Avestan, Pahlavi & Sogdian Scripts

The Avestan alphabet was created to write down the Avesta, the Zoroastrian scripture, which priests had passed on by heart for a thousand years. It was devised in the Sasanian period, most likely in the 5th or 6th century, by scholars who took the cursive Pahlavi letters of their own day and added a full set of vowel signs, so that every sound of the sacred recitation could be fixed on the page. The result is one of the most exact alphabets of the ancient world: 53 letters in Hoffmann's standard count, written from right to left, with a dot ⸱ between words. Unicode encodes 54 letters, because two historical variants of the same sound got separate code points.

The vowels come first, each short vowel with its long partner: 𐬀 a and 𐬁 ā, 𐬂 å and 𐬃 ā̊, the nasal 𐬄 ą and 𐬅 ą̇, the schwa 𐬆 ə and 𐬇 ə̄, 𐬈 e and 𐬉 ē, 𐬊 o and 𐬋 ō, 𐬌 i and 𐬍 ī, 𐬎 u and 𐬏 ū. Then the consonants: the velars 𐬐 k, 𐬑 x, 𐬒 xʸ, 𐬓 xᵛ, 𐬔 g, 𐬕 ġ, 𐬖 γ; the palatals 𐬗 c and 𐬘 j; the dentals 𐬙 t, 𐬚 ϑ, 𐬛 d, 𐬜 δ and the unreleased 𐬝 t̰ found at the ends of words; the labials 𐬞 p, 𐬟 f, 𐬠 b, 𐬡 β; eight nasals from 𐬢 ŋ to 𐬩 m̨; 𐬪 ẏ, 𐬫 y, 𐬬 v, 𐬭 r and the rare 𐬮 l; and the sibilants 𐬯 s, 𐬰 z, 𐬱 š, 𐬲 ž, 𐬳 š́, 𐬴 ṣ̌ with 𐬵 h at the end. Double vowels are written twice, so vairiiō has two 𐬌 in a row.

The language is Avestan, an Old Iranian tongue in two stages: Old Avestan, the language of the Gāϑās attributed to Zaraϑuštra himself, and Younger Avestan, used for the rest of the Yasna, the Yašts and the Vidēvdād. The name Ahura Mazdā is 𐬀𐬵𐬎𐬭𐬀⸱𐬨𐬀𐬰𐬛𐬁, Zoroaster is 𐬰𐬀𐬭𐬀𐬚𐬎𐬱𐬙𐬭𐬀, and the three-word summary of the faith, good thoughts, good words, good deeds, is 𐬵𐬎𐬨𐬀𐬙𐬀⸱𐬵𐬏𐬑𐬙𐬀⸱𐬵𐬎𐬎𐬀𐬭𐬱𐬙𐬀. The oldest surviving manuscripts date from the late 13th century and were copied by Parsi and Iranian Zoroastrian priests, who still recite these texts in fire temples in Mumbai, Gujarat, Yazd and Kerman. The same letters were also used for Pazand, Middle Persian written out phonetically in Avestan script, which is why some Zoroastrian prayer books look Avestan but read as Persian.

The manuscripts punctuate with stacked dots and rings: the three-dot marks 𐬺 𐬻 𐬼 in three sizes and the ring marks 𐬾 𐬿 close phrases, verses and sections, 𐬽 is the inverted form, and 𐬹 marks an abbreviation. The old-school name of the script in Pahlavi is dēn dibīrīh, 'religion writing'. Unicode added the block in version 5.2 (2009) together with the word separator ⸱, which lives in the Supplemental Punctuation block but belongs to this alphabet.

Pahlavi script and Parthian: Middle Persian on stone and coins

Pahlavi is the name for the scripts of Middle Persian, the language of the Sasanian Empire (224–651) and of the Zoroastrian books written after it. All of them descend from the Imperial Aramaic alphabet that the Achaemenid chancellery had used across Iran (the ancestor is in the Aramaic & Ancient Arabian Scripts collection), and all kept a strange Aramaic habit: common words were still spelled as Aramaic and read out in Persian. The Aramaic letters MLKA, 'king', were written and pronounced šāh; MLKAn MLKA, 𐭬𐭫𐭪𐭠𐭭 𐭬𐭫𐭪𐭠, is the šāhān šāh, King of Kings, of every Sasanian inscription. These spellings are called heterograms, and they are the first thing a student of Pahlavi learns to see through.

Inscriptional Parthian is the older cousin. Parthian was the language of the Arsacid dynasty (247 BCE–224 CE), and its script, 22 letters that still match the Aramaic abjad one for one, is known from the ostraca of Nisa in Turkmenistan, from coins, and from the first Sasanian royal inscriptions, which were cut in Parthian as well as Middle Persian: Shapur I's great account of his wars on the Ka'ba-ye Zartosht at Naqsh-e Rostam, and Narseh's inscription at Paikuli. The Parthian word for Iran is 𐭀𐭓𐭉𐭀𐭍, Aryān. The numbers 𐭘 1, 𐭙 2, 𐭚 3, 𐭛 4, 𐭜 10, 𐭝 20, 𐭞 100 and 𐭟 1000 are additive and, like the letters, run from right to left.

Inscriptional Pahlavi is the Middle Persian of those same monuments and of the coins of every Sasanian king from Ardashir I onward, and of the inscriptions of the high priest Kerdir. By then the letters had started to merge: one sign 𐭥 stands for waw, ayin and resh, and one 𐭬 for mem and qoph, which leaves 19 letters. Ērān, the name of the empire, is 𐭠𐭩𐭥𐭠𐭭. Psalter Pahlavi is a slightly later and more cursive hand known from one text: a Middle Persian translation of the Psalms, made from Syriac by the Church of the East and found in fragments at Bulayïq near Turfan in Chinese Turkestan, far from Iran. It has 18 letters, the section marks 𐮙 𐮚 and the four-dot signs 𐮛 𐮜, and numbers up to 𐮯 100.

The script people usually mean by 'the Pahlavi alphabet', the cursive Book Pahlavi of the Bundahišn, the Dēnkard and the Zoroastrian priestly literature, is not in Unicode. Its letters merged even further, to about a dozen distinct shapes, so that a single written word can be read several ways and the script cannot be encoded as a simple alphabet; proposals have been before the Unicode Technical Committee since 2007, the latest revised in 2024, and as of Unicode 18.0 it is still unencoded. Do not use the Inscriptional Pahlavi letters on this page as a stand-in for Book Pahlavi: they are a different script, and specialists will read them as the stone letters they are. Old Persian cuneiform, the Achaemenid script that came before all of these, has its own place in the Cuneiform Symbols collection.

Sogdian alphabet and Old Uyghur: the scripts of the Silk Road

The Sogdians were the merchants of the Silk Road. From Samarkand, Bukhara and Panjikent they ran the trade between Iran, India and China between the 4th and the 8th century, and their East Iranian language was the lingua franca of the caravan routes. They wrote it in their own Aramaic-derived alphabet, which Unicode splits into two blocks. Old Sogdian is the formal script of the earliest texts: the Ancient Letters, a bundle of merchants' letters from about 313 CE found by Aurel Stein in a ruined watchtower west of Dunhuang, the Kultobe inscriptions from southern Kazakhstan and rock inscriptions on the Upper Indus. Its word-final letters have shapes of their own, 𐼁 𐼃 𐼆 𐼏 𐼖 𐼛, and it has numbers from 𐼝 1 to 𐼥 100 and even a fraction, 𐼦 ½.

Sogdian proper is the cursive that developed from the 6th century and became the script of the Mugh documents, the archive of the ruler Dēwāštič hidden on Mount Mugh in Tajikistan in 722 as the Arab armies arrived, of the Bugut inscription in Mongolia, where the Türk khaganate used Sogdian as a language of state, and of thousands of Buddhist, Manichaean and Christian pages from Turfan and Dunhuang. Sogdian has no daleth: the letter 𐼹 lamedh is read δ, so the scribes made 𐽄 lesh for l and 𐽃 feth for f, and told similar letters apart with the combining dots and hooks 𐽆 𐽇 𐽈 𐽉 𐽌 𐽍. The word Samarkand is 𐼼𐼺𐼰𐽀𐼸𐼻𐼹, smʾrknδ. From the 6th century on, Sogdian was often written vertically in columns running left to right, the direction the Uyghurs then made the rule.

Old Uyghur is that vertical Sogdian cursive adopted by the Uyghur Turks of the Uyghur Khaganate and the Kingdom of Qocho in the Turfan oasis, from the 8th or 9th century. In it they wrote Buddhist sutras like the Altun Yaruq, Manichaean confessions and Christian texts, legal contracts and poetry; it was in use for a thousand years, from the Yuan dynasty chancelleries to the Timurid court of Herat, and lingered among the Yellow Uyghurs of Gansu into the 17th century. The letters merged still further: 𐽲 is both γ and x, 𐽰 aleph spells a and ä, 𐽳 waw spells o, u, ö and ü, and readers lean on the dots 𐾂 𐾃 𐾄 𐾅 and on context. Uyγur itself is 𐽰𐽳𐽶𐽲𐽳𐽾, and täŋri, heaven or god, is 𐾀𐽺𐽷𐽾𐽶.

In 1204 Genghis Khan had the Uyghur scribe Tata-tonga teach this script to his sons, and it became the Mongolian script, which is still written in Inner Mongolia today and has its own Traditional Mongolian Script collection; from Mongolian it passed to Manchu in 1599. The modern Uyghur language is written in Arabic letters (Uyghur Alphabet), and the Orkhon runes of the earlier Turkic empires are a different script altogether (Old Turkic Alphabet). Unicode encoded Old Sogdian and Sogdian in version 11.0 (2018) and Old Uyghur in 14.0 (2021); all three are stored right to left, and a vertical display is a matter of fonts and layout.

Manichaean, Chorasmian and Elymaic scripts

The Manichaean script was the book hand of the religion founded by Mani (216–274 or 277) in Sasanian Mesopotamia. His followers said he designed it himself; it is in any case a cousin of the Palmyrene and Syriac Aramaic scripts, written right to left with the Aramaic letters 𐫀 aleph to 𐫤 taw. Manichaeism spread east along the same roads as the Sogdians and was the state religion of the Uyghur Khaganate from 763 to 840, and the script was used for Middle Persian, Parthian, Sogdian, Bactrian, Tocharian and Old Uyghur. Almost everything that survives in it comes from Turfan, where German expeditions between 1902 and 1914 dug up fragments of exquisitely illuminated books. To write those languages the scribes added dotted and modified letters, 𐫂 𐫄 𐫊 𐫋 𐫌 𐫑 𐫒 𐫔 𐫕 𐫚 𐫜 𐫟 𐫠 𐫣, a whole-word sign 𐫈 for ud, 'and', the numbers 𐫫 1, 𐫬 5, 𐫭 10, 𐫮 20, 𐫯 100 and ornamental punctuation like the star 𐫰 and the fleuron 𐫱. Mani's own name is 𐫖𐫀𐫗𐫏. The block came in Unicode 7.0 (2014), with Psalter Pahlavi.

Chorasmian is the script of Khwarazm, the oasis land on the lower Amu Darya south of the Aral Sea, now Karakalpakstan in Uzbekistan and northern Turkmenistan. Khwarezmian was an East Iranian language related to Sogdian, and from about the 2nd century BCE to the 8th century CE it was written in this Aramaic-derived alphabet on coins, on the leather and wooden documents from the palace of Toprak-kala, on silver bowls and on the ossuaries of Tok-kala, where Zoroastrians kept the bones of their dead. After the Arab conquest the language lived on in Arabic letters until about the 14th century, and the great scholar al-Biruni, born in Khwarazm in 973, recorded its calendar and festivals. The script has 21 letters, including a 𐾱 small aleph and a 𐾷 curled waw, and numbers 𐿅 1 to 𐿋 100; Unicode added it in version 13.0 (2020).

Elymaic belongs to Elymais, the kingdom in the hills of Khuzestan, around Susa and the Zagros foothills, that kept a half-independence under the Parthians from the 2nd century BCE until the Sasanians ended it in the 220s CE. The language of its inscriptions is Aramaic, but the letters took a shape of their own, known from rock inscriptions at Tang-e Sarvak and Shimbar and from the legends of Elymaean coins. It keeps the full 22-letter Aramaic alphabet, 𐿠 to 𐿵, plus one ligature, 𐿶 zayin-yodh. Unicode 12.0 (2019) encoded it with 23 characters. For the alphabet all of these grew from, and for Hebrew and Syriac, which kept the same 22 letters, see Aramaic & Ancient Arabian Scripts, Hebrew Letters & Symbols and Syriac Alphabet & Letters.

How to type Avestan, Pahlavi and Sogdian, fonts and pitfalls

None of these scripts has a standard keyboard, and most people copy from a chart like this one. Tap a letter in the grid to copy it, or select a word from the text. To type by code point, Avestan runs from U+10B00 (𐬀) to U+10B3F, Inscriptional Parthian from U+10B40, Inscriptional Pahlavi from U+10B60, Psalter Pahlavi from U+10B80, Manichaean from U+10AC0, Old Sogdian from U+10F00, Sogdian from U+10F30, Old Uyghur from U+10F70, Chorasmian from U+10FB0 and Elymaic from U+10FE0; in Word type the hex and press Alt+X, on Linux Ctrl+Shift+U then the hex, on a Mac enable the Unicode Hex Input source. All ten are right-to-left scripts, so type words in reading order and let the text engine lay them out; the combos below are typed that way.

Fonts are the real obstacle. Every one of these scripts has a free Noto font from Google: Noto Sans Avestan, Noto Sans Inscriptional Parthian, Noto Sans Inscriptional Pahlavi, Noto Sans Psalter Pahlavi, Noto Sans Manichaean, Noto Sans Old Sogdian, Noto Sans Sogdian, Noto Sans Chorasmian, Noto Sans Elymaic and Noto Sans Old Uyghur, and nothing else is guaranteed. Windows 11 covers Avestan, Old Sogdian, Sogdian and Old Uyghur through its optional Sans Serif Collection; macOS, iOS and Android ship no fonts for these blocks, so on a phone a reader will often see boxes. If you post one of these letters, add the transliteration so the text still means something where the font is missing.

Pitfalls to know: Book Pahlavi is not encoded, so anything claiming to be 'Book Pahlavi Unicode' is either images or a misuse of the inscriptional letters. Manichaean, Sogdian and Old Uyghur letters join and change shape inside a word, so the isolated form in the grid is not always what appears in running text, and the Sogdian and Old Uyghur combining dots must follow a base letter or they float. In Old Sogdian the final forms 𐼁 𐼃 𐼆 𐼏 𐼖 𐼛 are separate characters that you choose yourself; nothing substitutes them for you. The Avestan word separator is ⸱ (U+2E31), not the Latin middle dot ·, and the Avestan letters 𐬀 a and 𐬁 ā, 𐬌 i and 𐬍 ī differ by a single stroke, so proofread a name before it goes on skin or stone.