The Proto-Sinaitic script

Despite spectacular archaeological progress in all sides of the Eastern Mediterranean during the 20th century, including deciphering the so-called Mycenean script (Linear B), the epigraphical evidence from the 2nd millennium Levant remains very scarce. People had been waiting for an archaeological confirmation of the Biblical story linking the Israelites with Egypt for the entire history. The English Egyptologist Sir William Matthew Flinders Petrie (1853 –1942) thought he had found the ‘proof’ that Israelites who came through Sinai to Egypt and passed back again could readily use writing by 1500 BC.

In 1905, while excavating various sites around the ancient turquoise mines in the Sinai Peninsula, Petrie came across a group of atypical, clumsy sculptures bearing the same inscription in unknown characters. He dated those objects to the 18th Dynasty based on the archaeological context, i.e., around 1500 -1300 BC (Flinders-Petrie and Currelly 1906). His hard evidence ends there. What follows is Petrie’s interpretations filled with logical jumps and speculation. Petrie was not expected to know all the hieroglyphic signs ever used by the ancient Egyptians. A comprehensive list of hieroglyphs was only published about half a century later (Gardiner 1957). What attracted his attention was the unusually casual style of the writing. He, therefore, concluded that the sculptures and the inscriptions were made by foreigners working in the mines. Based on the Bible, he hypothesized that the inscriptions are Semitic. This hypothesis ‘proves’ that the Bible is correct by circular reasoning. Therefore, the hypothesis must also be true. Why are the inscriptions Semitic? Because so says the Bible! Because – we know from the Bible – the Israelites passed from there, the language of the inscriptions must have been Semitic. How do we know that the Bible is accurate (and should be interpreted literally)? Because we found Semitic inscriptions in Egypt! However, Petrie also expressed his disposition to see the script he found as one of many ‘alphabets’ used continuously in Mediterranean lands long before the fixed ‘alphabet’ (consonantal system, to be accurate) was selected by the Phoenicians. He estimated a mass of linear signs to be around since 7000-6000 BC.

Of course, Petrie used the term ‘alphabets’ to mean abstract linear scripts comprising only a few characters instead of hundreds of elaborate figurative hieroglyphics. Today, the term alphabet applies to graphical representations of phonemes, including vowels. There are, however, several other problems in his interpretation. The sculptors could also be beginner trainees of Egyptian origin using free stones from the mine as training material. The Biblical story remains a myth until its historical relevance is independently validated. It is anti-scientific to claim that Israelites carved the stones because this is consistent with the Biblical story, and then that the Bible is right because we found stones inscribed by Israelites. Even if the sculptors were foreigners, they were not necessarily Israelites; these were likely not the only immigrants in Egypt. The other linear writing systems Petrie explicitly mentions are the Cretan, Anatolian, and Celtiberian, implying that any one of these could provide alternative explanations of the inscriptions he found. As to the dating of the linear signs around the Mediterranean, the 7th millennium BC was only his wild guess.

Petrie published a photograph showing only a part of the best-preserved copy of the inscription, known as the Serabit el-Khadim inscription, along with a broader plan of the statue on page 139 of his 1906 book. The close-up, now in the public domain, has been repeatedly used as the example of Proto-Sinaitic script (Fig. 1). The statue photograph, presumably still under copyright restrictions, clearly shows an additional figurative pictogram of a fish. Fish pictograms exist in both Egyptian and Cretan hieroglyphic scripts. All the signs shown in close-up may be interpreted either as Egyptian or Cretan hieroglyphs except one (Fig. 1 sign 4), which is confidently not in Gardiner’s list. While crosses, zigzag lines, simple curves, and oval signs are shared among all scripts, sign 4 is identical to Corpus Hieroglyphicarum Inscriptionum Cretae (CHIC) sign 056#113.b2 (Godart and Olivier 1996). It has never been found in an inscription from the Levant or elsewhere in Egypt (Goldwasser 2011a). Therefore, if the Serabit el-Khadim inscription is not Egyptian, the only reasonable alternative candidate is Cretan hieroglyphics.

Figure 1. The most famous Serabit el-Khadim inscription (circa 1500-1300 BC). Original artwork and dating by Flinders-Petrie (1906). The image is marked as public domain and processed by James R. Harris and Dann W Hone. All the signs may be interpreted either as Egyptian hieroglyphs (red), numbered according to (Gardiner 1957), Cretan hieroglyphs (blue) according to Corpus Hieroglyphicarum Inscriptionum Cretae (Godart and Olivier 1996), or Linear A (green) signs according to the SigLA database, except sign 4 which is only referenced in CHIC. A facsimile of the inscription by Gardiner is shown in the lower-left corner.

Still young, Gardiner attempted to substantiate Petrie’s hypothesis of an Egyptian origin of the Semitic alphabets (including the Phoenician) through Proto-Sinaitic (Gardiner 1916). He published all (11 in total) inscriptions from Serabit el-Khadim available at the time. He remarked, as Petrie did before, that 6 of them bear the same sequence of four letters (signs 4-7) with some variations which he considered insignificant. For example, sign 5 also appears as an eye with a pupil. Always according to Gardiner but about four decades later (Gardiner 1957), the Egyptian hieroglyphs V38 (vertical without pupil; determinant meaning bandage, without phonetic value), D21 (horizontal without pupil; mouth; /r/), and D4 (horizontal with the pupil; eye; /ir/[1]) are distinct signs. In the first cited work, Gardiner forces all the eye-like signs of being equivalent to the Semitic’ ayin (O; eye; [ʕ]), 𐨀 to be Lāmed (𐤋; goad; /l/) and sign 4, or analogous variants, to be Bēt (𐤁; house; /b/). He thus derives the later Phoenician word 𐤋𐤏𐤁 (/ba’al/; read from left to right), meaning owner, lord, master. He then assumes the mark + as an inflectional element rendering the female ba’alat, the lady, head of household, mistress. But the Hebrew name for mistress is not ba’alat but baʿalah[2]. Next, he assumes that because ba’alat repeatedly appears in the inscriptions, she must be a goddess, the Semitic equivalent to the Egyptian Hathor. With the same logic, the future archaeologist who will find multiple instances of the inscription TOILET in the ruins of a multi-storey public building may think that TOILET is a goddess. Nevertheless, Gardiner did warn that his suggestions can only be taken as ‘an unverifiable hypothesis.’ The phrase ‘[to the] lady’ is frequently attributed to the inscription by later reporters. Others go further to equate the entire sequence to MTLB’LT, transliterating Hebrew mat l’ba’alt, which means death to/for Ba’alt (Benner 2021). The stem mat would derive from the Semitic root mwt, to die (compare the English checkmate, mate, from Arabic māt, he has died, from earlier māta, to die, or Spanish matar, to kill)[3].

Thus, half a dozen unusual but not so novel characters fed scholars’ imagination and political motivations for an entire century. The arbitrary and hesitant initial interpretations of the Serabit el-Khadim inscription propagated by grapevine’s telegraph as proof that the script was the Semitic ancestor of the Phoenician alphabet, which, according to Herodotus, was given to the Greeks. In the relevant article of French Wikipedia, we read[4]:

A few kilometers from Serabit al-Khadim… numerous rock faces are engraved with a large number of signs. … These inscriptions are among the earliest traces of the Proto-Sinaitic alphabet, from which the Phoenician alphabet is derived. All these inscriptions present a particular feature: by the acrophonic principle (in) they use a derivation of Egyptian phonetic hieroglyphs to write a Semitic language. The authors of these two inscriptions used Egyptian signs, giving them the value of the first sound in their Semitic language of the word designated by the Egyptian hieroglyph. Thereby, the [Egyptian] pr pictogram representing a house in hieroglyphic script became *bēt in Semitic and was used to transcribe the phoneme /b/, i.e., the initial of *bēt. This name stuck to refer to the letter itself in the Hebrew alphabet and was so entrenched that it was passed down to the Greeks along with the alphabet (βῆτα; bēta, i.e., *bēt with a suffix -a). These inscriptions were therefore written by Semites who worked in the Egyptian mines of Sinai.

The above Wikipedia text is chock-full of exaggerations, anachronisms, arbitrary distortions, untestable assumptions, and logical fallacies. The ‘numerous rock faces’ were the two front sides of a statuette. The ‘large number of signs’ was 5 distinct, or 9, in total. The ‘acrophonic principle,’ which generally refers to the Phoenician abjad and later alphabets, has been backdated to explain the Proto-Sinaitic script, upgraded to an ‘alphabet’ of a loosely hypothetical language based on a single inaccurately spelled theonym (ba’alat instead of baʿalah). The fourth sign of the inscription, whatever it may represent, is confused with Gardiner’s hieroglyphs O1 (/pr/, house), O4 (/h/, room), and the alphabetic letters 𐤁 and B. The whole story is presented as a validated fact. Of course, the inscriptions were written by Semites because the language was Semitic! We know that the language was Semitic because they arbitrarily identified the consonants B, O, L, and T, to vaguely fit a Semitic theonym! Assuming the signs 4-7 were indeed equivalent to B, O, L, and T, respectively, and the writing system was consonantal, i.e., allowing for any vowels to be inserted anywhere, the sequence *B*O*L*T* fits not only Semitic but also innumerable Greek, Latin, Germanic, Finno-Ugric, and Romance words.

An English linguist may claim, for example, that *BOLT is akin to PIE *bʰeld-, to knock, strike, which gave the Lithuanian beldu, I knock, baldas, pole for striking, Dutch and West Frisian bout, German Bolz or Bolzen, Danish bolt, Swedish bult, Icelandic bolti, all cognates of English bolt, meaning door-stopper among other things. In Portuguese, balda means a structure fixed to a wall. In Greek, μπαλτάς (/baltas/; like the Semitic Ba’alt) is a cleaver, chopper, hatchet. The standard inscription *BOLT could, thus, indicate the use of the decorative stone statuettes as door-stoppers, wall fixtures, or the tool and technique used for sculpturing. If bolt is a cognate of Dutch and West Frisian bout, Proto-Sinaitic *BOLT could also be considered a root of French bout, meaning end, extremity, tip, bit, piece, or scrap. It could, thus, indicate the position of the decorative stone in a building, or simply that the piece is unfinished, discarded, junk, stone to be recovered, and should not be sold, taken away, or used. Yet, another French linguist could claim similarity between *BOLT and the colloquial French word boulot, meaning work, job, and frequently used in qualitative expressions such as beau boulot, good job, or petit boulot, odd job, temporary employment. In that case, the inscriptions might be taken as the instructor’s marks on the student’s work, or analogous to modern expressions such as lab-work, homework, handicraft, hand made, etc.

If the *B of the inscription was pronounced Greek, as /v/, the inscription could phonetically give /volt/. In English, the volt is green, like lime or some turquoise hues. The inscriptions could indicate the stone’s origin being the turquoise mine near Serabit el-Khadim. Alternatively, /volt/ could be at the origin of Old French volte and Modern French voûte, meaning vault, turn, cognate of Spanish vuelta, turn, and of Medieval and Modern Greek βόλτα, (volta, /volta/), turn, spiral, rotation, gyration, revolt, revolution.

These hypotheses assume that the ancient Egyptians thought and acted like we do. It takes religious bigotry to believe that everything they made and wrote was for unseen spirits rather than for earthly customers who could afford the price of their arduous work. Who would invent a computer programming language today for the sole purpose of writing letters to Santa Claus? Even the meanings of the Phoenician word 𐤋𐤏𐤁 (/ba’al/), owner, lord, master, do not need religious connotation. The master of a work is its maker; compare Greek μάστορας, mastoras, craftsman. Comparing the Proto-Sinaitic *BOLT to English, Lithuanian, Portuguese, Greek, or French words, does not require illiterate workers from all over Europe to have worked in the Serabit el-Khadim mines and have invented their respective alphabets there. Instead, the root *BOLT stemmed from Egypt and, from mouth to mouth, entered the vocabularies of modern languages, including Semitic, in one form or another, with one meaning or another.

