Monday, March 19, 2007


(Revised July.2015 and April 2021)

These little scraps of clay may be the bearers of the oldest-known West Semitic (Canaanite or 'Canaanian') inscriptions. They were found in Syria (at Umm el-Marra) and are dated to the Early Bronze Age (3000-2000), possibly in the 25th or 24th Century before the current era (BCE = BC).

The claim I am making for them as being the "oldest" Canaanian inscriptions relates to the script they are written in. West Semitic anti-snake-spells are said to be found in an Egyptian pharaoh's tomb, written in hieroglyphs; but I will argue that these examples are in the script designed especially for the Canaanian language.

The signs on the 'cylinders' are not numerous, and unfortunately two of them have the same text (2 and 3 : top right and bottom left), though one of them is damaged in the middle. It is not clear from the photographs whether any of the pieces was originally joined to any of the others. And we must lament the fact that no other such documents came to light from the same spot. Their broken state suggests that the tomb was invaded and these objects were deliberately smashed; they were made to have a cord threaded through them, presumably to be worn on the body as amulets; if so they were ineffectual as charms against harm; my reading of the inscriptions ("Saved" and "Established") implies that they were tickets for safe passage to the afterlife, miniature versions of "The Book of the Dead" that elite Egyptians carried with them to their last judgement, and entry into the Underworld.

The first question is: which way up? I am not sure, because even though I think I recognize the characters, many of them can be reversed and still be the same letter. The circle, for example (an eye with a dot for the pupil, or else the sun) does not have a top and a bottom. The /|\ would be a hand with three fingers shown; but it could still be the same object when inverted \|/.

From my experience with West Semitic scripts from the Bronze Age (see The Alphabet when young) and the Iron Age (see An ancient abagadary) I can tell that the script is Canaanian; but it could either be the proto-alphabet or the proto-syllabary. The syllabary has three times the number of signs as the proto-alphabet (because its characters are syllabograms, representing a consonant with one of the vowels a, i, or u), but they share many of their signs. Thus in the syllabary a picture of a snake represents NA (from nakhash 'snake') but simply N in the proto-alphabet; and the snake-sign is present in these inscriptions.

Assuming that 2 and 3 are the right way up, the first and the last letter in each could be a snake. But my choice would be to take the one on the right side as a cobra, and the bent line on the left side as a tusk, which is the sign for NI in the syllabary (from *nigh.atu).

The hand /|\ is usually in this stance in the syllabary, as KA (from kap 'hand'); it is also in the proto-alphabet as K (Greek Kappa).

The letter next to the K(A) is apparently the hook-sign, W(A) (from waw 'hook').

So my proposed reading is NIKAWANA or NAWAKANI.

After many years of labour on the documents from Byblos (in Lebanon), George Mendenhall was able to publish what I think is a plausible decipherment, in The syllabic inscriptions from Byblos (Beirut 1985). He found the sequence KAWANA in the long text on tablet D, and recognized it as the verb 'be' in Phoenician. Here, NIKAWANA could be the passive voice of the verb 'be' (a puzzling concept for English speakers), but it could mean something like 'it is established'.

To the left of the NI is another sign or part of a character: its shape is ) .

The other brief text is not so easy to explain. I am not sure how many characters there are in it.

On the far right we can detect a character resembling a reversed E; it is in fact the origin of Greek and Roman E, but it was a depiction of a person jubilating, with arms raised, and it says HI or H, from the root hll, 'celebrate', as in Hallelu-Yah, 'Celebrate Yahweh'. Notice the middle line (which seems to have a circle representing the head) extends to the right (the body of the person) and the stroke on fragment 4 might be the completion of that character.

If the bent line on the right has a vertical stroke projecting from the top, then it is another WA. However, if there is no projection then the sign as its stands would represent `U (the consonant being `ayin, and its origin is in the Egyptian character for 10 ).

The circle could be an eye, as already suggested, but in the syllabary and the proto-alphabet the eye is always depicted with an oval shape; the circle (with or without a central dot) was the standard form in the Iron Age, eventually becoming the letter O (a vowel, no longer a consonant) in the Greek and Roman alphabets. In the syllabary (by my calculations) the circle is the sun (shimshu, hence SHI); the Egyptian hieroglyph is N5 in Alan Gardiner's standard sign-list in his Egyptian Grammar; it regularly has a dot at its centre. This pair could thus produce a word shi`u, which might be related to a mysterious and rare West Semitic word meaning (apparently) 'incense' and 'perfume offering'.

