1314191911 Encyclopædia Britannica, Volume 1 — AlphabetPeter Giles

ALPHABET (see also Writing). By the word alphabet, derived from the Greek names for the first two letters—alpha and beta—of the Greek alphabet, is meant a series of conventional symbols each indicating a single sound or combination of sounds. The ideal alphabet would indicate one sound by one symbol, and not more than one sound by the same symbol. Symbols for a combination of sounds are not necessary, though they may be convenient as abbreviations. In the writing of some languages, e.g. Sanskrit, such abbreviations are carried to an extreme; in most Greek MSS. also they are of very frequent occurrence. These contractions, however, may prove too great a strain upon the eyesight or the memory, and thus become a hindrance instead of a help. This was apparently the case in Greek, for though the early printers cast types for all the contractions of the Greek MSS. these have now with one consent been given up. A consonant like x can only be regarded as an abbreviation; it expresses nothing that cannot as well be expressed by ks or gz, both of which combinations in different situations it may represent (see X). No alphabet corresponds exactly to the ideal which we have postulated, nor if it did, would it continue long so to do, as the sounds of most languages are continually changing. Hence in the case of dead languages or past forms of living languages, it is often very difficult to define with precision what the sounds of the past epoch were. The study of the history of English pronunciation occupied the late Dr A. J. Ellis for a large part of his life, and the results fill five large volumes. The sounds which are most difficult to define exactly are the vowels; a great variety may be indicated by the same symbol. In the New English Dictionary no fewer than thirteen different nuances of vowel sound are distinguished under the symbol A alone. In English, moreover, the vowel sounds tend to become diphthongs, so that the symbol for the simple sound tends to become the symbol for that combination which we call a diphthong. Thus the long ī in ride, wine, &c., has become the diphthong ai, and the name of the symbol I is itself so pronounced. In familiar, if vulgar, dialects, A tends in the same direction. In the “cockney” dialect, really the dialect of Essex but now no less familiar in Cambridge and Middlesex, the ai sound of ī is represented by oi as in toime, “time,” while ā has become ai in Kate, pane, &c. In all southern English ō becomes more rounded while it is being pronounced, so that it ends with a slight u sound. In the vulgar dialect already mentioned, the sound begins as a more open sound than in the cultivated pronunciation, so that no is really pronounced as naou. It is clear, therefore, that the best alphabet would not long indicate very precisely the sounds which it was intended to represent. See Phonetics.

But the history of the alphabet shows that at no time has it represented any European language with much precision, because it was an importation adapted in a somewhat rough and ready fashion to represent sounds different from those which it represented outside Europe. Wherever the alphabet may have originated, there seems no doubt that its first importation in a form closely resembling that with which we are familiar in modern times was from the Phoenicians to the Greeks. The Phoenicians were certainly using it with freedom in the 9th century B.C.; with so much freedom, indeed, that they must have been in possession of it for a considerable time before we can trace it. With the materials available up to August 1910 it would be idle here to attempt to trace its earlier history. Great discoveries in Cappadocia, Assyria and Egypt were then only at their beginning, and any statement was liable to be quickly disproved by the appearance of new evidence. The prevalent theory, universally accepted till a few years ago, was that of Vicomte Emmanuel de Rougé, first propounded to the Académie des Inscriptions in 1859, but unnoticed by the world at large till republished, after de Rougé’s death, by his son in 1874. According to this view the alphabet was borrowed by the Phoenicians from the cursive (hieratic) form of Egyptian hieroglyphics. The resemblances between some Egyptian symbols and some symbols of the Phoenician alphabet are striking; in other cases the differences are no less remarkable. As a matter of fact the Egyptians might have passed about thirty-five centuries B.C. from the picture writing of hieroglyphs to genuine alphabetic signs.[1] They did not, however, profit by their discovery, because, amongst the Egyptians, writing was clearly a mystery in both senses—only possible at that period for masters in the craft, and also something, like the writing of medical prescriptions at the present day in Latin, which was not to be made too easily intelligible to the common people. At all periods, moreover, hieroglyphic writing was a branch of decorative art, and it may have been that the ancient Egyptian, like the modern Turk, resented too much lucidity, and liked his literary compositions to be veiled in a certain obscurity. The alphabet devised by the Egyptians consisted of twenty-four letters. Egyptologists are at variance on the question whether this alphabet was the original, or had any influence upon the development of the Phoenician alphabet. “With the papyrus paper,” says Professor Breasted,[2] “the hand customarily written upon it in Egypt now made its way into Phoenicia, where before the 10th century B.C. it developed into an alphabet of consonants, which was quickly transmitted to the Ionian Greeks and thence to Europe.” On the other hand, Professor Spiegelberg,[3] writing soon after Professor Breasted, says that investigation has not as yet furnished proof that the Phoenician alphabet is of Egyptian origin, though he admits that in some respects the development of the two alphabets, both without vowel signs, is curiously parallel.

The most recent view is that of Dr A. J. Evans, who argues ingeniously that the alphabet was taken over from Crete by the “Cherethites and Pelethites” or Philistines, who established for themselves settlements on the coast of Palestine.[4] From them it passed to the Phoenicians, who were their near neighbours, if not their kinsfolk. Symbols like the letters of the alphabet have been found in European soil painted upon pebbles belonging to a stratum between the Palaeolithic and Neolithic age.[5] This was in France at Mas d’Azil on the left bank of the Arize. Elsewhere several series of such symbols resembling inscriptions have been found scratched on bones of the same period.[6] For the history of writing these may be important, but for the history of the alphabet, as we know it, they are not in question. The alphabet may have originated as Dr Evans thinks, but at present the proof is not conclusive. The Greek names of the letters, their forms, and the order of the symbols show that the Greek alphabet as we know it must have been imported by or from a Semitic people, and there is no evidence to contradict ancient tradition that this people was the Phoenicians. The view propounded by Deecke[7] in 1877, that the Phoenician alphabet had developed out of the late Assyrian cuneiform, never met with much acceptance and has really no evidence in its favour.

The earliest alphabetic document which can be dated with comparative certainty is the famous Moabite stone, which was discovered in 1868, and after a controversy between rival claimants which led to its being broken in pieces by the Arabs, ultimately reached the Louvre, where in a restored form it remains. The long inscription upon it celebrates the achievements of Mesha, king of Moab, who had been a tributary of Ahab, king of Israel, and rebelled after his death (1 Kings iii. 4, 5). Though the chronology of the period is somewhat uncertain, the date must be in the first half of the 9th century B.C. It is to be remembered, however, that important as this monument is for the development of the alphabet, and because it can be dated with tolerable accuracy, the dialect and alphabet of Moab are not in themselves proof for the Phoenician forms which influenced the peoples of the Aegean, and through them Western Europe. The fragment of a bronze bowl discovered in Cyprus in 1876, which bears round its edge an inscription dedicating it to Baal-Lebanon as a gift from a servant of Hiram, king of the Sidonians, is probably the oldest Phoenician document which we possess. This bowl, though perhaps a little earlier than the Moabite stone, in all probability is not more than a century older, while some authorities think it is even later. The earliest alphabet consisted of twenty-two letters, and bears a very close resemblance to the earliest Greek alphabet from Α to Τ. The symbols in the Greek alphabet from Υ to Ω, or in the numerical alphabet to Ϡ, are not found in the Phoenician alphabet.

As already mentioned, the twenty-two symbols of the Phoenician alphabet indicate consonantal sounds only. Greek did not possess so many consonants. The Phoenician alphabet possessed many more aspirates than were required in Greek, which tended more and more to drop all its aspirates. Before history begins it had also lost, except sporadically in out-of-the-way dialects, the semi-vowel i (approximately English y.) It therefore made the aspirates A, E, O and the semi-vowel I into vowels, and apparently converted the semi-vowel 𐤅 = w into the vowel Y = u, which it placed at the end of the alphabet and substituted for it as the sixth symbol of the alphabet the letter F with the old value of w. The superfluous sibilants were also adapted in various ways (see below).

