In and around the oasis of ancient Dadanâmodern-day al-Ê¿UlÄâin northwest Arabia, many rock surfaces can be found inscribed with texts written in the local script, called Dadanitic. Dadanitic falls under the umbrella of Ancient North Arabian (ANA) scripts,1 a group that also includes Taymanitic, Dumatic, Dispersed Oasis North Arabian (ONA),2 Safaitic, Hismaic, and Thamudic3 (Macdonald 2000, 32â37). M.C.A. Macdonald distinguishes ONA scripts, including Dadanitic,4 from the scripts that were used by nomadic populations. The ANA scripts are part of the South Semitic script family and are attested from the northern borders of Yemen to the southern Levant. They are usually negatively defined as those scripts that are not Ancient South Arabian (ASA). It has been hypothesized that the ANA scripts form one group, descending parallel to ASA from a putative proto-South Semitic script. However, a paleographic connection between all ANA scripts has yet to be demonstrated (Al-Jallad 2015, 10).
The corpus of Dadanitic inscriptions currently consists of about 2000 examples that are generally assumed to have been written between the sixth and first centuries BCE. While most are short graffiti, often not mentioning more than a personal name, there are also hundreds of texts dedicated to the local deity, á¸Å«Ä¡Äbat. Although most of the known ANA inscriptions are classified as graffiti, Dadanitic stands out as it is the only ANA variety in which we have clear attestations of commissioned texts (Al-Jallad 2018b, 2; and see Macdonald 2010a, 7 on craftsmanship and the use of writing in ancient society).
The skillfully carved relief style in which the inscription in Figure 1 is executed is, therefore, a typically Dadanitic feature within the ANA corpus. Other typical Dadanitic features of this inscription are its formulae and content. The inscription commemorates the performance of several rituals for the main deity of the oasis, á¸Å«Ä¡Äbat (á¸Ä¡bt in the inscription). One of its most striking features, in light of the present work, is its peculiar linguistic form: the inscription contains two causative verbs, each exhibiting a different morphological form (indicated in boldface in the transcription).
Figure 1
A dedicatory inscription in relief (AHÂ 197)
Photograph available on OCIANA
| AHÂ 197 (5â7) |
ḥggw / h-nq / w hÄ¡nyw / b-bt-hm / l- ----//tn / l-ḫrg / w-ʾáºlw / b-h-má¹£d /áºll / h-[nq] // l-á¸Ä¡bt âthey performed the pilgrimage and dedicated (lit. made increase wealth) at their temple for ⦠tn for ḫrg and they performed the áºll of the [nq] for á¸Ä¡btâ |
The linguistic variation attested in the Dadanitic inscriptions was already noted in the earliest publications of these texts (e.g., D.H. Müller 1889, 13â14; Grimme 1937, 300). Despite this early recognition, the phenomenon has not yet received any attention in its own right and no comprehensive explanation for the variation has been formulated to date. The grammar of the inscriptions continues to be of interest, however, with several descriptions of the grammar of the inscriptions published, usually accompanying an edition of texts. The most recent descriptions of Dadanitic are those of A. Sima (1999) and S. Farès-Drappeau (2005). However, while Simaâs (1999) description is very thorough, it only focuses on the formal inscriptions from the al-Ê¿Uá¸ayb area and does not deal with inscriptions from other areas or with graffiti, and although Farès-Drappeauâs (2005) work contains a grammatical sketch based on the whole corpus, she takes a very CAr.-oriented approach in her description.5 In this work, I aim to remedy the situation by approaching the language of the inscriptions on their own terms, offering both a more traditional description and a quantitative analysis of their linguistic variation.
Through this analysis, we will not only gain a better understanding of the linguistic variation in the Dadanitic inscriptions and its possible underlying causes, it will also improve our understanding of the linguistic situation at the oasis when these texts were composed. In addition, this study will shed light on the oasisâs scribal culture and the perennial question of literacy. While Macdonald has extensively discussed the interaction between writing material and variation in letter shapes (2015, 2010), this book will be the first study to systematically consider the interaction between writing practices and linguistic form.
1 Outline of the Present Work
This work is divided into eight chapters, preceded by an introduction that outlines the corpus and offers a methodological discussion on the use of epigraphic material to answer questions about its language and variation. Chapter 1 will provide a more in-depth discussion of the historical context of the Dadanitic inscriptions. It will focus on the dating of the corpus, with attention given to the different methods and types of evidence that have been used to try and establish a (relative) chronology of the Dadanitic inscriptions. Chapters 2 through 6 contain a description and grammatical analysis of the corpus, thereby helping to contextualize the variation discussed in Chapters 7 and 8 by giving an overview of the common and less common forms of the grammar, but also of the formulaic parts and writing styles. Chapter 2 discusses the script and different styles of inscribing used to carve the inscriptions. Chapter 3 contains an overview of the different genres that can be distinguished and the compositional formulae associated with them. Chapters 4 through 6 offer a description of the grammar of the Dadanitic inscriptions; specifically, Chapter 4 deals with the orthography and phonology of the inscriptions, Chapter 5 contains a description of its verbal morphology, and Chapter 6 addresses nominal and pronominal morphology.
Chapters 7 and 8 deal exclusively with the variation attested in the corpus, building on the description in the chapters that precede them. They offer a quantitative analysis of the variation in the corpus in an attempt to move beyond impressionistic accounts of its distribution and possible causes. Given the different methodological approach of Chapters 7 and 8 compared to the preceding chapters, Chapter 7 will will start with a methodological introduction, discussing in detail the statistical methods used to determine the interaction between different variables and how each variable was chosen. The aim of this analysis is to reveal patterns of co-occurrence between different varying features, which could offer insights into the reasons behind these variations.
For this method, the number of co-occurrences of two features within one inscription are added up and compared to the number of co-occurrences that would be expected to occur if there was no relation between the two features (i.e., if their distribution across the corpus was completely random and independent of each other). The null-hypothesis is that there is no relationship between the compared groups. The further the results of both calculations are apart, the less likely it is that the null hypothesis is true. If the chance of the given outcome occurring by chance is smaller than 5â¯%, the result is found to be significant, meaning it is unlikely to have happened by chance if the null hypothesis were true and there was no relationship between the two features. Chapter 7 will begin with an explanation of the statistical method used for the analysis, followed by a description of the variables included in the analysis. Chapter 8 will present the quantitative data, followed by a discussion in which the numbers will be interpreted. The chapter ends with a short conclusion summarizing the results.
2 The Corpus
As noted above, most Dadanitic inscriptions are found in and around the ancient oasis of Dadan, modern-day al-Ê¿UlÄ.6 The first western traveler said to have visited al-Ê¿UlÄ and MadÄʾin á¹¢Äliḥ (the Nabataean site ḤegrÄ, some 20â¯km northeast of the oasis of al-Ê¿UlÄ), was C.M. Doughty, who published an account of his travels in 1888. This work includes a description of both places and sketches of their surroundings. He also copied some of the inscriptions, which he thought to be Sabaic, or âHimyaricâ (Doughty 1888, 160). These copies were published already in the volume Documents épigraphiques recueillis dans le nord de lâArabie (Doughty 1884), which provided the basis on which J. Halévy offered a first attempt at their decipherment (Halévy 1884). According to D.H. Müller (1889, 15â21), the first decipherment of the script benefited greatly from the resemblance it had to the already deciphered Sabaic script.