The conglomeration of strokes to the left of these could be a single character. It could be Tta (t.a, my doubling of the letter indicating emphatic t, Hebrew and Phoenician Tet), originating in the Egyptian sign for beauty and goodness (Semitic t.ab 'good'). So, perhaps we have a label advertising 'good perfume'. And the other pieces give authentication to this: 'Guaranteed' ('It is established'). But it could be something entirely different, such as a person's name.

There is a possibility that we are in the presence of the sign NU, a bee (nubtu), hence nushi`u. This might mean 'saved' (from y-sh-`, the 'help and salvation' root that ultimately gives us the name Yeshu`, Jesus).

But the rear end of the presumed bee could be a separate character, an inverted Y, which is YI in the syllabary (from yimnu 'right hand'). The head of the bee might be human, and therefore RA (ra'ish 'head'), also alphabetic R . The Y might be BI (on its side), an eye with strokes radiating from it to depict teardrops (bikit 'weeping'). This view gives us RABI (as in Hebrew Rabbi): rab means 'great', 'much', 'chief'.

The possibilities are multiplying, and if the head is in fact attached to the horizontal stroke it becomes a rump with a tail (zanab), hence ZA, and a word for 'this'.

There is an important principle that I have not mentioned so far: from my experience of this writing system I see it is a logo-syllabary, which means that the signs can also function as logograms, so that they can represent the whole word that goes with the picture as well as just the first syllable. The snake-sign can be NA or the word nakhash. The same rule applies in the proto-alphabetic inscriptions, I find: N can also stand for nakhash (and also any word which has the same consonants n-kh-sh). Accordingly, we have to test each sign for the possibility that it is functioning as a logogram, not simply a syllabogram.

So, each character along the way in these sequences has to be tested for the possibility it is acting as a complete word, not just a syllable in another word. I can not see an example of this in the texts we are looking at here, but I may be overlooking something. For example, the HI could be a whole word, HLL: 'Celebrate!'.

If we are prepared to accept that this really is the ancient West Semitic syllabary (and the NI tusk shows it is the syllabary and not the proto-alphabet, as also the NU bee), and if we allow the estimated date of 2500 BCE, then we may make the claim that this is the oldest-known Canaanian inscription in a West Semitic script.

As noted earlier, a Semitic anti-snake spell (probably imported from Gubla/Byblos) has been recognized (by Richard Steiner) amid the writing on the wall inside the pyramid of Pharaoh Unas/Wenis at Saqqara, dating from the 24th century BCE. It is West Semitic language transcribed in Egyptian hieroglyphs. This is being hailed as possibly the most ancient Semitic inscription, given that the East Semitic cuneiform documents from Mesopotamia (`Iraq), and the North(?) Semitic cuneiform tablets from Ebla (Syria), are not earlier than the 24th century BCE.

If these new mini-tablets from Syria are really datable to the 25th century BCE, they are the-oldest known Semitic texts written in a Semitic script; but this would still apply if they belong in the 24th century (Schwartz 2021, 256-257), though the copper tablets from Gubla/Byblos are difficult to date, and might be older than the Tuba texts.

Further, this Early Bronze Age date for the West Semitic syllabary means that it was presumably the model for the later syllabaries of Crete (a set of acrophonic picto-syllabograms which developed into Linear A and Linear B) and Anatolia (the Luwian syllabary).

For background, see the Wikipedia article on the Byblos script, edited by HvD, which cautiously treats the problem as still unsolved (though not disregarding the views of Mendenhall and myself):

However, it is not simply the 'Byblos syllabary', since examples of this script have been found in Megiddo, Sinai, Egypt, Italy,  Jamaica, Puerto Rico, and Norway; and these additional inscriptions can be used to verify Mendenhall's decipherment.

See Brian Colless, The Canaanite Syllabary, Abr-Nahrain/Ancient Near Eastern Studies 35 (1998), 26-46. This lists most of the known West Semitic logo-syllabic inscriptions, and indicates where I have commented on each of them (in issues 30-34 of that same journal); it also provides a table of signs with my proposed sound-values for them (agreeing for the most part with Mendenhall's proposals).

My most recent work on the West Semitic logo-syllabary is available on a pdf at:

Here I would like to insert a point that Hoch makes in an article  (also the author of Semitic Words in Egyptian Texts of the New Kingdom..., 1994).

Hoch, James E. (1990). "The Byblos Syllabary: Bridging the Gap Between Egyptian Hieroglyphs and Semitic Alphabets". Journal of the Society for the Study of Egyptian Antiquities 20: 115–124.

Hoch (1990) points out that many of the signs seem to derive from Old Kingdom hieratic (simplified characters), rather than directly from hieroglyphic; as early as 2600 BC Egyptian influence in Byblos was strong.