The discovery of a large number of very archaic inscriptions in the island of Thera, which was made by Freiherr Hiller von Gärtringen in 1896, has shown that the earliest Greek alphabet was even more like the Phoenician than had been heretofore believed. The symbol for β in Relationship of Greek to Phoenician.Thera (Β) is nearer than any previously known to the Semitic letter (𐤁) though, as not infrequently happens in the transference of a symbol from one people to another, its position is inverted—a fate which in this alphabet has befallen also λ (Semitic 𐤋, Thera Λ), and possibly σ (Semitic 𐤑, Thera Ϻ). The era of excavation initiated by Dr Schliemann on the grand scale has increased our knowledge of Greek inscriptions beyond anything that was earlier dreamt of. Besides the excavations of Athens, Delos, Epidaurus and Delphi, the results of which are most important for the 5th century B.C. and later. the exploration of the sites of Olympia, of the Heraeum near Argos, of Naucratis in Egypt, and of various Cretan towns (above all the ancient Gortyn), has revolutionized our knowledge of the archaic alphabets of Greece. We can now see how long and laborious was the process by which the Greeks attained to uniformity in writing and in numeration. In no field, perhaps, was the centrifugal tendency of the Greeks more persistent than in such matters. In numeration, indeed, uniformity was not attained till at least the 2nd century of the Christian era. The differentiation of the local alphabets is found from the very beginning of our records. Unfortunately, as yet no record is preserved which can with any probability be dated earlier than the 7th century B.C., and the Phoenician influence had by then nearly ceased. How long this influence lasted we cannot tell. If in Crete a system of writing of an entirely different nature had been developed seven or eight centuries before, there must have been some very important reason for the entire abandonment of the old method and the adoption of a new. In Crete, at least, the excavations show that the old civilization must have ended in a social and political cataclysm. The magnificent palace of Minos—there seems no reason to withhold from it the name of the great prince whom Thucydides recognized as the first to hold the empire of the sea—perished by the flames, and it evidently had been plundered beforehand of everything that a conqueror would regard as valuable. The only force in Greek history which we know that could have produced this change was that of the Dorian conquest. As everywhere in the Peloponnese, except at Argos, there seems to have been a sudden break with the earlier civilization, which can have been occasioned only by the semi-barbarous Dorian tribes, so the same result seems to have followed from the same cause in Thera. The Dorians apparently were without an alphabet, and consequently when Phoenician traders and pirates occupied the place left vacant by the downfall of Minos’s empire, the people of the island, and of the sea coasts generally, adopted from them the Phoenician alphabet.[8] The Greeks who migrated to Cyprus, possibly as the result of the Dorian invasion, adopted a syllabary, not an alphabet (see Plate; also Writing.) That the alphabet was borrowed and adapted independently by different places not widely separated, and that the earliest Greek alphabets did not spread from one or a few centres in Greek lands, seem clear (a) from the different Greek sounds for which the Phoenician symbols were utilized; (b) from the different symbols which were employed to represent sounds which the Phoenicians did not possess, and for which, therefore, they had no symbols. The Phoenician alphabet was an alphabet of consonants only, but all Greek alphabets as yet known agree in employing A, E, I, O, Y as vowels. On the other hand, a table of Greek alphabets[9] will show how widely divergent the symbols for the same sound were. Except for a single Attic inscription (see Plate), the alphabets of Thera and of Corinth are the oldest Greek alphabets which we possess. Yet at Corinth alongside 𐌄 𐌄, which is found for the so-called spurious diphthong ει (i.e. the Attic ει, which does not represent an Indo-European ει, but arises by contraction, as in φιλεῖτε, or through the lengthening of the vowel sound as the result of the loss of a consonant, as in εἰρημένος for ϝεϝρημένος) the short ε sound is represented by B; ι is found at Corinth in its oldest form , and also as , while in Thera it is . In Thera the w sound of digamma (Ϝ) was entirely lost, and therefore is not represented. Both Thera and Corinth employ in the earliest inscriptions for ζ, not ξ, though in both alphabets the ordinary use as ξ is adopted, no doubt through the influence of trade with other states. On the other hand. at Cleonae. which is distant not more than 8 or 9 m. from Corinth, an ancient inscription written βουστροφηδόν has recently been discovered, which shows that though Cleonae for B wrote , like the Corinthian , and, as at Corinth, wrote 𐌁 for a vowel sound, the vowel thus represented was not short and long e (ε and η) as at Corinth, but η only, as in 𐌗𐌓𐌁𐌌𐌀, 𐌌𐌁 (χρῆμα μή). Here 𐌄 represents ε, and the spurious diphthong is represented by ει, as in 𐌄𐌉𐌌𐌄𐌍 (εἶμεν, Doric infinitive = εἶναι), a form which shows that ι has at Cleonae the more modern form 𐌉 as distinguished from the Corinthian .[10]

Regarding three other questions controversy still rages. These are: (a) how Greek utilized the four sibilants (Shin, Samech, Zain and Zade), which it rook over from the Phoenician; (b) what was the history of development in the symbols for φ, χ, ψ, ω (the history of ξ belongs to both heads); (c) the history of the symbol for the digamma Ϝ.

In the Phoenician alphabet Zain was the seventh letter, occupying the same position and having the same form approximately (𐤆) as the early Greek Ζ, while in pronunciation it was a voiced s-sound; Samech (𐤎‎) followed the symbol for n and was the ordinary s-sound, though, as we have seen, it is in different Greek slates at the earliest period ζ Greek use of Phoenician sibilants.as well as ξ; after the symbol for p came Ẓade (𐤑‎), which was a strong palatal s, though in name it corresponds to the Greek ζῆτα; while lastly Shin (𐤔‎) follows the symbol for r, and was an sh-sound. The Greek name for the sibilant (σίγμα) may simply mean the hissing letter and be a derivative from σίζω; many authorities, however, hold that it is a corruption of the Phoenician Samech. Unfortunately, it is not clear how many sibilants ere distinguished in Greek pronunciation, nor over what areas a particular pronunciation extended. There is, however, considerable evidence of the view that Greek σσ representing the sound arising from κy, χy, τy, θy was pronounced as sh (), while ζ representing gy, dy was pronounced in some districts zh ().[11]

On an inscription of Halicarnassus, a town which stood in ancient Carian territory, the sound of σσ in Ἁλικαρνασσέων is represented by Ͳ, as it is also in the Carian name Panyassis (ΠανυάͲιος, genitive), though the ordinary ΣΣ is also found in the same inscription. The same variation occurs at the neighbouring Teos and at Ephesus, while the coins of Mesembria in Thrace show regularly ΜΕͲΑ and ΜΕͲΑΜΒΡΙΑΝΩΝ, where Ͳ represents the sound which resulted from the fusion of θy, and which appears in Homer as σσ in μέσσος, while in later Greek it becomes μέσος.[12] This symbol Ͳ is in all probability the early form of the letter which was known to the Greeks as San (σάν) and in modern times as Sampi, and which is utilized as the numeral for 900 in the shape ϡ. According to Herodotus (i. 139), San was only the Dorian name for the letter which the Ionians called Sigma. This would bring it into connexion with the Phoenician 𐤔 (Shin), which, turned through a right angle, is possibly the Greek Σ, though some forms of Ẓade on old Hebrew coins and gems () equally resemble the Greek letter. From other forms of Sade, however, the other early form of σ, viz. , is probably derived. The confusion is thus extreme: the name Ẓade assimilated in Greek to the names ἦτα and θῆτα becomes ζῆτα, though the form is that of Zain; the name of Samech is possibly the origin of Sigma, while the form of Samech is that of Ξ which has not taken over a Phoenician name. It is probable that the form is an abbreviation in writing from right to left of the earlier , and of the four stroke . That the confusion of the sibilants was not confined to the Greeks only, but that pronunciation varied within a small area even among the Semitic stock, is shown by the difficulty which the Ephraimites found in pronouncing “shibboleth” (Judges xii. 6).

For the history of the additional symbols which are not Phoenician, we must begin with Y. There is no Greek alphabet in which the symbol is not represented. But the Phoenician form corresponding to it is the consonant w, and occupies the position of the Greek digamma as sixth in the series. Whence did the Greeks obtain the digamma? The History of the digamma.point is not clear, but probably the Greeks acted here as they did in the case of the vowel i and the consonant y, adopting the consonant symbol for the vowel sound. As, however, except in Cyprus, Pamphylia and Argos, the only y sound which survived in Greek—the glide between i and another vowel as in δυά=diya—is never represented, there was no occasion to use the Phoenician Jod in a double function. With Vau it was different; the u-sound existed in some form in all dialects, the w-sound survived in many far into historical times. The Phoenician symbol having been adopted for the vowel sound, whence came the new symbol Ϝ or for the digamma? Hitherto there have been two views. Most authorities have held that the new form was derived from E by dropping the lowermost crossbar; some have held that it developed out of the old Vau, a view which is not impossible in itself and has the similar development in Aramaic (Tema) in its favour. But as Dr Evans has found a form like the digamma among his most recent types of symbols, and as we have no intermediate forms which will prove the development of Ϝ from , though the form found at Oaxos in Crete, viz. shows a form sufficiently unlike Ϝ, it is necessary to suspend judgment.

The Greek aspirates were not the sounds which we represent by ph, th, ch (Scotch), but corresponded rather to the sound of the final consonants in such words as lip, bit, lick, the breath being audible after the formation of the consonant. It is not clear that Greek took over with this value, for in one Theran inscription ϴ𐌇 are found combined as equivalent Greek
aspirates,
&c.
to Τ–Η, while the regular representation of φ and χ is Π  and Κ , or Ϙ (koppa) respectively. In the great Gortyn inscription from Crete and occasionally in Thera, Π (in Crete in the form C) and Κ are used alone for φ and χ, just as conversely even in the 5th century the name of Themistocles has been found upon an ostrakon spelt Θεμισθοκλῆς. Such confusions show that even to Greek ears the distinction between the sounds was very small. To have recorded it in writing at all shows considerable progress in the observation of sounds. Such progress is more easily indicated by changes in the symbols among a people whose acquaintance with the art is not of long standing nor very familiar. English, though possessing sounds comparable to the Greek θ, φ, χ, has never made any attempt to represent them in writing. On the other hand, no doubt Athens in 403 B.C. officially adopted the Ionic alphabet and gave up the old Attic alphabet. The political situation in Athens, however, at this time was as exceptional as the French Revolution, and offered an opportunity not likely to recur for the adoption of a system in widely extended use which private individuals had been employing for a long time.

