CTLF Corpus de textes linguistiques fondamentaux • IMPRIMER • RETOUR ÉCRAN
CTLF - Menu général - Textes

Müller, Friedrich. Lectures on the Science of Language. Second Series – T06

Lecture VI.
On the Principles of Etymology.

Voltaire defined etymology as a science in which
vowels signify nothing at all, and consonants very
little. ‘L'étymologie,’ he said, ‘est une science où les
voyelles ne font rien, et les consonnes fort peu de chose
Nor was this sarcasm quite undeserved by those who
wrote on etymology in Voltaire's time, and we need
not wonder that a man so reluctant to believe in
any miracles should have declined to believe in
the miracles of etymology. Of course, not even
Voltaire was so great a sceptic as to maintain that
the words of our modern languages have no etymology,
i.e. no origin, at all. Words do not spring into
life by an act of spontaneous generation, and the
words of modern languages in particular are in many
cases so much like the words of ancient languages
that no doubt is possible as to their real origin and
derivation. Wherever there was a certain similarity
in sound and meaning between French words and
words belonging to Latin, German, Hebrew, or
any other tongue, even Voltaire would have acquiesced.
No one, for instance, could ever have
doubted that the French word for God, Dieu, was the
same as the Latin Deus; that the French homme, and
even on, was the Latin homo; the French femme, the
Latin femina. In these instances there had been no
change of meaning, and the change of form, though
238the process by which it took place remained unexplained,
was not such as to startle even the most sensitive
conscience. There was indeed one department
of etymology which had been cultivated with great
success in Voltaire's time, and even long before him,
namely, the history of the Neo-Latin or Romance
dialects. We find in the dictionary of Du Cange
a most valuable collection of extracts from mediæval
Latin writers, which enables us to trace, step by step,
the gradual changes of form and meaning from
ancient to modern Latin; and we have in the much-ridiculed
dictionary of Menage many an ingenious
contribution towards tracing those mediæval Latin
words in the earliest documents of French literature,
from the times of the Crusades to the Siècle of
Louis XIV. Thus a mere reference to Montaigne,
who wrote in the sixteenth century, is sufficient to
prove that the modern French gêner was originally
gehenner. Montaigne writes: ‘Je me suis contraint
et gehenné
,’ meaning, ‘I have forced and tortured
myself.’ This verb gehenner is easily traced back to
the Latin gehenna, *1 used in the Greek of the New
Testament and in the ecclesiastical writings of the
middle ages not only in the sense of hell, but in the
more general sense of suffering and pain. It is well
known that Gehenna was originally the name of the
valley of Hinnom, near Jerusalem (םהיג), the
Tophet, where the Jews burnt their sons and their
daughters in the fire, and of which Jeremiah prophesied
that it should be called the valley of slaughter:
for ‘They shall bury in Tophet till there be no place.’ 2
239How few persons think now of the sacrifices offered
to Moloch in the valley of Hinnom when they ask
their friends to make themselves comfortable, and
say, ‘Ne vous gênez pas.’

It was well known not only to Voltaire, but even to
Henri Estienne, *3 who wrote in the sixteenth century,
that it is in Latin we may expect to find the original
form and meaning of most of the words which fill the
dictionaries of the French, Italian, and Spanish languages.
But these early etymologists never knew of
any test by which a true derivation might be distinguished
240from a false one, except similarity of sound
and meaning; and how far this similarity might be
extended may be seen in such works as Perion's
Dialogi de Linguæ Gallicæ Origine’ (1557), or
Guichard's ‘Harmonie Etymologique des Langues
Hebraique, Chaldaique, Syriaque, Greque, Latine,
Italienne, Espagnole, Allemande, Flamende, Angloise

(Paris, 1606). Perion derives brébis, sheep (the
Italian berbice) from próbaton, not from the Latin
vervex, like berger from berbicarius. Envoyer he
derives from the Greek pémpein, not from the Latin
inviare. Heureux he derives from the Greek ourios.

Now, if we take the last instance, it is impossible
to deny that there is a certain similarity of form
and meaning between the Greek and French; and as
there can be no doubt that certain French words,
such as parler, prêtre, aumône, were derived from
Greek, it would have been very difficult to convince
M. Perion that his derivation of heureux was not quite
as good as any other. There is another etymology
of the same word, according to which it is derived
from the Latin hora. Bonheur is supposed to be
bona hora; malheur, mala hora; and therefore heureux
is referred to a supposed Latin form, horosus, in the
sense of fortunatus. This etymology, however, is no
better than that of Perion. It is a guess, and no
more, and it falls to the ground as soon as any of the
more rigid tests of etymological science are applied to
it. In this instance the test is very simple. There
is, first of all, the gender of malheur and bonheur,
masculine instead of feminine. Secondly, we find
that malheur was spelt in Old French mal aür, which
is malum augurium. (See Diez, ‘Etymologisches
Worterbuch der Romanischen Sprachen,’ 1858, s. v.)
241Thirdly, we find in Provençal agur, augur, and from
it the Spanish aguëro, an omen. Augurium itself
comes from avis, bird, and gur, telling, gur being
connected with garrire, garrulus, and the Sanskrit
gar or gṛî, to shout.

We may form an idea of what etymological tests
were in former times when we read in Guichard's
‘Harmonie Etymologique:’ *4 ‘With regard to the derivations
of words by means of the addition, subtraction,
transposition, and inversion of letters, it is
certain that this can and must be done, if we wish
to find true etymologies. Nor is it difficult to believe
this, if we consider that the Jews wrote from right to
left, whereas the Greeks and the other nations, who
derive their languages from Hebrew, write from left
to right.’ Hence, he argues, there can be no harm
in inverting letters or changing them to any amount.
As long as etymology was carried on on such principles,
it could not claim the name of a science. It
was an amusement in which people might display
more or less of learning or ingenuity, but it was
unworthy of its noble title, ‘The Science of Truth.’

It is only in the present century that etymology
has taken its rank as a science, and it is curious to
observe that what Voltaire intended as a sarcasm
has now become one of its acknowledged principles.
Etymology is indeed a science in which identity,
or even similarity, whether of sound or meaning,
242is of no importance whatever. Sound etymology
has nothing to do with sound. We know words
to be of the same origin which have not a single
letter in common, and which differ in meaning as
much as black and white. Mere guesses, however
plausible, are completely discarded from the province
of scientific etymology. What etymology professes
to teach is no longer merely that one word is derived
from another; but how to prove, step by step, that
one word was regularly and necessarily changed into
another. As in geometry it is of very little use to
know that the squares of the two sides of a rectangular
triangle are equal to the square of the hypotenuse, it
is of little value in etymology to know, for instance,
that the French larme is the same word as the English
tear. Geometry professes to teach the process by
which to prove that which seems at first sight so
incredible; and etymology professes to do the same.
A derivation, even though it be true, is of no real
value if it cannot be proved — a case which happens
not unfrequently, particularly with regard to ancient
languages, where we must often rest satisfied with
refuting fanciful etymologies, without being able to
give anything better in their place. It requires an
effort before we can completely free ourselves from
the idea that etymology must chiefly depend on
similarity of sound and meaning; and in order to
dispose of this prejudice effectually, it may be useful
to examine this subject in full detail.

If we wish to establish our thesis that sound etymology
has nothing to do with sound, we must
prove four points: —

1. That the same word takes different forms in
different languages

2. That the same word takes different forms in one
and the same language

3. That different words take the same form in
different languages

4. That different words take the same form in one
and the same language

In order to establish these four points, we should
at first confine our attention to the history of modern
languages, or, as we should say more correctly, to the
modern history of language. The importance of the
modern languages for a true insight into tire nature of
language, and for a true appreciation of the principles
which govern the growth of ancient languages, has
never been sufficiently appreciated. Because a study
of the ancient languages has always been confined to
a small minority, and because it is generally supposed
that it is easier to learn a modern than an ancient
tongue, people have become accustomed to look upon
the so-called classical languages — Sanskrit, Greek, and
Latin — as vehicles of thought more pure and perfect
than the spoken or so-called vulgar dialects of Europe.
We are not speaking at present of the literature of
Greece or Rome or ancient India, as compared with
the literature of England, France, Germany, and Italy.
We speak only of language, of the roots and words,
the declensions, conjugations, and constructions peculiar
to each dialect; and with regard to these, it must
be admitted that the modern stand on a perfect
equality with the ancient languages. Can it be supposed
that we, who are always advancing in art, in
science, in philosophy, and religion, should have
allowed language, the most powerful instrument of
the mind, to fall from its pristine purity, to lose its
vigour and nobility, and to become a mere jargon?
244Language, though it changes continually, does by no
means continually decay; or at all events, what we
are wont to call decay and corruption in the history
of language is in truth nothing but the necessary condition
of its life. Before the tribunal of the Science of
Language, the difference between ancient and modem
languages vanishes. As in botany aged trees are not
placed in a different class from young trees, it would
be against all the principles of scientific classification
to distinguish between old and young languages. We
must study the tree as a whole, from the time when
the seed is placed in the soil to the time when it bears
fruit; and we must study language in the same
manner as a whole, tracing its life uninterruptedly
from the simplest roots to the most complex derivatives.
He who can see in modern languages nothing
but corruption or anomaly, understands but little of
the true nature of language. If the ancient languages
throw light on the origin of the modern dialects, many
secrets in the nature of the dead languages can only
be explained by the evidence of the living dialects.
Apart from all other considerations, modern languages
help us to establish by evidence which cannot be
questioned the leading principles of the science of
language. They are to the student of language what
the tertiary, or even more recent formations, are to
the geologist. The works of Diez, his ‘Comparative
Grammar of the Romanic Languages’ and his ‘Lexicon
Comparativum Linguarum Romanarum’ are as valuable
in every respect as the labours of Bopp, Grimm,
Zeuss, and Miklosich; nay, they form the best introduction
to the study of the more ancient periods of
Aryan speech. Many points which, with regard to
Sanskrit, Greek, and Latin, can only be proved by
245inductive reasoning, can here be settled by historical

In the modern Romance dialects we have before
our eyes a more complete and distinct picture or
repetition of the origin and growth of language than
anywhere else in the whole history of human speech.
We can watch the Latin from the time of the first
Scipionic inscription (283 B. C.) to the time when we
meet with the first traces of Neo-Latin speech in Italy,
Spain, and France. We can then follow for a thousand
years the later history of modern Latin, in its six
distinct dialects, all possessing a rich and well-authenticated
literature. If certain forms of grammar are
doubtful in French, they receive light from the collateral
evidence which is to be found in Italian or
Spanish. If the origin of a word is obscure in Italian,
we have only to look to French and Spanish, and we
shall generally receive some useful hints to guide us
in our researches. Where, except in these modern
dialects, can we expect, to find a perfectly certain
standard by which to measure the possible changes
which words may undergo both in form and meaning
without losing their identity? We can here silence
all objections by facts, and we can force conviction by
tracing, step by step, every change of sound and sense
from Latin to French; whereas when we have to deal
with Greek and Latin and Sanskrit, we can only use
the soft pressure of inductive reasoning.

