Arsis & Thesis: How Chant Moves

The idea in one breath

Sing a sentence out loud — any sentence. Your voice naturally rises toward the word that matters and settles as the sentence ends. Chant works the same way, and the old teachers gave the two halves of that motion Greek names:

That is the whole idea. It is about direction — where the music is going — not about loud and soft. A phrase gathers, reaches its most important moment, and comes to rest, the way incense swings from a censer or a wave rolls to shore. Everything else on this page is detail in service of that wave.

Why does it matter? Because when every note is sung the same — same length, same weight — the result is what choir directors call the sewing machine: correct notes, dead prayer. The wave is what makes chant sound like speech addressed to God rather than typing.

Hear it and see it — right now

You do not have to imagine this. Every chant in the library is played with the wave built in:

• Listen. Open Salve Regina and press Play. The voice swells gently toward each phrase's summit and relaxes into each ending — that is arsis and thesis.
• Watch the light. The soft glow that travels with the melody brightens on the lift and dims into repose.
• Read the line. The colored line above each staff is the wave made visible: light gold where the music is gentle, deepening to red where it presses toward its summit — the singer's map of lift and landing (directors trace the same shape in the air; the gesture is called chironomy).

All three can be adjusted or turned off in the playback settings (the gear next to Play), including how strong the swelling is. How the site decides where the wave rises and falls is spelled out at the end of this page — but you can hear it correctly long before you can explain it.

Five rules for singers

These five rules fix most rhythmic problems without needing the full theory.

1. Not every note is equal. Ordinary notes are roughly the same length; dotted notes (mora) and certain neumes are longer.
2. Lengthen before a bar, not after it. The note before a vertical line broadens; the bar is a boundary, not a downbeat.
3. Accent Latin correctly — lightly. The stressed syllable lifts in pitch; it is not louder or longer. English speakers almost always accent too hard.
4. Follow the phrase, not a beat. Breathe at divisions after completing the note before them.
5. Shape the whole sentence. A phrase rises toward its most important word and settles toward the ending.

Singing lift & landing

Within a small group of notes, some lift toward a landing point (arsis) and one lands (thesis). The landing point in Solesmes editions is often marked with an ictus ('). The melody does not stop at each landing — it rebalances and gathers again, alternating gather → release across the line.

Arsic singing

• Direction: each arsic note moves toward the coming ictus or cadence.
• Density: faster internal motion in melismas (quilisma, approach notes).
• Energy: level or slightly rising — not a swell on every note.
• Common error: singing arsic notes heavier than the thesis, which inverts the gesture.

Thetic singing

• Completion: thetic notes close the gesture.
• Length: often broadened (episema, mora), especially at phrase endings.
• Afterward: energy moves forward into the next arsis — it does not collapse into silence.
• Common error: rushing the thesis to get to the next phrase, which destroys the cadence.

Note lengths

Chant duration is proportional: notes relate to each other by ratio, not to a metronome. Classic Solesmes teaching reduces this to two values — the ordinary note and the doubled note. The mora dot and the pressus produce doubling; the episema is a lighter broadening in between, not a third metric value.

These ratios are not metronome values. They shift with acoustic, tempo, and phrase size. The goal is contrast, not precision to a clock.

The rhythm marks themselves (mora, ictus, episema) are introduced with engraved glyphs in the reading tutorial. Here we focus on what they do to time. The single most important habit:

Latin accent

Chant is sung speech. If the Latin accent is wrong, the rhythm and melody fight the words. Getting the accent right — and light — is the foundation everything else sits on.

The Penultimate Law

For words of three or more syllables:

• If the penult (second-to-last syllable) is long → accent the penult: Re-gí-na.
• If the penult is short → accent the antepenult: mi-se-ri-cór-di-ae.

Three fixed rules cover the rest:

• Two-syllable words always accent the first: sál-ve, má-ter, Dé-us.
• Latin never accents the final syllable of an ordinary word.
• Enclitics (-que, -ve, -ne) pull the accent onto the syllable before them: Deus → Deús-que.

How it feels

In Church Latin the accent is a light lift — the pitch rises slightly, the volume does not. The accented syllable is not automatically longer, and it is arsic (a lift), not a thump. This is the single hardest adjustment for singers trained in stress-timed English, who instinctively make the accent louder and heavier.

Three things that are not the same

They often align on important words, and often diverge on weak syllables and approach notes. Do not "fix" a mismatch by pounding the wrong syllable.

Quilisma & pressus

Some neumes carry their own timing rules inside a melisma.

