One would scarcely expect to find anything like this in cultivated languages. Yet it does occur in both Sanscrit and Greek. The now meaningless particle _sma_ in Sanscrit when it follows the present changes it into a past, and in Greek a? alters the indicative into a subjunctive.

To form this general type, the Maipure makes use of the unchanged possessive p.r.o.noun, and treats nouns and verbs in the same manner. The noun must always be united to a possessive p.r.o.noun, a trait common to all the Orinoco tongues and many other American languages. In the 3d person sing., however, neither the verb nor the noun has such a p.r.o.noun, but it is to be understood; _nuani_, my son; _ani_, alone, not son, but "his son." The 3d pers. sing. of the verb is often the mere stem, without a personal sign, but that this peculiarity should also extend to the noun I have met only in this tongue. It is evident that a p.r.o.noun is considered as essential to a noun as to a verb, and although a similar usage is found in many tongues, yet it appears in none so binding. There are, indeed, some nouns which are free from the necessity of thinking them in connection with a person, but these have the suffix _ti_, which is dropped when the possessive p.r.o.noun is added; _java ti_, a hatchet, _nu java_, my hatchet. From this it is evident that _ti_ does not belong to the stem, and is incompatible with the use of a possessive, hence it is the sign of the substantive, in its independent condition. The same occurs in Mexican, and the chief termination of substantives, _tli_, is almost identical in sound with that in the Maipure.

In this respect the verbal, conjugated with the personal signs, differs nothing from the noun united to its possessive p.r.o.nouns.

Grammatically, the form first becomes a verbal one by the added particles of tense and mode. The signification of these can generally be clearly ascertained, and thus are united closely to the stem.

The particles which the language of the Abipones uses to form the general verbal type are quite different from the possessives. The tense and mode particles have elsewhere in the tongue independent meanings. Thus _kan_, the sign of the perfect, means a thing which has been, time that has past.

In the language of the Mocobis the personal signs consist merely in letters, prefixed and suffixed, and have no apparent relationship to the p.r.o.nouns. By affixing these letters, phonetic changes take place so that the stem is combined with them into one form.

Among the tense signs, a prefixed _l_ indicates a past time, a suffixed _o_, the future; but the others are independent particles, loosely attached to the stem.

I have already shown how the Mbaya language conjugates adjectives with the independent p.r.o.noun, and participles with the possessive p.r.o.noun.

The signs used in the conjugation proper of the attributive verb, do not appear elsewhere in the tongue, and must have descended from an older period of its existence.

In the tense and mode signs it is easily perceived how descriptive phrases pa.s.s into true forms. For the imperfect and pluperfect the speaker can choose among a number of particles, all of which indicate past time. The modes have definite signs, but these are merely appended, and some have separate significations. The future and perfect have not merely fixed particles, but these are worn down to one letter, so that the stem is actually incorporated with them.

2. In the languages heretofore considered the personal signs added to the word make up the conjugation, and the other signs are attached loosely and externally. The reverse of this, though not perfectly so, appears in the Lule language. The tense and mode signs, often of but one letter, are immediately and firmly attached to the stem, and the p.r.o.nouns are affixed to this to complete the conjugation. These p.r.o.nouns are, however, the ordinary possessives, so that noun and verb become in a measure identical; thus, _camc_ means both "I eat" and "my food;" _c.u.muee_, "I marry" and "my wife;" only in a few examples are the verbal p.r.o.nouns distinct from the possessives.

In this case, therefore, the personal signs are independent elements, occurring elsewhere in the language, while the tense and mode signs are true affixes.

The inflection-syllables form with the stem real verbal forms, and so far the conjugation of this language belongs to the third case. But each of the elements has its fixed position, and as soon as one has the key to the combination, he can recognize and separate them at once.

Reasons which it would require too much s.p.a.ce to set forth render it probable that all the tense signs are really auxiliary verbs or come from them. This is evident of the optative, as has already been shown.

The present only is simple, as it has no tense sign.

