P H I L O L O G I C A L SOCIETY.

VOL. 111. MARCH 26, 1847. No. 57.

GEORGESLOANE,
Esq. in the Chair.
Miss Anne Gurney, of NorthReppe, Norfolk, was elected a Member
of the Society.
A communication was then read :-
.
“ O n the Origin of the Demonstrative Pronouns, the Definite
Article, the Pronouns of the Third Person, the Relative, and the
Interrogative.” By T. Hewitt Key, Esq.
That the Greek interrogatives beginning with a T , as norepos, owe
their difference from the Ionic forms with an initial K , such as Koreppos,
solely to that variety of pronunciation which is called dialect, and
that they are substantially the same words, is commonly admitted,
and indeed is supported by the fact that to the Roman quicquid cor-
responded an Oscan pitpit. Again, it may also be assumed that
Grimm and others are right in treating quis and 11s as equivalent in
form as well as meaning. T h e changes in these words have been
repeatedly compared to those which have also taken place in T E T W * ,
~ U U Wor T E T T W , and coquo; again in ircavpes, reoaapes or rerrapes,
and quattuor ; and lastly in re and que. Moreover it will commonly be
found i n all instances of extreme change, that an intermediate variety
will present itself. Thus the late form r e m u occupies a middle place
between T B X W and T E T T W . So too the Slavonic tongues unite two of
the consonants in their interrogative kto. How easily noheros might
have passed through Trohcpos into a sound such as roXepos, is shown
by the ordinary English pronunciation of the name Ptolemy. I n the
next place it is commonly allowed that a guttural, such as that of
quis, has disappeared from the Latin derivatives uter, ubi, unde, ut,
usque, usquam, unquamf, much in the same way as in our own tongue,
quwhat. quwhere have gradually been softened into what, where, &c.
Indeed in who and whose the w itself has no vocal power, and in the
adverb how, the superfluous consonant has ceased to be written.
I n the Latin and English languages there is no substantial differ-
ence between the interrogative and relative forms. The same confu-
sion in fact is found in most languages. Thus in Stephanowitsch‘s
Serbian Grammar, translated by Grimm, we find it stated in a note
(p. 63), that all the modern Slavonic languages use their interroga-
tives as relatives ; and if he make an exception for the old Slavic, it
must be recollected how very limited in quantity and character are
the remains of that dialect.
If then the examples of other languages lead us to expect a con-
nexion of form between the interrogative and relative, and if further
* A form not actually occurring, but implied in many derivatives.
t The last three being connected with the compounds grcisque and quisyunm.
Thus, as quis : cum : : quisqram : (c)umquam.
VOL. 111. H
 58

we find that the initial guttural has a tendency to disappear, there

is nothing very startling in the position that the Greek relative 6s, 4,I ,
is a corruption of an older form beginning with a c. Indeed in the
Greek tongue itself we find T L S ,which is commonly an interrogative,
performing the office of a relative in burrs = quisquis.
So far we have been speaking only of the interrogative and rela-
tive. Let us now compare the latter with the so-called definite
article, or, to use the language of the older grammarians, let us com-
pare the postpositive and prepositive articles. In the Greek lan-
guage, taking the ordinary forms, we find that the fem. nom. of the
singular, and the maw. and fem. nom. of the plural have nothing
but an accent to distinguish the one from the other. Again, in the
masc. nom. of the singular 6s and 6, the sole difference as regards
the letters lies in the final s of the relative ; but this being the cha-
racteristic of the case itself, has of course nothisg to do with the
base of the word. Even those however who rely upon this differ-
ence are driven from their position by the fact, that in many combi-
nations the s preserves its place where the ordinary sense of the
relative is not to be found : for example, in q 6’ bs, said he; cat i)s,
and he, &c. Consistency moreover would require, that on the same
ground the Latin interrogative quis and the relative qri should be
held to be independent words.
In the neuters ro and ~a compared with 6 and i( there is indeed a
more marked difference in the prepositive appearing with an initial
T ; and the Doric nominatives rot, Tat, extend this difference still
further. But of what avail is this, when we find that the inter-
rogative itself has exchanged its K for a r in the form already men-
tioned, ris, rivos ? Nay, the interrogatives T O U and TY, at least in
their contracted form, are in no respect distinguishable from the
article. But the original identity of the prepositive and postpositive
articles seems placed beyond doubt by the two considerations, that we
see in the first, taken by itself, a twofold form, some parts having nn
initial T, some a mere aspirate : and secondly, that as regards meaning,
the forms commencing with a T are again and again used as relatives :
for example, in the ordinary language of Herodotus, as Ouovuc pev rp
r a p e v y rous re vauqyovs K ~ rovs
L a v Xa&at &c.; for no one will
venture to contend that the two words, identical as they are in form,
and connected too by the particles re, cat, are of different origin. (See
dso Buttmann, Gr. Gr. $ 75. Anmerk. 4.) Nor is it only in the
Greek language that we find the definite article performing the office
of a relative. Even in our own tongue the use of that for ‘ which’
is familiar to every one, and yet the immediate connection of this
word with the definite article is demonstrated by its German equi-
valent das, to which it corresponds, so far as the terminal consonant
is concerned, precisely as what to wus, and it to es. But in the
German language itself every form of the definite article does duty
as a relative, and in truth the pronoun der is in greater demand for
this purpose than the so-called relative itself. Thus within the
compass of five verses in the third chapter of St. John we have, Wir
reden das wir wissen, und zeugen das wir gesehen haben-Des
 59

