Text Box: Wabstractmath.org 

GLOSSARY  

 

Posted 27 December 2008

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

well-defined

Suppose a function F is  defined by an assertion “F(X) = E”, where X and E are some expressions in the symbolic language.  F is said to be well-defined if:

WD-1.  X denotes a specific element of the domain of F.

WD-2.  Let Y be an expression denoting the same element of the domain of F that X denotes.  Let E’ be the expression obtained from E by substituting every occurrence of X in E by Y.  Then E  and  denote the same mathematical object,

Note!

Although we say “F is well-defined”, in fact “well-defined” is a property of  the notation used in defining F, not of F itself.  

Example

For any integer n, let [n] denote the set of all integers of the same parity as n (in other words, the set of all even integers if n is even and the set of all odd integers if n is odd.)  For example [0] = [4] = [42] and [17] = [5].  Let  denote the set {[0], [1]} .  (This is fairly standard notation.  See congruence.)  Now define  by

                                                               

Then F is well-defined because an integer is even if and only if its square is even.  Here, X  is the expression “[n]” and E is the expression “  ”.  In fact, F is the identity function. 

Example

Using the same notation as in the preceding example, if we defined  G(n) = [the number of prime divisors of n],  then G is not well-defined by WD-2, since we would have [2] = [6] but G[2] = [1] and G[6] = [2] = [0], and .  Note that WD-1 does hold, since the number of prime divisors of an integer is an integer and the expression “[n]” is defined for every integer n. 

Example

Let  denote the set of all nonempty subsets of the set of nonnegative integers.  Thus  and  but . 

Define  by F(S) = the smallest element of S. Thus  and .

F is well defined:   WD-1 holds because there is a smallest element of S and it is an integer, and WD-2 holds because no matter how you write a set its smallest element is the same.  For example,

 

                                     

If we have left out the requirement that  consist of the nonempty subsets of the set of nonnegative integers, then F would fail WD-1 because there is no smallest element of the empty set.  Similarly if you set  to be the set of all nonempty subsets of , because for example the set of all odd integers has no smallest element.

Here is another example.

when

In math English, when may mean “if”.

Example

“When a function has a derivative, it is necessarily continuous.”

Usage

These phrases

¨  when and only when

¨  exactly when

¨  precisely when


all mean if and only if.

whenever

Whenever is also used to mean “if”.  It may be used after another “if” to avoid having two in a row, for example

 

“A relation  is symmetric if whenever  then  ”.

where

Where may be used in a special way in math English, sometimes without any connotation of location.

Examples

¨  Definition: An element a of a group is involutive if , where e is the identity element of the group.”  Note that where carries no connotation of location in this sentence.  The statement “where e is the identity element of the group” is a postcondition.

¨  “A point x where f'(x) = 0 is a critical point.” In contrast to the previous usage, this usage carries a connotation of location.

without loss of generality

A proof of an assertion involving two elements x and y of some mathematical structure S might appear to require consideration of two cases in which x and y are related in different ways to each other; for example for some assertion P, P(x, y) or P(y, x) could hold. However, if there is a symmetry of S that interchanges x and y, you may need to consider only one case. In that case, the proof may begin with a remark such as,

Without loss of generality, we may assume P(x, y).”

WLOG is an abbreviation for this phrase.

Example

“Theorem: Suppose m and n are two distinct elements of the closed interval of real numbers [a, b].  Then the closed interval determined by m and n is shorter than [a, b].   Proof:  Without loss of generality, assume m < n.  The length of [a, b] is b  a  and the length of [m, n] is n  m.  Since m and n are in [a, b], we now know that a < m < n < b.  Hence n  m < n  a < b  a, as required.”

Of course, this theorem is obvious if you draw a picture.  But this proof technique is useful in much more complicated circumstances.

 

Don’t be insulted by trivial examples. 

They confirm that you understand something.

 

 

witness

A witness to an existential statement of the form  is an object w for which P(w) is true.  For example, a witness to the statement “There is an integer x for which  ” is the integer 3 (among many others).  The integer 2 is not a witness to this statement because , which is not greater than 5.  On the other hand, the statement “There is an integer n for which  ” is false, so it has no witnesses at all.