Psalms in Our Time

Technical Notes: Leading the Singing

This material doesn't contain a connected history of the manner of hymn-singing, but certain themes seem to recur.

Witnesses everywhere mention a tendancy toward dragging the singing. Repeatedly, what started as a "moderately fast" pace compatible with normal speech patterns, within a few generations degenerated into a musical drawn-out drone and verbal incomprehensible groan. Calvin's "half-note per adult heartbeat", Luther's "medieval swing", Wesley's new tunes, Scottish psalm tunes -- all suffered the same dilatory fate. And congregations today are busily engaged in decelerating their own nineteenth-century gospel-song tunes.

Watts and Wesley both expostulated against the problem; musical reformers in Germany, England, Scotland, and America tried various solutions, often feeling compelled to replace virtually their entire repertoire of hymn tunes because people simply could not sing the old ones at any tolerable tempo. Abysmal singing was a contributing factor to the division in the Restoration movement.

Why is this?

I'd suggest three causes: one physical, one technical, and one physiological.

When a leader sings, his voice spreads through the room at roughly 800 feet per second.¹ As people in the room hear that sound, they respond by raising their own voice (after a delay of approximately one-fifth to one-tenth second). Their sound travels back towards the song leader at 800 feet per second; the song leader may also react to that sound.

This seems practically instantaneous, but consider the numbers. In a medium-size auditorium about 100 feet long, the sound of a note will take 1/8 second to make the round trip from podium to median pews. With 1/8 second of physiological delay, the song leader will hear notes 1/4 second after he sings them: that is, at a typical speaking pace of 120 quarter notes per minute, with a full eighth note's delay!

This is a consequence of fundamental physical facts, and cannot be avoided; a song leader must learn to accept it. Suppose he thoughtlessly treats it as an indication that the congregation simply cannot sing as fast as he wishes, and reduces his own tempo. The congregation must accordingly reduce theirs. He will then desperately slow down yet more: a grave error. Little Bo Peep may walk slowly or even backwards; her sheep's tails must remain a full sheep's length behind their heads! "Leave them alone" is sage advice. An alert song leader will set and maintain the pace throughout each song, knowing well that he will never hear the congregation seem to keep up with him; A careless leader will end up "leading" after the manner of the corpse at a wake. Better, indeed, to have no leader at all!

Some congregations have recognized the symptom, but assumed that the underlying problem was that the song leader was not sufficiently overpowering the congregation with his own voice; which could be solved if only he sang loud enough, or had enough acoustical support (whether a resonating mechanical instrument or an electronic amplifier). Not only does this approach strike at the heart of the concept of congregational singing, it is fundamentally flawed and catastrophically misguided: it merely enhances an existing source of delay. Loud sounds do not propagate any faster than soft sounds! In any case, the real problem is not how the congregation hears the song leader, but how he hears them.

We may at once note that the use of a keyboard instrument is no panacea: in addition to being susceptible to the same dilatory effect, organs and pianos have their own intrinsic delays and natural tempi which are not those of singing; historically, even skilled and conscientious musicians (e.g., Bach, Smart, and Vaughan Williams) failed to ameliorate, and even tended to exacerbate, the degeneration. The solution must lie in another direction altogether.

A far better approach would be for the song leader not to be heard at all (indeed, effective congregational singing by definition approaches that ideal.) Therefore the tempo should be given by hand motions, just as in Calvin's Genevan congregations. Alert singers will take their entry cues solely off those hand motions. This addresses the problem in three points, and a simple review of the analysis above shows that no better approach is possible:

1. The hand-beat completely eliminates the delay of the song leader's voice propagating, since light travels instantaneously. This is obviously optimal with respect to this source of delay.
2. Since the hand-motion is continuous and visibly progressive, singers can anticipate its turning points, and respond in almost perfect synchronization²: clearly optimal with respect to this source of delay.
3. Finally, the most invidious cause of delay can be eliminated if the song leader simply be taught to set and maintain the beat with the hand, and then to follow it just as if he were part of the singing congregation, as indeed he ought to be!

