The Dragon Revenant

Be sure to look for these other titles from Bantam Spectra

By Tom DeHaven
   Chronicles of the King’s Tramp
     Walker of Worlds
     The End-of-Everything Man
     The Last Humans

By Raymond E. Feist
     Magician: Apprentice
     Magician: Master
     Silverthorn
     A Darkness at Sethanon
     Prince of the Blood
     The King’s Buccaneer

By Katharine Kerr
     The Bristling Wood
     The Dragon Revenant
     A Time of Exile
     A Time of Omens

By Margaret Weis and Tracy Hickman
   The Death Gate Cyle
     Dragon Wing
     Elven Star
     Fire Sea
     Serpent Mage

By Angus Wells
   The Books of the Kingdoms
     Wrath of Ashar
     The Usurper
     The Way Beneath
   
The Godwars
     Forbidden Magic
     Dark Magic

I
N MEMORY OF
H
OWARD
“J
AKE
” J
ACOBSEN
1934–1988
H
E IS AND WILL BE SORELY MISSED
.

A
CKNOWLEDGMENTS

For my translations from the Llywarch Hen corpus I used Patrick Ford’s edition of the text in his
The Poetry of Llywarch Hen
, University of California Press, 1974. Since I was also swayed by his arguments in the introduction to that edition, I have translated
hen
in this context as “the ancestor.” Any errors in these translations are of course mine alone, as are such minor acts of magic as my turning winter into summer for the epilogue’s epigraph.

My special thanks go to:

John Boothe of Grafton Books for his support of and enthusiasm for this entire project,

Judith Tarr for sage advice and encouragement at the line of battle,

Eva, Jean, Linda, and Elaine of Future Fantasy Books in Palo Alto, California, for backing my books early on and for running a splendid bookshop, and, as always, my husband, Howard Kerr, for everything.

A N
OTE ON THE
P
RONUNCIATION OF
D
EVERRY
W
ORDS

The language spoken in Deverry is a member of the P-Celtic family. Although closely related to Welsh, Cornish, and Breton, it is by no means identical to any of these actual languages and should never be taken as such.

Vowels
are divided by Deverry scribes into two classes: noble and common. Nobles have two pronunciations; commons, one.

A
as in
father
when long; a shorter version of the same sound, as in
far
, when short.

O
as in
bone
when long; as in
pot
when short.

W
as the
oo
in
spook
when long; as in
roof
when short.

Y
as the
i
in
machine
when long; as the
e
in
butter
when short.

E
as in
pen
.

I
as in
pin
.

U
as in
pun
.

Vowels are generally long in stressed syllables; short in unstressed.
Y
is the primary exception to this rule. When it appears as the last letter of a word, it is always long whether that syllable is stressed or not.

Diphthongs
generally have one consistent pronunciation.

AE
as the
a
in
mane
.

AI
as in
aisle
.

AU
as the
ow
in
how
.

EO
as a combination of
eh
and
oh
.

EW
as in Welsh, a combination of
eh
and
oo
.

IE
as in
pier
.

OE
as the
oy
in
boy
.

UI
as the North Welsh
wy
, a combination of oo and
ee
. Note that OI is never a diphthong, but is two distinct sounds, as in
carnoic
(KAR-noh-ik).

Consonants
are mostly the same as in English, with these exceptions:

C
is always hard as in
cat
.

G
is always hard as in
get
.

DD
is the voiced
th
as in
thin
or
breathe
, but the voicing is more pronounced than in English. It is opposed to TH, the unvoiced sound as in
th
or
breath
. (This is the sound that the Greeks called the Celtic tau.)

R
is heavily rolled.

RH
is a voiceless
R
, approximately pronounced as if it were spelled
hr
in Deverry proper. In Eldidd, the sound is fast becoming indistinguishable from R.

DW, GW
, and
TW
are single sounds, as in
Gwendolen
or
twit
.

Y
is never a consonant.

I
before a vowel at the beginning of a word is consonantal, as it is in the plural ending-
ion
, pronounced
yawn
.

Doubled consonants
are both sounded clearly, unlike in English. Note, however, that
DD
is a
single letter
, not a doubled consonant.

Accent
is generally on the penultimate syllable, but compound words and place names are often an exception to this rule.

I have used this system of transcription for the Bardekian and Elvish alphabets as well as the Deverrian, which is, of course, based on the Greek rather than the Roman model. On the whole, it works quite well for the Bardekian, at least. As for Elvish, in a work of this sort it would be ridiculous to resort to the elaborate apparatus by which scholars attempt to transcribe that most subtle and nuanced of tongues. Since the human ear cannot even distinguish between such sound-pairings as B> and