Shortly after these initial attempts, the first edition of Dadanitic texts was D.H. Müllerâs (1889) work on the inscriptions brought back from the region of al-Ê¿UlÄ by J. Euting. Müller termed the inscriptions âLihjanischâ (1889, 4) after the attestation of several kings that are called mlk lḥyn âking of LiḥyÄnâ in the corpus (1889, 5). Müllerâs work includes a script table with his reading of the attested glyphs (1889, pl. X), and yet, despite these efforts, the reading of several letter shapes was amended in later works (Grimme 1926, 1932; Winnett 1937). The next step in the study of the Dadanitic inscriptions came when a large number of additional Dadanitic inscriptions became available after the publication of those collected by A. Jaussen and M. Savignac (1909). This expansion of the available corpus was of great importance to their further decipherment. Later, major contributions to the corpus were made by R. Stiehl (1971) and H. AbÅ« ʾl-Ḥasan (1997, 2002).7
Map 1
Showing the location of Dadan
Courtesy of Rohmer and Charloux (2015)
Another key moment in the study of the inscriptions came in the 1930s, with a change in terminology. H. Grimme (1932, 754) was the first to propose dividing the Dadanitic corpus into a Dadanite and LiḥyÄnite script type, using the term Dadanite to refer to the inscriptions Jaussen and Savignac had initially termed âold Thamudicâ (for more on the paleographic distinction between Dadanite and LiḥyÄnite see §â¯4 in Chapter 1 below). These labels referred to the two consecutive kingdoms that are mentioned in the inscriptions: the kingdom of Dadan and the kingdom of LiḥyÄn. However, rather than following Grimme, I have adopted the name âDadaniticâ to refer to the complete corpus by its geographical location. This usage is relatively recent and follows Macdonald (2000, 33), who replaced former âDadaniteâ and âLiḥyÄniteâ, arguing that, since we cannot make a clear distinction between two different types of script (see Chapter 1, §â¯4), dividing the corpus into two would not accurately reflect its variation. Moreover, even if a distinction could be made, we cannot directly attribute the change in script to a change in political power. Thus, connecting any phase of the script to a specific regime risks blurring our understanding of the history of the oasis (Macdonald 2000, 33).
Moving on to the current study, at the time the data for this study was collected and annotated, the Dadanitic corpus consisted of 1969 inscriptions, which are all brought together in the searchable online OCIANA database.8 The database was developed at the Khalili Research Center at the University of Oxford under the supervision of M.C.A. Macdonald and J. Johns.9 It includes all currently published and even some unpublished Dadanitic inscriptions. Each inscription has its own record containing available images, a transcription and translation when possible, and a complete bibliography.
The images provided on the OCIANA website have formed the basis for my analysis of the manners of inscribing, as outlined in Chapter 2. The reading of all inscriptions relevant to the analysis of the grammar or orthography was cross checked with the available photographs. Whenever the proposed reading or interpretation of an inscription relevant for my analysis does not follow the one given in OCIANA this will be mentioned in the discussion. In the glossary all readings that deviate from the interpretation in OCIANA are marked with a single asterisk when they are mine, otherwise they are followed by the reference to the publication in which they were proposed (see the glossary in the Appendix).
3 A Holistic Approach to the Epigraphic Object
While the principal concern of this work is the language of the Dadanitic inscriptions and the linguistic variation found within them, this cannot be studied without taking into account the character of the Dadanitic texts. In 1998, Macdonald, following P.A. Février (1989), emphasized the importance of studying the language and content of epigraphic material in the context of its physical form, location, and cultural/historical setting. All these elements contain information on the meaning and significance of a particular text and need to be taken into account in order to even begin to understand the already fragmentary epigraphic record. This approach finds strong parallels in the so-called New Philology or Material Philology, which was first promoted as such in a special issue of Speculum in 1990 (Nichols 1990), but clearly built on scholarship from previous decades, such as that of P. Zumthor (1972), D.F. McKenzie (1986) and, more directly, B. Cerquiglini (1989). The main overlap with the approach advocated by Macdonald (1998) is New Philologyâs focus on the materiality of a manuscript, how it was produced, how it was interacted with, and, importantly, how both the text and the object relate to their historical context (Driscoll 2010).
Therefore, the study of the language of the Dadanitic texts begins with the objects themselves. The features that make a text identifiable as Dadanitic are its script and to some extent its language, its location, and the particular genres and formulae associated with the Dadanitic writing culture. When considering the Dadanitic inscriptions as the product of local writing culture, for example, one needs to reckon with the fact that official inscriptions, which are not graffiti, do not represent natural speech but a written code, something which is often overlooked in epigraphic studies.10 The formalization of language for such purposes will have a standardizing effect that creates some distance between the spoken and written registers.
This chapter will discuss the foundational elements to a holistic approach to the epigraph, contextualizing the discussions surrounding script, genre, and the language of the inscriptions. Such an approach to the epigraphic object is the only way to move beyond simply describing its separate elements and to work towards an explanation for the seeming inconsistencies found in the inscriptions. Previous studies of the language of the inscriptions (e.g., Sima 1999; Farès-Drappeau 2005) have not been able to offer a comprehensive explanation for the variation attested in the inscriptions partly because they did not look beyond their transcribed text. Now that it has become clear that the explanation is not purely linguistic, it is understandable that the linguistic dimension alone does not contain the answers to the question of variation. For this reason, following the brief overview of the key varying features of the Dadanitic inscriptions, I will then turn to a discussion of the literary environment in which the inscriptions were produced, and the impact this has on how we should approach variation attested in them in more detail, before finally offering a consideration of graffiti and the spread of literacy.
To begin with the foundational elements, however, a brief overview of the key varying features of the Dadanitic inscriptions will be given. When studying variation in the Dadanitic corpus, there are three main elements of the inscriptions in which most variation can be found. These are the script, specifically the letter shapes and manner of inscribing; the genres of text that can be expressed and the different levels of formality that seem to have been associated with them; and finally, the language used to compose them. This section offers a treatment of each of these key elements of variation, ending with a thorough discussion of how to classify the language of the Dadanitic inscriptions and the impact surrounding cultures and languages it was in contact with, had on it.
3.1 Script
The Dadanitic corpus is primarily defined by its script, although the term is usually also employed to refer to the language of the Dadanitic inscriptions. The two do not always have to go together, as âany script can, of course, be used to express any language (more or less efficiently)â (Macdonald 2000, 37). A great example of this within the Dadanitic corpus is inscription AHUDÂ 1, which is written in the ASA script, which is associated with the Minaic inscriptions at the oasis, but is typically Dadanitic in its language and content and is, therefore, considered part of the Dadanitic corpus.
The following section will focus on the relationship between different manners of inscribing and the formality and genre of a text. The use of paleography to establish a relative chronology of the texts will be dealt with in more detail as part of the discussion on the dating of the inscriptions in Chapter 1. The script itself and the attested variation in letter shapes will be treated in more detail in Chapter 3.
3.1.1 Script Style and Socio-economic Status
Several different methods were used to inscribe the Dadanitic inscriptions, which seem to represent different levels of formality, and probably cost, associated with the execution of the inscription. While inscriptions in relief (as in Figure 1) are the most distinguishable, Dadanitic inscriptions were also deeply incised, chiseled, or pounded onto the rock (for a full discussion and examples of the different techniques see Chapter 2). The inscriptions using the first of these methods were deeply incised into a rock face or stone objects, while those executed in relief were made by skilled craftsmen who sometimes even signed their work (e.g., JSLih 082 and al-Ḫuraybah 06). These texts were probably commissioned (Macdonald 2010, 7). In contrast, chiseled and pounded inscriptions required less skill to produce. However, this does not mean that all pounded inscriptions are graffiti, as many of the dedicatory áºll inscriptions, which have an official character and were firmly entrenched in the writing culture of the oasis, were pounded onto the rock. In short, there clearly is not a one-to-one relationship between the purpose and content of a text and the level of execution of the inscription. That said, it is possible to see a general trend of using script registers that require less skill for graffiti while reserving more skilled manners of inscribing for more official inscriptions (see Chapter 2 and Chapter 7), despite the lack of a direct relationship between register and purpose.
3.1.2 Graffiti vs. Commissioned Texts
How, then, do we distinguish between graffiti and more official texts? Graffiti can generally be described as privately produced inscriptions or texts, made on private initiative, left in a public space (Macdonald 2015, 8; Milnor 2014, 5).11 This definition clearly distinguishes graffiti from commissioned texts, which were executed by a professional mason and their text probably drawn up by a professional scribe.12 This in turn has implications for the expected level of execution of the inscription, both in terms of its language and physical form.
In addition, the fact that graffiti are personal statements means that their content and formulations are in principle not bound to the same restrictions as the highly formulaic official inscriptions, even though they were clearly influenced by the structure and formulae of the official inscriptions. Therefore, graffiti often contain elements that do not occur in formal inscriptions, such as unique phrases or formulations (e.g., JaL 016 a) or references to the act of inscribing itself (e.g., AH 256). I will return to this point in Chapter 3, which offers a complete overview of commonly used phrases and the content of the inscriptions.
