Third person wanting. The accusatives mek, usik, Oek, eovik, are only in the oldest sources; accusative is usually same as dative.1

The genitive dual and plural ends in er as well as in ar. The possessive pronouns are adjectives with the above genitives for their stem, and declined in the strong declension.2

The third personal pronoun of three genders and the simple demonstrative are as follows:

The genitive masculine es is not found, sin being used instead; for ira, iru, sometimes irō occurs. In the Old High German simple demonstrative all the cases whose endings begin with a vowel insert i after d. Often in nominative plural neuter dei is found for diu; derā and deru vary between -a, -u, and -o.1

There are traces in Gothic of a demonstrative hi-, declined like i-. There are also the demonstratives declined like adjectives; Gothic yáins, Old High German gener, desēr, neuter diz; 2 Anglo-Saxon Ois, declined as follows: 3

Anglo-Saxon hva, neuter hvät, declined like se; Old High German huer, declined like der; Gothic hvaryis, which of many? declined like adjective second declension, is found again only in Old Norse; Gothic hvalar, Anglo-Saxon hväfer, Old High German huedar, which of two? declined regularly; Gothic hveleiks qualis, Anglo-Saxon hvilk, Old High German huelihher, New High German welcher.2

155. In the Gothic first personal pronoun the original a is reduced to i both in the root of the nominative singular and in that of the other singular cases. The genitive singular of the three personal pronouns takes an additional i to express the genitive relation, and subjoins na. This throws light on the Sanskrit genitive mama, tava, and the Zend mana, tava; for the Gothic -na seems to be arthritic (Def. 7) like the n of the weak declension (144), expressing attention directed to the person in the act of connecting it with what governs it (210); and Sanskrit -ma, -va probably expresses a second thought of the person in the act of connecting it with its correlative (8). 2 Ibid. i. p.

1 Grimm, Gram., i. pp. 790, 791.

3 Rask, Anglo-Saxon Gram., pp. 53, 56.

794-800.

The dative singular, dual, and plural of the personal pronouns is expressed by -s, which has probably come from bya. This element in the Sanskrit dative first person singular has become hya, and of the three spirants h, s, and v, s is the nearest to y, and y in coalescing with either of the others would naturally attract it to s; it would at the same time tend to make the preceding vowel i.

The accusative singular has -k. This has been explained by Bopp as equivalent to the Vedish particle -hā, -ga, which he identifies with ha in the Sanskrit first personal pronoun aham.1 Thus understood it would express personality, and would correspond to a second thought of the personal object which would be involved in the mental act of connecting it with what governs it.

The root of the nominative dual and plural of first person is vi (Sanskrit va), of second person yu. The -t of dual, Bopp takes as a residue of tva, two.2 These roots are too subjective for the other cases, and take an objective pronominal element n, to which probably w was subjoined for the dual and s for the plural; giving, with the genitive plural ending -ra (= Sanskrit -sām) vinwara, vinsara, yunwara, yunsara. Then followed by w may have been attracted by it so as to become ǹ, and changed w, to k in vinwara (vinkara), to qw in yunwara (yunqwara) on account of the preceding u, and the u might also have caused ns to be followed by w, and consequently softened to z in yunsara (yuzwara). Subsequently i and u imparted their voice to v and y, and were dropped, so that vi became u, and yu became i. It is remarkable that the dual has the plural endings, the duality being confined to the stem.

The accusative plural -s was probably -ks originally, as AngloSaxon and Old High German have -k, -h. The Anglo-Saxon stem of second plural dropped the z, so that the genitive became iuwara, çóver (134). In Old High German yu became i.

156. The feminine is expressed by i, in Gothic si, as it is also expressed by -i in Sanskrit and Greek in some adjectives and participles (5, 63).

This feminine i has become y in Gothic iyos and iya, but in iya it is also neuter, for the reduction of energy expressed by the close vowel corresponds to neuter as to feminine. The Gothic neuter ita corresponds to Latin id. The final a in ita, Oata, and in the accusative singular masculine and neuter of is and sa, must be a pronominal element referring to what the pronoun itself refers to. The pronominal genitive -zōs is Sanskrit -syās; -zái is Sanskrit -syai, of which the y has changed a to Oi; -mma is Sanskrit -sma(i); all pronominal; zē, zō, is Sanskrit -sā(m), zē retaining a sense of the y after s, which in Sanskrit has imparted itself in masculine and neuter to the preceding vowel (13); this y has changed a to i throughout the Gothic genitive; -m, -im, dative plural, is Sanskrit -b-, -ib'-.

Gothic so corresponds to Sanskrit sā, ō accusative singular to Sanskrit tā(m), 0ō neuter plural to Sanskrit tā(ni); this o of so is preserved in Anglo-Saxon seó feminine, and heó feminine. In Old High 1 Vergl. Gram., ii. p. 102; Curtius, Gr. Etym., p. 515.

2 Vergl. Gram., ii. p. 120.

German the o has become u, except in accusative singular feminine, where it is replaced by a; and in Anglo-Saxon accusative singular feminine a has been taken up by the stem Oā, Оās.

The Teutonic construction of the relative pronoun is remarkable. It is either a demonstrative pronoun representing the antecedent in the relative clause, in the case proper to that clause, and with an indeclinable demonstrative element subjoined to it which has a relative significance. Or if the antecedent is a personal pronoun it is that personal pronoun repeated in the relative clause in the proper case and with the indeclinable element subjoined to it. In Gothic this element is -ei, which Grimm identifies with the stem of Latin is, but Bopp with the Sanskrit relative ya. In Old High German the indeclinable element is dar, dir (Old High German dār= there); in Old Norse it is er.1 In Danish der, Frisian ther, which means where, is used by itself for the relative pronoun of all genders and both numbers.2

When the antecedent is a demonstrative pronoun it sometimes in Gothic takes the relative element, and is not repeated in the relative clause. And sometimes it is dropped as antecedent, and in the proper case in the relative clause takes the relative element.3

In the oldest High German the demonstrative is used by itself as relative; and the first and second personal pronouns when antecedent are repeated as relative.3

Gothic -uh

-quam.

=

Latin -que or -c; Gothic -hun =

Latin -cun or

Old Norse has a negative suffix -gi, attached to particles, nouns, and pronouns, to express not, so, neither, nothing, &c.1

Old and Middle High German have an indefinite element dih-, deh-, sih-, as deh ein ullus, and an indefinite pronoun eddes, ethes.5

The Gothic ei is used for ut, quod, and also makes other particles relative, svě = sva'ei so as, Oat ei = öri.

6

157. The Teutonic verb has only a present tense, and a past; but it has a subjunctive or ideal mood' as well as an indicative, the past as well as the present being thought either indicatively or in the other mood.

In Gothic and Old High German, the Greek and Latin future is rendered by the present. It was long after that its expression by auxiliaries became general."

In the original formation of the past, as it is seen in Gothic, there is either reduplication or change of the vowel of the root, or both. And the formations differ for past singular, past plural, and past participle. The vowel of the subjunctive past singular and plural is always the same as indicative past plural. The following table shows these vowel changes, the diphthong of the reduplication syllable put first:

1 Grimm, Gram., iii. p. 14-18; Bopp, Vergl. Gram., sect. 365. ? Grimm, Gram., iii. p. 174, note.

3 Ibid. iii. pp. 16, 17.
5 Ibid. iii. pp. 40, 41, 57.

8 Ibid. i. pp. 835, 837.

The vowel of the past participle seems to be the original radical vowel,1 except in XI. and XII., in which the radical i has yielded to the influence of a liquid following it,2 and become u; for the liquids have an affinity to u, being uttered, as u is, with a closure in the anterior part of the mouth, and unimpeded breath.

The reduplicating verbs all have a long radical vowel or diphthong, the a of I. being long by position, except in fahan and hahan, whose a is long by nature or position in all the other Teutonic languages.3

In VIII. to XII., i and u before h or r become aí, aú1 (133). Otherwise the radical vowel is short in the last six conjugations. These express the past in the singular by broadening the radical vowel; for even in VII. Gothic ō corresponds to a (133).

In all the conjugations there was perhaps a tendency to express the past by taking up the a of remotion as connected with the process of being or doing in the thought of the root (24, 27, 70, 86, 88), but there was not room in the first six conjugations to do this with sufficient expressiveness in the radical vowel on account of the long vowels or diphthongs of the root. It was therefore taken up in a reduplication syllable, but was still so associated with the process that it was lengthened to ai. In V. and VI. it also affected the radical vowels, changing them to ō=ā.

An increased sense of process in the present is expressed only in VIII. and IX. by an increase of the radical vowel.

In the plural of the past the process is less distinctly thought, because the subject, whose being or doing it is, is less distinct (24), and there is a tendency to think the past with loss of the process instead of with remotion, the being or doing as having ceased rather than as separated by an interval. This is expressed by reduction of the vowel of the present, and cannot be applied to the first six conjugations on account of the length of the radical vowel, nor to X. and XI. on account of the weakness of the vowel of the present. In the former, therefore, the plural of the past has the same stem as the singular, and in the latter it is expressed as remote, but diffused and lengthened by ē=ā (133). In VII. also the radical a has such an affinity for the a of remotion that it takes the latter in the plural as in the singular, ō in both corresponding to ā (133).

1 Bopp, Vergl. Gram., sect. 26. 3.
3 Ibid. i. p. 1023.

2 Grimm, Gram., i. p. 839. 4 Ibid. i. p. 843.