We will now leave the English language behind and concentrate on logical statements consisting only of variables and connectives. Keep in mind that variables are not limited to representing simple statements; they can represent compound statements as well.
Given a logical statement, view the variables as inputs and the truth value of the entire statement as an output. We would like a systematic way to determine how the truth value of the output changes as we vary the truth values of the inputs.
If a statement involves a finite number of variables, then since each variable can have one of only two possible truth values, there are a finite number of different combinations of input truth values for the statement. So we can test each combination one after the other to determine all possible outputs. Arrange this analysis in a table with all possibilities for the input variables on the left and the resulting outputs on the right.
The first two rows of the truth table for are consistent with the reading โif is true then is also true.โ Really, this reading of the conditional says nothing in the case that is actually false, but we cannot leave the โoutputโ column of the truth table blank for the corresponding rows where . Instead, the outputs in the last two rows of the truth table for are โdefaultโ values chosen to avoid inconsistencies. (See Exercise 1.6.3.)
Biconditional reads โ is true when is true, and only when is true.โ But this means that when is false, cannot be true, hence must by false, which explains the fourth row of the truth table.
We can now use the truth tables of the basic connectives to analyze more complicated statements. Liberal use of extra โintermediateโ columns to analyze substatements separately is highly recommended.