or ideograms (a more convenient term),-pictorial representations, expressing not merely visible objects, but also abstract ideas, and even actions; but each of these could also have the phonetic value of the name of the object which it depicted.

Distinct from these are the "letters "-in use, though not in origin. These have two parts-one, a symbol which was originally an ideogram, and which could still be used as such, but which in this particular combination lost its ideographic value, and retained only the phonetic value of the name of its object; the other, an ideogram, which laid aside its phonetic value, and only restricted to a particular class the phonetic symbol which it accompanied. Thus, for example, the ideogram of a ship had also the phonetic value tcheu-i.e., the name denoting ship in the spoken language; the ideogram of fire had the phonetic value hvó: these two symbols combined were still pronounced tcheu, and meant the flickering of flame. The second symbol dropped its phonetic value altogether, but kept the generic idea of fire the ship was lost, but the idea of undulating motion modified that of fire, and the complex symbol combined the two ideas, with the one sound tcheu. Similarly, the ideogram ship and speech combined expressed loquacity, and this in the spoken language was also tcheu, the phonetic value of the symbol for speech being dropped, just like that of the symbol fire above. In this way there are ten different ideas given by Endlicher (p. 10), all called in the spoken language tcheu, and all expressed to the eye by different complex symbols formed on this principle. These symbols, he reckons, form at least ths of the written language.

This is a very imperfect sketch of the Chinese system of writing, and into the history of the "keys," which indeed belong rather to Chinese lexicography, we do not propose to enter. But it is enough to throw light on some questions connected with our subject. First of all, we see ideography and phonetism existing side by side; and even the same symbol, having in most cases (not in all) either an ideographic or a phonetic value at will. Therefore, in this case the passage from the one system to the other may be considered as certain; but how it was made there is not sufficient evidence to show. It must have been earlier than the combination of pure ideograms mentioned above. It was probably greatly facilitated by the Chinese being a monosyllabic language; each syllable is a complete word in itself, expressing a complete notion: hence the idea of completeness and individuality would attach to such a combination of sound more easily than would be possible in polysyllabic language; and it would seem more natural to give that sound a symbol for itself, quite apart from its ideographic meaning. Further, as the whole number of single syllables of which the language consists is only 450, the effort of remembering the symbols could not be great, and the memory must have been already trained in that

The explanation of the cumbrous "letters" described above is simple; and it will show us, secondly, how so apparently monstrous a system of writing could be maintained, and has been in its essence maintained, down to the present day. With so few radical sounds in the language, it was inevitable that many different objects must have been expressed, as ideas grew and multiplied, by the same sound, as we saw above that there were eleven different ideas (including the ship itself) all called tcheu. These could be distinguished in the spoken language by tone or accent, and actually were so distinguished. But how were they to be distinguished in writing? Now, writing is but the visible exponent of language, and therefore is naturally formed under the same conditions-those conditions which, because the effect is obvious while the reason is often difficult to detect, we vaguely call the genius of the language: and it must accommodate itself to the defects as well as the strength of the language. There is an inherent evil in Chinese speech-inevitable in a monosyllabic language with a limited number of radicals-that the same combination of sound should serve to express many different ideas. A combination, therefore, of symbols is absolutely necessary, which shall represent to the mind through the eye the fact that the sound which is heard has changed its meaning to meet that of another sound which is not heard that tcheu no longer means a ship, but means the flickering of flame, or something else quite different. It would have been easy enough to have had different symbols for the different meanings of tcheu; but it would not practically have been so convenient, because it would not have represented so well the facts of the language. If the Chinese had chosen in their speech to do universally what they did occasionally, to form compounds like "ear-dooring" for "hearing" a thing, the native genius for pictorial representation would have produced a symbolism which might have supplied all its wants down to the present day. But that was not the bent of the language; and the writing therefore remains to the present day a mixture of ideography and phonetism, and is perhaps better so. Still, a great deal of confusion is possible. In modern writing, according to Endlicher, each syllable has several symbols, partly because of the extraordinary number of meanings belonging, as we have seen, to each combination of sound, partly from considerations of calligraphy, because it is not every symbol which will combine neatly with every other; and therefore for particular combinations a different symbol with the same phonetic value is required, so that the shapes of the mixed symbols increase in number. Also, the pictorial symbols being comparatively few, and many of these being employed phonetically for the same syllable, it is obvious that, with the growth of ideas, many new symbols must have been required. To meet this want, the mixed symbols so often mentioned were employed purely phonetically, each in new combination on the old principle with an ideogram, whose meaning was disregarded. Generally these symbols kept their phonetic worth, but sometimes in combination with particular ideograms they change. Thus we see a double evil arise in the language. Not only have we several symbols for each combination of sound, but also the same symbol can under certain circumstances have different phonetic values. But the difficulties thus caused seem greater to a stranger than to a native; and the Chinese have never been moved thereby to exchange their

