ON THE FORMS AND ORIGIN OF THE PRONOUNS OF THE FIRST AND SECOND PERSONS

BY DR. C. LOTTNER

[Extracted from Transactions of the Philological Society, 1859, pp. 34-66.]
 



Though linguistic science is unable to give any thing like satisfactory testimony on the question of the historical unity of mankind, yet in another, and in a far higher sense, it can establish the oneness of our race. It can prove the ideal unity of human nature, of which the sought-for historical unity is indeed but the shadow and the symbol. This it does by showing that, throughout the strange tumult of discordant voices, ever and anon we perceive the same psychological laws at work, leading nations widely different in origin along the same path in the formation of their languages. It was in this sense that Buschmann pointed out the similarity of the words meaning father and mother iu nearly all the languages of the globe. It was in this sense that Pott in his 'Zahlmethoden' proved the fact that, with few exceptions, all nations choose for the base of their numeral systems the number of the fingers of either one hand or both, or else of the fingers and toes taken together (5, 10, 20). It is still in this sense that I would call attention to the remarkable coincidences to be observed in the pronouns of the first and second person, which are it will be seen of such a nature as to afford ample proof of the identity of the human mind throughout the globe, while on the other hand they fail utterly to establish identity of origin.


PART I.

I shall first notice the phonetic similarity prevailing in [p.35] this class of words, and then I shall try to enter somewhat more deeply into the psychological process by which they are developed.

To begin with the phonetic part, I only remark by way of introduction that generally I shall give to one family of languages (such as the Indogerman) the right of producing only one witness, because the pronouns under consideration were mostly created before the separation of the families. From this rule I shall deviate only, when a language disagrees with its kindred tongues as to the main principle in the formation of these words. Thus, all the Indogerman forms of speech widely diffused and much diversified though they are now-a-days, go but for one single witness at the present inquest. By such a method the number of testimonies must become greatly lessened; and if we take further into consideration, that we have of but a few languages, comparatively speaking, anything like a reliable grammar and dictionary, while the greater number is only known by scanty vocabularies hastily drawn up and even, if furnishing terms for 'I' and 'thou' little to be relied upon in so delicate a matter, you will not wonder at the moderate number of witnesses brought before you.

In Koelle's Polyglotta Africana we have specimens of nearly two hundred languages of Central, and partly of Southern, Africa. To judge by the diversity or coincidence of their vocabulary, especially of their numerals, they seem to constitute about 60 different families of speech, a number which will appear less enormous if I add that a good many of these families comprise but one isolated language, resembling the Basque in that respect. I will not however deny the possibility that some of these apparently different families may some day turn out to be relations but even allowing that possibility, we still retain a sufficient number of families, radically disconnected, to start from in this inquiry. It is to be regretted that Koelle does not give one single second person of any verb, but only the first; yet even under these circumstances the information is extremely valuable. I give here the pronoun 'I' of these languages only abridging the [p.36] list a little and leaving out the Berber, the Arabic, and the Kaffer, which will occupy our attention afterwards; the Roman figures indicating the different families.

I. Fulup . . . ni XII. Legba . . . man, ma
  Filham . . . ni   Kiamba .. .  me, ma
II. Bola . . . . gi, yi XIII. Koama . . . me
  Sarar . . gi   Bagbalan . . . . me
  Pepel . . . . ndsir   Kasm, Yula . . . a
  Kanyop . . . man XIV. Isiele . . . na
III. Biafada . . . -mu (personal termination).   Sobo. . . me, mi
      Ihewe . . . mi
  Padsade . . -nde XV. Okul Oma (not given)
IV. Baga . . . in XVI. Nupe . . . ne
  Tinme . . . i   Kupa Esitako . . . me
V. Bulom . . . a    
  Mampa . . . ya XVII. Bornu . . . wu
VI. Kisi . . . i XVIII. Pika . . . na
VII. Mandingo .. .  mbara   Ngodsin . . . na
  Kabunga . . . na XIX. Ekamtulufu . . . me
  Vei . . . na   Udom . . . me
  Soso . . . inbada XX. Basa . . . ma
  Bambara . . . mbara   Kamuku . . . ma
  Tene . . . . me XXI. Dsuku . . . mi
  Mende . . . nge XXII. Wolof . . . -ma personal termination.
  Mano . . . . ma &c.    
VIII. Dewoi . . . . na XXIII. Bidsogo . . . i
  Kra . . . . . na   another dialect . . . me
  Krebo . . . ne XXIV. Gadsaga . . . n personal prefix.
IX. Adampe . . . me    
  Anfue . . . . me XXV. Banyun . . . ma
  Dahome . . . n (personal prefix). XXVI. Nalu . . . mi
    XXVII. Bulanda . . . ni
X. Aku . . . mo XXVIII. Limba . . . y, ya personal prefix.
  Igala . . . . na    
XI. Mose . . . me XXIX. Landoma . . . i personal prefix.
  Guresa . . . mi    
  Dselana . . . ma XXX. Asante . . . me

[p.37]

XXXI. Barba . . . na XLVI. Nki . . . me
XXXII. Boko . . . ma XLVII. Kambali . . . mi
XXXIII. Timbuktu . . . di XLVIIIL. Alege, Penin, Bute . . . me
XXXIV. Mandara not given by Koelle (compare however muksa-nga 'ma femme' with muksa-rica (sic?) 'ta femme' in Klaproth sur la langue Bornou). LI. Murundo . . . na
XXXV. Bagrmi . . . ma LII. Undaza . . . ma
XXXVI. Housa . . . na LIII. Ndob . . . mi
XXXVII. Piilo . . . mi LIV. Nkele . . . ma
XXXVIII. Yala . . . me LV. Konguan . . . nasi
XXXIX. Anan . . . me LVI. Mbarlke . . . me
XL. Dsarawa . . .  me LVII.

Tiwi . . . me

XLI. Koro . . . me LVIII. Boritsu . . . mi
XLII. Ham . . . mi LIX. Afudu . . . m personal prefix.
XLIII. Akurakura . . m personal prefix. LX. Mfut . . . m personal prefix.
XLIV. Okam . . . kin LXI. Nso . . . n personal prefix.
XLV. Yasgua . . . ma LXII. Mbe . . . ma

Glancing over this list, we remark at once, that the nasal sound (mostly the labial or dental, sometimes the guttural) strongly predominates, so that MA, MI or NA, NI are almost everywhere the roots of the pronoun meaning 'I'. Sometimes a language seems to present no nasal, but then generally the kindred tongues, if there are any, show that a phonetic alteration has taken place. Thus in V, the Timne has i but the cognate Baga in, showing that the i has been curtailed of its final consonant. Again, sometimes a compound is used which has obscured a little the personal characteristic. Thus in VII, the Mandingo has mbara, the Soso mbada, but not only do the cognate languages show simply a nasal sound, as Mano ma, Vei na, but from Koelle's grammar of the latter language we learn moreover, that bere means 'self,' and that m-bere is a compound term = myself. Many other apparent exceptions from the general rule would no doubt vanish, if we knew only cognate dialects or even the whole grammar of the special language. Thus [p.38] wu in Bornu seems to want the usual typical letter, yet we shall see further on that there are strong reasons for supposing that w is a euphonic alteration of m. Dismissing however such mere possibilities, and looking only at the actual state of things, another characteristic than the nasal sound is really met with only in V, VI, XXVIII, XXXIII; for, with regard to II, XLIV, LIV, LV, the case is not quite clear.

