INTEREST IN FORTUNE AND CIRCUMSTANCE. [SECT. VI. (this chap. III. 4, 5), to reduce a dependent verb to a verbal noun. But in Hungarian this connection is weaker. There is less adjustment of the verb to what it governs. And such a shortcoming in the adjustment of plan to fact and circumstance, arising probably from their favourable region dispensing with the necessity of it, accounts for the absence from Hungarian of gerund as well as of subjunctive (Gram. Sk., IV. 121). In Samoiede the gerund or verbal noun may take person endings which give it an appearance of subjectivity, but they are in truth possessive suffixes, and indicate close connection, but not subjective inherence (ibid. 98, 8). 6. In the Chinese group of languages there is no subjunctive mood. For those races have not sufficient originality of plan or design to adjust a fact or circumstance as aim or accessory part of a being or doing, carrying this subordination to the latter into the idea of the former, so as to affect its element of succession or process. They are, moreover, too realistic for the development of ideal moods, though they may express potentiality and such ideas as a matter of fact by the indicative of auxiliary verbs. In Japanese the verbal stem can take postpositions like a noun to express its government by another verb (ibid. V. 45), the weak subjectivity (this chap., III. 3) yielding to the subordination, so as to let the verb be treated as a noun, and the subordination corresponding to the degree of plan and combination shown by the race. 7. In Arabic, Ethiopic, and Amharic there is a subjunctive, in which the sense of realisation in the subject is reduced, and an ideal mood, in which in Arabic it is reduced further still (Gram. Sk., V. 55, 125, 136, 145). But neither of these is preserved in Hebrew, Syriac, Tamachek, or Haussa. In the desert the Arab needed contrivance and plan so far as objects and circumstances furnished materials for them, and when these could not be formed he had to wait on fortune, so that he had sufficient sense of object or aim and of condition to affect a verb with dependence as such on another verb, and to develop a subjunctive mood (ibid. 55), and sufficient sense of the imagined to develop an ideal mood. There was use too for a subjunctive and an ideal in Ethiopic and Amharic, for in Africa attention is attracted strongly to the external accessories of being and doing as well as to the gifts of fortune. But in Tamachek and Haussa there is not sufficient sense of the being or doing in the verb to maintain an ideal mood or a true verb in a dependent position, and it is apparently an infinitive or verbal noun that is used instead of the latter (ibid. 158, 161). In Palestine and in Syria life was easier than within the desert, and though thought tended more to external objects than in Arabia, there was less necessity for plan and contrivance, and less dependence on fortune. The contingent and ideal, therefore, was less thought. And the weaker sense of relations or dependence on the principal verb which arose from there being little plan accounts for the absence from Hebrew of the subjunctive mood. It explains also the more SECT. VI.] INTEREST IN FORTUNE AND CIRCUMSTANCE. verbal nature of the Hebrew than of the Arabic infinitive (ibid. 92); for the stronger sense of relation or dependence reduced the latter to a verbal noun. In Syriac the infinitive is very rarely used as a noun (ibid. 117); but there being less sense of the subjective process in the Syriac verb than in the Hebrew (this chap., IV. 13), it did not develop a more verbal as well as a less verbal infinitive. In Ethiopic, and in Amharic and Tamachek, there is a so-called verbal infinitive and a nominal infinitive, the former of the nature of a gerund, the latter a noun (Gram. Sk., V. 128, 145, 158). The nominal nature of both was due, probably to the sense of relation or government by the principal verb in these languages, which, however, though greater than in Hebrew, is less than in Arabic, for there are no case endings, and to this is due the more verbal nature of one of the infinitives compared with the Arabic infinitive. 8. The Indo-European races had such art and plan that in their conception relations are thought with more distinctness than by other races. The thought of a relation with them involves a sense of the two correlatives, but may be clear of both of them; whereas other races lose the true thought of a relation by losing the simultaneous sense of the correlatives or think it in connection with one correlative. Thus the Syro-Arabian tended to think a relation in connection with the second correlative; and in thinking one fact as a related part of another, the relation tended to be thought with the former, and to be carried into the idea of its verb, the subordination to the principal verb falling mainly on the verb of the subordinate sentence (ibid. 93). By the Indo-European, the relation was thought more distinctly from the subordinate fact, and this retained more sense of its own organisation. Its verb was less affected in the being or doing realised in its own subject, and was thought more strongly as the governing member of the subordinate sentence. Hence, when one sentence governed another through an expressed relation, it did not, except in Latin when the relation was close, so subordinate the latter to the former as a part of it, that it was expressed by a subjunctive mood; although when the governed sentence was direct object to the principal verb, its verb was reduced to the infinitive. This use by the Latin of a true subjunctive in relative sentences is a striking feature of the language, as it corresponds to the practical genius of the race, by virtue of which they had a stronger sense of the bearing of facts and circumstances as accessory to their beings and doings, and of the subordination as such of the former to the latter. With this also agrees their more matter of fact and less ideal character than that of the Greeks, in consequence of which they had less interest in the imagined, and had only one ideal mood, while the Greek had two. Sanskrit had less ideality than either; for it did not carry its one ideal mood into the past or the future so as to give it any tense except the present. Sanskrit was evidently affected by a Dravidian influence which lowered the life of its conception of fact. Hence came its reduced use of the tenses. And hence also came the loss of the second ideal mood which Zend had (ibid. VI. 52), and its large use of the ACTION THOUGHT IN ITS END. [SECT. VII. gerunds (ibid. 42), as well as its loss of the elements of relation thought separately from both correlatives. Perhaps it was the superior productiveness of his native region, enabling the Latin to supply his wants more independently of fortune, which made him more practical and less ideal than the Greek. But, however this may be, the fact that he was so is certain ; and his genius being such would lead him to note circumstance more strongly as subservient to his purposes, and to think less of the possibilities of the unknown. His development and use of moods, as compared with that of the Greek, is a strong confirmation of the principle of Book I., chap. iii., 6, which has been borne out by all the languages that have been examined. 9. Bask, too, in accordance with the strong sense of objects and conditions which is shown in the cases of the noun and the object elements of the verb, has a subjunctive as well as ideal moods (Gram. Sk., Bask, 10). VII.-Development of the passive verb, according to the tendency of the race to think action in its end; that of derivative verbs according to what gives interest to doing and being in the life. 1. The use of the passive verb is carried farther in Tagala than in any other language studied in this work. And, therefore, in that language its nature may be best seen. Now, its great use in Tagala arises from a tendency to think the fact in its end, as accomplished in the objects and with the conditions (Gram. Sk., III., 57); in consequence of which tendency, the fact is so generally thought, not from the standpoint of the agent, but from that of the object or condition. And there is a tendency of the same kind, though not nearly to the same degree, in Polynesian (ibid. 7), which also thinks fact as process to an end. In Tongan, though the verb passes to the object more immediately than in the purer Polynesian dialects, there is at the same time a stronger sense of the action of the subject which keeps the verb from being thought in its accomplished end, and no passive is formed (ibid. 16, 2, 3). But in Fijian the verb is thought with stronger reference to the object, and a passive is formed (ibid. 17). 2. In the Melanesian languages fact is thought less in its end, and is more tenacious of the subjective standpoint of the agent. But the languages of Maré and Lifu have a strong sense of the end of accomplishment, which they think as quiescent (ibid. 34, 37). And in them the verb has a passive construction, and there is a tendency towards this construction in subordinate or dependent verbs, which, when active, lose subjective energy, and are thought rather as states of action (ibid. 36, 3; 37). 3. Malay also has a strong sense of the end of action as a state of the object (ibid. 75, 76). 4. Among the Syro-Arabian languages the Arabic only, which only SECT. VII.] ACTION THOUGHT IN ITS END. had an accusative case ending, thought the verb sufficiently in relation to the object to be able to carry the simple verb into the object so as to think it completely from the standpoint of the object as a passive state of the object. Hebrew could do this only with the strong derived forms the causative (Hiphil) and the intensive (Piel), which from their nature have strong reference to the object. To the passive of the simple verb it could only approach by thinking it as a reflexive. The reflexive and the passive agree so far that the object of the action is in both the subject of the verb; but when the reflexive is used to express the passive, the subject realises the verb as thought from the standpoint of another who is the agent; whereas in the passive the subject realises the verb as thought from his own point of view. Into this point of view of the object the Hebrew could not enter with the simple verb, nor could the other Syro-Arabian languages, except Tamachek and Haussa, enter into it with any verb, all of them, with these exceptions, using reflexives for passives, because their verbs were not carried to the object as much as the Arabic verb. For when the action is thought in its end in the object, the mind passes more readily to the thought of it as realised by the object and seen from the object's point of view, this being the end, which is subsequent to the action. Tamachek and Haussa acquired under African influence a tendency to think the verb in connection with related objects. At least in Tamachek this is shown by the effort to form pronominal connections (Gram. Sk., V. 162). 5. So among the African languages a passive is formed by Kafir, in which the verb is thought with strong reference to the object, so as to take up a representative of it (ibid. I. 11); also, though less distinctly, by Mandingo and Susu, which have strong sense of the object (ibid. 33, 50), but not by Woloff, in which the verb has little reference to objects, nor in Bullom, in which the verb has to be supplemented by an additional element to carry it to the object, nor in Vei, in which, what the subject does or is forms so unimportant an element of fact (ibid. 36) that it is little thought as affecting an object, nor in Yoruba, which in its fracture of the verb shows that the object has not that attraction for the main body of the verbal root that it has in Mandingo (ibid. 22, 33). On the other hand, a passive is formed in Hottentot (ibid. 70), in Galla (ibid. III. 165), in Bari (ibid. 155), in Barea (ibid. 140), in Nubian (ibid. 131), in Egyptian, though not much used (ibid. 119), and in the Dinka auxiliaries of the past and future ti and bi by lengthening their vowel, the subjective element of the present being apparently too weak to admit the modification (ibid. 147); but Kanuri and Pul form only a passive participle (ibid. 178, 186). The nomadic life which belongs to the Hottentot and Galla belongs also in part to the Dinka (ibid. 142) and Bari (ibid. 151); and the material industry with which it is occupied leads thought strongly to the object and effect. As to the Barea, information is wanting. But in Nubian the elements of relation to the objects which are infixed in the verb (ibid. 131) show a strong tendency to think the verb in reference to ACTION THOUGHT IN ITS END. [SECT. VII. the object, and the same tendency in Egyptian is involved in the sense of the accomplishment which distinguishes it (ibid. 116). In Kanuri the classification of the verbs as more or less subjective (ibid. 176) shows that they are thought so strongly in connection with the subject that they have little reference to the object. And in Pul, while the verb takes subject prefixes like Kafir, showing close connection with the subject, it does not, like Kafir, take object infixes, showing that it is not thought in close connection with the object. 6. In the nomad languages Tartar, Mongolian and Tungusian, a passive is formed (ibid. IV. 7, 22, 41, 50, 56, 62). For the care of flocks and herds involves habitual attention to external objects and effects; and action is thought with strong reference to these. But in the more northern regions objects which may be useful for the purposes of life are scarce, and methods of procuring subsistence become necessary which involve persevering action, and which engage the interest of the race as their main occupations. The interest is thus drawn rather to courses of action than to objects, and the thought of action becomes associated with elements of continuity thought as defining what is to be accomplished. Accordingly, the remarkable feature appears in these languages of a surprisingly large development of derivative verbs with the absence of a passive distinct from a reflexive. This is the case in Samoiede (ibid. 96), in Ostiak (ibid. 105, 109), in Tscheremissian (ibid. 135), and in Sirianian (ibid. 145). But in regions of somewhat milder climate, in which useful objects of action were somewhat more abundant, a passive is found, as in Hungarian (ibid. 118), in Finnish (ibid. 151), and even in Lapponic (ibid. 161), the climate of Lapland being mitigated by the Gulf Stream. 7. Passing to the most northern region of America, we find in Eskimo also a great development of derivative verbs of process, with the absence of a passive distinct from the reflexive (ibid. II. 5, 15). But in Cree there is an intensely strong sense of the object and a passive form of the verb (ibid. 18, 27). In Dakota the verb is not thought in its reference to the object (ibid. 42), and there is no passive (ibid. 41). In Choctaw the verb seems to be thought in its end (this chap., IV. 11), and therefore in connection with its object, and there is a passive, which, however, involves no general passive element, and is developed by observation of the object, on account of the intense interest with which objects were observed (Gram. Sk., II. 47, 49). In Yakama the verb is not thought in close reference to the object (ibid. 56), and there is no passive (ibid. 56). In Selish the verb is thought so much in the object, that when this is plural the verb takes up the plurality (ibid. 64); and the root of a transitive verb may take the intransitive persons and verbal element, and become passive in its meaning (ibid. 63). But in Pima, although the verbal stem is thought in close connection with the object it is not thought in its end, there being a strong sense of the activity of the subject (ibid. 68). Hence there is no passive in Pima (ibid. 72). In Otomi the verb is thought in very close connection with the subject (ibid. 81, 82), and there is no indication of its being thought in |