Of the thousands of languages spoken before modern times, a handful developed scripts. The descendants of a few of those are still in use; of a handful of the others, traces have survived the millennia to be discovered by travelers or excavated by archeologists. If the modern descendants diverged considerably from the ancestors, or if there are none, the ancestors need to be deciphered. It is always a script, rather than a language, that is deciphered; if an (ancient) record can be pronounced, but not understood (as with Etruscan), the problem is one of interpretation rather than decipherment.
Decipherment of Scripts.
The prerequisite to decipherment is accurate reproduction of the enigmatic inscription. This may seem trivial now that photography has been commonplace for a century and a half, but before 1850 or so—during the quarter millennium when most of the ancient scripts were first found—only the artist's subjective eye served to disseminate an inscription's appearance. The artist's drawings further passed through the engraver's burin and, at the end of the period, the lithographer's crayon. Only with photography, perhaps, does it become apparent how subjective the artist's vision is: witness the depictions of the Sphinx at Giza or the Assyrian bulls that vary with stylistic fashion and the drafting skill of individual artists. The task is especially difficult when the artist is trying to reproduce unfamiliar writing, where apparently trivial details may distinguish two characters (compare the Roman letters C and G), but rather great differences may be completely insignificant (a Roman “t” may be made with or without the lower curve to the right, on the writer's whim). Successive attempts to delineate ancient inscriptions—but not with increasing accuracy—have been published (Daniels, 1988).
Given accurate copies of a graphic image, it must first be decided whether the image is in fact writing. The first cuneiform inscriptions, found on doorframes and such at Persepolis, were sometimes considered mere decoration. [See Persepolis.] Kufic Arabic inscriptions in mosques are so stylized as to be almost purely decorative rather than linguistic. [See Mosque Inscriptions.] The masons' use of South Arabian letters to guide the arrangement of paving stones at Hajar Kohlan provided the first clue to the canonical order of that script. [See South Arabian.] The letters' use was functional rather than linguistic, but their function derived entirely from an incidental property of scripts that their characters are learned in a fixed order. It is uncertain whether “potters' marks” dating to the era of the presumed development of Proto-Canaanite writing that might be letters of a script are in fact such or are arbitrary designs with some function or other. [See Proto-Canaanite.] Two modes of decipherment ensue when the marks are established as writing. In the exceptional case, the decipherer is very lucky and the new script occurs alongside an inscription in a known language—in a bilingual text. Usually, it is safe to assume that the two texts render the same content—one is a translation of the other (or both are translations of a third).
More often, the unknown script stands alone, and it is up to the ingenuity of the decipherer to discover a virtual bilingual: some stretch of text in a known language that can be presumed to equate to some stretch of text in the unknown one. Most commonly, virtual bilinguals have been at a phonetic level (personal or place names)—although a morphological/semantic level (numbers) figured in at least one case—and sometimes purely graphic (shapes of letters in a known script). This last category comes into play whenever a new text, even in a very familiar language, is unearthed because the ductus of no two scribes is identical. In addition, ancient inscriptions are sufficiently rare that normal variation in script or language will be magnified by the absence of intermediate forms. The history of decipherment is studded with examples of poorly chosen virtual bilinguals, however, and it is difficult to dissuade the proposers from their beliefs.
While the decipherer may formulate a hypothesis about the identity or relationship of the language concealed behind the enigmatic script, this has sometimes proved to be a misstep, misdirecting the investigator. (Most notably, the assumption that Linear B must be some pre-Greek substrate language such as Etruscan delayed success on that script.) It is more useful simply to identify and count the different characters: a small number, about thirty, suggests an abjad (consonantary) or an alphabet; greater variety, about one hundred or more, suggests a syllabary or an abugida, and several hundred or more, a logosyllabary (or a logography—no purely logographic script, if such has ever truly existed, has been deciphered; candidates include Proto-Elamite, Indus Valley script, and Easter Island script). [See Writing and Writing Systems.]
The surest results have been achieved by those who concentrated on proper names. Such cases include both the first, (Palmyrene, by Jean-Jacques Barthélemy, In 1756), and the two most familiar decipherments (Egyptian, by Jean-François Champollion, In 1823, and Linear B, by Michael Ventris, In 1953), as well as the one that is arguably the most important of all, the Old Persian and Mesopotamian cuneiform scripts, whose decipherment was begun by Georg Friedrich Grotefend (1802) and climaxed by Edward Hincks (1846–1852). [See Cuneiform.] The decipherments of Ugaritic and Himyaritic illustrate different kinds of virtual bilinguals (see Daniels, 1995; Daniels and Bright, 1996, sec. 9). [See the biography of Champollion.]
