An Important Example of an Equivalence Relation and Well-Definedness Suppose X and Y are sets and that f : X → Y is a function. Let us define a relation ∼ on X by x1 ∼ x2 if and only if f (x1 ) = f (x2 ). (1) Review Exercise 1. Show that ∼ is an equivalence relation. Proof. The definition of equivalence relation says that we must prove that ∼ is reflexive, symmetric and transitive. (i) Reflexive: Since f (x) = f (x) for every x ∈ X, the definition (1) implies that x ∼ x. (ii) Symmetric: If x1 ∼ x2 , then f (x1 ) = f (x2 ). But then f (x2 ) = f (x1 ), and hence x2 ∼ x1 , as desired. (iii) Transitive: If x1 ∼ x2 and x2 ∼ x3 , then f (x1 ) = f (x2 ) and f (x2 ) = f (x3 ). But then f (x1 ) = f (x3 ), and hence x1 ∼ x3 , as desired. Remark. Notice that in the proof above we used the fact that = is an equivalence relation, and hence is itself reflexive, symmetric and transitive. Review Exercise 2. If the function f that we started with is one to one, what can we say about ∼ ? Since ∼ is an equivalence relation, we define equivalence classes and the set of all equivalence classes for ∼ . Definition. (i) [x] = {x̂ ∈ X : x̂ ∼ x} is called the equivalence class of ∼ represented by x, or, less formally, the equivalence class of x. (ii) X̂ = {[x] : x ∈ X} is the set of all equivalence classes of ∼ . Remarks. (i) Note that a given equivalence class can have many representatives. In fact, [x1 ] = [x2 ] if and only if x1 ∼ x2 . (Show!) (ii) Precisely speaking, equivalence classes are subsets of X, so X̂ is a set of subsets of X. Usually, though, we just think of X̂ as a new set of elements and don’t stop to think of those elements as sets. Suppose we set fˆ([x]) = f (x) for [x] ∈ X̂. (2) Review Exercise 3. Show that fˆ: X̂ → Y is a well-defined, one to one function. Proof. We have two things to do here. First, we must show that fˆ is well-defined, i.e., that it makes sense. Isn’t that obvious? No! Looking carefully, we see that the definition of fˆ applied to [x] depends on x, the representative we chose for the equivalence class [x]. As remarked above, [x] can have more than one representative, i.e., we can have [x] = [x0 ] for x 6= x0 . If that’s true, can we be sure that fˆ([x]) is defined the same way as fˆ([x0 ])? Well, if [x] = [x0 ], then x ∼ x0 , as noted above. But the definition of ∼ then yields that f (x) = f (x0 ). But now using (2), we have fˆ([x]) = f (x) = f (x0 ) = fˆ([x0 ]), exactly what we need. What remains is to show that fˆ is one to one. Recall that this means that if fˆ([x1 ]) = fˆ([x2 ]), then [x1 ] = [x2 ]. But again using (2), we have that f (x1 ) = fˆ([x1 ]) = fˆ([x2 ]) = f (x2 ), and hence, using (1), that x1 ∼ x2 , and so [x1 ] = [x2 ], as desired. Remark. This example is very important in algebra and we will see it in action when we study the First Isomorphism Theorem a little later on. c R. Kubelka 2013