Music And Measure Theory

TL;DR

Harmony depends on closeness to simple rational ratios with small denominators, not on rationality alone.

Briefing Cornell Notes

Briefing

A ratio of musical frequencies can sound harmonious or cacophonous depending less on whether it is rational or irrational, and more on how well it matches simple frequency patterns—especially those with small denominators. The discussion starts with the familiar Pythagorean rule of thumb: rational ratios like 3/2 (a fifth) or 4/3 (a fourth) align with repeating beat patterns that the ear can lock onto as rhythm-like structure. But that neat dichotomy breaks down because many rational numbers are “too complicated” (for example, 211/198 or 1093/826), and the ear may fail to detect the corresponding beat pattern. The argument then shifts to approximation: even irrational ratios can sound good if they lie very close to a simple rational ratio. This is not just a theoretical point—pianos are tuned using the 12th root of 2, so common musical intervals correspond to irrational powers of 2 that happen to sit extremely near simple fractions (e.g., 2^(7/12) is close to 3/2). The upshot is a refined criterion: harmony is tied to closeness to rationals with small denominators, with “how close” depending on musical sensitivity.

That leads to a provocative thought experiment: imagine a musical savant who finds pleasure in every pair of notes with a rational frequency ratio, even the highly complex ones. Would such a person then find every ratio R between 1 and 2 harmonious—including the irrational ones? Since every real number is arbitrarily close to some rational, it’s tempting to say yes. But the second half of the material delivers a counterintuitive mathematical answer through a measure-theory style covering problem.

The challenge asks whether one can cover all rational numbers in (0,1) using open intervals whose total length is strictly less than 1, even though rationals are dense and every interval contains infinitely many rationals. The construction works by enumerating the rationals in reduced form (a systematic infinite list) and assigning to the nth rational an interval whose length is the nth term of a convergent series. By choosing a series that sums to any ε>0, the total length of all intervals can be made as small as desired—so small that most points in (0,1) lie outside the union of intervals, even though every rational is included. Geometrically, the intervals become tiny extremely fast for rationals that appear late in the list, which correspond to fractions with large denominators.

A visual zoom near √2/2 illustrates the intuition: rationals close to a given irrational may exist, but the intervals placed on them shrink faster than the rationals approach the irrational. In the ε=0.3 case, √2/2 lands in the uncovered 70%. When ε is reduced to 0.01, the “covered” 1% consists mostly of numbers near simple rationals—while the “cacophonous” set includes rationals with large denominators and irrationals that sit extremely close to them. The final connection reframes the earlier musical question: even if a savant loves all rational ratios, the set of ratios that can be treated as “harmonious” under a tolerance model can still be small in measure. The paradox—dense coverage with tiny total length—mirrors how a world full of rationals can still leave most real numbers effectively “uncovered,” and thus effectively cacophonous for that savant’s exponentially tightening tolerance.

Cornell Notes

Harmony is linked to how frequency ratios align with simple rational patterns, not merely to whether a ratio is rational. Many rationals are “too complicated” for the ear, and irrational ratios can still sound good when they approximate simple fractions with small denominators—an idea reinforced by piano tuning using the 12th root of 2. The measure-theory challenge shows a parallel phenomenon: all rationals in (0,1) can be covered by open intervals whose total length is as small as any ε>0. The trick is to enumerate rationals and assign the nth rational an interval of length equal to the nth term of a convergent series, making intervals for large-denominator rationals extremely tiny. This implies that even if every rational is “liked,” most real numbers can remain effectively “uncovered” under a tolerance model.

Why does “rational ratios sound harmonious, irrational ratios sound cacophonous” fail as a universal rule?

Rational ratios do create repeating beat patterns, but not all rationals are equally detectable. Fractions like 3/2 or 4/3 correspond to simple, low-denominator structures, so the ear can lock onto the pattern. By contrast, many rationals (e.g., 211/198 or 1093/826) have large denominators and produce beat patterns that are too intricate to perceive cleanly. Meanwhile, irrational ratios can still sound good if they lie very close to a simple rational ratio, meaning the ear effectively hears the nearby simple pattern.

How does piano tuning using the 12th root of 2 support the “approximation to simple rationals” idea?