If we wish to prove that the Latin coquo is the
same word as the Greek péptō, I cook, we have to
establish the fact that the guttural and labial tenues,
k and p, are interchangeable in Greek and Latin. No
doubt there is sufficient evidence in the ancient
languages to prove this. Few would deny the
246identity of pénte and quinque, and if they did, a
reference to the Oscan dialect of Italy, where Jive is
not quinque but pomtis, would suffice to show that
the two forms differed from each other by dialectic
pronunciation only. Yet it strengthens the hands of
the etymologist considerably if he can point to living
languages and trace in these exactly the same phonetic
influences. Thus the Gaelic dialect shows the
guttural where the Welsh shows the labial tenuis.
Five in Irish is coic, in Welsh pimp. Four in Irish
is cethir, in Welsh petwar. Again, in Wallachian,
a Latin qu followed by a is changed into p. Thus,
aqua becomes in Wallachian apà; equa, épà; quatuor,
patru. It is easier to prove that the French mime
is the Latin semet ipsissirnus, than to convince the
incredulous that the Latin sed is a reflective pronoun,
and meant originally by itself.

Where, again, except in the modern languages, can
we watch the secret growth of new forms, and so
understand the resources which are given for the
formation of the grammatical articulation of language?
Everything that is now merely formal in the grammatical
system of French can easily be proved to
have been originally substantial; and after we have
once become fully impressed with this fact, we
shall feel less reluctance to acknowledge the same
principle with regard to the grammatical system of
more ancient languages. If we have learnt how the
French future, j'aimerai, is a compound tense, consisting
of the infinitive and the auxiliary verb, avoir,
to have, we shall be more ready to admit the same
explanation for the Latin future in bo, and the Greek
future in . Modern dialects may be said to let out
the secrets of language. They often surprise us by
247the wonderful simplicity of the means by which
the whole structure of language is erected, and they
frequently repeat in their new formations the exact
process which had given rise to more ancient forms.
There can be no doubt, for instance, about the
Modern German entzwei. Entzweireissen does not
mean only to tear into two parts, but it assumes the
more general sense of to tear in pieces. In English,
too, a servant will say that a thing came a-two,
though he broke it into many pieces. Entzwei, in
fact, answers exactly the same purpose as the Latin
dis in dissolvo, disturbo, distraho. And what is the
original meaning of this dis? Exactly the same as
the German entzwei, the Low-German twei. In Low-German
mîne Schau sint twei means my shoes are torn.
The numeral duo, with the adverbial termination is,
is liable to the following changes: — Du-is may become
dvis, and dvis dbis. In dbis either the d or the b
must be dropped, thus leaving either dis or bis. Bis
in Latin is used in the sense of twice, dis in the sense
of a-two. The same process leads from duellum,
Zweikampf, duel, to dvellum, dbellum, and bellum;
from Greek dyis to dfis and dís (twice); from duiginti
to dviginti and viginti, twenty; from dyi-kosi to
dfi-kosi, fi-kosi, and eí-kosi.

And what applies to the form, applies to the meaning
of words. What should we say if we were told
that a word which means good in Sanskrit meant
bad in Greek? Yet we have only to trace the
Modern German schlecht back through a few centuries
before we find that the same word which now means
bad was then used in the sense of good, *5 and we are
248enabled to perceive, by a reference to intermediate
writers, that this transition was by no means so
violent as it seems to be. Schlecht meant right and
straight, but it also meant simple; simple came to
mean foolish; foolish, useless; useless, bad. Ekelhaft
is used by Leibniz in the sense of fastidious, delicate; *6
it now means only what causes disgust. Ingenium,
which meant an inborn faculty, is degraded into the
Italian íngannare, which means to cheat. Sælig,
which in Anglo-Saxon meant blessed, beatus, appears
in English as silly, and the same ill-natured change
may be observed in the Greek euḗthēs, guileless, mild,
silly, and in the German albern, stupid, the Old
High-German alawâr, verissimus, alawâri, benignus.

Thus, a word which originally meant life or time
in Sanskrit, has given rise to a number of words
expressing eternity, the very opposite of life and
time. Ever and never in English are derived from
the same source from which we have age. Age is
of course the French âge. This âge was in Old
French edage, changed into eage and âge. Edage,
again, represents a Latin form, ætaticum, which was
had recourse to after the original ætas had dwindled
away into a mere vowel, the Old French (Diez, s. v.).
Now the Latin ætas is a contraction of ævitas, as
æternus is a contraction of æviternus (cf. sempiternus).
Ævum, , again, corresponds by its radical, though
not by its derivative elements, to Greek aifṓn and
the Gothic aiv-s, time, and eternity. In Sanskrit,
we meet with a âyus, a neuter, which, if literally
249translated into Greek, would give as a Greek form
aîos, and an adjective, aiḗs, neut. aiés. Now, although
aîos does not survive in the actual language of Greece,
its derivatives exist, the adverbs aiés and aieí. This
aieí is a regular dative (or rather locative) of aiés,
which would form aiesi, aiei, like génesi and génei.
In Gothic, we have from aivs, time, the adverbs aiv,
ever, the Modern German je; and ni aiv, never, the
Modern German nie.

There is a peculiar charm in watching the various
changes of form and meaning in words passing down
from the Ganges or the Tiber into the great ocean of
modern speech. In the eighth century B. C. the Latin
dialect was confined to a small territory. It was but
one dialect out of many that were spoken all over
Italy. But it grew — it became the language of
Rome and of the Romans, it absorbed all the other
dialects of Italy, the Umbrian, the Oscan, the Etruscan,
the Celtic, and became by conquest the language
of Central Italy, of Southern and Northern Italy.
From thence it spread to Gaul, to Spain, to Germany,
to Dacia on the Danube. It became the language
of law and government in the civilized portions of
Northern Africa and Asia, and it was carried through
the heralds of Christianity to the most distant parts
of the globe. It supplanted in its victorious progress
the ancient vernaculars of Gaul, Spain, and Portugal,
and it struck deep roots in parts of Switzerland and
Walachia. When it came in contact with the more
vigorous idioms of the Teutonic tribes, though it
could not supplant or annihilate them, it left on their
surface a thick layer of foreign words, and it thus
supplied the greater portion in the dictionary of
nearly all the civilized nations of the world. Words
250which were first used by Italian shepherds are now
used by the statesmen of England, the poets of
France, the philosophers of Germany, and the faint
echo of their pastoral conversation may be heard in
the Senate of Washington, in the cathedral of Calcutta,
and in the settlements of New Zealand.

I shall trace the career of a few of those early
Eoman words, in order to show how words may
change, and how they adapt themselves to the changing
wants of each generation. I begin with the word
Palace, A palace now is the abode of a royal family.
But if we look at the history of the name we are soon
carried back to the shepherds of the Seven Hills.
There, on the Tiber, one of the seven hills was called
the Collis Palatinus, and the hill was called Palatinus,
from Pales, a pastoral deity, whose festival was celebrated
every year on the 21 st of April as the birthday
of Rome. It was to commemorate the day on
which Romulus, the wolf-child, was supposed to have
drawn the first furrow on the foot of that hill, and
thus to have laid the foundation of the most ancient
part of Rome, the Roma Quadrata. On this hill, the
Collis Palatinus, stood in later times the houses of
Cicero and of his neighbour and enemy Catiline.
Augustus built his mansion on the same hill, and his
example was followed by Tiberius and Nero. Under
Nero, all private houses had to be pulled down on the
Collis Palatinus, in order to make room for the emperor's
residence, the Domus Aurea, as it was called,
the Golden House. This house of Nero's was henceforth
called the Palatium, and it became the type of
all the palaces of the kings and emperors of Europe.

The Latin palatium has had another very strange
offspring — the French le palais, in the sense of palate.
251Before the establishment of phonetic rules to regulate
the possible changes of letters in various languages, no
one could have doubted that le palais, the palate, was
the Latin palatum. However, palatum could never
have become palais, but only pale. How palatium
was used instead is difficult to explain. It was a
word of frequent use, and with it was associated the
idea of vault (palais vouti). Now vault was a very
appropriate name for the palate. In Italian the palate
is called il cielo della bocca; in Greek ouranós, ouranískos.
Ennius, again, speaks of the vault of heaven
as palatum cœli. There was evidently a similarity
of conception between palate and vault, and vault and
palace; and hence palatium was most likely in vulgar
Latin used by mistake for palatus, and thus carried on
into French. *7

Another modern word, the English court, the
French cour, the Italian corte, carries us back to the
same locality and to the same distant past. It was on
the hills of Latium that cohors or cors was first used
in the sense of a hurdle, an enclosure, a cattle-yard.
The cohortes, or divisions of the Roman army, were
called by the same name; so many soldiers constituting
a pen or a court. It is generally supposed that cors
is restricted in Latin to the sense of cattle-yard, and
that cohors is always used in a military sense. This
is not so. Ovid (Fasti, iv. 704) used cohors in the
sense of cattle-yard:

‘Abstulerat multas ilia cohortis aves;’

and on inscriptions cors has been found in the sense
of cohors. The difference between the two words was
a difference of pronunciation merely. As nihil and nil,
252mihi and mi, nehemo and nemo, prehendo and prendo,
so cohors, in the language of Italian peasants, glided
into cors.

Thus cors, cortis, from meaning a pen, a cattle-yard,
became in mediæval Latin curtis, and was used, like
the German Hof, of the farms and castles built by
Roman settlers in the provinces of the empire. These
farms became the centres of villages and towns, and in
the modern names of Vraucourt, Graincourt, Liencourt,
Magnicourt, Aubignicourt, the older names of Vari
, Grani curtis, Leonii curtis, Manii curtis, Albini
, have been discovered. *8

Lastly, from meaning a fortified place, curtis rose
to the dignity of a royal residence, and became synonymous
with palace. The two names having started
from the same place, met again at the end of their
long career.

Now, if we were told that a word which in Sanskrit
means cow-pen had assumed in Greek the meaning of
palace, and had given rise to derivatives such as
courteous (civil, refined), courtesy (a graceful inclination
of the body, expressive of respect), to court (to
pay attentions, or to propose marriage), many people
would be incredulous. It is therefore of the greatest
use to see with our own eyes how, in modern languages,
words are polished down, in order to feel less
sceptical as to a similar process of attrition, in the
history of the more ancient languages of the world.

While names such as palace and court, and many
others, point back to an early pastoral state of society,
and could have arisen only among shepherds and husbandmen,
there are other words which we still use,
253and which originally could have arisen only in a seafaring
community. Thus government, or to govern,
is derived from the Latin gubernare. This gubernare
is a foreign word in Latin; that is to say, it was
borrowed by the Romans from the Greeks, who at a
very early time had sailed westward, discovered Italy,
and founded colonies there, just as in later times the
nations of Europe sailed farther west, discovered
America, and planted new colonies there. The Greek
word which in Italy was changed into gubernare was
kubernân, and it meant originally to handle the rudder,
or to steer. It was then transferred to the person or
persons entrusted with the direction of public affairs,
and at last came to mean to rule.