h4enschen Sohn, der im‘Himme1 ist-Alle die an Ihn glauben.

Again, in our old English poets we frequently find the adverbs, whose
form connects them with the definite article, used with the sense
which belongs to derivatives from the relative. The fact is familiar
to English scholars, but we may be permitted to quote two passages
from Chaucer’s Nonnes Priestes Tale :
For in an oxes stalk
ThiR night shall I be inordred there I lie.
Again :-
A col fox full of sleigh iniquitee
Into the yerd, there Chanticlere the faire
Was wont and eke his wives to repaire.
But there are yet other phrases still living in our tongue where the
article (or personal pronoun) seems to have the power of a relative ;
I mean those in which our ordinary grammars tell us that the rela-
tive is omitted or understood. Thus : ‘‘ The man you just saw is
the celebrated N- ;” “ The gentleman you were talking with, I do
not know ;” “ H i m I accuse the city ports by this hath entered.”
(Shakspere, Coriol. v. 5. 5.) This construction is found also in
the Swedish : as “ Den Herren du nyss sig, ar den beromde N-; ”
“ Den Herren du talade med, kanner jag ej. ” (Dieterich, Swed. Gr.

p. 208.) Now in each of these phrases, the first word, call i t what
you please, is virtually a relative. ‘‘ Quem virum mod0 vidisti est
clarissimus ille N- ;” “ Quem alloquebaris ego haud novi ;” “ Quem
accuso intravit jam portam.”
On the other hand, we have a remarkable instance of the relative
used for the article in the old Slavic. Dobrowsky (Instit. p. 608)
has this phrase : “ Utuntur interpretes relativo ad exprimendum
Grsrorum articulum i fi ro, quo carent Slavi.”
T h a t the so-called article in Greek was originally a demonstrative
pronoun, equivalent in power to an English this, is familiar to the
reader of Homer ; and we need only refer on this point to the pages
of Matthis’s Grammar, who quotes from the opening lines of the
Iliad : ra 8’ a r o i v a 6execrOac, ‘ and receive this ransom ;’ rvv S eyw
ou Xuaw, ‘ but this woman I will not let go.’ The same writer points
out traces of this usage of the mere article for a demonstrative in
many phrases which prevailed long after Homer’s time, as r p o TOU,
‘ before this ;’ r y , ‘ for this reason,’ &c. &c.
Words commencing like the Greek article with a t , and including
the notion of a demonstrative this, have been often pointed out in the
Latin language, as tum, tam, talis, tot, t6tus. tamen, and tandem
(= tamen-dem) ; and even in English, as today, tonight, tomorrow,
to-year, &c. But in the latter set of words the preservation of an
unaltered t is an anomaly, and at variance with the law of letter-
change which unites our tongue with the Latin. and requires a t h
to correspond to the Latin t , as seen in the more regularly-formed
the, then, they, thus, than, there, thence, thither. Again, our language
agrees with the Greek in having pronouns commencing with a mere
aspirate, thus : he, his, him, her, here, hence, hither. The difficulty
that at first sight presents itself, in giving the same origin to words
H 2
 60

