PE22

Sec.u;� C\"11\A. l,..n,'\SC't"tp'h'tw' .
�
l:r
Qh\��, w'h«"K. u.u.c.l t"k« -p.A-C!\�1-t'\ AS en· eeo'"'""'a""'
�'\tt""' \CIL11'1.fS'-'-CILt:f& of •1.\ f'h& ��(\; 1.�'\..Ut\\y W"Kttt.M. WU
'""' "'"'� -coAAf'"c .. .... of »t& f'&ftW'o�.. .... .,.'1,.,..,,

I
wc. . .,. ��-.1o 1•
c:cr.\.tecl t1« ·l'��'�Diocr•t:�n�·. "'n\... •• a ... e,..'\ud. t,.. ""'•
OftCo'K c�•..,.c.h""'-'"'11" t'h.c A1f'ka1nts. .H••• c•"''Y n..a __,.,.,

'"'"..' p •..,.,., '
t'k•• '\.t.lt.ll ,.., t\M• 11no�' a �v..,., •
'""C'he. �&'tt•�• t.Ucc\ ...,.,.1 �g '""�"'"-'bet" , .-ec.."tct\ .- ... t'h�
fot'let¥A"'ir ..-4..- . . �&'¥ """'W�-''' ..... r�"' '' " "'"·-'·"'l'h·- '
f-""""-·� ht \&low. "l':'h.c � of t'k.o \.tt:t- c:; t'-4 �"

of hl.o '"""""' : ..,.,,,,
le1h� 'G. � • (vl i t'"'f'" ..... . ..... l'h& l&Aifc""
e,.,ul\ 0,. '"""'"''""''""'J ..t co-hl"'""IY� ,Jh-.c� ccuo\ct 'IO.C'II•
OCr...A'\0" W\1\l-f"'\\)' � 1..,_ l'h.A� C .... t. \\u 1'\'""" l.tt'� --� !- f!tl\lctvcb
�'"" �l\\& J .., � •(-.,1 . o .. tlo& vu.hi.C •f 1 "a.,.' ,lu. vauJ. .

.At��·"··. A �GI'I"MQ\o/� 1 f•J.W.. � t alL.
. IQ, ,. ha.. 'T�ti"""'
� H
.. QZ.
'F 7ft ' tb. .• ··� :
ye( m4 "'"'"•
-mp· . '4f/." . ..
•\
a gn .. 6 11., ..-m ..J;r.; .. po •Y

t alltt ..#tl rt/ " "
.-
aralt.t
..
"'"41• .r .
"Jf '
.,c .• r; �& a.Jl •
-
c. Ka.i�w-,.;:,. t-it-·.
41'da ttl AIJA 4/IA dba ..� ...... , ...... -l"l' \� 1'1:··" !
�· _, ...
pARCDA elDALACDBeRON 1\All
Lhe �eanoRian Aiphabec ··

:
· p AKc 1
Quenya VeRb StRuccuRe

by J. R. R. coLKleN
PARMA ELDALAMBERON
The Book of Elven-tongues
··
:
Number22
J. R. R. LOlk.leN
Lhe �eanoRian Aiphaber

paRr 1
Quenya VeRb SmuccuRe
edired by
ChRisropheR c::;Rson
and
ARden R. Smirh
Copyright© 2015 The Tolkien Trust
MOUNTAIN VIEW, CALIFORNIA

The Feanorian Alphabet and Quenya Verb Structure J. R. R. Tolkien
Acknowledgements
We would like to thank our colleagues Carl F. Hostetter,

Patrick H. Wynne and Bill Welden, for their fellowship and
support in innumerable ways; Adam Victor Christensen for
acting as social media administrator; Johan Winge for the
use of his tengwar font; our printers Stanley De Passos and
Ellie Farrell of Windward Graphics; and Cathleen Blackburn
of Maier Blackburn and R. Andrew Boose of Davis Wright
Tremaine for their assistance with legal matters.
We are indebted to the Tolkien Trust for permission to

publish an edition of these writings; and we are grateful to
Christopher Tolkien, who provided us with copies of the
documents presented here and advice on the interpretation
of difficult points, and whose scholarship continues to be
crucial for understanding their interconnections.
CHRISTOPHER GILSON
ARDEN R. SMITH
April, 2015.
PARMA ELDALAMBERON No. 22 - Page 2

The Feanorian Alphabet and Quenya Verb Structure f. R. R. Tolkien
Table of Contents
The Feanorian Alphabet, Part 1
Introduction 6.
Version A
Part 1. Feanorian Alphabet- The Lindarin Use 9.
Part 2. Parmaqestarin Use 15.
Part 3. The Old Noldorin or Feanorian usage 24.
Part 4. The Beleriandic or Exilic Usage 30.
Version A'
Part 2. The Parmaqestarin Mode 42.
Qenya Spelling
Introduction 55.
Qenya Grammar- Spelling and Transcription 61.
On J£lfwine 's Spelling

Version 1 67.
Version 2 69.
Version 3 71.
Version 4 73.
Version 5 75.
Version 6 77.
Quenya Verb Structure

Introduction 80.
Quendian & Common Eldarin Verbal Structure 92.

Inflexion 92.
Verbal Stems (simple or strong verbs) 95.
Moods 97.
Formation of other Verbal Stems 98.
PARMA ELDALAMBERON No. 22- Page 3

Quenya Verbal System- Conjugation & Syntax 99.

Conjugation of simple or "strong" verbs
Aorist system 99.
The Present (Continuous or Imperfect) 100.
The Past & Perfect system 101.
The Future 104.
Imperative 105.
The Participles 106.
The reduplicating syllable 112.
Forms of weak verbs, or verbs with weak aorist-present 113.

Syntactical notes 118.
Expressions of "Mood" 120.
Irregular Verbs 122.

The Negative Verb 126.
Common Eldarin: Verb Structure 128.
Inflexion 128.
The Verbal Stems 129.

The verbal stems ofBasic Verbs 129.
Formation of derivative verbal stems 133.
Adjectival and Substantival forms of Verbal stems 136.
Imperative forms 139.

Negation 140.
Late Notes on Verb Structure 141.
List of Abbreviations 169.
PARMA ELDALAMBERON No. 22 -Page 4

The Feanorian Alphabet
Part 1
by J. R. R. Tolkien
Edited with introduction and commentary
by
Arden R. Smith

The Feanorian Alphabet ]. R. R. Tolkien
Introduction
The Emergence of the Feanorian Tengwar
The specimens of early Elvish writing systems that have been published in the pages of Parma
Eldalamberon thus far have shown a gradual evolution toward the scripts exemplified in The
Lord of the Rings. In a lengthy essay, entitled "The Feanorian Alphabet," the non-runic system
of stems and bows is finally attributed to the Noldo Feanor, and the letters receive their
Quenya name: tengwar.
As seen in the documents published in Parma Eldalamberon, no. 20, Tolkien used the "Qenya
Alphabet" (c. 1931) for the representation of English and other primary-world languages. In
the present essay he changes his focus, ignoring the use of the Feanorian script for the repre
sentation of primary-world languages and only describing applications used by the Quendi.
Tolkien introduces five main Elvish modes here, each of which he describes in detail,
discussing such fine points as different styles of vowelling and historical developments in the
orthography:
• The General or Phonetic Form, which was used "in works on linguistic history, or in the
description of foreign languages."
• The Lindarin Use or Old Valinorian Orthography used for Qenya, which was
abandoned by the Noldor and came to be used exclusively for the Lindarin dialect.
• The Parmaqestarin Use, a mode developed by the Noldor for the later, written form of
Qenya. Tolkien describes two main types of vowelling for this mode: the "normal use," in
which the vowel a was assumed after every consonantal letter, unless another vocalic letter
followed it or a dot (putta) was written below it, and the qanta-tenkele or "Full Writing," in
which all vowels are represented by full letters. This mode thus differs significantly from the
mode for Quenya exemplified by the calligraphic text of "Namarie" in The Road Goes Ever On, in
which the vowels are represented by diacritical marks (tehtar).l
• The Old Noldorin or Feanorian Usage, used for the representation of Old Noldorin (Kor
noldorin) before the Exile.
• The Beleriandic or Exilic Usage, used by the Noldor in Exile for the representation of
their own tongue. This variety is consistent with the Mode of Beleriand as exemplified in the
inscription on the West-gate of Moria in The Lord of the Rings and the calligraphic text of "A
Elbereth Gilthoniel" in The Road Goes Ever On. 2
In this essay, Tolkien also introduces some of the terminology associated with the Feanorian
script: tengwa 'letter', tehta 'mark', tema 'series', as well as names for the individual tengwar and
tehtar. These include phonetic names peculiar to the values in the Parmaqestarin, Old
Noldorin, and Beleriandic modes, but also the system of Qenya Tengwesser 'letter-names' or
Qantesser 'full-names', used to refer to the letters themselves, regardless of their application in
any particular mode. Many of the letter-names presented here differ in form or application
from those given in Appendix E to The Lord of the Rings and even from those appearing in the
manuscript of The Etymologies from the late 1930s.3 For example, the letter pis called r6men
1 The Road Goes Ever On, 1967, p. 57 (2nd ed. 1978, p. 65).
2 The Fellowship of the Ring, 1954, p. 3 1 9; The Road Goes Ever On, p. 62 (70).
3 Cf. The Return of the King, 1955, pp. 396, 400-1 , and Arden R. Smith, "Appendix Ill: The Tengwar in the Etymologies,"
Vinyar Tengwar 46 (July 2004), pp. 29-34, especially p. 31.
PARMA ELDALAMBERON No. 2 2 - Page 6

The Feanorian Alphabet J. R. R. Tolkien
'east' in Appendix E, Rana 'Moon' in The Etymologies, and riima 'wing' in "The Feanorian
Alphabet." The name ampano is applied in The Etymologies to F (umbar in Appendix E) but in
"The Feanorian Alphabet" to fn (ampa in Appendix E).4
Description of the Manuscripts
Tolkien wrote several essays of varying length and complexity with the title "The Feanorian
Alphabet," some with a direresis on either the e or the a of Feanorian. In order to distinguish
between them, the earliest and longest version, published in the present volume, will be
referred to as Version A. A separate manuscript, which is clearly a partial revision of Version
A and also published here, will be referred to as Version A'. We intend to publish the later
versions in future volumes of Parma Eldalamberon.
The manuscript of Version A consists of four distinct parts:
• Part 1 contains information on the Feanorian Alphabet in general and the Lindarin Use.
Tolkien wrote this section on seven sides of four sheets of "Oxford paper."5 The sides were
numbered in pencil 1-7, and the sheets were clipped together.
• Part 2 discusses the Parmaqestarin Use, full writing (qanta-tenkele), and the names of the
letters. This section was written on twelve sides of seven sheets of Oxford paper. The sides
were numbered in pencil 8-19, but the sheets were not clipped together. Though filed
separately from Part 1, it is clear that this is a continuation of the same text.
• Part 3 describes the Old Noldorin or Feanorian Usage. The first two pages of this section,
which are later insertions, were written on torn half-sheets of Oxford paper. The second of
these is the verso of half of a discarded page on Quendian base structure. The remainder of
this section was written on six sides (plus one deleted side) of four smaller sheets from a
writing pad. All these sheets were clipped together but not numbered. Tolkien filed Part 3
immediately after Part 2.
• Part 4 describes the Beleriandic or Exilic Usage. Tolkien wrote this section on seventeen
sides of nine sheets from the same type of writing pad as that used for Part 3. These pages
were numbered 1-16 in pencil, but the blank verso of the fourth sheet was numbered 8, those
numbered 9-16 were originally numbered 7-14, and the deleted verso of the eighth sheet was
not numbered, nor was the final side of the section. The sheets were clipped together and filed
immediately after Part 3.
Version A' is a revision of Version A, Part 2, covering the Parmaquestarin Mode, Phonetic
Names, and Quantesser (full-names) . This section was written on all sixteen sides of eight
sheets of Oxford paper. The verso of the third sheet consists of rough notes on "Changes
produced in hiatus," probably a deleted page from a different document but struck through in
any case, and the verso of the eighth sheet consists of rough notes on various features of the
Parmaquestarin Mode. With the exception of these two sides, the sides in this section have
been paginated a-n in pencil. These sheets were clipped together, but the first sheet was
separated from the others by six sheets of considerably later material (c. 1955-65).
4 In this respe ct the letter-names in The Etymologies correspond more closely to the values of the tengwar in the
Lindarin Use, where m= mba and F = mpa. They do not match the Lindarin Use completely, however: band eel,
for example, are called asto and ohta in The Etymologies but have the values tsa and ksa in the Lindarin Use.
5 This was unused paper from candidates' examination booklets, whi ch Tolkien used beginning in the summer of
1924 and continuing for many years thereafter; see The Lays ofBeleriand, edited by Christopher Tolkien, 1985, p. 81.
PARMA ELDALAMBERON No . 2 2 - Page 7

Date of the Manuscripts

Based on the paper, handwriting, and textual clues, Version A appears to have been written
sometime in the late 1930s. Spellings such as Qenya, qessetema, and Parmaqesta are usual,
though Quenya and Quenta appear once each even in the original layer of writing. The appear
ance of Numen6rea as a sample word presumably indicates that Version A is no earlier than the
inception of the legend of Numenor, for which Christopher Tolkien postulates a date of 1936.6
Version A also appears to be closely contemporary with a text entitled "Qenya Grammar.
Spelling and transcription" (hereafter referred to as "Qenya Spelling"), published elsewhere in
this volume. For example, disregarding later emendations, the names and sequence of the
tengwar given in "Qenya Spelling" are identical to those in Version A, Part 2, of "The
Feanorian Alphabet," with the exception of the variant letter d. hwai, which Tolkien omitted
from the former. In it Tolkien wrote: "Each letter also had a significant name, for these see
Alphabets." The reference there is presumably to Version A of "The Feanorian Alphabet."
Though Version A and "Qenya Spelling" present the material differently, the details are
largely consistent between the two texts, and the few deviations are minor. Version A has i
tehta and e-tehta, whereas "Qenya Spelling" has {-tehta and e-tehta. Where Version A has
pusta, later emended to putta, "Qenya Spelling" has putta as written. "Qenya Spelling" also
lists some tehta-names that do not appear in Versions A and A' of "The F eanorian Alphabet" at
all: lupe, sa-lupe, ai-tehta, an, and an-tehta.
Version A' probably dates from the early to mid 1940s. At this point Tolkien was frequently
using spellings like Quenya and Tarquesta, but u-less spellings are still the norm in technical
terminology and letter-names: qanta-tefikele, qesse, ufiqe, sa nuqema, etc. Quantesser appears as a
section heading, but qantesser in the text two lines below, and on the first page of the
manuscript we find Parmaquestarin twice and Parmaqestarin once.
There are no references in Version A' to persons, places, or concepts that arose during the
writing of The Lord of the Rings, but that could merely be due to this section's Quenyacentric
nature. In Versions B and C, however, we find references to such things as the Second and
Third Ages, Arnor, Gondor, and Westron, although these versions also predate the change of
Noldorin to Sindarin (c. 1951).7
Editorial Conventions
The texts are presented as initially written, with later emendations in the footnotes. Where it
is clear that the emendations were made in the course of writing, the earlier versions are given
in the footnotes.
Boldface is used for transliterations of tengwar and representations of sounds and sample
words in Elvish languages. Boldface italics are used for names of tengwar and tehtar. Other
words in Elvish languages are generally given in normal italics. These distinctions are purely
editiorial; Tolkien distinguishes such elements from the main text by underlining, if at all.
Square brackets [ ] around expansions of abbreviations or incomplete words are editorial;
other occurrences of square brackets in the text are Tolkien's own, unless noted otherwise.
Curly brackets { } indicate deletions in the manuscript.
�
6 The Lost Road and Other Writings, edited by Christopher Tolkien, 1987, pp. 7-13, especially p. 9.
7 No secondary-world dates appear in Version A'. Those that appear in Version A are all given in Years of the Sun
(Y.S.). See The War of the jewels, edited by Christopher Tolkien, 1994, pp. 3-4, 19-27.

[Version A, Part 1]
The Feanorian Alphabet.
This alphabet was devised, both as a general alphabet, and as applied to different tongues,
especially Qenya, by Feanor son of Finwe of the Noldor. Qenya had previously been written in a
different alphabet of various diverse modifications called the Valinorian, which afterwards
passed out of use among the Elves generally.1
The general Feanorian alphabet consisted of a scheme of 24 signs representing consonants,
together with several other consonantal signs that were not included in the scheme. To each
consonant, as in the alphabet of Rumil, the vowel a was attached, the other vowels being
represented by diacritic marks (usually placed above the consonantal sign). The writing ran
from left to right.
The scheme was based on the four primary signs: Y3 c:.1 n. er , each of which could be
similarly developed by lengthening the vertical up or down (h r ), or reduplicating the bow
(v.n c.ca). The application of the signs so obtained varied much according to the require
ments of the individual languages.
In origin it was probably ultimately derived from the Rumilian alphabet by taking separate
letters, and using the hanging bar as a side-stem: thus f = R. P-
p = R.�
f?' = R. f'?2
Feanorian Alphabet.
General or Phonetic form. This form was used in works on linguistic history, or in the
description of foreign languages.
p. t. k. q.
parmatema tinkotema kalmatema qessetema
p-series. t-series. k-series. q-series.
pa ta
r ka
� kwa
1.
p. �
2.
F ba
F da
"J
ga a:r gwa
3. h pha h tha d kha .cl khwa
4.
r fa3
r .l>a
cf xa
9 x wa
5.* ln oa b da cd 3a t:d 3wa
1 The phrase "or alphabet of Rumil" was inserted in pencil after the term Valinorian.
2 This paragraph was previously published as document excerpt R24a in The Alphabet ofRumil, edited by Arden R.
Smith, Parma Eldalamberon (PE), no. 13, p. 88.
3 Tolkien later connected the tengwar for pha and fa with a brace to the left. To the left of this brace he wrote
quickly with a fine-nibbed pen: "usually reverse b.. =f f-
=ph."

6. n. va, wa
{� ra
la
'0"' wa
7. Ul. ma Q:) na c..c..r ga a::r gwa

8.
{69 or 1'sa aza, ara
Other signs 1 ha, .::\.. ya (derived from Rumilian alphabet)4
The difference between a.. va and � wa was schematic since w, y belongs as 'weakest'
grade of both p and q series: but u. could be used to represent a v (labiodental or bilabial) as
distinct from a pure w. Later 't; a turned form of Vl = r became enlarged to G.
* When 'voiced aspirates' bh, dh, gh are required, this row is used to represent them; and
'voiced spirants' are represented by )n ba, )" da, a} 3a &c. In languages that had no
voiceless aspirates but had voiceless spirants (as Q.) row 3 could be used = fa &c.
The vowels were variously represented. The usual system, especially with Eldarin
languages, or others of fairly simple vocalic system, was as follows: The sign J was the vowel
carrier: normally = a. The sounds i, e, o, u were denoted by diacritical signs thus:
) i i r 1 (or f)
a i e o �
They were always so written (1) at beginning of words; (2) where no consonant preceded
and 2 consecutive vowels did not form a diphthong. Otherwise [a] was assumed to follow
every consonant sign; but if one of the above diacritics were placed above, that removed the [a]
and substituted the appropriate short vowel. Thus f = ta; "f = ata; fo = ti; �f = uti; j3 =
tu, &c. � = te a p{' = tao (2 syll[able]s.) &c.
(2 syllables);
Diphthongs were properly represented thus: by writing above and below the consonant, in
which case T stood for [�]: as jj' = tea; Tt = teo. But the only frequent diphthongs ended in
i, y: these were normally written with stopped wa, ya (see below): as _r. 1 pc::= tal, tay. 6
In Eldarin languages, however, vowel-groups ending i, u were always diphthongal (as ei, oi,
ai, ou, eu, au), and others (as ea, eo, oa) dissyllabic. In representing these there was no
attempt made to distinguish the two: f1
was therefore the normal way of writing tai
(diphthong).
When a consonant was final, or stood before another consonant, and was therefore followed
by no vowel, it was stopped thus: I3J? = mae
4 This line was previously published as excerpt R24b in The Alphabet ofRumil (see footnote 2).
5 Above this paragraph, in the top margin of the page, Tolkien later added in ink: "In alien (non-Eldarin) tongues,
where final consonants and combinations of consonants were frequent, ( (or"') was used for a."
6 Tolkien later struck through "normally" and wrote "correctly" above it.