Palmyra was an independent pagan monarchy on the eastern edge of the Roman Empire. It was fairly well known in the eighteenth century from classical historians. Monuments written in its language, known to be similar to Syriac, had been known in Europe since antiquity and had received the attention of antiquarians since the early 1600s. It was not until 1753 that adventurers traveled to the ruins of the city and brought back accurate copies of some of its inscriptions. Many of the inscriptions in the Palmyrene script and language were accompanied by Greek inscriptions, and it took no great insight to guess that the pairs were equivalent. Barthélemy (1716–1795), in fact, reports that it took him no more than two days to decipher the script. From the thirteen inscriptions that had been accurately published by Robert Wood and James Dawkins (English adventurers who traveled to the ruins of the city In 1753 and brought back accurate copies of some of its inscriptions), Barthélemy chose one in which the Greek parallel began with a proper name, Septimios. The seventh letter of the Palmyrene inscription was the same as the first, so (omitting short vowels) there seemed to be a correspondence; the next word, Ouorōdēn in the Greek, confirmed the 〈w〉 that spelled [o] and gave 〈r〉 and 〈d〉 as indistinguishable (a characteristic of Aramaic scripts of the period). When the twenty-two letters were identified, their resemblance to the corresponding Hebrew and Syriac forms became clear. A handful of verb forms were nearly identical to Syriac words that could translate the Greek, demonstrating that the language was Aramaic as well (Daniels, 1988). [See Palmyra; Palmyrene Inscriptions; Greek; Aramaic; Syriac; Hebrew Language and Literature.]
Grotefend (see above) used the virtual bilingual of Persian kings' names (known from Herodotus and from formulae in recently deciphered Sasanian inscriptions) to begin the decipherment of the Old Persian version of the trilingual inscriptions from Persepolis. Work on them was greatly advanced by such pioneers of Indo-European and Indo-Iranian studies as Rasmus Rask and Christian Lassen. Henry Creswicke Rawlinson replicated the decipherment of Old Persian, using the huge trilingual inscription at Bisitun that he copied at great personal risk, but he probably knew of Grotefend's insightful virtual bilingual. Rawlinson usually receives credit for deciphering Mesopotamian cuneiform, but his work lagged behind that of Hincks. Although Rawlinson was stationed in Baghdad, he was kept abreast of Hincks's discoveries by a quite efficient postal system. [See Sasanians; Persepolis; Bisitun; Cuneiform; and the biography of Rawlinson.]
Edward Hincks (1792–1866) was a Church of Ireland country pastor. Unlike the celebrated Champollion and Ventris (see above), he was not a young man when he achieved his decipherment. He was immensely learned in all areas of classical and Oriental (i.e., Near Eastern) studies, and in his first approach to Old Persian he hoped better to understand Egyptian hieroglyphs. In his first lecture on the subject, on 9 June 1846, at the Royal Irish Academy in Dublin (and published in its Transactions), he demonstrated that the Old Persian characters denoted consonants plus particular vowels and not a wide variety of consonants. He also showed that the second of the three languages was written with syllabic characters that could be used in the consonant-vowel pattern CVi-ViC (ta-aš, not ta-š). (The second language is what is called Elamite, and it was decades before real progress was made in interpreting it.) [See Elamites.] The most important discovery reported in Hincks's second lecture (30 November) was the equivalence between two styles of cuneiform, which he called lapidary and cursive. This made available a great mass of material, which was being brought to the British Museum and the Louvre in Paris by the first excavators in Mesopotamia. In a third lecture (11 January 1847), he refined his readings of many signs and presented the cuneiform system of numerals. The numerals were taken from extensive inscriptions copied in the late 1820s by F. E. Schulz in western Armenia (now eastern Turkey) but not published until 1840 in Paris, and not seen by Hincks until 1846.
It was Schulz's inscriptions, in a language now called Urartian, that proved to be the most important to the decipherment. [See Urartu.] The longest one is a royal annal covering thirteen years; each year is introduced by the same formula, but the formula is not always written identically. Hincks discovered certain signs to be optional and identified them as vowel signs. This enabled him to verify the vowel included in preceding CV signs. Moreover, he analyzed the Urartian texts grammatically; his results, including the readings of 110 signs, were published in the prestigious Journal of the Royal Asiatic Society (1848). Six months later, Hincks dealt with the inescapable fact that many signs had more than one phonetic value, and semantic values as well. At that point, he suggested the possibility that a different language could underlie the script. Two years later, on 29 July 1850, at the Edinburgh meeting of the British Association for the Advancement of Science, he asserted that the writing system was devised for a non-Semitic language. At that meeting, also, he showed grammatical patterning in the Assyrian language that was fully parallel to that in Hebrew.