Finally, some notes on the provenance and significance of these "perforated clay cylinders with written (?) symbols from Tomb 4, upper level".

The source for the pictures is:
That website reports on the Umm el-Marra project, directed by Glenn M. Schwartz (Baltimore) and Hans H. Curvers (Amsterdam).

Umm el-Marra is on the Jabbul plain in Syria, east of Aleppo and Alalakh, north-east of Ebla and Ugarit, north-west of Emar (on the Euphrates), and distant from Gubla/Byblos. A likely identification is TUBA, a city mentioned in the Ebla archives (24th C. BCE) later known in texts from Mari, Egypt, and Anatolia.

There are at least 6 brick-tombs in a necropolis on the acropolis of the city, dated 2500-2300 BCE. Possibly a 'royal' cemetery, the tombs contain skeletons of men, women, and infants, in wooden boxes, with equids and other animals buried in adjoining brick-installations.
Tomb 1 is dated c. 2300 (pottery = Ebla palace G finds)
Tombs 5, 6: c.2500 BCE
Tombs 2, 3, 4: no date suggested

If these objects are funerary amulets for the male personage in the tomb, then my proposed readings (made on the basis of no context whatsoever) now sound quite attractive:
NUSHI`U "saved"
HLL "Celebrate"
NIKAWANA "he is established"
This reminds me of the Egyptian Book of the Dead, with the deceased passing through the hall of judgement and emerging "justified of voice" (his claim to be innocent has been accepted by the divine judges).

However, there is still a faintly possible place for my "good incense" guess, since incense is mentioned as an ingredient in the mortuary rituals in the Pyramid Texts, the spells written on the interior walls of royal pyramid-tombs in Egypt (the anti-snake-spell is a Semitic example, as already noted).

It does not really matter whether these are the oldest-known Semitic inscriptions (if the tomb they were found in was not 25th century BCE but 24th, or even 23rd, which is the likeliest possibility): the significant point is that the West Semitic syllabary has now been attested in the Early Bronze Age (Late EBA), and thus prior to the proto-alphabet (the West Semitic consonantary, not attested before the Middle Bronze Age, 2000 BCE) as well as the Cretan and Luwian syllabaries; and so it may have inspired the invention of these three systems.

The time has come for reconsidering my earlier hypothesis on the 'evolution' or 'creation' of the West Semitic logo-syllabary (leading on to the proto-alphabet) as being influenced alike by the Mesopotamian cuneiform script (a vocalic logo-syllabary, with vowels included) and the Egyptian hieroglyphic script (a consonantal logo-syllabary, no vowels indicated). The use of the cuneiform system in Ebla involved reduction of the number of syllabic signs, and this possibly influenced the formation of the economical West Semitic pictosyllabary, found at Gubla/Byblos, and now also at Tuba (Umm el-Marra). However, if the West Semitic syllabary came before this simplified cuneiform, then it may have started the move towards simplification in the region, which ultimately resulted in a consonantal alphabet (the proto-alphabet, with acrophonic pictophonograms), and a cuneiform alphabet based on the characters of the proto-alphabet.

See further, Brian Colless, "The Egyptian and Mesopotamian contributions to the origins of the alphabet", in Cultural interaction in the ancient Near East, Abr-Nahrain supplement series 5 (1996) 67-75.
Brian E. Colless, The evolution of the Alphabet,

Glenn Schwartz (2010)  Early Non-Cuneiform Writing?: Third-Millennium BC Clay Cylinders from Umm el-Marra
in: S. Melville and A. Slotsky (eds.), Opening the Tablet Box: Near Eastern Studies in Honor of Benjamin R. Foster, Leiden: Brill, 375-95
(Schwartz is dismissive of my reading of the inscriptions, on the simplistic grounds that he does not think their signs look like the characters on the Byblos inscriptions we have.)

More recently, he has raised the question again, rejecting syllabic and promoting alphabetic:
Schwartz, Glenn M., 2021    Non-Cuneiform Writing at Third-Millennium Umm el-Marra, Syria: Evidence of an  Early Alphabetic Tradition?  Pasiphae XIII: 255-66;
Christopher Rollston has supported him:

My response to both of these scholars is contained in this large essay:

Richard C. Steiner (2011), Early Northwest Semitic Serpent Spells in the Pyramid Texts, Winona Lake, Indiana.

My gratitude is expressed here to Judith Weingarten (who first informed me about these important objects from Umm el-Marra), and to Beatrice Hopkinson and Rochelle Altman (who passed on the address of the website to me).

No comments: