Let us motivate some analysis of the real number line and its properties. In this section, I attempt to demonstrate the following:
- Set theory provides resolution to paradoxes in our real number system, and allows us to define various number systems and operations with mathematical rigor.
- Extending our number system from the naturals through to the reals ($\Bbb{N} \subseteq \Bbb{Z} \subseteq \Bbb{Q} \subseteq \Bbb{R}$) is required if
every length along the number line is to correspond to a value that we can actually define (Abbott, 2000 pg. 4).
This seems like a requirement of the standard model, where spacetime is taken as continuous.
- Functions may be generalized to a definition which is far less restrictive than the criteria proposed during
introductory courses in algebra and single-variable calculus. This is necessary when special functions (distributions) which appear in abundance in physics -- like Heaviside and Dirac-delta.
The Holes in our Rational Number Line
Consider this bizarre function proposed by Dirichlet:
$$
f(x) = \begin{cases}
1\ \ \text{if} \ x \in \Bbb{Q} \\
0\ \ \text{if} \ x \notin \Bbb{Q}
\end{cases} \tag{Eq. 1}$$
Plotting this function produces a figure that is highly discontinuous. As a result, f(x) pushes notions of being single-valued by appearing to produce two horizontal lines using inputs of x.
Resolution will not affect the plot's fine-structure or 'smoothness'. The Archimedean Property, held by rationals and reals, guarantees that there are infinite irrationals between any two elememts of $\Bbb{Q}$,
fitting this observation.
Fig 1. - A plot of f(x). This is a single function, coded numerically in MATLAB; this is not a plot of two functions stacked on top of one another.
Here are some important observations about $f(x)$:
- The irrational number line (situated at $f(x)=0$) is constructed from the "holes" in the rational number line above it (at $f(x)=1$).
- On a macroscopic scale, this demonstrates that elements of $\Bbb{Q}$ are dense in $\Bbb{R}$.
- Paradoxically, both sets of points consist of infinitely-many elements.
This property of $\Bbb{Q}$ to appear continuous, regardless of domain scope, is why the rational numbers form a field and support the identities for multiplication, division, addition and subtraction.
For example, every element of $\Bbb{Z}$ has the multiplication inverses to support division.
Providing Visualization for $\Bbb{Q}$'s Unique Density in $\Bbb{R}$
An interesting way to show that $\Bbb{Q}$ has density, as opposed to its subsets $\Bbb{N}$ and $\Bbb{Z}$,
is to plot an analogue to Dirichlet's function, but with integers at $f(x)=1$ and $x \in \Bbb{Q} \land x \notin \Bbb{Z}$ at f(x)=0. In this case, the empty intervals between the upper set of points ($\Bbb{Z}$)
will be filled by the rationals of $\Bbb{Q}$ -- analogous to (Fig. 1) where the holes of $\Bbb{Q}$ were filled by the irrationals:
$$
f_{\Bbb{Z}}(x) = \begin{cases}
1\ \ \text{if} \ x \in \Bbb{Z} \\
0\ \ \text{if} \ x \in \Bbb{Q} \land x \notin \Bbb{Z}
\end{cases} \tag{Eq. 2}$$
Fig 2. - A plot of $f_{\Bbb{Z}}(x)$ over a
domain that demonstrates how
empty intervals in $\Bbb{Z}$ are filled by densely-packed rationals.
Irrationals are absent from (Eq. 2), yet the set at $f_{\Bbb{Z}}(x)=0$ still looks continuous;
Boundedness of our Number System
A common approach to
the real number line begins with the Axiom of Completeness and the Archimedean Property.
These are leveraged to understand the rational set ($\Bbb{R}$), infinities, boundedness, and dense sets.
- Axiom of Completeness: Every nonempty set of real numbers that is bounded above has a least upper bound (Abbott, 2000, p.14).
- Archimedian Property (i): Given any number $x \in \Bbb{R}$, there exists an $n \in \Bbb{N}$ satisfying $n \gt x$
- Archimedian Property (ii): Given any number $y \in \Bbb{R}$, there exists an $n \in \Bbb{N}$ satisfying $1/n \lt y$ (Abbott, 2000, p.19)
Taking a surface-level look at our own number system through the context of these statements, familiar concepts fall out of [semi-]rigorous foundation:
- The natural numbers set $\Bbb{N}$ leads with 1, and thus is bounded below (infinum 1). It is not bounded above.
Absence of an upper-bound fits with layman notions for the infinite: the natural set begins but it never ends.
$$\Bbb{N} = \{1,2,3,4,\dots \}$$
- The integer set $\Bbb{Z}$ is not bounded above, but unlike $\Bbb{N}$ it is neither bounded below. However, like $\Bbb{N}$ its elements are ordered and countably infinite.
$$\Bbb{Z} = \{ \dots , -2, -1, 0, 1, 2, \dots \}$$
- The rational field $\Bbb{Q}$ contains the set $\Bbb{Z}$. Therefore, $\Bbb{Q}$ will not be bounded below or above, yet it is still ordered and countably infinite.
$$\Bbb{Q} = \{ p/q : p \land q \in \Bbb{Z},\ q \notin 0 \}$$
Defining $\Bbb{Q}$ requires "more" notation than its subsets. Despite being countably infinite like the naturals and the integers, it is impossible to list adjacent elements in the rational number set.
Applying this concept elsewhere, consider some set $A$ with elements that are ordered as such:
$$
A = \{r^2 \gt 2\ :\ x \in \Bbb{Q}\} \tag{Eq. 3}
$$
Set $A$ contains all rational numbers $r^2$ that are greater than 2, not including 2 itself.
- Rational numbers form an ordered set (field) that is countable,
and set A contains all rationals greater than 2
- Therefore, set A is also ordered and countable; these properties are inherited from $\Bbb{Q}$.
- A must be bounded below with a greatest lower bound (infinum) of 2, despite $2 \notin A$.
$$ \bbox[8px,border:1px solid black]{ \text{inf} A=2} $$