The history of the symbols φ and χ is altogether unknown. The very numerous theories on the subject have generally been founded on a principle which itself is in need of proof, viz. that these symbols must have arisen by differentiation from others already existing in the alphabet. The explanation is possible, but it is not easy to see why, for example, the symbol 𐌒 or 𐌘=Koppa, the Latin Q, should have been utilized for a sound so different as p-h; nor, again, why the symbol for θ (🜨) by losing its cross stroke should become φ, seeing that the sounds of θ and φ outside Aeolic (a dialect which is not here in question) are never confused. On the other hand, if we remember the large number of symbols belonging to the prehistoric script, it will seem at least as easy to believe that the persons who, by adding new letters to the Phoenician alphabet, attempted to bring the symbols more into accordance with the sounds of the Greek language, may have borrowed from this older script. It is now generally admitted that the improvements of the alphabet were made by traders in the interests of commerce, and that these improvements began from the great Greek emporia of Asia Minor, above all from Miletus. Symbols exactly like φ, χ, and ψ (𐊸, 𐊴, 𐊵) are found in the Carian alphabet, and transliterated by Professor Sayce[13] as v (and ü), h and kh respectively. If the Carian alphabet goes back to the prehistoric script, why should not Miletus have borrowed them from it? We have already seen that, in the earliest alphabets of Thera and Corinth, the ordinary symbol for ξ in the Ionic alphabet was used for ζ. This usage brought in its train another—the use of 𐊜, not for ψ as in Ionic, but for ξ in the name Άλεξαγόρα, and similarly in Melos, Πραξικύδεος.[14] This experiment, for it was no more, belongs apparently to the latter part of the 6th century, and was soon given up. As the Ionians kept the form , which the people of Thera used for ζ, in the same position in their alphabet as Samech occupied in the Phoenician alphabet, there can be no doubt as to its origin. The symbol which the Chalcidian Greeks used in the 6th century B.C. for ξ may be derived, according to the most widely accepted theory, from a primitive form of Samech 𐌎, which is recorded only in the abecedaria of the Chalcidian colonies in Italy. In this case the borrowing of the Greek alphabet must long precede any Phoenician record we possess. But it is not probable that the Ionic and Phoenician developed independently from the closed form. Kretschmer, however, in several publications[15] takes a different view. He thinks that the guttural element in ξ was a spirant, and therefore different from χ, which is an aspirate. He points out that in Naxos, in a 6th-century inscription,[16] ξ in Ναξίου, ἕξοχος and Φράξου is represented by , the first element in which he regards as a form of 𐌇h. As χ is found in the same inscription (in the form X), the guttural element must have been different, else ξ would have been spelt X. Attica and most of the Cyclades kept 𐌗 for the guttural element in ξ (written or ) and for χ as well. On the west of the Aegean a new symbol 𐌙 was invented for the aspirate value, and this spread over the mainland and was carried by emigrants to Rhodes, Sicily and Italy. The sign 𐌗 was kept in the western group for the guttural spirant in ξ, which was written X; but, as this spirant occurred nowhere else, the combination was often abbreviated, and X was used for X precisely as in the Italic alphabets we shall find that Ff develops out of a combination FH.

The development of symbols for the long vowels η and ω was also the work of the Ionians. The h-sound ceased at a very early period to exist in Ionic, and by 800 B.C. was ignored in writing. The symbol 𐌇 or H was then employed for the long open ē-sound, a use suggested by the name of the letter, which, by the loss of the aspirate, had passed from Heta to Eta. About the same period, and probably as a sequel to this change, the Greeks of Miletus developed Ω for the long open ō-sound, a form which in all probability is differentiated out of O. Centuries passed, however, before this symbol was generally adopted, Athens using only O for ο, ω and ου, the spurious diphthong, until the adoption of the whole Ionic alphabet in 403 B.C.[17]

The discoveries of the last quarter of the 19th century carried back our knowledge of the Latin alphabet by at least two centuries, although the monuments of an early age which have been discovered are only three. (a) In 1880 was discovered between the Quirinal and Viminal hills a little earthenware pot of a curious shape, being as it were, Latin alphabet.three vessels radiating from a centre, each with a separate mouth at the top.[18] Round the sides of the triangle formed by the three vessels and under the mouths runs an inscription of considerable length. The use for which the pot was intended and the purport of the inscription have been much disputed, there being at least as many interpretations as there are words in the inscription. The date is probably the early part of the 4th century B.C. Though found in Rome, the vessel is small enough to be easily portable, and might therefore have been brought from elsewhere in Italy.The Dvenos inscription. It is equally possible that the potter who inscribed the words upon it was not a native of Rome. One or two points in the inscription make it doubtful whether the Latin upon it is really the Latin of Rome. It is generally known as the Dvenos inscription, from the name of the maker who wrote on the vessel from right to left the inscription, part of which is DVENOS MED FECED (=fecit). (b) The second of these early records is the inscription on a gold fibula found at Praeneste and published in 1887. The inscription runs from right to left, and is in letters which show more clearly than ever that the Roman alphabet is borrowed from the alphabets of the Chalcidian Greek colonies in Italy. Its date cannot be later than the 5th and is possibly as early as the 6th century B.C. The words are MANIOS MED FHEFHAKED NVMASIOI, “Manius made me for Numasius.”The Praeneste fibula. The symbol for M has still five strokes, s has the angular form 𐌔, 𐌔. The inscription is earlier than the Latin change of s between vowels into r, for Numasioi is the dative of the older form which corresponds to the later Numerius. The verb form is remarkable. In the Dvenos inscription the perfect of facio is feced; here it is a reduplicated form with the same vowel as the present. The spelling also is interesting. The symbol K is still in ordinary use, and not merely used for abbreviations as in the classical age. But most remarkable is the representation of Latin F by FH. The reason for this is clear. The value of F in the Greek alphabet is w and not f as in Latin. Greek had no sound corresponding to Latin F, consequently an attempt is made by combining F and H to indicate the difference of sound. Etruscan uses FH in the same way. As Latin, however, made the symbol V indicate not only the vowel sound u, but also the consonant sound v (i.e. English w), the sign for the digamma F was left unemployed, and as FH was a cumbrous method of representing a sound which did not exist in Greek, the second element came to be left out in writing. Thus F came to be the representative of the unvoiced labiodental spirant instead of that for the bilabial voiced spirant. Whether the form fefaked was ever good Latin in Rome may be doubted, for the Romans, in spite of the few miles that separate Praeneste from Rome, were inclined to sneer at the pronunciation and idiom of the Praenestines (cf. Plautus, Trin. 609, Truc. 691; Quintilian i. 5, 56) (c) The last, and in some respects the most important, of these records was found in 1899 under an ancient pavement in the Comitium at the north-west corner of the Roman Forum. Forum inscriptions. It is engraved upon the four sides and one bevelled edge of a pillar, the top of which has been broken off. As the writing is βουστροφηδόν, beginning at the bottom of the pillar and running upwards and down again, no single line of the inscription is complete. Probably more than half the pillar is lost, so that it is not possible to make out the sense with certainty. The inscription is probably not older than that on the fibula from Praeneste, but has the additional interest of being undoubtedly couched in the Latin of Rome. The surviving portion of the inscription contains examples of all the letters of the early alphabet, though the forms of F and B are fragmentary and doubtful. As in the Praenestine inscription, the alphabet is still the western (Chalcidian) alphabet. K is still in use as an ordinary consonant, and not limited to a symbol for abbreviations as in the classical period. The rounded form of γ is found with the value of G in RECEI, which is probably the dative of rex. H has still the closed form 𐌇, M has the five-stroke form, S is the three-stroke 𐌔, tending to become rounded. R appears in the Greek form without a tail, and V and Y are both found for the same sound. The manner of writing up and down instead of backwards and forwards across the stone is obviously appropriate to a surface which is of considerable length, but comparatively narrow, a connected sense being thus much easier to observe than in writing across a narrow surface where, as in the gravestones of Melos, three lines are required for a single word. The form of the monument corresponds to that which we are told was given to the revolving wooden pillars on which the laws of Solon were painted. That the writing of Solon’s laws, which was βουστροφηδόν, was also vertical is rendered probable by the phrase ὁ κάτωθεν νόμος in Demosthenes’ speech Against Aristocrates, § 28, for which Harpocration is unable to supply a satisfactory explanation.