so opposed in meaning as here and there, hence and thence, &c., will
be considered presently.
W e return to the Latin language, where the pronoun hic claims
our attention. This little word is already a compound, and must
have originally existed without the suffix c or ce, just as ille and iste
preceded the nominatives illic and istic. I n fact, the plural forms hi,
h a , horum, his, hos, has, are found commonly without the suffix,
though not to its exclusion, as is seen in the nom. masc. pl. hice, the
fem. pl. haec, horunc, hisce. &c. Again, hodie and horsum are formed
from the simple pronoun, the Ro of the former being that ablative,
which with the suffix added became ho-c, and the ho of the latter
being that particle signifying ‘ hither,’ which with the snme addition
became huc, precisely as the adverbs ill0 and isto became illoc or
illuc, istoc or istuc. This suffix thrown aside, the analogy of ille and
iste brings us to he*, which hapi-ens to be the exact form of our own
personal pronoun in the masc. nominative. Again, the double form
of ipsus and apse, and the knowledge of the fact that the Latin lan-
guage has a strong tendency to get rid of final sibilants, lead us to
the conclusion that there must once have been a nom. hrs, which
would be the precise form in which the Greek 6s should appear in
Latin. An examination of the oblique cases would confirm a belief
in the connection of these Greek and Latin pronouns.
W e postpone for the present the consideration of the final con-
sonant of the word hic, and proceed to the pronoun is, eu, id.
Although is and id belong to a different form of declension from that
of ea, eo, &c., yet the difference is scarcely greater than that of
quis, quid, with the i compared to the forms quo, qua, &c., which
follow the analogy of bonus. W e may also compare the variety in
the declension of alius and the archaic alis, alid, or of the Greek T L S
and reo or TOU, r e p or T$. The sole peculiarity which yet remains to
the forms ea. eo, eum, &c., namely that they take an e rather than an i,
is exactly paralleled by the similar preference of an e before the same
strong vowels a, 0, u, in the conjugation of ire, ‘ to go.’ Thus we
have imus, iisse, iens, but eo, eam, euntis (Bopp, V. G. $ 361). Be-
tween hic, hcec, hoc, and is, ea, id, there is evidently a close connection,
though not an identity of meaning. Is, eu. id, i t must be admitted,
never refers to actual objects in nature. W e cannot say, pointing
to a hook, is liber for ‘ this book ;’ but on the other hand, though
hic liber may refer to a book on a table before one, it may also refer,
like is, to a book just mentioned in a previous clause or sentence.
There is then, we repeat, a near connection in meaning. As regards
form we may remark that the initial h must have been most faintly
pronounced by both Greeks and Romans, or it would have had the
power in their poetry of preventing that commingling of final and
initial vowels which is called elision. Again, but for the weakness
of the sound, the Greeks would never have degraded its ocular
symbol from a full letter H to the insignificant mark called the spiritus
asper. On the other hand, for the Latin we have the confirmation
Such coincidences are probably not altogether accidental. Compare the Ger-
man paiticle ILL,
‘from the speaker,’ with hin, the essential part of the Latin hinc.
 61

of the Italian Orazio, Omero, oggi, onore, &c. Secondly, the archaic
dative hibus tells us that hic, h m , hoc, like so many other pronouns,
had its declension in i, just as quibus tells us the same for the relative.
Thirdly, the pronoun is, ea, id, actually gives us an aspirate in the
plural Jbi and his, for such, rather than ii or iis, eis, &c., are com-
monly found in the best MSS. A familiar example occurs in the
titles of the Juristical writers, these as De his qui in potestate sunt in
Ulpian. That this his must be referred to is as a nominative will
be felt by those scholars who contrast the singular hic qui est in p o -
testate with is qui est in potestate; the first of which could only be
used in speaking of a definite person hefore one, while the latter,
like Ulyian’s phrase, is altogether indefinite. Lastly, our own pro-
noun it has also lost its aspirate, which was found in the A.-Sax.
hit, and-is still retained in the Dutch het. I t was also preserved
for a time in the neuter genitive his, for the little word its is allowed
to be of recent introduction, not having been known to Shakspere
or the translators of the Bible. Moreover we take a still greater
liberty with the plural them, which so often becomes em when used
as an enclitic : as in “ we found ’em all well.” These considerations
united seem to us to remove all difficulty arising from the varieties
of form, and to make it something more than possible that is, ea, id,
and hic, hcec, hoc, setting aside the suffix ce, may be one in origin.
W e must now go back to consider those forms of the interrogative,
relative, &c. which commence with an s. In aqpscov and aqrss the
a is generally admitted to represent the article ro (Ahrens de Dial.
Dor. p. 6 6 ) . The same writer points out that the Megarensian
phrase in Aristophanes--a& pciv;-has in its first element a dialectic
variety for ri or r i m . Again, in the non-enclitic Ion. u”aua,Att. tirra,
asButtmann( Gr. Gr. p.301 note) happily explains it, the first syllable
is the formal equivalent of the Latin ali in ali-quid and of et in the
German et-was ; and thus the remaining r u or ua is a variety of the
relative. The appearance of s in relative forms is seen in our
English whoso, which like torts and quisquis has a doubling of the
relative upon itself.
I n a paper read* before a Society bearing the same name as that
to which the present paper is addressed, the writer gave his reasons
for believing that our conjunction so was akin to the relative,
as also the Latin conjunction si, and thirdly, that the Latin adverb sic
was the very same word as regards the first two letters, the final c
being the same appendage which appears at the end.of hic, nunc, tune,
&c. I n confirmation of the argument, one reference, among others,
was to the use of sic as a correlative to si, for instance in Horace-
Sic igiiovisse putato
Me tibi si coenas hodie mecum-
where the two particles correlate with each other, and denote a con-
dition, precisely as ita and si so frequently do in Livy, &c. Further,
the use of so for ‘if’ in the German language was urged, as : S o
wir sagen wir haben keinen Sunde, SO verfiihren wir uns selbst und
* Nov. 16, 1840.
 62

