Ancient writing systems in the Mediterranean

A critical guide to electronic resources


- 1st century CE - present

edited by: Emiliano Fiori (translation revised by Melanie Rockenhaus)

  • Introduction
  • Index
  • Further information

The earliest (non-literary) written examples of Syriac can be traced back to the 1st century CE. It is an alphabetic script, comprising 22 graphemes; it is written leftwards. Throughout the centuries, the Syriac script diversified into a number of variants: the Estrangelo, the first to be found in manuscripts, culminating between the 7th and the 8th centuries, which as can be seen in the image to the left, reached China; the Serto, which is nowadays the most common kind of Syriac script, deriving from the ancient cursive form already in use in the first centuries; and the Chaldean or Nestorian (now rather defined “oriental”), a regional variant of Estrangelo, used in Eastern areas. Each of these scripts can be also used to write Arabic: in this case, they are called Karshuni or Garshuni script.

Go to the online resources.

Online resources


Hypotheses around the origins of the writing system

The origin of the Syriac script has long been a debated question. The most balanced presentation of the current scholarly consensus can be read in the crucial monograph by Healey and Drijvers 1999 on Syriac inscriptions and in the subsequent update by Healey 2000. 

A first hypothesis on the origins of the script linked it to the so-called “cursive Palmyrene” of the inscriptions found in the oasis of Palmyra (3rd century BCE-1st century CE). This script seemed to be very close to the earliest forms of Syriac. But the very existence of a “cursive Palmyrene” was challenged by Pirenne 1963. The solution proposed by the French scholar pointed towards the mountains between Turkey and Iran. The Syriac script has indeed a greater number of characters in common with the Aramaic scripts found on the northern course of the Tigris river, than with the Palmyrene script (but not with an inexistent “cursive Palmyrene”). In Pirenne’s opinion, it would be a compound of both the Aramaic and of the Palmyrene scripts, with a strongly original character: i.e. Syriac would have deviated from a line which it shared with Palmyrene due to the influence of the script of Northern Mesopotamia. According to the hypothesis which Joseph Naveh put forward twenty years later, the question should be solved by postulating the existence of a common origin of the two scripts (Syriac and Palmyrene), the “Seleucid Aramaic” (Naveh 1982, p. 149): this would have been used in the Seleucid kingdom of Syria. It is not found during the most ancient period of the kingdom, but in four inscriptions of the 1st and 2nd centuries CE, where an alphabet can be read which seems to contain graphemes which are now closer to Syriac, now to proper Palmyrene. Precisely because of this hybrid character, Naveh supposed that it may be identified as the common source of both scripts.

Earliest examples of a distinctly Syriac writing system

A proper Syriac alphabet, unmistakable with other scripts, appears in the first Syriac inscription known: it is written on stone and was found in Birecik on the Euphrates, 75 km west of Edessa. It is dated to 6 CE. The second most ancient datable inscription is of 73 CE, and comes from Serrin (south-west of Edessa). The characters of these inscriptions, as well as those of other very early inscriptions, cannot be identified with true Estrangelo (so-called after a Syriac word which means “round”), but it is commonly stated that they represent a primitive state of it. On a whole, Syriac stone inscriptions are spread over the whole history of this language, and have been found as far away as Central Asia and China.

A number of mosaics containing Syriac scripts have come down to us, these are entirely from the first three centuries CE.

The discovery of three early fragments on parchment has also been of great relevance for the study of the evolution of the Syriac script. A first one was found in Dura Europos in 1933 and two further fragments were found in the region of the middle course of Euphrates at the end of the 1980s. They are respectively dated to 243, 240 and 242. As will be seen below, these texts are mainly important for the identification of the early existence of a cursive variant of the square script. A regular calligraphic Estrangelo, similar to that which is still printed nowadays, appears with the first dated Syriac manuscript (411 CE).