We may therefore fairly assert, that in the Negro languages there is a strong tendency to use nasals for the first person. Nor is this tendency restricted to them; for we find the same thing also in the Basque ni, in Quasi-Qumuq (Caucasus) na, in the Georgian me (compare the cognate Lazian ma), in Japanese m, Korean nai, in Zamuca nu, Aymara na, Chiquita ni, the Mandan mi, and perhaps also in the Quichua noca. The most elementary form of the pronoun would be the Bushmen's mm 'I' (compare inng 'my'), as given by Lichtenstein, but it is rather strange, that the cognate Hottentot, which is now very well known, presents forms differing toto clo from the one just cited.

So far the similarity of human speech throughout the globe in the pronoun 'I' is undeniable; for, though there are exceptions, they do not annihilate the rule. It would not be the least wonderful, if the similarity vanished totally in the second person. The opposite is however the case; for even with our present scanty information we succeed in finding some salient traces of relationship in the principles on which the pronoun thou is formed. Namely:

Some languages use a nasal sound both in the first and second person, indicating the difference of the two either by a different vowel, or by employing nasals of different organs, or by both these means together. Such are:

In Africa: I. Tumali ngi 'I'     ngo 'thou'
                II. Wolof na          ngo
                III. Bornu wu        ni.

If we compare with this the possessive suffixes of the same language ni 'my', nem 'thy', it seems that a nasal [p.39] belongs by right to the first person, and that wu consequently stands for mu.

In Asia: IV. The BOTIYA FAMILY, especially in the Lohitic branch, as:

Burmese            nga 'I'      meng 'thou' (towards equals)
                                          men 'thou' (towards inferiors)
Dophlas             ngo          no
Abor-Miri          ngo          no
Mikir                  ne            nang

and also occasionally in the Transhimalayan branch, as in

Horpa                 gna           ni

V. MUNDA.
            Mundala Kol         ng             am
            Sontal Kol             inge          umge
            Sinbhum Kol         aing           um.

VI. DRAVIDIAN.
            Tamul                     nan            ni
            Canarese                nanu, na    nin. ninu, ni
            Telugu                     nenu         nivu
                cas. obl.               na             ni

VII. Kassia, generally, but wrongly, considered to be a member of the Thai family
                                            nga            me.

VIII. Chinese                       ngo            ni.

IX. Susian of Behistun
                -mi 'my'         -ni 'thy' (possessive suffixes according to Norris)
      comp. niku 'we'         ni 'thou' (accord, to Norris).

X. AUSTRALIAN
                ngai 'I'             ninna 'thou' (acc. to Schumann)
                ngatoa             ngintoa (New-South-Wales, acc. to Threlkeld).

In America:
XI. Aruncanese         ni         mi (pronominal adj.)
XII. Chiquito             ni         ni
XIII. Tarahumara     ne         mu
XIV. Othomi             ma 'my' ni 'thy'

[p.40]

XV. Dakota                 miye         niye
XVI. Kizh                     no ma      oma
        Netela                   no            ma
XVII. TSIHAILI-SELISH
        Shushwap             n                 an pronominal pref.
        Selish                    in                an  "                 "
XVIII. Sahaptin            in                im
XIX. Tshinuk                naika 'I'               maika 'thou'
        comp. also             ndaika 'we two'  mdaika 'you two'
                                       ntcaika 'we'        mcaika 'you'.

To these a few instances may be added of languages framing their two first personal pronouns in the manner just described, while the family to which they respectively belong follows in general other principles. Such are:

Languages of the Finnish stock
    Ugro-ostiac        ma 'I'         nen 'thou'
                                m               n (possessive affixes)
    Vogul                   am             nan

Finnish proper only in the suffixed possessive pronouns plural, as
                                kala-mme 'our fish'     kala-nne 'your fish'
                          but olemme 'we are'         olette 'you are'

Also the Turkish suffixed pronouns seem to follow the same principle; compare baba-m, my father, with baba-n, thy father. Further, in the Tonga language we have a similar opposition between

ma-ua, ma 'we two' and mo-ua, mo 'you two',

and the same is remarked in some other Polynesian dialects, though in general the Malay languages seem to follow a different principle.

Widely spread though this mode of marking the difference between the first and second persons is, it is far from being general. For a good many languages reserve the nasal for the first person only, while in the second they have one of the three tenues, p, k, t, or sometimes, though not often, a medial or aspirate. They may be conveniently subdivided into three minor classes according to their presenting in the [p.41] 2d person one or other of the three tenues. Thus we have, first:

Languages with a nasal in the first, and p in the second person.

In America: I. Moxa
            nu-ti, nu 'I'        pi-ti, pi 'thou'

(compare also the cognate Maipura language:
            nuja 'I'                piya 'thou').

II. The CORA FAMILY.
    Cora         neapue 'I'           apue 'thou'
                      ne                      pe (as signs of person before the verb)
    Tepeyuana ane                    ape
    Pima           ani                     api
    (comp. also Pima umua 'us', upua 'you').

In Africa: III. The FANTE-FAMILY, as
        Akra         mi         bo
        Akcambu  mi
        Yebu         mi         wo
        Fante       mi         ao
        Yoruba     mo         o.

These different forms are very interesting, because they low the gradual softening down of the hard labial sound of the second person into a mere vowel.

In Asia: IV. Kassia     nga         pha.

Secondly we have languages in which the character of the first person is a nasal, and that of the second the guttural, sometimes softened down a little. Thus we have in America:

I. The DELAWARE-FAMILY.
        Delaw. proper         ni 'I'         ki 'thou'
        Shawanese               nelah       kelah
        Mohegan                  neah        keah
        Penobscot                 nia          kia
        Algonkin                    nin         kin
        Chippeway                 nin         keen
        Knistenaux                 nitha      kitha

[p.42]

II. Maya         en 'I'        ech 'thou.'

In Africa:
III. Houssa     ni 'I'        ka msc. , 'thou'
                                      ki fem. 'thou'
                       mu 'we'   ku 'you'.

