diff --git a/components/liturgy/LiturgySettings.tsx b/components/liturgy/LiturgySettings.tsx
index 84efefa..00d5957 100644
--- a/components/liturgy/LiturgySettings.tsx
+++ b/components/liturgy/LiturgySettings.tsx
@@ -1,4 +1,4 @@
-import React, { createContext, useContext, useEffect, useState } from 'react';
+import React, { createContext, useContext, useEffect, useRef, useState } from 'react';
 
 /**
  * Page-level reader preferences. Lives in localStorage so the reader's
@@ -87,9 +87,31 @@ export function useLiturgySettings() {
 export const SettingsButton: React.FC = () => {
   const { settings, setSettings } = useLiturgySettings();
   const [open, setOpen] = useState(false);
+  const wrapRef = useRef<HTMLDivElement>(null);
+
+  // Close the popover on a click outside it, or on Escape. Without this
+  // the only way to dismiss it is the gear toggle — clicking the chant
+  // body left it stuck open.
+  useEffect(() => {
+    if (!open) return;
+    const onDown = (e: MouseEvent) => {
+      if (wrapRef.current && !wrapRef.current.contains(e.target as Node)) {
+        setOpen(false);
+      }
+    };
+    const onKey = (e: KeyboardEvent) => {
+      if (e.key === 'Escape') setOpen(false);
+    };
+    document.addEventListener('mousedown', onDown);
+    document.addEventListener('keydown', onKey);
+    return () => {
+      document.removeEventListener('mousedown', onDown);
+      document.removeEventListener('keydown', onKey);
+    };
+  }, [open]);
 
   return (
-    <div className="fixed top-4 right-6 z-40">
+    <div ref={wrapRef} className="fixed top-4 right-6 z-40">
       <button
         type="button"
         onClick={() => setOpen((v) => !v)}
diff --git a/components/liturgy/shapes/TripleScriptWitness.tsx b/components/liturgy/shapes/TripleScriptWitness.tsx
index 52fb311..e07cfba 100644
--- a/components/liturgy/shapes/TripleScriptWitness.tsx
+++ b/components/liturgy/shapes/TripleScriptWitness.tsx
@@ -192,7 +192,7 @@ const TransliterationLine: React.FC<{
     if (!respelling) return null;
     return (
       <div
-        className="relative z-0 text-slate-500 italic text-sm mt-1 leading-relaxed select-text tracking-wide"
+        className="relative z-[5] bg-slate-950 text-slate-500 italic text-sm mt-1 leading-relaxed select-text tracking-wide"
         style={{ fontFamily: SCRIPT_FONT.Latn }}
         aria-label={`Pronunciation respelling of ${variant.label}`}
       >
@@ -204,7 +204,7 @@ const TransliterationLine: React.FC<{
   if (!variant.transliteration) return null;
   return (
     <div
-      className="relative z-0 text-slate-500 italic text-sm mt-1 leading-relaxed select-text"
+      className="relative z-[5] bg-slate-950 text-slate-500 italic text-sm mt-1 leading-relaxed select-text"
       style={{ fontFamily: SCRIPT_FONT.Latn }}
       aria-label={`Transliteration of ${variant.label}`}
     >
@@ -468,7 +468,11 @@ const HoverSpan: React.FC<{
       data-hover-span="true"
       data-morpheme-idx={morphemeIdx}
       data-concept-ids={conceptAttr}
-      className={`relative inline-block cursor-help border-b border-dotted border-emerald-700/40 hover:border-emerald-300 hover:text-emerald-100 transition-colors ${
+      // Each morpheme gets its own underline + tiny horizontal padding so
+      // adjacent morphemes don't merge visually. mn10 pattern: the eye
+      // sees per-segment breaks (kar · aṇī · yam) rather than one long
+      // continuous underline under the whole word.
+      className={`relative inline-block cursor-help px-[2px] border-b border-dotted border-emerald-700/40 hover:border-emerald-300 hover:text-emerald-100 transition-colors ${
         bold ? 'font-semibold' : ''
       }`}
       onMouseEnter={() => setOpen(true)}
@@ -704,7 +708,16 @@ const EnglishLine: React.FC<{
    * attestations.
    */
   witnessBy?: string;
-}> = ({ text, accentByEnIdx, witnessBy }) => {
+  /**
+   * Witness's per-English-word mapping to Pāli surface position. When
+   * an entry is -1, the English word is "glue" — connective tissue
+   * English needs that has no Pāli counterpart ("This is what should be
+   * done" → "is", "what", "be" carry no Pāli, only "done" maps). The
+   * mn10 reader dims those words so the eye lands on content words.
+   * Without alignTo, all words render at full opacity.
+   */
+  alignTo?: number[];
+}> = ({ text, accentByEnIdx, witnessBy, alignTo }) => {
   const { settings } = useLiturgySettings();
   const tokens = tokenizeEnglish(text);
   let engIdx = -1;
@@ -717,6 +730,11 @@ const EnglishLine: React.FC<{
         const accentClass = settings.showAccents && accent ? ACCENT_CLASS[accent] : '';
         const concepts = conceptsForToken('en', 'Latn', t, witnessBy);
         const conceptAttr = concepts.length > 0 ? concepts.join(' ') : undefined;
+        // Glue word: English-only scaffolding with no Pāli source. mn10
+        // renders these at 0.55 opacity (Legend.tsx uses 0.3 for "ghost
+        // words"; we settle higher because liturgy glue is more often
+        // unavoidable English syntax than fully supplied content).
+        const isGlue = alignTo !== undefined && alignTo[engIdx] === -1;
         return (
           <span
             key={i}
@@ -725,6 +743,7 @@ const EnglishLine: React.FC<{
             // bg + z-10 to hide alignment-line strokes behind the word (see
             // PaliLine sibling above for the same treatment).
             className={`relative z-10 inline-block bg-slate-950 ${accentClass}`}
+            style={isGlue ? { opacity: 0.55 } : undefined}
           >
             {t}
           </span>
@@ -830,7 +849,8 @@ type Line = {
 
 function computeAlignmentLines(
   container: HTMLDivElement,
-  alignTo: number[] | undefined
+  alignTo: number[] | undefined,
+  morphemeAlignTo?: (number | null)[]
 ): Line[] {
   if (!alignTo) return [];
   const cRect = container.getBoundingClientRect();
@@ -862,8 +882,12 @@ function computeAlignmentLines(
       if (!enEl) continue;
 
       // Three positioning strategies, in order of preference:
-      //   1. Authored morpheme spans exist → anchor on morpheme #i
-      //      (clamped to last morpheme if more English than morphemes).
+      //   1. Authored morpheme spans exist → anchor on a morpheme.
+      //      Which morpheme: if the witness authored `morphemeAlignTo`
+      //      for this English token, use that index (lets a curator fix
+      //      crossed arrows when English reorders the morphemes). Else
+      //      fall back to the positional heuristic (i-th English → i-th
+      //      morpheme, clamped to the last).
       //   2. No morpheme spans but the group has >1 English mapping to
       //      this word → distribute proportionally along the word's
       //      width so the arrows fan into separate landing zones
@@ -874,11 +898,15 @@ function computeAlignmentLines(
       let y1: number;
       let subIdx: number | undefined = undefined;
       if (morphemeEls.length > 0) {
-        const clamped = Math.min(i, morphemeEls.length - 1);
-        const mr = morphemeEls[clamped].getBoundingClientRect();
+        const authored = morphemeAlignTo?.[engIdx];
+        const target =
+          typeof authored === 'number'
+            ? Math.min(Math.max(authored, 0), morphemeEls.length - 1)
+            : Math.min(i, morphemeEls.length - 1);
+        const mr = morphemeEls[target].getBoundingClientRect();
         x1 = mr.left + mr.width / 2 - cRect.left;
         y1 = mr.bottom - cRect.top;
-        subIdx = clamped;
+        subIdx = target;
       } else if (engIndices.length > 1) {
         const xOffset = ((i + 0.5) / engIndices.length) * wordRect.width;
         x1 = wordRect.left + xOffset - cRect.left;
@@ -933,7 +961,7 @@ const AlignmentLines: React.FC<{ lines: Line[]; containerWidth: number }> = ({
   const visible = lines;
   return (
     <svg
-      className="absolute inset-0 pointer-events-none"
+      className="absolute inset-0 pointer-events-none z-0"
       style={{ width: '100%', height: '100%', overflow: 'visible' }}
       aria-hidden="true"
     >
@@ -1052,7 +1080,7 @@ const SegmentRow: React.FC<{
         setLines([]);
         return;
       }
-      setLines(computeAlignmentLines(containerRef.current, currentWitness?.alignTo));
+      setLines(computeAlignmentLines(containerRef.current, currentWitness?.alignTo, currentWitness?.morphemeAlignTo));
     };
     compute();
     const raf = requestAnimationFrame(compute);
@@ -1166,17 +1194,34 @@ const SegmentRow: React.FC<{
     // which would paint the line's source endpoint at the viewport corner.
     // (See task #73, user-reported mobile Devanāgarī bug.)
     if (!containerRef.current.contains(hovered.element)) return [];
-    const fresh = computeAlignmentLines(containerRef.current, currentWitness?.alignTo);
+    const fresh = computeAlignmentLines(containerRef.current, currentWitness?.alignTo, currentWitness?.morphemeAlignTo);
     const cRect = containerRef.current.getBoundingClientRect();
     const r = hovered.element.getBoundingClientRect();
 
-    // Step 3 — idx match
-    const idxMatched = fresh.filter((l) =>
-      hovered.kind === 'pali' ? l.paliIdx === hovered.idx : l.engIdx === hovered.idx,
-    );
-
-    // Step 4 — concept overlap
+    // Step 3 — idx match. When the user is hovering a specific morpheme
+    // within a Pāli word (the inner HoverSpan emits `data-morpheme-idx`),
+    // narrow to lines that anchor at that morpheme. Without this, every
+    // arrow for the whole word stays visible regardless of which morpheme
+    // the cursor is on — and the per-morpheme tooltips feel decoupled
+    // from the arrow shown. See screenshot feedback (verse 1 karaṇīyam).
     const hoveredEl = hovered.element as HTMLElement;
+    const hoveredMorphemeStr = hoveredEl.dataset.morphemeIdx;
+    const hoveredMorphemeIdx =
+      hovered.kind === 'pali' && hoveredMorphemeStr !== undefined
+        ? parseInt(hoveredMorphemeStr, 10)
+        : null;
+    const idxMatched = fresh.filter((l) => {
+      if (hovered.kind === 'pali') {
+        if (l.paliIdx !== hovered.idx) return false;
+        if (hoveredMorphemeIdx !== null && l.morphemeIdx !== undefined) {
+          return l.morphemeIdx === hoveredMorphemeIdx;
+        }
+        return true;
+      }
+      return l.engIdx === hovered.idx;
+    });
+
+    // Step 4 — concept overlap (hoveredEl already declared above)
     const hoveredConceptStr = hoveredEl.dataset.conceptIds;
     const hoveredConcepts = hoveredConceptStr
       ? new Set(hoveredConceptStr.split(/\s+/).filter(Boolean))
@@ -1270,6 +1315,7 @@ const SegmentRow: React.FC<{
               text={currentWitness.text}
               accentByEnIdx={accentByEnIdx}
               witnessBy={currentWitness.by}
+              alignTo={currentWitness.alignTo}
             />
           </div>
         </div>
diff --git a/data/liturgy/bodhi-heart-sutra.ts b/data/liturgy/bodhi-heart-sutra.ts
index 920b527..dd6e273 100644
--- a/data/liturgy/bodhi-heart-sutra.ts
+++ b/data/liturgy/bodhi-heart-sutra.ts
@@ -152,7 +152,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               gloss: 'the Noble Lord Avalokiteśvara — the bodhisattva of compassion',
               citations: [wikipediaCitation('Avalokiteśvara')],
               morphemes: [
-                { text: 'Ārya', type: 'prefix', gloss: 'noble — an honourific', pronunciation: 'AHR-yah', citations: [wikipediaCitation('Ārya')] },
+                { text: 'Āry', type: 'prefix', gloss: 'noble — an honorific (*ārya*; its final vowel merges with the *ā-* of the next piece)', pronunciation: 'AHR-yah', citations: [wikipediaCitation('Ārya')] },
                 { text: 'āvalokit', type: 'stem', gloss: 'one who looks down (with compassion) on the world', pronunciation: 'ah-vah-loh-KEET' },
                 { text: 'eśvaro', type: 'suffix', gloss: 'lord — the title-ending', pronunciation: 'AYSH-vah-roh' },
               ],
@@ -330,7 +330,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               citations: [wikipediaCitation('Pāramitā')],
               morphemes: [
                 { text: 'pāram', type: 'stem', gloss: 'the other shore — what awakening looks like from this shore', pronunciation: 'PAH-rahm' },
-                { text: 'itā', type: 'suffix', gloss: 'gone (past participle, feminine) — turning the noun into "having gone"', pronunciation: 'ee-TAH' },
+                { text: 'itā', type: 'suffix', gloss: 'gone — the "X-ed" form, turning the word into "having gone"', pronunciation: 'ee-TAH' },
               ],
             },
             {
@@ -338,11 +338,11 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'चरमाणो',
               scriptAlts: { 'zh-Hant': '行', 'ja-Jpan': '行', 'bo-Tibt': 'སྤྱོད་པ' },
               pronunciation: 'chah-rah-MAH-noh',
-              etymology: '√car "to move, practise" + middle present participle',
+              etymology: '√car "to move, practise" + the "-ing" ending (the form that means doing it for oneself)',
               gloss: 'practising, moving in — the bodhisattva is *doing* this, not just thinking it',
               morphemes: [
                 { text: 'cara', type: 'root', root: '√car', gloss: 'to move, walk, practise', pronunciation: 'CHAH-rah', conceptIds: ['concept.practice-carya'] },
-                { text: 'māṇo', type: 'suffix', gloss: 'middle present participle — the one doing [the moving] for themselves', pronunciation: 'MAH-noh' },
+                { text: 'māṇo', type: 'suffix', gloss: 'the "-ing" form — here, the one doing [the moving] for themselves', pronunciation: 'MAH-noh' },
               ],
             },
           ],
@@ -472,7 +472,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               form: 'tāṃś',
               scriptAlt: 'तांश्',
               pronunciation: 'TAHM-sh',
-              etymology: '*tān* (accusative plural masculine of *tad*) + sandhi *-ś*',
+              etymology: '*tān* — the "those" object-form of *tad* — with a sandhi *-ś*',
               gloss: 'those — refers back to the five skandhas',
             },
             {
@@ -601,7 +601,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'रूपान्',
               scriptAlts: { 'zh-Hant': '色', 'ja-Jpan': '色', 'bo-Tibt': 'གཟུགས' },
               pronunciation: 'ROO-pahn',
-              etymology: '*rūpa* "form" + ablative ending',
+              etymology: '*rūpa* "form" + the "-from" ending',
               gloss: 'from form — "X is not different *from* form"',
               accent: 'amber',
               citations: [wikipediaCitation('Rūpa')],
@@ -717,7 +717,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'शून्यताया',
               scriptAlts: { 'zh-Hant': '空', 'ja-Jpan': '空', 'bo-Tibt': 'སྟོང་པ' },
               pronunciation: 'SHOON-yah-TAH-yah',
-              etymology: '*śūnyatā* "emptiness" + ablative ending',
+              etymology: '*śūnyatā* "emptiness" + the "-from" ending',
               gloss: 'from emptiness — "X is not different from *emptiness*"',
               accent: 'sky',
             },
@@ -1088,7 +1088,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
                 ],
                 'bo-Tibt': [
                   { text: 'ཤཱ་རི', type: 'phonetic', pronunciation: 'shā-ri', gloss: 'phonetic: *śāri* (the mother\'s name)' },
-                  { text: 'འི', type: 'semantic', pronunciation: "'i", gloss: 'genitive — "of"' },
+                  { text: 'འི', type: 'semantic', pronunciation: "'i", gloss: 'the particle that marks "of" — possession' },
                   { text: 'བུ', type: 'semantic', pronunciation: 'bu', gloss: 'son — together: "son of Śāri"' },
                 ],
               },
@@ -1275,7 +1275,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'अनुत्पन्ना',
               scriptAlts: { 'zh-Hant': '不生', 'ja-Jpan': '不生', 'bo-Tibt': 'མ་སྐྱེས་པ' },
               pronunciation: 'ah-noot-PAHN-nah',
-              etymology: '*a-* (not) + *utpannā* "arisen" (past participle of √pad)',
+              etymology: '*a-* (not) + *utpannā* "arisen" — the "X-ed" form of √pad',
               gloss: 'unarisen, not having come into being',
               morphemes: [
                 { text: 'an', type: 'prefix', gloss: 'not — the negation', pronunciation: 'ahn' },
@@ -1287,7 +1287,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'अनिरुद्धा',
               scriptAlts: { 'zh-Hant': '不滅', 'ja-Jpan': '不滅', 'bo-Tibt': 'མ་འགགས་པ' },
               pronunciation: 'ah-nee-ROOD-dhah',
-              etymology: '*a-* (not) + *niruddhā* "stopped, ceased" (past participle of √rudh)',
+              etymology: '*a-* (not) + *niruddhā* "stopped, ceased" — the "X-ed" form of √rudh',
               gloss: 'unceased, not stopped',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
@@ -1555,7 +1555,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'तस्माच्',
               scriptAlts: { 'zh-Hant': '是故', 'ja-Jpan': '是故', 'bo-Tibt': 'དེ་ལྟ་བས་ན' },
               pronunciation: 'tahs-MAHCH',
-              etymology: 'ablative of *tat* "that" — "from that", "therefore"',
+              etymology: 'the "from" form of *tat* "that" — "from that", "therefore"',
               gloss: 'therefore — drawing the consequence from what came before',
             },
             {
@@ -1569,7 +1569,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'शून्यतायाम्',
               scriptAlts: { 'zh-Hant': '空', 'ja-Jpan': '空', 'bo-Tibt': 'སྟོང་པ་ཉིད' },
               pronunciation: 'shoon-yah-TAH-yahm',
-              etymology: 'locative of *śūnyatā* "emptiness" — "in emptiness"',
+              etymology: 'the "in" form of *śūnyatā* "emptiness" — "in emptiness"',
               gloss: 'in emptiness',
               accent: 'sky',
             },
@@ -1585,7 +1585,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               scriptAlt: 'रूपम्',
               scriptAlts: { 'zh-Hant': '色', 'ja-Jpan': '色', 'bo-Tibt': 'གཟུགས' },
               pronunciation: 'ROO-pahm',
-              etymology: '*rūpa* "form, appearance" (nominative neuter)',
+              etymology: '*rūpa* "form, appearance"',
               gloss: 'form, appearance — the first of the five skandhas',
               accent: 'amber',
             },
@@ -1838,7 +1838,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
             { form: 'na', scriptAlt: 'न', pronunciation: 'nah', gloss: 'no' },
             { form: 'raso', scriptAlt: 'रसो', scriptAlts: { 'zh-Hant': '味', 'ja-Jpan': '味', 'bo-Tibt': 'རོ' }, pronunciation: 'RAH-soh', gloss: 'taste, flavour' },
             { form: 'na', scriptAlt: 'न', pronunciation: 'nah', gloss: 'no' },
-            { form: 'spraṣṭavyaṃ', scriptAlt: 'स्प्रष्टव्यम्', scriptAlts: { 'zh-Hant': '觸', 'ja-Jpan': '觸', 'bo-Tibt': 'རེག་བྱ' }, pronunciation: 'sprahsh-TAHV-yahm', gloss: 'touch — that which can be touched (gerundive of √spṛś)' },
+            { form: 'spraṣṭavyaṃ', scriptAlt: 'स्प्रष्टव्यम्', scriptAlts: { 'zh-Hant': '觸', 'ja-Jpan': '觸', 'bo-Tibt': 'རེག་བྱ' }, pronunciation: 'sprahsh-TAHV-yahm', gloss: 'touch — that which can be touched' },
             { form: 'na', scriptAlt: 'न', pronunciation: 'nah', gloss: 'no' },
             { form: 'dharmāḥ', scriptAlt: 'धर्माः', scriptAlts: { 'zh-Hant': '法', 'ja-Jpan': '法', 'bo-Tibt': 'ཆོས' }, pronunciation: 'DHAR-mah', gloss: 'mental objects, dharmas — the objects of the mind-faculty, sixth in the sense-object schema' },
           ],
@@ -2314,18 +2314,18 @@ export const bodhiHeartSutra: LiturgyDoc = {
             },
           ],
           words: [
-            { form: 'tasmād', scriptAlt: 'तस्माद्', scriptAlts: { 'zh-Hant': '故', 'ja-Jpan': '故' }, pronunciation: 'tahs-MAHD', gloss: 'therefore (ablative of *tat* "that")' },
+            { form: 'tasmād', scriptAlt: 'तस्माद्', scriptAlts: { 'zh-Hant': '故', 'ja-Jpan': '故' }, pronunciation: 'tahs-MAHD', gloss: 'therefore — literally "from that"' },
             {
               form: 'aprāptitvāt',
               scriptAlt: 'अप्राप्तित्वात्',
               scriptAlts: { 'zh-Hant': '無所得', 'ja-Jpan': '無所得', 'bo-Tibt': 'ཐོབ་པ་མེད་པ' },
               pronunciation: 'ah-PRAHP-teet-vaht',
-              etymology: '*a-* "not" + *prāpti* "attainment" + *-tva* "-ness" + *-āt* (ablative) — "from-the-not-being-of-attainment"',
+              etymology: '*a-* "not" + *prāpti* "attainment" + *-tva* "-ness" + *-āt* "due to" — "from-the-not-being-of-attainment"',
               gloss: 'due to non-attainment — because there is nothing to attain',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
                 { text: 'prāpti', type: 'stem', gloss: 'attainment', pronunciation: 'PRAHP-tee' },
-                { text: 'tvāt', type: 'suffix', gloss: '-ness, due to (ablative)', pronunciation: 'TVAHT' },
+                { text: 'tvāt', type: 'suffix', gloss: 'forms an abstract noun ("-ness") and marks "due to"', pronunciation: 'TVAHT' },
               ],
             },
           ],
@@ -2429,7 +2429,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               etymology: '*citta* "mind" + *āvaraṇa* "obscuration, covering"',
               gloss: 'mental obscurations — the veils that hide things as they are',
               morphemes: [
-                { text: 'citta', type: 'stem', gloss: 'mind', pronunciation: 'CHEET-tah' },
+                { text: 'citt', type: 'stem', gloss: 'mind (the stem *citta*; its final vowel merges with the *ā-* that follows)', pronunciation: 'CHEET-tah' },
                 { text: 'āvaraṇa', type: 'suffix', gloss: 'obscuration, cover', pronunciation: 'AH-vah-rah-nah' },
               ],
             },
@@ -2503,12 +2503,12 @@ export const bodhiHeartSutra: LiturgyDoc = {
               form: 'nāstitvād',
               scriptAlt: 'नास्तित्वाद्',
               pronunciation: 'NAHS-teet-vahd',
-              etymology: '*na* "not" + *asti* "is" + *-tva* "-ness" + *-ād* ablative ending',
+              etymology: '*na* "not" + *asti* "is" + *-tva* "-ness" + *-ād*, the "due to" ending',
               gloss: 'due to the absence (of) — literally "from the not-being-ness of"',
               morphemes: [
                 { text: 'nā', type: 'prefix', gloss: 'not (the negation)', pronunciation: 'nah' },
                 { text: 'sti', type: 'stem', gloss: 'being (root √as "to be")', pronunciation: 'STEE' },
-                { text: 'tvād', type: 'suffix', gloss: '"-ness, from" — forms the abstract noun and marks ablative ("due to")', pronunciation: 'TVAHD' },
+                { text: 'tvād', type: 'suffix', gloss: 'forms an abstract noun ("-ness") and marks "due to"', pronunciation: 'TVAHD' },
               ],
             },
           ],
@@ -2601,7 +2601,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