Slight differences are found between the personal signs of some tenses, so that these tenses can be distinguished by them, a trait usually seen only in tongues so far cultivated that the grammatical forms have undergone such changes as no longer to present simple and uniform combinations. Equally curious is the regular omission of the tense sign of past time in the third person plural only. Although, except in this case and that of the present, each tense has its definite sign, inserted between the stem and the personal sign, yet there are, besides these, various particles expressing past time, which can accompany the usual tense form, so that there is a double sign of time, one in the word itself and one loosely attached to it.

The languages of the Mbayas, Abipones, Mocobis and Lules are closely allied both in words and in some grammatical forms. It is all the more extraordinary, therefore, to find the last-mentioned pursuing a method in the structure of its verb which is almost totally opposed to that in the other three tongues.

_Case 3d._

The languages of this cla.s.s approach in their conjugations those of the more cultivated tongues, in which each verbal inflection has a fixed and independent form. Both the person, the tense and the mode signs are united to the stem, in such a manner that none of the three can be said to be either less or more loosely attached than the others.

All the conjugations about to be discussed lack, however, that fixity of form which grammatically satisfies the mind.

The elements are placed definitely and regularly one by the other, but are not incorporated into each other, and are therefore readily recognizable.

They are found, moreover, outside of the verb elsewhere in the language either without any change or with slight differences of sound; the personal signs as p.r.o.nouns, the other affixes as particles.

The composition of the verb is separable, and may receive into itself other parts of speech.

No American language is free from these drawbacks to perfection of form in the conjugations. In some all three are found; in most the first and last. In really grammatically developed tongues, as in the Sanscrit, Greek, Latin and German, none of these imperfections exists.

The verb includes in itself no part of its object, the affixes modifying the stem have lost all independent life, and the a.n.a.lysis of the formal elements becomes a difficult philological task, which often fails and only rarely can be fully proved.

I shall discriminate in regard to the conjugations about to be considered that which is an approach toward a fixed form from the intentional separation of the form to insert a governed word.

_1. Approach toward a Fixed Form._

In the Mixteca language, the personal sign is the unchanged possessive p.r.o.noun. If the verb is governed by a noun in the third person, the possessive is dropped. It is left to the speaker to choose whether he designates the person, either by prefixing the personal p.r.o.noun or suffixing the possessive. The tense signs are prefixed syllables, but the perfect and future signs are altogether different from those of the present, and materially alter the verbal stem.

The Beto language prefixes the personal signs and also the possessive p.r.o.nouns to the nouns. As the latter are not fully known, we cannot judge of their ident.i.ty with the verbal p.r.o.nouns. The latter do not seem to differ much from the personal p.r.o.nouns. The tense signs are easily recognized suffixes.

Another conjugation of the same language, by the suffixed p.r.o.noun without tense signs, and with the verb omitted, has been mentioned above (I, 1), as forming a substantive verb.

A second substantive verb arises from the conjugation above explained, with the tense signs.

These two forms may also be combined, and this ill.u.s.trates with what superfluous fullness grammatical forms spring up even among rude nations. The conjugation with the tense sign is changed by a participial suffix into a verbal, and then the p.r.o.noun is suffixed, as in the conjugation without the tense sign. The latter, therefore, stands twice in the form. The p.r.o.noun used in the conjugation with tense signs may also be prefixed to a simple adjective, and the p.r.o.noun used in the conjugation without tense sign is suffixed to this, and the participial ending is then added. This is treated as a verb with the substantive verb understood. But sometimes the verb "to be" in the form without tense signs is added, and then the whole form contains the p.r.o.noun three times, without gaining thereby any additional meaning.

The Carib conjugation seems to have arisen from the forms of many dialects or epochs, and is therefore more complicated and formal, and less easy to a.n.a.lyze.

The personal signs are prefixed. In the substantive verb there are two cla.s.ses, of which only one is also common to attributive verbs. The other indicates in the verb "to be" also the connection of persons with the infinitive and gerund, and is therefore of the nature of a possessive. It may also be that when it is combined with other tenses, the notion among these nations is altogether a substantial one, as we have already seen with the subjunctive.