die wahrheit is nicht in uns ; S o wir aber unsere Sunden bekennen,

so ist er getreu und gerecht dass, &c." A similar use of so in English
was noticed by a gentleman present at the discussion, as " So you
admit the principle, we care n& for the form." Shakspere too has
' SO please you,' for ' if you please,' &c.
Again, the Gothic words sd-s (qui) and sd-s (SUE), noticed by
Grimm (D. G. iii. p. 22. Q 7), are alreadyrelatives in the first element
though strengthened in that sense by the relative suffix s ( e s ) , just
as our where-as" also unites two relatives. The same writer refers
to the use of sem in Icelandic and som in Danish and Swedish with
relative power. W e ourselves have preserved the latter in our now
vulgar how-som-ever. Moreover our indefinite pronoun some, which
by the way existed with the same sense in all the old German
dialects (Grimm, D. G . iii. p. 4), must be the same word, just as ris
in Greek, and in like manner the German was, is at once an interro-
gative and an indefinite pronoun. The Latin pronouns sum, sum, sas,
as used in the time of Ennius for eum, earn, eas, have the same initial
sibilant ; and lastly our own such is traceable through the German
solch-er = so-lich-er to the Latin ta-li-s, precisely as our which+
(quwhilk) through the German welch-er = we-lich-er to the Latin
qua-li-s, for the 1%'of the Latin words has lost its final guttural, pre-
cisely as our own manly has, compared with man like, Germ. mannlzch.
In proceeding to the discussion of the pronouns of the third
person, we must once more remember that the classical languages
most freely use the masculine and feminine of the article or demon-
strative pronoun as personal pronouns, having in fact no other forms ;
secondly, that the r , the ordinary initial of the Greek article, must
be expected to appear in English as th, according to the regular law
of interchange between the languages ; thirdly, tliat the Greek, be-
sides its rou, rqs, rou, has other cases with a mere aspirate, and at
times substitutes a s for a r , as in q p e p o v , &c. ; fourthly, that the
Latin has dropped even the aspirate inmost of the cases of is, en, id.
These four considerations seem abundantly sufficient to establish that
he, she, it, and they are branches from one common stem ; and not,
as many grammarians tell us, unrelated words accidentally brought
into connection. The little word the (now called the definite pro-
noun) was, it is well known, in the earlier language susceptible of
declension with the sense ' this,' and is therefore well-entitled to
claim, as belonging to it, all the forms just given of our personal
pronoun. Thoso. who attribute an innate notion of the masculine
to the initial h of he, or of the feminine to the s or sh of she 1, or of
plurality to the th of they, will have all their ideas upset by a perusal
of the Anglo-Saxon and Frisian pronouns here given in their full
declension from Rask :-
* " St. Albans, whereas (= where) the king and queen do mean to hawk " (Sh.
Hen. VI. Part 11. 1. 2). On the other hand, in Act III. Scene 3, we have where
signifying whereas: ' Where from thy sight I should be raging mad.'
t If which has now lost the modification of meaning belonging to the suffix like,
sn also has the French que2. .
In the Lithnanian, sh, or as it is there written, sz,runs throngh the declension
of the masculine (Bopp, 1'. G. 358).
 63

ANQLO-SAXON. FRISIAN.
SING. PLUR. SING. PLUR.
m. f. n. m.f.n. m. f. n. m.f.n.
N. he he6 hit hi thi se thet se
A. hine hi hit hi thene se thet se
G. his hire his hira thes there thes th6ra
D. him hire him him th6m there thbm thiim.
 64
W e have just said that the K of rcqvos presents no difficulty. We
may go further and say that it tends to prove instead of disproving
the connection, for the modern Italian quello, quel, colui, exhibit the
pronoun ille with a guttural.
It appears then that our pronoun began with a guttural, that the
middle vowel was an e or 0 , and that the final consonant was a liquid,
the Latin and Italian pleading for an I , but the English, German and
Greek for an n. And the claims of the latter liquid are strongly
supported by the united voices of the Slavonic dialects, as will be
seen in the following extracts, where for other reasons we include the
demonstrative pronouns which begin with a t and signify for the
most part ‘this.’ The forms are given in the order m. f. n.
OLD SL-AVIC. BOHEMIAN.
ty, ta, t o ; ‘this.’ ten (olim sen), ta, to ; ‘ this.’
on, ona, ono; ‘ that.’ onen & ow ; ‘ that.’
on; ‘he,’ &c. on, ona, ono ; ‘ he, she, it.’
Dobrowsky, Instit. p. 341. Dobrowsky, Gr. p. 90.
ILLYRIAN. UPPER-LUSATIAN.

ti, t.e, t a ; ‘this.’ ton (or to), te, t a ; ‘ this.’

