An Etymology of Latin and Greek (1882)

mmmm

Riip

m:\

I ««-';fi«irKi>.m

f^v:^^!'-,

ii.

G.Boone.

U/3

Digitized by the Internet Archive in

2008

with funding from

IVIicrosoft

Corporation

http://www.archive.org/details/etymologyoflatinOOhalsrich

AN

ETYMOLOGY

LATIN^ Al^TD GREEK.

BY

CHAELES

S.

HALSEY, A.M.

BOSTON: PUBLISHED BY GINN, HEATH, & 1882.

CO.

Entered, according to Act of Congress, in the year 1882, by

CHARLES

S.

HALSEY,

in the office of the Librarian of Congress, at Washington.

GiNN, Heath, J. S.

&

Co., Printers:

Gushing, Supt., ioi Pearl Street,

Boston.

PEEFACE. The following work had its origin in Many students of the classical languages,

a all

felt

want.

along the

early part of their course, use text-books provided with vocabularies.

These vocabularies, from the necessity of

their limits, are brief little into

and imperfect, and they enter but

Even when afterwards

the subject of etymology.

the lexicon

is

used, the etymology

is

often studied only for

separate words as they occur in reading ; and the scattered

and fragmentary information given in the lexicons produces a corresponding state of knowledge in the mind even of a diligent student.

No

connected, systematic, or

thorough knowledge of etymology the

is

grammar something may be done

mology

;

thus acquired.

In

for historical ety-

but the requirements of other topics in a school

grammar must always prevent

this subject

there the full treatment which

its

from receiving

importance demands.

There remain the larger works expressly devoted to the them in German, excellent when one gets to them and is prepared for them, but by their style subject, nearly all of

and fulness, as well as size and cost, not adapted to the wants of an American school-room or of the ordinary stuThey will be studied only by the few, and the dent. benefit to be derived from them will generally come only when the student is far advanced in his course, and after years of study of other works.

54? 244

PREFACE.

IV

But

historical etymology, that gives the original

and words themselves together, unites them by the natural bond of central

their

meaning

of related words, and, gathering the

common

should

it

origin, should not be so long deferred, nor be pursued only as a higher range of study.

Itself the historical foundation of all the structure of lan-

guage, certainly

it

should form a prominent part in the

foundation of the course of study. form,

Presented in a simple to be made furnish can a large vocabulary of the

it

most practical words, and these not arranged for comparison in the separate language merely, Greek with Greek, Latin with Latin, but placed side by side, each language throwing light upon the other. A wider comparison en-

mind a deeper comprehension of the laws of progress in language reveals

riches with knowledge and enlarges the

new and

;

interesting truth, arousing curiosity and stimu-

lating to further investigations. It has

we have

been urged against the study of etymology that not within our reach

suflBicient

material to furnish

the basis of the science, and that etymologists, proceeding

upon any well-ascertained general principles, but and even roaming wild excursions of fanciful associations, have pro-

often not

upon

superficial resemblance of words,

off in

duced such results as to bring the study into deserved condemnation. We must always bear in mind that historical

etymology

is

not specially concerned with the

absolute origin of language.

It is concerned to ascertain

the early forms, wherever they are traceable. are

many words which we

True, there

cannot trace to their early

many words, and these the most important, that we can trace, and of their etymology our knowledge is as reliable as any in the whole range of It must be acknowledged, too, that the work language. forms; but there are also very

V

PREFACE. of

many

discredit

professed etymologists did in former times bring

But the

upon the study.

The general

case

is

now widely

and methods according to which all scientific etymological research must proceed, are now thoroughly established and recognized. The application of these principles requires a wide and careful different.

principles

comparison of kindred words. always going on and becoming

As still

this

comparison

wider and more

is

dis-

criminating, the special results attained, relating either to single

words or to the rules deduced, must always be held any modification which may be reasonably

as open to

required by continued investigation.

For a long period of time, extending

to the year 1876,

the views of etymologists in regard to the rules of Indo-

European phonetics were

in substantial agreement.

Be-

ginning with that year, certain important modifications

were proposed in some of the rules of the Indo-European phonetic system and these modifications are now generally ;

accepted

among

the

German

Part IV., Chaps. I.-III.