Minister meant, etymologically, a small man; and
it was used in opposition to magister, a big man.
Minister is connected with minus, less; magister with
magis, more. Hence miníster, a servant, a servant
of the Crown, a minister. From minister came the
Latin ministerium, service; in French contracted
into métier, a profession, A minstrel was originally
a professional artist, and more particularly a singer
or poet. Even in the Mystery Plays, the theatrical
representations of portions of the Old or New Testament
story, such as still continue to be performed at
Ammergau in Bavaria, mystery is a corruption of
ministerium; it meant a religious ministry or service,
and had nothing to do with mystery. It ought to be
spelt with an i, therefore, and not with a y.

There is a background to almost every word which
we are using; only it is darkened by ages, and requires
to be lighted up. Thus lord, which in modern
English has become synonymous with nobleman, was
in Anglo-Saxon hlâf-ord, which is supposed by some
254to mean ord, the origin of hlâf, loaf; while others
look upon it as a corruption of hlâf-weard, the warder
of bread. *9 It corresponds to the German Brotherr,
and meant originally employer, master, lord. Lady
in Anglo-Saxon is hlæfdige, and likewise means ‘she
who looks after the loaf,’ the mistress; unless it is a
corruption of hlâf-weardige, the feminine of hlâf-weard.
Earl, the same as the Danish Jarl, was, I
believe, originally a contraction of elder; earl, therefore,
and alder in alderman were once the same word.
In Latin, an elder would be senior, and this became
changed into seigneur, sieur, and at last dwindled
down to sir. Duke meant originally a leader; count,
the Latin comes, a companion; baron, the mediæval
Latin baro, meant man; and knight, the German
Knecht, was a servant. Each of these words has risen
in rank, but they have kept the same distance from
each other.

As families rose into clans, clans into tribes, tribes
into confederacies, confederacies into nations, the
elders of each family naturally formed themselves
into a senate, senatus meaning a collection of elders.
The elders were also called the grey-headed, or the
Greys, and hence the German Graf, gravio, originally
der Graue. But at the head of such senates the
German nations at an early time placed a king. In
Latin the king is called rex, the Sanskrit râjan, in
Maharája, and this rex, the French rot, meant originally
steersman, from regere, to steer. The Teutonic nations,
on the contrary, used the name König, or King,
and this corresponds to the Sanskrit janaka. What
did it mean? It simply meant father, the father of a
255family, ‘the king of his own kin,’ the father of a clan,
the father of a people. Need I add what was the
original, and what is still the true meaning of queen?
In German we have simply formed a feminine of König,
namely, Königin. In English, on the contrary, the old
word for mother has been retained. In the translation
of the Bible by Ulfilas, in the fourth century, we meet
with qens and qino, meaning wife and woman. In the
eleventh century we read in Notker, Sol chena iro
charal furhten unde minnon
, ‘a wife shall fear and
love her husband.’ After the fifteenth century the
word is no longer used in High German, but in the
Scandinavian languages the word still lives, karl and
kona still meaning man and wife.

We thus see how languages reflect the history of
nations, and how, if properly analysed, almost every
word will tell us of many vicissitudes through which
it passed on its way from Central Asia to India or to
Persia, to Asia Minor, Greece, and Italy, to Russia,
Gaul, Germany, the British Isles, America, New Zealand;
nay, back again, in its world-encompassing
migrations, to India and the Himalayan regions from
which it started. Many a word has thus gone the
round of the world, and it may go the same round
again and again. For although words change in
sound and meaning to such an extent that not a
single letter remains the same, and that their meaning
becomes the very opposite of what it originally was,
yet it is important to observe, that since the beginning
of the world no new addition has ever been made to the
substantial elements of speech, any more than to the
substantial elements of nature. There is a constant
change in language, a coming and going of words;
but no man can ever invent an entirely new word.
256We speak to all intents and purposes substantially the
same language as the earliest ancestors of our race;
and, guided by the hand of scientific etymology, we
may pass on from century to century through the
darkest periods of the world's history, till the stream
of language on which we ourselves are moving
carries us back to those distant regions where we
seem to feel the presence of our earliest forefathers,
and to hear the voices of the earth-born sons of Maim.

Those distant regions in the history of language
are, no doubt, the most attractive, and, if cautiously
explored, full of instructive lessons to the historian
and the philosopher. But before we ascend to those
distant heights, we must learn to walk on the smoother
ground of modern speech. The advice of Leibniz
that the science of language should be based on the
study of modern dialects, has been but too much
neglected, and the results of that neglect are visible
in many works on Comparative Philology. Confining
ourselves therefore for the present chiefly to the
modern languages of Europe, let us see how we can
establish the four fundamental points which constitute
the Magna Charta of our science.

1. The same Word takes different Forms in different

This sounds almost like a truism. If the six
dialects which sprang from Latin have become six
independent languages, it would seem to follow that
the same Latin word must have taken a different form
in each of them. French became different from
Italian, Italian from Spanish, Spanish from Portuguese,
because the same Latin words were pronounced
257differently by the inhabitants of the countries
conquered or colonized by Rome, so that, after a
time, the language spoken by the colonists of Gaul
grew to be unintelligible to the colonists of Spain.
Nevertheless if we are told that the French même is
the same as the Italian medesimo, and that both are
derived from the Latin ipse, we begin to see that even
this first point requires to be carefully examined, and
may help to strengthen our arguments against all
etymology which trusts to vague similarity of sound
or meaning.

How then can French même be derived from Latin
ipse? By a process which is strictly genealogical,
and which furnishes us with a safer pedigree than that
of the Montmorencys or any other noble family. In
Old French même is spelt meïsme, which comes very
near to Spanish mismo and Portuguese mesmo. The
corresponding term in Provençal is medesme, which
throws light on the Italian medesimo. Instead of
medesme, Old Provençal supplies smetessme. In order
to connect this with Latin ipse, we have only to consider
that ipse passes through Old Provençal eps into
Provençal eis, Italian esso, Spanish ese, and that the
Old Spanish esora represents ipsâ horâ, as French
encore represents hanc horam. If es is ipse, essme would
be ipsissimum, Provençal medesme, metipsissimum,
and Old Provençal smetessme, semetipsissimum *10

To a certain point it is a matter of historical rather
than of philological inquiry, to find out whether the
English beam is the German Baum. Beam in Anglo-Saxon
is beám, , Frisian bâm, Old Saxon bâm and bôm,
Middle High-German boum, Modern High-German
Baum. It is only when we come to Gothic bagms that
258philological arguments come in, in order to explain
the loss of g before m. This must be explained by a
change of beagm into beawm, and lastly into beam. *11

If we take any word common to all the Teutonic
dialects, we shall find that it varies in each, and that
it varies according to certain laws. Thus, to hear is
in Gothic hausjan, in Old Norse heyra, in Old Saxon
horian, in Anglo-Saxon hyran, in Old High-German
horran, in Swedish höra, in Danish hore, in Dutch
hooren, in Modern German hören.

We have only to remember that English ranges, as
far as its consonants go, with Gothic and Low-German,
while Modern German belongs to the third or High-German
stage, in order to discover without difficulty
the meaning of many a German word by the mere
application of Grimm's Law. Thus: —

tableau I. | II. | III. | Drei is three | Zehn is ten | Tag is day | Du is thou | Zagel is tail | Trommel is drum | Denn is then | Zahn is tooth | Traum is dream | Durch is through | Zaun is town | T(h)euer is dear | Denken is to think | Zinn is tin | T(h)au is dew | Drang is throng | Zerren is to tear | Taube is dove | Durst is thirst | Zange is tong | Teich is dough

If we compare tear with the French larme, a mere
consultation of historical documents would carry us
from tear to the earlier forms, taer, tehr, teher, tæher,
to Gothic tagr. The A. S. tæher, however, carries us
back, even more simply than the Gothic tagr, to the
corresponding form dákry in Greek, and (d)aśru in
Sanskrit. We saw in our last Lecture how every
Greek d is legitimately represented in Anglo-Saxon
by t, and k by h. Hence tosher is dákry. In the
259same manner there is no difficulty in tracing the
French larme back to Latin lacruma. The question
then arises, are dákry and lacruma cognate terms?
The secondary suffix ma in lacruma is easily explained,
and we then have Greek dákry and Latin
lacru, differing only by their initials. Here a phonetic
law must remove the last difference. D, if
pronounced without a will, is apt to lapse into L.
Dákry, therefore, could become lacru, and both can
be derived from a root dak, to bite. *12 Only let it
be borne in mind that although an original d may
dwindle down to l, no l in the Aryan languages
was ever changed into d, and that it would be wrong
to say that l and d are interchangeable.

The following table will show at a glance a few of
the descendants of the Latin preposition ante

ante, before.
It. anzi; Sp. antes; Old Fr. ans, ains (ainsné = aîné, elder).

ante ipsum.
Old Fr. ainçois, before.
It. anziano; Sp. anciano; Fr. ancien, old.

abante, from before.
It. avanti, Fr. avant, before.
It. avanzare; Sp. avanzar; Fr. avancer, to bring forward.
It. vantaggio; Sp. ventaja; Fr. avantage, advantage.

It. davanti; Fr. devant, before.
Fr. devancer, to get before.