Long vowels were expressed by the 'long-carrier' with the appropriate signs. Since in
J
Eldarin languages a + long vowel seldom occurred, the preceding consonant was assumed to
have no [a]: thus P.J = ta; pj = te. But another method and more in harmony with the short
vowels was to use only initially and after hiatus: as Jl%t = ama; P'JY5 = ticire. Otherwise the
J
vowel marks were placed above the 'long-sign': thus - -e-m = lama; � = tt�ma.
Other diacritics.
When 'ya' U. !] immediately followed a consonant, it could be denoted by-' (yatehta), thus:
r& = tya, � = nya, f = sya.
When 'wa' [w, y] immediately followed a consonant, in cases not provided for in the
qessetema, it could be denoted by v or •, thus "a:t:; = alwa instead of full writing �.
When a consonant was long or double, this could be denoted by sign -l• or v. Thus pP =
patta; ri{ = lasse. The writing of two signs: as f"f.f (pat-ta) was only used when a word
division, or division between elements of a recognized compound, intervened.
A preceding nasal could be denoted by � as a:( = ngwa = �a:r.
The letters were named as above with affixed a: but in monosyllables the a was usually long
in speaking of a letter: as en ye ta = that is a tee. aza was so named because z did not occur
initially in Eldarin. It was in fact later called ara, owing to Q. change of z > r.
Vowel signs and diacritics were called tehta (pl. tehtar) 'marks'. Thus 1 = a-tehta, •• = e
tehta &c. The under-stop was called putta ('stop'). :. was called and6lama 'long consonant'.
- was called andatehta 'long mark'. / was called yatehta, � was called wa-tehta. ,_, was
called nengwetehta 'nasal sign'.
Note. The voiceless r and I that existed in Qenya initially (mainly derived from sr, si) were
represented either by preceding h: !Q 1c. transcribed hr, hi, or by addition of an s
hook: � transcribed rh {and called rha), c; transcribed lh (and called lha).8 The latter
method was later, and after development of z ( )!) to r.9
The Lindarin Use.

The Lindar also applied this script specially to the representation of Qenya. In this case the full
list of primary signs: n. Ill.,� 1 pl- 1 1:n, b.,�� }n-:
that is 8 in each of the 4 series, 32 in
all, was employed, but differently allotted: only the voiceless-stops remained the same. A fifth
or kya (tya-) series was made by adding a small hook to the ka-series: Ef = kya, tya.
7 The tengwar and the transliteration are clearly written here, but in this mode the tengwar should be trans
literated as vat.
8 The phrase "in Qenya" » "in archaic Qenya."
9 The phrase "was later" » "was devised later."

Since the minimal grades�. n, er, er were now used for simple nasals, but er [gwa] was
not required, this u- was used for [wa], and a modification o or a was used for [va]. Qenya
required both. 10
This arrangement reveals the fact that the established form of Qenya was already reached,
for it recognizes the Q. coalescence of th, }J/ s in s - b = sa; and the coalescence of z, r - y =
ra. The Q. coalescence of ty-ky, ndy-ngy, had already occurred also: the use of the ka-series
as base of the signs was due simply to graphic convenience - 'f was more convenient than
�·11
This is the Old Valinorian orthography of Qenya.12 It is also called the Lindarin orthography,
because this use of the letters was abandoned by the Noldor, long before the Exile, but was long
specially kept up by the Lindar, in writing Qenya or their own colloquial language. It is now
only used for Lindarin; Qenya is written according to the later Parmaqesta usage, described
below.
The values of the letters in the following table, and in the tables succeeding, are given
according to the transcription devised by the Noldor of Eressea.
The Lindarin Use.

Table of
the
Old Valinorian or Lindarin orthography.13
The values are given in the usual transcription applied to
the Noldorin orthography by the Noldor themselves in Eressea.
(a) p. (b) t. (c) ty. (d) k (e) q.
1. u � F � [er ]14
ma na nya na* ls
2. m. �
� cc:.r cx::l'
mba nda ndya nga16 iigwa1 7
10 This paragraph was emended in red ink to read: "Since the minimal grades �, n, er, er were now used
for simple nasals, signs were required for va and wa. For wa 0 was used; for va sr or S...."
11 This paragraph was emended in red ink to read: "This arrangement reveals the fact that the established form of
Qenya was not yet reached, for in its proper form it does not recognize the Q. coalescence of th, I>Is in s - b }>a,
but ' sa. The coalescence of z, r - "! ra is recognized except in early documents where l z. The Q.
=
= = =
coalescence of ty-ky, ndy-ngy, had already occurred: the use of the ka-series as base of the signs was due
simply to graphic convenience - 'f was more convenient than �·"
1 2 "Qenya" » "Old Quenya (or Lindarin)" in red ink.
1 3 "Old Valinorian or Lindarin orthography" » "Old Valinorian or Old Lindarin orthography" in red ink. Tolkien later
scrawled in pencil diagonally across this table: "Revise needed."
14 The brackets around this letter were struck through and the value "*iiwa" added below in red ink, probably at
the same time that 1:f'" was changed to 0 for wa below.
1 5 na » iia in red ink.
1 6 nga » iiga in red ink.

The Feanorian Alphabet f. R. R. Tolkien
3.
p. � ff er ecr
pa ta tya ka qa
4. p:l- � fCf "f cq-
mpa nta ntya nka nqa
5. h k �[:X,] el -d
fa sals hya ha hwa19
6. 1n. m [fcl] eel [-a:l]
psa tsa ksa
7.
r � fl 4 [�]
pta [sta] used ssa20 htya hta
8. }b. r F4 «) -crj
sqa
spa sta21 stya ska
t:r22
9. $1,2.. :s',t; � L
va ra la ya 'a wa
10 . 23 2f rda S lda > these signs earlier = rha, lha
(voiceless r, I)
* chiefly initial. In strict Q. phonetically na. Medially it is often confused with nga. 24
The old s-sign 6

displaced by the use of b was still sometimes used (1) not infrequently as
a variant of b, especially in monosyllables with no other consonant, as 6 = sa, or finally as
pcpC = peltas instead of jiGf�.25
The signs 2f � now became used for rd, Id since in the Lindarin and general pronun-
ciation of spoken Qenya rh, lh had become voiced. In writing pure Qenya (Parmaqesta),
1 7 ngwa » iigwa in red ink.

1 8 sa » "tha, pa [later sa]" in red ink.
1 9 An alternate form of the tengwa, d., was added later in square brackets to the right.
20 These values were struck through and replaced by "sta later ss" in pencil.
2 1 The value sta was struck through and replaced by "zda later sta after zd > st" in pencil.
.
22 The tengwa was deleted and replaced by 0 in red ink.
(
2 3 In column (a) of this line, Tolkien later added " sa" in pencil, then wrote " ' t'
sa " over it in red ink.
24 This note appears in the left margin of the manuscript page, to the left of line 2.
2 5 Above this paragraph, in the top margin of the manuscript, Tolkien later added in black ink: "Originally in
Qenya [p] from older [th] had remained distinct from [s] . The latter had become [z] in voiced neighbourhood,
except in absolute final position, before [p] developed; and went on after to [r] . The former became [s] but was
not voiced. Hence \, became the usual sign for [s]." He also emended the beginning of the original paragraph in
red ink and pencil (the latter marked by square brackets) to read: "In later Lindarin after coalescence of p / s = s
[after p th > s]: b was preferred; but the old s-sign G was still sometimes used (1) not infrequently as a variant
of b."
PARMA ELDALAMBERON No. 22 - Page 1 3

however, the distinction was still observed by writing h before r, 1, so �'! � or placing it in
abbreviated form above r, 1 ' t;.26
The invention of .\ hya (instead of �), a tall form of 1\- ya, was late, and not accepted by
the Lindar.
Other variations, also chiefly used by the Noldor, were:
(1) Reversal of functions of/ yatehta, and!:. e-tehta: thus p = tya; r5 = te. In this case the
r-series was used, not er-series, as a rule, and the Lindarin � J �etc. given up.
(2) cl a variant hwa beside d.
(3) For -! watehta � was frequent and ..£'used for vowel [u].27
Examples:
tengwesta qenyava mahtya anaristya ma1wa erde
phonetic, general f'�� ;�n � ,m.,� mc;cr
or � f' 0
or we;
Lindarin
....
older
� � � ·�� et; 'ft.
�
,_
cr�u
later
J�� '�� (
�.
26 This paragraph was written at the bottom of the page, with an arrow showing where it was supposed to be
inserted. Tolkien also added the word "later" above "now."
27 _7 was emended from !:. , probably in the course of writing.
PARMA ELDALAMBERON No. 22 -Page 1 4
[Version A, Part 2]
Parmaqestarin Use.
The Noldor later developed two uses of the Feanorian letters, distinct from the Lindarin use, and
proceeding independently from the original form, the values of which were in general more
closely adhered to. The first of these was a special adaptation for the denotation of the
parmaqesta or written form of Qenya; the second was employed in writing their own language,
Kor-noldorin.
This adaptation to Qenya is called the parmaqestarin orthography, and with small variations
in different times and places remains the standard manner of writing Qenya. On it is based the
transcription into Latin letters devised by the Noldor of Eressea, which is used in this account.
The principal changes introduced were: (1) the rejection of the letters with doubly extended
stem c}� � etc.), which necessitated certain changes in the distribution of values; (2) the
development of the independent 'letters' for most of the vowel-sounds. For convenience and
brevity the old method of assuming [a] to follow every consonant, that was not 'stopped' (as p>
= t), or followed by a vowel (so p1. = ti not ta+i), was retained. Most of the usual Qenya
consonant combinations were therefore expressed by individual letters or by diacritics; and
several signs no longer required in Qenya were used for this purpose: thus cl b were used
for ht, st; and the series f"-, pn, etc. used for mb, nd, since b, d, g mainly appeared in these
combinations. 28
The Noldorin origin of the system is shown by the use of the sign CA in value [o]. 2 9 This was
a loan from the Feanorian or Noldorin use (described next).30 It was due to the Noldorin
change of Eldarin [a] > [o] . The sign CA originally had the value [3], and was called [3a], in N.
with loss of [3] and change of [a] > [o], it was called [o]. Hence, as in Q. no sign for [3] was
needed, but a convenient sign for [o] was required, c.r was given the value [o].
2 8 Using red ink Tolkien later expanded "thus cl b were used for ht, st" » "thus later (after invention of 1 h =
and change of):> > s) cl b were used for ht, st."

Below this paragraph, at the bottom of the page, Tolkien wrote in blue ink over a pencil draft: "Since this
system was devised for pure Qenya in 'classical' written form, signs were still required for the voiceless initial I, r
that occurred in that language when it was correctly spelt. lh was then usually represented by 2,., and rh by f;
but the Lindarin method C 1 {11} ! rr:. could also be used."
In the left margin he wrote in red ink: "In early use initial h-breath was distinguished as 1 while medial h ( = x)
was written el. No distinction was made, however, between medial and init ... " This was struck through and
replaced by the following, also in red ink: "Originally e.I h [x] product of kh etc. But since this acquired the
=
sound [h] initially, 1 was devised for it. No similar distinction was drawn between initial hw and medial, which
early coalesced in [}u] - still earlier hy. The use of cf ht instead of ep is late and never universal."
=
29 Tolkien later added "see below" at the end of this sentence.

3 0 "Noldorin" » "Old Noldorin."

Table of the Parmaqestarin letters & values

The values are not strictly phonetic, but given according to the Eressean transcription: thus l:f
= q, phonetically [kw], but q is used in the Eressean transcription owing to the accidental
similarity of the letters in form & function.
A. Parmatema, p-series: 1. fL p. 2. b.. f. 3. t1. v. 4. 1%1 m. 5. ln. mp. 6. F mb.

7. f.. ps.
B. Tinkotema, t-series: 1. f t. b st.31 3. 6 8.32 4. � r. 5. � n.
2. 6. m nt. 7. F nd.
8. 1' rh.33 9. 1' rd. 10. '(; V4 11. <::; ld. 12. t: ss.35 13. Jt ts.
C. Kalmatema, k-series: 1. er k. 2. d ht.36 3. 1 h.37 4. ccr ij [n]. 5. c.d iik. 6. "1 fig.
7. tf ks.38
D. Qessetema, q-series: 1. 9' q. 2 . .J hw. 3. � w.39 4. "CC:�" ijw [iiw]. 5. t::ef Dq.40 6. "CXf
E. Napannar, additional letters: 1. � hy. 3. J short vowel carrier. 4.

carrier. 5. c.r o. 6. 0 u.42
2. A y.
J long vowel
A later variation is to include each vowel separately in the Napannar, thus: 1 X. 2 A. 3 c- a.
4 J a. 5 i i. 6 j i. 7 c e. 8 j e. 9 c.r o. 10 0 or er u.43
3 1 st » ")> (th)" in red ink, with a footnote added at the bottom of the page, also in red ink: "Later after TQ change
of )> > s, b was sometimes used = st, cf. eA. = ht."
32 Added in red ink: "also final unvowelled
V�·"
e
33 rh » r in red ink, and "8 [ 4a. hr]" added in blue ink.
34 Added in blue ink: "[lOa. 2 hi]."
3 5 ss » st. Another item was added in the right margin in red ink: "12 � = ns."
36 ht » h and "medial orig. [x]" added in red ink.
3 7 Added in red ink: "initial only."
38 In red ink, Tolkien added "8. cp ht," along with a footnote: "Later in some uses, c£ = ht as 1 = h in all
positions." Then he added, also in red ink: "4. � {Da} 3a > 'a," and changed earlier 4-8 to 5-9 respectively.
Finally, he added a note to new item 4 in the left margin, also in red ink: "Later when {even} reduced to clear or
glottal beginning often reduced to'-. Later lost when clear beginning was lost. Then et > value o, see below."
3 9 Tolkien added "or 0" in red ink, but then deleted it.
4 0 A faintly penciled tilde appears above the D.
4 1 Tolkien later added "7. Ef skw (sq)."
4 2 Items 5 and 6 were later struck through.
43 Tolkien later emended this sentence in blue ink and added a sentence before it in red ink, resulting in the
following: "This alphabet in its oldest and most Classical form expressed vowels by diacritics as in primal
Feanorian mode: 1 a,� i, { e, j-:» o, f u (notl). See above. But a later variation was to use er (disused by loss
of [3]) and 0 = o, u. Each vowel was then separately included in the Napannar, thus ... "
PARMA ELDALAMBERON No. 22- Page 16

Variations of form: for d [hw] a variant d. as in the N. orthography of their own language
was occasionally used, especially by Noldor. For J as sign of long [a] a variant ' was and is
frequent, especially initially as r� = .Arien. Similarly instead of'· a form r (joined to
next letter) is frequent in function [a], as cop = atta. The letter o frequently has the form er,
but this is avoided before letters of the q-series. Final t::: sometimes appears as -c or "G.
Diacritics not included in the alphabet. These were
(1) •• written above or below a letter to indicate a following y [! ]: as p or J?. = [ty]; J??, cJ, �
= [ny, hty, sty] ; y or!? = [ry]; t; [Iy]. Writing above was usual in all cases save P. � ·
=
(2) �above only, indicating a following w [tJ]: as� = [nw]; ? or� = [rw], etc.
(3) -the 'long mark' used with consonants to indicate length, or doubling, as A22m =
yavanna; and With er, t:f tO indicate length aS U I 0 = [0], [ii],44
(4) In the older system the C) and/ distinguishing [i] and [e] were regarded as diacritics and
1, £, j, j were given no special alphabetic name or place.
(5) In place of t;
•
an s-hook or flourish was sometimes used over the consonant which s
preceded; but st had a special sign and �p. [ p.:-'or ,2J = sp; � [ cf or f'J; � [ cf
or tf\were not frequent in Qenya: sq occurred naturally but not frequently; sp, sk were in
pure Q. transposed > ps, ks, and so only occurred in borrowed or dialectal words. For [sq]
er is also found. 45
(6) The dot or point below was still used, where a consonant was not followed by any vowel, i.e.
finally and before another consonant_46 It was chiefly used in case of final � n, c:; I, f? t,
� s. [r] had 2 forms of which]' was used only as vowelless form. The chief combinations
to possess no special sign were [I, r] + t, p, k, q, m, and [rn] .47 Here']' and C, [sometimes
1::'] were used: as t't!D:1- or� = alma.48 Special combinations had to be provided for
[rd], [Id] as no sign for uncombined [d] existed.49 Inflexional final s was often expressed by
the hook � as 9" = kas.
44 Added later: "For & '§=ss &t, is also found."

45 Item (5) was later struck out in red ink.
4 6 This sentence was later emended to read: "The dot or point below was still used, even in later use with express
vowels, where a consonant was not followed by any vowel, i.e. finally and before another consonant."
47 This sentence was later emended to read: "The chief combinations to possess no special sign were [1, r] + p, t, k,
f, }>, h, ty, and [rn], also sp sk rare in pure Q."
4 8 The bracketed note is Tolkien's.
49 A note in the left margin was inserted at this point: "In consequence 1/ and -;- now became fixed in the values
rd, Id. The archaic distinction (when still observed in writing) between initial rh, lh (voiceless) and r, 1 was
represented by J1lc or by 'tfte." To this was later appended a note in red ink: "signs .2.. p."
Writing ofVowels and Diphthongs

In normal use [a] was assumed after every consonantal sign, unless a vocalic letter followed:
thus 'f = [ki]; but c..r = [ki], not [ka1]. Where [i] occurred without a preceding consonant,
initially and after another vowel, it was denoted by J the short 'vowel-carrier' without
diacritic: thus �= [atar]; u{r = [vea]. Long [i] was represented in all positions by the long
'vowel-carrier' without diacritic: thus J�. 'Y.'J'Ill, �'Ill- = [ira], [rima], [ereima]. In
functions [a, i] the vowel-carriers ., were often given, especially initially, the shapes C" '
J
(see above).
The diphthongs in Qenya were ai, au; eu, oi; iu, ui. Since these vocalic sequences were
always diphthongs, and other sequences, ea, eo, ia, ie, io, oa, uo, ue, ua, always dissyllabic, it
was not necessary to special[ly] denote conjunctions of vowels as diphthongs. And especially
in the earliest usage full writing was not uncommon: n = [al], t"D = [ay], c.,ti = [01], etc.50 But
the [ l] and [y] diacritics could also be employed: thus •i or!. =[a}], (=[ay], q.. or<;:= [ol],
o = [w]. The forms '-t = [eu], "\ = [1u ] were however rare; also the variants -z,
� �
/
�.
.
The
normal representation was
ai, au; oi, ui; eu, iu
' C.
-
"
..
�0
[Note: For '"i a form

is sometimes found. Occasionally a 'phonetic' spelling with stopped
..·i
�· ty =1, y is found: then �. � alone =a}, ay after consonants.