               accent: 'emerald',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not — the negation', pronunciation: 'ah' },
-                { text: 'trasto', type: 'stem', gloss: 'afraid (past participle)', pronunciation: 'TRAHS-toh' },
+                { text: 'trasto', type: 'stem', gloss: 'afraid — the "X-ed" form', pronunciation: 'TRAHS-toh' },
               ],
             },
           ],
@@ -2693,8 +2693,8 @@ export const bodhiHeartSutra: LiturgyDoc = {
               gloss: 'having crossed beyond inverted views — past the four distortions (taking the impure as pure, the impermanent as permanent, etc.)',
               accent: 'emerald',
               morphemes: [
-                { text: 'viparyāsa', type: 'stem', gloss: 'inverted view — seeing things upside-down (impermanent as permanent, suffering as pleasure, etc.)', pronunciation: 'vee-pahr-YAH-sah', citations: [wikipediaCitation('Viparyasa')] },
-                { text: 'atikrānto', type: 'suffix', gloss: 'having crossed beyond, transcended', pronunciation: 'ah-tee-KRAHN-toh' },
+                { text: 'viparyās', type: 'stem', gloss: 'inverted view — seeing things upside-down (impermanent as permanent, suffering as pleasure, etc.). The stem is *viparyāsa*; its final vowel merges with the *a-* that follows.', pronunciation: 'vee-pahr-YAH-sah', citations: [wikipediaCitation('Viparyasa')] },
+                { text: 'ātikrānto', type: 'suffix', gloss: 'having crossed beyond, transcended (*atikrānto*; it carries the long *ā* formed where the two words\' vowels merged)', pronunciation: 'ah-tee-KRAHN-toh' },
               ],
             },
           ],
@@ -2778,7 +2778,7 @@ export const bodhiHeartSutra: LiturgyDoc = {
                 'bo-Tibt': [
                   { text: 'མྱ་', type: 'semantic', pronunciation: 'mya', gloss: 'ill, afflicted' },
                   { text: 'ངན་', type: 'semantic', pronunciation: 'ngan', gloss: 'bad — together མྱ་ངན ("mya ngan") = "sorrow, grief, affliction"' },
-                  { text: 'ལས་', type: 'semantic', pronunciation: 'las', gloss: 'from, beyond (ablative particle)' },
+                  { text: 'ལས་', type: 'semantic', pronunciation: 'las', gloss: 'from, beyond — the particle that marks "from"' },
                   { text: 'འདས་པ', type: 'semantic', pronunciation: "'das pa", gloss: 'having passed beyond — together: "passed-beyond sorrow" = *nirvāṇa*' },
                 ],
               },
diff --git a/data/liturgy/bodhicitta-dedication.ts b/data/liturgy/bodhicitta-dedication.ts
index 0711aee..6b9a099 100644
--- a/data/liturgy/bodhicitta-dedication.ts
+++ b/data/liturgy/bodhicitta-dedication.ts
@@ -170,7 +170,7 @@ export const bodhicittaDedication: LiturgyDoc = {
             {
               form: 'kyé pa',
               scriptAlts: { 'bo-Tibt': 'སྐྱེས་པ' },
-              etymology: 'Tibetan *skyes pa* — past participle of *skye* "arise"',
+              etymology: 'Tibetan *skyes pa* — the "arisen" form of the verb *skye* "to arise"',
               gloss: 'having arisen, the arisen one — together with *ma* = "not-arisen"',
             },
             {
@@ -293,7 +293,7 @@ export const bodhicittaDedication: LiturgyDoc = {
             {
               form: 'né',
               scriptAlts: { 'bo-Tibt': 'ནས' },
-              etymology: 'Tibetan ablative particle *nas* — "from"',
+              etymology: 'Tibetan *nas* — the particle that marks "from"',
               gloss: 'from — *gong né* = "from higher"',
             },
             {
diff --git a/data/liturgy/heart-sutra.ts b/data/liturgy/heart-sutra.ts
index aa9faf7..bd3d5dc 100644
--- a/data/liturgy/heart-sutra.ts
+++ b/data/liturgy/heart-sutra.ts
@@ -164,7 +164,7 @@ export const heartSutra: LiturgyDoc = {
               gloss: 'the Noble Lord Avalokiteśvara — the bodhisattva of compassion',
               citations: [wikipediaCitation('Avalokiteśvara')],
               morphemes: [
-                { text: 'Ārya', type: 'prefix', gloss: 'noble — an honourific', pronunciation: 'AHR-yah', citations: [wikipediaCitation('Ārya')] },
+                { text: 'Āry', type: 'prefix', gloss: 'noble — an honorific (*ārya*; its final vowel merges with the *ā-* of the next piece)', pronunciation: 'AHR-yah', citations: [wikipediaCitation('Ārya')] },
                 { text: 'āvalokit', type: 'stem', gloss: 'one who looks down (with compassion) on the world', pronunciation: 'ah-vah-loh-KEET' },
                 { text: 'eśvaro', type: 'suffix', gloss: 'lord — the title-ending', pronunciation: 'AYSH-vah-roh' },
               ],
@@ -342,7 +342,7 @@ export const heartSutra: LiturgyDoc = {
               citations: [wikipediaCitation('Pāramitā')],
               morphemes: [
                 { text: 'pāram', type: 'stem', gloss: 'the other shore — what awakening looks like from this shore', pronunciation: 'PAH-rahm' },
-                { text: 'itā', type: 'suffix', gloss: 'gone (past participle, feminine) — turning the noun into "having gone"', pronunciation: 'ee-TAH' },
+                { text: 'itā', type: 'suffix', gloss: 'gone — the "X-ed" form, turning the word into "having gone"', pronunciation: 'ee-TAH' },
               ],
             },
             {
@@ -350,11 +350,11 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'चरमाणो',
               scriptAlts: { 'zh-Hant': '行', 'ja-Jpan': '行', 'bo-Tibt': 'སྤྱོད་པ' },
               pronunciation: 'chah-rah-MAH-noh',
-              etymology: '√car "to move, practise" + middle present participle',
+              etymology: '√car "to move, practise" + the "-ing" ending (the form that means doing it for oneself)',
               gloss: 'practising, moving in — the bodhisattva is *doing* this, not just thinking it',
               morphemes: [
                 { text: 'cara', type: 'root', root: '√car', gloss: 'to move, walk, practise', pronunciation: 'CHAH-rah', conceptIds: ['concept.practice-carya'] },
-                { text: 'māṇo', type: 'suffix', gloss: 'middle present participle — the one doing [the moving] for themselves', pronunciation: 'MAH-noh' },
+                { text: 'māṇo', type: 'suffix', gloss: 'the "-ing" form — here, the one doing [the moving] for themselves', pronunciation: 'MAH-noh' },
               ],
             },
           ],
@@ -484,7 +484,7 @@ export const heartSutra: LiturgyDoc = {
               form: 'tāṃś',
               scriptAlt: 'तांश्',
               pronunciation: 'TAHM-sh',
-              etymology: '*tān* (accusative plural masculine of *tad*) + sandhi *-ś*',
+              etymology: '*tān* — the "those" object-form of *tad* — with a sandhi *-ś*',
               gloss: 'those — refers back to the five skandhas',
             },
             {
@@ -613,7 +613,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'रूपान्',
               scriptAlts: { 'zh-Hant': '色', 'ja-Jpan': '色', 'bo-Tibt': 'གཟུགས' },
               pronunciation: 'ROO-pahn',
-              etymology: '*rūpa* "form" + ablative ending',
+              etymology: '*rūpa* "form" + the "-from" ending',
               gloss: 'from form — "X is not different *from* form"',
               accent: 'amber',
               citations: [wikipediaCitation('Rūpa')],
@@ -729,7 +729,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'शून्यताया',
               scriptAlts: { 'zh-Hant': '空', 'ja-Jpan': '空', 'bo-Tibt': 'སྟོང་པ' },
               pronunciation: 'SHOON-yah-TAH-yah',
-              etymology: '*śūnyatā* "emptiness" + ablative ending',
+              etymology: '*śūnyatā* "emptiness" + the "-from" ending',
               gloss: 'from emptiness — "X is not different from *emptiness*"',
               accent: 'sky',
             },
@@ -1100,7 +1100,7 @@ export const heartSutra: LiturgyDoc = {
                 ],
                 'bo-Tibt': [
                   { text: 'ཤཱ་རི', type: 'phonetic', pronunciation: 'shā-ri', gloss: 'phonetic: *śāri* (the mother\'s name)' },
-                  { text: 'འི', type: 'semantic', pronunciation: "'i", gloss: 'genitive — "of"' },
+                  { text: 'འི', type: 'semantic', pronunciation: "'i", gloss: 'the particle that marks "of" — possession' },
                   { text: 'བུ', type: 'semantic', pronunciation: 'bu', gloss: 'son — together: "son of Śāri"' },
                 ],
               },
@@ -1287,7 +1287,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'अनुत्पन्ना',
               scriptAlts: { 'zh-Hant': '不生', 'ja-Jpan': '不生', 'bo-Tibt': 'མ་སྐྱེས་པ' },
               pronunciation: 'ah-noot-PAHN-nah',
-              etymology: '*a-* (not) + *utpannā* "arisen" (past participle of √pad)',
+              etymology: '*a-* (not) + *utpannā* "arisen" — the "X-ed" form of √pad',
               gloss: 'unarisen, not having come into being',
               morphemes: [
                 { text: 'an', type: 'prefix', gloss: 'not — the negation', pronunciation: 'ahn' },
@@ -1299,7 +1299,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'अनिरुद्धा',
               scriptAlts: { 'zh-Hant': '不滅', 'ja-Jpan': '不滅', 'bo-Tibt': 'མ་འགགས་པ' },
               pronunciation: 'ah-nee-ROOD-dhah',
-              etymology: '*a-* (not) + *niruddhā* "stopped, ceased" (past participle of √rudh)',
+              etymology: '*a-* (not) + *niruddhā* "stopped, ceased" — the "X-ed" form of √rudh',
               gloss: 'unceased, not stopped',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
@@ -1567,7 +1567,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'तस्माच्',
               scriptAlts: { 'zh-Hant': '是故', 'ja-Jpan': '是故', 'bo-Tibt': 'དེ་ལྟ་བས་ན' },
               pronunciation: 'tahs-MAHCH',
-              etymology: 'ablative of *tat* "that" — "from that", "therefore"',
+              etymology: 'the "from" form of *tat* "that" — "from that", "therefore"',
               gloss: 'therefore — drawing the consequence from what came before',
             },
             {
@@ -1581,7 +1581,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'शून्यतायाम्',
               scriptAlts: { 'zh-Hant': '空', 'ja-Jpan': '空', 'bo-Tibt': 'སྟོང་པ་ཉིད' },
               pronunciation: 'shoon-yah-TAH-yahm',
-              etymology: 'locative of *śūnyatā* "emptiness" — "in emptiness"',
+              etymology: 'the "in" form of *śūnyatā* "emptiness" — "in emptiness"',
               gloss: 'in emptiness',
               accent: 'sky',
             },
@@ -1597,7 +1597,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'रूपम्',
               scriptAlts: { 'zh-Hant': '色', 'ja-Jpan': '色', 'bo-Tibt': 'གཟུགས' },
               pronunciation: 'ROO-pahm',
-              etymology: '*rūpa* "form, appearance" (nominative neuter)',
+              etymology: '*rūpa* "form, appearance"',
               gloss: 'form, appearance — the first of the five skandhas',
               accent: 'amber',
             },
@@ -1850,7 +1850,7 @@ export const heartSutra: LiturgyDoc = {
             { form: 'na', scriptAlt: 'न', pronunciation: 'nah', gloss: 'no' },
             { form: 'raso', scriptAlt: 'रसो', scriptAlts: { 'zh-Hant': '味', 'ja-Jpan': '味', 'bo-Tibt': 'རོ' }, pronunciation: 'RAH-soh', gloss: 'taste, flavour' },
             { form: 'na', scriptAlt: 'न', pronunciation: 'nah', gloss: 'no' },
-            { form: 'spraṣṭavyaṃ', scriptAlt: 'स्प्रष्टव्यम्', scriptAlts: { 'zh-Hant': '觸', 'ja-Jpan': '觸', 'bo-Tibt': 'རེག་བྱ' }, pronunciation: 'sprahsh-TAHV-yahm', gloss: 'touch — that which can be touched (gerundive of √spṛś)' },
+            { form: 'spraṣṭavyaṃ', scriptAlt: 'स्प्रष्टव्यम्', scriptAlts: { 'zh-Hant': '觸', 'ja-Jpan': '觸', 'bo-Tibt': 'རེག་བྱ' }, pronunciation: 'sprahsh-TAHV-yahm', gloss: 'touch — that which can be touched' },
             { form: 'na', scriptAlt: 'न', pronunciation: 'nah', gloss: 'no' },
             { form: 'dharmāḥ', scriptAlt: 'धर्माः', scriptAlts: { 'zh-Hant': '法', 'ja-Jpan': '法', 'bo-Tibt': 'ཆོས' }, pronunciation: 'DHAR-mah', gloss: 'mental objects, dharmas — the objects of the mind-faculty, sixth in the sense-object schema' },
           ],
@@ -2033,7 +2033,7 @@ export const heartSutra: LiturgyDoc = {
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
                 { text: 'vidyā', type: 'stem', root: '√vid', gloss: 'knowledge — *avidyā* = "not-knowing", the first link of dependent origination', pronunciation: 'VEED-yah' },
-                { text: 'kṣayo', type: 'stem', gloss: 'ending, exhaustion (nominative)', pronunciation: 'KSHAH-yoh' },
+                { text: 'kṣayo', type: 'stem', gloss: 'ending, exhaustion', pronunciation: 'KSHAH-yoh' },
               ],
             },
           ],
@@ -2125,7 +2125,7 @@ export const heartSutra: LiturgyDoc = {
               morphemes: [
                 { text: 'jarā', type: 'stem', gloss: 'aging', pronunciation: 'jah-RAH' },
                 { text: 'maraṇa', type: 'stem', root: '√mṛ', gloss: 'death — twelfth and final link of dependent origination', pronunciation: 'mah-RAH-nah' },
-                { text: 'kṣayaḥ', type: 'stem', gloss: 'ending, exhaustion (nominative)', pronunciation: 'KSHAH-yahh' },
+                { text: 'kṣayaḥ', type: 'stem', gloss: 'ending, exhaustion', pronunciation: 'KSHAH-yahh' },
               ],
             },
           ],
@@ -2336,18 +2336,18 @@ export const heartSutra: LiturgyDoc = {
             },
           ],
           words: [
-            { form: 'tasmād', scriptAlt: 'तस्माद्', scriptAlts: { 'zh-Hant': '故', 'ja-Jpan': '故' }, pronunciation: 'tahs-MAHD', gloss: 'therefore (ablative of *tat* "that")' },
+            { form: 'tasmād', scriptAlt: 'तस्माद्', scriptAlts: { 'zh-Hant': '故', 'ja-Jpan': '故' }, pronunciation: 'tahs-MAHD', gloss: 'therefore — literally "from that"' },
             {
               form: 'aprāptitvāt',
               scriptAlt: 'अप्राप्तित्वात्',
               scriptAlts: { 'zh-Hant': '無所得', 'ja-Jpan': '無所得', 'bo-Tibt': 'ཐོབ་པ་མེད་པ' },
               pronunciation: 'ah-PRAHP-teet-vaht',
-              etymology: '*a-* "not" + *prāpti* "attainment" + *-tva* "-ness" + *-āt* (ablative) — "from-the-not-being-of-attainment"',
+              etymology: '*a-* "not" + *prāpti* "attainment" + *-tva* "-ness" + *-āt* "due to" — "from-the-not-being-of-attainment"',
               gloss: 'due to non-attainment — because there is nothing to attain',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
                 { text: 'prāpti', type: 'stem', gloss: 'attainment', pronunciation: 'PRAHP-tee' },
-                { text: 'tvāt', type: 'suffix', gloss: '-ness, due to (ablative)', pronunciation: 'TVAHT' },
+                { text: 'tvāt', type: 'suffix', gloss: 'forms an abstract noun ("-ness") and marks "due to"', pronunciation: 'TVAHT' },
               ],
             },
           ],
@@ -2451,7 +2451,7 @@ export const heartSutra: LiturgyDoc = {
               etymology: '*citta* "mind" + *āvaraṇa* "obscuration, covering"',
               gloss: 'mental obscurations — the veils that hide things as they are',
               morphemes: [
-                { text: 'citta', type: 'stem', gloss: 'mind', pronunciation: 'CHEET-tah' },
+                { text: 'citt', type: 'stem', gloss: 'mind (the stem *citta*; its final vowel merges with the *ā-* that follows)', pronunciation: 'CHEET-tah' },
                 { text: 'āvaraṇa', type: 'suffix', gloss: 'obscuration, cover', pronunciation: 'AH-vah-rah-nah' },
               ],
             },
@@ -2525,12 +2525,12 @@ export const heartSutra: LiturgyDoc = {
               form: 'nāstitvād',
               scriptAlt: 'नास्तित्वाद्',
               pronunciation: 'NAHS-teet-vahd',
-              etymology: '*na* "not" + *asti* "is" + *-tva* "-ness" + *-ād* ablative ending',
+              etymology: '*na* "not" + *asti* "is" + *-tva* "-ness" + *-ād*, the "due to" ending',
               gloss: 'due to the absence (of) — literally "from the not-being-ness of"',
               morphemes: [
                 { text: 'nā', type: 'prefix', gloss: 'not (the negation)', pronunciation: 'nah' },
                 { text: 'sti', type: 'stem', gloss: 'being (root √as "to be")', pronunciation: 'STEE' },
-                { text: 'tvād', type: 'suffix', gloss: '"-ness, from" — forms the abstract noun and marks ablative ("due to")', pronunciation: 'TVAHD' },
+                { text: 'tvād', type: 'suffix', gloss: 'forms an abstract noun ("-ness") and marks "due to"', pronunciation: 'TVAHD' },
               ],
             },
           ],
@@ -2623,7 +2623,7 @@ export const heartSutra: LiturgyDoc = {
               accent: 'emerald',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not — the negation', pronunciation: 'ah' },
-                { text: 'trasto', type: 'stem', gloss: 'afraid (past participle)', pronunciation: 'TRAHS-toh' },
+                { text: 'trasto', type: 'stem', gloss: 'afraid — the "X-ed" form', pronunciation: 'TRAHS-toh' },
               ],
             },
           ],
@@ -2715,8 +2715,8 @@ export const heartSutra: LiturgyDoc = {
               gloss: 'having crossed beyond inverted views — past the four distortions (taking the impure as pure, the impermanent as permanent, etc.)',
               accent: 'emerald',
               morphemes: [
-                { text: 'viparyāsa', type: 'stem', gloss: 'inverted view — seeing things upside-down (impermanent as permanent, suffering as pleasure, etc.)', pronunciation: 'vee-pahr-YAH-sah', citations: [wikipediaCitation('Viparyasa')] },
-                { text: 'atikrānto', type: 'suffix', gloss: 'having crossed beyond, transcended', pronunciation: 'ah-tee-KRAHN-toh' },
+                { text: 'viparyās', type: 'stem', gloss: 'inverted view — seeing things upside-down (impermanent as permanent, suffering as pleasure, etc.). The stem is *viparyāsa*; its final vowel merges with the *a-* that follows.', pronunciation: 'vee-pahr-YAH-sah', citations: [wikipediaCitation('Viparyasa')] },
+                { text: 'ātikrānto', type: 'suffix', gloss: 'having crossed beyond, transcended (*atikrānto*; it carries the long *ā* formed where the two words\' vowels merged)', pronunciation: 'ah-tee-KRAHN-toh' },
               ],
             },
           ],
@@ -2800,7 +2800,7 @@ export const heartSutra: LiturgyDoc = {
                 'bo-Tibt': [
                   { text: 'མྱ་', type: 'semantic', pronunciation: 'mya', gloss: 'ill, afflicted' },
                   { text: 'ངན་', type: 'semantic', pronunciation: 'ngan', gloss: 'bad — together མྱ་ངན ("mya ngan") = "sorrow, grief, affliction"' },
-                  { text: 'ལས་', type: 'semantic', pronunciation: 'las', gloss: 'from, beyond (ablative particle)' },
+                  { text: 'ལས་', type: 'semantic', pronunciation: 'las', gloss: 'from, beyond — the particle that marks "from"' },
                   { text: 'འདས་པ', type: 'semantic', pronunciation: "'das pa", gloss: 'having passed beyond — together: "passed-beyond sorrow" = *nirvāṇa*' },
                 ],
               },
@@ -2900,7 +2900,7 @@ export const heartSutra: LiturgyDoc = {
               etymology: '*tri-* "three" + *adhva* "way, time-period" + *vyavasthita* "established, situated"',
               gloss: 'established in the three times — past, present, and future',
               morphemes: [
-                { text: 'tri', type: 'prefix', gloss: 'three', pronunciation: 'tree' },
+                { text: 'try', type: 'prefix', gloss: 'three (*tri-*; the *i* turns to *y* before the vowel that follows)', pronunciation: 'tree' },
                 { text: 'adhva', type: 'stem', gloss: 'time, way, road', pronunciation: 'AHD-vah' },
                 { text: 'vyavasthitāḥ', type: 'suffix', root: '√sthā', gloss: 'established, dwelling (plural)', pronunciation: 'vyah-vahs-thee-TAHH' },
               ],
@@ -3000,8 +3000,8 @@ export const heartSutra: LiturgyDoc = {
                 ],
               },
               pronunciation: 'prahj-NYAH-pah-rah-mee-TAHM',
-              etymology: '*prajñā* "wisdom" + *pāramitā* "perfection" (accusative)',
-              gloss: 'the perfection of wisdom (accusative — object of "relying on")',
+              etymology: '*prajñā* "wisdom" + *pāramitā* "perfection"',
+              gloss: 'the perfection of wisdom — here the object of "relying on"',
               accent: 'sky',
               morphemes: [
                 { text: 'prajñā', type: 'stem', gloss: 'wisdom', pronunciation: 'prahj-NYAH' },
@@ -3088,7 +3088,7 @@ export const heartSutra: LiturgyDoc = {
                   { text: '阿', type: 'phonetic', pronunciation: 'ā', gloss: 'phonetic: *a-* (the negation)' },
                   { text: '耨', type: 'phonetic', pronunciation: 'nòu', gloss: 'phonetic: *-nu-*' },
                   { text: '多', type: 'phonetic', pronunciation: 'duō', gloss: 'phonetic: *-tta-*' },
-                  { text: '羅', type: 'phonetic', pronunciation: 'luó', gloss: 'phonetic: *-rām* (accusative)' },
+                  { text: '羅', type: 'phonetic', pronunciation: 'luó', gloss: 'phonetic: *-rām*, the object-marking ending' },
                 ],
               },
               pronunciation: 'ah-NOOT-tah-rahm',
@@ -3096,7 +3096,7 @@ export const heartSutra: LiturgyDoc = {
               gloss: 'unsurpassed, supreme',
               morphemes: [
                 { text: 'an', type: 'prefix', gloss: 'not (assimilated *a-* before vowel)', pronunciation: 'ahn' },
-                { text: 'uttarāṃ', type: 'stem', gloss: 'higher, superior (accusative) — "than which nothing is higher"', pronunciation: 'OOT-tah-rahm' },
+                { text: 'uttarāṃ', type: 'stem', gloss: 'higher, superior — "than which nothing is higher"', pronunciation: 'OOT-tah-rahm' },
               ],
             },
             {
@@ -3163,16 +3163,16 @@ export const heartSutra: LiturgyDoc = {
             { by: 'Thich Nhat Hanh (2014)', text: 'Therefore Shariputra, it should be known that the Insight that brings us to the Other Shore', alignTo: [0, -1, -1, -1, -1, 1, -1, -1, 2, -1, -1, -1, -1, -1, -1, -1], url: 'https://plumvillage.org/sutra/the-other-shore', license: 'Plum Village translation, quoted with attribution' },
           ],
           words: [
-            { form: 'tasmāj', scriptAlt: 'तस्माज्', scriptAlts: { 'zh-Hant': '故', 'ja-Jpan': '故' }, pronunciation: 'tahs-MAHJ', gloss: 'therefore (ablative of *tat* "that")' },
+            { form: 'tasmāj', scriptAlt: 'तस्माज्', scriptAlts: { 'zh-Hant': '故', 'ja-Jpan': '故' }, pronunciation: 'tahs-MAHJ', gloss: 'therefore — literally "from that"' },
             {
               form: 'jñātavyaṃ',
               scriptAlt: 'ज्ञातव्यं',
               scriptAlts: { 'zh-Hant': '知', 'ja-Jpan': '知' },
               pronunciation: 'GYAH-tahv-yahm',
-              gloss: 'should be known (gerundive)',