If instead of a Latin we take a Sanskrit word, and
follow it through all its vicissitudes from the earliest
to the latest times, we see no less clearly how inevitably
260one and the same word assumes different
forms in different dialects. Tooth in Sanskrit is
dat (nom. dantaḥ, but genitive of the old base, dataḥ).
The same word appears in Latin as dens, dentis, in
Gothic as tunthus, in English as tooth, in Modern
German as Zahn. All the changes are according
to law, and it is not too much to say that in the
different languages the common word for tooth could
hardly have appeared under any form but that in
which we find it. But is the Greek odoús, odóntos,
the same word as dens? And is the Greek odóntes,
the Latin dentes, a mere variety of edontes and edentes,
the eaters? I am inclined to admit that the o in
odóntes is a merely phonetic excrescence, for although
I know of no other well-established case in Greek
where a simple initial d assumes this prosthetic
vowel, it would be against all rules of probability to
suppose that Greek had lost the common Aryan term
for teeth, danta, and replaced it by a new and independent
word so exactly like the one which it had
given up. Prosthetic vowels are very common in
Greek before certain double consonants, and before
r, l, n, m *13 The addition of an initial o in odóntes
may provisionally be admitted. But if so, it follows
that odóntes cannot be a mere variety of edontes. For
wherever Greek has these initial vowels, while they
are wanting in Sanskrit, Latin, &c., they are, in the
true sense of the word, prosthetic vowels. They are
not radical, but merely adscititious in Greek, while if
odóntes were derived from the root ed, we should have
to admit the loss of a radical initial vowel in all the
261members of the Aryan family except Greek — an
admission unsupported by any analogy. *14

In languages which possess no ancient literature
the charm of tracing words back from century to
century to its earliest form is of course lost. Contemporary
dialects, however, with their extraordinary
varieties, teach us even there the same lessons, showing
that language must change and is always changing,
and that similarity of sound is the same unsafe guide
here as elsewhere. One instance must suffice. Man
in Malay is orang; hence orang utan, the man of the
forest, the Orangutang. This orang is pronounced in
different Polynesian dialects, rang, oran, olan, Ian,
ala, la, na, da, ra. 15

We now proceed to a consideration of our second

2. The same Word takes different Forms in the same

There are, as you know, many Teutonic words
which, through two distinct channels, found their way
twice into the literary language of Chaucer, Shakespeare,
and Milton. They were imported into England
at first by Saxon pirates, who gradually dislodged the
Roman conquerors and colonists from their castra
and coloniæ, and the Welsh inhabitants from their
villages, and whose language formed the first permanent
stratum of Teutonic speech in these islands.
They introduced such words as, for instance, weardian,
to ward, wile, cunning, wise, manner. These words
were German words, peculiar to that soft dialect of
262German which is known by the name of Low German,
and which was spoken on those northern coasts from
whence the Juts, the Angles, and Saxons embarked
on their freebooting expeditions.

Another branch of the same German stem was the
High German, spoken by the Franks and other Teutonic
tribes, who became the conquerors of Gaul, and
who, though they adopted in time the language of
their Roman subjects, preserved nevertheless in their
conversational idiom a large number of their own
home-spun words. The French or Frankish language
is now a Romanic dialect, and its grammar is but a
blurred copy of the grammar of Cicero. But its
dictionary is full of Teutonic words, more or less
Romanized to suit the pronunciation of the Roman
inhabitants of Gaul. Among warlike terms of German
origin, we find in French guerre, the same as war;
massacre, from metzeln, to cut down, or metzgen, to
butcher; maçon, Metze, Stein-metze, i.e. stone-cutter;
auberge, Italian albergo, the German Herberge, barracks
for the army, Old High-German heriberga; bivouac,
the German Beiwacht; boulevard, German Bollwerk;
bourg, German Burg; breche, a breach, from brechen;
havresac, German Hafersack; haveron, Old High-German
habaro, oats; canapsa, the German Knapp-sack,
Ess-sack, from knappen, knabern, or Schnapp-sack; *16
éperon, Italian sperone, German Sporn; héraut,
Italian araldo, German Herold, i.e. Heerwalt, or from
Old High-German haren, French harer, to call; maréchal,
Old German mariscalco.

Many maritime words, again, came from German,
263more particularly from Low German. French chaloupe
= Sloop, Dutch sloep; cahute = Dutch kajuit,
German Kaue, or Koje; stribord, the right side of
a ship, English starboard, Anglo-Saxon steorbord,
Steuerbord; hâvre, Hafen; Nord, Sud, Est, Ouest, all
come from German.

But much commoner words are discovered to be
German under a French disguise. Thus, haie, hedge,
is Hecke; hair, to hate, Anglo-Saxon hatian; hameau,
hamlet, Heim; hater, to haste; honnir, to blame,
Gothic háunjan, höhnen; harangue, (h)ring, as in ring-leader.
The initial h betrays the German origin of
all these words. Again, choisir, to choose, is kiesen,
A. S. ceósan, Gothic kiusan, or Gothic kausjan, to examine;
danser, tanzen; causer, to chat, kosen; dérober,
to rob, rauben; éier, to spy, späen; gratter, kratzen;
grimper, to climb, klimmen; grincer, grinsen, or Old
High-German grimisôn; gripper, greifen; rôtir, rösten;
tirer, to tear; tomber, to tumble; guinder, to wind;
déguerpir, to throw away, werfen *17

It was this language, this Germanized Latin, which
was adopted by the Norman invaders of France, themselves
equally Teutonic, and representing originally
that third branch of the Teutonic stock of speech
which is known by the name of Scandinavian. These
Normans, or Northmen, speaking their newly-acquired
Franco-Roman dialect, became afterwards the victors
of Hastings, and their language, for a time, ruled
supreme in the palaces, law-courts, churches, and
colleges of England. The same thing, however, which
had happened to the Frank conquerors of Gaul and
the Norman conquerors of Neustria happened again
264to the Norman conquerors of England. They had to
acquire the language of their conquered subjects;
and as the Franks, though attempting to speak the
language of the Roman provincials, retained large
numbers of barbaric terms, the Normans, though
attempting to conform to the rules of the Saxon
grammar, retained many a Norman word which they
had brought with them from France.

Thus the German word wise was common to the
High and the Low branches of the German language;
it was a word as familiar to the Frank invaders of Gaul
as it was to the Saxon invaders of England. In the
mouths of the Roman citizens of France, however, the
German initial W had been replaced by the more guttural
sound of gu. Wise had become guise, and in this
new form it succeeded in gaining a place side by side
with its ancient prototype, wise. By the same process
guile, the Old French guile, was adopted in English,
though it was the same word originally as the Anglo-Saxon
wile, which we have in wily. The changes
have been more violent through which the Old High-German
wetti, a pledge (Gothic vadi), became changed
into the mediæval Latin wadium or vadium *18 Italian
gaggio, and French gage. Nevertheless, we must recognise
in the verbs to engage or disengage Norman
varieties of the same word which is preserved in the
pure Saxon forms to bet and to wed, literally to bind
or to pledge.

There are many words of the same kind which
have obtained admittance twice into the language of
England, once in their pure Saxon form, and again
in their Roman disguise. Words beginning in Italian
265with gua, gue, gui, are almost invariably of German
origin. A few words are mentioned, indeed, in
which a Latin v seems to have been changed into g.
But as, according to general usage, Latin v remains v
in the Romance dialects, it would be more correct to
admit that in these exceptional cases Latin words had
first been adopted and corrupted by the Germans, and
then, as beginning with German w, and not with
Latin v, been readopted by the Roman provincials.

These exceptional cases, however, are very few, and
somewhat doubtful. It was natural, no doubt, to
derive the Italian guado, a ford, the French gué, from
Latin vadum. Yet the initial gua points first to
German, and there we find in Old High-German wat,
a ford, watan, to wade. The Spanish vadear may be
derived from Latin, or it may owe its origin to a
confusion in the minds of those who were speaking
and thinking in two languages, a Teutonic and a
Romanic. The Latin vadum and the German wat
may claim a distant relationship.

Guère in je ne crois guère was for a time traced
back to parum, varium, valide, avare, or grandem rem,
the Provençal granren. But, like the Italian guari,
it comes from wâri, true, which gradually assumed
the meaning of very. *19 The Latin verus changes to
vero and vrai.

Guastare, French gâter, has been traced back to
Latin vastare; but it is clearly derived from Old
High-German wastjan, to waste, though again a confusion
of the two words may be admitted in the minds
of the bilingual Franks.266

Guêpe, wasp, is generally derived from vespa; it
really comes from the German Wespe. *20

It has frequently been pointed out that this very
fact, the double existence of the same word (warden
and guardian, &c.), has added much to the strength and
variety of English. Slight shades of meaning can thus
be kept distinct, which in other languages must be
allowed to run together. The English brisk, frisky,
and fresh, all come from the same source. 21 Yet
there is a great difference between a brisk horse, a
frisky horse, and a fresh horse — a difference which it
would be difficult to express in any other language.
It is a cause of weakness in language if many ideas
have to be expressed by the same word, and fresh in
English, though relieved by brisk, and frisky, embraces
still a great variety of conceptions. We hear
of a fresh breeze, of fresh water (opposed to stagnant),
of fresh butter, of fresh news, of a fresh hand,
a freshman, of freshness of body and mind; and such
a variation as a brisk fire, a brisk debate, is therefore
all the more welcome. Fresh has passed through a
Latin channel, as may be seen from the change of its
vowel, and to a certain extent from its taking the
suffix ment in refreshment, which is generally, though
not entirely, restricted to Latin words. 22 Under a thoroughly
foreign form it exists in English as fresco, in
267fresco-paintings, so called because the paint was applied
to the walls whilst the plaster was still fresh or damp.

The same process explains the presence of double
forms, such as ship and skiff, the French esquif; from
which is derived the Old French esquiper, the Modern
French équiper, the English to equip. Or again, sloop
and shallop, the French chaloupe.

Thus bank and bench are German; banquet is German

Bar is German (O. H. G. para); barrier is Romanized.
Cf. Span, barras, a bar, French embarras,
and English embarrassed.

Ball is German; balloon Romanized.

To pack is German; bagage Romanized.

Ring, a circle, is German; O. H. G. hring. To harangue,
to address a ring, to act as a ringleader, is
Romanized; It. aringa, Fr. la harangue.

Sometimes it happens that the popular instinct
of etymology reacts on these Romanized German
words, and, after tearing off their foreign mask, restores
to them a more homely expression. Thus the
German Krebs, the O. H. G. krebiz, is originally the
same word as the English crab. This krebiz appears
in French as écrevisse; it returned to England in this
outlandish form, and was by an off-hand etymology
reduced to the Modern English crayfish.

Thus filibuster seems to be derived from the Spanish
filibote or flibote, but the Spanish word itself was a
corruption of the English fly-boat.

And as the German elements entered into the English
language at various times and under various
forms, so did the Latin. Latin elements flowed into
England at four distinct periods, and through four
distinct channels.268

First, through the Roman legions and Roman
colonists, from the time of Cæsar's conquest, 55 B. C.,
to the withdrawal of the Roman legions in 412: e. g.
colonia = coln; castra = chester; stratum = street.

Secondly, through the Christian missionaries and
priests, from the time of St. Augustine's landing in
597 to the time of Alfred: e. g. candela = candle;
Kyriake = church; diaconus = dean; regular = rule; corona =
crown; discus = dish; uncia = inch.

Thirdly, through the Norman nobility and Norman
ecclesiastics and lawyers, who, from the days of Edward
the Confessor, brought into England a large
number of Latin terms, either in their classical or in
their vulgar and Romanized form.

Fourthly, through the students of the classical
literature of Rome, since the revival of learning to the
present day. These repeated importations of Latin
words account for the coexistence in English of
such terms as minster and monastery. Minster found
its way into English through the Christian missionaries,
and is found in its corrupt or Anglicized
form in the earliest documents of the Anglo-Saxon
language. Monastery was the same word, only pronounced
with less corruption by later scholars, or
clergymen, familiar with the Latin idiom. Thus
paragraph is the Latin paragraphus, but slightly
altered; pilcrow, pylcrafte, and paraf, are vulgar corruptions
of the same word. *23 In a similar way, the
verb to blame became naturalized in England through
the Norman Conquest. The original Latin or Greek
word from which the French blâmer was derived
kept its place in the form of to blaspheme in the
269more cultivated language of the realm. Triumph was
a Latin word, naturally used in the ecclesiastical
and military language of every country. In its degraded
form, la triomphe, it was peculiar to French,
and was brought into England by the Norman nobility
as trump, trump card. *24 We can watch the same
process more fully in the history of the French
language. That language teems with Latin words
which, under various disguises, obtained repeated admittance
into its dictionary. They came first with
the legions that settled in Gaul, and whose more or
less vulgar dialects supplanted the Celtic idiom of the
country. They came again in the track of Christian
missionaries, and not unfrequently were smuggled in
for the third time by the classical scholars of a later
age. The Latin sacramentum, in its military acceptation,
became the French serment; in its ecclesiastical
meaning it appears as sacrement. Redemptio, in its
military sense, became the French rançon, ransom; in
its religious meaning it preserved the less mutilated
form of redemption. Other words belonging to the
same class are acheter, 25 to buy, accepter, to accept, both
derived from the Latin acceptare. Chétif, miserable,
captif, both from Latin captivus. Chose, a thing,
cause, a cause, both from Latin causa. Façon and
faction, from Latin factio; meaning originally the
manner of doing a thing, then peculiarity, then party.
Both fraile and fragile come from fragilis. On and
l'homme, from homo. Noël, Christmas, and natal, from
natalis. Naïf and natif from nativus. Parole and
parabole from parabola. Penser, to weigh or ponder
270in one's mind, and peser, to weigh on scales, both come
from Latin pensare. Pension also is derived from
pensum. In Latin, too, expendo is used in the sense of
spending money, and of weighing or considering.

The Latin pronoun ille exists in French under two
different forms. It is the il of the pronoun of the
third person, , and the le of the definite article. Of
course it must not be supposed for a moment that by
any kind of agreement ille was divided into two parts,
il being put aside for the pronoun, and le for the
article. The pronoun il and elle in French, egli and
ella in Italian, el and ella in Spanish, are nothing but
provincial varieties of ille and illa. The same words,
ille and illa, used as articles, and therefore pronounced
more rapidly and without an accent, became gradually
changed from il, which we see in the Italian il to el,
which we have in Spanish; to lo (illum), which exists
in Provençal and in Italian (lo spirito); and to le,
which appears in Provençal *26 dialects and in French.

As there are certain laws which govern the transition
of Latin into French and Italian, it is easy to
determine whether such a word as opera in French is
of native growth, or imported from Italian. French
has invariably shortened the final a into e, and a
Latin p in the middle of words is generally changed
into French b or v. This is not the case in Italian.
Thus the Latin apis, a bee, becomes in Italian ape,
in French abeille. 27 The Latin capillus is the Italian
capello, the French cheveu. Thus opéra has become
271œuvre in French, whereas in Italian it remained opera *28
Spanish obra.

There is a small class of words in French which
ought to be mentioned here, in order to show under
how many disguises words have slipped in again
and again into the precincts of that language. They
are words neither Teutonic nor Romance, but a
cross between the two. They are Latin in appearance,
but it would be impossible to trace them
back to Latin unless we knew that the people who
spoke this Latin were Germans who still thought in
German. If a German speaks a foreign tongue, he
commits certain mistakes which a Frenchman never
would commit, and vice versâ. A German speaking
English would be inclined to say to bring a sacrifice;
a Frenchman would never make that mistake. A
Frenchman, on the contrary, is apt to say that he
cannot attend any longer, meaning that he cannot
wait any longer. Englishmen, again, travelling abroad,
have been heard to call for Wächter, meaning the
waiter; they have declared, in German, Ich habe einen
grossen Geist Sie nieder zu klopfen
, meaning they had
a great mind to knock a person down; and they have
announced in French, J'ai changé mon esprit autour
272de cette tasse de café
, meaning that they had changed
their mind about a cup of coffee.

There are many more mistakes of that kind, which
grammarians call Germanisms, Gallicisms, or Anglicisms,
and for which pupils are constantly reproved
by their masters.

Now the Germans who came to settle in Italy and
Gaul, and who learnt to express themselves in Latin
tant bien que mal, had no such masters to reprove
them. On the contrary, their Roman subjects did
the best they could to understand their Latin jargon,
and, if they wished to be very polite, they would probably
repeat the mistakes which their masters had
committed. In this manner the most ungrammatical,
the most unidiomatic phrases would, after a time,
become current in the vulgar language.

No Roman would have expressed the idea of entertaining
or amusing by intertenere. Such an expression
would have conveyed no meaning at all to Cæsar or
Cicero. The Germans, however, were accustomed to
the idiomatic use of unterhalten, Unterhaltung, and
when they had to make themselves understood in
Latin they rendered unter by inter, halten by tenere,
and thus formed entretenir, a word owned neither by
Latin nor German.

It is difficult, no doubt, to determine in each case
whether words like intertenere, in the sense of entertaining,
were formed by Germans speaking in Latin
but thinking in German, or whether one and the same
metaphor suggested itself both to Romans and Germans.
It might seem at first sight that the French
circonstance, circumstance, was a barbarous translation
of the German Umstand, which expresses the same
273idea by exactly the same metaphor. But if we consult
the later Latin literature, we find there, in
works which could hardly have experienced any influence
of German idiom, circumstantia, in the sense
of quality or accident, and we learn from Quintilian,
v. 10, 104, that the word had been formed in Latin as
an equivalent of the Greek perístasis.

In some cases, however, it admits of no doubt that
words now classical in the modern languages of Europe
were originally the unidiomatic blunders of Germans
attempting to express themselves in the Latin of their
conquered provinces.

The future is called in German Zukunft, which
means ‘what is to come.’ *29 There is no such word
in ancient Latin, but the Germans again translated
their conception of future time literally into Latin,
and thus formed l'avenir, what is to come, ce qui est
à venir

One of the many German expressions for sick or
unwell is unpäss. It is used even now, unpässlich,
Unpässlichkeit. The corresponding Latin expression
would have been æger, but instead of this we find
the Provençal malapte, It. malato, Fr. malade. Mal-apte
is the Latin male-aptus, meaning unfit, again an
unidiomatic rendering of unpass. What happened was
this. Male-aptus was at first as great a mistake in
Latin as if a German speaking English were to take
unpass in the sense of unpassend, and were to say,
‘that he was unfit,’ meaning he was unwell. But as
there was no one to correct the German lords and
masters, the expression male-aptus was tolerated, was
274probably repeated by good-natured Roman physicians,
and became after a time a recognised term.

One more word of the same kind, the presence of
which in French, Italian, and English it would be
impossible to explain except as a Germanism, as a
blunder committed by people who spoke in Latin,
but thought in German.

Gegend in German means region or country. It is
a recognised term, and it signified originally that
Avhich is before or against, what forms the object of
our view. Now in Latin gegen, or against, would be
expressed by contra; and the Germans, not recollecting
at once the Latin word regio, took to transLating
their idea of Gegend, that which was before them, by
contratum, or terra contrata. This became the Italian
contrada, the French contrée, the English country. *30275

And here, in discussing words which, though originally
distinct in origin and meaning, have in the
course of time become identical or nearly identical in
sound, I ought not to pass over in silence the name of
a scholar who, though best known in the annals of the
physical sciences, deserves an honourable place in the
history of the Science of Language. Roger Bacon's
views on language and etymology are strangely in
advance of his age. He called etymology the tale of
truth, *31 and he was probably the first who conceived,
the idea of a Comparative Grammar. He uses the
strongest language against those who proposed derivations
of words in Latin, Greek, and Hebrew without
a due regard to the history of these languages.
‘Brito,’ he says, ‘dares to derive Gehenna from the
Greek ge, earth, and ennos, deep, though Gehenna is a
Hebrew word, and cannot have its origin in Greek.’ 32
As an instance of words becoming identical in the
course of time, he quotes kenon as used in many
276mediæval compounds. In cenotaph, an empty tomb,
ceno represents the Greek κενός, empty. In cenobite,
one of a religious order living in a convent, ceno is
the Greek κοινός, common. In encenia, festivals kept
in commemoration of the foundation of churches, &c.,
cenia answers to the Greek καινός, new, these festivals
being intended as renewals of the memory of pious
founders. *33 Surely this does honour to the thirteenth

Accidents like those which we have hitherto discussed
are, no doubt, more frequent in the modern
history of speech, because, owing to ethnic migrations
and political convulsions, the dialects of neighbouring
or distant races have become mixed up
together more and more with every century that has
passed over the ethnological surface of Europe. But
in ancient times also there had been migrations, and
wars, and colonies, causing a dislocation and intermixture
of the various strata of human speech, and
the literary languages of Greece and Rome, however
uniform they may seem to us in their classical writings,
277had grown up, like French or English, by a constant
process of absorption and appropriation, exercised on
the various dialects of Italy and Greece. What
happened in French happened in Latin. As the
French are no longer aware that their paysan, a
peasant, and païen, a pagan, were originally but
slight dialectic varieties of the same Latin word
paganus, a villager, the citizen of Rome used the two
words luna, moon, and Lucina, the goddess, without
being aware that both were derived from the same
root. In luna the c belonging to the root lucere, to
shine, is elided; not by caprice or accident, but
according to a general phonetic rule which requires
the omission of a guttural before a liquid. Thus
lumen, light, stands for lucmen; examen for exagmen;
flamma, flame, for flagma, from flagare, to burn;
flamen for flagmen, the lighter, the priest (not brahman);
lanio, a butcher, if derived from a root akin to
lacerare, to lacerate, stands for lacnio. Contaminare,
to contaminate, is certainly derived from the same verb
tango, to touch, from which we have contagio, contagion,
as well as integer, intact, entire. Contaminare,
therefore, was originally contagminare. This is in fact
the same phonetic rule which, if applied to the Teutonic
languages, accounts for the change of German
Nagel into nail, Zagel into tail, Hagel into hail, Riegel
into rail, Regen into rain, Pflegel into flail, Segel into
sail; and which, if applied to Greek and Latin, helps
us to discover the identity of the Greek láchnē, wool,
and Latin lâna; of Greek àráchnē, a spider, and Latin
arânea. Though a scholar like Cicero *34 might have
278been aware that ala, a wing, was but an abbreviated
form of axilla, the arm-pit, the two words were as
distinct to the common citizen of Borne as païen and
paysan to the modern Frenchman. Tela, a web,
must, on the same principle, be derived from texela,
and this from the verb texere, to weave. Thus mala,
the cheek, is derived from maxilla, the jawbone, and
velum, a sail or. veil, from vexillum, anything flying or
moved by the wind, a streamer, a flag, or a banner.
Once in possession of this rule, we are able to discover
even in such modern and corrupt forms as subtle, the
same Latin root texere, to weave, which appeared in
tela. From texere was formed the Latin adjective
subtilis, that which is woven under or beneath, with
the same metaphor which leads us to say fine spun;
and this dwindled down into the English subtle.