Thus: p..., ptF =pai, pau;
� = pui; P('f" =teu &c.]5 1
Later modifications and variations

In cursive writing final � is often reduced to a flourish with or without 'stop': thus
, " L"_Q
p.�, ft'"�· p�l:. r =peItas.
s2
Final � [n] is occasionally written as an understroke: P.]1E-• "fLY� =parman.

The only source of [g], outside the combinations nk, ng etc., and of the combination [gw] in
Qenya, was the initial [g, gw] representing ancient Eldarin initial [gg-], [ggw-] . This [g] was
still so pronounced at the time of the establishment of the orthography, hence retention of the
signs cu-, ea- with values [g], [gw] . But this [g] disappeared in all the Koreldarin languages
later; consequently in pronouncing words with initial CAr, ea- [n] was substituted in the
50 Tolkien made an insertion before this sentence and modified the beginning of this sentence in bluer ink thus:
"In the oldest use with all vowels as diacritics the diphthongs were represented thus: tai, tau tp� pCZ toi, teu
p.fl (t } p 'i p� toi, teu {.lt � tui jli tiu pi but � i. could be substituted for ·; 'l. In the
earliest form of the later usage full writing was not uncommon: n [al], t"D =[ay], c.,ti [ot], etc."
= =
5 1 The bracketed note is Tolkien's.

5 2 Inserted in bluer ink at the bottom of the page: "In late or 'incorrect' use, 6 was used for all s including old tha.
)., {o(} then = st. J was used for h and o( ht. = + 6
was then not used, and was {often} used [for] ss."

ordinary tarqesta.53 Thus cc.r�

�
= :goldo, Gnome; me = :gwalya-, to pain, were pro
nounced noldo, nwalya, and n is the usual transcription. 54
In Parmaqesta owing to tradition c.c..r, ea- were normally retained in the etymological
places; but the use of a::r for [nw] of any origin instead of � is frequent. Among the Noldor
this use of ur is now usual; while cc.c- is replaced initially by phonetic � . cc.c- is thus
disused, except in such writers as use it for c.r + c..r = o instead of �.55
'Full writing'
There exists also a system called qanta-tenkele or 'full writing'. In this system the
assumption of [a] is abandoned, and since 'stopped' consonants are no longer needed, '! alone
is used for [r]. The sign v:> is given value [a]. The vowel system is thus:
a e i 0 u
, J or j " or 1. c..r 0
i e i 0 ii
/ /
� J ( " 6
ai au eu oi ui iu
..
-y:) � 'J -
CA
...
0
�
"
The following exemplary words are thus represented :-
1 Usual
transcription Noldo, nwalyar, Numen6rea, arien, anar, erde, laure
2 Normal
Parmaqesta cc.r�, ccccp. �om.i�cm(r,
Jmf�. �·�£,a�
3 Variant �czscr, [ mt'Y )56 ��ccrr.n'r- •
J� '� �"'
4 Qanta mc..�OCJit�,rno'��Q, �i)� . .,.,,��. V�
1 kaimas, parman, teuka, niule, muile, poinesse, tc�ma
2
� p-ym.� p<O'f �lot;' dl.oc;' �tt,(. pjcn-
3 crt� � ��crz UBcc � moid F��: r-?IXI
,,,. s �-c' )
4 �� '"· Pl'F· �, m.OcJ . �£) rf�
5 3 "all the Koreldarin" » "Lindarin" in red ink.
54 Added in red ink: "The Exiled Noldor who possessed initial fi still used it."
55 This paragraph was emended in red ink thus: "In Pannaqesta owing to tradition c.c..r, ea- were normally
retained in the etymological places; but the use of a::r for [nw] of any origin instead of � is frequent, since
medially only figw � (not fiw) is found. Among the Noldor this use of -ar became usual." The rest of the
paragraph was struck out.
56 The brackets are Tolkien's.
PARMA ELDALAMBERON No . 22 - Page 19

1 talat,
mahtya, istya, ohta, tengwesta, Finwe, Earendel
� c.n:l ,�'h hi,�.( (�P"�
� · Ul.d,
2
3 J�t? � .. 'l�.. � p{"f"'� hi.ari t'l;)i...-p:n.'C' or -� /
L
I' ,
4 �P-��. � c.rc:h;, .PJ�� b.CUJ "7JP;,; c

.. )
The letters were named and recited in the following order :-
1 2 3 4 5 6 7
p- h. 0.. lll.. h }XL r;
pii fii57 vii mii ampa amba apsa58
8 9 10 11 12 13 14
p k � }';) � 1�
F
tii asta59 sii rii nii anta anda
15 16 17 18 19 20
"f � c r; t It
ar arda lii alda assa atsa
21 22 23 24 25 26 27
er cl 1 c.c.f' cJ
"[ tf
kii ahta hii yii anka anga aksa
nii-yiira
o
an6
28 29 30 31 32 33 34 35
� et td >v
9'" er �
I
qii hwii I hwai 6 1 wii l)Wii anqa angwa : hyii
nwii 62
36 37 38 39 40
1
>..
J CA" 0
yii telko andatelko 0 ii

('stem') (long stem)
57 A new name phii was added below this, and a new letter p with the name apta was inserted to the right of the
number 3, all in faint pencil.
5 8 Tolkien added "older psii" below this in faint pencil.
59 Tolkien changed the name of this letter to thii and reassigned the name asta to the letter p. all in faint pencil.
60 These three names are connected in the manuscript by a brace on the left.
6 1 Numbers 29 and 30 are connected in the manuscript by a horizontal brace above them.
62 These two names are connected in the manuscript by a brace on the left.

Note (1) 24 c.c..r was properly called [gi], but since in tarqesta this was pronounced [ni], as
the name of Yn , the title nii-yara 'old ni' was usual. [an] was a later name, on analogy of ar,
ri, and was partly responsible for use of C!C.A""' as final stopped [n].
Note (2) The variant d. was called hwai (after its name in the Noldorin use (see below);
also hwii noldorinwa or 'Noldorin hwi'. It was usually omitted in reciting the alphabet, which
was regarded as having 39 signs.
Diacritics: � was called tikse 'dot'. The dot above as in -=i was called i-tehta 'i-mark'. The dot
below as in 1:::; was called pusta 'stop'.63
·•
above or below was called: ya-tehta 'y-mark'.
!:::!.. was called: wa-tehta 'w-mark'.
/
j
as in .( , was called e-tehta 'e-mark'
or simply tekko 'stroke, accent'.
was called anda-tehta 'long mark'.
,......, _? ---:::> &c. signs of s were called sa-rinke 's-hook'.
The final -e = s was called sat.
Where names were required for the letter forms P,. }' c] -r} , no longer used in
Parmaqesta, they were called fa-taina, asta-taina, ahta-taina, hwa-taina: i.e. prolonged f etc.
The forms ')xL , p� , cc.l , 'u:( were similarly called ampataina etc.; but more often ampamba,
antanda, ankanga, anqangwa.
These names, usual in grammatical, linguistic, or palaeographical writing, became the
established names of the letters as marks, and were used often irrespective of the values in
different systems. They are still used in Eressea, even by the Noldor, rather than the phonetic
names assigned in their own old Komoldorin use (see below).
Thus even in the Full Writing system the names remain unchanged and the vowel letters
� ) , c.r o are called:
[a] [e] [i] [o] [u]
rii andatelko telko o ii rather than
a e i o u
But there grew up another system of letter-names which is now general in tarqesta as the
names of the signs, however employed. This originated from the fact that many of the "phonet
ic" or alphabetic names were actually real words: thus 4 mii = hand, 6 amba = up, 7 apsa = meat,
9 asta resembled asto 'dust', 13 anta = jaw, 14 anda = long, 15 arda = day (period of 12 hours), 18
alda = tree, 19 assa = hole, 26 anga = iron, &c.64 Significant names, often retaining the
alphabetic name where this was significant, were then devised for all the letters and signs, and
also for the usual diacritical forms [as p = ty] .65 The sound was the first sound or combination
of the word where this could be initial in Qenya; if not, it was the medial combination of word
beginning with a vowel. The names were all nouns (except for a few of the diacritical letters as
i; alya rich).
63 pusta » putta.
64 In this list the glosses 'jaw" » "face"; and "day (period of 12 hours)" » "realm."
65 The bracketed note is Tolkien's.

These names were called Tengwesser 'letter-names', or Esse-qantar 'full names'.66 They were
as follows :-
1. p- parma book 2. h falma wave 3. 1:1. Vala god
4. m. Manwe (name) 5. ln.. ampa crook 6. p:1- umbar doom

ampano building
7. � apsa meat 8. p tinlco metal 9. '6 Este (name)
asto dust
/'
10. 0 silme moonlight 11. .o rana moon 12. � nengwe nose

13. b anta face67 14. r ando gate 15. 1' riima wing
antaro high mountain
16. � arda realm 17. "'G lambe tongue 18. Z5 alda tree
19. 6 esse name 20. � atsa claw 21. 'l kalma light
22. c..l ohta war 23. 1 harma treasure 24. c::cr yoldo Gnome
yolwe lore
25. cd ankajaw 26. CCf anga iron 27 . � aksa ravine
ankale Sun (poetic name)
28. er qesse feather 29 . .cJ hwesta breeze 30. cl hwinde eddy
31. 1%' winge foam 32. ccr ywalme torment 33. ..cJ unqe cavity
unqale agony
34. -c1:f ungwe gloom 35. � hyalma conch 36. J\. yanta bridge
37. 1. tellco stem* 38. anda tellco long stem* 39. cc Osse (name)
J
40. 0 Ulmo (name).
* These are descriptive names not
representative of sounds.
For the series n + voiceless stop ( c:...c:I 1-;r., b. cd ) a long and a shorter name are found:
either all are long ampano, antaro, ankale, unqale, or all short ampa, anta, anka, unqe. The name
of b is usually Este (goddess), but was originally asto .
66 Esse-qantar » Quantesser » Qantesser in blue ink.

67 Tolkien wrote "jaws" faintly in pencil above "face." Probably at the same time he struck out "jaw" as the gloss
of anka, also in pencil.

The Feanorian A lphabet J. R. R. Tolkien
The letters b = forrnen North; l'Y.) = numen West;? = romen East; )v = hyarrnen South are
usually used to denote points of the compass: b. � "'! = N S E W. Consequently these
t"::Y.)
names are often used in denoting these letters, instead of falma, hyalma, rama, nen. *68 Note
that "/'· even in the system where it is used only as 'stopped r' , does not have a name
containing this, since the single consonants are all represented by words beginning with their
sound-value.
* In naming points of Compass in Q. order is always W E S N: '"lA'h- . Since West =
Valinor and the face is turned thither; hyarrnen 'left hand' is thus South; forrnen 'right
hand' is North, and placed last, because of Morgoth, h... is thus called 'the ill omened
letter'.
Other names not having an alphabetic place are often used, as follows:-
1. p ty tyelpe silver. }? sty istyar scholar.69 ;:, or 1 ry arya day (12 h[ou]rs). r'-' ny nyelle
bell. � nty intya guess, notion. F ndy indyo grandson. 1:; ly alya rich.
2. r tw atwa double. � or 1 rw arwa having, possessing. -t; lw olwa branch or alwa well
-.,
grown. l?'3 nw anwa true.
3. OS nn anna gift. a. mm amme mother. p tt atto father. Cf kk ekko point, spine,
thorn. r5 or t o"o! alas! � 11 alla hail!
4. C" [= a] anar sun. ? [ = i] iira dawn. t Ingwe (name). j (or C ) = [i] {re longing. (, (or J) =
[e] Elwe (name). J = [e] ele star-ray, beam. a = [o] oma voice. 0' = [ii] ur fire. C.Z = [au]
Au.le (name). •;. , ..a. , t = [ai] aire sea/0
68 nin » nengwe.
69 "scholar" >> "wizard, lore-master."
70 aire » airen in pencil.

[Version A, Part 3]
The different phonetic character of Noldorin necessitated a system quite unlike the
Parmaqesta. 7 1 In writing their own language, therefore, the Noldor made another alphabetic
arrangement (with phonetic names).72 This never became fixed as did the Parmaqesta, but was
constantly modified to fit the changing Noldorin language. There were however two main
forms: (1) the Feanorian or Old Noldorin used before the Exile, and now obsolete: (2) the
Beleriandic or Exilic adapted to the later forms of Noldorin, and still used for Eressean Noldorin. 73
The Old Noldorin or Feanorian usage.

Alphabet.
After the letter is given its ON name in transcription; under the letter is given the usual
transcription followed in square brackets by the phonetic value.
1. 2. 3. � kii : the shape q is rarely used.

p. pu
r tO
p [p] t [t] k [k]

4.
)Do bu 5. )-n dO 6.
CIJ gii : the shape "J is rarely used.
b [b] d [d] g [g]
7. b. phu 8. h thO 9. cl khu : sometimes the shape -J is used.
ph [f bilab.)14 th [l>Y5 kh [xY6
10. b. ampa 11. m anta 12. c.cl anka : sometimes the shape � is used.
mp [mp, mph] nt [nt, nth] nk [gk, gkh]
7 1 Noldorin » "Beleriandic" in pencil, but subsequently struck through.

72 "own language" » "own later colloquial language" in pencil, but the insertion was later struck through.
73 This paragraph was written on a half-sheet, which Tolkien clipped to the front of this section. Below this,
probably at the same time that he made the penciled emendations, he wrote in pencil: "altered [?if] O[ld]
Noldorin Quenya." This page is followed by another half-sheet:
=
1')..
I ! tP
Foll[owing] pages require revision. This application is rather suitable for Beleriandic.
N[oldorin] kept cw, gw initially.
au, {ai (ae) »} ae, ui y

ei, ai c:e = o > e
c:ei = > ei
Note Pengolod was of mixed Telerin (Doriath) and Noldorin ancestry, though living in Gondolin.
d :zr a
The paragraph beginning with "Foll. pages ... " was later struck through in ink, and again with a wavy pencil line.
This page was written on the verso of half of a discarded page on Quendian base structure.
74 "[f bilab.]" » "[ph > late ON f bilab.]."
75 "[)J]" » "[th > late )J)."
76 "[x]" » "[kh > late x)."

13. p. spii 14. r sto 15. � skii : sometimes the shape � is used.
sp [sph] st [sth] sk [skh]
16. f. azba 17. }n azda 18. c4 azga : sometimes the shape � is used.
zb [zb, ab] zd [zd, ad] zg [zg, ag]
19. 'D1 amba 20. � anda 21. -ea anga : the shape c.c.c is rarely used.
mb [mb] nd [nd] ng [gg]
22. 1:1. mu 23. � no 24. er' 0 25. -a- wai
m [m] n [n] o, o [ol w [w]
26. 0 ii 27. C. astinta 28. A. estinta 29. 1. i
u, a [lil (and before cl� c:J, cc) c, ) e [e] i, i [i ]

a [a]
30. 31. )., hyo 32. cl hwai 33. t so
J yo
y [j, ! ] hy [b later xi]

'
hw [lu , later X¥] s [s]
34. �- essa 35. ti lo 36. 7 ro [37. 7 , 7 nurra
ss [ss] l [l] r [r] a [a]
38. l ho ]
h [h].
Other letters in occasional use were: 1f rho [voiceless r]; -;"' lho [voiceless l]; &. mhii
[voiceless m]; • nhO [voiceless n]; and in very early use also }Is azma [zm], 'l'J azna [zn].
Notes.
(1) The alphabet consisted normally of 36 letters. nurra 37 dropped out of use early in ON
with the conversion of [a], only occurring finally, into a. It is transcribed a since this
indicates its later development and normal ON form, and is distinctive, actual a re
quiring no quantitative mark: it was in ON always short. The sign J was reintroduced
from the original Feanorian general alphabet only in late ON to denote h (at first absent
from ON) redeveloped from initial kh, hy and intervocalic s.
(2) Owing to the labialization of the kw-series in ON the letter-series of Cf and q- became
interchangeable. Where the stem did not extend above the letter the closed-top forms
Cif• Clllf'" _. were almost exclusively used. But where the stem did extend upwards the
open forms were preferred: thus cl , � , c.cl, c4 . The use of the closed forms in these
cases was limited normally to occurrence after c. [a] : thus c.d = akh, to distinguish this
from ..l [nk]. But the distinction was often accomplished by writing e as '1: thus qcJ .
(3) But already before adaptation to Old Noldorin c.r-, a"' had ceased to be consonantal,
owing to the disappearance of the nasals [g, g'] for which they originally stood, and
had become used as signs for o, y. (w) . This use persisted in ON.

(4) l. . c:1 already invented before in other applications of the Feanorian syllabary were
regularly employed in ON for the voicelessj [b]
transcribed hy, and the voiceless w [h>]
transcribed hw, derived from PQ sy, sw. Initially [g]
endured until the end of the ON
period. On phonetic changes affecting these sounds and consequently their spelling,
see below.
(5) The sound [:g]occurred in ON principally in the groups :gk,
:gg, which were denoted by
the special letters eel. c:a . The PQ combinations nm, nn (from D, 3 + m, n) appear to
have given gm, gn, falling together with gm, gn from PQ k, g + m, n. Hence ON
standard spelling apa-. � · Later here ar g became [:g],
as bm, bn > mm, mn; dm, dn
> nm, nn; but ON having no special sign for [:g]
retained the gm, gn spelling. The
nasality is indicated by occasional spellings .._.. , _, = ngm, ngn.
(6) The sound [z]only occurred in zd, zg and rare zb. For each of these ON used a single
letter J- , cc] and )-. . The phonetic changes that overtook these groups do not there
fore normally find expression in ON.77 The exilic development (to aw, Bj)
suggests that
)a , � became during ON [ab, Bg], but these letters continued in use. zd, however,
became d, with lengthening of the preceding vowel, and in consequence in late ON �
is often substituted for ')- .7 8 Since a sign for [z]
was not required, the old z-sign t was
employed in ON for the very frequent group ss, since G was not convenient for the
addition of the length sign / (andatekta).7 9
Compendia and diacritics.