+              gloss: 'should be known',
               morphemes: [
                 { text: 'jñā', type: 'root', root: '√jñā', gloss: 'to know', pronunciation: 'gyah' },
-                { text: 'tavyaṃ', type: 'suffix', gloss: 'gerundive suffix — "should/must be Xed"', pronunciation: 'TAHV-yahm' },
+                { text: 'tavyaṃ', type: 'suffix', gloss: 'the "should-be-Xed" ending — turns the verb into an instruction', pronunciation: 'TAHV-yahm' },
               ],
             },
             {
@@ -3180,7 +3180,7 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'प्रज्ञापारमिता',
               scriptAlts: { 'zh-Hant': '般若波羅蜜多', 'ja-Jpan': '般若波羅蜜多' },
               pronunciation: 'prahj-NYAH-pah-rah-mee-TAH',
-              gloss: 'the Perfection of Wisdom (nominative)',
+              gloss: 'the Perfection of Wisdom',
               accent: 'sky',
               morphemes: [
                 { text: 'prajñā', type: 'stem', root: '√jñā', gloss: 'wisdom — *pra-* "forth" + *√jñā* "to know"', pronunciation: 'prahj-NYAH' },
@@ -3217,7 +3217,7 @@ export const heartSutra: LiturgyDoc = {
               gloss: 'the great mantra',
               morphemes: [
                 { text: 'mahā', type: 'prefix', gloss: 'great', pronunciation: 'mah-HAH' },
-                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra, instrument of thought (nominative)', pronunciation: 'MAHN-trahh' },
+                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra — an instrument of thought', pronunciation: 'MAHN-trahh' },
               ],
             },
           ],
@@ -3251,7 +3251,7 @@ export const heartSutra: LiturgyDoc = {
               morphemes: [
                 { text: 'mahā', type: 'prefix', gloss: 'great', pronunciation: 'mah-HAH' },
                 { text: 'vidyā', type: 'stem', root: '√vid', gloss: 'knowledge, illumination — same root as Chinese 明 (míng "bright, clear")', pronunciation: 'veed-YAH' },
-                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra (nominative)', pronunciation: 'MAHN-trahh' },
+                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra', pronunciation: 'MAHN-trahh' },
               ],
             },
           ],
@@ -3285,7 +3285,7 @@ export const heartSutra: LiturgyDoc = {
               morphemes: [
                 { text: 'an', type: 'prefix', gloss: 'not (assimilated form of *a-* before vowel)', pronunciation: 'ahn' },
                 { text: 'uttara', type: 'stem', gloss: 'higher, superior', pronunciation: 'OOT-tah-rah' },
-                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra (nominative)', pronunciation: 'MAHN-trahh' },
+                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra', pronunciation: 'MAHN-trahh' },
               ],
             },
           ],
@@ -3320,7 +3320,7 @@ export const heartSutra: LiturgyDoc = {
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
                 { text: 'sama', type: 'stem', gloss: 'equal — first occurrence: "no equal"', pronunciation: 'SAH-mah' },
                 { text: 'sama', type: 'stem', gloss: 'equal — second occurrence: "equal to [the unequaled]"', pronunciation: 'SAH-mah' },
-                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra (nominative)', pronunciation: 'MAHN-trahh' },
+                { text: 'mantraḥ', type: 'stem', root: '√man', gloss: 'mantra', pronunciation: 'MAHN-trahh' },
               ],
             },
           ],
@@ -3354,7 +3354,7 @@ export const heartSutra: LiturgyDoc = {
               morphemes: [
                 { text: 'sarva', type: 'prefix', gloss: 'all, every', pronunciation: 'SAHR-vah' },
                 { text: 'duḥkha', type: 'stem', gloss: 'suffering, dis-ease — the First Noble Truth term', pronunciation: 'DOOH-khah' },
-                { text: 'praśamanaḥ', type: 'stem', root: '√śam', gloss: 'calming, stilling, allayer (nominative)', pronunciation: 'prah-SHAH-mah-nahh' },
+                { text: 'praśamanaḥ', type: 'stem', root: '√śam', gloss: 'calming, stilling, allayer', pronunciation: 'prah-SHAH-mah-nahh' },
               ],
             },
           ],
@@ -3378,8 +3378,8 @@ export const heartSutra: LiturgyDoc = {
             { by: 'Thich Nhat Hanh (2014)', text: 'This is the truth that can not be doubted,', alignTo: [-1, -1, -1, 0, -1, 1, -1, -1, -1], url: 'https://plumvillage.org/sutra/the-other-shore', license: 'Plum Village translation, quoted with attribution' },
           ],
           words: [
-            { form: 'satyam', scriptAlt: 'सत्यम्', scriptAlts: { 'zh-Hant': '真實', 'ja-Jpan': '真實' }, pronunciation: 'SAHT-yahm', etymology: '*sat* "being, true" + *-ya* abstract suffix — "truth itself"', gloss: 'true, real (neuter accusative)' },
-            { form: 'amithyatvāt', scriptAlt: 'अमिथ्यत्वात्', scriptAlts: { 'zh-Hant': '不虛', 'ja-Jpan': '不虛' }, pronunciation: 'ah-MEETH-yaht-vaht', etymology: '*a-* "not" + *mithyā* "false" + *-tva* abstract + *-āt* ablative — "due to not being false"', gloss: 'because it is not false (ablative — gives the reason)' },
+            { form: 'satyam', scriptAlt: 'सत्यम्', scriptAlts: { 'zh-Hant': '真實', 'ja-Jpan': '真實' }, pronunciation: 'SAHT-yahm', etymology: '*sat* "being, true" + *-ya* abstract suffix — "truth itself"', gloss: 'true, real' },
+            { form: 'amithyatvāt', scriptAlt: 'अमिथ्यत्वात्', scriptAlts: { 'zh-Hant': '不虛', 'ja-Jpan': '不虛' }, pronunciation: 'ah-MEETH-yaht-vaht', etymology: '*a-* "not" + *mithyā* "false" + *-tva* abstract + *-āt* "due to" — "due to not being false"', gloss: 'because it is not false — the ending carries the "because"' },
           ],
         },
 
@@ -3406,13 +3406,13 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'प्रज्ञापारमितायाम्',
               scriptAlts: { 'zh-Hant': '般若波羅蜜多', 'ja-Jpan': '般若波羅蜜多' },
               pronunciation: 'prahj-NYAH-pah-rah-mee-TAH-yahm',
-              etymology: 'locative of *prajñāpāramitā* — "in / from the Perfection of Wisdom"',
-              gloss: 'in the Perfection of Wisdom (locative)',
+              etymology: 'the "in / from" form of *prajñāpāramitā* — "in / from the Perfection of Wisdom"',
+              gloss: 'in the Perfection of Wisdom',
               accent: 'sky',
               morphemes: [
                 { text: 'prajñā', type: 'stem', root: '√jñā', gloss: 'wisdom', pronunciation: 'prahj-NYAH' },
                 { text: 'pāramitā', type: 'stem', gloss: 'perfection, the going-beyond', pronunciation: 'pah-rah-mee-TAH' },
-                { text: 'yām', type: 'suffix', gloss: 'locative case ending — "in / from"', pronunciation: 'yahm' },
+                { text: 'yām', type: 'suffix', gloss: 'the "-in / -from" ending', pronunciation: 'yahm' },
               ],
             },
             {
@@ -3420,11 +3420,11 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'उक्तो',
               scriptAlts: { 'zh-Hant': '說', 'ja-Jpan': '說' },
               pronunciation: 'OOK-toh',
-              etymology: 'past participle of √vac "to speak"',
+              etymology: 'the "X-ed" form of √vac "to speak"',
               gloss: 'spoken, declared',
               morphemes: [
-                { text: 'ukt', type: 'root', root: '√vac', gloss: 'spoken (past participle of *√vac* "to speak")', pronunciation: 'ookt' },
-                { text: 'o', type: 'suffix', gloss: 'masculine nominative singular ending', pronunciation: 'oh' },
+                { text: 'ukt', type: 'root', root: '√vac', gloss: 'spoken — the "X-ed" form of *√vac* "to speak"', pronunciation: 'ookt' },
+                { text: 'o', type: 'suffix', gloss: 'the ending that marks it the subject — one thing', pronunciation: 'oh' },
               ],
             },
             {
@@ -3432,11 +3432,11 @@ export const heartSutra: LiturgyDoc = {
               scriptAlt: 'मन्त्रः',
               scriptAlts: { 'zh-Hant': '咒', 'ja-Jpan': '咒' },
               pronunciation: 'MAHN-trahh',
-              etymology: '*man-* "to think" + *-tra* instrument suffix — "instrument of thought"',
-              gloss: 'the mantra (nominative)',
+              etymology: '*man-* "to think" + *-tra* "instrument-of" — "instrument of thought"',
+              gloss: 'the mantra',
               morphemes: [
                 { text: 'man', type: 'root', root: '√man', gloss: 'to think', pronunciation: 'mahn' },
-                { text: 'traḥ', type: 'suffix', gloss: '*-tra* instrument suffix + *-aḥ* nominative — "instrument of [thought]"', pronunciation: 'TRAHH' },
+                { text: 'traḥ', type: 'suffix', gloss: '*-tra* "instrument-of" + the subject-marking *-aḥ* — "instrument of [thought]"', pronunciation: 'TRAHH' },
               ],
             },
           ],
diff --git a/data/liturgy/metta-sutta.ts b/data/liturgy/metta-sutta.ts
index 28651bc..3b014f3 100644
--- a/data/liturgy/metta-sutta.ts
+++ b/data/liturgy/metta-sutta.ts
@@ -57,375 +57,1545 @@ export const mettaSutta: LiturgyDoc = {
       shape: 'triple-script-witness',
       large: true,
       segments: [
+        // ── v1a: karaṇīyam-attha-kusalena ──
         {
-          id: 'verse-1',
-          pali: 'Karaṇīyamatthakusalena, yaṁ taṁ santaṁ padaṁ abhisamecca; sakko ujū ca suhujū ca, suvaco c\'assa mudu anatimānī.',
+          id: 'v1a-karaniya',
+          pali: 'karaṇīyam-attha-kusalena',
           scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'karaṇīyam-attha-kusalena' },
+          ],
+          witnesses: [
+            {
+              by: 'Amaravati',
+              text: 'This is what should be done by one who is skilled in goodness,',
+              alignTo: [0, 0, 0, 0, 0, 0, -1, 2, -1, -1, 2, -1, 1],
+              // karaṇīyam: "done" → √kṛ root (morpheme 0); the rest → -yam.
+              // kusalena: "skilled" → kusal stem (0); "one" → -ena ending (1).
+              morphemeAlignTo: [1, 1, 1, 1, 1, 0, null, 1, null, null, 0, null, null],
+              url: AMARAVATI_URL,
+            },
+            {
+              by: 'Sujato (SuttaCentral)',
+              text: 'This is what should be done by one skilled in the good',
+              alignTo: [0, 0, 0, 0, 0, 0, -1, 2, 2, -1, -1, 1],
+              morphemeAlignTo: [1, 1, 1, 1, 1, 0, null, 1, 0, null, null, null],
+              url: 'https://suttacentral.net/snp1.8/en/sujato',
+              license: 'CC0',
+            },
+            {
+              by: 'Thanissaro (Access to Insight)',
+              text: 'This is to be done by one skilled in aims',
+              alignTo: [0, 0, 0, 0, 0, -1, 2, 2, -1, 1],
+              morphemeAlignTo: [1, 1, 1, 1, 0, null, 1, 0, null, null],
+              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
+              license: 'CC BY-NC',
+            },
+          ],
+          words: [
+            {
+              form: 'karaṇīyam',
+              pronunciation: 'kah-rah-NEE-yahm',
+              etymology: 'from the verb meaning "to do, make" — turned into "what ought to be done"',
+              gloss: 'should be done — the opening word frames the whole sutta as instruction: "Here is what you do."',
+              morphemes: [
+                { text: 'karaṇī', type: 'root', root: '√kṛ', gloss: 'from the verb "to do, make" — same root that gives English speakers *karma* (literally "the doing")', pronunciation: 'kah-rah-NEE' },
+                { text: 'yam', type: 'suffix', gloss: 'the "-yam" tail turns the verb "do" into "should be done". Similar to English adding "-able" to make "doable", but with a *must* flavour rather than a *can*.', pronunciation: 'yahm' },
+              ],
+              citations: [dpdCitation('karaṇīya')],
+            },
+            {
+              form: 'attha',
+              pronunciation: 'AHT-tah',
+              etymology: 'one of the most range-y Pāli nouns — covers everything from "the good" to "what something means"',
+              gloss: 'benefit, welfare, good, purpose, aim, meaning — a single word holding all these senses at once. Here: "the good", the worthwhile aim that the skilled person is pursuing.',
+              accent: 'amber',
+              citations: [dpdCitation('attha')],
+            },
             {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Karaṇīyamatthakusalena, yaṁ taṁ santaṁ padaṁ abhisamecca; sakko ujū ca suhujū ca, suvaco c\'assa mudu anatimānī.',
+              form: 'kusalena',
+              pronunciation: 'koo-sah-LAY-nah',
+              etymology: '*kusala* "skilful, wholesome" — the stem\'s final "a" merges with the "-ena" ending',
+              gloss: 'by one who is skilled. *Kusala* is the same word Buddhist texts use for "wholesome" states of mind — skilfulness is moral as well as practical.',
+              morphemes: [
+                { text: 'kusal', type: 'stem', gloss: 'skilful, wholesome — the same word Pāli texts use to mark a "good" mind-state (the opposite of *akusala*, "unwholesome").', pronunciation: 'koo-SAHL' },
+                { text: 'ena', type: 'suffix', gloss: 'the "-ena" tail marks the *doer* of the action — "by this person". English wedges in the word "by" to do the same job; Pāli changes the tail of the word itself.', pronunciation: 'AY-nah' },
+              ],
+              citations: [dpdCitation('kusala')],
             },
           ],
+        },
+        // ── v1b: yaṁ taṁ santaṁ padaṁ abhisamecca ──
+        {
+          id: 'v1b-santam-padam',
+          pali: 'yaṁ taṁ santaṁ padaṁ abhisamecca',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'yaṁ taṁ santaṁ padaṁ abhisamecca' },
+          ],
           witnesses: [
             {
               by: 'Amaravati',
-              text: 'This is what should be done by one who is skilled in goodness, and who knows the path of peace: let them be able and upright, straightforward and gentle in speech, humble and not conceited.',
+              text: 'and who knows the path of peace:',
+              alignTo: [-1, -1, 4, -1, 3, -1, 2],
               url: AMARAVATI_URL,
             },
             {
               by: 'Sujato (SuttaCentral)',
-              text: 'This is what should be done by one skilled in the good who has comprehended the state of peace. They\'d be capable, sincere, and upright, easy to speak to, gentle, and not proud,',
+              text: 'who has comprehended the state of peace.',
+              alignTo: [-1, 4, 4, -1, 3, -1, 2],
+              morphemeAlignTo: [null, 2, 2, null, null, null, null],
               url: 'https://suttacentral.net/snp1.8/en/sujato',
               license: 'CC0',
             },
             {
               by: 'Thanissaro (Access to Insight)',
-              text: 'This is to be done by one skilled in aims who wants to break through to the state of peace: Be capable, upright, and straightforward, easy to instruct, gentle, and not conceited,',
+              text: 'who wants to break through to the state of peace:',
+              alignTo: [-1, -1, -1, 4, 4, -1, -1, 3, -1, 2],
+              morphemeAlignTo: [null, null, null, 2, 1, null, null, null, null, null],
               url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
               license: 'CC BY-NC',
             },
           ],
           words: [
+            { form: 'yaṁ', pronunciation: 'yahm', gloss: 'which, that — a pointer word; in this verse it points to *padaṁ* ("the state") coming up' },
+            { form: 'taṁ', pronunciation: 'tahm', gloss: 'that one — pairs with *yaṁ* the way English pairs "which" with "that"' },
             {
               form: 'santaṁ',
-              gloss: 'Peaceful, calmed. From *√śam* "to calm, quiet". The "path of peace" the worker of goodness is heading toward.',
+              pronunciation: 'SAHN-tahm',
+              etymology: 'from a verb root meaning "to calm, quiet down"',
+              gloss: 'peaceful, calmed — describes the *padaṁ* ("state") that comes next. The same root *√śam* gives the word *śamatha* (tranquility meditation).',
               accent: 'sky',
+              morphemes: [
+                { text: 'sant', type: 'root', root: '√śam', gloss: 'from the verb "to calm, quiet". *Sant* is the form for something that has been calmed — like English "settled" or "stilled".', pronunciation: 'sahnt' },
+                { text: 'aṁ', type: 'suffix', gloss: 'the "-aṁ" tail flags this word as the *thing being acted on* — here, the "peaceful state" that\'s being comprehended. Pronounced as a soft nasal close, like "um" in English "hum".', pronunciation: 'ahm' },
+              ],
               citations: [dpdCitation('santa')],
             },
+            {
+              form: 'padaṁ',
+              pronunciation: 'PAH-dahm',
+              etymology: 'literally "foot, step", broadened to "foothold, place, state, condition" — like English "footing" or "standing"',
+              gloss: 'state, ground, footing — what the commentaries take as *nibbāna*, the peaceful destination. Same root gives English speakers the *-pad* in compound words like *quadruped*.',
+              citations: [dpdCitation('pada')],
+            },
+            {
+              form: 'abhisamecca',
+              pronunciation: 'ah-bhee-sah-MAYCH-cha',
+              etymology: 'three pieces: *abhi-* (toward) + *sam-* (together, fully) + *√i* (to go) — "having gone all the way to it"',
+              gloss: 'having come to fully understand it — literally "having gone right up to" the peaceful state. The word frames understanding as a journey completed, not a thought entertained.',
+              morphemes: [
+                { text: 'abhi', type: 'prefix', gloss: 'a prefix meaning "toward, up to" — adds a reaching-for flavour to whatever follows', pronunciation: 'ah-bhee' },
+                { text: 'sam', type: 'prefix', gloss: 'a prefix meaning "together, fully". Combined with *abhi-* it intensifies into "going all the way".', pronunciation: 'sahm' },
+                { text: 'ecca', type: 'suffix', root: '√i', gloss: 'from the verb "to go". The tail of the word means "having Xed" — so *abhi-sam-ecca* = "having gone fully up to [it]", i.e. having grasped it completely.', pronunciation: 'AYCH-cha' },
+              ],
+            },
           ],
         },
+        // ── v1c: sakko ujū ca suhujū ca ──
         {
-          id: 'verse-2',
-          pali: 'Santussako ca subharo ca, appakicco ca sallahukavutti; santindriyo ca nipako ca, appagabbho kulesu ananugiddho.',
+          id: 'v1c-sakko-uju',
+          pali: 'sakko ujū ca suhujū ca',
           scripts: [
-            {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Santussako ca subharo ca, appakicco ca sallahukavutti; santindriyo ca nipako ca, appagabbho kulesu ananugiddho.',
-            },
+            { lang: 'pi-Latn', label: 'Pāli', text: 'sakko ujū ca suhujū ca' },
           ],
           witnesses: [
             {
               by: 'Amaravati',
-              text: 'Contented and easily satisfied, unburdened with duties and frugal in their ways. Peaceful and calm and wise and skillful, not proud or demanding in nature.',
+              text: 'let them be able and upright,',
+              alignTo: [-1, -1, -1, 0, -1, 1],
               url: AMARAVATI_URL,
             },
             {
               by: 'Sujato (SuttaCentral)',
-              text: 'content and unburdensome, living lightly with few duties, sensible and prudent, not arrogant or fawning over families.',
+              text: 'They\'d be capable, sincere, and upright,',
+              alignTo: [-1, -1, 0, 3, -1, 1],
               url: 'https://suttacentral.net/snp1.8/en/sujato',
               license: 'CC0',
             },
             {
               by: 'Thanissaro (Access to Insight)',
-              text: 'content and easy to support, with few duties, living lightly, with peaceful faculties, masterful, modest, and no greed for supporters.',
+              text: 'Be capable, upright, and straightforward,',
+              alignTo: [-1, 0, 1, -1, 3],
               url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
               license: 'CC BY-NC',
             },
           ],
           words: [
             {
-              form: 'Santussako',
-              gloss: 'Easily content, satisfied with little. *sam* "fully" + *√tuṣ* "be pleased". A core monastic virtue.',
-              accent: 'amber',
+              form: 'sakko',
+              pronunciation: 'SAHK-koh',
+              etymology: 'from the verb *√śak* "to be able" — the first of the worker-of-goodness\'s inner qualities',
+              gloss: 'able, capable — having the strength to follow through',
+              citations: [dpdCitation('sakka')],
+            },
+            {
+              form: 'ujū',
+              pronunciation: 'OO-joo',
+              etymology: 'plain meaning "straight" — both bodily (straight posture) and morally (straight conduct)',
+              gloss: 'upright, straight — neither bent in posture nor crooked in dealings',
+              citations: [dpdCitation('uju')],
+            },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+            {
+              form: 'suhujū',
+              pronunciation: 'SOO-hoo-joo',
+              etymology: '*su-* (the prefix "well, properly") + *uju* (straight). The *s* + *u* combination triggers a soft *h-* in front of *uju* — sound-smoothing.',
+              gloss: 'really, properly straight — *ujū* doubled down. Fully sincere, with no hidden crookedness.',
+              morphemes: [
+                { text: 'su', type: 'prefix', gloss: 'a prefix meaning "well, easily, properly". Like English "well-" in "well-spoken".', pronunciation: 'soo' },
+                { text: 'hujū', type: 'stem', gloss: '*uju* "straight" — with an *h-* added in front to smooth the join with *su-*. Same word as the previous *ujū*, just intensified by the prefix.', pronunciation: 'HOO-joo' },
+              ],
             },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
           ],
         },
+        // ── v1d: suvaco c'assa mudu anatimānī ──
         {
-          id: 'verse-3',
-          pali: 'Na ca khuddhaṁ samācare kiñci, yena viññū pare upavadeyyuṁ; sukhino vā khemino hontu, sabbe sattā bhavantu sukhitattā.',
+          id: 'v1d-suvaco-mudu',
+          pali: 'suvaco c\'assa mudu anatimānī',
           scripts: [
-            {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Na ca khuddhaṁ samācare kiñci, yena viññū pare upavadeyyuṁ; sukhino vā khemino hontu, sabbe sattā bhavantu sukhitattā.',
-            },
+            // tokens hint keeps the sandhi compound c'assa as one hover unit;
+            // without it the default Latin tokenizer splits on the apostrophe.
