Insight Editions · No. 81
Metaphysics: The Key Concepts
Helen Beebee, Nikk Effingham and Philip Goff
An alphabetical guide to metaphysical concepts and debates, including ontology, modality, time, causation, personal identity, free will, and the arguments associated with major philosophers.
A study guide to Metaphysics: The Key Concepts by Helen Beebee, Nikk Effingham and Philip Goff. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
The Book in Brief
The design is alphabetical with cross-references. Short entries define a term in a few lines ("Aleph", "Overlap", "Singleton"); medium entries explain a concept with an example; long entries, with section headings, survey a debate (free will, fission, modal realism, perdurantism, presentism, truth). Bold cross-references knit these into six major topics, which a reader can follow as connected chapters.
The six major topics that the book's cross-references knit together.
The authors' voice is plain, careful and lightly comic: Marty catches a cold in 1955, everyone in a possible world sings Livin' la Vida Loca, Hulk Hogan was once 2′ tall, and McTaggart's breakfasts served no food. The jokes carry examples that recur across entries: Old Jack and Wee Johnny for personal identity, the telephone box and the Coke can for universals, Jane for free will, the bridge bombers for pre-emption. The book is even-handed: positions are presented with their motivations and objections, and the authors rarely take sides.
| Topic | Core entries | Central question |
|---|---|---|
| Free will | Compatibilism, consequence argument, Frankfurt, libertarianism, luck | Can free action coexist with determinism, or with indeterminism? |
| Personal identity | Animalism, psychological continuity, fission, survival | What makes Old Jack the same person as Wee Johnny? |
| Universals | Universals, tropes, class and resemblance nominalism | What do two red things have in common? |
| Realism | Global realism, relativism, scepticism, truth | Is reality independent of our minds? |
| Time | Presentism, eternalism, A and B theories, McTaggart | Do past and future exist? Is tense real? |
| Causation and laws | Counterfactual theory, DTA, Ramsey-Lewis, pre-emption | What is causation, and what makes a law? |
The six major topics and where to find them.
Since 2011 the field has moved in three directions that matter for readers. First, grounding and fundamentality have become central tools, often replacing supervenience as the way to say one thing is "nothing over and above" another. Second, powers and dispositional essentialism have strengthened the anti-Humean side of the causation and laws debate. Third, metametaphysics and experimental philosophy have matured into fields of their own, testing the armchair methods the Introduction describes. Two figures the book describes as living, Armstrong (died 2014) and Kripke (died 2022), have since died, as has Parfit (2017), whose views on survival the book discusses.
Who wrote it, and how it shows. The three authors were colleagues in the British metaphysics community: Helen Beebee (then at Birmingham, a leading writer on causation, laws and free will), Nikk Effingham (Birmingham, on ontology and time travel) and Philip Goff (on consciousness, later widely known for his defence of panpsychism). The book's strongest entries, on causation, laws, free will, ontology and time, track these interests. Because entries are unsigned, the reader cannot tell who wrote what, and the voice is admirably consistent.
Three reading paths. A reader can use the book in three ways. As a dictionary, look up a term and follow one cross-reference. As a topic course, start from one of the six major entries (free will, personal identity, universals, realism, time, causation and laws) and follow its bold cross-references in turn; Chapter 2's Table 1 and Chapter 3's neuron diagrams are natural anchors. As a survey, read straight through, noticing how the same examples (Old Jack, Jane, the telephone box, the bridge bombers) recur across the alphabet.
Three sizes of entry, joined by cross-references.
Argument map: how the alphabet hangs together
From vocabulary to dispute. Early entries fix the shared tools: abstract vs concrete, analytic vs synthetic, a priori vs a posteriori, class, existence, essence. Those tools then serve the six topics. A reader who meets "necessary a posteriori" in Chapter 3 without the Chapter 1 map of analytic, a priori and necessary will miss why Kripke's examples matter. Likewise "Humean supervenience" in Chapter 2 presupposes the Chapter 1 contrast between categorical and dispositional properties and the Chapter 1 sketch of laws.
Cost-benefit as method. Across the book the authors rarely declare a winner. They set out rival positions, list arguments and counterexamples, and weigh theoretical virtues: simplicity, strength, conformity to common sense, fit with science. Lewis's genuine modal realism is the clearest case: offered for its reductive benefits despite the "incredulous stare". The same method appears in the choice among theories of holes, of composition and of personal identity after fission.
How the four alphabetical blocks accumulate tools.
| Block | Alphabet | Gist |
|---|---|---|
| Ch 1 | Abstract to Existence | Ontology vocabulary; opens free will, identity, causation |
| Ch 2 | Experimental philosophy to Natural kinds | Free will table; holes; Humeanism; metametaphysics; modal realism |
| Ch 3 | Natural necessity to Psychological continuity | Modality; Libet; presentism/perdurantism; pre-emption; PAP |
| Ch 4 | Quasi-memory to Zeno | Realism; Ramsey-Lewis; truth; tropes; time travel; method |
Table O1. Four alphabetical blocks and what each carries.
About this book
Metaphysics: The Key Concepts by Helen Beebee, Nikk Effingham and Philip Goff, 2011. This is a chapter-by-chapter study of the book, with summaries, maps, tables and diagrams; short quotations keep their page references. It is shared free for personal study, is not for sale, and is not endorsed by the author or publisher. Please read and buy the original.
Insight Editions · rajeshrmahapatra.com/insight-editions
How to use the A closer look and literature boxes
A closer look. Each section ends with flagged slips: wrong Big Bang age, Stella McCartney's father named "John", Michelson-Morley overstated, misspellings (Pierce, Rudolph, Ghenghis), internal inconsistencies. Contested interpretive claims (whether Quine is an austere nominalist; whether GMR is parsimonious) are labelled contested rather than errors.
Literature boxes as field snapshots. Boxes headed "Literature update (2026)" report PhilPapers Survey 2020 splits and named post-2011 books. They are not claims that the 2011 authors endorsed those results, and they are not invented laboratory science. Where the book itself cites empirical figures (Libet timings, Chernobyl half-life, cosmological age), verification notes sit in Critical-lens or time boxes.
Recommended path through each alphabetical section.
What Is Metaphysics?
Physicists, biologists, sociologists and archaeologists also study reality. Metaphysics differs in two related ways. First, it asks questions that science does not (fully) answer. Whether we have free will depends partly on what "free will" means: must the decision be undetermined right up to the moment of choice, or is a certain kind of control enough? Such conceptual questions call for conceptual analysis. Likewise, physicists may say what is possible (nothing outruns light), but a theory of modality, or of what causation or truth is, is not their remit.
Yet many metaphysicians insist that metaphysics is not restricted to concepts: we want to know what causation is, and what "cause" means is only a route to that. That leads to ontology: do numbers exist? Are there universals, features such as being furry genuinely shared by different things? Does the truth of "there are infinitely many primes" give us reason to believe in numbers, and "how can we be sure that the number 537,582,201.8 really exists"?
Why "key concepts". The authors' choice of a concept-by-concept format is itself a thesis about metaphysics. Many metaphysical questions turn on what a concept requires: whether free will requires the ability to do otherwise, whether causation requires a law, whether persistence requires being wholly present. A dictionary of concepts is therefore not a mere aid to studying metaphysics but a large part of it. At the same time, the authors insist that concepts are a route to reality, not a substitute for it: the question whether there are universals is not settled by analysing the word "property".
Two kinds of metaphysical question, both answered from the armchair.
Two kinds of question, six topics
Conceptual questions ask what a term requires. Does free will require that the decision was undetermined right up to the choice, or is a certain kind of control enough? Does causation require a law connecting cause and effect, or only a counterfactual dependence? Does persistence require being wholly present at each time, or only having a temporal part then? The authors treat such questions as the natural entry point: until the senses of the key terms are fixed, ontological disputes talk past each other.
Ontological questions ask what exists. Do numbers exist? Are there universals, features such as being furry genuinely shared by different dogs? Does the truth of "there are infinitely many primes" give reason to believe in numbers? Metaphysicians often answer such questions "from the armchair", but many insist that the analysis of concepts is only a route to what reality contains, not a substitute for it. The introduction's joke about being sure that "the number 537,582,201.8 really exists" marks the gap between mathematical practice and ontological commitment.
Introduction map: concepts feed six ongoing disputes.
Reference and crash course. As a reference, the book separates senses of contested terms so that parties are not "using a key term to mean different things". As a crash course, the longer entries with section headings carry the six topics; short entries are often cross-references. Further reading favours accessible sources; the suggested textbooks are Blackburn (1999), Conee and Sider (2007) and Carroll and Markosian (2010).
Abstract vs Concrete to Existence
Abstract vs concrete to Ancestral relation
Abstract vs concrete. Concreta "are not, as you might expect, those that are mainly made of cement". They include tables, mountains, stars, you and me; also immaterial objects if any (ghosts, souls), events (the Second World War, your lunch yesterday), immanent universals, tropes and, for believers, holes. Roughly, concreta are in space and time; abstracta (platonic universals, numbers, classes) are not: "you would not find the number six lurking around somewhere in Liverpool". Two debates follow: whether abstracta exist (platonism), and how to define the terms at all (Swoyer 2008).
Agent causation. Some hold that only an "uncaused cause" of one's actions is ultimately responsible. Most philosophers take the causal relata to be events, facts or states of affairs (Jane's being hungry at t caused her decision to order pizza). On agent causation, Jane herself, a substance, causes the decision. It is usually tied to indeterminism, though some agent causalists are compatibilists. Criticisms: acting rationally seems to need reasons as causes (Clarke 1993 replies that agent causation can allow this), and it is "mysterious" what gives agents alone this power. Agent causalists answer that the mystery is the familiar one of being the source of what one does, and that event-causal theories simply relocate the problem: if Jane's decision is caused only by prior events, it is hard to see where Jane comes in.
Causal relata for action: events, facts, or the agent as substance.
Analytic vs synthetic truths. Analytic truths are true "merely in virtue of the meaning of words" ("all bachelors are unmarried"); synthetic truths are the rest ("some bachelors are tall"). Intuitively the analytic truths are the a priori ones, but Kripke's necessary a posteriori truths complicate the map (see Chapter 3). Ancestral relation: given a relation R, the relation of standing in a chain of R relations is its ancestral, or transitive closure; ancestor of is the ancestral of parent of, hence the name. Aleph names sizes of infinity.
Animalism to Armstrong
A priori/a posteriori. A claim is knowable a priori if its justification "need not rely on experience in any way" (the infinity of primes). A child may still learn that 2 + 5 = 7 by counting objects; what matters is whether experience is required. Aristotle (384-322 BC), Plato's pupil, brought Plato's Forms "down to Earth": a universal is wholly present in each of its instances (immanent realism). David Armstrong, born in Melbourne in 1926, spent most of his career at Sydney, led the re-popularising of metaphysics after the linguistic turn, defended immanent universals, the DTA view of laws, Australian materialism about the mind, an ontology of states of affairs, truthmakers and a combinatorial theory of modality he later rejected.
Backwards causation in the Marty example.
Austere nominalism to Bodily continuity
Austere nominalism. The realist explains why a telephone box and a Coke can are both red by a shared universal. The nominalist can give another explanation (classes, tropes, resemblance) or deny that any is needed: "explanation has to end somewhere". Quine took this second line; Armstrong mocked it as "ostrich nominalism", refusing "to answer a compulsory question on the examination paper". The dispute is tied to Quine's criterion of ontological commitment: "There is something that is a telephone box and is red" commits us to a red box, not to redness.
Being. Most metaphysicians avoid the term. It survives where some distinguish subsistence from existence (non-existent objects "have being"), and in Locke's definition of essence, "the being of a thing, whereby it is what it is", which many find less clear than the notion it defines. George Berkeley (1685-1753), Bishop of Cloyne, the second of the great British Empiricists, wrote the Principles (1710) and the Three Dialogues; see idealism.
Bodily continuity theories ground personal identity in the identity of bodies or brains: Old Jack (2010) is Wee Johnny, who ran in the 1952 Helsinki Olympics, if they share one body, though not its atoms. A whole-body version may simply be animalism. A brain version looks like an "unstable halfway house": we favour the brain because it supports psychological continuity, yet the theory denies survival when psychology continues without the brain, and a human can survive with one hemisphere.
Three rival grounds of personal identity in the section.
Bradley's regress to Bundle theories
Bradley's regress: each added relation could exist without Plato being wise.
Brains in a vat (BIV). Putnam's (1981) update of Descartes' evil demon: we are disembodied brains fed experiences by a computer, as in The Matrix (1999), except that there the whole body is envatted. Sceptics use it to argue we lack knowledge of the external world. Putnam argues instead that "I am a brain in a vat" is self-refuting: envatted brains cannot refer to real brains and vats, since their word "vat" is not causally connected to vats. If the hypothesis were true, it could not be truly stated; so whoever states it truly is not a brain in a vat. Critics reply that this shows at most that the sceptic cannot express the worry in the envatted language, not that the worry is empty. The entry ties this to global realism and anti-realism.
Bundle theory of the self concerns synchronic unity: introspection reveals a diversity of thoughts and feelings; is there also a unified I? Bundle theorists (Hume) say no. Bundle theory vs substance-attribute theory concerns objects: strip away an apple's shape, colour and juiciness: is anything left? Bundle theorists say no (properties bound by compresence); substance-attribute theorists posit a substratum. Bundles of universals cannot allow two qualitatively identical objects (a world with only two electrons), so bundle theorists prefer tropes; substratum theorists face unobservability and Bradley's regress.
Canberra Plan to Chance
The Canberra Plan, step by step.
Carving nature at its joints. A language with "grue" (green before midday, blue after) and "bleen" could be as truthful as ours, yet "blue" and "green" seem to reflect reality better. For David Lewis (1983), natural properties ground objective similarity (grass and leaves are alike in being green; morning grass and afternoon Smurfs are not alike in being grue) and figure in causation and laws. Some think naturalness comes in degrees; some think the notion unintelligible.
Categorical and dispositional properties. Fragility and solubility are dispositions ("powers"), defined via conditionals: if struck, it would break. Being square is categorical. Many say fundamental physics characterises mass, charge and spin dispositionally; Humeans insist dispositions are grounded in categorical properties, so science may leave us ignorant of reality's categorical nature.
Causal relata. For type-level causation ("smoking causes cancer") the relata are properties. For token-level causation the favourite is events (Lewis, Kim, Davidson); Mellor (1995) prefers facts ("E because C"). Facts handle causation by absence ("the lack of oxygen caused Joe to fall unconscious") better, since few accept negative events, though negative facts are suspect too. Some metaphysicians therefore deny that absences cause at all, and treat "the lack of oxygen caused Joe to fall unconscious" as loose talk for a counterfactual: had there been oxygen, he would not have fallen unconscious. Causation and laws: the modern debate descends from Hume, who denied natural necessity; Lewis leads the Humeans, Armstrong the anti-Humeans.
Chance is a specific, "rock-bottom" or single-case probability, between 0 and 1 only if indeterminism is true: the chance that a strontium-90 atom decays in the next 15 years, unlike the probability that the next card is an ace (already 1 or 0).
Class to Conceptual vs logical truths
Class. Given some things, there is a class of them: all goats, or "just myself and Brad Pitt's left arm". Classes are identical iff they have the same members; membership is not transitive. Class nominalism says an object is red in virtue of membership of the class of red things. Problem: coextensive properties (every creature with a heart has kidneys) would be identical; Lewis answers with classes across possible worlds. Intuitively, though, the ball is in the class of round things because it is round, and on this view it has no nature of its own.
Compatibilism holds free will (and moral responsibility) compatible with determinism. Its chief opponent is the consequence argument. Hume: liberty is "a power of acting or not acting, according to the determinations of the will". Harry Frankfurt's hierarchical view: acting freely needs only first-order desires (a dog can act freely), but freedom of the will needs effective second-order desires: the non-addict's desire to drink tracks whether he wants to want it; the addict's does not.
Compatibilist analyses and the incompatibilist challenge.
Concept. "The ingredients of thought". Locke, Berkeley and Hume called them "ideas", largely mental images, a view that fails for concepts with no image. Conceptual analysis, dominant for much of the twentieth century, gives necessary and sufficient conditions that hold in all possible worlds ("S performs A freely iff X"), tested by counterexamples: Frankfurt's neurosurgeon against necessity of the ability to do otherwise; random acts against its sufficiency.
Conceptual scheme. Anti-realists say we can think only of reality as constrained by our scheme, roughly our entire stock of concepts (German "Schnee" and "snow" express one concept). Davidson ("On the very idea of a conceptual scheme", 1974) defines distinct schemes as mutually untranslatable languages and argues they are impossible. Conceptual vs logical truths: "all bachelors are unmarried" is true by meaning; "if snow is white, then snow is white" by logical form.