The stem often receives the addition _r_ or _ri_, the meaning of which is not known.

The structure of the Tamanaca conjugation also reveals a combination of at least two separate structures. Some tenses use as their personal signs entire p.r.o.nouns, almost identical with the personals. Other tenses merely change the initial letter of the verb, while there is little similarity between these affixes and the p.r.o.nouns. In the plural some of the persons insert a syllable between the verb and the tense sign.

The tense signs are suffixed, and consist merely of terminal letters or syllables, except two true particles, which distinguish the continued present from the present aorist.

There are an initial _y_ and a _t_ occasionally appearing in all persons, of which we can only say that they are not radicals.

The conjugation of this language, therefore, consists of elements not readily a.n.a.lyzed.

The Huasteca language prefixes the possessive p.r.o.nouns as personal signs. It may also drop them, and use in their stead the independent p.r.o.nouns; or may combine both; or may use abbreviated personals; so that there is a prevailing arbitrariness in this part of the verbal form.

The tense signs are usually suffixes; but in the future they are prefixes, which are incorporated with the personal sign placed between them and the stem. They consist of simple sounds, of no independent signification. But the particles of the imperative are so separable that when this mode is preceded by an adverb, they attach themselves to it.

The Othomi language does not make use of the possessive p.r.o.nouns in the conjugation, but suffixes abbreviated forms of the personals, or else prefixes others of special form, but identical in many letters and syllables with the personals. In the present condition of the language the suffixes are used only with the substantive verb; in the attributive verb, however, they may have been driven forward by the governed p.r.o.nouns suffixed. Every verbal inflection may also take, besides its p.r.o.nominal prefix, also the unabreviated[TN-9] personal p.r.o.noun in front, or the abbreviated one after it.

The tense signs consist princ.i.p.ally of single vowels, by means of which the p.r.o.nominal prefixes are attached to the stem. The imperfect and pluperfect alone have besides this a loosely attached particle.

The past tenses possess a prefix, which we have already seen appears to have been derived from an auxiliary verb.

In the third person of some tenses in certain verbs the stem undergoes a change of its initial letters, which appears to transform these inflections into verbal adjectives, an instance of the confusion of the ideas of noun and verb common in all these languages.

The Mexican language possesses a peculiar cla.s.s of verbal p.r.o.nouns which form the personal signs. This p.r.o.noun is similar to the personal in its consonants, but has a vowel of its own. It is a prefix. The plural is marked by the accent, or by a special termination. This personal sign is inseparable from the verb, but the speaker may also prefix the independent personal p.r.o.noun.

The tense signs are all without signification, being single letters or syllables. The perfect is marked not so much by an affix, as by changing, the termination of the verb in various ways, but chiefly by shortening and strengthening the sound. All tense designations are placed at the end of the word, except the augment for past time. If by augment we mean a vowel sound prefixed to the verb in certain tenses in addition to their usual signs, then the Mexican is the only American language which possesses one.

The modes are designated by loosely attached particles, also by a different structure of the tenses, and in the second person a peculiar p.r.o.noun.

Thus the Mexican conjugation consists of true verbal forms, not of separate parts of speech of independent significance; but the elements of these forms are easily recognizable, and can be reached without difficulty.

The most difficult to a.n.a.lyze, and hence the most nearly approaching our conjugations, is that of the Totonaca language.

The personal signs differ from the p.r.o.nouns. That of the 2d pers.

sing. is not easily recognized, and several forms of it must be a.s.sumed. Its position as a prefix or suffix differs, and it is variously located with reference to the other verbal signs. Still more difficult is it to distinguish the tense signs. There are three different systems of prefixes and suffixes in the conjugation, and the plan on which these are combined with each other serves to distinguish the tense. But only a few of these affixes really appear to designate tense; of the others this may be suspected at best, and of others again it is improbable.

Thus there are verbal affixes which cannot be considered to designate either persons, modes or tenses.