Diacritics.
The following signs, not counted in the alphabet, were commonly employed in ON writing.
1. ..L called (andatektha) 'long-mark'. This was used to denote length of the sound repre
sented by the letter below it. It was principally used above 10 u. "t; [n], [m], [1], [r].
1 Its use
above p [t], [p], [k]
p. er was rare after the earlier period, since ON kk becamett, pp, tth, pph,
kkh , for which compendia were used: see below; b, d, g did not occur long, and ss had a special
sign. Length of vowels was only exceptionally indicated by /. Quantity in ON vowel-signs is
assumed to be long unless otherwise marked: see (b) next. 80 But when a long vowel stressed
came to stand in a final syllable owing to the loss of nurra [a]
in ON after I, n, s, then r,
andatektha was used to mark the length, and also by implication the abnormal ultimate stress:
thus *galiso 'he grows' > *gal5sa > *galas written �.
2. T called (stintasse) 'shortness'. This was only used under the vowel-signs o er [u], [o],
t [i] - thus 9 , c;. ) - to denote that they were short. It was mainly employed in stressed
syllables (thus by implication marking the incidence of the main stress); never in final syllables
which were normally unstressed, and contained short-vowels except when marked / as
77 This sentence was revised in the course of writing from: "The phonetic changes that overtook these groups do
not therefore find expression in ON. At the end of the ON period they became ad, Bg, ab."
78 These last two sentences were emended to: "The exilic development suggests that ')- , )a , � became during
ON [a, ab, ag], but these letters continued in use." Then "[8, ab, Bg]" » " [i d, 1Jb, Ig]" in pencil.
79 "(andatekta)" » "(andatektha)."
80 "(b)" here means the following paragraph, numbered "2."

a
described in 1 . 8 1 The long values ii, o, i were assumed where ; was absent. � , Jt.. [ ] , [e] were
always short, since PQ i, e had become o, i: hence their names a-stinta, e-stinta 'short i, short
e'. In transcription the ON method is employed, where the exact ON spelling is represented.
But in the grammar and Etymologies long vowels are marked with the macron (as o, ii), short
vowels are unmarked; i represents ON [ a] nurra, which became a.
The chief exceptions were i in standard ON (from at), long and distinct from ) (i); also ii
occurred finally distinct from ii. c.r in final syllables was short unless marked 1 .82
n' n m
3. ,_ called (amba-nO') 'raised denoted an or following the vowel sign above which it
was written. It was seldom used in formal writing. Since the most frequent groups with nasal
+ consonant had special letters (as 1'" nt,
,,� nd),
while nn, mm
were denoted by .J, .J. , it
was used chiefly as an abbreviation as in ""- = ana.
It was specially frequent in later ON to
n:
express final as �1:' = aran
'king'; or ns �� kanse.
as = Since ns
became ss, while final
n was usually lost in early Exilic, this use probably indicates recognition of a weakening of n in
these positions. 83 The sign was in ON seldom used above consonantal signs: note that when so
used � = nm not nn.
4. � called (amba-wai) 'raised w'. This was frequently used above letters to denote a
following w: thus � = nw, =
� = khw, ; rw, etc. 84 Its most regular use was to express the
diphthongs ay, and the archaic oy from PQ 9 which later also became ay:
thus -c;. [ y] (and a
archaic a [oy]). These were seldom written c.c..,, or ea- .
5 . .!!. called (amba-yo) 'raised y'.
This was the most frequently used diacritic. y [j]
after a
consonant was frequent, and almost always expressed by � : thus !j [ry] , :C [ly], 'f [ky], .p,
[dy], Y.. [ny], etc. After consonants j was chiefly used to denote division of a compound: as
�c.��. -
= Elyanme'sky-bridge' se. 'rainbow'. But in less formal writing [y] at the beginning
of words or in compounds or even within groups of connected words was often expressed by :!
as 'L;; �.._ = in yanmi'the bridges'. The diphthongs aj
and archaic ej
(from PQ re), which
later became ai,
were written E (A.) [ai
was never written c:l , but ; was sometimes used in
the archaic period] .85
6. 6 [s] after a consonant was usually reduced to a hook -, t::' called in ON gambasse (in Q.
called salenka).86 Thus p., = [ps], � = [ks]; more rarely used were � [ns], lf [rs], GO [Is] &c.
r- was in older period used = [ts], but this group later became [t.i>], and so fell with tth from
tt: written }l (for f'h) : see compendia. In informal writing final s was always written ..., as
1'"� = dis.87
8 1 "never" >> "rarely"; "and contained" >> "and usually contained."

82 This paragraph is a note written at the bottom of the manuscript page.
83 The first s in ss appears to have been struck through in pencil.
84 "� = nw" was deleted.
86 salenka » sarinke.
87 This sentence is a replacement for that written at the beginning of the deleted following page: "In informal
writing final (, was always expressed by --, , as Cli:fCO� = galos; yni..., = dis."

Compendia. 88
For the frequent later combinations of ON tth, kth the combined letters [ }=I = p+ 'b tth, and
cp = Cf + \, kth] are sometimes found. 89 Less common abbreviations are:
= }!- =p.b pth, tJ =
crJ kkh )l ph pph.

, =
After the development of ON > ts t.l>

and development of generally to I>, the ON 1� = th ts
became t.l>
and was given a special shape 1'\.. = t.l>,
and from it were made 3r = (z t = kth ks),
�= pth
(}b ps).90 ps, ks
having at same time become
= and were written \ eJ. phs ps, xs
Signs ofsound-change in ON spelling.
A marked change of ON was the aspiration of t, p, k after all consonants. Since sp, st, sk and
mp, nt, nk had special letters, this did not receive spelling-recognition in ON, except that in
very late ON absolute initial sp, st, sk �, cJ , p became occasionally written \.. 1 },. cl , i.e. ph, th,
kh, or bh. (\l), "" (t/), clcl (cl'), indicating a change > sf, si>, sx > ff, .1:>1>, xx.91 This does not
occur initially (and here sp, st, sk remained in Exilic). la. b, , � remained unchanged in use,
1
and here the aspiration probably did not proceed to spirantalization until the Exilic period.
The letters p. are transcribed sp etc., and 1u. mp without reference
. to exact phonetics (initial
� )t c.1 are sometimes transcribed sph, sth, skh) .
But in other cases the aspiration became early marked, hence earlier ON [pt] pp , [tt] pp or
I, [kt] 'I' and [pp] pp or -1-. [kk] err or � are normally spelt as described above P" etc. or
compendiously Jt., . Hence in normal transcription pth, tth, kth, pph, kkh are used. Similarly
88 The remainder of the deleted page mentioned in the previous note, struck through with a wavy line in crayon,
consists of a rejected version of this section, which reads thus:
Compendia.
The groups t-th, p-th, k-th, derived from various sources but all having in later ON the values [tlJ, plJ,
klJ], were frequent. Less frequent were pph, kkh. These became written, when the aspiration (and later
the resulting spirantization) of the second element became established, pb t-th (for PI', f tt);
ef
= =
� = p-ph (for pp. . ,l pp); = = k-th (for cyp kt), pn

= = p-th (for � pt); cri k-kh (for
= =
err I cf = kk). By combining these letters, with one stem, the following compendia became usual: p =
tth, tl>; 2J kkh, kx; + kth, klJ. Less usual were ). pph; }!- pth. For tth 'B or 1'-- was also used
= = = =
since [ts], which it originally denoted, became tth.

Thus makta uc.y c. later > 1ac.4oc. maktha (or in full c��). kapta c:rrr c later >
= ,;,c.
= kaptha (or in full crerl,C). }atta, D&tse �C: I
1'1) �A, later > �C (or t;ertr >• ,.
(or n�) = lattha, natthe. kappa CfC ,lc. later > � (or in full er"'p.k) = kappha. lukko
1:0� later > 't;�GI"" (or in full '"t:."cyclc.r> = lukkho. The using of the long stem possibly
indicates that the whole of these groups were becoming spirantal: i.e. lattha > lalJlJa, lukkho > Iuxxo.
Such at any rate is the Exilic development.
90 "became" >> "now."
9 1 "sf, slJ, sx" was corrected in the course of writing from "sf, sth, sk."

in rth, lth. The older ts is often transcribed ts, although it early also became tth = t!>, as is
seen in the spelling }J for ts, and still more in the employment of rz:, = tth.
The sound h did not exist in the older period. It redeveloped in ON (i) from cl = kh > X
when this stood in absolute initial position in a word group; (ii) from older l. = [b] in the same
position; (iii) for s between vowels.92 The emp loyment of this l is not consequent or
consistent, the traditional spellings with J 1 '1. . t remaining always in use. But it is notable
that it only appears in certain cases, and so indicates that the beginnings of the differentiation
of certain initial consonant sounds, from which the Exilic "mutations" proceeded, goes back to
the late ON period. Thus 1 h only appears for c:l or l.. in absolute initial position in a word
group, and not after a closely connected word ending in a vowel. So l(u- hiro for khiro but
i.cl'?c.r- (not ,1,yc..r- ) for ikhiro 'the master'. On the other hand 1 h for 6 s only appears
medially, or rarely when preceded by a closely connected word beginning with a vowel, but not
in absolute initial position.
Signs of the change 'l. [g ] > X1 medially are seen in the confusion in late ON of medial tl =
khy and l , and in the occasional appearance of initial -a.l after the article 'i. [i] 'the' for l .
Similarly a sign of the change of d. [lu] > x ' is seen in the use of this sign instead of '!I or J.«
for the rare khw (derived from PQ khm).
The initial PQ combinations nd, ng,
mb were anciently preserved in ON. In late ON for
"'" arl G2. appear }'" 1 «f . pt- , i.e.
• d, g,
b indicating a change by loss of the nasal in absolute
initial position or after a consonant. But the nasal forms are preserved after vowels (as still
occasionally in archaic Gondolic).
ON zd -d
> (with lengthening of the preceding vowel), zb
> ab (<Jb), > zg 8g:
see note 6
above. The change zd d
> is indicate[ d) in late ON 1" for l" .
92 Item (i) was originally written as "from cl = kh > x when this stood in absolute initial position in a word group,
or after a final con ... " but was left incomplete, and the words after the comma were struck out.

[Version A, Part 4]
The Beleriandic or Exilic Usage
Alphabet.
1. J1 pui 2. )=' tau 3. Cf' c{ 4. ,.- cti
p [p ] t [t] e [k] (cw) cu [kw, ku]

5. )D- bui 6. � dau 7. "5 gi 8. � gd or gwae
b [b] d [d] g [g] gw, gu [gw, gu]
9. h phui 10. 'b thau 11. cl achau, acho 12. c.d chwae
ph [f] th [l>] eh [x] ehw [xw]
13. ln afui 14. 1» a8au, a8o 15. cd *aghau, au-feleg93 16. .r:J awae-feleg
f [v] a [a] (gh*) [3, later au] w [w]
17. p. espui 18. r estau, esto 19. � esci 20 . ., escu
sp [sp, initially esp] st [st, initially est] se [sk, initially esk] scu [sku, initially
se-w esku]94
21. 1:1. mui 22. � nau 23. er odew 24. er awae-dithen
m [m] [ ] n no [oJ w [w]
25. D1 ambui, ammui 26. ,-n andau, anno 27. cu' engi 28. a:r angwae
mb or mm nd or nn [gg gg, g] ng > ngw, ngu
[mb mm, m > ] [nd > nn, n] [ggw, ggu]
[29. "!:, or ra.. amhui 30. "'C alau, alo 31. � lhau 32. .., arau, aro
mh [v] later obsolete] 95 1 [1] lh [voiceless !] r [r]
33. if rhau 34. � sau 35. ' essau, esso9 6 36. 'l, hau
rh [voiceless r] s [s ] ss [ss ,> s] h [h]
37. cL (hwae) chwae 38. J gasdil 39. 0 u 40. 9 jdew97
Gondoliaren , ['] u [i'iJ y [j]
hw orwh [h>] ,
41. c,, before � .1 &c.
• 42. "' edew 43. \ 1 44. j iau
c, adew e [e] i [1 ] i [t]
a [a] 98
93 au-feleg was corrected in the course of writing from au-veleg.

94 "[sku, initially esku]" was emended, probably in the course of writing, from "[sku]."
95 mh >> ni .
96 essau, esso assau,
« asso,
presumably before the section labeled The names (below) was written.
97 jdew was emended, presumably before the next page was written, from jdew.
98 Below this table, Tolkien later wrote in ink, overwriting a similar note in pencil: "In addition for voiceless n /!,
(nhui ) was used, and [for voiceless] m .L (mhui) [was used]." Then "(nhui )" »
"(anhau)," and "(mhui )" »
" (amhui )."

The diacritics not given alphabetical place were :-

•
peg eriol 'single dot' .
.!!.. peg dadol 'double dot', or { orthannen 'raised i' .
.l teith 'stroke' or andeith 'long-mark'.
et. nau orthannen 'raised n' .
awae orthannen 'raised w' .
"'
- edew orthannen 'raised e'.
thinnas 'shortening', sign of short vowel.
The names. au was derived from ON o in such names as tau = ON to. The ui in such names as
pui, bui and the remainder of the labials was altered to provide a distinct vowel for each
series.99 ae in chwae, awae, etc. was the regular development of ON ai after w. Since, for
reasons given below, the old Feanorian 3rd and 4th series were again distinguished, the 3rd
series Cf etc. were given the vowel i, thus d.* The vowel a was prefixed to all such letters as
did not occur in absolute initial position in Exilic words; the ending au, i, ii becoming thus
unstressed was later reduced to o, i, u, thus acho, aDo, asso, anno, alo, aro. But the letters p
esto, etc. which did not in N. occur initially were given a prefix e- after esso, altered from ON
name essa. In compendious writing they were thus used for est, etc. initially; and for initial est
etc. in Doriathrin or Doriathrin loan-words (where est etc. was the normal form taken by old
PQ initial st). engi shows i-affection from *angi.
Since the name 'aghau' by loss of gh [3] became au, this letter, c.cl , was called au feleg 'large
au' after ...1 awae-feleg 'large awae', so called in distinction from a- awae, or awae dithen 'little
awae'. Only those vowels that when long were descended from ON long vowels, and were not
dependent on later lengthenings, were named by their simple long values ii, i. The remainder
were called by the normal short value prefixed to tew 'letter' (archaic teiw < ON tegma, Q
tengwa) : odew, adew, edew, ydew.
* But au, o after the spirants as *achau, -o, *aghau.
Notes on the use ofthe letters in Exilic.

The Old Noldorin usage was rearranged, partly by deliberate redistribution of the signs to
provide notation for new sounds, developed in Exilic, or found in the Ilkorin and Danian
dialects of Beleriand (which adopted the Noldorin letters), and partly by shift of values owing
to the retention of spelling after change of pronunciation - thus :).. becomes [h], since hy
which it originally denoted became h initially. The latter process plays a small part, however,
since the Gnomes knew the history of their own alphabet and of its various applications,
having among them Feanor himself and his sons, and brought with them and continued to use
and write not only their own changeable tongue but also the Valinorian "Quenya," the
parmalambe or 'book-language'.
Thus the changes whereby the distinction of the 3rd or er -series and the 4th or .,--series
was revived; and the letters In, 'b.,. c.cl etc. were restored to the notation of voiced spirants -
absent from ON but redeveloped, either at the very end of the pre-Exilic period or in early
Exilic - were simply returns to the original Feanorian scheme.
99 "altered to provide" » "altered from ON ii to provide."

The Feanorian A lphabet ]. R. R. Tolkien
The 4th-series. Since ON w became initially gw in Exilic, Cif'" was normally employed as gw.
The sound ngw also occurred in Exilic (i) as the analogical initial nasal 'mutation' of gw (for
historical nw): as in gwaea 'bond', pl. i ngweia 'the bonds'; (ii) in compound words as ang-wea
'iron bond, fetter', or in words ending in ng + suffixes as wen, wed. For this Cllll"' was employed
- though in analyzed compounds not made with a suffix separate writing was no more usual
=
in formal writing: thus ..� i-ngweia; but usually � (or � ) = ang-wea.
Similarly -t:J was employed for the rare collocations of medial se +w. But � cw did not occur
at all in Exilic Noldorin; and this letter is mainly used in Doriathrin and Danian, in which cw
was common. But ,.. was used in certain words derived from Ilkorin contain[ing] cw, which
was rendered cu in Noldorin. Thus Falathrin: cwen 'a sea-bird' (PQ *kw�ne 'petrel', Telerin
pine, Q qene, not found in ON where form would have been *paine) became Exilic N. euen,
spelt cp.o . 10° For this reason the letters ,. etc. of the fourth series were in compendious
=
writing frequently used as syllabic signs di, g\i, esg\i, ang\i. This was specially the case with
Cf•ee\ • which were rarely otherwise employed in Noldorin. But the sound-group [3w], properly
denoted by .cJ , did not occur in N., nor in Doriathrin or Danian, hence this sign was seldom
used at all, though retaining an alphabetic place and name. It was used in Gondolic sometimes
for the initial w that was the apparent mutation of gw. Since g by vocalic mutation > ' (nil) by
loss of [3], it was assumed that gw by vocalic mutation > 3w > w; historically w was normal and
only gave gw in absolute initial position.
Since the voiced spirants v, a and archaically 3, developed from ON b, d, g between vowels,
became common in Exilic and were denoted as originally in the Feanorian system by tn., m ,
cd , the nasal groups mp, nt, ne had no special letters. They were denoted either by full
writing ap. , "'' · '"'f - rarely and archaically - or usually by placing above the consonant the
=
diacritic - 'nau orthannen': thus 'ji , jS, Ef mp, nt, ne. NB. in ON - had denoted a nasal
following the sign above which it stood. This use was still preserved in archaic Gondolic, where
- was placed above vowel sounds to denote the nasalization which occurred in the process of
the change from m, mh > v, w. Thus glaiw (�) < ON glaibe 'salve'; but �� sa�w. =
from *sa�mh, *sa3mh < ON sagma 'poison'.101 This use soon became obsolete, since the
nasality was lost early in all dialects.
The voiceless character of absolute initial l, r was often undenoted in early Gondolic, and
c. 7 were used in all positions. Rarely �. "1')... were used for the voiceless forms. Generally
during the Gondolic period, and since the voiceless forms - chiefly initial, but also derived
from contact of l, r + h medially - were written with the modified letters "tO . 7f , named lhau,
rhau, transcribed lh, rh. Since in late Exilic and Toleressean medial lth r:J. became voiceless
ll; medial lh from l + h (s) is often written unhistorically -d- .
The letter j , though it remained as a separate letter in the alphabet, with the name iau
(from ON yo), was in fact only used as a variant form of t [i], with the same functions. It was
=
favoured in use, instead of l : finally; in j i 'the' prefixed to following words; and after vowels
100 Below "*paine" Tolkien later wrote p�ne in pencil.

1 0 1 The word glaiw was written twice in the manuscript; the letter a appears to have been emended from
something else in the first instance. The tengwar spelling of sa�w was preceded by another spelling, which was
deleted with a heavy, wavy line. Finally, "*sa3mh" was emended from "*sagmh."

as the final element of diphthongs: in the last case cj was always used, not cl . But later in
diphthongs � was usually employed: as o , older o; or more frequently oj = [uj]. This
change in use was due to the loss of all spirantal quality by ON y = [j ] in Exilic, and to its
frequent conversion into vocalic [i].
The raised � 'peg eriol' was used in Exilic as in ON as a mere definer of -. , J to facilitate
analysis of letters. It was also, in archaic Gondolic period only, used as a modifier of vowels as
er, � (beside a., � ) = re, y: see Vocalic notation. In later Gondolic, and so since, it was
frequent[ly] placed just before or above consonants that had suffered a recognized initial
mutation. It was thus especially employed together with the prefixed article i. In compounds,
regarded as one vowel, it was seldom employed. Thus Y��-ld�� = i vegli vorn 'the
black bear'; but the name Meglivorn is written IU"'(C-L "T"" ·10 2 Similarly �� dor 'land';
)•mc7 HJ6r 'the land'; j·1� i· nnjr 'the lands'; ''fe.•o �i�'G-.1 aran thithlum
'King of Hithlum'.103
Transcription.
The transcription used by J£lfwine, based mainly on Old English, and partly on Irish use, was
variable. Hence e for k.
The voiceless spirants 1o [l>], ll. [f], cl [x ], --d [xw], cl. [hw].104 These he represented thus:
b by l>, a (not initially), or th. Medially where Old English had not in his time a voiceless [l>]
normally, he often used th, or even l>l>, although the sound in Noldorin was never long in his
f,
time. ll. by ph. Medially where OE had not normally voiceless [f], he often used ph or ff,
although [f] was never long in Noldorin in his time. cl by h, especially when final, since OE h
in his time represented [x] finally. But initially and medially, where [x] was not usually found
in OE (except in ht, a combination not found in Noldorin), he used eh, and medially sometimes
hh. But initial eh in N. was only found as a 'mutation' of h, and this mutation he often
neglected, writing h: as in aran Hithlum = N. [aran xil>lum] 'King of Hithlum'. 4 [lu] did not
appear in N. in his time, but he used hw for -d [xw], beside ehw (especially medially). ccJ, &
[3, v] were obsolete in his time, but if he found them in archaic spellings, he represented c.c.,
by g, h (or gh), and -. by m, or mh.
f,
The voiced spirants ""� 'bL he represented by l>, a, d and by u and b (rarely).105 Initially
where l�. ba. appeared only as mutations of ,.,.. p:L d, b, he never used l>, f,
but often
neglected mutation, writing d, b. If the mutation was noted, he used a, and u (v) - rarely b.
f.
Medially he used a and For 11" [w] he used the contemporary OE letter p , transcribed w.
In later Toleressean this has been regulated thus:
b 'h cl � are transcr[ibed] ph, th, eh, ehw.
la.. I. are transcr[ibed] f, a - b sometimes being used (instead of f) as the init[ial]
mutation of b.
tr is transcribed w .
..._ , c.J in archaic forms are transcr[ibed] mh, gh.
1 02 The tengwar here actually read meglivurn.

103 A note in faint pencil appears below the words aran Chithlum: "no, genitives were not mutated".
1 o4 " [hw] " >> "[h:>]."
105 .. __ -- " << .. __ -- - _, , .
1� u;a. 1 � Ea;lloo c:c� m th e course o f wn•t•mg.

In anglicizing Noldorin names and words - in the Quenta and the Silmarillion and other
allied histories - this transcription is modified. Partly for practical reasons. The use of e is
retained as this is characteristic; hence such names as Celegom must be pronounced with e =
f
k (not s). 106 But = [v] is usually represented by v, as in Dagor-vregedur, Meglivorn, &c. But
f
finally, where v is avoided in English, (as in of) is usually employed, although it is not
f
voiceless. The voiceless is thus (finally only) represented by ph, or ff (although it is not long).
th, eh are retained, but must be given voiceless spirantal sounds: thus th only as in thin, path,
athlete (not as in this, breathe); eh as [x] in German acht, Scottish loch (not as in church).
The voiced dental spirant 1� [a], as in modern English this, breathe is properly represented
by a or dh, but is often represented simply by d.
In general the older Gondolic spelling or sound (in transcription) is retained in old names.
Thus nd is retained instead of late Exilic nn, and medial mb appears in a few old names,
though tsL is usually transcribed mm: as in lhammas 'linguistic history'. Thus Gondolin,
Grond, Beleriand, Angband, Talagand, &c.
Note the diphthongs ae, au, ei, oe, ui should be given the approximately equivalent
diphthongal sounds occurring in English: die (N dae); now (N naur) or mouth (N mauth); clay (N
dei); coin (N eoen); ruin (N rhuin).107
Changes in use due to Exilic sound changes.

The sound [}u ] became [xw]. This probably occurred in all positions - possibly already in late
ON. At any rate ON [hw] a. and [khw] � . clcr" [derived only from PQ khm] appear generally
in all Exilic dialects as xw, spelt normally "c1 . 108 But in the N.W. dialects (Hithlum and
Gondolin) - either by a preservation of a special initial pronunciation, or by a weakening in
absolute initial position similar to that of x > h, the sound was a voiceless w [m] still in absolute
initial position. This was naturally written a. , called hwae - or owing to its absence from
other dialects and from post-Gondolic, chwae gondoli6ren: 'Gondolic chwae'. Since it was also
normal as an initial sound in the other Beleriandic dialects, it was also called later chwae
feleriandren 'Beleriandic chwae'.109 In late Gondolic and post-G[ondolic] Exilic it became ehw,
and so fell out of use in Noldorin.
Early in the third century the change (beginning to appear at the end of the second) was
complete, whereby the voiced spirant [3] derived from ON g became vocalized, and was lost
between vowels. It was long retained in Gondolic spelling [cci transcribed gh], especially
where it had not become a syllabic vowel as in tara 'tough' < targh, ON targa, and its original
place was still marked by hiatus.110 But later in Gondolic spelling the use of u.T became rare.
It was sometimes retained to mark the mutation of initial g: it is then transcribed gh or " since
no consonantal sound was actually preserved: thus Clllf""C"'"" golodh 'Gnome', ic.clc.reu-\r,
igholodh or i"oloa 'the Gnome'. The old h-sign 1 (no longer required as h since this function
was exercised by ')v) was reintroduced and called gas-dil 'stop-gap', and employed to mark
hiatus caused by loss of g, 3 and so (since the medial-hiatus was usually got rid of by
106 "reasons. The" << "reasons, the".

107 au >> "au (aw)".
108 The note in square brackets is Tolkien's.
109 "Beleriandic dialects" » "Beleriandic languages."
110 The note in square brackets is Tolkien's.

contraction) usually a sign of the mutation of g: as �,� i'olo<J 'the Gnome', gasdil is thus
transcribed ['] . 111
The old letter � retained, nonetheless, its alphabetic name and place. Since its original
name 'aghau' became phonetically a'au > au, it was called au-feleg 'large au' (to distinguish it
from � al}). It was thus sometimes used or abbreviated.11 2
The ON sound m became opened to an (originally nasal) spirant. Since this change
originally affected medial m in every case except in the combinations mb, mp (which were
denoted by UL , jL) the old sign was still used: thus u. normally = spirant m (mh).1 1 3 But m in
absolute initial position remained unchanged. Therefore the spirantal value, especially when
this was grammatically a mutation of m, as in malt 'gold', i· mhalt, was represented when
required by the modified sign � or '1A..J : as � � . Later [mh] ...., became
denasalized, at first becoming b, or 1} with nasalization of a preceding vowel, and ultimately
with loss of this nasality becoming identical with f [v], w. The sign w- then went out of use,
being longest preserved, even after the sound was [v], as the grammatical mutation of m. �
thus only appeared initially in archaic Gondolic.
The groups mb, nd, gg written t:J4. . ""' 1 � have now become [m]; [nn] medially and [n]
finally; [gg] medially and [g] finally. 11 4 The spelling has largely remained unchanged. The
process was as follows:
Final [mb] became [m] earliest in Gondolic, probably before Y.S. 300. Hence 'W. and 12.
interchange finally, w:a. transcribed m predominating: thus Gond. lham 'tongue', Nargothrondic
and Feanorian dialects lhamb until later.
Final [ng] similarly, but perhaps later, also became [g] before the end of the Gondolic
period, but since the alphabet possessed no sign for isolated [g], the spelling with t:er ,
transcribed ng, remained unchanged.
Medial and final [nd] remained unchanged until late Exilic. Hence such spellings in the
histories as Grond, Angband, Beleriand, Gondolin, &c.
In later Gondolic (after Y.S. 300) medial mb > mm. The letter 'n1.. thus became used = [mm]
of any origin (PQ pm, bm, sm, mm and mb ); and is usually transcribed mm. Thus [bar] fU�
'home' transcribed bar, pl. i.•m.J\1.]' [imme1r later ime1r] transcribed i· mmeir. Before or
after consonants in compounds, &c., [mm] > [m], usually recognized in spelling: so archaic
amb 'up' + rhiin 'rising' > [ambriin > amriin] : written c.�&,;,, transcribed amrun.
9I In post-Exilic, Toleressean, [mm] generally > [m] but this is not recognized usually in
spelling.
In late Exilic (after Y.S. 307) [nd] > [nn]. The letter ,.,., thus became used for [nn] of any
origin (PQ tn, dn, sn, nn, nd); and is usually transcribed nn, except in ancient historical
Gondolic names: as Gondolin. This [nn] became [n] finally and before consonants, and this is
m Note that Tolkien erroneously used o instead of c.� in the tengwar spelling of i'ololJ.
11 2 This last sentence was later replaced with: "It is occasionally found in late Exilic used for final [aw] in mono
syllables instead of en- : as l"' ccf for ,.,er daw."
11 3 "medial m in every case" « "m in every case," probably before the next sentence was written.
11 4 This sentence was revised in the course of writing from: "The groups mb, nd, gg written t:J4. • ""' 1 �
became during the fourth century Y.S. mm, nn, gg. This change earliest affected mb, which in Gondolic became
mm, and in unstressed syllables m, before Y.S. 300, though it at first remained medially. During Gondolic after
300 [nd] remained except finally in unstressed syllables; [Dg] became [DD] and in unstressed syllables [D]."

recognized in spelling, except that "" = nn is retained in stressed monosyllables as a sign of

the short quantity of the preceding vowel. Thus � =gonn, Gondolic gond 'rock' .
<f Note ndr > nr not dr (as nr, nnr, see Consonants): So Belerian(d)ren 'Beleriandic ' but
GondolilJren 'Gondolic ' . In post-Exilic [nn] > [n] before a main stress and so also in
initial mutation (of d); but this is not recognized usually in spelling: the mutation of 1.,.,
[d] is always written '" [nn or nd archaically]. 11 5 Thus p•�l d6r 'land ' , pl. e.•wo.Jy
i· nn,Yr [inyr] 'the lands '.116
In late Exilic [:gg] suffered no further change.117 ec.r thus represents [:gg] except finally,
where it is [:g]. [:gg] has since remained before r, l, but became [:g] before other consonants, as
Angband = [a:gban] . Before a main stress [:gg] is sometimes in Toleressean reduced to [:g], but
this is not normal.
Note that mb, nd, remaining, because analysed or separated syllabically, or redeveloped
as in compounds of various date - were, after the development of ,.,.., , us. as signs for [nn,
mm], written i;;, p:a.. . Similarly [:gg] in which [:g, n] belongs to one word, or component, is
usually written Uf .
p, p, cJ , sp, st, se are in later orthography often replaced by full writing � , 4-,
The letters
'r · For f.4r there also appear 'F· kr =sb, sg. The phonetic intention is voiceless s +
voiceless media b, g [ p, c] - since in these combinations p, c have not their slight[ly] aspirated
normal pronunciation. In st, t is normal and hence 4' st is written.
The notation of Vowels.

In Exilic the Feanorian or ON notation was in the main retained: words being re-spelt as
successive sound-changes occurred.
Thus o = [u] . c.r = [ o] . c or c; before Cj &c. = [a] .
A = [e].* t (or j see above) =[i].
* When written large, this was written � to distinguish it from A. = hau [h]. This
shape was later often employed 'X as a small letter.
The old 'long' diphthongs and the 'long' diphthongs produced (a) by vowels + vocalized
consonants; (b) by long vowels + epenthetic j; (c) by the diphthongization of ON o > ou > au:
were usually denoted by two vowel signs.11 8 Thus archaic eo = [au], uo = [ou], en or tt1 =
[oi]; .Al or A} = [ei]; CA- = [a�] . CO\. = [oe]; "' , '!) =[ui]. The modifier !.! peg dadol 'double dot' ,
in ON called ambayo 'raised y ' , was used thus ..A = e.t produced by epenthesis of l to e: as in
bar 'home ' , pl. beJr (later by lengthening before sonorous consonants bar, heir) written
'F'1 · �1 ·
11 5 The note in square brackets is Tolkien's.

11 6 Note that Tolkien erroneously used c instead of a in the tengwar spelling of d6r.
11 7 Tolkien revised this sentence from: "In late Exilic [Dg] > [gg]" in the course of writing. The hooked descender
added to the D opens to the right, so the resulting letter looks more like Q than g.
118 "old 'long' diphthongs" « "old diphthongs," probably in the course of writing.

The new sounds [re ] and [y] produced by i-affection were denoted in various ways: thus [re]
by CA- and by �. [y] by 2,; and 9· [re!] from epenthesis was denoted if-. In those hands
which used � for [re], er occasionally denotes [re!], while 0 = [y] .
Length of vowel was denoted sometimes by double writing as 1"" � = dor 'land', more
often by the andeith or long sign: as T"c.7d6r. But ' andeith' only stood on stressed long
=
vowels (Accent II) . In the archaic period of Gondolic the Feanorian method of marking only
short vowels, with ' thinnas' -; was kept up, but this was soon abandoned.
Later [re ], [re!] became [e], [e1] while the quantity of vowels was largely regulated by stress
and the following consonant.119 The andeith became thus less used, and principally when
words of otherwise identical spelling were thus distinguished: as n\1-.J = nith (ON nitse) but
...,1, = nith [nip] from ON nikthe . The sign 9 for [y] endured, but � (or in some hands .!. )
was restored to the denotation of 1 in diphthongs: e i , ui were usually written :A., (; . In
imitation of this the diphthongs ae, oe were usually written €., t;,.. with .A- = [e] above. The
diacritic � in the archaic period little used except over consonants as in � = [nw] was re
employed, and [at}] was commonly written � .1 2 0
But the later diphthongs due to loss of final vowels or other causes ew, iw were written
XIE, tcr . Similarly final at} - from aw + lost vowel, or from at} + lost m, b, w, or from
stressed final at} in monosyllables - was always written e.&r : transcribed aw. 1 2 1 Hence final
au, after reducting of unstressed it} > o, always appears as aw, although where later
unstressed (in Acc[ent] III) it is in fact only a normal short diphthong at}: thus ON Araume (PQ
*Or9me, Q Orome) > *Ar61}mh > *Arauo > Araw > Araw '1c..cr'
= [ ar al} ] .
The mutation of a was distinguished from e only in the earliest Gondolic documents, where
the spelling c. or e transcribed [re] appears.1 22
11 9 The two breves in this sentence were written over macrons.

1 20 The words "was reintroduced" following "little used" were deleted in the course of writing.
121 The words "was always written" following "lost m, b, w" were deleted in the course of writing.
1 22 "mutation" >> "i-affection."

Table of Vowel Notations in Exilic Noldorin

Transcriptions are given below the N. letters.
Phon[etic] Value. Origin. Notation Archaic. Later Gondolic. Late Exilic.

[a] ON i � or e � e
i a a a
[a] lengthening of prec. e or l l {,
a
,
a a
, a
[ie] affection of i
- .. > [e] , q.v.
c... , c.
ie
[ e] ON e, i -'\.. or A.
1
A
e e e
/
[e] length. of prec. or � 1..
e, e,
A
e
[o J ON ii, o Cjr or a- er
0 0 0 0
[o] length. of prec. u- or c!f 6 &

0 6 6 6
[&] * affection of o
.. .. [e] , q.v.
et', CA", � � >
[ii] ON u, o , or o
O 0 0
ii u u u
[ii] ON ii, occasionally or 6 6 6
length. of ii
o
u
,
u u
,
[y] affection of u, o
••
o, ,, , 9
•
9 9
ii y y y
[Y]** length. of prec. did not occur 9, '
y y
* This was never long. In final stressed syllables - where short vowels were lengthened
- it did not occur, since here affection of o was > y. It occurred before ON ya as orya;
but here rei developed: retr > etr.
** only occurred after archaic period from lengthening of stressed y before voiced final
consonants as in dyr < d}lr 'lands'. 1 2 3
1 2 3 This table originally began thus:

Table of Vowel notations. Transcriptions below.
[a] from ON a. Gondolic
Tolkien then struck this out with a wavy line in blue colored pencil and began a second version of the table on
another sheet:

The diphthongs.
Old 'short' epenthetic diphthongs.
[e! ] e or affected a + epenthetic 1 . X later became either [e] or becoming long was
identified with [ei].
[ re! ] affected o, ii in unstressed syllables + epenthetic 1 . £} rei later > e1, and so to [e] or [ei] .
The 'long' diphthongs - ON diphthongs, diphthongization of ON o, or new diphthongs

from short vowel + vowel (in contractions or in contact with vocalized 3, x), or from
long vowels + epenthetic i .
[6U] ON au, o or o + 3, X · archaic c:ro later > [au] , q.v.
[oi] ON ai. archaic c0 later > [oe], q.v.
[all.] older au as above: Gondolic eo or C' later usually e .
[ii] ON ai after w, or a + i in contraction. archaic 9 later > [ae], q.v.
[ae] older ai, as above; or a + 3, x. � later G t - usually t (beside �)
[c>e] older oi, as above; or o + e in contraction. � later G tz - usually � (beside Ct1\ )
[ei] e + 3, x or affected a + 3, x; l
the same + i in contraction. � ...U , ":) late Gond. X - usually X (beside Ai )
Later length[ ening] of ei j
Tabulated list of vocalic notations.

archaic Gondolic. later. later Exilic.
[a] from ON a " [a] or e [a] C [a] C. [a]
[a] length. of " e [a] or t [a] cl [aJ l [aJ
• ••
[ie] affection of a c or e [al] > [e), q.v.
[eJ ON e � [e] or � [e]
,
� [e] � e also shape X £
/
[e) length. of prec. � [e] or )\.. [e] ;( [eJ :A. e / ,
�£
[oJ ON u, o er [o] - c.r [o] er [o] er [o]
[o] length. of prec. c...r- [o] - ec-
./
[6] & [6] � [6]
[re] affection of o, u u� er I V - tC" [re] > [e], q.v.
[re] length. of prec.* (did not occur) ff > [e], q.v.
* re, y were orig. short, since vowels were not lengthened before consonant + l (last stage of vanishing 1
which caused affection), while if the conson[ant] vanished a diphthong rei, ui was formed. But later in
stressed ultimate syll[able]s vowels were lengthened before single voiced consonants, hence y. [re] did not
occur, since the unrounding of re > e began at about the time of the later lengthening. Also re could only
occur in final stressed syllables before ON ya which ...
The note was left unfinished and the entry for [re ] struck through with two lines in ink. Then Tolkien struck out
both the table and the note with wavy lines in blue colored pencil and wrote the final version of the table on the
verso of the sheet on which he had begun the first version.

[ui] ON ui; affect[ed] o, u + 3, X l

or + i in contraction; ON � ai, oj late Gond. o - usually :; (beside oj)
ii + epenthetic l ; also ON J
au, o + epenthetic j or + i
in contraction. In Nargo
throndic dialect affection
of oe.
[rei] Only from affected o + i in tij CQ rare later > [ei] , q.v.
contraction as *nusina > noen
pl. nrein (*nre(h)ini)
The overlong diphthongs, especially from diphthongs standing or coming to stand

finally in a stressed syllable.
[it} ] from aw + lost vowel; final at} in monosyll[able]s, final stressed au (ou) + mh, v &c.
This is written c.cr and length t.- rarely marked. Only in most archaic period is v, mh
preserved as cyc.o.._, (araumh) =later c.1ur Araw.
In case of other diphthongs as ae, oe, ui, ei in stress[ed] monosyll[able]s, length is

occasionally denoted by full writing: as "'"' dae (=da�).124
1 24 At the end of this sentence Tolkien started to write "or pto" but then struck it through.

� uc'� m,�,� Cf,�t'f'<' '1 ur �� 125

Gondolin nirnaeth arnediad Araw Meglivorn
'f'Dt'C.W.CI)'lot 1 2 6 ��J\,.., ��.-;,.M� 1-J �
ro-'Turin
r'� turammarth Morwen Eledwen Thingol
UM��cl. 1 2 7 �c.. ., � 1� "1 �"' fJ J c.c.rtref!'�
Menegroth Angband Morgoth orch i· Ngrelreid
J'"'·:r:;Jt.l,, �1\..n c.6 ,��oJ�� e�n., c.,OQ 'F�t:?-1 i�
i· Ngele<J Pennas na Ngelea andrand Beleriand
l"Cf "'F,� �� i'4"cc.t( 1 28 c,u.l /,.,� 1 2 9 '7cl c.cc.�
Nargothrond Phelagund Inglor Amriin Orchalammoth
r�� c.a'4�'?'�( Xc.>'A-,"�r; �"''""AL7

Talagant Cristorondor Earendel aearendel
&';ten� )O:t� Wl�l,-'"8 130 b��"fot::lu.,., 13 1 c:.rn
Sirion Dagor Vregedur Phingolphin aran [Chithlum
cl,"t;.u� ��,-cr,J:t&C-n' e-yii),.,�bi-=' 1 32 c::.r"r"""'r? '"='
Chithlum Ered-wethion rerreid-wethion Celegorn
ro� ,��
Taur na Phuin.133
1 2 5 The tengwar actually spell Meglivurn instead of Meglivorn.

1 26 The third letter was originally written as " but changed to 'E·
1 27 The tengwar actually spell Menegrath instead of Menegroth.
1 28 The tenwgar actually spell Inglur instead of Inglor.
1 2 9 Tolkien struck out the andeith in this name but did not delete the macron in the transliteration.
1 30 The tengwar version of Dagor actually reads Dagur. The final letter of Vregedur was originally written as "
but changed to 'E ·
1 3 1 The tengwar actually spell Fingulfm instead offingolfm.
1 32 A now illegible tehta was deleted above the initial letter of this name. It is difficult to tell whether the
penultimate tengwa in this name is a or an erroneous o .
1 33 This group of sample words was written below the deleted: "Tabulated list of vocalic notations" (second
version) and above the associated deleted footnote, set off from both by horizontal lines in blue colored pencil.
Tolkien later penciled asterisks before Phingolphin and Phuin, referring to a penciled footnote: "*anglicized
Fingolfin, Fuin."

[Version A' (revision of Version A, Part 2)]

Revision
The Parmaqestarin Mode
The Noldor later developed two uses of the Feanorian letters, distinct from the Lindarin Use, and
proceeding independently from the original form of the alphabet, the values of which were
more closely adhered to.
The first of these was a special adaptation for the writing of their own tongue in its ancient
form: Kor-noldorin. That is described below. But the most important was the second, a simpli
fied form of the preceding, fitted to the somewhat simpler phonetic character of Quenya.134
This became the standard method of writing Quenya and the Lindarin use became obsolete,
though lingering on among the Lindar for the writing of their own colloquial dialect.
The state of the language for which the Parmaqestarin use was devised was the classical one,
after the special Quenya development of s > z > r had occurred, but before various later
changes. * 1 35 In general, in spite of certain later variations of pronunciation due to the alter
ation of sounds according to the habits of the native languages of users of PQ - the so-called
Tarquesta changes - the spelling of PQ remained fixed by the Parmaquestarin mode. 1 36
* Most notably th ()?) is distinguished from s; also ii [g] still occurs as an initial sound; r
from z, s still distinguished from [r]; voiceless initial r, I still recognized.137
Table of Parmaquestarin Letters and Values.

The transcription is not strictly phonetic, but is given according to the Eressean method: thus
'Cf as q [phonetically kw], since in the Eressean method q was used owing to its accidental
similarity to the Feanorian letter. 1 3 8
A. Parmatema : p-series.
1
p. 2 h.,. 3
� 4 U2, 5 au,· 6
p2-
p ph, f pht, pt m mp mb
7 11- 8
J3
V ps
B. Tinkotema : t-series.
1
p 2 b 3
p 4 � 5 ln 6
F
t th [)?] st n nt nd
1 34 "phonetic character" « "phonetic form" in the course of writing.

1 35 This sentence was emended in ink to: "The state of the language for which the Parmaqestarin use was devised
was the classical one, of a somewhat archaic character, before various later changes."
1 36 "PQ" here stands for Parmaquesta.
1 37 "r from z, s still distinguished from [r]" was emended in pencil to " z is still distinguished from s. "
1 38 The note in square brackets is Tolkien's.
PARMA ELDALAMBERON . No. 22 - Page 42

t .2
7 � 8
� 9
rh, hr1 40
10 c; 11
lh, hl 1 4 1
12
JP
rd1 42
rn 9 ts l
13
�4
14
�f144J 15 �14 ?
ld1 3 *z , 8 5
c. Kalmathna : k-series.
1
Cf
2 el 3
cJ 4 � 5 cd 6 c�
k h 1 46 he 47 ii iik fig
7 * c:.r 1 48 8
Ef 9 1
(3) ' ks h (see note)
D. Qessetema : q-series. 1 49
1
� 2 d 3 � 4 -ed 5
�
q kw]
[ hw iiw iiq [iikw] iigw
V
6
w
7
�
sq [ skw]
1 39 r » "f (from d)" in pencil.

1 40 "rh, hr" » r. Since Tolkien made this emendation by striking through each h in pencil, the resulting value
actually reads "r, r."
14 1 This entry (numeral, tengwa, and values) was deleted in pencil.
1 4 2 rd » "hr, rh" in pencil. Tolkien also rewrote and underlined rd below this in pencil, followed by an asterisk,
also in pencil. There is also an uncertain character in the margin to the right of the tengwa, written and deleted
in pencil. The numeral 12 was also struck through in pencil, due to the deletion of entry 11, but this entry and
those following were not renumbered.
143 ld » lh in pencil.
1 44 Tolkien drew arrows in ink to show that entry 14 should follow entry 15. The asterisk was written in red ink, as
was an addition to the right of "*z, f," which appears to have been "= ss." This was struck out and replaced by a
new note in red ink: "classical = ss." This was in turn struck out, with the exception of the first s of ss, which was
allowed to stand as the value for G ' in entry 15. Tolkien later struck out the asterisk and f in pencil, so that the
final value given in this entry is simply z.
145 Tolkien originally wrote no phonetic value for this entry, the notes in red ink for entry 14 having intruded into
that space. See the preceding note.
146 h » "kh [h]" in pencil.
1 47 ht » "kht [ht]" in pencil.
14 8 The asterisk was written in red ink.
1 49 The numbering in this series has been editorially corrected. Tolkien originally numbered only the first entry,
but later numbered the rest in pencil. In doing so, however, he skipped 'b:J iigw, so only six entries were
numbered.

E. Napannar: additional letters.1 50

1 1. 2 A.- 3 1 c 4
JJ
5 0
hy y short carrier long carrier w

or a or a
The forms J 9 were usually employed for { ( when vowel or other diacritics
were superimposed: thus � sa, §' so.1 5 1
= =
The alphabet has above been given in its most archaic and (somewhat pre-classical) form.1 52
The following modifications (mainly the result of sound-changes) occurred in the classical
form. [Other later or "incorrect" forms, due to the recognition of Tarquesta pronunciations,
will be noted below.]1 53
In all correct "classical" spelling the distinction between \a th (}>) and � is maintained.
But the distinction between 6 z and w::» disappeared before classical period; hence �
replaced �. The latter was then used ss.1 54 =
Initial cl [h] [x] became already in classical period breath h. It is thus permissible to use 1
=
= cl initially (only). But the better use is cl throughout.

ea- was devised for [3]: vanishing remains of CE g, which in early classical still remained as
['], clear beginning initially.155
In best classical use ea- is no longer employed.1 56 [The use of ea- as a fuller form of ' C'
(short carrier), espec. in expressing diphthongs as cr, � ; and as initial form as c.rr atta, =
&c/ = ohta, may be noted.]1 5 7

When the clear beginning requires denoting as in e· anta 'he gives', it is marked
by raised dot or stroke $• eh, e· anta. =
1 50 Above this line, at the top of the page, Tolkien wrote a barely legible note: "revised[;] a letter for f < d is
required."
1 5 1 This note was written in red ink.
1 52 The word "and" was struck through.
1 53 The brackets are Tolkien's.
1 54 This paragraph was revised in ink to read: "In all correct 'classical' spelling the distinction between la th (�).
�. ' z is maintained. But the distinction between them disappeared after the classical period in TQ
pronunciation; hence ' was then used ss. " Additionally, "then" was emended to "later" in pencil. A nearly
=
illegible sentence was added in the left margin in faint pencil: "Strictly � /s fell together but z remained for those
who could make [the] sound but since z only occurred a[t] end of words and before vowels no special signs were
used and in fact ... word ... " Christopher Tolkien assisted greatly in deciphering this sentence, adding the
comment: "But this is the sort of thing that my father himself couldn't interpret soon after he had written it"
(personal communication, 8 July 2013).
1 55 "in early classical" » "in the early classical period" in ink.
1 56 "But this early vanished and" was inserted before this sentence in ink, joining this paragraph to the one
preceding.
1 57 The brackets are Tolkien's. He started to write "may be noted" immediately after the semicolon, but then
struck it out in order to add the second set of examples.

The sound zd present in earliest period became before classical period st (together with zg >
sk/ks; zb > sp/ps; bb, dd, gg > pp, tt, kk). Hence the letters � = zd, )a = zb, ..J zg fell
out of use. 1 58
Sound changes which occurred after the 'classical' fixing of the spelling, and which
sometimes resulted in spelling modifications ("incorrect" and unclassical) are the following :-
1. Confusion of v/w. See phonology. The classical PQ retained the distribution of d.. v / -cr
w according to etymology as a rule, but the TQ confusion was often reflected either in use
of l4.. throughout for both - the cases of retained w (se. in tw, nw, iigw, kw, lw, rw) being
denoted by separate letters ("C( = kw, oar = iigw) or by the diacritic c.., f.», ;;, ' .e. �
= tw, nw, lw, rw. [On CZZ' = nw see below.] 1 59 The earliest distinction to be given up was
that between lv, rv (< lb, rb) and lw, rw, which early coalesced in lw, rw . Here (: , � for
�. � (l - ) is 'classical'. 1 60
2. The TQ colloquial pronunciation of initial ps, ts, ks as s is never 'correctly' represented as
' (though this occasionally appears).
3. In colloquial Lindarin and casual writings representing this, and in TQ pronunciation \:. 'I>'
became identical in sound with �. A Lindarin habit of confusion of h / b thus grew up,
and 6 being generalised, b was used for p = st. At same time 1 (h) was used for cl [h , x ]
and cl used for c...J = ht. P,. = pt was abandoned as this became colloquially ht. The long
legged (up and down) letters were thus got rid of.
BP The Noldor continued to keep :1>/s distinct, since they had the sound [:1>] in their own
language.
4. Initial voiceless I (lh), r (rh) became 1, r and the signs 2. , '! were abolished. 2. went out
of use altogether, but '2 became used as [r] where no vowel followed. Medial lh, rh
original = Ix, rx became U , rr but the spelling ,.-c.\ (or ·d\ �cl (or yl) was
maintained.
5. The colloquial development of ts > ss occasionally appears in use of � for N·
The initial ii, iiw became in Lindarin pronunciation (owing to there being no initial ii in
later Lindarin) represented by n-, nw. This was widely adopted even by Noldor, in whose
language fig-, iigw- occurred initially, but only in certain syntactic circumstances. The
spelling was maintained in correct classical use; but in less formal writing �C , ,..., are
confused initially, while or was often used = nw medially instead of � .
In TQ iin > nn, hence ..,., often appears = � . 1 6 1
1 58 --4 zg .)a zg in the course of writing.

«
1 59 The brackets are Tolkien's.

! 60 u� [._,. p ) " >> � ' km m .
v
•
• •
16 1 Tolkien followed this sentence with a section heading: "Other minor devices." He then struck this out with a
series of diagonal pen strokes.

The vowelling. In classical use and for the writing of formal documents the vowelling used was a
form of the ancient Feanorian: - expression by diacritics placed above the consonant preceding
the vowel-sound.
The signs used were: [a] ' , [i] i , [e] { , [o] f' , [u] ..(
or C) (never � ).
The consonant without diacritic was deemed to contain vowel [a]. [a] was expressed by
adding J or J . So y.> = ta, � = ta. Other vowels were written above the consonant if short:
p = te; or above J if long: y;,{ = te. � NB. not J'J which = tea. Initial [a] [a] or these
sounds following no consonant (or hiatus) were expressed by ' )• initially often c- J• so .1J'
= ata, er� = atar. Similarly with other vowels: f-1' = tie.
Where no vowel followed a consonant, it was 'stopped' by the sign -:- or -:;- beneath, so
'f? = at. The chief cases where this was required were 1, r + p, t, k; ph, th, h; ty, q, m and fin,
mn, rn; since most of the consonant-combinations were expressed by separate signs, or by
diacritics (as f = ty: see below) . Also the finals: t, 1, r, s, n, nt.
Final s was commonly represented by a hook � � , so p.� .-.)2-"-'� =
parmas.l62
Final n similarly by ;::;;;- or ;::::: as p10� = parman.
Final r was commonly, and as the plural inflection normally, denoted -;jO : �� =
parmar.
-t, -nt since they became obsolete phonetically (> s, ns/s) and also functionally disused (loss
of dual) only appeared in formal writing and had no abbreviations other than n �.
� 'Incorrect' and unclassical is the use of 7 = rh, after loss of that sound, as an equivalent
of ..,.,
. ,
Stopped and final -1 was often written '"G" : � = a1ma, a!'t»CS = menel.
�
The diphthongs could be expressed with � , � .163 Thus � = tai, p� = toi, y:»tq- =
tau. But they were normally expressed by the diacritics "i' = following [y]; 1 = following
[w].164 So -p·t• = tai; �..,..., = toina; U� = 1aure. [But since vowel-pairs ending in i, u
were always diphthongs and others always not, the most usual method was to express the i, u
vocalically.165 p3 . pf, pi , P, , f� pi.� p� = tai, tau, tei, tea, teo, toi, teu.
[Hence the usual transcription ai etc.]]166
1 62 Tolkien emended this sentence in ink when he wrote the sentence on final r below. It originally read: "Final s
was commonly represented by a hook � � or � · so � � parmas." Another spelling was=
added lower down on the page, in the left margin: �.

1 63 "were usually" » "could be" in ink, probably in the course of writing the following sentence.
1 64 "normally" >> "more commonly" >> "normally" in ink.
1 65 "the most usual method was to express the i, u vocalically" » "it is not uncommon to find i, u expressed
vocalically" in ink.
1 66 Both sets of brackets are Tolkien's.

Following y, w after consonant (or vowel carrier) was expressed by � 2, so f = ty, �

= tw, � = nw.
@'" - - = thy, �
... was often written below the 'full' consonants: h F•
= sty .16 7
A long or double consonant was never written twice unless to mark division in such a
compound as at-tul- "come back, return". Doubling was marked by - : as in �'OYeJ =
yavanna.
@'" For � or � later � was commonly employed. This is a classical usage, since � = z, r
coalesced with � r before the classical period.
Some variations. In hands which did not use 9 = �•) was frequently used for J (especially
initially) as ,tt3� = irien. As noted above incorrectly er was often used = t 1
esp[ecially] initially as �"'? = atar.168 This was archaic but correct in certain words where
original g had vanished as csro = 'alda (N galadh); but in words with old vocalic beginning it
was incorrect.
Noldor who in their own language distinguished between hw [= ro] initially and medial chw
[ = x11], frequently introduced the same difference into their TQ, though already in classical
Quenya hw, hy had developed same sound medially and initially [ro, b ] . The use of cl for [ro]
initially is a Noldorism.
Other vowel systems. A form used especially in ornamental script or incised inscriptions was the
so-called cirateiikele (not "daywriting" but Writing of A, since the a-sign ·r was called cira in
the full names, see below). In this system a was not assumed but expressed by iJ and
stopping of vowelless consonants was not required. So j'A = ara, �� valin6re.169

Variant vocalizations etc.
The Noldor often employed a form of vocalization halfway between the old diacritical
method, and the so-called qanta-teiikele or "full writing" in which every vowel had a separate
letter.
By this system o and o u
-
CA' o " a ii
... . -
u = oi ,
0 = ui .
1 67 Tolkien began to write .
st . . instead of thy . Below this sentence, at the bottom of the page, there are some
doodles, including the words " � = nyarre. � � ·" Rough notes on "Changes produced in hiatus"
appear on the verso of the preceding page, struck through with a single diagonal line. Presumably this was a
rejected page from a different document, since it has nothing to do with tengwar.
1 68 "As noted above incorrectly" » "As noted above" in ink.
1 69 The tengwar for valin6re actually spell valin6e. Below this line, Tolkien deleted the following incomplete
note: "BF NB. in this system i = e, { = ... "

The Feanorian A lphabet f. R. R. Tolkien
The other vowels were expressed as described above but were not placed above the consonant,
but always on a carrier: 1 [i], j [i], t [e], J [e], ' [a], ' [i].170 The preceding consonant was
held to contain [a] unless a vowel sign immed[iately] followed: 1-JCA'" = to (not tao). Initial [a,
i] or after hiatus was expressed by 1 J.
The diphthongs were expressed:
c., ••
a.. , c..f" , �' G. ' te

ai au oi ui eu iu.
The Noldorin origin of this is seen in use of c:..r = o. This is derived from the Komoldorin use.
Thus 3ii name of � became with loss of [3] i > o with N. change of i > 9 . Hence CJr called [o]
but having no function was adopted as sign = [o] . U = u is a modification of Cl .
The qantatenkele is an elaboration from the beginning. (It is never used in strictly classical
language.) In this system y is used = r and » becomes a vowel a. The vowels are all written
out thus: the sign / being now used to denote length :-
• .
� J mJ Jml � 0
a e i 0 u
"
� {e ( & ()
i i 0 ii
••
••
rO � 1
••
CA' " <z
ai au eu oi ui iu.171
"Phonetic Names"
The letters were recited and named in the following order (that already given).
1
11' 2 b., 3 � 4 '122,...- 5 tu. 6
�7 "'11.... 8
�
pii phii aphta mii ampa amba vii apsa
fii [fiipa] older psii
9 p lO b up 12 -Jn 13 b, 14 Jn 15 � 16 �
tii thii asta nii anta anda rii atsa
[siita] older tsii
17
l' 18 c 19 2 20 � 21 'S 22 69 23
£7
(older rhii)17 2 lii lhii arda alda sii [aza > ara]
[ara] [ala] assa
1 70 " ' [a], ' [i]" was struck through with three horizontal lines in ink.
1 7 1 Below this table Tolkien wrote the following doodles: ) 9 -:\ ) DP C <r
1 72 "(older rhii)" » rhii in ink.

24 c.r 25 cl 26 4 27 ca" 2s cd. 29 CCf 3o er 31 £f 32 J

,_
kii hii ahta iiii aiika aiiga a aksa hatelko
[later aha] [hiika] [later o] older ksii [later hii]
33 � 34 '-d 35 t:a" 36 � 37
lxJ 38 � 39
�
qii hwii iiwii aiiqa aiigwa wii asqa
40 � 4 1 A, [ ' J 0 d.. ]
hyii yii [ telko anda ii hwai or
telko hwii noldorinwa. ] 173
Notes
apsa, atsa, aksa became more commonly used since though classical Q. tolerated initial
ps, ts, ks, these became in TQ (earliest ts > s) s. After change of th ()?) > s, b was called
yiira sii (old sii) so usually even when it replaced p in function.174 When rhii became
pronounced rii, 1 was called ii.rii.; and !,t alii.. )
J were distinguished as sii nuqema
(s inverted), assa nuqerna. c..t::� after change to pronunciation nii was called yiira nii
(old nii) but o::r- remained with name iiwii (pron[ounced] nwii). After coalescence of
wii/vii in sound vii, IL was called ii.vii. or va nuqerna. 175
:A, yii was reckoned the last letter of the alphabet and other signs though named were
given no places in the order.176
Other names were: -
� tikse 'dot' or as in i. = [i] i-tehta 'i-mark'
-:- below was called putta 'stop'
� above or below : ya-tehta
.!z.. " " " : wa-tehta

-
- tekko 'stroke' or as in l e-tehta
/
- above was called anda-tehta 'long mark'

� etc. were called sa-riiike 's-hook'
-.F' was called ar or ar-riiike.
1 73 Tolkien deleted the square brackets before the last four entries and numbered them 42 to 45 in pencil. In the
manuscript, there is a curly bracket to the left of the first line of the table (1-8) and another joining the next two
lines (9-23).
1 74 "so usually even when it replaced p in function" « "unless it replaced p and took its name" in the course of
writing.
1 75 Tolkien has clearly written va nuqerna instead of the expected vii nuqerna here.
1 76 "the alphabet" » "the classical alphabet" in pencil.

Where names were required for }tt. � u.J a:f "':J , not used in PQ; or for the obsolete
p.. , p, c.J , they were called177
� ampataina "extended mp " or ampamba
,
po antataina
,
antanda
.:cf aiikataina
,
ankaiiga
-J qataina
,
hwiiqa
-cJ anqataina anqangwa
p. older apta (aphta) pataina or {apa

, ,
r ,
asta tataina
,
sata178
4 ahta kataina haka.
Quantesser
The following names were also employed and called the tengwesser "letter-names," or
qantesser "full names." They were so well known and recognized that for brevity the letter
could stand for its name (sometimes marked •I • : so • p • = parma 'book').
1 p. parma book. b. formen north, or falma wave.179 3 p. uphto pit.180 4 m. Manwe

2
(name) or malta gold. 5 b ampa crook. 6 p1- umbar doom. 7 11. Vala god. 8 )i apsa
meat.181 9 � tinko metal. 10 h thiile, siile breath.182 1 1 p Este (name) or asto dust_l83
12 1':Y.) niimen west, or nengwe nose.184 13 b.., anta jaws.185 14 ando gate. 15 romen
F
r3
east, or rana moon.186 1 6 � atsa claw/87 17 p rhama. rama wing. 18 c; lambe tongue.
1 77 A superfluous P.:, was deleted before ):.s.. ,and ":1 was inserted.
1 78 The names in this line were corrected from ahta, laltaina, and halca. Tolkien appears to have written s over the
initial t of tdtaina, but sataina would be inconsistent with the other names in the series, which are formed on the
basis of phonetic names for the voiceless stops, not the voiceless fricatives.
1 79 "formen north, or" was inserted in ink, probably before the page was completed.
1 80 "uphto pit" » "opto back" in ink. The final o of uphto appears to have been changed from e.
1 8 1 "meat" » '1uice" in ink. Tolkien wrote 'VPAS" above the letter-name, also in ink.
1 82 Tolkien wrote "spirit" in pencil above "breath," but did not delete the latter.
1 83 "asto dust" >> "osto fort" in ink.
1 84 "niimen west, or" was inserted in ink, probably before the page was completed.
1 85 "antajaws" >> "anto mouth" in ink.
1 86 " ri»nen east, or" was inserted in ink, probably before the page was completed.
1 87 "claw" » "tassel, fryse, fringe" in ink. Both "tassel" and "fryse" are uncertain; the Oxford English Dictionary lists
the latter as an obsolete spelling of frieze, defined as (1) 'a kind of coarse woollen cloth, with a nap', and (2) 'the
nap or down on a plant, or a tuft of the same'. Tolkien wrote 'VTAs" above the letter-name, also in ink.

19 2 lhinda, linda Elf of Lindarin kindred. 20 2g arda region. 2 1 � 'alda tree. 22 6 silme
moonlight. 23 � esse name.188 24 er kalma light. 25 ...r harma treasure, or aha rage.189
c
26 c:J ohta war. 27 CU iioldo, noldo "Gnome," Elf of Noldorin kindred. 28 W aiika jaws,
rows of teeth. 29 � anga iron. 3 0 c.r 'anne thought, purpose.190 3 1 � aksa ravine.191
32 ) hana post [or halla tall] .192 33 .,.. qesse feather. 34 � hwesta breeze. 35 � iiwalme
torment. 36 o.c::cl ufiqe cavity. 37 -c:.::J uiigwe (spider) gloom. 38 1:1:'" wiiige foam. 39 lf
usqe dusk, twilight. 40 � hyarmen south, or hyalma conch, shell. 41 � yanta bridge.
Notes. h. formen, � ntimen, t"3 r6men, � hyarmen, through use to denote points of
compass as N W E S, usually replaced the older pictorial names: falma, nengwe, rana,
hyalma.
� In naming points of compass the order in Eldarin languages is always W S E N
�la.. since West = Valinor and the face i s turned thither, South i s the next
most propitious, E[ast] unpropitious because of hostile Men, and N[orth] evil because of
Morgoth.
Since the face is West, formen: for- "right-hand" becomes North; and hyarmen: hyar
"left-hand" becomes South - this is the opposite of human languages (as a rule). Also
the associations of right and left are reversed: it is the forya or dexter that has "sinister"
associations. (This is not contradicted by physical conditions since the Qendi are
normally ambidextrous, and have no such specialization as we have.)
It will be observed that some of the pictorial names coincide with the phonetic: such as •
ampa, � apsa, 1v anta, p atsa, tc; alda, '1)1 arda, c.c.r aiiga, ff aksa, or nearly so as � ando,
p asto. It is probably from such examples as S' alda 'tree' that the device of giving pictorial
noun-names (only one, 1 halla, is an adjective) to all letters set out. It proved, however, useful
amid the varying usages and values; and these tengwesser were and are employed as the names
of the shapes irrespective of usage. The only alterations to fit phonetics were the substitution
of apsa, atsa, aksa for obsolete psara 'peel'; tsette 'fly'; ksenna 'odour'; and aha representing
medial [x] for harma (the latter however never went wholly out of use).
The use of the divine names U1 Manwe, � Este was archaic; similarly with Ulmo as name
of sign o beside iir fire, heat.
Other names are ) telko stem, leg; J anda telko long leg, which are descriptive.
1 88 "c:ize > are or" was inserted in ink before esse, but these names were left unglossed.
1 89 "harma treasure, or" was inserted in ink, probably before the page was completed.
1 90 'anne >> 'enne in pencil.
1 91 "ravine" >> "edge" in ink.
1 92 The bracketed addition is Tolkien's.

For the series 1»., m CCf c:cr longer names are also found: ampano 'construction,
building'; antaro 'high mountain, peak'; "1 ankale 'the very bright', a name of the Sun; uiiqale
'death agony'.
Other names for letters modified and not included in alphabet are :-
j5 tyelpe silver; }?. istyar wizard, lore-master; � arya day (12 hours); m nyelle bell
(small); ];;, intya guess, notion; f" indyo grandson; "C alya rich.
j3 atwa double; � arwa possessing; � olwa branch; � anwa true.193
tw, lw, rw,
r � 1
Senna. vowels were a
assumed to follow all consonants
preconsonantal cons[onants] when not otherwise expressed were stopped - .
�cc
•
= vanwa
r 9
longs were ) J
� ·
) J ) a/e/i/o/u but 11" "
were often doubled and sometimes ''
diphthongs { -r} ai { ...'}, or tp.J � or rarely[?] { r �}
r-t '? If"" F l r•c:c or rarely rt'
other [m] [????] s o r - tea. tr } f- = koa. · �
, .. c, .. .. ...,
foll[owing] y /w were - - {[?]} r = tya . r = twa
f =tyo { � } � lf'1 4 satwi. 1r ketwo

� - 196
:= was used [????] = nn {and[?]} c=;; J = ns

1M.. = mp. r- r st. c1 = ht.
1 93 Written below this, but subsequently deleted, was: "The vowels in Quantateiikele are sometimes named :
t: anna gift; ] ara dawn; ; imbe dell; 1 ire longing; � ."
1 94 Beginning with this line, the remainder of the text was written very roughly on the final page of the
manuscript.
1 95 The sense of the words in this line seems to be "other combinations not diphthongs"; but if that is what Tolkien
wrote, the illegible words were heavily abbreviated.
1 96 This word and the untransliterated word in the previous line both read Manwe.

{rk, rt} rk {(rh)} rt rs rn rp rv rm rkw rw

Ik It Is Ip Iv Im Ikw Iw
� was used = 7. finally only.
d s r
{ � z} = r
V
z r
\ I
coalesced d r s/z r
V
r
r
V
r z =� = 11
\ I
r z final r
\ I \ I
r 11'.:)
}'

Qenya Spelling
Qenya Grammar - Spelling and Transcription
On J£lfwine's Spelling
by ]. R. R. Tolkien
Edited with an introduction by
Christopher Gilson and Arden R. Smith

Qenya Spelling ]. R. R. Tolkien
Introduction
J. R. R. Tolkien in "The Feanorian Alphabet," after outlining its General or Phonetic form,
describes four different usages among the Qendi: the Lindarin, Parmaqestarin, Noldorin and
Beleriandic uses. The second of these was "a special adaptation for the denotation of the
parmaqesta or written form of Qenya," and it "remains the standard manner" of writing the
language and the basis for its "transcription into Latin letters." Contemporary with this
description of the Feanorian Alphabet is a manuscript comprising the first two sections of a
"Qenya Grammar," one describing the Spelling and transcription of the language, and the other
the Qenya Consonants, in which the usage of the tengwar is this same 'parmaqestarin' standard.
The manuscript was composed on both sides of unlined sheets of 6 Ys by 8 Ys inch paper with
pen-nibs of various widths, the wider ones used for writing the examples of the tengwar as
well as the title "Qenya Grammar." The pages were numbered in pencil from 1 to 9, except for
the last page. The manuscript text ends at the bottom of this tenth page in mid-sentence, and
there was probably a continuation that is no longer extant. Placed before the manuscript is a
cover sheet (the back of the cover of a writing-pad of the same size as the sheets of the
manuscript), which contains the following text written in red ball-point pen:
Phonetic
Structure
of C. Quenderin
& CE.
In need of historical revision.
The first "C" has been written over an original "E" that was probably a false start on "Eldarin."
Placed right behind this cover sheet is another somewhat larger (7 % by 9 � inch) page with the
following text in ball-point:
c.
Quenderin &
Eldarin
Structure and Basic
Grammar
Partly complete
With unrevised notes
&
matter
Neither of these texts seems to refer to the "Qenya Grammar" or its excerpted beginning as
such, although either might suggest the larger context in which Tolkien intended to include it,
each at different times.
For convenience we will refer to the manuscript excerpt as "Qenya Spelling" (or QS). Its
composition was begun sometime after (though probably not long after) the composition of
"The Feanorian Alphabet." This is shown by the fact that, in that earlier description of the
Parmaqestarin usage, Tolkien tentatively assigned the value rh to the letter 1'· then decided in

Qenya Spelling J. R. R. Tolkien
the course of composition that it would be used instead for the sound r when not followed by a
vowel; while in QS the latter is already the assigned value of this letter from the start of its
composition. A fragment survives of an earlier version of this grammar called Tengwesta
Qenyava, which begins with a description of "the special application to Q. of the Valinorian
alphabet of Feanor" (see PE 21, p. ix). This has an earlier conception of the alphabet, where for
instance p:t. is used for mp, while later in QS b1. is used for mp and F for mb. Such changes
to the writing system and its more elaborate description in a separate "book concerning the
alphabets" may have been the principal reasons for Tolkien revising the beginning of the
Qenya Grammar.
Tolkien's earliest application of a writing system to the represention of Qenya employed
"RWnil's Alphabet," which he described as "a selection from the universal Eldarin alphabet
system" (PE 13, pp. 62-3). This explanation of the Elves' representation of Qenya originally
constituted the beginning of the manuscript version of the Early Qenya Grammar from the
1920s. Placed with it and numbered continuously with the rest of the manuscript grammar
was a typescript page which tabulates the Alphabet of transcription.1 That table gave the same
set of transcribed consonant sounds which are given here in the table beginning the section on
Qenya Consonants, although the arrangement is slightly different. For example, the earlier table
lists the consonants w and y in the same row as the voiced spirant v and has a separate row for
the "liquids" l and r, but the later table gives v as the only voiced spirant and lists w, � r and y
together in a row of "sonants."
The table of the Alphabet of transcription is followed by a phonetic interpretation of the
"native descriptions of the sounds as reported by J£lfwine." This is a series of descriptive notes
similar to but more exhaustive than the corresponding phonetic notes for the Early Qenya
Grammar. For example, a note on the application of Rumil's Alphabet says that t and n "are
both (as in English) somewhat retracted from the true dental position" (PE 13, p. 63). QS states
more comprehensively: "The 'dentals' t, d, n, l, r are 'point' consonants, but made as in English
just behind the teeth." Tolkien may have intended to give an explanation of all of the
consonant sounds of Qenya, but the manuscript ends in the middle of the explanation of the
sound transcribed ng and representing the back nasal [:g] . The extant manuscript of QS was
very carefully written with only occasional corrections in the original ink, probably mostly in
the course of composition. A few additions were made later in pencil, and these have been
included in the text presented here and indicated as such in the footnotes.
In Tolkien's original description of the Parmaqestarin use of the Feanorian Alphabet he said:
"On it is based the transcription into Latin letters devised by the Noldor of Eressea, which is
used in this account." In describing the Beleriandic or Exilic usage, on the other hand, he
employed "the transcription used by J£lfwine, based mainly on Old English, and partly on Irish
use," giving a detailed account of this from J£lfwine's point of view (see above, pp. 15, 33).
J£lfwine's role in this conception is expanded in QS where the transcription of Qenya used in
the grammar is also attributed to him, and said to be founded "primarily on Latin spelling
(according to the pronunciation of his day) rather than on his native Old English." Tolkien
seems to have given a good deal of thought at this time to the conception of J£lfwine's
transcription of the different Elvish languages; and he placed with the manuscript of QS six
versions of a short text, four of which bear the same title, On JElfwine's Spelling. We have
designated these "Version 1" through "Version 6" (abbreviated JES 1 to JES 6) based on the
1 Early Qenya Grammar, edited by earl F. Hostetter and Bill Welden, PE 14, pp. 41-2.

apparent order of composition. Each version takes up two sides of a single sheet of unlined
paper. JES t, JES 4, lES S and JES 6 are on sheets measuring approximately 9 � by 7 � inches,
while JES 2 and JES 3 are on sheets of about tO by 8 inches, and the edges of these larger sheets
of paper are now foxed and torn in places.
Each of these texts begins with a variation on the same sentence, the final version of which
is: "JElfwine, in his own narrative and in his translations of Eldarin legends and chronicles into
English, often spelt the names of Eldarin origin after the manner of the English of his day,
following as nearly as he could the pronunciations that he heard in Eressea." JElfwine's own
language was Old English (or Anglo-Saxon), the orthography of which differed from that of our
modern usage not only in the shapes of some letters but also in the normal repesentation of
certain sounds. The two main points that affected JElfwine's transcriptions of Elvish were that
(t) the voiceless vs. voiced spirants, such as those we represent by f and v were . not distin
guished in Old English spelling, since their difference is predictable from context, and so for
instance only the letter f was used, which is pronounced [v] when between vowels or next to
another voiced consonant, but pronounced [f] elsewhere; and (2) the letter y was not used in
Old English, the sound being represented by the letter-combination ge.
As an example of how these spelling habits influenced JElfwine's representation of names in
his translations, Tolkien cites the Noldorin name Icfrvael 'the Bloodstained' which JElfwine spelt
gearfael, and he writes this out in Anglo-Saxon orthography in the opening paragraphs of five
of the texts, JES t to JES 5. This name Iarvael was also cited in The Etymologies, under the base
YAR 'blood', where it was subsequently emended to Iarwath, and in this form was also cited
under the base WA3 'stain, soil'.2 This would suggest that "On JElfwine's Spelling" dates from
the period when Tolkien was working on The Etymologies around t937 and t938, or shortly after
this when he seems to have continued making "rather desultory entries" (cf. Etym. 344-5).
JES t also cites the names Rohan and Galadriel which emerged in the composition of The Lord of
the Rings in the early t940s.3
The two pages of JES 1 are designated by the letters "c" and "d," which suggests that it is a
surviving excerpt from a longer work, perhaps the contemporary grammar of Noldorin, since
the examples cited in it are all from that language. It consists primarily of a list of all of the
Elvish sounds for which JElfwine employed more than one spelling or used a two-letter
combination, followed by a shorter list of the spellings "he recommends" when making careful
notes on the Elvish languages� Tolkien also explains his own practice of "Anglicizing names in
narrative," such as the use of f for the sound [v] when it occurs finally as in nindalf = nindalv,
and adds a final remarkable note that the Anglicized spellings correspond to the Toleressean
pronunciation of these names.
JES 2 gives a tabulation similar to that in JES t, summarizing "JElfwine's system and his
actual practice, in the matter of the difficult letters or sounds." This table is introduced by a
detailed explanation of the features of Old English orthography and phonology behind the
usages listed in the table, to show that JElfwine gave "considerable thought to the
transcription of Eldarin letters." For instance, Tolkien explains how the Old English letter o
"was used merely as a variant of p without phonetic distinction," either letter being used for
2 The Etymologies (Etym.) in The Lost Road, ed. Christopher Tolkien, 1987, pp. 339-400. Also cf. "Addenda and
Corrigenda to the Etymologies," ed. Carl F. Hostetter and Patrick H. Wynne, VT, nos. 45 and 46.
3 cf. The Return of the Shadow, ed. Christopher Tolkien, 1988, pp. 422, 434-5; The Treason oflsengard, ed. Christopher
Tolkien, 1989, pp. 246, 261. Also see the Etym. bases ROK (whence N roch 'horse'), GALAD 'tree', and RIG (N rhf
'crown'), which underlie these names, though the names themselves are not cited there.

the voiceless or voiced dental spirant and the pronunciation determined by context, similarly
to the pronunciations of the letter f. Since the voicing of spirants was not determined
contextually in Noldorin, J£lfwine proposed the use of ph and th for the voiceless spirants and
to restrict f and p to their voiced values, with a special use of o to represent the voiced dental
spirant word-initially in grammatical variants of Noldorin words beginning with d. And in
similar fashion J£lfwine also "suggested that a letter b, 'after the manner of the Old Saxons,'
should be used initially instead of f, where this f [= v] was a grammatical variation of b." He
would have employed this name (OE Eald-Seaxan) to refer to the Saxons that still remained in
northern Germany, and in their earliest surviving writings (from around the 9th century) the
voiced labial spirant is often represented by a crossed-b. 4 The implication seems to be that
J£lfwine, in addition to being "acquainted with Latin, and to some degee with Irish," knew
about Old Saxon writing as well .
.JES 3 begins with a revision of the first two and a half paragraphs of 1£5 2, with various
improvements in wording, some made after the original layer of writing, and incorporating
examples of Old English words and their modern equivalents to exemplify the usage of the
letter f with both voiceless and voiced ponunciations. This ends near the bottom of the recto
of the manuscript, with the explanation of J£lfwine's use of f, p ; ph, th ; and b, o. The verso
includes an alphabetical list of all of the Latin and Old English letters or letter combinations
used for transcribing the sounds of either Noldorin or Quenya, followed by a rough chart of
these arranged phonetically.
Below the chart and (probably later) above the list Tolkien added three progressively more
roughly written paragraphs of further revisions to the text of 1£5 2, mainly regarding the
spelling of the sounds [s] and [z]. These were both represented by the letter s in Old English,
and one letter was adequate for transcribing Noldorin, which had only the voiceless sound. In
spoken Quenya, however, either voiceless [s] or voiced [z] could occur medially. As explained
in 1£5 2, "the voiceless [s] medially was derived from older Parmaquesta p (th)," and since the
Noldor had this sound in their own language they "tended to preserve the archaic p, s as such,
though the Lindar represented them as s, z. " J£lfwine represented the Noldorin pronunciation
of these sounds using the same spellings he used for Noldorin itself, but a difficulty arose if he
wanted to represent the Lindarin pronunciation of [s] vs. [z] as Tolkien would explain in 1£5 4:
"For this distinction J£lfwine had no available signs: z, if used at all, denoted ts in the usages
with which he was familiar."5 Consequently he devises his own spelling sh to transcribe the
voiceless medial [s] of Quenya, although he recognizes that this is "a novel and not very
acceptable combination." This sh appears to be present already (though not yet explained) in
the alphabetical listing and phonetic chart of spellings in 1£5 3 .
.JES 4 up to the discussion of s and z is for the most part a 'fair copy' of the beginning of 1£5 3,
again with some improvements in wording and arrangement. The same six examples of the
Old English use of f are cited, but they are collected together into a footnote. J£lfwine's
"translations of linguistic lore" are now described more specifically as his "translation of
Noldorin linguistic lore," an idea that perhaps was suggested by the fact mentioned in 1£5 2
"that his teachers were Noldor." In 1£5 4 Tolkien abandons the assertion made in 1£5 2 and 1£5 3
that in choosing a way to transcribe the Eldarin spirants J£lfwine wanted to avoid "a system
4 See F. Holthausen, Altsiichsiches Elementarbuch, Heidelberg, 1921, §220, pp. 75-6.