A final division between formal inscriptions and graffiti is supported by how the inscriptions are spread across the landscape. As can be seen in Map 2, the non-graffiti are clustered around a few specific places mostly close to the oasis itself. In contrast, graffiti seem much less bound to specific places in the landscape.13
Map 2
The approximate distribution of different types of inscriptions in the landscape
Map adapted from Rohmer and Charloux (2015)
3.2 Genre
Another point of variation in the Dadanitic corpus is the genre of the text, which is determined by its content and is closely associated with particular formulae. This closely follows the definition of genre as outlined in D. Biber and S. Conrad (2009, 2), who, when considering the genre of a text, take its purpose and situational context into account. They also include the conventional structures that are part of a specific variety of text in their analysis. In this work, the conventional structures will be referred to as a textâs formulaic parts. It should be noted, however, that in order to be able to say anything meaningful about the interaction between genre and linguistic forms attested in it, linguistic variables cannot be used to determine the genre of a text. Whether an inscription includes an h-causative or a ʾ-causative is not used to determine whether it should be considered a graffito or a dedicatory inscription, for example.14
While the general structure of the textsâstarting with personal names, followed by a dedication, and ending in a petition for protection from a deityâis comparable to that found in other ANA and Ancient South Arabian (ASA) corpora (Al-Jallad 2015, 201â221; Avanzini 2017, 97â98), the specific formulae, the áºll ritual, and the deity á¸Ä¡bt are unique to the Dadanitic inscriptions.15 The main distinction in genre is that between graffiti and more formal inscriptions, as mentioned above. Within the more formal inscriptions, several different types can be distinguished within the Dadanitic corpus, such as dedicatory, áºll, building, and legal inscriptions. Each genre and associated formulae will be discussed in detail in Chapter 3.
Using the category of genre as one of the variables in the quantitative analysis in Chapter 8 mostly functions as a control category for the hypothesis that the more complex carving techniques were reserved for more formal and expensive inscriptions. If this hypothesis is correct, the formality of the inscription might also have a relationship to the register of the language used in it. In other words, if genre and the manner of inscribing always have a similar relationship to the linguistic variants occurring in the inscriptions, this is likely owing to the same or a similar cause: most likely register.16 However, one can also imagine that specific genres came with their own formulae, possibly containing specific linguistic features as well. This is clearly demonstrated by the ná¹r inscriptions, for example. In Chapter 7, §â¯2.2 the relationship between genre and register will be treated in more detail.
3.3 Language and Linguistic Variation
Finally, there is significant linguistic variation in the Dadanitic corpus. For example, the verb áºll âto perform the áºll ceremonyâ17 has been attested in no fewer than four variant forms, ʾáºll, háºll, ʾáºl, and háºl (Sima 1999, 93), which all appear in identical contexts; numbers between 10 and 20 can either be expressed by a âteen-and-digitâ or a âdigit-teenâ system (Sima 1999, 118â120; but cf. Macdonald 2008, 213); and both h(n)- and ʾ(l)-forms of the definite article are attested (Farès-Drappeau 2005, 65). The language expressed in the Dadanitic texts does not seem to be a homogeneous entity, raising the question of whether it is even possible to talk about the Dadanitic language in the first place.
Therefore, a description of the language of the Dadanitic inscriptions must grapple with this variation. Some of the earliest scholars working on Dadanitic show the variation in their grammatical sketches or editions of the inscriptions, but do not attempt any explanation (e.g., D.H. Müller 1889, 13â14; Grimme 1937, 300). One of the reasons that even those scholars who believed Dadanitic should be divided into several different phases based on paleography (see Chapter 1) did not connect the linguistic variation to the variation in letter shapes they found, is that the varying forms did not neatly line up with the proposed paleographic divide. Grimme, for example, placed JSLih 063 (containing ʾdq), JSLih 062 (containing hdq), and JSLih 049 (containing hwdq) in his later LiḥyÄnite phase, based on their content and letter shapes (Grimme 1937, 300).
More recent discussions of Dadanitic seem to offer two main explanations for the linguistic variation in the Dadanitic inscriptions. One is the development of the language over time (Sima 1999), while the other focusses more on synchronic linguistic variation (e.g., Farès-Drappeau 2005). For the former, A. Sima (1999, 117) suggests that the h-causatives were probably âon their way outâ, based on their low number of occurrences, implying a chronological development. For the latter, which takes a different approach, Farès-Drappeau proposes that the variation in the form of the definite article is due to synchronic linguistic diversity at the oasis (2005, 65â66), but she does not offer any explanation for the other points of variation.18 Macdonald also recognizes multilingualism at the oasis and classifies several inscriptions as mixed Arabo-Dadanitic texts (e.g., JSLih 071 and JSLih 276 in Macdonald 2000, 52â53), reflecting substrate influence of Arabic within the Dadanitic inscriptions. The main distinguishing feature in the inscriptions termed Arabo-Dadanitic by Macdonald is the use of the definite article ʾl-, as opposed to the more common h-. This is somewhat problematic, as the form of the definite article is not a reliable way to classify a language as Arabic (Huehnergard 2017, 22â23; Al-Jallad 2018b, 6).
3.3.1 Classifying the Dadanitic Language
This leads to the question of how to determine what defines the language of an inscription as Arabic or Dadanitic. To determine this, we need to establish a set of diagnostic features, ideally morphological innovations, that set one language or group of languages apart from others within the same larger language family. Thus, it is only when languages share such diagnostic features that they can be said to share a direct ancestor (Al-Jallad 2018b, 5; Huehnergard and Rubin 2011).
A commonly cited feature to relate different epigraphic languages to Arabic, as mentioned above, is the form of the definite article. This is not unproblematic, however. The use of the definite article seems to be an innovation from Proto-Semitic, as it is absent in Akkadian and Classical Ethiopic. Many of the Central Semitic languages have a definite article which is remarkably similar in its syntax across languages, but it displays a wide range of forms and is not found in all. This indicates that while it was not a Central Semitic innovation, it is likely an innovation that spread areally (Huehnergard and Rubin 2011, 269). Similarly, in Arabic the definite article is not found in all varieties (Al-Jallad 2018b, 12)19 and, therefore, it cannot be considered a feature of Proto-Arabic, nor is it a reliable feature for linguistic classification. In his 2017 article, J. Huehnergard listed 14 Proto-Arabic innovations. A. Al-Jallad (2018a) added another five features to this list. Of these 19 features, Dadanitic only shares two with Proto-Arabic: the merger of s1 and s3 (Huehnergard 2017, 13; Al-Jallad 2018b, 6) and the use of the complementizer ʾn, as in AH 203 (Al-Jallad 2018a, 320).
Aside from these two features, Al-Jallad (2018b, 21â23) shows that there are five Proto-Arabic innovations in which Dadanitic did not participate, or did so only partially. First, several of the Dadanitic inscriptions retain the archaic h-causative, demonstrating that the earliest stages of Dadanitic did not participate in the Proto-Arabic innovation that led to the ʾ-causative (Al-Jallad 2018b, 21). Second, Dadanitic still uses the third-person singular pronoun anaphorically and does not seem to have replaced it with a form derived from the demonstrative base (Al-Jallad 2018b, 22). Third, Dadanitic did not level the -at allomorph for the feminine ending (e.g., qrt /qarÄ«t/ âvillageâ instead of *qryt /qariyat/) (Al-Jallad 2018b, 22). Fourth, whenever dual forms are used, they seem to have retained archaic complexity, using a -y ending in the pronouns and -h /-Ä/ on verbs (Al-Jallad 2018b, 22). Finally, Dadanitic seems to use the preposition ʾdky instead of the typically Arabic ḥattÄ (Al-Jallad 2018b, 22â23).