However, despite their rationality and relative simplicity, these hypotheses are just as untestable as the ba’alat hypothesis. Following a thesis book (Brown 1985), Ray Brown has published a series of web pages with examples explaining how easy it is to ‘prove’ that an inscription belongs to a language if one starts with the hypothesis that it belongs to that language (Brown 2021a; 2021b; 2021c). It is impossible to translate an inscription without recognizing the language or some coherent translation, e.g., from bilingual inscriptions. By ‘coherent’ translation, I mean one that fits the archaeological context, is long enough to be specific and predicts the meaning of other inscriptions in a way that fits the archaeological record each time.

The occurrence of the digram bn does not prove Semitic origin. The root bn does exist in Semitic languages as /b(i)n/ or /b(ē)n/, meaning son, son of, and is frequently followed by a personal name[5], like Bin Laden or Ben Laden (French spelling), also spelled Ibn Lādin. When found in an inscription followed by an incomprehensible sequence of glyphs, we usually presume that the inscription is Semitic, bn means son of, and what follows is a patronym. However, in Ancient Egyptian, bn is the negative particle not (Dickson 2006; p45). What follows may be an unknown verb, adjective, or adverb, and the inscription – if Egyptian – may represent a prohibition or any kind of negation. There are about 350 English and about as many French words on the pattern *bVn*, where V is any vowel. Nevertheless, it would sound absurd to claim that bn is Proto-English, Proto-French, or Proto-whatever, simply because it fits a modern word. Unless one can safely deduce the language of an inscription otherwise, a digram has no specific meaning on its own.

I am also wary of the decipherment of Linear B as Greek for the same reason. This transcription required enormous distortion of Greek phonetics and grammar, and excessive reliance on proper nouns and toponyms, only to translate a small part of the existing texts into Greek (Ventris and Chadwick 1973). Masses of untranslated texts that do not fit the Greek Linear B theory are arbitrarily defined as Linear A. If a Minoan or Mycenean inscription fits Classical Greek, it is Linear B. If it doesn’t, it is classified as Linear A. Yet, Linear A was written with the same characters, at the same places and times as Linear B. One way to explain this paradox is to assume that two ethnically different populations, speaking different languages, coexist for long without mixing. Another explanation is that there was only one language and one script, but our understanding is incomplete. Orly Goldwasser admits that all early texts known to date, referring to the Proto-Sinaitic, are very short and consist mainly of theonyms, personal names, and titles (Goldwasser 2011b), i.e., meaningless words.

Goldwasser’s arguments supporting a Proto-Semitic alphabet hypothesis of the Serabit el-Khadim inscriptions are debated (Colless 2014), but she explains the acrophony principle of an alphabet quite well. Take the example of sign 4 in the inscription of Fig. 1. To the eyes of an illiterate worker, she says, sign 4 resembles a house. To my eyes, it resembles a bottle or a box, but never mind. No matter how a literate Egyptian would pronounce and interpret that sign, the illiterate Semite would call it bayt, house. The first phoneme of bayt, /b/, is attributed to sign 4 and any sign that looks like a house, while the signified ‘house’ is discarded. Sign 4 loses, thus, its semantic value. It no longer means house but gains a phonetic value, /b/. Wherever sign 4 is encountered, it means nothing but /b/. Acrophony is a phonocentric principle by excellence. It assumes spoken words, like /bayt/, preceding written ones, like bayt. The phonocentric theory does not explain the origin or the meaning of spoken words. From a graphocentric perspective, sign 4 would have retained its intuitive meaning of a bottle-like shape. It would have, only later, gained its phonetic value /b/ imitating the sound of a bottle opener with the characteristic lip gesture producing a voiced bilabial plosive.

Of course, one may argue that a bilabial plosive gesture was performed to signify a bottle-like object before the invention of sign 4, which could graphically transcribe the phoneme. Nevertheless, neither a phoneme nor a grapheme representing stenosis (bottleneck) would have made sense before the invention of bottles or the observation of bottleneck phenomena. Sounds and graphemes that do not represent something (non-words, non-iconic) are useless. Language follows observation and technology; it does not precede them; it cannot be older than ‘civilization’.

Having gone through the above exercise to trace the potential meanings of the pattern *B*O*L*T* (extendable as * [B, V] * [O, U] * [L, I] * [T, X] *) from the Serabit el-Khadim inscription, we note that this pattern occurs in the names of objects consisting of a closed hollow space with a small opening on one side. A house (Semitic bayt) is one of them. The English bottle, box, boat, boot, built (from build), boutique, about (on every side), belt, button, ballot (voting), Baltic, balloon, or the Greek βάλτος (baltos; /valtos/; swamp), also have the notions of closed space with a small opening. The sign 5 of the Serabit el-Khadim inscription, if interpreted as *‘ayin (eye), would corroborate with the notion of a small opening in the otherwise closed space introduced by sign 4. An eye is a small opening that opens and closes at will. It is possible that the signs did not lose their semantic values but linearly summed them up to make more precise sense. This would be the graphocentric view. Subsequent phonetic variation, merely due to lousy imitation, may be semantically irrelevant. For example, the choice of phonetic gesture to represent a bottleneck, /b/ in English, /v/ in Greek, may be arbitrary, but passing air through closed lips is not random.

A second two-line inscription classified as Proto-Sinaitic, known as the Wadi el-Hol inscription, was published almost one century after the first (Darnell et al. 2005). This one was discovered during1994-1995 excavations in the heart of literate Egypt, on an ancient road between Thebes and Abydos. It was dated to 1900-1800 BC. To celebrate the discovery, New York Times wrote (Wilford 1999):

The first experiments with alphabet thus appeared to be the work of Semitic people living deep in Egypt, not in their homelands in the Syria-Palestine region, as had been thought. Although the two inscriptions have yet to be translated, other evidence at the discovery site supports the idea of the alphabet as an invention by workaday people that simplified and democratized writing, freeing it from the elite hands of official scribes. As such, alphabetic writing was revolutionary in a sense comparable to the invention of the printing press much later.

And continues citing the authors of the discovery:

This gives us 99.9 percent certainty… It was the accidental genius of these Semitic people who were at first illiterate, living in a very literate society. Only a scribe trained over a lifetime could handle the many different types of signs in the formal writing. So, these people adopted a crude system of writing within the Egyptian system, something they could learn in hours, instead of a lifetime. It was a utilitarian invention for soldiers, traders, merchants. The scholars who have examined the short Wadi el-Hol inscriptions are having trouble deciphering the messages, though they think they are close to understanding some letters and words. A few of these signs just jump out at you, at anyone familiar with proto-Sinaitic material. They look just like one would expect. Scholars said they could identify shapes of letters that eventually evolved from the image of an ox head into A and from a house, which looks more like a 9 here, into the Semitic B, or bayt. The only words in the inscriptions the researchers think they understand are, reading right to left, the title for a chief in the beginning and a reference to a god at the end.

According to the authors, associated scholars, and journalists, the alphabet did not develop gradually by a worldwide literate community seeking better expression. It was invented by a handful of illiterate but ingenuous immigrant workers. Unfortunately, nobody seems to have listened to them, since the traditional hieroglyphic writing continued to be in use for many centuries – unlike the global explosion of alphabetic writing at the beginning of the 1st millennium – and none of their signs was finally adopted, even in their homeland. The new script was good for soldiers, traders, and merchants, but only illiterate miners seem to have used it again some four hundred years later in the Serabit el-Khadim mines, Sinai. The scholars admit having trouble deciphering the signs but are 99.9% certain that the text is Semitic! They see a title of a chief and the name of a god but do not know what language that is!

Figure 2. The Wadi el-Hol inscription (1900-1800 BC; marked as public domain) transliterated using Egyptian and Cretan hieroglyphs or Linear A signs (Gardiner, CHIC, or SigLA numbered, respectively): 1=CHIC 001 #058.d; 2=Gardiner V28 or SigLA A339; 3=CHIC 061 #062.a; 4=SigLA A326 or CHIC 071 #059aB; 5=CHIC 011 #072.a; 6=CHIC 001 #041.b; 7=CHIC 062 #062.bB or #062.cB (rotated); 8=CHIC 007 #070 (SigLA A338); 9=CHIC 061 #058.a; 10=CHIC 001 #041.b; 11=CHIC 061 #058.a; 12=CHIC 062 #062.a; 13=CHIC 061 #058.a; 14=SigLA A370; 15=CHIC 035 #115.a or SigLA AB58; 16=CHIC 001 #058.d; 17=CHIC 061 #062.a (091) or Gardiner N35; 18=CHIC 169 #108.a (or 077) or SigLA A311; 19=CHIC 070 #112.a; 20=CHIC 001 #058.d; 21=CHIC 001 #041.b; 22=Gardiner W3; 23=Gardiner T2; 24=CHIC 070 #112.a; 25=CHIC 007 #070 (SigLA A338); 26=CHIC 169 #108.a (or 077) or SigLA A311; 27=CHIC 011 #072.a; 28=SigLA A370.

Instead, we can be 99.9% certain that the scholars who examined the Wadi el-Hol inscription did not consider Cretan or other scripts. Fig. 2 shows the Wadi el-Hol inscription with transliteration using attested Cretan and Egyptian signs. The characters 2, 22, and 23 are common Egyptian hieroglyphs absent from the Cretan script databases I have searched. All the other signs, i.e., about 90% of the inscription, are identical to contemporary Linear A signs found in Minoan Crete or present the minor idiosyncratic variation expected between scribes. In such casual writing, it is impossible to tell if, for example, the differences among the little fellows at positions 6, 10, and 21, and their Cretan counterparts, are of any linguistic relevance. Some of the signs are simply rotated. In their original publication of the inscription (Darnell et al. 2005), the authors meticulously compare each sign to Egyptian hieroglyphics to demonstrate the continuity of the pictographic forms. But the Egyptian phonetic values of corresponding hieroglyphics do not produce meaningful sequences in either Egyptian or Semitic language. Either Egyptian phonetics are not applicable, or the language is unknown or both. Based on a Proto-Sinaitic-like inscription found in Timna in 2009, Wimmer estimates that the so-called Proto-Sinaitic script dates after the late Eighteenth Dynasty (New Kingdom), i.e., it only starts in the 14th century and lasts until the 12th century BC (Wimmer 2010).

Claims

The so-called Proto-Sinaitic script consists of characters seen in Egypt and much more frequently in Crete. Its relations to later Canaanite scripts, the Phoenician alphabet, and Biblical stories are made up without evidence.

References

Benner, Jeff A. 2021. “Serabit El-Khadim Inscription.” Ancient Hebrew Research Center.

Brown, Raymond A. 1985. Evidence for Pre-Greek Speech on Crete from Greek Alphabetic Sources. Amsterdam: A M Hakkert.

———. 2021a. “Some Purported ‘Translations’ of the Lemnos Stele.” Eteocretan Language Pages.

———. 2021b. “Some ‘Translations’ of the Epioi Fake.” Eteocretan Language Pages.

———. 2021c. “Welcome to These Pages.” Eteocretan Language Pages.

Colless, Brian E. 2014. “The Origin of the Alphabet: An Examination of the Goldwasser Hypothesis.” Antiguo Oriente 12: 71–104.

Darnell, John Coleman, F W Dobbs-Allsopp, Marilyn Lundberg, P Kyle McCarter, and Bruce Zuckerman. 2005. Two Early Alphabetic Inscriptions from the Wadi El-Hôl: New Evidence for the Origin of the Alphabet from the Western Desert of Egypt. The Annual of the American Schools of Oriental Research. Vol. 59. Boston, MA: American Schools of Oriental Research.

Dickson, P. (2006). Dictionary of Middle Egyptian. Open source.

Flinders-Petrie, William Matthew, and Charles Trick Currelly. 1906. “The Lesser and Foreign Monuments.” In Researches in Sinai, 122–41. New York, NY: E.P. Dutton and Company.

Gardiner, Alan Henderson. 1916. “The Egyptian Origin of the Semitic Alphabet.” The Journal of Egyptian Archaeology 3 (1): 1-16 (23 pages).

———. 1957. Egyptian Grammar: Being an Introduction to the Study of Hieroglyphs. 3rd ed. Oxford: Griffith Institute.