Page 191 - HOW TO PROVE IT: A Structured Approach, Second Edition
P. 191

P1: PIG/  P2: OYK/
                   0521861241c04  CB996/Velleman  October 20, 2005  2:54  0 521 86124 1  Char Count= 0






                                                       Relations                       177
                            Proof. We will prove 1, 2, and half of 4, and leave the rest as exercises. (See
                            exercise 6.)

                                                                            −1 −1
                            1. First of all, note that R −1  is a relation from B to A,so(R )  is a relation
                                                             −1 −1
                               from A to B, just like R. To see that (R )  = R, let (a, b) be an arbitrary
                               ordered pair in A × B. Then
                                                  −1 −1
                                         (a, b) ∈ (R )  iff (b, a) ∈ R −1  iff (a, b) ∈ R.
                                                −1
                            2. First note that Dom(R ) and Ran(R) are both subsets of B. Now let b be
                               an arbitrary element of B. Then
                                                −1
                                                                     −1
                                      b ∈ Dom(R )iff ∃a ∈ A((b, a) ∈ R )
                                                    iff ∃a ∈ A((a, b) ∈ R)iff b ∈ Ran(R).
                            4. Clearly T ◦ (S ◦ R) and (T ◦ S) ◦ R are both relations from A to D. Let
                               (a, d) be an arbitrary element of A × D.
                                 First, suppose (a, d) ∈ T ◦ (S ◦ R). By the definition of composition, this
                               means that we can choose some c ∈ C such that (a, c) ∈ S ◦ R and (c, d) ∈
                               T . Since (a, c) ∈ S ◦ R, we can again use the definition of composition
                               and choose some b ∈ B such that (a, b) ∈ R and (b, c) ∈ S. Now since
                               (b, c) ∈ S and (c, d) ∈ T , we can conclude that (b, d) ∈ T ◦ S. Similarly,
                               since (a, b) ∈ R and (b, d) ∈ T ◦ S, it follows that (a, d) ∈ (T ◦ S) ◦ R.
                                 Now suppose (a, d) ∈ (T ◦ S) ◦ R. A similar argument, which is
                               left to the reader, shows that (a, d) ∈ T ◦ (S ◦ R). Thus, T ◦ (S ◦ R) =
                               (T ◦ S) ◦ R.

                                                                 −1 −1
                            Commentary. Statement 1 means ∀p(p ∈ (R )  ↔ p ∈ R), so the proof
                            should proceed by introducing an arbitrary object p and then proving p ∈
                              −1 −1
                                                              −1 −1
                            (R )   ↔ p ∈ R. But because R and (R )  are both relations from A to
                            B, we could think of the universe over which p ranges as being A × B,so
                            p must be an ordered pair. Thus, in the preceding proof we’ve written it as
                            an ordered pair (a, b) from the start. The proof of the biconditional statement
                                     −1 −1
                            (a, b) ∈ (R )  ↔ (a, b) ∈ R uses the method, introduced in Example 3.4.4,
                            of stringing together a sequence of equivalences.
                              The proofs of statements 2 and 4 are similar, except that the biconditional
                            proof for statement 4 cannot easily be done by stringing together equivalences,
                            so we prove the two directions separately. Only one direction was proven.
                            The key to this proof is to recognize that the given (a, d) ∈ T ◦ (S ◦ R)isan
                            existential statement, since it means ∃c ∈ C((a, c) ∈ S ◦ R and (c, d) ∈ T ), so
                            we should introduce a new variable c into the proof to stand for some element
                            of C such that (a, c) ∈ S ◦ R and (c, d) ∈ T . Similarly, (a, c) ∈ S ◦ R is an
   186   187   188   189   190   191   192   193   194   195   196