Saturday, October 20, 2007


Gordon J. Hamilton has theorized on the beginnings of the letters of the alphabet in The Origins of the West Semitic Alphabet in Egyptian Scripts (2006), 433 pages. He argues that all the letters of the alphabet derive from Egyptian hieroglyphs. His main source of data is the corpus of West Semitic inscriptions from the Egyptian turquoise mines in the Sinai Peninsula; these have long been known as the 'Proto-Sinaitic' inscriptions, but both of us have set that term aside; other inscriptions of the same family have turned up elsewhere, and whether they are found in Egypt, Sinai, or Syria-Palestine, they are 'Proto-Canaanite', in language and script. I have added the term 'proto-alphabetic' to the discussion, meaning that they are connected with the 'proto-alphabet', the original prototype of the alphabet. (Note that I prefer to say 'Canaanian' rather than the traditional 'Canaanite', which carries pejorative bias from the Bible.)

In the decades he was working on this thesis (1983-2005) Hamilton was unaware of the existence of more inscriptional evidence from ancient Thebes. We can test his case against this additional material (published by William Flinders Petrie in 1912 as the frontispiece of his book The Formation of the Alphabet). The most important of these is a copy of the proto-alphabet (Thebes 1), reproduced here.

My drawing shows most of what I see amid the faded paint-marks, but I have a bias: I have been working on this question throughout the same period, and I think these tablets give support to my proposed system (though my paradigm was constructed before I knew of the existence of this controlling evidence). My own table of values and identifications for the Egyptian hieroglyphs behind the original signs of the West Semitic alphabet is now available on the web, and the latest version is at the end of this article. Note that in my view not every letter of the proto-alphabet had an Egyptian counterpart; and all certainly did not survive into the Greco-Roman alphabet.

But that is not the only witness we have: another copy of the letters of the proto-alphabet is found on two ostraca from the same collection (Thebes 2 + 3):

At the outset (p.1) Hamilton states his three aims, for studying each of the original letters:
[1] isolating its graphic prototype(s) in Egyptian writing
(D: O31, door; also K1, fish);
[2] ascertaining its typologically earliest graphic forms
(door with post; also fish with fins);
[3] establishing the meaning of its acrophonic name(s)
(dalt, door; also dag, fish).

The examples I have provided show that Hamilton believes two different characters were used for D. In what follows, I will call such doublets 'allographs' (if this word is already known to you with another meaning, then I apologize, and I will change it here if a better term is offered to me). Incidentally, I will argue that the fish-sign is not D but S, and I am not alone in holding this opinion.

An important detail in his case is the acceptance of 'the acrophonic principle' (see his third aim) as the main factor in the formation of the proto-alphabet; the term 'acrophony' ('summit sounding', Greek akron 'peak', phone 'voice, sound') is not found in ordinary dictionaries, but the word 'acronym' helps to explain it; an acronym is a name or noun made up from the initial letters of a set of words (NATO, laser); in acrophony the top sound of  a word that describes the picture or symbol (D from dalt 'door', which still shows its door shape, and its Hebrew name is Dalet, its Greek name is Delta; no hint of a fish). When a syllabary is being constructed, the first syllable of the word is sounded (DA from dalt 'door' in the West Semitic syllabary).

Denial of the role of acrophony in the formation of West Semitic and other scripts was a feature of the I. J. Gelb's manual of 'Grammatology' (A Study of Writing, 1963), a word that is worth retaining, I think; ironically the syllabograms of the Luwian logo-syllabary, or 'Hittite hieroglyphs', which he helped to decipher, are obviously acrophonic. In the footsteps of Gelb, Barry Powell (2009, 175-186, 255) continues to regard acrophony as 'a discredited theory', which is to be discarded. Powell  has a good understanding of the Egyptian writing system (a logo-consonantary, though he insists on calling it a logo-syllabary, after Gelb), and he knows that in the Iron Age the Phoenician consonantary (again he bows to Gelb and regards it as a syllabary) became the Greek alphabet; but  he pontificates on the West Semitic scripts and inscriptions of the Bronze Age from a standpoint of ignorance and inexperience, denying any connection or continuity between the proto-alphabetic inscriptions (from Canaan, Sinai, and Egypt) and the Phoenician script. As for the West Semitic syllabary, the Byblian 'pseudo-hieroglyphic' script, which he accepts as  an undeciphered syllabary, 'with no clear forbears and no successors', he wonders whether Aegean writing provided a model for it (p. 186). We know better than that: the West Semitic syllabary goes back at least to 2300 BCE, and Powell has the Cretan syllabary possibly arising around 2100 BCE (p. 129), and the Luwian syllabary about 1300 BCE; and in my view these are acrophonic systems, and the pattern for them was provided by the ubiquitous Canaanites.

Since Gordon Hamilton first submitted his thesis, at Harvard University in 1985, two new Proto-Canaanian graffiti have been found in southern Egypt, in the Wadi el-Hol, near Thebes, and this discovery gave his work a new impetus. One is horizontal, the other is vertical. The signs are reproduced on the right side of the Sinai-Egypt column on my table.

However, Hamilton has not only overlooked the other inscriptions, from Thebes (see above), but he has also neglected to consider the Canaanite syllabary, which contains many of the signs of the proto-alphabet, and these had been undergoing their own development before they changed from being syllabic to consonantal (for example: BA to consonantal B). These are shown on the BS column of my table.

The cuneiform alphabet has obviously been designed on the basis of the pictographic proto-alphabet, and this is another useful aid in testing our proposed identifications for Proto-Canaanian signs; but Hamilton does not utilize it in his blinkered approach. They appear on the Canaan column of my table.

As already stated, at the end of this article I have appended a detailed table of my own solution to the problem, and the reader could print out a copy to consult as a useful and reliable map while traversing the rough terrain I have created below.

We will work our way through the proto-alphabetic letters, relating Hamilton's opinions to mine, and to the Thebes proto-alphabet tablets. We should start with the uncontroversial characters. If I have found an equivalent 'syllabogram' in the West Semitic syllabary, I will note that in square brackets.

