Any mathematical system must have a starting point; we cannot create something out of nothing. The starting point of a mathematical system (or any logical system, for that matter) is a collection of basic terminology accompanied by a collection of assumed facts about the things the terminology describes.
In the axiomatic system of Axiomย 8.1.1, Axiomย 1 is redundant as we may infer from Axiomย 4 that there exist three distinct woozles. But there is no harm in including this axiom for clarity. As well, we will later investigate the effect of altering it.
By Axiomย 4, there exists a trio \(w_1,w_2,w_3 \) of distinct woozles that snarf no dorple in common. Breaking this trio into various pairs and applying Axiomย 3, we see that there exists a dorple \(d_1 \) that \(w_1 \) and \(w_2 \) both snarf in common, there also exists a dorple \(d_2 \) that \(w_1 \) and \(w_3 \) both snarf in common, and there also exists a dorple \(d_3 \) that \(w_2 \) and \(w_3 \) both snarf in common. These snarfing relationships are illustrated in the diagram below.
As this would contradict our initial assumption, it must be the case that \(d_1 \) and \(d_2 \) are distinct. Similar arguments allow us to also conclude that \(d_1 \ne d_3 \) and \(d_2 \ne d_3 \text{.}\)
Suppose \(d_1,d_2 \) are snarf buddies. By contradiction, suppose they snarf more than one woozle in common: let \(w_1,w_2 \) be distinct woozles both snarfed by \(d_1 \) and \(d_2 \text{.}\) By Axiomย 2, each of \(w_1,w_2 \) snarfs each of \(d_1,d_2 \text{.}\) But this contradicts Axiomย 3, as two distinct woozles cannot snarf more than one dorple in common.
In the new, modified axiomatic system, our previous two theorems (Theoremย 8.1.3 and Theoremย 8.1.5) remain true, because it is still true that there exist at least three distinct woozles. But we can now also prove the following.
A nonsense system like the one in Exampleย 8.1.1 is just that โ nonsense โ and not much use unless there are actual examples to which the developed theory can be applied.
a system obtained by replacing the primitive terms in an axiomatic system with more โconcreteโ terms in such a way that all the axioms are true statements about the new terms
If we agree that the axiom statements are still all true with the new terms, then any theorems proved under the abstract system are still valid in the new model system.
Example8.1.7.A model for the woozel-dorple system.
Again consider the axiomatic system of Exampleย 8.1.1, still using the modified version of Axiomย 1. Let the three distinct woozles be the points \((0,0) \text{,}\)\((1,1) \text{,}\) and \((2,0) \) in the Cartesian plane. Let dorple now mean line in the plane, and let snarf now mean lies on. Convince yourself that the axioms of the system are all true with this interpretation of the primitive terms.
Using nonsense terms like woozle, dorple, and snarf for the primitive terms in an axiomatic system is usually not a good idea, as it takes all intuition out of the process of discovering statements that can be deduced from the axioms. It would have been much better if we had used the words point instead of woozle, line instead of dorple, and lies on instead of snarfs as our primitive terms, to be able to use our intuition about how such objects interact. In such a case, the axioms we choose should be a reflection of our idea of the simplest possible properties about the primitive terms, properties that everyone could reasonably agree are โtrueโ without proof. However, for the theorems deduced from such an axiomatic system to have the widest possible applicability, we should leave the words point and line as truly primitive, undefined terms โ that is, point and line should not be taken to mean point in the plane and line in the plane, as in the example above, but rather just left as some abstract, intuitive idea of point and line.