green numbers give full details.     |    back to list of philosophers     |     unexpand these ideas

---

Ideas of David Lewis, by Text

[American, 1941 - 2001, Pupil of Willard Quine. Professor at Princeton University.]

1966	An Argument for the Identity Theory

§I	p.100	7441	Experiences are defined by their causal role, and causal roles belong to physical states
			Full Idea: The definitive characteristic of any experience is its causal role, its most typical causes and effects; but we materialists believe that these causal roles which belong by analytic necessity to experiences belong in fact to certain physical states.
			From: David Lewis (An Argument for the Identity Theory [1966], §I)
			A reaction: This is the Causal version of functionalism, which Armstrong also developed. The word 'typical' leads later to a teleological element in the theory (e.g. in Lycan). There are other things to say about mental states than just their causal role.

§II n6	p.101	7442	'Pain' contingently names the state that occupies the causal role of pain
			Full Idea: On my theory, 'pain' is a contingent name - that is, a name with different denotations in different possible worlds - since in any world, 'pain' names whatever state happens in that world to occupy the causal role definitive of pain.
			From: David Lewis (An Argument for the Identity Theory [1966], §II n6)
			A reaction: Better to say that 'pain' (like 'sound') is ambiguous. It is indiscriminately used by English-speakers to mean [1] the raw quale that we experience when damaged, and [2] whatever it is that leads to pain behaviour. Maybe frogs have 2 but not 1.

1968	Counterpart theory and Quant. Modal Logic

	p.45	16994	Counterpart theory is bizarre, as no one cares what happens to a mere counterpart
			Full Idea: Probably Humphrey could not care less whether someone else, no matter how much resembling him, would have been victorious in another possible world. Thus Lewis's view seems even more bizarre that the usual transworld identification it replaces.
			From: comment on David Lewis (Counterpart theory and Quant. Modal Logic [1968]) by Saul A. Kripke - Naming and Necessity notes and addenda note 13
			A reaction: I begin to see this as a devastating reply to a theory I previously found quite congenial.

I	p.28	11974	Counterparts are not the original thing, but resemble it more than other things do
			Full Idea: Your counterparts resemble you closely in content and context in important respects. They resemble you more closely than do the other things in their worlds. But they are not really you.
			From: David Lewis (Counterpart theory and Quant. Modal Logic [1968], I)
			A reaction: It is a dilemma. If my counterpart were exactly me, I couldn't contemplate possibly losing a leg, or my sanity. But if my counterpart isn't exactly me, then I don't have much interest in its fate. Only essences can save us here. Cf. me tomorrow.

I	p.29	11975	If the closest resembler to you is in fact quite unlike you, then you have no counterpart
			Full Idea: If whatever thing in world w6 it is that resembles you more closely than anything else in w6 is nevertheless quite unlike you; nothing in w6 resembles you at all closely. If so, you have no counterpart in w6.
			From: David Lewis (Counterpart theory and Quant. Modal Logic [1968], I)
			A reaction: This is the nub, because the whole theory rests on deciding whether two things resemble sufficiently 'closely'. But then we need a criterion of closeness, so we must start talking about which properties matter. Essences loom.

III	p.32	11976	Aristotelian essentialism says essences are not relative to specification
			Full Idea: So-called 'Aristotelian essentialism' is the doctrine of essences not relative to specifications.
			From: David Lewis (Counterpart theory and Quant. Modal Logic [1968], III)
			A reaction: In other words, they are so-called 'real essences', understood as de re. Quine says essences are all de dicto, and relative to some specification. I vote for Aristotle.

III	p.35	11977	Essential attributes are those shared with all the counterparts
			Full Idea: An essential attribute of something is an attribute it shares with all its counterparts.
			From: David Lewis (Counterpart theory and Quant. Modal Logic [1968], III)
			A reaction: I don't like this. It ties essence entirely to identity, but I think essence precedes identity. Essence is a nexus of causal and explanatory powers which bestows an identity on each thing. But essence might be unstable, and identity with it.

Post B	p.41	11979	It doesn't take the whole of a possible Humphrey to win the election
			Full Idea: Even if Humphrey is a modal continuant, it doesn't take the whole of him to do such things as winning.
			From: David Lewis (Counterpart theory and Quant. Modal Logic [1968], Post B)
			A reaction: This responds to Kripke's famous example, that people only care about what happens to themselves, and not to some 'counterpart' of themselves.

V	p.37	11978	Causal necessities hold in all worlds compatible with the laws of nature
			Full Idea: Just as a sentence is necessary if it holds in all worlds, so it is causally necessary if it holds in all worlds compatible with the laws of nature.
			From: David Lewis (Counterpart theory and Quant. Modal Logic [1968], V)
			A reaction: I don't believe in the so-called 'laws of nature', so I'm not buying that. Is there no distinction in Lewis's view between those sentences which must hold, and those which happen to hold universally?

1970	Anselm and Actuality

	p.83	16456	For modality Lewis rejected boxes and diamonds, preferring worlds, and an index for the actual one
			Full Idea: Lewis was suspicious of boxes and diamonds as regimenting ordinary modal thought, …preferring a first-order extensional theory including possible worlds in its domain and an indexical singular term for the actual world.
			From: report of David Lewis (Anselm and Actuality [1970]) by Robert C. Stalnaker - Mere Possibilities 3.8

1970	How to Define Theoretical Terms

Intro	p.78	15527	Defining terms either enables elimination, or shows that they don't require elimination
			Full Idea: To define theoretical terms might be to show how to do without them, but it is better to say that it shows there is no good reason to want to do without them.
			From: David Lewis (How to Define Theoretical Terms [1970], Intro)

Intro	p.78	15526	There is a method for defining new scientific terms just using the terms we already understand
			Full Idea: I contend that there is a general method for defining newly introduced terms in a scientific theory, one which uses only the old terms we understood beforehand.
			From: David Lewis (How to Define Theoretical Terms [1970], Intro)
			A reaction: Lewis is game is to provide bridge laws for a reductive account of nature, without having to introduce something entirely new to achieve it. The idea of bridge laws in scientific theory is less in favour these days.

II	p.81	15528	A Ramsey sentence just asserts that a theory can be realised, without saying by what
			Full Idea: If we specify a theory with all of its terms, and then replace all of those terms with variables, we can then say that some n-tuples of entities can satisfy this formula. This Ramsey sentence then says the theory is realised, without specifying by what.
			From: David Lewis (How to Define Theoretical Terms [1970], II)
			A reaction: [I have compressed Lewis, and cut out the symbolism]

III	p.83	15529	It is better to have one realisation of a theory than many - but it may not always be possible
			Full Idea: A uniquely realised theory is, other things being equal, certainly more satisfactory than a multiply realised theory. We should insist on unique realisation as a standard of correctness unless it is a standard too high to be met.
			From: David Lewis (How to Define Theoretical Terms [1970], III)
			A reaction: The point is that rewriting a theory as Ramsey sentences just says there is at least one realisation, and so it doesn't meet the highest standards for scientific theories. The influence of set-theoretic model theory is obvious in this approach.

III	p.86	15530	A logically determinate name names the same thing in every possible world
			Full Idea: A logically determinate name is one which names the same thing in every possible world.
			From: David Lewis (How to Define Theoretical Terms [1970], III)
			A reaction: This appears to be rigid designation, before Kripke introduced the new word.

V	p.89	15531	The Ramsey sentence of a theory says that it has at least one realisation
			Full Idea: The Ramsey sentence of a theory says that it has at least one realisation.
			From: David Lewis (How to Define Theoretical Terms [1970], V)

1972	Psychophysical and theoretical identifications

	p.9	9409	Laws are the best axiomatization of the total history of world events or facts
			Full Idea: The Mill-Ramsey-Lewis theory takes laws to be axioms (or theorems) of the best possible systematizations of the world's total history, where such a history is a history of events or facts.
			From: report of David Lewis (Psychophysical and theoretical identifications [1972]) by Stephen Mumford - Laws in Nature 1.3

	p.45	9423	If simplicity and strength are criteria for laws of nature, that introduces a subjective element
			Full Idea: Lewis's simplicity and strength criteria introduce an element of subjectivity into the laws, because the best system seems to be determined by what we take to be simple and strong in a system.
			From: comment on David Lewis (Psychophysical and theoretical identifications [1972]) by Stephen Mumford - Laws in Nature 3.5
			A reaction: [Mumford cites Armstrong 1983:67 for this]

	p.45	9424	A number of systematizations might tie as the best and most coherent system
			Full Idea: Since the best system view is a coherence theory, the possibility could not be ruled out that a number of different systematizations of the same history might be tied for first place as equally best.
			From: comment on David Lewis (Psychophysical and theoretical identifications [1972]) by Stephen Mumford - Laws in Nature 3.5
			A reaction: [Mumord cites Armstrong 1983:70] Personally I am a fan of coherence theories, and this problem doesn't bother me.

1973

Causation

	p.37	9476	If dispositions are more fundamental than causes, then they won't conceptually reduce to them
			Full Idea: Maybe a disposition is a more fundamental notion than a cause, in which case Lewis has from the very start erred in seeking a causal analysis, in a traditional, conceptual sense, of disposition terms.
			From: comment on David Lewis (Causation [1973]) by Alexander Bird - Nature's Metaphysics 2.2.8
			A reaction: Is this right about Lewis? I see him as reducing both dispositions and causes to a set of bald facts, which exist in possible and actual worlds. Conditionals and counterfactuals also suffer the same fate.

	p.148	4795	Lewis's account of counterfactuals is fine if we know what a law of nature is, but it won't explain the latter
			Full Idea: Lewis can elucidate the logic of counterfactuals on the assumption that you are not at all puzzled about what a law of nature is. But if you are puzzled about this, it cannot contribute anything towards resolving your puzzlement.
			From: comment on David Lewis (Causation [1973]) by L. Jonathan Cohen - The Problem of Natural Laws p.219
			A reaction: This seems like a penetrating remark. The counterfactual theory is wrong, partly because it is epistemological instead of ontological, and partly because it refuses to face the really difficult problem, of what is going on out there.

	p.161	17524	Lewis has basic causation, counterfactuals, and a general ancestral (thus handling pre-emption)
			Full Idea: Lewis's basic account has a basic causal relation, counterfactual dependence, and the general causal relation is the ancestral of this basic one. ...This is motivated by counterfactual dependence failing to be general because of the pre-emption problem.
			From: report of David Lewis (Causation [1973]) by Alexander Bird - Causation and the Manifestation of Powers p.161
			A reaction: It is so nice when you struggle for ages with a topic, and then some clever person summarises it clearly for you.

	p.162	10392	It is just individious discrimination to pick out one cause and label it as 'the' cause
			Full Idea: We sometimes single out one among all the causes of some event and call it 'the' cause. ..We may select the abnormal causes, or those under human control, or those we deem good or bad, or those we want to talk about. This is invidious discrimination.
			From: David Lewis (Causation [1973])
			A reaction: This is the standard view expressed by Mill - presumably the obvious empiricist line. But if we specify 'the pre-conditions' for an event, we can't just mention ANY fact prior to the effect - there is obvious relevance. So why not for 'the' cause as well?

	p.162	17525	The counterfactual view says causes are necessary (rather than sufficient) for their effects
			Full Idea: The Humean idea, developed by Lewis, is that rather than being sufficient for their effects, causes are (counterfactual) necessary for their effects.
			From: report of David Lewis (Causation [1973]) by Alexander Bird - Causation and the Manifestation of Powers p.162

	p.409	8397	Counterfactual causation implies all laws are causal, which they aren't
			Full Idea: Some counterfactuals are based on non-causal laws, such as Newton's Third Law of Motion. 'If no force one way, then no force the other'. Lewis's counterfactual analysis implies that one force causes the other, which is not the case.
			From: comment on David Lewis (Causation [1973]) by Michael Tooley - Causation and Supervenience 5.2
			A reaction: So what exactly does 'cause' my punt to move forwards? Basing causal laws on counterfactual claims looks to me like putting the cart before the horse.

	p.454	8405	A theory of causation should explain why cause precedes effect, not take it for granted
			Full Idea: Lewis thinks it is a major defect in a theory of causation that it builds in the condition that the time of the cause precede that of the effect: that cause precedes effect is something we ought to explain (which his counterfactual theory claims to do).
			From: report of David Lewis (Causation [1973]) by Hartry Field - Causation in a Physical World
			A reaction: My immediate reaction is that the chances of explaining such a thing are probably nil, and that we might as well just accept the direction of causation as a given. Even philosophers balk at the question 'why doesn't time go backwards?'

p.193	p.193	8420	A proposition is a set of possible worlds where it is true
			Full Idea: I identify a proposition with the set of possible worlds where it is true.
			From: David Lewis (Causation [1973], p.193)
			A reaction: As it stands, I'm baffled by this. How can 'it is raining' be a set of possible worlds? I assume it expands to refer to the truth-conditions, among possibilities as well as actualities. 'It is raining' fits all worlds where it is raining.

p.193	p.193	8419	The modern regularity view says a cause is a member of a minimal set of sufficient conditions
			Full Idea: In present-day regularity analyses, a cause is defined (roughly) as any member of any minimal set of actual conditions that are jointly sufficient, given the laws, for the existence of the effect.
			From: David Lewis (Causation [1973], p.193)
			A reaction: This is the view Lewis is about to reject. It seem to summarise the essence of the Mackie INUS theory. This account would make the presence of oxygen a cause of almost every human event.

p.194	p.194	8421	Regularity analyses could make c an effect of e, or an epiphenomenon, or inefficacious, or pre-empted
			Full Idea: In the regularity analysis of causes, instead of c causing e, c might turn out to be an effect of e, or an epiphenomenon, or an inefficacious effect of a genuine cause, or a pre-empted cause (by some other cause) of e.
			From: David Lewis (Causation [1973], p.194)
			A reaction: These are Lewis's reasons for rejecting the general regularity account, in favour of his own particular counterfactual account. It is unlikely that c would be regularly pre-empted or epiphenomenal. If we build time's direction in, it won't be an effect.

p.195	p.195	8423	My counterfactual analysis applies to particular cases, not generalisations
			Full Idea: My (counterfactual) analysis is meant to apply to causation in particular cases; it is not an analysis of causal generalizations. Those presumably quantify over particulars, but it is hard to match natural language to the quantifiers.
			From: David Lewis (Causation [1973], p.195)
			A reaction: What authority could you have for asserting a counterfactual claim, if you only had one observation? Isn't the counterfactual claim the hallmark of a generalisation? For one case, 'if not-c, then not-e' is just a speculation.

p.196	p.196	8424	Determinism says there can't be two identical worlds up to a time, with identical laws, which then differ
			Full Idea: By determinism I mean that the prevailing laws of nature are such that there do not exist any two possible worlds which are exactly alike up to that time, which differ thereafter, and in which those laws are never violated.
			From: David Lewis (Causation [1973], p.196)
			A reaction: This would mean that the only way an action of free will could creep in would be if it accepted being a 'violation' of the laws of nature. Fans of free will would probably prefer to call it a 'natural' phenomenon. I'm with Lewis.

p.197	p.197	8425	For true counterfactuals, both antecedent and consequent true is closest to actuality
			Full Idea: A counterfactual is non-vacuously true iff it takes less of a departure from actuality to make the consequent true along with the antecedent than it does to make the antecedent true without the consequent.
			From: David Lewis (Causation [1973], p.197)
			A reaction: Almost every theory proposed by Lewis hangs on the meaning of the word 'close', as used here. If you visited twenty Earth-like worlds (watch Startrek?), it would be a struggle to decide their closeness to ours in rank order.

p.200	p.200	8426	One event causes another iff there is a causal chain from first to second
			Full Idea: One event is the cause of another iff there exists a causal chain leading from the first to the second.
			From: David Lewis (Causation [1973], p.200)
			A reaction: It will be necessary to both explain and identify a 'chain'. Some chains are extremely tenuous (Alexander could stop a barrel of beer). Go back a hundred years, and the cause of any present event is everything back then.

p.203	p.203	8427	I reject making the direction of causation axiomatic, since that takes too much for granted
			Full Idea: One might stipulate that a cause must always precede its effect, but I reject this solution. It won't solve the problem of epiphenomena, it rejects a priori any backwards causation, and it trivializes defining time-direction through causation.
			From: David Lewis (Causation [1973], p.203)
			A reaction: [compressed] Not strong arguments, I would say. Maybe apparent causes are never epiphenomenal. Maybe backwards causation is impossible. Maybe we must use time to define causal direction, and not vice versa.

1973	Counterfactuals

	p.213	8434	In good counterfactuals the consequent holds in world like ours except that the antecedent is true
			Full Idea: According to Lewis, a counterfactual holds when the consequent is true in possible worlds very like our own except for the fact that the antecedent is true.
			From: report of David Lewis (Counterfactuals [1973]) by Paul Horwich - Lewis's Programme p.213
			A reaction: Presumably the world being very like our own would make it unlikely that there would be anything else to cause the consequent, apart from the counterfactual antecedent.

	p.224	14361	Lewis says indicative conditionals are truth-functional
			Full Idea: Unlike Stalnaker, Lewis holds that indicative conditionals have the truth conditions of material conditionals.
			From: report of David Lewis (Counterfactuals [1973]) by Frank Jackson - Conditionals 'Further'
			A reaction: Thus Lewis only uses the possible worlds account for subjunctive conditionals, where Stalnaker uses it for both. Lewis is defending the truth-functional account for the indicative conditionals.

3.3	p.73	9419	A law of nature is a general axiom of the deductive system that is best for simplicity and strength
			Full Idea: A contingent generalization is a law of nature if and only if it appears as a theorem (or axiom) in each of the true deductive systems that achieves a best combination of simplicity and strength.
			From: David Lewis (Counterfactuals [1973], 3.3)

1973	Possible Worlds

	p.288	15789	Lewis's distinction of 'existing' from 'being actual' is Meinong's between 'existing' and 'subsisting'
			Full Idea: I suggest that Lewis's view in fact is just Meinong's view. ...Meinong distinguishes between 'existing' and merely 'subsisting', Lewis between 'being actual' and merely 'existing'.
			From: comment on David Lewis (Possible Worlds [1973]) by William Lycan - The Trouble with Possible Worlds 06
			A reaction: Lewis attempts to make actuality purely 'indexical' in character, like distinguishing the world 'here' from the world 'elsewhere', but Lycan seems right that he is committed to more than that.

	p.295	15790	Lewis can't know possible worlds without first knowing what is possible or impossible
			Full Idea: Lewis's knowledge of what possible worlds there are and of other general truths about worlds is posterior, not prior, to his knowledge of what things are possible and what things are impossible.
			From: comment on David Lewis (Possible Worlds [1973]) by William Lycan - The Trouble with Possible Worlds 07
			A reaction: This elementary objection seems to me to destroy any attempt to explain modality in terms of possible worlds. It is a semantics for modal statements, but that doesn't make it an ontology. To assess possibilities, study actuality.

	p.297	15791	What are the ontological grounds for grouping possibilia into worlds?
			Full Idea: Lewis must seek some ontological ground for the grouping of possibilia into disjoint worlds.
			From: comment on David Lewis (Possible Worlds [1973]) by William Lycan - The Trouble with Possible Worlds 07
			A reaction: I do love people like Lycan who ask the simple commonsense questions about these highly sophisticated systems that students of philosophy are required to study. If a proposition is a 'set of worlds', understanding a proposition is beyond me.

1973

works

	p.5	4398	An event causes another just if the second event would not have happened without the first
			Full Idea: Lewis gives an account of causation in terms of counterfactual conditionals (roughly, an event c causes an event e iff if c had not happened then e would not have happened either).
			From: report of David Lewis (works [1973]) by Stathis Psillos - Causation and Explanation Intro
			A reaction: This feels wrong to me. It is a version of Humean constant conjunction, but counterfactuals are too much a feature of our minds, and not sufficiently a feature of the world, to do this job. Tricky.

1976	The paradoxes of time travel

	p.194	23019	The interesting time travel is when personal and external time come apart
			Full Idea: Lewis says a journey only counts as interesting time travel if there is a disparity between personal time and external time. …It is also necessary that the person at the beginning of the journey is the same as the person at the end of the journey.
			From: report of David Lewis (The paradoxes of time travel [1976]) by Baron,S/Miller,K - Intro to the Philosophy of Time
			A reaction: Other cases would presumably be the whole of time accelerating or slowing, or going into reverse.

	p.202	23021	Lewis said it might just be that travellers to the past can't kill their grandfathers
			Full Idea: Lewis denied that the grandfather paradox makes time travel impossible. Time travel does not entail that it is possible to kill your grandfather. Maybe the traveller fails in their attempt every time they try to do it.
			From: report of David Lewis (The paradoxes of time travel [1976]) by Baron,S/Miller,K - Intro to the Philosophy of Time 8.2
			A reaction: The problem with this answer is that killing the grandfather is only an extreme case of making any change in the past which ripples forward into the present. Virtually all changes by the time traveller would be impossible (though not contradictory).

1976	Probabilities of Conditionals

	p.14	14283	A conditional probability does not measure the probability of the truth of any proposition
			Full Idea: Lewis was first to prove this remarkable result: there is no proposition A*B such that, in all probability distributions, p(A*B) = pA(B) [second A a subscript]. A conditional probability does not measure the probability of the truth of any proposition.
			From: report of David Lewis (Probabilities of Conditionals [1976]) by Dorothy Edgington - Conditionals (Stanf) 3.1
			A reaction: The equation says the probability of the combination of A and B is not always the same as the probability of B given A. Bennett refers to this as 'The Equation' in the theory of conditionals. Edgington says a conditional is a supposition and a judgement.

1979	Attitudes De Dicto and De Se

	p.89	18415	The actual world is just the world you are in
			Full Idea: Lewis equates knowing which world is actual with knowing which world one is in.
			From: report of David Lewis (Attitudes De Dicto and De Se [1979]) by Cappelen,H/Dever,Josh - The Inessential Indexical 05.1
			A reaction: [This view is not, of course, Actualism, but an alternative treatment of actuality, within a multitude of possibilities].

	p.90	18416	Attitudes involve properties (not propositions), and belief is self-ascribing the properties
			Full Idea: Lewis suggests that we take attitudes to have properties, rather than propositions, as contents. To stand in the belief relation to a property is to self-ascribe that property.
			From: report of David Lewis (Attitudes De Dicto and De Se [1979]) by Robert C. Solomon - Erotic Love as a Moral Virtue 05.1
			A reaction: This is the sort of convoluted suggestion that Lewis has to come up with, in pursuit of his project of a wholly consistent metaphysics. Examine Lewis's account of properties before you judge this proposal! Self-ascribing is joining a set!

	p.108	18418	A theory of perspectival de se content gives truth conditions relative to an agent
			Full Idea: Lewis's theory of a perspectival 'de se' content ...delivers truth conditions not absolutely, but only relative to a choice of agent/center.
			From: report of David Lewis (Attitudes De Dicto and De Se [1979]) by Cappelen,H/Dever,Josh - The Inessential Indexical 05.7
			A reaction: The proposal rests on a theory of 'centred' possible worlds, specifying the viewpoint of some agent within the whole system. It relies on accepting the idea that indexicals are special, which Cappelen and Dever reject.

	p.249	16390	Lewis's popular centred worlds approach gives an attitude an index of world, subject and time
			Full Idea: Many philosophers now prefer Lewis's centred worlds framework for indexicals …It is two-dimensional, saying an attitude only has a truth-value when evaluated with respect to a contextual index, containing a subject and time, as well as a world.
			From: report of David Lewis (Attitudes De Dicto and De Se [1979]) by François Recanati - Mental Files 18.2
			A reaction: [compressed; this is said to have largely ousted the older Kaplan-Perry view] You only begin to understand the possible worlds game when you see how many problems find proposed 'solutions' there.

	p.255	16392	A content is a property, and believing it is self-ascribing that property
			Full Idea: For Lewis, a belief mode is analysed by saying that to believe a content (analysed as a property) is for the subject of thought to 'self-ascribe' that property.
			From: report of David Lewis (Attitudes De Dicto and De Se [1979]) by François Recanati - Mental Files 18.3
			A reaction: Lewis is weird. I would have thought you only self-ascribe the 'property' when you find yourself believing it. Lewis seems desperate to eliminate mental language. Belief can be a primitive concept without being primitive in ontology.

1979	Counterfactual Dependence and Time's Arrow

	p.209	8433	There are few traces of an event before it happens, but many afterwards
			Full Idea: Lewis claims that most events are over-determined by subsequent states of the world, but not by their history. That is, the future of every event contains many independent traces of its occurrence, with little prior indication that it would happen.
			From: report of David Lewis (Counterfactual Dependence and Time's Arrow [1979]) by Paul Horwich - Lewis's Programme p.209
			A reaction: Lewis uses this asymmetry to deduce the direction of causation, and hence the direction of time. Most people (including me, I think) would prefer to use the axiomatic direction of time to deduce directions of causation. Lewis was very wicked.

1980	Mad Pain and Martian Pain

§II	p.123	7443	A theory must be mixed, to cover qualia without behaviour, and behaviour without qualia
			Full Idea: To pass our test it seems that our theory will have to be a 'mixed' theory, to account for the Madman (whose pain has odd causes, and odd effects) and also for the Martian (who has normal causes and effects, but an odd physical state).
			From: report of David Lewis (Mad Pain and Martian Pain [1980], §II) by PG - Db (ideas)
			A reaction: A statement that 'pain' is ambiguous (qualia/causal role) would help a lot here. Martians have the causal role but no qualia, and the madman has the qualia but lacks the causal role. I say lots of different qualia might have the same causal role.

§III	p.124	7444	Type-type psychophysical identity is combined with a functional characterisation of pain
			Full Idea: The materialist theory Armstrong and I proposed joins claims of type-type psychophysical identity with a behaviourist or functionalist way of characterising mental states such as pain.
			From: David Lewis (Mad Pain and Martian Pain [1980], §III)
			A reaction: Armstrong has backed off from 'type-type' identity, because the realisations of a given mental state might be too diverse to be considered of the same type. Putnam's machine functionalism allows the possibility of dualism.

§III	p.125	7445	The application of 'pain' to physical states is non-rigid and contingent
			Full Idea: The word 'pain' is a non-rigid designator; it is a contingent matter what state the concept and the word apply to. (Note: so the sort of theory Kripke argues against is not what we propose).
			From: David Lewis (Mad Pain and Martian Pain [1980], §III)
			A reaction: I like the view that a given quale is necessarily identical to a given mental state, but that many mental states might occupy a given behavioural role. The smell of roses might occupy the behavioural role of pain. Frog pain isn't quite like ours.

1980	Subjectivist's Guide to Objective Chance

p.124	p.46	9425	Lewis later proposed the axioms at the intersection of the best theories (which may be few)
			Full Idea: Later Lewis said we must choose between the intersection of the axioms of the tied best systems. He chose for laws the axioms that are in all the tied systems (but then there may be few or no axioms in the intersection).
			From: comment on David Lewis (Subjectivist's Guide to Objective Chance [1980], p.124) by Stephen Mumford - Laws in Nature

1983	Extrinsic Properties

	p.4	14979	Being alone doesn't guarantee intrinsic properties; 'being alone' is itself extrinsic
			Full Idea: The property of 'being alone in the world' is an extrinsic property, even though it has had by an object that is alone in the world.
			From: report of David Lewis (Extrinsic Properties [1983]) by Theodore Sider - Writing the Book of the World 01.2
			A reaction: I always choke on my cornflakes whenever anyone cites a true predicate as if it were a genuine property. This is a counterexample to Idea 14978. Sider offers another more elaborate example from Lewis.

I	p.111	15454	Extrinsic properties come in degrees, with 'brother' less extrinsic than 'sibling'
			Full Idea: Properties may be more or less intrinsic; being a brother has more of an admixture of intrinsic structure than being a sibling does, yet both are extrinsic.
			From: David Lewis (Extrinsic Properties [1983], I)
			A reaction: I suppose the point is that a brother is intrinsically male - but then a sibling is intrinsically human. A totally extrinsic relation would be one between entities which shared virtually no categories of existence.

I	p.112	15455	Total intrinsic properties give us what a thing is
			Full Idea: The way something is is given by the totality of its intrinsic properties.
			From: David Lewis (Extrinsic Properties [1983], I)
			A reaction: No. Some properties are intrinsic but trivial. The 'important' ones fix the identity (if the identity is indeed 'fixed').

1983	Introduction to Philosophical Papers I

p.1	p.-5	21461	I tried to be unsystematic and piecemeal, but failed; my papers presuppose my other views
			Full Idea: I should have like to be a piecemeal, unsystematic philosopher, offering independent proposals on a variety of topics. It was not be. I succumbed too often to the temptation to presuppose my views on one topic when writing on another.
			From: David Lewis (Introduction to Philosophical Papers I [1983], p.1)
			A reaction: He particularly mentions his possible worlds realism as a doctrine which coloured all his other work. A charming insight into the mind of a systematic thinker (called by someone 'the most systematic metaphysician since Leibniz').

1983	Survival and Identity, with postscript

	p.610	16079	De re modal predicates are ambiguous
			Full Idea: Lewis is perhaps the most prominent proponent of the view that de re modal predicates are ambiguous.
			From: report of David Lewis (Survival and Identity, with postscript [1983]) by Lynne Rudder Baker - Why Constitution is not Identity n25

1983	New work for a theory of universals

	p.18	10717	Natural properties figure in the analysis of similarity in intrinsic respects
			Full Idea: Lewis argues that there are natural properties, which makes various analyses possible, especially of similarity in intrinsic respects. Naturalness comes in degrees, with perfectly natural properties being the limiting case.
			From: report of David Lewis (New work for a theory of universals [1983]) by Alex Oliver - The Metaphysics of Properties 4
			A reaction: This sounds to be the wrong way round. We don't start with similarities and work back to natural properties. We encounter natural properties (through their causal action), and these give rise to the similarities.

	p.18	21961	Physics aims to discover which universals actually exist
			Full Idea: For Lewis, we can see the purpose of physics as being to discover what universals there actually are.
			From: report of David Lewis (New work for a theory of universals [1983]) by A.W. Moore - The Evolution of Modern Metaphysics Intro
			A reaction: It seems that Lewis uses the word 'property' to mean predicates, which consist of a multitude of sets, while universals are the properties that naturally exist and cut nature at the joints . Infuriating, because the other way around seems better.

	p.29	14499	Properties are classes of possible and actual concrete particulars
			Full Idea: Lewis has a preference for a nominalist conception of properties as classes of possible and actual concrete particulars.
			From: report of David Lewis (New work for a theory of universals [1983]) by Kathrin Koslicki - The Structure of Objects II.3
			A reaction: I'm sympathetic to nominalism, but still can't swallow the idea that a property like redness is nothing more than a collection of particulars, the red things. This class will include all sorts of non-red features.

	p.98	16217	Lewisian natural properties fix reference of predicates, through a principle of charity
			Full Idea: For Lewis natural properties are important for their role in making language and thought determinate: principles of charity or humanity tell us to attribute natural properties to predicates wherever possible, break underdetermination of their reference.
			From: report of David Lewis (New work for a theory of universals [1983]) by Katherine Hawley - How Things Persist 3.8
			A reaction: Lewis always seems to find reasons in semantics or logic for his metaphysics, instead of in the science. Lewis ends up with 'folk' natural properties, instead of accurate ones.

	p.128	7031	Lewis says properties are sets of actual and possible objects
			Full Idea: David Lewis has produced an important theory of properties as sets of actual and possible objects.
			From: report of David Lewis (New work for a theory of universals [1983]) by John Heil - From an Ontological Point of View §12.2
			A reaction: The notion that a property is an 'object' sounds wrong, as it is too passive. It also seems to allow for the possibility of uninstantiated properties existing, where properties are presumably always 'of' something.

	p.211	15120	Lewisian properties have powers because of their relationships to other properties
			Full Idea: According to Lewis's conception, the causal powers of a property are constituted by its patterned relations to other properties in the particular Humean mosaic that is the actual world.
			From: report of David Lewis (New work for a theory of universals [1983]) by John Hawthorne - Causal Structuralism Intro
			A reaction: I just can't grasp this as a serious proposal. Relations cannot be the bottom line in explanation of the world. What are the relata? I take powers to be primitive.

'1 Ov Many'	p.198	8605	In addition to analysis of a concept, one can deny it, or accept it as primitive
			Full Idea: There are three ways to give an account: 1) 'I deny it' - this earns a failing mark if the fact is really Moorean. 2) 'I analyse it thus'. 3) 'I accept it as primitive'. Not every account is an analysis.
			From: David Lewis (New work for a theory of universals [1983], '1 Ov Many')
			A reaction: I prefer Shoemaker's view (Idea 8559). Personally I think 1) should be employed more often than it is (it is a very misunderstood approach). 3) has been overused in recent years (e.g. by Davidson and McGinn).

'1 Ov Many'	p.199	8576	The One over Many problem (in predication terms) deserves to be neglected (by ostriches)
			Full Idea: The transformed problem of One over Many (in terms of predication, rather than sameness of type) deserves our neglect. The ostrich that will not look at it is a wise bird indeed.
			From: David Lewis (New work for a theory of universals [1983], '1 Ov Many')
			A reaction: This is aimed at Armstrong, and defends Quine. The remark moves Ostrich Nominalism from the category of joke to the category of respectable. I think I side with Armstrong. How is predication primitive if it has two components?

'Cont of L'	p.220	8585	Reference partly concerns thought and language, partly eligibility of referent by natural properties
			Full Idea: Reference consists in part of what we do in language or thought when we refer, but in part it consists in eligibility of the referent. And this eligibility to be referred to is a matter of natural properties.
			From: David Lewis (New work for a theory of universals [1983], 'Cont of L')
			A reaction: This is a surprising conclusion for Lewis to reach, having started from properties as any old set members (see Idea 8572). There are references to intentional objects, such as 'there should have been someone on duty'.

'Cont of L'	p.220	8613	Objects are demarcated by density and chemistry, and natural properties belong in what is well demarcated
			Full Idea: Where my cat (Bruce) ends, there the density of matter, the relative abundance of chemical elements, abruptly change. Bruce is also a locus of causal chains, which traces back to natural properties. Natural properties belong to well demarcated things.
			From: David Lewis (New work for a theory of universals [1983], 'Cont of L')
			A reaction: This is an amazingly convoluted way to define natural properties in terms of the classes they generate, but it seems obvious to me that the properties are logically prior to the classes.

'Cont of L'	p.221	8586	Natural properties tend to belong to well-demarcated things, typically loci of causal chains
			Full Idea: One thing that makes for naturalness of a property is that it is a property belonging exclusively to well-demarcated things (like my cat Bruce, who is a locus of causal chains).
			From: David Lewis (New work for a theory of universals [1983], 'Cont of L')
			A reaction: Compare Idea 8557. Well-demarcated things may also have gerrymandered properties that are parts of 'arbitrary Boolean compounds' (Lewis). Why not make use of the causal chains to identify the properties?

'Cont of L'	p.224	8615	We need natural properties in order to motivate the principle of charity
			Full Idea: We need natural properties, so that the principle of charity will impute a bias towards believing that things are green rather than grue, and towards a basic desire for long life, rather than long-life-unless-one-was-born-on-a-Monday....
			From: David Lewis (New work for a theory of universals [1983], 'Cont of L')
			A reaction: Lewis always seems to be approaching things from the wrong end. We don't need properties so that we can attribute charity, so that we can interpret. We interpret, because we can be charitable, because we all experience natural properties.

'Cont of L'	p.224	8614	A sophisticated principle of charity sometimes imputes error as well as truth
			Full Idea: Unlike principles of crude charity, sophisticated principles of charity call for imputations of error in the subject if he has lived in deceptive conditions.
			From: David Lewis (New work for a theory of universals [1983], 'Cont of L')
			A reaction: This begs lots of questions about how you decide conditions are 'deceptive' if you have not yet embarked on your radical interpretation of the subject. Davidson's point still stands, that imputing truth must be the normal procedure.

'Cont of L'	p.227	8589	For us, a property being natural is just an aspect of its featuring in the contents of our attitudes
			Full Idea: The reason natural properties feature in the contents of our attitudes is that naturalness is part of what it is to feature therein. We aren't built to take a special interest in natural properties, or that we call them natural if they are interesting.
			From: David Lewis (New work for a theory of universals [1983], 'Cont of L')
			A reaction: Evolution never features in Lewis's metaphysics. I would have thought we were very much built to focus on natural properties. This sounds odd, and gives no help in distinguishing natural properties from all our other daft contents.

'Dup,Sup,Div'	p.203	15727	Physics aims for a list of natural properties
			Full Idea: Physics aspires to give an inventory of natural properties.
			From: David Lewis (New work for a theory of universals [1983], 'Dup,Sup,Div')
			A reaction: The sort of beautifully simple remark by which philosophers ought to earn a good living in the intellectual community. Come on physicists - this is all we want! Presumably the inventory will include an account of how they all work.

'Dup,Sup,Div'	p.205	8607	Supervenience is reduction without existence denials, ontological priorities, or translatability
			Full Idea: Supervenience is a stripped down form of reductionism, unencumbered by dubious denials of existence, claims of ontological priority, or claims of translatability.
			From: David Lewis (New work for a theory of universals [1983], 'Dup,Sup,Div')
			A reaction: Interesting. It implies that the honest reductionist (i.e. me) should begin by asserting supervience, and only at a second stage go on to deny a bit of existence, loudly affirm priorities, and offer translations. Honest toil.

'Dup,Sup,Div'	p.205	8606	A supervenience thesis is a denial of independent variation
			Full Idea: A supervenience thesis is a denial of independent variation.
			From: David Lewis (New work for a theory of universals [1983], 'Dup,Sup,Div')
			A reaction: Not everyone agrees on this. This says if either A or B change, the change is reflected in the other one. But the other view is of one-way dependence. A only changes if B changes, but B can also make changes that don't affect A.

'Dup,Sup,Div'	p.208	8608	Counterfactuals 'backtrack' if a different present implies a different past
			Full Idea: A counterfactual can be said to 'backtrack' if it can be said that if the present were different a different past would have led up to it (rather than if the present were different, the same past would have had a different outcome).
			From: David Lewis (New work for a theory of universals [1983], 'Dup,Sup,Div')
			A reaction: A nice clear definition of a concept which is important in Lewis's analysis of causation. In the current context he is concerned with elucidation of determinism and materialism. I would say (intuitively) that all counterfactuals backtrack.

'Intro'	p.188	8569	I suspend judgements about universals, but their work must be done
			Full Idea: I suspend judgement about universals themselves; I only insist that, one way or another, their work must be done.
			From: David Lewis (New work for a theory of universals [1983], 'Intro')
			A reaction: This seems surprising (but admirable) in a great metaphysician, but I suppose it is symptomatic of the Humean approach to metaphysics. In the light of Ideas 3989 and 3990, I would have expected Lewis to deny universals. He probably did.

'Laws and C'	p.215	8611	A law of nature is any regularity that earns inclusion in the ideal system
			Full Idea: A law of nature is any regularity that earns inclusion in the ideal system (or, in case of ties, in every ideal system).
			From: David Lewis (New work for a theory of universals [1983], 'Laws and C')
			A reaction: Reminiscent of Peirce's view of truth (Idea 7661). This wouldn't seem to eliminate the danger of regularities with underlying causes ending up as laws (day causes night). Or very trivial regularities ending up as laws.

'Laws and C'	p.217	8584	Causal counterfactuals must avoid backtracking, to avoid epiphenomena and preemption
			Full Idea: My counterfactual analysis of causation needs counterfactuals that avoid backtracking; else the analysis faces fatal counterexamples involving epiphenomenal side-effects or cases of causal preemption.
			From: David Lewis (New work for a theory of universals [1983], 'Laws and C')
			A reaction: The concept of true epiphenomena (absolutely no causal powers) strikes me as bogus.

'Min Mat'	p.210	8579	Psychophysical identity implies the possibility of idealism or panpsychism
			Full Idea: Psychophysical identity is a two-way street: if all mental properties are physical, then some physical properties are mental; but then all physical properties might be mental, or every property of everything might be both physical and mental.
			From: David Lewis (New work for a theory of universals [1983], 'Min Mat')
			A reaction: I suspect that this is the thought that has impressed Galen Strawson. The whole story seems to include the existence of 'mental properties' as a distinct category. This line of thought strikes me as a serious misunderstanding.

'Min Mat'	p.212	8580	Materialism is (roughly) that two worlds cannot differ without differing physically
			Full Idea: Final definition of 'Materialism': Among worlds where no natural properties alien to our world are instantiated, no two differ without differing physically; and two such worlds that are exactly alike physically are duplicates.
			From: David Lewis (New work for a theory of universals [1983], 'Min Mat')
			A reaction: This would presumably allow for an anomalous monist/property dualist view of mind, but not full dualism. But if there are no psychophysical laws, what stops the mental changing while the physical remains the same?

'Min Mat'	p.212	8581	Physics discovers laws and causal explanations, and also the natural properties required
			Full Idea: Physics must not just discover laws and causal explanations. In putting forward as comprehensive theories that recognise only a limited range of natural properties, physics proposes inventories of the natural properties instantiated in our world.
			From: David Lewis (New work for a theory of universals [1983], 'Min Mat')
			A reaction: Physics does this job extremely well, offering things like force, spin, charge that are the building blocks for their theories. There is metaphysics at the heart of physics, unavoidably.

p.355-7	p.130	15460	All perfectly natural properties are intrinsic
			Full Idea: Lewis proposed that all perfectly natural properties are intrinsic.
			From: report of David Lewis (New work for a theory of universals [1983], p.355-7) by David Lewis - Defining 'Intrinsic' (with Rae Langton) IX
			A reaction: Depends what you mean by 'natural', 'property' and 'intrinsic'! Presumably there are natural extrinsic facts, in naturally necessary relationships. If all natural properties are powers, they would have to be intrinsic. Extrinsics would be derivative.

'Un and Prop'	p.189	8570	To have a property is to be a member of a class, usually a class of things
			Full Idea: To have a property is to be a member of a class, usually a class of things. (Note: this resembles the doctrine of Class Nominalism, but I do not claim to solve the One Over Many problem by this means, far from it).
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: Lewis remains neutral about the traditional question of whether universals exist. What does he mean by "is" in his assertion? Identity, predication or class membership? I think Lewis is open to many of the objections to Class Nominalism.

'Un and Prop'	p.190	8571	Universals are wholly present in their instances, whereas properties are spread around
			Full Idea: Universals and properties are different because a universal is supposed to be wholly present wherever it is instantiated. A property, by contrast, is spread around. The property of being a donkey is partly present wherever there is a donkey.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: No mention of tropes. The claim that universals are widespread, and yet must be instantiated, is dealt with by Lewis's commitment to the existence of possible donkeys.

'Un and Prop'	p.191	8572	Any class of things is a property, no matter how whimsical or irrelevant
			Full Idea: Any class of things, be it ever so gerrymandered and miscellaneous and indescribable in thought and language, and be it ever so superfluous in characterizing the world, is nevertheless a property.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: I much prefer, at the very least, the sparse approach of Armstrong, and in fact would vote for Shoemaker's highly physical view. Lewis proceeds after this to try to pick out the properties that really matter.

'Un and Prop'	p.192	15726	Natural properties fix resemblance and powers, and are picked out by universals
			Full Idea: Perhaps we could call a property 'perfectly' natural if its members are all and only those things that share some one universal, ...where the natural properties would be the ones whose sharing makes for resemblance, and the ones relevant to causal powers.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: This is Lewis fishing for an account of properties that does a bit better than the mere recourse to set theory (which he intuitively favours) seems to do. He remains neutral about the ontological status of a universal (though he prefers nominalism).

'Un and Prop'	p.192	18433	There are far more properties than any brain could ever encodify
			Full Idea: There are so many properties that those specifiable in English, or in the brain's language of synaptic interconnections and neural spikes, could only be an infinitesimal minority.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: Thus there are innumerable properties that must lack predicates. But there are also innumerable predicates that correspond to no real properties. I conclude that properties and predicates have very little in common. Job done.

'Un and Prop'	p.192	8573	Most properties are causally irrelevant, and we can't spot the relevant ones.
			Full Idea: Properties do nothing to capture the causal powers of things. Almost all properties are causally irrelevant, and there is nothing to make the relevant ones stand out from the crowd.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: Shoemaker, who endorses a causal account of properties, has a go at this problem in Idea 8557. The property of being massive is more likely to be causal than existing fifty years after D-Day. Lewis attempts later to address the problem.

'Un and Prop'	p.194	8604	We need properties as semantic values for linguistic expressions
			Full Idea: We need properties, sometimes natural and sometimes not, to provide an adequate supply of semantic values for linguistic expressions.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop')
			A reaction: A characteristically twentieth century approach, which I find puzzling. We don't need a Loch Ness Monster in order to use the term 'Loch Ness Monster'. Lewis appears to have been a pupil of Quine... He was not, though, a Predicate Nominalist.

'Un and Prop' n9	p.194	8574	Class Nominalism and Resemblance Nominalism are pretty much the same
			Full Idea: Moderate Class Nominalism and Resemblance Nominalism (in its present form) seem to me to be a single theory presented in different styles.
			From: David Lewis (New work for a theory of universals [1983], 'Un and Prop' n9)
			A reaction: Lewis has earlier endorsed a cautious form of Class Nominalism (Idea 8570). Which comes first, having a resemblance, or being in a class? Quine seems to make resemblance basic (Idea 8486), but Lewis seems to make the class basic (Idea 8572).

1984	Putnam's Paradox

'Glob Desc'	p.60	14209	Descriptive theories remain part of the theory of reference (with seven mild modifications)
			Full Idea: Description theories of reference are supposed to have been well and truly refuted. I think not: ..it is still tenable with my seven points, and part of the truth of reference [7: rigidity, egocentric, tokens, causal, imperfect, indeterminate, families].
			From: David Lewis (Putnam's Paradox [1984], 'Glob Desc')
			A reaction: (The bit at the end refers to his seven points, on p.59). He calls his basic proposal 'causal descriptivism', incorporating his seven slight modifications of traditional descriptivism about reference.

'What Is'	p.77	14215	Causal theories of reference make errors in reference easy
			Full Idea: Whatever happens in special cases, causal theories usually make it easy to be wrong about the thing we refer to.
			From: David Lewis (Putnam's Paradox [1984], 'What Is')
			A reaction: I suppose the point of this is that there are no checks and balances to keep reference in focus, but just a requirement to keep connected to an increasingly attenuated causal chain.

'What Might'	p.65	14210	A gerrymandered mereological sum can be a mess, but still have natural joints
			Full Idea: The mereological sum of the coffee in my cup, the ink in this sentence, a nearby sparrow, and my left shoe is a miscellaneous mess of an object, yet its boundaries are by no means unrelated to the joints of nature.
			From: David Lewis (Putnam's Paradox [1984], 'What Might')
			A reaction: In that case they do, but if there are no atoms at the root of physics then presumably their could also be thoroughly jointless assemblages, involving probability distributions etc. Even random scattered atoms seem rather short of joints.

'Why Anti-R'	p.71	14213	Anti-realists see the world as imaginary, or lacking joints, or beyond reference, or beyond truth
			Full Idea: Anti-realists say the only world is imaginary, or only has the parts or classes or relations we divide it into, or doubt that reference to the world is possible, or doubt that our interpretations can achieve truth.
			From: David Lewis (Putnam's Paradox [1984], 'Why Anti-R')
			A reaction: [compression of a paragraph on anti-realism] Lewis is a thoroughgoing realist. A nice example of the rhetorical device of ridiculing an opponent by suggesting that they don't even know what they themselves believe.

'Why Model'	p.68	14212	A consistent theory just needs one model; isomorphic versions will do too, and large domains provide those
			Full Idea: A consistent theory is, by definition, one satisfied by some model; an isomorphic image of a model satisfies the same theories as the original model; to provide the making of an isomorphic image of any given model, a domain need only be large enough.
			From: David Lewis (Putnam's Paradox [1984], 'Why Model')
			A reaction: This is laying out the ground for Putnam's model theory argument in favour of anti-realism. If you are chasing the one true model of reality, then formal model theory doesn't seem to offer much encouragement.

1986	Causal Explanation

	p.237	4809	Lewis endorses the thesis that all explanation of singular events is causal explanation
			Full Idea: Lewis endorses the thesis that all explanation of singular events is causal explanation.
			From: report of David Lewis (Causal Explanation [1986]) by Stathis Psillos - Causation and Explanation p.237
			A reaction: It is hard to challenge this. The assumption is that only nomological and causal explanations are possible, and the former are unobtainable for singular events.

I	p.215	15552	We only pick 'the' cause for the purposes of some particular enquiry.
			Full Idea: Disagreement about 'the' cause is only disagreement about which part of the causal history is most salient for the purposes of some particular inquiry.
			From: David Lewis (Causal Explanation [1986], I)
			A reaction: I don't believe this. In the majority of cases I see the cause of an event, without having any interest in any particular enquiry. It is just so obvious that there isn't even a disagreement. Maybe there is only one sensible enquiry.

I	p.215	15551	Ways of carving causes may be natural, but never 'right'
			Full Idea: There is no one right way - though there may be more or less natural ways - of carving up a causal history.
			From: David Lewis (Causal Explanation [1986], I)
			A reaction: This invites a distinction between the 'natural' causes and the 'real' causes. Presumably if any causes were 'real', they would have a better claim to be 'right'. Is an earthquake the 'real' (correct?) cause of a tsunami?

I	p.216	15553	Causal dependence is counterfactual dependence between events
			Full Idea: I take causal dependence to be counterfactual dependence, of a suitably back-tracking sort, between distinct events.
			From: David Lewis (Causal Explanation [1986], I)
			A reaction: He quotes Hume in support. 'Counterfactual dependence' strikes me as too vague, or merely descriptive, for the job of explanation. 'If...then' is a logical relationship; what is it in nature that justifies the dependency?

II	p.217	14321	To explain an event is to provide some information about its causal history
			Full Idea: Here is my main thesis: to explain an event is to provide some information about its causal history.
			From: David Lewis (Causal Explanation [1986], II)
			A reaction: The obvious thought is that you might provide some tiny and barely relevant part of that causal history, such as a bird perched on the Titanic's iceberg. So how do we distinguish the 'important' causal information?

III	p.223	15554	A disposition needs a causal basis, a property in a certain causal role. Could the disposition be the property?
			Full Idea: I take for granted that a disposition requires a causal basis: one has the disposition iff one has a property that occupies a certain causal role. Shall we then identify the disposition with its basis? That makes the disposition cause its manifestations.
			From: David Lewis (Causal Explanation [1986], III)
			A reaction: Introduce the concept of a 'power' and I see no problem with his proposal. Fundamental dispositions are powerful, and provide the causal basis for complex dispositions. Something had better be powerful.

IV	p.225	15556	Science may well pursue generalised explanation, rather than laws
			Full Idea: The pursuit of general explanations may be very much more widespread in science than the pursuit of general laws.
			From: David Lewis (Causal Explanation [1986], IV)
			A reaction: Nice. I increasingly think that the main target of all enquiry is ever-widening generality, with no need to aspire to universality.

IV	p.225	15555	Explaining match lighting in general is like explaining one lighting of a match
			Full Idea: Explaining why struck matches light in general is not so very different from explaining why some particular struck match lit. ...We may generalize modestly, without laying claim to universality.
			From: David Lewis (Causal Explanation [1986], IV)
			A reaction: A suggestive remark, since particular causation and general causation seem far apart, but Lewis suggests that the needs of explanation bring them together. Lawlike and unlawlike explanations?

V	p.228	15558	A good explanation is supposed to show that the event had to happen
			Full Idea: It is said that a good explanation ought to show that the explanandum event had to happen, given the laws and circumstances.
			From: David Lewis (Causal Explanation [1986], V)
			A reaction: I cautiously go along with this view. Given that there are necessities in nature (a long story), we should aim to reveal them. There is no higher aspiration open to us than successful explanation. Lewis says good explanations can reveal falsehoods.

V	p.228	15559	Does a good explanation produce understanding? That claim is just empty
			Full Idea: It is said that a good explanation ought to produce understanding, ...but this just says that a good explanation produces possession of that which it provide, so this desideratum is empty. It adds nothing to our understanding of explanation.
			From: David Lewis (Causal Explanation [1986], V)
			A reaction: I am not convinced by this dismissal. If you are looking for a test of whether an explanation is good, the announcement that the participants feel they have achieved a good understanding sounds like success.

V n7	p.226	15557	Verisimilitude has proved hard to analyse, and seems to have several components
			Full Idea: The analysis of verisimilitude has been much debated. Some plausible analyses have failed disastrously, others conflict with one another. One conclusion is that verisimilitude seems to consist of several distinguishable virtues.
			From: David Lewis (Causal Explanation [1986], V n7)
			A reaction: Presumably if it is complex, you can approach truth in one respect while receding from it in another. It seems clear enough if you are calculating pi by some iterative process.

VI	p.230	15560	We can explain a chance event, but can never show why some other outcome did not occur
			Full Idea: I think we are right to explain chance events, yet we are right also to deny that we can ever explain why a chance process yields one outcome rather than another. We cannot explain why one event happened rather than the other.
			From: David Lewis (Causal Explanation [1986], VI)
			A reaction: This misses out an investigation which slowly reveals that a 'chance' event wasn't so chancey after all. Failure to explain confirms chance, so the judgement of chance shouldn't block attempts to explain.

1986

Events

I	p.241	15561	The events that suit semantics may not be the events that suit causation
			Full Idea: There is no guarantee that events made for semantics are the same as events that are causes and effects.
			From: David Lewis (Events [1986], I)
			A reaction: This little cri de couer could be a motto for a huge amount of analytic philosophy, which (for some odd reason) thought that mathematics, logic, set theory and formal semantics were good tools for explaining nature.

I	p.242	15562	Causation is a general relation derived from instances of causal dependence
			Full Idea: Causation is the ancestral of causal dependence: event c causes event e iff either e depends on c, or e depends on an intermediate event which in turn depends on c, or....
			From: David Lewis (Events [1986], I)
			A reaction: This is Lewis making sure that we don't postulate some huge bogus thing called 'Causation' which is supposed to be in charge of Nature. Good point.

II	p.245	15564	An event is a property of a unique space-time region
			Full Idea: I propose to identify an event with a property, or in other words with a class, a unique spatio-temporal region corresponding to where that event occurs.
			From: David Lewis (Events [1986], II)
			A reaction: [I've run together two separate bits, on p.244 and 245] Lewis cites Montague's similar view, that events are properties of times.

II n2	p.244	15563	Properties are very abundant (unlike universals), and are used for semantics and higher-order variables
			Full Idea: Properties are abundant, numbering at least beth-3 for properties of individuals alone; they are suited to serve as semantic values of arbitrarily complex predicates and gerunds, and higher-order variables. (If there are universals, they are sparse).
			From: David Lewis (Events [1986], II n2)
			A reaction: To me this is an outrageous hijacking of the notion of property which is needed for explaining the natural world. He seems to be talking about predicates. He wants to leave me with his silly universals - well I don't want them, thank you.

III	p.247	15565	Events have inbuilt essences, as necessary conditions for their occurrence
			Full Idea: Events have their essences built in, in the form of necessary conditions for their occurrence.
			From: David Lewis (Events [1986], III)
			A reaction: Revealing. He thinks the essence of an event is something which precedes the event. I take it as obvious that if an event has an essence, it will be some features of the event that occur in it and during it. They need to be intrinsic.

V	p.258	15566	Events are classes, and so there is a mereology of their parts
			Full Idea: If events are classes, as I propose, then they have a mereology in the way that all classes do: the parts of a class are its subclasses.
			From: David Lewis (Events [1986], V)
			A reaction: Lewis says events are properties, which he regards as classes. It is not clear that events are strictly mereological. Could one happening be two events? Is WWII a simple sum of its parts? [see p.260]

VI	p.261	15567	Some events involve no change; they must, because causal histories involve unchanges
			Full Idea: Not all events involve change. We cannot afford to count the unchanges as nonevents, for the unchanges may be needed to complete causal histories.
			From: David Lewis (Events [1986], VI)
			A reaction: You end up calling non-changes 'events' if you commit to a simplistic theory that all causal histories consist of events. Why not allow conditions as well as events? Lewis concedes that he may be abusing language.

1986	Introduction to Philosophical Papers II

	p.-10	9426	The world is just a vast mosaic of little matters of local particular fact
			Full Idea: The world is a vast mosaic of local matters of particular fact, just one little thing and then another.
			From: David Lewis (Introduction to Philosophical Papers II [1986])
			A reaction: Basing laws on this picture is what Lewis calls 'Humean Supervenience'.

p.ix-x	p.-10	16210	Humean supervenience says the world is just a vast mosaic of qualities in space-time
			Full Idea: Humean supervenience says the world is a vast mosaic of local matters of particular fact. We have a geometry of external relations of spatio-temporal distance between points, and local qualities at points. …In short: we have an arrangement of qualities.
			From: David Lewis (Introduction to Philosophical Papers II [1986], p.ix-x)
			A reaction: [compressed] This is the key fundamental tenet of David Lewis's philosophy. He names it after Hume because it contains no necessary connections. It is 'supervenient' because all worldly truths reduce to and depend on the mosaic. His thesis is contingent.

1986	Comment on Armstrong and Forrest

p.109	p.109	15452	We could not uphold a truthmaker for 'Fa' without structures
			Full Idea: We could not, without structures, uphold the principle that every truth has a truthmaker. If Fa is true, the truthmaker is not F, not a, nor both together; not their mereological sum; not a set-theoretic construction. These would exist just the same.
			From: David Lewis (Comment on Armstrong and Forrest [1986], p.109)
			A reaction: This point ought to trouble Lewis, as well as Armstrong and Forrest. If we assert 'Fa', we must (in any theory) have some idea of what unites them, as well as of their separate existence. It must a fact about 'a', not a fact about 'F'.

p.110	p.110	15453	The main rivals to universals are resemblance or natural-class nominalism, or sparse trope theory
			Full Idea: The leading rivals to a theory of universals are resemblance or natural-class nominalism, or sparse trope theory.
			From: David Lewis (Comment on Armstrong and Forrest [1986], p.110)
			A reaction: If that is the complete menu, I choose resemblance nominalism. All discussion of properties in terms of classes is wildly misguided (because properties come first). Why not 'natural' tropes?

1986	On the Plurality of Worlds

	p.7	10470	There are only two kinds: sets, and possibilia (actual and possible particulars)
			Full Idea: Lewis's multi-purpose ontology seems to have only two kinds: sets and possibilia (actual and possible particulars).
			From: report of David Lewis (On the Plurality of Worlds [1986]) by Alex Oliver - The Metaphysics of Properties 3
			A reaction: This is awfully like the ontology of his teacher Quine, but with the wicked addition of modal properties. It is no wonder that Lewis was a bit vague about the concrete boundary, as both of his kinds seem to be abstract. His Achilles' Heel?

	p.7	16441	Lewis rejects actualism because he identifies properties with sets
			Full Idea: It is the identification of properties with sets that rules out, for Lewis, an actualist account of possible worlds.
			From: report of David Lewis (On the Plurality of Worlds [1986]) by Robert C. Stalnaker - Mere Possibilities 1.1
			A reaction: I suppose the sets which are the properties have to include all the possible red things as well as the actual one. This escapes the renate/cordate problem.

	p.79	7690	If sets exist, then defining worlds as proposition sets implies an odd distinction between existing and actual
			Full Idea: If sets exist, then the conventional concept of a logically possible world as a proposition set requires a counterintuitive distinction between existence and actuality, between what exists and what is actual.
			From: comment on David Lewis (On the Plurality of Worlds [1986]) by Dale Jacquette - Ontology Ch.2
			A reaction: This pinpoints the obvious difficulty that most people have with Lewis's claim that possible worlds exist. Russell's claim that universals 'subsist' (Idea 5409) is a similar attempt to have two different sorts of existence in your ontology.

	p.112	14404	The counterpart relation is sortal-relative, so objects need not be a certain way
			Full Idea: Lewis takes the counterpart relation to be sortal-relative, so he (no less than Quine) denies that objects, qua existing, are necessarily a certain way.
			From: report of David Lewis (On the Plurality of Worlds [1986]) by Trenton Merricks - Truth and Ontology 5.III n10
			A reaction: Does this mean that there could be two different versions of the same possible world (certainly not!), or that worlds are entirely created by our concepts rather than by what is actually possible.

	p.115	5441	Why should statements about what my 'counterpart' could have done interest me?
			Full Idea: If I only have counterparts in possible worlds who are not identical to me, statements about what I could have done will seem irrelevant to me, because they will be about someone else.
			From: comment on David Lewis (On the Plurality of Worlds [1986]) by Thomas Mautner - Penguin Dictionary of Philosophy p.115
			A reaction: We might rephrase the statement as "I could have been the person who did x". Presumably my counterpart is not just any stranger, but someone I could have been. "I could have been a brick" - now that seems irrelevant to me!

	p.115	5440	A counterpart in a possible world is sufficiently similar, and more similar than anything else
			Full Idea: 'Jack could have been taller' implies a different Jack in a different world, so Lewis defines a counterpart in a possible world as an individual sufficiently similar to Jack, and more similar to Jack than anything else in that world.
			From: report of David Lewis (On the Plurality of Worlds [1986]) by Thomas Mautner - Penguin Dictionary of Philosophy p.115
			A reaction: If we say something like "I could have been twins" or "I could have been a genius" in another world, it would need an odd concept of my personal identity for it to remain identical in those counterfactual situations. Lewis has a point.

	p.127	12255	For Lewis there is no real possibility, since all possibilities are actual
			Full Idea: Lewis-style modal realism eliminates all real possibility since on his account everything is actual relative to its own world.
			From: comment on David Lewis (On the Plurality of Worlds [1986]) by David S. Oderberg - Real Essentialism 6.2
			A reaction: Since it is possible for me to be in New York and in Chicago, but not both at once, his possibilities have to be kept apart, even though they are actual. I expect my visit to Chicago to remain as only a possibility.

	p.194	9219	Lewis posits possible worlds just as Quine says that physics needs numbers and sets
			Full Idea: Lewis's argument for possible worlds parallels Quine's for the existence of sets: our best overall empirical theory, mathematical physics, quantifies over real numbers, so we have reason to posit real numbers, or the sets to which they may be reduced.
			From: report of David Lewis (On the Plurality of Worlds [1986]) by Theodore Sider - Reductive Theories of Modality 3.6
			A reaction: They both strike me as suspect. Indeed, the extreme implausibility of Lewis's conclusion throws doubt on Quine's original strategy. I'm happy to work with sets and possible worlds, and only worry about ontological commitment at a later stage.

	p.248	15022	If possible worlds really exist, then they are part of actuality
			Full Idea: The familiar complaint against Lewis is that if his worlds existed, they'd be part of actuality.
			From: comment on David Lewis (On the Plurality of Worlds [1986]) by Theodore Sider - Writing the Book of the World 11.5
			A reaction: Sider presents that as rather superficial, but it sounds a pretty good objection to me. Lewis would note that only our world has the indexical features which he says pick out actuality. Real possible worlds might lack indexical features?

§1.5	p.269	15399	The property of being F is identical with the set of objects, in all possible worlds, which are F
			Full Idea: Lewis thought that the property of being F was identical with the set of objects, in all possible worlds, which are F.
			From: report of David Lewis (On the Plurality of Worlds [1986], §1.5) by Ross P. Cameron - Intrinsic and Extrinsic Properties
			A reaction: I can't make head or tail of a theory which says that a property is a set of objects. I'll show you a room full of objects and tell you they are a property. How are you going to work out what the property is? 'Being F' is a predicate, not a property!

1.2	p.15	9650	Supervenience concerns whether things could differ, so it is a modal notion
			Full Idea: The idea of supervenience is when there could be no difference of one sort without difference of another sort. ..Clearly this 'could' indicates modality, and without modality we have nothing of interest.
			From: David Lewis (On the Plurality of Worlds [1986], 1.2)
			A reaction: This might explain why philosophers are going to be more at home with the concept than neuroscientists would be.

1.2 n3	p.7	16132	On mountains or in worlds, reporting contradictions is contradictory, so no such truths can be reported
			Full Idea: To tell the alleged truth about contradictory things that happen on a mountain is just contradicting yourself, but you can't tell the truth by contradicting yourself. There is no mountain where contradictions are true, and impossible worlds are no better.
			From: David Lewis (On the Plurality of Worlds [1986], 1.2 n3)
			A reaction: [compressed] He says this works for any 'restricted' domain like a mountain or a real world, but that it wouldn't apply in an unrestricted fictional world.

1.2 n3	p.7	16133	Possible worlds can contain contradictions if such worlds are seen as fictions
			Full Idea: If worlds were like stories or story-tellers, there would indeed be room for worlds according to which contradictions are true.
			From: David Lewis (On the Plurality of Worlds [1986], 1.2 n3)
			A reaction: Most existing fictions contain tiny contradictions, but we might ask whether that thereby disqualifies them from depicting genuinely 'possible' worlds.

1.3	p.24	9651	Verisimilitude might be explained as being close to the possible world where the truth is exact
			Full Idea: We might explain the closeness to the truth (or 'verisimilitude') in terms of closeness of possible worlds. A theory is close to the truth to the extent that our world resembles some world where that theory is exactly true.
			From: David Lewis (On the Plurality of Worlds [1986], 1.3)
			A reaction: [Lewis cites Risto Hilpinen for this thought] I am always puzzled why Lewis and co. talk of whole worlds in their accounts. If I am close to the truth about cooking a good omelette, what has the rest of the world got to do with it?

1.4	p.38	9652	To just expect unexamined emeralds to be grue would be totally unreasonable
			Full Idea: Some beliefs are just unreasonable in a strong sense. Think of the man who, for no special reason, expects unexamined emeralds to be grue.
			From: David Lewis (On the Plurality of Worlds [1986], 1.4)
			A reaction: This is a nice converse way of seeing the point that 'grue' is such an totally artificial predicate. I still say that the most illuminating point is that grue is not a colour, so seeing a grue thing is no confirmation at all.

1.5	p.22	10723	A property is the set of its actual and possible instances
			Full Idea: Lewis proposes that a property is the set of its actual and possible instances.
			From: report of David Lewis (On the Plurality of Worlds [1986], 1.5) by Alex Oliver - The Metaphysics of Properties 10
			A reaction: I just can't make sense of any proposal that a property is a set. Things fall into natural sets because they have properties. Only a philosopher would believe such a weird proposal as this one. Triangular and trilateral?

1.5	p.50	9653	It would be easiest to take a property as the set of its instances
			Full Idea: The simplest plan for properties is to take a property just as the set of all its instances.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: I find this a weird and counterintuitive proposal. I suppose if you think maths has been reduced to set theory, you might want to reduce everything else, and then we can all go home. I thought things were in sets because of their properties.

1.5	p.51	15733	Accidentally coextensive properties come apart when we include their possible instances
			Full Idea: In modal realism, 'accidentally coextensive' properties are not coextensive. They only appear so when we ignore their other-worldly instances. If we consider all instances, then it never can happen that two properties are coextensive but might have been.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: It is not clear why this 'can never happen'. Maybe even God can't make a hearted creature without kidneys. Lewis is aware of this question.

1.5	p.51	15732	Properties don't seem to be sets, because different properties can have the same set
			Full Idea: The usual objection to taking properties as sets is that different properties may happen to be coextensive. Creatures with hearts are creatures with kidneys. Talking donkeys are flying pigs (since there are none). Yet the properties differ.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: This is the difficulty which Lewis proposes to solve by defining properties across possible situations as well as actual ones, so that the properties can come apart. Nice move.

1.5	p.52	15734	If a property is relative, such as being a father or son, then set membership seems relative too
			Full Idea: A property that is instantiated in a relative way (such as being a father or a son) could not be the set of its instances? Is the thing to be included in the set or not?
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: He says philosophers contrive ways to define properties as functions, but he prefers to call such properties 'relations', and define them that way. It never even occurred to me that 'being a son' was one of my properties, but what do I know?

1.5	p.53	15735	Properties don't have degree; they are determinate, and things have varying relations to them
			Full Idea: I have made no place for properties that admit of degree, so that things may have more or less of the same property. There are plain properties, and then there are relations to them.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: An interesting question, little discussed. Elsewhere, Lewis ascribes all vagueness to our inadequate predicates, rather than to the world, which I find quite persuasive.

1.5	p.53	9654	A proposition is a set of entire possible worlds which instantiate a particular property
			Full Idea: I identify propositions with certain properties - namely, with those that are instantiated only by entire possible worlds. Then if properties generally are the sets of their instances, a proposition is a set of possible worlds.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: I don't get this. How can the proposition that tomatoes are edible be an entire set of possible worlds? The proposition seems to be about tomatoes, and nothing else. Should we talk of 'possibilities', rather than of 'possible worlds'?

1.5	p.53	15736	A proposition is the property of being a possible world where it holds true
			Full Idea: I identify propositions with properties that are instantiated only by entire possible worlds. If properties are the sets of their instances, a proposition is a set of possible worlds. A proposition is the property of being a world where it holds.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: This is so far away from my concept of a proposition (as a truth-evaluable representational mental event) that I struggle to compute it. So the proposition that I am sitting here is the property of 'being the actual world'. Eh?

1.5	p.55	9655	Trilateral and triangular seem to be coextensive sets in all possible worlds
			Full Idea: We can't take a property as sets of this-worldly instances, because two properties may be coextensive. Some say it is just as bad in all possible worlds, if the property is necessary, as when all triangles are trilateral, which seem different.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: [compressed] Renate/cordate is the standard example of the first problem. Lewis seems to equivocate over exactly what is meant by a property. I take the example to be a powerful objection to treating properties as sets.

1.5	p.55	15737	To be a 'property' is to suit a theoretical role
			Full Idea: To deserve the name of 'property' is to be suited to play the right theoretical role. It is wrong to speak of 'the' role associated with the word 'property', as if it were fully and uncontroversially settled.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Once again I see a chicken-and-egg problem. Surely something has a theoretical role because of its intrinsic character, or its prior definition? How could you formulate a theory if you lacked properties? We don’t meet properties as gaps in theories.

1.5	p.57	15738	Propositions can't have syntactic structure if they are just sets of worlds
			Full Idea: If it is central to 'proposition' that there be quasi-syntactic structure, so that there are subject-predicate, or negative, or conjunctive, or quantified propositions, then sets of possible worlds will not do.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: He proposes 'more complicated set-theoretic constructions out of possibilia' instead. I am very much committed to propositions having quasi-syntactic structure.

1.5	p.59	15739	There is the property of belonging to a set, so abundant properties are as numerous as the sets
			Full Idea: The abundant properties far outrun the predicates of any language we could possibly possess. ...Properties are as abundant as the sets, because for any set whatever, there is the property of belonging to that set.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: The idea of calling such things 'properties' strikes me as preposterous, but it is interesting that we confront truths which outrun our predicates. We can't have all of these predicates together, but there is no impediment to any one of them.

1.5	p.59	9656	The 'abundant' properties are just any bizarre property you fancy
			Full Idea: Properties are 'sparse' or 'abundant'. The abundant properties may be as extrinsic, as gruesomely gerrymandered, as miscellaneously disjunctive, as you please. They pay no heed to the qualitative joints, but carve things up every which way.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: This seems to be a logician's idea, which is needed for the notion of a set as pure extension, but it has very little to do with what I understand by the word 'property'. Better to call it a 'categorization'. E.g. filing George W. Bush under 'jackal'.

1.5	p.60	14996	Natural properties give similarity, joint carving, intrinsicness, specificity, homogeneity...
			Full Idea: Sharing of [perfectly natural properties] makes for qualitative similarity, they carve at the joints, they are intrinsic, they are highly specific, the sets of their instances are ipso facto not highly miscellaneous.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: All this sounds like just what I want, but when I read Lewis he seems to be arriving at these natural properties by the wrong route. Too much Hume, too much extensionalism.

1.5	p.61	15741	All of the natural properties are included among the intrinsic properties
			Full Idea: It cannot be said that all intrinsic properties are perfectly natural, ...but it can plausibly be said that all perfectly natural properties are intrinsic.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Idea 15742 give the example he uses to support this claim. I like the concept of 'intrinsic' properties, but cannot currently see any use for the concept of 'natural' ones.

1.5	p.61	15742	A disjunctive property can be unnatural, but intrinsic if its disjuncts are intrinsic
			Full Idea: A property can be unnatural by reason of disjunctiveness, as the property of being tripartite-or-liquid-or-cubical is, and it is intrinsic if its disjuncts are.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: This strikes me as being utterly, shockingly and disgracefully wrong. A disjunction can't possibly be a property. Can a person have the property of being 'fat or thin'? A disjunction offers candidates for properties, not the properties themselves.

1.5	p.63	15743	Defining natural properties by means of laws of nature is potentially circular
			Full Idea: Shall we say that natural properties are the ones that figure in laws of nature? - Not if we are going to use naturalness of properties when we draw the line between laws of nature and accidental regularities.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Personally I wouldn't dream of defining anything by saying that it figured in laws of nature. The laws, if there be such (see Mumford) are built up from more fundamental components, such as (perhaps) properties.

1.5	p.63	15744	We can't define natural properties by resemblance, if they are used to explain resemblance
			Full Idea: Shall we say that natural properties are the ones whose instances are united by resemblance? - Not if we are going to say that resemblance is the sharing of natural properties.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: The target of this appears to be the proposal of Quinton. By now I have totally given up on so-called 'natural' properties. Lewis says the circularity (also in Idea 15743) is a reason to treat 'natural' here as primitive (though he rejects that).

1.5	p.64	15745	Universals recur, are multiply located, wholly present, make things overlap, and are held in common
			Full Idea: One and the same universal recurs; it is multiply located; it is wholly present in both instances, a shared common part whereby the two instances overlap. Being alike by sharing a universal is 'having something in common' in an absolutely literal sense.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: A helpful spelling out of the commitment involved (in Armstrong and others) in belief in universals. To me this is a convenient list of reasons why the whole proposal is nonsense. Why does Lewis take them seriously?

1.5	p.65	15746	If particles were just made of universals, similar particles would be the same particle
			Full Idea: We cannot say that a particle is composed entirely of its several universals, because then another particle exactly like it would have the very same universals, and yet the two particles would not be the same.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: This is an argument either (implausibly) for haecceities or characterless substrata, or else for tropes (which are all separate, unlike universals). Particles as bundles of universals is not a theory I take seriously.

1.5	p.65	15748	Trope theory needs a primitive notion for what unites some tropes
			Full Idea: Trope theory needs a primitive notion - 'instantion' in yet another sense - to say how the tropes that comprise a particle are united.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Any theory which says that objects are just 'bundles' of tropes is asking for trouble. But if you say (with Lowe and others) that tropes are 'modes' of existing entities, you still have to give an account of the entity.

1.5	p.65	15749	Trope theory (unlike universals) needs a primitive notion of being duplicates
			Full Idea: Trope theory has the drawback that we need the primitive notion of duplicate tropes, whereas with universals we just say that it is one and the same universal through some charged particles.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: The normal term for this primitive is 'perfect resemblance', though that seems to make close resemblance a bit complicated and puzzling. I'm not sure if I understand resemblance as a feature of the world, rather than of our minds.

1.5	p.65	15750	Tropes need a similarity primitive, so they cannot be used to explain similarity
			Full Idea: Trope theory cannot analyse similarity, because duplication of tropes is itself a primitive relation of similarity.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: A reasonable reply to this one, I think, is that no one can explain or analyse similarity. To say that the same universal (or bunch of graded universals) is instantiated explains nothing. Maybe type-identity must be primitive in any theory?

1.5	p.65	15747	Universals aren't parts of things, because that relationship is transitive, and universals need not be
			Full Idea: It cannot be said that a universal is instantiated by anything that has it as a part, since the relation of part to whole is transitive. If charge is part of a particle, which is part of an atom, then charge is part of the atom, but an atom isn't charged.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Given the total mystery involved in 'instantiation', it wouldn't surprise me if someone appealed to the part-whole relation, but all moves to explain instantiation are desperate. Make it a primitive, if you must, then tiptoe away.

