31 ideas
14273 | Conditional Proof is only valid if we accept the truth-functional reading of 'if' [Edgington] |
Full Idea: Conditional Proof seems sound: 'From X and Y, it follows that Z. So from X it follows that if Y,Z'. Yet for no reading of 'if' which is stronger that the truth-functional reading is CP valid, at least if we accept ¬(A&¬B);A; therefore B. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.2) | |
A reaction: See the section of ideas on Conditionals (filed under 'Modality') for a fuller picture of this issue. Edgington offers it as one of the main arguments in favour of the truth-functional reading of 'if' (though she rejects that reading). |
10301 | The axiom of choice is controversial, but it could be replaced [Shapiro] |
Full Idea: The axiom of choice has a troubled history, but is now standard in mathematics. It could be replaced with a principle of comprehension for functions), or one could omit the variables ranging over functions. | |
From: Stewart Shapiro (Higher-Order Logic [2001], n 3) |
10588 | First-order logic is Complete, and Compact, with the Löwenheim-Skolem Theorems [Shapiro] |
Full Idea: Early study of first-order logic revealed a number of important features. Gödel showed that there is a complete, sound and effective deductive system. It follows that it is Compact, and there are also the downward and upward Löwenheim-Skolem Theorems. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.1) |
10298 | Some say that second-order logic is mathematics, not logic [Shapiro] |
Full Idea: Some authors argue that second-order logic (with standard semantics) is not logic at all, but is a rather obscure form of mathematics. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.4) |
10299 | If the aim of logic is to codify inferences, second-order logic is useless [Shapiro] |
Full Idea: If the goal of logical study is to present a canon of inference, a calculus which codifies correct inference patterns, then second-order logic is a non-starter. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.4) | |
A reaction: This seems to be because it is not 'complete'. However, moves like plural quantification seem aimed at capturing ordinary language inferences, so the difficulty is only that there isn't a precise 'calculus'. |
10300 | Logical consequence can be defined in terms of the logical terminology [Shapiro] |
Full Idea: Informally, logical consequence is sometimes defined in terms of the meanings of a certain collection of terms, the so-called 'logical terminology'. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.4) | |
A reaction: This seems to be a compositional account, where we build a full account from an account of the atomic bits, perhaps presented as truth-tables. |
10290 | Second-order variables also range over properties, sets, relations or functions [Shapiro] |
Full Idea: Second-order variables can range over properties, sets, or relations on the items in the domain-of-discourse, or over functions from the domain itself. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.1) |
10590 | Up Löwenheim-Skolem: if natural numbers satisfy wffs, then an infinite domain satisfies them [Shapiro] |
Full Idea: Upward Löwenheim-Skolem: if a set of first-order formulas is satisfied by a domain of at least the natural numbers, then it is satisfied by a model of at least some infinite cardinal. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.1) |
10292 | Downward Löwenheim-Skolem: if there's an infinite model, there is a countable model [Shapiro] |
Full Idea: Downward Löwenheim-Skolem: a finite or denumerable set of first-order formulas that is satisfied by a model whose domain is infinite is satisfied in a model whose domain is the natural numbers | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.1) |
10296 | The Löwenheim-Skolem Theorems fail for second-order languages with standard semantics [Shapiro] |
Full Idea: Both of the Löwenheim-Skolem Theorems fail for second-order languages with a standard semantics | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.3.2) |
10297 | The Löwenheim-Skolem theorem seems to be a defect of first-order logic [Shapiro] |
Full Idea: The Löwenheim-Skolem theorem is usually taken as a sort of defect (often thought to be inevitable) of the first-order logic. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.4) | |
A reaction: [He is quoting Wang 1974 p.154] |
10294 | Second-order logic has the expressive power for mathematics, but an unworkable model theory [Shapiro] |
Full Idea: Full second-order logic has all the expressive power needed to do mathematics, but has an unworkable model theory. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.1) | |
A reaction: [he credits Cowles for this remark] Having an unworkable model theory sounds pretty serious to me, as I'm not inclined to be interested in languages which don't produce models of some sort. Surely models are the whole point? |
10591 | Logicians use 'property' and 'set' interchangeably, with little hanging on it [Shapiro] |
Full Idea: In studying second-order logic one can think of relations and functions as extensional or intensional, or one can leave it open. Little turns on this here, and so words like 'property', 'class', and 'set' are used interchangeably. | |
From: Stewart Shapiro (Higher-Order Logic [2001], 2.2.1) | |