For the other Proto-Arabic innovations we simply have no Dadanitic data for comparison. For example, there are no clear attestations in the Dadanitic corpus of feminine plural verbs in the suffix conjugation, making it impossible to tell whether Dadanitic participated in the Proto-Arabic innovation of realizing these with the suffix -na (Huehnergard 2017, 13 also see Chapter 5, §â¯1.5 for further discussion of the fem.pl. in Dadanitic). Other features are obscured by the Dadanitic script, which does not display final short vowels, which makes it impossible to tell for sure whether the Dadanitic subjunctive would have been realized with a final -a (Al-Jallad 2018a, 319). Based on the features that are shared between Dadanitic and Arabic, in combination with the lack of participation in some of the Proto-Arabic innovations, Al-Jallad (2018, 21â24) concludes that Dadanitic is most likely a sister language of Arabic and not a direct descendant from Proto-Arabic.
To classify the language of a specific inscription in the Dadanitic script as a form of Arabic, it needs to contain at least one of the innovative features of Proto-Arabic that we do not otherwise find in the Dadanitic inscriptions. One such text is JSLih 384, which was already classified as linguistically Arabic in previous publications (Müller 1982, 32â33; Macdonald 2000, 50; Fiema et al. 2015, 409). The language of this inscription is considered a form of Old Arabic,20 based on the usage of the feminine relative pronoun ʾlt (compare Car. ʾallatÄ«). This is the only inscription in Dadanitic script to use this form of the relative pronoun. The common Dadanitic form of the relative pronoun was masculine Ḡand feminine á¸t (see Chapter 6, §â¯5.1). Thus, even though Dadanitic cannot be classified as a form of Arabic as a whole, the presence of Old Arabic in the Dadanitic corpus shows that Arabic substrate influence cannot be excluded as a possible cause for some of the variation attested in the corpus, such as the use of the definite article ʾ(l)- (Al-Jallad 2018b, 23â24).
3.3.2 Dadanitic in Its Multilingual Context
Dadanitic writing practice did not exist in a cultural and linguistic vacuum, with interaction between the Dadanitic language and writing practice and its surrounding written and spoken languages. As the Lingua Franca of the Near East in this period, it is not surprising that the presence of Aramaic can be felt in the Dadanitic cultural sphere. For example, in JSLih 384, the differences are clear enough that we can say that the language of this text was Old Arabic, written in Dadanitic script. Aside from the grammatical divergence it shows from Dadanitic, however, the text also does not follow the common Dadanitic formulae. Most Dadanitic inscriptions, whether they are graffiti or monumental inscriptions, follow the general structure that is common to most ANA varieties: starting with the name or lineage of the author(s), followed by a verb which agrees with the persons mentioned in the first element, and ending in a blessing formula or a curse on whoever might damage the inscription (see Chapter 4 for an in-depth discussion of compositional formulae). JSLih 384, on the other hand, starts with the object of the verb (nfs¹ âfunerary monumentâ), followed by the person the funerary monument was made for, only then giving the verb (bnh âshe builtâ)21 and the name of the woman who set up the momument. Interestingly, this structure is the same as that of Aramaic dedicatory inscriptions found at TaymÄʾ (Stein 2013, 35; e.g., the inscription labeled Teima 2). This shows that the author of JSLih 384 not only broke with Dadanitic linguistic conventions, but also with its formulaic conventions, likely drawing on the Aramaic writing culture for the formulaic structure she used.
This is not the only example of contact between Dadanitic and Aramaic written culture. There are several Aramaic inscriptions left at TaymÄʾ dated to LiḥyÄnite kings. While this does not mean that these kings knew Aramaic themselves, as these texts were likely commissioned, it does show that they were aware of the presence of Aramaic and its prestige outside of Dadan. They clearly felt that while Dadanitic was the preferred mode of expression for their rituals and inscriptions at home, outside Dadan, Aramaic was much better suited to their communicative goals.
Inside the Dadanitic inscriptions we can also find several lexical items that probably came from Aramaic. These roughly fall into two categories: terms for administrative personnel and architectural terms. Regarding the first, in one inscription, we encounter the title fḥt âgovernorâ. The originally Assyrian term seems to have entered the region and, ultimately, Dadanitic through its usage in Aramaic. Another term that might have entered Dadanitic through Aramaic is ná¹r âguardâ or âhe guardedâ(Abu al-Hasan 2002, 260 considers it to be of Aramaic origin). If this is indeed a loan from Aramaic, that would explain the shift of *Ạto á¹ in this word. However, this sound change can be found sporadically in other lexical items in Dadanitic making it unclear whether this is an internal development or due to contact with a language that had already undergone this sound change, such as Aramaic (see Chapter 4, §â¯6.3 for a full discussion of this sound change in Dadanitic).
Turning to the second category, in the realm of architecture and dedicatory objects the following lexical items seem to be of Aramaic origin: mgdl âtowerâ, ʾrbÊ¿w âsanctuary, square structureâ, and mḥrw âincense burnerâ. The word mgdl is firmly attested in North-West Semitic languages, such as Hebrew, Ugaritic, and Aramaic. It also occurs as mijdal in Arabic, but this seems to be a loan as the mifÊ¿al pattern is mostly reserved for tools in Arabic and not used for places. Given the prominence of Aramaic in the region in the period the Dadanitic inscriptions were carved, this seems to be the most likely source language for the word in Dadanitic. There is a single attestation of mgdl in a Minaic inscription from Dadan (M 315), indicating that it probably entered Minaic in the cultural context of the Minaean trading post at Dadan. The word ʾrbÊ¿w is attested in two building inscriptions (JSLih 059 and U 008) and can probably be connected with Nabataean ʾrbÊ¿n, referring to architectural structures in dedicatory inscriptions, which L. Nehmé suggests comes from the root RBÊ¿ âfourâ (Nehmé 2003, 25 and see Chapter 4, §â¯2.2 for a more in-depth discussion). Dadanitic mḥrw finds a parallel in mḥrn in an Aramaic inscription on an incense altar currently at the TaymÄʾ Museum (TM.IA.017, Macdonald and Al-Najem 2021, 19, also see Chapter 4, §§â¯2.2 and 6.6 for a discussion of the phonological form of ʾrbÊ¿w and mḥrw).
In addition, al-Ḫuraybah 12 and JSLih 035 may contain the verb Ê¿bd with the meaning âto do, to makeâ, which is generally considered an Aramaic isogloss (Huehnergard 1995, 276). In both texts, the verb occurs in the phrase Ê¿bd l-mrʾ-h. In OCIANA, the phrase is translated as âhe made [this] for his lordâ in al-Ḫuraybah 12, following the Aramaic meaning of the root. In the record of JSLih 035, on the other hand, OCIANA offers the translation âhe served for his lordâ22 following the common Semitic meaning of the root.23 The most commonly attested verb to express âto make, to doâ in Dadanitic is fÊ¿l (e.g., U 039; AH 038; al-Ḫuraybah 06; Al-SaÊ¿Ä«d 2011.1 and .2). While this might be taken as an argument that Ê¿bd more likely had the meaning âto serveâ in Dadanitic, it should be noted that Al-SaÊ¿Ä«d 2011.1 and 2011.2 contain the parallel phrase PN fÊ¿l l-DN âPN made (it) for DNâ (al-Said 2011). Moreover, an interpretation of Ê¿bd as âto makeâ would work well with the self-identification of the person preceding the verb as h-á¹£nÊ¿ âthe artisanâ in both al-Ḫuraybah 12 and JSLih 035, turning them into signatures of the inscription. If this interpretation is correct, the duplication of the verb âto do, to makeâ would further support the status of Ê¿bd with this meaning as a borrowing.
Finally, some interaction with Minaic writing can also be observed in the Dadanitic corpus. The Minaean presence at Dadan and the close cultural ties between Minaeans and Dadanites will be discussed in more detail in Chapter 1, §§â¯3 and 6. The most common form of interactions between the two writing cultures can be seen in the mixing of Minaic and Dadanitic lettershapes in graffiti (e.g., JSLih 220). More intensive interaction between the two written languages can be observed in AHUD 1, an inscription in Minaic script, following Dadanitic formulae and linguistic forms, and two Minaic inscriptions (JSMin 145 and JSMin 166). The two Minaic inscriptions were written by the same author, who identifies as the artisan of Wadd (the main deity of the Minaeans). The inscriptions contain roughly the same content, giving the name and title of the author and mentioning an offering he made. Both texts are written in Minaic script and mostly follow Minaic linguistic conventions (such as the use of an s¹-causative), but they also contain a nominal form of the Dadanitic ʾ-causative, and JSMin 145 also seems to contain the definite article ʾ-, attested in Dadanitic (Kootstra 2018a, 24). This all points to quite intense contact between both writing traditions, and at least some degree of bilingualism. Most of the cross-pollination between both corpora seems to have been from Dadanitic to Minaic, which would fit well with a situation where Dadanitic was the main and most prestigious written register at the oasis.
