§6 Sabaean Inscriptions

The Sabaean alphabet is given below in Figs.5a,b. Now that we have the Sabaean alphabet, let us return to the Sabaean inscription in Fig.2 (in the section Code Breaking).
We see that the second row is just:

s¹flyn

That seems close enough to Safalia!

Note that the consonant y is apparently being used as a vowel. This is done in English too: y in "yes" is a consonant and y in "easy" is used as a vowel. In Hebrew the letter for y, namely yod, is a consonant but is sometimes used as a vowel. But we are speculating here: perhaps the y in the Sabaean inscription is really used as a consonant and the vowel that goes with it is just not written, as they often are not in semitic languages.

Note also from Fig.5a that we there are 4 different forms of the letter s in Sabaean, each with somewhat different pronunciation. I don't know the difference between them, but it is not unusual for a language to have different s sounds. In English, the s in "sing" is different from the s in "easy", but are represented by the same letter, even though the letter z could also represent the second of these s sounds. Some other languages use different letters for the different s sounds. For example, there are five s sounds in Hebrew, namely sin, shin, samech, sade and zayin. Jenni [Jenn] transliterates these five sounds respectively by . The transliterations of s sounds in the Sabaean alphabet in Fig.5a are respectively s¹ s² s³ , and possibly also z and . I don't know whether the systems for transliterations are completely standardized, so I can't assume that, for example, the sound transliterated as z in Hebrew is the same as the sound transliterated as z in Sabaean, or that the sounds transliterated as are the same. Jenni [Jenn] doesn't say whether he is using a standard set of phonetic symbols for transliteration. Nor does the Sabaean dictionary [Bee]. How, then, can we compare these s sounds or compare the Hebrew alphabet with the Sabaean alphabet?

I would also like to refer to Fig.4 in §5 and to point out that the script identified in the time line as being Old South Arabian script looks remarkably like the letters of the Sabian alphabet. This is no coincidence: look in Fig.4 at the Old South Arabian alphabet. The first 5 letters given there are (reading from right to left, of course) are in the Sabaean alphabet as given in Fig.5b, and are respectively ' b g d h. An examination of the various alphabets in Fig.4 shows that in each case, the first few letters of each alphabet are given in the alphabetical order of that alphabet. From this we conclude that the Sabaean alphabet as given in Fig.5a is not in alphabetical order. Instead, it seems to be given in a logical ordering according to the phonetic values of the letters. Note also that ' b g d h would also be a perfectly good way of transliterating the first five letters of the Hebrew alphabet. Finally, note that the Sabaean letter which we identified as Dh is transliterated in Fig.5a as . Thus, we have also taken a step towards deciphering the phonetic notation that is being used in [Bee]. Finally, let us contemplate the fact that the alphabet which in Fig.5b is called the Mino-Sabaean alphabet is called the Old South Arabian alphabet in the time line. This suggests that there was a general South Arabian alphabet which was used by various peoples, including the Sabaeans. Actually, there is some information on this in [Jens], p.337. There we are told that south-Semitic inscriptions are written in 5 basic kinds of scripts which fall into two main groups:

(1) North Arabian scripts, which include (a) Thamudic (b)Safatene (c)Lityanite
(2) South Arabian scripts, which consists of (a)(Mino-)Sabaean (b)Old Abyssinian, together with Ethiopic

So Sabaean is one of various South Arabian scripts.

Thus, having access to the Sabaean alphabet made it possible for me to appreciate more fully the MMA exhibits in Figs.1, 2. In general, we can look forward to similar benefits from suitably chosen reference materials, but at the same time we are reminded that every source of information also raises new questions.

In Fig.6 above, I show a drawing of an inscription from the catalogue [Lou] which, according to the description, is written in Sabaean. Using the Sabaean alphabets above, can you verify that it says Baray shams?

According to the catalogue, this incense burner contains the inscription, "baray Shams, offered to the sun goddess". How do you think they arrived at this translation when all the inscription seems to contain is the name "baray Shams"? First see if you can guess.

