Imperial Dictionary; and after fifteen months' labour, had the satisfaction of seeing every character in the dictionary derived from another, classed under its proper primitive. The result of his search is now laid before the candid reader. Exclusive of the two hundred and fourteen elements, the number of characters from which another is formed, amounts to three thousand eight hundred and sixty-seven.* From these by the addition of a single element to each, is formed the great body of the language, in nearly the same manner as the great mass of the Greek language is formed from about 3500 primitives, and that of the Sungskrit language, from about 1700 dhatoos or roots. The greatest number of Derivatives which spring from any one of these, is seventy four, and the least, one; the addition of a single element to the primitive, forms each derivative, which in general expresses an idea in some measure distinct from that of the primitive character, but still bearing some relation thereto. It is however proper to observe, that the term "primitive" is not applied to them on account of their origin, but merely with reference to their use. In the former sense few could be properly termed primitives beside the 214 elements; nor indeed all of them, as we have already seen that some of them are evidently compounded of two or three others. It is merely on account of their office in the language therefore, that the name is given. Thus maì, to sell, contains three elements, and produces no less than thirty-five derivatives by combining itself separately with that number of elements; Etsai, ability, produces nine; file wooh, not, kin, now, no less than sixty-two.

twenty-five; and

*It will be easily understood, that in all the numbers given here, the possibility of mistake is allowed. This however is an immaterial circumstance: no numbers are given which the writer has not examined himself, and many of them more than once. It is presumed therefore, that few mistakes will be found so great as to affect any conclusion drawn from the number.

These 3867 primitives however, are not all equally prolific; more than seventeen hundred of them produce only one derivative each; and as they themselves are in general derivatives formed from some of the other primitives, they scarcely deserve the name. Were we to rank among the primitives every Greek word which produces another, the number of Greek primitives would be swelled far beyond that of these Chinese primitives. Thus σvenaμeave to collect into one, by dropping a preposition can be easily reduced to aμbavw; but this latter does not exalt itself to the rank of a primitive by producing two derivatives. We properly ascend higher, to the root λaμbavo, which produces above fifty. We may therefore exclude from the rank of primitives, not only the 1726 which produce each only one philological shoot, but even those that produce only two; which will be found to be four hundred and fifty-two of the remaining 2141. This will leave one thousand six hundred and eighty-nine characters as forming the great mass of the language, which is evident from comparing them; the 1726 which produce one derivative each, can of course produce only 1726; and the 452 producing two each, only 904; taken together, 2630. So that if we estimate the number produced by the 3867 primitives at twenty-five thousand (five-sixths of thirty thousand,) 2178 of these primitives, if we may thus term them, will produce only 2630 derivatives, while the remaining 1689 will produce 22,370. These then are the real primitives of the language; few indeed, yet sufficiently numerous for the purpose of forming it; for were we to divide the twenty-two thousand derivatives equally among these sixteen hundred and ninety primitives, this would give scarcely fifteen to each of them; a much fewer number than a Greek primitive in general produces; some of which,

E 2

as 27 for example, produce more than two hundred derivatives, which is more than double the number produced by any Chinese primitive.*

Were we further to deem the elements themselves primitives, which office they really fill, as well as that of formatives, (since of two elements united, the one to which the forming element is added, must be the primitive,) we should find, that as there are scarcely 3600 characters produced by the union of two with each other, each of the 214 could not on an average claim a greater number of derivatives than sixteen each. It would indeed be easy to shew by examples that this mode of classing the elements in their union with each other, has much the advantage in elucidating the sense from the connection of the two elements. If we then add the 214 elements to the 1689 primitives, we shall have one thousand nine hundred and three characters producing nearly the whole language, and this by associating with themselves as formatives, 214 of their own number. The manner in which this simplifies the language, is too apparent to need pointing out. By beThe following statement, which is the result of repeated examination, will shew pretty accurately what share each of these primitives has in producing the language:

coming fully acquainted with these nineteen hundred characters, a man is in possession of all the materials of which the language is formed; and if the principles already mentioned, run through the formation of the whole language, his noticing the effect produced by adding the various formatives to a few of them, must assist him in giving a pretty shrewd guess at their effect when united with the rest.

It is somewhat singular, that in the number of its primitives, the Chinese language, widely different as it is in its structure, should correspond so nearly with both the Greek and the Sungskrit languages. Of the two kinds of primitives which Nugent gives as the basis of the Greek language, the first and most important part contains about 2100, and the second part, deemed less important, about 1500. Thus, without taking into the account derivatives which may happen to produce one or two other words, the number of Greek primitives, nearly equals the largest number of Chinese primitives, even if we include these which produce only one derivative; while the first and most important part of them exceeds in number the most important part of the Chinese primitives. The 1760 Sungskrit dhatoos also, exceed the most important part in number, and if we were to call these primitives too, which after receiving one preposition, still form two or three words by associating anew other prepositions, the number would fully equal the number of Chinese characters here termed primitives.

OF THE VARIOUS CLASSES OF THE PRIMITIVES.

Having ascertained the existence and the number of the Chinese primitives,

it will perhaps throw further light on the language if we examine minutely the characters which compose them. We might indeed, as has been already observed, apply this name to the elements themselves, as most of them really form derivatives by receiving other elememts. Nor do they form a profusion; in this respect they are equalled by these we are now treating of as primitives. But in one respect they differ from these; the elements almost exclusively perform the office of formatives, very few of the 1689 primitives serving as formatives, and these only in an instance or two. Leaving them therefore, we come to the real Primitives; and on examining these it will be found that the greater part of them consist of those characters already described as probably the first formed, or the original characters of the language. They may be included under three classes:

I. The first class consists of those characters which are formed from an element by some addition that, taken alone, has no meaning. They are therefore not formed by the union of two ideas, for if divided, the two parts do not convey each a distinct idea. These then are the next remove from the elements themselves. Among them are shee, an age, the present age or state of things; which is placed underneath yih, one, as its key; but the other part conveys no distinct idea. From this, which is a character much in use, are formed twenty-two derivatives, by its receiving separa'ely as many of the elements. Another character of the same kind ismin, the people, the multitude, which is formed from shee, a tribe, by an addition, however, which of itself means nothing. This primitive produces twenty-eight derivatives The character Akin, now, is another of these; which is formed by placing two strokes beneath the elementyin, man, but these two strokes have no meaning of themselves. This character, which is in common use,