Looking at the traces left by Arabic, Aramaic, and Minaic on the Dadanitic inscriptions, it becomes apparent that we cannot see Dadanitic in a linguistic vacuum and further highlights the problems surrounding attempts to distinguish âtheâ Dadanitic language. Was the shift of *Ạto á¹ an internal development, or triggered by contact with Aramaic, or should we consider the ná¹r forms as a borrowing from Aramaic? What about the introduction of the ʾ-causative, can we see that as evidence of the influence of an Arabic substrate or as a diachronic development, as suggested by Sima (1999, 117)?
In light of this situation, the present study will depart from a usage-based definition of the Dadanitic language and analyze the distribution of recurring linguistic variation across the corpus and in relation to other features of the inscriptions that define them as âDadaniticâ. This will allow for variation to be incorporated in the description and analysis of the language of the inscriptions, while simultaneously setting up some clearly defined boundaries as to what can be considered Dadanitic. JSLih 384, for example, falls outside the scope of Dadanitic proper; that is, while it belongs to the corpus because it was found at Dadan and was written in the Dadanitic script, it breaks with both the formulaic and linguistic conventions of Dadanitic. Such exceptional inscriptions are important in aiding our understanding of the Dadanitic corpus, and although they mostly define the edges of the tradition rather than its inner workings, they do offer a glimpse of the cultural context that is otherwise largely left unwritten.
4 Scribal School and Variation
Now that we have established an inclusive working definition of Dadanitic that can establish some boundaries to the corpus while also allowing for the presence of variation, we can turn to the evidence for the presence of a local scribal culture, its influence on the production of the texts, and the analysis of the variation attested in them. While the preceding paragraphs explored some of the key characteristics of the Dadanitic inscriptions and the cultural environment in which they were producedâidentifying those as script, genre, and languageâwe still might ask how these features inform our understanding of literacy at the oasis. Who was writing in Dadan and how were they taught to write? Understanding the status of literacy at the oasis and how the inscriptions were produced is crucial to our approach to the language used in them. Even though there is little direct evidence available to answer such questions, this section will bring together available information to sketch a picture of the status and use of literacy and scribal culture in ancient Dadan.
The development of the Dadanitic script (see Table 1 in Chapter 1, §â¯4.1) and the contact through commerce with other literate societies from the south of the peninsula, Egypt, Mesopotamia, and the Levant make it very likely that writing on perishable materials was part of day-to-day life in Dadan. Therefore, Dadan can probably be considered a literate society (Macdonald 2010, 14) in the sense that it relied on reading and writing for the functioning of its government and commerce (Macdonald 2005, 49).24 Even though no documents on perishable materials have been found to date, contracts, letters, and administrative documents were probably drawn up regularly at the oasis.
This probable existence of different types of texts, ranging from private letters to official legal documents, which would have been written for different purposes (formal and informal), may also explain the point of contact between the different forms of written language that led to the mixing of forms in the inscriptions, similar to Macdonaldâs (2015, 7) suggested process for the mixing of letter shapes, which will be discussed in more depth below in the discussion of paleography (Chapter 1, §â¯4). That is, if people who were more used to writing private documents, like letters or private notes, attempted to carve a graffito they might try to imitate the linguistic style associated with the monumental inscriptions, comparable to trying to use the formal script for an inscription. As we know from the Sabaic material, for example, the linguistic register used for private letters is often a lot more progressive than that used for monumental inscriptions (Stein 2011, 1048). The accidental combination of both registers in the writings of those not used to writing on stone may explain how different layers of historical forms ended up in the same register. It is interesting, however, if at some point different registers of both script and language existed, how their mixing became widespread enough to become acceptable, even for the more formal registers. To answer this question, it is helpful to turn to the spread of literacy and how people were trained to read and write.
If the use of writing was indeed so widespread in ancient Dadan, this would suggest that professional scribes were employed and trained at the oasis. Even though it has been argued that learning to read and write an alphabetic writing system is simple enough not to require any formal education (e.g., Jamieson-Drake 1991, 9:154; but cf. Rollston 2010, 92),25 setting up formal documentsâsuch as letters, contracts, and deedsâwould require expertise beyond âjustâ knowing the letters.26 For example, based on comparisons with scribal education in Mesopotamia and Egypt, K. van der Toorn (2007, 98) argues that Hebrew scribes not only learned the 22 letters of the Hebrew alphabet, but were also trained to become familiar with the language use and terminology associated with specific fields, such as notary documents and documents for litigation (Van der Toorn 2007, 99â100). In addition, scribes were likely responsible for bookkeeping, which would not only require them to know how to write but also to know how to draw up a balance sheet and to perform some basic arithmetic (Van der Toorn 2007, 100).
Besides skills in drawing up different kinds of documents, scribes would also learn to prepare their writing materials. In the case of scribes in ancient Israel these would include reed pen, papyrus or parchment, and a stylus (Rollston 2010, 112). Since we do not know what perishable materials were used for writing in Dadan, these materials were not necessarily the same, but anyone writing regularly would need some knowledge of how to produce some of these tools, unless all writing materials were imported ready to use, which would have been very costly.
In summary, even though the acquisition of basic reading and writing skills may have been relatively easy for an alphabetic script such as Dadanitic, a society in which writing was used for more than personal notes and lists would still require some professionally trained scribes to produce the more formal documents and prepare writing equipment. In this light, it may be useful to think about just how many scribes an oasis like Dadan would need. W.M. Schniedewind reminds us that âwriting is fundamentally a luxury goodâ (2013, 118) and that it needs a degree of economic and cultural support to thrive. How many documents, then, would a society like Dadanâs need and how many people could afford to commission a text? Given the relatively small reach of the kingdoms of Dadan and LiḥyÄn, the demand for writing was probably not massive and one family of scribes in which knowledge would be passed on from father to son may have been sufficient to supply the oasis with the official writing it needed; clearly this cannot be compared to the bureaucratic apparatus needed by the Babylonians or ancient Egyptians.27
Additional support for the presence of scribes at Dadan may be found in the inscriptions themselves. Based on the attestation of the word h-s¹fr âthe writerâ in Dadanitic, there seems to have been a distinct difference between the mason who produced the objects of the inscription and the scribe who set up the text, at least for some of the commissioned inscriptions. There are two inscriptions that mention h-s¹fr âthe writerâ of the inscription and their name at the end, alongside âthe artisanâ h-á¹£nÊ¿ who presumably cut the inscription (JSLih 082 and AH 220).28 This slot in the formulae is usually reserved for mentioning the artisans involved in the production of the inscription, who are always mentioned separately from those who dedicated it, whose names are given at the start of the text (see Chapter 3, §â¯3.4). This supports the reading of h-s¹fr in this position as a professional title. Furthermore, the fact that h-s¹fr and h-á¹£nÊ¿ are both used in the same inscription tells us that the one who cut the inscription was apparently a different person to the one who wrote the textâthat is, unless we assume that the fact that the writer is mentioned in only these two inscriptions points to the unusual circumstances under which they were made. Since it is quite common, however, not to mention the artisan who cut the inscription either, even in inscriptions executed in relief, it seems unlikely that anything only mentioned sporadically was necessarily out of the ordinary.29
As for the training of individuals to read, write, and produce inscriptions, there is only one abecedary attested in the Dadanitic script (JSLih 158). The inscription contains several badly formed glyphs and repetitions of the same sequence of letters, which led Macdonald to conclude that it probably represents a writing exercise (1986, 113). The presence of repeating glyphs and badly formed letters in several other inscriptions on the same rock face led Macdonald to suggest that it was used as a practice site (1986, 115). Since this is the only such practice site known so far, it can tell us little about the organization of schooling in the oasis, unfortunately. It is unclear whether this site was used for the official training of masons, or by private persons.