Pianos are tuned in equal half-steps, multiplying frequency by 2^(1/12) each step. That makes common intervals correspond to irrational ratios: a fifth is 2^(7/12), a fourth is 2^(5/12). These irrational numbers are very close to simple rationals—2^(7/12) is near 3/2, and 2^(5/12) is near 4/3. The chromatic scale works well because powers of 2^(1/12) repeatedly land within about a 1% margin of simple fractions.

What is the covering challenge asking, precisely?

It asks for open intervals whose union contains every rational number in (0,1), while the sum of their lengths is strictly less than 1. Since rationals are dense, every open interval contains infinitely many rationals, so it feels impossible to avoid covering “too much” of (0,1). The twist is that infinitely many very small intervals can still cover all rationals while leaving most real numbers outside the union.

How can all rationals in (0,1) be covered with total interval length ε, where ε can be arbitrarily small?

First, list the rationals in reduced form in an infinite sequence (e.g., 1/2, 1/3, 2/3, 1/4, 3/4, then all reduced fractions with denominator 5, then 6, and so on). Next, assign the nth rational an open interval of length equal to the nth term of a convergent series that sums to ε (for example, scale a series like 1/2 + 1/4 + 1/8 + … so it totals ε). Because the series converges, the total length of all intervals is ε, yet every rational gets at least one interval containing it. Intervals for rationals that appear late in the list (typically large denominators) become extremely small.

What does the zoom near √2/2 illustrate about “dense sets” versus “measure”?

Even though rationals cluster everywhere, the intervals placed over rationals shrink so rapidly that an irrational like √2/2 can still fall outside most of them. In the ε=0.3 setup, √2/2/2 (as discussed) lies in the uncovered 70%: rationals approaching it exist, but the intervals covering those rationals are tiny. This formalizes the intuition that only rationals with relatively simple structure (small denominators) get intervals large enough to “reach” nearby points in a measurable way.

How does the math covering result connect back to the musical savant question?

The savant loves all rational ratios, but a tolerance model matters: if the savant’s notion of harmony requires being within an interval around each rational, then the set of “harmonious” ratios can still have small measure. When ε is small, the covered 1% is dominated by numbers near simple rationals and by irrationals extremely close to high-denominator rationals. So even a savant who likes every rational can still experience 99% of ratios as cacophonous under exponentially shrinking tolerance for more complex fractions.

Review Questions

What role do small denominators play in turning beat patterns into perceived harmony, and how does approximation change the rational/irrational distinction?
Describe the construction that covers all rationals in (0,1) with total interval length ε. Why does it not contradict density of rationals?
In the analogy to music, what does it mean for an irrational number to land outside the union of intervals, and how does that relate to “tolerance” for complex rationals?

Key Points

1
Harmony depends on closeness to simple rational ratios with small denominators, not on rationality alone.
2
Irrational intervals can sound good because they are tuned to lie near simple fractions (as with 2^(7/12) near 3/2).
3
A convergent series lets infinitely many open intervals cover every rational in (0,1) while keeping total length equal to any chosen ε>0.
4
Enumerating rationals and assigning the nth rational an interval of length equal to the nth series term forces intervals for large-denominator rationals to become extremely small.
5
Dense sets can still have “small coverage” in the sense of measure: every rational is covered, yet most real numbers remain outside the union.
6
The musical savant thought experiment fails under a tolerance model: even if all rationals are liked, most irrationals can remain effectively cacophonous when tolerances shrink fast for complex fractions.

Highlights

Piano tuning uses the irrational ratio 2^(1/12), yet common musical intervals still work because these irrational powers land very near simple rationals like 3/2 and 4/3.

All rationals in (0,1) can be covered by open intervals whose total length is as small as any ε>0—despite rationals being dense.

The key mechanism is assigning interval lengths via a convergent series after enumerating rationals, making intervals for complicated fractions shrink faster than nearby points approach.

The “dense but small-measure” covering mirrors how a savant’s exponentially tightening tolerance can leave most ratios effectively cacophonous.

Topics

Frequency Ratios
Rational Approximation
Measure Theory
Open Interval Coverings
Musical Tuning