If not, keep reading. Surprisingly, the answer was found after further study of the artifact displayed in Fig.7 and will be discussed there.

In Fig.7 below is a stele of a camel driver. Using the Sabaean alphabets in Fig.5 above, can you identify the letters in it? Which of the two versions shown there is most representative of the letters you actually see in the inscription? Is there one of these alphabets you could do without entirely in identifying the letters of the inscription?

One of the wonderful things about [Lou] and many other catalogues is that it contains bibliographic references to the scholarly literature. For the stele in Fig.7, the description contains a bibliographic reference to the following source: "[CIS] IV 445". This reference didn't make any sense to me, but when I consulted the bibliography on p.111 of [Lou], I learned that [CIS] stands for "Corpus Inscriptionum Semiticarum" and that it is published by the Académie des Belles Lettres in Paris.

I called up the Académie des Belles Lettres and asked where I could see a copy of the book. They told me to go to a bookstore named Boccard on Rue Medicis by the Jardins du Luxembourg. When I arrived I learned that the Corpus Inscriptionum Semiticarum consists of dozens of volumes. The proprietor was willing to show me some volumes, but asked if I could be somewhat specific. I showed him the reference to [CIS] IV 445 and he couldn't tell from that what volume was being referred to.

As the title suggests, [CIS] is written in Latin. It is in 5 parts, each part being subdivided into volumes, each volume being subdivided into two books, one full of pictures (planches) and the other containing learned commentary on the pictures (text). Eventually, with essential help from Mary Wolinski, I was finally able to examine some of the volumes in the Bibliothèque Nationale and in the Bibliothèque d'Art et Archeologie in Paris.

The stele in Fig.7 appears in Plate XIII of [CIS] IV Part IV Tome II and is discussed there in Article XIV, which is entitled ', pp.143-144. I didn't know what "" was but from the translation of the inscription given in [Lou] and [CIS] IV 445, I concluded that it was one of the local gods.

Fig.7
Stele of
a caravan
leader, 2-3
century AD
Drawing
based on
[Lou], p.38
Cat.# AO 1029

The discussion in [CIS] IV is quite detailed. It is written in Latin but uses Sabaean, Arabic and Hebrew fonts for Sabaean transcription, comparison with Arabic words and transliteration into Hebrew phonetic values, respectively. Finally, it uses Latin for the translation of the inscription. It gives the precise dimensions of the stele as 55 cm high, 29 cm wide and 9 cm thick. It compares the stele with others in the volume. It says the stele was first studied in the articles of [Mor], which also contain an accurate drawing and a detailed description, which they quote.

In the picture we notice a stringed instrument. In [CIS], there is a reference to [Al-H](a) and [Mas] for a discussion of it. The sources are not merely cited but criticized. For example, they say they can't see in the picture something that Mordtmann claims to see. They say that the artifact was next studied by J.and H. Derenbourg, [Der](a),(b).

They transcribe the text in Sabaean as follows:

Incidentally, the vertical lines that appear every few letters are separators between words. They were not added by CIS; they actually occur in the original inscription, as you can easily verify. I tried to transcribe it myself and was able to do most of it, but I had trouble with some of the letters. After comparing the troublesome letters with their transcription, I was convinced of what CIS said except in the case of the letter(s) at the beginning (i.e. right) of the top row, which apparently involved some restorative guesswork.

For reasons that are not clear to me, CIS transliterates all of the inscriptions into Hebrew letters. I noticed the same practice in the book [Bu1] for giving the phonetic values of Egyptian hieroglyphs. Apparently, Hebrew phonetics are treated as a standard for Afro-Asiatic languages, which include all semitic languages as well as ancient Egyptian and other groups. But is that really a good idea? There are more letters in the Sabaean alphabet than in the Hebrew alphabet and apparently more sounds. Of course, there are variants of the Hebrew consonants, indicated by adding dots to them, and CIS does appear to use them. Perhaps that is sufficient, but the matter is not clear to me at the moment. What is the system used by CIS for transliterating Sabaean and other scripts into Hebrew phonetic values? If you find out, please let me know.