IV. The Kafir-languages. In these the original character seems to be a little obscured. For we find for instance in the Kafir properly so called

mi-na, I         we-na, thou,

and similar forms are met with also in the cognate languages. Hence Appleyard in his Grammar (p. 89) is perfectly right in saying, that w and u are the characteristics of the second person. The same author however immediately adds, that ku is used as a verbal prefix of the second person. The view, that k must be the characteristic letter of this person, is borne out by the possessive forms of the Suaheli, where we have 'niumba ya-ngo, my house', but 'niumba ya-ko, thy house' (the ya being merely the sign of relation). We find also in the same language expressions like ame-ni-penda 'he has loved me,' but ame-ku-penda-we 'he has loved thee,' the latter phrase presenting both k and w as characteristics of the second person. Thus it is sufficiently clear, that while a nasal always appears in the first person, we must acknowledge k to have been the sign of the second. Whether the w, being the other sign of that person, is radically different from the k, or merely developed out of it by gradual euphonic softening, I shall not venture to decide.

In Asia we find a guttural sound in the second person in the

V. Old Chinese ngo 'I' ghou 'thou'.

VI. BOTIYA in the Trans- and Sub-himalayan branches, as
            Tibetan     gnyo     khye
            Guruny     gna       ken
            Limbu       inga       khene
            &c.

VII. The Malay languages. The case of these is somewhat doubtful. The following forms however are in favour of their being arranged under this head;

[p.43]

        we             thou         you
            Tonga
        mau-tolu,  man             koe                    
                Newzealandic       ma-tou         koe           kou-tou.                            

Here the tolu of the Tonga = New-Zealandic tou is the numeral three, and the difference of the two pronominal bases consists chiefly in the m and k. There are however reasons for considering the k of the second person merely a euphonic change of t. Namely, the plural of the first person cited just now has the meaning of 'we' in the strict sense, that is, the person or persons addressed are not included. There is however another plural, meaning 'we, you included' which is in Tonga tau-tolu, tau, in New-Zeal, a-tou. Now it has been remarked that in other languages which have such a double plural of the first person, the inclusive plural is often derived from the second person, as for instance in Delaware, ni-luna, we exclusive, is formed from ni, I, but ki-luna, we inclusive, is formed from ki, thou. Assuming the same principle for the Polynesian forms, we should get a base ta for the pronoun of the second person. It must however be confessed, that nowhere in this family languages do we really find a singular pronoun 'thou' presenting a t-base: for the Kavi ki-ta, thou, which apparently furnishes a form of that sort (the ki being a mere prefix), is scarcely any thing but a plural used civilly for a singular, at least the Malayan ki-ta, and Tagalian qui-ta, mean only we (incl.), that is , thou and I. Whether therefore the t of the second person as usually found in the Malay and Polynesian languages be a euphonic change of t, or the latter be a different root, I shall not venture to decide. At all events the Malay languages have also some claim to be reckoned amongst those

Languages which have a nasal in the first, and a dental tenuis in the second. Such are

In America:
    I. Huasteca     nana 'I'     tata 'thou'
    II. Mexican     nehuatl    tehuatl
                            ni             ti (before the verb).
    III. Poconchi   in             at.

[p.44]

In Asia:
    IV. Jukagir     matak 'I'     tat 'thou'.
    V. The KORJAK-FAMILY, compare
                            I         thou         we         you
Korjak             gomma gyttsche     muyo     tuyo
Kamshadalian kymma kyse           buse (?)  suse
Tshuktshian     gim        gyr             muri       turi.

(It is clear that in the singular a particle ky, go, is prefixed. By stripping this off, we get the bases m and which are confirmed by the plural forms. The latter is softened down in some of the forms to tsh, and ultimately to s, r.)


The evidence now brought before the reader will I think be sufficient to establish the following conclusions:

1. There is a remarkable family likeness between most languages in the formation of the two first pronouns.
2. This likeness is however of such a nature that it does not necessitate, but rather excludes, the assumption of the historical unity of all these languages; whilst
3. It affords a strong proof of the psychological unity of mankind.

It ought not to be concealed that there are languages which present pronouns of the first and second persons not in accordance with any of the above classes. But we must not overrate the difficulty which their dissenting voice creates, because it is not at all unlikely, that many of the most glaring contradictions to the common systems would vanish, if we had but the power of discriminating between what are really old personal pronouns and those which are mere substitutes of civility, such as Monsieur in French, which often enough approaches to the real pronoun vous. Besides, a good many euphonic changes may have impaired the original forms which would come under our knowledge if we had older documents or kindred languages ready for comparison. Thus for instance the Dhimal, a dialect of the Lohitic branch of the Botiya class, offers ka for I, and na for thou, which seems contrary to all the rules laid down, [p.45] but if we compare only a few of the languages nearest of kin, as Mikir na, Burmese nya, we see at once, that the original nasal has passed by degrees through n, ng, into k. Similar euphonic changes may have occurred where from want of information they escape our knowledge. I consider therefore the main result of the previous survey of human speech safely established.

Before we proceed further, we have yet to see how far those languages which interest us most of all, namely the Finno-Tartaric, Semitic, and Indogerman, are in accordance with the laws just discovered.

What is usually called the Finno-Tartarian family of languages, are rather five families, the right of which to be considered of kindred origin is not yet fully established,at least, so great an authority as Boethlingk stoutly controverts the current opinion on that subject. I deem it therefore safest to treat separately each of the five families. They are:

I. The Tungusian including the Mandshu. In this latter have bi, I, gen. mini, plur. be, gen. meni, in which we clearly perceive the usual m changed into b in the nominative (as in Greek, [Greek], when compared with Sansk. mrta). Besides there is an inclusive plural mu-se. In the second person we have si, gen. sini. plur. sue, gen. sueni. The s is probably a euphonic change of t (compare the termination of the second person present in Tungusian proper, which is ndi), an assumption which would be certain if the relationship of the Tungusian with Finnish and Mongolian was more firmly established.

II. The Mongolian. Bi, I, min-u, of me, na-da (u and la are case-particles), to me. 'Thou' is tsi, plural ta. It is clear, that the sibilant in the singular is only developed y the influence of the i out of the pure dental tenuis remained unimpaired in the plural. Bi is of course for mi, and the n in the dative is equally put for m.

III. The Turkish. Ben, I, biz, we, in the Ottoman, are for men, miz, at least the possessive affixes are m, my, muz, our, and some Tartarian dialects actually give us men [p.46] for 'I'. With regard to the second person sen, thou, siz, we, it is probable that their s is a softened t, and for 'probable' we should say 'certain', if the affinity of the Turkish with the other so called Finno-Tartarian languages was established beyond doubt.

IV. The Samoyedic presents most clearly m in the first person; t in the second. Compare
                                I                 thou
        Ostjak             man             tan
        Tawgyan         mannang     tannaug
        Kamassian      man             than
        Jenissei           modi            todi.

V. The Finnish. Here also m and t, the two characteristics, stand out in full relief. Compare
                                    I             thou             we             you
        Finnish proper    mina       sina             me             te
        Lappic                 mon        ton              miye, mi    tiye, ti
        Syrjanic               me          te                 mi             ti
        Hungarian           en            te                mi             ti.