These views

philologists.

be found stated and explained in Part

I.,

will

Chap. VL, and

In presenting them I

am much

indebted to Prof. Maurice Bloomfield, with whose cordial

approval I have given the statement of those chapters

condensed mainly from his paper on the Greek Ablaut,

American Journal of Philology" for September, 1880. The Preliminary Statement of the same views is condensed from his article in the Journal of Depublished in the

cember, 1881.

'^

The

roots,

arranged in accordance with this

system, are given by themselves near the close of the volume, so that the use of

them

will not lead to

any

confusion.

In the preparation of the present work, the author has endeavored to conform to the latest investigations of the highest authorities.

In general, doubtful or disputed ety-

PEEFACE.

VI

mologies have been omitted,

or, in

the few cases given, they

The table of vowel-scales is from ^^ Comparative Grammar." It is assumed that Schleicher's any student who may use this Etymology is already proare

.

marked

doubtful.

vided with a suitable

grammar

work does not

therefore, this

which are given

suffixes

of Latin or

Greek

;

state in full the prefixes

grammars.

in the

and,

and

Neither does

aim to present in full the processes of inflection, which would require a larger treatise upon comparative grammar. it

The

object of this

work

is

to present, within the limits

of a school-book, the most needful etymological information

that

is

lexicon.

not adequately furnished by the

Even within

grammar

or the

these limits, some things are stated

that are not intended to be learned in the early part of a student's course,

e.g.,

later reference.

They

the Sanskrit forms.

because they illustrate the subject, and

may

are given

be used for

Great prominence has been given to the

Many

derivation of English words.

of the cognate

words

here treated have descended to us through the French, or

A

through the Teutonic family.

complete index

is

fur-

nished for the Latin, the Greek, and the cognate English

words.

The study

of etymology, as here presented,

may

advan-

tageously begin at an early stage in the study of Latin;

and

it

should continue, in some form, throughout the

The present work may be

course of classical education.

used for regular daily lessons in connection with the study of the classical text,

and

and may

be employed

facility,

also,

for

with equal advantage

reference

on individual

words. C. S.

Schenectady

:

April, 1882.

HALSEY.

PRELIMINARY STATEMENT OF THE NEW SYSTEM OF INDO-EUROPEAN PHONETICS.

The

clianges proposed

chiefly to the vowels.

by the new system have reference European vowels,

It is held that the

had been previously supposed, later modian original Indo-European '^, but are themselves original Indo-European vowels. The theory of vowel-increase has been abandoned the consequence is roots of the form a^s, ^,

€p,

Formerly the roots were inconsistently arpv, and 8tfc, thus set down as ccr, €p, fxeu, SepK, TTcvO, but allowing the e a function in the one case and denying it the same in another which is perfectly parallel. These roots have ot, 8opK, ttovO, in addition to the form with ablaut a° (Greek o etc.) a weak form, which differs from the strong by the lack This reduced form cr, t, o-pv, ^p, /xv, St/c, 8p/c, irvO, of this e (o)

/i,€v,

Sct/c,

SepK, irevO).

t,

:

:

may tions

assumed to have stood originally only in formawhich had the word-tone on some non-radical syllable,

safely be

— thus naturally bringing about a

less distinct

pronunciation

The graphical representation of this weakened utterance is root minus the e-o vowel. The recognition of these weak root-forms leads irresistibly to the assumption oi Indo-European lingual and nasal vowels; Indo-European r, (J), n, m, represented in Greek by ap or pa (aX or Aa) for the lingual, and a and av, a and a/A, for the of the root-syllable.

nasal vowels.

Strange in external appearance are the Indo-European and Greek groundforms or explanatory symbols which are the result: ^Tn-wpiai for ravu/xat

;

"^

f^n-ua for fiaivio =^ venio.

The

PEELIMINARY STATEMENT OF THE

Vlll

Greek groundform

for icj^OdpaTo

tainly does not seem as

if

would be * i-cj^Op-yro,

one of the acquisitions of

It cer-

gram-

tlie

matical science of to-day were simplicity of method in representing its processes. We will, however, gladly put up with a cumbrous system of symbols,

if we are compensated for it by such symbols help to convey to the reader the exact meaning of the writer. This quality the signs, which

exactness

—

if

may

be gleaned from the examples above, in general possess i is the designation for semi vocalic or semiconsonantal y in distinction from the full consonant (spirant)

to a high degree,

the same is true for u. When we examine the symbolgroup ^nsma% there can be no doubt as to the exact value

y

;

represented by

it

:

m-

is

carries the syllable tone

The

a syllable in which the element that in the main nasal (a nasal vowel).

is

vocalic color of this nasal vowel the

undertake to express, and sentations of

widely en,

:

it is

symbol does not

indeed unknown.