I did try naively to transliterate the passage into Hebrew phonetics and you might find it challenging to do the same. For most of the letters, there was no difficulty, but for others I had to examine their answer. Rather than repeat it without comprehension, I prefer to transliterate according to the phonetic values in Fig.5a, which are perhaps more useful for an English speaking reader anyway. I obtain:

Looking naively at this, knowing nothing of the Sabaean language, the word bn stands out in the first line. This is a typical semitic word meaning "son of". One finds it in Arabic as "Ibn" and in Hebrew as "ben". From this, it is tempting to suppose that this has something to do with "`glm, son of s<sup>1</sup>`dlt". Now let us look at their translation. I suppose that knowing that they transliterate Sabaean into Hebrew phonetics, it will not surprise anyone to learn that they also translate it into Latin. Their Latin translation is as follows:

We will tentatively translate the Latin into English as:

Thus, our naive guess regarding the first line seems to be on the mark! The name `. is recognizable in the second line without its vowels as .

This translation is consistent with the synopsis of the inscription given in [Lou]. The transcription, transliteration and translation are discussed minutely in [CIS], which cites references to other authors who have made critical studies of the text. For just these two lines, they cite [IbnH], [Al-H](b), [Der](c), [Web], [Hom] and [Hal]. I have not yet succeeded in making sense of some of these citations, although I have entered them into the bibliography ofthis article. Any help would be welcome.

Let us pause for a moment to consider our progress. We began by drawing pictures of artifacts in museums, noting the catalogue numbers and trying to check the blurbs. This led us to consider the museum catalogues in the hope of resolving details we were unable to check ourselves. The catalogues in turn referred us to more scholarly literature, such as the great compendium [CIS]. There we found that the details that were not obvious to us were part of a discussion long in progress in the published literature and obtained pointers to the texts of that discussion. We have left behind the world of grand synopses which paint the world in a broad brush and turned our magnifying glasses instead on the fine brushstrokes. We have, in effect, caught a glimpse of one small chamber of the labyrinth, from which we can discern the openings of innumerable passages leading elsewhere.

It seems that the paths multiply endlessly, and they do, but one also learns to find one's way around as some of the different paths begin to converge. For example, one of our rewards for pursuing Fig.7 into [CIS], and thereby becoming acquainted with [CIS], is that we can now address the question of the sun goddess in Fig.6 posed earlier.

In the top half of the incense burner in Fig.6, we saw a picture in the center. It seems to be a person wearing a gown of some kind, with hands outstretched towards the sky and with a circle for a head, disconnected from the body. At least that is how it appeared to me when I saw it, but being accustomed to the diversity ofways in which artists represent their subjects, I simply accepted it as a human body.

But then while I was reading in [CIS] about the stele in Fig.6, I found a similar artifact in [CIS] in which the figure on the top is much simpler. A drawing based on the picture in [CIS] appears in Fig.9, where we see the sun and the crescent. The drawing in Fig.6 is clearly a variation on this theme. This common image is what pertains to the sun goddess. But this too is only the tip of an iceberg, or so I imagine.

Fig.9
Another incense burner for the sun goddess.
Drawing of [CIS] II 362 based on [CIS] Pars
II Tome I Plate XLII Sabaean inscription

§7 Nabatean Inscription

In Fig.10 below, I show an inscription which the catalogue [Lou] says is written in Nabatean. In Fig.11, I show the Nabatean alphabet. This is my primary reference for Nabatean. Using it, I am myself unable to confirm that this is Nabatean, nor that it contains the name Tawelos. Quite by accident, however, I found that this artifact is discussed in [CIS], even though the blurb in [Lou] makes no reference to [CIS]. It is [CIS] II 167, appearing in Plate XXIII and its discussion appearing in Vol.1, p.199.

