Tabrizi Jali typeface:
reviving history through letterforms
by Borna Izadpanah

Tabrizi Jali is the first type revival based on the early Persian Naskh types of Iranian presses from the nineteenth century. It is a product of several years of extensive research into the history of printing and typefounding, and practical experiments. Thus, it is a great pleasure to announce its release on the first day of spring and the Persian New Year (Nowruz), the 20th of March, 2023. A year that also marks 200 years since the publication of the earliest known Persian book in Tehran.

Although the gloomy and dreadful events of the last months in Iran do not provoke a positive atmosphere of cheerfulness and joy, the courage and determination of young, educated, and politically astute generations provide strong reasons to remain hopeful for a better tomorrow. In accordance with that spirit and to express my support and solidarity with women-led protest movements in Iran, I am pleased to offer a free license to individuals to use Tabrizi Jali typeface.

Woman, Life, Freedom



Arabic script printing with movable type was introduced to Iran in the 1810s. [1] The first known Persian book – written in modified Arabic script – was printed in 1818 in the northwest city of Tabriz. [2] Within only three years since the inception of printing in that city, at least three founts of type in the Naskh style were produced. The best specimens of type-making and printing from this period are displayed in types and publications attributed to a highly skilled and ingenious local craftsman Mirza Zayn-ol-Abedin Tabrizi. He was responsible for producing at least two founts of type in Tabriz and two types in the capital city of Tehran, where he printed the first Persian book in 1823. [3]

Fig. 1. Persian Naskh types of Mirza Zayn-ol-Abedin Tabrizi with their approximate date and place of production.

During this early phase, all metal types produced in these cities were in the Persian flavour of the Naskh style, which, apart from the more popular Nasta’liq style, was the most used writing style in manuscripts at the time. The choice of the Naskh style as a typographic model in Iran could also be explained by the genres of early Iranian publications, most of which had religious subjects. Naskh was also a more versatile style for representing various text genres in different languages, particularly the religious and Arabic language texts, which were mostly written in that style.

Fig. 2. A line from the Persian book Zad al-Ma’ad in Persian Naskh style from an 1819 manuscript (top), an 1836 lithograph edition printed in Tabriz (middle) and an 1829 typographic edition printed by Mirza Zayn-ol-Abedin in Tehran (bottom). (Personal collection)

The most accomplished and refined type-making project of Mirza Zayn-ol-Abedin is the second Persian Naskh type he produced in Tehran to be used exclusively for printing the Qur’an. Completed by 1827, this was his largest Naskh type and was first used to print the earliest known Qur’an with movable type in a Muslim country. [4] Tabrizi Jali typeface was designed based on the impressions of Mirza Zayn-ol-Abedin’s type (henceforth TL1) in this very edition of the Qur’an, from which only two surviving copies have so far been identified.

Fig. 3. An excerpt from ‘Surah al-Baqarah’ in the 1827 Tehran edition of the Qur’an printed by Mirza Zayn-ol-Abedin. Border width 14.5 cm (Personal collection)

Design Process: the original type and its revival

To capture high-quality and accurate images of the printed impressions of TL1 in the 1827 Qur’an, I followed the digitisation method devised by my esteemed colleague Riccardo Olocco. [5] However, since the notion of the baseline is less evident in the Arabic script (especially in cascading compositions of letters) compared to the Latin script I had to put extra efforts into capturing precise images. An additional challenge was imposed by some pages’ poor condition, making it difficult to produce focused and sharp images. After some experiments, I conducted proper assessments and produced accurate and consistent photographs of printed impressions. Next, I surveyed and selected the best available impression of each type sort from several specimens on different pages.

Fig. 4. Riccardo Olocco’s method of using a compact camera on top of a measuring magnifier which I used to produce enlarged images of printed impressions of the TL1 type.

The character set of TL1 is cleverly planned, allowing close imitation of the written forms of the Persian Naskh style without the typical distortions and oddities evident in many earlier Arabic script types – particularly those produced in Europe. Another quality that sets this type apart from many earlier examples is that it consists mostly of single letterforms, individual components (partial letterforms), and only a few conjuncts (type sorts consisting of two or more characters). [6]

Despite the fairly horizontally aligned feel of TL1 – compared to manuscript models – its printed impressions show a close accordance with Rasm al-Khatt [script rules] of the Persian Naskh style. This fount of type retains the ‘cascading’ notion of the writing style (particularly evident in different forms of jim, tcheh, hah, khah letters in the medial and final positions), but in some cases, the cascade is reduced by giving preference to more horizontally aligned variants of letters like mim in their medial and final positions. Thus, it successfully achieves some degree of typographic simplification without needlessly distorting letterform proportions and breaking the script rules.