Other words in Latin, the difference of which must
be ascribed to the influence of local pronunciation, are
cors and cohors, nil and nihil, mi and mihi, prendo and
prehendo, prudens and providens, bruma, the winter
solstice, and brevissima, scil. dies, the shortest day. *35
Thus, again, susum stands for sursum, upward, from
sub and versum. Sub, it is true, means generally
below, under; but, like the Greek hypó, it is used in
the sense of ‘from below,’ and thus may seem to
have two meanings diametrically opposed to each
other, below and upward. Submittere means to place
below, to lay down, to submit; sublevare, to lift from
below, to raise up. Summus, a superlative of sub,
hýpatos, a superlative of hypó, do not mean the lowest
279but the highest. *36 As sub-versum glides into sursum
and susum, so retroversum becomes retrorsum, retrosum,
and rursum. Proversum becomes prorsum, originally
forward, straightforward; and hence oratio
, straightforward speech or prose, opposed to
oratio vincta, fettered or measured speech, poetry. 37

Now as we look upon Æolic and Doric, Ionic and
Attic, as dialects of one and the same language, as we
discover in the Romance languages mere varieties of the
Latin, and in the Scandinavian, the High German, and
Low German, only three branches of one and the same
stock, we must learn to look upon Greek and Latin,
Teutonic and Celtic, Slavonic, Sanskrit, and the ancient
Persian, as so many varieties of one and the same
original type of speech, which were fixed in the end
as the classical organs of the literature of the world.
Taking this point of view, we shall be able to understand
how what happens in the modern, happened in
the ancient periods of the history of language. The
same word, with but slight dialectic variations, exists
in Greek, Latin, Gothic, and Sanskrit, and vocables
which at first sight appear totally different, are
separated from each other by no greater difference
than that which separates an Italian word from its
cognate term in French. There is little similarity to
the naked eye between pen and feather, yet if placed
under the microscope of comparative grammar, both
words disclose exactly the same structure. Both are
derived from a root pat, which in Sanskrit means to
fly, and which is easily recognised in the Greek pétomai,
I fly. From this root a Sanskrit word is derived by
280means of the instrumental suffix tra, pat-tra, or patatra,
meaning the instrument of flying, a wing, or a
feather. From the same root another substantive
was derived, which became current in the Latin dialect
of the Aryan speech, patna or petna, meaning equally
an instrument of flying, or a feather. This petna
became changed into penna — a change which rests
not merely on phonetic analogy, but is confirmed by
Festus, who mentions the intermediate Italian form,
pesna *38 The Teutonic dialect retained the same derivative
which we saw in Sanskrit, only modifying its
pronunciation by substituting aspirated for hard consonants,
according to rule. Thus patra had to be
changed into phathra, in which we easily recognise
the English feather. Thus pen and feather, the one
from a Latin, the other from a Teutonic source, are
established as merely phonetic varieties of the same
word, analogous in every respect to such double words
as those which we pointed out in Latin, which we saw
in much larger numbers in French, and which impart
not only the charm of variety, but the power of
minute exactness to the language of Chaucer, Shakespeare,
and Milton.

3. Different Words take the same Form in different

We have examined in full detail two of the propositions
which serve to prove that in scientific etymology
identity of origin is in no way dependent on identity
of sound or meaning. If words could for ever retain
their original sound and their original meaning,
281language would have no history at all; there would
have been no confusion of tongues, and our language
would still be the language of our first ancestors.
But it is the very nature of language to grow and to
change, and unless we are able to discover the rules of
this change, and the laws of this growth, we shall never
succeed in tracing back to their original source and
primitive import the manifold formations of human
speech, scattered in endless variety over all the
villages, towns, countries, and continents of our globe.
The radical elements of language are so extremely
few, and the words which constitute the dialects
of mankind so countless, that unless it had been
possible to express the infinitesimal shades of human
thought by the slightest differences in derivation or
pronunciation, we should never understand how so
colossal a fabric could have been reared from materials
so scanty. Etymology is the knowledge of the
changes of words, and so far from expecting identity,
or even similarity of sound in the outward appearance
of a word, as now used in English, and as used
by the poets of the Veda, we should always be on
our guard against any etymology which would fain
make us believe that certain words which exist in
French existed in exactly the same form in Latin, or
that certain Latin words could be discovered without
the change of a single letter in Greek or Sanskrit.
If there is any truth in the laws which govern the
growth of language, we can lay it down with perfect
certainty, that words of identically the same sound in
English and in Sanskrit cannot be the same words.
And this leads us to our third proposition. It does
happen now and then that in languages, whether
282related to each other or not, certain words appear of
identically the same sound and with some similarity
of meaning. These words, which former etymologists
seized upon as most confirmatory of their views,
are now looked upon with well-founded mistrust.
Attempts, for instance, are frequently made at comparing
Hebrew words with the words of Aryan
languages. If this is done with a proper regard to
the immense distance which separates the Semitic
from the Aryan languages, it deserves the highest
credit. But if instead of being satisfied with pointing
out the faint coincidences in the lowest and most
general elements of speech, scholars imagine they can
discover isolated cases of minute coincidence amidst
the general disparity in the grammar and dictionary
of the Aryan and Semitic families of speech, their
attempts become unscientific and reprehensible.

It is surprising, considering the immense number
of words that might be formed by freely mixing the
twenty-five letters of our alphabet, that in languages
belonging to totally different families, the same ideas
should sometimes be expressed by the same or very
similar words. Dr. Rae, in order to prove some kind
of relationship between the Polynesian and Aryan
languages, quotes the Tahitian pura, to blaze as a fire,
the New Zealand kapura, fire, as similar to Greek
pyr, fire. He compares Polynesian ao, sunrise, with
Eos; Hawaian mauna with mons; Hawaian ike, he saw
or knew, with Sanskrit îksh, to see; manao, I think,
with Sanskrit man, to think; noo, I perceive, and
noo-noo, wise, with Sanskrit jnâ, to know; orero or
orelo, a continuous speech, with oratio; kala, I proclaim,
with Greek kaleîn, to call; kalanga, continuous
283speech, with harangue; kani and kakani, to sing, with
cano; mele, a chaunted poem, with mélos. *39

It is easy to multiply instances of the same kind.
Thus in the Kafir language to beat is beta, to tell is
tyelo, hollow is uholo. 40

In Modern Greek eye is mati, a corruption of ommation;
in Polynesian eye is mata, and in Lithuanian
matau is to see.

And what applies to languages which, in the usual
sense of the word, are not related at all, such as
Hebrew and English, or Hawaian and Greek, applies
with equal force to cognate languages. Here, too, a
perfect identity of sound between words of various
dialects is always suspicious. No scholar would now-a-days
venture to compare to look with Sanskrit
lokayati; to speed with Greek speúdō; to call with
Greek kaleîn; to care with Latin cura. The English
sound of i which in English expresses an eye, oculus,
is used in German in the sense of egg, ovum; and it
would not be unreasonable to take both words as
expressive of roundness, applied in the one case to an
egg, in the other to an eye. The English eye, however,
must be traced back to the Anglo-Saxon edge, Gothic
augô, German Auge, words akin to Sanskrit akshi, the
Latin oculus, the Greek ὄsse; whereas the German Ei,
which in Old High-German forms its plural eigir, is
identical with the English egg, the Latin ovum, the
Greek ōfon, and possibly connected with avis, bird.
284This Anglo-Saxon edge, eye, dwindles down to y in
daisy, and to ow in window, supposing that window is
the Old Norse vindauga, the Swedish vindöga, the Old
English windor. *41 In Gothic a window is called augadauro,
in Anglo-Saxon, eágduru, i.e. eye-door. In
island (which ought to be spelt iland), the first portion
is neither egg nor eye, but a corruption of Gothic
ahva, i.e. aqua, water; hence Anglo-Saxon eóland, the
Old Norse aland, waterland.

What can be more tempting than to derive ‘on
the whole
’ from the Greek kath hólon, from which Catholic? 42
Buttmann, in his ‘Lexilogus,’ has no misgivings
whatever as to the identity of the Greek hólos and
the English hale and whole and wholesome. At present,
a mere reference to ‘Grimm's Law’ enables any tyro
in etymology to reject this identification as impossible.
First of all, whole, in the sense of sound, is
really the same word as hale. Both exist in Anglo-Saxon
under the form of hâl, in Gothic as hail,
German heil. 43 Now, an initial aspirate in Anglo-Saxon
or Gothic presupposes a tenuis in Greek, and
if, therefore, the same word existed in Greek, it could
only have been hólos, not hólos.

In hólos the asper points to an original s in Sanskrit
and Latin, and hólos has therefore been rightly identified
with Sanskrit sarva and Latin salvus and sollus, in
sollers, sollemnis, solliferreus, &c.

There is perhaps no etymology so generally
acquiesced in as that which derives God from good.
In Danish good is god, but the identity of sound
285between the English God and the Danish god is merely
accidental; the two words are distinct, and are kept
distinct in every dialect of the Teutonic family. As
in English we have God and good, we have in Anglo-Saxon
God and god; in Gothic, Guth and god; in Old
High-German, Cot and cuot; in German, Gott and gut;
in Danish, Gud and god; in Dutch, God and goed.
Though it is impossible to give a satisfactory etymology
of either God or good, it is clear that two words
which thus run parallel in all these dialects without
ever meeting, cannot be traced back to one central
point. God was most likely an old heathen name of
the Deity, and for such a name the supposed etymological
meaning of good would be far too modern,
too abstract, too Christian. *44 In the Old Norse, Gođ is
actually found in the sense of a graven image, an
idol, and is then used as a neuter, whereas, in the
same language, Guđ as a masculine, means God.
When, after their conversion to Christianity, the
Teutonic races used God as the name of the true
God, in the same manner as the Romanic nations
retained their old heathen word Deus, we find that in
Old High-German a new word was formed for false
gods or idols. They were called apcot, as if ex-gods.
The Modern German word for idol, Götze, is but a
modified form of God, and the compound Oelgötze,
which is used in the same sense, seems actually to
point back to ancient stone idols, before which, in the
days of old, lamps were lighted and incense burned.
Luther, in transLating the passage of Deuteronomy,
286‘And ye shall hew down the graven images of their
gods,’ uses the expression, ‘die Götzen ihrer Götter.’

What thus happens in different dialects may happen
also in one and the same language; and this leads us to
the consideration of our fourth and last proposition.