The order of presentation is this:
[1]'A [2]R [3]` [4]Kh [5]Dh [6]L [7]M [8]N [9]B [10]H [11]Y [12]K [13]G [14]P [15]Q [16]Ss [17]S [18]D [19]Hh [20]Tt [21]W [22]Sh [23]Th [24]T [25]Z [26]Zz [27]Gh [28](Dd)

'A[1] B[9] G[13] D[18] H[10] W[21] Z[25] Dh[5] Hh[19] Kh[4] Tt[20] Zz[26] Y[11] K[12] L[6] M[7] N[8] S[17] `[3] Gh[27] P[14] Ss[16] Q[15] R[2] Sh[22] Th[23] T[24]

[1] 'Alep (No 1: p. 29-38) [Syllabic 'A]
Greek Alpha is certainly from Canaanian 'alep ('alpu 'ox'); when A is inverted, the horns and snout are clearly visible. So, aim 3 is covered already: the initial consonant of the word is ' (glottal stop). With regard to aim 1, he says that the letter "derives from the Egyptian hieroglyph F1", which depicts the profiled head of an ox. The tables I have published say the same thing, but in his working out of aim 2, Hamilton tries to find an Egyptian prototype for all the variations of ox-head we see in the Canaanian script. The Egyptian model has horns, an eye, an ear, and little or no indication of a mouth. The two Wadi el-Hol examples lack the ear, and have a clearly delineated mouth. I would prefer to believe that a Canaanian scribe was merely expected to draw a recognizable ox-head. That is what we see on the Thebes proto-alphabet tablets (Thebes 1, top left corner; Thebes 2, in a similar position, but notice it has two eyes, a unique feature, apparently, not found on Egyptian hieroglyph F1; and it has a neck).

[2] R (24: 221-231) [Syllabic RA]
R (resh, ra'sh) "derives from the Egyptian hieroglyph D1", showing a human head in profile. Agreed, but no Canaanian example matches exactly with any Egyptian prototype, and the three Wadi el-Hol heads (1.1, 1.16, 2.4) are all different from one another. It seems to me that all the instances cited are rough-hewn heads obeying no artistic or epigraphic laws. The R on the Thebes 1 proto-alphabet is on the left side, the third sign from the top. It is lying down, having a single-line neck, no mouth, but an eye, and, apparently, a hair line. The presumed R on Thebes 2 is at the bottom of the left side.

[3] `ayin (19: 180-188) [Syllabic `A]
The source is hieroglyph D4, 'eye'. Again, surely the idea was simply to draw an eye, with or without the pupil. The `ayin on the Thebes 1 proto-alphabet is in the bottom left corner. Does it have a pupil, or is that the stem of the cactus-plant (it is meant to be a hand, K) sticking into it? In the Egyptian examples that Hamilton presents, I cannot see a perfect match for the Wadi el-Hol horizontal eye, or the Sinai 345 and Lahun vertical characters (182-183); actually there are two cases on the sphinx (S 345), and the one under the Egyptian inscription has a counterpart in the middle of Thebes 3 [ (-|] (the curved part does not quite meet the straight line).

[4] Kh (4: 57-60)
This is a letter that disappeared from the alphabet, but it derives from V28, which has the value Hh (h.) in Egyptian.  It looks like a twisted thread, a hank of yarn, and I have suggested the name Khayt ('thread') for it. Hamilton argues that it is a wick of twisted flax, and offers a picture of such a wick (Fig. 2.14). He invokes an Ethiopic name for it, kharam. Whatever its precise reference, it did exist, and it can be found in the bottom right corner of Thebes 1, and in the middle of Thebes 3 (below the eye, with some of its lines faded), in both instances apparently with two loops, not three as at Wadi el-Hol; both forms are also found in Egyptian writing, and in the Sinai inscriptions, though Hamilton has overlooked the three-looped example on 376, and has misconstrued DY as Kh on 365b. His assertion on the stance of Kh (59), that the Egyptian hieroglyph and the letter are always upright, is not supported by the horizontal instance on the Thebes tablets.

[5] Dh (16: 145-154) [Syllabic allograph DhA/ZA]
Another lost alphabetic letter. In the Iron Age the sounds dh and z coalesced, and the Z-sign then covered both (holding its position following Waw, while Dh disappeared, allowing M and N to stand side by side). There is no difference of opinion about the character representing Dh: it is a pair of parallel lines, and it is found on the Theban tablet 1 in the first column on the left, above the head (though they are not strictly parallel). As regards its hieroglyphic connection, in 1988 I suggested either Z4 ('duality') or D13 (eyebrow(s), Semitic *dhayp). Hamilton rejects my eyebrow hypothesis, because two wavy lines never occur in the Proto-Canaanian inscriptions; this is true, but I am intrigued by a character on another of Petrie's ostraka from Thebes, consisting of a stroke above an eye (lower left region of the tablet on the left, Thebes 2), though it may be merely an `ayin with an extra flourish. However, the cuneiform alphabet's Dh-sign consists of a simple wedge together with an angle-wedge (as used for the `ayin-sign, representing an eye), and this is a perfect match for that character on Thebes ostrakon 2.

Hamilton relates the duality sign acrophonically to Semitic *dhayn ('these two'). A case where the sign is on the right slant for Z4 (\\) is on the Sinai bi-lingual sphinx, and the scribe 'Asa obviously knew Egyptian (note the hieroglyphic text on the statuette).

[6] L (14: 126-137)
"The letter lamed has only one definite source in Egyptian writing: V1, 'coil of rope'." (Hamilton, 126). Really? Some of us have thought that it was S39, 'crook', but many of the examples do have very curly (or 'coily') tops. L may have had both origins (or neither!), and the two versions would therefore be allographs (alternative signs for the same sound; see notes on S, for example). The root l-m-d (in the letter's name) refers to 'training' and 'learning'; the rope or the stick could both have such an association. The example on the Thebes 1 proto-alphabet is in the same column as the ox (by coincidence!), to the left of the eye, looking exactly like an italic l, and apparently representing a walking-stick rather than a rope. Thebes 2 has a crook for L, bottom line (with one of the many mysterious dots accompanying it). Another document from Egypt shows letters of the proto-alphabet, with a Semitic word accompanying each character, and for its curled L it has rw, which could represent Semitic lawi, which is the Ethiopic name for L, Lawi. Romain Butin (Harvard Theological Review 1932: 146, and Colless, Abr-Nahrain 1988: 44) noted a possible connection with the name Lawa, and he pointed to a root lawa ‘wind, coil’ (as in Arabic); Hamilton records this in a footnote (2006: 136, n. 157) but he rejects the other possibility of a shepherd’s crook (S38, S39) as the prototype (2006: n. 148). However, on the two alphabets from Thebes, neither  L has a coil: the one on Thebes 2 has a crook (like S38) and the other (on Thebes 1) has an inverted example of S39 (looking just like our italic l). It could be that they are allographs: either the coil and the crook are both original, or else one developed from the other.