1.5	p.66	9657	You must accept primitive similarity to like tropes, but tropes give a good account of it
			Full Idea: If you will not countenance primitive similarity in any form, then trope theory is not for you. But if you will, then duplication of tropes is an especially satisfactory form of primitive similarity to take.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: This presents the question about tropes in a beautifully simple form. Perfect similarities seem fine, but partial similarities (red and pink things) are hard, and abstract reference (pinkness and redness) is even harder.

1.5	p.67	15751	Surely 'slept in by Washington' is a property of some bed?
			Full Idea: That the most noteworthy property of this bed is that George Washington slept in it - surely this is true on some legitimate conception of properties?
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Wrong! This example is a nice clear test case. This is an absurd slippery slope. A bed he once looked at? I would have thought this was a relation the bed once entered into, and a relation isn't a property.

1.5	p.73	9658	An explanation tells us how an event was caused
			Full Idea: An explanation, I think, is an account of etiology: it tells us something about how an event was caused.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5)
			A reaction: Will this cover mathematical explanations? Numbers would have to have causal powers.

1.5 n37	p.51	15731	Quantification sometimes commits to 'sets', but sometimes just to pluralities (or 'classes')
			Full Idea: I consider some apparent quantification over sets or classes of whatnots to carry genuine ontological commitment to 'sets' of them, but sometimes it is innocent plural quantification committed only to whatnots, for which I use 'class'.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5 n37)
			A reaction: How do you tell whether you are committed to a set or not? Can I claim an innocent plurality each time, while you accuse me of a guilty set? Can I firmly commit to a set, to be told that I can never manage more than a plurality?

1.5 n44	p.60	15740	I don't take 'natural' properties to be fixed by the nature of one possible world
			Full Idea: Some people suppose my natural properties are distinguished by nature, and hence natural in one world and not another. I intend properties to be natural or unnatural simpliciter, not relative to one or another world.
			From: David Lewis (On the Plurality of Worlds [1986], 1.5 n44)
			A reaction: This is an important warning for the likes of me. I've have begun to doubt the utility of the term 'natural' property, and this reinforces my view.

1.5 n47	p.67	15752	We might try defining the natural properties by a short list of them
			Full Idea: We might try defining the natural properties by a short list, of the mass properties, charge properties, quark properties and flavours...
			From: David Lewis (On the Plurality of Worlds [1986], 1.5 n47)
			A reaction: He rejects this because of possible natural properties in other possible worlds. Defining anything by a list seems like cheating. Does John, Paul, George and Ringo 'define' something?

1.6	p.73	10469	A world is a maximal mereological sum of spatiotemporally interrelated things
			Full Idea: A world is a maximal mereological sum of spatiotemporally interrelated things.
			From: David Lewis (On the Plurality of Worlds [1986], 1.6)
			A reaction: (I wonder what Lewis's account of space was?) A mereological sum is "the least inclusive thing that includes all the parts" (p.69). It is maximal when all 'worldmates' are parts. But then 'worldmates' are defined as parts, so it threatens circularity.

1.6	p.78	9659	Causation is when at the closest world without the cause, there is no effect either
			Full Idea: If it is the case at world W that if event C had not occurred, E would not have occurred either, then the counterfactual means that at the closest worlds to W at which C does not occur, E does not occur either.
			From: David Lewis (On the Plurality of Worlds [1986], 1.6)
			A reaction: This is a very Humean account, though updated, which sees nothing more to causation than transworld regularities. To me that is just describing the evidence for causation, not giving an account of it (even if the latter is impossible).

1.7	p.82	8901	Abstraction is usually explained either by example, or conflation, or abstraction, or negatively
			Full Idea: Abstraction is usually explained in one of four ways: the Way of Example (cf. donkeys and numbers), the Way of Conflation (same as sets), the Negative Way (non-spatial and non-causal) or the Way of Abstraction (incomplete descriptions).
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: [Compressed; a footnote dismisses Dummett's fifth way] Example has blurred boundaries, and explains nothing. Gerrymandered sets don't produce concepts. Negative accounts explain nothing. So it's the Way of Abstraction!

1.7	p.82	8938	The Way of Example compares donkeys and numbers, but what is the difference, and what are numbers?
			Full Idea: The Way of Example says concrete entities are things like donkeys and puddles, but abstract entities are things like numbers. That gives us little guidance. There are no uncontroversial accounts of numbers, and donkeys and number differ in too many ways.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: That demolishes that fairly swiftly. It may be unfair to demand an agreed account of numbers, but the respect(s) in which donkeys and numbers differ needs to be spelled out before anything useful has been said.

1.7	p.83	8903	Abstracta can be causal: sets can be causes or effects; there can be universal effects; events may be sets
			Full Idea: Is it true that sets or universals cannot enter into causal interaction? Why can't we say that a set of things causes something, or something causes a set of effects? Or positive charge has characteristic effects? Or an event is a sort of set?
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: This idea, and 8902, form a devastating critique of attempts to define abstraction in a purely negative way, as non-spatial and non-causal. Only a drastic revision of widely held views about sets, universals and events could save that account.

1.7	p.83	8902	If abstractions are non-spatial, then both sets and universals seem to have locations
			Full Idea: If abstract entities are not located, then a set of things does seem to have a location, though perhaps a divided one; and universals, if they are wholly present in each particular, are where their instances are, so negation can't define abstraction.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: He admits that non-spatial accounts of sets and universals are possible, but the jury is out on both of them, and more cautious theories, even if they are realist, will give them both locations. A good argument.

1.7	p.84	8904	The Way of Abstraction says an incomplete description of a concrete entity is the complete abstraction
			Full Idea: The Way of Abstraction says abstract entities are abstractions from concrete entities; they result from somehow subtracting specificity, so that an incomplete description of the original concrete entity is a complete description of the abstraction.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: Defined like this, it rather looks as if abstractions would be entirely verbal - which may well be the correct situation, except that higher animals seem capable of minimal levels of abstraction. This Way is denigrated by Frege and Geach.

1.7	p.85	8906	If we can abstract the extrinsic relations and features of objects, abstraction isn't universals or tropes
			Full Idea: Why can't we abstract a highly extrinsic aspect of something, say its surname, or its spatiotemporal location, or its role in a causal network, or its role in some body of theory? But these are unsuitable candidates for being genuine universals or tropes.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: (This is a criticism of the proposal in Idea 8905) Obviously we can abstract such things. In particular the role in a causal network is a function, which is a central example of an abstract idea. Russell keeps reminding us that relations are universals.

1.7	p.85	8905	If universals or tropes are parts of things, then abstraction picks out those parts
			Full Idea: A theory of non-spatiotemporal parts of things, whether recurring universals or non-recurring tropes, makes good sense of some abstractions. Unit negative charge is a universal common to particles, and an abstraction by being part of them.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: He cautiously refers to 'some' abstractions. It is one of Donald Williams's proud boasts concerning his trope theory that it will handle this problem well. I'm not sure that we should be saying that abstractions are actually concrete bits of things.

1.7	p.85	8908	For most sets, the concept of equivalence is too artificial to explain abstraction
			Full Idea: Most sets cannot be regarded as abstractions by equivalence: most sets are equivalence classes only under thoroughly artificial equivalence. (And the empty set is not an equivalence class at all).
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: [Recorded for further investigation..] My intuitions certainly cry out against such a thin logical notion giving a decent explanation of such a rich activity as abstraction.

1.7	p.85	8907	The abstract direction of a line is the equivalence class of it and all lines parallel to it
			Full Idea: We can abstract the direction of a line by taking the direction as the equivalence class of that line and all lines parallel to it. There is no subtraction of detail, but a multiplication of it; by swamping it, the specifics of the original line get lost.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: You can ask how wide a line is, but not how wide a direction is, so a detail IS being subtracted. I don't see how you can define the concept of a banana by just saying it is 'every object which is equivalent to a banana'. 'Parallel' is an abstraction.

1.7	p.86	8909	Abstractions may well be verbal fictions, in which we ignore some features of an object
			Full Idea: The inevitable hypothesis is that abstractions are verbal fictions. We say we are speaking about abstractions when we are speaking abstractly about the original thing. We are ignoring some features, not introducing a new thing lacking those features.
			From: David Lewis (On the Plurality of Worlds [1986], 1.7)
			A reaction: Thus Lewis ends up pretty close to Locke and the traditional view. This makes abstraction not a feat of platonic perception, in which magical non-material objects are spotted, but a feat of counterfactual imagination.

1.8	p.90	9660	The impossible can be imagined as long as it is a bit vague
			Full Idea: Imaginability is a poor criterion of possibility. We can imagine the impossible provided we do not imagine it in perfect detail and all the time.
			From: David Lewis (On the Plurality of Worlds [1986], 1.8)
			A reaction: In general I agree, but Williamson nicely opposes this view. The fact is that we derive most of our understanding of what is possible from imagination. We just have to realise that we can get it wrong, and so must attend to detail.

2.3	p.106	16278	A particular functional role is what gives content to a thought
			Full Idea: Anything that is a thinker at all has a thought with a certain content in virtue of being in a state which occupies a certain functional role.
			From: David Lewis (On the Plurality of Worlds [1986], 2.3)
			A reaction: So often Lewis seems to get things the wrong way round. Maybe if you invert his entire (fabulously consistent) philosophy, you get the right answer? I take the content to be what makes the role possible.

2.4	p.109	16279	General causal theories of knowledge are refuted by mathematics
			Full Idea: Causal accounts of knowledge are all very well in their place, but if they are put forward as general theories, then mathematics refutes them.
			From: David Lewis (On the Plurality of Worlds [1986], 2.4)
			A reaction: You might have some sort of notion of an abstraction from causation which mimicked it in the mathematics case. Lots of things seem to be 'forced' in mathematics. Call it 'ersatz causation'. Necessities are enforcers.

2.5	p.116	9661	Induction is just reasonable methods of inferring the unobserved from the observed
			Full Idea: I use the word 'induction' broadly, to cover all the methods we deem reasonable for forming beliefs about the unobserved parts of our world on the basis of experience with the observed parts.
			From: David Lewis (On the Plurality of Worlds [1986], 2.5)
			A reaction: Good. Attempts to be precise about it seem to be hopeless and invite paradoxes. Personally I just define it as 'learning from experience', because that makes what we do continuous with the behaviour of other sensible animals.

2.7	p.131	16280	Often explanaton seeks fundamental laws, rather than causal histories
			Full Idea: Sometimes the pursuit of explanation is more the pursuit of unified and general fundamental laws than of information about the causal histories of events.
			From: David Lewis (On the Plurality of Worlds [1986], 2.7)
			A reaction: It is hard to disagree, given the 'sometimes'. I don't think that Newton's Law of Gravity (say), with its lovely equation, actually explained anything at all about gravity. Finding the law closes the quest for an accurate description of what happens.

2.7	p.133	16274	If the well-ordering of a pack of cards was by shuffling, the explanation would make it more surprising
			Full Idea: Suppose you find in a hotel room a pack of cards in exactly standard order. Not surprising - maybe it's a new deck, or someone arranged them. Not so. They got that way by being fairly shuffled. The explanation would make the explanandum more surprising.
			From: David Lewis (On the Plurality of Worlds [1986], 2.7)
			A reaction: [compressed] A lovely Lewisian example, that instantly makes big trouble for the (implausible) view that a cause is something which increases the likelihood of a thing.

2.8	p.135	16281	Honesty requires philosophical theories we can commit to with our ordinary commonsense
			Full Idea: The maxim of honesty: never put forward a philosophical theory that you yourself cannot believe in your least philosophical and most commonsensical moments.
			From: David Lewis (On the Plurality of Worlds [1986], 2.8)
			A reaction: I take it as important that this test is according to the philosopher's commonsense, and not according to some populist idea. This would allow, for example, for commonsense to be sensitive to scientific knowledge, or awareness of the logic.

239-	p.155	11903	Extreme haecceitists could say I might have been a poached egg, but it is too remote to consider
			Full Idea: Lewis's strategy for defending extreme haecceitism is that supposed impossibilities (that I might have been a poached egg) could be reconstrued as genuine possibilities that are so remote from reality that they are ignored.
			From: report of David Lewis (On the Plurality of Worlds [1986], 239-) by Penelope Mackie - How Things Might Have Been 9.3
			A reaction: Not a promising route. Wiggins asks: if you think I could have been a poached egg, start by defining more precisely this 'I' to which we are referring. The definition will blatantly exclude any possibility of my poachedegghood.

248-263	p.18	13793	An essential property is one possessed by all counterparts
			Full Idea: For Lewis, if a property possessed by a given individual or kind is missing in some of the contextually relevant counterparts, that property is accidental to the individual or kind; if it is possessed by all of them, that property is essential.
			From: report of David Lewis (On the Plurality of Worlds [1986], 248-263) by Crawford L. Elder - Real Natures and Familiar Objects 1.4
			A reaction: This is a sophisticated version of the idea that essential properties are just necessary properties. It might make sense with a very sparse view of properties (mainly causal ones), but I think of essences as quite different from necessities.

3.1	p.136	16282	Ersatzers say we have one world, and abstract representations of how it might have been
			Full Idea: The ersatzers say that instead of an incredible plurality of concrete worlds, we can have one world only, and countless abstract entities representing ways that this world might have been.
			From: David Lewis (On the Plurality of Worlds [1986], 3.1)
			A reaction: Put me down as an ersatzer. They seem to be the same as Actualists. Are worlds other possible worlds, or ways 'this world might have been'? Not the same. Does actuality constrain what is possible? (Barcan formula?)

3.1	p.138	16283	For me, all worlds are equal, with each being actual relative to itself
			Full Idea: For me, all the worlds are on an equal footing in that each is actual relative to itself and none is actual relative to any other.
			From: David Lewis (On the Plurality of Worlds [1986], 3.1)
			A reaction: Lewis says the world we call 'actual' is simply a matter of how our indexicals refer. That sounds the wrong way round to me (as so often with Lewis).

3.1	p.141	16284	Ersatz worlds represent either through language, or by models, or magically
			Full Idea: I distinguish three principal ways ersatz worlds represent: linguistic, in which they are like stories or theories; pictorial, like pictures or isomorphic scale models; or magical, in which it is just their nature to represent.
			From: David Lewis (On the Plurality of Worlds [1986], 3.1)
			A reaction: I think I incline to the 'model' view. The linguistic version means animals can't assess possibilities. I take modelling to be basic to what a mind is, and what a mind is for.

3.2	p.145	16286	Linguistic possible worlds need a complete supply of unique names for each thing
			Full Idea: There are two difficulties with Carnap's taking possible worlds as linguistic. Everything must have a name, or our state-descriptions will be silent about nameless things, and nothing may have two names, or we may affirm and deny a predicate of one thing.
			From: David Lewis (On the Plurality of Worlds [1986], 3.2)
			A reaction: The idea of possible worlds as linguistic has no appeal for me, so this problem doesn't surprise or bother me, but it sounds fairly terminal for the project.

3.2	p.151	16287	Maximal consistency for a world seems a modal distinction, concerning what could be true together
			Full Idea: An ersatz world must be maximally consistent (hence destroyed by an additional sentences), …but that is prima facie a modal distinction: a set of sentences is consistent iff those sentences, as interpreted, could all be true together.
			From: David Lewis (On the Plurality of Worlds [1986], 3.2)
			A reaction: This is indicative of Lewis's motivation for his project, which is to eliminate modal facts from the world. Only a vast multitude of non-modal concrete worlds can satisfy all the contraints. Cf many-worlds quantum mechanics for non-locality.

3.2	p.154	16288	Analysis reduces primitives and makes understanding explicit (without adding new knowledge)
			Full Idea: The object of analysis is to reduce our burden of primitive notions, and to make tacit understanding explicit - not to bootstrap ourselves into understanding what we didn't understand at all beforehand.
			From: David Lewis (On the Plurality of Worlds [1986], 3.2)
			A reaction: I am particularly keen on the idea of 'making tacit understanding explicit'. I connect this with faith in intuition, and with the coherence view of justification.

3.2	p.165	9662	Linguistic possible worlds have problems of inconsistencies, no indiscernibles, and vocabulary
			Full Idea: Linguistic representations of possible worlds have three problems: some descriptions are inconsistent (which worlds cannot be); we cannot have indiscernible descriptions (though some worlds might be so); and descriptions are limited by vocabulary.
			From: David Lewis (On the Plurality of Worlds [1986], 3.2)
			A reaction: Lewis is wonderful at getting problems clearly on the table. I take the idea of possible worlds as linguistic entities to be a non-starter, because (as usual) animals do it too, when they think of possibilities, which even the dimmest ones must do.

3.3	p.173	16289	We can't account for an abstraction as 'from' something if the something doesn't exist
			Full Idea: We cannot really be talking about the things whence an abstraction-like entity is abstracted if there are no such things.
			From: David Lewis (On the Plurality of Worlds [1986], 3.3)
			A reaction: Sounds like a killer blow, but I don't think so. I can't think of a concept which doesn't have a possible basis in reality, assuming that it might be a complex assemblage of abstracted components.

3.4	p.189	16290	I believe in properties, which are sets of possible individuals
			Full Idea: I believe in properties. That is, I have my candidates for entities to play the role and deserve the name. My principal candidates are sets of possible individuals.
			From: David Lewis (On the Plurality of Worlds [1986], 3.4)
			A reaction: I am bewildered by any claim that a property is a 'set'. The property of being a teaspoon is just a large pile of teaspoons (one pile in each possible world)? This is the tyranny of first-order logic in philosophy. Are sets more real than properties?

4.1	p.192	15968	Identity is simple - absolutely everything is self-identical, and nothing is identical to another thing
			Full Idea: Identity is utterly simple and unproblematic. Everything is identical to itself; nothing is ever identical to anything except itself. There is never any problem about what makes something identical to itself; nothing can ever fail to be.
			From: David Lewis (On the Plurality of Worlds [1986], 4.1)
			A reaction: I have great problems with expressing this concept as a thing being 'identical to itself'. I will always say that it 'has an identity'. But then it is problematical, because what constitutes an identity? When do dispersing clouds lose it?

4.1	p.193	15969	Two things can never be identical, so there is no problem
			Full Idea: There is never any problem about what makes two things identical; two things can never be identical.
			From: David Lewis (On the Plurality of Worlds [1986], 4.1)
			A reaction: This expresses Lewis's preference for usage of the word 'identity', rather than a simple solution. It pays no attention to type-identity, which is an obvious phenomenon. In some sense, it is just obvious that two electrons are 'identical'.

4.1	p.197	16291	In counterpart theory 'Humphrey' doesn't name one being, but a mereological sum of many beings
			Full Idea: For the counterpart theorist, the trick is to say that 'Humphrey' names not the Humphrey of our world, and not the Humphrey of another, but rather the trans-world individual who is the mereological sum of all those local Humphreys.
			From: David Lewis (On the Plurality of Worlds [1986], 4.1)
			A reaction: On Lewis's perdurantism Humphrey is a 'spacetime worm' across his lifetime. Now we are adding all the possible Humphreys to the sum. I'm losing track of Humphrey's shape.