+            { lang: 'pi-Latn', label: 'Pāli', text: 'suvaco c\'assa mudu anatimānī', tokens: ['suvaco', "c'assa", 'mudu', 'anatimānī'] },
           ],
           witnesses: [
             {
               by: 'Amaravati',
-              text: 'Let them not do the slightest thing that the wise would later reprove. Wishing: in gladness and in safety, may all beings be at ease.',
+              text: 'straightforward and gentle in speech, humble and not conceited.',
+              alignTo: [-1, -1, 2, -1, 0, 3, -1, 3, 3],
+              morphemeAlignTo: [null, null, null, null, null, 2, null, 0, 2],
               url: AMARAVATI_URL,
             },
             {
               by: 'Sujato (SuttaCentral)',
-              text: 'They\'d not do the slightest thing that others might blame them for. Their thought is: \'May all beings be happy and safe! May all beings be happy!\'',
+              text: 'easy to speak to, gentle, and not proud,',
+              alignTo: [0, 0, 0, 0, 2, -1, 3, 3],
+              morphemeAlignTo: [0, 0, 1, 1, null, null, 0, 2],
               url: 'https://suttacentral.net/snp1.8/en/sujato',
               license: 'CC0',
             },
             {
               by: 'Thanissaro (Access to Insight)',
-              text: 'Do not do the slightest thing that the wise would later censure. Think: Happy, at rest, may all beings be happy at heart.',
+              text: 'easy to instruct, gentle, and not conceited,',
+              alignTo: [0, 0, 0, 2, -1, 3, 3],
+              morphemeAlignTo: [0, 0, 1, null, null, 0, 2],
               url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
               license: 'CC BY-NC',
             },
           ],
+          words: [
+            {
+              form: 'suvaco',
+              pronunciation: 'SOO-vah-choh',
+              etymology: '*su-* "well" + *vaca* "speech" (root *√vac*) — "easily-spoken-to"',
+              gloss: 'easy to speak to, amenable to correction',
+              morphemes: [
+                { text: 'su', type: 'prefix', gloss: 'well, easily', pronunciation: 'soo' },
+                { text: 'vaco', type: 'stem', root: '√vac', gloss: '*vaca* "speech" — together with *su-*: "spoken-to, easily-addressed"', pronunciation: 'VAH-choh' },
+              ],
+              citations: [dpdCitation('suvaca')],
+            },
+            {
+              form: "c'assa",
+              pronunciation: 'CHAHS-sah',
+              etymology: '*ca* "and" + *assa* "(may) be" — a run-together of two words; *assa* is the wishing form of *√as* "to be"',
+              gloss: 'and may [one] be — this wishing voice carries through the next several verses',
+            },
+            {
+              form: 'mudu',
+              pronunciation: 'MOO-doo',
+              etymology: 'Pāli *mudu* "soft, gentle" (Skt *mṛdu*)',
+              gloss: 'gentle, soft, mild',
+              citations: [dpdCitation('mudu')],
+            },
+            {
+              form: 'anatimānī',
+              pronunciation: 'ah-nah-tee-MAH-nee',
+              etymology: '*an-* "not" + *ati-* "over, excessive" + *māna* "pride" + *-ī* possessive — "not having excessive pride"',
+              gloss: 'not arrogant, not conceited — the opposite of *atimāna* (excessive self-regard)',
+              morphemes: [
+                { text: 'an', type: 'prefix', gloss: 'not (assimilated *a-* before vowel)', pronunciation: 'ahn' },
+                { text: 'ati', type: 'prefix', gloss: 'over, beyond, excessive', pronunciation: 'AH-tee' },
+                { text: 'mānī', type: 'stem', root: '√man', gloss: 'one having pride (*māna* + possessive *-ī*)', pronunciation: 'MAH-nee' },
+              ],
+              citations: [dpdCitation('māna')],
+            },
+          ],
+        },
+        // ── v2a: santussako ca subharo ca ──
+        {
+          id: 'v2a-santussako',
+          pali: 'santussako ca subharo ca',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'santussako ca subharo ca' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Contented and easily satisfied,', alignTo: [0, -1, 2, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'content and unburdensome,', alignTo: [0, -1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'content and easy to support,', alignTo: [0, -1, 2, 2, 2], morphemeAlignTo: [null, null, 0, 0, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'santussako',
+              pronunciation: 'sahn-TOOS-sah-koh',
+              etymology: '*saṁ* (fully) + *tussako* (one who is pleased — from the verb *√tuṣ* "to be pleased, satisfied"). Same *√tuṣ* gives English *santosha* (contentment) in popular usage.',
+              gloss: 'easily contented — pleased with what one has; the temperament of one not chasing more',
+              accent: 'amber',
+              morphemes: [
+                { text: 'san', type: 'prefix', gloss: 'the prefix *saṁ-* "fully, together" (the *ṁ* assimilates to *n* before *t*)', pronunciation: 'sahn' },
+                { text: 'tussako', type: 'stem', root: '√tuṣ', gloss: 'one who is pleased — from the verb "to be satisfied". The full word means "fully content, easily pleased".', pronunciation: 'TOOS-sah-koh' },
+              ],
+            },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+            {
+              form: 'subharo',
+              pronunciation: 'SOO-bah-roh',
+              etymology: '*su-* (easily) + *bhara* (carrying, supporting — from *√bhṛ* "to bear, carry")',
+              gloss: 'easy to support — needing little from those who give. Used of monastics who don\'t make a fuss about food, robes, lodging.',
+              morphemes: [
+                { text: 'su', type: 'prefix', gloss: 'a prefix meaning "well, easily". Same prefix as in *suvaco* (verse 1).', pronunciation: 'soo' },
+                { text: 'bharo', type: 'stem', root: '√bhṛ', gloss: 'from the verb "to bear, carry". *Subharo* = "easy to carry, easy to support" — undemanding.', pronunciation: 'BAH-roh' },
+              ],
+            },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+          ],
+        },
+        // ── v2b: appakicco ca sallahukavutti ──
+        {
+          id: 'v2b-appakicco',
+          pali: 'appakicco ca sallahukavutti',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'appakicco ca sallahukavutti' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'unburdened with duties and frugal in their ways.', alignTo: [-1, 0, -1, 0, 2, 2, -1, -1], morphemeAlignTo: [null, 0, null, 0, 0, 1, null, null], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'living lightly with few duties,', alignTo: [2, 2, -1, 0, 0], morphemeAlignTo: [1, 0, null, 0, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'with few duties, living lightly,', alignTo: [-1, 0, 0, 2, 2], morphemeAlignTo: [null, 0, 1, 1, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'appakicco',
+              pronunciation: 'AHP-pah-keech-choh',
+              etymology: '*appa* (little, few) + *kicca* (duty, what should be done — from the same verb *√kṛ* as *karaṇīyam* in verse 1)',
+              gloss: 'having few duties — light schedule, few external obligations',
+              morphemes: [
+                { text: 'appa', type: 'prefix', gloss: 'a stem meaning "little, few". Used as a first member of compounds to negate or minimise.', pronunciation: 'AHP-pah' },
+                { text: 'kicco', type: 'stem', root: '√kṛ', gloss: 'duty, what should be done — same verb-root as *karaṇīyam* in verse 1. *Appa-kicco* = "few-duties".', pronunciation: 'KEECH-choh' },
+              ],
+            },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+            {
+              form: 'sallahukavutti',
+              pronunciation: 'sahl-LAH-hoo-kah-VOOT-tee',
+              etymology: '*sallahuka* (very light — *sa(ṁ)* intensifier + *lahuka* "light") + *vutti* (mode of life, conduct — from *√vṛt* "to turn, go on")',
+              gloss: 'living lightly — way-of-life with little baggage, easy of habits',
+              morphemes: [
+                { text: 'sallahuka', type: 'stem', gloss: 'very light, easy — *sa-* intensifies *lahuka* "light" (related to English "light" through Indo-European cousins)', pronunciation: 'sahl-LAH-hoo-kah' },
+                { text: 'vutti', type: 'stem', root: '√vṛt', gloss: 'conduct, way of going on — from a verb meaning "to turn, proceed". Same root as English *vortex*.', pronunciation: 'VOOT-tee' },
+              ],
+            },
+          ],
+        },
+        // ── v2c: santindriyo ca nipako ca ──
+        {
+          id: 'v2c-santindriyo',
+          pali: 'santindriyo ca nipako ca',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'santindriyo ca nipako ca' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Peaceful and calm and wise and skillful,', alignTo: [0, -1, 0, -1, 2, -1, 2], morphemeAlignTo: [0, null, 0, null, null, null, null], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'sensible and prudent,', alignTo: [0, -1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'with peaceful faculties, masterful,', alignTo: [-1, 0, 0, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'santindriyo',
+              pronunciation: 'sahn-TEEN-dree-yoh',
+              etymology: '*santa* (peaceful — same *santaṁ* as verse 1) + *indriya* (faculty — the senses, plus mind, treated as a sixth)',
+              gloss: 'with calmed senses — eyes, ears, mind etc. all at peace; not yanked around by sense-impressions',
+              accent: 'sky',
+              morphemes: [
+                { text: 'sant', type: 'root', root: '√śam', gloss: 'calmed, at peace — the same root as *santaṁ* in verse 1', pronunciation: 'sahnt' },
+                { text: 'indriyo', type: 'stem', gloss: 'the senses (sight, hearing, smell, taste, touch, plus mind treated as a sixth). Literally "belonging to Indra" — the powers of the lord of the senses.', pronunciation: 'EEN-dree-yoh' },
+              ],
+            },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+            {
+              form: 'nipako',
+              pronunciation: 'NEE-pah-koh',
+              etymology: 'related to *√pac* "to mature, ripen" — a person whose understanding has fully ripened',
+              gloss: 'wise, prudent — the kind of wisdom that comes from maturity, not book-learning',
+            },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+          ],
+        },
+        // ── v2d: appagabbho kulesu ananugiddho ──
+        {
+          id: 'v2d-appagabbho',
+          pali: 'appagabbho kulesu ananugiddho',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'appagabbho kulesu ananugiddho' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'not proud or demanding in nature.', alignTo: [0, 0, -1, 0, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'not arrogant or fawning over families.', alignTo: [0, 0, -1, 2, -1, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'modest, and no greed for supporters.', alignTo: [0, -1, 2, 2, -1, 1], morphemeAlignTo: [null, null, 0, 2, null, null], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'appagabbho',
+              pronunciation: 'AHP-pah-GAHB-bho',
+              etymology: '*appa* (little, not) + *gabbha* (impudent, rude — from a different sense of the word that also means "womb, interior")',
+              gloss: 'not impudent — not pushy with families, not putting on airs',
+              morphemes: [
+                { text: 'appa', type: 'prefix', gloss: 'a stem meaning "little, not" — same prefix as in *appakicco*', pronunciation: 'AHP-pah' },
+                { text: 'gabbho', type: 'stem', gloss: 'impudent, forward, intrusive. *Appa-gabbho* = "not-pushy".', pronunciation: 'GAHB-bho' },
+              ],
+            },
+            {
+              form: 'kulesu',
+              pronunciation: 'koo-LAY-soo',
+              etymology: '*kula* (family, clan, household) — the "-esu" tail means "in/among"',
+              gloss: 'in/among families — i.e., the lay families a monastic visits for alms',
+              morphemes: [
+                { text: 'kul', type: 'stem', gloss: 'family, household. *Kula* covers the extended sense — clan, lineage, the people in a house.', pronunciation: 'kool' },
+                { text: 'esu', type: 'suffix', gloss: 'the "-esu" tail marks "in/among" with a plural — like English saying "in those families".', pronunciation: 'AY-soo' },
+              ],
+            },
+            {
+              form: 'ananugiddho',
+              pronunciation: 'ah-nah-noo-GEED-dho',
+              etymology: '*an-* (not) + *anu-* (along, after) + *giddha* (greedy, attached — from *√gṛdh* "to be greedy")',
+              gloss: 'not running-after with greed — not chasing the families for what they can give',
+              morphemes: [
+                { text: 'an', type: 'prefix', gloss: 'a prefix meaning "not"', pronunciation: 'ahn' },
+                { text: 'anu', type: 'prefix', gloss: 'a prefix meaning "along, after" — adds a "chasing after" flavour', pronunciation: 'AH-noo' },
+                { text: 'giddho', type: 'stem', root: '√gṛdh', gloss: 'greedy, attached — from a verb meaning "to be greedy". *An-anu-giddho* = "not chasing-after-greedily".', pronunciation: 'GEED-dho' },
+              ],
+            },
+          ],
+        },
+        // ── v3a: na ca khuddhaṁ samācare kiñci ──
+        {
+          id: 'v3a-na-khuddham',
+          pali: 'na ca khuddhaṁ samācare kiñci',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'na ca khuddhaṁ samācare kiñci' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Let them not do the slightest thing', alignTo: [-1, -1, 0, 3, -1, 2, 4], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: "They'd not do the slightest thing", alignTo: [-1, 0, 3, -1, 2, 4], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'Do not do the slightest thing', alignTo: [3, 0, 3, -1, 2, 4], morphemeAlignTo: [2, null, 2, null, null, null], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'na', pronunciation: 'nah', gloss: '"not". Combines with the next verb to negate the whole sentence.' },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and. Here it links this verse to the qualities listed in verse 2.' },
+            {
+              form: 'khuddhaṁ',
+              pronunciation: 'KHOOD-dahm',
+              etymology: '*khudda* "small, petty, mean" + the *-ṁ* tail marking it as the object of the action',
+              gloss: 'anything petty or mean — not just small in size, but small-minded',
+            },
+            {
+              form: 'samācare',
+              pronunciation: 'sah-MAH-chah-ray',
+              etymology: '*saṁ-* (together, fully) + *ā-* (toward) + *√car* (to walk, go, act). The whole verb means "to conduct oneself, behave".',
+              gloss: 'should do, should conduct oneself in. The ending makes it "let one do" — a gentle imperative.',
+              morphemes: [
+                { text: 'sam', type: 'prefix', gloss: 'a prefix meaning "fully, together" (the *ṁ* assimilates to *m* before a vowel)', pronunciation: 'sahm' },
+                { text: 'ā', type: 'prefix', gloss: 'a prefix meaning "toward, at"', pronunciation: 'ah' },
+                { text: 'care', type: 'root', root: '√car', gloss: 'from the verb "to walk, conduct oneself, behave". Same root as *caryā* (practice, conduct).', pronunciation: 'CHAH-ray' },
+              ],
+            },
+            { form: 'kiñci', pronunciation: 'KEEN-chee', gloss: '"anything (at all)". Pairs with the negation *na* — "not anything", i.e. "nothing".' },
+          ],
+        },
+        // ── v3b: yena viññū pare upavadeyyuṁ ──
+        {
+          id: 'v3b-yena-vinnu',
+          pali: 'yena viññū pare upavadeyyuṁ',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'yena viññū pare upavadeyyuṁ' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'that the wise would later reprove.', alignTo: [0, -1, 1, -1, -1, 3], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'that others might blame them for.', alignTo: [0, 2, -1, 3, -1, -1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'that the wise would later censure.', alignTo: [0, -1, 1, -1, -1, 3], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'yena',
+              pronunciation: 'YAY-nah',
+              gloss: 'by which, on account of which. A pointer back to the petty thing in 3a — "[the thing] on account of which the wise would reprove".',
+            },
+            {
+              form: 'viññū',
+              pronunciation: 'VEEN-yoo',
+              etymology: '*vi-* (apart, fully) + *√jñā* (to know) — "one who fully discerns"',
+              gloss: 'the wise — specifically, those with discerning knowledge. Same *√jñā* as English *gnostic*, *recognise*.',
+              accent: 'sky',
+              morphemes: [
+                { text: 'vi', type: 'prefix', gloss: 'a prefix meaning "apart, fully" — adds a discerning, analytical flavour', pronunciation: 'vee' },
+                { text: 'ññū', type: 'root', root: '√jñā', gloss: 'from the verb "to know". *Viññū* = "one who discerns, the wise".', pronunciation: 'NYOO' },
+              ],
+            },
+            {
+              form: 'pare',
+              pronunciation: 'PAH-ray',
+              gloss: 'others — the other people, the rest of the community. (Plural of *para* "other".)',
+            },
+            {
+              form: 'upavadeyyuṁ',
+              pronunciation: 'oo-pah-vah-DAYY-oom',
+              etymology: '*upa-* (near, against) + *√vad* (to speak) — "to speak against, reprove". The "-eyyuṁ" tail makes it "they would speak against".',
+              gloss: 'would reprove, would blame — the conditional "would" plus a plural "they".',
+              morphemes: [
+                { text: 'upa', type: 'prefix', gloss: 'a prefix meaning "near, against" — adds a "speaking-against" flavour to the verb that follows', pronunciation: 'OO-pah' },
+                { text: 'vad', type: 'root', root: '√vad', gloss: 'from the verb "to speak". Same root as English *vow* via Indo-European cousins.', pronunciation: 'vahd' },
+                { text: 'eyyuṁ', type: 'suffix', gloss: 'the "-eyyuṁ" tail says "they would (do this)". A conditional plural — *vad-eyyuṁ* = "they would speak".', pronunciation: 'AYY-oom' },
+              ],
+            },
+          ],
+        },
+        // ── v3c: sukhino vā khemino hontu ──
+        {
+          id: 'v3c-sukhino-khemino',
+          pali: 'sukhino vā khemino hontu',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'sukhino vā khemino hontu' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Wishing: in gladness and in safety,', alignTo: [3, -1, 0, 1, -1, 2], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: "'May all beings be happy and safe!", alignTo: [3, -1, -1, 3, 0, 1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'Happy, at rest,', alignTo: [0, -1, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'sukhino',
+              pronunciation: 'SOO-khee-noh',
+              etymology: '*sukha* (happiness, ease) + *-in* (possessive) — "having happiness"',
+              gloss: 'happy ones, those-with-ease. *Sukha* is the everyday opposite of *dukkha* (suffering) — it just means "the good kind of feeling".',
+              accent: 'amber',
+            },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or — though here it almost reads "and" in the listing.' },
+            {
+              form: 'khemino',
+              pronunciation: 'KHAY-mee-noh',
+              etymology: '*khema* (safety, security) + *-in* (possessive) — "having safety"',
+              gloss: 'safe ones, those-out-of-danger. *Khema* is the word for the safety of a city behind walls, or a mind out of reach of harm.',
+              accent: 'sky',
+            },
+            {
+              form: 'hontu',
+              pronunciation: 'HOHN-too',
+              etymology: 'from *√hū / √bhū* "to be". The "-ntu" tail makes it "may they be" — a wish or invocation.',
+              gloss: 'may [they] be. The wishing voice — the heart of the metta sutta starts here.',
+            },
+          ],
+        },
+        // ── v3d: sabbe sattā bhavantu sukhitattā ──
+        {
+          id: 'v3d-sabbe-satta',
+          pali: 'sabbe sattā bhavantu sukhitattā',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'sabbe sattā bhavantu sukhitattā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'may all beings be at ease.', alignTo: [-1, 0, 1, 2, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: "May all beings be happy!'", alignTo: [-1, 0, 1, 2, 3], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'may all beings be happy at heart.', alignTo: [-1, 0, 1, 2, 3, -1, 3], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
           words: [
             {
               form: 'sabbe',
-              gloss: 'All, every. The first appearance of the universalising move that runs through the sutta.',
+              pronunciation: 'SAHB-bay',
+              etymology: 'plural of *sabba* "all, every" (Skt *sarva*)',
+              gloss: 'all, every. The first appearance of the universalising move — the wish covers *everyone*, with no exception.',
               accent: 'rose',
             },
             {
               form: 'sattā',
-              gloss: 'Beings, sentient creatures. *sabbe sattā* = "all beings", the recurring address of the metta wish.',
+              pronunciation: 'SAHT-tah',
+              etymology: 'plural of *satta* "being, sentient creature" — from *√as* "to be", literally "those who are"',
+              gloss: 'beings, sentient creatures. *Sabbe sattā* = "all beings" is the recurring address of the metta wish.',
               accent: 'rose',
               citations: [dpdCitation('satta')],
             },
+            {
+              form: 'bhavantu',
+              pronunciation: 'bah-VAHN-too',
+              etymology: 'from *√bhū* "to be, become" — the "-ntu" tail makes it "may they be"',
+              gloss: 'may [they] be. Same wishing voice as *hontu* in 3c; the verb just changes register.',
+              morphemes: [
+                { text: 'bhav', type: 'root', root: '√bhū', gloss: 'from the verb "to be, become". Same root that gives English *be* through Indo-European cousins.', pronunciation: 'bahv' },
+                { text: 'antu', type: 'suffix', gloss: 'the "-antu" tail says "may they (be)". A wishing form — *bhav-antu* = "may they be".', pronunciation: 'AHN-too' },
+              ],
+            },
+            {
+              form: 'sukhitattā',
+              pronunciation: 'soo-khee-TAHT-tah',
+              etymology: '*sukhita* (made-happy) + *-attā* (state, condition — from *attan* "self")',
+              gloss: 'happy of heart, in a state of happiness. The "-attā" tail names it as an inner state, not just a passing feeling.',
+              accent: 'amber',
+              morphemes: [
+                { text: 'sukhit', type: 'stem', gloss: '"made happy" — from *sukha* (happiness) with a participle ending', pronunciation: 'soo-KHEET' },
+                { text: 'attā', type: 'suffix', gloss: 'the "-attā" tail names the inner state. *Sukhit-attā* = "happy-of-self", happy from the inside.', pronunciation: 'AHT-tah' },
+              ],
+            },
           ],
         },
+        // ── v4a: ye keci pāṇabhūtatthī ──
         {
-          id: 'verse-4',
-          pali: 'Ye keci pāṇabhūtatthī, tasā vā thāvarā vā anavasesā; dīghā vā ye mahantā vā, majjhimā rassakāṇukathūlā.',
+          id: 'v4a-ye-keci',
+          pali: 'ye keci pāṇabhūtatthī',
           scripts: [
-            {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Ye keci pāṇabhūtatthī, tasā vā thāvarā vā anavasesā; dīghā vā ye mahantā vā, majjhimā rassakāṇukathūlā.',
-            },
+            { lang: 'pi-Latn', label: 'Pāli', text: 'ye keci pāṇabhūtatthī' },
           ],
           witnesses: [
+            { by: 'Amaravati', text: 'Whatever living beings there may be,', alignTo: [0, 2, 2, -1, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'Whatever creatures there are—', alignTo: [0, 2, -1, -1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'Whatever beings there may be,', alignTo: [0, 2, -1, -1, -1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'ye', pronunciation: 'yay', gloss: 'which (plural) — pointer word, paired with *keci* below to sweep "whoever".' },
+            { form: 'keci', pronunciation: 'KAY-chee', gloss: '"some, any (plural)". Together *ye keci* = "whoever there is", a sweeping inclusion.' },
             {
-              by: 'Amaravati',
-              text: 'Whatever living beings there may be, whether they are weak or strong, omitting none, the great and the mighty, medium, short or small.',
-              url: AMARAVATI_URL,
+              form: 'pāṇabhūtatthī',
+              pronunciation: 'PAH-nah-boo-TAHT-thee',
+              etymology: '*pāṇa* (living being, breath) + *bhūta* (existing) + *atthī* (one who is). A piled-up compound emphasising "every kind of being that breathes and exists".',
+              gloss: 'every kind of breathing, existing being — Pāli triples the synonym to leave no creature uncovered',
+              accent: 'rose',
+              morphemes: [
+                { text: 'pāṇa', type: 'stem', gloss: 'living being, breath. From *√pan* "to breathe". Same root as English *pneumatic*.', pronunciation: 'PAH-nah' },
+                { text: 'bhūt', type: 'root', root: '√bhū', gloss: '"having become, existing" — past form of the verb "to be". Same root as *bhavantu*.', pronunciation: 'BOOT' },
+                { text: 'atthī', type: 'suffix', gloss: '"one who is" — adds "the being-one" sense. Final word means "every breathing-existing-being-er".', pronunciation: 'AHT-thee' },
+              ],
             },
+          ],
+        },
+        // ── v4b: tasā vā thāvarā vā anavasesā ──
+        {
+          id: 'v4b-tasa-thavara',
+          pali: 'tasā vā thāvarā vā anavasesā',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'tasā vā thāvarā vā anavasesā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'whether they are weak or strong, omitting none,', alignTo: [-1, -1, -1, 0, 1, 2, 4, 4], morphemeAlignTo: [null, null, null, null, null, null, 1, 0], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'none excepted, weak or strong,', alignTo: [4, 4, 0, 1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'weak or strong, without exception,', alignTo: [0, 1, 2, -1, 4], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'tasā', pronunciation: 'TAH-sah', gloss: 'tremblers — beings that startle, the still-frightened (creatures still subject to fear). Sometimes rendered "weak".' },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
