Epigraphy in Crisis—Dating Ancient Semitic Inscriptions

042

Joan Scheuer’s fascinating volunteer report in the January/February issue (“Searching for the Phoenicians in Sardinia,” BAR 16:01) was especially interesting to me as a student of paleography. She describes a very important Phoenician inscription, known as the Nora Fragment, and explains how Professor Frank Cross of Harvard is able to date the Phoenician presence on Sardinia to the 11th century B.C. because he dates the Nora Fragment to this time.

Relying on Professor Cross’s publication in a scholarly journal,¹ Scheuer explains that “Cross has been able to date the inscription on the basis of the shape and stance of the letters”; that is, he dates the inscription paleographically—on the basis of a typology of letter forms of the Semitic alphabet in which Phoenician is written. Based on Cross’s paleographic analysis, Scheuer declares the Nora Fragment to be “the oldest Semitic inscription ever found in the central or Western Mediterranean. It strongly supports the theory [she says, again relying on Cross] that Phoenicians were present in Sardinia in the 11th century, or even earlier.”

May I present a dissenting view?

Based on the corpus of relevant inscriptions presently available, it would be extremely unwise to date a West Semitic inscription to a particular century from the 11th to ninth centuries B.C. on the basis of paleography alone. The reason is simple: Many of the letter forms persisted for a period of at least 200 years. So it is impossible to tell on the basis of these letter forms alone when the inscription was made—whether in the 11th century or the ninth century.

Moreover, we must also bear in mind that stone was not the customary material on which to write. A model of the inscription to be incised in the stone would be given to an often illiterate artisan who could easily vary the “pure” form on the model simply because of his ineptness.

To date one of these inscriptions more closely than “11th to 9th centuries B.C.,” we must therefore consider a number of other factors: the origin of the inscription, the artifact on which it is carved, the iconography with which it is associated, historical associations with the content of the inscription, etc.

The failure to do so can—and has—led to several scholarly errors.

Take the case of the justly famous and important Tell Fakhariyah inscription.^a It was carved on the skirt of a life-size black basalt statue of a king. On the front side of the skirt is a long text in cuneiform letters in the Akkadian language; on the back side is the same text in an early form of the Semitic alphabet, written in the Aramaic language. It is the latter inscription that concerns us here.

The Tell Fakhariyah inscription was found accidentally by a farmer plowing his field near the headwaters of the Khabur River, in northern Syria, near the Turkish border. The inscription was first fully published in 1982.² In this editio princeps, the editors dated the inscription to the mid-ninth century B.C., based on historical allusions in the text, as well as on iconographic features of the statue and some stylistic affinities of the inscription’s cuneiform version with Assyrian inscriptions of Ashurnasirpal II, who ruled in the mid-ninth century B.C. (883–859 B.C.). They arrived at this dating despite the paleographic peculiarities of the Aramaic inscription that would allow its script to be dated to the 11th century B.C., according to the evolution of alphabetic writing as known to us before the discovery of the Tell Fakhariyah statue.

The debate that ensued can easily be envisaged: Do we base the date of the statue and its inscriptions on the content of the text and its historical allusions, on the stylistic characteristics of the statue and other such factors? Or do we ignore this evidence and date the statue on paleographic grounds alone.

043

Joseph Naveh,³ an Israeli authority in the field of West Semitic paleography, is inclined to defend the paleographic “evidence” against the conclusions based on other grounds. On the other hand, a doctoral dissertation written in Tubingen, West Germany, by Helene S. Sader (under the supervision of W. Rollig) dates the Tell Fakhariyah inscription “most likely” to the eighth century B.C.⁴ Not surprisingly, she relies mainly on late iconographic parallels. (In my view, the iconographic parallels she relies on are most likely a century older than she dates them. They should be dated to the ninth century. But this is less important for the problem we are discussing here.)

If we conclude that the Tell Fakhariyah statue dates to the ninth—or eighth—century B.C., we must also conclude that the early letter forms, called “proto-Canaanite,” which we previously thought could well be found no later than the 11th century B.C., were in fact used as late as the ninth or perhaps the eighth century B.C.

This conclusion confronts the paleographers with a dilemma. Do they stand by their 11th century B.C. date for these proto-Canaanite letter forms or do they accede to the ninth- to eighth-centuries B.C. date implied in the non-paleographical evidence in the Tell Fakhariyah inscription?

In short, West Semitic paleography is at a crossroads. It must either admit that it alone has too meager means for dating early inscriptions with a higher degree of precision than within the limits of two or even three centuries, or question the historical, iconographic and factual evidence.

The Tell Fakhariyah inscription is especially important because it is long and therefore contains numerous examples of the letters that are considered to be, for typological purposes, the most archaic. These letters are the daleth (D, d), the kaph (K, k), the lamed (L, l), the mem (M, m), the sade (TS, x) or the ‘ayin (a guttural sound, [). All these archaic letter forms would indicate, according to the kind of paleographic analysis often favored before the discovery of the Tell Fakhariyah inscription, that the Tell Fakhariyah inscription was from the 11th century B.C. But the non-paleographic evidence for dating the Tell Fakhariyah inscription now undermines this analysis.

Moreover, other inscriptional evidence casts further grave doubt on the early dating of the Tell Fakhariyah inscription. For example, an inscription on a stone slab known as the Bir Hadad Stele was found about five miles north of Aleppo, Syria, in the 1930s.⁵ The archaic letter ‘ayin is a circle with a dot in the center. It was thought that the dot was no longer used after the 11th century, so an inscription with a dot in the ‘ayin could be dated to the 11th century B.C. But the Bir Hadad Stele clearly belongs to the late ninth century B.C., and it has an ‘ayin with a dot in the center. What this shows is that this archaic letter form was used as late as the ninth century B.C.

The ‘ayin with a dot in the center also appears on the Nora Fragment from Sardinia; Cross uses this letter form to date the Nora Fragment to the 11th century. But I believe the ‘ayin with a dot in the center simply continued in use right into the ninth century, as both the Bir Hadad Stele and the Tell Fakhariyah inscription demonstrate. (The Nora Fragment is probably contemporaneous with the Nora Stone, a longer Sardinian inscription that Cross himself dates to the ninth century, as discussed in Scheuer’s article.)

Cross makes another argument for the 11th-century date of the Nora Fragment. Because it is so fragmentary—it consists of only nine letters—Cross is able to read it upside down; that is upside down compared to the way the scholar who originally published the Nora Fragment read it, followed by the editors of the Corpus Inscriptionum Semiticarum I, no. 145.⁶ Then Cross reads the first line from right to left and the second from left to right.⁷ This is known as boustrophedon writing; the name means “as the ox plows,” which is how some early inscriptions are written, changing direction with each line. This practice died out in Canaanite or Phoenician scribal circles toward the middle of the 11th century B.C., so Cross uses this to buttress his argument that the Nora Fragment dates back to the 11th century B.C. But Cross can read the fragment in boustrophedon fashion only because he reads it upside down. There is no doubt, however, that the slightly damaged letter read by Cross as a left-looking lamed is a normal ninth- or eighth-century ‘aleph that Cross has read upside down. Therefore, the inscription cannot antedate the ninth century.

Accordingly, the Nora Fragment cannot be used as evidence that the Phoenicians were in Sardinia in the 11th century B.C. But the point is broader: We must be very careful in 049dating ancient Semitic inscriptions on the basis of paleography alone. We are still in the process of developing a reliable typology of the development of letter forms before the ninth century B.C. More material is badly needed before we can create a letter typology on firm ground. When, if ever, this will come, we do not know. Unfortunately, it depends on luck and on the vagaries of the archaeologist’s spade.

This analysis also has important implications for the date of the Greek borrowing of the Semitic alphabet. Some have recently argued that because of the appearance of what were thought to be very archaic letter forms in the earliest Greek inscriptions (from the eighth century B.C.), the borrowing had to have taken place before these archaic characteristics disappeared from the Semitic alphabet. On this basis, the borrowing was assigned to the 11th century B.C. Now we know that these archaic characteristics did not disappear after the 11th century, but in fact persisted in the Semitic alphabet for several centuries thereafter. So the argument for an 11th-century borrowing is undermined. Indeed, the Greek letters that preserve the archaic letter forms in inscriptions not earlier than the eighth century B.C. indicate that the Tell Fakhariyah script was still in use in the ninth to eighth centuries B.C., not only at this site 200 miles east of the Mediterranean, but also at the fringes of the Greek world. If this is true, it becomes much more likely that the Greeks borrowed the Semitic alphabet not in the 11th century B.C., but sometime in the eighth century B.C.

Joan Scheuer’s fascinating volunteer report in the January/February issue (“Searching for the Phoenicians in Sardinia,” BAR 16:01) was especially interesting to me as a student of paleography. She describes a very important Phoenician inscription, known as the Nora Fragment, and explains how Professor Frank Cross of Harvard is able to date the Phoenician presence on Sardinia to the 11th century B.C. because he dates the Nora Fragment to this time. Relying on Professor Cross’s publication in a scholarly journal,1 Scheuer explains that “Cross has been able to date the inscription on the basis of the shape and stance of […]