4.2	p.202	9663	A thing 'perdures' if it has separate temporal parts, and 'endures' if it is wholly present at different times
			Full Idea: Something 'perdures' iff it persists by having different temporal parts, or stages, at different times, though no one part of it is wholly present at more than one time; whereas it 'endures' iff it persists by being wholly present at more than one time.
			From: David Lewis (On the Plurality of Worlds [1986], 4.2)
			A reaction: Only a philosopher would come up with a concept like perdurance. I'm thinking about this one, and will get back to you in a later-numbered idea... He compares perdurance to the way a road persists through space.

4.2	p.203	9664	Endurance is the wrong account, because things change intrinsic properties like shape
			Full Idea: The principal and decisive objection to endurance, as an account of the persistence of ordinary things, is the problem of temporary intrinsics. Persisting things change their intrinsic properties, such as their shape. My own shape keeps changing.
			From: David Lewis (On the Plurality of Worlds [1986], 4.2)
			A reaction: Presumably if something was going to endure through time it would need a shape. If it has no particular shape, it lacks identity? Lewis discusses the problem at length. Why is a precise shape essential to anything?

4.2	p.204	9665	There are three responses to the problem that intrinsic shapes do not endure
			Full Idea: The problem for the endurance view of temporary intrinsic properties like shape is met be either saying shape is a disguised relational property, or only present properties are intrinsic, or the shapes belong to different things (perdurance).
			From: David Lewis (On the Plurality of Worlds [1986], 4.2)
			A reaction: [compressed] It is certainly implausible to deny that shape is a feature of all physical objects. Or it appears to be. Shapes are hard to pin down at the quantum level. How do you sharply divide moments for the perdurance view? ...

4.2	p.207	9666	It is quite implausible that the future is unreal, as that would terminate everything
			Full Idea: It is hard to believe that any philosopher means it when they say the future is unreal. If anyone is right that there is no future, that moment is their last, and it is the end of everything.
			From: David Lewis (On the Plurality of Worlds [1986], 4.2)
			A reaction: A bit simplistic. I might say 'there will be a future time, but it doesn't exist now'. That's the peculiar thing about time. If I say New York doesn't exist, then clearly I can't visit it. The London 2012 Olympic Stadium is going to exist.

4.3	p.211	9667	Mereological composition is unrestricted: any class of things has a mereological sum
			Full Idea: I claim that mereological composition is unrestricted: any old class of things has a mereological sum. Whenever there are some things, even out of different possible worlds, there is something composed of just those things.
			From: David Lewis (On the Plurality of Worlds [1986], 4.3)
			A reaction: To say the least, a rather unusual usage for the English word 'thing'. I presume that Lewis is in the grip of a slippery slope problem - that there is no way to define the borderline between things and non-things. Presumably 'class' is unrestricted too.

4.4	p.221	15129	Haecceitism implies de re differences but qualitative identity
			Full Idea: If two worlds differ in what they represent de re concerning some individual, but do not differ qualitatively in any way, I shall call that a haecceitist difference. Haecceitism, then, says there are at least some haecceitist differences between worlds.
			From: David Lewis (On the Plurality of Worlds [1986], 4.4)
			A reaction: Lewis bases this view on Kaplan. My brief summary of this is that 'identity may be hidden'. If all electrons are different, what distinguishes them?

4.4	p.230	9669	There are no free-floating possibilia; they have mates in a world, giving them extrinsic properties
			Full Idea: There are no free-floating possibilia. Every possibility is part of a world - exactly one world - and thus comes surrounded by worldmates, and fully equipped with extrinsic properties in virtue of its relations to them.
			From: David Lewis (On the Plurality of Worlds [1986], 4.4)
			A reaction: This is a key claim in the possible worlds understanding of modality, contrary to the more common sense and normal language claim that a possibility is an isolated thing.

4.4	p.239	9670	Extreme haecceitism says you might possibly be a poached egg
			Full Idea: The most extreme version of haecceitism says that anything could possibly have any qualitative character; for instance, there is a world according to which you are a poached egg.
			From: David Lewis (On the Plurality of Worlds [1986], 4.4)
			A reaction: Presumably a plausible haecceitist view would have to be combined with essentialism, given that the possibility that I might be a poached egg is beyond my intuitive grasp.

4.4 n32	p.244	9057	Vagueness is semantic indecision: we haven't settled quite what our words are meant to express
			Full Idea: I regard vagueness as semantic indecision: where we speak vaguely, we have not troubled to settle which of some range of precise meanings our words are meant to express.
			From: David Lewis (On the Plurality of Worlds [1986], 4.4 n32)
			A reaction: But that seems to leave the problem of how you are going to decide the boundaries of 'heap' or 'bald', if we all agree to become more precise. In law precise boundaries are often drawn a bit arbitrarily, simply because a boundary is needed.

4.5	p.251	9671	Whether or not France is hexagonal depends on your standards of precision
			Full Idea: Say that France is hexagonal, and you thereby set the standards of precision low, and you speak the truth; say that France is not hexagonal (preferably on some other occasion) and you set the standards high, and again you speak the truth.
			From: David Lewis (On the Plurality of Worlds [1986], 4.5)
			A reaction: This is very persuasive. It fits with my views on justification, which are to do with how high I (or more often 'we') decide to set the standards, thereby defining knowledge for that occasion. Hm. Has Lewis cracked vagueness? [P.S. NO!]

4.5	p.252	19280	I can ask questions which create a context in which origin ceases to be essential
			Full Idea: If I ask how things would be if Saul Kripke had come from no sperm and egg, but was brought by a stork, that makes sense. I create a context that makes my question make sense, which is a context that makes origin not to be essential.
			From: David Lewis (On the Plurality of Worlds [1986], 4.5)
			A reaction: I'm not clear why delivery by a stork doesn't just count as a different origin, and hence it turns out to be essential to Kripke. If Kripke were a necessary being (and he's a good candidate), then he would have no origin.

p.202-4	p.95	14737	Properties cannot be relations to times, if there are temporary properties which are intrinsic
			Full Idea: The problem of 'temporary intrinsics' is that in one model we think of properties as relations to times (I am 'bent' relative to now), but change sometime involves intrinsic properties. I am just plain bent, not bent with respect to something else.
			From: report of David Lewis (On the Plurality of Worlds [1986], p.202-4) by Theodore Sider - Four Dimensionalism
			A reaction: [I've compressed Sider's summary] The question of whether intrinsic properties endure over time runs in parallel with the question of whether objects endure over time, and the two issues cannot be separated.

p.212-3	p.121	13268	There are no restrictions on composition, because they would be vague, and composition can't be vague
			Full Idea: Lewis says that if not every class has a fusion then there must be a restriction on composition. The only plausible restrictions would be vague ones, which is impossible, because then whether composition occurs would be vague. So every class has a fusion.
			From: report of David Lewis (On the Plurality of Worlds [1986], p.212-3) by Theodore Sider - Four Dimensionalism 9.1
			A reaction: This is Lewis's key argument in favour of unrestricted composition, his Vagueness Argument. Why can't composition be vague? If you gradually reassemble a broken mirror, at what point does the mirror acquire its unity?

p.60-	p.85	16262	Sparse properties rest either on universals, or on tropes, or on primitive naturalness
			Full Idea: Lewis surveys three accounts of sparse properties: a set of objects instantiating a single universal; a set of objects having as parts duplicates of some trope; and a set distinguished by a further unanalyzable, primitive characteristic of naturalness.
			From: report of David Lewis (On the Plurality of Worlds [1986], p.60-) by Tim Maudlin - The Metaphysics within Physics
			A reaction: The very idea of suggesting that a property is some set of objects strikes me as bizarre. I present you with a table full of objects and say that is the complete set of some property. You then have to study the objects to find out what the property is.

p.61-2	p.268	15397	If a global intrinsic never varies between possible duplicates, all necessary properties are intrinsic
			Full Idea: Lewis defines a globally intrinsic property as one that never varies between duplicates across possible worlds. This has the immediate problem that any property that is necessarily had, or necessarily lacked, by every thing will be intrinsic.
			From: comment on David Lewis (On the Plurality of Worlds [1986], p.61-2) by Ross P. Cameron - Intrinsic and Extrinsic Properties 'Analysis'
			A reaction: [He also cites Langton and Lewis 1998] To me this is the sort of tangle you get into when you equate properties with predicates. The problem seems to concern necessary predicates (but those may not be necessary properties).

p.61-2	p.268	15398	Global intrinsic may make necessarily coextensive properties both intrinsic or both extrinsic
			Full Idea: If a globally intrinsic properties are those that never vary between duplicates across possible worlds, then necessarily coextensive properties will either be both intrinsic or both extrinsic.
			From: comment on David Lewis (On the Plurality of Worlds [1986], p.61-2) by Ross P. Cameron - Intrinsic and Extrinsic Properties 'Analysis'
			A reaction: Presumably this problem would arise if some intrinsic property entailed an extrinsic property (or, less likely, vice versa). These sorts of problems arise when you try to define everything extensionally (even across possible worlds).

1986	Modal Realism at Work: Properties

	p.23	4038	Properties are sets of their possible instances (which separates 'renate' from 'cordate')
			Full Idea: Lewis agrees that properties cannot be sets of their actual instances, but claims they can be sets of their possible instances. This would distinguish coextensive properties like being cordate and renate, since they might be separated.
			From: report of David Lewis (Modal Realism at Work: Properties [1986]) by DH Mellor / A Oliver - Introduction to 'Properties' §10
			A reaction: Sounds wrong. Two properties could be obviously different even if they could never be separated. In this world a creature might briefly survive without kidneys.

1986	Against Structural Universals

'Intro'	p.79	15433	Tropes are particular properties, which cannot recur, but can be exact duplicates
			Full Idea: Tropes are supposed to be particularized properties: nonspatiotemporal parts of their instances which cannot occur repeatedly, but can be exact duplicates.
			From: David Lewis (Against Structural Universals [1986], 'Intro')
			A reaction: Russell's objection is that 'duplication' appears to be a non-trope universal. The account seems wrong for very close resemblance, which is accepted by everyone as being the same (e.g. in colour, for football shirts).

'The magical'	p.100	15448	The 'magical' view of structural universals says they are atoms, even though they have parts
			Full Idea: The 'magical' conception of structural universals says 'simple' must be distinguished from 'atomic'. A structural universal is never simple; it involves other, simpler, universals, but it is mereologically atomic. The other universals are not its parts.
			From: David Lewis (Against Structural Universals [1986], 'The magical')
			A reaction: Hence the 'magic' is for it to be an indissoluble unity, while acknowledging that it has parts. Personally I don't see much problem with this view, since universals already perform the magical feat of being 'instantiated', whatever that means.

'The magical'	p.101	15449	If 'methane' is an atomic structural universal, it has nothing to connect it to its carbon universals
			Full Idea: What is it about the universal carbon that gets it involved in necessary connections with methane? Why not rubidium instead? The universal 'carbon' has nothing more in common with the universal methane than the universal rubidium has!
			From: David Lewis (Against Structural Universals [1986], 'The magical')
			A reaction: This is his objection to the 'magical' unity of structural universals. The point is that if methane is an atomic unity, as claimed, it can't have anything 'in common' with its components.

'The pictorial'	p.90	15439	The 'pictorial' view of structural universals says they are wholes made of universals as parts
			Full Idea: On the 'pictorial' conception, a structural universal is isomorphic to its instances. ...It is an individual, a mereological composite, not a set. ...It is composed of simpler universals which are literally parts of it.
			From: David Lewis (Against Structural Universals [1986], 'The pictorial')
			A reaction: I'm not clear why Lewis labels this the 'pictorial' view. His other two views of structural universals are 'linguistic' and 'magical'. The linguistic is obviously wrong, and the magical doesn't sound promising. Must I vote for pictorial?

'The pictorial'	p.91	15441	The structural universal 'methane' needs the universal 'hydrogen' four times over
			Full Idea: What is wrong with the pictorial conception is that if the structural universal 'methane' is to be an isomorph of the molecules that are its instances, it must have the universal 'hydrogen' as a part not just once, but four times over.
			From: David Lewis (Against Structural Universals [1986], 'The pictorial')
			A reaction: The point is that if hydrogen is a universal it must be unique, so there can't be four of them. To me this smacks of the hopeless mess theologians get into, because of bad premisses. Drop universals, and avoid this kind of stuff.

'The pictorial'	p.91	15440	A whole is distinct from its parts, but is not a further addition in ontology
			Full Idea: A whole is an extra item in our ontology only in the minimal sense that it is not identical to any of its proper parts; but it is not distinct from them either, so when we believe in the parts it is no extra burden to believe in the whole.
			From: David Lewis (Against Structural Universals [1986], 'The pictorial')
			A reaction: A little confusing, to be 'not identical' and yet 'not different'. As Lewis says elsewhere, the whole is one, and the parts are not. A crux. Essentialism implies a sort of holism, that parts with a structure constitute a new thing.

'The pictorial'	p.93	15443	Mathematicians abstract by equivalence classes, but that doesn't turn a many into one
			Full Idea: When mathematicians abstract one thing from others, they take an equivalence class. ....But it is only superficially a one; underneath, a class are still many.
			From: David Lewis (Against Structural Universals [1986], 'The pictorial')
			A reaction: This is Frege's approach to abstraction, and it is helpful to have it spelled out that this is a mathematical technique, even when applied by Frege to obtaining 'direction' from classes of parallels. Too much philosophy borrows inappropriate techniques.

'Uninstantiated'	p.103	15450	Maybe abstraction is just mereological subtraction
			Full Idea: We could say that abstraction is just mereological subtraction of universals.
			From: David Lewis (Against Structural Universals [1986], 'Uninstantiated')
			A reaction: This only works, of course, for the theories that complex universals have simpler universals as 'parts'. This is just a passing surmise. I take it that abstraction only works for a thing whose unity survives the abstraction.

'Uninstantiated'	p.107	15451	I assume there could be natural properties that are not instantiated in our world
			Full Idea: It is possible, I take it, that there might be simple natural properties different from any that instantiated within our world.
			From: David Lewis (Against Structural Universals [1986], 'Uninstantiated')
			A reaction: Interesting. Fine for Lewis, of course, for whom possibilities seem (to me) to be just logical possibilities. Even a scientific essentialist, though, must allow that different stuff might exist, which might have different intrinsic properties.

'Variants'	p.95	15444	Different things (a toy house and toy car) can be made of the same parts at different times
			Full Idea: Different things can be made of the same parts at different times, as when the tinkertoy house is taken apart and put back together as a tinkertoy car.
			From: David Lewis (Against Structural Universals [1986], 'Variants')
			A reaction: More important than it looks! This is Lewis's evasion of the question of the structure of the parts. Times will individuate different structures, but if I take type-identical parts and make a house and a car simultaneously, are they type-identical?

'Variants'	p.96	15445	Butane and Isobutane have the same atoms, but different structures
			Full Idea: The stuctural universal 'isobutane' consists of the universal carbon four times over, hydrogen ten times over, and the universal 'bonded' thirteen times over - just like the universal 'butane'.
			From: David Lewis (Against Structural Universals [1986], 'Variants')
			A reaction: The point is that isobutane and butane have the same components in different structures. At least this is Lewis facing up to the problem of the 'flatness' of mereological wholes.

'Variants'	p.97	15446	Composition is not just making new things from old; there are too many counterexamples
			Full Idea: Not just any operation that makes new things from old is a form of composition! There is no sense in which my parents are part of me, and no sense in which two numbers are parts of their greatest common factor.
			From: David Lewis (Against Structural Universals [1986], 'Variants')
			A reaction: One of those rare moments when David Lewis seems to have approached a really sensible metaphysics. Further on he rejects all forms of composition apart from mereology.

'What are'	p.81	15434	Structural universals have a necessary connection to the universals forming its parts
			Full Idea: There is a necessary connection between the instantiating of a structural universal by the whole and the instantiating of other universals by its parts. We can call the relation 'involvement', a nondescript word.
			From: David Lewis (Against Structural Universals [1986], 'What are')
			A reaction: In the case of a shape, I suppose the composing 'universals' [dunno what they are] will all be essential to the shape - that is, part of the very nature of the thing, loss of which would destroy the identity.

'Why believe'	p.82	15435	If you think universals are immanent, you must believe them to be sparse, and not every related predicate
			Full Idea: Any theorist of universals as immanent had better hold a sparse theory; it is preposterous on its face that a thing has as many nonspatiotemporal parts as there are different predicates that it falls under, or different classes that it belongs to.
			From: David Lewis (Against Structural Universals [1986], 'Why believe')
			A reaction: I am firmly committed to sparse universal, and view the idea that properties are just predicates as the sort of nonsense that results from approaching philosophy too linguistically.

'Why believe'	p.82	15436	Universals are meant to give an account of resemblance
			Full Idea: Perhaps the main job of a theory of universals is to give an account of resemblance.
			From: David Lewis (Against Structural Universals [1986], 'Why believe')
			A reaction: This invites the quick reply, popular with some nominalists, of taking resemblance as primitive, and hence beyond explanation.

'Why believe'	p.85	15437	We can't get rid of structural universals if there are no simple universals
			Full Idea: We can't dispense with structural universals if we cannot be sure that there are any simples which can be involved in them.
			From: David Lewis (Against Structural Universals [1986], 'Why believe')
			A reaction: Lewis cites this as Armstrong's strongest reason for accepting structural universals (and he takes their requirement for an account of laws of nature as the weakest). I can't comprehend a world that lacks underlying simplicity.

'Why believe'	p.86	15438	We can add a primitive natural/unnatural distinction to class nominalism
			Full Idea: To class nominalism we can add a primitive distinction between natural and unnatural classes.
			From: David Lewis (Against Structural Universals [1986], 'Why believe')
			A reaction: Lewis explores this elsewhere, but this looks like a very complex concept to play the role of a 'primitive'. Human conventions seem to be parts of nature.

1988	Rearrangement of Particles

1	p.188	15540	You can't deny temporary intrinsic properties by saying the properties are relations (to times)
			Full Idea: To say that properties are really relations to times is to treat temporary intrinsics (such as my changing shape) as a matter of relations, but then 'intrinsic properties' would not deserve the name, and it is untenable if it denies temporary intrinsics.
			From: David Lewis (Rearrangement of Particles [1988], 1)
			A reaction: [I have compressed a paragraph; he refers to his 1986:204] If a property is meant to be a 'relation to a time', I am not sure what the relata are meant to be, and I agree with Lewis that this seems a long way from properties.

1988	Vague Identity: Evans misunderstood

p.318	p.318	16458	Semantic vagueness involves alternative and equal precisifications of the language
			Full Idea: If vagueness is semantic indeterminacy, then wherever we have vague statements, we have several alternative precisifications of the vague language involved, all with equal claims of being 'intended'.
			From: David Lewis (Vague Identity: Evans misunderstood [1988], p.318)

1990	Noneism or Allism?

p.152	p.152	15532	'Allists' embrace the existence of all controversial entities; 'noneists' reject all but the obvious ones
			Full Idea: An expansive friend of the controversial entities who says they all exist may be called an 'allist'; a tough desert-dweller who says that none of them exist may be called a 'noneist'.
			From: David Lewis (Noneism or Allism? [1990], p.152)
			A reaction: Lewis gives examples of the obvious and the controversial entities. Lewis implies that he himself is in between. The word 'desert' is a reference to Quine.

p.159	p.159	15533	We can quantify over fictions by quantifying for real over their names
			Full Idea: Substitutionalists simulate quantification over fictional characters by quantifying for real over fictional names.
			From: David Lewis (Noneism or Allism? [1990], p.159)
			A reaction: I would say that a fiction is a file of conceptual information, identified by a label. The label brings baggage with it, and there is no existence in the label.

p.159	p.159	15534	We could quantify over impossible objects - as bundles of properties
			Full Idea: We can quantify over Meinongian objects by quantifying for real over property bundles (such as the bundle of roundness and squareness).
			From: David Lewis (Noneism or Allism? [1990], p.159)

p.163	p.163	15535	We can't accept a use of 'existence' that says only some of the things there are actually exist
			Full Idea: If 'existence' is understood so that it can be a substantive thesis that only some of the things there are exist, we will have none of it.
			From: David Lewis (Noneism or Allism? [1990], p.163)
			A reaction: Lewis is a strong advocate, following Quine, of the univocal sense of the word 'exist', and I agree with them.