onaj, ono, ona ; ‘ that.’ won, &c. ; ‘ that.’
on, ona, on; ‘he, she, it.’ won, wona, won0 ; ‘ he, she, it.’
. Babukitsch, Gr. p. 52. Jordan, Gr. p. 70.
RUSSIAN. SERVIAN.

tom, &c., ‘that,’ ‘the,’ ‘the other.’ taj, ta, to ; ‘*this.’

ony, ona, onoe ; ‘ that,’ or ‘ this.’ onaj, ona, ono ; ‘that.’
on, ona, ono ; ‘ he, she, it.’ on, ona, ono ; ‘ he, she, it.’
Hamonihre, Gr. p. 120. Stephanowitsch‘sGr. pp. 54-60.
From the evidence thus produced, it seems probable that n* rather
than 1 was the final consonant of the pronoun we have been ex-
amining; and thus we have the syllable ken or kon for the base of
the word.
But the meaning requires further consideration. The examples
quoted from the Slavonic tongues show that the pronoun containing
the syllable on was freely used for the third personal pronoun, which
in sense seems more closely related to the demonstrative ‘this’ than to
‘ that’ ; and secondly, Hamonikre gives to the Russian pronoun ad-
jective on the distinct meaning of ‘ ceci’ as opposed to ‘ cela.’ The
same difference of meaning seems to have prevailed in Greek, fdr the
old grammarims appear to have agreed that rqros, rqvos, K ~ L V O Sand
C K W 0s were but dialectic varieties of one word (see Ahrens de Dial.
Dor. p. 267); and Ahrens himself, while he wishes to make out that
S~VOS is of a different origin from the three others, pet admits that at
times it is used like eKElVOS for a distant object, though more commonly
for one that is near. But there is no difficulty in the supposition that
* The Turkish pronoun 81, used alike for the personal pronouns We, she, and for
the demonstrative that, pleads for the 1 in the nom. sing., but for the n in the
other cases, as Bn-un, &c. (A. L. David’s Gr. in French, p. 23.)
 65

one and the same pronoun may have denoted both ‘ this’ and ‘ that.’
Our own word that is an example in point, for there can be no doubt
that originally it was immediately connected with the pronoun the,
corresponding as we have already said to the German das, the neuter
of der. Again, in the Latin epigram, where it is said of Dido, in re-
lation to the two objects of her afTection-
Hoc pereunte fugit, hoc fugiente perit-
the latter hoc might, but for t h s metre, have given place to illo.
As Bopp observes (V. G. Q 371 Transl.), ‘‘ That which in Sanscrit
signifies ‘ this,’ means also for the most part ‘ that,’ the mind sup-
plying the place whether near or remote.” W e have already con-
tended that here and there, hither and thither, &c., are in origin the
same words. Again, the Latin alius or alis has commonly the sense
of ‘other,’ denoting difference rather than similarity. B u t this
sense seems to be one which is not well-adapted to explain the
meaning of the compound a l i - p i s , ‘ some one.’ The various forms
of alius in the most important of the kindred tongues may be seen
in Bopp (V.G. Q 374), where it appears that n rather than I was
the original element. But instead of the explanation of it given by
the German scholar, we think* that the numeral one (Scot. ane) is
the base of the word, as of the German einig, Eng. any, so that the
literal translation of ‘ Aliud est maledicere, aliud accusare,’ would
perhaps be, ‘ It is one thing to abuse, one to accuse,’ where for the
second one, the word another might equally stand. But though the
same word, with the help of the finger pointing to different quarters
in succession, might thus come to be used in senses apparently op-
posite, yet there is some inconvenience in it ; and no better mode of
avoiding this inconvenience can well be proposed than that of adopt-
ing for the remoter object some dialectic variety of the same word.
Thus in Mantchou ere is ‘ this’ (celui-ci) and tere ‘ that ’ (celui la),
two words which stand to each other in the same relation of form
and meaning as our own here and there. The same difference
distinguishes in the same language ouba ‘ceci’ and touba ‘cela,’
enteke ‘ hujusmodi’ and tenteke ‘ illiusmodi’ (Gabelentz, Gr. Mand-
choue, p. 38). The f o r m ere, tere, are found with the same variety
of sense in the Mongolian (I. J. Schmidt, Gr. p. 46).
Hitherto our remarks upon the pronouns, so far as their external
shape is concerned, have been directed to the initial consonant, which
seems to have been in origin a guttural such as k, and to the vowel
which takes various forms. W e next ask the question, whether
there was a final consonant in the earliest root. The Greek inter-
rogative ris, rivos places before us in the body of the oblique cases a
consonant which is too often neglected by philologists. The writer
in dealing with other parts of language has more than once given his
reasons for believing that the Greek verbs, &c., which appear at one
* Still both views will be united if one itself be in origin a demonstrative, a thing
far from impossible, because if language creates for itself several varieties of a de-
monstrative, the one most in use will be first used in an enumeration, the second
most in use will be employed in the second place, and thus they will acquire in
time the notions of ‘ first’ and ‘ second,’ or ‘ one’ and ‘ two.’
 66