This does introduce another skill into the duty of song-leading: but it is by far the simplest skill that has been proposed to address the problem. Many effective song leaders who would be hard-pressed to puzzle out an unfamiliar tune from the position of the black dots, can beat time accurately all day from the little flags on the note stems. But how many congregations have instead spent untold sums purchasing instruments and hiring trained musical performers, simply in order to not address an underlying problem totally unrelated to the competance of the leader's voice? How much easier just to wave the hand over the congregation, and just sing together?

Generational Degeneration of Tempo

I would tentatively suggest that there is another phenomenon that causes the tempi used within a specific congregation to degrade over a generation: the change in heartbeat rate through the human life cycle. Children learning the patterns of worship may have a heartbeat of 140-160; it may drop to 60-70 in old age. Each song leader sets a pace roughly as he learned the song, gauged by his own heartbeat. So the Genevan tunes he perceived as a child (at a rate of three heartbeats per half note) as led by a young adult (at a rate of 1 quarter note per heartbeat) will seem quite fast (3 notes per 2 heartbeats) to an older adult. No matter: as a young man he will lead those tunes at three heartbeats per half note, and his children will learn them at 1 quarter note per heartbeat. Every generation will sing 20-50% slower, simply because of this effect, unless the song leader, conscious of this phenomena, leads all songs significantly faster than he remembers learning them. This may not be possible: but if the older people in the congregation don't think the songs are being sung "too fast" -- this kind of degenerative decay is almost certainly occurring.

Beating time does not address the generational degeneration problem. But a congregation that is actively seeking to learn new hymns, and that has a large enough repertoire for each tune to remains fresh, should be less susceptible to this problem. And the simple fact is that we cannot expect to solve problems for the next generation -- attempts to do so have generally misfired catastrophically -- any more than we can expect to have our problems solved by the last generation. If we can view tradition as our tentative dialogue with fallible ancestors, without forgetting that the Bible is our authoritative dialogue with God, then we'll be prepared to deal with recurrent issues.

Notes

¹The numbers are rounded off for computational convenience; a 200-millisecond delay is very noticeable, but the difference between a 200-millisecond and a 300-millisecond delay will hardly be perceptible. Jack Boyd of Abilene Christian University gives a similar but incomplete analysis of this in his book Leading the Lord's Singing.

^1aAfter I had provided this analysis, an internet newsgroup provided this unsolicited testimony from a participant in a Welsh hymn sing:
"Unaccompanied singing is a great thing, but I don't feel we were particularly handicapped by accompaniment ... One thing I *have* noticed about singing with organ vs singing unaccompanied is due to the innate lag in a church of any significant size. There are always some areas of the church where the organ is heard later than in others, and therefore people standing there are singing a syllable a bit later; then the organ slows down to keep them company, and so forth ad infinitum (and ad nauseum). That doesn't seem to happen so much with just voices, even though the congregation may be nominally following a choir."

^1bThe organ doesn't solve the problem, but merely transfers it: now the organist must set and maintain the pace, regardless of the sound coming from the congregation. Is this an improvement? No less an authority than Vaughan Williams said "The speed indications should not be judged at the pianoforte."

^1cSee his arrangement of "Ein Feste Berg". His own arrangements perpetuated the "isorhythmic" form into which the chorale tunes had fallen because they were sung so excruciatingly slowly. See Liemohn, The Chorale Through 400 Years.

^1dHenry Smart, see the discussion under "Regent Square".

^1eRalph Vaughan Williams, generally regarded as the greatest English composer since the Renaissance, and music editor of the 1906 and 1933 English Hymnal, the most musically influential hymnal in the English-speaking world. But his metronome markings therein are generally regarded as much too slow.

²The synchronization error is on the order of milliseconds: since it is as likely to err in either direction, it is zero for all practical purposes. This is trivial to verify with a simple reaction-time program running on a computer (or programmable pocket calculator.) Forty years ago, this was a standard demo program for the RECOMP II computer in the lab at Middle Tennessee State University.

DRAFT COPY: CIRCULATED FOR REVIEW
Copyright © 2002,2003,2004, Stephen Hutcheson
Please circulate copies to elicit comments, but do not broadcast or publish. Feedback would be received with gratitude.