"

Some fundamental concepts of logic

If you encountered propositional logic before, you may know that there are three large classes of formulas one may distinguish. Usually, these are called contingent, tautology, and contradiction. In our semantics, all this means is the following:

Definition 5.1. A formula XX of β„’0\mathcal{L}_0 is a tautology iff for every structure 𝐒\mathbf{S}, π’βŠ¨X\mathbf{S} \models X. A formula XX is a contradiction iff there is no structure 𝐒\mathbf{S} such that π’βŠ¨X\mathbf{S} \models X. Finally, a formula XX is contingent if it is neither a tautology, nor a contradiction.

Remark 5.1. Since tautologies are true in every structure, and contradictions are true in no structure, contingent formulas are those that are true in some structures, false in others.

Now one thing to note at the beginning is that atomic formulas are all contingent. That is, for each atomic formula PP, there is a structure 𝐒\mathbf{S} such that π’βŠ¨P\mathbf{S} \models P, and there is a structure 𝐒′\mathbf{S}' such that π’β€²βŠ¨ΜΈP\mathbf{S}' \not\models P. On the other hand, some complex formulas are contingent, some are tautologies, and some are contradictions.

What makes a formula a tautology or a contradiction? It is the way the connectives work. Notice that by definition, each connective must obey some specific rules in any structure regarding the truth-values of the formulas on which it operates. For example, X∧YX \wedge Y is true in any 𝐒\mathbf{S} iff XX and YY are both true individually in 𝐒\mathbf{S}. However, some formulas may say things that go against these rules. This means that they won’t be true in any structure 𝖲\mathsf{S}, since again, every structure must abide by these rules, while the formula says otherwise.

Let’s take the simplest example: X∧¬XX \wedge \neg X. In fact, this is not a formula of the language, but a class of formulas of form X∧¬XX \wedge \neg X. But as it turns out, no matter what formula you take for XX, it will always be the case that for any structure 𝐒\mathbf{S}, π’βŠ¨ΜΈX∧¬X\mathbf{S} \not\models X \wedge \neg X. In other words, this formula schema cannot have a true instance.

Now again, the reason for this is the way structures are defined. In particular, in every structure 𝐒\mathbf{S}, either π’βŠ¨X\mathbf{S} \models X, or π’βŠ¨ΜΈX\mathbf{S} \not\models X. This is sometimes called the Law of Excluded Middle, and it falls out of our definition of a structure. What this means is that every formula XX is either true in a structure, or false in a structure, and not a third thing (the β€˜excluded middle’).

We also have that it is never the case that π’βŠ¨X\mathbf{S} \models X and π’βŠ¨ΜΈX\mathbf{S} \not\models X. This is sometimes called the Law of Non-contradiction. Again, this falls out of our definition of what it means for something to be a structure. Now adding these two laws together, you simply get that every formula XX is either true or false, and not both or neither in a structure.

So take X∧¬XX \wedge \neg X. By the above, either π’βŠ¨X\mathbf{S} \models X, or π’βŠ¨Β¬X\mathbf{S} \models \neg X. Those are the only two options we have. We can consider each of them in turn:

  1. Suppose π’βŠ¨X\mathbf{S} \models X. Then, by definition of Β¬\neg, π’βŠ¨ΜΈX\mathbf{S} \not\models X. But π’βŠ¨X∧¬X\mathbf{S} \models X \wedge \neg X iff π’βŠ¨X\mathbf{S} \models X, and π’βŠ¨Β¬X\mathbf{S} \models \neg X, by definition of ∧\wedge. So since the latter half fails, π’βŠ¨ΜΈX∧¬X\mathbf{S} \not\models X \wedge \neg X.

  2. Suppose π’βŠ¨ΜΈX\mathbf{S} \not\models X. Then, by definition of Β¬\neg, π’βŠ¨Β¬X\mathbf{S} \models \neg X in this case. But again, π’βŠ¨X∧¬X\mathbf{S} \models X \wedge \neg X iff π’βŠ¨X\mathbf{S} \models X, and π’βŠ¨Β¬X\mathbf{S} \models \neg X, by definition of ∧\wedge. Moreover, in this case, we already assumed that π’βŠ¨ΜΈX\mathbf{S} \not\models X. So once again, π’βŠ¨ΜΈX∧¬X\mathbf{S} \not\models X \wedge \neg X.

This means that there is no structure 𝐒\mathbf{S} in which X∧¬XX \wedge \neg X is true. So every instance of X∧¬XX \wedge \neg X is false in every structure, no matter what you take XX to be.

At first, this type of talk might be confusing, but all we did was precisely represent what is usually illustrated by the following truth table:

XX ¬X\neg X X∧¬XX \wedge \neg X
TT FF FF
FF TT FF

In particular, a formula is contradictory, as represented in a truth table, if all values in the column underneath the formula are false. Again, the truth table rows just represent two large (jointly exhaustive) classes of structures, one in which XX is true, the other in which XX is false, and then show that neither of these classes make X∧¬XX \wedge \neg X true.

It will be important in what’s to come that every contradictory formula hides a claim of the form X∧¬XX \wedge \neg X for some formula XX. In fact, every contradictory formula hides a claim of the form P∧¬PP \wedge \neg P, for some atomic formula PP. In other words, the only contradictory formulas are those that try to say that something both is and is not the case (though not necessarily in exactly those words).

On the other hand, we may find formulas of the form X∨¬XX \vee \neg X, which are not contradictory, but the exact opposite, tautological. Again, the reason is because of how the connectives are defined in a structure. Again, in every structure, we have that either π’βŠ¨X\mathbf{S} \models X, or π’βŠ¨ΜΈX\mathbf{S} \not\models X. So:

  1. If π’βŠ¨X\mathbf{S} \models X, then π’βŠ¨X∨¬X\mathbf{S} \models X \vee \neg X, since by definition of ∨\vee, S⊨X∨¬XS \models X \vee \neg X iff it models either XX or Β¬X\neg X.

  2. If π’βŠ¨ΜΈX\mathbf{S} \not \models X, then by definition of Β¬\neg, π’βŠ¨Β¬X\mathbf{S} \models \neg X, so by definition of ∨\vee, π’βŠ¨X∨¬X\mathbf{S} \models X \vee \neg X (because of the latter half of the disjunction now).

Thus, X∨¬XX \vee \neg X must be true in every structure 𝐒\mathbf{S}, no matter which formula we take for the variable XX. And again, this is just an explanation of the following truth-table:

XX ¬X\neg X X∨¬XX \vee \neg X
TT FF TT
FF TT TT

As you can see, unlike with contradictions, tautologies have all values as true in the column underneath them.

Now we can formulate one of the most significant connections between tautologies and contradictions. As such, we will give it the proper form of a theorem.

Proposition 5.1. A formula XX of β„’0\mathcal{L}_0 is a tautology iff its negation Β¬X\neg X is a contradiction.

Proof. From left to right, suppose XX is a tautology. Then, for every 𝐒\mathbf{S}, π’βŠ¨X\mathbf{S} \models X. By definition, π’βŠ¨X\mathbf{S} \models X iff π’βŠ¨ΜΈΒ¬X\mathbf{S} \not\models \neg X. So since every π’βŠ¨X\mathbf{S} \models X, no π’βŠ¨Β¬X\mathbf{S} \models \neg X. By definition, Β¬X\neg X is then a contradiction.

From right to left, suppose Β¬X\neg X is a contradiction. Then, there is no 𝐒\mathbf{S} such that π’βŠ¨Β¬X\mathbf{S} \models \neg X. So every 𝐒\mathbf{S} is such that π’βŠ¨ΜΈΒ¬X\mathbf{S} \not\models \neg X. But any π’βŠ¨ΜΈΒ¬X\mathbf{S} \not\models \neg X iff π’βŠ¨X\mathbf{S} \models X. So every π’βŠ¨X\mathbf{S} \models X, and thus XX is a tautology.Β β—»

What is the symbol ‘β—»’? This symbol is used at the end of proofs of propositions as a sign that the proposition one set out to prove was indeed shown to hold. The practice of signifying the end of a proof in mathematics has a long history. Before boxes, mathematicians used to write ‘Q.E.D.’, the initials for the Latin phrase ‘quod erat demonstrandum’, meaning ‘which was to be demonstrated’. As often, proofs are extremely terse in presentation, it is not always immediate on first read that what was to be demonstrated has been demonstrated. Accordingly, signifying the end of a proof marks the place from where one should start backtracking if they missed the point.

Remark 5.2. Notice the form of this proof. Since we are proving a biconditional (an iff statement), we need to prove that the left hand side entails the right hand side, and the right hand side entails the left hand side. In order to prove that one side entails the other, we have to assume that one side holds, and derive that the other side then must also hold. Thus, we assumed the left hand side and derived the right, then assumed the right hand side, and derived the left.

As before, inverting this biconditional by negating both sides also gets us a true biconditional. Thus:

Proposition 5.2. A formula XX of β„’0\mathcal{L}_0 is not a tautology iff its negation Β¬X\neg X is not a contradiction.

Here is an additional proposition that may help here:

Proposition 5.3. A formula XX of β„’0\mathcal{L}_0 is contingent iff Β¬X\neg X is contingent.

Proof. From left to right, if XX is contingent, it is true in some structure 𝐒\mathbf{S}, and false in some structure 𝐒′\mathbf{S}'. Now if π’βŠ¨X\mathbf{S} \models X, then π’βŠ¨ΜΈΒ¬X\mathbf{S} \not\models \neg X, so Β¬X\neg X is false in 𝐒\mathbf{S}. On the other hand, if π’β€²βŠ¨ΜΈX\mathbf{S}' \not\models X, then π’β€²βŠ¨Β¬X\mathbf{S}' \models \neg X, so Β¬X\neg X is true in 𝐒\mathbf{S}. So Β¬X\neg X is also contingent.

From right to left, the reasoning is almost exactly the same.Β β—»

Exercise 5.1. Finish the proof for Proposition 5.3.

Now given the above two propositions, we can see that if a formula XX of β„’0\mathcal{L}_0 is not a tautology, then Β¬X\neg X is not a contradiction, because if XX is not a tautology, it is either a contradiction or contingent, so Β¬X\neg X will either be a tautology (hence not a contradiction), or contingent (hence not a contradiction). And similarly the other way around.

In fact, the following also follows:

Proposition 5.4. Β¬X\neg X is a tautology iff XX is a contradiction.

Proof. By Proposition 5.1, YY is a tautology iff Β¬Y\neg Y is a contradiction. Here is the trick. Since YY can be any formula, we substitute Β¬X\neg X for YY. Thus, Β¬X\neg X is a tautology iff ¬¬X\neg \neg X is a contradiction. But ¬¬X\neg \neg X is a contradiction iff XX is also a contradiction. This is because for any 𝐒\mathbf{S}, if π’βŠ¨Β¬Β¬X\mathbf{S} \models \neg \neg X, then π’βŠ¨ΜΈΒ¬X\mathbf{S} \not\models \neg X, then π’βŠ¨X\mathbf{S} \models X (and similarly the other way around). So Β¬X\neg X is a tautology iff XX is a contradiction. Which is what we wanted to show.Β β—»

Some of what the above propositions tell us is summarized in Table 5.1.

Relations between tautologies, contradictions and contingencies
if, and only if
XX is tautology ⇔\Leftrightarrow Β¬X\neg X is contradiction
XX is contradiction ⇔\Leftrightarrow Β¬X\neg X is tautology
XX is contingent ⇔\Leftrightarrow Β¬X\neg X is contingent

Here is another thing that will be relevant presently. Suppose you have a formula XX, and it is not a tautology (so it is either a contradiction or contingent). Then, its negation Β¬X\neg X will not be a contradiction (it will either be a tautology or contingent). Thus, if XX is not a tautology, there is at least one structure 𝐒\mathbf{S} such that π’βŠ¨Β¬X\mathbf{S} \models \neg X, and vice versa. This is usually what is called a counterexample. In other words:

Proposition 5.5. XX is not a tautology iff there is a structure 𝐒\mathbf{S} such that π’βŠ¨Β¬X\mathbf{S} \models \neg X.

Exercise 5.2. Using the reasoning with counterexamples sketched above, show for each of the following formulas that they are not tautologies.

  1. (¬Y∧¬X)∧¬X(\neg Y \wedge \neg X) \wedge \neg X

  2. (X→Y)→X(X \rightarrow Y) \rightarrow X

  3. (X→Y)→Y(X \rightarrow Y) \rightarrow Y

  4. (X∨Y)β†’X(X \vee Y) \rightarrow X

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Learning Logic Backwards Copyright © by Peter Susanszky is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.