Part I
Truth-Functional Logic
Chapter 1
The Basics
1 Deductively Valid Inference
There is much more to logic than the question of what makes inferences deductively valid or invalid, but to most people that is what logic is all about, so that is where we shall begin. One of the most basic features of these inferences is that they seem to be composed of declarative sentences, that is to say sentences which make assertions capable of being true or false. âBoris Yeltsin became President of Russia in 1993â, âAll hydrogen atoms have one proton in their nucleusâ and âMichelangelo painted the ceiling of the Sistine Chapelâ are declarative sentences, and (we believe) true ones at that. âShut that door!â and âIs Hanoi in Scotland?â are not. Neither is Chomskyâs funny example âColourless green ideas sleep furiouslyâ, which has the grammatical form, but only the form, of a fact-stating sentence.
So far, so good. The sentences composing an inference are its premises and conclusion, the latter usually signalled by the prefix âthereforeâ (for which, for brevity, we shall often use the symbol â´). If the inference is deductively valid the conclusion is called a deductive or logical consequence, or simply consequence, of the premises. What else do we know? Well, one of the most familiar facts about deductively valid inferences, and the one which probably goes farthest towards explaining the importance they have always been accorded, is that it is impossible for their conclusions to be false if their premises are true: if anything is basic to the notion of deduction, that surely is. Consider this example, known as a disjunctive syllogism:
Itâs raining or itâs snowing.
Itâs not raining.
â´ Itâs snowing.
Clearly, you donât have to know whether itâs actually raining or not, or snowing or not, to know that if the premises are true, so too is the conclusion. Not only is the conclusion true if the premises are. The conclusion must be true if the premises are true; there is no possibility of its being false.
Not only is this the most important property of deductively valid inferences; it is difficult to think of any other that has that same generality. This being so, we might as well take it as the defining property, and accordingly frame the following
Provisional definition: a valid deductive inference is one whose premises cannot all be true and conclusion false.
The definition is provisional because the word âcannotâ itself rather obviously needs a definition, and providing an adequate one is not trivial: most of this book will be occupied in the task. But one thing we do know is that âcannotâ here has nothing at all to do with empirical fact, as it does in the statement that water cannot unaided run uphill. âCannotâ in this context refers to a logical impossibility. It is a logical, not merely a physical, impossibility that âItâs snowingâ is false if both âItâs either raining or itâs snowingâ and âItâs not rainingâ are true (assuming sameness of spatio-temporal reference in premises and conclusion). Here are two more examples to consider.
If Lev is in Moscow then Irina is in Kiev.
Lev is in Moscow.
â´ Irina is in Kiev.
Cain was hairy and Abel was his victim.
â´ Cain was hairy.
It is intuitively clear that these remain deductively valid, in the sense of the definition above, whatever sentences are substituted for âCain was hairyâ, âAbel was his victimâ, âLev is in Moscowâ, âIrina is in Kievâ and, in the disjunctive syllogism, âItâs rainingâ and âitâs snowingâ. Another way of putting it is to say that if we replace these sentences by the letters A, B, C, D, E and F, the respective formal representations (or formalisations) of these inferences
E or F
not E
â´ F
If A then B
A
â´ B
C and D
â´ C
will always generate deductively valid inferences when the letters A, B, C, D, E and F are replaced by any sentences.
An explanation of why this is so will plausibly rest on an analysis of the logical role played by the particles âandâ, âorâ, ânotâ, and âif⌠thenââ. Now, a common method of analysing some phenomenon is to construct a model of it and see whether the model behaves in a way sufficiently resembling what it is supposed to model. This will be our procedure. The model, which will be presented in a systematic form in Chapter 3, is called a propositional language. âAndâ, âorâ, ânotâ, etc. are basic syntactical items of these languages, and in the following sections we shall describe the way they are used to form compound truth-functional sentences, and the rules which determine how truth and falsity should be ascribed to these. (The syntax of a language is the set of rules which determine its formal structure, that is to say the way its basic vocabulary is organised into well-formed expressions, among which are the sentences of the language; the rules by which the sentences are equipped with truthconditions constitute the languageâs semantics.)
2 Syntax: Connectives and the Principle of Composition
âAndâ, âorâ, âifâŚthenââ are structural items, called connectives by logicians, which articulate sentences into further sentences. âCain was hairyâ and âAbel was his victimâ are said to be conjoined by âandâ to yield the conjunction âCain was hairy and Abel was his victimâ; the two sentences forming the conjunction are its conjuncts. âNotâ operates on the sentence âItâs rainingâ to generate its negation âItâs not raining.â âItâs rainingâ and âitâs snowingâ are disjoined by âorâ to form the disjunction âItâs raining or itâs snowingâ of those two sentences, which are called the disjuncts. The sentences âLev is in Moscowâ and âIrina is in Kievâ are combined into the conditional sentence âIf Lev is in Moscow then Irina is in Kievâ; âLev is in Moscowâ is the antecedent, and âIrina is in Kievâ is the consequent.
These connectives play such a fundamental role that they have been given special symbols by logicians. The following are now standard:
Connective | Symbol |
not | ÂŹ |
and | ⧠|
or | ⨠|
ifâŚthen | â |
Because they operate on pairs of sentences to generate other sentences, â§, ⨠and â are binary connectives; ÂŹ is unary, because it operates on single sentences. In what follows we shall refer, as just now, to â§, â¨, and â directly as connectives rather than as connective symbols.
In our model the basic items out of which its sentences are built are these connectives and a stock of capital letters A, B, C, D, etc., called sentence letters, from the beginning of the Roman alphabet. These are intended to represent some given set of English sentences with whose internal structure we are not concerned. The sentence letters are often called the atomic sentences of the model, because all its other sentences are compounded from them, using the connectives. The first level of composition consists of the negations, disjunctions, conjunctions of sentence letters, and conditionals formed from them. Each of these compounds ...