[7] M (15: 138-144) [Syllabic MU, and MI]
Hieroglyph N35 stands for 'water' (mu, maym), with its zigzag line representing waves (with many peaks and troughs). It is undoubtedly the inspiration for M in the alphabet. M is found on the Thebes 1 proto-alphabet at the bottom of the tablet; start with the drawing, and then look for the thin faint line on the photograph. A search may also be made at the bottom of Thebes 3 (bottom left corner); the double triangle on Thebes 2, top line, resembles an Arabian form of M, with a line closing it off (like M underlined); but this is most likely the B (see 9 below). M is one of the characters that shows that the inventer of the proto-alphabet really did borrow Egyptian hieroglyphs. Hamilton notes there are also vertical forms (N35B) in both systems, and it should be noted here that in the West Semitic syllabic usage, as I see it, the vertical syllabogram stands for MI (presumably from 'rain', and a simplification of N4 representing the sky with four vertical wavy lines for the rain). Hamilton (143) and I agree on a principle of economical use of space operating in the Wadi el-Hol inscriptions: horizontal M on the column, vertical M on the line. The same applies to the snakes representing N, but the serpent between `ayin and T on the vertical inscription (in the divine name `aNaT) is W in Hamilton's view. Further confirmation is provided by the vertical forms of B (house), P (mouth), and Th (breast) on this horizontal line. When the number of peaks is reduced to two (/\/\) there is a possibility of confusion with Sh/Th (\/\/) and this might explain why M was closed off at the bottom or the top (according to its stance in a particular case) in the Arabian script.

[8] N (17: 154-171) [Syllabic NA]
Even in its Greco-Roman form, N shows its origin as a snake (nahhash, or nahas); the original sign has two graphic variants, representing two kinds of snake: I9 (cobra) and I10 (viper). It is an erect cobra that we see on the Thebes 1 proto-alphabet (top right ). There is an example on Thebes 3, which is just like N (though in reverse).

[9] B
(2: 38-52) [Syllabic BA]
The name Beyt (Greek Beta) applied to the B-sign shows its 'house' connections (bayt). Hamilton relates the graphic variants (or perhaps allographs) to O1 'house', and O4 'hut'. Neither of these cover the house with a porch, though Hamilton adduces a drawing of an Egyptian soul-house with such an entrance. The Wadi el-Hol B seems to be a hut; it has a vertical stance to save space on the horizontal line. Another puzzle seems to have been solved through Hamilton's method: the standard Egyptian form of O1 is rectangular with a doorway in the middle of the bottom line, but the normal proto-alphabetic shape is square, usually with a gap in a bottom corner; hieroglyph O1B could account for this. The example on the Thebes tablet 1 (top right) comes as a surprise, though it is certainly B, because it corresponds to a form of B in the Phoenician alphabet and one version of BA in the syllabary (as shown on my table); but it does not have an Egyptian counterpart in the evidence Hamilton adduces. A similar form is at the top of Thebes 2, next to a door (D). James Hoch (The Byblos Syllabary, Journal of the SSEA 20, 1990, 118-119) produces an Egyptian form from Giza in the Old Kingdom era, and another from Hatnub in the Middle Kingdom period.

[10] H (6: 76-86) [Syllabic HI]
Here Hamilton and I are agreed that the basic hieroglyph is A28, 'man with both arms raised'. It is a determinative marker for 'joy' and 'high', so I relate it to *hillulu, which carries ideas of exultation and exaltation (as in Halleluyah). Hamilton plumps for interjections, such as Hoy. He has overlooked the cases where the sign is inverted, as in Sinai 358, and for a model we could look to A29, 'man upside-down' (standing on hands and head). Also, on the three Wadi el-Hol examples, two of the figures have one arm pointing down (2.5, 1.11). Invoking A1, 'seated man' seems odd, given that both his arms are pointing forward; an obvious choice would be A32, 'man dancing' (one arm up, the other across his chest), likewise a determinative for 'joy' and 'jubilation'. However, on the Thebes 1 tablet the H (under the B) is not clearly discernible; on the photograph we can make out a head and an upraised arm, and the other arm (meeting the tail of the snake) may or may not be pointing downwards; there is a stick-body, and legs apparently pointing outwards. Ultimately, of course, the body will drop away, leaving what became Greco-Roman E. Searching for H on Thebes 2 + 3 raises a dilemma: the pitchfork figure on Thebes 2, at the end of the middle line (with three dots on one side and two on the other) would be H, on the model of A29 (as in Sinai 358); it would not be K, which is on Thebes 3 (top, centre).

[11] Y (11: 108-116)
We are now confronted with the problem of distinguishing Y and K. My rough rule is that Y (Yod, 'hand') is a forearm with a hand, while K (Kap, 'palm, hand') is a hand with fingers shown. Another way of looking at it is: Y-signs resemble Y, and K-signs resemble K. Hieroglyph D36 is the basic model for Y: an arm viewed from the side, with the elbow included at one end and the thumb raised at the other. Hamilton complicates the situation by adding D47, 'hand with curved palm' (showing the wrist but not the arm); this is a variant of D46, the basic model for K. If we need an Egyptian prototype at all, D36 (on my table) is sufficient: for Wadi el-Hol 2.8b (which has been ignored by everyone except Hamilton and myself); for the very simple Y in the middle of Thebes 3; and for the Y on the Thebes 1 proto-alphabet (bottom left, under the eye), which is like the Y on the Sinai sphinx (but an ax and Z for Hamilton); in all these cases the elbow has simply been omitted. The Sinai example where the elbow is intact (S365B) has not been recognized by Hamilton, but in combination with an adjacent D it has been consigned to his collection of Kh-signs (see 4 above).