The sina of the Suomi itself is of course a softened form caused by the following i (compare Suomi isa, father, with Hungarian atya), but the t is retained in the plural and the verbal termination, as ole-t, thou art, rakasta-t, thou lovest. In Hungarian we have the m of the first person in the plural, and also in the possessive suffixes and verbal terminations of the singular, as hala-m, my fish, lato-m, I see (it). According to analogy we should expect in the isolated form a singular mi; the en which we find instead is very puzzling, not on account of its n, which might easily have been substituted for m, but on account of the long vowel at the beginning, as no cognate language knows of any similar form, except perhaps the Vogul, in which, according to Klaproth, they say am for I.

To sum up, it has been shown, that the five families generally called by one name, the Finno-Tartarian, all present m as the characteristic of the first person, while for the second we find t distinctly in three of them, and [p.47] even the remaining two have s, apparently a euphonic alteration of the same letter.

In turning to the Semitic family, I must remark that this family stands in a decided relationship with the Coptic, the Berber, the SahoGalla (these two latter languages being members of the same family), and perhaps even with some other African families of speech. The form of these three sister-families, as I would call them, of the Semitic proper, do not fail to throw considerable light on the Semitic itself, in this as well as in other matters. In all these languages here is at first sight a considerable difference between the isolated pronouns and the possessive affixes, the latter being he more simple of the two. The suffixed pronouns of the second person are:

                                    sing. msc. fem. plur. msc. fem,
                Egyptian                k         t                 ten
                Coptic                    k     - (ti, i)             ten
                Berber                    k         m       kun        kunt
Semitic    Arabic                   ka        ki        kum        kunna
proper     Hebrew                  ka        ki        kem         ken

From this, there would result a k as the characteristic of the second person, but the Egyptian forms seem to authorize a belief, that either there was another base t, or else that one of the two is a euphonic change of the other. The latter seems to be the more likely. At all events both the verbal terrminations or prefixes and the isolated pronouns of all these languages show decidedly a t-sound as the character of this person.

Compare the terminations of the so called Semitic Perfect

                        sing. 2. msc.     fem.     pl. 2. msc.     fem.
Arabic                     ta                 ti             tum         tunna
Hebrew                   ta                 t (ti)         tem         ten
with the isolated pronouns
Arabic                     an-ta         an-ti         an-tuni     an-tunna
Hebrew                    at-ta         at-t(at-ti)   at-tem     at-tem

[p.48]

Here we find t most decidedly as the character of the second person, the difference between the terminations and the isolated pronouns consisting merely in the latter having the syllable an prefixed, the n of which is assimilated in Hebrew. In Egyptian and Coptic we have

                    2. sing.         masc.         fem.     pl. comm.
    Egyptian                     ento-k         ento         ento-ten
    Coptic                         entho-k       entho       entho-ten

The k in sing. msc. and the ten in the pl. are clearly the pronominal suffixes, and en is the prefix corresponding to the Arabian an. Thus there remains TO as the really significant part of the word. The prefix en, or rather an, which is not sufficiently clear in the Semitic proper, finds its explanation in the Coptic. For there an is the pronoun indefinite, which is however used as a sort of article before some words, to give them a more substantial, or rather substantive, character; as an ζ (ζ is the Greek numeral) 'a seven, a week.' Sometimes it seems to impart a collective: character, as tou, a mountain, an-tou, [Greek]. Again, it takes the form en in shashi, bitter, en-shashi, something bitter, bitterness.

While we recognize easily in the Semitic and its cognate languages a T (or k) as the character of the second person, the question, what may be the proper characteristic of the first, is fraught with considerable difficulty. The isolated pronouns are:

                                                            Semitic proper
             Coptic     Berber     Galla     Arabic     Hebrew
        (I)  an-ok     nakki         ani         ana           anokhi, ani
      (we) an-on     nakni         nu         naχ-nu      anaχnu (naχnu)

The possessive affixes are:

                                                            Semitic proper
            Coptic     Berber     Galla      Arabic     Hebrew
    (my)     i             yu             ko              i             i
    (our)     n          nagh         kefa             na         nti

It is clear, that we have in the isolated forms the prefix an just mentioned, and the final i in the singular, the final [p.49] ni. nu in the plural are probably nothing but the possessive suffixes superadded as in the case of the Egyptian second person mentioned above. Thus we would retain the syllable AK, OK as the true characteristic of the first person, and must assume that in those forms where the K is apparently wanting, it has been lost. In the suffixed pronouns however only the Galla shows a k, while all the others have i, y in the singular, n in the plural. It is impossible to explain the latter as an alteration of k and it would be equally impossible to maintain that the ni nu were shortened forms of the isolated pronouns, because, as we have seen, the latter themselves suppose the pre-existence of such a suffix-form. Thus we are evidently driven to the conclusion, that N was, at least legitimately and originally, the characteristic, or one of the characteristics, of the first person plural. Now there is in addition one instance at least of the same in the singular: namely, both in Hebrew and Arabic the accusative suffix meaning 'me' is (not i, but) ni.

Therefore finally, though not without some hesitation, I set down the Semitic and its cognate tongues amongst those languages 'which present the nasal in the first, and T in the second, person'.


Comes lastly our own Indogerman family. Here we must look, not to the nominative but to the accusative, in order find the common type again. And on doing so it appears clearly enough in:

            Sanskr.     Greek     Latin     Gothic     Slavonic     Lithuanian
(me)     ma-m         υε           me         mi-k         me                 man (dative)
(thee)    tva-m        τε, σε     te           thu-k         te                  tav   (dative)

However, on more minute enquiry, some difficulties present themselves. First, we have in the nom. of the first person early a different series in Sanskr. aham. Greek [Greek], Goth. ik, &c. which necessitate the supposition of an original agham; secondly, from the Sanskrit it is clear, that the base of the second person is not properly speaking ta, but tua, and this is also borne out by the nominative form of all [p.50] the other dialects: Gr. [Greek], Lat., Lith. tu, Slav. ty, Goth. thu; the u of which is evidently a contraction of VA, as shown by the Sanskr. nom. tvam. A still greater difficulty arises from the circumstance that both the M and the T are only found in the singular. There are indeed in the first person plural and dual, forms beginning with n (Skr. nas, us, nau, us two, Gr. [Greek], Lat. nos, Sl. ny, us, &c.), and sometimes even m (Lith. mes, we, = Slav, my), but they are rather exceptions than the rule, and not totally free from the suspicion of having originated from erroneous analogy. In the second person moreover, there is scarcely a trace of a t in the dual and plural, unless it be in the Greek [Greek], and Welsh chwi, which both imply an original initial SV, that might possibly stand for TV.

