One of these features is the l-prefix of the defi nite article. It may be true that the l- is not common among the Semitic languages in general, and it is not common among all the modern dialects of Arabic. In some modern Yemeni dialects m- or n- function as definite article prefixes. In fact, in the majority of Egyptian dialects m- as a prefix appears in an unproductive manner with some words such as ’imbāriḥ , ‘yesterday.’
The same sound phenomenon was also recorded in pre-Islamic dialects, especially in the Hijaz region and Ṭayyi’. In the ancestors of some old dialects of Arabic, Central Asian dialects, the article is missing altogether. In this latter case, it is not quite clear if the absence of the prefix came as a result of contact with areal languages or was originally missing from the source dialects in the peninsula. Another morphological phenomenon that does not exist except in Arabic, but not in all varieties, is form IX if‘alla . This verbal form does not exist in large parts of North African Arabic dialects. The verbal form if‘āl exists instead.
Admittedly, these are very few and sporadic examples. I adduced them, however, to indicate that in fact what we now call Arabic is an amalgamation of complex varieties that share some features with the rest of the Semitic languages, share other features with one branch and not the others, share some features with another branch and share some internally unstable features with different branches. The features that some varieties of Arabic may share with different Semitic languages are not always common among all the dialects that we collectively call Arabic. What this may indicate is simple: what we now call Arabic may indeed have been a complex of neighboring Semitic dialects.
The revelation of the Qur’ān and the establishment of the Arab state may have been instrumental in the consolidation of these dialects into a language. This view represents a challenge to the traditional view of language relationships as that of a family that can be represented by a tree, or stammbaum, where every language is represented on the tree by an independent branch in isolation from the other sibling branches with which it shares a trunk. These examples and many others indicate that the Semitic languages may have not been branches stemming out of a proto-Semitic form but rather a continuum of isoglosse.
There are recent noteworthy attempts that use new data and new analysis of existing data to connect the pre-Islamic varieties of Arabic in the peninsula to the Semitic varieties of northwestern Arabia. We will report here on one such study to illustrate the potential contribution to the field. Al-Jallad published a grammar of the Safaitic texts in which he makes the connection between Arabic and northwest Arabian Semitic varieties from the beginning of the Common Era. The Safaitic texts are a group of texts between the fi rst and fourth centuries written in a group of varieties that came to be collectively called Ancient North Arabian. The language of these texts should be considered varieties of Old Arabic because there are some important isoglosses on all levels of linguistic analysis.