[12] K (12: 116-123) [Syllabic KA]
I have said that K (Kap) shows a hand with fingers, and that is what we find in the texts; but no Egyptian hieroglyph has a model for that, though Hamilton and I both refer to D46, simply because it is a hand; but Hamilton points out that some examples in its history mark the fingers with lines (numbered D46D). Unfortunately, there is no K in the Wadi el-Hol texts. Hamilton cites only a few Proto-Canaanian instances; the main reason is that he has inadvertently placed several of them in other boxes, notably Ç/Ss (as papyrus plants) and Y. The example on the Thebes tablet 1 is above the eye and to the left of L; its stem is poking into the eye; it has three digits; the middle one is the longest; it represents a hand and wrist. Thebes 3 (top, middle) has a K with deformed fingers, we might say, but approaching the form of Greco-Roman K.  The West Semitic syllabary has a character that looks exactly like K (see the BS column of my table), and apparently has the value KI (presumably from kippat, 'palm-branch'). Consequently, there may be two allographs for K in the proto-alphabet: palm-of-hand and palm-branch.

We are only halfway through the proto-alphabet, and already there are serious differences arising between Hamilton's theory and my suggested paradigm. Great chasms will now open up between his proposals for identifying the letters and the confirming evidence available to us. In my judgement, Hamilton's remaining identifications will be either faulty or false.

[13] G (3: 53-57) [Syllabic GA]
G is equivalent to hieroglyphs T14 and T15, 'throw-stick', or 'boomerang'; the West Semitic word for it is gaml. Agreed. Yet Hamilton asserts that there are no known instances from the Bronze Age. On the contrary, there are numerous examples (see my table): in the syllabary on the one hand, and in the Sinai texts on the other (for example, gn 'garden', six times). Hamilton has perversely consigned them to the P-box (20: 188-196), where they are hypothetically tied to a word *pi, supposedly meaning 'wall-corner'. Yet, this character rarely appears as a true right angle; it is mostly obtuse, as boomerangs usually are; it would thus appear that the Sinai scribes could draw square houses with four right angles, but not right-angled wall-corners, even though the Egyptian sign is rigidly right-angled. Hamilton has taken up a speculative idea that was created by twentieth-century scholars and has doggedly defended it, while describing it as "one of the least transparent combinations of acrophone and graphic images in the early alphabet" (196). There is a throw-stick on each of the Wadi el-H.ol graffiti, which Hamilton regards as cases of P. On the Thebes tablet 1, G stands above K and next to R, as a right angle; but the true P-sign, I submit, is over to the left of H. There is a nice boomerang in the centre of Thebes 2.

[14] P (195) [Syllabic PU]
Hamilton admits that the name of the letter goes back to a West Semitic word for 'mouth', and yet he does not support this (and hieroglyph D21 mouth) as the original sign. On the Wadi el-Hol graffiti (1 horizontal, 2 vertical) he chooses the two boomerangs (2.9, 1.9) for P, overlooking the obvious mouth (1.13), though it is in a vertical stance, (|), and has an unusual line separating the lips, to be compared with this scribe's practice of marking the ox's mouth with a line (1.12, right next to 1.13), and also according to the principle (acknowledged by us both) that long signs can have a vertical stance in horizontal lines (likewise M, N, B, Th in this context). On the Thebes 1 proto-alphabet, P is represented by a mouth (top lip a straight line, bottom lip rounded), to the left of H; it seems to have a tongue, as does the clearer counterpart on Thebes 2, bottom right. Another clear example (though without the tongue) is on Thebes ostrakon 4.

Here it combines with D/Dh (double lines) to produce the word pad, 'fine gold'. Note also, from the left, L (the sign probably has a longer line than shown in my drawing, but it could perhaps be considered as a throw-stick, though the acute angle would negate that possibility), Z (double triangle), Q (cord wound on stick), W (or Tet if there is a crossbar on the stem), and K (note the longer middle digit, and the short arm, indicating that this is not H, as a person standing on the hands, like the example on Thebes 2; see 10 above).

[15] Q [Syllabic Q-]
I have constantly argued that Q has its origin in a cord wound around a stick, the 'line' (Hebrew qaw) used by builders (Colless 1988, 49). This is the form that Q has in the South Arabian alphabet. The sign corresponds to Egyptian V24, as in Thebes ostrakon 4, above; 

On Thebes ostrakon 1, Q is below the P, and one of its top lines is touching the tongue poking out from the mouth; this form is like hieroglyph V25, where the additional stroke represents the end of the cord; an equivalent Q is in a prone position on the left side of Thebes 3 (with a horizontal stem, a dot and <) below N; another such example is found on the Sinai sphinx in the word nqy, 'my offering'. Hamilton (23: 209-221) has completely ignored this possibility. Instead he has chosen the Ss-sign as Q, because scholars before him have thought a word in the Sinai inscriptions was nqbn, 'borers' (meaning 'miners') and not n-ss-b-n, 'prefects' or 'overseers'. He has taken up the Hebrew name Qop, 'monkey', and turned a bag into a baboon. This is certainly worth trying, given that an ox, a snake, and a fish also appear in the proto-alphabet; but one is in danger of making a monkey for one's back, when the evidence for the true Q is so patent.

[16] Ss
In 1988 (48-49) I published my support for the idea first suggested by A. van den Branden that the model for the sign we identified as Ss (S.adé) was hieroglyph V33, a tied bag (Semitic ss-r-r). Hamilton (197, n 254) dismisses this suggestion as "bizarre", but his rejection of it may prove to be much more so. Looking at the Thebes 1 proto-alphabet, Hamilton will need to fit the sign below the 'alep (ox-head) into his system; he can not claim the K below it as a clump of papyrus (M15, his choice for Ss); this is certainly a bag tied at the top, and we seem to be stuck with it for Ss. It produces six plausible words in the Sinai inscriptions (Colless 1990, 5). In isolation, it can be mistaken for the fish, but they are easily differentiated on this tablet (the fish is right next to it). On Thebes 2, it is in the top right corner, inverted, while the fish is to the left of the boomerang, in the middle line.