I know that several attempts have been made to show that not only the bases employed by the Indogermans in the plural and dual, but even the nominative singular of the first person, descend from the same roots MA and TVA, but I confess that I lack the desperate courage needed for such identifications, and, whatever truth or untruth may be in these jeux d'esprit, one thing is certain, that even before the separation of the tribes, the nominative singular of the first, and the plural and dual of both the first and second, presented the same want of harmony with the rest of the forms, as for instance they do in the actual Sanskrit historically known. What this disharmony arose from, and whether it actually was developed out of original unison, it is impossible to say, but nevertheless we can affirm, that in a still older period M was the only characteristic of the first, and T (not TV) the only sign of the second. This I derive from the personal terminations of the verb, which are in the present

                    1. sing.     l. pl.     2. s.     2. d.      2. pi.
        Sanskrit  mi         mas         si         thas     tha
        Greek  [Greek]
        Gothic   (m)         m             s         ts         th
        Latin      (m)         mus         s                     tis
&c.

[p.51]

In the first person the M is apparent, and so is the T of the second in the dual and plural. The si of the singular is no doubt a euphonic alteration, while the Skr. dhi, Gr. [Greek] in the imperative, presents another form of the dental in that person. The second sing, of the reduplicated preterite has retained the original sound, Skr. -tha, Goth, -t, Lat. (is-)ti, for the th of the Sanskrit is here, as nearly always, very modern.

Thus we discern two different periods in the prehistorical Indogerman times; first, a period when M and T were the characteristics of the two persons, secondly, a period when the former became restricted to the oblique cases of the singular, while the latter in its new shape TVA was also confined to the singular, extending however its dominion to the nominative. This second period had been gone through before our ancestors separated. In the historical times the difference of singular and plural, and of the nominative 'I' as opposed to the oblique case me is upon the whole, preserved, but here and there we observe a tendency to come back to the pre-primitiveif I may say sosimplicity. This tendency to increase the domain of the M and T has three different modes of appearance, namely first: the M enters the nominative case. This we have in some of the English dialects which replace I by me, in the moi of the French, in the men of the New-Persian, which is the only form in that language, and even from the oldest times historically known in the me of the Old Irish and the other Keltic languages. As it is very likely that the Etruscan is one of the Indogerman family, the mi of that language is probably the oldest instance of the M usurping the nominative in an Indogerman language. Secondly we find the m of the first person extended to the plural. This we have in some German and Norse dialects (mer, mir for wir is common about Thuringia, mer is also sometimes used in Old Norse), in the New-Iranian languages almost throughout (we is in New-Persian ma, Armenian meq, Ossetian maχ), and in modern Greek [Greek]. I am aware, that in most of these languages there are euphonic laws at work, which [p.52] may partly explain the origin of those forms, but I think that the above-proven tendency towards uniformity goes for something. It is very true, for instance, that the Romaic [Greek] is [Greek] shortened in the same way as [Greek] for [Greek], [Greek] for [Greek], but I think the apheresis was favoured at least by the circumstance that they got by it the wished for oneness of sound. The third phase of this retrograde movement towards simplicity we have, when the dental of the second is introduced into the plural. This is the case partly already in the Old Norse, where p-id, both of you, p-er, you, are almost the only forms ever used, though the cognate Swedish and Danish retain the form without a dental even to this very day. On the Faero-islands they introduce the dental even into the oblique cases, and say tiur, you (nom.), tiara, of you, tiun, to you.Da for 'ihr' is used in some of the middle parts of Germany, [Greek]; is in Romaic the legitimate plural of [Greek], and in Armenian we have du-q, you, corresponding to du, thou. Furthest however in this tendency of simplifying the complicate variety of the old Indogerman languages go the modern dialects of Hindustan. For instance, in Bengali we have:

                                        I         we         thou         you
Pronouns of equality     ami     amra      turni     tomara
Pronouns of inferiority
orin the second person
of superiority             mui     mora       tui         tora.

Thus the Bengali has completed the circular movement, and ultimately come back to the simple state which we have reason to suppose that our common Indogerman ancestors once started from.

To sum up shortly, we have up to this point seen, that similar phonetical principles are observable in the formation of the two first pronouns, as well in the Indogerman, the Semitic, and the Finno-Tartaric languages, as in other families of speech, thus confirming the inward unity of all mankind, though not at all attesting community of origin.

There remains now the more difficult enquiry, by what psychological process these words are developed.


[p.53]

PART II.

In a mere metaphysical point of view there seems to exist a vast gulph between the Me, the individual consciousness, on the one hand, and the It, the outward world of objects, on the other. If language reflected this bipartition cherished by a good many philosophers, we should expect to meet at the best with a pronoun of the first, and a pronoun of the third, person. However language, following therein the example of nature, or rather a law of existence according to which the individual mind enters into possession of its own self only through the contact of, and in community with, other individual minds, language introduces between the Me and the far off situated It or That an intermediate link, a pronoun which is both me and not me, in short, the word Thou. It is not my intention to dive at present into all the depths which are contained in this little word, (a subject which W. v. Humboldt has fully developed in his essay on the dual number), on the contrary, I content myself with the indisputable assertion, that whatever the actual nature of the Thou may be, it cannot be overlooked, that in a mere abstract metaphysical point of view, it is but one of the many cases of non-ego, and that therefore it is not altogether unreasonable to expect, that language should treat it as such, in other words that the pronoun of the second person should somehow be a variety strongly marked indeed by individual characteristics of the pronoun of the third person. If we now turn to the real languages to see how far they justify such an expectation, it is first worthy of being remarked, that in only one single language have I found a form for thou which may be clearly analysed to mean, and even in the present historical state of society must be felt by the speakers to mean, 'thy it', tuum illud. This is the Jurako-Samojedic pudar, which is a compound of puda, he, she, it, and r, the suffixed pronoun of the second person singular. But not only does the said lan- [p.54] guage in this respect differ from its nearest of kin, the other Samojedic dialects, but even this very pudar 'thy it' would be impossible without the previous existence of the suffix of the second person r. Even here then the subsumption of the second person under the third is, to say the least, incomplete, if not altogether delusive.

Nevertheless on earnest search we remark in several of those families of languages the records of which are before us nearly complete, an indisputable similarity of form in the pronoun of the second person and certain pronouns of the third, or (as for the future I shall call them) demonstratives. That I am justified in substituting the latter denomination for the former, will I hope scarcely be denied by any one who is acquainted with any of the more ancient forms of Indo-European speech, the mere knowledge of the way in which the Latin uses is, ille, iste, being amply sufficient to establish the identity of the demonstrative and the so called pronoun of the third person. The similarity of form between the second person and the demonstrative is evident in the following families of languages:

1) INDOGERMAN. It has been shown, that the tva,which is in all the Indogerman languages historically known the base of the second person, and was so already before the scission of languages,that this tva must nevertheless have been in a still more early period ta. Ta again is the well known base of the Skr. tam, eum, Gr. [Greek] , Slav. tu, Lith. ta, Goth, thana, O.H.G. den, &c. Thus the identity of form is complete.