4.1 Graffiti and the Spread of Literacy
While there seems to be some evidence for the existence of a scribal class at Dadan, the presence of a large number of graffiti in and around the oasis30 suggests that literacy was much more widespread than a small cadre of professional scribes. In the context of ancient Israel, Schniedewind points to the growing number of graffiti and attestations of writing in an administrative context, such those as found on seals, seal impressions, weights, and economic texts in the period between the eighth and sixth centuries BCE to argue for what he calls âthe democratization of writingâ (2013, 99â105). He links this spread of the ability to write to a loss of a strictly controlled written standard; that is, as writing was no longer confined to a small scribal elite, the ability to control the written standard diminished, as evidenced by the increase in inconsistencies in grammar and spelling in this period in Ancient Hebrew writing (Schniedewind 2013, 100). In the Dadanitic situation there is currently no way to tell whether there was a process of democratization or whether literacy was relatively widespread within the community from the beginning of the production of the inscriptions, as we cannot date the inscriptions relative to each other. There are some interesting parallels, however, between the situation as described by Schniedewind and what we see in the Dadanitic corpus; specifically, in the existence of large amounts of graffiti accompanied by a remarkable amount of variation in grammar, letter shapes, and orthography.
While the presence of large amounts of graffiti in the area of Dadan suggests that literacy probably spread beyond a small group of professional scribes, this does not mean that everyone at the oasis could read and write, or that everyone reached the same or a similar level of literacy. Here the concept of craftmanâs literacy seems helpful. W.V. Harris defines craftmanâs literacy as âthe condition in which the majority ⦠of skilled craftsmen are literate, while women and unskilled laborers and peasants are mainly notâ (1989, 8). Within such a larger group of literate individuals within society the majority may only have achieved what Schniedewind defines as âsignature literacy or craft literacyâ, a level of literacy sufficient for practical purposes such as signing oneâs name, writing lists and receipts, and possibly the ability to read short letters (2013, 105). This level of literacy is not comparable to that of a trained scribe but would be sufficient to leave a short graffito.31
The amount of variation that began to occur in Ancient Hebrew writing between the eighth and sixth centuries BCE lead Schniedewind to conclude that there was no strong Hebrew scribal institution present in Iron Age Judah (Schniedewind 2013, 117). When compared to the Dadanitic situation, this may lead us to conclude there was likewise no strong scribal tradition in Dadan either, as we have plenty of variation in all layers of writing. However, this is based on the supposition that the goal of any scribal tradition would be uniformity, which may not have been the case. Relatively widespread literacy can help explain, however, how a certain amount of variation entered the written norm in the first place and enabled it to develop and maintain some connection to the spoken language. The incorporation of more progressive linguistic forms such as the ʾ-causative and the collapse of word final triphthongs which we see reflected in the spelling of rá¸-h âmay he please himâ (see Kootstra 2019 and discussed further in Chapter 4, §â¯3.2) most likely followed these developments in the spoken language and eventually became the most common form in writing as well. However, this does not mean that the written language was simply a transcription of the spoken language, as we can see for example from the occasional spelling of á¹ for *Ạ(e.g., in AH 009.1). The loss of *Ạseems to have been a feature of at least some spoken register at the oasis, but it clearly was not the most favored in the written language.
Even though the influence of literate individuals from outside the scribal elite may have opened up the written tradition, allowing it to incorporate a certain amount of variation and flexibility, this does not necessarily mean that there was no scribal tradition present at all. An interesting point of comparison to this may be the monumental Sabaic inscriptions from the south of the Peninsula, which can be divided into an Early, Middle, and Late period. Evidence from Sabaic letters written on palm sticks, however, shows that the spoken language changed at a more rapid pace. There we find, for example, that the glyph Ḡis often used to represent *áº, while they were consistently kept apart in monumental writing until the end of the tradition (Stein 2011, 1048). Thus, despite the gradual implementation of linguistic changes in the monumental tradition, the private documents on perishable material confirm that it was quite far removed from the spoken language.
Besides the occasional use of á¹ for *Ạin Dadanitic, there are several other specific forms to be found in the corpus that suggest that the author of the text was aiming for a written standard they had not quite mastered. For example, in the inscription in which both an h-causative and a ʾ-causative occur (Al-SaÊ¿Ä«d 1419/1999: 4â24, no. 1, side 1â2), the inscription in which two dedicants agree with a dual verb but plural resumptive pronouns are used (U 019), and the inscription which is completely in the singular except for the resumptive pronoun in the blessing formula in the dual (AH 120). Such inconsistencies seem like hypercorrections, which suggests a certain distance between the spoken and written registers of Dadanitic that the authors of the inscriptions were quite aware of for these linguistic features.
Even though literacy seems to have been too widespread for a small scribal class to have maintained complete control of the written standard, people were clearly not simply transcribing their spoken language, and there was some form of written standard present. From the likely use of writing for bureaucracy and the possibility of employing a special scribe to set up an inscription (as evidenced by JSLih 082 and AHÂ 220), it seems clear that there were trained scribes present at the oasis. They would have been trained in the âproperâ use of the language, including grammar and orthography. Their knowledge of the language and the highly formulaic nature of the inscriptions in general would probably have been enough to establish a core scribal code, which likely included knowledge of less common grammatical forms and phrases. At the same time, there was probably also a group of people outside this cadre of professionally trained scribes that knew how to read and write to some extent (as evidenced by the large amount of graffiti present at the oasis). The common use of writing by such less highly trained individuals may explain how variation entered the written record and possibly even pushed it to become more flexible and incorporate more progressive linguistic and paleographic forms, while the presence of highly trained individuals ensured the continued presence and knowledge of more archaic linguistic forms within the written code. Considering the language of the inscriptions as a written and learned standard, which differed in some ways from the spoken register, has methodological implications for the approach of the description of the language of the inscriptions and how to deal with the variation found within them. This will be discussed in more detail below.
4.2 Expected Patterns of Variation in a Written Code
The historical context and evidence from the content and language of the inscriptions make it likely that not only the physical production of the text on stone was commissioned, but also the drawing up of the text itself could be taken care of by a trained professional scribe. While there are two inscriptions that mention both the scribe and the mason of the text, it is unclear whether they represent the usual division of labor. Both having a trained scribal class and the use of commonly known standard formulae would have a standardizing effect on the language.32 One of the expected effects of using standardized language and formulae to write is that the language becomes resistant to change and will likely develop at a slower rate than the spoken language, creating an environment in which the spoken and written registers can become separated from each other to some degree.33
The assumption that the variation in the corpus is due to the archaic nature of the language of the inscriptions implies that most of the inscriptions will contain more archaic linguistic forms, with occasional interference from more progressive linguistic forms from the spoken language. However, this is not the general distribution of the linguistic variants in Dadanitic. There we see that the linguistically more progressive forms are the most common, while most of the variant âinfiltratingâ forms are linguistically archaic.34 Since the oasis of Dadan was an important trading hub, it was a multilingual place, as evidenced by the presence of both Minaic and Aramaic writing at the oasis in addition to the Dadanitic inscriptions. Given the multilingual nature of Dadanitic society, this distribution of linguistic forms could indicate that the written standard was based on a more linguistically progressive language than (one of the) spoken language(s) at the oasis.35
If the variation in the Dadanitic corpus is indeed due to such a difference between written and spoken language, the higher prestige forms should occur relatively often in more expensive and higher register texts. Presumably, someone who could pay for a good mason to produce a beautiful inscription would also want the language of the inscription to be sophisticated and would therefore employ an individual that could be trusted to produce a good text. Infiltration from the spoken language, in this case reflecting the more archaic linguistic forms, is then expected to occur more often in informal inscriptions, where the formality of the language is of less concern, or in more poorly made inscriptions. However, upon closer inspection, the opposite seems to be the case again. While the archaic forms are indeed the less common forms, they are more closely associated with higher register inscriptions than with graffiti.