Fig.10 Basalt basin, betw. 1st cent. BC and 1st cent. AD
Nabatean Inscription: faithful Tawelos
Drawing based on [Lou], p.61
Catalog number: AO 4993 of [Lou], p.61

Fig.11 The Nabatean alphabet
Drawing based on [Jens], p.322

The discussion in [CIS] relies essentially on [Vog]. According to the English in italics in [Lou], the basin was offered by Tawel to the temple of Beelshamen at Si'a. The longer description in French is a little more cautious, saying that it was probably left as an offering at the sanctuary of Si'a by the faithful Tawel. But [CIS] is the most circumspect of all. It transcribes the inscription as:

It transliterates the text in Hebrew phonetics as

The first word of the inscription is partially missing. All we can see of it are two letters and a fragment of a third one. They say that [Vog] completes this to the word , which they say means (in Latin) "obtulerunt", which we recognize as the perfect of the verb offerre.

So we see how the notion that this is an offering arose: it is based on Vogüé's guess as to how to complete the first word. The next four letters , they say, constitute the name Tawel, which they describe as a name with -el ("nomen cum "). As for the last two letters visible on the artifact, [CIS] says that perhaps ("forsan") it is like a certain arabic word meaning "obedient". From this, we apparently get the "faithful" in the description in [Lou], i.e. from a guess in [CIS]. Of course, [CIS] states it as a guess, whereas [Lou] adopts it and proclaims it unquivocally as a fact. [CIS] also mentions another guess as to the last word, this one offered by [Vog], but rejects it as "improbable". It appears that one is better informed by reading [CIS] than by reading [Lou], at least in this instance. But we must also acknowledge that we would not have known about [CIS] if not for [Lou].

§8 Thamudic inscription

The Old and New Thamudic alphabets are given in Fig.12b below. Note that each letter often has several variants.

Fig.12b The Thamudic alphabet, based on [Jens]

I am unable to confirm that the letters of the inscription in Fig.12a are Thamudic letters. It is alleged that the inscription says: "by Sagag, son of 'Abd-Osiris, son of Qursan". I can't confirm that either. If you can do so, please let me know. This inscription is cited as an example of religious syncretism. Syncretism refers to the fusing of elements of different religions. For example, a Channukah bush might be regarded as an example of syncretism. In the case of this semitic inscription, the foreign element is the name of the Egyptian god Osiris. For more information about this inscription, the catalogue refers us to [Mac].

§9 Why is this Nabatean?

In Fig.13 below, I show an inscription from Tayma currently on display at the Institut du Monde Arabe (IMA).

According to the catalogue [Lou], the inscription in Fig.13 is written in what it describes as a "local variant of Nabatean". In Figs.14 and 15, I show two other drawings. Fig14 is the original drawing I made when I examined the artifact at close range at IMA. At that time, I had no idea what the Nabatean alphabet looked like. In Fig15 is a drawing I made of the plate in [CIS] that pertains to this artifact. That plate is, however, itself a drawing. My own copy of this plate is imperfect, but suffices to show that different copyists looking at this drawing are inclined to reach rather different conclusions. It also suggests that the drawing in [CIS] is not faithful to the photograph. The problem and the experience of transcribing something written in an unknown alphabet gives one some insights into the problems of writing optical character recognition programs. Very simply, if one has some reason to believe one already knows what the possible forms of the letters are, one copies very differently than an artist copying shapes and forms would. The discussion in [CIS] II vol.1 (1889) 299-300 of this tablet ([CIS] II 336) refers primarily to [Eut] and [Hub]. The font used by [CIS] for transcribing Nabatean is quite different from the "font" on the tablet, so their transcription doesn't resemble what is on the tablet. Their font does, however, resemble the one in Fig.10 of §7. This raises the question: how does one conclude that this is Nabatean writing if it doesn't look like the Nabatean writing? [CIS] doesn't address this point. However, if we wish to have conffidence in their conclusions, we must consider this point and to do so, we will have to examine the works [Eut] and [Hub]. This we have not done and will not do in this article.

CopyRight 1999