2) SEMITIC FAMILY, together with its sister-families. In these we have seen before the base of the second person is either tha (tho) or ta, to which the syllable an, Egyptian an, en, is prefixed (above, p. 48). The same tha, ta, with the same prefix en is apparent in Egyptian en-to-f, he, en-to-s, she, Koptic entho-f, entho-s, plur. entho-u, they, where the f, s, u are the possessive affixes of the same persons. Compare also the Koptic ti, tal, hie, illic. The Berber has natta, he, natta-f, she, nufni, ii, nufanti, eae, in which the same prefix with the vowel inverted (na, nu) is easily recog- [p.55] nized, leaving as the base ta, fa1. In the suffixed pronouns the Berber has preserved the t-sound in the accusative form (-f, him, -t, her, -fan, them), while it is changed into s in the dative forms -s, him, her, -san, -sant. The same change into s has evidently taken place in the Egyptian suffixes -s, her, hers, -sen, their (the latter replaced by n in Koptic, that is, by the mere termination of the plural occupying the place of the lost suffix). The t, th as the elementary sound of the pronoun of the third person seems at first entirely wanting in the Semitic proper, but if we remember the essential identity of the pronoun 'he' with the demonstrative bases, we shall not hesitate to recognize it in the common pronoun demonstrative, Arabic da,2 or ta, hic, di or ti, haec, Chaldean den, hic, da, haec, Hebrew zeh, fem. zot, compare also the Arabic tamma, hic, hoc loco = Chald. tam. Hebr. sham. The common forms of the pronoun of the third person in Semitic proper begin always with A, and so do the suffixes, only in Assyrian we find a sh (according to Oppert, according to others it is s) at the beginning of the suffixes -shu, his, -sha, hers, -shun, eorum, shin, earum. Whether this sibilant may be considered as an intermediate stage between the t, th, of the Egyptian and Berber, and the common Semitic h. I will not venture to decide, because this euphonic change of s into h is very rare, if at all known, in this family, but the existence of a demonstrative base da, ta, nearly identical in form with the base of the second person even in Semitic proper, is established.

3) In the FINNISH languages we again found t as the characteristic of the second person, and the t as a demonstrative base is sufficiently indicated by the Finnish tuo, ille = Estonian to, Mordvinian te, hic, tona, ille, Lapponian tat, ille, pl. tah, Tsheremissian ta, hic. In this family there is also a demonstrative base S, which most frequently furnishes the pronoun of the third person properly so called, Tsheremissian [p.56] so, he = Mordv., Lappon. son, which becomes han in Suomi, the s being very often liable to this change in that language. Whether this S be originally only a modification of the T, others may decide. For our present purpose, the fact that t is here also the characteristic of both the second person and the demonstrative, is sufficiently established.

4) THE SAMOJEDIC. As to the t in the second person, see above. The demonstrative t we have in Jurakian tuky, hic, tiky, iste, taky, ille, in which Castren considers ky a particle. Add the Ostiakian tam, tau, tap, hic, to, iste, Kamassian di, is, du, hic. In Ostiakian also 'he' is tep, tap, in Kamassian du, di.

5) MONGOLIAN. The t of the second person finds an echo in, or rather is the echo of, the demonstrative of the Buriatian dialect tere, ille, where re must be a suffix, as it is cast off in the plural te-de, the de being the plural particle.

6) How far we were right in ascribing a t as the original character of the second person to the MANDSHU language, depends in a manner on the general view we take regarding the relationship of this family to the other Finno-Tartarian languages; a question not to be decided in this place. Assuming however that we have done right in claiming the said character of the second person for it, we may further remark, that in this language tshe (for tye, te?) means 'they', and that the Tungusian proper shows the t-sound most clearly in its suffix of the third person plural -tin.

7) The certainty of the t as the characteristic of the second personal pronoun in the TURKISH languages depends on the previous settlement of the same question which we raised concerning the Mandshu. The demonstrative t seems to make its appearance in the demonstrative tigi.3

To sum up,the base of the pronoun of the second person is identical with a demonstrative base, or nearly so, in Indogerman, Semitic and its sister-dialects, Finnish, Samo- [p.57] yedic, Mongolian and perhaps Mandshu and Turkish. I forbear to extend the enquiry over other families of languages, for the simple reason that the difficulty attending it forbids a single man to enter into labours affording ample scope for, and necessitating the aid of, a great many scholars, but which would never come to an end were I to do it all alone. Thus then the question arises, whether the identity of form between thou and that, du and das, can possibly be the result of mere chance, and as this can scarcely be admitted, the similarity becoming only more striking the farther we go back to antiquity, the more carefully we analyze the facts, what must we think of this apparently undeniable identity?

It is now time to introduce the interesting and truly philosophical remarks put forth by W. v. Humboldt in his essay 'Ueber den Zusanunenhang der Ortsadverbia mit den Personalpronominibus.' The main substance of this essay is the proposition: "Some languages have constituted a relation between the notion of space and the pronominal notions''.

Such are, first, the Tonga. In this language mei signifies the motion towards the speaker, atu the motion from the speaker to the person spoken to, lastly angi from the speaker to a person at present not spoken to. Now these three adverbs are very often put, while the pronouns or verbs are left out. Especially the verb 'to give' is almost invariably suppressed, nay seems to have been lost altogether. Thus they say:

mei ia giate au = hither this to me give me this,
                                  teu atu ia giate coy = shall I thither this to thee = I shall give you this.

In these examples the pronouns (au, I, coy, thou) are preserved, but they may be dispensed with in such sentences, as:

bea behe mei he tunga fatine = when spoke hither the several women = when several women spoke to me or us;
taila, 'to tell,' tala mei 'to tell hither' th. i. me or us, tala tu 'to tell thither' th. i. thee or you.

Secondly, Old Chinese. In this language nai means 'thou,' [p.58] but according to Neumann it indicates originally a relation in space. The same scholar adduces places, in which nai really means 'ille', as nai yan 'illud verbum', nai tong 'like him'.

Thirdly. In Japanese the words kono, sono, ano = kore, sore, are, mean hic, iste, ille, and konata, sonata, anata are the corresponding answers to the question where? that is, they mean: here, there (istic), yonder, respectively. But the two first are raised to the signification 'from my, thy, side, as to me, thee'. Sonata originally 'istic' has become quite a pronoun of the second 'thou', and at present is strictly different from its etymon sono 'iste'; konata 'here' seems to be employed as a pronoun of the first person only in the limited sense 'as to me', and is used besides as a pronoun of the second person addressed to superiors.

Fourthly. In Armenian there is a suffix s, which originally means 'this' or 'here', but takes the meaning of I and my. Thus gam-s, I am here, or I am now, and hair-s, this father, the father here, I a father, my father.

Thus far the resume of Humboldt's paper, which amply shows, that demonstrative adverbs like there, here, may become the substitutes of the pronouns thou and I. One thing is remarkable in this excellent essay, namely, that Humboldt while bringing forward out of the storehouse of his immense learning examples taken from the most remote languages, should have overlooked an instance of the same mode of speaking which might have been had almost at home; I mean the Italian d and vi substituted for 'us' and 'you' in the accusative case, though or because they originally mean 'here' and 'there'.