The differentiation of the Roman alphabet from the Greek is brought about (a) by utilizing the digamma for the unvoiced labiodental spirant F; (b) by dropping out the aspirates θ, φ, χ (𐊜 in the Chalcidian alphabet, whence the Roman is derived) from the alphabet proper and employing them only as numerals, θ () being gradually modified till it was identified with C as though the initial of centum, 100. Differenti-ation of Roman from Greek alphabet. Similarly 🜕 became in time identified with M as though the initial of mille, 1000, and the side strokes of χ in the above form were flattened out till it became , and ultimately L, 50. (c) After 350 B.C., at latest, there was in Latin no sound corresponding to Z, which was therefore dropped. In the Chalcidian alphabet the symbol for x was placed after the symbols common to all Greek alphabets, a position which X retains in the Latin (and also in the Faliscan) alphabet. K in time passed out of use except as an abbreviation, its place being taken by C, which, as we have seen, is in the earliest inscription still g. Three points here require explanation: (1) Why K fell into disuse: (2) Why C took the place of K; (3) why the new symbol G was put in the place of the lost Z. It is clear that C must have become an equivalent of K before the latter fell out of use. There is some evidence which seems to point to a pronunciation of the voiced mutes which, like the South German pronunciation of g, d, b, but slightly differentiated them from the unvoiced mutes, so that confusion might easily arise. The Etruscans, who were separated from the Romans only by the Tiber, gradually lost the voiced mutes. But another cause was perhaps more potent. C and IC, as k was frequently written, would easily be confused in writing, and Professor Hempl (Transactions of the Chalcidian Philological Association for 1899, pp. 24 ff.) shows that the Chalcidian form of ζ𐠳 developed into shapes which might have partaken of the confusion. Owing to this confusion, the new symbol G, differentiated from C, took the place of the useless 𐠳. In abbreviations, however, C remained as before in the value of G, as in the names Gaius and Gnaeus. Y and Z were added in the last century of the republic for use in transliterating Greek words containing υ and ζ.[19]

The dialect which was most closely akin to Latin was Faliscan. The men of Falerii, however, regularly took the side of the Etruscans in wars with Rome, and it is clear that the civilization of the old Falerii, destroyed for its rebellion in 241 B.C., was Etruscan and not Roman in character. Peculiar to this alphabet is the form for f𐋇. Much more important than the scanty remains of Faliscan is the Oscan alphabet. The history of this alphabet is different from that of Rome. It is certain from the symbols which they develop or drop that the people of Campania and Samnium borrowed their alphabet from the Etruscans, who held dominion in Campania from the 8th to the 5th century B.C. Previous to the Punic wars Campania had reached a higher stage of civilization than Rome. Unfortunately, the remains of that civilization are very scanty, and our knowledge of the official alphabet outside Capua, and at a later period Pompeii, is practically confined to two important inscriptions, the tabula Agnonensis, now in the British Museum, and the Cippus Abellanus, which is now kept in the Episcopal Seminary at Nola. Of Etruscan origin also is the Umbrian alphabet, represented first and foremost in the bronze tablets from Gubbio (the ancient Iguvium). The Etruscan alphabet, like the Latin, was of Chalcidian origin. That it was borrowed at an early date is shown by the fact that most of its numerous inscriptions run from right to left, though some are written βουστροφηδόν. That it took over the whole Chalcidian alphabet is rendered probable by the survival in Umbrian and Oscan, its daughter alphabets, of forms which are not found in Etruscan itself. This mysterious language, despite the existence of more than 6000 inscriptions, and the publication in 1892 of a book written in the language and handed down to us by the accident of its use to pack an Egyptian mummy, remains as obscure as ever, but apparently it underwent very great phonetic changes at an early period, so that the voiced mutes B, D, G disappeared. Of the existence of the vowel O there is no evidence. If it ever existed in Etruscan, it had been lost before the Oscans and Umbrians borrowed their alphabets. On the other hand, both of their alphabets preserve B and Umbrian G in the form >. Etruscan also retained this symbol in the form Ɔ, and utilized it exactly as Latin did to replace 𐌊. Oscan, in order to represent D, introduced later a form 𐊯, thus creating confusion between the symbols for d and for r. This form was adopted for d because D had already been borrowed from Etruscan as the symbol for r, although 𐌓 is also found on Etruscan inscriptions. For the Greek digamma Etruscan used both and 𐌅, but the former only was borrowed by the other languages. Etruscan, like Latin, used 𐌅𐌇 (from right to left) to represent the sound of Latin F, but, unlike Latin, adopted 𐌇 not 𐌅 as the single symbol. This form it then wrote as two lozenges , whence developed a later sign, 𐌚, which is used also in Umbrian and Oscan. As the old digamma was kept, this new sign was placed after those borrowed from the Chalcidian alphabet. Similarly it used and 𐠳 for the Chalcidian ζ; Umbrian borrowed the first, Oscan the second form. The form for h was still closed 𐌇, which Etruscan passed on to Oscan, while Umbrian modified it to 🜕. The form for m has five strokes; from a later form 𐋅 the Oscan form was borrowed. Of the two sibilants, 𐌑 and 𐌔 or S, Oscan adopted only 𐌔, Umbrian both 𐌑 and the rounded form S. 𐌒 is found on Etruscan inscriptions, but not in the alphabet series preserved; neither Umbrian nor Oscan has this form. T appears in Etruscan as У, , and X; of these Umbrian borrows the first two, while Oscan has a form T like Latin. Etruscan took over the three Greek aspirates, θ, φ, χ in their Chalcidian forms; θ survives in Umbrian as Ꙩ, the others naturally disappear. Both Umbrian and Oscan devised two new symbols. Umbrian took over from Etruscan perhaps the sign 𐌛, but gave it the new value of a spirant which developed out of an earlier d-sound, but which is written in the Latin alphabet with rs. The second Umbrian symbol was 𐌜, which was the representative of an s-sound developed by palatalizing an earlier k. In Oscan, which had an o-sound, but no symbol for it, a new sign was invented by placing a dot between the legs of the symbol for u𐌞. This, however, is found only in the best-written documents, and on some materials the dot cannot be distinguished. The symbol 𐌝 was invented for the open ĭ-sound and close ē-sound.[20] At a much later epoch it was introduced into the Latin alphabet by the emperor Claudius to represent y, and the sound which was written as i or u in maximus, maxumus, &c.

Besides the Italic alphabets already mentioned, which are all derived from the alphabet of the Chalcidian Greek colonists in Italy, there were at least four other alphabets in use in different parts of Italy: (1) the Messapian of the south-east part of the peninsula, in which the inscriptions of the Illyrian dialect in use there were written, an alphabet which, according to Pauli (Alt-italische Forschungen, iii. chap. ii.) was borrowed from the Locrian alphabet; (2) the Sabellic alphabet, derived from that of Corinth and Corcyra, and found in a few inscriptions of eastern-central Italy; (3) the alphabet of the Veneti of north-east Italy derived from the Elean: (4) the alphabet of Sondrio (between Lakes Como and Garda), which Pauli, on the insufficient ground that it possesses no symbols corresponding to φ and χ, derives from a source at the same stage of development as the oldest alphabets of Thera, Melos and Crete.

From the fact that upon the Galassi vase (unearthed at Cervetri, but probably a product of Caere), which is now in the Gregorian Museum of the Vatican, a syllabary is found along with one of the most archaic Greek alphabets, and that a similar combination was found upon the wall of a tomb at Colle, near Siena, it has been argued that syllabic preceded alphabetic writing in Italy. But a syllabary where each syllable is made by the combinations of a symbol for a consonant with that for a vowel can furnish no proof of the existence of a syllabary in the strict sense, where each symbol represents a syllable; it is rather evidence against the existence of such writing. The syllabary upon the Galassi vase indicates in all probability that the vase, which resembles an ink-bottle, belonged to a child, for whose edification the syllables pa, pi, pe, pu and the rest were intended. The evidence adduced from the Latin grammarians, and from abbreviations on Latin inscriptions like lubs for lubens, is not sufficient to establish the theory.

It has been argued that the runes of the Teutonic peoples have been derived from a form of the Etruscan alphabet, inscriptions in which are spread over a great part of northern Italy, but of which the most characteristic are found in the neighbourhood of Lugano, and in Tirol near Innsbruck, Botzen and Trent. The Danish scholar L. F. A. Teutonic runes.Wimmer, in his great work Die Runenschrift (Berlin, 1887), contends that the resemblance, though striking, is superficial. Wimmer’s own view is that the runes were developed from the Latin alphabet in use at the end of the 2nd century A.D. Wimmer supports his thesis with great learning and ingenuity, and when allowance is made for the fact that a script to be written upon wood, as the runes were, of necessity avoids horizontal lines which run along the fibres of the wood, and would therefore be indistinct, most of the runic signs thus receive a plausible explanation. The strongest argument for the derivation from the Latin alphabet is undoubtedly the value of f attaching to ; for, as we have seen, the Greek value of this symbol is w, and its value as f arises only by abbreviation from FH. On the other hand, several of Wimmer’s equations are undoubtedly forced. Even if we grant that the Latin symbols were inverted or set at an angle (a proceeding which is paralleled by the treatment of the Phoenician signs in Greek hands), so that represents Latin 𐌖, , Latin 𐌄, Latin 𐌋, and Latin 𐌃; while the symbol for the voiced spirant ð is doubled, , 𐌃𐌃, it is difficult to believe that the symbol for the spirant g, viz. , represents a Latin 𐌊 (which was of rare occurrence), or again , >< a Latin N, or that the symbol for ng, , represents < = c doubled. Moreover, the date of the borrowing seems too late. The runes are found in all Teutonic countries,and the Romans were in close contact with the Germans on the Rhine before the beginning of the Christian era. We hear of correspondence between the Romans and German chieftains in the early days of the empire. It is strange, therefore, if the Roman alphabet, which formed the model for the runes, was that of two whole centuries later, and even then the formal alphabet of inscriptions. By that time the Teutons were likely to have more convenient materials than wood whereon to write, so that the adaptation of the forms would not have been necessary. That the Germans were familiar with some sort of marks on wood at a much earlier period is shown by Tacitus’s Germania, chap. x. There we are told that for purposes of divination certain signs were scratched on slips of wood from a fruit-bearing tree (including, no doubt, the beech; cp. book, German Buch, and Buchstabe, a letter of the alphabet); the slips were thrown down promiscuously on a white cloth. whence the expert picked them up at random and by them interpreted fate. In these slips we have the origin of the Norse kefli, the Scots kaivel, which were and are still used as lots. The fishermen of north-east Scotland, when they return after a successful haul, divide the spoil into as many shares as there are men in the boat, with one share more for the boat. Each man then procures a piece of wood or stone, on which he puts a private mark. These lots are put in a heap, and an outsider is called in who throws one lot or kaivel upon each heap of fish. Each fisherman then finds his kaivel, and the heap on which it lies is his. This system of “casting kaivels,” as it is called, is certainly of great antiquity. But its existence will not help to prove an early knowledge of reading or writing, for in order that everything may be fair, it is clear that the umpire should not be able to identify the lot as belonging to a particular individual. It has, however, been contended that a system of primitive runes existed whence some at least of the later runes were borrowed, and the ownership marks of the Lapps, who have no knowledge of reading and writing, have been regarded as borrowed from these early Teutonic runes.[21] Be this as it may, the resemblances between the runic and the Mediterranean alphabets are too great to admit of denial that it is from a Greek alphabet, whether directly or indirectly, that the runes are derived. That Wimmer postdates the introduction of the runic alphabet seems clear from the archaic forms and method of writing. It is very unlikely that a people borrowing an alphabet which was uniformly written from left to right should have used it in order to write from right to left, or βουστροφηδόν. Hence Hempl contends[22] that Wimmer’s view must be discarded, and that the runes were derived about 600 B.C. from a western Greek alphabet which closely resembled the Formello alphabet (one of the ancient Chalcidian abecedaria) and the Sabellic and North Etruscan alphabets. He thus fixes the date at the same period as Isaac Taylor had done in his Greeks and Goths and The Alphabet. Taylor, however, derived the runes from the alphabet of a Greek colony on the Black Sea. Hempl’s initiative was followed by Professor Gundermann of Giessen, who announced in November 1897[23] that he had discovered the source of the runic alphabet, the introduction of which he declares preceded the first of the phonetic changes known as the “Teutonic sound-shifting,” since = g is used for k, = χ for g, a Theta-like symbol for d, while zd is used for st. If this view (which is identical with Taylor’s) be true, we have a parallel in the Armenian alphabet, which is similarly used for a new value of the sounds. Hempl, on the other hand, contends that the sound-shifting had already taken place, and, arguing that several of the symbols have changed places (e.g. f and a, u and B b, because at this time b was a bilabial spirant and not a stop), ultimately obtains an order—a b d e f z kgw h i j ƕ p r s t u l m n ð o. As neither Gundermann nor Hempl has published the full evidence for his view, no definite conclusion at the moment is possible.

Inscribed Pebbles from Mas d’Azil.


Prehistoric Linear Script from Crete.

Gold Fibula from Praeneste, with Early Latin Inscription.
Right to left.
Vadstena Pendant, with Runic Alphabet;
about a.d. 600.

Cyprian Inscription (4th century b.c.) from Curium (British Museum Excavations, p. 64). Below are (1) the transliteration of the symbols; (2) the Greek words, both like the Cyprian reading from right to left.

Oldest Attic Inscription. From a Dipylon Vase probably of 8th century b.c. Right to left.

Inscription on Buddha Vase, perhaps 4th century b.c. In one of the earliest runic records which we possess, the pendant found at Vadstena in Sweden in 1774, and dating from about A.D. 600 (see Plate) the signs are divided up into three series of eight (the twenty fourth, , being omitted for want of room). Upon the basis of this division a Ogam writing.system of cryptography (in the sense that the symbols are unintelligible without knowledge of the runic alphabet) was developed, wherein the series and the position within the series of the letter indicated, were each represented by straight strokes, the strokes for the series being shorter than those for the runes or the series being represented by strokes to the left the runes by strokes to the right of a medial line.[24] From this system probably developed the ogam writing employed among the Celtic peoples of Britain and Ireland. The ogam inscriptions in Wales are frequently accompanied by Latin legends, and they date probably as far back as the 5th and 6th centuries A.D.. Hence the connexion between Celt and Teuton as regards writing must go back to a period preceding the Viking inroads of the 8th century. Taylor, however, conjectures (The Alphabet, ii. p. 227) that the ogams originated in Pembroke, “where there was a very ancient Teutonic settlement, possibly of Jutes, who as is indicated by the evidence of runic inscriptions found in Kent, seem to have been the only Teutonic people of southern Britain who were acquainted with the Gothic Futhoro.” However this may be, the ogam alphabet shows some knowledge of phonetics and some attempt to classify the sounds accordingly. The symbols are as follows[25]:—

Symbols of Ogam Alphabet.

The form of the ogam alphabet made it easy to carve hastily; hence in the old sagas, when a hero is killed we find the common formula “His grave was dug and his stone was raised, and his name was written in ogam.” According to Sophus Müller (Nordische Altertumskunde, ii. p. 264), it was from Britain that the use of runes upon gravestones was derived a use which, to judge from the number of bilingual inscriptions in Britain, the Celts derived from the Romans.

The special forms of the alphabet—the Cyrillic and the Glagolitic—which have been adopted by certain of the Slavonic peoples are both sprung directly from the Greek alphabet of the ninth century A.D., with the considerable additions rendered necessary by the greater variety of sounds in Slavonic as compared with Greek. Apart from other evidence, the use of B with the value of v, of H as well as I with the value of ī, of Φ with the value of f and X with that of the Scotch ch, would be proof that the alphabet was not borrowed till long after the Greek classical period, for not till later did β, φ, χ become spirants and η become identified with ι. The confusion of β with v necessitated the invention of a new symbol Б in the Cyrillic, Ⰱ in the Glagolithic for b, while new symbols were also required for the sounds or combinations of sounds ž (zh), dz, št (sht), č (ts), c (ch in church), š (sh), ŭ, ĭ, y (u without protrusion of the lips), ĕ (a close long e sound), for the combination of o, a and e with consonantal I (English y) and for the nasalized vowels , (nasalized o in pronunciation) and the combinations je̬ and ja̬ (English ye̬, ya̬). In all these matters Glagolitic differs very little from Cyrillic; it has only one symbol for ja (ya) and ĕ because both in this dialect were pronounced the same. It has also only one symbol for e and je (ye) for the phonetic reason that je always appears in the old ecclesiastical Slavonic, for which the alphabets were fashioned, at the beginning of words and after vowels: cp. the English use of the symbol u in unspoken and uniform. Glagolitic has a symbol for the palatalized g (Ӡ), but it is used only in the transcription of Greek words, γ having become y early between vowels in the popular dialects.

Such an elaborate alphabet could hardly have been invented except by a scholar, and tradition, probably rightly, has attached the credit for its invention to Cyril (originally Constantine), who along with his brother Methodius proceeded in A.D. 863 to Moravia from Constantinople, for the purpose of converting the Slavonic inhabitants to Christianity. The only question which concerns us here is which of the two alphabets was the earlier in use, and after much discussion authorities on Slavonic seem generally agreed that it was the Glagolitic (the name is derived from the Old Bulgarian, i.e old ecclesiastical Slavonic glagolŭ, “word”). According to Professor Leskien (Grammatik der altbulgarischen (altkirchenslavischen) Sprache, Heidelberg, 1909, p. xxi.), Cyril had probably made a prolonged and careful study of Slavonic before proceeding on his missionary journey, and probably in the first instance with a view to preaching the Gospel to the Slavs of Macedonia and Bulgaria, who were much nearer his own home, Thessalonica, than were those of Moravia. The Glagolitic was founded upon the ordinary Greek minuscule writing of the period, as was shown by Dr Isaac Taylor,[26] though the writing of the letters separately without abbreviations and an obvious attempt at artistic effect has gradually differentiated it from Greek writing. This alphabet, which is much more difficult to read than the bolder Cyrillic founded on the Greek uncial, survived for ordinary purposes in Croatia and in the islands of the Quarnero till the 17th century. The Servians and Russians apparently always used the Cyrillic, and its advantages gradually ousted the Glagolitic elsewhere, though the service book in the old ecclesiastical language which is used by the Roman Catholic Croats is in Glagolitic.[27]

While the Carian and Lycian were probably independent of the Greek in origin, so, too, at the opposite end of the Mediterranean was the Iberian. On the other hand, the Phrygian was very closely akin to the Greek in alphabet as well as in linguistic character. The Greek alphabet, with which it was most closely connected, was the Western, for the evidence is strongly in Phrygian.favour of the form 𐌙 having the value of χ, not ψ, in Phrygian, as it certainly has in the Etruscan inscription found on Lemnos in 1886, which is in an alphabet practically identical.

To a much later era belongs the Armenian alphabet, which, according to tradition, was revealed to Bishop Mesrob in a dream. The land might have been Grecized had it not, about A.D. 387, been divided between Persia and Byzantium, the greater part falling to the former, who discouraged Greek and favoured Syriac, which the Christian Armenians did not understand. Armenian.As those within Persian territory were forbidden to learn Greek, an Armenian Christian literature became a necessity. Taylor contends that the alphabet is Iranian in origin, but the circumstances justify Gardthausen and Hübschmann in claiming it for Greek. That some symbols are like Persian only shows that Mesrob was not able to rid himself of the influences under which he lived.

Of the later development of Phoenician amongst Phoenician people little need be said here. It can be traced in the graffiti of the mercenaries of Psammetichus at Abu Simbel in Upper Egypt, where Greeks, Carians and Phoenicians all cut their names upon the legs of the colossal statues. Still later it is found on the stele of Byblos, and on the sarcophagus of Eshmunazar (about 300 B.C.). The most numerous inscriptions come from the excavations in Carthage, the ancient colony of Sidon. One general feature characterizes them all, though they differ somewhat in detail. The symbols become longer and thinner; in fact, cease to be the script of monuments and become the script of a busy trading people. While the Phoenician alphabet was thus fertile in developing daughter alphabets in the West, the progress of writing was no less great in the East, first among the Semitic peoples, and through them among other peoples still more remote. The carrying of the alphabet to the Greeks by the Phoenicians at an early period affords no clue to the period when Semitic ingenuity constructed an alphabet out of a heterogeneous multitude of signs. If it be possible to assign to some of the monuments discovered in Arabia by Glaser a date not later than 1500 B.C., the origin of the alphabet and its dissemination are carried back to a much earlier period than had hitherto been supposed. Next in date amongst Semitic records of the Phoenician type to the bowl of Baal-Lebanon and the Moabite stone comes the Hebrew inscription found in the tunnel at the Pool of Siloam in 1881, which possibly dates back to the reign of Hezekiah (700 B.C.). The only other early records are seals with Hebrew inscriptions and potters’ marks upon clay vessels found in Lachish and other towns.[28] Like the Phoenician, these Hebrew signs are distinctly cursive in character but, as the legend on the coins of the Maccabees shows, became stereotyped for monumental use, while the Jews after the exile gradually adopted the Aramaic writing, whence the square Hebrew script is descended. The Samaritans alone stuck fast to the old Hebrew as part of their contention that they, and not the Jews, were the true Hebrews.

The oldest records in Aramaic were found at Sindjirli, in the north of Syria, in 1890, and date to about 800 B.C. At this epoch the Aramaic alphabet, or at any rate the alphabet of these records, is but little different from that shown upon the Moabite stone. Either two sounds are confused under one Aramaic.symbol, or these records represent a dialect which, like Hebrew and Assyrian, shows sh, z, and ç, where the ordinary Aramaic representation is , d and the Arabic tḥ, dh, and ṭh. The Aramaic became in time by far the most important of the northern Semitic alphabets. Even while long and important documents in Assyria were still written on clay tablets, in cuneiform, a docket or précis of the contents was made upon the side in Aramaic, which thus became the alphabet of cursive writing—a fact which explains its later development. Two changes, the inception of which is early, but the completion of which belongs to the Persian period, gave the impulse which Aramaic obeyed in all its later developments. These were (a) the opening of the heads of letters, so that beth , daleth , and resh become respectively , and , while 𐤏‎ becomes first U and ultimately V. In the later development the heads tend to be reduced in size, and finally to disappear. (b) As was natural in cursive writing, angles tend to become rounded, and the tails of the letters, which in, Phoenician are very long are curved round in the middle of words so as to join on to the succeeding letter. These characteristics were naturally emphasized in the Aramaic writing on papyrus which, beginning about 500 B.C., during the Persian sovereignty in Egypt, lasted on there till about 200 B.C. The gradual development of this script into the square Hebrew, and the more ornamental writing of Palmyra, may be traced in the works of Berger and Lidzbarski.[29]

In the land of the Nabataeans, a people of Arabian origin, the Aramaic alphabet was employed in a form which ultimately developed into the modern Arabic alphabet. Probably the earliest example of the Aramaic script in Arabia is the stele of Tema, in north-western Arabia, whereon is commemorated Arabic.the establishment of a worship of an Aramaic divinity. This monument, now in the Louvre, is not later than the 5th century B.C. In it the writing preserves its ancient form, the heads of the closed letters being only very slightly opened. The Nabataean inscriptions belong to a different epoch and a different style. They were first discovered by Charles Doughty in 1876–1877, who was followed between 1880 and 1884 by Huber and Euting, to whom a complete collection of these records is due. The records are fortunately dated, and belong to the period from 9 B.C. to A.D. 75. A further development can be traced in the graffiti with which pilgrims adorned the rocks of Mount Sinai down to the 2nd or 3rd century A.D. By the help of these inscriptions it is possible to trace the development of the modern Arabic where so many of the forms of the letters have become similar that diacritic points are essential to distinguish them, the original causes of confusion being the continuous development of cursive writing and the adoption of ligatures. Arabic writing, as known to us from documents of the early Mahommedan period, exhibits two principal types which are known respectively as the Cufic and the nashki. The former soon fell into disuse for ordinary purposes and was retained only for inscriptions, coins, &c.; the latter, which is more cursive in character, is the parent of the Arabic writing of the present day. Another form of the Aramaic alphabet, namely, the so-called Estrangela writing which was in use amongst the Christians of northern Syria, was carried by Nestorian missionaries into Central Asia and became the ancestor of a multitude of alphabets spreading through the Turkomans as far east as Manchuria.

There still remains a branch of the Semitic languages which, except for one or two of the languages belonging to it, was practically unknown till recent years. This is the South Semitic. Till the 19th century the earliest form known of this alphabet was the Ethiopian or Geez, in which Christian documents have been preserved from the early centuries of our era, South Semitic.and which is still used by the Abyssinians for liturgical purposes. The travels of two English naval officers, Wellsted and Cruttenden, through Yemen in southern Arabia in 1835, first called attention to the earlier monuments of Arabia. Fulgence Fresnel first established the importance of the inscriptions discovered by these Englishmen, and in 1843, when French consul at Jeddah, obtained through a French traveller, Francois Arnaud, information about other monuments of the same kind. In 1869 Joseph Halevy brought back nearly seven hundred inscriptions from Yemen, and this number has been increased from other quarters by several thousands, through the energy of several adventurous scholars, but chiefly by Eduard Glaser’s repeated journeys. The south Arabian inscriptions to which the terms Himyaritic and Sabaean are applied fall into two groups, the Sabaean proper and the Minaean. These are distinguished by differences in grammar and phraseology rather than in alphabet. The relative age of the Minaean and Sabaean monuments is a matter of dispute amongst Semitic scholars. Inscriptions in a kindred dialect were brought from El-Ola, in the north of the Hedjaz, by Professor Euting. To these D. H. Müller[30] gave the title of Lihyanite, from the name of the tribe (Liḥjān) to which they belong. Their date is supposed to be earlier than that of the Sabaean and Minaean. Minaean inscriptions were found at the same place, the Minaeans having had a trading station there. In 1893 J. Theodore Bent copied carefully at Yeha in Abyssinia a few inscriptions, some of which had been already copied in 1814 by the English traveller Salt. These inscriptions are of the greatest importance, because they demonstrate, according to D. H. Müller,[31] that the Sabaeans had colonized Abyssinia as early as 1000 B.C. Other inscriptions copied by Bent at Aksum belong to the 4th century A.D. and later. Two of the earliest are written in Sabaean characters, but in the language which is known as Geez or Ethiopic. From about A.D. 500 Ethiopic was written in an alphabet which according to Müller was no gradual growth but an ingenious device of a Greek scholar of this period at the court of Abyssinia. The Sabaean, like other Semitic, inscriptions are generally written from right to left, but a few are βουστροφηδόν; the Ethiopic is written from left to right, and makes a marked advance upon the ordinary Semitic manner of writing by indicating the vowels. This is done by varying the form of the consonant according to the vowel which follows it. The Ethiopic system is thus rather a syllabary than an alphabet. It is noticeable that the changes thus established were made upon the basis of the old Sabaean script, which in its oldest form is evidently closely related to the old Phoenician, though it would be premature to say that the Sabaean alphabet is derived from the Phoenician. It is as likely, considering the date of both, that they are equally descendants from an older source. The characteristics of the Sabaean are great squareness and boldness in outline. It has twenty-nine symbols, whereby it is enabled to differentiate certain sounds which are not distinguished from one another in the writing of the northern Semites. As we have seen, it is a tendency in northern Semitic to open the heads of letters, and therefore it is possible that the Sabaean form for Jod 𐩺 may be older[32] than the Phoenician . Similarly if means mouth, Hommel is right in contending that the Sabaean 𐩰 is more like the object than the Phoenician , if we suppose the form, like 𐩦 or the Phoenician 𐤔‎ and 𐩣 for the Phoenician 𐤌‎ turned through an angle of 90°. So also if Kaf corresponds to the Babylonian Kappu, “hollow-hand,” the Sabaean form 𐩫 which Hommel[33] interprets as the outline of the hand with the fingers turned in and the thumb raised is a better pictograph than the various meaningless forms of k (𐤊‎, &c.).

The rock inscriptions in the wild district of Safah near Damascus which have been collected by Halévy are also written in an Arabic dialect, but, owing chiefly to their careless execution, they are to a large extent unintelligible. The character appears to be akin to the Sabaean. It has been suggested that they were the work of Arabs who had wandered thus far from the south.

There still remain for discussion the alphabets of the Indo-European peoples of Persia and India from which the other alphabets of the Farther East are descended. When Darius in 516 B.C. caused the great Behistun inscription to be engraved, it was the cuneiform writing, already long in use for the languages of Mesopotamia, that was adopted for this purpose. We have seen that Persia.at Babylon itself the Aramaic language and character were well known. It is probable therefore, a priori, that from the Aramaic alphabet the later writing of Persia should be developed. The conclusion is confirmed by the coins, the only records with Iranian script which go back so far; but the special form of Aramaic from which the Iranian alphabet is derived must at present be left undecided. The later developments of the Iranian alphabet are the Pahlavi and the Zend, in which the MSS. of the Avesta are written. Of these manuscripts none is older than the 15th century A.D. The Pahlavi is properly the alphabet of the Sassanid kings who ruled in Persia from A.D. 226 till the Arab conquest in the 7th century A.D. Under the Sassanids the old Persian worship, which had fallen with the Achaemenid dynasty in Alexander’s time, and had been neglected by the subsequent Arsacid line, was revived and the remains of its liturgical literature collected. The name is, however, also applied to the alphabet on the coins of the Parthian or Arsacid dynasty, which in its beginnings was clearly under Greek influence; while later, when a knowledge of Greek had disappeared, the attempts to imitate the old legends are as grotesque as those in western Europe to copy the inscriptions on Roman coins. The relationship between the Pahlavi and the Aramaic is clearest in the records written in the “Chaldaeo-Pahlavi” characters; the most important of these documents is the liturgical inscription of Hadji-abad, where the Arsacid and Sassanian alphabets are found side by side. Taylor (The Alphabet, ii. p. 248 f.) regards the former as probably derived from the “ancient alphabet of Eastern Iran, a sister alphabet of the Aramaean of the satrapies,” while the Sassanian belongs to a later stage of Aramaic.

Table I.

Table II.—Cyrillic and Glagolitic Symbols not given above.

The alphabets of India all spring from two sources: (a) the Kharoṣṭhī, (b) the Brāhmī alphabet. The history of the former is fairly clear. It was always a local alphabet, and never attained the importance of its rival. According to Bühler,[34] its range lay between 69° and 73° 30′ E. and 33° to 35° N., India.a conclusion which is not invalidated by the fact that some important modifications are found beyond this area, nor by Dr Stein’s discovery of a great mass of documents in this alphabet at Khotan in Turkestan, for, according to tradition, the ancient inhabitants of Khotan were emigrants banished in the time of King Açoka from the area to which Bühler assigns this alphabet (see Stein’s Preliminary Report, 1901, p. 51). Rapson[35] has pointed out that both Kharoṣṭhī and Brāhmī letters are found upon Persian silver sigloi, which were coined in the Punjab and belong to the period of the Achaemenid kings of Persia. As Bühler shows in detail, the Kharoṣṭhī alphabet is derived from the alphabet of the Aramaic inscriptions which date from the earlier part of the Achaemenid period. The Aramaic alphabet passed into India with the staff of subordinate officials by whom Darius organized his conquests there. The people of India already possessed their Brāhmī alphabet, but had this other alphabet forced upon them in their dealings with their rulers. The Kharoşţhī is then the gradual development under local conditions of the Aramaic alphabet of the Persian period. As Stein’s explorations show, both alphabets may be found on opposite sides of the same piece of wood.

The history of the Brāhmī alphabet is more difficult. In its later forms it is so unlike other alphabets that many scholars have regarded it as an invention within India itself. The discovery of earlier inscriptions than were hitherto known has, however, caused this view to be discarded, and the problem is to decide from which form of the Semitic alphabet it is derived. Taylor (The Alphabet, ii. p. 314 ff.), following Weber, argues that it comes from the Sabaeans who were carrying on trade with India as early as 1000 B.C. Even if the alphabet had not reached India till the 6th century B.C., there would be time, he contends, for the peculiarities of the Indian form of it to develop before the period when records begin. The alphabet, according to Taylor, shows no resemblance to any northern Semitic script, while its stiff, straight lines and its forms seem like the Sabaean. Bühler, on the other hand, shows from literary evidence that writing was in common use in India in the 5th, possibly in the 6th, century B.C. The oldest alphabet must have been the Brāhmī lipi, which is found all over India. But he rejects Taylor’s derivation of this alphabet from the Sabaean script, and contends that it is borrowed from the North Semitic. To the pedantry of the Hindu he attributes its main characteristics, viz. (a) letters made as upright as possible, and with few exceptions equal in height; (b) the majority of the letters constructed of vertical lines, with appendages attached mostly at the foot, occasionally at the foot and at the top, or (rarely) in the middle, but never at the top alone; (c) at the tops of the characters the ends of vertical lines, less frequently straight horizontal lines, still more rarely curves or the points of angles opening downwards, and quite exceptionally, in the symbol ma, two lines rising upwards. A remarkable feature of the alphabet is that the letters are hung from and do not stand upon a line, a characteristic which, as Bühler notes (Indian Studies, iii. p. 57 n.), belongs even to the most ancient MSS., and to the Asoka inscriptions of the 3rd century B.C. When these specially Indian features have been allowed for, Bühler contends that the symbols borrowed from the Semitic alphabet can be carried back to the forms of the Phoenician and Moabite alphabets. The proof deals with each symbol separately; as might be expected of its author, it is both scholarly and ingenious, but, it must be admitted, not very convincing. Further evidence as to the early history of this alphabet must be discovered before we can definitely decide what its origin may be. That such evidence will be forthcoming there is little doubt. Even since Bühler wrote, the vase, the top of which is reproduced (see Plate), has been discovered on the borders of Nepal in a stupa where some of the relics of Buddha were kept. The inscription is of the same type as the Asoka inscriptions, but, in Bühler’s opinion (Journal of the Royal Asiatic Society, xxx., 1898, p. 389), is older than Asoka’s time. It reads as follows: iyaṃ salilanidhane Budhasa bhagavate sakiyanam sukitibhatinaṃ sabhagiṇikanaṃ saputadalanaṃ. “This casket of relics of the blessed Buddha is the pious foundation (so Pischel, no doubt rightly, Zeitsch. d. deutsch. morg. Gesell. lvi. 158) of the Sākyas, their brothers and their sisters, together with children and wives.”

How this alphabet was modified locally, and how it spread to other Eastern lands, must be sought in the specialist works to which reference has already been made. Its extension to new and hitherto unknown languages was in 1910 in process of being rapidly demonstrated by English and German expeditions in Chinese Turkestan.

Authorities.—Owing to the rapid increase of materials, all early works are out of date. The best general accounts, though already somewhat antiquated, are: (1) The Alphabet (2 vols., with references to earlier works), by Canon Isaac Taylor (1883), reprinted from the stereotyped plates with small necessary corrections (1899); and (2) Histoire de l’écriture dans l’antiquité, by M. Philippe Berger (Paris, 1891, 2nd ed. 1892). An excellent popular account is The Story of the Alphabet, by E. Clodd (no date, about 1900). Faulmann’s Illustrierte Geschichte der Schrift (1880) is a popular work with good illustrations. For the beginnings of the alphabet, Dr A. J. Evans’s Scripta Minoa (vol. i., 1909) is indispensable, whether his theories hold their ground or not. The Semitic alphabet is excellently treated by Lidzbarski in the Jewish Encyclopaedia (1901); his Nordsemitische Epigraphik (1898) has excellent facsimiles and tables of the alphabets, and there are many contributions to the history of the alphabet in the same writer’s Ephemeris für semitische Epigraphik (Giessen, since 1900). See also “Writing” (by A. A. Bevan) in the Encyclopaedia Biblica, and “Alphabet” (by Isaac Taylor) in Hastings’ Dictionary of the Bible. A very good article, now somewhat antiquated, is Schlottmann’s “Schrift und Schriftzeichen” in Riehm’s Handwörterbuch des biblischen Altertums (1884, reprinted 1894). For Greek epigraphy the fullest and also most recent work is W. Larfeld’s Handbuch der griechischen Epigraphik (vol. ii., 1902; vol. i., 1907) (see especially Herkunft und Alter des griechischen Alphabets, i. 330 ff.). For the history of the Greek alphabet the fundamental work was A. Kirchhoff’s Studien zur Geschichte des griechischen Alphabets (4th ed., 1887): his theories were adopted and worked out on a much larger scale in E. S. Roberts’s Introduction to Greek Epigraphy, pt. i. “The Archaic Inscriptions and the Greek Alphabet” (1887), pt. ii. (with E. A. Gardner) “The Inscriptions of Attica” (1905). See also Salomon Reinach’s Traité d’épigraphie grecque (1885). in Iwan von Müller’s Handbuch der klassischen Altertumswissenschaft important articles on both Greek and Latin epigraphy and alphabets have appeared (Greek in edition 1 by G. Heinrichs, 1886; in edition 2 by W. Larfeld, 1892; Latin by Emil Hübner). See also “Alphabet,” by W. Deecke, in Baumeister’s Denkmäler des klassischen Altertums (1884), and by Szanto (Greek) and Joh. Schmidt (Italic) in Pauly’s Realencyclopädie edited by Wissowa (1894). Mommsen’s Die unteritalischen Dialekte (1850) is not without value even now. Other literature and references to fuller bibliographies in separate departments have been given in the notes. Elsewhere in this edition of the Encyclopaedia Britannica the articles on the various languages and under the headings Inscriptions, Palaeography, Writing, &c., should be consulted, while separate articles are given on each letter of the English alphabet. The writer is indebted to Dr A. J. Evans for a photograph of the Cretan linear script, and to Professors A. A. Bevan and Rapson of Cambridge, and to Mr F. W. Thomas, librarian of the India Office, for help in their respective departments of Semitic and Indian languages.  (P. Gi.) 


  1. Breasted, History of Egypt (1906), p. 45.
  2. Op. cit. p. 484.
  3. Die Schrift und Sprache der alten Ägypter (1907), p. 24.
  4. Scripta Minoa, i. (1909), § 10, pp. 77 ff.
  5. E. Piette, L’Anthropologie, vii. (1896) pp. 384 ff.
  6. E. Piette, L’Anthropologie, xvi. (1905) pp. 8-9. The apparent inscriptions of this period are conveniently collected and figured together in Dechelette’s Manuel d’archéologie préhistorique celtique et gallo-romaine, i. (1908) p. 233.
  7. Der Ursprung des alt-semitischen Alphabets aus der neu-assyrischen Keilschrift (ZDMG. xxxi. pp. 102 ff.). A still more sweeping theory of the same nature is propounded by the Rev. C. J. Ball in the Proceedings of the Society of Biblical Archaeology, xv. 1893) pp. 392 ff.
  8. In an excellent summary of the different views held as to the origin of the alphabet (Journal of the American Oriental Society, vol. xxii, first half, 1901), Dr J. P. Peters agrees (pp. 191 ff.) that the best test is the etymology of the names of the letters. He shows that twelve of the letter-names are words with meanings [in the northern dialects of Semitic], all of them indicating simple objects, six of the twelve being parts of the body. The objects denoted by the other six names—ox, house, valve of a door, water, fish and mark or cross—clearly do not belong to any people in a nomadic state, but to a settled, town-abiding population. . . . Six of the letter-names are not words in any known tongue, and appear to be syllables only. Four letter-names are triliterals, and resemble in their form Semitic words.” As 11 of the 12 which have meanings are to be found in the Assyrian-Babylonian syllabaries, he suggests a possible Babylonian origin. Different views with regard to some of these symbols are expressed by Lidzbarski, Ephemeris für semitische Epigraphik, ii. pp. 125 ff. (1906). The earliest tradition of the names is discussed by Nöldeke in his Beiträge zur semitischen Sprachwissenschaft (1904), pp. 124 ff.
  9. See, for example, the tables at the end of Roberts’s Introduction to Greek Epigraphy (1887); or Kirchhoff’s Studien zur Geschichte des grieschischen Alphabets (4th ed. 1887); or Larfeld’s Handbuch der grieschischen Epigraphik, vol. i. (1907).
  10. Cp. Fränkel, Corpus inscriptionum Graecarum Pelopennesi, i., No. 1607.
  11. See Witton, in American Journal of Philology, xix. pp. 420 ff., and Lagercrantz, Zur griechischen Lautgeschichte (Upsala, 1898).
  12. See Foat, “Tsade and Sampi” (Journal of Hellenic Studies, xxv. pp. 338 ff., xxvi. p. 286). A number of ingenious points often uncertain are raised by A. Gercke, “Zur Geschichte des ältesten griechischen Alphabets” (Hermes, xli., 1906, pp. 540 ff.).
  13. See especially Proceedings of the Society of Biblical Archaeology for 1895, p. 40; cf. also Kalinka, Neue Jahrbücher für Philologie, iii. (1899), p. 683. Similar forms are also found in the Safa inscriptions (South Semitic) with similar values, and Praetorius argues (Z.D.M.G. lvi., 1902, pp. 677 ff., and again lviii., 1904, pp. 725 f.) that these were somehow borrowed by Greek in the 8th century B.C., while in lxii. pp.283 ff. he argues that the reason why the Greeks borrowed Θ for the aspirated t was its form, the cross in 𐌈 being regarded as T and the surrounding circle as a variety of an occasional form of 𐌇 the aspirate. Here also (p. 287) as in his Ursprung des kanaanäischen Alphabets, pp. 13 f., he argues that the two forms of the digamma Ϝ and , and also the South Semitic 𐩥ω, could all have developed from the Cyprian 𐠳we. But proof is impossible without evidence of the intermediate steps.
  14. Inscriptiones Graecae, xii., fasc. iii. Nos. 811, 1149.
  15. See especially Athenische Mitteilungen, xxi. p. 426.
  16. Figured in Roberts’s Introduction to Greek Epigraphy, p. 65.
  17. Details of the history of the individual letters will be found in separate articles.
  18. It is figured most accessibly in Egbert’s Introduction to the Study of Latin Inscriptions, p. 16.
  19. Gardthausen “Ursprung und Entwicklung der grieschen-lateinischen Schrift” (Germanisch-romanische Monatsschrift, i. (1909), pp. 337 ff.) argues for a “proto-Tyrrhenian” alphabet from which Etruscan, Umbrian and Oscan descended as one group, and Faliscan and Latin as the other. Evidence in favour of such a position for the Latin alphabet is not forthcoming.
  20. For further details of these alphabets, see Conway, The Italic Dialects, ii. pp. 458 ff. The recent discovery by Keil and Premerstein (Denkschriften der Wiener Akademie, liii., 1908) of Lydian inscriptions containing the symbol 8 suggests that the old derivation of the Etruscans from Lydia may be true and that they brought this symbol with them (see article on F.) But the inscriptions are not yet deciphered, so that conclusive proof is still wanting.
  21. R. M. Meyer, Paul Braune und Sievers’ Beiträge, xxi. (1896), pp. 162 ff.
  22. In a paper published in the volume of Philologische Studien, presented as a “Festgabe” to Professor Sievers in 1896, and in a second paper in the Journal of Germanic Philology, ii. (1899), pp. 370 ff.
  23. See Literaturblatt für germanische und romanische Philologie for 1897, col. 429 f.
  24. A species of cryptography exactly like this, based upon the “abjad” order of the Arabic letters, is still in use among the Eastern Persia is (E. G. Browne, A Year amongst the Persians, p. 391 f.).
  25. Cf. Rhŷs, Outlines of Manx Phonology, p. 73 (publications of the Manx Society, vol. xxxiii.); Rhŷs and Brynmor Jones, The Welsh People, pp. 3, 502. An interpretation of the oldest ogam inscriptions is given by Whitley Stokes in Bezzenberger’s Beiträge, xi (1886), p. 143 ff. Besides the collections of ogams by Brash (1879) and Fergeson (1887), a new collection by Mr R. A. S. Macalister is in course of publication (Studies in Irish Epigraphy, 1897, 1902, 1907). Professor Rhŷs , who at one time considered runes and ogam to be connected, now thinks that ogam was the invention of a grammarian in South Wales who was familiar with Latin letters.
  26. Archiv für slavische Philologie, v. 191 ff., where the Glagolitic and the cursive Greek, the Cyrillic and the Greek uncial are set side by side in facsimile.
  27. For further details and references to literature see the introduction to Leskien’s Grammatik (not to be confused with his Handbuch), from which this is abbreviated.
  28. These are figured most accessibly in Lidzbarski’s article on the alphabet in the Jewish Encyclopaedia, vol. i. (1901); see also his table of symbols added to the 27th edition of Gesenius’ Hebraische, Grammatik (1902).
  29. See Berger’s Histoire de l’écriture dans l’antiquité, p. 252 ff.; Nordsemitische Epigraphik, p. 186 ff., from whom this summary is taken. Lidzbarski’s second volume and G. A. Cooke’s Textbook of North-Semitic Inscriptions (Oxford, 1903) contain the most convenient collections of Northern Semitic inscriptions for the student’s purposes.
  30. Müller, Epigraphische Denkmäler aus Arabien (Vienna, 1889).
  31. Epigraphische Denkmäler aus Abessinien (Vienna, 1894). Praetorius (Z.D.M.G. lviii. p. 724) holds that the oldest Sabaean inscriptions may date from about 700 B.C., that the Lihyan inscriptions are at earliest of the Hellenistic period and the Safa inscriptions still later.
  32. Praetorius (Z.D.M.G. lviii. p. 461 f.) attempts to trace the development of the Sabaean form from the Phoenician.
  33. Hommel, Süd-arabische Chrestomathie (Munich, 1893), p. 5.
  34. Bühler, Indian Studies, iii. (2nd. ed., 1898), p. 93. The account of these alphabets is drawn from this work and from the same author’s Indische Paläographie in the Grundriss der indo-arischen Philologie, to which is attached an atlas of plates (Strassburg, 1896), and in which a full bibliography is given.
  35. For a coin and a gild token with inscriptions see Rapson’s Indian Coins (in Grundriss d. ind.-ar. Phil.), Plate I.