Alternatively, variation may be due to diachronic change. Many of the linguistic variants display a form that is linguistically more archaic and one that is more developed. It is therefore also a logical possibility that the variation in the corpus reflects diachronic change rather than synchronic variation. If this is the case, we would expect to find that archaic linguistic forms cluster together, possibly even to the exclusion of some of the more progressive forms, in case one form ceased to be productive before another developed. It seems indeed the case that certain archaic linguistic forms tend to occur together within individual inscriptions. In addition to giving new insight into the mechanisms underlying variation in the corpus, this may also cast new light on previous proposals about the chronological development of the script.
4.3 Methodological ConcernsâAnalyzing the Language of a Scribal School
In this introduction, the question was posed whether it is even possible to speak of the Dadanitic language, given the amount of linguistic variation attested in the Dadanitic corpus. As just discussed, the variation indeed suggests that the spoken languages of the authors of the inscriptions were not homogeneous, probably due to both synchronic variation and diachronic change. However, it seems possible to distill a written language from the bulk of the inscriptions. The presence of trained scribes at the oasis means that a description of the language of the inscriptions is really a description of the language of the writing tradition.36
In this light, then, deviations from standard conventions form invaluable evidence for the linguistic background of the person who composed the inscription and the spoken language(s) at the oasis. The investigation of the language of the Dadanitic inscriptions in Chapters 3 through 7 will therefore aim to identify both the most commonly used forms in the writing tradition and the less common varying forms, both in grammatical features and formulaic parts. Whenever there are two variants of what has to be the same form (e.g., the h- and ʾ-causatives) there is always one form that is clearly the most common, in terms of number of attestations, and one that is the variant. These more common forms are the frame of reference for our understanding of the core of the Dadanitic writing tradition and as such they can anchor the discussion of any variant forms.
Forms that fall on the periphery of the writing tradition include unique words, or forms that are used in uncommon or unique contexts, and personal names. Since the writing formulae are part of the writing tradition of the oasis, the spellings of the forms that fall within the common formulae were also likely a part of the tradition. This is observable in the relative consistency of the orthography in these inscriptions. This also makes it more likely that when we do see repeating alternative forms in these formulaic environments, they do not reflect random variation due to uncertainty about the existing spelling conventions but represent phonologically or morphologically different forms.
Another issue that needs to be kept in mindâespecially when describing the orthography and phonology of Dadanitic, as discussed in Chapter 4âis that we have no transcriptions of Dadanitic language in other scripts, like Greek, for example.37 This sets the Dadanitic corpus apart from corpora like Nabataean and Safaitic. Consequently, a description of the phonology of Dadanitic must rely solely on the orthography of the inscriptions to make inferences about the phonology. The use of matres lectionis to write final long vowels, for example, has implications for the status of the (word final) triphthongs. However, the interpretation of the value of the possible matres lectionis -y and -w also depends on our understanding of the development of word final diphthongs and triphthongs. To avoid circularity, we therefore need additional evidence; for instance, from the use of matres lectionis in environments where they do not represent an etymological diphthong or triphthong (e.g., the use of -y to represent the first-person possessive suffix /-Ä«/). Whenever such conclusive forms are not attested in the corpus, the available evidence will be provided as completely as possible, and the different possible explanations of the data will be discussed.
4.3.1 Evidence from Personal Names
Personal names cannot tell us about the synchronic grammar of the language, but they can shed light on the orthographic practice. As Macdonald has thoroughly discussed, names do not necessarily reflect the language of their bearer, as they are often linguistically archaic and can âtravelâ from language to language (1999, 254â257). This also has implications for the use of personal names to say anything about the phonology or orthography of a language. Even though we can use names to say something about the phonology of the language of their bearer,38 the fact that they are often borrowed from other languages still needs to be taken into account.39 Someone mentioned in a Dadanitic inscription with the name ná¹r (JSLih 079), for example, does not necessarily tell us that the language this person spoke had merged *Ạand *á¹, as it is also possible that the name was taken from another language that had merged the two, like Aramaic.
There are other examples, however, when the spelling of a name does reveal something about the orthographic practices of the script used to write it. A good example is the female name ʾmtktbh, in which the etymological ending *-ay of the feminine elative is represented by -h. Since it cannot reflect an etymological spelling or an archaic pronunciation here, the -h must represent a final -Ä (see Chapter 4, §â¯2.1) and so it informs us on the use of matres lectionis in Dadanitic.
Finally, it needs to be remembered that the vocalization of a name as we find it in the epigraphic record is often far from clear (see Macdonald 1999, 271 for a discussion of how to interpret the consonantal skeleton of a name). This makes the interpretation of glides in personal names highly problematic. For example, when we find both zd and zyd, this does not necessarily indicate that these forms show a difference in the spelling of the diphthong in the name Zayd, as it could equally be the case that zd represents the name Zayd while zyd represent the name ZiyÄd with the y representing a consonant. Whenever relevant, examples from personal names will be used to illustrate points about phonology and orthography. In most cases, however, for the reasons outlined above, these examples will not provide any conclusive evidence on the matter discussed, but merely additional support or a side note to possible variation.
Of course, there are many factors involved in variation. Having established the methodological approach on which this study is based, it is first necessary to offer some background information locating the Dadanitic inscriptions in time and space, which I will do in Chapter 1. With this background scope established, Chapters 2 through 6 of this work can then turn to the key work of providing a complete picture of the written practice of ancient Dadan, including both common variation and linguistic and formulaic outliers, in order to be able to say something about the âwhatâ and âhowâ of variation, which will be treated in depth in Chapters 7 and 8.
See Macdonald (2000) for an excellent overview and nomenclature, and Al-Jallad (2018) for a linguistically oriented overview.
M.C.A. Macdonald classifies Dispersed ONA as all texts written in ONA scripts used in sedentary environments that cannot be classified as Taymanitic, Dadanitic, or Dumaitic. This subgroup includes texts from Mesopotamia that were formerly called âChaldeanâ (Macdonald 2000, 29).
Thamudic includes a number of different scripts that have not fully been identified (Macdonald 2000, 32).
Besides Dadanitic, ONA includes Taymanitic, Dumatic, and Dispersed ONA (Macdonald 2000, 29).
But see Al-Jallad (2018, 21â23), where he shows that Dadanitic is probably a sister language of Arabic rather than a direct descendant of Proto-Arabic.
Several Dadanitic inscriptions have been found further away from the oasis: four were found about 85â¯km north of al-Ê¿UlÄ in Jabal ThadrÄ (JaL 171; 174) and two in the area northeast of WÄdÄ« Ramm, in southern Jordan (Hidalgo-Chacón DÃez 2015). A few Dadanitic inscriptions have been found in the vicinity of the nearby oasis of TaymÄʾ (Hayajneh 2016). Several inscriptions in the Aramaic script mentioning the king of LiḥyÄn have also been found at TaymÄʾ (JSNab 334; 335; 337).
For a more elaborate overview of the decipherment of the Dadanitic script see Farès-Drappeau (2005, 31â33) and (36â41) for a discussion of the history of scholarship on Dadanitic. For an overview of contributions to the study of Dadanitic following Caskelâs (1954) edition see Sima (1999, 3â4). All the inscriptions including available photographs and bibliography are available on the OCIANA database
22-04-2018
For more information about the project and staff see
This has of course long been recognized for ancient languages such as Akkadian, written in cuneiform script, or even in alphabetic writing traditions such as Nabataean, where the difference between the written Aramaic and the substrate of Arabic, of which traces can be found in the written language, is more immediately apparent. However, in ASA and ANA epigraphy this plays a much less prominent role in the approach to their language, probably partly due to the large number of graffiti found in scripts that fall under these umbrella terms.
While K. Milnor uses this general description of graffiti, she also cautions that the category âgraffitiâ should be evaluated in its cultural and historical context. As ideas of authorship and public and private property change over time, so do graffiti, both in their appearance and in what can be understood to make up the category in the first place (Milnor 2014, 4).
For a more elaborate discussion on the possible role of a scribal school at the oasis see the discussion on scribal school and variation below in the Introduction.