+            { form: 'thāvarā', pronunciation: 'TAH-vah-rah', gloss: 'standfast — beings firm, stable, no longer trembling. The pair *tasā / thāvarā* sweeps "the still-fearful and the unshaken".' },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'Whatever creatures there are— none excepted, weak or strong, long or large, medium-sized, short, slender or thick,',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'anavasesā',
+              pronunciation: 'ah-NAH-vah-SAY-sah',
+              etymology: '*an-* (not) + *avasesa* (remainder, what is left over)',
+              gloss: 'with-no-remainder — leaving none out. Like saying "every last one" in English.',
+              morphemes: [
+                { text: 'an', type: 'prefix', gloss: 'a prefix meaning "not"', pronunciation: 'ahn' },
+                { text: 'avasesā', type: 'stem', gloss: '"remainder, leftover" — *an-avasesā* = "without leftover, with nothing left out"', pronunciation: 'ah-vah-SAY-sah' },
+              ],
             },
+          ],
+        },
+        // ── v4c: dīghā vā ye mahantā vā ──
+        {
+          id: 'v4c-digha-mahanta',
+          pali: 'dīghā vā ye mahantā vā',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'dīghā vā ye mahantā vā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'the great and the mighty,', alignTo: [-1, 0, -1, -1, 3], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'long or large,', alignTo: [0, 1, 3], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'long, large,', alignTo: [0, 3], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'dīghā', pronunciation: 'DEE-gah', gloss: 'long ones. Pāli *dīgha* "long" gives the *Dīgha Nikāya* (the Long Discourses) its name.' },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
+            { form: 'ye', pronunciation: 'yay', gloss: 'which (plural) — opens a new "those who are…" clause' },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'Whatever beings there may be, weak or strong, without exception, long, large, middling, short, subtle, blatant,',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'mahantā',
+              pronunciation: 'mah-HAHN-tah',
+              etymology: '*mahā* "great" (Skt and Pāli) + *-nt* participle ending — "the great-being ones"',
+              gloss: 'great ones, large. Same *mahā-* as in *mahāmantraḥ* in the Heart Sutra.',
             },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
           ],
         },
+        // ── v4d: majjhimā rassakāṇukathūlā ──
         {
-          id: 'verse-5',
-          pali: 'Diṭṭhā vā yeva addiṭṭhā, ye ca dūre vasanti avidūre; bhūtā vā sambhavesī vā, sabbe sattā bhavantu sukhitattā.',
+          id: 'v4d-majjhima',
+          pali: 'majjhimā rassakāṇukathūlā',
           scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'majjhimā rassakāṇukathūlā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'medium, short or small.', alignTo: [0, 1, -1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'medium-sized, short, slender or thick,', alignTo: [0, 1, 1, -1, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'middling, short, subtle, blatant,', alignTo: [0, 1, 1, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'majjhimā', pronunciation: 'MAHJ-jhee-mah', gloss: 'middling ones. *Majjhima* is "middle" — gives the *Majjhima Nikāya* (the Middle-Length Discourses) its name. Same idea as "medium".' },
             {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Diṭṭhā vā yeva addiṭṭhā, ye ca dūre vasanti avidūre; bhūtā vā sambhavesī vā, sabbe sattā bhavantu sukhitattā.',
+              form: 'rassakāṇukathūlā',
+              pronunciation: 'RAHS-sah-KAH-noo-kah-THOO-lah',
+              etymology: 'a three-part compound: *rassakā* (short) + *aṇukā* (fine, subtle — same word as in *aṇuvīhī*, "atomic") + *thūlā* (gross, coarse). The whole compound sweeps short-fine-thick — every size.',
+              gloss: 'short ones, fine-subtle ones, thick-coarse ones — three sizes piled together to leave none uncovered',
+              morphemes: [
+                { text: 'rassakā', type: 'stem', gloss: 'short, brief. Plural ending makes it "the short ones".', pronunciation: 'RAHS-sah-kah' },
+                { text: 'ṇuka', type: 'stem', gloss: 'subtle, fine, minute (*aṇuka*; the final vowel shortens before the next piece). *Aṇu-* gives English speakers *atomic* through Indo-European cousins (Sanskrit *aṇu* = atom).', pronunciation: 'noo-kah' },
+                { text: 'thūlā', type: 'stem', gloss: 'gross, thick, coarse — the opposite of *aṇuka*. Together *aṇu-thūla* covers the whole subtle-to-gross axis.', pronunciation: 'THOO-lah' },
+              ],
             },
           ],
+        },
+        // ── v5a: diṭṭhā vā yeva addiṭṭhā ──
+        {
+          id: 'v5a-dittha-adittha',
+          pali: 'diṭṭhā vā yeva addiṭṭhā',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'diṭṭhā vā yeva addiṭṭhā' },
+          ],
           witnesses: [
+            { by: 'Amaravati', text: 'The seen and the unseen,', alignTo: [-1, 0, 1, -1, 3], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'those who are seen or unseen,', alignTo: [-1, -1, -1, 0, 1, 3], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'seen or unseen,', alignTo: [0, 1, 3], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              by: 'Amaravati',
-              text: 'The seen and the unseen, those living near and far away, those born and to-be-born. May all beings be at ease!',
-              url: AMARAVATI_URL,
+              form: 'diṭṭhā',
+              pronunciation: 'DEET-tah',
+              etymology: 'past form of *√dṛś* "to see" — "the seen ones" (creatures we can directly perceive)',
+              gloss: 'the seen — visible beings we can point to. Same root that gives *dassana* (vision, view) later in the sutta.',
+              morphemes: [
+                { text: 'diṭ', type: 'root', root: '√dṛś', gloss: 'from the verb "to see"', pronunciation: 'deet' },
+                { text: 'ṭhā', type: 'suffix', gloss: 'the "-ṭhā" tail makes the verb into "the seen ones" (plural past form)', pronunciation: 'tah' },
+              ],
             },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
+            { form: 'yeva', pronunciation: 'YAY-vah', gloss: '"just, indeed" — emphasis particle, like English "also" or "as well".' },
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'those who are seen or unseen, those living far or near, those born or to be born— may all beings be happy!',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'addiṭṭhā',
+              pronunciation: 'AHD-deet-tah',
+              etymology: '*a-* (not) + *diṭṭhā* (seen) — "the unseen ones"',
+              gloss: 'the unseen — creatures hidden from us (insects in the earth, beings in other realms). The pair *diṭṭhā / addiṭṭhā* sweeps "everything we see and everything we don\'t".',
+              morphemes: [
+                { text: 'ad', type: 'prefix', gloss: 'the prefix "not" (assimilated *a-* before *d*)', pronunciation: 'ahd' },
+                { text: 'diṭṭhā', type: 'stem', root: '√dṛś', gloss: 'same "seen" stem as the previous word — paired here with the negation prefix', pronunciation: 'DEET-tah' },
+              ],
+            },
+          ],
+        },
+        // ── v5b: ye ca dūre vasanti avidūre ──
+        {
+          id: 'v5b-dure-avidure',
+          pali: 'ye ca dūre vasanti avidūre',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'ye ca dūre vasanti avidūre' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'those living near and far away,', alignTo: [0, 3, 4, 1, 2, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'those living far or near,', alignTo: [0, 3, 2, 1, 4], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'near or far,', alignTo: [4, 1, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'ye', pronunciation: 'yay', gloss: 'which (plural) — opens "those who…"' },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
+            { form: 'dūre', pronunciation: 'DOO-ray', gloss: 'far away. Same word as Skt *dūra*, related to English *durable* through "lasting/distant" semantics.' },
+            {
+              form: 'vasanti',
+              pronunciation: 'vah-SAHN-tee',
+              etymology: 'from *√vas* "to dwell, live" — "they dwell"',
+              gloss: 'they dwell, they live. The "-anti" tail makes it "they (do)".',
+              morphemes: [
+                { text: 'vas', type: 'root', root: '√vas', gloss: 'from the verb "to dwell, live in a place"', pronunciation: 'vahs' },
+                { text: 'anti', type: 'suffix', gloss: 'the "-anti" tail says "they (do this)" — plural present', pronunciation: 'AHN-tee' },
+              ],
             },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'seen or unseen, near or far, born or seeking birth: May all beings be happy at heart.',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'avidūre',
+              pronunciation: 'ah-vee-DOO-ray',
+              etymology: '*a-* (not) + *vi-* (apart, away) + *dūre* (far) — "not-far-away"',
+              gloss: 'nearby — literally "not-far-away". The double negation gives a softer "near" than a direct word.',
             },
           ],
+        },
+        // ── v5c: bhūtā vā sambhavesī vā ──
+        {
+          id: 'v5c-bhuta-sambhavesi',
+          pali: 'bhūtā vā sambhavesī vā',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'bhūtā vā sambhavesī vā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'those born and to-be-born.', alignTo: [-1, 0, 1, 2], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'those born or to be born—', alignTo: [-1, 0, 1, -1, -1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'born or seeking birth:', alignTo: [0, 1, 2, 2], morphemeAlignTo: [null, null, 2, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
           words: [
+            {
+              form: 'bhūtā',
+              pronunciation: 'BOO-tah',
+              etymology: 'past form of *√bhū* "to become" — "those who have become, the existing ones"',
+              gloss: 'those who have come into being. *Bhū* gives English speakers *be* through Indo-European cousins.',
+              accent: 'sky',
+            },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
             {
               form: 'sambhavesī',
-              gloss: '"Those still seeking birth", those yet to be born. Paired with *bhūtā* "already become" — the wish covers both the born and the yet-to-be-born.',
+              pronunciation: 'sahm-BAH-vay-see',
+              etymology: '*saṁ-* (together, fully) + *bhava* (being, coming-to-be) + *-esī* (one-who-seeks, from *√iṣ* "to seek")',
+              gloss: 'those still seeking to come-into-being — beings yet to be born. Paired with *bhūtā* "already become" — the wish covers both the born and the yet-to-be-born.',
               accent: 'sky',
+              morphemes: [
+                { text: 'sam', type: 'prefix', gloss: 'the prefix "fully, together" (the *ṁ* assimilates to *m*)', pronunciation: 'sahm' },
+                { text: 'bhav', type: 'root', root: '√bhū', gloss: '"becoming, coming-to-be" — same root as *bhūtā*', pronunciation: 'bahv' },
+                { text: 'esī', type: 'suffix', root: '√iṣ', gloss: 'from the verb "to seek". *Sam-bhav-esī* = "one who is seeking-to-come-into-being".', pronunciation: 'AY-see' },
+              ],
             },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
           ],
         },
+        // ── v5d: sabbe sattā bhavantu sukhitattā (refrain — same as v3d) ──
         {
-          id: 'verse-6',
-          pali: 'Na paro paraṁ nikubbetha, nātimaññetha katthaci naṁ kanci; byārosanā paṭighasaññā, nāññamaññassa dukkhamiccheyya.',
+          id: 'v5d-sabbe-satta',
+          pali: 'sabbe sattā bhavantu sukhitattā',
           scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'sabbe sattā bhavantu sukhitattā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'May all beings be at ease!', alignTo: [-1, 0, 1, 2, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'may all beings be happy!', alignTo: [-1, 0, 1, 2, 3], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'May all beings be happy at heart.', alignTo: [-1, 0, 1, 2, 3, -1, 3], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'sabbe', pronunciation: 'SAHB-bay', gloss: 'all, every — same word as in verse 3, the refrain that recurs through the sutta.', accent: 'rose' },
+            { form: 'sattā', pronunciation: 'SAHT-tah', gloss: 'beings — same word as in verse 3. *sabbe sattā* is the metta sutta\'s recurring address.', accent: 'rose', citations: [dpdCitation('satta')] },
+            { form: 'bhavantu', pronunciation: 'bah-VAHN-too', gloss: 'may [they] be — the wishing voice from verse 3, repeated here as the refrain closes the stanza.' },
+            { form: 'sukhitattā', pronunciation: 'soo-khee-TAHT-tah', gloss: 'happy of heart, happy from the inside — same word as in verse 3.', accent: 'amber' },
+          ],
+        },
+        // ── v6a: na paro paraṁ nikubbetha ──
+        {
+          id: 'v6a-na-paro-param',
+          pali: 'na paro paraṁ nikubbetha',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'na paro paraṁ nikubbetha' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Let none deceive another,', alignTo: [-1, 0, 3, 2], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'Let none deceive another,', alignTo: [-1, 0, 3, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'Let no one deceive another', alignTo: [-1, 0, 1, 3, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'na', pronunciation: 'nah', gloss: 'not. Negates the verb at the end of the line.' },
+            { form: 'paro', pronunciation: 'PAH-roh', gloss: 'one person, "the other one" (subject form). *Para* = "other" — same root as English *par* / *peer*.' },
+            {
+              form: 'paraṁ',
+              pronunciation: 'PAH-rahm',
+              etymology: '*para* "other" + *-ṁ* (object-of-the-action ending)',
+              gloss: '"the other one" (object form). Pāli repeats *paro/paraṁ* — "one-person, the-other-person" — instead of saying "anyone…anyone".',
+            },
             {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Na paro paraṁ nikubbetha, nātimaññetha katthaci naṁ kanci; byārosanā paṭighasaññā, nāññamaññassa dukkhamiccheyya.',
+              form: 'nikubbetha',
+              pronunciation: 'nee-KOOB-bay-tah',
+              etymology: '*ni-* (down, against) + *√kub* (a verb of crookedness, "to cheat") + the "-etha" tail (third-person "let him/her"). Together: "let one deceive".',
+              gloss: 'should deceive, should cheat. With *na* in front: "let no one deceive".',
+              morphemes: [
+                { text: 'ni', type: 'prefix', gloss: 'a prefix meaning "down, against" — adds an against-someone flavour', pronunciation: 'nee' },
+                { text: 'kubb', type: 'root', root: '√kub', gloss: 'from a verb meaning "to be crooked, to cheat"', pronunciation: 'KOOB' },
+                { text: 'etha', type: 'suffix', gloss: 'the "-etha" tail says "let one (do this)" — a gentle command', pronunciation: 'AY-tah' },
+              ],
             },
           ],
+        },
+        // ── v6b: nātimaññetha katthaci naṁ kanci ──
+        {
+          id: 'v6b-natimannetha',
+          pali: 'nātimaññetha katthaci naṁ kanci',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'nātimaññetha katthaci naṁ kanci' },
+          ],
           witnesses: [
+            { by: 'Amaravati', text: 'or despise any being in any state.', alignTo: [-1, 0, 3, 3, -1, 1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'nor despise anyone anywhere,', alignTo: [-1, 0, 3, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'or despise anyone anywhere,', alignTo: [-1, 0, 3, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              by: 'Amaravati',
-              text: 'Let none deceive another, or despise any being in any state. Let none through anger or ill-will wish harm upon another.',
-              url: AMARAVATI_URL,
+              form: 'nātimaññetha',
+              pronunciation: 'NAH-tee-MAHN-yay-tah',
+              etymology: '*na* (not) + *ati-* (over, excessive) + *maññetha* (one should think/regard — from *√man* "to think"). The whole verb: "should not think-down-upon", "should not despise".',
+              gloss: '"should not despise" — should not regard someone as less than oneself. The same *ati-māna* "excessive self-regard" we met in verse 1 as *anatimānī*, but now phrased as an action.',
+              morphemes: [
+                { text: 'nā', type: 'prefix', gloss: 'the word *na* "not". Its vowel has lengthened — *na* + *ati* runs together as *nāti*, a common Pāli sound-join where two short *a*\'s merge into one long *ā*.', pronunciation: 'nah' },
+                { text: 'ti', type: 'prefix', gloss: 'the tail of *ati* "over, excessive" — the opening *a-* of *ati* was absorbed into the *nā* before it. Same *ati* as in *anatimānī* (verse 1).', pronunciation: 'tee' },
+                { text: 'maññetha', type: 'stem', root: '√man', gloss: 'from the verb "to think, regard". The "-etha" tail makes it "let one think/regard".', pronunciation: 'MAHN-yay-tah' },
+              ],
             },
+            { form: 'katthaci', pronunciation: 'KAHT-tah-chee', gloss: '"anywhere" (the *-ci* suffix makes pronouns sweeping — like English "-ever" in "wherever").' },
+            { form: 'naṁ', pronunciation: 'nahm', gloss: 'him (object form of "this one"). Refers back to the *paro* introduced in 6a.' },
+            { form: 'kanci', pronunciation: 'KAHN-chee', gloss: '"anyone" — *ka* "who" + *-ci* sweeping. Together with *katthaci*: "anywhere, anyone".' },
+          ],
+        },
+        // ── v6c: byārosanā paṭighasaññā ──
+        {
+          id: 'v6c-byarosana-patighasanna',
+          pali: 'byārosanā paṭighasaññā',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'byārosanā paṭighasaññā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Let none through anger or ill-will', alignTo: [-1, -1, -1, 0, 1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'out of anger or hostility.', alignTo: [-1, -1, 0, 1, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'or through anger or irritation', alignTo: [-1, -1, 0, 1, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'Let none deceive another, nor despise anyone anywhere, nor wish harm for one another out of anger or hostility.',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'byārosanā',
+              pronunciation: 'byah-ROH-sah-nah',
+              etymology: '*vi-* (apart) + *ā-* (toward) + *√ruṣ* "to be angry" — "anger turned against someone"',
+              gloss: 'out of anger — the hot kind of anger that flares outward toward a specific target',
             },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'Let no one deceive another or despise anyone anywhere, or through anger or irritation wish for another to suffer.',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'paṭighasaññā',
+              pronunciation: 'pah-TEE-gah-SAHN-yah',
+              etymology: '*paṭi-* (back, against) + *gha* (striking — from *√han* "to strike") + *saññā* (perception, sense)',
+              gloss: 'with a sense of resistance — the cold pushed-back-against feeling of ill-will, distinct from hot anger. *Paṭigha* is the technical term for aversion in Buddhist psychology.',
+              morphemes: [
+                { text: 'paṭi', type: 'prefix', gloss: 'a prefix meaning "back, against" — adds a "pushing-back" flavour', pronunciation: 'pah-TEE' },
+                { text: 'gha', type: 'stem', root: '√han', gloss: 'from the verb "to strike". *Paṭi-gha* = "strike-back" — the recoiling, push-back of aversion.', pronunciation: 'gah' },
+                { text: 'saññā', type: 'stem', gloss: 'perception, sense, recognising. Adds an "experiencing-as" flavour — *paṭighasaññā* = "the experience of pushing-back".', pronunciation: 'SAHN-yah' },
+              ],
             },
           ],
         },
+        // ── v6d: nāññamaññassa dukkhamiccheyya ──
         {
-          id: 'verse-7',
-          pali: 'Mātā yathā niyaṁ puttaṁ, āyusā ekaputtamanurakkhe; evampi sabbabhūtesu, mānasaṁ bhāvaye aparimāṇaṁ.',
+          id: 'v6d-aniccheyya',
+          pali: 'nāññamaññassa dukkhamiccheyya',
           scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'nāññamaññassa dukkhamiccheyya' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'wish harm upon another.', alignTo: [1, 1, -1, 0], morphemeAlignTo: [1, 0, null, null], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'nor wish harm for one another', alignTo: [-1, 1, 1, -1, 0, 0], morphemeAlignTo: [null, 1, 0, null, 1, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'wish for another to suffer.', alignTo: [1, -1, 0, -1, 1], morphemeAlignTo: [1, null, null, null, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'nāññamaññassa',
+              pronunciation: 'NAHN-yah-MAHN-yahs-sah',
+              etymology: '*na* (not) + *aññamañña* "one-another" (the Pāli word for reciprocity) + *-assa* (object marker, "to/for")',
+              gloss: '"not for one-another" — the line\'s subject. *Aññamañña* literally pairs two "other"s: "this-other and that-other", i.e. reciprocally.',
+              morphemes: [
+                { text: 'nā', type: 'prefix', gloss: 'the word *na* "not". Its vowel has lengthened — *na* + *aññamañña* runs together as *nāññamañña*, two short *a*\'s merging into one long *ā*.', pronunciation: 'nah' },
+                { text: 'ññamañña', type: 'stem', gloss: 'the Pāli word for reciprocity — *añña* "other" doubled to mean "one-another, mutually" (its opening *a-* merged into the *nā* before it).', pronunciation: 'AHN-yah-MAHN-yah' },
+                { text: 'ssa', type: 'suffix', gloss: 'the "-(a)ssa" tail marks "to" or "for" — *aññamaññassa* = "for one another"', pronunciation: 'sah' },
+              ],
+            },
             {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Mātā yathā niyaṁ puttaṁ, āyusā ekaputtamanurakkhe; evampi sabbabhūtesu, mānasaṁ bhāvaye aparimāṇaṁ.',
+              form: 'dukkhamiccheyya',
+              pronunciation: 'DOOK-kah-meech-CHAYY-yah',
+              etymology: '*dukkha* (suffering — the First Noble Truth term) + *iccheyya* (one should wish — from *√iṣ* "to wish, desire")',
+              gloss: 'should wish suffering — the verb the line negates with *na*. Together: "should not wish suffering for one-another".',
+              morphemes: [
+                { text: 'dukkham', type: 'stem', gloss: 'suffering — the First Noble Truth term. The "-am" tail marks it as the object of the wishing.', pronunciation: 'DOOK-kahm' },
+                { text: 'iccheyya', type: 'stem', root: '√iṣ', gloss: 'from the verb "to wish, desire". The "-eyya" tail says "should/would wish".', pronunciation: 'eech-CHAYY-yah' },
+              ],
             },
           ],
+        },
+        // ── v7a: mātā yathā niyaṁ puttaṁ ──
+        {
+          id: 'v7a-mata-yatha',
+          pali: 'mātā yathā niyaṁ puttaṁ',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'mātā yathā niyaṁ puttaṁ' },
+          ],
           witnesses: [
+            { by: 'Amaravati', text: 'Even as a mother', alignTo: [-1, 1, -1, 0], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'As a mother', alignTo: [1, -1, 0], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'As a mother', alignTo: [1, -1, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              by: 'Amaravati',
-              text: 'Even as a mother protects with her life her child, her only child, so should one protect the boundless heart that loves all beings.',
-              url: AMARAVATI_URL,
+              form: 'mātā',
+              pronunciation: 'MAH-tah',
+              etymology: 'Pāli *mātā* "mother". Same root that gives English *mother*, Latin *mater*, Greek *meter* through Indo-European cousins.',
+              gloss: 'a mother — the subject of the line. The simile that the next three lines unpack.',
             },
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'As a mother would protect with her life her one and only child, so they\'d cultivate a heart of love without limit for all sentient beings.',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'yathā',
+              pronunciation: 'yah-TAH',
+              gloss: '"as, like" — the simile word. *Yathā… evampi…* = "as… even so…" frames the whole verse.',
             },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'As a mother would risk her life to protect her child, her only child, even so should one cultivate the heart limitlessly with regard to all beings.',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'niyaṁ',
+              pronunciation: 'NEE-yahm',
+              gloss: '"own, of-one\'s-own". Same root as Skt *nija* "innate". Emphasises that the child is the mother\'s own.',
+            },
+            {
+              form: 'puttaṁ',
+              pronunciation: 'POOT-tahm',
+              etymology: 'Pāli *putta* "son, child" + the *-ṁ* tail (object of the action)',
+              gloss: 'a child, a son (object form). Same root as Skt *putra*.',
+            },
+          ],
+        },
+        // ── v7b: āyusā ekaputtamanurakkhe ──
+        {
+          id: 'v7b-ayusa-ekaputta',
+          pali: 'āyusā ekaputtamanurakkhe',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'āyusā ekaputtamanurakkhe' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'protects with her life her child, her only child,', alignTo: [1, -1, 0, 0, 1, 1, 1, 1, 1], morphemeAlignTo: [3, null, null, null, 3, 1, 1, 0, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'would protect with her life her one and only child,', alignTo: [-1, 1, -1, 0, 0, 1, 1, -1, 1, 1], morphemeAlignTo: [null, 3, null, null, null, 3, 0, null, 0, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'would risk her life to protect her child, her only child,', alignTo: [-1, -1, 0, 0, -1, 1, 1, 1, -1, 1, 1], morphemeAlignTo: [null, null, null, null, null, 3, 3, 1, null, 0, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'āyusā',
+              pronunciation: 'AH-yoo-sah',
+              etymology: '*āyu* "life, lifespan" + *-sā* (ending marking "by means of" — "with her life")',
+              gloss: '"with her life" — the cost the mother pays. *Āyu* is the same word in *Āyurveda* ("knowledge of life").',
+            },
+            {
+              form: 'ekaputtamanurakkhe',
+              pronunciation: 'AY-kah-POOT-tah-mah-noo-RAHK-khay',
+              etymology: '*eka* (one) + *putta* (child) + *anurakkhe* (would protect — *anu-* "along, after" + *√rakṣ* "to protect")',
+              gloss: '"would protect her one-and-only child". A four-piece compound: *one* + *child* + *along-with* + *protect*. The whole verse pivots on this image — the limit-case of love.',
+              accent: 'amber',
+              morphemes: [
+                { text: 'eka', type: 'prefix', gloss: '"one, only" — emphasises the only-child framing', pronunciation: 'AY-kah' },
+                { text: 'putta', type: 'stem', gloss: '"child, son" — same word as *puttaṁ* in 7a', pronunciation: 'POOT-tah' },
+                { text: 'manu', type: 'prefix', gloss: 'sandhi form of *anu-* ("along, after") — the *m* at the join is a sound-bridging consonant', pronunciation: 'mah-noo' },
+                { text: 'rakkhe', type: 'root', root: '√rakṣ', gloss: 'from the verb "to protect, guard". The "-e" tail makes it "would protect" — the wishful voice.', pronunciation: 'RAHK-khay' },
+              ],
+            },
+          ],
+        },
+        // ── v7c: evampi sabbabhūtesu ──
+        {
+          id: 'v7c-evampi-sabbabhuta',
+          pali: 'evampi sabbabhūtesu',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'evampi sabbabhūtesu' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'so should one protect the boundless heart that loves all beings.', alignTo: [0, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: "so they'd cultivate a heart of love without limit for all sentient beings.", alignTo: [0, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1], morphemeAlignTo: [null, null, null, null, null, null, null, null, null, null, 0, 1, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'even so should one cultivate the heart limitlessly with regard to all beings.', alignTo: [0, 0, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1], morphemeAlignTo: [1, 0, null, null, null, null, null, null, null, null, null, 0, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            {
+              form: 'evampi',
+              pronunciation: 'AY-vahm-pee',
+              etymology: '*evaṁ* "thus, in this way" + *pi* "too, also"',
+              gloss: '"even so", "in just this way too". The simile-pivot — *yathā* in 7a was "as"; *evampi* here completes "as… so too…".',
+              morphemes: [
+                { text: 'evam', type: 'stem', gloss: '"thus, in this way" — Pāli\'s pointer-back word', pronunciation: 'AY-vahm' },
+                { text: 'pi', type: 'suffix', gloss: '"too, also" — emphasises the parallel: "even so, just so"', pronunciation: 'pee' },
+              ],
+            },
+            {
+              form: 'sabbabhūtesu',
+              pronunciation: 'SAHB-bah-BOO-tay-soo',
+              etymology: '*sabba* (all) + *bhūta* (being, existing — same word as in verse 5) + *-esu* (plural "in/among")',
+              gloss: '"in/toward all beings" — the wide target of the cultivation. The "-esu" tail marks plural location: like English "among all beings".',
+              accent: 'rose',
+              morphemes: [
+                { text: 'sabba', type: 'stem', gloss: '"all, every" — same root as *sabbe* in verse 3', pronunciation: 'SAHB-bah' },
+                { text: 'bhūt', type: 'root', root: '√bhū', gloss: '"being, having become" — same root as *bhūtā* in verse 5', pronunciation: 'BOOT' },
+                { text: 'esu', type: 'suffix', gloss: 'the "-esu" tail marks "among/in" with a plural — "among all beings"', pronunciation: 'AY-soo' },
+              ],
             },
           ],
+        },
+        // ── v7d: mānasaṁ bhāvaye aparimāṇaṁ ──
+        {
+          id: 'v7d-manasam-bhavaye',
+          pali: 'mānasaṁ bhāvaye aparimāṇaṁ',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'mānasaṁ bhāvaye aparimāṇaṁ' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'so should one protect the boundless heart that loves all beings.', alignTo: [-1, -1, -1, -1, -1, 2, 0, -1, -1, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'cultivate a heart of love without limit', alignTo: [1, -1, 0, -1, -1, -1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'cultivate the heart limitlessly', alignTo: [1, -1, 0, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
           words: [
             {
               form: 'mānasaṁ',
-              gloss: '"The heart-mind". *manas* + accusative. The thing to be cultivated boundlessly.',
+              pronunciation: 'MAH-nah-sahm',
+              etymology: '*manas* (mind, heart-mind) + the *-aṁ* tail (object form). The thing-to-be-cultivated.',
+              gloss: 'the heart-mind — *manas* covers what English splits into "mind" + "heart". The thing the line says to cultivate.',
               accent: 'sky',
               citations: [dpdCitation('manas')],
             },
             {
               form: 'bhāvaye',
-              gloss: '"Let one cultivate, let one bring into being". Optative of *√bhū* "to be", causative — to *cause to become*. The technical term for meditative cultivation.',
+              pronunciation: 'BAH-vah-yay',
+              etymology: 'wishing form of the verb *bhāveti* "to cultivate, bring into being". *Bhāveti* is the same root *√bhū* "to be", in its causative form: "to *cause to become*".',
+              gloss: '"let one cultivate, let one bring into being". The technical term for meditative cultivation — *bhāvanā* (development) shares this root. Whatever the action is here, it\'s a *bringing into being*, not just a feeling.',
               accent: 'amber',
+              morphemes: [
+                { text: 'bhāv', type: 'root', root: '√bhū', gloss: 'from the verb "to be, become" — same root as *bhavantu* in verse 3. Here in causative: "to make-become, to cultivate".', pronunciation: 'bahv' },
+                { text: 'aye', type: 'suffix', gloss: 'the "-aye" tail says "would/should (do this)" — a wish or gentle instruction', pronunciation: 'AH-yay' },
+              ],
               citations: [dpdCitation('bhāveti')],
             },
+            {
+              form: 'aparimāṇaṁ',
+              pronunciation: 'ah-PAH-ree-MAH-nahm',
+              etymology: '*a-* (not) + *parimāṇa* (measure, limit — from *pari-* "around" + *√mā* "to measure")',
+              gloss: '"limitless, without measure" — describing the *mānasaṁ* (heart-mind) the line says to cultivate. Same root *√mā* gives English speakers *measure* through Indo-European cousins.',
+              accent: 'rose',
+              morphemes: [
+                { text: 'a', type: 'prefix', gloss: '"not"', pronunciation: 'ah' },
+                { text: 'pari', type: 'prefix', gloss: '"around, all-around" — adds an enclosing flavour to the next part', pronunciation: 'PAH-ree' },
+                { text: 'māṇaṁ', type: 'stem', root: '√mā', gloss: 'from the verb "to measure". *Pari-māṇa* = "all-around-measure" = limit. *A-pari-māṇa* = "without limit, boundless".', pronunciation: 'MAH-nahm' },
+              ],
+            },
           ],
         },
+        // ── v8a: mettañca sabbalokasmiṁ ──
         {
-          id: 'verse-8',
-          pali: 'Mettañca sabbalokasmiṁ, mānasaṁ bhāvaye aparimāṇaṁ; uddhaṁ adho ca tiriyañca, asambādhaṁ averaṁ asapattaṁ.',
+          id: 'v8a-metta-sabbaloka',
+          pali: 'mettañca sabbalokasmiṁ',
           scripts: [
-            {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Mettañca sabbalokasmiṁ, mānasaṁ bhāvaye aparimāṇaṁ; uddhaṁ adho ca tiriyañca, asambādhaṁ averaṁ asapattaṁ.',
-            },
+            { lang: 'pi-Latn', label: 'Pāli', text: 'mettañca sabbalokasmiṁ' },
           ],
           witnesses: [
+            { by: 'Amaravati', text: 'Radiating kindness over the entire world:', alignTo: [-1, 0, -1, -1, 1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'With love for the whole world,', alignTo: [-1, 0, -1, -1, 1, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'With good will for the entire cosmos,', alignTo: [-1, 0, 0, -1, -1, 1, 1], morphemeAlignTo: [null, 0, 0, null, null, null, null], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              by: 'Amaravati',
-              text: 'Radiating kindness over the entire world: spreading upwards to the skies, and downwards to the depths; outwards and unbounded, freed from hatred and ill-will.',
-              url: AMARAVATI_URL,
+              form: 'mettañca',
+              pronunciation: 'MAYT-tahn-chah',
+              etymology: '*mettā* (loving-kindness, friendliness) + *ca* "and" — the consonant clusters together for chant rhythm',
+              gloss: 'and loving-kindness. *Mettā* is the name-virtue of the sutta — not affection or warmth alone, but a steady goodwill toward every being without exception. Related to *mitta* "friend".',
+              accent: 'rose',
+              citations: [dpdCitation('mettā')],
+              morphemes: [
+                { text: 'mettañ', type: 'stem', gloss: '*mettā* "loving-kindness" — the sutta\'s name-virtue. Sound-merged with the next word *ca*.', pronunciation: 'MAYT-tahn' },
+                { text: 'ca', type: 'suffix', gloss: '"and" — same word as elsewhere; here joined to the previous word by chant rhythm', pronunciation: 'chah' },
+              ],
             },
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'With love for the whole world, they\'d cultivate a heart that\'s limitless, upwards, downwards and side-to-side, unbounded, free of enmity and hate.',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'sabbalokasmiṁ',
+              pronunciation: 'SAHB-bah-LOH-kah-smeem',
+              etymology: '*sabba* (all) + *loka* (world) + *-smiṁ* (the "in/at" ending)',
+              gloss: '"in the whole world" — the wide target. *Loka* is "world" in the cosmological sense, covering all realms of existence.',
+              morphemes: [
+                { text: 'sabba', type: 'stem', gloss: '"all, every" — same root as *sabbe* in verses 3 and 5', pronunciation: 'SAHB-bah' },
+                { text: 'loka', type: 'stem', gloss: '"world, realm" — covers all the realms of existence in Buddhist cosmology', pronunciation: 'LOH-kah' },
+                { text: 'smiṁ', type: 'suffix', gloss: 'the "-smiṁ" tail says "in/at" — *sabba-loka-smiṁ* = "in the whole world"', pronunciation: 'smeem' },
+              ],
             },
+          ],
+        },
+        // ── v8b: mānasaṁ bhāvaye aparimāṇaṁ (refrain — same as v7d) ──
+        {
+          id: 'v8b-manasam-bhavaye',
+          pali: 'mānasaṁ bhāvaye aparimāṇaṁ',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'mānasaṁ bhāvaye aparimāṇaṁ' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'spreading upwards to the skies,', alignTo: [-1, -1, -1, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: "they'd cultivate a heart that's limitless,", alignTo: [-1, 1, -1, 0, -1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'cultivate a limitless heart:', alignTo: [1, -1, 2, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'mānasaṁ', pronunciation: 'MAH-nah-sahm', gloss: 'the heart-mind — same word as in verse 7. The refrain returns: cultivate this thing, boundlessly.', accent: 'sky', citations: [dpdCitation('manas')] },
+            { form: 'bhāvaye', pronunciation: 'BAH-vah-yay', gloss: 'let one cultivate — same wishing form as in verse 7. The technical term for meditative cultivation.', accent: 'amber', citations: [dpdCitation('bhāveti')] },
+            { form: 'aparimāṇaṁ', pronunciation: 'ah-PAH-ree-MAH-nahm', gloss: 'boundless, without measure — same word as in verse 7. The line completes the refrain.', accent: 'rose' },
+          ],
+        },
+        // ── v8c: uddhaṁ adho ca tiriyañca ──
+        {
+          id: 'v8c-uddham-adho',
+          pali: 'uddhaṁ adho ca tiriyañca',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'uddhaṁ adho ca tiriyañca' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'and downwards to the depths;', alignTo: [2, 1, -1, -1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'upwards, downwards and side-to-side,', alignTo: [0, 1, 2, 3], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'above, below, & all around,', alignTo: [0, 1, 2, 3, 3], morphemeAlignTo: [null, null, null, 0, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'uddhaṁ', pronunciation: 'OOD-dahm', gloss: 'upward — the direction-word for "above". The radiating-out begins.' },
+            { form: 'adho', pronunciation: 'AH-doh', gloss: 'downward — paired with *uddhaṁ* "upward". Vertical sweep.' },
+            { form: 'ca', pronunciation: 'chah', gloss: 'and' },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'With good will for the entire cosmos, cultivate a limitless heart: above, below, & all around, unobstructed, without enmity or hate.',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'tiriyañca',
+              pronunciation: 'TEE-ree-yahn-chah',
+              etymology: '*tiriyaṁ* "across, horizontally" + *ca* "and" — sound-merged together',
+              gloss: '"and across" — completing the three-axis sweep: up + down + sideways. *Tiriya* is "transverse, horizontal" — the metta covers all directions.',
+              morphemes: [
+                { text: 'tiriyañ', type: 'stem', gloss: '"across, horizontally" — the third axis, completing up + down + side', pronunciation: 'TEE-ree-yahn' },
+                { text: 'ca', type: 'suffix', gloss: '"and" — joined to the previous word', pronunciation: 'chah' },
+              ],
             },
           ],
+        },
+        // ── v8d: asambādhaṁ averaṁ asapattaṁ ──
+        {
+          id: 'v8d-asambadham-averam',
+          pali: 'asambādhaṁ averaṁ asapattaṁ',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'asambādhaṁ averaṁ asapattaṁ' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'outwards and unbounded, freed from hatred and ill-will.', alignTo: [-1, -1, 0, -1, -1, 1, -1, 2], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'unbounded, free of enmity and hate.', alignTo: [0, 1, 1, 2, -1, 1], morphemeAlignTo: [null, 0, 0, null, null, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'unobstructed, without enmity or hate.', alignTo: [0, 1, 1, -1, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
           words: [
             {
-              form: 'Mettañca',
-              gloss: '*mettā* + *ca* "and". *Mettā* is loving-kindness, friendliness; not affection or warmth alone but a goodwill toward all beings without exception. The name-virtue of the sutta.',
-              accent: 'rose',
-              citations: [dpdCitation('mettā')],
+              form: 'asambādhaṁ',
+              pronunciation: 'ah-sahm-BAH-dahm',
+              etymology: '*a-* (not) + *sambādha* (crowded, narrow, constricted)',
+              gloss: '"uncrowded, unconstricted" — the metta-heart has no walls penning it in. *Sambādha* is "narrow", *asambādha* is "wide open".',
+              morphemes: [
+                { text: 'a', type: 'prefix', gloss: '"not"', pronunciation: 'ah' },
+                { text: 'sambādhaṁ', type: 'stem', gloss: '"narrow, constricted, crowded". *A-sambādhaṁ* = "without constriction, wide-open".', pronunciation: 'sahm-BAH-dahm' },
+              ],
+            },
+            {
+              form: 'averaṁ',
+              pronunciation: 'ah-VAY-rahm',
+              etymology: '*a-* (not) + *vera* (hatred, enmity)',
+              gloss: '"without hate". *Vera* is the technical term for enmity that perpetuates itself across lifetimes — the kind of grudge that won\'t let go. *A-vera* is the heart with no grudge.',
+              accent: 'sky',
+              morphemes: [
+                { text: 'a', type: 'prefix', gloss: '"not"', pronunciation: 'ah' },
+                { text: 'veraṁ', type: 'stem', gloss: '"hatred, enmity, grudge" — the technical Buddhist term for hate that perpetuates itself. *A-vera* = "without grudge".', pronunciation: 'VAY-rahm' },
+              ],
+            },
+            {
+              form: 'asapattaṁ',
+              pronunciation: 'ah-sah-PAHT-tahm',
+              etymology: '*a-* (not) + *sapatta* (enemy, rival)',
+              gloss: '"without enemy" — without anyone the heart sets itself against. *Sapatta* is "rival, adversary" — *asapatta* is the heart with no adversary.',
+              morphemes: [
+                { text: 'a', type: 'prefix', gloss: '"not"', pronunciation: 'ah' },
+                { text: 'sapattaṁ', type: 'stem', gloss: '"rival, enemy". *A-sapatta* = "with no rival, with no one to oppose".', pronunciation: 'sah-PAHT-tahm' },
+              ],
             },
           ],
         },
+        // ── v9a: tiṭṭhaṁ caraṁ nisinno vā ──
         {
-          id: 'verse-9',
-          pali: 'Tiṭṭhaṁ caraṁ nisinno vā, sayāno vā yāvat\'assa vigatamiddho; etaṁ satiṁ adhiṭṭheyya, brahmametaṁ vihāraṁ idhamāhu.',
+          id: 'v9a-tittham-caram',
+          pali: 'tiṭṭhaṁ caraṁ nisinno vā',
           scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'tiṭṭhaṁ caraṁ nisinno vā' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'Whether standing or walking, seated', alignTo: [-1, 0, 3, 1, 2], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'Standing, walking, sitting,', alignTo: [0, 1, 2], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'Whether standing, walking, sitting,', alignTo: [-1, 0, 1, 2], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Tiṭṭhaṁ caraṁ nisinno vā, sayāno vā yāvat\'assa vigatamiddho; etaṁ satiṁ adhiṭṭheyya, brahmametaṁ vihāraṁ idhamāhu.',
+              form: 'tiṭṭhaṁ',
+              pronunciation: 'TEET-tahm',
+              etymology: 'present-form of *√sthā* "to stand" — "standing, while-standing"',
+              gloss: 'standing — the first of the four postures. Same *√sthā* gives English speakers *stand* through Indo-European cousins.',
             },
+            {
+              form: 'caraṁ',
+              pronunciation: 'CHAH-rahm',
+              etymology: 'present-form of *√car* "to walk, go" — same root as *samācare* in verse 3',
+              gloss: 'walking — the second posture.',
+            },
+            {
+              form: 'nisinno',
+              pronunciation: 'nee-SEEN-noh',
+              etymology: '*ni-* (down) + past form of *√sad* "to sit" — "sat down"',
+              gloss: 'seated — the third posture. The *ni-* prefix adds "down".',
+              morphemes: [
+                { text: 'ni', type: 'prefix', gloss: '"down" — same prefix as in *nikubbetha* in verse 6', pronunciation: 'nee' },
+                { text: 'sinno', type: 'root', root: '√sad', gloss: 'from the verb "to sit". Same root that gives English *settle* and *sediment* through Indo-European cousins.', pronunciation: 'SEEN-noh' },
+              ],
+            },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
+          ],
+        },
+        // ── v9b: sayāno vā yāvat'assa vigatamiddho ──
+        {
+          id: 'v9b-sayano-vigatamiddho',
+          pali: "sayāno vā yāvat'assa vigatamiddho",
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: "sayāno vā yāvat'assa vigatamiddho", tokens: ['sayāno', 'vā', "yāvat'assa", 'vigatamiddho'] },
           ],
           witnesses: [
+            { by: 'Amaravati', text: 'or lying down, free from drowsiness,', alignTo: [1, 0, 0, -1, -1, 3], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: "or lying down— as long as they're not drowsy—", alignTo: [1, 0, 0, -1, 2, 2, -1, -1, 3], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'or lying down, as long as one is alert,', alignTo: [1, 0, 0, 2, 2, 2, -1, -1, 3], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'sayāno', pronunciation: 'sah-YAH-noh', etymology: 'from *√śī* "to lie down"', gloss: 'lying down — the fourth posture. Together with the three above, the verse covers every bodily state.' },
+            { form: 'vā', pronunciation: 'vah', gloss: 'or' },
+            { form: "yāvat'assa", pronunciation: 'YAH-vaht-AHS-sah', gloss: '"as long as [one] is" — a clause-opener. *Yāvat* "as long as" + *assa* "would be".' },
             {
-              by: 'Amaravati',
-              text: 'Whether standing or walking, seated or lying down, free from drowsiness, one should sustain this recollection. This is said to be the sublime abiding.',
-              url: AMARAVATI_URL,
+              form: 'vigatamiddho',
+              pronunciation: 'vee-GAH-tah-MEED-doh',
+              etymology: '*vi-* (away) + *gata* (gone) + *middha* (drowsiness, torpor)',
+              gloss: '"with drowsiness gone away" — alert, free of the sluggish-mind state. *Middha* is one of the five hindrances (drowsiness) in Buddhist psychology.',
+              morphemes: [
+                { text: 'vigata', type: 'prefix', gloss: '"gone away, departed" — *vi-* "away" + *gata* "gone"', pronunciation: 'vee-GAH-tah' },
+                { text: 'middho', type: 'stem', gloss: '"drowsiness, torpor" — one of the five hindrances. *Vi-gata-middho* = "the one whose drowsiness has departed".', pronunciation: 'MEED-doh' },
+              ],
             },
+          ],
+        },
+        // ── v9c: etaṁ satiṁ adhiṭṭheyya ──
+        {
+          id: 'v9c-etam-satim',
+          pali: 'etaṁ satiṁ adhiṭṭheyya',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'etaṁ satiṁ adhiṭṭheyya' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'one should sustain this recollection.', alignTo: [-1, -1, 2, 0, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'they would commit to this kind of mindfulness;', alignTo: [-1, 2, 2, -1, 0, -1, -1, 1], morphemeAlignTo: [null, 1, 1, null, null, null, null, null], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'one should be resolved on this mindfulness.', alignTo: [-1, 2, -1, 2, -1, 0, 1], morphemeAlignTo: [null, 1, null, 1, null, null, null], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'etaṁ', pronunciation: 'AY-tahm', gloss: 'this (object form) — refers back to the cultivated heart-mind from verse 8.' },
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'Standing, walking, sitting, or lying down— as long as they\'re not drowsy— they would commit to this kind of mindfulness; this is what they call a divine meditation in this life.',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'satiṁ',
+              pronunciation: 'SAH-teem',
+              etymology: '*sati* (mindfulness, remembering) + *-ṁ* (object form). From *√smṛ* "to remember".',
+              gloss: 'mindfulness, recollection (object). The central Buddhist faculty — keeping something in present awareness. Same root as Skt *smṛti*.',
+              accent: 'sky',
             },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'Whether standing, walking, sitting, or lying down, as long as one is alert, one should be resolved on this mindfulness. This is called a sublime abiding here & now.',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'adhiṭṭheyya',
+              pronunciation: 'ah-DEET-tay-yah',
+              etymology: '*adhi-* (over, upon) + *√sthā* (to stand) — "to stand upon, take a firm stand on, resolve". The "-eyya" tail says "should/would".',
+              gloss: '"should be resolved on, should stand-firmly-on". *Adhiṭṭhāna* (resolve, determination) is one of the ten *pāramī* (perfections).',
+              accent: 'amber',
+              morphemes: [
+                { text: 'adhi', type: 'prefix', gloss: '"over, upon" — adds a "standing-upon" flavour to the verb', pronunciation: 'AH-dee' },
+                { text: 'ṭṭheyya', type: 'root', root: '√sthā', gloss: 'from the verb "to stand". *Adhi-ṭṭha* = "stand-upon, resolve". Same root as *tiṭṭhaṁ* (standing) in 9a.', pronunciation: 'TAY-yah' },
+              ],
             },
           ],
+        },
+        // ── v9d: brahmametaṁ vihāraṁ idhamāhu ──
+        {
+          id: 'v9d-brahma-vihara',
+          pali: 'brahmametaṁ vihāraṁ idhamāhu',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'brahmametaṁ vihāraṁ idhamāhu' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'This is said to be the sublime abiding.', alignTo: [0, -1, 2, -1, -1, -1, 0, 1], morphemeAlignTo: [1, null, null, null, null, null, 0, null], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'this is what they call a divine meditation in this life.', alignTo: [0, -1, -1, 2, 2, -1, 0, 1, 2, -1, -1], morphemeAlignTo: [1, null, null, 1, 1, null, 0, null, 0, null, null], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'This is called a sublime abiding here & now.', alignTo: [0, -1, 2, -1, 0, 1, 2, -1, 2], morphemeAlignTo: [1, null, 1, null, 0, null, 0, null, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
           words: [
             {
               form: 'brahmametaṁ',
-              gloss: '*brahma* + *etaṁ* "this is divine". The *brahmavihāra* — divine abiding — is the meditator\'s dwelling-place. *Brahma* here doesn\'t name the deity but the highest, sublime register.',
+              pronunciation: 'BRAH-mah-MAY-tahm',
+              etymology: '*brahma* (divine, supreme) + *etaṁ* "this" — sound-merged together',
+              gloss: '"this divine [one]". The *brahmavihāra* — divine abiding — names the meditator\'s dwelling-place. *Brahma* here doesn\'t name the deity but the highest, sublime register.',
               accent: 'sky',
+              morphemes: [
+                { text: 'brahma', type: 'stem', gloss: '"divine, supreme, highest" — same word in Skt and Pāli. *Brahmavihāra* is the technical term for the four sublime states (loving-kindness, compassion, joy, equanimity).', pronunciation: 'BRAH-mah' },
+                { text: 'metaṁ', type: 'stem', gloss: '*etaṁ* "this" — joined to *brahma* by chant rhythm. The whole word means "this divine-state".', pronunciation: 'MAY-tahm' },
+              ],
+            },
+            {
+              form: 'vihāraṁ',
+              pronunciation: 'vee-HAH-rahm',
+              etymology: '*vi-* (apart) + *√hṛ* "to carry, dwell" — "dwelling, abiding"',
+              gloss: 'an abiding, a dwelling-place. *Vihāra* gives the word for a monastery (the place one dwells), and in compounds names the four sublime abidings (*brahmavihāra*).',
+            },
+            {
+              form: 'idhamāhu',
+              pronunciation: 'EE-dah-MAH-hoo',
+              etymology: '*idha* "here" + *āhu* "they say, it is said"',
+              gloss: '"here they call [it]" — closes the verse with the traditional formula *iti āhu* "thus they say". The community\'s naming of the state.',
+              morphemes: [
+                { text: 'idha', type: 'stem', gloss: '"here, in this place" — anchors the divine abiding in *this* life, not a future heaven', pronunciation: 'EE-dah' },
+                { text: 'māhu', type: 'stem', gloss: '*āhu* "they say, it is said" — the traditional Pāli formula that closes a teaching with "thus they say"', pronunciation: 'MAH-hoo' },
+              ],
             },
           ],
         },
+        // ── v10a: diṭṭhiñca anupaggamma ──
         {
-          id: 'verse-10',
-          pali: 'Diṭṭhiñca anupaggamma, sīlavā dassanena sampanno; kāmesu vineyya gedhaṁ, na hi jātuggabbhaseyya puna retīti.',
+          id: 'v10a-ditthi-anupaggamma',
+          pali: 'diṭṭhiñca anupaggamma',
           scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'diṭṭhiñca anupaggamma' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'By not holding to fixed views,', alignTo: [-1, 1, 1, -1, 1, 0], morphemeAlignTo: [null, 0, 2, null, 2, null], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'Avoiding wrong views,', alignTo: [1, 1, 0], morphemeAlignTo: [0, 2, null], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'Not taken with views,', alignTo: [1, 1, -1, 0], morphemeAlignTo: [0, 2, null, null], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              lang: 'pi-Latn',
-              label: 'Pāli',
-              text: 'Diṭṭhiñca anupaggamma, sīlavā dassanena sampanno; kāmesu vineyya gedhaṁ, na hi jātuggabbhaseyya puna retīti.',
+              form: 'diṭṭhiñca',
+              pronunciation: 'DEET-tee-nyah-chah',
+              etymology: '*diṭṭhi* (view) + *ca* "and" — sound-merged together. From *√dṛś* "to see".',
+              gloss: 'and views. *Diṭṭhi* is "view" — most often *wrong* view in technical Pāli: a fixed position, a doctrine clung to. Letting go of clinging to views is the closing instruction.',
+              accent: 'amber',
+              citations: [dpdCitation('diṭṭhi')],
+              morphemes: [
+                { text: 'diṭṭhiñ', type: 'stem', root: '√dṛś', gloss: '"view, position" — same root as *diṭṭhā* (seen) in verse 5. Here it means a held position, a doctrine.', pronunciation: 'DEET-tee-nyah' },
+                { text: 'ca', type: 'suffix', gloss: '"and" — joined to the previous word', pronunciation: 'chah' },
+              ],
+            },
+            {
+              form: 'anupaggamma',
+              pronunciation: 'ah-noo-PAH-gahm-mah',
+              etymology: '*an-* (not) + *upa-* (near, up to) + *√gam* (to go) — "not going near, not approaching"',
+              gloss: '"not falling into, not going near". The wise person doesn\'t collapse into the views they encounter. Same *√gam* as English *come* through Indo-European cousins.',
+              morphemes: [
+                { text: 'an', type: 'prefix', gloss: '"not"', pronunciation: 'ahn' },
+                { text: 'upa', type: 'prefix', gloss: '"up to, near" — adds an approaching-flavour to the verb', pronunciation: 'OO-pah' },
+                { text: 'ggamma', type: 'root', root: '√gam', gloss: 'from the verb "to go" (the *g* doubles at the prefix-join, a common Pāli sound-strengthening). *Upa-gamma* = "approach, fall into". *An-upa-gamma* = "not approaching, not falling into".', pronunciation: 'GAHM-mah' },
+              ],
             },
           ],
+        },
+        // ── v10b: sīlavā dassanena sampanno ──
+        {
+          id: 'v10b-silava-dassanena',
+          pali: 'sīlavā dassanena sampanno',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'sīlavā dassanena sampanno' },
+          ],
           witnesses: [
+            { by: 'Amaravati', text: 'the pure-hearted one, having clarity of vision,', alignTo: [-1, 0, 0, -1, 2, -1, 1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'ethical, attaining vision,', alignTo: [0, 2, 1], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'but virtuous & consummate in vision,', alignTo: [-1, 0, -1, 2, -1, 1], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
             {
-              by: 'Amaravati',
-              text: 'By not holding to fixed views, the pure-hearted one, having clarity of vision, being freed from all sense desires, is not born again into this world.',
-              url: AMARAVATI_URL,
+              form: 'sīlavā',
+              pronunciation: 'SEE-lah-vah',
+              etymology: '*sīla* (virtue, ethical conduct) + *-vā* (possessive — "having")',
+              gloss: '"virtuous, ethical". *Sīla* is the technical term for the moral training of Buddhist practice — the precepts. *Sīla-vā* = "the one who has *sīla*".',
+              accent: 'sky',
+              morphemes: [
+                { text: 'sīla', type: 'stem', gloss: '"virtue, ethical conduct" — the technical term for the moral training. The Five Precepts and Eight Precepts are forms of *sīla*.', pronunciation: 'SEE-lah' },
+                { text: 'vā', type: 'suffix', gloss: 'the "-vā" tail means "having" — *sīla-vā* = "one who has virtue, the virtuous one"', pronunciation: 'vah' },
+              ],
             },
             {
-              by: 'Sujato (SuttaCentral)',
-              text: 'Avoiding wrong views, ethical, attaining vision, having removed desire for sensual pleasures, they would never come to lie in a womb again.',
-              url: 'https://suttacentral.net/snp1.8/en/sujato',
-              license: 'CC0',
+              form: 'dassanena',
+              pronunciation: 'DAHS-sah-NAY-nah',
+              etymology: '*dassana* (vision, sight, view) + *-ena* (the "by means of" ending)',
+              gloss: '"with vision, by means of seeing". *Dassana* is from the same *√dṛś* root as *diṭṭhi* in 10a — but where *diṭṭhi* is a *held* view, *dassana* is direct seeing.',
+              morphemes: [
+                { text: 'dassan', type: 'root', root: '√dṛś', gloss: 'from "to see" — same root as *diṭṭhi*. *Dassana* = direct seeing, vision (as opposed to *diṭṭhi* = held position).', pronunciation: 'DAHS-sahn' },
+                { text: 'ena', type: 'suffix', gloss: 'the "-ena" tail says "by means of" — same suffix as *kusalena* in verse 1', pronunciation: 'AY-nah' },
+              ],
             },
             {
-              by: 'Thanissaro (Access to Insight)',
-              text: 'Not taken with views, but virtuous & consummate in vision, having subdued desire for sensual pleasures, one never again will lie in the womb.',
-              url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html',
-              license: 'CC BY-NC',
+              form: 'sampanno',
+              pronunciation: 'sahm-PAHN-noh',
+              etymology: '*saṁ-* (fully) + past form of *√pad* "to attain" — "fully attained, accomplished"',
+              gloss: '"perfected in, fully attained". *Sampanno* with the previous word: "perfected in vision".',
             },
           ],
+        },
+        // ── v10c: kāmesu vineyya gedhaṁ ──
+        {
+          id: 'v10c-kamesu-vineyya',
+          pali: 'kāmesu vineyya gedhaṁ',
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'kāmesu vineyya gedhaṁ' },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'being freed from all sense desires,', alignTo: [-1, 1, -1, -1, 0, 0], morphemeAlignTo: [null, null, null, null, 0, 0], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'having removed desire for sensual pleasures,', alignTo: [-1, 1, 2, -1, 0, 0], morphemeAlignTo: [null, null, null, null, 0, 0], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'having subdued desire for sensual pleasures,', alignTo: [-1, 1, 2, -1, 0, 0], morphemeAlignTo: [null, null, null, null, 0, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
           words: [
             {
-              form: 'Diṭṭhiñca',
-              gloss: '*diṭṭhi* + *ca*. *Diṭṭhi* is "view" — most often *wrong* view in technical Pāli: a fixed position, a doctrine clung to. Letting go of clinging to views is the closing instruction.',
+              form: 'kāmesu',
+              pronunciation: 'KAH-may-soo',
+              etymology: '*kāma* (sensual pleasure, desire) + *-esu* (plural "in/among" ending)',
+              gloss: '"in/toward sensual pleasures". *Kāma* covers all sense-pleasure attachment — same word in *Kāma Sūtra* (which is actually a manual on the *first* of three life-aims, *kāma*).',
+              morphemes: [
+                { text: 'kām', type: 'stem', gloss: '"sense-pleasure, desire" — broader than just sex; covers all attachment to sense-objects', pronunciation: 'kahm' },
+                { text: 'esu', type: 'suffix', gloss: 'the "-esu" tail says "in/toward" with a plural — same suffix as *kulesu* in verse 2', pronunciation: 'AY-soo' },
+              ],
+            },
+            {
+              form: 'vineyya',
+              pronunciation: 'vee-NAYY-yah',
+              etymology: '*vi-* (apart, away) + *√nī* "to lead" — "to lead away, dispel"',
+              gloss: '"having led-away, having dispelled". The "-eyya" tail makes it the wishing voice — "would lead away".',
+              morphemes: [
+                { text: 'vi', type: 'prefix', gloss: '"apart, away" — adds a "leading-away" flavour', pronunciation: 'vee' },
+                { text: 'neyya', type: 'root', root: '√nī', gloss: 'from the verb "to lead". *Vi-nī* = "lead away, dispel".', pronunciation: 'NAYY-yah' },
+              ],
+            },
+            {
+              form: 'gedhaṁ',
+              pronunciation: 'GAY-dahm',
+              etymology: '*gedha* (greed, attachment — same root *√gṛdh* as *ananugiddho* in verse 2) + the *-ṁ* tail (object form)',
+              gloss: '"greed, attachment" (the object dispelled). Same root as the *ananugiddho* "not chasing greedily" of verse 2 — the closing verse returns to the opening virtues.',
               accent: 'amber',
-              citations: [dpdCitation('diṭṭhi')],
+            },
+          ],
+        },
+        // ── v10d: na hi jātuggabbhaseyya puna retīti ──
+        {
+          id: 'v10d-na-jatu',
+          pali: "na hi jātu g(g)abbhaseyya puna retīti",
+          scripts: [
+            { lang: 'pi-Latn', label: 'Pāli', text: 'na hi jātu gabbhaseyya puna retīti', tokens: ['na', 'hi', 'jātu', 'gabbhaseyya', 'puna', 'retīti'] },
+          ],
+          witnesses: [
+            { by: 'Amaravati', text: 'is not born again into this world.', alignTo: [-1, 0, -1, 4, -1, -1, -1], url: AMARAVATI_URL },
+            { by: 'Sujato (SuttaCentral)', text: 'they would never come to lie in a womb again.', alignTo: [-1, 0, 2, -1, -1, 3, -1, -1, 3, 4], morphemeAlignTo: [null, null, null, null, null, 1, null, null, 0, null], url: 'https://suttacentral.net/snp1.8/en/sujato', license: 'CC0' },
+            { by: 'Thanissaro (Access to Insight)', text: 'one never again will lie in the womb.', alignTo: [-1, 0, 4, -1, 3, -1, -1, 3], morphemeAlignTo: [null, null, null, null, 1, null, null, 0], url: 'https://www.accesstoinsight.org/tipitaka/kn/snp/snp.1.08.than.html', license: 'CC BY-NC' },
+          ],
+          words: [
+            { form: 'na', pronunciation: 'nah', gloss: 'not — together with *hi* below: "not at all, certainly not".' },
+            { form: 'hi', pronunciation: 'hee', gloss: 'an emphatic particle: "indeed, certainly". *Na hi* = "certainly not".' },
+            { form: 'jātu', pronunciation: 'JAH-too', gloss: '"ever, at any time". Pairs with the negation: "not ever, never again".' },
+            {
+              form: 'gabbhaseyya',
+              pronunciation: 'GAHB-bah-SAY-yah',
+              etymology: '*gabbha* (womb) + *seyya* (lying-down — from *√śī* "to lie")',
+              gloss: '"lying in a womb" — the technical phrase for rebirth in mammalian form. Same *√śī* as *sayāno* "lying down" in verse 9 — but here the lying-down is the womb-state.',
+              morphemes: [
+                { text: 'gabbha', type: 'stem', gloss: '"womb, inner chamber". Same word as *gabbho* in verse 2 but in a different sense (rude/intrusive there, womb here).', pronunciation: 'GAHB-bah' },
+                { text: 'seyya', type: 'root', root: '√śī', gloss: 'from "to lie down" — same root as *sayāno* in verse 9', pronunciation: 'SAY-yah' },
+              ],
+            },
+            { form: 'puna', pronunciation: 'POO-nah', gloss: '"again". *Punarbhava* (re-becoming) is the technical term for rebirth — here the line says *na puna* = "not again".' },
+            {
+              form: 'retīti',
+              pronunciation: 'RAY-tee-tee',
+              etymology: '*eti* (one goes, comes — from *√i*) + *iti* (the quotation closure: "thus")',
+              gloss: '"comes [to it], thus [it is said]". The closing *iti* marks the end of the discourse — like English closing quotation marks. The whole verse-end *eti iti* = "[one] comes [there], thus [is said]".',
+              morphemes: [
+                { text: 'retī', type: 'root', root: '√i', gloss: 'from the verb "to come, go" — sandhi-form of *eti*', pronunciation: 'RAY-tee' },
+                { text: 'ti', type: 'suffix', gloss: 'the closing *iti* = "thus", marking the end of the recited text. Like English closing quote-marks.', pronunciation: 'tee' },
+              ],
             },
           ],
         },
diff --git a/data/liturgy/morning-chants.ts b/data/liturgy/morning-chants.ts
index 456ef4c..d17bfb1 100644
--- a/data/liturgy/morning-chants.ts
+++ b/data/liturgy/morning-chants.ts
@@ -105,7 +105,7 @@ const REFUGE_WORDS: WordGloss[] = [
   {
     form: 'buddhaṁ', scriptAlt: 'बुद्धं', pronunciation: 'BOO-dhang', root: '√budh',
     accent: 'sky',
-    etymology: '√budh "to wake" → *buddha* + accusative ending',
+    etymology: '√budh "to wake" → *buddha*, with the ending that marks it the thing one goes to',
     gloss: 'to the Buddha (the one I am going toward)',
     citations: [dpdCitation('buddha'), pronCite],
     morphemes: [
@@ -117,7 +117,7 @@ const REFUGE_WORDS: WordGloss[] = [
   {
     form: 'dhammaṁ', scriptAlt: 'धम्मं', pronunciation: 'DHUM-mang', root: '√dhṛ',
     accent: 'amber',
-    etymology: '√dhṛ "to hold" → *dhamma* + accusative ending',
+    etymology: '√dhṛ "to hold" → *dhamma*, with the ending that marks it the thing one goes to',
     gloss: 'to the Dhamma — the teaching, truth, the way',
     citations: [dpdCitation('dhamma'), pronCite],
     morphemes: [
@@ -141,7 +141,7 @@ const REFUGE_WORDS: WordGloss[] = [
   },
   {
     form: 'saraṇaṁ', scriptAlt: 'सरणं', pronunciation: 'SAH-rah-nang', root: '√śri',
-    etymology: '√śri "to lean on" → *saraṇa* + accusative',
+    etymology: '√śri "to lean on" → *saraṇa*, with the ending that marks it the destination',
     gloss: 'refuge, shelter, protection',
     citations: [dpdCitation('saraṇa'), pronCite],
     morphemes: [
@@ -508,13 +508,13 @@ export const morningChants: LiturgyDoc = {
               citations: [dpdCitation('adinnādāna'), pronCite],
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not (negation)', pronunciation: 'ah' },
-                { text: 'dinnā', type: 'stem', gloss: 'given (past participle of "to give")', pronunciation: 'DEEN-nah' },
+                { text: 'dinnā', type: 'stem', gloss: 'given — the "X-ed" form of the verb "to give"', pronunciation: 'DEEN-nah' },
                 { text: 'dānā', type: 'stem', gloss: 'taking, grasping', pronunciation: 'DAH-nah', citations: [dpdCitation('ādāna')] },
               ],
               scriptMorphemes: {
                 'pi-Deva': [
                   { text: 'अ', type: 'prefix', pronunciation: 'ah', gloss: 'not, negation' },
-                  { text: 'दिन्ना', type: 'stem', pronunciation: 'DEEN-nah', gloss: 'given (past participle of "to give")' },
+                  { text: 'दिन्ना', type: 'stem', pronunciation: 'DEEN-nah', gloss: 'given — the "X-ed" form of the verb "to give"' },
                   { text: 'दाना', type: 'stem', pronunciation: 'DAH-nah', gloss: 'taking, grasping' },
                 ],
               },
@@ -543,7 +543,7 @@ export const morningChants: LiturgyDoc = {
               gloss: 'in sensual pleasures, in sexual matters',
               citations: [dpdCitation('kāma'), pronCite],
               morphemes: [
-                { text: 'kāma', type: 'stem', gloss: 'sensual pleasure, sense-desire — broader than sex; includes pleasures of taste, sound, sight, touch', pronunciation: 'KAH-mah', citations: [dpdCitation('kāma')] },
+                { text: 'kām', type: 'stem', gloss: 'sensual pleasure, sense-desire — broader than sex; includes pleasures of taste, sound, sight, touch (the stem *kāma*; its final vowel merges into the ending)', pronunciation: 'KAH-mah', citations: [dpdCitation('kāma')] },
                 { text: 'esu', type: 'suffix', gloss: '"in [these pleasures]" — locates the conduct', pronunciation: 'AY-soo' },
               ],
             },
@@ -771,7 +771,7 @@ export const morningChants: LiturgyDoc = {
           words: [
             {
               form: 'etaṁ', scriptAlt: 'एतं', pronunciation: 'AY-tang',
-              etymology: '*eta-* "this" + accusative ending',
+              etymology: '*eta-* "this" + the ending that marks it the object of the verb',
               gloss: 'this',
               citations: [dpdCitation('etaṁ'), pronCite],
               morphemes: [
diff --git a/data/liturgy/om-mani-padme-hum.ts b/data/liturgy/om-mani-padme-hum.ts
index af664c1..3b9be8b 100644
--- a/data/liturgy/om-mani-padme-hum.ts
+++ b/data/liturgy/om-mani-padme-hum.ts
@@ -175,8 +175,8 @@ export const omManiPadmeHum: LiturgyDoc = {
                 ],
               },
               pronunciation: 'PUHD-meh',
-              etymology: '*padma* "lotus" + locative or vocative ending *-e*',
-              gloss: 'in the lotus (locative) — or "O Lotus One" (vocative) — represents *wisdom* ([[prajñā]])',
+              etymology: '*padma* "lotus" + the *-e* ending — which can mean either "in the …" or "O …!" (addressing)',
+              gloss: '"in the lotus" — or, read as a call, "O Lotus One" — represents *wisdom* ([[prajñā]])',
               accent: 'sky',
               morphemes: [
                 {
@@ -190,7 +190,7 @@ export const omManiPadmeHum: LiturgyDoc = {
                 {
                   text: 'e',
                   type: 'suffix',
-                  gloss: 'either "in the [lotus]" (locative) or addressing "O [Lotus One]" (vocative) — Sanskrit grammars allow both readings',
+                  gloss: 'the *-e* ending — either "in the [lotus]" or, read as a call, "O [Lotus One]!". Sanskrit grammars allow both readings',
                   pronunciation: 'eh',
                   citations: [ungroundedCitation('Sanskrit case morphology — needs grammar reference')],
                 },
diff --git a/data/liturgy/sho-sai-myo-kichijo-darani.ts b/data/liturgy/sho-sai-myo-kichijo-darani.ts
index 8b3245c..83ec1c2 100644
--- a/data/liturgy/sho-sai-myo-kichijo-darani.ts
+++ b/data/liturgy/sho-sai-myo-kichijo-darani.ts
@@ -116,12 +116,12 @@ export const shoSaiMyoKichijoDarani: LiturgyDoc = {
                 'zh-Hant': [
                   { text: '母', type: 'phonetic', pronunciation: 'mǔ', gloss: 'phonetic: transliterates *bu*/*mu* (semantic "mother" unrelated)' },
                   { text: '䭾', type: 'phonetic', pronunciation: 'tuó', gloss: 'phonetic: transliterates *dhā*' },
-                  { text: '喃', type: 'phonetic', pronunciation: 'nán', gloss: 'phonetic: transliterates *nāṃ* (genitive plural ending)' },
+                  { text: '喃', type: 'phonetic', pronunciation: 'nán', gloss: 'phonetic: transliterates *nāṃ*, the "of-those" ending' },
                 ],
               },
               pronunciation: 'bood-DHAH-nahm',
-              etymology: '*buddha* "awakened one" + *-ānāṃ* (genitive plural)',
-              gloss: 'of the buddhas (genitive plural)',
+              etymology: '*buddha* "awakened one" + *-ānāṃ*, the "of-those" ending',
+              gloss: 'of the buddhas — the ending carries "of-those", a plural possessive',
             },
           ],
         },
@@ -171,12 +171,12 @@ export const shoSaiMyoKichijoDarani: LiturgyDoc = {
                 ],
               },
               pronunciation: 'ah-prah-tee-HAH-tah',
-              etymology: '*a-* "not" + *prati-* "against" + *hata* "struck" (past participle of √han)',
+              etymology: '*a-* "not" + *prati-* "against" + *hata* "struck" — the "X-ed" form of √han "to strike"',
               gloss: 'unimpeded, unobstructed — literally "not struck against"',
               morphemes: [
                 { text: 'a', type: 'prefix', gloss: 'not', pronunciation: 'ah' },
                 { text: 'prati', type: 'prefix', gloss: 'against, toward', pronunciation: 'prah-tee' },
-                { text: 'hata', type: 'stem', root: '√han', gloss: 'struck (past participle)', pronunciation: 'HAH-tah' },
+                { text: 'hata', type: 'stem', root: '√han', gloss: 'struck — the "X-ed" form of √han "to strike"', pronunciation: 'HAH-tah' },
               ],
             },
             {
@@ -190,16 +190,16 @@ export const shoSaiMyoKichijoDarani: LiturgyDoc = {
                   { text: '舍', type: 'phonetic', pronunciation: 'shě', gloss: 'phonetic: transliterates *-na-*' },
                   { text: '娑', type: 'phonetic', pronunciation: 'suō', gloss: 'phonetic: transliterates *(śā)sa-*' },
                   { text: '曩', type: 'phonetic', pronunciation: 'nǎng', gloss: 'phonetic: transliterates *-nā-*' },
-                  { text: '喃', type: 'phonetic', pronunciation: 'nán', gloss: 'phonetic: transliterates *-nām* (genitive plural)' },
+                  { text: '喃', type: 'phonetic', pronunciation: 'nán', gloss: 'phonetic: transliterates *-nām*, the "of-those" ending' },
                 ],
               },
               pronunciation: 'shah-sah-NAH-nahm',
-              etymology: '*śās* "to teach, instruct" → *śāsana* "teaching" + *-ānām* (genitive plural)',
-              gloss: 'of the teachings (genitive plural) — together: "of-those-with-unimpeded-teachings"',
+              etymology: '*śās* "to teach, instruct" → *śāsana* "teaching" + *-ānām*, the "of-those" ending',
+              gloss: 'of the teachings — together: "of-those-with-unimpeded-teachings"',
               morphemes: [
                 { text: 'śās', type: 'root', root: '√śās', gloss: 'to teach, instruct', pronunciation: 'shahs' },
-                { text: 'ana', type: 'suffix', gloss: 'noun-forming — yields *śāsana* "teaching"', pronunciation: 'ah-nah' },
-                { text: 'ānām', type: 'suffix', gloss: 'genitive plural ending — "of those …"', pronunciation: 'AH-nahm' },
+                { text: 'an', type: 'suffix', gloss: 'noun-forming — yields *śāsana* "teaching" (the suffix is *-ana*; its final vowel merges into the ending that follows)', pronunciation: 'ah-nah' },
+                { text: 'ānām', type: 'suffix', gloss: 'the "-ānām" tail marks "of those …" — a possessive plural', pronunciation: 'AH-nahm' },
               ],
             },
           ],
diff --git a/data/liturgy/ti-sarana.ts b/data/liturgy/ti-sarana.ts
index 4dc12bb..098a01f 100644
--- a/data/liturgy/ti-sarana.ts
+++ b/data/liturgy/ti-sarana.ts
@@ -27,8 +27,8 @@ const pronCite = suttaCentralPronunciationCitation();
 const REFUGE_OBJECT_WORDS: { [key: string]: WordGloss } = {
   buddhaṁ: {
     form: 'Buddhaṁ', scriptAlt: 'बुद्धं', pronunciation: 'BUD-dahng', accent: 'amber',
-    etymology: '√budh "to wake up" + past participle',
-    gloss: 'the Awakened One — accusative (object of "I go to")',
+    etymology: '√budh "to wake up" — *buddha* is the "awakened" form of that verb',
+    gloss: 'the Awakened One — here it is what one goes to: the object of "I go to"',
     citations: [dpdCitation('buddha'), pronCite],
   },
   dhammaṁ: {
@@ -48,7 +48,7 @@ const REFUGE_OBJECT_WORDS: { [key: string]: WordGloss } = {
 const SARANAM: WordGloss = {
   form: 'Saraṇaṁ', scriptAlt: 'सरणं', pronunciation: 'SAH-rah-nahng',
   etymology: '√śri "to take shelter, lean on"',
-  gloss: 'refuge, shelter — accusative (the destination)',
+  gloss: 'refuge, shelter — the destination one goes to',
   citations: [dpdCitation('saraṇa'), pronCite],
 };
 
diff --git a/data/liturgy/way-of-compassion.ts b/data/liturgy/way-of-compassion.ts
index 56efd65..06f3d15 100644
--- a/data/liturgy/way-of-compassion.ts
+++ b/data/liturgy/way-of-compassion.ts
@@ -159,8 +159,8 @@ export const wayOfCompassion: LiturgyDoc = {
             { by: 'Thanissaro (Access to Insight)', text: 'As a mother would risk her life to protect her child, her only child, even so should one cultivate the heart limitlessly with regard to all beings.', url: THANISSARO_URL, license: 'CC BY-NC' },
           ],
           words: [
-            { form: 'mānasaṁ', gloss: '"The heart-mind". *manas* + accusative. The thing to be cultivated boundlessly.', accent: 'sky', citations: [dpdCitation('manas')] },
-            { form: 'bhāvaye', gloss: '"Let one cultivate, let one bring into being". Optative of *√bhū* "to be", causative — to *cause to become*. The technical term for meditative cultivation.', accent: 'amber', citations: [dpdCitation('bhāveti')] },
+            { form: 'mānasaṁ', gloss: '"The heart-mind" — *manas* with the ending that marks it the thing acted on. What is to be cultivated boundlessly.', accent: 'sky', citations: [dpdCitation('manas')] },
+            { form: 'bhāvaye', gloss: '"Let one cultivate, let one bring into being" — the wishing form of *√bhū* "to be", in its make-it-happen sense: to *cause to become*. The technical term for meditative cultivation.', accent: 'amber', citations: [dpdCitation('bhāveti')] },
           ],
         },
         {
diff --git a/docs/sutta-studio/DATA_FAILURE_MODES.md b/docs/sutta-studio/DATA_FAILURE_MODES.md
new file mode 100644
index 0000000..a4bacca
--- /dev/null
+++ b/docs/sutta-studio/DATA_FAILURE_MODES.md
@@ -0,0 +1,216 @@
+# Liturgy / Sutta data — failure modes
+
+> **Status:** Drafted 2026-05-20 from the Metta Sutta depth pass + the
+> cross-chant QC sweep that followed it.
+> **Companions:** `CURATION_PROTOCOL.md` (how a phase gets curated),
+> `tests/components/liturgy/liturgy-data-quality.test.ts` (the machine
+> guards described below).
+
+This catalogues the classes of error found while hand-curating the
+liturgy chant data, **so a future generation pipeline (LLM curation,
+where no human reviews every word) can avoid producing them — or
+detect them automatically when it does.**
+
+The organising distinction is **how the error announces itself**:
+
+- **Silent** — nothing crashes, no test fails, the page renders. The
+  reader just gets a degraded artifact. These are the dangerous ones:
+  without a guard they survive indefinitely.
+- **Loud** — a test fails or the build breaks. Cheap to catch.
+- **Judgment** — no mechanical check is decisive; needs a taste call.
+  A pipeline can *flag* these for review but not auto-resolve them.
+
+Every silent class below now has a guard (see the test file). The
+goal for any generator: **emit data that passes all guards, and when
+it cannot, degrade cleanly rather than emit broken data.**
+
+---
+
+## 1. Morpheme reconstruction failure — *silent*
+
+**What.** A word carries a `morphemes[]` array whose `text` values,
+concatenated, do not equal the surface `form`. The renderer's
+`splitByMorphemes` walks the surface slicing one morpheme at a time;
+if the pieces don't reconstruct it returns `null` and the word
+**silently** falls back to a single whole-word hover. The per-morpheme
+tooltips and arrows just vanish — no error, no warning.
+
+**Why it happens.** Pāli/Sanskrit *sandhi* at the morpheme join: two
+vowels merge (`na`+`ati` → `nāti`, `a`+`ā` → `ā`), a vowel shortens
+before the next piece, or `i`→`y` before a vowel. A curator writes the
+morphemes in their *citation* form (`na`, `ati`) instead of their
+*surface* form (`nā`, `ti`).
+
+**Found in.** 13 words across Metta Sutta, morning-chants, Sho Sai,
+Heart Sutra, Bodhi Heart Sutra.
+
+**Guard.** `morphemes[]` (and per-script `scriptMorphemes[]`) must
+concatenate back to the surface form. Test enforces it across every
+word in every chant.
+
+**Pipeline rule.** A generator emitting morphemes **must self-check
+reconstruction before emitting**. If the morphemes it produced don't
+reconstruct the surface, it must either (a) fix the split to the
+surface form, naming the sandhi in the gloss, or (b) emit *no*
+`morphemes` at all — whole-word hover is a clean degradation; a
+broken split is not. Never emit morphemes you haven't reconstruction-
+checked.
+
+---
+
+## 2. Internal-ID leak into reader text — *silent*
+
+**What.** A gloss or etymology string contains a curator's internal
+identifier — segment-ID shorthand like `v7d`, `v1a` ("same word as
+v7d"). Meaningless to a reader; it only ever made sense to the person
+authoring the file.
+
+**Why it happens.** Cross-reference glosses ("same root as X in …")
+are written while the curator is thinking in segment IDs.
+
+**Found in.** 26 glosses in the Metta Sutta.
+
+**Guard.** Reader-facing text must not match `/\bv\d+[a-z]\b/`.
+
+**Pipeline rule.** The generator should never be *given* internal IDs
+in a context where they can leak into prose — or a post-filter strips
+them. Cross-references in reader text must resolve to human-facing
+labels ("verse 7"), never internal handles. This generalises: **no
+internal identifier of any kind belongs in reader-facing strings.**
+
+---
+
+## 3. Alignment-array shape errors — *loud*
+
+**What.** `alignTo[]` length ≠ the witness's whitespace-split English
+word count (off-by-one shifts every downstream arrow); or an entry
+points past the segment's Pāli word count.
+
+**Why it happens.** English text edited without re-counting; Pāli
+re-segmented without re-indexing.
+
+**Guard.** `alignment-audit.test.ts` — pre-existing; this class was
+already loud. It fired ~15× during the verse-by-verse authoring and
+each was fixed before commit.
+
+**Pipeline rule.** Trivially machine-checkable; a generator should run
+the same length + range check on its own output before emitting. This
+is the cheap class — make every silent class look like this one.
+
+---
+
+## 4. Crossed morpheme arrows — *judgment, now authorable*
+
+**What.** When several English words map to one Pāli word, the
+renderer distributes their arrows across that word's morphemes *by
+position* (1st English → 1st morpheme). When English reorders the
+morphemes, the arrows cross — e.g. `kusalena` = `kusal` (skilled) +
+`ena` (by-an-agent), rendered "one … skilled", so the heuristic sends
+`kusal`'s arrow to "one".
+
+**Why it's judgment.** Which English word belongs to which morpheme
+is a semantic decision; position is only a guess.
+
+**Mitigation.** The `Witness.morphemeAlignTo` field lets a curator
+state the pairing explicitly. 41 witnesses authored for Metta.
+
+**Pipeline rule.** A generator that produces `alignTo` should produce
+`morphemeAlignTo` in the same pass — it already had to decide the
+word↔word mapping; the word↔morpheme mapping is the same reasoning one
+level finer. Falling back to the positional heuristic is acceptable
+*only* when English and Pāli morpheme order agree.
+
+---
+
+## 5. Grammar jargon in plain-register text — *judgment, tripwired*
+
+**What.** A gloss names a grammatical concept with a term a reader
+with no Pāli training doesn't know — "gerundive", "accusative
+singular", "past participle", a bare "(nominative)" appended as noise.
+
+**Why it's judgment.** `CURATION_PROTOCOL.md` §3.4's pay-rent rule:
+a technical term *may* stay if it does work plain English cannot AND
+is glossed in the same sentence. So it's not an absolute ban — but the
+default is to *show* the idea ("the object of 'I go to'", "the 'X-ed'
+form") rather than *name* it.
+
+**Found in.** 84+ hits across 9 chant files.
+
+**Guard.** A jargon tripwire in the test — *not* an absolute ban. A
+term that genuinely pays rent goes in `JARGON_ALLOWLIST` with a
+rationale. (The guard's case-insensitive scan already caught one hit
+— `Optative` — that a case-sensitive grep had missed.)
+
+**Pipeline rule.** Generate glosses for the stated default reader (a
+thoughtful adult, no Indic-linguistics training). A generator should
+run the jargon scan on its own output and, for any hit, either rewrite
+to plain English or justify the term inline. Treat a jargon hit as a
+*signal to re-read*, exactly as a human curator would.
+
+---
+
+## 6. Framing / commentary prose — *judgment*
+
+**What.** Curator-voice paragraphs above the chant body explaining
+"what this chant is about" — often AI-generated, displacive, and not
+what the reader came for.
+
+**Found in.** Heart Sutra, Metta Sutta, Way of Compassion, Om Mani,
+sangha descriptions.
+
+**Why no guard.** Whether a prose block earns its place is a pure
+taste call; a regex can't decide it. A pipeline can *flag* every
+`prose-commentary` section for human review but should not auto-strip.
+
+**Pipeline rule.** Default to *not* generating framing prose. The
+chant body, word glosses, and witness translations carry the meaning;
+a separate prose paragraph usually duplicates them. If commentary is
+generated, it must be sourced, not asserted.
+
+---
+
+## 7. Tokenization mismatches — *silent*
+
+**What.** The renderer's tokenizer splits a word the data did not
+expect — e.g. `c'assa` splits on the apostrophe into `c` + `assa`;
+a hyphenated compound splits on the hyphen. The `alignTo`/`words`
+indices then refer to the wrong tokens.
+
+**Mitigation.** The `tokens` hint on a `ScriptVariant` overrides
+tokenization for that line.
+
+**Pipeline rule.** A generator must tokenize with the *same* tokenizer
+the renderer uses (or emit an explicit `tokens` hint) and index
+`alignTo`/`words` against *that* token stream — never against a
+naive whitespace split when the text contains apostrophes, hyphens,
+or script-specific separators.
+
+---
+
+## 8. Script-mixed transliteration — *judgment*
+
+**What.** One transliteration line bundles two scripts' romanizations
+(Sino-Japanese romaji + Mandarin pinyin under the same Hanzi), so when
+the reader cycles to one script they see both readings layered.
+
+**Pipeline rule.** A transliteration belongs to exactly one script
+variant. When the source script is Hanzi, show pinyin; when Japanese,
+show romaji — never both on one line.
+
+---
+
+## The meta-lesson
+
+The two genuinely dangerous classes (**1** and **2**) were *silent* —
+they survived from first authoring until a human happened to hover the
+wrong word. The fix was not "curate more carefully"; it was **make the
+silent classes loud**: a `test()` per word, per chant, that re-derives
+the invariant the renderer depends on.
+
+For an auto-generation pipeline the principle is the same and sharper:
+**a generator must run, on its own output, every guard the renderer's
+correctness silently assumes** — reconstruction, index shape, token
+agreement — and degrade cleanly (emit less, not broken) when it cannot
+satisfy one. The judgment classes (4–6, 8) it cannot resolve alone;
+those it should *flag*, with the specific reason, for the smallest
+possible human review — never silently guess and move on.
diff --git a/tests/components/liturgy/liturgy-data-quality.test.ts b/tests/components/liturgy/liturgy-data-quality.test.ts
new file mode 100644
index 0000000..4017dd9
--- /dev/null
+++ b/tests/components/liturgy/liturgy-data-quality.test.ts
@@ -0,0 +1,127 @@
+/**
+ * Liturgy data-quality audit — catches two *silent* failure classes that
+ * neither crash nor fail a render, they just quietly degrade the reader.
+ *
+ * 1. Morpheme reconstruction.
+ *    A word's `morphemes[]` (and per-script `scriptMorphemes[]`) must
+ *    concatenate back to the surface form. The renderer's `splitByMorphemes`
+ *    walks the surface slicing one morpheme at a time; if the pieces don't
+ *    reconstruct it returns `null` and the word silently falls back to a
+ *    single whole-word hover — the per-morpheme tooltips/arrows just vanish
+ *    with no error. Pāli/Sanskrit sandhi at the morpheme join (two vowels
+ *    merging, a vowel shortening, i→y before a vowel) is the usual cause.
+ *
+ * 2. Segment-ID leaks.
+ *    Cross-reference glosses sometimes use the curator's internal segment-ID
+ *    shorthand ("same word as v7d") — meaningless to a reader. Reader-facing
+ *    text must never contain a `v<number><letter>` token; say "verse N".
+ */
+
+import { describe, it, expect } from 'vitest';
+import { LITURGY_DOCS_BY_SANGHA } from '../../../data/liturgy';
+import type { LiturgyDoc, TripleScriptWitnessSection } from '../../../types/liturgy';
+
+const ALL_DOCS: LiturgyDoc[] = Object.values(LITURGY_DOCS_BY_SANGHA).flatMap(
+  (docsForSangha) => Object.values(docsForSangha)
+);
+
+function tripleScriptSections(doc: LiturgyDoc): TripleScriptWitnessSection[] {
+  return doc.sections.filter(
+    (s): s is TripleScriptWitnessSection => s.shape === 'triple-script-witness'
+  );
+}
+
+// The curator's internal segment-ID shorthand: v + digits + a trailing
+// letter (v1a, v7d, v10c). Near-zero false-positive risk in real prose.
+const SEGMENT_ID_SHORTHAND = /\bv\d+[a-z]\b/;
+
+// Grammar jargon that CURATION_PROTOCOL.md §3.4 flags as failure-tone in
+// the plain-prose register. A glossed word should *show* the grammatical
+// idea ("the object of 'I go to'", "the 'X-ed' form") rather than name it
+// with a term a no-Pāli-training reader has to already know.
+//
+// This is a tripwire, not an absolute ban: the protocol's pay-rent rule
+// does allow a technical term IF it does work plain English can't AND is
+// glossed in the same sentence. If a future gloss legitimately earns one,
+// add it to an explicit allowlist here with the rationale — don't just
+// delete the test. Today every liturgy gloss is plain, so the list is bare.
+const JARGON = /\b(gerundive|accusative|nominative|genitive|locative|ablative|optative|vocative|declension|declensional|instrumental case|past participle|present participle)\b/i;
+const JARGON_ALLOWLIST: string[] = [];
+
+for (const doc of ALL_DOCS) {
+  const sections = tripleScriptSections(doc);
+  if (sections.length === 0) continue;
+
+  describe(`liturgy data quality: ${doc.sangha}/${doc.slug}`, () => {
+    for (const section of sections) {
+      for (const segment of section.segments) {
+        for (const word of segment.words ?? []) {
+          // ── 1. Morpheme reconstruction ──────────────────────────────
+          if (word.morphemes && word.morphemes.length > 0) {
+            it(`${segment.id} · ${word.form}: morphemes reconstruct the surface form`, () => {
+              const concat = word.morphemes!.map((m) => m.text).join('');
+              expect(
+                concat.toLowerCase(),
+                `morphemes [${word.morphemes!.map((m) => m.text).join('+')}] ` +
+                  `do not reconstruct "${word.form}" — splitByMorphemes will ` +
+                  `return null and the word degrades to whole-word hover`
+              ).toBe(word.form.toLowerCase());
+            });
+          }
+
+          // Per-script morphemes must reconstruct that script's surface.
+          // A multi-token script word (e.g. `賀多舍 娑曩喃`) carries
+          // token separators — ASCII space, Tibetan tsek `་` — that are
+          // not morpheme content; the renderer splits per-token, so the
+          // morphemes reconstruct the separator-free string. Strip those
+          // before comparing.
+          const stripSep = (s: string) => s.replace(/[\s་]/g, '');
+          for (const [lang, morphs] of Object.entries(word.scriptMorphemes ?? {})) {
+            const surface = word.scriptAlts?.[lang];
+            if (!surface || !morphs || morphs.length === 0) continue;
+            it(`${segment.id} · ${word.form} [${lang}]: scriptMorphemes reconstruct the script surface`, () => {
+              const concat = stripSep(morphs.map((m) => m.text).join(''));
+              expect(
+                concat,
+                `${lang} scriptMorphemes do not reconstruct "${surface}"`
+              ).toBe(stripSep(surface));
+            });
+          }
+
+          // ── 2. Segment-ID leaks in word-level reader text ───────────
+          const wordText = [word.gloss, word.etymology].filter(Boolean) as string[];
+          for (const morph of word.morphemes ?? []) {
+            if (morph.gloss) wordText.push(morph.gloss);
+          }
+          for (const morphs of Object.values(word.scriptMorphemes ?? {})) {
+            for (const m of morphs ?? []) {
+              if (m.gloss) wordText.push(m.gloss);
+            }
+          }
+          for (const text of wordText) {
+            it(`${segment.id} · ${word.form}: gloss text carries no internal segment ID`, () => {
+              expect(
+                SEGMENT_ID_SHORTHAND.test(text),
+                `reader-facing text contains a segment-ID shorthand — ` +
+                  `use "verse N" instead: "${text}"`
+              ).toBe(false);
+            });
+
+            // ── 3. Grammar jargon in plain-register gloss text ────────
+            it(`${segment.id} · ${word.form}: gloss text stays in plain register`, () => {
+              const match = text.match(JARGON);
+              const flagged =
+                match !== null && !JARGON_ALLOWLIST.includes(match[0].toLowerCase());
+              expect(
+                flagged,
+                `reader-facing text uses grammar jargon "${match?.[0]}" — ` +
+                  `show the idea in plain English (CURATION_PROTOCOL §3.4), ` +
+                  `or allowlist it with a pay-rent rationale: "${text}"`
+              ).toBe(false);
+            });
+          }
+        }
+      }
+    }
+  });
+}
diff --git a/types/liturgy.ts b/types/liturgy.ts
index 7d6e3ae..d1c97d0 100644
--- a/types/liturgy.ts
+++ b/types/liturgy.ts
@@ -40,6 +40,20 @@ export type Witness = {
    * e.g., "Fully Self-Enlightened One" all aligns to *sammā-sambuddhassa*).
    */
   alignTo?: number[];
+  /**
+   * Optional per-token morpheme target. Parallel-indexed to `alignTo`.
+   * `morphemeAlignTo[i]` is the morpheme index (within the Pāli word that
+   * `alignTo[i]` points to) that English word `i` should anchor its arrow
+   * to. `null`/absent → the renderer falls back to its positional heuristic
+   * (i-th English token mapped to a word → i-th morpheme).
+   *
+   * Why this exists: when English reorders the morphemes of a Pāli word,
+   * the positional heuristic crosses the arrows. Example: `kusalena` =
+   * `kusal` (skilled) + `ena` (by-an-agent); Amaravati renders it "one …
+   * skilled", reversing the order, so the heuristic sends `kusal`'s arrow
+   * to "one". Authoring `morphemeAlignTo` fixes the pairing explicitly.
+   */
+  morphemeAlignTo?: (number | null)[];
 };
 
 /**