[17] S (18:172-180, column) [Syllabic SA] (5b: 66-75. fish)
The fish stands for S, and there can be no equivocation about this; but because samak ('fish') is only attested in Arabic, and because dag is the common West Semitic word for 'fish', W. F. Albright and his school have insisted that the fish represents D. Typically, Hamilton follows this traditional line, but he recognizes that the argument for D as a door (dalt, Greek Delta) is compelling, so he accepts them both as allographs for D ('alternate pictographs'; cp. Cross 2003, 316). Not so. The fish is S, and its allograph is the S that Hamilton recognizes, a column (samk), derived from R11 (now known to be a spinal column, rather than 'a bundle of stalks tied together'). Unfortunately, the purported instances cited by Hamilton are of his own manufacture; only the fish is attested in  proto-alphabetic texts from Sinai, though both are attested in Canaan (see the two rows for S on my table). The column (--|-|-| or --|-|) is present in the West Semitic syllabary (though not the fish); and it is found on the Lakish Dagger (incomprehensibly transcribed as T by Hamilton, 391); it became the standard S in the Phoenician alphabet. The door and the fish can not be allographs for the same sound, as they are found together in Sinai 376 (in the name 'Asa and in dwt, 'sickness'); and yet Hamilton (p. 379) transcribes both as d. In the Thebes proto-alphabet tablets both signs occur, in proximity: on Thebes 2 the door (D) is at the top, and the fish (S) is in the middle line; on Thebes 1 the fish (S) is situated in the centre (top), with the door (D) to the right of it; I realize that Hamilton could see this as an ax (T7), and hence Z in his system (92-97); but we have already observed (on Thebes ostrakon 4, above) that Z consists of two triangles, and Z will receive more attention below. Note that in the cuneiform alphabet the two forms are both represented (as shown on my table): the normal S (two small wedges atop a longer vertical wedge, presumably the fish); and `S (seven wedges clearly depicting the 'telegraph pole', the spinal column), which was used for transcribing Hurrian words. It seems that both are also found on Thebes 2: the enigmatic figure above the fish and beside the door is the spinal S; if its 2 dots are joined we have two cross bars on a stem with a trumpet-bell base (as on the hieroglyph R11). On the far right of Thebes 1 there are marks (but the photograph has cut off the full picture), and the other Samek might be here. The Arabian S (S3) is obviously derived from the fish.

[18] D
(5a:61-66, door) [Syllabic DA]
The door (dalt, Greek Delta) that represents D has a post, and may have two or more panels; all these details accord with the forms of the Egyptian model, O31; the cuneiform D (comprising 6 wedges) also conforms to this pattern. Greek Delta is a triangle, in line with Phoenician D; this could simply have developed out of the rectangular D (note that one version of the Arabian letter D is a triangle attached to a vertical line, the doorpost), but it has been suggested (by H. Jensen in his book on scripts) that it came out of a another kind of door, namely a tent-door, a triangular flap. In this regard, I have proposed that the puzzling triangular character on Sinai 357 (no 19) is a tent-door, given that the workers lived in tents (Sinai 365). The door on the Thebes 1 proto-alphabet is square, with a long post, to the left of the B, in the top right area; the example on Thebes 2 (top line, centre) is rectangular (and not joined securely to its post). This predominant form, with its defining post, could not be confused with H in the proto-alphabetic period, and yet Hamilton (63) supposes that a fish-D was introduced for this reason; he needs this hypothesis only because he and others have not discerned the original H, but have put a fence in its place.

[19] H
The field is now narrowed, and we can recognize H (h.) as the divided rectangle in the top left corner of the Thebes 1 proto-alphabet; it is more elusive on Thebes 2 and 3, perhaps the smudged marks in an identical position on ostrakon 2, or a larger ghostly edifice on 3, above Y. Hamilton (9: 97-102) defends a surmise beloved of the Albright school that Het is a fence, and he offers hieroglyph O42 as the Egyptian origin for it. I cannot refute this, and it is very tempting, but if it is true then I would have to admit it as an allograph. The only two Sinai examples he adduces are taken from an inscription that is rather illegible (375a), but I am happy to accept them as H, though as a dwelling with three compartments, rather than a fence (one is in a vertical stance, the other horizontal). This sound (h, h.) should have more occurrences in a collection of West Semitic inscriptions; by my calculations, it appears in Ugaritic texts in the 16th position of frequency, out of 26; and my choice for H in the Sinai corpus achieves 12th place, with seven appearances (including the two alleged fences): two have a rectangle divided into two squares, and one of the squares is also divided [360, 361]; one has the double-square with a semi-circular courtyard [380]; the other two have a square with a round courtyard [353, 356]. Hamilton deposits my examples in his B-box, which is understandable, since this is another form of house, with the addition of a courtyard, rounded or square. I do not connect this with an Egyptian hieroglyph, because it is a West Asian style of mansion (attested in the Hyksos domain in the Nile Delta); but if pressed I would refer to O6 (rectangular enclosure seen in plan), Egyptian h.wt, 'mansion', and likewise used for the consonant H. My proposal for the Semitic name of the original letter is, 'court', 'mansion', applied to the home built for Ba`al. The example on the Thebes 1 tablet may have had a curved wall on the left side, but it is now unclear. Eventually some of the walls will fall away, leaving H, in the Grecian and Roman alphabets.

[20] Tt [Syllabic TtA]
The Phoenician Tet, and the early form of Greek Theta, is a cross within a circle: (x) or (+). Hamilton (10:103-108) takes this as his starting point (extracting a dubious instance from the obscurity of Sinai 375a). He calls in O49, 'crossroads', the determinative for 'town'. This is a rare letter, and it could well be absent from the Sinai corpus. However, I have identified it as the cross-plus-circle sign in Sinai 351 (Hamilton sees it as H on its side): +o. This is a significant piece for the argument that the proto-alphabet was closely connected with the Egyptian system: it is the nfr sign denoting 'good' and 'beautiful', and therefore perfectly matching Semitic t.ab, 'good/beautiful'. It is also found in the syllabary, for TtA (T.A). It is the tall letter to the left of Q and below the fish on the Thebes 1 proto-alphabet; Thebes 3 has a large example (top left, o+). The rarity of this character makes it difficult to decide whether +o and (+) are allographs with different acrophonic origins, or the one has developed into the other, with the cross simply moving into the circle.

Hamilton (106-108) toys with the name T.êt. and the word t.ît., 'clay' of the streets, connecting it with the Egyptian character O49 (town with streets). Allowing more than one sign for Tt (as with S: fish and spine), other possibilities for (+) are: t.awar (Hbr t.îrâh) 'enclosure with wall'; t.ene' 'basket' (from Egyptian dnyt), later Hbr t.nî (I would invoke the Cretan sign for KA, which has the same form, and represents a cane-basket, in my personal view; the Egyptian wickerwork-basket character V30 is shown in side-view and gives no indication of cross-weaving).