1991	Parts of Classes

	p.80	10191	Set theory reduces to a mereological theory with singletons as the only atoms
			Full Idea: Lewis has shown that set theory may be reduced to a mereological theory in which singletons are the only atoms.
			From: report of David Lewis (Parts of Classes [1991]) by Fraser MacBride - Review of Chihara's 'Structural Acc of Maths' p.80
			A reaction: Presumably the axiom of extensionality, that a set is no more than its members, translates into unrestricted composition, that any parts will make an object. Difficult territory, but I suspect that this is of great importance in metaphysics.

	p.118	18395	Sets are mereological sums of the singletons of their members
			Full Idea: Lewis pointed out that many-membered classes are nothing more than the mereological wholes of the classes formed by taking the singleton of each member.
			From: report of David Lewis (Parts of Classes [1991]) by David M. Armstrong - Truth and Truthmakers 09.4
			A reaction: You can't combine members to make the class, because the whole and the parts are of different type, but here the parts and whole are both sets, so they combine like waterdrops.

	p.137	14244	Lewis only uses fusions to create unities, but fusions notoriously flatten our distinctions
			Full Idea: Lewis employs mereological fusion as his sole method of making one thing out of many, and fusion is notorious for the way it flattens out and thereby obliterates distinctions.
			From: comment on David Lewis (Parts of Classes [1991]) by Oliver,A/Smiley,T - What are Sets and What are they For? 3.1
			A reaction: I take this to be a key point in the discussion of mereology in ontological contexts. As a defender of intrinsic structural essences, I have no use for mereological fusions, and look for a quite different identity for 'wholes'.

	p.372	10566	Lewis prefers giving up singletons to giving up sums
			Full Idea: In the face of the conflict between mereology and set theory, Lewis has advocated giving up the existence of singletons rather than sums.
			From: report of David Lewis (Parts of Classes [1991]) by Kit Fine - Replies on 'Limits of Abstraction' 1

Pref	p.-4	15496	We can build set theory on singletons: classes are then fusions of subclasses, membership is the singleton
			Full Idea: The notion of a singleton, or unit set, can serve as the distinctive primitive of set theory. The rest is mereology: a class is the fusion of its singleton subclasses, something is a member of a class iff its singleton is part of that class.
			From: David Lewis (Parts of Classes [1991], Pref)
			A reaction: This is a gloriously bold proposal which I immediately like, because it cuts out the baffling empty set (which many people think 'exists'!), and gets mathematics back to being about the real world of entities (as the Greeks thought).

Pref	p.-3	15497	We can replace the membership relation with the member-singleton relation (plus mereology)
			Full Idea: Given the theory of part and whole, the member-singleton relation may replace membership generally as the primitive notion of set theory.
			From: David Lewis (Parts of Classes [1991], Pref)
			A reaction: An obvious question is to ask what the member-singleton relation is if it isn't membership.

1.2	p.4	15498	We can accept the null set, but there is no null class of anything
			Full Idea: There is no such class as the null class. I don't mind calling some memberless thing - some individual - the null 'set'. But that doesn't make it a memberless class.
			From: David Lewis (Parts of Classes [1991], 1.2)
			A reaction: The point is that set theory is a formal system which can do what it likes, but classes are classes 'of' things. Everyone assumes that sets are classes, reserving 'proper classes' for the tricky cases up at the far end.

1.2	p.5	15500	Classes divide into subclasses in many ways, but into members in only one way
			Full Idea: A class divides exhaustively into subclasses in many different ways; whereas a class divides exhaustively into members in only one way.
			From: David Lewis (Parts of Classes [1991], 1.2)

1.2	p.5	15499	A subclass of a subclass is itself a subclass; a member of a member is not in general a member
			Full Idea: Just as a part of a part is itself a part, so a subclass of a subclass is itself a subclass; whereas a member of a member is not in general a member.
			From: David Lewis (Parts of Classes [1991], 1.2)
			A reaction: Lewis is showing the mereological character of sets, but this is a key distinction in basic set theory. When the members of members are themselves members, the set is said to be 'transitive'.

1.2	p.8	15501	We have no idea of a third sort of thing, that isn't an individual, a class, or their mixture
			Full Idea: As yet we have no idea of any third sort of thing that is neither individual nor class nor mixture of the two.
			From: David Lewis (Parts of Classes [1991], 1.2)
			A reaction: You can see that Lewis was a pupil of Quine. I quote this to show how little impression 'stuff' makes on the modern radar. His defence is that stuff may not be a 'thing', but then he seems to think that 'things' exhaust reality (top p.8 and 9).

1.4	p.11	15502	There are four main reasons for asserting that there is an empty set
			Full Idea: The null set is a denotation of last resort for class-terms that fail to denote classes, an intersection of x and y where they have no members in common, the class of all self-members, and the real numbers such that x^2+1=0. This is all mere convenience.
			From: David Lewis (Parts of Classes [1991], 1.4)
			A reaction: A helpful catalogue of main motivations for the existence of the null set in set theory. Lewis aims to undermine these reasons, and dispense with the wretched thing.

1.4	p.13	15503	We needn't accept this speck of nothingness, this black hole in the fabric of Reality!
			Full Idea: Must we accept the null set as a most extraordinary individual, a little speck of sheer nothingness, a sort of black hole in the fabric of Reality itself? Not really.
			From: David Lewis (Parts of Classes [1991], 1.4)
			A reaction: We can only dream of reaching the level of confidence that Lewis reached, to make such beautiful fun of a highly counterintuitive idea that is rooted in the modern techniques of philosophy.

1.8	p.20	15504	Atomless gunk is an individual whose parts all have further proper parts
			Full Idea: A blob can represent atomless gunk: an individual whose parts all have further proper parts.
			From: David Lewis (Parts of Classes [1991], 1.8)
			A reaction: This is not the same as 'stuff', since gunk is a precise fusion of all those parts, whereas there is no such precision about stuff. Stuff is neutral as to whether it has atoms, or is endlessly divisible. My love of stuff grows. Laycock is a hero.

2.1	p.31	15507	Set theory has some unofficial axioms, generalisations about how to understand it
			Full Idea: Set theory has its unofficial axioms, traditional remarks about the nature of classes. They are never argued, but are passed heedlessly from one author to another. One of these says that the classes are nowhere: they are outside space and time.
			From: David Lewis (Parts of Classes [1991], 2.1)
			A reaction: Why don't the people who write formal books on set theory ever say things like this?

2.1	p.31	15506	If we don't understand the singleton, then we don't understand classes
			Full Idea: Our utter ignorance about the nature of the singletons amounts to sheer ignorance about the nature of classes generally.
			From: David Lewis (Parts of Classes [1991], 2.1)

2.1	p.32	15508	If singletons are where their members are, then so are all sets
			Full Idea: If every singleton was where its member was, then, in general, classes would be where there members were.
			From: David Lewis (Parts of Classes [1991], 2.1)
			A reaction: There seems to be a big dislocation of understanding of the nature of sets, between 'pure' set theory, and set theory with ur-elements. I take the pure to be just an 'abstraction' from the more located one. The empty set has a puzzling location.

2.1 n2	p.33	15509	Some say qualities are parts of things - as repeatable universals, or as particulars
			Full Idea: Some philosophers propose that things have their qualities by having them as parts, either as repeatable universals (Goodman), or as particulars (Donald Williams).
			From: David Lewis (Parts of Classes [1991], 2.1 n2)
			A reaction: He refers to 'qualities' rather than 'properties', presumably because this view makes them all intrinsic to the object. Is being 'handsome' a part of a person?

2.2	p.37	15511	If singleton membership is external, why is an object a member of one rather than another?
			Full Idea: Suppose the relation of member to singleton is external. Why must Possum be a member of one singleton rather than another? Why isn't it contingent which singleton is his?
			From: David Lewis (Parts of Classes [1991], 2.2)
			A reaction: He cites Van Inwagen for raising this question, and answers it in terms of counterparts. So is the relation internal or external? I think of sets as pairs of curly brackets, not existing entities, so the question doesn't bother me.

2.3	p.39	15512	In mereology no two things consist of the same atoms
			Full Idea: It is a principle of mereology that no two things consist of exactly the same atoms.
			From: David Lewis (Parts of Classes [1991], 2.3)
			A reaction: The problem with this is screamingly obvious - that the same atoms might differ in structure. Lewis did refer to this problem, but seems to try to wriggle out of it, in Idea 15444.

2.5	p.43	15513	Maybe singletons have a structure, of a thing and a lasso?
			Full Idea: Maybe the singleton of something x is not an atom, but consists of x plus a lasso. That gives a singleton an internal structure. ...But what do we know of the nature of the lasso, or how it fits? We are no better off.
			From: David Lewis (Parts of Classes [1991], 2.5)
			A reaction: [second bit on p.45]

2.6	p.50	15514	A huge part of Reality is only accepted as existing if you have accepted set theory
			Full Idea: The preponderant part of Reality must consist of unfamiliar, unobserved things, whose existence would have gone unsuspected but for our acceptance of set theory.
			From: David Lewis (Parts of Classes [1991], 2.6)
			A reaction: He is referring to the enormous sets at the far end of set theory, of a size that had never been hitherto conceived. Excellent. Daft to believe in something entirely because you have accepted set theory, with no other basis.

2.6	p.53	15515	To be a structuralist, you quantify over relations
			Full Idea: To be a structuralist, you quantify over relations.
			From: David Lewis (Parts of Classes [1991], 2.6)

2.7	p.55	15516	A property is any class of possibilia
			Full Idea: A property is any class of possibilia.
			From: David Lewis (Parts of Classes [1991], 2.7)

2.8	p.58	15517	Giving up classes means giving up successful mathematics because of dubious philosophy
			Full Idea: Renouncing classes means rejecting mathematics. That will not do. Mathematics is an established, going concern. Philosophy is as shaky as can be.
			From: David Lewis (Parts of Classes [1991], 2.8)
			A reaction: This culminates in his famous 'Who's going to tell the mathematicians? Not me!'. He has just given four examples of mathematics that seems to entirely depend on classes. This idea sounds like G.E. Moore's common sense against scepticism.

3.2 n2	p.63	15518	I like plural quantification, but am not convinced of its connection with second-order logic
			Full Idea: I agree fully with Boolos on substantive questions about plural quantification, though I would make less than he does of the connection with second-order logic.
			From: David Lewis (Parts of Classes [1991], 3.2 n2)
			A reaction: Deep matters, but my inclination is to agree with Lewis, as I have never been able to see why talk of plural quantification led straight on to second-order logic. A plural is just some objects, not some higher-order entity.

3.5	p.80	15520	Existence doesn't come in degrees; once asserted, it can't then be qualified
			Full Idea: Existence cannot be a matter of degree. If you say there is something that exists to a diminished degree, once you've said 'there is' your game is up.
			From: David Lewis (Parts of Classes [1991], 3.5)
			A reaction: You might have thought that this was so obvious as to be not worth saying, but as far as I can see it is a minority view in contemporary philosophy. It was Quine's view, and it is mine.

3.5	p.80	15519	Trout-turkeys exist, despite lacking cohesion, natural joints and united causal power
			Full Idea: A trout-turkey is inhomogeneous, disconnected, not in contrast with its surroundings. It is not cohesive, not causally integrated, not a causal unit in its impact on the rest of the world. It is not carved at the joints. That doesn't affect its existence.
			From: David Lewis (Parts of Classes [1991], 3.5)
			A reaction: A nice pre-emptive strike against all the reasons why anyone might think more is needed for unity than a mereological fusion.

3.6	p.81	15521	Given cats, a fusion of cats adds nothing further to reality
			Full Idea: Given a prior commitment to cats, a commitment to cat-fusions is not a further commitment. The fusion is nothing over and above the cats that compose it. It just is them. They just are it. Together or separately, the cats are the same portion of Reality.
			From: David Lewis (Parts of Classes [1991], 3.6)
			A reaction: The two extremes of ontology are that there are no objects, or that every combination is an object. Until reading this I thought Lewis was in the second camp, but this sounds like object-nihilism, as in Van Inwagen and Merricks.

3.6	p.87	15522	The one has different truths from the many; it is one rather than many, one rather than six
			Full Idea: What's true of the many is not exactly what's true of the one. After all they are many while it is one. The number of the many is six, whereas the number of the fusion is one.
			From: David Lewis (Parts of Classes [1991], 3.6)
			A reaction: Together with Idea 15521, this nicely illustrates the gulf between commitment to ontology and commitment to truths. The truths about a fusion change, while its ontology remains the same. Possibly this is the key to all of metaphysics.

3.6	p.87	14748	The many are many and the one is one, so they can't be identical
			Full Idea: What is true of the many is not exactly what is true of the one. After all they are many while it is one. The number of the many is six, whereas the number of the fusion is one. The singletons of the many are distinct from the singleton of the one.
			From: David Lewis (Parts of Classes [1991], 3.6)
			A reaction: I wouldn't take this objection to be conclusive. 'Some pebbles' seem to be many, but a 'handful of pebbles' seem to be one, where the physical situation might be identical. If they are not identical, then the non-identity is purely conceptual.

3.6	p.87	15523	Set theory isn't innocent; it generates infinities from a single thing; but mathematics needs it
			Full Idea: Set theory is not innocent. Its trouble is that when we have one thing, then somehow we have another wholly distinct thing, the singleton. And another, and another....ad infinitum. But that's the price for mathematical power. Pay it.
			From: David Lewis (Parts of Classes [1991], 3.6)

4.6	p.111	15524	Zermelo's model of arithmetic is distinctive because it rests on a primitive of set theory
			Full Idea: What sets Zermelo's modelling of arithmetic apart from von Neumann's and all the rest is that he identifies the primitive of arithmetic with an appropriately primitive notion of set theory.
			From: David Lewis (Parts of Classes [1991], 4.6)
			A reaction: Zermelo's model is just endlessly nested empty sets, which is a very simple structure. I gather that connoisseurs seem to prefer von Neumann's model (where each number contains its predecessor number).

4.7	p.120	15525	Plural quantification lacks a complete axiom system
			Full Idea: There is an irremediable lack of a complete axiom system for plural quantification.
			From: David Lewis (Parts of Classes [1991], 4.7)

p.81	p.17	10660	A commitment to cat-fusions is not a further commitment; it is them and they are it
			Full Idea: Given a prior commitment to cats, a commitment to cat-fusions is not a further commitment. The fusion is nothing over and above the cats that compose it. It just is them. They just are it.
			From: David Lewis (Parts of Classes [1991], p.81), quoted by Achille Varzi - Mereology 4.3
			A reaction: I take this to make Lewis a nominalist, saying the same thing that Goodman said about Utah in Idea 10657. Any commitment to cat-fusions being more than the cats, or Utah being more than its counties, strikes me as crazy.

p.84-7	p.20	6129	Lewis affirms 'composition as identity' - that an object is no more than its parts
			Full Idea: Lewis says that the parts of a thing are identical with the whole they compose, calling his view 'composition as identity', which is the claim that a physical object is 'nothing over and above its parts'.
			From: report of David Lewis (Parts of Classes [1991], p.84-7) by Trenton Merricks - Objects and Persons §I.IV
			A reaction: The ontological economy of this view is obviously attractive, but I don't agree with it. You certainly can't say that all identity consists entirely of composition by parts, because the parts need identity to get the view off the ground.

1992	Armstrong on combinatorial possibility

p.207	p.207	14399	Presentism says only the present exists, so there is nothing for tensed truths to supervene on
			Full Idea: Presentism says that although there is nothing outside the present, yet there are past-tensed and future-tensed truths that do not supervene on the present, and hence do not supervene on being.
			From: David Lewis (Armstrong on combinatorial possibility [1992], p.207)
			A reaction: Since I rather like both presentism and truth supervening on being, this observation comes as rather a devastating blow. I thought philosophy would be quite easy, but it's turning out to be rather tricky. Could tensed truths supervene on the present?

'The demand'	p.203	15545	Armstrong's analysis seeks truthmakers rather than definitions
			Full Idea: I suggest that Armstrong has an unfamiliar notion of analysis, as not primarily a quest for definitions, but as a quest for truth-makers.
			From: David Lewis (Armstrong on combinatorial possibility [1992], 'The demand')
			A reaction: This is not a dichotomy, I think, but a shift of emphasis. A definition will probably refer to truthmakers; a decent account of truthmakers would approximate a definition.

'The demand'	p.204	15546	Predications aren't true because of what exists, but of how it exists
			Full Idea: Predications seem, for the most part, to be true not because of whether things are, but because of how things are.
			From: David Lewis (Armstrong on combinatorial possibility [1992], 'The demand')
			A reaction: This simple point shows that you get into a tangle if you insist that truthmakers just consist of what exists. Lewis says Armstrong offers states of affairs as truthmakers for predications.

'Truth'	p.206	15548	Say 'truth is supervenient on being', but construe 'being' broadly
			Full Idea: I want to say that 'truth is supervenient on being', but as an Ostrich about universals I want to construe 'being' broadly.
			From: David Lewis (Armstrong on combinatorial possibility [1992], 'Truth')
			A reaction: [His slogan is borrowed from Bigelow 1988:132-,158-9] This seems much more promising that the more precise and restricted notion of truthmakers, as resting on the existence of particular things. Presentism is the big test case.

'What is there'	p.200	15543	How do things combine to make states of affairs? Constituents can repeat, and fail to combine
			Full Idea: To me it is mysterious how a state of affairs is made out of its particular and universal constituents. Different states of affairs may have the very same constituents, and the existence of constituents by no means entails the existence of the states.
			From: David Lewis (Armstrong on combinatorial possibility [1992], 'What is there')
			A reaction: He is rejecting the structure of states of affairs as wholes made of parts. But then mereology was never going to explain the structure of the world.

1993	Many, but almost one

'A better solution'	p.174	15539	Basic to pragmatics is taking a message in a way that makes sense of it
			Full Idea: The cardinal principle of pragmatics is that the right way to take what is said, if at all possible, is the way that makes sense of the message.
			From: David Lewis (Many, but almost one [1993], 'A better solution')
			A reaction: Thus when someone misuses a word, suggesting nonsense, we gloss over it, often without even mentioning it, because the underlying sense is obvious. A good argument for the existence of propositions. Lewis doesn't mention truth.

