The Bat Creek Inscription: Did Judean Refugees Escape to Tennessee?

046

047

At the beginning of the First Jewish Revolt against Rome (66–73 or 74 A.D.), “the inhabitants of Caesarea massacred the Jews who resided in their city,” says the first-century Jewish historian Flavius Josephus. “Within one hour, more than 20,000 were slaughtered, and Caesarea was completely emptied of Jews.”¹ Is it possible that a few of these Caesarea Jews escaped by sea from this famous port city and eventually ended up in Tennessee?

In 68 A.D., the Jews of Joppa, “finding themselves cut off from the country, which had passed into the enemy’s hands, … built themselves a fleet of piratical ships and made raids on the traffic [in the eastern Mediterranean].”² When the Roman infantry and cavalry attacked Joppa, these people took refuge in their ships just out of range of the Romans’ weapons. But a gale accomplished what the Romans had failed to do by dashing the fleet against the rocky shore, killing 4,200 Jews. Although Josephus implies that none survived, he notes that the fleet attempted to make it to the open sea.³ Thus we may legitimately wonder whether any of these ships escaped, and, if so, where they could have gone.

Clearly, Jewish refugees from these cities had both the opportunity and the incentive to flee by sea from Roman Judea. However, the whole Mediterranean was dominated by Rome and therefore would have offered no safe haven. Did they seek refuge beyond the Strait of Gibraltar? Did they land in America? A remarkable find known as the Bat Creek inscription, dating to approximately the same period as these events, provides some evidence that they did.

In 1889,⁴ the Smithsonian Institution’s Bureau of American Ethnology, in the course of its Mound Survey Project, discovered a small inscribed stone in an undisturbed burial mound on the Little Tennessee River.⁵ Located at the mouth of Bat Creek, some 40 miles south of Knoxville, Tennessee, the mound was the smallest in a group of three; it measured 28 feet in diameter and 5 feet high. The excavator was John W. Emmert, described in the final report as a “regular assistant” to the project.⁶ Emmert found “nothing of interest” until he reached the bottom of the mound, “where nine skeletons were found lying on the original surface of the ground … ”⁷ The badly decayed skeletons lay in two parallel rows, with two in one row and seven in the other. Eight of the skeletons lay with their heads to the north, but one skeleton, in the group of two, lay with its head to the south. It was under the skull of this particular skeleton that Emmert found the inscribed stone.

About 4.5 inches long and 1.75 inches wide, the light-tan siltstone is encrusted on the engraved side with a dark brown layer of iron oxide.⁸ Most of the extant eight characters of the inscription have been scratched through the dark crust to expose the light interior color, although a few do not penetrate in places. The lower left corner of the stone is broken off, obliterating most of a ninth character. Two vertical strokes at the top left of the stone, sometimes shown in drawings of the inscription, are shallow test scratches that were made after the stone’s discovery.⁹

In his 1894 final report, project director Cyrus Thomas published a photo of the inscription and identified it as being “beyond question letters of the Cherokee alphabet said to have been invented by George Guess (or Sequoyah), a half-breed Cherokee, about 1821.”¹⁰ He apparently based his conclusion more on the find’s location near the heart of historical Cherokee territory than on any actual similarity to the Cherokee syllabary.¹¹

The stone then lay virtually forgotten in storage in the National Museum of Natural History in Washington, D.C., until the 1960s, when a few observers noticed that, if the published photograph is held upside down, several of the letters look much more like Phoenician and related Canaanite scripts than like Cherokee, when held in either orientation.¹² Joseph B. Mahan sent a photo of the stone to Professor Cyrus Gordon, a Semitic languages scholar renowned 050for his work on Ugaritic, a Canaanite language written in cuneiform. Gordon was also interested in possible pre-Columbian contacts between the Old and New Worlds.

After examining the photo, Gordon declared in 1970 that the signs on the Tennessee stone were in fact paleo-Hebrew letters of the first or second century A.D.¹³

Paleo-Hebrew is the Phoenician-like alphabet that was used to write Hebrew and other Semitic languages in First Temple times (before 586 B.C.). When the Jewish exiles returned from Babylon, beginning in 538 B.C., they replaced these old letters, for most purposes, with the Aramaic based, square-Hebrew alphabet that is still in use today. However, paleo-Hebrew continued to be the preferred script for coin legends whenever Judea threw off foreign domination. In addition, a handful of the Dead Sea Scrolls were written entirely in paleo-Hebrew;¹⁴ and several of the scrolls that were otherwise written in the post-exilic, square-Hebrew script used the old letters to write the unutterable name of God, YHWH (Yahweh or Jehovah). Except in its distinctively Samaritan form, the paleo-Hebrew script is not known to have been used since 135 A.D., the date when Rome finally suppressed the Second Jewish Revolt.¹⁵

Gordon argued that the letters on the Bat Creek stone most closely resemble those on coins of the First and Second Jewish Revolts against Rome. He therefore dated the inscription to that approximate period.

The inscription’s eight complete letters are arranged with seven on the top line and a single letter standing alone below this line. A comma-like word divider appears after the first two letters of the top line (reading from the right, as Hebrew is written). Gordo identified the five letters to the left of the word divider as reading lyhwd, meaning “for Judea.”

The yod (y, letter 4 in the drawing in the sidebar) is a highly distinctive letter. The particular stance found here is precisely that of the Jewish Revolt coins, but not that of the First Temple period (pre-586 B.C.). Finding it flanked as it is here by a clear-cut paleo-Hebrew lamed (1, letter 3) and heh (h, letter 5, also spelled he) constitutes striking evidence of paleo-Hebrew.

Letter 6, which resembles an elongated “s” with a cross-bar, is a cursive-influenced variant of the waw (w) that appears on many of the Jewish Revolt coins. The precise Bat Creek form appears on a fourth-century B.C. seal of a governor of Samaria.¹⁶ Letter 7 is not well made but is still recognizable as daleth (d, also spelled dalet).¹⁷

Gordon proposed that the remnant of the broken letter, at the left end of the inscription, belonged to a mem (m, letter X). With this addition, the word would become plural, so that the letters following the word-divider would read lyhwdm, or “for the Judeans.” The association of this word with Judea is therefore not altered by the additional letter.¹⁸

Gordon admitted that the other three letters on the Bat Creek stone are not so positively identifiable. Following a suggestion by Robert R. Stieglitz,¹⁹ he tentatively identified letter 2, to the right of the word divider, as a qoph (q) and the rightmost letter (letter 1) as zayin (z). The first line would then read, “A comet for the Judeans.” Gordon argued that this would make sense as a personal title for a Judean leader in view of the Star Prophecy, which refers to a “star” that would arise from Jacob to annihilate Israel’s foes (Numbers 24:17). Gordon tentatively identified letter 9, the lone letter below the others, as an aleph (’) and suggested that this may be a numeral, indicating “Year 1” of some era, as on the coins of the First Jewish Revolt.

It is my own observation that letter 1 looks more like a backwards resh (r) than a zayin. However, while rq does mean “but” in Hebrew, the phrase “But for the Judeans” does not make much sense without any 051context. The lone letter in the second line could, in my view, just as easily be a second type of waw, or even a samekh (s), as an aleph. As waw, it would indicate “Year 6, ” rather than “Year 1.”²⁰

Incredulous scholars, who had learned well in school that there were absolutely no Old World contacts with the New World before Christopher Columbus’ 1492 voyage, apart perhaps from a minimal Viking settlement at the northern-most tip of Newfoundland, generally greeted Gordon’s reading of the Bat Creek stone with a stone wall of silence. For two decades, no Semitic scholar came forward in print either to agree with Gordon’s reading or to explain what, if anything, is wrong with it.

Indeed, very few scholars so much as mentioned in print that such a thing as the “Bat Creek inscription” even exists. One notable exception was the late Glyn Daniel, former editor of the prestigious journal Antiquity, who summarily wrote off Gordon in a lead editorial as a “nonscholar” for his “belief” in the Bat Creek stone, among other things. Daniel did not find it necessary to give any explanation of what he or anyone else found wrong with Gordon’s reading of the inscription.²¹

Fortunately, a few other artifacts were found with the same burial as the Bat Creek stone, including a pair of cupreous bracelets and some wood fragments.²²

When first found, the bracelets were classified as “copper.” Artifacts made of pure native copper from Lake Superior are not unusual in ancient mound burials in the eastern United States, so nothing more was thought of them until Gordon’s 1970 announcement. At that time, the Smithsonian analyzed them and found that they were not domestic native copper at all, but rather copper alloyed with approximately 27 percent zinc and 3.3 percent lead. In other words, they were heavily leaded yellow brass.²³ Brass is a synthetic alloy believed not to have been made in the New World before Columbus.

Brass was well known in the ancient Mediterranean world, particularly after 45 B.C., when Roman brass or orichalcum coins became plentiful. In 1970, when Bat Creek bracelets were analyzed, however, it was believed that Romans brass containing more than 22 percent zinc never contained more than 1 percent lead. An information sheet distributed by the Smithsonian’s anthropology department therefore concluded that “recent tests by our Conservation Laboratory on the brass bracelets found in the same grave [ as the Bat Creek stone] definitely established that they are 18-19th century trade goods and do not have the chemical composition of the Roman or early Semitic periods.”²⁴

In 1977 and 1978, however, Paul Craddock of the British Museum published a series of articles on ancient copper alloys, citing numerous examples of ancient artifacts with essentially the same composition as the Bat Creek bracelets-including a first-century B.C. Egyptian statuette of “Hermes Leading a Lady,” from the British Museum’s own collection.²⁵ However, Craddock noted that brass with such a high zinc composition became Substantially less common after the second century A.D. Although on purely metallurgical grounds the Bat Creek bracelets could be modern, they therefore could equally well be ancient; and if ancient, they are most likely of the same approximate period that Gordon chose for the inscription on paleographical grounds, without knowledge of the precise composition of the bracelets or its significance.

The bracelets are somewhat unusual in that they were wrought from a flat bar that was folded lengthwise into a circular cross section before being bent into a c-shape. This contrasts with most mass-produced, modern trade goods, which were drawn or cast. Thus each one of the Bat Creek bracelets has a seam and a hollow core that run its entire length. 052Otherwise, the bracelets are completely plain and therefore provide no clue as to which of the many Old World cultures could have been responsible for their manufacture.²⁶ The only evidence we have for this is the inscription itself.

The wood fragments found with the inscribed stone and with the brass bracelets were too small to be radiocarbon dated by the traditional method available in 1970. Thus, it was anyone’s guess at that time how old they might be. Archaeologist Marshall McKusick, in another of the rare published acknowledgments of Gordon’s reading of the Bat Creek stone, claimed in a 1979 letter to Biblical Archaeologist that the wood’s failure to decay in the moist environment of the mound demonstrated that the burial must date from the period of historic European contact, and therefore would be much too recent to contain an ancient Hebrew inscription, as Gordon claimed.²⁷

However, McKusick overlooked the fact that the wood fragments were found saturated with copper salts from the bracelets²⁸ and that organic material thus preserved by copper occasionally survives for thousands of years without decaying.²⁹ In the 1970s, scientists developed the new AMS method of radiocarbon dating that requires only a few milligrams of carbon. In 1988, with funding from the Institute for the Study of American Cultures (a private foundation in Columbus, Georgia), the Smithsonian allowed such a test to be performed, using the ETH linear accelerator in Zurich (the same machine on which the recent carbon dating of the Dead Sea Scrolls was performed). The wood dated to the period 32 to 769 A.D. The unusually broad time span in this dendrochronologically calibrated dating was due to the small amount of carbon recovered from the 30 milligrams of material the Smithsonian submitted to the laboratory. Retesting with more material—5.5 grams is available—would give greater precision.³⁰

Since the brass bracelets were found in an undisturbed context in close association with the wood fragments, the carbon date resolves the question of the bracelets’ age. These artifacts taken together provide conclusive physical evidence of either some sort of a pre-Viking—not to mention pre-Columbian—Old World contact with the New World, or the equally remarkable independent invention in the New World of the trick of making brass. Together the bracelets and the carbon-14 date point to the period 32 to 200 A.D. This correlates remarkably well with Gordon’s paleographic dating of the inscription to the first or second centuries A.D.

All this still leaves unanswered the question of how these hypothetical Semitic voyagers could have ended up in eastern Tennessee rather than closer to the coast. The Tennessee River itself is not naturally navigable past Muscle Shoals in northern Alabama, so at least one major portage would have been necessary to have reached the Little Tennessee from the Gulf coast via the Mississippi, Ohio and Tennessee rivers. Moreover, the Great Smoky Mountains pose a formidable barrier to overland travel to this region from the east coast.