[21] W [Syllabic WA]
Waw seems to be basically a circle on a stem; there is an example above the Tt on Sinai 351 (acknowledged as W by Hamilton, 343). Albright had suggested this was a mace, and Hamilton (7: 86-92) has no doubt that the letter derives from T3, a mace. The fact that it has a knob on top leads him to bring in some examples of Q, not realizing that the one or two projections at the top of Q (cord on stick) are what distinguish proto-alphabetic Q from W. Later, the Q will take the form of W, and the circle of W will open up: --( . Already in the syllabary WA was an open hook (see the BS column on my table). The word waw occurs in the Hebrew Bible (used for the hooks from which the curtains in the Tent-Temple were suspended). Where is it on the Thebes proto-alphabet tablets? On Thebes 1 it must be the square sign with a short stem and a jagged edge, on the left side, next to K; another example with a square top, but a long stem, is on Thebes 2, middle line, right side, next to H. Is it a weapon or a curtain-hook? I find an example (1.3) on one of the Wadi el-Hol graffiti (horizontal): rb wn, 'much wine'. My feeling is that there was no hieroglyph available for this word waw, and so the Canaanite scribes simply drew the object they knew. Hamilton (91) wants to derive the word waw ('hook', 'nail', 'pin', or 'peg') from the name of the letter (the mace came to look like a 'peg' and so it became the word for it); an unnecessary surmise, given that very few West Semitic nouns start with w, and this one was ideal for representing WA and W. 

[22] Sh
[Syllabic ShI/ThI]
Where do we start on this thorny problem? Hamilton (13: 123-126), and others, can find only one Proto-Canaanite attestation of Sh/S: the triangular sign on Sinai 357 (number 19 on my drawing), which he relates, plausibly enough, to hieroglyph M44, 'thorn'. (His attempts to find other instances on 375 and 376 are not acceptable, as he admits, 124, n 144). At this point we should remember an important detail to assist our search: when Sh and Th coalesced in the Iron Age, the sign that survived under the name Shin was Th, not Sh, because Th was in the position near the end of the alphabet, whereas Sh was in the middle, between K and L. But is this thorn really the Sh that was lost? I have suggested that it is a variant of D (see 18 D, above), and I can not find it on the Thebes proto-alphabet tablets. My proposed identification for Sh is the sun (shimsh), based on hieroglyph N5 (a circle, with or without a dot in the centre) for the syllabary (ShI/ThI), and for the proto-alphabet N6 (a circle with a serpent). In the cuneiform alphabet, both forms for Sh appear as allographs. But if it is true that N6 did not exist before the New Kingdom (16th century BCE), then the Sinai and Wadi el-Hol inscriptions must date from that period. However, the way to look at it Is to invoke N6B, which has two serpents guarding the sun-disc. Here is an example of this sign in a West Semitic (syllabic) inscription (Thebes ostrakon 6).

The sun-sign (circle with snake) is the fifth from the left. Before I had ever seen this character used in West Semitic writing, I had guessed (in Colless 1988, p. 51) that the Proto-Canaanite Sh was this sun-sign with the sun-disc omitted. The Albright school takes that character to be a 'composite bow', for which a word *thann is concocted, and the value Th is imposed on it. I am being sceptical, though our confirmatory source could well provide comfort to this hypothesis: when we look for the sign on the Thebes proto-alphabet we have a choice between a fairly clear character between Q and Kh, and an obscure one below Kh (both in the lower right section).

The latter sign would be Sh, by my criteria, as a sun-serpent without the sun disc, and yet it certainly looks like hieroglyph Aa32 (alias J32A), 'archaic bow'. Can I make room for an allograph of Sh or Th? Or is it a coincidence that in its development this sun-sign came to look like a bow-sign? I will continue to maintain that this is the Sh-sign that disappeared, and that the other similar character to the left of Kh is the one that survived as Shin (also encompassing Th).

Looking ar Thebes Ostrakon 3 (on the right), we can find both Sh and Th. On the far right we see Sh (faint but clear, next to T, a cross); it has a vertical stance; its central peak is rounded; there is no obvious sun-disc, but there are possible traces of one. The breast-sign for Th is in the middle section; it has a sharp central peak, and this is the form that later Sh/Th will have. (Incidentally, the fact that these two are missing from Thebes 2, on the left, suggests that they form a single document.)

Can we seek assistance from the Wadi el-Hol lines of letters? I suggest that 2.2 and 2.10 (vertical) are forms of the sun-sign, Sh. The shape is unusual: the sun-disc and the serpent are there, but the tail is missing; this form is found as a symbol on stelas in Sinai, dating from the Middle Kingdom. We can see how this could become the Arabian Th [o-o]; but there has been a reversal here, and Arabian Sh [\/\/] would go back to Proto-Canaanite Th, which I connect with thad, 'breast'.

At the Timna copper mines, we see an example of the sun with one serpent (N6A, with its tail and the sun intact) and another with two snakes (N6B, but the sun is absent!) 

[23] Th [Syllabic ThA/ShA]
As stated in the Sh-section, Hamilton (25: 231-244) has a bow (J32A and also T10) for Th, with a hypothetical *thann as the acrophonic source. For my part, I favour a derivation from thad, 'breast', which has no 'double-breasted' counterpart in the Egyptian inventory. Hamilton (237) presents two breast signs as bows: Wadi el-Hol 1.10 (horizontal) and Sinai 375 (line 3, twice in th-l-th, 'three'). He also cites the Megiddo signet ring (which is clearly syllabic) and its example of ShA/ThA (237, 241). (Elsewhere, 252, he mistakes the character next to it [ShU, 'sceptre'] for a T, but eventually he will have to recognize that this inscription is not consonantal proto-alphabetic but syllabic.) The Thebes proto-alphabet breast sign, which will become the letter Shin in the Iron Age, is to the left of Kh. As with W, Z, and T, the letter Th did not have an Egyptian prototype, in my view.

[24] T [Syllabic TU]
Hamilton (27: 246-253), to achieve a full total of Egyptian sources, produces Z11, 'crossed planks', to go with the simple cross that is the universal signature for the illiterate. The word that goes with it is certainly taw or tu, 'mark', but whereas Proto-Canaanite T can be + or x, Egyptian Z11 is invariably -|- (and it does not need to be brought into the picture at all). As noted above in the Sh-section, there is an example on the far right of Thebes 3. But where is it on Thebes 1? My drawing shows two possibilities: one on the H sign, the other (very small) between the M and the Sh signs. The latter will be my choice, because I need the other for Z.