'The paradox'	p.167	15537	If cats are vague, we deny that the many cats are one, or deny that the one cat is many
			Full Idea: To deny that there are many cats on the mat (because removal of a few hairs seems to produce a new one), we must either deny that the many are cats, or else deny that the cats are many. ...I think both alternatives lead to successful solutions.
			From: David Lewis (Many, but almost one [1993], 'The paradox')
			A reaction: He credits the problem to Geach (and Tibbles), and says it is the same as Unger's 'problem of the many' (Idea 15536).

'The problem'	p.164	15536	We have one cloud, but many possible boundaries and aggregates for it
			Full Idea: Many surfaces are equally good candidates to be boundaries of a cloud; therefore many aggregates of droplets are equally good candidates to be the cloud. How is it that we have just one cloud? And yet we do. This is Unger's (1980) 'problem of the many'.
			From: David Lewis (Many, but almost one [1993], 'The problem')
			A reaction: This is the problem of vague objects, as opposed to the problem of vague predicates, or the problem of vague truths, or the problem of vague prepositions (like 'towards').

'Two solutions'	p.170	15538	Semantic indecision explains vagueness (if we have precisifications to be undecided about)
			Full Idea: Semantic indecision will suffice to explain the phenomenon of vagueness. [note] Provided that there exist the many precisifications for us to be undecided between. If you deny this, you will indeed have need of vague objects.
			From: David Lewis (Many, but almost one [1993], 'Two solutions')
			A reaction: [He mentions Van Inwagen 1990:213-83] There seem to be three solutions to vague objects: that they really are vague, that they are precise but we can't know precisely, or Lewis's view. I like Lewis's view. Do animals have any problem with vagueness?

1993	Mathematics is Megethology

p.03	p.3	10807	Mathematics reduces to set theory, which reduces, with some mereology, to the singleton function
			Full Idea: It is generally accepted that mathematics reduces to set theory, and I argue that set theory in turn reduces, with some aid of mereology, to the theory of the singleton function.
			From: David Lewis (Mathematics is Megethology [1993], p.03)

p.03	p.3	10806	Megethology is the result of adding plural quantification to mereology
			Full Idea: Megethology is the result of adding plural quantification, as advocated by George Boolos, to the language of mereology.
			From: David Lewis (Mathematics is Megethology [1993], p.03)

p.03	p.3	10808	Mathematics is generalisations about singleton functions
			Full Idea: We can take the theory of singleton functions, and hence set theory, and hence mathematics, to consist of generalisations about all singleton functions.
			From: David Lewis (Mathematics is Megethology [1993], p.03)
			A reaction: At first glance this sounds like a fancy version of the somewhat discredited Greek idea that mathematics is built on the concept of a 'unit'.

p.05	p.5	10809	We can accept the null set, but not a null class, a class lacking members
			Full Idea: In my usage of 'class', there is no such things as the null class. I don't mind calling some memberless thing - some individual - the null set. But that doesn't make it a memberless class. Rather, that makes it a 'set' that is not a class.
			From: David Lewis (Mathematics is Megethology [1993], p.05)
			A reaction: Lewis calls this usage 'idiosyncratic', but it strikes me as excellent. Set theorists can have their vital null class, and sensible people can be left to say, with Lewis, that classes of things must have members.

p.07	p.7	10810	I say that absolutely any things can have a mereological fusion
			Full Idea: I accept the principle of Unrestricted Composition: whenever there are some things, no matter how many or how unrelated or how disparate in character they may be, they have a mereological fusion. ...The trout-turkey is part fish and part fowl.
			From: David Lewis (Mathematics is Megethology [1993], p.07)
			A reaction: This nicely ducks the question of when things form natural wholes and when they don't, but I would have thought that that might be one of the central issues of metaphysicals, so I think I'll give Lewis's principle a miss.

p.09	p.9	10812	The null set is not a little speck of sheer nothingness, a black hole in Reality
			Full Idea: Should we accept the null set as a most extraordinary individual, a little speck of sheer nothingness, a sort of black hole in the fabric of Reality itself? Not that either, I think.
			From: David Lewis (Mathematics is Megethology [1993], p.09)
			A reaction: Correct!

p.09	p.9	10811	The null set plays the role of last resort, for class abstracts and for existence
			Full Idea: The null set serves two useful purposes. It is a denotation of last resort for class abstracts that denote no nonempty class. And it is an individual of last resort: we can count on its existence, and fearlessly build the hierarchy of sets from it.
			From: David Lewis (Mathematics is Megethology [1993], p.09)
			A reaction: This passage assuages my major reservation about the existence of the null set, but at the expense of confirming that it must be taken as an entirely fictional entity.

p.12	p.12	10813	What on earth is the relationship between a singleton and an element?
			Full Idea: A new student of set theory has just one thing, the element, and he has another single thing, the singleton, and not the slightest guidance about what one thing has to do with the other.
			From: David Lewis (Mathematics is Megethology [1993], p.12)

p.13	p.13	10814	Are all singletons exact intrinsic duplicates?
			Full Idea: Are all singletons exact intrinsic duplicates?
			From: David Lewis (Mathematics is Megethology [1993], p.13)

p.16	p.16	10815	We don't need 'abstract structures' to have structural truths about successor functions
			Full Idea: We needn't believe in 'abstract structures' to have general structural truths about all successor functions.
			From: David Lewis (Mathematics is Megethology [1993], p.16)

p.18	p.18	10816	We can use mereology to simulate quantification over relations
			Full Idea: We can simulate quantification over relations using megethology. Roughly, a quantifier over relations is a plural quantifier over things that encode ordered pairs by mereological means.
			From: David Lewis (Mathematics is Megethology [1993], p.18)
			A reaction: [He credits this idea to Burgess and Haven] The point is to avoid second-order logic, which quantifies over relations as ordered n-tuple sets.

1994	Lewis: reduction of mind (on himself)

p.412	p.412	3990	The whole truth supervenes on the physical truth
			Full Idea: The whole truth supervenes on the physical truth.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.412)
			A reaction: This seems to me the central truth about brains, and we should not be lured into abandoning it. We should not, however, exclude the possibility that there is a non-physical reality.

p.412	p.412	3989	I am a reductionist about mind because I am an a priori reductionist about everything
			Full Idea: My reductionism about mind began as part of an a priori reductionism about everything.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.412)
			A reaction: He says this is 'a priori' to avoid giving hostages to fortune, but I think is the best explanation of the total evidence facing us

p.414	p.414	3991	Where pixels make up a picture, supervenience is reduction
			Full Idea: In the case of millions of pixels making up a picture on a computer screen, the supervenience is reduction.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.414)
			A reaction: Since 'supervenience' seems a suspect relationship about which no one is clear, this is a point very much worth making.

p.416	p.416	3992	Folk psychology makes good predictions, by associating mental states with causal roles
			Full Idea: Folk psychology is a powerful instrument of prediction, …which associates with each mental state a typical causal role.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.416)
			A reaction: This seems a good account of why we should take folk psychology very seriously, even if it is sometimes wrong (e.g. about people who are mentally ill).

p.419	p.419	3993	Arguments are nearly always open to challenge, but they help to explain a position rather than force people to believe
			Full Idea: Philosophical arguments are never incontrovertible - well, hardly ever. Their purpose is to help expound a position, not to coerce agreement.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.419)
			A reaction: A bit over-cautious, perhaps. Most philosophers are converted to a position when they hear a single key argument, though it is probably 'tipping the balance' of previous discussions.

p.420	p.420	3994	Human pain might be one thing; Martian pain might be something else
			Full Idea: Human pain might be one thing. Martian pain might be something else.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.420)
			A reaction: A key suggestion in support of type-type physicalism, and against the multiple realisability objection to the identity theory

p.421	p.421	3995	A mind is an organ of representation
			Full Idea: A mind is an organ of representation.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.421)
			A reaction: This does not seem to necessarily involve awareness, so it seems to put intentionality at the centre of things. It is a good slogan.

p.422	p.422	3996	Folk psychology doesn't say that there is a language of thought
			Full Idea: I don't believe that folk psychology says there is a language of thought.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.422)
			A reaction: This is aimed at Jerry Fodor. Certainly folk psychology is a strong theory, but a so-called 'language of thought' (the brain's machine code) seems a much weaker one.

p.424	p.424	3998	If you don't share an external world with a brain-in-a-vat, then externalism says you don't share any beliefs
			Full Idea: If the famous brain in a bottle is your exact duplicate in brain states, but only experiences the computer's virtual reality, so that you share no objects of acquaintance, then according to externalists you share no beliefs whatsoever.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.424)
			A reaction: A very nice reductio ad absurdum of the idea that all concepts and beliefs have external meaning.

p.424	p.424	3997	Nothing shows that all content is 'wide', or that wide content has logical priority
			Full Idea: There is nothing to support the thesis that wide content is the only kind of content, or that it is any way pre-eminent or basic.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.424)
			A reaction: The idea that all content is 'wide' seems quite wrong. We can't all be wrong about the meaning of a word, because the underlying facts have not yet been discovered.

p.425	p.425	3999	A spontaneous duplicate of you would have your brain states but no experience, so externalism would deny him any beliefs
			Full Idea: According to externalists, Davidson's 'swampman' is your exact duplicate in brains states, but hasn't had time to become acquainted with much, so he has virtually no beliefs.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.425)
			A reaction: An implausible fantasy, but it does highlight the fact that beliefs and concepts are primarily internal states.

p.430	p.430	4000	Wide content derives from narrow content and relationships with external things
			Full Idea: Wide content is derivative, a product of narrow content and relationships of acquaintance with external things.
			From: David Lewis (Lewis: reduction of mind (on himself) [1994], p.430)
			A reaction: I would say: content is a mental state, but it is created and fixed by a community, and wide content is the part fixed by experts in the community. We can all be wrong about meanings, and occasionally most of us are wrong about a specialised meaning.

1995	Should a materialist believe in qualia?

p.327	p.327	8209	Part of the folk concept of qualia is what makes recognition and comparison possible
			Full Idea: The concept of qualia (a part of the folk concept) is the concept of properties of experiences apt for causing abilities to recognize and to imagine experiences of the same type.
			From: David Lewis (Should a materialist believe in qualia? [1995], p.327)
			A reaction: I presume the other part of the folk concept would be what it is about qualia that makes this possible, namely that they 'look/sound/feel.. the same'. Lewis emphasises the functional aspect, which could not possibly be the whole story.

1996	Elusive Knowledge

	p.68	12895	Knowing is context-sensitive because the domain of quantification varies
			Full Idea: The context-sensitivity of 'knows' is a function of contextual restrictions on the domain of quantification.
			From: report of David Lewis (Elusive Knowledge [1996]) by Stewart Cohen - Contextualism Defended p.68
			A reaction: I think the shifting 'domain of quantification' is one of the most interesting features of ordinary talk. Or, more plainly. 'what are you actually talking about?' is the key question in any fruitful dialogue. Sophisticated speakers tacitly shift domain.

	p.80	19562	We have knowledge if alternatives are eliminated, but appropriate alternatives depend on context
			Full Idea: S knows P if S's evidence eliminates every alternative. But the nature of the alternatives depends on context. So for Lewis, the context sensitivity of 'knows' is a function of contextual restrictions ln the domain of quantification.
			From: report of David Lewis (Elusive Knowledge [1996]) by Stewart Cohen - Contextualism Defended (and reply) 1
			A reaction: A typical modern attempt to 'regiment' a loose term like 'context'. That said, I like the idea. I'm struck by how the domain varies during a conversation (as in 'what we are talking about'). Domains standardly contain 'objects', though.

	p.421	12898	Justification is neither sufficient nor necessary for knowledge
			Full Idea: I don't agree that the mark of knowledge is justification, first because justification isn't sufficient - your true opinion that you will lose the lottery isn't knowledge, whatever the odds; and also not necessary - for what supports perception or memory?
			From: David Lewis (Elusive Knowledge [1996])
			A reaction: I don't think I agree. The point about the lottery is that an overwhelming reason will never get you to knowing that you won't win. But good reasons are coherent, not statistical. If perceptions are dubious, justification must be available.

p.419	p.419	12897	To say S knows P, but cannot eliminate not-P, sounds like a contradiction
			Full Idea: If you claim that S knows that P, and yet grant that S cannot eliminate a certain possibility of not-P, it certainly seems as if you have granted that S does not after all know that P. To speak of fallible knowledge just sounds contradictory.
			From: David Lewis (Elusive Knowledge [1996], p.419)
			A reaction: Starting from this point, fallibilism seems to be a rather bold move. The only sensible response seems to be to relax the requirement that not-P must be eliminable. Best: in one epistemic context P, in another not-P.

p.429	p.429	12899	The timid student has knowledge without belief, lacking confidence in their correct answer
			Full Idea: I allow knowledge without belief, as in the case of the timid student who knows the answer but has no confidence that he has it right, and so does not believe what he knows.
			From: David Lewis (Elusive Knowledge [1996], p.429)
			A reaction: [He cites Woozley 1953 for the timid student] I don't accept this example (since my views on knowledge are rather traditional, I find). Why would the student give that answer if they didn't believe it? Sustained timid correctness never happens.

1997	Finkish dispositions

I	p.134	15461	A 'finkish' disposition is real, but disappears when the stimulus occurs
			Full Idea: A disposition which would straight away vanish if put to the test is called 'finkish'. A finkishly fragile thing is fragile so long as it is not struck. But if it were struck, it would straight away cease to be fragile, and it would not break.
			From: David Lewis (Finkish dispositions [1997], I)
			A reaction: There are also 'antidotes'. Finks kill the disposition, antidotes kill the effect. These cases are problems for the simple conditional analysis of a disposition - because we never achieved the consequent.

I	p.136	15462	Backtracking counterfactuals go from supposed events to their required causal antecedents
			Full Idea: 'Backtracking' counterfactual reasoning runs from a counterfactually supposed event to the causal antecedents it would have to have had.
			From: David Lewis (Finkish dispositions [1997], I)
			A reaction: Why not call it a 'transcendental' counterfactual? Presumably you go thisworld>> counterfactualevent>> worldneededtocauseit. It conjures up two possible worlds instead of one.

II	p.140	15463	All dispositions must have causal bases
			Full Idea: Prior, Pargetter and Jackson have argued convincingly for the thesis that all dispositions must have causal bases.
			From: David Lewis (Finkish dispositions [1997], II)
			A reaction: [Their paper is 1982] This key thesis is tackled by modern defenders of powers. The question is not who has the best arguments, but who offers the most coherent picture. What is a 'causal basis'? What sort of thing could be primitive or fundamental?

II	p.144	15464	The distinction between dispositional and 'categorical' properties leads to confusion
			Full Idea: To avoid the danger of claiming that dispositions are their own categorical bases, we do better to eschew the alleged distinction between dispositional and 'categorical' properties altogether.
			From: David Lewis (Finkish dispositions [1997], II)
			A reaction: Since I have been unable to form any intuitive notion of what a 'categorical' property is, I like this, though not necessarily for his reason.

1998	Defining 'Intrinsic' (with Rae Langton)

	p.272	15400	We must avoid circularity between what is intrinsic and what is natural
			Full Idea: Lewis revised his analysis of duplication because he had assumed that as a matter of necessity perfectly natural properties are intrinsic, and that necessarily how a thing is intrinsically is determined completely by the natural properties it has.
			From: report of David Lewis (Defining 'Intrinsic' (with Rae Langton) [1998]) by Ross P. Cameron - Intrinsic and Extrinsic Properties 'Analysis'
			A reaction: [This compares Lewis 1986:61 with Langton and Lewis 1998] I am keen on both intrinsic and on natural properties, but I have not yet confronted this little problem. Time for a displacement activity, I think....

IV	p.121	15458	A property is 'intrinsic' iff it can never differ between duplicates
			Full Idea: A property is 'intrinsic' iff it never can differ between duplicates; iff whenever two things (actual or possible) are duplicates, either both of them have the property or both of them lack it.
			From: David Lewis (Defining 'Intrinsic' (with Rae Langton) [1998], IV)
			A reaction: This leaves me wondering how one could arrive at a precise definition of 'duplicates'. Can it be done without mentioning that they have the same intrinsic properties?

IV	p.121	15457	Interdefinition is useless by itself, but if we grasp one separately, we have them both
			Full Idea: All circles of interdefinition are useless by themselves. But if we reach one of the interdefined pair, then we have them both.
			From: David Lewis (Defining 'Intrinsic' (with Rae Langton) [1998], IV)

V	p.122	15459	Ellipsoidal stars seem to have an intrinsic property which depends on other objects
			Full Idea: The property of being an ellipsoidal star would seem offhand to be a basic intrinsic property, but it is incompatible (nomologically) with being an isolated object.
			From: David Lewis (Defining 'Intrinsic' (with Rae Langton) [1998], V)
			A reaction: Another nice example from Lewis. It makes you wonder whether the intrinsic/extrinsic distinction should go. Modern physics, with its 'entanglements', doesn't seem to suit the distinction.

1998	A world of truthmakers?

p.220	p.220	15549	If it were true that nothing at all existed, would that have a truthmaker?
			Full Idea: If there was absolutely nothing at all, then it would have been true that there was nothing. Would there have been a truthmaker for this truth?
			From: David Lewis (A world of truthmakers? [1998], p.220)
			A reaction: This is a problem for Lewis's own claim that 'truth supervenes on being', as well as the more restricted truthmakers invoked by Armstrong.

2001	Forget the 'correspondence theory of truth'

	p.279	10847	Truthmakers are about existential grounding, not about truth
			Full Idea: Instances of the truthmaker principle are equivalent to biconditionals not about truth but about the existential grounding of all manner of other things; the flying pigs, or what-have-you.
			From: David Lewis (Forget the 'correspondence theory of truth' [2001])
			A reaction: The question then is what the difference is between 'existential grounding' and 'truth'. There wouldn't seem to be any difference at all if the proposition in question was a simple existential claim.

p.276	p.276	10845	To be true a sentence must express a proposition, and not be ambiguous or vague or just expressive
			Full Idea: Sentences or assertions can be derivately called true, if they succeed in expressing determinate propositions. A sentence can be ambiguous or vague or paradoxical or ungrounded or not declarative or a mere expression of feeling.
			From: David Lewis (Forget the 'correspondence theory of truth' [2001], p.276)
			A reaction: Lewis has, of course, a peculiar notion of what a proposition is - it's a set of possible worlds. I, with my more psychological approach, take a proposition to be a particular sort of brain event.

p.277	p.277	10846	Truthmaker is correspondence, but without the requirement to be one-to-one
			Full Idea: The truthmaker principle seems to be a version of the correspondence theory of truth, but differs mostly in denying that the correspondence of truths to facts must be one-to-one.
			From: David Lewis (Forget the 'correspondence theory of truth' [2001], p.277)
			A reaction: In other words, several different sentences might have exactly the same truthmaker.

2003	Things qua Truthmakers

p.26	p.91	14401	Every proposition is entirely about being
			Full Idea: Every proposition, no matter what subject matter it may have, is entirely about being.
			From: David Lewis (Things qua Truthmakers [2003], p.26), quoted by Trenton Merricks - Truth and Ontology 4.VI
			A reaction: Since Lewis's propositions are sets of possible worlds, I presume this claim is made because proposition relates to all of the possible worlds, which therefore presumably embody Being.