"But"I hear already some sagacious reader ask"what have we to do with all these Humboldtian remarks here? Let it be true, that words meaning there may be used for 'thou', but will that justify the supposition, that the mere pronoun 'iste, ille' can be also used for 'thou'? There is a vast difference between a pronoun demonstrative and a demonstrative adverb." Here I must remark that I by no means meant to say, that the pronoun 'this' or 'that' might [p.59] give rise to 'their,' but only, that the demonstrative base became at once the have of the pronoun of the second person, which is vastly different too. For the stage of speech here brought under consideration lies so far backward in the development of the mind that we are carried by it into a time when there were only roots. Any step further depends therefore on the preliminary condition that we should view the nature of a root in the right light. First of all. in a more highly developed state of language such as is presented by Sanskrit, Greek or Latin, the root is nothing but the complex of sounds, from which according to grammatical laws a number of words originate, but it is no word itself. Hence it is neither noun nor verb, though both are developed out of it. If I take the words [Greek],being verbs, substantives, adjectives, adverbs,they may be compared to the different parts of a plant rising as branches into the air, stretching as roots into the ground, quivering as leaves in the wind, shining as blossoms in the sun, while the radix [Greek] is only if I may so the point of indifference between them, that which is neither branch nor root nor flower, nor fruit nor leaf, but the spring of life in all of them.4 This power of life spread in the fully developed tree all over its different parts is contained before the development begins in the small compass of the germ, and to this phase of vegetable existence the so called radix may be compared in those languages which like the Chinese are devoid of grammatical forms, and wherever else the grammatical forms are yet to come. This applies especially to the way in which children use language when they begin to speak. Then, every word of theirs is a root, or rather a germ, which in time may become a family of words branching off into nouns and verbs &c.; but as yet it is nothing of all that. If a child says 'pa, ma' it is only our adult reason misled by our own way of speaking which starts the question, whether that be a noun or verb. It is neither. [p.60] Pa is every thing which the father does, is, and has, a teeming conception brimful still of the original riches of overwhelming sensations. Moreover, if the power of desire is uppermost in the infant soul, then 'pa' is not only every thing which the father does and is, but also every thing he is wanted to do or to become. It requires some degree of intellectual labour to realize in a manner this psychological state and a good deal of skill to describe it, and I doubt myself whether I have succeeded in the attempt at doing so. However, unless we get some faint idea of it, the problem of what a root or the germ of a family of words is, must remain a mystery for ever.

If we now come to demonstrative roots, it is clear, that they also must be considered in that undeveloped state as the germ of all sorts of pronominal forms, adjective, substantive, adverbial, not as the reality of one peculiar class of them. The correctness of this view is borne out by the fact, that in languages that are totally foreign to our highly organized grammatical system, the pronouns have changing powers, which, though they excite astonishment, and even a sort of dismay, in the souls of Indogerman grammarians, are perfectly natural in their stage of development. Thus the more attentive of my readers have no doubt already remarked, that the Chinese nai (as given above in the passage referring to Humboldt's essay) means 'ille, illius, illic' that is, is at once a possessive, a substantive, and an adverbial pronoun. Another interesting feature of the demonstratives in their germlike state is exemplified by Threlkeld's grammar of the languages of Australia. There we hear of an 'absolute pronoun ta, it is. Not merely declarative, but absolute. It is derived (?) from the verb substantive, verb assertive'. [Rather: it is also the verb substantive.] Again, when mentioning its plural form ta-ra, he says: ta-ra, they 'are, the things, the plural of 'it is', these, those'. These are startling and shocking facts for our Indogerman ears. But still more startling is it, that in the Negro-English the actual Indogerman forms are employed in a manner approaching nearly to the queer description which Threlkeld [p.61] gives of the corresponding Australian demonstrative. Namely, the English there becomes in Negro-English de, which besides the sense of the English word has also the meaning of the substantive verb as: mi de 'I am', literally 'I there,' na hoso, at home, de na hoso 'he is at home.' Again, that becomes da, but with a slightly altered pronunciation we get da, which takes the place of the verb substantive, as mi da joe brara = me there you[r] brother = I am your brother.5

From all this we may learn that the original demonstrative is neither a substantive nor an adverb, but all of this, and even a verb; or rather, that it is a whole embryonic sentence, which according to circumstances may be differently translated into our fully developed language. The child sees an object and says ta! I do not insist of course on this very sound being invariably produced; we may translate this by 'there (it is)' or 'that it is' or 'carry me thither' or 'give me it' , and by a variety of expressions besides, but the truth is, that every one of these interpretations is wrong, because it replaces the teeming fullness of the infantine word by a clearer but less rich expression of our more abstract language. Yet if a choice between the different translations must be made, I trust that few of my readers will refuse me their consent when saying: 'there' the adverb is by far the most adequate. So much then for the nature of the original demonstrative, which we see is such, that the difficulty of its being turned to a pronoun of the second person without the previous formation of an adverb of place, that this difficulty vanishes, because the original demonstrative partakes itself of the nature of a local adverb.

Consequently, as the historical facts forcibly pointed to the formal identity of the pronoun of the second person [p.62] with a demonstrative, we may venture to consider this identity as established, and therefore say, that the Thou is really developed out of the 'It or There'.

But Humboldt's remarks as given above carry us still farther. For from them it resulted, that 'I' also was in several languages expressed by a demonstrative adverb, originally meaning 'here,' and consequently we may expect, that the pronoun 'I, me,' may have a similar origin to that of 'thee, thou', a view which is supported by the phonetical similarity of the two first pronouns observable in so many languages, as shown before. At present however I am able to give something like a proof of this supposition only for the Indo-European languages. Leaving the two bases of the nom. sing. and plural out of the account, the pronominal forms of the first person in our family are reducible to three main bases: 1) MA, the base of all the cases of the singular, except the nominative, and which in a pre-primitive time even extended over the plural, as the verbal terminations show (s. p. 50.) 2) NA in several of the dual and plural forms (Skr. nas, nau, Lat. nos, etc.). 3) A in all the oblique cases of the plural, where it is however always compound with SMA, thus giving rise to ASMA (Sanskr. asman, Gr. [Greeek]). All of these three are also demonstrative bases: MA is found as such in the Sanskr. compound ima- 'hic', not used in the nominative, and isolated in the Greek [Greek]. NA is found in the Greek [Greek] and in a good many compounds such as Sanskr. ana-, whence anena 'by him' = Lith. an(a)s, Slav. onu, and Sanskr. ena, Old Lat. oinos i.e. unus, Goth, ains, &c. Lastly, A is a well known and nearly fully declined demonstrative base in Sanskrit, of which several forms almost coincide with the pronoun 'we'; thus a-sma-t 'ab eo' and a-sma-t 'a nobis' are only different in the quantity of the last vowel sound.6

[p.63]

Thus we come to the result, that, out of the original demonstratives, a base pointing more to the distance has been taken to signify the pronoun of the second, while another base referring to a spot near at hand has acquired the power of I. (This distinction of nearness and distance entirely rests I must confess on Humboldt's essay. It would be impossible show with our present means, that in Indogerman for instance the demonstrative base ta, which has given rise to the pronoun 'thou,' points more to distance than the other MA, NA, A.)

