West Semitic Syllabary


In the ruins of the ancient Phoenician city of Gubla or Gebal (known to the Greeks as Byblos) written evidence shows that the Phoenician alphabet was used there, in the Iron Age; the two inscriptions from the tomb of King Ahiram are famous examples.

With regard to the Bronze Age, a number of documents on stone and metal testify to the use of another script at Gubla: it has dozens of characters (whereas the Phoenician alphabet had only twenty-two letters). It has signs that closely resemble Egyptian hieroglyphs, and it is therefore commonly known as the Byblian pseudo-hieroglyphic script; this is not a complete misnomer, but examples of this writing system have turned up not only in Lebanon but also in Syria-Palestine and Egypt; and the 'pseudo' element seems harsh, suggesting that this is a phony phonic system. The inventory of characters was clearly constructed from borrowed Egyptian hieroglyphs, as was the alphabet, and seventeen of the seventy or so signs were the same as in the Phoenician alphabet. The number suggests that this was a syllabary, with three vowels represented (as in Arabic: a, i, u) and about twenty-two consonants (as in the Phoenician and Hebrew alphabet). 

Logo-syllabary is the word I apply to this script, because (in my experience of it) the signs could be employed not merely as syllabograms (for example: a door representing DA, from daltu 'door'), but also as logograms (a door representing the whole word daltu 'door'). 

The proto-alphabet functioned in a similar way, as a logo-consonantary, with the letters representing consonants, as 'consonantograms' (G, a boomerang, gamlu; P, a mouth, pu; N, a snake, nakhashu); and each letter could also act as a  logogram for the word from which it was derived,  (B could stand for 'house', baytu); but there was a third function that I have detected in the proto-alphabetic inscriptions: all the consonants in the word could be used in forming other words (NT [snake + cross] = n-kh-sh-t 'copper').

Both the syllabary and the consonantary were built acrophonically: they were based on the acrophonic principle, whereby the top part of the word (Greek akron 'peak') provides the phonic component, this being the initial consonant for the  alphabetic letters (Wawu 'hook'), and the first syllable for the syllabograms (WAwu). Thus the signs in both systems are 'acrophonograms'; but they also used the rebus principle, to function as logograms and 'rebograms' (constituent parts of words).

The language they were designed for was Canaanite, the West Semitic language of Kana`an (Syria-Palestine), the ancestor of  Phoenician and Hebrew. That is the language found in the texts available to us, in my readings of them.

However, all the handbooks on writing systems insist that the proto-alphabet and the syllabary are still 'undeciphered'. I humbly beg to differ. In my opinion their inscriptions can be understood. 

We have an ancient copy of the proto-alphabet from Egypt: http://cryptcracker.blogspot.com/2006/07/alphabet-when-young-above-is.html

Most of the West Semitic proto-alphabetic signs were indeed Egyptian hieroglyphs:

The proto-alphabetic inscriptions from the Sinai turquoise mines are able to be interpreted:

George Mendenhall published a credible decipherment of The Syllabic Inscriptions from Byblos in 1985. Critics were sceptical because he did not explain fully the steps he had taken to achieve his results.

I endeavoured to refine his work and to interpret all the documents,
in the journal Abr-Nahrain (Ancient Near Eastern Studies), from 1993 to 1998.

My latest statement on the West Semitic logo-syllabary is here:

None of these articles are easy reading, but they could eventually prove to be significant. 

I have now produced a booklet setting forth my own decipherment attempt, as a prologue to my published studies.

Download the syllabary document here.
[PDF, 1.6MB, 37 pages] (Right click; on Mac, Control-click to download file to your computer.)

There my approach is to examine the longest document, Tablet D:

[1] Recognize the signs of the proto-alphabet that are already in the
syllabary, and put their consonantal values into the text.

[2] Then, on the probable assumption that the syllabary is an acrophonic
system, as is the proto-alphabet, I look for Egyptian hieroglyphs among the
characters (example: the peculiar symbol for 'night') and match a West
Semitic word with it (laylu, to give LA), and compare its relative frequency
with the same syllable in Ugaritic texts (Ugarit 8th, Gubla D 5th). Another
example: the most frequent sign is a sickle, which would go with maggalu,
and MA is the most frequent syllable on the Ugaritic frequency table. These
plus K (the character that looks just like K) detect the M-L-K root (king and kingship).

[3] Animals, and body-parts, and other objects are matched with WS words.

[4] Meaningful words and phrases emerge, and the text finally yields most of
its meaning (requiring the recognition a few signs functioning also as
logograms, as in the proto-alphabet).

[5] The resulting table (which is very close to George Mendenhall's results)
can then be applied to all the other available texts, not only from Byblos
and elsewhere in Syria-Palestine but also from Egypt and Italy, and even Jamaica.

The inscription from northern Italy was discovered by Giovanni Garbini,
who has certainly learned to recognize West Semitic syllabic inscriptions,
and without his discoveries I would be short of material to work on.
But he seems disinclined to get inside the system, and test
Mendenhall's results himself. However, to his further credit, his critical
review was the only one that looked closely at specific details of
Mendenhall's transcriptions of the texts, rather than merely sniping from
the sidelines. I think I answered all his points satisfactorily in
Abr-Nahrain 35 (1998) (The Canaanite Syllabary, 28-46): 29, 38-44.

Here is his present position, as stated in his *Introduzione all'epigrafia
(2006) 65: "... nel 1985 è apparsa una monografia con la
decifrazione suggerita dall'americano G.E. Mendenhall; questa ha trovato un
unico seguace nell'australiano B.E. Colless. La datazione delle iscrizioni
al XXIV sec. a.C. e un riduzione, del tutto arbitraria, del numero dei segni
attestati non depongono a favor del tentativo di Mendenhall."

Thus, the proposed decipherment of the American Mendenhall has only found
one follower, namely the Australian Colless.
[For those who might wonder about me: born 1936, Australian citizen,
and a committed resident of Aotearoa/New Zealand, non-salaried reasearcher
attached to Massey University, Palmerston North campus.

Is there no one else prepared to take a stand on this issue?

Here are Garbini's two stated reasons for doubt, both of which have
collapsed, I think:

[1] Number of signs
Dunand's edition lists over a hundred signs, but when you look closely at
the characters, it is not hard to detect graphic variants of particular
characters. This culling is not 'arbitrary', not random, but judicious.

As soon as we recognize the bee (NUbtu) we know A5,6,7,8 are the same sign.

The eagle-vulture (apparently RUh.amu) has many variant shapes A1-4[wrongly
A14 on my table on the pdf!], A9 (recognized as a variant by Dunand), and a
simplified alternative form (A11, A19, E9) that goes back to the Egyptian
hieroglyphic system; Mendenhall discovered this equation through his method
but did not know of the Egyptian connection.

That takes eight or so away from the total.

My results reduce the table to 62 (with some syllables still lacking), three
times the size of the Phoenician alphabet (3 x 22), suggesting that Dh and
Z, H. and H_, Sh and Th, T. and Z.,`ayin and Ghayin, were not distinguished.

      (Click on each image to see an enlargement of the tables)

[2] Date of invention
Garbini dismisses Mendenhall's date of 24th C. BCE, and there has been no
means of verification for that until recently. Now from the city of Tuba
(Umm el-Marra) in Syria we have a few examples of the script (c.2300 or even 2350 BCE)


My account of this important evidence is here:


Brian Colless (January 2009)