However, the region in which the Bat Creek stone was found is not as out-of-the-way as it first appears. The current view is that the very first Old World visitor to the interior of the eastern United States after Columbus, namely Hernando De Soto from 1539 to 1543, simply left his ships in Florida, scrambled over the mountains to the north of the Smokies and came down the Tennessee River to camp for six days on an island at the mouth of the Little Tennessee River, just 12 miles downstream from the mouth of Bat Creek. De Soto’s main force remained there during this period, but he is believed to have sent cavalry up the Little Tennessee to explore its valley.³¹ Such a foray would have passed within a pike’s throw of the Bat Creek mound group.

053

For whatever reason, the lower Little Tennessee Valley evidently attracted early modern explorers and may have been equally attractive to explorers in the first or second century A.D.

My 1988 publication of the Bat Creek radiocarbon date and bracelet composition in Tennessee Anthropologist elicited a major response in the same journal by anthropologists Robert Mainfort and Mary Kwas.³² For analysis of the Hebrew aspects of the find, their 1991 paper relies entirely on the views of Frank Moore Cross, former Hancock Professor of Hebrew and Other Oriental Languages at Harvard University, as privately communicated to them. Cross’ comments, as reported by Mainfort and Kwas, constitute the first published critique by any Semitic language scholar of the Gordon-Stieglitz reading of the inscription.

In his carefully considered letter to Mainfort, Cross objected, among other things, to Gordon’s reading of lyhwd as “for Judea,” on these grounds: “In Palaeo-Hebrew, Judah (Judea) is spelled yhwdh not yhwd. The latter is the Aramaic designation and appears only in Aramaic [not paleo-Hebrew] scripts.”³³

Evidently Cross forgot about the 1972 discovery of the “Abba” inscription, a late Second Temple paleo-Hebrew tomb inscription cut into the bedrock of Jerusalem. Epigraphers Joseph Naveh and E. S. Rosenthal have both identified the language of this seven-line inscription as Aramaic, not Hebrew.³⁴ Unless Naveh and Rosenthal are both mistaken about the language, Cross is wrong to assert that paleo-Hebrew was never used to write the Aramaic language. In fact, the Abba inscription actually contains the very word yhwd in paleo-Hebrew letters. Thus Cross is wrong in saying that yhwd in paleo-Hebrew letters, as in the Bat Creek inscription, is “anomalous.”

In the Abba inscription, yhwd happens to be a personal name, but it is a matter of minor consequence whether the reference in the Bat Creek inscription is to a person named Judah, to the land of Judah, or to the people of Judah, or whether its language is Hebrew per se or Aramaic. In any case, it indicates that someone from Roman-era Judea ended up in the New World.

In an Associated Press interview, Cross went on to declare that the Bat Creek inscription “doesn’t fit any paleo-Hebrew script that I know,” and in a communication to me stated that even “had the stone been found in a Roman level in Jerusalem I would publish it as an enigmatic graffito and pronounce it undecipherable.”³⁵ Cross is, of course, one of the world’s foremost experts on early Hebrew scripts, and as such his opinion carries great weight.

My own conviction is that the Bat Creek inscription is a rustic, and therefore imperfect, specimen of paleo-Hebrew. As such, however, it does not by itself indicate anything more than a minimal contact with the New World by a few Hebrew sailors, who may have had no impact on New World culture, who may not have left either descendants or communicable diseases behind, who may have had no idea where they were going when they left and who may never have returned to the Old World to tell where they had been.

At the beginning of the First Jewish Revolt against Rome (66–73 or 74 A.D.), “the inhabitants of Caesarea massacred the Jews who resided in their city,” says the first-century Jewish historian Flavius Josephus. “Within one hour, more than 20,000 were slaughtered, and Caesarea was completely emptied of Jews.”1 Is it possible that a few of these Caesarea Jews escaped by sea from this famous port city and eventually ended up in Tennessee? In 68 A.D., the Jews of Joppa, “finding themselves cut off from the country, which had passed into the enemy’s hands, … built themselves a fleet […]