Fig. 5. This illustration shows the composition of the Arabic word Sahih [correct] with Mirza Zayn-ol-Abedin’s TL1 type (right) and the Naskh type of the first complete printed Persian translation of the New Testament in St Petersburg in 1815 (left). This comparison demonstrates that Zayn-ol-Abedin’s more economical approach (using 4 pieces of type) produces a more accurate rendition of the word according to the script rules, whereas the 6 pieces of type used in the New Testament not only do not maintain the cascading notion of the Naskh style but also ‘incorrectly’ adds two extra ‘teeth’ characters (2 & 4) to the word ‘correct’.

As shown in figure 3, most connecting strokes of TL1 are smooth and properly joined, leaving only a few visible gaps between the characters. However, minor breaks are spotted when carefully scrutinising the impressions and close-up images. These breaks mostly occur when certain contextual alternates are recycled in different instances. For example, in the variants of letters like lam, teh and feh that are connected to the final forms of yeh. (fig. 6) Though this is often disguised by ink spread, typical ‘imperfections’ of letterpress impressions, and the texture of the paper.

Fig. 6. See the gap and inaccuracies in these connections of the initial form of letters lam, teh and feh (from left to right) to the final form of the letter yeh.

An important advantage of making such hardly noticeable compromises is in producing a more economical type system both from the point of view of production costs and typesetting speed and accuracy. In the digital revival, however, reusing alternates would pose an issue since digital typefaces are used in various sizes, making the gaps and defective connections immediately evident. Therefore, designing a few more contextual alternates was one of the few necessary modifications I decided to make in the revival.

Fig. 7. Same conjuncts in figure 6 set in Tabrizi Jali typeface.

I redrew this type three times. In the first round, I added several more contextual and stylistic variants and tried to make the stroke modulations and proportions more consistent as an ‘improvement’ of the original design. However, I decided this approach was anachronistic since I was imposing contemporary ideas of aesthetics and consistency onto a historical model. In the second attempt, I tried to remain closer to the original proportions but compared the type forms with manuscript models from the same period to better inform my design decisions. However, I decided to abandon this approach due to the historical significance of the original design and the necessity of retaining its characteristics. As I continued to scrutinise and assess the exceptional qualities of TL1, I developed a better appreciation for the type-makers ‘design’ decisions and even some of the ‘unusual’ characteristics of the type. This led me to view my revival as a tribute to original design rather than an opportunity to merely showcase my self-proclaimed drawing and design skills.

Fig. 8. The TL1 impressions (top row). Early sketches where I tried to give the characters more consistent stroke modulation and proportion (middle row). The final design follows the original metal-type impressions more faithfully (bottom row).

In the final version, I only made minor additions and modifications when I felt it was necessary. For example, I included the modern form of the Persian letter gaf (hard g), which was not commonly used in that period. I designed more uniform and larger numerals since the numerals of TL1 are fairly small and slightly irregular and only used for numbering folios and the publication date in the colophon.

Fig. 9. Examples of TL1’s figures used for numbering folios.

I also included commonly used punctuations which did not exist in TL1 in accordance with the manuscript conventions of the time. Word spacing is another feature available in the revival since modern readers have grown accustomed to it, but it did not exist then. Users can still follow the historical models and use the zero-width non-joiner (ZWNJ) to avoid word spacing. [7]

Fig. 10. The same text with and without word spacing.
Top: two lines of Persian Naskh text by Mohammad Sadeq Golpayegani (Shams-ol-kottab) in a lithograph publication. Middle: Same text in Tabrizi Jali without word spacing. Bottom: Same text in Tabrizi Jali with word spacing. Also, note the different composition of some characters and words in the handwritten specimen and Tabrizi Jali typeface, which follows the TL1 type.

The Tabrizi Jali typeface includes all the vowel marks and swash forms – which are beautifully kerned in TL1. However, mark (and dot) positioning and kerning of swash forms displayed in printed impressions of TL1 are extremely challenging to imitate accurately. Especially since there seem to be many instances where manual adjustments and alterations are made in typesetting relative to specific instances and letterforms compositions. Therefore, to achieve the same result when typesetting with Tabrizi Jali, the kerning of swash forms and possible clashing of vowel marks should be modified manually. However, I put much effort into fitting the typeface and contextual kerning to achieve the best result that my time and technical skills allowed. Finally, the same vertical and horizontal arrangements of the two dots in TL1 are incorporated in my revival.

