# A question on the truth-functional connectives

Hi guys I'm new to sentential logic and here is the problem:

Suppose that there is a truth-functional connective "#", and we only know that (P # (P # P)) is a tautology. Now we are asked to say something, if any, about the truth-table of "#".

The answer given is : The truth-value of (P # P) must be T. I cannot understand this thoroughly. What I have learned about the truth-functional connective is only its definition, which states that the truth-value of the new sentence constructed by it depends only on the truth-value of its components sentence. Then I know if we can write down the truth table for a connectives like "$\neg$", it would be a truth-functional connective since the value of "$\neg$P" depends only on P.

This structure (P # (P # P)) looks a bit complicated for me, and can anyone help me to explain its meaning?

• I'm not sure I believe that answer. What if $\#$ is the NAND operator, $P\#Q=\neg(P\land Q)$? Seems to me that $P\#(P\#P)$ is a tautology, because $T\#(T\#T)=T\#F=T$ and $F\#(F\#F)=F\#T=T$ but $P\#P$ is not a tautology because $T\#T=F.$ Or did I completely misunderstand the problem? – bof Jan 10 '17 at 4:08
• Maybe provide a name and page number of the textbook the problem is from? Your question appears to be lacking some context; or at least the redundant context a human often needs to understand a question. – user400188 Jan 10 '17 at 9:28

Once you know the definition of $\#$, you evaluate something like $(P \# (P \# P))$ just like other mathematical statements: first evaluate the inside $(P \# P)$, and once you know what that is (say, $(P\#P)$ evaluates to $Q$), then evaluate $(P \# Q)$.

For example, suppose # is defined by the following truth table:

\begin{array}{c|c|c} P & Q & (P \# Q) \\ \hline T & T & F \\ T & F & T \\ F & T & T \\ F & F & T \\ \end{array}

Then we can work out $(P \# (P \# P))$ as follows:

\begin{array}{c|c|c} P & (P \# P) & (P \# (P \# P))\\ \hline T & F & T \\ F & T & T \\ \end{array}

And so now we see that $(P \# (P \# P))$ is a tautology ... meaning that $(P \# (P \# P))$ can be a tautology without $(P \# P)$ being true. Indeed, I just defined the $\#$ as the NAND which was already pointed out to be a counterexample to the claim the book apparently makes.

OK, so is there maybe something else that we can say about the definition of $\#$? Let's consider all of the 16 possible definitions of $\#$:

\begin{array}{c|c|c|c|c|c|c|c|c|c|c|c|c|c|c|c|c} P & Q & \#_1 & \#_2 & \#_3 & \#_4 & \#_5 & \#_6 & \#_7 & \#_8 & \#_9 & \#_{10} & \#_{11} & \#_{12} & \#_{13} & \#_{14} & \#_{15} & \#_{16}\\ \hline T & T & T & T & T & T & T & T & T & T & F & F & F & F & F & F & F & F\\ T & F & T & T & T & T & F & F & F & F & T & T & T & T & F & F & F & F\\ F & T & T & T & F & F & T & T & F & F & T & T & F & F & T & T & F & F\\ F & F & T & F & T & F & T & F & T & F & T & F & T & F & T & F & T & F\\ \end{array}

And now let's evaluate $\varphi_i=(P \#_i (P \#_i P))$ for each of these 16 possible definitions (I'll flip the table so I have some more room):

\begin{array}{c|c|c} \# & P = T & P = F\\ \#_1 & (T \#_1 (T \#_1 T)) = (T \#_1 T) = T & (F \#_1 (F \#_1 F)) = (F \#_1 T) = T\\ \#_2 & (T \#_2 (T \#_2 T)) = (T \#_2 T) = T & (F \#_2 (F \#_2 F)) = (F \#_2 F) = F\\ \#_3 & (T \#_3 (T \#_3 T)) = (T \#_3 T) = T & (F \#_3 (F \#_3 F)) = (F \#_3 T) = F\\ \#_4 & (T \#_4 (T \#_4 T)) = (T \#_4 T) = T & (F \#_4 (F \#_4 F)) = (F \#_4 F) = F\\ \#_5 & (T \#_5 (T \#_5 T)) = (T \#_5 T) = T & (F \#_5 (F \#_5 F)) = (F \#_5 T) = T\\ \#_6 & (T \#_6 (T \#_6 T)) = (T \#_6 T) = T & (F \#_6 (F \#_6 F)) = (F \#_6 F) = F\\ \#_7 & (T \#_7 (T \#_7 T)) = (T \#_7 T) = T & (F \#_7 (F \#_7 F)) = (F \#_7 T) = F\\ \#_8 & (T \#_8 (T \#_8 T)) = (T \#_8 T) = T & (F \#_8 (F \#_8 F)) = (F \#_8 F) = F\\ \#_9 & (T \#_9 (T \#_9 T)) = (T \#_9 F) = T & (F \#_9 (F \#_9 F)) = (F \#_9 T) = T\\ \#_{10} & (T \#_{10} (T \#_{10} T)) = (T \#_{10} F) = T & (F \#_{10} (F \#_{10} F)) = (F \#_{10} F) = F\\ \#_{11} & (T \#_{11} (T \#_{11} T)) = (T \#_{11} F) = T & (F \#_{11} (F \#_{11} F)) = (F \#_{11} T) = F\\ \#_{12} & (T \#_{12} (T \#_{12} T)) = (T \#_{12} F) = T & (F \#_{12} (F \#_{12} F)) = (F \#_{12} F) = F\\ \#_{13} & (T \#_{13} (T \#_{13} T)) = (T \#_{13} F) = F & (F \#_{13} (F \#_{13} F)) = (F \#_{13} T) = T\\ \#_{14} & (T \#_{14} (T \#_{14} T)) = (T \#_{14} F) = F & (F \#_{14} (F \#_{14} F)) = (F \#_{14} F) = F\\ \#_{15} & (T \#_{15} (T \#_{15} T)) = (T \#_{15} F) = F & (F \#_{15} (F \#_{15} F)) = (F \#_{15} T) = F\\ \#_{16} & (T \#_{16} (T \#_{16} T)) = (T \#_{16} F) = F & (F \#_{16} (F \#_{16} F)) = (F \#_{16} F) = F\\ \end{array}

So, we see that $(P \# (P \# P))$ is a tautology for $\#_1$, $\#_5$, and $\#_9$. What do these definitions have in common? Well, one thing to note is that for all of them it holds that $(P \# Q)$ is true whenever $P$ is false. And that, I think, is pretty much the most interesting thing you can really point out for this problem.

To say that $P\#(P\#P)$ is a tautology means that $F\#(F\#F)=T$ and $T\#(T\#T)=T.$

Note that $F\#F=F$ implies $F\#(F\#F)=F\#F=F.$ Thus, in order for $F\#(F\#F)=T$ to hold, we must have $F\#F=T$ and $F\#T=T.$

$T\#(T\#T)=T$ holds if either $T\#T=T$ (and we don't care about $T\#F$) or else $T\#T=F$ and $T\#F=T.$

So there are three possibilities for the connective $\#$: $$P\#Q=T$$ or $$P\#Q=P\rightarrow Q$$ or $$P\#Q=\neg(P\land Q).$$