The ná¹r inscriptions are a notable exception. They are found almost exclusively at Jabal Iṯlib, a rock outcrop to the northeast of the oasis, except for one that was found at WÄdÄ« MuÊ¿tadil. Given that the authors of the inscriptions seem to have acted as guards, it is not surprising that their inscriptions all cluster in a favorable look-out place such as Jabal Iṯlib, although recent discoveries of the MadÄʾin SÄlih archaeological project suggest they may be associated with a funerary context (Nehmé et al. 2021, 14â19 and see Chapter 1, §â¯6). They do not only stand out as a group due to the content and location of the inscriptions, but they also all share the merger of Ạand á¹ in the verb and 5 of the 19 ná¹r inscriptions share a particular style of engraving (see Chapter 2).
This is similar to the use of genre by I. Taavitsainen (2001, 140), who proposes to distinguish a linguistically based category, âtext typeâ, from a non-linguistically based âgenreâ. Even though the use of specific linguistic features seems to have been preferred in certain genres of inscriptions (see Chapter 7), there are no features that are exclusively used with specific genres. Therefore, there is no clear difference between a linguistically motivated âtext typeâ distinct from a content- and formula-based âgenreâ. Because of this, I will not use the separate category âtext typeâ in my analysis of Dadanitic.
See Chapter 3 for a complete overview of all specific formulae used in the Dadanitic inscriptions.
Here, register will be used to refer to the social hierarchy of the inscriptions. In register studies and sociolinguistics the term is generally used to indicate âsituational language useâ (Taavitsainen 2001, 141), in other words, how peopleâs language use changes depending on the situation in which they use it, which can include different social dynamics, different media (written or spoken language), etc. (Ferguson 1994, 16). Since the Dadanitic corpus only reveals the language use in one specific medium, register will primarily relate to the level of formality of the inscription.
For reasons of space and brevity, I adopt here the neutral translation of (h/ʾ)áºll as âto perform the áºll ceremonyâ offered in the OCIANA database. The interpretation of the ritual has long been debated. Previous interpretations have suggested the ritual included the construction of sunroofs for a religious ceremony (Stiehl 1971, 5â7), or the construction and maintenance of a subterranean canal system (Van den Branden 1969; Sima 1999, 49â50). Recently, a new interpretation of the form háºl from the same root in Sabaic has been suggested, which links it to the act of writing, rather than shade (see
Farès-Drappeau also refers to Robinâs (2001) proposal that there might have been a north Arabian koiné that developed from the trade contacts with the south of the Peninsula (Farès-Drappeau 2005, 65â66) again referring to multilingualism in the region.
Hismaic and some Safaitic inscriptions lack the definite article (Al-Jallad 2018a, 322).
Al-Jallad identifies the variety more specifically as Old ḤigÄzÄ«, which he identifies based on the use of the relative pronoun formed using a demonstrative form based on the use of (h)alla+DEM, such as CAr. ʾallaá¸Ä« (Al-Jallad 2015, 13â14; 2020).
The common form of the third-person feminine singular verb in the suffixing conjugation would be bnt in Dadanitic. See Chapter 5, §â¯1.2 for a full discussion of this verbal form.
Note that Jaussen and Savignac also translated Ê¿bd as the verb âto doâ, âlike in the Nabataean inscriptionsâ, in their publication of JSLih 035 (1909â1912, 363).
Based on the records in OCIANA,
Macdonald uses this this term to distinguish it from societies such as those which produced the Safaitic and Hismaic inscriptions, which he termed non-literate societies. These societies seem to have had high rates of individual literacy, as evidenced by the large amounts of inscriptions left in these scripts, but they seemed to employ writing primarily for purposes that were not related to the functioning of society, such as record keeping, drawing up of contracts, etc. He proposes to use the term illiterate only for individuals who could not read or write and not for societies as a whole. This distinction is a very useful one related to the use of writing in oasis towns and by nomadic groups in pre-Islamic Arabia. This does not mean that Macdonald would suggest a sharp divide between literacy and illiteracy within literate societies. This divide has been challenged for decades in literacy studies, as can be seen, for example, in E. Chamberlinâs analysis of hunting practices of hunter gatherers as reading practice (Chamberlin 2002). For an overview of the development of the field of literacy studies see Street (2009).
The material attested at Ugarit clearly shows a discrepancy between the amount of teaching materials and practice texts for Akkadian cuneiform as opposed to those in the Ugaritic alphabetic cuneiform, with the Akkadian cuneiform material forming the vast majority of that recovered. This seems to confirm that learning the alphabetic script took less effort, which would open up the potential for the development of rudimentary literacy (Schniedewind 2013, 105). Also, among psycholinguists the issue of learnability of writing in different types of writing systems is controversial. Sebba gives an overview of psycholinguistic studies to show learnability is a complex issue and difficult to establish with certainty. His overview suggests, however, that more phonemic writing systems seem to favor quick learning for new learners, who will rely more on phonology, but are not necessarily the most friendly for smooth reading for more âmatureâ users, who will rely more on lexicon (Sebba 2007, 18â23).
Viewing literacy as more than just the memorization of a script is central to the approach of the New Literacy Studies which developed in the 1980s. This approach to literacy views reading and writing as a set of social and cultural practices and presumes one would not only need to learn the script but also how to use it in different social and cultural contexts (Sebba 2007, 13; and see Barton 2007, 22â28 for an overview of the development of the field of literacy studies). In the Dadanitic context, the high level of formularity of even graffiti indeed suggests that even rudimentary levels of literacy included learning the basic formulaic structure of a graffito (or an inscription more generally), besides the script.
See Van der Toorn (2007, 54â73) for an overview of scribal practices and training in Mesopotamia and Egypt.
An often-used argument for the existence of a Hebrew scribal class is the use of the term sÅpÄr (âS¹FR) to indicate someoneâs profession (Van der Toorn 2007, 78â81).
The verbs s¹á¹r (JaL 061 f) and s¹fr (Ǧabal Iṯlib 08; JSLih 128), both meaning âto writeâ, are mentioned in inscriptions as well, but do not seem connected to the professional production of a text, so much as to the act of inscribing a graffito.
In fact, graffiti make up the bulk of the attested inscriptions: 1462 of 1871 inscriptions of which the genre could clearly be identified are graffiti (see Chapter 7, §â¯2).
In literacy studies the phenomenon of being able to perform certain literacy events, but not others, is tied to the idea of literacy domains or literacies. A literacy âis a stable, coherent, identifiable configuration of practicesâ (Barton 2007, 38). In other words, filling out your tax forms, reading a book to a child, and skimming the headlines of the newspaper are all literacy events in the sense that you interact with written language, but they each require different skills. One of the uses of approaching literacy through literacy domains is to move away from viewing different uses of literacy on a scale from simple to complex, but to see them as different uses of literacy.
Note that it seems that a writing culture with standard compositional formulae can also develop without the existence of scribal schools, as it did in the Safaitic and Hismaic inscriptions, for example (Al-Jallad 2015, 3).
We can see this, for example, in the variation attested in the Aramaic from the Achaemenid period as described by Folmer (1995). A clear example can be found in the letters belonging to the Yedaniah archive, where some scribes diverge more from the archaic standard (for which Folmer used the Arsham letters on leather) than others (Folmer 1995, 693). The higher degree of formality of more archaic forms is further supported by variation in language use across different genres of text, with legal texts, for example, containing more archaic spelling than private letters (Folmer 1995, 696).
For an overview of the absolute number of occurrences of variant forms see Chapter 7.
An example of a situation where the local language is more archaic than the high prestige written language can be found in the Hermopolis letters, which display influence of the more linguistically progressive Achaemenid imperial Aramaic. The distribution of the varying forms is different, however, with the infiltrations of the high prestige form the minority (Gzella 2011, 582â583).
See, e.g., Barton (2007, 33â50) for a practice approach to literacy and his introduction of the metaphor of an âecology of written languageâ.
There may be two Minaic inscriptions at Dadan that include several borrowings or code-switches to Dadanitic: JSMin 145 and JSMin 166 (Kootstra 2018b).
Even though the name Michael comes from Hebrew originally, its English pronunciation can tell us, for example, that [i] came to be pronounced as [Éɪ] in modern English.
Even though the Dutch equivalent of the name Michael âMichielâ [mi