The

repre-

in the various languages of the family diverge

it

Greek and Sanskrit a and an but German un, Latin in. In the same way r is an element mainly ;

Lithuanian

of a lingual character, bearing the tone of the syllable

rendering of

in the

;

the Sanskrit at least coincides with the symbol

it

the other languages again vary greatly Zend Greek ap and a\ Latin and German or (ur) Lithuanian ir. The remainder Tna^ is practically identical with Greek /xe. The symbol does not, however, profess to define (Sanskrit r)

^T^

;

:

;

;

;

the value of the Indo-European vowel, which

it

renders, quite

vowel sound lying somewhere bebut without quite reaching a in the same way

so closely; a^ expresses a

tween

and a, sound between

e

;

and a which does not quite reach a. Nevertheless it is becoming more and more common to write simply e and o for a^ and a*' even at the expense of perfect exactness and in the present work the more simple forms are preferred, so that in Indo-European roots and words e may be found where a* could also be written, and o where a" could a° is a

;

also be written.

o

:

:::

NEW SYSTEM OF INDO-EUROPEAN The

'

in

*

new

writers of the

reihen,'

'

:

PHONETICS.

IX

school treat the vowel-phenomena

vocalreihen,' an expression which, like

many

Ger-

man

grammatical terms, can be rendered but inadequately into English by 'vowel series.' Parallel with the three vocalic forms presented in the a^-reihe (form with a", form with a°,

and form without

this

there appear three other series

a^'-a'')

— the

^-series

0, 0, dj

justifying the following proportion for the

e,

:

o,

the a-series

e,

TABLE

An by

;

€-series

c

:

o

:

~

rj

:

0)

:

€

d-series

d

:

cd

:

= = a =

o)-series

w

:

w

:

o

presented by

17-series is

of the d-series

Si-8(D'fJii,

I.

77-series

example of the

TL-Oe-fjiaL

by

and the o-series Greek

a, o, d,

:

^rj-ixi, cfxD'in],

rt-Orf-fit, Bay-fio-gy

:

^a-/xeV

In order to understand the origin of these to refer to the

assumes that

'

all

series, i.e.,

the

to their recognition, it will be necessary

method which led

a* (e)

of the w-series

;

Se-Sco-Ka, So-Tos.

Theory of Sonant Coefficients.' This theory Indo-European roots can have but one vowel,

varying with a° (0)

;

all

other seemingly vocalic elements

are in reality semiconsonants, which assume the function of

vowels only when this semiconsonant

is

e-o

called

'

has for some reason been sonant

coefficient.'

the root does not possess such a sonant coefficient, vowelless

with X^Vi

all

Se/o,

with

lost

it

the facts in the case of roots of the a^-series

TTOT, Sol, etc.,

Xcitt,

this

remains

This agrees incontrovertibly

(Trer-o/xau i-TTT'O/jirjv).

orrcX, fcev,

;

In cases where

i\€v6, Scp/c, irevO, etc.,

but only upon the

;

ttct, Set,

can interchange

loss of this e or o

do the

semiconsonantal elements contained in these roots assume the function of vowels

The

:

St,

A, /xv, AtTr, i\vO, SpK, iryO, etc. x^, Sp, err

possible sonant coefficients of roots of the a*-series are

accordingly

:

i,

u^

r,

(^),

n,

m

;

and

if

we add

these to the

:

:

PRELIMINARY STATEMENT OF THE we

real vowels of the a*-series,

obtain the following five (or

within the a^-series

six) series

TABLE

= = p =

€L

I

OL

I

L

€v

:

ov

:

V

ep: op:

In Greek the roots

about 250, and roots

(cX €v €fl

made according

o\

'^)=

ov

'

Y

op.

:

p.

^^

to these

models are

in verbal formations are of this class.

represented

same plan does not boldly made.

As

coefficients to c-o

form

to

;

may be found

is

t,

v, p, (A), v, /x

are the sonant

as these are forced in the reduced root(t, v,

p, (A), v,

a sonant coefficient

p\

which

(c),

so in Table is

lost

;

i.e., 77

stands for

ce

;

w

for o€

;

into

€0, 00,

and

o.

We

;

e-o

has

same way the oa for w, and a

in the

vocalism of the d-series goes back to ea for d the sonant coefficient

I.

performing

the function of a vowel, because the real root- vowel

been

then,

constructed on the

removed, and the attempt has been

lie far

in Table II.,

play the part of vowels

of the T^-series

The thought,

Sanskrit and Gothic).

(e.g.,

that the remaining roots also

is

In

other languages also these roots are preponderatingly

the

€

:

probable that more than one-half of the

it is

which occur

II.

;

so also the w-series is to be resolved

could then add to Table

II.

three per-

fectly parallel series

TABLE

From L

and

€€

:

oe

:

€

ca

:

oa

:

a

€0

:

00

:

= =

o

the standpoint of the phonetist

objection can be urged as

III.

;

c,

a,

and

o

it is

can be

v (Sievers, Phonetik, p. 123)

:

*

believed that no

consonans

'

as well

the contractions with

the root-vowels into the vowel-forms actually occurring

would

;

NEW SYSTEM also pass criticism,

OF INDO-EUROPEAN PHONETICS.

though

it

is

XI

to be noted that in the first

perpendicular column of Table III. the semiconsonantal ele-

ments impress their vocalic color on the result (€€, ea, €o rj, d, co), while in the second perpendicular column the semiconsonantal :

element succumbs, and the result of the contraction (w) has the vocalic color of the real root-vowel

From

(o).

the standpoint, however, of the history of the Indo-

European languages, we are not

present warranted in

at

accepting these results (shown in immediate connection with

Table

No one

III.).

language shows even a single instance in

which the elements supposed

we

to underlie the contraction occur

This, to be sure,

uncontracted.

is

no

final

condemnation

accustomed more and more to view the

are becoming

immediate historic background of the separate Indo-European the Indo-European parent language, as a real languages,

—

language devoid of unnatural regularity, presenting in

many

— phenomena

phenomena of a very secondary nature, which had a long history before them and the these contractions must not be absolutely denied. respects

;

possibility of

Practically,

however, they cannot as yet he recognized in that form. This theory has, nevertheless, yielded one result that we may safely adopt, namely, the recognition of the fact that the

and dunder which

the

97-

It will

series €

varies with

be interesting

vocalic material

is

yj

and d

of

vary with w under the same circumstances o.

now

to see

what

vocalic

and semi-

furnished for the Indo-European parent

speech.

The a*-series yields two real vowels a* and a° (e and 0) and the following sounds wavering between consonantal and vocalic function y and i v and u r and r {I and I) n and and m perhaps also the nasals corresponding to the two 72, Indo-European guttural series, which could be designated by ft and fij and n and n. Its diphthongs would be ei, oi^ eu, ou^ (in a wider sense of the term also er, or (el, ol); en, on; em, om, and even m, on; en, on). :

:

m

;

;

;

;

;

:

NEW SYSTEM

Xll

OF INDO-EUEOPEAN PHONETICS.

e and & (so designated to differentiate two following series) and e. The a-series yields a and & and a. The o-series yields & and & and o. Of diphthongal material in which the first part is a long vowel there appears certainly at least du in the stem nduIonic (not pan-hellenic)

The

it

e-series yields

from the

:

of the

o's

:

;

:

:

vy]v- H

i

z j

Palatal

Lingual

Labial

Series.

Series.

Series.

interruption.

o o ,

d

Sonant. Surd.

-

Semivowels.

The diphthongs

Double Consonants.

are (a^), ae,

ei, (oi), oe,

ui,

au,

the forms inclosed in parenthesis being found only in

early Latin.

The diphthongs

follows: ai as ai in aisle

;

are sounded approximately as

ae originally sounded as (Roman)

(Roman) e ; ei as ei in eight; oi as oi in oil; oe German oe in Oel, or Eng. o in world ; ui as uee in queen; au as ou in house; eu as eu mfeud; ou as ou in group. Of the consonants, 5, cZ, p^ t^ r, m, w, A are sounded as in de\ later as

nearly as

Z,

PRINCIPLES OF ETYMOLOaY.

14 English

as

^' ;

y

w

in yes; 8 as in so; v like

in wait;

f

as in

fate; g as in get ; c, k, q b>s c in can; ch, th, ph, as c, p, with the slight addition of A-sound, as in the words, hlochhead^ ?^,

The letters y and z were introhothouse, uphill; x as in mix. duced into the Latin language after the time of Cicero, and were used only in words taken from Greek, y being employed to represent the Greek v, and z to represent the Greek C Latin y has the sound of French u, and for this reason its position in the scheme is between u and i z may be sounded ;

like dz in adze or like z in zone.

OHAPTEE

V.

PHONETIC CHANGE. Throughout the history of language, changes of sound are going on. In comparing one language with any of its kindred, we must first ascertain to what sounds of the latter the sounds We then have a guide of the former regularly correspond. for the regular etymological comparison of words.

what

tration of this appears in

is

An

illus-

called (from its discoverer)

"Grimm's Law of Permutation of Consonants," which exhibits, with some exceptions not necessary here to be shown, the regular interchange between (1) Sanskrit, Greek, and Latin,

Gothic and Low German dialects High German and its stock (including This law may be expressed by the following

taken as one group (including English)

modern German). formula

:

—

(1) Sanskrit,

;

(2)

;

(3)

Aspirate

Greek, Latin

Low German

(2)

Gothic and

(3)

High German

English)

Sonant

Surd

(including

Sonant

Surd

Aspirate

Surd

Aspirate

Sonant

:

PHONETIC CHANGE. It

may

15

—

be illustrated by the followir[g table

I.

^ 1 1

Greek

9

.

.

.

Latin

f

.

.

.

2.

English d

3.

German

.

.

th =

t or

.

t

dvyoLT'qp

Brip

dvpa

fera

fores

fiedv

daughter

deer

door

mead

tochter

thier

thor

meth

II. 1

^ '

1

Greek

8

.

.

.

oBovs

dufjLau

Bvo

edeiv

vScop

Latin

d

.

.

.

dens

dvx)

edere

unda

.

.

.

two

eat

water

.

.

zahn

domare tame zdhmen

zwei

essen

wasser

rv

TpeTs

2.

English t

3.

German

z or s

tooth

III.

^ 1 '

1

Greek

t

.

.

.

Latin

t

.

.

.

tic

tres

tenuis

is-tud

.

.

.

thou

three

thin

that

brother

.

.

.

du

drei

diXnn

das

hruder

2.

English th

3.

German d

{(Tv)

t6 frater

General Table of Grrimm's Law.

A Latin.

Sanskrit. Greek.

X Aspirates

.

.

TK [PH

\

G Sonants

Surds

.

.

.

-

.

t.f(g,v) f(d,b)

C High

Low Germ.

Germ.

k

g

t

f(b)

d b

7

g

k

ch

S

d

t

zz

i8

b

P

f,

P

i

g(j) d

[b

b

K

k

K

c,

t

T

t

th,

P

TT

P

f,

"^ .

dh (h) bh(h)

B Gothic and

J

T P

q

b,

g

V

d

ph

(f)

d f,

V

PRINCIPLES OF ETYMOLOGY.

16

PEINCIPLE OF PHONETIC CHANGE. The

which underlies the greater part of phonetic to ease of utterance. In using the or-

principle

change

the tendency

is

we

gans of speech, effort, to

and

naturally tend to economize or diminish

reduce the distance between one sound and another,

so to

make each

necessary step in utterance as short and

Accordingly, the general direction of phofrom the extremes toward the middle of the

easy as possible.

change

netic

is

alphabetic scheme,

movement

in the opposite direction being

only exceptional or from special causes.

EESULTS OF PHONETIC CHANGE. The forms

change appear chiefly under two

results of phonetic I.

:

Weak

Articulation; II. Indistinct Articulation,

Weak

I.

Articulation.

"Weak Articulation appears under four forms I. SubstituThese II. Loss; III. Assimilation; IV. Dissimilation. four forms are applied to vowels and to consonants. :

tion;

Vowel-Ohange.

By

Substitution.

I.

may

Original a

be made.

substitution the following changes

may

be changed,

—

This 1. In Greek and Latin to c, e: V^ad, eSo9, sedes. change was very extensive even in the Graeco-Italic period. 2. In Greek and. Latin, to o, o: ->/dam, 80/A09, ddmus. 3.

In Greek, tot:

4.

In Latin, to

second

member Loss.

II.

i,

of a

Ind.o-'EnY.,

compound word

:

:

V^^^'

Greek

:

Ti-Orj-fiL,

especially in the capio, accipio.

In the following examples, the vowel

closed in parenthesis.

Latin

dd-dhd-mi; Greek,

— a very frequent change,

lost is en-

yty(€)vo)u,at, €(r(€);)(ov, 7raT(€)/oos.

(e)sum, gig(e)no, disc ip(u) Una.

::

PHONETIC CHANGE.

changed

u

having

made

The

resulting vowel

e in

genosis, genoris^

Two

generis (from genus).

vowels in contact

may approximate

pulsus

pello^

;

may be an easier utter-

difficult, it

for that position

one form of assimilation.

is

in flagro, fulgor

each other

a vowel closely connected with

utterance thereby

its

to a vowel,

This

ance. is

When

Assimilation.

III.

a consonant has

17

:

Two

* (e)syam, * siam, * siem.

vowels separated from each other only by a consonant sometimes assimilate :

hone, bene.

IV. Dissimilation. The

prevent

obj ect of dissimilation is to

same vowel.

repetition of the

Thus, sequontur was a form

retained instead of sequuntur ; aliinus became alienus.

Oonsonant-Ohange. I.

Substitution,

1.

In Greek and Latin

surd to sonant;

;

orig.

y disappears,

or

is

as'-P;' s is retained at the s before

a vowel

is

g:

i

II.

and

pagus.

its

effects

;

5,

and

v appears

end of roots and words, but

initial

generally changed to the rough breathing.

r,

s,

y,

v are generally re-

and y and v are interchanged

u.

Loss.

This

and Latin an fi€iSdo),

iryjyvvfjiL,

to the original letters, y,

In Latin, the original letters

3.

a change from original -y/'BAK,

seen only in

tained, but often s passes into

with

y,

Greek shows an aversion

2.

V

we have

k to

e.g., orig.

TYilTor

may

be

initial,

original initial 5 or ;

medial, or -y

is

-y/vARK, VALK, VLAK,

loss is not so frequent,

final.

sometimes poLKos,

very rare in Greek

:

In Greek

lost

lacev.

ip€(T)L,

:

->/^^^>

Medial /ActXo(v)a,

occurs most frequently before y, s, and v It occurs also before nasals di{c)sco, raa{g)ior, sua(d)vis.

fi€L^(a.

In Latin,

it

and before t and d: tor(c)tus, i{s)dem. Loss end of a word affects single consonants or combinations In Greek, when several consonants end a of consonants. word, they are sometimes all dropped, as in yaXa(/cT); but lu{c)na, lu{c)men;

at the

generally the last only

is

retained,

and the preceding vowel

is

PRINCIPLES OF ETYMOLOGY.

18 then lengthened

;

as, tlO€(vt)^,

of several consonants

may end

nOeU.

In Latin, a combination

a word, as in ferunt, urbs

in the older Latin, final consonants, especially

s,

?7i,

t,

;

but were

frequently dropped.

Assimilation.

III.

lation of consonants in

The most important rules for Greek and Latin are given

assimi-

in the

grammars. IV. Dissimilation.

Examples are

The

rules are given in the

a^riovy aa-riov; i-Ov-OrjVy iTvOrjv

;

grammars.

Ot-Orj-fjuL,

tlOtjixi;

claudtrum, claustrum, II.

Indistinct Articulation.

In general, the immediate cause of indistinct articulation is A part of the an excessive tendency to ease of utterance. needful sound of a word is slurred or omitted then some indistinct or indefinite sound is added on and this, afterwards becoming more definite, may lead to the utterance of a sound even more difiicult than the original one which had thus Indistinct articulation appears under the following suffered. forms: I. Lahialism ; 11. Dentalism ; III. Parasitic Bound; ;

;

lY. Aspiration.

Labialism.

I.

^ and

This

a change from ^ to

is

tt

and p, or from

pronounced lazily, a slight t6;-sound and then, if the is apt to be produced immediately after it lips be nearly closed, an indistinct labial sound is produced.

^

to

If the h is

h.

;

For examples, see Nos. 496-515. II.

to 0.

Dentalism. This is a change from h For examples, see Nos. 516-520.

III.

come C

Parasitic Sound.

may regularly bemay have a slight sound

of S (here called parasitic) uttered before C-

from gh

In Greek, Sy

Initial y, if uttered lazily,

may become

to r, or

it,

and then the hy

Thus, for original y in -yjY^G,

we

find I in

PHONETIC CHANGE.

19

Aspiration. Examples are