time with a nasal at the end of a syllable, at another time without

it, originally possessed that letter as an essential part of the word*.
And surely that man would be a bold philologist who would contend
that /3rhr:ous is older than Pekrtoves, or the common acc. Xoyous than
the Cretan hoyovs. A final nasal occurs also in the Icelandic rela-
tive sem, Dan. and Sw. som, as well as in our indefinite pronoun
some. Again, our own howsomever leads us to consider the middle
syllable of quicumque as deriving its m (or n ) from the same source.
The Slavonic forms ten and sen in the Bohemian, ton in Upper-Lu-
satian, tom in Russian, exhibit a similar nasal in the demonstrative
signifying ‘this,’ while a large majority of the allied languages
assign the same final letter to the other demonstrative signifying
‘ that.’ But in the northern dialects of the German tongue we find
still stronger evidence. The Swedish grammarian Dieterich is of-
fended at finding hwemt used as a nominative when his ‘theory
would require it to be a dative, and still greater is the offence he
takes at the genitive hwems. But if the view we are now drawing
attention to, be correct, the liquid is entitled to take its place both
in the nominative and the oblique cases. Again, for the personal
pronouns we find in Icelandic Masc. N. hann:, A. hann, G . hans,
D. honum; and for the Fem. N. hbn, A. hana, G. hennar, D. henni
(Rask, Transl. p. 94). So in the same tongue the definite article
not only has a final n in the m. and f.nom., but it carries this n into
every oblique case of all genders and both numbers. The Swedish
presents a similar peculiarity in both the nom. and gen. of its definite
article en, its demonstrative den, and its personal pronouns m. Ban
and f. hon. So again the Finnish has N. Ian, G. Wan-en, &c. (Vhael,
Gr. p. 52) ; and the northern dialect of the Lappish agrees to some
extent in the same peculiarity, as N. sodn, G. sun, D. sunji (Fiell.
strom’s Gr. p. 39 ; see also Rask’s Gr. p. 79). Lastly, the Turkish
relative is both kim and kih (David’s Gr. pp. 25, 26), the Finnish
either ken or cu, and the Mongolian ken (Schmidt, p. 144).
All our investigations then, whether as regards the interrogative,
or relative, or demonstrative pronouns, including under the latter
term the words which signify the, he, she, it, this and that, lead to a
strong probability that from a form such as ken$ every one has been
* Observe too the n in the Spanish quien.
t So Dopp with t h e nom. kim in Sanscrit (V. G. 390).
I: This is also used as a relative (Rask,.p. 100).
4 A final n is commonly convertible with the several consonants r, s, t. Com-
pare the Greek neuters in p u r with the Latin in men, as ovopn and nomen ; also
such forms as +rap Ijrar-os, repas rEpar-os,fri~usfrigOr-is,&c. Hence we must
not be surprised to find the lcelandic relative taking the shape h e r (instead of ken)
for its base, the letter r being continued through the cases of all genders and both
numbers, and taking a second r for the usual Icelandic suffix of the nominative
case. Hence also in all probability the r of therein, herein, wherein, &c. &c., rather
than, as was suggested in the paper read Dec. 8, 1843, from the German nomina-
tive er, ‘he.’ Hence olso, if the writer is not mistaken, the t and s final of the so-
called neutera, that, whct, das, was ; for thus the German stock of languages will be
brought into harmony with the Greek and Latin, which commonly represent the
neuter by the mere base without any special suffix ; and hence perhaps the expla-
nation of the Danish der being used for all genders and both numbers, as well as
 67
derived. If however we admit such a primaval root as the common
source of all, there still remains the question, what was its signi-
fication? W e are aware that Bopp in his V. G. divides all roots
into verbs or pronominal forms, but such a division seems unphilo-
sophical, hecause pronouns in their very nature cannot have belonged
to language in its first form. Their title alone tells us that they are
but deputies, and most certainly language might have been very in-
telligiljle. though not possessed of a single pronoun. Moreover,
those who like the present writer maintain the principle that the
imitation of the actual sounds of nature was the original source of
all language, will admit that action-for without action there cannot
be noise-must have been the first object of language, and the gram-
matical term for action is a verb. W e believe then that the word
ken was a verb, and it so happens that our own tongue still preserves
such a verb with the sense which is the one of all others most
adapted to our purpose, namely ‘see.’ Such a word is precisely
that which might well accompany the act of pointing to an object ;
and the demonstrative sense of the pronouns is the one from which
all the other notions readily flow, and moreover the one which has
the greatest historical antiquity. On the latter point we need only
refer again to the first sections of Matthize’s Syntax. But a few
more words will be required to trace the changes of signification.
W e hare already shown how the notion of that ’ may be derived
from the notion of ‘ this’; and the connection of the ideas expressed
by the,’ as also by ‘ he, she, it,’ is too obvious to require explana-
tion. The relative, which next presents itself, has for its main ob-
ject to denote the identity of that which is the subject of two different
predicates. If the said thing be present, the speaker may point to it.
For example, suppose him desirous of saying he entertains a high opi-
nion of a h r s e which he bought the day before, he can say, “ This
horse I bought yesterday, this horse I highly value.” But even if
the horse be not in view, still he may imagine it to exist in the form
of any stick or stone at hand. Thus the lawyers of ancient Rome,
in conveying a distant estate in land, took up any clod of earth, and
spoke of it as though it were the actual land then selling. And this
view of the relative seems to explain that remarkable construction in
the Greek writers, where, after a so-called relative clause, the apo-
dusis or main clause, not satisfied with a pronoun correlating with the
relative, adds thereto a conjunction &* or re$ or Kai (see Kuhner’s

