History of the Pashto script: from the Roshani manuscripts to Unicode

At a glance

• Pashto is written in the Arabic script, with letters added or reshaped for sounds Arabic lacks: ټ ډ ړ ڼ ځ څ ښ ږ ې ۍ and others. [11, 12]

• The oldest Pashto work that survives is the Khayr ul-bayan of Bayazid Ansari (Pir Roshan), who lived from 1524/25 to 1572/73. Its oldest dated copy is from 1651, and a Makhzan ul-islam of 1649 is older. [5]

• Raverty, a British officer who published Pashto books in 1860, counted twenty-nine letters in the old alphabet, five of them peculiar to Pashto. [1]

• Scholars disagree on where the special letters were first made. Penzl argued for Kandahar and Morgenstierne for the north-east. [6, 10]

• Spelling was discussed at meetings in Kabul in 1942, 1948 and 1958, and at Bara Gali in Swat in 1990. [7, 8, 13]

• All the Pashto letters have been in Unicode since version 1.1. [11]

Arabic letters for a language with other sounds

Raverty, a British officer, wrote in his grammar of 1867 that Pashto is written in the naskh hand of Arabic. He held that the original alphabet had twenty-nine sounds, that nine Arabic letters never occur in pure Pashto words, and that "the language really contains but twenty-nine letters, including five peculiar ones". He named them by sound: ts, dz, a trilled or retroflex r, sh and a nasal n. [1] These correspond to the letters څ, ځ, ړ, ښ and ڼ. [1, 11]

Counts differ by writer. Morgenstierne, writing for the Encyclopaedia Iranica, counted thirty-one consonants in the "classical" orthography recorded since the sixteenth century, including q, f and ʿayn that occur only in loanwords. [10] MacKenzie counted thirty-six consonant signs in the standard alphabet of 1959, seven of them almost only in Arabic loanwords. [6] Ishida's orthography notes list thirty-two basic consonant letters and eight more for loanwords. [12]

The Unicode Standard's chart for the Arabic block marks these letters as used for Pashto: ټ (teh with ring), ډ (dal with ring), ړ (reh with ring), ڼ (noon with ring), ځ (hah with hamza above), څ (hah with three dots above), ښ (seen with dot below and dot above), ږ (reh with dot below and dot above), ۍ (yeh with tail) and ې (the letter "e"). [11] The four ringed letters write the retroflex sounds, which Morgenstierne lists as ṭ, ḍ, ṇ and ṛ. The chart says ځ stands for /dz/ and څ for /ts/. [11] Ishida adds that Pashto has special letters for word-final diphthongs, which usually do grammatical work as well as show a sound. [12] The sounds themselves are described in The sounds of Pashto and The Pashto alphabet.

The oldest Pashto writing

Mikhail Pelevin, who described the Pashto manuscripts of the Berlin State Library in 1994, calls Bayazid Ansari's Khayr ul-bayan ("Best manifest") the oldest of all the works in Pashto preserved to the present day. [5] Bayazid is better known as Pir Roshan, the founder of the Roshani movement. [5] His life and his quarrel with Akhund Darweza are told in Pir Roshan and Akhund Darweza.

One account, reported in a Nangarhar University article of 2025, puts the first attempt to adapt letters for Pashto much earlier, under Sultan Mahmud of Ghazni (387–421 AH). The authors say Qazi Khairullah Nowshari recorded it in the introduction to his Khair ul-lughat and gave no source, and they could find none. [8] No manuscript from that period is cited. What survives starts in the seventeenth century.

The Khayr ul-bayan and its explanation of letters

The copy Pelevin describes is Ms. or. fol. 4093 in Berlin. It was finished on 20 Ramazan 1061 (6 September 1651) by a scribe who calls himself a follower of Pir Roshan. Its first eight folios, 1b to 8a, are an introduction "with the explanations of the graphical system used by the author". The introduction is written in four languages: Arabic, Persian, Pashto and a fourth that Pelevin marks as Panjabi with a question mark. The main text is in Pashto with Arabic insertions, in rhymed prose. [5] Pelevin does not list the letters.

Another copy, dated 1668, is in the Salar Jung Museum in Hyderabad. It is complete, in all four languages, and every language except Persian is vowelled. [5] Pelevin also records that the oldest dated manuscript of the Makhzan ul-islam of Akhund Darweza, in the same museum, is dated 20 Safar 1059 (5 March 1649). It is older than the Berlin Khayr ul-bayan. [5]

The Nangarhar authors read Bayazid's introduction as a treatise on letters for Pashto's special sounds. They say he chose twelve letters and called them thirteen because he used the sign for g in two places, and that his script was not taken up outside his followers. They also say Bayazid saw the added letters as his work by God's command. [8] These are the authors' readings of Roshani texts and of the poet Dawlat Lawani. Pelevin, who saw the manuscript, only says the introduction explains a graphic system. [5] For Dawlat Lawani and the other Roshani poets, see Daulat Lawanai, Arzani Khweshki and Wasil Roshani.

Akhund Darweza and the spelling of the Makhzan

Akhund Darweza's Makhzan, which Raverty says he wrote against Pir Roshan's opinions, was written with his son Karimdad and also dealt with letters. [1, 5] John Leyden, writing in 1810, says it contains "mystical explanations of the Arabic and Pashtu letters of the alphabet". Karimdad added a dissertation on "the peculiar letters of the Afghans, as distinguished from those of the Arabs". Leyden translates his apology for the writing of Afghan poems: [3]

Know gentle reader, that in Afghan poems, the authors are accustomed to pay no great attention to the correctness and similarity of the rhyme, or the equal proportion of the lines to each other, in a couplet; nor are they more careful, with respect to uniformity of orthography.

Bernhard Dorn quoted the same passage in the preface to his Pashto reader, printed at St Petersburg in 1847. [4] Pelevin finds in a Berlin copy of the Makhzan a final part "about the peculiarities of the Afghan graphic system and the shortcomings of Pashto versification". [5]

Raverty went further. He wrote that the orthography "followed for the last three centuries or more" was "first arranged by Akhund Darwezah". He gave no source in the note. [1] The Nangarhar authors say Akhund Darweza's followers had a script of their own, in some respects like Bayazid's. [8] These accounts do not say who fixed which letter.

Whose alphabet: Kandahar or the north-east

The special letters follow the sound system of one dialect. In Kandahar speech the five sounds ts, dz, ž, ṣ̌ and ẓ̌ are all kept apart. MacKenzie's table shows that Yusufzai speakers in the north-east say ts as s, dz as z, and the last two as x and g. [6] Herbert Penzl observed that the Kandahar dialect alone has a sound system matching the usual spelling, and concluded, in MacKenzie's quotation of him, that the special signs must have been created in the Kandahar area: "Kandahar appears to be the cradle of the Pashto alphabet". [6] In his own article of 1959 he wrote that the standard orthography follows the distinctions of Kandahar, and that speakers of other dialects who have lost them still write by the Kandahar rule. [7]

Georg Morgenstierne had held the opposite. When the orthography was fixed in the sixteenth century, he wrote, the distinction between ṣ̌ and ẓ̌ and x and g was "still preserved even among the north-eastern tribes, who were probably the creators of Pashto literature". [6]

MacKenzie thought the question could not be decided "on orthographic evidence alone", and he brought one more fact. The manuscript of 1651, which he called the earliest known, spells dz and two other sounds with signs that later disappeared. They were still used in a diwan of Mirza dated 1690. In a diwan of the poet Najib dated 1696 to 1697 the copyist, Gul Muhammad Peshawari, changed the old signs to the new ones in his model. MacKenzie draws two conclusions: the older tradition was shared beyond the Roshanis, and the new letters still stood for separate sounds in the north-east at that date. [6]

Hands and spellings of the eighteenth century

Pelevin describes a Berlin copy of the diwan of Kazim Khan Shaida (1725 to 1778), a descendant of Khushal Khan Khattak, written by Faiz Ali of Delhi and bought at Agra in 1796. The hand is a large calligraphic nastaliq. The scribe used what Pelevin calls the Khattak system: a hamza over a letter marks the retroflex ṭ, ḍ, ṛ and ṇ, a hamza below with dots marks ts (two dots) and dz (one dot), and the sounds ṣ̌ and ẓ̌ are written with the plain letters for sh and zh because of the Khattak dialect. [5] See Kazim Khan Shaida.

The Nangarhar authors call this the script of Khushal Khan Khattak himself, "his own invention", and say his works and those of his sons and grandsons are written in it. [8] Pelevin does not say who made it, and the article does not cite a manuscript in Khushal's own hand. Khushal's poetry is in Khushal Khan Khattak.

Raverty, in 1867, wrote that modern Pashto books varied greatly in orthography and that no manuscript later than the reign of Ahmad Shah or his son Timur Shah could be trusted, since it was almost impossible to find two copies agreeing, unless one person wrote them. He adopted the spelling of the Makhzan and of an Aurangzeb-period prose work he had found at a pawn shop in Bombay. [1] He also noted books in which the copyist wrote ordinary Arabic letters in place of the Pashto ones. [1] The Nangarhar authors describe an "Ahmad Shahi" script used until the reign of Amir Sher Ali Khan that had no standard form and depended on each writer. [8]

Printing: St Petersburg, Hertford, Lahore and Peshawar

Raverty mentions a Pashto New Testament made by the Serampore missionaries in 1818. [1] Dorn printed his Chrestomathy of the Pushtu or Afghan Language, with an Afghan and English glossary, in 1847. The glossary opens with an "Afghan alphabet" and notes that Afghans, like Persians and Turks, do not always write the vowel letters. [4]

Raverty's Gulshan-i-Roh, a Pashto reader, was printed at Hertford by Stephen Austin in 1860. His preface says Pashto manuscripts of any age were scarce, and offers a book printed "in a type which, for clearness and beauty, cannot be equalled". He meant it to let a European orientalist learn the language at home. [2]

Pelevin writes that, after Dorn's and Raverty's work on Rahman Baba, "more than ten lithographic editions" of his diwan appeared in India and Pakistan, and he names a Peshawar edition of 1960. He gives no dates for the lithographs. [5] Editions of the Khayr ul-bayan appeared at Peshawar in 1967 and at Kabul in 1975. [5] Bellew's dictionary was issued in Lahore in 1901 by the Mufid-i-Am Press of Munshi Gulab Singh and Sons. [9]

One spelling for every dialect

Pashto Academy members discussed spelling at meetings in Kabul in 1942, 1948 and in the summer of 1958. Penzl, writing in 1959, says they took up questions of orthographic practice and the standing of certain forms used by writers. [7] Khalid Khan Khattak, a linguist, told a workshop at the Pashto Academy in Peshawar in December 2016 that standardisation began in the late 1930s. In 1939, he said, an Afghan minister, Gul Mohammad Wazir Mohmand, had proposed important changes to the script that could not be adopted. In 1942 scholars at the Kabul Pashto Academy made standardisation possible. [13]

Penzl found one local feature in Peshawar publications: the diphthong ay was often written with the letter ې (majhul ye), because Peshawar speakers say it as e. The orthography conference of the Pashto Academy condemned the practice in 1958. [7]

A three-day seminar at Bara Gali in Swat, held on 10 July 1990 on the initiative of the Pashto Academy of the University of Peshawar, brought together about a hundred writers and linguists. A working committee reached unanimous decisions, among them that Arabic words that entered Pashto should keep their Arabic spelling, and that Arabic letters not be written in Pashto words. [8] Zalmai Hewadmal, the chief investigator quoted in the same article, said the decisions had technical difficulties. [8] Khattak said the two-day Academy workshop of December 2016 had ratified about 17 decisions taken at the 1990 seminar. Yousaf Jazab, a young linguist, told it that the script had many shortcomings and no place for semivowels. [13] The history of these meetings is in Standard written Pashto and Pashto academies.

On 23 August 2026 the Pajhwok Afghan News agency reported a five-day workshop at the National Museum in Kabul on "Correction and Unification of Pashto and Dari Writing Systems", held by a commission of the Supreme Council of Languages. [14]

Pashto in Unicode

Every Pashto letter named above is in the Unicode Arabic block, and the Unicode character database's age file dates each of them to version 1.1. [11] Ishida's notes on Pashto orthography, last updated on 27 April 2026, explain how to write it. He says the Afghan and Peshawar authorities standardise a few details differently, so the code points used differ slightly: the word-final schwa is written ۀ in Peshawar and ه elsewhere. [12] He also says the Pashto standards recommend ک (U+06A9) over the Arabic ك (U+0643). Common fonts do not show the difference, and search suffers when the two are mixed. [12]