
• Q: How would you answer the brain-in-a-vat sceptic? A: Several ways: Moore's appeal to what is more certain than the sceptical premise, Kant's transcendental argument that objective experience presupposes objects, Wittgenstein's view that such hinge propositions are not candidates for doubt, and Putnam's semantic argument that a vat-brain's words could not even refer to vats.
• Q: Does Indian philosophy have an epistemology comparable to the Western one? A: Yes — pramāṇa-śāstra debates precisely which means of knowledge are valid (perception, inference, comparison, testimony, postulation, non-apprehension), how error arises and whether knowledge validates itself; Nyāya's treatment of inference and testimony is as technical as anything in analytic epistemology.
• Q: Is the mind-body problem relevant to artificial intelligence? A: Directly — functionalism implies a suitably organised machine could have mental states, while Searle's Chinese Room and the hard problem of consciousness argue that syntax and behaviour leave understanding and subjective experience unexplained.

Common mistakes

• Treating metaphysics and ontology as synonyms: ontology is one branch of metaphysics, alongside cosmology, rational theology and philosophical psychology.
• Collapsing the three contrasts — a priori/a posteriori (epistemic), analytic/synthetic (semantic), necessary/contingent (metaphysical) — and so missing Kant's and Kripke's central claims.
• Describing Gettier cases as showing that justification or truth is unnecessary; they show the three conditions are not *sufficient*.
• Confusing a *definition* of truth with a *criterion* of truth, or presenting coherence as a definition when the question asked for a test.
• Confusing determinism with fatalism, and indeterminism with freedom — random choice is no more free than a caused one.
• Stating Locke's memory criterion of personal identity without Reid's brave-officer objection, or giving the objection without the standard repair (overlapping psychological continuity).
• Writing an answer as a parade of names and dates with no argument: marks go to numbered premises, the strongest objection, a reply, and a stated position.
• Straying into territory the modern-philosophy, Greek or existentialism questions own — use Descartes, Hume and Kant here only for the *problem*, not as a history of their systems.

Related topics

Greek philosophy: Socrates, Plato and Aristotle · Modern Western philosophy: rationalism, empiricism and Kant · Existentialism and phenomenology · Logic: deduction, induction and fallacies · Indian epistemology: pramāṇas and theories of error · Philosophy of mind and artificial intelligence