[25] Z
Hamilton (8: 92-97) derives Z from hieroglyph T7, 'ax', and *zayn, 'weapon'. It is certainly a double triangle, and Hamilton could be right. My proposal is that it represents manacles (handcuffs), for which a word ziqqu exists, but no hieroglyphic prototype. I can find only one instance in the Sinai corpus, also recognized by Hamilton as Z: on the far right of 375a. There are examples of Z on Thebes 3 and 4 (easy to find).

But where is Z on Thebes ostrakon 1?

I have not shown it on my drawing, but a large Z can be made out on the photograph: follow the black line to the right of Kh in the bottom right area; turn north-west along a black line to meet the H figure; go east along a white line for a short distance; then return southwestwards along a white line to Kh. The triangles are not equal, but I think the Z-sign is there. This, not Dh, is the letter that became our Greco-Roman Z.

Lurking somewhere in the dim regions of the Thebes tablets may be the letters Z. and Gh, which are known from the cuneiform alphabet. I think that we may be able to find both of the missing letters there, between Q and Th, and above M.

[26] Z.
Transcribing this letter is problematic, when it is not possible here to put a dot under the consonant; but on the analogy of my Hh, Tt, Ss, it could be Zz. Possibilities for Zz/Z. (a rare sound) are: Z.BY, 'gazelle', or Z.R, 'back'. But the one I have always favoured as a hypothesis is ZzL (Ugaritic; Hebrew s.el 'shade, shadow'). The relevant Egyptian hieroglyph is S35, a sunshade: --|), used as an 'ideogram' in writing shwt 'shadow, shade'. The cuneiform alphabet has three wedges for Zz: =<. The wedge on the right represents a circular element in the original glyph. A small example of Zz is on Thebes 1 above M. An instance of Zz appears on Thebes 3, on the left side, next to N.
[27] Gh
The faint image (next to Zz) that I suggested was a face may be seen as a a vine-stand with grapes hanging from it, hieroglyph M43. Many years ago (1988, 63) I noted that the South Arabian Gh, "a baseless square with a diagonal line projecting down from its top left corner", resembles the Egyptian vine-hieroglyph (but with the grapes missing, as on hieroglyph M43A). The acrophonic word would be GhNB, 'grapes'. A possible example appears in the middle of line 3, below the fish, on Thebes 2. But on Thebes 3, under Tet and above N, we can perhaps see a pair of grapes on a horizontal line, which might indicate that the whole tablet should be inverted. If this is indeed Gh, what do we do with the character on Thebes 2, which has the shape of Arabian S1?

[28] (Dd) That covers all the letters. Surely this puzzle has not been solved? What about D./Dd? The Arabian alphabet has a sign for this sound, looking like a door with its post knocked off, and so an adapted form of D? The cuneiform alphabet seems to lack Dd; and in Sinai inscription 356 we find the word SsRHh (s.rh.) "excavation chamber", which has an Arabic counterpart d.arih. "grave". Accordingly it appears that Dd/d. was not one of the West Semitic consonants.

Returning to the book under review, let me say that I am deeply impressed by the massive amount of labour that Gordon Hamilton has put into this project. It was certainly a task well worth undertaking. My big disappointment is that he apparently emerges from it as affirming that if he (or William Foxwell Albright or Frank Moore Cross) thinks a certain character looks like a particular object (and a similar Egyptian hieroglyph) then it must be so. I would have been interested to see him consider all my suggestions along the same lines, rather than dismiss them as bizarre without really testing them.

Realistically, we should not start out on this exercise if we do not know what the original letters of the alphabet are. We could have achieved this knowledge through deciphering the Sinai Proto-Canaanite texts, employing the proper procedure (cryptanalysis). However, without the Thebes proto-alphabet to refer to, we are all merely making stabs in the dark, playing a game of guesswork. And yet, if we compare the signs on the five Theban Proto-Canaanite ostraka (showing 27 characters) with a table of the Phoenician alphabet (22 letters) we can match most of the corresponding signs with ease and determine which ones have fallen out of use (as I see it: Dh, Kh, Sh, Gh, Zz); further comparison with such 'missing links' as the Izbet Sartah abagadary would clarify the fish and column allographs for S (samk), and the evolution of D (dalt, door); the difference between W and Q; the development of Ttet from +o to (+); the obvious simplification of the mouth for P, and the deflating of the tied bag for Ss .

If Gordon Hamilton intends to continue in this field he will need to take account of the six Theban ostraka (they could not be forgeries, by the way, since the West Semitic consonantary and syllabary were unknown to Western scholarship when Petrie published the photographs). He should also look into the West Semitic syllabary, so that he can learn to distinguish the two related systems (at present he still thinks the Megiddo signet ring is consonantal not logo-syllabic, as noted above). His comparison procedure could be applied to the pictosyllabograms of the syllabary; after all it has long been known as "the Byblos pseudo-hieroglyphic script", because it obviously uses Egyptian characters. And, if he believes the Proto-Canaanite alphabet was so dependent on the Egyptian logo-consonantary, then he should ponder whether the proto-alphabet might also be a logo-consonantary, with its letters functioning not only as consonantal acrophonograms, but also as logograms.

The erroneous Albright chart of the original letters of the proto-alphabet is the one you see in museums, on websites, and in books on the alphabet. My feeling is that Hamilton's thesis represents the last dying gasp of the Albrightian system. I myself was once glad to be inside that straitjacket, following the hidebound tradition and propagating its dogmas. But I broke out of it when I discovered the work of Romain Butin on the Proto-Canaanite inscriptions, and from there I found my own way. Ironically, Gordon Hamilton has dedicated his book to the memory of Romain Butin, S.M. (1871-1937).