picturesque but unwieldy system. The impure syllabism marked out for them by the genius of their language has been their furthest development. It was reserved for the Japanese to borrow the Chinese characters, and, expelling all ideographic associations, to employ them simply as syllables, thus advancing to a pure syllabic writing. This borrowing and extension of a system by a foreign nation will be more fully dwelt upon hereafter. It should perhaps be added that the expression of many different senses by one symbol, which has so largely modified the Chinese writing, is not peculiar to monosyllabic language. It is found in all languages, though not to the same extent: roots of different sense have been worn down by phonetic decay till they reach the same form, and this cause may have operated to some extent in China, though it cannot have been very important.

The cuneiform writing, so called from the wedge-like shape of the characters, or which compose it, was employed by different nationalities. It was first deciphered by Grotefend on inscriptions of Persepolis, and was found to be the exponent of the Aryan spoken by the conquering Persians, which belonged, as is well known, to the Indo-European family of languages. But cuneiform inscriptions in three languages were found on a monument at Behistun: the first was the Persian, and much the simplest in form; the second and third were composed of elements of the same shape in much more unwieldy combinations.1 It was obvious that the three inscriptions were identical in meaning, but in different languages; and principally by the help afforded by recurring proper names, whose value could be compared with the known values in Persian, the characters of the last two inscriptions were deciphered, and found to belong, one to the language of the Assyrian and Babylonian subjects of Darius, the other to the old Scythian population of Media, who used a Turanian speech. Other languages, the old Armenian and that of Susa, were found afterwards to be represented by the same characters; and to these different systems the collective name Anarien (i.e. non-Aryan) has been given by French writers (Oppert, &c.), to distinguish them from the Aryan-Persian, which is a purely phonetic character.

It seems clear that the origin of this system was Turanian, and that it was borrowed by the Semitic races who used it. It was originally hieroglyphic, though the stiff combinations of wedges give but little indication of such an origin. But both in Assyrian and Babylonian there is an older character and a newer one, and the older forms can again be traced back to a still more archaic shape, which was unquestionably the original of both, and which is not cuneiform, but composed of straight lines only.2 These show little of the brilliancy of invention of the Chinese; they seem to appeal to the reason rather than to the eye; they are obviously intended to recall the image of the object, but they must have been first explained in order to be intelligible at all, and then they might be remembered. For example, a house was denoted by ; a town by

. Neither of these are symbols which will be intelligible as soon as seen by a person who has not been taught them. This is probably due to the fact that they were produced, not by the hair-pencil of the Chinese, but by the chisel; they were intended to be written on rock, and for this straight lines are more convenient; and the wedge shape which they assumed afterwards may be explained

A part of this trilingual inscription is printed in De Rosny's Ecritures Figuratives, p. 70.

For specimens, see Oppert, vol. ii., p. 63.

by the ease with which it can be made by two strokes of the chisel-perhaps no other figure so clear can be produced with such facility.

This system seems to have reached syllabism before it was adopted by the Aramaic peoples. But the syllabism was still mixed up with ideography, just as we have seen was the case in China-that is, the same symbol denoted ideographically the object, and phonetically the sound, of the name of the object; as though in English we should denote by the symbol B both the insect bee and the sound be. But there is a difference between this idiom and the Chinese; it was polysyllabic, whereas Chinese was syllabic. When, then, the name of the object contained more than one syllable, the first alone was taken to be denoted phonetically by the symbol. The evidence for this is small in quantity, owing to the scanty remains of the language of that Turanian element of the Chaldee nation from which the cuneiform writing was borrowed. To this language the name Accadian has been given by Dr Hincks, and this name seems to be now generally received. But the Medo-Scythic, mentioned above, which is a closely-connected dialect, supplies us with forms sufficiently close to the old Scythian spoken originally by all the Turanian stock in that part of Asia. Thus one symbol in Assyrian denotes ideographically God and phonetically an; now the name for God in Medo-Scythic is Annap. Another denotes a city and but; batin is a city in Scythian. Another is a father and at; in Scythian a father is atta. (Oppert, ii. 79; Lenormant, i. 41.) This evidence will doubtless be strengthened with time, but even now it is conclusive; and the principle thus established, the arbitrary selection of the first part of a name to have a particular phonetic value, seems to be exactly the principle which we should a priori have expected to find if we had tried to conceive the possible ways in which ideography could pass into phonetism.

The confusion which was occasioned by the imperfection of Assyrian writing was immensely increased by the fact of their characters being borrowed, not indigenous, as in China. There is first of all the obvious difficulty of adapting Turanian symbols to a Semitic language, in which the short vowels were not written, and the meaning of the radical group of consonants in any particular place had to be determined by the context. Instead of being able to retain the same symbol to express a root in its modified forms, e.g. in the conjugation of a verb, a new symbol would be necessary for each person-form, which could be expressed by mere vowel change in the root, and these symbols might be totally unconnected, so that all sense of the connection of different parts of a verb would be lost. This is bad enough, but it is an evil inherent in the borrowing of such a system of writing to express a language whose genius was so essentially different. But there was another evil, much greater, which might have been avoided, and was not. This is polyphony-the expression of many different sounds by the same symbol. When the Assyrians took an Accadian symbol, they should have taken only its phonetic value, or one of them, if it had more than one, and in this way they might have acquired a purely syllabic character, as the people of Susa afterwards actually did. But, as was not unnatural at the time, they took it with all its values, ideographic and phonetic, and added more of their A striking example given by Oppert (ii. 85) will make this plain. In Accadian this symbol ideogram for an open hand, doubtless originally in a more elaborate form. In the spoken language a hand was called kurpi, and therefore, by the principle mentioned above, this symbol had also the phonetic value kur. But by a metaphor the hand symbol had the further ideographic

own.

was the

in which a symbol could be transferred from one object to another, because the names of the two had the same sound in the spoken language, are not very distinct, and have not been fully examined; on this point we may hope for more light from M. Lenormant. He points out that the same symbol denotes "holiness" and a "slave." No metaphorical explanation seems possible here; but both are sounded hen in the spoken language, and the community of symbol becomes at once intelligible. In such a practice as this we see at once a cause of great confusion, especially when the same symbol was employed to denote two things the names of which were not exact homophones, and yet sufficiently near in sound to allow themselves to be expressed by the same symbol; e.g., when the circle which denoted the sun was also taken to denote the idea of day, the sun was called ra, the day hru, and so the symbol became a polyphone; it had two not very different sounds. It is true that here the application of the symbol for the sun to denote the day was not caused only by the similarity of sound in the two words-it was probably employed at first metaphorically; but there can be little doubt that it was helped to its double use by the indistinctness of the Egyptian vowel-sounds, which caused the two words to be sounded nearly alike. From this and similar causes arose that polyphony which necessitated the use of the determinatives described above. Vestiges of the syllabic stage in Egyptian exist beyond a doubt, and they point to a slowlyeffected transition from the older to the newer form of writing. Thus the symbol of a fish represented at the syllabic stage the syllable an; later on, the letter a alone came to be denoted by a reed, and n by a waving line. Now we find the syllable an represented not merely by its own simple exponent, the fish, but also by the reed and fish together, that is, in phonetic value, by A. an; by the reed

above the waving line (an); and even by all three (4.

an

n

(Lenormant, ii. 44). This surely points to a stage at which the alphabetic values of the reed and line were not yet so firmly fixed that the writer could dispense with the older and more familiar sign of the fish to specialise the other two. Of Egyptian alphabetism proper it is not necessary to give examples; we are sufficiently acquainted with the use of letters pure and simple, and their use in Egypt is

not denied.

To what cause are we to assign the progress of the Egyptian beyond the Assyrian method of writing? What circumstances enabled the one nation to develop at least an imperfect alphabetism, while the other never advanced beyond syllabism? No certain answer can be given; but at least a probable suggestion is made by M. Lenormant. The Egyptian vowel sounds were indistinct: the consonants were clear and definite. Therefore it was natural (as Lepsius pointed out) that in each syllable the consonant should come to be regarded as the important element, and should finally extrude the following vowel altogether. Thus a large number of symbols, which originally represented syllables beginning with the same consonant but followed by different vowels, would become in time absolutely identical in value, the different representatives of the same consonant. And a great abundance of such homophones is actually found in Egyptian. The method, therefore, which was followed in passing from the syllable to the mere alphabetic sign, was identical with that which we have already pointed out in Assyrian, by which the symbol of a polysyllabic word was taken to have the phonetic value of the first syllable of that word; in each case it denoted the first element of the name the syllable in Assyrian, the single sound in Egyptian. And in each language the symbol thus applied to a new use still retained for a long time its old value as the hieroglyphic or at least conventional exponent of a

material object or of an idea. Thus in Egypt nefer meant good. This word in writing is expressed in two ways: first, by a single symbol-which had originally been the pictorial representation of some material object, but was afterwards the conventional symbol of the idea of goodness; secondly, by this same symbol followed by two others, which had also, from being originally hieroglyphs, acquired the phonetic values of ƒ and r; that is to say, one symbol could at will express the whole word nefer and its initial letter n. This is the natural, perhaps the only possible way of eliminating the single sound; but it is obvious that great difficulties would attend it at the outset. There could be at first no convention to restrict the symbol for n to that of the particular word nefer; any other beginning with n would have served. There was no law to prevent a writer taking as many symbols for n as took his fancy; and in fact each letter in this way did have several different symbols.

some.

It follows that while Egypt must be credited with having first invented an alphabetic system, and must for ever claim for this the gratitude of the world, yet that system was far too imperfect to become the instrument of a popular literature. It suffered equally from the opposite diseases of homophony and polyphony, from the expression of the same sound by many different symbols, and from the use of one symbol to denote many different syllables. And each of these evils was only aggravated by time. The earlier Egyptian writing is much more simple than the later, wherein homophones increased to a degree to which there was practically no limit except the strength of the memory; and the numerous phonetic devices to unravel the confusions of polyphony must have been equally burdenIt might have been expected that polyphony at least would have become extinct with time; that the different symbols for the same syllable would all have been worn down into single letters, and thus, though homophony might have multiplied, polyphony would have perished. This might have been the case if these symbols had ever become perfectly clear of their originally pictorial or conventional origin. But this was never the case. To the last, the employment of a symbol to express an object or idea continued side by side with its employment as a single letter. The spirit of hieroglyphism, real if not apparent, could not be vanquished by alphabetism; and in order that ideography may be finally expelled, it would seem that circumstances are needed more favourable than can be often found combined at any period of any nation's history. In fact, a purely phonetic alphabet is most likely to be produced when one nation borrows from another such portion of that nation's symbols as it requires for its own needs, and rejects that superfluity which only leads to confusion. We have already seen indications of this fact.

It

Many circumstances combine to render it difficult for a nation to reach of itself pure phonetism in writing. There is the strong disinclination to change, of which we have before spoken. It is always easier to put up with difficulties to which we have been accustomed all our life than to make any radical change, especially when that change causes at once serious difficulties at every moment. was easier for the Egyptians to retain the odd mixture of ideographic, syllabic, and alphabetic writing, and occasionally to add some new key for unlocking the difficulties to the formidable list which was already in use. genuity of these grammatical devices almost surpasses belief. We can only refer the curious to the hieroglyphic grammar in the fifth volume of Bunsen's Egypt's Place in Universal History. In the second place, a good deal must be allowed to the restraining influence of religion. It is well known that most of the ancient nations ascribed a divine origin to their systems of writing. It might well

The in

seem to them to be too wonderful a thing for the result of human ingenuity. Thus in one of the Assyrian lists of the different values of syllables, published, as has been already mentioned, by royal authority, Sardanapalus V. states that the god Nebo has revealed to the kings, his ancestors, the cuneiform writing, which he thus endeavours to simplify for the better understanding of his people (Oppert, ii. 53). The Sanskrit character, which is now known to be due to a Phoenician source, was called Devanagarî, “belonging to the city of the gods," unless, as Prof. M. Müller suggests (Sanskrit Grammar, p. 1), we are to understand by the gods here only the Brahmans; but whatever the name may mean, their belief in its divine origin is certain enough. And M. Lenormant points out (i. 80) that the native Egyptian term for writing meant "writing heavenly words." Now it is clear that no nation among which this belief lingered in any degree would be likely to alter fundamentally the spirit of their system of writing. Lastly, it is possible, though, as we have suggested above, not very probable, that the obscurities of the existing system may have recommended it to the priests. These reasons may suffice to show that it was not in Egypt that we should expect to find the development of a purely phonetic system. But just as the Japanese took the Chinese characters, and gave them a development which they have never had in the land of their creation-just as the people of Susiana took the cuneiform writing and made it purely phonetic, without any remnant of ideography, so the work of extracting order out of the chaos of Egyptian writing was reserved for the Phoenicians.

The Phoenicians were peculiarly fitted to perform this inestimable part in the history of human development. An active and enterprising nation, they were early brought into commercial relations with Egypt, and must of necessity have learnt something of their system of writing; they could see its advantages and its perfectly remediable faults; the advantage of one definite symbol for one sound, and the disadvantage of a dozen; the desirability that this symbol should signify that sound only, and the undesirability of its denoting a horse or a man as well. And the religious scruples which may have affected the Egyptians need have no weight for strangers. If the characters were divine for the priests of Isis, they were a convenient instrument to supply every-day wants for the sailors of Tyre.1

These considerations do not, of course, amount to a proof that the Phoenician alphabet was derived from Egypt. It is of course possible that it disengaged itself by degrees out of an earlier hieroglyphic system at home. But of such a system no vestiges remain; and the correspondence between the Phoenician characters and those of the earlier Egyptian hieratic is sufficiently striking to warrant us in regarding it as at least provisionally true that what was natural and perfectly possible did actually take place. The general testimony of the early Greek and Roman writers, that the alphabet was invented in Phoenicia, must then be limited to the sense in which Tacitus says that the Phoenicians had this credit-tanquam repererint, quæ acceperant.

It cannot be known with certainty whether the Phoenicians took, together with the Egyptian symbols, the phonetic values which they had in Egypt, or whether they totally disregarded those values, and simply assigned to the symbols the value of their own sounds at will. The first view, however, seems clearly the more probable. The Phoenicians could only become acquainted with the Egyptian

1 M. Lenormant (p. 83) will have it that the Phoenicians must have been "trés peu religieux, et au fond presque athée." They may have been so, but surely not merely in order to borrow an alphabet from Egypt. It is enough that that alphabet could have had no sanctity for them.

2 For evidence of this, see plate, p. 600.

symbol and sound together; the one would naturally suggest the other; and we should expect that they would first take the symbols belonging to those sounds which exactly corresponded in Egyptian and Phoenician, then the symbols of other Egyptian sounds which did not exactly correspond to their own, but which seemed in each case the most analogous to them; but that there would never be any violent rupture between the symbo. and its old sound. Yet it seems quite certain that there is no connection between the names which the letters bore in Phoenicia and the original object of which the Egyptian character is the debased representation. Thus the first letter of the Phoenician alphabet (corresponding to the Hebrew aleph) was named from its fancied resemblance to an ox's head, the second (Hebrew beth) to a house, and so on. But the symbol which strangely seemed to the Phoenicians like an ox is only the form, rapidly drawn, of an eagle; beth, in like manner, is the quickly-drawn figure of a crane. It would seem, then, that the Phoenicians borrowed sound and symbol, but no name. They cared nothing for the history of the symbols; and when they found it convenient to have a name for each symbol they chose some object whose name began with the letter in question; and we should have said that it was totally impossible that any similarity in form between the letter and the object whose name it borrowed could have helped to give currency to the nomenclature, did we not see evidence of similar and apparently equally impossible fancies in the names of the constellations, let the origin of those names be what it may.

Such, very briefly traced, seems to have been the origin of the Phoenician alphabet, the parent of almost every alphabet, properly so called, existing on the earth. For the main ramifications of this alphabet in subsequent times we cannot do better than translate the summary of an author already often referred to, M. François Lenormant. He distinguishes (p. 110) five main stems. These are

1. The Semitic stem, wherein the values of the letters have remained exactly the same as those of the Phoenicians, except in a few derived alphabets framed in Persia and the countries immediately adjacent, which being employed to write Indo-European languages, turn the soft breathings of the Phoenician into genuine vowels. This stem subdivides itself into two main branches-the Hebræo-Samaritan and the Aramaic.

2. The Central stem, whose province includes Greece, Asia Minor, and Italy. The transformation of the symbols of the smooth, and even of the rough, breathings into symbols of vowels is here the invariable rule. This stem contains first the different varieties of the Hellenic alphabet, then the alphabets derived from Asia in the same sense as the old Greeks did), and the Italian. In the Greek, including three families-the Albanian, Asiatic (taking the Asiatic family we distinguish two groups-one for the Phrygian alphabet only, which is made up of elements whose origin is exclusively Greek, the other containing the Lycian and Carian, where these elements are mixed up with Cypriote characters. The Italian family must also be subdivided into an Etruscan group and a Latin group, between which stands the Faliscan alphabet, of a mixed character.

3. The Western stem, containing the systems of writing which resulted from the spread of the alphabet by the colonists of Tyre among the indigenous inhabitants of ancient Spain. This stem reckons but one single family. It has, as that which precedes it, for its fundamental character the change of the value of the Phoenician breathings. But the direction in which the forms of the letters vary is signally different.

4. The Northern stem, containing only one branch, the runes of the Teutonic and Scandinavian peoples, who were settled at a particular epoch in the north of Europe, but had arrived from Asia, where they still lived during a part of historic time, and where they must have had imparted to them the alphabet produced by the Phoenicians. Some elements in the runic writing seem to point to a direct reception of the writing from the seamen of Canaan; others, on the contrary, bear a certain stamp of Greek influence.

3 The only two alphabets, in the strict sense of the term, which M. Lenormant cannot classify as of Phoenician origin are the Cypriote and the Persian cuneiform-the former still imperfectly deciphered, but seemingly to some extent syllabic; the latter perhaps not pure alphabetic, but retaining certain ideograms.