A variety of writing systems

It would be simplistic to state that Estrangelo was the original Syriac script, from which the others later stemmed: it is possible to find a remarkable variety within the Estrangelo script itself, for example in texts on parchment of the 3rd century. Moller 1998 has thus rightly opposed the view according to which the script of the first of these fragments, coming form Dura Europos and described by Torrey 1935 as presenting a primitive form of Estrangelo, can be properly defined as Estrangelo, since it contains a style which is already highly cursive. By studying the evolution of the script throughout the first written examples, especially those in mosaics, Drijvers and Healey 1998 conclude that Syriac had had a cursive variant since its origins; this was mostly used for practical purposes (and indeed the document from Dura Europos is a bill of sale). It is true that, as Healey 2000, p. 63 reminds us, the variety of scripts also depends on the different quality of the materials on which Syriac was written. The stone inscriptions usually present a style of Estrangelo which resembles the one later found in the first complete dated manuscript (411 CE); the mosaics (until the 3rd century) and the parchment fragments, on the contrary, allowed a more fluid script by their own nature. The first manuscripts, for the same reason, manifested cursive variants from the beginning, though written in Estrangelo. The cursive variant always went alongside the square script which evolved into Estrangelo, occasionally appearing since the manuscript of 411, and also linking documents far away from one another in time with surprising affinity: for instance, the cursive of a 6th century manuscript can be very similar to that of a parchment of the 3rd century. It is precisely from this cursive variant that the Serto script developed, Serto being a formalized and calligraphic version of the ancient variant; its first proper evidence in manuscripts can be traced back, according to Hatch 2002, to the 8th century.

Hatch dates the rise of the oriental or “Nestorian” script to about the beginning of the 7th century. This script is commonly regarded as a regional variant of Estrangelo, to which it is actually akin.

A vast geographic spread

A typical peculiarity of the Syriac language and scripts is their wide dissemination over the whole Asian continent. Though born as Mesopotamic and Mediterranean script, by the end of the 8th century Syriac had reached China and Tibet thanks to the missionary activity of the Eastern Syriac Church, in particular of the patriarch Timotheus I (780-823). This is mainly demonstrated by a large number of central Asian inscriptions and by a small but relevant group of Chinese documents, especially the very famous stele of Xi'an Fu (8th century). It must also be mentioned that Estrangelo influenced the form of the script of other languages, as e.g. Sogdian, a northern middle-Iranian language originating from the valley of the Zarafshan river (today in Uzbekistan and Tajikistan) but widespread over all the Silk Road: during some centuries, it was actually the latter’s lingua franca. Through Sogdian, the Estrangelo script also influenced two important Altaic languages: Uighur and Mongolian.

Diacritical signs

Another major characteristic of the Syriac script is the great variety of diacritical signs, which are used for a wide range of functions. For a detailed illustration of the vocalization system, the tables of symbols can be consulted; here we provide a synthetic description of the typology and the evolution of the other diacritical signs, mostly relying on Segal 1953. Segal himself (p. 5) divides these symbols into four categories: 1) the plural sign, 2) the sign which distinguishes dolat from rish, 3) the proper diacritic and 4) the accent. We will not dwell on the latter, which is necessary for the liturgical performance of the sacred texts, since it is too complex a topic to fit an introductory account.

1) and 2) The most ancient diacritical signs are beyond any doubt that of plural (two dots, usually horizontally lying above the letter) and the dots (respectively above and below) which distinguish dolat from rish.

3) The first example of diacritical points is already in the first dated manuscript (411 CE). Segal has suggested dividing the development of the diacritical signs and the signs of vocalization into periods according to the following distribution:

-a first “common” period, until ca. 600 CE, when all the basic points appear which distinguish homograph words of different meaning from one another (a wide phenomenon in Syriac when it is written without vowels – and a very common feature of many Semitic languages). This is indeed the original function of the diacritical signs in Syriac. 

-In the second period (7th-12th centuries) a distinction between Eastern-Syrian and Western-Syrian diacritical systems appears. The most relevant feature of this period is the appearance within the Eastern script of a system of dots which is exclusively intended for vocalization. An exam of the extant manuscripts shows the evolution towards a complex system of dot-vowels was gradual and is to be seen as the result of a generalization and a specialization of the distinctive function of the diacritical points which was proper to the first period. Since the 11th century, then, it is possible to notice a systematization of this evolution by Syriac grammarians, who acknowledge the actual situation of their day.

