The Prohibition

Parallel fifths do not sound wrong. They sound like fewer voices than exist. The rule that bans them does not improve counterpoint -- it makes counterpoint counterpoint.

humanitysociety

A perfect fifth is the interval between C and G --- a frequency ratio of 3:2. Strike both notes and the overtone series of each align with unusual precision: the third harmonic of the lower note occupies the same frequency as the second harmonic of the upper. Two sounds whose spectral signatures overlap so closely that the auditory system faces a parsing problem.

Carl Stumpf named the phenomenon Verschmelzung in 1890 --- melting. The modern term is tonal fusion. When two tones sound at a perfect fifth, the aligned overtones are consistent with two independent sources, but they are equally consistent with the harmonics of a single complex tone. The brain defaults toward the simpler interpretation. In 1987, Lola DeWitt and Robert Crowder tested this experimentally: subjects heard either single tones or pairs of tones at various musical intervals and reported how many sounds they heard. At consonant intervals --- fifths and octaves especially --- subjects made significantly more errors, hearing one tone where two were sounding. The fusion is not inattention. It is the auditory system doing what it evolved to do: interpreting aligned overtone spectra as evidence of a single source.

Two voices sounding a fifth is not a problem. It is consonance --- one of the most stable intervals in Western harmony. The problem begins when two voices move in parallel fifths: both ascending or descending together, maintaining the 3:2 ratio across successive beats. The overtone alignment persists through time. The auditory system’s single-source interpretation locks in. What had been two independent melodic lines becomes, perceptually, one line with altered timbre. The texture goes from N voices to N-1. Not by removing a voice. By merging two into one.

This is why counterpoint prohibits parallel fifths. Not because they sound wrong. They do not sound wrong. They sound like fewer things than exist.

In 1991, David Huron analyzed J.S. Bach’s polyphonic keyboard works and found the prohibition operating as a measurable compositional principle. The prevalence of different vertical intervals in Bach’s writing is directly correlated with their degree of tonal consonance --- with one exception. Intervals that promote tonal fusion are avoided in direct proportion to the strength with which they promote it. Unisons appear less frequently than octaves. Octaves less frequently than perfect fifths. Perfect fifths less frequently than any other interval. The avoidance tracks a single variable: the degree to which two simultaneous voices risk being heard as one.

Huron later derived the traditional rules of voice-leading from six perceptual principles: toneness, temporal continuity, minimum masking, tonal fusion, pitch proximity, pitch co-modulation. Five govern how voices relate to each other. One governs whether voices exist as separate voices at all.

The prohibition is not regulatory. It is constitutive. Remove it and you do not get worse polyphony. You get homophony --- multiple voices moving as one, melodically varied perhaps, structurally collapsed into a single line. Polyphony is not the condition of multiple voices sounding together. Any hymn has that. Polyphony is the condition of multiple voices maintaining independence while remaining in coherent harmonic relationship --- hearing each other, responding to each other, arriving at different melodic conclusions about the same harmonic material. The prohibition on parallel fifths maintains that independence as a structural property of the music. It does not improve counterpoint. It is what makes counterpoint counterpoint.

A democratic system is polyphonic in this specific sense. Its value is not that multiple institutions exist --- many governance arrangements have multiple institutions. Its value is that they process the same reality from genuinely different orientations while remaining in coherent constitutional relationship. A judiciary evaluating legality. A legislature evaluating preference. An executive evaluating feasibility. A press evaluating all three. The capacity is not agreement. It is the maintained independence of the evaluating.

When institutions begin moving in parallel --- courts routinely tracking executive preference, legislatures routinely confirming executive proposals, press coverage routinely adopting executive framing --- tonal fusion begins. The institutional count remains. Each retains its name, its building, its procedures. But the system has begun producing what functions as one evaluation distributed across several letterheads. The count stays at N. The capacity does not.

Tonal fusion does not require a composer who intends it. Two voices can fuse because they were deliberately written in parallel, or because performers drifted into alignment over time, or because the acoustic properties of the room favored consonance over independence. The cause is irrelevant to the perceptual result. Democratic institutions can align because an authoritarian executive coerced them into compliance. They can also align because good-faith actors in each institution, confronting the same evidence, independently arrived at compatible conclusions across domains over years. Incremental convergence produces the same structural deficit as designed capture. Not because the individual assessments are wrong. A court that agrees with the legislature on one question may have reached the correct judgment independently. That is a single consonant chord. Institutions that systematically reach compatible conclusions across domains, over time --- that is motion in parallel fifths. The fusion has the same acoustic signature either way.

The deficit is identical in both cases. The system has lost the capacity to generate genuinely independent assessments of the same material. And the capacity cannot be restored by adding more institutions. You do not fix tonal fusion by adding a third voice in parallel. You fix it by restoring the independence of the voices that merged.

Tonal fusion cannot be identified from a single chord. Two voices sounding a fifth at one moment is consonance. The same two voices maintaining fifths through a sequence of beats is fusion. The test is pattern, not instance --- and pattern is only available retrospectively. By the time the diagnosis is clear, the independent voice has already been absent for some time.

What polyphony loses when voices merge is not quality. It is category. The music has not become worse polyphony. It has become a different thing --- one that may sound full, may even sound correct, but no longer contains the structural property that made it polyphony rather than elaborated unison. The prohibition does not tell voices what to sing. It identifies the one movement that would cost them their separate existence. Everything else is theirs.

Whether a governance system is the kind of system it claims to be is answered the same way. Not by what the institutions decided. By whether they moved independently while deciding it.

Sources

- Solen