Some Early Arabic Inscriptions from Al-Ḥanākiyya, Saudi Arabia

DURING the spring of 1974, Dr. Michael Willis had occasion to visit a rock outcropping at al-Hanakiyya, Saudi Arabia, located about 110 km east-northeast of Medina. The rock outcropping is in the form of a large red sandstone butte standing isolated in the midst of otherwise rather flat terrain just to the left of the main highway through al-Hanakiyya as one travels towards the Najd. A second, smaller hill of similar appearance is located nearby. The steep rock faces of the larger butte contain numerous pre-Islamic inscriptions which appear to be known to those working on such materials but have not, to the best of my knowledge, found their way to

the Safaitic Bedouin (New Haven, 1968), pp. 2-3: "If a basalt stone is broken, the fresh surface is gray; but on prolonged exposure to the air, the stone, through the chemical action of precipitation and atmospheric gases with its iron and manganese constituents, develops a patina which turns in the course of time to a dull red-orange and then to a brownish or purplish black ... ." 3 I have tried to bear in mind the warnings expressed by Grohmann in EPRL, p. xxi: ". . . dating according to seemingly typical forms of letters-which look old but in fact occur throughout nearly the whole period in which the old monumental angular style was used, e.g., open CAin or triangular MTm or H--might be considered purely haphazard. Well known examples ... show the danger of any such attempt. However, an obvious relationship between the style of writing, together with certain isolated forms on the one hand, and the style or writing and forms in certain dated parallels on the other, occasionally allow these texts to be dated." doubtless be possible to clarify some of the problems I have encountered in rendering the texts. One of the photographs provided was of an inscription that was apparently so weathered, or shallow in its engraving, that little more than an occasional word could be made out, although the inscription itself was relatively long (perhaps eight lines?). This inscription has not been included in the present collection; from its script, it appears to be somewhat more recent than the others considered here.

Note on the Figures
All the original photographs taken by Dr. Willis were 35 mm color transparencies. From these slides were made a set of ten 8" x 10" prints, most in color, and it was from these prints, in the main, that I worked, although for a time I consulted the original slides to attempt to resolve difficulties in inscription W 5. The relationship of the prints to the plates published here and to the inscriptions is as follows:

Print Number Figure Number
Inscription Number 1 (black and white) 1 W 1 2 (color) 2 W 2 3 (black and white, same as 2) -W 2 4 (color) 3 W 3 5 (color) 4 W 4 6 (color-light photo) 5 W 5 7 (color-dark photo) W 5 8 (color) 6 W 6-W 12 9 (color) 7 W 13-W 14 10 (color) W 15 (omitted) Prints number 2 and 3 were identical (made from the same slide original), except that number 2 was in color, number 3 in black and white, and not as clear as number 2. Print number 7, like number 6, showed inscription W 5, but in much poorer light, and is too dark to merit publication here, although ironically it turned out that by candling it against a bright light some of the inscription could be recovered from it more easily than from print 6. As noted above, the inscription W 15 was too indistinct in the photograph to warrant publication here.     1-2. The name is clearly cAsim ibn cAll, even though the final ypd in cAll has been omitted, for we have in W 3 another inscription with the same name in which CAll is written with the final yd'. The name has a double nisba, the second separated from the first by the word thumma ("then") and referring to a particular clan or lineage within the larger tribal group designated by the first nisba. Such double nisbas are encountered fairly frequently in the literary sources for the early Islamic period.13 Since the inscription is written without diacritical points, the first nisba could be read either as al-ThaclabT or as al-Taghlibi, both of which are fairly common tribal names.14 The latter possibility can be excluded for several reasons. First, the tribe of Taghlib  1. On the imperfect used as optative, see EPRL, no. Z 2. The imperative mood, or the perfect used to express the optative, is much more common. 1-2. The name is problematic. "Al-Hulw" is rare but attested as a masculine proper name.3 However, there appears to be a diacritical dot below the letter r, and if this is in fact the case, then we must assume that the name is an incomplete rendition of some other, such as Jalwan or Jilwan32 or JiluwT-but all of these lack the definite article. JiluwT is, or course, a fairly common name in the Saudi family.33 For the name Arbad, see IK/Caskel, s.v., and Tab. index, s.v. J. J. Hess, Beduinennamen, p. 25, records the feminine form of the name, Rabda. My restoration to "al-Hulw" is conjectural.

