# about a solution of a question in symbolic logic

I cannot understand the following solution from my tutorial note.

The question is like this:

Let $P(x)$ be a predicate with universe of discourse $\{a,b,c\}$. The quantifier $\exists!$ is used to assert that there is a unique element of the universe of discourse which makes a predicate true. Now express $\exists ! xP(x)$ using only the operators $\land$, $\lor$ and $\lnot$.

$$\exists ! xP(x) \equiv P(a) \lor P(b) \lor P(c) \lor [\lnot P(a) \land P(b) \land \lnot P(c)] \lor [P(a) \land \lnot P(b) \land \lnot P(c)]\lor [\lnot P(a) \land \lnot P(b) \land P(c)]$$

What is the meaning of the last three parts?

$$\lor [\lnot P(a) \land P(b) \land \lnot P(c)] \lor [P(a) \land \lnot P(b) \land \lnot P(c)]\lor [\lnot P(a) \land \lnot P(b) \land P(c)]$$

-
I typeset your question to make it easier to read - please check I didn't make any mistakes in the notation. – Matthew Pressland Oct 8 '12 at 16:56
"The answer" is incorrect -- for example it is true for the interpretation where $P$ hold s for all three individuals, simply because then $P(a)$ is true and therefore $P(a)\lor\mathit{whatever}$ is automatically true too. – Henning Makholm Oct 8 '12 at 17:04
yes, that correct thank for the correction – Samuel Oct 9 '12 at 4:26

$$\exists!xP(x)\equiv [\lnot P(a) \land P(b) \land \lnot P(c)] \lor [P(a) \land \lnot P(b) \land \lnot P(c)]\lor [\lnot P(a) \land \lnot P(b) \land P(c)]$$
The first term in square brackets says that $b$ has the property $P$ and $a$ and $c$ do not; the second says that $a$ has the property and $b$ and $c$ do not; and the third says that $c$ has the property and $a$ and $b$ do not. Since $a,b$, and $c$ are the only objects in the universe of discourse, these are the only three ways that exactly one of them can have the property $P$.
It should not include $P(a)\lor P(b)\lor P(c)$: that part alone is equivalent to $\exists xP(x)$, and $\exists xP(x)\lor\exists!xP(x)$ is equivalent to $\exists xP(x)$, not to $\exists!xP(x)$.
But suppose it is true that $a = b$ ... then the l.h.s. could be true and the r.h.s. false. – Peter Smith Oct 8 '12 at 18:07
@Peter: Clearly we are to assume that the objects $a,b$, and $c$ are distinct. – Brian M. Scott Oct 8 '12 at 18:10