Consequence argument to Direction of time
Contingent truths are true but not necessarily true ("Barack Obama won the 2008 US presidential election"); necessary truths could not have been false (2 + 2 = 4). The counterfactual theory of causation says c causes e iff e would not have occurred without c. Pre-emption breaks the simple version: a back-up shooter who would have fired means my bleeding does not depend on your shot. Lewis defines causation as the ancestral of causal dependence, but trumping, late pre-emption and overdetermination spawn ever more complex theories. The entry presents the dialectic as a cycle of counterexample and refinement, which is also a fair picture of how the analytic method works in practice: each new case tests whether the analysis captures the concept or only its central cases.
Pre-emption: death would have occurred without your shot.
Determinables and determinates. Scarlet is a determinate of red, red of coloured; x is a determinate of y iff having x is a specific way of having y. Some reduce determinable facts to determinate ones. Determinism: the state of the universe at one time P plus the laws L logically entail its state at every other time, as when a pool player knows exactly where the red will go.
Direction of time. The Battle of Hastings, the American Revolution and the Boer War have a natural order; Paris, Moscow and Beijing can be ordered either way. Boltzmann's thermodynamic answer: the direction of time is fixed by increasing entropy. But entropy only tends to increase; a dispersed gas could by chance re-gather, and it seems absurd that time would then run backwards.
Dretske-Tooley-Armstrong to Existence
Two answers to what makes a law: regularity or a relation between universals.
Empiricism: all knowledge of the external world is a posteriori, and (a second thesis) all concepts derive from experience (Locke, Berkeley, Hume, the tabula rasa); meaning-empiricism (Hume) rejects ideas with no source in experience, such as natural necessity and the self. Empiricism sits uneasily with armchair metaphysics. Endurantism is best taken as the denial of perdurantism: objects lack temporal parts and are "wholly present". Endurantists differ on temporary intrinsics and coincidence: the "Standard Account" lets a statue and lump share a place, facing double counting: if statue and lump are two objects each weighing a kilogram, why does the scale read one kilogram? Endurantists answer that they share their matter, so their weights are not additive, but critics ask what then makes them two.
Essence. "The being of a thing, whereby it is what it is" (Locke). Kind essences (what makes Dobbin a horse) differ from individual essences (what makes him Dobbin); biologists dispute whether species have essences (cladism or DNA). Modalist accounts: a is essentially F iff a is F at every world where a exists. This is "thin", fitting Locke's nominal essence, our classificatory features. Anti-modalist ("thick") accounts reverse the order: things have their modal profile because of their essence; only kinds that carve nature at its joints (gold, electron, perhaps horse, but not watches) have essences, fixed independently of mind and discoverable a posteriori (water is H₂O).
Three readings of essence.
Eternalism: all times exist, tenselessly: the dinosaurs, "the Big Bang that happened 15 billion years ago", Martian outposts. It answers presentism's problems with past-tense singular truths (
Worked comparisons from the section
Bundles, substrata and the regress. Three entries, bundle theory vs substance-attribute theory, Bradley's regress and austere nominalism, interlock. The substance-attribute theorist needs something that ties the substratum to its properties; Bradley's regress threatens any tie that is itself a further entity. The bundle theorist needs something that ties properties to one another (compresence) and faces the same threat one level down. The austere nominalist avoids both by refusing to treat "a is F" as needing any constituent analysis at all, which is exactly why Armstrong calls the position ostrich-like. The book lets the reader see that the choice is between kinds of primitive: an unanalysed tie, an unanalysed compresence, or unanalysed predication. Armstrong's own answer, that instantiation is a "non-relational tie", appears in Chapter 2.
Backwards causation, bilking and time travel. The backwards causation entry deliberately separates the causal question from time travel: a magic button in 2010 could give someone a cold in 1955 without anyone travelling. That matters because the standard objections differ. The bilking argument is epistemic: we could never be justified in believing in a backward cause, since we could always try to prevent the earlier event after observing it. The grandfather paradox is logical: if I could kill my grandfather before my parent's birth, a contradiction would follow. Lewis's (1976) response, which the time travel entry (Chapter 4) summarises, is that I can in one sense kill him (I have the gun and the skill) but in another sense cannot (it is not compossible with the past), so I am bound to fail in some ordinary way.
Three views of laws. The DTA entry is best read alongside Ramsey-Lewis (Chapter 4) and regularity (Chapter 4). The naive regularity theory says laws are just true universal generalisations, which lets in accidents such as the lutetium case. The best-system view filters regularities by simplicity and strength; DTA instead posits a second-order relation between universals. The trade-off is clear: DTA explains why regularities hold (they are necessitated), but faces the "inference problem", how a relation between universals can entail anything about particulars, while the best-system view needs no extra entities but makes lawhood depend on standards of simplicity that may seem ours rather than nature's.
Two senses of "determinism". The section's determinism and chance entries prepare one of the book's clearest corrections of a popular confusion. Determinism is a thesis about entailment by the laws plus a complete state at a time, not about predictability in practice: a deterministic world may be chaotic and unpredictable, while an indeterministic world may be highly predictable in aggregate. And determinism is an empirical thesis, so, as the book says, "it is not up to philosophers to take a stand on whether or not the Universe is really deterministic".
Reading the entries together: three threads
Thread one: the vocabulary of ontology. The section's short definitional entries are not filler. Abstract vs concrete, class, being, existence and event together fix the terms in which later debates are conducted. The book's approach is to distinguish senses before taking sides: "abstract" is first explained by examples, then by the rough criterion of being outside space and time, and the reader is warned that the criterion is itself disputed. The same pattern recurs in essence, where Locke's phrase is quoted and then divided into kind and individual essence and into modalist and anti-modalist readings, and in existence, where the three "orthodox" features are set out before their dissenters. A reader who learns these distinctions early is protected against the failure the Introduction warns of: two parties "using a key term to mean different things".
Thread two: free will from both sides. Four entries in this stretch, agent causation, compatibilism, conceptual analysis and the consequence argument, already lay out the whole free will debate that Chapter 2 completes. The consequence argument gives the incompatibilist's strongest case, resting on a principle about what we can render false. Compatibilism answers in two ways: deny that freedom requires the ability to do otherwise (Frankfurt, previewed in conceptual analysis), or analyse that ability so it fits determinism. Agent causation offers the libertarian's positive story, and the book notes honestly that it is "mysterious" why agents alone should have such powers. The method entry, conceptual analysis, shows how such disputes proceed: proposed necessary and sufficient conditions meet counterexamples aimed at either half.
Thread three: causation as a family of problems. Causal relata, causation and laws, causation by absence, chance, the counterfactual theory, determinism and the DTA view of laws form a compact introduction to the causation debate. The organising contrast is Humean against anti-Humean: Lewis analyses causation through counterfactual dependence and laws through the best system, while Armstrong posits a real relation of necessitation between universals. The entries also show why the relata matter: whether absences can cause depends on whether there are negative events or facts, and whether indeterministic causation is possible depends on how chance is understood. Many of these threads are completed in Chapter 3, where the neuron diagrams give the counterfactual theory's problem cases a visual form.
What the section models. The entries are written for students, and the prose is direct, with a fondness for vivid examples: Marty's 1955 cold, Jane's pizza, a telephone box and a Coke can, Brad Pitt's left arm. These examples do real work. They make abstract principles testable and give later entries a shared stock of cases. The cost is occasional looseness, as when a figure for the age of the universe or a statement of physical law is given casually; the closer look below flags those instances.
Key arguments reconstructed
The consequence argument, step by step. (1) If determinism is true, the remote past together with the laws entails every proposition about what I do. (2) No one can render false a proposition about the past before their birth. (3) No one can render false a law of nature. (4) If no one can render p false, and p entails q, then no one can render q false (the transfer principle). With the premise that free action requires the ability to do otherwise, it follows that no one acts freely. Compatibilists most often attack (4), arguing that "can" in the relevant sense does not transfer across entailment, or attack the link between freedom and the ability to do otherwise.
The bilking argument. (1) Suppose a later event B is claimed to cause an earlier event A. (2) After A has occurred (or failed to occur) and been observed, we could try to prevent B. (3) If we succeeded in preventing B on occasions when A occurred, the claim of backward causation would be refuted; if we always failed, that pattern would call for explanation. (4) Either way we would never be justified in believing that B causes A. The argument is epistemic: it does not show backwards causation impossible, only that its believers face a test they cannot pass. Dummett's shaman shows that a community could rationally keep dancing only if it does not try to bilk.
Putnam's anti-sceptical argument. (1) Words refer to what they are suitably causally connected to. (2) A brain in a vat has no causal connection to real vats and brains, only to computer signals or images. (3) So its word "vat" does not refer to vats, and its sentence "I am a brain in a vat" does not say that it is a brain in a vat. (4) Hence if I am a brain in a vat, my sentence "I am a brain in a vat" is false; if I am not, it is false too. The hypothesis cannot be truly asserted. Critics accept the semantics but deny the anti-sceptical conclusion, since I cannot know which language I speak without first knowing that I am not envatted.
Why bundles of universals cannot be twins. (1) On bundle theory an object is nothing but its properties. (2) If properties are universals, two objects with exactly the same properties are bundles of numerically the same items. (3) So they are one object, not two. (4) But a world with two qualitatively identical electrons seems possible. So either bundle theory is false, its properties are tropes, or the identity of indiscernibles is necessarily true. The entry adds that the argument assumes bundle theory is, if true, necessarily true; a contingent bundle theorist might accept that twins are impossible only in our world.
How the entries have aged
Still current. Most entries in this section remain accurate introductions. The definitions of analytic, a priori, ancestral relation, class and determinism are textbook-standard, and the accounts of the consequence argument, Bradley's regress and the counterfactual theory match how the debates are still taught. The Canberra Plan entry is, if anything, more relevant now, since conceptual engineering (asking what our concepts should be, not only what they are) has grown out of the same Australian tradition; the 2020 survey found 39.5% accepting conceptual engineering as a method.
Shifted emphasis. Three entries would be written differently today. Essence would give more space to the anti-modalist view, which after Fine has become the dominant framework, and would connect essence to grounding. Categorical and dispositional properties would note the rise of powers metaphysics and of "neo-Aristotelian" metaphysics generally (Tahko, ed., Contemporary Aristotelian Metaphysics, 2012). Causation by absence would note that causal modelling treats absences as values of variables, removing much of the ontological worry about negative events.
Dated details. The section's dated facts are few: the age of the universe, Armstrong's death in 2014, and Barack Obama's presidency (2009-17), used as an example in contingent truths and event. None affects the philosophy.
What the section asks of the reader. The entries assume no prior philosophy but do assume patience with distinctions. A reader meeting "analytic", "a priori" and "necessary" for the first time should note that they are defined in terms of meaning, knowledge and modality respectively, and that the whole interest of later entries (Kripke's necessary a posteriori, Kant's synthetic a priori) lies in cases where the three come apart. Likewise "abstract", "universal" and "class" overlap without coinciding: platonic universals are abstract, immanent universals are not, and classes are abstract but not universals. Keeping these vocabularies straight is the price of entry to the rest of the book.
The section's further reading
Each entry ends with one to three suggestions, chosen for accessibility. For abstracta the authors send readers to Swoyer (2008); for agent causation to Clarke (1993, 2008); for animalism to Olson's work; for backwards causation to Black (1956) and Dummett (1964); for brains in a vat to Putnam (1981, Chapter 1); for class nominalism to Armstrong (1989) and Loux (2006); for categorical and dispositional properties to Fara (2009); for determinables to Sanford (2008); and for essence to Locke's Essay (Book III, Chapter 3). The section's reading list leans on the Stanford Encyclopedia of Philosophy, which has continued to update most of these entries since 2011, so a reader today can follow the same pointers to newer versions.
Three orthodox features of existence, each with dissenters.
| Term | Meaning or cross-reference |
|---|---|
| Analysis | See conceptual analysis |
| Anti-criterialism | The simple view of personal identity |
| A-theory, B-theory | See tensed and tenseless theories of time |
| Bilking argument | See backwards causation |
| Criterion of ontological commitment | See ontological commitment |
| Diachronic identity | Identity over time |
| Dispositional properties | See categorical and dispositional |
| Exemplification | The instantiation relation |
Short and cross-reference entries in this section.
Depth: animalism, agent causation and the consequence argument
Animalism, full structure. Animalists ground personal identity in biological facts: each person on this planet is, as a contingent matter, identical to a human animal. For Old Jack to be Wee Johnny is for them to be the same human animal. The view does not say that all persons must be animals (other planets or future machines could host persons), nor that all human animals are persons (a foetus or someone in a permanent vegetative state is a human animal). Eric Olson treats "person" as a phase sortal like "child": we are not essentially persons. That fits the common belief that "she was once a foetus" and might become "a vegetable". The remaining task is the diachronic identity of human animals, which animalists take to be more tractable than the identity of persons as such.
Brain-transplant cost. The main intuitive cost is the brain-transplant case: intuition seems to follow the person with the brain (or psychology), while animalism follows the animal that keeps most of the organism. Olson replies that we confuse identity with what matters for responsibility and prudence: we might still punish Cameron-with-Obama's-brain for Obama's crimes without those persons being identical. The entry thus separates the metaphysical question (who is numerically identical to whom?) from the normative question (whom may we hold responsible?).
Olson-style animalism in three theses.
Agent causation, full dialectic. Most philosophers take the causal relata to be events, facts or states of affairs: Jane's being hungry at t caused her decision to order pizza. On agent causation, Jane herself, a substance, causes the decision. The view is usually tied to indeterminism: the agent must be an "uncaused cause", and the decision must not be determined by past plus laws. Some agent causalists (Markosian 1999) are compatibilists. Criticisms: acting rationally seems to need reasons as causes (Clarke 1993 replies that agent causation can allow reasons); and it is "mysterious" what gives agents alone this power. Dualist agent causalists answer that an immaterial substance makes the difference from dogs and rocks; not all agent causalists require dualism.
Consequence argument, transfer principle. Van Inwagen (1975) characterises the ability to do otherwise as the ability to render a proposition false. Jane made tea at 8 am on 1 January 2009 (time t); could she have rendered "Jane made tea at t" false? Under determinism that proposition is entailed by the remote past plus the laws. No one can render false a proposition about the past before their birth, and no one can render false a law. By the transfer principle (if no one can render p false, and p entails q, then no one can render q false), no one can render false any proposition about what they do. With PAP, free will is incompatible with determinism. Compatibilists attack transfer, or attack PAP (see Frankfurt in Chapter 2).
Depth: compatibilism, carving joints, bundles and chance
Compatibilism beyond the conditional analysis. Hume's "hypothetical liberty" is the power of acting or not according to the will, belonging to "every one who is not a prisoner and in chains". Freedom of action needs only effective first-order desires (a dog can act freely). Freedom of the will needs effective second-order desires: the non-addict's desire to drink tracks whether he wants to want it; the unwilling addict's does not. Fischer separates guidance control (acting for one's own reasons, sufficient for responsibility) from regulative control (genuine alternatives). In a Frankfurt case, Jones has guidance control without regulative control, and is still responsible.
Carving nature at its joints, Lewis 1983. A language with "grue" (green before midday, blue after) and "bleen" could be as truthful as ours. Still, "blue" and "green" seem to reflect reality better. For Lewis, natural properties ground objective similarity (grass and leaves are alike in being green; morning grass and afternoon Smurfs are not alike in being grue) and figure in causation and laws. Naturalness may come in degrees. The entry ties the notion to metaontology: whether rival languages carve equally well is part of the dispute between Hirsch and Sider. Armstrong's Universals and Scientific Realism is structured around the need for some account of natural classes.
Frankfurt's hierarchical model of free will.
Bundle theory vs substance-attribute, full cost. Strip an apple's shape, colour and juiciness: is anything left? Bundle theorists say no (properties bound by compresence); substance-attribute theorists posit a substratum. Bundles of universals cannot allow two qualitatively identical objects (a world with only two electrons), so bundle theorists prefer tropes. Substratum theorists face unobservability and Bradley's regress. Armstrong's alternative treats instantiation as a "non-relational tie" inside a world of states of affairs, so that focusing on Plato or on wisdom is focusing on aspects of an irreducibly unified fact.
Bradley's regress, viciousness. Plato and wisdom both existing does not make Plato wise. Add an instantiation relation: still not enough, since it could relate Socrates to wisdom instead. Add instantiation* relating Plato to instantiation, and so on. At each stage the posited entities could exist without Plato being wise, so the regress looks vicious for substance-attribute realism, and perhaps for trope versions too. The entry's lesson is that any relational analysis of predication risks an open-ended demand for further ties.
Chance as single-case probability. Chance is rock-bottom or single-case probability, between 0 and 1 only if indeterminism is true: the chance that a strontium-90 atom decays in the next 15 years, unlike the probability that the next card is an ace (already 1 or 0 once the deck is fixed). The entry prepares the later link between chance, probabilistic causation and the problem of luck for libertarianism.
Depth: Canberra Plan, class, conceptual analysis and essence
Canberra Plan, step by step. Collect the platitudes about C (the folk theory): pain is typically caused by bodily damage (hot iron, pin, falling brick) and causes behaviour ("ouch!"). Ask what in the world plays the role. If a brain state does, physicalism about pain looks plausible. Menzies (1996) applied the Plan to causation, with platitudes such as "causes typically raise the chances of their effects" and "causation is an intrinsic relation"; the role-occupant might be energy transfer. The method bridges conceptual analysis to ontology: the analysis gives the role, science supplies the occupant. Critics doubt that there is a determinate folk theory, and note that platitudes often conflict.
Class and class nominalism. Given some things, there is a class of them: all goats, or "just myself and Brad Pitt's left arm". Classes are identical iff they have the same members; membership is not transitive. Class nominalism says an object is red in virtue of membership of the class of red things. Problem: coextensive properties (every creature with a heart has kidneys) would be identical; Lewis answers with classes across possible worlds. Intuitively the ball is in the class of round things because it is round, so the view seems to reverse explanation: the object has no nature of its own.
From folk theory to worldly realiser.
Conceptual analysis and its tests. Dominant for much of the twentieth century, analysis gives necessary and sufficient conditions that hold in all possible worlds ("S performs A freely iff X"), tested by counterexamples. Frankfurt's neurosurgeon targets the necessity of the ability to do otherwise for responsibility; random or coerced acts target its sufficiency. The entry is also a method entry: it shows how the whole book proceeds, by proposing analyses and trying to break them.
Essence, modalist and anti-modalist. Locke: "the being of a thing, whereby it is what it is". Kind essences (what makes Dobbin a horse) differ from individual essences (what makes him Dobbin). Biologists dispute whether species have essences (cladism or DNA). Modalist account: a is essentially F iff a is F at every world where a exists (thin, near Locke's nominal essence). Anti-modalist ("thick") accounts reverse the order: things have their modal profile because of their essence; only kinds that carve nature at its joints (gold, electron, perhaps horse, not watches) have essences, fixed independently of mind and discoverable a posteriori (water is H₂O).
Eternalism's motivations. All times exist tenselessly: dinosaurs, the Big Bang, Martian outposts. Eternalists stress they do not exist now, just as things on the other side of the planet do not exist here. Motivations: presentism's trouble with past-tense singular truths (
Depth: animalism's thinking-animal problem and transplant replies
Too many thinkers. If you are a person and not identical to the human animal in your chair, yet that animal thinks and reads, then two thinkers occupy your chair. That is weird, and it raises an epistemological puzzle: how do you know you are the person rather than the animal? Animalism avoids the double count by identifying you with the animal. Critics either deny that the animal thinks in the full sense, or accept a temporary coincidence of thinkers, or reject the assumption that persons are substances distinct from animals.
Animalism's double-thinker pressure on dualism of person and organism.
Transplant intuition. The brain (or cerebrum) transplant intuition says you go with your psychology into another body. Animalists may (i) deny the intuition, (ii) treat the shaved-down brain as still the animal "whittled" and rebuilt, or (iii) follow Olson in separating identity from what matters for survival and responsibility: we might punish the being with Obama's brain for Obama's crimes without numerical identity. The entry thus forces a choice among identity facts, practical concerns and biological continuity.
Depth: DTA inference problem, gunk, sets and negative facts
DTA inference problem. If N(F, G) is a contingent relation between universals, why does it entail that particular F's are G's? Armstrong needs the necessitation relation to underwrite the regularity without becoming a further unexplained posit. Critics say N is a "dormitive virtue" label unless its modal force is explained. Friends of DTA reply that N is primitive but still explanatory relative to bare regularity.
Gunk and atomless matter. "Gunk" is matter every part of which has proper parts: no simple atoms. The possibility of gunk pressures views that require a foundation of mereological atoms (some nihilisms; some trope bundle stories). It also interacts with supertasks and continuum metaphysics in later entries.
Further Chapter 1 devices used later in the book.
Sets, classes and ontological commitment. Set theory supplies the vocabulary for class nominalism and for ersatz worlds. The entry reminds readers that Quinean commitment tracks the values of bound variables: talking about sets in the metalanguage may already incur ontology. Negative facts and absences (the lack of a hippopotamus in the room) challenge truthmaker maximalism and feature in causal theories that invoke prevention. INUS conditions (insufficient but non-redundant parts of unnecessary but sufficient conditions) offer an older Mackie-style analysis of causation that the counterfactual and process entries later compete with.
Depth: abstract/concrete, causation menu and endurantism
Abstract vs concrete. Numbers, sets and universals (on some views) are abstract; tables, electrons and events are concrete. Tests proposed: spatiotemporal location, causal efficacy, mind-independence. None is uncontroversial (immanent universals are located; some deny numbers are causal). The distinction structures ontology: Quinean commitment to numbers via mathematics is a commitment to abstracta unless paraphrased away.
Causation theories, preview of the neuron cases. Counterfactual dependence (Lewis): c causes e if e would not have occurred without c (refined for pre-emption via chains). Probabilistic: causes raise the chance of their effects (Eva/Eric coin cases). Process / conserved quantity: causation is the transfer of energy or mark transmission. The book's Figures 2 to 4 (neuron diagrams) test these against early and late pre-emption and trumping. No single analysis wins cleanly in the 2011 text; interventionist descendants (Woodward) appear in literature notes as later developments.
Three causation menus tested by the neuron diagrams.
Endurantism. Objects persist by being wholly present at each time. Paired often with presentism, but combinable with eternalism if temporary intrinsics are handled relationally. Endurantists reject temporal parts as the deep story of persistence. The Hogan height problem and the statue/clay puzzle are shared battlegrounds with perdurantism.
Depth: Bradley's regress and the non-relational tie
Why each stage fails. Plato exists; wisdom exists; yet Plato need not be wise. Add Instantiation relating them: Instantiation could relate Socrates to wisdom instead, so the three entities do not yet yield Plato's wisdom. Add Instantiation* relating Plato to Instantiation: again the entities could exist without the predication. At every finite stage the inventory is compatible with the predication's failure, so the regress looks vicious for any relational analysis of having a property.
Armstrong's exit. Treat the state of affairs Plato's being wise as an irreducibly unified hunk of reality. Focusing on Plato or on wisdom is focusing on an aspect of that unity; there is no further relation that must be glued on. Critics ask whether the "non-relational tie" is only a label. Friends reply that some stopping point is inevitable, and that states of affairs earn their keep in truthmaker theory and in DTA. Trope bundle theories face a cousin regress about compresence unless compresence is likewise non-relational or primitive.
Consequence argument core transfer.
Tests for abstract versus concrete.
Two persistence models.
Three law ontologies introduced by E.
Canberra Plan applied to pain.
Experimental Philosophy to Natural Kinds
Experimental philosophy to Free will
Experimental philosophy once meant what we now call science (Bacon, the Royal Society, Newton, Boyle; Hume's Treatise subtitle promises "the experimental method of reasoning"). Today it means testing philosophical intuitions by questionnaire, typically on undergraduates: is Joe, whose bank robbery was fully determined, morally responsible? A more ambitious form, close to psychology, explains why we have the intuitions: describing a case abstractly or concretely and personally can flip answers about determinism and responsibility.
Facts. Some mean only true propositions; others take facts as real constituents of nature, as did the early Wittgenstein: "The world is the totality of facts, not of things". Facts serve as truthmakers and as causal relata. Fictionalism treats unwanted entities as fictions: it is literally false that Holmes lived at 221b Baker Street, but true-in-the-fiction. "There is a prime number between 5 and 11" is true according to the platonist fiction (Field 1980). Hermeneutic fictionalism says that is what we always meant; revolutionary fictionalism says it is what we should say. "Modal fictionalism" (Rosen 1990) is a misnomer: modal claims stay true; possible worlds become fictional.
Fission: two hemispheres, two persons, one original.
Three ways to answer fission.
Flow of time. Space does not drag everything eastwards, but we advance through time "second after second" whatever we do. Flow differs from direction: you can swim upstream in a river, not in time. Tensed theorists and presentists take flow as evidence; eternalists and tenseless theorists explain the appearance by accumulating memories. Critics ask how fast time flows: "one second per second" seems empty. Form and matter: Aristotle's toaster is metal (matter) organised as a toaster (form). Four-dimensionalism may mean perdurantism, eternalism, or both together.
Free will. "Freedom of the will" comes from faculty psychology; many now speak of freedom of action (including choosing and deciding). Incompatibilists say determinism and free will cannot be reconciled; whether determinism holds is for science, which has not settled it, leaving us unsure whether we are ever morally responsible. Since few think determinism likely, hard and soft determinism are now unpopular. A further worry is luck: if our decisions are undetermined, do we control them? The entry stresses that "the" problem of free will has evolved into several problems, with different solutions, so the reader should ask which problem a given position addresses.
| Position | Determinism? | Compatible with free will? | Agents sometimes act freely? |
|---|---|---|---|
| Compatibilism | ? | Y | ? |
| Incompatibilism | ? | N | ? |
| Libertarianism | N | N | Y |
| Soft determinism | Y | Y | Y |
| Hard determinism | Y | N | N |
| Illusionism | ? | N | N |
| Agent causalism | ? | ? | ? |
Table 1 (redrawn): Y = committed to the thesis, N = committed to its denial, ? = neutral.
Fusion to Holes
Mereological fusion: x is the fusion of the ys iff the ys compose x (not the opposite of fission, though Parfit uses it that way). God's eye view: global anti-realists such as Putnam say realism presupposes a "God's Eye point of view", or "a No Eye view of truth", which they reject as unintelligible.
Presentism, the growing block and eternalism.
Guidance vs regulative control (Fischer): guidance control (deciding when to set your alarm, then getting up) fits determinism; regulative control needs alternatives. Gunk: every part has a proper part; that differs from infinite divisibility. Haecceitism: identity across worlds is brute, so a world could differ from ours only in that Elvis and Napoleon swap qualities (Napoleon born 1935 in Mississippi, Elvis in 1769 in Corsica); haecceitism is not belief in haecceities. Hard determinism holds incompatibilism plus determinism, so that nobody is ever morally responsible; its popularity has waned as philosophers accept that indeterminism may be true; P. F. Strawson argued that the reactive attitudes it would abolish give reason to reject it.
Figure 1 (redrawn): enlarging a hole can shorten its lining (shown in orange), so the hole cannot simply be the lining.
Three realist theories of holes.
Hume to Instantiation
David Hume (1711-76), Scottish philosopher and historian, best known in life for his six-volume History of England; the last and most hard-hitting British Empiricist, denying grounds for believing that impressions resemble external objects. Humeanism eschews natural necessity or accepts Humean supervenience: regularity and counterfactual theories of causation and the Ramsey-Lewis view of laws. Humean supervenience (Lewis): "all there is to the world is a vast mosaic of local matters of particular fact, just one little thing and then another ... an arrangement of qualities"; laws, chance and causation supervene on it. Its chief threat is quantum entanglement, which seems non-local; Lewis himself conceded that if physics required irreducibly non-local facts, Humean supervenience would be false, and defended only its contingent truth, its compatibility with a world like ours.
Lewis's Humean supervenience.
Qualitative versus numerical identity.
Identity. Two guests in "the same dress" share qualitative identity; squeezing into the very same dress would be numerical identity. Confusing them wrecks the personal identity debate. Identity over time (diachronic): the child in the old photograph is you (the entry jokes that saying you are identical to some "things" is a grammatical slip, since you are only ever identical to one thing); you are not Genghis Khan, a young Mel Gibson or a seventeenth-century table. It concerns persistence conditions, change and how things persist (endurance or perdurance).
Illusion of free will. Not every false belief is an illusion (thinking Sydney is Australia's capital is just false). Free will may be an illusion because it seems we could do otherwise or are uncaused causes; or in Freud's sense, a belief motivated by wish-fulfilment, as Smilansky (2000) holds, recommending we keep it. Immanent realism: universals are wholly present in their instances, so none is uninstantiated (Aristotle, Armstrong).
Incompatibilism. You made tea at nine o'clock; deliberating between tea and coffee presupposes you could do either. Determinism removes that ability (the Consequence Argument). A second, causal chain argument says determinism places the ultimate source of action outside us. Indeterminism: quantum mechanics is often taken to refute determinism: strontium-90, spread over Europe by Chernobyl (1986), has a 28-year half-life, a single-case chance rather than ignorance, as with a poker player's estimate. Indexicals ("I", "here") shift reference with context (Angela Merkel in Berlin). They matter to metaphysics because tensed language ("now") is indexical, and B-theorists argue that the special status of the present is like the special status of here: a feature of perspective, not of reality.
Classical and modern views of the infinite.
Instantiation (exemplification) relates an object to its properties; taken as a relation it invites Bradley's regress, so Armstrong calls it a "non-relational tie".
Intrinsic properties to Lewis
Intrinsic vs extrinsic. Mass is intrinsic; weight varies with surroundings (none in deep space, more near Jupiter). Being an uncle is extrinsic: a birth across the world changes you extrinsically. The distinction matters for Humean supervenience (a mosaic of intrinsic qualities) and for the problem of temporary intrinsics. Ordinary language uses "intrinsic" more strongly, for essential features; the philosophical sense is weaker. Definitions via duplicates and loneliness (Langton and Lewis) are contested. INUS conditions (Mackie 1965): placing a bomb (b) and the timer reaching zero (z) are each an Insufficient but Necessary part of a condition (b & z) that is Unnecessary but Sufficient for the bridge's collapse (rust could also do it).
Immanuel Kant (1724-1804) never left Königsberg (now Kaliningrad). The Critique of Pure Reason (1781), by his own account "dry, obscure ... and moreover long-winded", and the Prolegomena (1783) answer Hume, who "struck a spark". Kant asks how synthetic a priori truths are possible: metaphysics studies the necessary structure of experience.
Saul Kripke, a teenage publisher in modal logic, founded possible worlds semantics and wrote Naming and Necessity (1980; Princeton lectures, 1970), source of necessity of origin and the necessary a posteriori. Laws of nature are non-accidental generalities: e = mc² versus "all lumps of solid gold are under ten metres". Leibniz's Law: the indiscernibility of identicals is uncontroversial; the identity of indiscernibles is contested (Black's two spheres). David Lewis (born 1941, Oberlin, Ohio) switched from chemistry after hearing Ryle, studied under Quine, taught at Princeton, loved Australia (its football and bush songs), and pursued Humean supervenience across modal realism, counterfactual causation, the Ramsey-Lewis laws, perdurantism and time travel, writing until his death in 2001.
Libertarianism to Natural kinds
Linguistic turn: mid-twentieth-century attention shifted from reality to language. John Locke (1632-1704), first British Empiricist, author of the Essay (1690), shaped debates on personal identity, essence, primary and secondary qualities and universals. J. M. E. McTaggart (1866-1925), lecturer at Trinity College, Cambridge, held philosophy breakfasts with no food, attended by Moore and Russell; see the unreality of time (Chapter 4). Meinongians say some things do not exist. The memory theory (Locke): later y is earlier x iff y remembers x's experiences.
Mereology, the study of parts and wholes, was formalised by Leśniewski and spread by Leonard and Goodman (1940s). Systems differ on transitivity, gunk vs simples, and when composition occurs (the Special Composition Question).
Metaphobes and the road back to metaphysics.
Modality studies possibility and necessity: "2 + 2 = 4" is necessary; "Barack Obama is President" (true in 2011) contingent. Modal realists believe possible worlds exist. Genuine (concrete) modal realism (Lewis): worlds are universes spatiotemporally disconnected from ours, including one where everyone dons clown outfits to sing Ricky Martin's Livin' la Vida Loca. Benefits: reducing modality to what exists, and identifying propositions and properties with sets of worlds or possibilia. Costs: the "incredulous stare", and whether other worlds' inhabitants are relevant to what I could do. Lewis answers the second worry with counterpart theory: I could have been a fisherman because a counterpart of mine, someone relevantly similar, is a fisherman in another world. Ersatz modal realists make worlds abstract (maximal propositions or states of affairs), at the price of primitive modality.
Moorean facts are common-sense propositions so well justified that no philosophical argument against them should be taken seriously (Moore raising his hands against scepticism). The concept gives mild conservatism about common sense a sharper form: some premises are more certain than any philosophical premise could be, so an argument against them is better read as a reductio of its own assumptions. Moral responsibility may be metaphysically deep (originating causation, incompatibilist ability) or shallow (fit for praise and blame, Strawson's reactive attitudes). The moving spotlight combines eternalism with the tensed theory: all times exist, and the present moves along them like a spotlight down a street. Natural kinds carve nature at its joints: electrons, water and perhaps tigers, but not "ravens-or-writing-desks"; for some (Ellis 2001) they are universals, for others kinds with necessary a posteriori identifications ("water = H₂O", "gold is the element with atomic number 79").
Worked comparisons from the section
Three ways to avoid an entity. Fictionalism, the paraphrase strategy (applied in holes) and austere nominalism (Chapter 1) are three anti-realist techniques, and the section lets the reader compare them. The paraphraser claims that a sentence apparently about holes really means something about perforated objects; the difficulty is finding paraphrases for hard sentences and, worse, the lack of a general method for producing them. The fictionalist concedes that the sentence is literally false and explains its usefulness by its truth in a story; the cost is that we then assert falsehoods all the time, and must explain why the story is so useful. The austere nominalist concedes nothing to the question at all. Each strategy is a way of avoiding an ontological commitment on Quine's criterion, and the choice among them depends on which costs one finds least bad.
Frankfurt and the ability to do otherwise. The nefarious neurosurgeon is a thought experiment about moral responsibility, not directly about freedom, but because freedom is taken to be necessary for responsibility, a case of responsibility without alternatives is also a case of freedom without alternatives. That is why the case threatens the Consequence Argument's premise that free action requires the ability to do otherwise. The incompatibilist's best replies either find a residual alternative (the "flicker" of beginning to decide otherwise) or argue that Frankfurt cases presuppose determinism: Black can predict Jones's decision only if it is determined by prior signs, in which case the case begs the question.
Hume, Humeanism and natural necessity. The section has three entries with Hume's name: the man, Humeanism and Humean supervenience. Hume himself argued that we have no impression of necessary connection between causes and effects, only of constant conjunction, and (arguably) concluded that the idea of necessity is projected by the mind. Modern Humeans, led by Lewis, convert this into an ontological thesis: there is no natural necessity in the world, and causation, laws and chance must supervene on the mosaic of particular facts. Anti-Humeans (Armstrong, dispositional essentialists) reply that the mosaic cannot explain why regularities hold. The issue returns in Chapter 3 as natural necessity and in Chapter 4 as quidditism.
Infinity across the book. The infinity entry supplies tools used elsewhere: Aristotle's distinction between actual and potential infinity, which bears on Zeno and supertasks, and Cantor's sizes of infinity (the "aleph" numbers of Chapter 1), which bear on how many possible worlds or classes there are. The entry's lesson is that mathematical progress changed what metaphysicians could coherently say: after Cantor, "infinite" names a family of precise sizes, where before it named only the endless.
Reading the entries together: a map of the free will debate
The positions. Table 1 is the section's centrepiece, and the entries around it explain each row. The traditional question is whether free will is compatible with determinism. Compatibilists say yes; incompatibilists say no. Combining each answer with a verdict on determinism yields the classical positions: soft determinism (compatible, and determinism is true), hard determinism (incompatible, and determinism is true, so no free will) and libertarianism (incompatible, determinism false, and we are free). Illusionism adds that belief in free will is an illusion, either perceptual (it seems we could do otherwise) or Freudian (we want it to be true). Agent causalism cuts across the grid: it is a view about the kind of causation involved in free action, which is why its row is all question marks.
The arguments. Two incompatibilist arguments organise the field. The Consequence Argument (Chapter 1) says determinism removes the ability to do otherwise; Frankfurt cases reply that responsibility does not need that ability. The causal chain or source argument says determinism places the ultimate origin of our actions in the remote past; source libertarians and agent causalists answer by making the agent an originator. On the other side, the problem of luck (Chapter 3) argues that indeterminism does not help: if nothing about Jane settles which way she decides, it seems a matter of chance. The debate is thus a dilemma: determinism seems to rule freedom out, and indeterminism seems not to rule it in.
The method. The free will entries display the book's attention to concepts. "Freedom of the will" is traced to faculty psychology; guidance and regulative control (Fischer) separate two things "control" might mean; moral responsibility is split into "deep" and "shallow" conceptions. Many apparent disagreements turn out to be disagreements about which concept is in play. The book also notes when a question belongs to science: whether determinism is true is not for philosophers to settle, and experimental philosophy asks what ordinary people actually think.
Arguments that organise the free will debate.
Reading the entries together: time, identity and modality
Time. Flow of time, four-dimensionalism, growing block and moving spotlight set out the ontology of time that Chapter 3 and 4 complete with presentism, tensed and tenseless theories and McTaggart. The section's lesson is that two questions must be kept apart: which times exist (presentism, growing block, eternalism) and whether there are objective tensed facts (A-theory versus B-theory). The combinations are not all equally natural, but the moving spotlight shows that eternalism and the A-theory can go together.
Identity. Identity, identity over time, fission, Leibniz's Law and the memory theory build the formal background of the personal identity debate. The key formal facts are that identity is one-one, transitive and obeys the indiscernibility of identicals. Fission exploits the first two; the statue and lump puzzle (Chapter 1) exploits the third. Once these formal properties are clear, the reader sees why every theory of personal identity must say something about branching cases.
Modality and method. Modal realism, haecceitism, Kripke, Lewis and metametaphysics connect the section to the question of what metaphysics can know. Lewis's genuine modal realism is offered for its theoretical benefits, not its intuitive appeal; the incredulous stare is a cost to be weighed. Metametaphysics then asks whether such cost-benefit reasoning can deliver knowledge of mind-independent reality at all: the metaphobe's challenge from Hume to Carnap, and the metaphile's revival after Quine.
Key arguments reconstructed
Fission against psychological continuity. (1) Person x at t1 is identical to y at t2 iff x is psychologically continuous with y. (2) After fission, both the person with Obama's body (L) and the person with Cameron's body (R) are psychologically continuous with the Queen. (3) So the Queen = L and the Queen = R. (4) By symmetry and transitivity, L = R. (5) But L and R are two people, leading separate lives. So (1) is false as stated. The no-branching clause blocks (3) by denying either is the Queen; the price is that whether L is the Queen depends on whether R exists, an extrinsic fact about something that happens elsewhere, which seems absurd.
The dead past objection. (1) On the growing block, past times and their contents exist, including Genghis Khan in AD 1220 thinking "this is the present". (2) His belief is false: the present has moved on. (3) My epistemic situation, a thinking being with experiences as of the present, is just like his. (4) So I have no reason to think my time is present rather than past. The growing block theorist must say what makes the present special in a way detectable from inside, or must accept the sceptical result. The objection does not touch presentism (no past thinkers exist) or the B-theory (no special present to be wrong about).
Frankfurt against PAP. (1) Jones decides on his own to kill Smith; Black's device never acts. (2) Jones is morally responsible for the killing. (4) So responsibility does not require the ability to do otherwise; PAP is false. The incompatibilist's main reply, the dilemma defence, asks how Black knows what Jones will decide.
Cantor and the sizes of infinity. (1) Two collections have the same size iff their members can be put in one-to-one correspondence. (2) The natural numbers and the even numbers can be so paired (1 with 2, 2 with 4, and so on), so a collection can be the same size as a part of itself. (3) But no pairing of the natural numbers with the real numbers can be complete: given any list of reals, Cantor's diagonal construction defines a real not on the list. (4) So there are more reals than naturals: infinities come in different sizes. Metaphysicians use the result when counting possible worlds, properties or classes.
How the entries have aged
Still current. The section's central entries, free will, incompatibilism, libertarianism, Frankfurt, fission, identity and modal realism, remain accurate. Table 1 still maps the classical positions, though a modern version would add a row for free will scepticism (no free will in the basic-desert sense, whether or not determinism is true), which now has more defenders than hard determinism.
Shifted emphasis. Experimental philosophy has moved from a challenge to the armchair into a broad empirical programme, including large cross-cultural studies of intuitions about knowledge, reference and responsibility. Metametaphysics has become a recognised subfield with its own handbooks; the Carnap-Quine debate the entry describes is now continued by easy ontology (Thomasson), quantifier variance (Hirsch) and structure (Sider). Humean supervenience is now discussed mainly through the best-system account of laws and chances, and through debates over whether quantum entanglement refutes it.
Dated details. Kripke (born 1940) died in 2022; the Queen and the politicians in the fission example belong to their time (Elizabeth II died in 2022; Cameron left office in 2016). Kant's university has been renamed again. The strontium-90 remark in indeterminism, true in 2011, is no longer true: more than half of the Chernobyl strontium has now decayed.
What the section asks of the reader. This is the book's longest stretch of argument, and it rewards reading the free will entries in the order Table 1 suggests: free will, incompatibilism, Frankfurt, libertarianism, illusion of free will, then (in Chapter 3) PAP, neuroscience and luck. Readers should also notice how often the section returns to identity in its formal sense: fission, identity over time, haecceitism and Leibniz's Law all turn on what it is for something to be one and the same thing, and the section's examples (the same dress, Elvis and Napoleon, the Queen's hemispheres) are designed to make that formal notion vivid.
The section's further reading
The entries' suggestions include O'Connor (2008), Blackburn (1999, Chapter 3) and Beebee and Dodd (2006, Chapter 2) for free will; Casati and Varzi (2009) for holes; Mumford (2007) for Armstrong; Hall (2010) and Nolan (2005) for Lewis; Ahmed (2007) for Kripke; Mele (2008) for neuroscience and free will (in Chapter 3); and Field (1980) and Rosen (1990) for fictionalism. Readers of the free will entries will also find Kane's Oxford Handbook of Free Will (2nd edn, 2011) an authoritative survey.
| Term | Meaning or cross-reference |
|---|---|
| Extrinsic properties | See intrinsic vs extrinsic |
| Frankfurt-style cases | See Frankfurt's neurosurgeon |
| Gedankenexperiment | German for thought experiment |
| General causal sentences | Type-level causal claims |
| Grandfather paradox | See time travel |
| Identity of indiscernibles | See Leibniz's Law |
| Improper part | See part |
| Linguistic framework | See metametaphysics |
| Metaphysical realism | Most things exist mind-independently |
| Multiple occupancy thesis | Two people in one body before fission |
Short and cross-reference entries in this section.
Depth: Frankfurt cases, free will map and holes
Frankfurt's nefarious neurosurgeon, full case. Jones decides to kill Smith. Black, a neurosurgeon, has implanted a device that will intervene only if Jones begins to decide otherwise: it will force the decision to kill. As it happens, Jones decides on his own; the device never acts. So the Principle of Alternative Possibilities (PAP) is false: responsibility does not require alternatives. Because freedom is taken as a precondition of responsibility, the case also threatens the ability-to-do-otherwise premise of the Consequence Argument.
Dilemma defence. How does Black know what Jones will decide? If a prior neurological sign determines the decision, the case assumes determinism and begs the question against the incompatibilist. Compatibilists reply that responsibility still looks present even when the case is cleaned of deterministic assumptions; incompatibilists insist the cleaned case either restores alternatives or removes responsibility.
Frankfurt case: device idle, responsibility present, PAP under threat.
Free will positions, reading Table 1. Soft determinism: free will is compatible with determinism, and determinism is true. Hard determinism: incompatible, determinism true, so no free will. Libertarianism: incompatible, determinism false, and we are free. Illusionism: belief in free will is an illusion (perceptual or Freudian). Agent causalism is a thesis about the kind of causation in free action, so its row in the table is largely question marks: it can combine with different answers on determinism. Guidance control (Fischer) is enough for responsibility; regulative control is the genuine ability to do otherwise. Frankfurt cases separate the two.
Holes, three theories and Figure 1. Common sense says holes exist (Swiss cheese has holes; golfers aim for them). Realists about holes posit either hole linings as immaterial particulars or holes as regions of space. Paraphrasers rewrite "there are two holes in this cheese" as a claim about the cheese being doubly perforated, without quantifying over holes. Fictionalists say hole-talk is useful fiction. Casati and Varzi's lining diagram (redrawn as Figure 1) shows why linings are not just the material around the hole: the same material can surround different holes if the topology differs. The entry is a miniature of the book's method: take an ordinary noun, ask what ontology it commits us to, and compare avoidance strategies.
Depth: Humeanism, infinity, identity and immanent realism
Hume and Humeanism. Hume argued that we have no impression of necessary connection between causes and effects, only of constant conjunction; the idea of necessity is projected by the mind. Modern Humeans, led by Lewis, convert this into ontology: there is no natural necessity in the world. Causation, laws and chance must supervene on the mosaic of particular categorical facts (Humean supervenience). Anti-Humeans (Armstrong, dispositional essentialists) reply that the mosaic cannot explain why the regularities hold, and that science discovers real necessities (the half-life of caesium-137 after Chernobyl is not a mere pattern).
Infinity, Aristotle to Cantor. Aristotle distinguished potential infinity (you can always add one more) from actual infinity (a completed infinite collection). Cantor showed that infinite collections come in different sizes: the naturals and the even numbers can be paired one-to-one, yet no pairing of the naturals with the reals is complete (diagonal argument). Metaphysicians use the result when counting possible worlds, properties or classes. The entry's lesson: after Cantor, "infinite" names a family of precise sizes.
From Hume's psychology of causation to Lewis's ontology.
Identity of indiscernibles and Leibniz's Law. Leibniz's Law (indiscernibility of identicals): if a = b, then a and b share all properties. The identity of indiscernibles is the converse: if a and b share all properties, a = b. Max Black's two-sphere world challenges the converse: two qualitatively identical spheres in an otherwise empty space seem possible, yet they are two. Haecceitists allow identity and distinctness that outrun qualitative character. The entry also separates numerical identity (one and the same) from qualitative similarity, a distinction reused in the soldiers' uniforms versus brothers' father analogy for tropes.
Immanent vs Platonic realism. Both are realisms about universals. Immanent (Aristotelian) realists locate universals in their instances: redness exists only where red things do. Platonic realists allow uninstantiated universals (a shade of blue never realised). Immanent realism fits naturalism and the claim that science discovers universals; it struggles with laws about uninstantiated properties and with the possibility of alien properties. Platonic realism is freer with such cases but faces the "third man" and epistemic access worries.
Depth: modal realism, metametaphysics and moral responsibility
Genuine modal realism, benefits and costs. Lewis's possible worlds are concrete universes, spatiotemporally disconnected from ours, including one where everyone dons clown outfits to sing Ricky Martin's Livin' la Vida Loca. Benefits: modality reduces to what exists (p is possible iff p is true at some world); propositions become sets of worlds; properties become sets of possibilia. Costs: the "incredulous stare"; and whether other worlds' inhabitants are relevant to what I could do. Lewis answers with counterpart theory: I could have been a fisherman because a counterpart of mine is a fisherman elsewhere. Ersatz realists make worlds abstract (maximal propositions or states of affairs) and keep modality as primitive.
Metametaphysics: metaphiles and metaphobes. Metaphiles think a priori enquiry yields substantive truths about mind-independent reality. Metaphobes think metaphysical questions are meaningless, unanswerable or only about our conceptual scheme. Hume would commit metaphysics "to the flames"; Kant confines it to experience; logical positivists called it unverifiable; Wittgenstein saw linguistic confusion; Strawson (1959) preferred "descriptive" over "revisionary" metaphysics. From the 1970s traditional metaphysics revived, helped by Quine's reply to Carnap on internal vs external questions. The entry frames the whole book: cost-benefit reasoning (Lewis) is itself a metaphile method under metaphobe challenge.
Three readings of moral responsibility.
Moral responsibility, deep and shallow. On a deep conception, responsibility requires originating causation or incompatibilist ability: you must be the ultimate source of the act. On a shallow (Strawsonian) conception, responsibility is fitness for praise and blame under the reactive attitudes (resentment, gratitude), which can survive determinism. Frankfurt cases push toward the shallow reading if they show responsibility without alternatives. Illusionists claim the deep sense is incoherent and the shallow sense is a useful fiction. The entry links back to Table 1: which row one occupies depends partly on which conception of responsibility one adopts.
Fission against psychological continuity (preview). (1) Person x at t1 is identical to y at t2 iff x is psychologically continuous with y. (2) After fission, both L and R are continuous with the Queen. (3) So Queen = L and Queen = R. (4) By symmetry and transitivity, L = R. (5) But L and R lead separate lives. The no-branching clause blocks (3) at the price of making identity extrinsic: whether L is the Queen depends on whether R exists. Chapter 3 develops the full dialectic; Chapter 4's survival and simple-view entries are opposite responses.
Depth: genuine vs ersatz modal realism, full ledger
GMR's reductive payoffs. Under genuine modal realism, a proposition is the set of worlds where it holds:
Ersatz costs. Ersatz worlds are abstract: sets of propositions, sentences, universals or states of affairs. Because consistency of a set of propositions is already a modal notion (two propositions are consistent iff they could both be true), ersatzism cannot reduce modal facts to non-modal ones. Nor can it identify propositions with sets of worlds without circularity if worlds are themselves sets of propositions. The ersatzer pays with primitive modality what Lewis pays with concrete plenitude.
Genuine versus ersatz theories of what worlds are.
Lewis's answer is counterpart theory: de re possibility is about counterparts, not about me existing in another world (I am world-bound). Counterpart relations are similarity relations, context-sensitive: different counterpart relations may be in play for "I could have been a banker" and "I could have been a poached egg". Critics find the relativity of de re modality a further cost; Lewis treats it as a feature that matches linguistic flexibility.
Haecceitism link. If two worlds can differ only in which individuals play which qualitative roles, haecceitism is true. GMR with counterparts typically rejects haecceitistic differences that outrun qualitative similarity; some ersatz views allow them. The entry thus connects modal realism to the identity-of-indiscernibles debate in the same section.
Depth: metametaphysics comeback and method
From Hume's bonfire to Quine. Hume would commit much of traditional metaphysics "to the flames". Kant asked how synthetic a priori truths are possible and confined metaphysics to the conditions of experience. Logical positivists treated metaphysical claims as unverifiable and meaningless. Wittgenstein saw them as linguistic confusion. Strawson (1959) preferred descriptive metaphysics (mapping our conceptual scheme) to revisionary metaphysics (remaking it to fit reality). From the 1970s traditional metaphysics returned. Quine's earlier rejection of the analytic/synthetic distinction undercut the positivist firewall between logic/maths and metaphysics: if that line blurs, metaphysics cannot be dismissed while maths is kept.
Carnap internal/external and the Hirsch-Sider dispute. Carnap distinguished questions internal to a linguistic framework (answerable) from external questions about the framework itself (pragmatic). Quine collapsed the distinction into a single web of belief. Later, Hirsch treats rival ontological languages as often equally good (quantifier variance); Sider replies that only one meaning of the quantifier carves nature at the logical joints. The dispute is metametaphysics about whether ontology discovers structure or negotiates frameworks.
Metametaphysics: the fall and return of traditional ontology.
What returned is not unchanged. Many contemporary metaphysicians no longer treat the subject as purely a priori. Engagement with physics (relativity and presentism; quantum chances), with linguistics and with experimental philosophy reshapes method. The book's own cost-benefit style (Lewis) is metaphile: it assumes theoretical virtues can track mind-independent structure. Metaphobes read the same ledgers as negotiations about concepts. Readers should notice which stance an entry presupposes when it "weighs costs".
Depth: experimental philosophy, extrinsic properties and free will illusionism
Experimental philosophy. Surveys and vignettes test whether philosophical intuitions are stable across cultures, orders and demographics. The book's entry treats X-phi as a challenge to armchair confidence. Later replications (noted in the literature box) are mixed; the metamethodological question remains whether disagreement among folk undercuts expert intuition or merely maps its borders.
Extrinsic vs intrinsic. An intrinsic property is had in virtue of how a thing is in itself; an extrinsic property involves relations to distinct things (being an uncle; being six metres from a giraffe). Duplicate tests and loneliness tests try to regiment the distinction. Fission's no-branching clause makes identity extrinsic in a way many find absurd: whether L is the Queen depends on whether R exists elsewhere.
Intrinsic/extrinsic cut used in fission and duplication tests.
Illusionism can combine with hard determinism or with the claim that the deep sense of responsibility is incoherent while the shallow sense is a useful social practice. It occupies a row of Table 1 distinct from soft determinism and libertarianism.
Depth: Frankfurt cases beyond the basic script
Responsibility without freedom? Fischer's view allows that Jones is responsible via guidance control even if, lacking alternatives, he did not act freely in the regulative sense. Other compatibilists keep the word "free" for guidance control itself. Incompatibilists who accept the dilemma defence say that once Black's knowledge is made indeterminism-friendly, either a flicker of an alternative remains or responsibility lapses. Buffer cases (agents must pass through an intermediate state before deciding) and blockage cases (all alternatives are closed in advance) are later variants noted in the literature box; the 2011 entry's core is the original idle-device script.
Link to Dennett. Dennett (1984) also denies that the ability to do otherwise is required for responsibility, and argues that everyday control (thermostats; skilled action) is compatible with determinism.
Cost-benefit over hole theories.
Control notions separated by Frankfurt cases.
Counterpart theory for world-bound individuals.
Identity principles in the E to N block.
Metaphobe-friendly versus ambitious projects.
Natural Necessity to Psychological Continuity
Natural necessity to Necessity of origin
Natural necessity is "intra-world" necessity. For Lewis and all Humeans, all necessity is inter-world: truth in all possible worlds, not an intrinsic feature of any world. Non-Humeans find natural necessity in three places: fundamental dispositions (an electron with disposition D must repel another nearby electron), causation (an intrinsic tie between c and e; under indeterminism perhaps "tendency"), and laws (Armstrong's N, contingent necessitation).
Kripke's necessary a posteriori.
Necessity is how things must be: □P, P is true at all worlds; kinds include logical, metaphysical and nomological necessity. Necessity of origin (Kripke): an object's actual origin is essential to it. Stella McCartney could not have had different biological parents; a world where John Lennon and Yoko Ono have a daughter called Stella who designs Madonna's wedding dress is not a world where Stella is theirs, but one where someone else is like her.
Negative facts to Nominalism
Negative facts: there are no unicorns, no air in deep space; I am not on fire. If facts are real constituents of the world, it is odd to think that the world contains, in addition to the positive facts, a further fact of something's absence. Trivially facts; as constituents of reality they are suspect, though some truthmaker theorists need them for negative truths, and others offer totality facts (Armstrong).
Figure 2 (redrawn): two neurons fire and both stimulate a third; also two bricks smashing one window.
Figure 3 (redrawn): d prevents b (reverse arrowhead); dotted b and its link show what would have fired. Read as the two bombers, e is the bridge collapsing.
Figure 4 (redrawn): the same pre-emption in alternative notation: shaded circles fire, the unfilled b does not, and a line ending in a dot marks inhibition.
Libet's timeline as the book reports it.
Nihilism usually means mereological nihilism (nothing ever composes anything) and rarely ontological nihilism (nothing exists), whose literature is "almost without exception very challenging to read". Nominalism may deny abstract objects (Field) or universals (Quine later). Armstrong is a nominalist in the first sense, not the second; Quine in the second, not the first. Worries about universals: redness wholly present in the lamp and the curtains would be a metre away from itself; and there is no criterion of identity: "no entity without identity".
Non-existent objects to Overdetermination
Non-existent objects: Russell (1872-1970) held that abstracta subsist; Meinongians hold that Santa Claus, Sherlock Holmes, the monster under the bed and round squares do not exist in any sense; Priest (2005) defends a modern noneist version.
Ockham's razor (William of Ockham, 1288-1348): "do not believe in something unless there is a really, really good reason to". It is generally, but not universally, held to be a theoretical virtue; the question of why parsimony should indicate truth is taken up in Chapter 4. One-to-one relations: marriage (in British law); not being on the left of (John is left of George, Paul, Ringo and the Eiffel tower). Identity is one-one, which drives the fission problem. Ontological commitment: "Jack went to the cinema" commits to Jack and a cinema; "a leprechaun mugged me" to leprechauns; Becher's phlogiston theory to phlogiston, though it was false. Ontology names one's commitments ("ghosts are no part of my ontology") and the sub-discipline studying categories of being (not the computer-science sense).
Three views of what numbers are.
Overdetermination: two bricks pass through a window at once; each is a redundant cause with equal claim, unlike pre-emption (the washing dries because you hung it out, though someone else would have). The simple counterfactual theory says neither caused the shattering. Some bite the bullet; others count the fusion of the throws as the cause.
Paraphrase to Platonism
Paraphrase: "the average family has 2.7 children" seems to commit us to an average family, "and, more gruesomely, 70 per cent of a child"; anti-realists paraphrase it away. Part: the proper part is ordinary "part" (a table leg); everything is an improper part of itself. A particular occupies one place at a time; tropes are particulars (Socrates' wisdom is distinct from Plato's). Particulars and universals asks what it is to be an object at all (bundles vs substrata), and whether properties are universals, tropes or classes.
Three answers to the problem of temporary intrinsics.
Persistence: existing at more than one time (David Cameron in 1980 and 2010). Personal identity: what makes Old Jack in 2010 the same person as Wee Johnny of the 1952 Olympics; see animalism, bodily and psychological continuity. A phase sortal ("kitten", "child", "pensioner") applies for part of a thing's existence.
How the two anti-realisms keep the banana in the cupboard.
Platonic heaven once meant the place of the Forms that souls passed through; now it means outside space and time. Platonism today means belief in abstract objects, named after Plato's Forms in which things participate.
Possibilia to Primitive
Possibilia are merely possible things: hobbits, the tree you would have had, time-travelling DeLoreans. Quine (1948) mocked them: is the possible fat man in the doorway the same as the possible bald man?
Possibility, ◊P, is truth at some world. Postmodernism: Foucault, Derrida, Baudrillard and Rorty, "suspicious of classical notions of truth, reason, identity and objectivity" (Eagleton 1996). Pragmatism (C. S. Peirce, William James, John Dewey) ties truth to practical difference; if Freudian psychoanalysis works, is its theory of ego, superego and id thereby true?
Possible worlds sort modalities by which worlds count.
Pre-emption: you plant your bomb before I arrive at 10 am, so I go home; your bomb destroys the bridge at midday, though without yours mine would have. Early pre-emption is handled by Lewis's chains of dependence; late pre-emption (Suzy's and Billy's rocks) is harder. Presentism: only present concrete things exist (George Bush and dolphins, but not George Washington or dinosaurs); some presentists accept abstract ersatz times. Motivations: intuition, a reply to McTaggart, compatibility with endurantism (Crisp 2003). Problems: truthmakers for the past, cross-temporal relations and special relativity.
Primary and secondary qualities: Locke (after Galileo and Descartes) says extension, shape and solidity exist independently of perception and resemble our ideas; colours and smells are powers to cause sensations, response-dependent. Berkeley argued the distinction collapses. A primitive is a predicate a theory leaves unanalysed ("__ is located in space"); fewer primitives is a virtue.
PAP to Psychological continuity
Since freedom is a precondition of responsibility, Frankfurt cases against PAP also tell against the ability-to-do-otherwise premise of the Consequence Argument. Most incompatibilists accept a weaker PAP: drunk Sam cannot avoid hitting a pedestrian at that speed, but is derivatively responsible because she could have chosen not to drive.
Probability-raising: Eric's swap counts as a cause.
Probability: the classical view (favourable over total possibilities; a die gives 1 in 6) needs "equally possible" cases, which is circular (spontaneous combustion would be 1 in 2). Rivals: frequentism, propensity and subjective (credence) interpretations, plus Lewis's Principal Principle linking chance and credence.
Problem of luck: if a decision is undetermined, a world identical up to that moment, with the same laws, has the other decision. Kane's (1999) Jane, torn between a business meeting and helping an assault victim in an alley, seems no more in control than a coin toss. Process theories (Salmon; Dowe 2000) make causation the transfer of energy or a conserved quantity: billiard balls, switches and lights. They handle pre-emption (only Suzy's rock transfers energy to the bottle) but struggle with causation by absence.
Properties are characteristics (redness, solidity); austere nominalists deny redness exists, while universals, tropes and class or resemblance nominalism give rival realist accounts. Propositions are truthbearers that would exist even with no sentences (the world would still be round), "sentences written down in God's Big Book of Reality".
Reading the entries together: modality, causation and the self
Modality from three directions. The section approaches modality through natural necessity (necessity inside a world), necessary a posteriori truths and necessity of origin (Kripke's metaphysical necessity), and possible worlds (the semantic machinery). The entries teach the reader to separate three questions often run together: what is necessary (the extension of necessity), how we know it (a priori or a posteriori), and what makes it so (worlds, essences or natural necessity). Kripke's examples show that the first two can come apart; Lewis's reduction and Armstrong's necessitation offer opposite answers to the third.
Causation as a battleground of counterexamples. Overdetermination, pre-emption, the neuron diagrams, probabilistic theories and process theories form a dialectic in which every theory is tested against a small set of cases: two bricks, two bombers, Eva and Eric's coin, billiard balls, absences. The counterfactual theory struggles with redundancy; probabilistic theories with probability-lowering causes (a golf ball that hits a branch but still goes in); process theories with absences. The neuron diagrams make the shape of each case visible, and the reader can see why no single analysis has won: each handles some diagrams and fails others.
The self from three directions. Personal identity, phase sortal, PAP and psychological continuity join the earlier entries on animalism and bodily continuity. The psychological continuity entry is the most worked-out in the section: it shows how Locke's memory theory was repaired against Butler and Reid (continuity for transitivity, q-memory for circularity) and why the repaired theory still faces fission. Read with survival (Chapter 4), it shows how far philosophers were willing to go: Parfit concluded that identity is not what matters.
Three analyses tested against the neuron diagram cases.
Worked comparisons from the section
Platonism, formalism and fictionalism about number. The numbers entry is a model of how the book handles an ontological dispute. It begins from a truth we all accept ("there are at least five prime numbers bigger than 2"), notes that it appears to assert the existence of numbers and is knowable a priori, and asks what numbers could be to make both true. Platonism explains the necessity and a priority of arithmetic but leaves access and application mysterious (the thirteen oranges). Formalism explains knowledge by derivation from axioms but seems to make arithmetic about nothing, which sits badly with its use in science. Fictionalism keeps the platonist's semantics and denies the ontology. The entry's contrast with "there are at least five cities bigger than Birmingham" makes the epistemic puzzle vivid.
Perdurance against its rivals. The perdurantism entry gives three motivations; the one it develops in detail is temporary intrinsics. The Hulk Hogan case is useful because it isolates the problem: change requires one thing to have incompatible properties, and the perdurantist, relationist and presentist each remove a different element. The perdurantist multiplies things (temporal parts), the relationist multiplies relations (heights become relations to times), and the presentist removes the past (Hogan is not 2′ tall at all). Lewis argued that relationism makes shape not intrinsic and presentism denies the past, so perdurance wins; endurantists reply that temporal parts are themselves an extravagant cost.
Idealism without God. The phenomenalism and idealism entry explains why the views were attractive despite their oddity: they promise to defeat scepticism by closing the gap between experience and world, since if the banana just is its appearances, our experience cannot be systematically misleading about it. The price is the counterfactual analysis of unobserved objects, whose truthmakers are unclear: what makes it true that I would have a tree-experience if I opened my eyes, if not the tree?
Memory connections link each stage to the next; continuity is their ancestral.
Key arguments reconstructed
From rigid names to necessary identity. (1) "Hesperus" and "Phosphorus" are rigid designators: each refers to the same object in every world in which it exists. (2) In fact both refer to Venus. (3) So in every world where Venus exists, "Hesperus" and "Phosphorus" both refer to Venus, and "Hesperus is Phosphorus" is true. (4) So it is necessarily true. (5) Yet ancient astronomers could not know it from meanings alone; it took observation. So it is necessary and a posteriori. The argument relies on Kripke's rejection of descriptivism: if "Hesperus" meant "the evening star", the identity would be contingent, since the evening star might have been Mars.
Libet's inference and where it fails. (1) The readiness potential begins about 350 ms before subjects report the urge to flex. (2) The readiness potential is the initiation of the action. (3) So the action is initiated before the conscious will. (4) So conscious will does not initiate action, and free will is at most a veto. The entry's objections target (2) and the step to (4): the reported awareness may be the forming of an intention that itself causes the flex, and the inference assumes that a free act must have no physical precursors, an assumption most compatibilists reject.
The luck argument. (1) Suppose Jane's decision to help the victim is undetermined. (2) Then a world exactly like ours up to the decision, with the same laws, contains Jane deciding to go to the meeting. (3) Nothing about Jane, her reasons, character or effort, differs between the worlds before the decision. (4) So nothing about Jane explains why she decided one way rather than the other: the difference is a matter of luck. (5) We are not responsible for what is a matter of luck. Libertarian replies include Kane's dual efforts (Jane is trying to make both choices, so either outcome is her doing) and agent causation (Jane herself settles the matter).
The argument from temporary intrinsics. (1) Hogan was 2′ tall and is 6′7″ tall. (2) Heights are intrinsic properties, not relations. (3) Nothing can have incompatible intrinsic properties. (4) If Hogan is wholly present at both times, one and the same thing has both heights. So either heights are relations to times (deny 2), only the present exists (deny that he has the past height), or Hogan has distinct temporal parts that bear the heights (deny 4). Lewis (1986) took the third option and called it the only one that keeps both intrinsic shape and the reality of the past.
How the entries have aged
Still current. Necessary a posteriori truths, necessity of origin, possible worlds, pre-emption, overdetermination, PAP, probability and psychological continuity remain accurate introductions. The neuron diagrams are still the lingua franca of the causation literature, though they now sit alongside structural equation models, which represent the same cases with variables and equations and handle pre-emption by asking what happens when some variables are held fixed.
Shifted emphasis. The neuroscience and free will entry would be revised most. Later work reinterpreted the readiness potential as the build-up of spontaneous neural fluctuations (Schurger and colleagues, 2012) and found it absent or much weaker for deliberate, meaningful choices (Maoz and colleagues, 2019); many neuroscientists now regard Libet-style data as irrelevant to free will in the sense philosophers discuss. The presentism entry would add the large literature on truthmakers for the past and on presentism in relativistic physics. Platonism about numbers would mention neo-Fregean abstractionism and mathematical structuralism, which the 2020 survey found was the single most popular view of foundations (21.2%).
Dated details. The presentism entry's "George Bush" (the president at the time of writing) dates the entry to before 2009, as the authors acknowledge with "at the time of me writing this".
What the section asks of the reader. The neuron diagrams repay slow reading: before looking at Figure 3, try to say in words what happens in the two-bomber case, then check whether the diagram captures it. The same habit helps with the modal entries. The section's lesson is that pictures and formal tools are aids to judgement, not replacements for it.
The section's further reading
Pointers include Kripke (1980) for the necessary a posteriori and origin; Mele (2008) for Libet; Armstrong (1989) and Loux (2006) for nominalism; Dowe (2000) for process theories; Kane (1999) for luck; Shoemaker (1970) and Parfit (1971) for psychological continuity; and Lewis's papers for neuron diagrams and pre-emption, now collected with replies in Collins, Hall and Paul (eds), Causation and Counterfactuals (2004).
| Term | Meaning or cross-reference |
|---|---|
| Natural law | Law of nature |
| Natural property | Properties that carve nature at its joints |
| Ostrich nominalism | Pejorative for austere nominalism |
| Overlap | Sharing a part |
| Participation | See instantiation |
| Point particle | See simple |
| Proper class, proper part | See set; see part |
| Only x and y principle | See fission |
Short and cross-reference entries in this section.
Depth: natural necessity, necessary a posteriori and neuroscience of free will
Natural necessity vs logical necessity. Logical necessities hold in every possible world ("2+2=4"). Natural necessities hold in every world with the same laws: a massive body near Earth accelerates at about 9.8 m/s², but there are worlds with different laws. Humeans deny genuine natural necessity; anti-Humeans (Armstrong's N, dispositional essentialists) affirm it. The distinction matters for counterfactuals: "if the match had been struck, it would have lit" seems to track natural, not merely logical, connections.
Necessary a posteriori (Kripke). Some truths are necessary yet knowable only empirically. "Water is H₂O" and "gold has atomic number 79" are the stock examples: once the actual nature is discovered, we judge that nothing that is not H₂O could be water. The result breaks the traditional alignment of necessary/a priori/analytic. It also feeds essentialism about natural kinds and the claim that science discovers essences. Contingent a priori truths (the metre bar in Paris, on some readings) complete the reopening of the square.
Kinds of necessity the modality entries separate.
Neuroscience of free will (Libet). Libet's experiments timed readiness potentials against the subject's reported urge to flex: the brain activity began roughly 350 ms before the conscious wish. Some took this to show that conscious will is epiphenomenal. Critics note that the task (spontaneous flexing) is unlike ordinary deliberation; that the clock-report method is noisy; and that a veto or "free won't" might still be conscious. The entry treats the science as evidence that must be interpreted through the same conceptual distinctions (guidance vs regulative control, deep vs shallow responsibility) already mapped in Chapters 1 and 2, not as a free-standing refutation of free will.
Depth: possible worlds, presentism, perdurantism and pre-emption
Possible worlds as tool and as ontology. Possible worlds are used to define necessity (true at all worlds), possibility (true at some), counterfactuals (true at the closest worlds where the antecedent holds) and propositions (sets of worlds). The ontological question is separate: are worlds concrete (Lewis), abstract ersatz (Plantinga, Stalnaker), or mere useful fictions? The entry's point is that almost every modern theory of modality, laws and causation uses world-talk; the dispute is what the talk commits us to.
Presentism: motivations and problems. Only present concrete things exist: George Bush and dolphins, not George Washington or dinosaurs. Some presentists accept abstract ersatz times. Motivations: intuition, a reply to McTaggart's paradox, compatibility with endurantism (Crisp 2003). Problems: truthmakers for past-tense truths; cross-temporal relations (I admire Socrates); and special relativity, which seems to deny a unique global present. Growing-block theorists keep the past and lose the future; eternalists keep both.
Which times exist, before A/B theory is added.
Epistemic objection to the growing block. (1) On the growing block, past times exist, including Genghis Khan in AD 1220 thinking "this is the present". (2) His belief is false: the present has moved on. (3) My epistemic situation is just like his. (4) So I have no reason to think my time is present rather than past. The growing-block theorist must say what makes the present detectable from inside, or accept the sceptical result. Presentism (no past thinkers) and B-theory (no special present) escape the objection.
Perdurantism vs endurantism. Perdurantists say objects persist by having temporal parts: the whole four-dimensional worm is the object, and what exists at t is a stage. Endurantists say objects are wholly present at each time they exist. Perdurantism pairs naturally with eternalism and offers a clean account of change (different stages have different properties). Endurantism pairs with presentism for some authors, though others combine endurantism with eternalism. The statue and clay, and fission cases, are pressed on both sides.
Pre-emption, early and late. You plant your bomb before I arrive at 10 am, so I go home; your bomb destroys the bridge at midday, though without yours mine would have (early pre-emption). Lewis's chains of counterfactual dependence handle early cases. Late pre-emption (Suzy's rock shatters the bottle; Billy's would have a split-second later) is harder: the effect does not counterfactually depend on Suzy's throw, yet Suzy caused it. Process and conserved-quantity theories hope to do better; "trumping" (Schaffer's sergeant and major) is pre-emption with no intervening chain.
Depth: PAP, probability, properties and psychological continuity
Frankfurt cases attack PAP. Most incompatibilists accept a weaker PAP: drunk Sam cannot avoid hitting a pedestrian at that speed, but is derivatively responsible because she could have chosen not to drive. The distinction between direct and derivative responsibility lets incompatibilists absorb some "no alternatives now" cases without surrendering the link between freedom and alternatives for the key choices.
Probability: chance, credence and frequency. The entry separates subjective probability (credence), frequentist probability (limiting relative frequency) and chance (single-case propensity or Lewisian chance). Chance links to laws under Humean best-system accounts and to the problem of luck for libertarians: if nothing about Jane settles which way an undetermined decision goes, the outcome looks like chance rather than control. Probabilistic causation uses probability-raising as a mark of causes, with well-known counterexamples from common causes and pre-emption.
Psychological continuity built from overlapping links.
Properties: the menu. Realists about universals, trope theorists, class nominalists, resemblance nominalists and austere nominalists all answer "what are properties?" differently. Sparse (natural) properties carve joints; abundant properties correspond to any predicate. Intrinsic vs extrinsic, categorical vs dispositional, and quidditistic vs structural conceptions cut across the menu. Chapter 3's property entry is the hub that sends the reader back to Chapter 1's bundle theory and forward to Chapter 4's tropes and universals.
Psychological continuity, full structure. The Lockean tradition analyses personal identity in terms of overlapping chains of memory, intention and character. Psychological continuity is the ancestral of strong psychological connectedness: enough links of memory and character, possibly overlapping. Fission is the central objection (see Chapter 2's argument map). Replies: the no-branching clause (identity requires uniqueness); the claim that identity is not what matters (Parfit); the simple view that identity is primitive; and animalism's biological substitute. Quasi-memory (Chapter 4) tries to remove the circularity Butler alleged in memory-based accounts.
Depth: possible worlds semantics and kinds of modality
Dropping the modal word. "The Queen could play the banjo" becomes "there is a possible world at which the Queen plays the banjo". "Necessarily 2+2=4" becomes "at all possible worlds, 2+2=4". "It is impossible for me to fly" becomes "at no possible world do I fly". The paraphrase is useful even for philosophers who deny that worlds are concrete; the ontological question is separate from the semantic convenience.
Nested kinds of possibility. Start with logically possible worlds. Restrict to those with our physical laws: physical (nomological) possibility. Restrict to those compatible with what an agent knows: epistemic possibility. Relative to what I know, it may be epistemically necessary that the Queen does not play the banjo even if it is metaphysically possible. Confusing these grades generates fallacies: physical impossibility is not logical impossibility; epistemic openness is not metaphysical contingency.
Possible-worlds semantics before ontology is settled.
Actuality. "The actual world" is the world we inhabit. For Lewis, actuality is indexical: every world is actual from the standpoint of its inhabitants. For many ersatzers, actuality is a primitive property of one abstract world (or of the maximal consistent set that is realised). The indexical reading fits GMR; the primitive reading fits actualist ersatzism. Both must explain the modal force of "actually p" in embedded contexts.
Depth: perdurantism's three motivations and fission's clauses
Temporary intrinsics. Hulk Hogan was 2 ft tall as a child and is now 6 ft 7 in. Those heights are incompatible if had simpliciter. Perdurantism: different temporal parts have different heights. Relational endurantism: Hogan bears being 6 ft 7 in-at-t rather than being 6 ft 7 in simpliciter. Presentism: only present heights are had, so past shortness is not a present contradiction. Lewis prefers temporal parts because he wants intrinsic properties had simpliciter.
The argument from spacetime analogy and from possibility. Just as objects have spatial parts, they may have temporal parts; denying the latter while accepting the former seems arbitrary once eternalism is granted. Lewis also imagines a world of short-lived person-stages intrinsically matching each instant of Napoleon's life; if that world is possible, and ours is qualitatively like it moment by moment, then either our world also has stages or a brute difference in persistence structure must be accepted. Perdurantists take the first option.
Three formal replies inside reductive theories.
Fission: no-branching vs best candidate vs multiple occupancy. No-branching: identity requires a unique continuer; at fission the original ceases. Best candidate: the most continuous continuer inherits identity if one leads; if tied, perhaps neither. Multiple occupancy (often with perdurantism): two persons already shared the pre-fission stages; fission reveals them. Parfit's further move: the identity question can be empty; survival (connectedness) is what matters. Animalists and simple-view theorists reject the shared reductive premise that identity must analyse into continuity facts.
Depth: presentism's logical and relativistic pressures
Truthmakers and singular terms. "
Relativity. Special relativity makes simultaneity frame-relative. If what is present is what is simultaneous with me, and existence is presence, then what exists threatens to become frame-relative: travellers with different velocities would not share a ontology. Presentists explore foliations, privileged frames, or denying that relativity's geometry settles the metaphysics of tense. Eternalists treat the relativistic block as home territory.
Logic and relativity as costs for present-only ontology.
Motivations restated. Beyond intuition, presentists claim a clean reply to McTaggart (no A-series contradiction if only the present is real), a fit with endurantism (Crisp 2003), and a solution to temporary intrinsics (only present properties are had). Opponents argue that the logical and relativistic costs outweigh these gains, and that growing-block or eternalist A-theories can keep tensed change without presentist ontology.
Depth: numbers, overdetermination and the problem of luck
Numbers. Platonism: numbers exist as abstract objects. Formalism: mathematics is rule-governed symbol play without ontology. Fictionalism: mathematical talk is useful fiction. The indispensability argument (Quine-Putnam) pushes from scientific practice toward platonism; nominalist paraphrases and fictionalist attitudes push back. The entry sits beside possible worlds because both are abstract-entity disputes with heavy theoretical utility.
Overdetermination. Two sufficient causes of one effect (two bullets each lethal) threaten theories that ban systematic overdetermination (some physicalist exclusion arguments). Compatibilists about mental and physical causation sometimes accept harmless overdetermination; others deny the mental is distinct enough to overdetermine.
Why indeterminism may not deliver libertarian freedom.
Problem of luck for libertarians. If Jane's undetermined decision could have gone either way given the same past and laws, nothing about Jane settles the outcome. The decision looks lucky rather than controlled. Agent-causalists try to make the agent the settler; event-causal libertarians add efforts of will or contrastive explanations. Compatibilists treat the luck problem as showing that indeterminism does not deliver the control incompatibilists wanted.
Depth: necessity of origin and Ockham's razor
Necessity of origin (Kripke). A person originating from different parents, or a table from different wood, would not be that person or table. The thesis supports essentialism about origins and feeds modal arguments about what is possible for an individual. Critics allow qualitatively matching but numerically distinct origin stories, or treat origin claims as contingent on referential practices.
Ockham's razor. Do not multiply entities beyond necessity. In this book the razor appears whenever realist and anti-realist packages are scored: holes, universals, possible worlds, temporal parts. "Beyond necessity" is the disputed phrase: Lewis multiplies concrete worlds to reduce modal ideology; nominalists multiply ideology (primitive resemblance) to reduce ontology. Cost-benefit reasoning is the razor made comparative rather than absolute.
Nested grades of possibility.
Neuroscience timeline and rival interpretations.
Why chains help early cases more than late ones.
Necessity of origin in outline.
Why the growing block faces a sceptical present.
Quasi-Memory to Zeno
Quasi-memory to Rationalism
Quasi-memory and quasi-states. Butler (1736) charged that memory presupposes identity: genuine memory requires that you are the person who had the experience; likewise I can only intend my own actions (I cannot intend that you study for exams, only desire it). The answer (Shoemaker 1970, Parfit 1971) is q-memory: an apparent memory of an experience someone had, caused in the right way by that experience, without requiring that it was mine. Q-intending and q-states generally follow. Ordinary memory is then q-memory plus identity, and the analysis is no longer circular.
Quasi-memory answers Butler's circularity charge.
Quidditism (no relation to Harry Potter's sport) says a property's identity is brute, independent of the dispositions it confers: there is a world where mass behaves as charge does here, and vice versa. It is the analogue of haecceitism for properties and seems forced by Humean supervenience, since fundamental properties are categorical. Bird (2007) takes that as reason to reject categoricalism; Lewis (2008, "Ramseyan humility") bites the bullet, accepting that we cannot know which quiddities play which roles.
W. V. Quine (1908-2000), from Akron, Ohio, turned to philosophy on a fellow poker player's suggestion, wrote a Harvard PhD on Principia Mathematica, served three years in the Navy in the Second World War, and gave metaphysics its criterion of ontological commitment, its attack on the analytic/synthetic distinction and its naturalism.
The Ramsey-Lewis criteria for laws.
Rationalism (Descartes 1596-1650, Spinoza 1632-77, Leibniz 1646-1716) is not the Rationalist Association's creed of "reason and evidence"; it holds that some truths about reality (every event has a cause, God exists) are knowable a priori, and that some knowledge is innate.
Realism and anti-realism to Relata
Realism and anti-realism. The terms are slippery. On one usage realism about X adds irreducibility or mind-independence: a regularity theorist is then an anti-realist about causation, and a response-dependence theorist about colour. The book's chosen usage is simply that realism about X says X exists: so the reductionist is a realist about causation (plenty of "c caused e" truths) but an anti-realist about fundamental causal relations, and the response-dependence theorist agrees London buses are red.
Two scales of realism dispute.
Reductionism: Ks exist but are wholly accounted for by a description that does not mention Ks. In philosophy of mind "reductive" means type identity, contrasted with non-reductive supervenience physicalism; in the book's metaphysical sense both count as reductionist. Redundant causation covers pre-emption (your flatmate would have hung out the washing) and overdetermination.
