Gurmukhi Script
Gurmukhi carries Punjabi in India and is the exclusive script of Sikh scripture. It shares the series introduction's syllable-centric logic with Devanagari, but its everyday complexity looks different: there's no continuous headline, and the two features that define a well-built Gurmukhi font are the addak gemination mark and a small set of subjoined (pairī) consonant forms. This article covers its letters, those two features, and what a font needs to render it correctly.
Gurmukhi, in one screen
An abugida descended from Brahmi via the earlier Laṇḍā scripts, standardized in the 16th century, used for Punjabi in India and for the entire text of the Guru Granth Sahib.
Gurmukhi builds its vowels differently from Devanagari. Three consonant-shaped letters act as vowel bearers, and the ten independent vowel sounds are formed by attaching a vowel sign to one of the three. Traditional counting, the Painti ("the thirty-five"), treats those three bearers as part of the consonant count alongside two standalone letters and five varga groups of five. Shown separately here for clarity, the pieces still add up to the same thirty-five.
Devanagari letters share a top stroke that fuses into one continuous shirorekha. Gurmukhi letters also carry top strokes, but they stay independent: set two consonants side by side and their tops don't join into a single bar. The result is a broken, rhythmic silhouette along the top of a word rather than one solid line.
Without a headline to anchor the eye, that job falls to the baseline instead, which raises the stakes for the zones above and below it. Everyday Punjabi text is dense with above-base marks (addak, the top vowel signs, bindi and tippi) and it also uses below-base vowel signs. A Gurmukhi typeface holds together because those marks stay legible and collision-free at text size, not because of a single wide connecting stroke the way a Devanagari design does.
With no headline to draw, three bands do the work instead: an ascender zone above for addak, dots, and the top vowel signs, the baseline itself, and a descender zone below for the two below-base vowel signs.
Notice that, unlike the Devanagari diagram in the previous article, there's no strong line running through the letter tops here: the tippi over ਪੰ and the vowel sign in ਬੀ both sit above the baseline, but nothing connects them into a bar. The baseline, drawn strong here, is doing the job the shirorekha did for Devanagari.
The series introduction showed a Devanagari consonant and vowel sign reordering at render time. The same reordering rule applies to Gurmukhi's pre-base vowel sign, alongside a post-base sign that doesn't move and a combined sign built from two visual pieces.
The sihārī (vowel sign I) in ਕਿ is typed after the consonant but drawn to its left, the same reordering the series introduction covered. The vowel sign in ਕਾ stays on the right where it's typed, no reordering involved. The vowel signs in ਕੋ and ਕੌ are each a single Unicode character, but like Devanagari's split matras they're drawn as two pieces: a small hook above the consonant and a vertical stroke to its right, coordinated as one glyph.
Gurmukhi's everyday complexity isn't tall conjunct stacks. It's two narrower features: a gemination mark that has to share crowded headroom with other above-base marks, and a small set of consonants that compress into reduced companion forms instead of standing at full size.
Addak marks that the following consonant is doubled: ਪੱਤਾ ("leaf") is pronounced with a doubled ਤ, and the mark itself sits in the ascender zone, where it has to stay visually distinct from top vowel signs and dot marks rather than merging into them. Pairī ("at the foot") forms are Gurmukhi's version of a conjunct: instead of stacking a full consonant, a small set, ਰ, ਵ, and ਹ, reduces to a below-base companion form, encoded as the base consonant plus virama plus the subjoined letter. A fourth consonant, ਯ, takes a post-base form instead, sitting after the base rather than beneath it.
The series introduction covered the general Indic feature set. Gurmukhi is a lighter user of it than Devanagari: the OpenType spec makes half and rphf available for typographic preference, but ordinary Gurmukhi text doesn't use half-forms or reph the way Devanagari does. What it does lean on is below-base and post-base substitution, plus nukta handling for the six loanword letters.
| Feature | Tag | What it builds here |
|---|---|---|
| Nukta forms | nukt | Substitutes the precomposed forms of ਸ਼, ਖ਼, ਗ਼, ਜ਼, ਫ਼, and ਲ਼, or positions a separate nukta mark on its base. |
| Below-base forms | blwf | Builds the subjoined (pairī) forms of ਰ, ਹ, and ਵ. |
| Vattu ligatures | vatu | Ligates a below-base pairī form directly onto the preceding consonant, where the font provides a combined shape. |
| Post-base forms | pstf | Builds the post-base pairī form of ਯ. |
| Localized forms | locl | Swaps in Punjabi-specific glyph variants where they differ from the script default. |
| Above-base marks | abvm | Positions addak, tippi, bindi, and the above-base vowel signs over the base glyph. |
| Below-base marks | blwm | Positions the below-base vowel signs and the nukta mark under the base glyph. |
Applied in order, these features substitute the pairī and nukta forms the syllable needs, then position every mark relative to whatever base glyph resulted. None of it depends on the reph or half-form machinery Devanagari relies on, even though a font is technically free to declare those features too.
| Symptom | Likely cause | Fix |
|---|---|---|
| Addak collides with a top vowel sign or dot mark | Missing or weak above-base mark-to-mark anchors | Build explicit mark-to-mark anchors for addak against every top vowel and dot combination |
| Sihārī appears on the wrong side or overlaps the previous letter | Indic reordering isn't running, or the shaping engine isn't invoking it for this script | Confirm the app is actually running Indic shaping for Gurmukhi, not rendering glyphs in stored order |
| Pairī rakar or vava lands in the wrong spot | Below-base anchor missing or mispositioned on the base glyph | Audit below-base anchors across the full consonant set, not just the common bases |
| ਸ਼, ਖ਼, ਗ਼, ਜ਼, ਫ਼, or ਲ਼ render as a base letter plus a stray dot | Font lacks the precomposed glyph or the nukt substitution | Add explicit nukta substitutions or precomposed glyphs for all six loanword letters |
| Below-base vowel signs get clipped in UI | Line-height or container too tight for the descender zone | Allow real descender space in UI text and test on the platforms most likely to clip it |
- 1Draw consistent top strokes across the consonant set, and confirm the baseline, not a connecting line, is what carries the word.
- 2Build the addak mark and test it against every top vowel sign and dot mark for collisions.
- 3Implement pairī forms for ਰ, ਵ, and ਹ, plus the post-base form for ਯ, and test each against consonants with different top-stroke shapes.
- 4Add precomposed glyphs or nukta substitutions for all six modern loanword letters.
- 5Position every vowel sign, including combined ones like ਕੋ and ਕੌ, against every consonant shape.
- 6Test across shaping engines and at multiple sizes and weights before calling the script done.
Copy these into a test document. Each targets a different part of the shaping pipeline covered above.