our own what and that. The ready interchange of f n and s is well seen in the
first element of the Icelandic pronoun for ‘ this,’ in the acc. m. hen-nu, f. bes-sa,
n. bet-ta. Compare also the perf. part. in the same 1anguage;as kaldin, ‘ holden,’
Nom. m. haldin-n, f. haldin, n. haldit, where Rask would seem to be wrong in
supposing that the n falls away before t in the neuter (Icel. Gr. p. 86). Of
course the d of quid, quod, i d , &c. must be of the same origin as our t in what, it,
&c., and is therefore also explained.
* Oi 6’ apa M q h v q v Kal eaUpUK1qV 6V6pOV70,
T w v 86 @iXorrqrqsqpx6v. 11. p. 716.
+ ‘ 0 s C R &OLS 6 m d 3 q r a i paXa r’ ~ K X U aurov. O~ 11. a. 218.
AXX’ b r E 89 p’ E K 7010 8 U O J 8 E K a r l ) YEVET I ) W S ,
Knc T o r 6 89 r p o s O h u p ~ o viaav f h l . 11. a. 493.
 68

Gr. $0 723, 738); as though the original form of the relative sentence
had been, “This horse I bought yesterday and this horse I highly
value.” The writer has elsewhere proposed a similar explanation of
the use of atque after alius.
That the interrogative should be explained through the medium of
the relative, and not directly from the demonstrative, is suggested
by its remarkable agreement in form with the relative. All we re-
quire besides is that the so-called indirect use of the interrogative may
claim precedence of the direct use. The step is but a short one from
‘‘ Monstra eum qui fecit” to “Dic quis fecit” or “ fecerit,” for the
original signification of dic-ere, or in the oldest orthography deic-ere,
is the same as that of the Greek ~ E ~ I - Y U Y‘to
~ ,show,’ and the use of
the indicative in the indirect question is not uncommon in the older
writers. Secondly, the passage from ‘ Dic quis fecit’ to ‘Quis fecit?’
would be an abbreviation of n o unusual kind, the very tone in which
a question is put rendering the use of the verb Dic a mere idle form,
the more so as this word would, in the full phrase, be the invaria-
ble, and for that very reason almost superfluous attendant on every
question.
The consideration of the suffix ce in hic was postponed. This
also we would refer, not, as is commonly done, and we ourselves too
have elsewhere done, to the particle ewe, but to the same verb ken
from which the pronouns themselves have been deduced. The mean-
ing is well-suited, inasmuch as the particle is never added but to
demonstratives ; and what seems still more decidedly to establish the
proposition, is the circumstance that hic then hecomes a reduplicated
form, a virtual repetition of the same root, which would be in agree-
ment with the well-known formation of the corresponding pronoun in
other languages, as o h o s , the Germ. die-ser, Icelandic pes-si, pet-ta,
Swed. den-ne, Bohem. ten-to, Lithuan. szit-tas, drc. &c. On the other
hand, the degradation of ken to ke in an enclitic syllable is precisely
what is to be expected. The Homeric particle of precisely the same
form, h e y , has undergone the same corruption into K C . So the En-
glish article an and Greek privative riv readily drop their nasal. So
again, the moment that the Latin conjunction vel becomes an encli-
tic, it is robbed of its final consonant and becomes ve. Compare also
the interrogative particles num and the enclitic ne. Lastly, our own
verb look, when ceasing to be used as a formal verb, takes the lighter
shape of 20.
To avoid an unnecessary complication of the subject, no notice has
hitherto been taken of those forms in which the initial t of the demon-
strative pronoun has been supplanted by the dental nasal n ; whence
num in etiamnum, nunc, V U Y , now, ncim, and the German noch. Clough
and Bopp hare directed notice to the appearance of this liquid in the
demonstrative pronouns of the Sanscrit and other Eastern lan-
guages (V. G. 5 369). It repeatedly makes its appearance in the
Finnish. The interrogative particle num, ‘ whether,’ is probably
from the same source, the more so as the Turkish uses ne for its
neuter interrogative (David, p. 26), and the Chinese has na for its
ordinary interrogative (Endlicher’s Gr. p. 278). And having men-
 69
timed this language, we may be permitted to notice that its pro-
nominal words give a strong confirmation to the general principles
supported in this paper. Thus so (itself, as we have already no-
ticed, a relative in our European tongues) is used a t once as an in-
terrogative (Endlicher, Gr. p. 273. $ 2), as a relative (p. 271. $ 2),