In the Western-Syrian context a different vocalization system evolved, which is the most universally known today: the vocalization by means of Greek letters. Already in the 7th century the polymath Jacob of Edessa had (unsuccessfully) suggested a system of vowels shaped on the Greek ones, and imagined that they would be written in the very body of the word. This would put vowels on the same graphic level of consonants. It is not possible to know why this highly Hellenophile project had no success (not even a single extant manuscript, except for those containing his grammar, displays a positive assumption of his proposal). His intuition was anyhow right in the sense that a vocalization system by means of Greek vowels became prominent later on in a Western Syrian context. It is not clear when this system became predominant, since the first datable manuscripts to use it are of the 13th century; it is nonetheless sure that it was established before the 11th century, since the first Syriac grammarians, in that century, treat it as a well-established system. But such a system, which comprised five signs only, was less complete than the Eastern Syrian one, for in only one case did it take into account the different quality (opening degree) of vowels.

In the 13th century, the two syntheses by the Western Syrian intellectuals Severus bar Shakko and Bar Hebraeus are still worth mentioning: in their grammars of the Syriac language, a remarkable space is devoted to diacritical signs and to the comparative study of the Eastern and Western systems.

A digression: Palmyrene

It has been remarked that the earliest forms of Syriac script have a number of features in common with the Palmyrene script. Palmyrene was a western Aramaic dialect whose script is attested in the first three centuries CE, not only in the eponymous Syrian oasis but also in the farthest areas of the Roman world. Isolated epigraphic examples have been found as far as England, Egypt, Algeria, Italy, Hungary and Romania. The first datable inscription is of 44 BCE, the last of not much before 272, when Palmyra was destroyed by the Romans.

The majority of Palmyrene inscriptions are carved on stone, are of a votive or funerary nature and are written in the so-called monumental Palmyrene, which is mainly akin to square Hebrew. What is most interesting now is the variant which some have called “cursive” (Chabot 1922), which can be observed in mural graffiti at Dura Europos, Palmyra and in the northern region of Palmyra: this variant shows indeed a remarkable affinity with Syriac. It is moreover used in the inscriptions found in Europe. It is also a late variant, which is not attested before the 2nd century CE; it was likely used on perishable materials like papyrus or parchment, so that no examples thereof have come down to us. But Pirenne 1963, p. 118-119, opposes the view that this script is actually a “cursive” script, since it is also found in monuments; according to her, it cannot even be properly defined as Palmyrene, since its peculiarity depends on the fact that it is contaminated by extraneous elements.

Examples have also been found of a script which is not possible to identify as either Syriac or Palmyrene. These are in four inscriptions: the inscription of Amassamses from Deir Ya'qub, close to Edessa (2nd century CE), the short inscription which was found on the sarcophagus of “Queen Sadan” (maybe a queen of Adiabene, 1st century CE), the longer inscription of  Dura Europos (32 BCE) and the one discovered in el-Mal in southern Syria (7-6 BCE). It has already been noted that Pirenne 1963 and Naveh 1982 interpreted these data: the latter by calling this hybrid script “Seleucid Aramaic” and by identifying it with a common ancestor of Syriac and “cursive Palmyrene”, used by the Arameans of the Seleucid kingdom; Naveh thus accepted the existence of a cursive variant of Palmyrene. Pirenne, on the contrary, as we have seen, did not admit it, but she anticipated Naveh’s conclusion by stating that this compound, called by her “Syro-Palmyrene” and Syriac should have a common source. She thus explained the divergence of Syriac from this common origin by means of the strong influence of the scripts used along the northern course of the Tigris, such as those of the inscriptions of  Assur, Hatra and the Tur 'Abdin (2nd century BCE-2nd century CE). Syriac is indeed more similar to these scripts than to Palmyrene.