4. Different Words may take the same Form in one
and the same Language.

The same causes which make words which are
perfectly distinct in their origin to assume the same,
or very nearly the same sound in English and German,
may produce a similar convergence between two words
in one and the same language. Nay, the chances are,
if we take into account the peculiarities of pronunciation
and grammar in each dialect, that perfect
identity of sound between two words, differing in
origin, will occur more frequently in one and the
same than in different dialects. It would seem to
follow, also, that these cases of verbal convergence are
more frequent in modern than in ancient languages;
for it is only by a constant process of phonetic corruption,
by a constant wearing off of the sharp edges
of words, that this verbal assimilation can be explained.
Many words in Latin differ by their terminations
only; these terminations were generally omitted in the
modern Romance dialects, and the result is, that these
words are no longer distinguishable in sound. Thus
novus in Latin means new; novem, nine; the terminations
being dropped, both become in French neuf.
Suum, his, is pronounced in French son; sonum, sound,
is reduced to the same form. In the same manner
tuum, thine, and tonus, tone, become ton. The French
feu, fire, is the Latin focus; feu, in the sense of late,
287is not exactly Latin — at least, it is derived from Latin
in the most barbarous way. In the same manner as
we find in Spanish somos, sois, son, where sois stands
ungrammatically for Latin estis; as in the same language
a gerund siendo is formed which would seem
to point to a barbarous Latin form, essendo, so a past
participle fuitus may have been derived from the Latin
perfect fui, I was; and this may have given rise to
the French feu, late. Hence we find both feu la reine
and la feue reine.

It sometimes happens that three Latin words are
absorbed into one French sound. The sound of mer
conveys in French three distinct meanings; it means
sea, mother, and mayor. Suppose that French had
never been written down, and had to be reduced to
writing for the first time by missionaries sent to Paris
from New Zealand, would not mer, in their dictionary
of the French language, be put down with three distinct
meanings — meanings having no more in common
than the explanations given in some of our old
Greek and Latin dictionaries? It is no doubt one of
the advantages of the historical system of spelling that
the French are able to distinguish between la mer,
mare, le maire, major, la mère, mater; yet if these
words produce no confusion in the course of a rapid
conversation, they would hardly be more perplexing
in reading, even though written phonetically.

There are instances where four and five words, all
of Latin origin, have dwindled away into one French
term. Ver, the worm, is Latin vermis; vers, a verse,
is Latin versus; verre, a glass, is Latin vitrum; vert,
green, is Latin viridis; vair, fur, is Latin varius.
Nor is there any difference in pronunciation between
the French mat, the month of May, the Latin majus;
288mais, but, the Latin magis; mes, the plural of my,
Latin mei; and la maie , a trough, perhaps the Latin
mactra; or between sang, blood, sanguis; cent, a
hundred, centum; sans, without, sine; sent, he feels,
sentit; s'en, in il s'en va, inde.

Where the spelling is the same, as it is, for instance,
in louer, to praise, and louer, to let, attempts have not
been wanting to show that the second meaning was
derived from the first; that louer, for instance, was
used in the sense of letting, because you have to praise
your lodgings before you can let them. Thus fin,
fine, was connected with fin, the end, because the end
occasionally expresses the smallest point of an object.
Now, in the first instance, both louer, to let, and louer,
to praise, are derived from Latin; the one is laudare,
the other locare. In the other instance we have to
mark a second cause of verbal confusion in French.
Two words, the one derived from a Latin, the other
from a German source, met on the neutral soil of
France, and, after being divested of their national
dress, ceased to be distinguishable from each other.
The same applies to the French causer. In one sense
it is the Latin causare, to cause; in another, the Old
German chôsôn, the Modern German kosen. As French
borrows not only from German, but also from Greek,
we need not be surprised if in le page, page, we meet
with the Greek paídion, a small boy, whereas la page
is the Latin página, a page or leaf.

There are cases, however, where French, Italian, and
Spanish words, though apparently invested with two
quite heterogeneous meanings, must nevertheless be
referred to one and the same original. Voler, to fly, is
clearly the Latin volare; but voler, to steal, would seem
at first sight to require a different etymology. There
289is, however, no simple word, whether in Latin, or
Celtic, or Greek, or German, from which voler, to steal,
could be derived. Now, as we observed that the same
Latin word branched off into two distinct French
words by a gradual change of pronunciation, we must
here admit a similar bifurcation, brought on by a
gradual change of meaning. It would not, of course,
be satisfactory to have recourse to a mere gratuitous
assumption, and to say that a thief was called volator,
a flyer, because he flew away like a bird from his
pursuers. But Professor Diez has shown that in Old
French, to steal is embler, which is the mediæval Latin
imbulare, used, for instance, in the Lex Salica. This
imbulare is the genuine Latin involare, which is used
in Latin of birds flying down, *45 of men and women
flying at each other in a rage, 46 of soldiers dashing
upon an enemy, 47 and of thieves pouncing upon a thing
not their own. §48 The same involare is used in Italian
in the sense of stealing, and in the Florentine dialect
it is pronounced imbolare, like the French embler. It
was this involare, with the sense of seizing, which
was abbreviated to the French voter. Voter, therefore,
meant originally, not to fly away, but to fly upon, just
as the Latin impetus, assault, is derived from the root
pat, to fly, in Sanskrit, from which we derived penna and
feather. A complete dictionary of words of this kind
290in French has been published by M. E. Zlatagorskoi,
under the title, ‘Essai d'un Dictionnaire des Homonymes
de la Langue Française’ (Leipzig, 1862), and
a similar dictionary might be composed in English.
For here, too, we find not only Romance words
differing in origin and becoming identical in form,
but Saxon words likewise; nay, not unfrequently we
meet with words of Saxon origin which have become
outwardly identical with words of Romance origin.
For instance: —

I. to blow | A. S. blâwan, the wind blows
to blow | A. S. blôwian, the flower blows
to cleave | A. S. clífian, to stick
to cleave | A. S. clúfan, to sunder
a hawk | A. S. hafuc, a bird; German Habicht
to hawk | A. S. to offer for sale, German höken
to last | A. S. gelæstan, to endure
last | A. S. latost, latest
last | A. S. hlæst, burden
last | A. S. lást, mould for making shoes
to lie | A. S. liegan, to repose
to lie | A. S. leogan, to speak untruth
ear | A. S. eáre, the ear; Lat. auris
ear | A. S. eár, the ear of corn; Gothic ahs; German Ähre

II. count | Latin comes
to count | Latin computare
to repair | Latin reparare
to repair | Latin repatriare
tense | Latin tempus
tense | Latin tensus
vice | Latin vitium
vice | Latin vice

III. corn | A. S. corn, in the fields
corn | Latin cornu, on the feet
sage | A. S. salwige, a plant
sage | Latin sapius
to see | A. S. seohan
291see | Latin sedes
scale | A. S. scalu, of a balance
scale | A. S. scealu, of a fish
scale | Latin scala, steps
sound | A. S. sund, hale
sound | A. N. sund, of the sea, from swimman
sound | Latin sonus, tone
sound | Latin subundare, to dive *49

Although, as I said before, the number of these
equivocal words will increase with the progress of
phonetic corruption, yet they exist likewise in what
we are accustomed to call ancient languages. There
is not one of these languages so ancient as not to disclose
to the eye of an accurate observer a distant past.
In Latin, in Greek, and even in Sanskrit, phonetic
corruption has been at work, smoothing the primitive
asperity of language, and now and then producing
exactly the same effects which we have just been
watching in French and English. Thus, Latin est is
not only the Sanskrit asti, the Greek esti, but it likewise
stands for Latin edit, he eats. Now, as in German
ist has equally these two meanings, though they are
kept distinct by a difference of spelling, elaborate
attempts have been made to prove that the auxiliary
verb was derived from a verb which originally meant
to eat — eating being supposed to have been the most
natural assertion of our existence.

The Greek ίόs means both arrow and poison; and
here again attempts were made to derive either arrow
from poison, or poison from arrow. 50 Though these
292two words occur in the most ancient Greek, they are
nevertheless each of them secondary modifications of
two originally distinct words. This can be seen by
reference to Sanskrit, where arrow is ishu, whereas
poison is visha, Latin virus. It is through the influence
of two phonetic laws peculiar to the Greek
language — the one allowing the dropping of a sibilant
between two vowels, the other the elision of the initial
v, the so-called digamma — that ishu and visha converged
towards the Greek ίόs.

There are three roots in Sanskrit which in Greek
assume one and the same form, and would be almost
undistinguishable except for the light which is thrown
upon them from cognate idioms. Nah, in Sanskrit,
means to bind, to join together; snu, in Sanskrit,
means to flow, or to swim; nas, in Sanskrit, means to
come. These three roots assume in Greek the form

Néō, fat. nḗsō (the Sanskrit nah), means to spin,
originally to join together; it is the German nähen,
to sew, Latin nere. Here we have only to observe
the loss of the original aspirate A, which reappears,
however, in the Greek verb nḗthō, I spin; and the
former existence of which can be discovered in Latin
also, where the c of necto points to the original guttural

snu, snauti, to run, appears in Greek as néō.
This néō stands for snefō. S is elided as in mikrós
for smikrós *51 and the digamma disappears, as usual,
between two vowels. It reappears, however, as soon
as it stands no longer in this position. Hence fut.
293neúsomai, aor. éneusa. From this root, or rather from
the still simpler and more primitive root nu, the
Aryan languages derived their word for ship, originally
the swimmer; Sanskrit naus, nâvas; Greek naûs,
nēós; Latin navis; and likewise their word for snow,
the Gothic snaivs, the Latin nix, but nivis, like vivo,
vixi. Secondary forms of nu or snu are the Sanskrit
causative snavayati, corresponding to the Latin nare,
which grows again into natare. By the addition of a
guttural, we receive the Greek nḗchō, I swim, from
which nêsōs, an island, and Náxos, the island. The
German Nachen, too, shows the same tendency to
replace the final v by a guttural.

The third root is the Sanskrit nas, to come, the
Vedic nasati. Here we have only to apply the Greek
euphonic law, which necessitates the elision of an s
between two vowels; and, as our former rule with
regard to the digamma reduced nefō to néō, this will
reduce the original nésō to the same néō. Again, as
in our former instance, the removal of the cause removed
the effect, the digamma reappearing whenever
it was followed by a consonant, so in this instance the
s rises again to the surface when it is followed by
a consonant, as we see in nóstos, the return, from

If, then, we have established that sound etymology
has nothing to do with sound, what other method is to
be followed in order to prove the derivation of a word
to be true and trustworthy? Our answer is, We must
discover the laws which regulate the changes of
letters. If it were by mere accident that the ancient
word for tear took the form aśru in Sanskrit, dákry
in Greek, lacruma in Sanskrit, tagr in Gothic, a
scientific treatment of etymology would be an impossibility.
294But this is not the case. In spite of the
apparent dissimilarity of the words for tear in English
and French, there is not an inch of ground
between these two extremes, tear and larme, that
cannot be bridged over by Comparative Philology.
We believe, therefore, until the contrary has been
proved, that there is law and order in the growth
of language, as in the growth of any other production
of nature, and that the changes which we
observe in the history of human speech are not the
result of chance, but are constrained by general and
ascertainable laws.295

1* Molière says, ‘Je sens de son courroux des gênes trop cruelles.’