Regularity theory of causation: c causes e iff (i) c is contiguous with e, (ii) c precedes e, (iii) all events like c are followed by events like e, a close paraphrase of Hume's "first definition" (1739-40: 172), though it is unclear Hume meant to define causation. Mackie's INUS refinement and the problems of common causes and accidental regularities follow. The regularity theory of laws says it is a law that all Fs are Gs iff all Fs are Gs: counterexamples include "all the carpets in my house are beige" and the vacuous "all unicorns are made of cheese". Relata are what a relation relates: Jack and Jim, 20 metres apart; Jack and Jill, if he hates her. The notion matters most in the causation debate, where the choice of relata (events, facts, properties) shapes what a theory can say about absences and about type-level claims.
Relativism to Response-dependence
Relativism makes truth relative to something about humans, and since "p" is true iff p, it relativises what is the case ("God exists" relative to a society's beliefs makes God's existence mind-dependent). Forms vary in breadth (only moral or aesthetic truth, not the speed of light or the Earth's circumference) and strength: thoroughgoing subjectivism ("true for me") dissolves disagreement but also seems self-refuting. The entry asks whether relativism about meaning is the better reading.
Special relativity in the book's three steps.
Resemblance nominalism: an object is red in virtue of resembling other red things, taking resemblance as primitive; slightly less parsimonious but more intuitive than class nominalism. Problems include imperfect community and companionship (Goodman), and whether resemblance is itself a universal (Russell's regress). Response-dependence: being red is partly constituted by looking red to us; one is then a realist about red things but an anti-realist about mind-independent colours.
Scepticism to Supervenience
Scepticism is a philosophical position, not doubt (as about whether Labour will win): some class of claims, usually about the external world, cannot be known, or there is no reason to believe them. Descartes' (1641) evil demon feeds a disembodied mind false experiences; the brain in a vat is its modern form. Since our evidence comes entirely from the senses, we seem to have no reason to reject the hypothesis; responses include Moore, Putnam's semantic externalism, inference to the best explanation and contextualism. Restricted forms include inductive scepticism (will the sun rise tomorrow?) and scepticism about other minds. The entry stresses how worrying the global form is: if it is true, it is no more rational to believe in trees and chairs than in fairies and little green men.
Set: a class; some set theories forbid certain classes from being members to avoid paradox (Russell's); these are proper classes. A simple has no proper parts, perhaps quarks or superstrings; whether simples can be extended is disputed. The simple view of personal identity rejects all reductions: identity is primitive, most naturally if persons are souls or Cartesian egos, though some materialists hold it too. A singleton has one member: {Brad Pitt}, {{Brad Pitt}}.
Singularism about causation has several readings: causation without a covering law, causation not reducible to laws, or not reducible to relations between types. Soft determinism: compatibilism plus determinism. Sortals take numerical modifiers ("two cats", not "two waters"). Sparse properties (Lewis) are the natural few; abundant properties the rest.
Answers to van Inwagen's Special Composition Question.
States of affairs: Jack Smart said believing in both facts and things is as dangerous as mixing your drinks. Armstrong builds his ontology from states of affairs (a's being F). Subsistence: being without existence (Meinong, Russell). Substance: in philosophy an irreducible particular (seven souls, not seven magnesiums). A substratum bears properties. A sum is a fusion. A supertask completes infinitely many tasks in finite time (counting infinite sand grains in a lifetime).
Supervenience is necessary co-variation: no A-difference without a B-difference. If Jack thinks about dinner and Joe multiplies 6 by 23, some physical difference must account for it. Varieties differ by strength of necessity (logical, metaphysical, nomological) and scope (local or global); supervenience alone does not explain why the co-variation holds, which is one motive for grounding.
Survival to Theoretical virtues
Temporal parts: a football match has two halves; for perdurantists I have a part that exists for my first year, growing from newborn to one-year-old size and then ceasing. Synchronic identity is identity at a time.
Tensed and tenseless theories of time.
Thought experiment to Truthmakers
Thought experiments: Locke's cobbler and prince swap "consciousness" in their sleep. Who wakes in the cobbler's bed, missing his servants? Who is to blame for the prince's murder of a courtier? Locke says the person in the cobbler's body, so bodily continuity is not required for personal identity. Williams (1970) showed how redescribing the case can reverse intuitions; experimental philosophy studies such effects.
Time became a major sub-discipline after McTaggart, though it goes back to Zeno. Time travel raises the grandfather paradox: kill your grandfather before he meets your grandmother and you are never born. Lewis (1976) says you can travel but will fail: the gun misfires, or you slip on a banana peel. Time travel challenges presentism (where would you go?) and, some say, endurantism (Sider 2001), others perdurantism (Gilmore 2007). Token-level causation holds between particular events ("John's smoking caused his cancer"). A transitive relation: taller than; at least six feet away from is not.
Lewis on the grandfather paradox.
Trope theory is the nominalism closest to realism: properties are constituents of objects but unrepeatable. Our books share "the same property" as two soldiers share "the same uniform", not as brothers share "the same father". Similarity becomes resemblance between tropes, perhaps primitive. Trumping is pre-emption with no intervening chain (Schaffer's sergeant and major).
Three theories of truth surveyed.
Truthbearers: usually propositions, so that "snow is white" and "Schnee ist weiss" express the same one. Truth conditions: "snow is white" is true iff snow is white. Truthmakers: every truth has an entity whose existence necessitates it. Barack Obama makes
Type and tokens to Zeno
Type and tokens: "war is war" contains two word types and three tokens. Type-level causation ("smoking causes cancer") concerns properties or populations.
McTaggart's argument (1908) for the unreality of time: time requires A-facts; but every event is past, present and future (Hastings was future in 1065, present in 1066, now past), which is contradictory. The obvious reply, "at different times", McTaggart says yields a vicious regress. Others drop A-facts (tenseless theorists) or non-present times (presentists).
McTaggart's argument for the unreality of time.
Zeno: not Zeno of Citium (344-262 BC), founder of Stoicism, but Zeno of Elea (born about 490 BC), defending Parmenides' monism. In the dichotomy, a runner covering 10 yards must first run 5, then 2½, then 1¼, and so on infinitely, so "nothing ever moves". The others are Achilles and the tortoise, the arrow, and the stadium. Modern limit theory shows an infinite series of distances can have a finite sum, though whether that answers Zeno, and whether supertasks are possible, is still argued.
Reading the entries together: realism, time and method
Local and global realism. The section distinguishes two kinds of dispute that share a name. Local disputes ask whether some particular kind of entity exists (numbers, universals, holes, causal relations); the book's chosen usage makes realism about X simply the claim that X exists. Global disputes ask whether there is a mind-independent reality at all, and here the anti-realist positions (idealism, pragmatism, relativism, Putnam's internal realism) are of a different order. The truth entry shows how the two connect: correspondence theories are often thought to go with global realism, coherence and pragmatist theories with anti-realism, while deflationism claims neutrality. Scepticism sits at the junction: the global realist must admit that reality could outrun all our evidence, which is exactly the sceptic's opening.
Time, completed. Relativity, tensed and tenseless theories, temporal parts, time travel, unreality of time and Zeno complete the time cluster. The relativity entry is the section's most technical, and its point is philosophical: if simultaneity is relative, there is no unique present, which seems to tell against presentism and the growing block, and may favour perdurantism. Defenders of the present reply by privileging a frame or by appeal to quantum theory. McTaggart's argument shows that the A-theory has its own internal difficulty; Zeno shows how old the puzzles are.
Method, completed. Theoretical virtues, thought experiments, supervenience and reductionism describe the toolkit used throughout. The book's honest admission is that which features count as virtues is "highly contentious", and that thought experiments can mislead as easily as they illuminate. Metaphysical disputes are typically settled, when they are settled at all, by weighing virtues against vices across a whole theory, not by a single decisive argument.
Universals, tropes and austere nominalism on resemblance.
Worked comparisons from the section
Ramsey-Lewis against DTA. The two main theories of laws in the book disagree about what grounds the difference between laws and accidents. On the best-system view it is a matter of the systematic role a regularity plays in the simplest and strongest summary of the world. On DTA it is a matter of an extra ingredient, necessitation between universals. The Ramsey-Lewis entry clarifies a common misunderstanding: the best system is not current science (Ptolemaic astronomy did not make geocentrism a law in ancient Greece) but the ideal theory scientists would reach with unlimited time and resources.
Tropes, universals and resemblance nominalism. Read together, trope theory, resemblance nominalism and universals complete the particulars-and-universals cluster begun with austere and class nominalism. All agree that the Coke can and the telephone box resemble each other; they disagree about what the resemblance consists in. The universals realist has one entity in two places; the trope theorist has two exactly resembling tropes; the resemblance nominalist takes resemblance between the whole objects as primitive. The soldiers' uniforms and the brothers' father make the contrast between qualitative and numerical sameness concrete, linking back to the identity entry of Chapter 2.
Survival and the simple view. The survival and simple view entries are opposite responses to fission. The simple view says personal identity is primitive, so there is always a fact about which (if either) of the fission products is me, even if we cannot know it. Parfit says the question of identity can be empty, and that what matters, psychological connectedness, is present in both. The cobbler and prince of the thought experiment entry show where the debate began: Locke's confidence that consciousness, not body, carries the person.
Key arguments reconstructed
McTaggart's argument, step by step. (1) Time requires change, and change requires A-facts (events moving from future to present to past); B-facts never change. (2) Being past, present and future are incompatible properties. (3) Every event has all three (Hastings was future in 1065, present in 1066, and is past). (4) So every event has incompatible properties, a contradiction. (5) So there are no A-facts, and so no time. The obvious reply rephrases (3): every event is past, was present and was future. McTaggart says the tensed copulas themselves invoke A-facts at a higher level, generating a regress; whether that regress is vicious is the crux.
Parfit's argument that identity does not matter. (1) If Freddy's operation produces one New-Freddy, Freddy survives and has what matters. (2) If it produces two, Freddy (on the no-branching view) does not survive. (3) But the relation between Freddy and each New-Freddy is intrinsically the same in both cases; the only difference is whether another, similar person also exists. (4) Whether I get what matters cannot depend on such an extrinsic fact. (5) So what matters in the two-person case is present too, and it is not identity: it is psychological connectedness and continuity. Parfit drew practical conclusions, about death and about concern for one's future self, that remain controversial.
The sceptical argument from the evil demon. (1) If I know there is a chair before me, I know I am not deceived by a demon (or am not a brain in a vat). (2) I do not know I am not so deceived, since my evidence would be the same either way. (3) So I do not know there is a chair. The argument is valid, so responses must reject a premise. Moore reversed it: I know there is a hand here, so I know I am not deceived. Contextualists say "know" shifts standards with context; externalists deny that knowledge requires evidence that rules out the demon.
Relativity against presentism. (1) Presentism says that exactly the present things exist. (2) If special relativity is true, which events are simultaneous with a given event depends on the frame of reference. (3) So "the present" is frame-relative, and there is no absolute fact about which things are present. (4) But what exists cannot be frame-relative. So presentism is false, or there is a privileged frame that physics does not detect. Presentists have taken both routes: some accept a hidden preferred foliation (suggested by some interpretations of quantum theory), others relativise existence.
How the entries have aged
Still current. Ramsey-Lewis, realism and anti-realism, relativity, scepticism, the Special Composition Question, supervenience, tensed and tenseless theories, time travel, tropes, truth, universals and Zeno are accurate introductions that reflect the state of their debates. The relativity entry's account of simultaneity is sound even where its account of the 1887 experiment by Michelson and Morley is overstated.
Shifted emphasis. Supervenience would now be presented alongside grounding, which many metaphysicians think captures "nothing over and above" better, since supervenience is symmetric in some cases and cannot express explanation. Truthmakers would discuss "truthmaker maximalism" and its critics more fully. The Special Composition Question would give more space to priority monism (Schaffer), mentioned only briefly, and to the argument from vagueness for universalism, which has become a standard topic. The survival entry would note Parfit's later work, On What Matters (2011, 2017), which turned to ethics.
Dated details. Quine (died 2000) and Lewis (died 2001) were already dead in 2011, as the entries say. The truthmakers entry's example "
What the section asks of the reader. The final chapter is the most varied, and its entries are best read in clusters: realism (realism and anti-realism, global realism, relativism, response-dependence, scepticism, truth, truthmakers); time (relativity, temporal parts, tensed and tenseless theories, time travel, unreality of time, Zeno); properties (quidditism, resemblance nominalism, tropes, universals); and method (supervenience, theoretical virtues, thought experiments). Read this way, the chapter completes each of the book's six topics and returns, at the end, to the question of method with which the Introduction began.
The section's further reading
Pointers include Ramsey (1928) and Lewis (1973) for laws; Bird (2007) and Lewis (2008) on quidditism; Sklar (1992, Chapter 2) and Hawley (2009) for relativity; Dummett (1960) and Lowe (1987) on McTaggart; Lewis (1976) and Dainton (2001, Chapter 8) on time travel; van Inwagen (1990) and Schaffer (2007) on composition; and Parfit (1971, 1984) on survival.
Zeno's dichotomy.
| Term | Meaning or cross-reference |
|---|---|
| Psychologism | See concept |
| Reactive attitudes | See moral responsibility |
| Regulative control | See guidance control |
| Ship of Theseus | See identity over time |
| Singular causal sentences | Token-level causal claims |
| Source incompatibilism | See incompatibilism |
| Split brain case | See fission |
| Temporary intrinsics | See perdurantism |
| Transitive closure | See ancestral relation |
| Unrestricted composition | Universalism |
Short and cross-reference entries in this section.
Depth: quasi-memory, quidditism, Ramsey-Lewis and realism
Quasi-memory and Butler's charge. Butler (1736) charged that memory presupposes identity: genuine memory requires that you are the person who had the experience; likewise I can only intend my own actions. Shoemaker (1970) and Parfit (1971) answer with q-memory: an apparent memory of an experience someone had, caused in the right way by that experience, without requiring that it was mine. Q-intending and q-states generally follow. Ordinary memory is then q-memory plus identity, and the analysis is no longer circular. The device lets psychological-continuity theories state their claim without building identity into the analysans.
Quidditism. If properties are quiddities, two worlds can differ only by a systematic swap of the roles of mass and charge: the same causal structure, different identity of the properties playing the roles. Structuralists and dispositional essentialists deny that such worlds are distinct: a property's identity is exhausted by its causal or nomic role. The debate reconnects to Humean supervenience (Chapter 2): Humeans typically accept quidditistic differences; anti-Humeans often reject them.
Shoemaker-Parfit device against Butler's circularity charge.
Ramsey-Lewis best-system theory of laws. Laws are the axioms or theorems of the deductive systems that best balance simplicity and strength in systematising the mosaic. The theory is Humean: laws supervene on particular facts. It improves on naive regularity by excluding accidental generalisations (all the gold spheres are less than a mile across) when they do not earn their keep in the best system. Problems: the standards of simplicity and strength seem mind-dependent; and the theory struggles with probabilistic laws and with the intuition that laws govern.
Realism and anti-realism, local and global. Local realism about Fs says Fs exist mind-independently. Global realism says mind-independent reality has a determinate structure. Anti-realist options include idealism, phenomenalism, verificationism, and relativism. The brains-in-a-vat and evil-demon scenarios test whether our evidence distinguishes realism from sceptical alternatives. Putnam's semantic externalism argues that a brain in a vat could not refer to vats, so cannot truly say "I am a brain in a vat"; critics reply that the sceptical hypothesis can be redescribed.
Depth: relativism, scepticism, survival and the simple view
Relativism's breadth and strength. Relativism makes truth relative to something about humans. Since "p" is true iff p, it relativises what is the case: "God exists" relative to a society's beliefs makes God's existence mind-dependent. Forms vary in breadth (only moral or aesthetic truth, not the speed of light) and strength: thoroughgoing subjectivism ("true for me") dissolves disagreement but also seems self-refuting. A softer reading relativises meaning or assertibility rather than truth itself. The entry asks which reading, if any, survives the self-refutation charge.
Scepticism: global and restricted. Scepticism is a position, not mere doubt: some class of claims cannot be known, or there is no reason to believe them. Descartes' (1641) evil demon feeds a disembodied mind false experiences; the brain in a vat is the modern form. Since our evidence comes from the senses, we seem to have no reason to reject the hypothesis. Responses: Moore (raising his hands), Putnam's externalism, inference to the best explanation, contextualism. Restricted forms: inductive scepticism (will the sun rise tomorrow?) and other minds. If global scepticism is true, believing in trees is no more rational than believing in fairies.
How far sceptical hypotheses reach.
Survival vs identity (Parfit). After fission, the question "which one am I?" may be empty: there is no further fact. What matters for prudence and responsibility is psychological connectedness, present in both products. So identity is not what matters in survival. The simple view replies that personal identity is primitive: there is always a fact about which (if either) product is me, even if we cannot know it. Locke's prince and cobbler thought experiment opened the debate by locating the person with consciousness rather than body; animalism and the bodily criterion are later counters.
Depth: tensed and tenseless theories, tropes, truth and time travel
Tensed (A) vs tenseless (B) theory. A-theory: there are irreducible tensed facts (it is raining now; the Battle of Hastings is past). B-theory: only tenseless relations (earlier than, later than) and tenseless facts; tense is linguistic or indexical. McTaggart argued that the A-series is contradictory (every event is past, present and future) and that the B-series alone cannot give change. Replies distinguish A-properties held at times from simultaneous possession, or reject the demand that change require A-properties. Presentism and the growing block are typically A-theoretic; eternalism can go either way (moving spotlight = eternalism + A).
Trope theory. Properties are constituents of objects but unrepeatable. Our books share "the same property" as two soldiers share "the same uniform", not as brothers share "the same father". Similarity becomes resemblance between tropes, perhaps primitive. Tropes promise the benefits of realism (properties as constituents) without repeatable universals, and they help bundle theory avoid the two-electron problem. Critics ask whether resemblance must itself be a trope, restarting a regress, and whether science needs genuine shareable properties.
Four theories of truth in the long entry.
Truth: correspondence, coherence, deflationism, identity. Correspondence: a truth-bearer is true iff it corresponds to a fact or to the world. Coherence: truth is membership in a coherent system of beliefs. Deflationism (redundancy, disquotational): saying "p is true" adds nothing beyond saying p; the truth predicate is a device for generalisation. The identity theory of truth identifies true propositions with facts. Each view faces stock objections: correspondence needs an account of facts and of the correspondence relation; coherence struggles with multiple coherent systems; deflationism may be too thin for truthmaker theory; identity theories must explain falsehood. The entry is among the book's longest because truth sits under realism, time (truthmakers for the past) and laws.
Time travel and consistency. If time travel is possible, the grandfather paradox asks whether Tim can kill his grandfather before his father is conceived. Consistency constraints (Novikov; Lewis on what Tim can and cannot do) say that the past includes Tim's failure: he can try, but the events already include his not succeeding. Relativity's block universe makes the metaphysics of "visiting" another time less exotic, while still leaving the causal loops and free-will puzzles. The entry links back to eternalism, causation and the ability to do otherwise.
Depth: universals, vagueness, Zeno and closing method
Universals: one over many. The Coke can and the telephone box are both red. Realists say they share a universal, redness, either immanent or Platonic. Nominalists deny shared universals and offer classes, tropes or primitive resemblance. The "one over many" argument is the classic motivation; the companionship and imperfect-community problems press resemblance and class views; Bradley's regress presses relational accounts of instantiation. Armstrong's states of affairs try to stop the regress with a non-relational tie.
Vagueness. "Heap", "bald" and "tall" admit borderline cases. Epistemicism says there is a sharp cut-off we cannot know. Supervaluationism says borderline claims are neither true nor false on all precisifications. Many-valued and degree theories assign intermediate values. Ontic vagueness locates vagueness in the world, not only in language. Metaphysicians care because composition, persistence and personal identity may be vague: is there a fact about when the ship of Theseus ceased to be the same ship?
Particulars-and-universals path through Q to Z.
Zeno's paradoxes and modern replies. Achilles and the tortoise: to catch the tortoise Achilles must complete infinitely many tasks (halfway, three-quarters, ...). The dichotomy: to reach the door you must first reach the halfway point, and so on. Modern calculus treats the infinite sum of decreasing intervals as finite; some metaphysicians still ask whether a supertask (infinitely many actions in finite time) is possible. Aristotle's potential vs actual infinity returns here: if only potential infinities are real, the completed series of tasks never exists as a whole.
Method closing the alphabet. The Q-Z block returns the reader to cost-benefit reasoning, to the metaphile/metaphobe dispute, and to the six topics. Truth, tropes and universals finish particulars-and-universals; tensed/tenseless and time travel finish time; Ramsey-Lewis finishes laws; quasi-memory and survival finish personal identity; relativism and scepticism finish realism. The alphabetical format thus ends where the introduction began: concepts as the route into debates about what reality contains.
Depth: truth as property, deflationism and truthmakers
Is truth a metaphysically thick property? Everyday talk treats "true" as unproblematic; metaphysics asks whether truth needs a nature. Because "... is true" is a predicate, truth seems a property of sentences, propositions or beliefs. Class nominalists can allow the property while saying it is only membership in the class of truths. Deflationists go further: the truth predicate is a device for generalisation and endorsement ("everything the Pope says is true" compresses an infinite conjunction), not a name for a deep relation.
Correspondence and realism. Metaphysical realism (reality independent of our minds) is often paired with correspondence: p is true iff p corresponds to the facts. Correspondence needs an account of facts and of corresponding. Coherence theories risk relativism if multiple maximal coherent systems disagree. Identity theories equate true propositions with facts and must still explain falsehood and logical complexity.
Deflationism and truthmaker theory as rival pressures.
Truthmakers without truthmaker maximalism. Truthmaker theory says truths are true because of being. Maximalists want a truthmaker for every truth; others allow that "Mary rarely goes out without her umbrella" is made true by a collection of particular goings-out, without a special rarity fact. Presentism's struggle with past-tense truthmakers and negative existentials ("there are no unicorns") are central stress tests. The entry's length reflects truth's role under realism, time and laws.
Depth: Ramsey-Lewis best systems and the language objection
Simplicity vs strength. Theory A: a single weak axiom. Theory C: a list of every truth (maximal strength, insane complexity). Theory B balances both. Laws are the axioms or theorems of the best balance over the mosaic. Accidental regularities ("all gold lumps are under a mile across") typically lose to simpler, stronger rivals and so are not laws, unlike naive regularity theory.
The language / "grue" pressure. Suppose another culture's predicates make a gerrymandered theory look simple. If "best" depends on language, laws would seem language-relative, which is absurd for governing nature. Lewis answers by restricting the competition to systems whose predicates express natural properties (carving joints). The best-system story thus depends on the naturalness hierarchy from Chapter 1. Without natural properties, Humean laws threaten to become conventional; with them, the theory regains objectivity at the price of an extra metaphysical posit.
Ramsey-Lewis scoring of candidate systematizations.
Probabilistic laws and governance. Best systems can include chance axioms (Lewis's "chance"-friendly Humeanism), but fitting probabilities into simplicity/strength scoring is delicate. Anti-Humeans also object that systematisation describes the mosaic without governing it: laws should explain the patterns, not summarise them. The entry leaves that governance intuition as a standing cost against RLV.
Depth: special composition, supervenience and substance
Special Composition Question (van Inwagen). When do some things compose a further thing? Nihilism: only simples; talk of tables is convenient fiction or plural reference. Universalism: any things compose something. Restrictivism: composition under restricted conditions (life, bonding, contact). Each answer reshapes ontology's size. The 2011 text presents nihilism and universalism as common poles; later survey data (literature box) show moderate restrictivism as frequent among respondents.
Supervenience. A-properties supervene on B-properties when no A-difference without a B-difference. Humean supervenience is a global claim about the mosaic. Supervenience is necessary but often thought too weak for "nothing over and above"; grounding talk (literature notes) later tries to supply asymmetric explanation.
Van Inwagen's Special Composition Question, three poles.
Substance. Enduring particular that bears properties; contrasts with bundles of tropes or universals and with events. Agent causation needs substances as causes. Bundle theories try to do without substrata. The entry closes a loop with Chapter 1's bundle vs substance-attribute dispute and Bradley's regress.
Depth: correspondence, coherence and the Pope generalisation
Why deflationists still want "true". "Everything the Pope says is true" cannot be replaced by a finite list of conditionals, one per possible utterance. The truth predicate enables generalisation and endorsement. That linguistic role, deflationists say, exhausts truth's nature; correspondence theorists say the predicate works because of a world-relating property.
Coherence and rival systems. If truth is coherence with a system, two maximal coherent systems can disagree, making the same sentence true in one and false in the other unless further constraints (ideal science; convergence) are added. Those constraints often smuggle correspondence-like contact with the world. Correspondence, for its part, must say what facts are and how sentences latch onto them without invoking truth again.
Identity theory. True propositions just are facts. Attractive for truthmaker theorists who want no gap between bearer and maker; awkward for falsehood, for contingent truths about the future on some views, and for logically complex truths that do not look like atomic facts.
Special Composition Question poles.
Grandfather paradox under Lewisian consistency.
Tensed versus tenseless structure of time.
Three attitudes to borderline heaps.
Zeno dichotomy meeting the calculus reply.
Main Points
- Distinguish the senses first. "Realism", "nominalism", "essence", "fusion", "four-dimensionalism" and "experimental philosophy" each have two or more meanings; many disputes are partly verbal.
- Free will is a dilemma, not a single problem. Determinism seems to remove the ability to do otherwise; indeterminism seems to make choices a matter of luck. Every position in Table 1 is a way through.
- Identity's logic drives the personal identity debate. Because identity is one-one and transitive, fission forces a choice between no-branching clauses, multiple occupancy, the simple view and Parfit's claim that identity is not what matters.
- Properties pose a question about sameness. Universals, tropes, classes and resemblance are rival accounts of what the telephone box and the Coke can share; austere nominalism refuses the question.
- Time has two questions. Which times exist (presentism, growing block, eternalism) is separate from whether tense is objective (A-theory versus B-theory); relativity bears on both.
- Causation splits Humeans from anti-Humeans. Regularity, counterfactual and best-system theories avoid natural necessity; DTA, powers and singularism embrace it.
- Modality has three faces. What is necessary, how we know it, and what makes it so come apart, as Kripke's necessary a posteriori truths show.
- Counterexamples are the engine of analysis. Pre-emption, overdetermination, Frankfurt cases and fission each target one half of a proposed necessary and sufficient condition.
- Theory choice is cost-benefit analysis. Parsimony, simplicity, explanatory power and common sense are weighed, and which features count as virtues is itself contested.
- Metaphysics must answer the metaphobe. From Hume's flames to Carnap's frameworks, the question whether armchair reasoning can reveal mind-independent reality hangs over every entry.
The single most useful habit the book teaches is to ask, of any metaphysical claim, which sense of its key terms is in play, which argument supports it, and what it costs. A reader who carries those three questions through the alphabet can enter any of the debates surveyed here, and most of those that have arisen since 2011, from grounding to powers to free will scepticism.
| Theory | What it says | Use it when |
|---|---|---|
| Quine's criterion | Commitments are values of bound variables | Testing a paraphrase |
| Conceptual analysis | Necessary and sufficient conditions | Defining freedom or cause |
| Canberra Plan | Collect platitudes, find the occupant | Mind, causation, colour |
| Possible worlds | Modal claims quantify over worlds | Necessity, counterfactuals |
| Counterfactual analysis | Cause as difference-maker | Token causation |
| Best system | Laws are axioms of the ideal theory | Laws without necessity |
| Supervenience | No A-difference without B-difference | Nothing over and above |
| Theoretical virtues | Weigh costs and benefits | Choosing between theories |
The book's method, applied across its six topics.
How the six topics connect. The book's cross-references reveal a dense web beneath the alphabet. Free will depends on causation (is agent causation possible?) and on laws (does determinism hold?). Personal identity depends on identity over time, which depends on the persistence debate (endurance or perdurance), which depends on the ontology of time (presentism makes temporal parts hard to have). Universals connect to laws through DTA and to modality through Lewis's class nominalism across possible worlds. Realism and anti-realism frame them all: an anti-realist about mind-independent reality will treat every other debate differently. Reading the entries in this order, following the cross-references, turns the dictionary into a short course.
What the book leaves out. The book is avowedly a book of concepts, so some areas receive little space. Philosophy of mind appears mainly through Armstrong and the Canberra Plan; philosophy of religion through idealism's God; social metaphysics (race, gender, social construction) not at all, and has since become one of the field's growth areas (Haslanger, Ásta, Mason). Grounding appears only implicitly in supervenience and reductionism. Quantum mechanics is mentioned in indeterminism and Humean supervenience but its interpretations, which bear on determinism and locality, are not surveyed.
Six questions to test understanding. A reader who has worked through the book should be able to answer these. (1) Why does the consequence argument need a transfer principle, and how does Frankfurt's case bear on its other premise? (2) Why does fission threaten psychological continuity theories but not, in the same way, the simple view? (3) What is the difference between a trope theorist and a resemblance nominalist in explaining why the telephone box and the Coke can are both red? (4) How can a truth be necessary yet knowable only a posteriori? (5) Why does pre-emption refute the naive counterfactual theory, and what do Figures 3 and 4 show? (6) What would have to be true for presentism to survive special relativity? Each question draws on entries from at least two chapters, which is the point: the book's concepts are interdependent.
Each entry insists on saying what a term means, what positions it allows and what can be said for and against each. That discipline is exactly what newer debates about grounding, social construction and conceptual engineering need, and a student who has absorbed it is well equipped to read them.
reading order and open questions
Suggested path through the six topics. For free will: Consequence Argument → Compatibilism → Frankfurt → PAP → Libertarianism → Problem of luck → Agent causation. For personal identity: Locke/thought experiments → Psychological continuity → Fission → Quasi-memory → Survival/Parfit → Simple view → Animalism. For particulars and universals: One over many → Class/resemblance/austere nominalism → Tropes → Immanent/Platonic realism → Bradley's regress → Bundle vs substratum. For time: McTaggart → Presentism/Growing block/Eternalism → A/B theory → Relativity → Time travel. For causation and laws: Counterfactual dependence → Pre-emption → Process theories → DTA → Ramsey-Lewis → Natural necessity → Humean supervenience. For realism: Scepticism → Brains in a vat → Phenomenalism → Truth theories → Relativism → Metametaphysics.
Disputes the 2011 text leaves live.
What the 2011 text leaves open. The authors rarely pick winners. Live disputes include: whether Frankfurt cases refute PAP; whether fission shows identity is not what matters; whether Humean supervenience can ground laws and chances; whether presentism can supply truthmakers and survive relativity; whether tropes or universals better serve science; whether metametaphysical quietism undercuts the whole project.
Coverage checklist after the depth pass
Free will cluster. Agent causation, compatibilism (Hume, Frankfurt, Fischer, Dennett), consequence argument, Frankfurt cases, PAP, libertarianism, problem of luck, illusionism, moral responsibility deep/shallow: each has argument structure, main replies and cross-links.
Identity cluster. Animalism, bodily and psychological criteria, fission clauses, quasi-memory, survival/Parfit, simple view, thought experiments (prince/cobbler, teletransport): each states the analysis and the decisive objection.
Universals cluster. Austere/class/resemblance nominalism, tropes, immanent and Platonic realism, bundle vs substratum, Bradley's regress, natural properties: each positions the others.
Time and modality. Presentism, growing block, eternalism, A/B theory, McTaggart, relativity, time travel; possible worlds, GMR/ersatz, kinds of modality, necessary a posteriori.
Causation and laws. Counterfactual, probabilistic, process; pre-emption early/late/trumping; DTA; Ramsey-Lewis; natural necessity; Humean supervenience.
Realism. Scepticism, brains in a vat, phenomenalism/idealism, relativism, theories of truth, truthmakers, metametaphysics.
Depth-pass tallies and reading advice
Mid-length entries gained the missing framework items (kinds of modality; composition poles; luck problem; too-many-thinkers). Short cross-reference entries remain in glossary tables.
Literature boxes are field snapshots (PhilPapers 2020; named post-2011 books), not replacements for the 2011 arguments. Diagrams now track mechanisms entry by entry; tables hold free-will positions and block maps.
Synthesis checklist of the book's spines.
Method to carry away after the alphabet.
- Distinguish the senses first. "Realism", "nominalism", "essence", "fusion", "four-dimensionalism" and "experimental philosophy" each have two or more meanings; many disputes are partly verbal.
- Free will is a dilemma, not a single problem. Determinism seems to remove the ability to do otherwise; indeterminism seems to make choices a matter of luck. Every position in Table 1 is a way through.
- Identity's logic drives the personal identity debate. Because identity is one-one and transitive, fission forces a choice between no-branching clauses, multiple occupancy, the simple view and Parfit's claim that identity is not what matters.
- Properties pose a question about sameness. Universals, tropes, classes and resemblance are rival accounts of what the telephone box and the Coke can share; austere nominalism refuses the question.
- Time has two questions. Which times exist (presentism, growing block, eternalism) is separate from whether tense is objective (A-theory versus B-theory); relativity bears on both.
- Causation splits Humeans from anti-Humeans. Regularity, counterfactual and best-system theories avoid natural necessity; DTA, powers and singularism embrace it.