Future Development Plans and Potentials

For this release, I kept adding extra characters to a minimum to remain more faithful to TL1. The current glyph set of Tabrizi Jali covers the required characters of typesetting in Arabic and Persian languages. In future versions of the typeface, I intend to expand the language support, but due to the extremely tight fitting and kerning in this typeface, supporting languages like Urdu is a highly demanding task. Especially due to the additional work required to correctly represent characters like Bari Ye and the limitations of the current text-shaping technologies. It is also important to find and select suitable stylistic models for additional letterforms in different languages, which requires serious research and several rounds of testing.

Tabrizi Jali is the first member of a typeface family, which will ultimately include other Persian Naskh types of Mirza Zayn-ol-Abedin produced in Tabriz and Tehran. Tabrizi Jali can be viewed as the ‘display’ style of the Tabrizi type family, and other type styles will include a ‘matn’ (text), ‘ketabat’ (book) and ‘khafi’ (caption) style. After completing this typeface family, I intend to design a more elaborate and accurate typographic rendition of the Persian Naskh style based on the carefully compiled manuscript specimens and particularly a unique document from 1907 by one of the most renowned and influential Iranian calligraphers of the late Qajar and Early Pahlavi period, Mohammad-Hoseyn Sayfi Qazvini, better known as Emad-ol-Kottab (1866–1936).

Fig. 11. A spread from the recently discovered handwritten foundry-type character set devised by Emad-ol-Kottab. Evidently, this document was produced to serve as the basis of a new Persian Naskh type consisting of 412 characters which probably never materialised. Image courtesy of Dr Hamidreza Ghelichkhani (Private collection in Tehran)

One of the significant outputs of working on historical revivals and research-informed typeface designs like Tabrizi Jali is reintroducing highly evolved and elaborated textual conventions in various manuscript traditions. For instance, the mixed-use of writing styles to emphasise, highlight or contrast a section of the text. The following example shows the use of Naskh and Nasta’liq styles to differentiate the text in different languages. This provides a model for suitable and functional use of secondary styles instead of the out-of-context but commonly used ‘slanted Arabic’ or ‘italic Arabic’ that is imposed by Latin script typographic conventions.

Fig. 12. Top: detail from a nineteenth-century lithograph publication shows the effective use of the Naskh style to highlight embedded Arabic sections in a Persian text written in a more fluid Nasta’liq style. Bottom: Tabrizi Jali typeface and my underdevelopment Nasta’liq typeface (based on lithograph book from Delhi) used to create a similar text configuration seen in the lithograph example above.

The Naskh types of Mirza Zayn-ol-Abedin, particularly TL1, are the benchmark of Persian Naskh and Iranian typography. Tabrizi Jali is an attempt to reintroduce a genre-defining model which can be used as a reference for researchers, designers and students, similar to the Roman types of master type-makers such as Nicolas Jenson, Claude Garamond or John Baskerville, which play such roles in Latin typography.


Download Tabrizi Jali


[1] For a more detailed account of nineteenth-century printing and type-making in Iran see my ‘Persian and Arabic printing with movable type in Qajar Iran (1818–1900),’ in Arabic Typography: History and Practice, niggli, 2023, 75–153. [^]

[2] Resaleh-ye Jahadiyeh, written by Isa ebn Hasan (Mirza Bozorg) Qa’em Maqam Farahani and printed in 1818 by Mohammad Ali ebn Haji Mohammad Hossein Ashtiyani in Tabriz. [^]

[3] Mohreq-ol-Qolub, written by Mohammad Mahdi ebn Abi-Zarr Naraqi and printed in 1823 by Mirza Zayn-ol-Abedin Tabrizi in Tehran. [^]

[4] Until the 1850s several editions of the Qur’an were printed with this type in Tehran. [^]

[5] Design Type Revivals [^]

[6] The term ‘ligature’ is frequently but wrongly used to describe such conjuncts in the context of the Arabic script since unlike in Latin script the fuses of two or more characters in a single type sort (or glyph) in Arabic are in most cases not optional and required by script rule. [^]

[7] This can be done by typing shift+space in Persian keyboards. [^]

previous: ILT’s Favorite Fonts of 2022

next: Hot New Fonts: spring 2023