2 Jeremiah vii. 31-32.

3* Henri Estienne, Traicte de la Conformité du Langage Français
avec le Grec
, 1566. What Estienne means by the conformité of
French and Greek refers chiefly to syntactical peculiarities,
common to both languages. ‘En une epistre Latine que je
mi l'an passé audevant de quelques miens dialogues Grecs, ce
propos m'eschappa, Quia multo majorem Gallica lingua cum
Græca habet affinitatem quam Latina; et quidam tantum (absit
invidia dicto) ut Gallos eo ipso quod nati sint Galli, maximum ad
linguæ Græcæ cognitionem προτέρημα seu πλεονέκτημα aflferre
putem.’ Estienne's etymologies are mostly sensible and sober;
those which are of a more doubtful character are marked as such
by himself. It is not right to class so great a scholar as H. Estieune
together with Perion, and to charge him with having ignored the
Latin origin of French. (See August Fuchs, Die Romanischen
, 1849, p. 9.) What Estienne thought of Perion may be
seen from the following extract (Traicte de la Conformité, p. 139):
‘Il trouvera assez bō nombre de telles en un livre de nostre
maistre Perion: je ne di pas seulemēt de phantastiques, mais de
sottes et ineptes, et si lourdes et asnieres que n'estoyent les
autres temoignages que ce poure moine nous a laissez de sa
lourderie et asnerie, on pourroit penser son œuvre estre supposé.’
Estienne is wrongly charged with having derived admiral, French
amiral, from άλμυρός. He says it is Arabic, and so it is. It is
the Arab Emir, prince, leader, possibly with the Arabic article.
French amiral; Span. admirante; It. almiraglio, as if from admirabilis.
Hammer's derivation from amîr al bahr), commander of
the sea, is untenable.

4* ‘Quant à la derivaison des mots par addition, substraction,
transposition, et inversion des lettres, il est certain que cela se
peut et doit ainsi faire, si on veut trouver les étymologies. Ce qui
n'est point difficile à croire, si nous considérons que les Hebreux
escrivent de la droite à la senestre, et les Grecs et autres de la
senestre à la droite.’

5* ‘Er (Got) enwil niht tuon wan slehtes,’ God will do nothing
but what is good. Fridank's Bescheidenheit, in M. M.'s German
, p. 121.

6* Not mentioned in Grimm's Dictionary.

7* See Diez, Lexicon Comp. s. v.

8* Mannier, Etudes sur les Noms des Villes. Paris, 1861, p. xxvi.

9* See Grimm, Deutsches Wörterbuch, s. v. Brotherr.

10* Diez, Grammatik and Lexicon, s. v.

11* Grimm, Deutsche Grammatik, ii. 66; i. 261.

12* See M. M. in Kuhn's Zeitschrift, v. 152. Pott, Etymologische
, ii. 58-60, 442, 450.

13* Curtius, Grundzüge der Griechischen Etymologie, ii. 291.
Savelsberg, in Hofer's Zeitschrift, iv. p. 91.

14* See Schleicher, Compendium, § 43.

15 Logan, Journal of Indian Archipelago, iii. p. 665.

16* Danneil, Wörterbuch der Altmärkisch-plattdeutschen Mundart,
1859, s. v.

17* See Diez, Grammatik der Romanischen Sprachen, passim.

18* Diez, Lexicon Comparaticum, s. v.

19* Biez, Lexicon Comp., s. v., second edition, proposes weiger
instead of wâri.

20* In Ital. golpe and volpe. Span. vulpeja, Fr. goupil, Lat.
vulpecula, and a few more words of the same kind, mentioned by
Diez (p. 267), the cause of confusion is less clear; but even if
admitted as real exceptions, they would in no way invalidate the
very general rule.

21 Grimm, Deutsche Grammatik, ii. 63, friskan, frask, fruskun;
O. H. G. friscing, victima (caro recens), frischling, porcellus.

22 After Saxon verbs, ment is found in shipment, easement,
fulfilment, forebodement.

23* See Promptorium Parvulorum, p. 398.

24* Trench, On Words, p. 156.

25 Fuchs, p. 125.

26* Diez, Romanische Grammatik, ii. 35.

27 Diez, Rom. Gram. i. 177. There are exceptions to this rule;
for instance, Italian riva, for ripa; savio, for sapio; and in
French, such words as vapeur, stupide, capitaine, Old French

28* Diez, ii. 20. Opera is not the Latin opus, used as a feminine,
but the plural of opus. Such neutral plurals were frequently
changed into Romance feminines, and used in the singular. Thus
Latin gaudia, plural neut., is the French joie, fem, sing., Italian
gioja. A diminutive of the French joie is the Old French joel,
a little pleasure; the English jewel, the French, joyau.

Latin arma, neut. plur. | Italian and Sp. arma | Fr. l'arme
folia — It. foglia — Fr. feuille
vela — It. and Sp. vela — Fr. voile
batualia — It. battaglia — Fr. bataille

29* In Claus Groth's Fiv nie Leder ton Singn un Beden vœr
, 1864, tokum, i. e. to come, is used as an
adjective: ‘Se kamt wedder to tokum Jahr.’

30* Cf. M. M., Ueber Deutsche Schattirung Romanischer Worte,
in Kuhn's Zeitschrift, v. 11.

I take this opportunity of stating that I never held the
opinion ascribed to me by M. Littré (Journal des Savants, avril
1856; Histoire de la Langue Française, 1863, vol. i. p. 94), with
regard to the origin of the Romance languages. My object was to
explain certain features of these languages which, I hold, would
be inexplicable if we looked upon French, Italian, and Spanish
merely as secondary developments of Latin. They must be
explained, as I tried to show, by the fact that the people in whose
minds and mouths these modern dialects grew up, were not all
Romans or Roman provincials, but tribes thinking in German
and trying to express themselves in Latin. It was this additional
disturbing agency to which I endeavoured to call attention, without
for a moment wishing to deny other more normal and generally
admitted agencies which were at work in the formation of
the Neo-Latin dialects, as much as in all other languages advancing
from what has been called a synthetic to an analytic
state of grammar. In trying to place this special agency in its
proper light, I may have expressed myself somewhat incautiously,
but if I had to express again my own view on the origin of the
Romance languages, I could not do it more clearly and accurately
than in adopting the words of my eminent critic: ‘A mon tour,
venant, par la série de ces études, à m'occuper du débat ouvert,
j'y prends une position intermédiaire, pensant que, essentiellement,
c'est la tradition latine qui domine dans les langues romanes,
mais que l'invasion germanique leur a porté un rude coup, et que
de ce conflit ou elles ont failli succomber, et avec elles la civilisation,
il leur est resté des cicatrices encore apparentes et qui sont,
àa un certain point de vue, ces nuances germaniques signalées par
Max Müller.’

31* Roger Bacon, Compendium Studii, cap. 7 (ed. Brewer, p.
449): ‘quoniam etymologia est sermo vel ratio veritatis.’

32 l. c. cap. 7, p. 450. ‘Brito quidem indignissimus auctoritate,
pluries redit in vitium de quo reprehendit Hugutionem et Papiam.
Nam cum dicit quod Gehenna dicitur a ge, quod est terra, et
ennos, quod est profundum, Hebræum vocabulum docet oriri ex
Græco; quia ge pro terra est Græcum, et gehenna est Hebræum.’

33* l. c. cap. 7, p. 457. ‘Similiter multa falsa dicuntur cum istis
nominibus, cenobium, cenodoxia, encenia, cinomia, scenophagia,
et hujusmodi similia. Et est error in simplicibus et compositis,
et ignorantia horribilis. Propter quod diligenter considerandum
est quod multa istorum dicuntur a κενᾠ Græco, sed non omnia.
Et sciendum quod cenon, apud nos prolatum uno modo, scribitur
apud Græcos tribus modis. Primo per e breve, sicut kenon, et
sic est inane seu vacuum, a quo cenodoxia, quæ est vana gloria.
… Secundo modo scribitur per diphthongum ex alpha et iota,
sicut kainon, et tune idem est quod novum; unde encænia, quod
est innovatio vel dedicatio, vel nova festa et dedicationes ecclesiarum.
… Tertio modo scribitur per diphthongum ex omicron et
iota, sicut koinos…. Unde dicunt cenon, a quo epicenum, communis
generis…. Item a cenon, quod est commune, et bios, quod
est vita, dicitur cenobium, et cenobitæ, quasi communiter viventes.’

34* ‘Quomodo enim vester Axilla Ala factus est nisi fugâ literæ
vastioris, quam literam etiam e maxillis et taxillis et vexillo et
paxillo consuetudo elegans Latini sermonis evellit.’ — Cicero, Orat.
45, § 153.

35* Pott, Etymologische Forschungen, i. p. 645.

36* The Sanskrit upa and upari correspond to Greek ὺπό and
ὺπέρ, Latin sub and super, Gothic uf and ufar.

37 Quint. 9, 4, ‘oratio alia vincta atque contexta, alia soluta.’

38* Cf. Greek έρετμός, Latin resmus and remus. Triresmos occurs
in the inscription of the Columna Rostrata.

39* See M. M., Turanian Languages, p. 95, seq. Pott, in
Deutsche Morgenländische Gesellschaft, ix. 430, containing an
elaborate criticism on M. M.'s Turanian Languages. The same
author has collected some more accidental coincidences in his
Etymologische Forschungen, ii. 430.

40 Appleyard, Kafir Language, p. 3.

41* Grimm, Deutsche Grammatik, ii. pp. 193, 421.

42 Pott, Etymol. Forschungen, i. 774, seq. ‘Sollum Osce totum
et solidum significat.’ — Festus.

43 Grimm, Deutsche Grammatik, i. pp. 389, 394.

44* In the language of the gipsies, devel, meaning God, is connected
with Sanskrit deva. Kuhn, Beiträge, i. p. 147. Pott, Die
, ii. p. 311.

45* ‘Neque enim debent (aves) ipsis nidis involare; ne, dum
adsiliunt, pedibus ova confringant.’ — Col. 8, 3, 5.

46 ‘Vix me contineo, quin involem in capillum, monstrum.’ —
Ter. Eun. 5, 2, 20.

47 ‘Adeoque improvisi castra involavere.’ — Tac. H. 4, 33.

48§ ‘Remitte pallium mihi meum quod involasti.’ — Cat. 25, 6.
These passages are taken from White and Riddle's Latin-English
, a work which deserves the highest credit for the
careful and thoughtful manner in which the meanings of each
word are arranged and built up architecturally, story on story.

49* Large numbers of similar words in Mätzner, Englische
, i. p. 187; Koch, Historische Grammatik der Englischen
, i. p. 223.

50 The coincidence of τόξον, a bow, and τοξικόν, poison for smearing
arrows (hence intoxication) is curious.

51* Cf. Mehlhorn, § 54. Also σϕάλλω, fallo; σϕόλλος fungus.
Festus mentions in Latin, smitto and mitto, stritavus and tritavus.