A reaction: Important. Students of the metaphysics of properties, who arrive with limited experience of logic, are bewildered by this attitude. Note that the metaphysics is left wide open, so never let logicians hijack the metaphysical problem of properties. |
14281 | A thing works like formal probability if all the options sum to 100% [Edgington] |
Full Idea: One's degrees of belief in the members of an idealised partition should sum to 100%. That is all there is to the claim that degrees of belief should have the structure of probabilities. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 3.1) |
14284 | Conclusion improbability can't exceed summed premise improbability in valid arguments [Edgington] |
Full Idea: If (and only if) an argument is valid, then in no probability distribution does the improbability of its conclusion exceed the sum of the improbabilities of its premises. We can call this the Probability Preservation Principle. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 3.2) | |
A reaction: [Ernest Adams is credited with this] This means that classical logic is in some way probability-preserving as well as truth-preserving. |
14270 | Simple indicatives about past, present or future do seem to form a single semantic kind [Edgington] |
Full Idea: Straightforward statements about the past, present or future, to which a conditional clause is attached - the traditional class of indicative conditionals - do (in my view) constitute a single semantic kind. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 1) | |
A reaction: This contrasts with Idea 14269, where the future indicatives are group instead with the counterfactuals. |
14269 | Maybe forward-looking indicatives are best classed with the subjunctives [Edgington] |
Full Idea: According to some theorists, the forward-looking 'indicatives' (those with a 'will' in the main clause) belong with the 'subjunctives' (those with a 'would' in the main clause), and not with the other 'indicatives'. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 1) | |
A reaction: [She cites Gibbard, Dudman and 1988 Bennett; Jackson defends the indicative/subjunctive division, and recent Bennett defends it too] It is plausible to say that 'If you will do x' is counterfactual, since it hasn't actually happened. |
14275 | Truth-function problems don't show up in mathematics [Edgington] |
Full Idea: The main defects of the truth-functional account of conditionals don't show up in mathematics. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.3) | |
A reaction: These problems are the paradoxes associated with the material conditional ⊃. Too often mathematical logic has been the tail that wagged the dog in modern philosophy. |
14274 | Inferring conditionals from disjunctions or negated conjunctions gives support to truth-functionalism [Edgington] |
Full Idea: If either A or B is true, then you are intuitively justified in believe that If ¬A, B. If you know that ¬(A&B), then you may justifiably infer that if A, ¬B. The truth-functionalist gets both of these cases (disjunction and negated conjunction) correct. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.1) | |
A reaction: [compressed version] This summarises two of Edgington's three main arguments in favour of the truth-functional account of conditions (along with the existence of Conditional Proof). It is elementary classical logic which supports truth-functionalism. |
14276 | The truth-functional view makes conditionals with unlikely antecedents likely to be true [Edgington] |
Full Idea: The truth-functional view of conditionals has the unhappy consequence that all conditionals with unlikely antecedents are likely to be true. To think it likely that ¬A is to think it likely that a sufficient condition for the truth of A⊃B obtains. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.3) | |
A reaction: This is Edgington's main reason for rejecting the truth-functional account of conditionals. She says it removes our power to discriminate between believable and unbelievable conditionals, which is basic to practical reasoning. |
14290 | Doctor:'If patient still alive, change dressing'; Nurse:'Either dead patient, or change dressing'; kills patient! [Edgington] |
Full Idea: The doctor says "If the patient is still alive in the morning, change the dressing". As a truth-functional command this says "Make it that either the patient is dead in the morning, or change the dressing", so the nurse kills the patient. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 5) | |
A reaction: Isn't philosophy wonderful? |
14271 | Non-truth-functionalist say 'If A,B' is false if A is T and B is F, but deny that is always true for TT,FT and FF [Edgington] |
Full Idea: Non-truth-functional accounts agree that 'If A,B' is false when A is true and B is false; and that it is sometimes true for the other three combinations of truth-values; but they deny that the conditional is always true in each of these three cases. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.1) | |
A reaction: Truth-functional connectives like 'and' and 'or' don't add any truth-conditions to the values of the propositions, but 'If...then' seems to assert a relationship that goes beyond its component propositions, so non-truth-functionalists are right. |
14272 | I say "If you touch that wire you'll get a shock"; you don't touch it. How can that make the conditional true? [Edgington] |
Full Idea: Non-truth-functionalists agree that when A is false, 'If A,B' may be either true or false. I say "If you touch that wire, you will get an electric shock". You don't touch it. Was my remark true or false? They say it depends on the wire etc. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.1) | |
A reaction: This example seems to me to be a pretty conclusive refutation of the truth-functional view. How can the conditional be implied simply by my failure to touch the wire (which is what benighted truth-functionalists seem to believe)? |
14282 | On the supposition view, believe if A,B to the extent that A&B is nearly as likely as A [Edgington] |
Full Idea: Accepting Ramsey's suggestion that 'if' and 'on the supposition that' come to the same thing, we get an equation which says ...you believe if A,B to the extent that you think that A&B is nearly as likely as A. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 3.1) |
14278 | Truth-functionalists support some conditionals which we assert, but should not actually believe [Edgington] |
Full Idea: There are compounds of conditionals which we confidently assert and accept which, by the lights of the truth-functionalist, we do not have reason to believe true, such as 'If it broke if it was dropped, it was fragile', when it is NOT dropped. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 2.5) | |
A reaction: [The example is from Gibbard 1981] The fact that it wasn't dropped only negates the nested antecedent, not the whole antecedent. I suppose it also wasn't broken, and both negations seem to be required. |
14287 | Does 'If A,B' say something different in each context, because of the possibiites there? [Edgington] |
Full Idea: A pragmatic constraint might say that as different possibilities are live in different conversational settings, a different proposition may be expressed by 'If A,B' in different conversational settings. | |
From: Dorothy Edgington (Conditionals (Stanf) [2006], 4.1) | |
A reaction: Edgington says that it is only the truth of the proposition, not its content, which changes with context. I'm not so sure. 'If Hitler finds out, we are in trouble' says different things in 1914 and 1944. |
8430 | Causal statements are used to explain, to predict, to control, to attribute responsibility, and in theories [Kim] |
Full Idea: The function of causal statements is 1) to explain events, 2) for predictive usefulness, 3) to help control events, 4) with agents, to attribute moral responsibility, 5) in physical theory. We should judge causal theories by how they account for these. | |
From: Jaegwon Kim (Causes and Counterfactuals [1973], p.207) | |
A reaction: He suggests that Lewis's counterfactual theory won't do well on this test. I think the first one is what matters. Philosophy aims to understand, and that is achieved through explanation. Regularity and counterfactual theories explain very little. |
8396 | Many counterfactuals have nothing to do with causation [Kim, by Tooley] |
Full Idea: Kim has pointed out that there are a number of counterfactuals that have nothing to do with causation. If John marries Mary, then if John had not existed he would not have married Mary, but that is not the cause of their union. | |
From: report of Jaegwon Kim (Causes and Counterfactuals [1973], 5.2) by Michael Tooley - Causation and Supervenience | |
A reaction: One might not think that this mattered, but it leaves the problem of distinguishing between the causal counterfactuals and the rest (and you mustn't mention causation when you are doing it!). |
8429 | Counterfactuals can express four other relations between events, apart from causation [Kim] |
Full Idea: Counterfactuals can express 'analytical' dependency, or the fact that one event is part of another, or an action done by doing another, or (most interestingly) an event can determine another without causally determining it. | |
From: Jaegwon Kim (Causes and Counterfactuals [1973], p.205) | |
A reaction: [Kim gives example of each case] Counterfactuals can even express a relation that involves no dependency. Or they might just involve redescription, as in 'If Scott were still alive, then the author of "Waverley" would be too'. |
8428 | Causation is not the only dependency relation expressed by counterfactuals [Kim] |
Full Idea: The sort of dependency expressed by counterfactual relations is considerably broader than strictly causal dependency, and causal dependency is only one among the heterogeneous group of dependency relationships counterfactuals can express. | |
From: Jaegwon Kim (Causes and Counterfactuals [1973], p.205) | |
A reaction: In 'If pigs could fly, one and one still wouldn't make three' there isn't even a dependency. Kim has opened up lines of criticism which make the counterfactual analysis of causation look very implausible to me. |
4781 | Many counterfactual truths do not imply causation ('if yesterday wasn't Monday, it isn't Tuesday') [Kim, by Psillos] |
Full Idea: Kim gives a range of examples of counterfactual dependence without causation, as: 'if yesterday wasn't Monday, today wouldn't be Tuesday', and 'if my sister had not given birth, I would not be an uncle'. | |
From: report of Jaegwon Kim (Causes and Counterfactuals [1973]) by Stathis Psillos - Causation and Explanation §3.3 | |
A reaction: This is aimed at David Lewis. The objection seems like commonsense. "If you blink, the cat gets it". Causal claims involve counterfactuals, but they are not definitive of what causation is. |