1

Verify that $\bigl(p\to(q\to r)\bigr)\to \bigl((p\to q)\to (p\to r)\bigr)$ is a tautology.

I am confused on this whole tautology even after looking at examples both in my book and on-line. I started a truth table and this is what I have so far. Can someone please explain this to me in a way that I can understand.

\begin{array}{ccc} p & q & r & ? & ? & ? & ?\\ \hline T & T & T\\ T & T & F\\ T & F & T\\ T & F & F\\ F & T & T\\ F & T & F\\ F & F & T\\ F & F & F \end{array}

dfeuer
  • 9,069
  • 3
    Do you have any idea what the other columns of the truth table should represent, i.e., what the column headers should be? – Brian M. Scott Dec 16 '13 at 07:12
  • The first statement means that $p$ plus $q$ imply $r$, so if $p$ implies $q$, then $p$ implies $p$ plus $q$ which implies $r$. Hope my English is good enough for understanding. – asatzhh Dec 16 '13 at 07:19
  • @BrianM.Scott, are $\implies$ and $\to$ really different here? I'm a bit of a logic newbie. – dfeuer Dec 16 '13 at 07:54
  • @dfeuer: They are: $\to$ is the usual formal connective, going along with $\land,\lor,\leftrightarrow$, and $\neg$ in the standard repertoire. (It also has the virtue of matching the OP’s original notation.) – Brian M. Scott Dec 16 '13 at 07:56
  • @BrianM.Scott, so $\implies$ and $\iff$ are what, then? They are also used with $\land$, $\lor$, and $\lnot$.... – dfeuer Dec 16 '13 at 07:58
  • @dfeuer: Typically used at the meta level or informally, rather than within the formal language. – Brian M. Scott Dec 16 '13 at 07:59
  • NO i DO NOT THIS IS ALL THAT i WAS GIVEN. I SAW THAT TOO AND WAS JUST THINKING OF PLUGGING THEM IN MYSELF. – HammerTime Dec 16 '13 at 08:22

5 Answers5

2

A completely different approach (called a tableau proof, and a bit on the informal side for one). Feel free to ignore.

1) Assume $p \implies (q \implies r)$.

2) Assume $p\implies q$.

<blockquote>
  <p>3) Assume $p$.</p>

  <blockquote>
    <p>4) By (2) and (3) and modus ponens, $q$.</p>

    <p>5) By (1) and (3) and modus ponens, $q \implies r$.</p>

    <p>6) By (4) and (5) and modus ponens, $r$.</p>
  </blockquote>

  <p>7) By (3)&ndash;(6) and implication, $p \implies r$.</p>
</blockquote>

<p>8) By (2)&ndash;(7) and implication, $(p\implies q)\implies (p\implies r)$.</p>

9) By (1)–(8) and implication, $\bigl(p\implies (q\implies r)\bigr)\implies \bigl((p\implies q)\implies (p\implies r)\bigr)$.

dfeuer
  • 9,069
  • This doesn't show that [(p⟹(q⟹r))⟹((p⟹q)⟹(p⟹r))] is a tautology. It only shows that if one assumes "implication" as a rule of inference along with modus ponens, that we can derive [(p⟹(q⟹r))⟹((p⟹q)⟹(p⟹r))]. – Doug Spoonwood Oct 19 '14 at 12:30
  • @DougSpoonwood, people usually accept an introduction rule to go along with an elimination rule. – dfeuer Oct 19 '14 at 14:34
2

The following solution does not use a full truth table, so may not be the intended solution.

How could $\bigl(p\to(q\to r)\bigr)\to \bigl((p\to q)\to (p\to r)\bigr)$ be false?

The sentence $\bigl(p\to(q\to r)\bigr)$ would have to be true, and the sentence $\bigl((p\to q)\to (p\to r)\bigr)$ would have to be false.

How could $\bigl((p\to q)\to (p\to r)\bigr)$ be false? The sentence $p\to q$ would have to be true, and $p\to r$ would have to be false. To make $p\to r$ false, we must have $p$ true and $r$ false. But if $p$ is true, to make $p\to q$ true, we need $q$ true.

To sum up, the only way that $\bigl((p\to q)\to (p\to r)\bigr)$ can be false if $p$ and $q$ are true and $r$ is false.

However, under these conditions, $\bigl(p\to(q\to r)\bigr)$ is false.

Thus there is no assignment of truth values to $p$, $q$, and $r$ such that $\bigl((p\to q)\to (p\to r)\bigr)$ is false and $\bigl(p\to(q\to r)\bigr)$ is true.

André Nicolas
  • 507,029
  • That is what I was thinking as well and assume that this cannot be verified as a tautology. – HammerTime Dec 16 '13 at 08:34
  • It looks as if the above is not clear enough. It is a proof that the sentence is a tautology. Instead of checking line by line in the truth table whether it was T for all possible assignments of truth values to $p,q,r$, the above showed there was no truth assignment under which the sentence is F. – André Nicolas Dec 16 '13 at 15:47
  • The direct proof in my answer seems less mind-twisty. Do you know a more readable way to present it? – dfeuer Dec 16 '13 at 17:11
  • Your solution is good, efficient. The only issue might be that it is logic-oriented rather than truth-table oriented. – André Nicolas Dec 16 '13 at 17:20
  • @AndréNicolas If there is no truth assignment under which the sentence is F, and there exists only two truth values, then under all truth assignments, the sentences is not F. Not F is true, and consequently his argument ends up working. – Doug Spoonwood Dec 18 '13 at 22:55
  • The argument here assumes that [(p→(q→r))→((p→q)→(p→r))] is false, and then derives that [(p→(q→r))→((p→q)→(p→r))] is true under the scope of that assumption. The assumption then can get discharged into a conditional. Finally, the law of Clavius CCNppp can get used to infer the result. – Doug Spoonwood Oct 19 '14 at 12:47
2

A great way to see results like this is with the help of the method of analytic tableaux. You take the negation of your formula and apply a series of contradiction-hunting rules to get

enter image description here,

which is closed (meaning that each of its paths end in contradictions), which in turn implies that your original formula is indeed a tautology.

Can you see how this relates to the tables in the other answers?

Shaun
  • 44,997
1

A tautology is a statement that is always true. For propositions $a$ and $b$, $a\rightarrow b$ is logically equivalent to $\neg a\vee b$. If $a=b$, then $\neg a\vee b$ is logically equivalent to $\neg b\vee b$, which is a tautology. So one way to prove that the given statement is a tautology is to consider $$ a\equiv p\rightarrow\left(q\rightarrow r\right) $$ and $$ b\equiv\left(p\rightarrow q\right)\rightarrow\left(p\rightarrow r\right) $$ and show that $a$ and $b$ are logically equivalent. We can do this by checking the truth table:

\begin{array}{ccccc} p & q & r & p\rightarrow\left(q\rightarrow r\right) & \left(\left(p\rightarrow q\right)\rightarrow\left(p\rightarrow r\right)\right)\\ 0 & 0 & 0 & 1 & 1\\ 0 & 0 & 1 & 1 & 1\\ 0 & 1 & 0 & 1 & 1\\ 0 & 1 & 1 & 1 & 1\\ 1 & 0 & 0 & 1 & 1\\ 1 & 0 & 1 & 1 & 1\\ 1 & 1 & 0 & 0 & 0\\ 1 & 1 & 1 & 1 & 1 \end{array}

(0 is false and 1 is true in the above)

parsiad
  • 25,154
  • 1
    Note that the truth table can be expanded to show subexpressions. – dfeuer Dec 16 '13 at 07:51
  • It's not the same as yours but this is what I came up with. [p→(q→r)] T F F F F F F F I believe that this problem is not a tautology. As all statements should be true correct? – HammerTime Dec 16 '13 at 08:26
  • Nope, this proposition is a tautology. Perhaps you need to review what a truth table is, or review $\rightarrow$ ? – parsiad Dec 16 '13 at 08:41
  • Ok I will do that and thank you for the tip. – HammerTime Dec 16 '13 at 09:19
-1

Your well-formed formula is an instance of axioms schema 3 of the Łukasiewicz axiomatization of propositional logic. By the completeness of propositional logic, your wf is a tautology.

sdfsadf
  • 55
  • Completeness says that if |=$\alpha$, then $\vdash$$\alpha$. Soundness says that if $\vdash$$\alpha$, then |=$\alpha$. So, one could demonstrate that the self-distribution formula of the OP (sometimes called "Frege") is a tautology by invoking soundness. However, one won't know for which model it could qualify as sound. One won't know that it holds for two-valued logic. – Doug Spoonwood Oct 19 '14 at 12:22