I could conclude with this, but before doing so, I shall make myself my own critic, an office which I have exercised here and there even in the body of the Paper.

I refer the two first pronouns to demonstrative bases, but it might be justly said, that although I have spoken a good deal about the characteristic nature of the demonstrative, I have not explained their origin. To this I might answer, that such at present was not the subject of my enquiry, but prefer to declare that in the main I agree with Schwartze, who says on p. 339 of his Coptic grammar with regard to he original demonstrative: "Every object is to the child a living palpable thing. When it cannot reach anywhere with its hand, then instinctively it utters a cry, in order to cause to approach that which has awakened its interest." I add, when the soul becoming aware of this cry issuing forth from its own interior, takes it up as a sign for the indefinite outward reality which is the object of its desire, and shapes it into an articulate sound, then we have a pronoun demonstrative in the sense explained above.

Secondly it may be said, that the view taken here of the two first pronouns is materialistic, in as much as they are opposed to be nothing but demonstratives, by which the spiritual nature of Me and Thou could not be expressed. To this I answer, that I do not mean to say that I and Thou are not materially different from the outward world of That and There, but simply, that when the mental de- [p.64] velopment has reached the point where the consciousness of a person (a Thou), and of the individual mind's own personality (the Me) is aroused, then the matter for shaping the requisite words is taken from the great storehouse of the Here and There, of the demonstratives. Why the mind ever should come to the feeling of personality, either another person's or its own, this, the deepest of all psychological and metaphysical mysteries, remains of course unsolved by the above observations.

Lastly, it will have been remarked that I always put the primitive historical state of language in parallelism with the state in which we meet with language in the mouth of infants. I wish it to be understood that I do so, not from any confusion of ideas, but on principle, because I hold that the last difficulties attending the question of the creation of language can only, and only, be elucidated by psychological enquiries to be made at present, since the present repeated creation of language by children differs only in so far from the primitive one, that the first manufacturers of language had to do both, to go through the psychological development, and to invent, or rather to shape, the articulate sounds, while our children find the latter ready coined, the moment their mind has entered such or such a state of development in which a certain word or class of words is requisite. But this does not constitute any material difference.

In order therefore to throw further light on the matter of the development of the two first pronouns out of demonstratives, new psychological observations are very much needed; especially it would be of paramount importance to learn, whether children begin first to use the word thou (you) or I, or whether the two make their appearance at the same time. The second supposition seems very improbable, the third I presume will be found in accordance with the facts. At least Humboldt's essay tended to show that there is always a reciprocity between the Me and Thou, the one emerging from the Here, the other from the There.

[p.65]

P.S. Since I finished this Paper. I have found another language fully in accordance with the views developed in it, which at the same time furnishes a striking instance how much the real character of the personal pronouns may be obscured by peculiarities of grammatical structure, so that it requires nothing less than a thorough acquaintance with the anatomy of the language concerned, in order to extricate the true characteristic out of its veils and encumbrances. He who hears for the first time that in Greenlandish uvanga means 'I', and ivdlit 'thou', will probably think this language a glaring exception to all the rules laid down in this Paper. And such was my impression too. However, from Kleinschmidt's Greenlandish grammar, it appears that the Greenlandish chimes in with these very rules most beautifully.

Namely -ga, my, -t, thy, are in Greenlandish the possessive suffixes of nouns, as igdlu-ga, my house, igdlu-t, thy house. Here we have t in the second person; and that ga in the first is developed out of a nasal appears from the fact that in the verb the objective suffix of the first person is invariably nga (= me). It will further be seen by any one who takes the trouble of examining carefully all the various forms of the Greenlandish suffixes in the singular, dual, plural, subjective, objective, possessive, that, throughout the whole, G, NG, is the unmistakable characteristic of the first, and T (sometimes it seems changed into S) of the second.

But what then of the queer forms of the isolated pronouns? This mystery has been solved with much skill by Kleinschmidt himself. The forms are as follows:

uta-nga. I                     itdli-t, thou
vra-guk, both of us      ili-vilik, both of you (k denotes the dual)
uta-gut, we                   ili-vse, you (t denotes the plural)

Now t, vtik, vse are the common possessive affixes of the second person singular, dual, plural, and nga, guk, gut, the common accusative affixes of the first person added to verbs. I agree with Kleinschmidt that the latter three are probably here somewhat irregularly used as possessives. It [p.66] is clear then that the real personal character is only in these suffixes, and that the remaining bases UVA and ILE are merely employed for the purpose of giving them a support, since the Greenlandish is averse to the separate use of pronominal bases and always wants to affix them to something. Kleinschmidt's further assumption, that the said two bases UVA, ILE, are identical with the two demonstrative roots UV, here, IK (IV), yonder, is highly probable as to the latter, and as to the former I consider it certain. Consequently the learned missionary is right in saying that uvanga, ivdlit mean originally "meine hierheit, deine dortheit", "my here-ness, thy there-ness".

It is already remarkable that in these Greenlandish absolute personal pronouns we should again trace a connection between me and here, thee and there, but still more remarkable is it, that the NG, T, the characteristics of the first and second person find an echo in, or are the echo of, two pronominal bases simply demonstrative. Namely on p. 21 of Kleinschmidt's grammar we find a demonstrative root MA, here, and another TASS, there. Besides there is an enclitic demonstrative base ta, which may be prefixed to all the other demonstrative bases, except Tass, and which I because of this very exception consider with Kleinschmidt as the original root of that same tass. Hence it would appear that exactly as the root MA 'here' is to TA 'there', so is NG, the root of the first person, to T, the root of the second; or, according to my view, the two latter at the two former transformed into personal pronouns by the process described in my Paper.


NOTES

1 The i at the end of the fem. sing., the ti at the end of the fem. plur. are the characteristics of that gender.

2 đ represents nearly the English th in the.

3 Having no Turkish grammar at hand, I repeat here Castren's statement in his Samoyedic grammar.

4 It may be seen from this that the grammatical expression 'root' is unfortunately grounded on an inadequate simile.

5 Traces of a use of the demonstrative similar to the one described a the text, we have even in Indogerman languages. For instance, the German da! when accompanying the act of giving, and the Greek [Greek] Sophron used even the plural [Greek]) resemble somewhat the Negro mode of speaking.

6 It is curious, that MA, NA, A are at the same time also the buses of the negations (comp. Skr. ma, Gr. [Greek], Skr. na 'not' = Greek [Greek], Lat. n Goth. t, and the privative a of both Greek and Sanskrit. This is again nothing but another development of their original demonstrative force of pointing to some object distant in space and consequently removed from the thoughts of the Me.