How Liouville Constructed the First Transcendental Number
In 1844, French mathematician Joseph Liouville solved a centuries-old mystery in number theory by proving the existence of transcendental numbers. While mathematicians had long suspected that certain numbers could not be the solutions to any algebraic equation, no one had been able to prove it. This article provides a clear look at how Liouville defined the boundaries of algebraic numbers using rational approximations and how he exploited those boundaries to artificially construct the Liouville constant, the first proven transcendental number in history.
To understand Liouville's achievement, one must first understand the difference between algebraic and transcendental numbers. An algebraic number is any number that can be the solution (or root) of a polynomial equation with integer coefficients. For example, all rational fractions are algebraic, and irrational numbers like the square root of 2 are also algebraic because it is the solution to the equation \(x^2 - 2 = 0\). A transcendental number, on the other hand, "transcends" algebra; it is not the root of any integer polynomial equation. Before 1844, mathematicians suspected that famous constants like \(e\) and \(\pi\) were transcendental, but proving a negative—that no possible polynomial equation could ever yield these numbers—was immensely difficult.
Instead of trying to prove that a naturally occurring number like \(\pi\) was transcendental, Liouville took a brilliant reverse approach. He decided to study the shared properties of algebraic numbers, find a strict limitation that they all obey, and then custom-build a brand-new number designed specifically to break that rule.
Liouville focused on how closely irrational numbers can be approximated by rational fractions (written as \(p/q\), where \(p\) and \(q\) are integers). He discovered a fundamental speed limit to how well a fraction can approximate an irrational algebraic number. Known today as Liouville’s Theorem on Diophantine Approximation, his proof showed that if an irrational number is the root of a polynomial of degree \(n\), there is a strict limit to how close a fraction \(p/q\) can get to it. Specifically, the error between the algebraic number and the fraction cannot be smaller than a certain value tied to \(q^n\). In simple terms, algebraic numbers resist being approximated too closely by simple fractions.
Armed with this theorem, Liouville realized that if he could create a number that was irrational, yet could be approximated extremely closely by fractions—closer than his theorem allowed for any degree \(n\)—that number could not possibly be algebraic. It would have to be transcendental.
To achieve this rapid approximation, Liouville needed a number whose decimal expansion contained massive, ever-increasing gaps of zeros. In 1851, he presented his most famous example, now known as the Liouville constant. He constructed the number by placing a \(1\) in the decimal places corresponding to factorials (\(n!\)), and \(0\) everywhere else.
The factorials are calculated by multiplying a number by all positive integers below it:
- \(1! = 1\)
- \(2! = 1 \times 2 = 2\)
- \(3! = 1 \times 2 \times 3 = 6\)
- \(4! = 1 \times 2 \times 3 \times 4 = 24\)
- \(5! = 120\)
Placing \(1\)s at the 1st, 2nd, 6th, 24th, and 120th decimal places produces the Liouville constant: \(0.110001000000000000000001000...\)
First, Liouville knew this number was irrational because the gaps of zeros grow infinitely larger, meaning the decimal never repeats. Second, because the gaps of zeros grow so fast, cutting the decimal off right after any \(1\) provides a fraction that is an astonishingly close approximation to the true value of the infinite number.
For example, if you stop at the third \(1\) (the 6th decimal place), the fraction is \(110001 / 1000000\). The error between this fraction and the true Liouville constant is microscopic because the next \(1\) does not appear until the 24th decimal place, leaving 17 zeros in between. If you stop at the fourth \(1\), there are 95 zeros before the next \(1\) appears at the 120th place.
As you go further down the number line, the error between the truncated fraction and the actual number shrinks at a rate dictated by factorials, which grow vastly faster than the \(q^n\) limit established in Liouville's theorem. Because this constructed number allows for rational approximations that are infinitely tighter than what is permitted for any algebraic number, it violates Liouville's theorem.
By sidestepping the famous constants of his day and instead engineering a number from scratch using factorials, Joseph Liouville successfully proved that transcendental numbers were real, laying the groundwork for other mathematicians to later prove the transcendence of \(e\) in 1873 and \(\pi\) in 1882.