The last group of letters in this line is not clear in the photograph. The first letter
following the wdw may be a lam, in which case we would read, "and (may God also forgive) . . ." followed by another name or designation of someone. No acceptable reading for a name, etc., seems possible on the basis of the letters given, however, and the reconstruction provided, though tentative, seems as likely as any. The verb amadda most commonly means "to assist" or "to reinforce" in a military sense, but it can also mean to aid by giving foodstuffs, etc.; indeed, it is used in this sense several times in the QurDan (e.g., Q. 52:22, wa-'amdadnd-hum bi-fiakihatin wa lahmin mimmal yash-tahi~n, "and We succored them with fruit and meat that they desired").  (line 1), the form of final cayn with open top and sharply bent tail (line 3), the large,  rectilinear dal (line 2), and the small, tightly-curved rd' (lines 1 and 2).

As in W 1-W 3, the letter forms of this inscription suggest a fairly early date, although the brevity of the text requires us to be a bit cautious. A date within the first three centuries A.H., however, seems assured by such features as the open medial cayn
Once The inscription is in general difficult to read, being very crudely inscribed on a rough surface that was apparently already partly covered with what appear to be, in part, animal drawings (note especially the form drawn below the word "God" in line 1). Numerous drawings of this kind not reproduced in the facsimile sketch can be found in the photograph to the upper right of the inscription. Other marks, probably tribal wusum, are interspersed amid the words of the inscription, notably above line 2. Poor lighting makes the whole left side of the inscription very difficult to make out in the two photographs, and the lower left corner of the inscription has evidently flaked off. Even in places where the writing seems to be quite legible in the photographs, the letter groups sometimes yield no satisfactory reading. 1. The lower part of the name is obscured by an area of chips and scratches, and the reading should be considered only tentative. 2. The verb nasara, "to assist," is frequently used in the Qur dn, but usually in the sense of God assisting believers. If we take al-rahman to be the subject of the verb, however, ("whomsoever the Merciful assists .. ."), we would expect the verb to have an object pronoun (man yansuru-hti al-rahmanu). Since the pronoun is lacking, the reading "whosoever helps God . . " seems indicated. The notion of believers assisting God is less common in the Qur'dn than the reverse but nonetheless attested (e.g., Q. 59:8 and Q. 47:7). The tribal mark (?) above yansuru is much lighter than the other markings and presumably is much more recent than the inscription and other drawings; the W-shaped mark between al-rahman and I1, on the other hand, is much darker, and may antedate the inscription, which seems to have been written around the chipping beneath it. The last two words on the line are difficult to discern in the photographs, and the solution proposed for them, while appearing to conform generally to the shapes visible, is not entirely satisfactory. The word rendered here as yujanninu appears to have two barbs between jTm and final nian in the lighter photograph, but only one in the darker. The verb/janna cald (form I) is used in the sense of "to overshadow" or "to cover (something) with darkness" in Q. 6:76, and form IV (a/anna) can also be used in this way. My reading, however, requires that form II (or the emphatic of I, yajunnanna) be used and must remain tentative. Probably only close examination of the original inscription will clarify this passage. 3. Unlike the preceding line, the readings here seem quite straightforward. The inscription appears at this point to be written around the designs scratched in the rock surface, and hence to be later than those designs. On the other hand, the fact that part of man near the beginning of the line has been effaced suggests that further flaking or chipping of the surface took place after the inscription was written. The mark at the far left of line 3 may or may not represent the beginning of another word which is mostly effaced by flaking. 4. The way in which the line is curved around the designs above it again implies that the designs are earlier than the inscription. As in line 2, many of the words here do not resolve themselves into readily acceptable readings, and satisfactory solutions may have to await examination of the inscription itself. The word istamadda occurs in the dictionaries but is not attested in the Qur'an--nor is any other plausible reading of the apparently clear letters (istqfida/yastafid, etc.). The word following IT in this line is even more vexing. The article and initial letter tda, with diacritical dots, seem clear enough, as do the two barbs and final nin. But none of the possible readings that can be derived from these apparently clear letters (of which the proposed tabaiyyun is by far the most common word) makes much sense of the sentence. The maca-hiC at the left of the line is very tentative; the other words, however, being written on a smoother spot in the stone, are more secure. 5. If the text is indeed to read "his defender" (hamT-hT), the scribe has dropped the ,yd' that should stand before the enclitic pronoun. After yadhidu can one would expect, of course, something in need of protection; is hami-hT an error for hima-hCi, "his private pasture"? The reading bi-yad toward the end of the line is conjectural. 6. The last word before the break seems legible enough but can hardly be satisfactorily restored without further context. The text of this inscription, being less formulaic than most, is among the most interesting of all in this collection. Despite the difficulty encountered in reading several parts of it, the general structure of the text is readily grasped. It takes the form of a series of parallel religious and/or ethical maxims of the form, "he who does A, does/suffers B; he who does C, does/suffers D; .. ." Maxims of this kind are found in the wisdom literature of the ancient Near East34 and in the biblical wisdom literature. Single maxims of this form occur in many passages in the Qur'dn, such as Q. 6:48, "... whoever believes and acts justly, no fear shall be upon them ...
." A few Greek maxims of this form, but different in content, circulated in Arabic translation.35 The closest analogue to W 5 in both structure and content is a set of parallel maxims found in EPRL, no. Z 137 (pp. 87-88), which reads, "whosoever trusts in God, God will spare him; and whoseover is spared by God, God will grant him security ...." Paleographic analysis of this inscription is complicated by the fact that it is inscribed very crudely on a rough stone surface, so that we may be dealing with distorted and atypical letter forms. Generally, however, the letter forms suggest a fairly early date, although probably not quite as early as some of the other inscriptions in this group. Several letter forms are common to the first two centuries A.H., such as those of final rd' as a tight curve near the base line, the angular dal, the isolated kdf, and the medial hd'. Other letters, however, suggest a date after the first century A.H.; final nin, for example, is more curved than those in first-century inscriptions, and initial hd", with its overhang (lines 2 and 5 . 6) example, the word .j•iz, "current," is written in reverse, and in a few other instances of copper coins from the second half of the first century A.H. the mint names "Palestine" and "Aleppo" are reversed.39 Such reverse writing on coins may not be fully analogous to the reverse inscription in W 7, however, because we may suspect that on coins the reversal was the inadvertent result of accidentally reversing part of the stamps used in making the die; for one thing, other Arabic phrases on the same coins are read normally, from right to left. The argument that reversed inscriptions on coins are essentially accidents rather than conscious efforts to write from left to right seems strengthened, moreover, by one coin from Damascus in which the initial dal of Dimashq is engraved upside down and to the left of the letter group m-sh-q, which is engraved properly.40 In the case of W 7, on the other hand, it is certain that the decision to write from left to right rather than vice-versa was an intentional one. Whether this phenomenon reflects a phase in the development of the Arabic script so early that even the proper direction of the script was not yet firmly set, or (more likely) merely the whim of the engraver, or even some neurological flaw in his brain, we cannot say.
Despite their reversal, the letters conform closely to the archaic style of writing found in most of the other inscriptions we have examined above. We may note the open medial cayn (lines 3 and 5); the final mTm with short tail (lines I and 2); rd' as a short, tight hook; and the final nins of lines 3 and 5, with their straight, sloping stroke ending in a fairly sharp hook. The final nins of lines I and 4, on the other hand, are more smoothly and fully curved. Trailing diagonally across the rock face, this inscription is too small and indistinct in the photograph to read beyond the initial words.
Paleographically, the letters visible appear similar to those of W 6 and W 7. The final nan of al-rahman, barely visible, appears to be short and pulled under the line to the right, with no hook, closer to the Syriac nan than to the usual Arabic form, even of an early date. I know of no other published inscriptions where this form is attested, but it is found in lines 5, 6, and 8 of the papyrus PERF, no. 558, which is dated