The shape of a phrase

Rhythm works at several nested levels. You do not need the Latin names to sing well, but they help a director talk precisely about where to breathe and how much to lengthen.

The lengthening (mora) grows with the rank of the bar, and each level keeps its own repose — a period ending does not cancel the smaller morae within it. The bars themselves are boundaries, not beats; the four bar glyphs are engraved in the reading tutorial.

Above all the levels sits the whole phrase's wave — Mocquereau's grand rythme: the line rises toward its summit, usually the most important word (often the melodic apex), and settles from that summit to the cadence. If every small group sounds identical and the line has no summit, you get the sewing-machine effect: correct notes, dead rhythm.

Worked example

Salve Regina (simplex) — opening with morae, ictus, and a comma division. Notice the mora on -ve and -nae, ictus marks throughout, and the half bar after misericordiae.

Kyrie XVII is another good study piece — mora, ictus, episema, and double bar in one line. See the reading tutorial.

The rhythmic foot

A rhythmic foot (pied rythmique) is the smallest unit of arsis–thesis alternation: a group of two or three notes with one thesis (landing). Feet chain across the line to build phrases. They are not measures — they do not imply a fixed tempo or a strong/weak beat hierarchy.

Why only two and three? Because of the 2/3 spacing rule (below): consecutive ictus must fall two or three notes apart, so feet can only be binary or ternary. Larger groups are subdivided; a lone note is never left between two ictus.

The ictus

The ictus (') marks the thesis — the note where a foot lands. It is a light rhythmic touch, a point of placement, not a dynamic accent. Mocquereau, who systematized this in the early 20th century, was explicit: do not crescendo into the ictus as if it were a downbeat in 4/4.

How to sing it

• Articulation: a light touch — a moment of rhythmic definition, not a punch.
• Duration: may be slightly broadened (especially with an episema); it is not necessarily shorter.
• Dynamics: avoid swelling into it.

Finding the ictus when it is unmarked

Editions print some ictus, not all. Apply these rules in order:

1. Printed mark. A vertical episema ' always carries an ictus.
2. Doubled note. A mora dot or a pressus carries an ictus.
3. First note of a compound neume, when nothing above applies.
4. 2/3 spacing. The remaining ictus fall so that consecutive ictus are two or three notes apart — never adjacent, never four or more.

Foot spacing on a real phrase

Applying the 2/3 spacing rule to the opening of Salve Regina shows how landing points distribute. This is a teaching illustration, not a claim about a specific edition — in real chant the notes, not the syllable count, govern where the ictus falls.

SÁL ve  |  Re GÍ na  |  MÁ ter  |  mi se ri  CÓR di ae
 (2)         (3)          (2)        (3)        (3)

Landing points fall every two or three notes — never adjacent, never four or more apart.

When ictus and accent disagree

Mora vocis

Mora vocis means "delay of the voice" — the lingering at points of repose that gives chant its breathing, unhurried quality. It comes in two related senses:

• The mora dot (.): an explicit, local doubling of a single note.
• The broader concept: the general broadening at any point of repose — dotted notes, the note before any bar, and some thetic notes with an episema at the end of an incise.

In every case the voice lingers on the written pitch. It is sustained tone, not a gap of silence — the vocal sound continues; only the duration expands. This is where ensemble precision matters most: the schola must hold the same pitch together and release together.

Not a modern fermata

Too little, too much

The rules this site sings by

Chironomy is ultimately an art — but the classic Solesmes method taught it through rules precise enough for any parish schola to apply from the score. This site encodes those rules as a standardized reading: every note of every chant receives an intensity derived only from the notation and its text. That one value drives the playback dynamics, the traveling glow, and the chironomy curve. The seven rules, exactly as the engine applies them:

1. The incise is the unit. The score is segmented at every divisio, line break, and asterisk; each incise is shaped on its own. Intensity never carries across a bar.
2. Every incise is an arch, and strength is always risen into. Intensity rises from the beginning toward the melodic peak (arsis) and relaxes from the peak to the cadence (thesis). Cadences settle lower than the phrase began, and the opening notes of every incise are an approach — even when the melody opens on its summit, the sound climbs into that strength rather than striking it cold after the previous cadence.
3. Higher leans arsic, lower thetic. The pitch contour, normalized within the incise, is blended with the arch — the arch gives the overall shape (55%), the contour lets local rises breathe inside it (45%).
4. The Latin accent is arsic — a crest of the wave. Every note of an accented syllable is lifted, and the lift is guaranteed locally: the accented syllable rides above the syllables beside it, so the line rises into the accent and falls away from it, never through it. The one exception is the very opening of an incise: rule 2 wins there — an initial accent (Gló-ri-a) keeps its lift but is never forced above what follows, so the entrance remains an approach. This matters most on recitations, where the melody is flat and the accent is the only relief. The lift is applied after rule 7's smoothing so it cannot be averaged away. The scores carry the accents in their lyrics (Dó-mi-nus), so what you see under the staff is exactly what the engine reads.
5. Repose signs soften; cadences decay. Episema and mora notes — already lengthened by the player — also ease in intensity, and the final syllable of every incise decays. Longer, softer, lower: the signature Solesmes ending.
6. The quilisma stays light. Slightly reduced, so the ornament is never struck (the note before it is broadened by the timing rules).
7. Chant undulates; it never lurches. The result is averaged with its neighbors and clamped away from the extremes — a wave, matching the continuous conducting gesture.

One engine, three schools

These seven rules are the Solesmes reading, and the Solesmes playback style applies them in full. The other two interpretation styles adapt the same engine to their own doctrine of where expression lives: Equalist flattens the arch (Pothier's expression flows from the words — the Latin accent lifts more, the repose signs are ignored, phrase ends still settle), and Mensural keeps loudness nearly even, letting the long notes carry their own agogic weight — a slight lean, the opposite of rule 5's softening — and treats the accent as secondary (a small lift, no guaranteed crest). What never changes: the incise is the unit, and cadences settle.

A caution for directors: the engine derives its shape from the melodic contour, the Latin accent, and the written signs — it does not compute Mocquereau's binary/ternary feet from ictus placement. The two agree far more often than not, but where an edition's ictus grid would pull against the melodic arch, the engine follows the arch.

Hearing it: the range control

In playback, intensity becomes loudness within a chosen window — Subtle (±4 dB), Natural (±8 dB, the default), or Expressive (±12 dB), all in the playback settings. Setting Dynamics to Off (flat) restores perfectly even loudness. Downloaded WAV audio follows the same settings. The glow and the chironomy curve have their own toggles beside it.

The numbers on a real phrase

For the opening of Salve Regina (intensity from 0 to 1, 0.5 neutral): the rise c–e–g–h reads 0.12 → 0.39 → 0.68 → 0.75 — the arsis climbing to the summit near the accented gí — then the descent relaxes, and the final cadence note lands at 0.05: longer and softer, exactly as Rule 5 demands.

What stays interpretive

These rules standardize the common ground. They do not decide ictus-placement debates where editions disagree, the manuscript nuances semiology reads (below), or the taste of a director leaning a phrase for the day. Where schools differ, the blended rules average through the disagreement rather than taking sides loudly — treat the rendered dynamics as a well-prepared default, not a verdict.

The weights mirror apps/web/cantus/interpret/dynamics.js; the two are kept in sync.

Solesmes & semiology

You can sing well without mastering this — it explains why scholas sometimes disagree.

Solesmes / Mocquereau

Dom André Mocquereau systematized binary/ternary feet, ictus as thesis marker, and chironomy (hand gestures). The ictus marks in your Liber Usualis come from this theory.

Semiology / Cardine

Dom Eugène Cardine read early neume shapes as carrying their own articulation. Semiology is skeptical of imposing a regular ictus grid on every passage.

Where they agree: chant is free and proportional; notes are not all equal; phrase endings need real length.

Problems come from inconsistency — e.g. claiming semiology while conducting in 4/4.

When it sounds wrong

Diagnosis by ear — for directors, but also for pastors and listeners: you do not need to read neumes to tell whether a schola is singing chant well. Words intelligible, phrases like spoken sentences, calm broadened endings — right. Mechanical ticking, pounding on barlines, accents on final syllables, clipped endings — wrong.

Rehearsal steps (directors)

1. Read the Latin aloud with correct accent — no melody.
2. Mark phrase endings: , ; : ::.
3. Mark every mora (dots and notes before bars).
4. Trace feet and ictus if using Solesmes interpretation.
5. Add neume rules (quilisma, pressus, liquescents).
6. Sing with morae and phrase shape at a moderate tempo.
7. Unify tempo last — tempo serves the phrase, not a grid.

Quick reference

Next steps

• Reading square notation — if you have not yet learned the neumes and clefs.
• Browse the library — open a chant, watch the chironomy curve, and press Play.
• Study Kyrie XVII or Salve Regina — ideal first repertoire for rhythm and text together.