It remains to be seen whether this book will live up to the reputation which J. Day has bestowed on it already: it "contains a wealth of information and provides a benchmark for future work on the subject" (JSOT 31.5, 2007, 330). Frank Moore Cross (Leaves, 2003, 329) praises it as an "excellent study". In their edition (without translation) of the Wadi el-Hol early alphabetic inscriptions (2005, 93, n 20) Darnell, Dobbs-Allsopp, Lundberg, McCarter, and Zuckerman (for all of whom I have great respect) describe "this important study" as "by far the most thorough and reliable discussion of the paleography of early alphabetic inscriptions known to us". Actually, it is the only one of its kind, if 'paleography' means not simply the study of ancient writing, or of manuscripts written in ink in particular, but refers rather to dating the characters of writing systems by comparing their forms, and setting up a chronological typology. That is what Cross does, but I tend to think it is not possible for the proto-alphabet. It is certain that the inventer of the West Semitic logo-consonantary, that is, the consonantal proto-alphabet, had the hieroglyphs of the Egyptian logo-consonantary in mind; but that does not mean that all who used this new tool had to conform to the changes each sign went through in its cursive development (hieroglyphic > hieratic). To write R a scribe drew a human head; for G a boomerang; for D a door (or a tent-flap if you lived in a tent at the Sinai mines); for S any kind of fish, or else a spinal column in some regions; for T a cross (x or +, it did not matter); for H a person rejoicing with both arms raised, or dancing with one arm down, or standing on his hands; for N any kind of snake.

However, the first conclusion of this book (Chapter 3) is that "the writers of the early alphabet" (not the first deviser of the West Semitic logo-consonantary) "adopted and adapted thirty-three Egyptian signs". That seems exorbitant for 26 or 27 sounds (especially when set against my supposition that five or more of the letters did not have an Egyptian original).

Hamilton's Table A (270-271) sets out the range of the 33 Egyptian forms borrowed (notably incised hieroglyphic and hieratic). "The major conclusion of this study is that West Semites borrowed and reutilized a pre-existent range of both hieroglyphic and hieratic forms of Egyptian signs for use as letters in their monoconsonantal alphabetic system of writing." In a footnote he adds that this conclusion was also reached by "the epigraphic team of Darnell et al.(2005: esp. 86) based on their very insightful work with the Wadi el-H.ol alphabetic texts". Unfortunately their work also started from a position of unsound knowledge of the original Semitic characters. Their "paleographic chart" (Plate 10, 124) tabulates 14 signs (as against the 16 that Gordon Hamilton recognizes, and the 17 that I can find there) including 6 incorrect identifications, in my view. Hamilton and I have noticed the Y beside the T on the vertical inscription, and everyone else who has written on this subject has overlooked it.

My identifications of the Wadi el-Hol characters are available elsewhere. I think that the two inscriptions form a single text, inscribed at the same time by one person (his signature is in the two similar ox-heads: 1.12, 2.11). The entire text can be viewed as running leftwards. In my interpretation the first word is MShT (Hebrew mishteh) 'drinking party' or 'drinking place', and the sequence at the start of the horizontal line is RB WN, 'plenty of wine'; both ideas are consonant with Egyptian inscriptions on the site, which speak of soldiers having celebrations with music and drink for the goddess Hhat-Hhor. The West Semitic deity is `Anat, named (`NT, eye, snake, cross) and depicted on the vertical text, holding a 'handkerchief' (hieroglyph S29, a piece of folded cloth seen in the hands of statues, presumably a mark of distinction). The dancing figure (H) above her name would be a logogram for hillulu, 'celebration'.

In passing, I should mention something fishy: John Darnell and the team of Semitic epigraphers (Dobbs-Allsopp, Lundberg, McCarter, Zuckermann) have excluded from their bibliography the scholars who espouse the identification of the fish-sign as S (Samek) not D (still known as Dalet, 'door'): Butin, van den Branden, Puech, Colless (though Tropper and Briquel-Chatonnet [with an extra -n- between a and t] get through the net. Those who have seen it as D (dag, 'fish') are given due deference: Albright, Cross, Sass, Hamilton. As noted above (under D) Hamilton chooses to have it both ways: the fish and the door are doublets ('alternate pictographs').

Note that my 'linguistic decipherment' (a term that Hamilton uses) for the Proto-Canaanite consonantal and syllabic texts is being published on this site, beginning with "Sinai sphinx speaks" (Sinai 345). In the light of the information I have provided here, it might be worth reconsidering.

Gordon Hamilton's labours have certainly yielded fruit (notably in the study of the letter B), and the approach should be extended further, not abandoned. The trouble is that his misidentifications have led to some shaky generalizations.

My dear epigraphical colleagues, we try to convince ourselves that we are practising objective science, but the subjective results we all present look like sophistry spiced with speciosity.

COLLESS, Brian E., "Recent Discoveries Illuminating the Origin of the Alphabet", Abr-Nahrain, 26 (1988), pp. 30-67. A preliminary attempt to construct a table of signs and values for the proto-alphabet, and to make sense of some of the inscriptions from Sinai and Canaan.
COLLESS, B.E., "The Proto-alphabetic Inscriptions of Sinai", Abr-Nahrain, 28 (1990), pp. 1-52. An interpretation of 44 inscriptions from the turquoise-mining region of Sinai.
COLLESS, B.E., "The Proto-alphabetic Inscriptions of Canaan", Abr-Nahrain, 29 (1991), pp. 18-66. An interpretation of 30 brief inscriptions from Late-Bronze-Age Palestine.
COLLESS, B.E., 1996, "The Egyptian and Mesopotamian Contributions to the Origins of the Alphabet", in Cultural Interaction in the Ancient Near East, ed. Guy Bunnens, Abr-Nahrain Supplement Series 5 (Louvain) 67-76.
COLLESS, B.E., 1992, "The Byblos Syllabary and the Proto-alphabet", Abr-Nahrain 30 (1992), 15-62.

And my other articles on the Canaanite syllabary ("Byblos pseudo-hieroglyphic script") in Abr-Nahrain (now Ancient Near Eastern Studies) from 1993 to 1998, culminating in:
COLLESS, Brian E., "The Canaanite Syllabary", Abr-Nahrain 35 (1998) 28-46.

All except 1988 are available at the Peeters website.

CROSS, F.M., Leaves from an Epigrapher's Notebook (2003). Collected articles

DARNELL, John et al, "Two early Alphabetic Inscriptions from the Wadi el-H.ôl", The Annual of the ASOR 59 (2005) 63-124.

GELB, I. J., A Study of Writing: The Foundations of Grammatology, 2nd en (Chicago 1963).

HAMILTON, Gordon J., The Origins of the West Semitic Alphabet in Egyptian Scripts (Washington 2006) XVI +433 pages.

POWELL, Barry B., Writing: Theory and History of the Technology of Civilization (2009)

SASS, B., The Genesis of the Alphabet (Wiesbaden 1988)

This is my chart showing the development
of the proto-alphabet.

Click on it to see the enlarged picture.

No comments: