Reading Nastaʿlīq and Older Print Styles

Nastaʿlīq preserves the same logical Persian letters as modern type, but it does not keep them on one even baseline. A joining group may descend diagonally, one letter may sit above another, and the dots may be placed only after the main strokes are complete. Shikasta-Nastaʿlīq pushes those freedoms farther; lithography reproduces a scribe’s hand; early movable type pulls the script back toward repeatable metal sorts.

The reliable method is structural: locate a line’s right edge, recover joining groups, read the undotted skeleton, assign dots, then normalize spelling in a separate layer. Do not “modernize while looking.” The source image, diplomatic transcription, and normalized reading answer three different questions.

Four technologies, visibly compared

Nastaʿlīq manuscript handShikasta-Nastaʿlīq
1879 Nastaʿlīq siyah-mashq with heavily scaled and stacked strokes
Free scaling, overlap, and diagonal composition
1898–99 Shikasta-Nastaʿlīq Hafez passage with compressed joins
Compressed joins and displaced dots
Lithographic bookOlder movable type
1901 lithographic Shahnameh spread reproducing a scribal hand
Printed frame and columns, but handwritten letter variation
1829 Richardson dictionary page with repeatable Persian movable type
Even lines and mechanically repeated sorts

This comparison is diagnostic, not decorative: Nastaʿlīq and Shikasta change the geometry of a hand; lithography reproduces that changing geometry on many copies; movable type assembles a line from repeatable pieces.

Nastaʿlīq as calligraphic composition

Nastaʿlīq siyah-mashq by Mirza Gholamreza Esfahani, 1879

Nastaʿlīq siyāh-mashq by Mirza Gholamreza Esfahani, ۱۲۹۶ AH/1879, Malek National Library and Museum. Public-domain source and catalog transcription.

This practice composition is not a modern paragraph: scale, overlap, and repetition are part of the design. The inscription is catalogued diplomatically without imposing modern punctuation. The normalized column below regularizes boundaries for reading but does not pretend that the Arabic quotations are modern conversational Persian.

Diplomatic catalog transcriptionNormalized reading layer
توکلت علی الله ففرو الی الله مشقه الفقیر غلام رضا سنه ۱۲۹۶توکلت علی‌الله؛ ففرّوا الی‌الله. مشقه الفقیر غلام‌رضا، سنه‌ی ۱۲۹۶.

The huge black strokes make a crucial point: ink position is not character order. Read the inscription’s logical sequence from the rightmost joining group; do not begin at the visually highest or largest stroke.

Shikasta-Nastaʿlīq: compressed joins and mobile dots

Shikasta-Nastaʿlīq by Ali Akbar Golestaneh, 1316 AH

A Hafez passage written in Shikasta-Nastaʿlīq by Ali Akbar Golestaneh, ۱۳۱۶ AH/1898–99, public domain. Full attribution, transcription, and licence. Hafez’s poem is fourteenth-century and public domain.

Read the upper right first, then continue leftward through the written line. The catalog and established text let us keep diplomatic observation apart from modern editorial normalization:

Diplomatic layerNormalized reading
دوش می آمد و رخساره بر افروخته بوددوش می‌آمد و رخساره برافروخته بود
تا کجا باز دل غمزده سوخته بودتا کجا باز دل غم‌زده‌ای سوخته بود
کفر زلفش ره دین می زد و آن سنگین دلکفر زلفش ره دین می‌زد و آن سنگین‌دل

دوش می‌آمد و رخساره برافروخته بود.

dush mi-âmad va rokhsâre bar-afrukhte bud.

Last night the beloved came, face aflame. (literary; Hafez)

تا کجا باز دل غم‌زده‌ای سوخته بود؟

tâ kojâ bâz del-e gham-zade-i sukhte bud?

Who knows which sorrow-stricken heart had been burned again? (literary; Hafez)

The second line contains a genuine reading problem at scan scale. The small material corresponding to ‌ای can be swallowed visually by the final sweep and displaced dots, so the local shape invites غمزده. The established Hafez text and the metre support غم‌زده‌ای. A responsible transcription records that reasoning; it does not silently claim that every character was equally plain.

Lithography: printed, but still a hand

A 1901 lithographic Shahnameh spread

A lithographic *Shahnameh spread dated 1901, principally scribed by Muhammad Hussein; public-domain facsimile. Catalog record and licence. Ferdowsi’s work dates to about 1010.*

Lithography transfers the drawn page, so its letters retain sloping baselines, handwritten compression, and irregular dot placement. The repeated columns and printed frame do not turn the writing into movable type. Read each column at its own upper-right corner, and treat illustrations, captions, and marginal blocks as separate reading zones.

نسخه‌ی چاپ سنگی شبیه دست‌خط است.

noskhe-ye châp-e sangi shabih-e dast-khat ast.

A lithographic edition resembles handwriting.

Older movable type: repeatable sorts

Richardson dictionary page in early Persian and Arabic movable type, 1829

Page 709 of John Richardson’s Persian-Arabic-English dictionary, enlarged by Francis Johnson and printed by J. L. Cox in London, 1829; British Library IOR/R/15/5/397 via Qatar Digital Library. Public-domain record and IIIF source.

Here the repeated letter bodies, upright text lines, and fixed spacing reveal metal type. Joins are more modular and the dramatic Nastaʿlīq cascade is absent. Some glyph designs look Arabic or old-fashioned beside modern Iranian type, but the reading procedure remains Persian: identify the skeleton, then the word. Do not infer an Arabic pronunciation merely from the typographic style.

💡
A “printed” page may be a photographically or lithographically reproduced hand. Diagnose the production method visually: repeated identical sorts and an even baseline suggest movable type; freely changing strokes and cascading joins suggest a transferred hand.

Recover logical order from a stack

Take normalized برافروخته. Its logical letters are ب + ر + ا + ف + ر + و + خ + ت + ه, even when the written word descends. Non-connecting ر، ا، ر، و break the pen path, giving the practical groups بر | ا | فر | و | خته. The groups remain ordered right to left; vertical stacking never reverses the encoded sequence.

For a line, use this sequence:

  1. mark the right and left limits of the text line;
  2. split at obvious non-connectors and spaces;
  3. trace each joining group’s main stroke before considering dots;
  4. count and assign dots by plausible letter skeletons;
  5. test the resulting words against syntax, meaning, metre, and parallel witnesses;
  6. create a separate normalized layer with modern ک، ی, ZWNJ, and punctuation.

این سطر را از راست به چپ بخوانید.

in satr râ az râst be chap bekhânid.

Read this line from right to left.

ابتدا اسکلت واژه را پیدا کنید.

ebtedâ eskelet-e vâzhe râ peydâ konid.

First find the word’s skeleton.

نقطه‌ها بالای سطر جابه‌جا شده‌اند.

noqte-hâ bâlâ-ye satr jâbe-jâ shode-and.

The dots have been displaced above the line.

Terminal sweeps: compare families, not silhouettes

J. Swaine's engraved Persian handwriting plate, 1828

J. Swaine, engraved Plate 4 for the 1828 edition of William Jones’s *A Grammar of the Persian Language; public domain. Source record. The plate visibly places alternate terminal sweeps in comparable sequences.*

Use the plate as a contrast exercise. In the first two bands, cover the dots and compare the black skeletons; then uncover them.

Terminal familyWhat the sweep concealsWhat decides the reading
س / ص / نAll can finish in a broad leftward bowl or sweep; teeth may flatten, and ن may resemble a long bowl.ص begins with its heavier ṣād body; س retains the sīn tooth rhythm when visible; ن requires one dot, though that dot may float far above.
ح / ع / قThe descending terminal bowl can dominate the small entry stroke.Trace the entry: ح belongs to the ج-family skeleton, ع has its own hooked head, and ق is confirmed by its two dots even when displaced.
ل / یBoth may end in an exceptionally long return sweep.ل begins with a tall upright; ی has the yeh skeleton and historically variable dots, which are often omitted in Persian hands.

In Golestaneh’s Shikasta specimen, final ن in دین and سنگین stretches beneath neighboring material, while ل in دل carries a terminal sweep. Dots above سنگین are not necessarily centered over the letters they identify. Count candidates only after tracing the join; otherwise a displaced dot can make you invent a letter that the main stroke cannot support.

حروف را جداگانه بازنویسی نکنید.

horuf râ jodâgâne bâz-nevisi nakonid.

Do not rewrite the letters separately.

قرائت مشکوک را میان دو قلاب بگذارید.

qerâ'at-e mashkuk râ miyân-e do qollâb begozârid.

Put an uncertain reading between square brackets.

💡
A missing dot is evidence of uncertainty, not permission to guess. Preserve the undotted skeleton first; use grammar, collocation, metre, and other witnesses to rank readings.

Diplomatic transcription is not normalization

A diplomatic transcription preserves source spelling, word division, abbreviations, and uncertainty as the project can represent them. A normalized reading supplies declared modern conventions. A critical edition may add a third layer that chooses among witnesses. For modern encoding conventions, consult orthographic variation before and after standardization.

صورت اصلی را کنار صورت معیار نگه دارید.

surat-e asli râ kenâr-e surat-e me'yâr negah dârid.

Keep the original form beside the standard form.

هر تغییر را در یادداشت نسخه ثبت کنید.

har taghyir râ dar yâddâsht-e noskhe sabt konid.

Record every change in the textual note.

Common Mistakes

1. Following visual height instead of logical right-to-left order

❌ د‌و‌ب

Incorrect — this reverses the logical letters of بود because a sweep was followed left to right.

✅ بود

bud

was

2. Breaking every stacked form into separate letters

❌ م ی آ م د

Incorrect — ordinary spaces destroy the word and its joining groups.

✅ می‌آمد

mi-âmad

was coming

3. Copying Arabic kaf and yeh into normalized Persian

❌ اين كاتب قديمی است.

Incorrect normalized encoding — Arabic ي and ك have been copied into Persian text.

✅ این کاتب قدیمی است.

in kâteb qadimi ast.

This scribe is old.

4. Reproducing a manuscript gap as a modern full space

❌ می آمد و می نوشت.

Incorrect modern normalization — both productive prefixes require ZWNJ.

✅ می‌آمد و می‌نوشت.

mi-âmad va mi-nevesht.

He/She would come and write.

5. Treating an omitted short vowel as a missing letter

❌ بیرافروخته بود.

Incorrect — the short vowel in bar-afrukhte is not written with ی.

✅ برافروخته بود.

bar-afrukhte bud.

It was ablaze. (literary)

Key Takeaways

Read calligraphic Persian by logical groups, not by the picture’s apparent geometry. Nastaʿlīq stacks; Shikasta compresses and displaces dots; lithography preserves a hand; movable type repeats fixed sorts. Keep a diplomatic transcription beside a normalized one, flag real ambiguity, and distinguish terminal families by entry stroke, skeleton, and dot count rather than by the length of the final sweep.

Now practice Farsi

Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.

Start learning Farsi

Related Topics