Rawlinson's treatment of the Assyrian version of the great Bisutun inscription was published In January 1852. Hincks's last article pertaining to the decipherment per se is dated 24 May 1852: it gives values to 252 characters and identifies elements of Rawlinson's publication that were taken from Hincks's work without acknowledgment. It also includes the first mention of ancient textual material that signaled the close of the decipherment phase of the interpretation of cuneiform. This was a fragment of a lexical list, which Hincks saw comprised a list of signs accompanied by a rendition of their pronunciation. Henceforth, most signs were to be identified from their appearance in such documents, of which thousands of fragments have been recovered from throughout the cuneiform world. Hincks's accomplishment was recognized in his own time; his undeserved eclipse by Rawlinson may be attributable to the personalities of the two men. Hincks was shy and unambitious, with no academic position, rarely leaving the village of Killyleagh, in Ulster. Rawlinson was energetic, well placed in the military and the British Museum, and endowed with dual hagiographers in his brother, a prolific historian, and the immensely powerful E. A. Wallis Budge, who ruled British Oriental studies for decades. Both authors downplay—and distort—Hincks's role. Subsequent chroniclers and historians have not known of the existence of the Transactions of the Royal Irish Academy (for details and references, see Cathcart and Donlon, 1983; Daniels, 1994).
It is Ugaritic that was deciphered by means of numerals, and also by the use of grammatical information. The first texts were excavated at Ugarit In 1929 and published immediately; Charles Virolleaud, the epigrapher, supposed from the location and date of the site and the finds that their language was likely to be Northwest Semitic. He was fortunate enough to discover a tablet that resembled Mesopotamian accounting documents and, in a column that might contain numbers (their names were spelled out), a sequence of symbols like XYX appeared. About the only common word in Semitic with such a consonant pattern is ṯalāṯ, “three.” Meanwhile, the eminent scientist Hans Bauer analyzed the letters that appeared at the beginnings or ends of words, or both, and compared them with the limited inventory of affixes in Semitic. Édouard Dhorme noted that a phrase at the beginning of what looked like a letter was identical to a phrase on a sword, with the addition of one letter at the beginning. He surmised that this was the letter l, representing the preposition to. The work of the three scholars converged, and the Ugaritic script was soon worked out (Corré, 1966). The unique use of three letters for ả, ỉ, and ủ was established by Johannes Friedrich. [See Ugarit; Ugaritic; Ugarit Inscriptions; and the biography of Virolleaud.]
A Near Eastern script deciphered via the shapes of the characters is the Himyaritic (i.e., Sabaean), as it was called by early researchers. Emil Rödiger discovered two unrelated Arabic manuscripts that contained lists of letters said to be musnad, “Himyaritic.” The two lists were strikingly similar, suggesting that they might preserve a genuine tradition; indeed, when the first inscriptions from South Arabia were published In 1837 and 1838, Rödiger was able to interpret them fairly successfully, using the manuscript abecedaries. His teacher, the eminent philologist Wilhelm Gesenius, mistrusting that evidence and relying more on comparative philology and limited data from a newly discovered South Arabian language, had an earlier success but went further astray. Unfortunately, the first somewhat extensive text available (ten lines) contained almost nothing but proper names (Daniels, 1986). [See Himyar.]
Interpretation of Languages.
With the decipherment of ancient scripts came the discovery that they could be used for languages other than those for which they were devised. The recovery of many intermediary forms has enabled the demonstration that all the West Semitic scripts (including Palmyrene, Ugaritic, and Himyaritic, as well as Hebrew and Arabic) belong to a single tree of descent. A productive assumption, as has been seen, was that the languages associated with the unknown scripts were similar to the known Semitic languages: Hebrew, Syriac (Christian Aramaic), Jewish Aramaic, and Arabic. The affinity of the Semitic languages had been recognized since the Middle Ages (previously called Oriental, they did not receive the name “Semitic” until 1781). The four classical Semitic languages are Hebrew, Arabic, Syriac, and Ethiopic, or Ge῾ez; the language of the Aramaic portions of the Bible was called Chaldee, as were the languages of the rabbinic writings. It is not yet clear when the Aramaic languages were identified as a group (characterized by the phonological change of Semitic interdentals to stops and by a particular selection of the ancestral verb stems); Barthélemy does not use the term of Palmyrene, but it was established by the mid-nineteenth century.