and as a demonstrative (p. 272, Anmerkung). And again, the word
tcP signifies sometimes ‘ this’ (p. 268), sometimes ‘ who’ or ‘ which’
(p. 270).
The views we have been led to entertain appear to explain many
constructions in our own and other languages. Thus the little
word as* is used with the power at one time of a prepositive, a t
another of a postpositive particle. For example, in the combination
‘ as well as,’ the first as is the equivalent of the Latin ‘ tam,’ the
second of ‘ quam.’ While this very word quam is the ordinary par-
ticle after the comparative in Latin, the English language employs
for this office a derivative from the prepositive, viz. than, or in Old-
English then, and so the German uses denn. Again, when we come
across such a phrase as the German ‘ so lange’ = ‘ so long as,’ our
first impression will probably be, that some German particle equiva-
lent to the English ‘as,’ may have fallen out ; but the difficulty dis-
appears when we look upon the German so in this phrase as the
representative of the Latin quam.’ Thus ‘so lange’ will correspond
to ‘ quamdiu.’ O n the other hand, the conjunctions of the Latin
language which consist of a preposition and some suffix of relative
form, as quam, quod, or ut (for example postquam, propterea-quod,
prout), admit of explanation, if we may translate the suffix by ‘ this.’
the pronoun being in appoeition, as the grammarians say, to the
sentence attached ; and this view is in agreement with the formation
of the German conjunctions nach-dem, in-dem, &c. Again, such
English phrases as ‘ the longer, the better,’ now lose their strange-
ness, although other languages commonly combine in this construc-
tion a demonstrative and relative particle, ‘ quo diutius, eo satius.’
So also the apparent contradiction of meaning between our English
particle though, ‘ quamvis,’ and its German analogue ‘ doch,’ ‘ tamen,’
is accounted f o r t .
* Its use as a relative should not be forgotten : “ Him as prigs what isn’t hisn,”
&c., though a vulgar phrase, may be defended in all its unusual forms. As regards
the word as, see the use of es, el, &c. for the relative in Grimm, iii. 22, and the relative
particle z of the Slavonic (Dobrowsky, Bohem. GI. 5 1 4 4 ) . The him itself is a legi-
timate nominative, and hence perhaps in the passage quoted above from the Corio-
lanus the construction was less offensive. Lastly of him the first three letters
constitute the base of the pronoun, and the final n the genitival suffix (see the pre-
cedingpaper, No. 5G).
f This suggests that the particle yet is probably of pronominal origin. I t could
hardly be otherwise, seeing that doch, noch, tamen, adhuc, so similar to it in sense,
are admitted to be drawn from such a source. The a o r d yet with its y, like yon
and yesterday, leads one to expect a guttural in kindred tongues; and such is
found probably in pqkwi, a compound of pq and I C ~ T Lalthough , the simple word
e n has lost its initial guttural, as we have seen in many instances of pronominal
words. The Latin at too seems to be only a shorter form of E T I , at any rate the
translation ‘yet’ is better adapted than ‘ but’ to the greater number of passages
where it occurs. The final vowel would of course disappear in Latin. Compare
e ~ i U, H O , & H O , 6ri, with oh, ab, sub, ut.
 70
[In the paper read next after this by Mr. Guest, will be found a
remarkable confirmation of the view taken in this paper,-a con-
firmation the more valuable as that gentleman was not present on
the 26th of March, and in fact had written his paper on the Chinese
vocabulary before that date. It is therefore a testimony altogether
new and independent when he points out that the Chinese possess
the very verb ken in the sense of ‘behold,’ and when he deduces
from it the Chinese pronoun signifying this. See also a preceding
paper (sol. i. p. 287) by the same gentleman, in which he draws at-
tention to the use of whom as a nominative in Old-English, and to
the appearance of an n at the end of the Dutch relative. The
present writer happened not to have seen this paper in time, or he
would have noticed it in its proper place. It might have been ob-
served that the Latin particle en, ‘ behold,’ seems to be a corruption
of the supposed verb ken. Lastly, it may be right to state that
abundant evidence might have been deduced from the Sanscrit, had
the space of the paper permitted.]