The Nature of Definition
in Mathematics and Science
In
natural language, a definition is intended to give sense of meaning and context
to words. This is generally done by
providing synonyms and examples of usage.
Our paradigm for language definitions is a dictionary. Webster’s
says a definition is “explanation of the meaning or meanings of a word”; The American Heritage Dictionary calls
it “a statement of precise meaning”. As
a second item, American Heritage adds
that to define is to “describe the
basic qualities of”. This latter notion
is consistent with definition in sciences.
Science and mathematics define objects and properties by providing
criteria for identification for assessing whether an example meets, or fails to
meet, the criteria to be that object or have that property.
In
taxonomy, the categorization of living things is made according to presence or
absence of specified basic qualities. Mammals, for example, are warm-blooded
vertebrate animals that nourish their young with milk. All these characteristics are necessarily
present in any mammal. When we note
that mammals include such animals as cats, dogs, humans, mice, whales, we are
giving examples of mammals, not defining them. All mammals have some hair; this provides
an equivalently strong (necessary) identification criterion, only mammals have
hair. (It is a characteristic less pertinent to the name, though.) Hairy
animals have all the defining characteristics of mammals. When we note that most mammals bear their
young alive, we are providing a stronger criterion that can identify some
mammals that will then have all the defining characteristics of mammals
In
mathematics, an even number is an integer that has no remainder upon division
by the number 2. This provides testable characteristics;
every number meets or fails these defining criteria. Examples of even numbers include 2, 4, 6, 0,
-2, and 324. When we note that even
numbers always have last digit (in the ones column) 0, 2, 4, 6, or 8, we are
noting an equivalent condition that can identify even numbers. When we note that for counting numbers, a set
of an even number of objects can always be arranged in pairs, we are providing
a stronger criterion that can identify some of the even numbers.
The
“concrete operational” response to “what is a mammal?” or “what is an even
number?” will be a list of examples.
Asking for clarification, “how do you know?” will generally result in a
strong identifying criterion (sufficient condition). The “formal operational” stage will allow for
elicitation of defining characteristics or their equivalent. The table below shows our two examples
through this progression.
|
|
Mammal |
Even
Number |
|
What
is a… |
cat,
dog, human |
2,4,6,
…, 10, 12 |
|
How
do you know? |
has
live babies, not eggs like birds or reptiles or fish |
that
many objects can be evenly paired |
|
How
is it defined? |
females
have mammary glands that provide nourishment for young |
division
by 2 results in remainder of 0 |
|
How
else can you identify it? |
all
mammals, and only mammals have hair …
somewhere |
last
digit must be 0,2,4,6, or 8 |
In
mathematics, we would call these characterizations examples, sufficient criterion, defining criterion, and identifying criterion, respectively. In mathematics texts, generally a definition
is given, then some examples, then some propositions or theorems that provide
alternative criteria.
Here
is a typical algebra example.
Definition: An additive
identity is a number that, when added to another number yields that
number. For any number x, x
+ (additive identity) = x.
Example: Zero is an additive identity; 0+x=x.
Proposition: Zero is the additive identity;
0 = 0 + (additive
identity) = (additive
identity).
property of identity from example