As the decipherment of cuneiform proceeded, several scholars came to believe that cuneiform Assyrian was a Semitic language. This was an assumption rather than a demonstration; however. It was not until Hincks identified the similarity in grammar as well as vocabulary that the identification could be regarded as certain. The other cuneiform languages are more difficult. Those for which the most materials are available are Sumerian, Elamite, Hittite, Hurrian, and Urartian. [See Sumerian; Hittite; Hurrian.]
The interpretation of Sumerian was aided somewhat by the discovery, early on, of tablets containing grammatical paradigms. The first interpretations coincided with a vogue for typological classification of language. The German-English philologist Friedrich Max Müller assigned nearly all the non-Indo-European, non-Semitic languages of Eurasia to a phylum he called Turanian (encompassing many more languages than the modern term Altaic), largely on the basis of their so-called agglutinative structure. Early Sumerologists, finding similar patterning in Sumerian, believed it, too, belonged in the phylum. They found a worthy opponent in Joseph Halévy (who was actually familiar with a number of Turanian languages). He, however, was of the opinion that Sumerian was not a real language at all, but a code devised by Assyrian priests to maintain the secrecy of the cult (Cooper, 1991). The discovery in the late nineteenth century of archives, notably at Telloh, of the most mundane documents written in Sumerian disproved the notion. Over the following century, attempts have been made to connect Sumerian with virtually every language family on the globe—most recently the Nostratic superphylum encompassing Indo-European and most of its neighbors. So little information is accessible to nonspecialists, however, that no credence can be given to these attempts.
The limited amount of Elamite material and its formulaic nature—even though it crosses several periods in nearly three millennia—make it difficult to analyze. The proposal by David McAlpin to connect it with the well-studied Dravidian family has attracted much notice and entered the handbooks, but it is not accepted by the small corps of Elamitologists. A dispassionate linguist will be immediately suspicious because McAlpin seems to have a Dravidian etymology for every Elamite word, an unthinkable occurrence. With cuneiform Hittite, the interpretation is certain. The Czech scholar Bedřich Hrozný (Daniels and Bright, 1996) noted a word that could be transcribed as watar in a context that demanded the meaning “water” and successfully interpreted the language as Indo-European. Urartian (so important in the decipherment of cuneiform) resisted interpretation for decades. Extensive studies were published by Archibald H. Sayce in the 1880s, but the ergative type of language was not yet familiar to philologists. Hurrian is known from essentially one document—physically, the largest cuneiform-inscribed clay tablet known—and from many enigmatic ritual bilingual texts with Hittite equivalents. All that can be said is that Hurrian and Urartian are related. It has been suggested they are ancestral to some languages of the Caucasus. Lexical connections with (Indo-European) Armenian have been proposed.
- Cathcart, Kevin J., and Patricia Donlon. “Edward Hincks, 1792–1866: A Bibliography of His Publications.” Orientalia 52 (1983): 325–356.
- Cooper, Jerrold S. “Posing the Sumerian Question: Race and Scholarship in the Early History of Assyriology.” Aula Orientalis 9 (1991): 47–66.
- Corré, Alan D. “Anatomy of a Decipherment.” Proceedings of the Wisconsin Academy of Sciences, Arts, and Letters 55 (1966): 11–20.
- Daniels, Peter T. “‘To Prove Him with Hard Questions’: The Decipherment of Himyaritic.” Appendix 1 of “How to Decipher a Script,” in Writing/Écriture, edited by Pierre Swiggers and Willy Van Hoecke. Louvain, forthcoming.
- Daniels, Peter T. “‘Shewing of Hard Sentences and Dissolving of Doubts’: The First Decipherment.” Journal of the American Oriental Society 108 (1988): 419–436.
- Daniels, Peter T. “Edward Hincks's Decipherment of Mesopotamian Cuneiform.” In The Edward Hincks Bicentenary Lectures, edited by Kevin J. Cathcart, pp. 30–57. Dublin, 1994.
- Daniels, Peter T. “The Decipherments of Near Eastern Scripts.” In Civilizations of the Ancient Near East, vol. 1, edited by Jack M. Sasson et al., pp. 81–93. New York, 1995.
- Daniels, Peter T., and William Bright, eds. The World's Writing Systems. New York, 1996.
- Friedrich, Johannes. Extinct Languages. Translated by Frank Gaynor. New York, 1957.
- Gordon, Cyrus H. Forgotten Scripts: Their Ongoing Discovery and Decipherment. 2d ed. New York, 1982.
- Pope, Maurice. The Story of Archaeological Decipherment from Egyptian Hieroglyphs to Linear B. New York, 1975.
Peter T. Daniels