5 Cf. the discussion of the letter Z in the Oxford English Dictionary : "Z was used in O.E. in the spelling of alien words,
and (with or without t or d) in certain loan words ... with the value (ts)." Also see joseph Wright and Elizabeth
Mary Wright, Old English Grammar, London, 1925, §7, p. 9.

completely unfamiliar to his own countrymen." On the other hand, Tolkien is careful to point
out that J£lfwine's invention of sh for spelling the voiceless medial s of Quenya would have had
"no such associations as it has in modern English," and he gives an explanation of how J£lfwine
may have come up with this idea, in part by analogy with his use of ph, th, eh. Following this
Tolkien describes J£lfwine's suggested use of o and o to represent the Noldorin voiced spirants
when they are grammatical variants of the voiced stops b and d, essentially as this is presented
in 1£5 3; but Tolkien adds the observation that this distinction was unnecessary "since in N.
voiced th [o] and v only occurred initially as variations of d and b," and that J£lfwine did not
usually follow his own suggestion in practice.
There is a break in the text of lES 4 after this description, followed by a more hastily written
paragraph that begins mid-sentence, and might have been added much later. It concerns the
use of i to represent the sound of modern English y at the beginning of Noldorin words. In J£S 2
Tolkien had said that this was J£lfwine's practice, "especially in linguistic writings or notes,"
but that he often used 9 or ge to spell this sound as in Old English; but Tolkien had added that
the latter usage "is not fortunate" since in Noldorin the back stop [g] could occur before or
after any vowel. In this final paragraph of J£S 4 he notes that using i for this sound "was not
entirely satisfactory" because in Noldorin i could also be a word-initial vowel, as in dissyllabic
ior as distinct from 1 or (yor).
The first three paragraphs of JES 5 are very close to those of J£S 4, again with a few revisions.
For instance, there is an explicit statement that J£lfwine was "under the instruction of the
Noldor." In the next two paragraphs Tolkien elaborates the conception of J£lfwine's invention
of the spelling sh for the voiceless medial s of Quenya, explaining that this would not have the
associations that modern English sh has for us because that sound was represented by se or see
in Old English. But ultimately Tolkien would reject this conception in a later revision of lES S;
and instead he merely quotes J£lfwine as observing that in the tongue of the Eldar of Valinor
"in the inner part of a word� is softened as is s in our tongue." The implication seems to be
that, having no letters of his own available for the distinct pronunciations of this Elvish letter,
J£lfwine did not attempt to represent the difference in his transcriptions, and was content
with describing it.
Tolkien silently abandons the conception that J£lfwine used o and o for the Noldorin voiced
spirants when they are grammatical variants of b and d, having already made the observation
in J£S 4 that this usage was unnecessary. In the next paragraph of lES S he describes how
J£lfwine represented the sounds of Noldorin and Quenya that correspond to the consonantal w
and y of modern English. The conception is essentially the same as that described in J£S 2; but
regarding w (which is Tolkien's transcription of J£lfwine's use of Old English p), there is a new
observation that " u often appears instead, notably in cu, gu, tu for initial cw, gw, tw, " and in line
with this Tolkien notes that when transcribing Quenya J£lfwine "as a rule employed qu," and
this is associated with his view of that language being "as it were the Latin of the Eldar." This
concept of an 'Elf-latin' was already present in the description of J£lfwine's usage in the
"Qenya Grammar," although there it is said that "he notably diverged from Latin usage: he
nearly always used q rather than qu to represent [kw], probably because he was struck by the
(actually fortuitous) similarity of the Feanorian sign 'Cf for the group [kw]." This was clearly
a fictional reflection of Tolkien's own usage that goes all the way back to the Qenya Lexicon, and

a sort of rationale for it.6 So the change in conception may reflect Tolkien's own increasing
tendency around this time to use qu where before he would have used q.
The carefully written text of lES S ends with a sentence introducing "lElfwine's alphabet"
which is said to be "arranged both in conventional and in phonetic order." Tolkien must have
meant to include a list and a chart of letters and letter combinations similar to those in lES 3. A
final rough and hastily written paragraph seems to consider lElfwine's different treament of
Noldorin and Quenya, and mentions his view of the latter as "the Elf-latin." These points are
incorporated in the introductory paragraph of the next and final version of the text.
JES 6 has no explicit title and as with 1ES1 Tolkien may have intended to include it in a larger
work. In the first paragraph the cited example of lElfwine's use of Old English spellings is now
geafanna, transcribing the name Yavanna, a Quenya transcription presumably being preferable
to a Noldorin one. In lES 2 Tolkien said that using g(e) to represent the consonantal y-sound of
Noldorin "is not fortunate" since the voiced stop 9 can occur before any vowel. In lES S he said
"such spellings are inconvenient" and in lES 6 he says that "this usage was unsuitable to
Noldorin" and that "in his more careful spellings and transcriptions J£lfwine avoids it." In
J£S 6 Tolkien also adds a clarification that lElfwine's system of transciption is really "two
systems: one for the transcription of Quenya, and another for Noldorin." For Quenya he
attempted to follow "the spelling and pronunciation of Latin" but for Noldorin "the usage of
letters in the English of his own time." The former is explicitly motivated by lElfwine's
assertion that "the Quenya is indeed the Latin of the J£lfe (Elves)."
The remainder of 1ES 6 begins with a heading Noldorin, and it describes lElfwine's system for
transcribing that language in essentially the same way as in lES s, with some improvements in
wording, and without the mention of any specific points relating to Quenya. This may be
connected with the fact that mentions of lElfwine's transcription of names are given in the
Outline of Phonetic Development and the Outline of Phonology, detailed descriptions of the sound
system of Quenya dating from the 1940s and early 19SOs. In both of these Tolkien states that
"k, kw, ks are used for lElfwine's c, qu (q, cw), x, " and in the latter also mentions "his ambiguous i,
or g, ge," and his name relflreden 'elf-latin' for Quenya/
The text of lES 6 is very carefully written with a nib-pen and has only two corrections made
in the course of composition. Later Tolkien revised the first eight out of the ten mentions of
the "Noldorin" language to "Sindarin" using a fountain-pen. This reflects the major change in
Tolkien's conception of the Eldarin languages, wherein the language inherited by the Noldor
was reimagined as the language they adopted from the Sindar in Beleriand, a change which
emerged around 19S1 during the composition of the Grey Annals.8 This shows that 1ES 6 must
have been composed before this date, and also that the conception of lElfwine as narrator and
translator of Eldarin legends and chronicles, and as transcriber of Eldarin languages "under
the instruction of the Noldor" in Eressea, survived Tolkien's renaming and reimagining of the
genetic relationships of their languages.
6 cf. "The Sounds of Qenya," in Qenyaqetsa, the Qenya Phonology and Lexicon, ed. Christopher Gilson, Carl F. Hostetter,
Patrick H. Wynne and Arden R. Smith, PE 12, p. 21.
7 Quenya Phonology, ed. Christopher Gilson, PE 19, pp. 30, 68.
8 Cf. The War ofthe jewels, ed. Christopher Tolkien, 1994, pp. 19-21, 25-8.

. Qenya Grammar .
Spelling and transcription.

Qenya, when used as the parmaqesta or common written language of all the Qendi, was usually
written in that modification of the Feanorian alphabet which is called the 'parmaqestarin'.
This is described in the book concerning the alphabets. Here only the most usual form, that
used in this book, is given.
The letters used were 39 in number, recited in the following order. Their names'll are given
in transcription, for which see below. The value
of the letter is the first letter of the name:
thus 1:2. va = [v]; except where the letter expressed a consonantal combination which could
not occur initially: in that case the name begins with a followed by the value, as ampa = [mp].
On the value of 1 'ar', see vowels.
'li These are the 'plain names' used in grammatical and similar works. Each letter also had
p
a significant name, for these see Alphabets. Thus was called parma 'book'; tinko t
'metal'; k kalma 'light'; q qesse 'feather' - hence the names parmatema, p-series (=
labials), etc.9
Alphabet. A. Parmah�ma, p-series: 1 fL 2 'b. 3 11 4 m. 5 m.. 6 F 7 a :

pa fa va ma ampa amba apsa
B. Tinkotema, t-series: 8 p 9 b 10 6 11 Q 1 2 � 1 3 � 14 F 15 y
ta asta sa ra na anta anda ar
16 Jf 17 t; 18 � 19 � 20 Jl � :
arda la alda assa atsa ans or ansa 10
C. Kalmatema, k-series: 21 CJ 22 d 23 J 2 4 «<' 2 5 eel 2 6 CCJ 27 £f :
ka ahta ha nga (na-yara)11 afika afiga aksa
D. Qessetema, q-series: 28
� 29 d 30 "'0" 3 1 u:a 32 a:J 33 a::r :
qa hwa wa ngwa (fiwa) afiqa afigwa
E. Napannar, additional letters. 34 "' 35 � 3 6 J 37 J 38 u 39 0 :·
hya ya 'telko'* 'andatelko'* 6 u.
* These are names of the shapes meaning 'stem' and 'long stem' not descriptive of the
values: for which see the Vowels.
9 This note is a footnote in the manuscript; and similarly for the subsequent three notes below marked by 'll .
10 This item was added later in pencil, written hastily in the space below the heading for the next series.
11 The alternative name nci-ycira was written below ngci in the original ink; later fief was added in pencil above ngci,
but the original form was not deleted. Probably at the same time the tilde was added in pencil in each of the
forms afika, afiga, fiwci, afiqa and afigwa.

Vowels. It will be noted that there are only 2 vocalic letters: er [o] and 0 [u]. These in spite
of their names which are long (because Qenya does not have stressed short vowels finally)
express short o and it The other vowels are expressed by diacritics placed above 1 when
short, and above J when long: without diacritic these signs express [a, a]; marked j ) they
express [i, I]; and marked ( { they express [e, e]. The full vowel list is thus
1 J ( " 0 J J f Cl' o .
a e o u a f e 6 u.
In value [a] 1 is frequently written e' initially, and in the value [a] ) is frequently written )
initially.
In 'parmaqestarin' use, however, every consonant is assumed to be followed by [a], unless a
vocalic sign immediately follows. Thus pt; is read tala; but pft;i is read tili. The expression
of [a] is thus only required initially and before or after another vowel. Thus tf = ata, ph = tea,
.
pn = taz.
.
Where a consonant stood finally or before another consonant without a following vowel it
was marked by a point beneath as 'l22.f! = mat. But most of the consonantal combinations of
Qenya had separate letters: as p1 = mb. The chief combinations that had no separate letter
were: l, r + t, p, k, q, m; and m. In these cases the letter "f, which specially represents r with no
following vowel (and hence is named ar), is used: Tf' = rta; while l is usually expressed by a
reduced and dotted form "'t!' (which is not however regarded as a separate letter with an
alphabetic place): as nDD- (or � ) = alma. Qenya only tolerated finally n, r, s, t, l and nt, ns:
these are expressed by t:p 11, � , f? , C: (or G ), �, � . 1 2
Long or double consonants are marked by superimposed - : thus ij5 = atta. The same sign
is used (see above) to mark the length of El 1 c; = o, u.'ll
'li In the older system all vowels were expressed by the diacritics on the telko: thus r = o,
r
"1 = u; = o, ; = 11. 13
Following [y] was expressed by ,. superimposed in all cases except those letters which had a
high stem: thus p = tya, !! = stya, l5 = nya.
Following [w] was expressed by ..., where no separate letter existed: thus �
= lwa, ,":g, =
nwa.
The Qenya diphthongs: ai, au; ui, oi, were usually expressed by means of these diacritics. 1 4
Thus Ji', jif,
po , pu in the 'parmaqestarin' use = tai, tau, tui, toi. Since i, e were already
12 In the list of consonants tolerated finally, ns was added later in pencil. The combination nt was not originally
given tengwar expression, but later � was added in pencil; next to this Tolkien first gave � for ns, then
replaced this with � , both in pencil. (See footnote 10.)
1 3 The manuscript has o and u as the transcription for both pairs of vowels, but the long carriers in second pair
}) �
show that the lengthened vowels are intended. For the tengwar versions of u and il Tolkien added L and J
below in pencil, although he did not delete the earlier versions.

expressed by diacritics (l , ( ) the diphthongs iu, eu were usual[ly] written in full \o = [iu]; (o
= [eu]. The same full writing could be applied to the other diphthongs: as &it = [oi]. For ; = ai
as a diphthong the more distinctive form 'f is often employed.
In pure Qenya words sp, sk do not occur, and sq (skw) is rare. These can be expressed by � or
by a small hook above where they occur in borrowed words: thus �p. = spa or p.15 For 'f
(sqa) the form � is sometimes used, although this should express qs - cf. Ef• l\• 'Jt = ksa,
psa, tsa which really contain the s-hook though counted as separate letters. Thus usqe is
written oy..q.t or or � "2f". A similar hook is also sometimes used for final 'stopped' s,
especially as an inflexional ending: as � for -pfm.4 = parmas.
For final 'stopped' n in inflexions a stroke _ or curve ,_ is often placed under the
preceding letter: thus f'l't!!:. for -pyma� = parman.
These diacritics though not regarded as alphabetic have names. These are: a diacritic is
called tehta'mark'. The stopping point as in (l = t is called putta
'stop'. The mark - as in
iW = nna, C4 = oo (o) is called andatehta
'long mark'. The single superior dot (in \ = [i]) is
calledi-tehta 'i-mark'; � as in ( is called e-tehta
'e-mark'; •• is called ya-tehta
'y-mark'; ..,
wa-tehta 'w-mark'; the s-hook was called sa-rinke
's-hook' when below; but when above in
shape /"' is called lope
'plume' or sa-lope.
•' is called ai-tehta
'ai-mark'. The understroke -
expressing final stopped n is called or an, an-tehta.
Note: A completely different system of vowelling is sometimes used especially in
ephemeral writing. This is called qanta-tenkele
or "full-writing." The assumption of [a] is
abandoned, and every vowel has an independent sign. Since 2 forms of [r] are thus otiose 'r
alone is used for [r] and the sign tCJ (resembling c:.t already used for [o]) is employed for [a].
Length is expressed by a diacritic I or - .
This vowel system in full is :-
» 1. J (or ' ) er 0
a i e 0 u
q ' fe, 4 0
a
, l
, , ,
0 u
••
B it 0 a 1 l
ai oi ui au eu iu.
This system is not used in this grammar.
r
In older classical Q. the sounds voiceless (rhci) and I (lhci) had existed initially, where they
were principally derived from original sr, sl and less frequently from khr, thr, etc. (see below).
For these sounds the signs i'f , 'l; were originally devised. But rh, lh became in Lindarin and
1 4 Tolkien originally included the diphthongs iu and eu in this list, but deleted them heavily in ink, probably in the
course of composition.
1 5 Tolkien wrotef here, but that would presumably be read as sta.

in the common tarqesta pronunciation r, l, while single signs were needed for rd, ld which
could not be spelt out, since no separate sign for d existed. In consequence i! , "£' became used
for rd, Id and the voiceless r, l (if represented at all) where written ,� , h : or � '? � � r- . 16
The transcription used in the above description is based on that of lElfwine. lElfwine re
garded Qenya as the equivalent of Latin in the Quendian world: the spoken and written
language of learning and ceremony common to peoples of diverse Vernaculars. He often calls
it celf-lceden 'Elf-latin' in his writings. He consequently founded his transcription primarily on
Latin spelling (according to the pronunciation of his day) rather than on his native Old English.
In one point he notably diverged from Latin usage: he nearly always used q rather than qu
to represent [kw] , probably because he was struck by the (actually fortuitous) similarity of the
Feanorian sign '9' for the group [kw] . This peculiarity is here retained, because it is
convenient, although it is unsystematic - for the other letters of the qessetema have to be
expressed analytically, and spelt out: as ngw.
lElfwine also uses k far more frequently than it is used in Latin or in Anglo-Saxon -
especially before e, i, and so doubtless to avoid the pronunciation of [c] as [ts] or [tS].
In the present transcription the chief divergences from lElfwine consist of regularization of
his not wholly uniform methods; and of the following special points.
(1) The substitution of y, w for i, u in consonantal function. Note: for i = y lElfwine
occasionally uses g ( "5') before e (especially initially), after the contemporary pronunciation of
9 in English. y is here preferred to j, since j is useful for the representation of the sound [ dz], as
in modern English joy, which occurs in certain Eldarin languages.
(2) the uniform use of k instead of c; and of ks instead of J£lfwine's variable x, ks, cs. This
also helps to characterize Qenya words and spellings and to differentiate them from Noldorin,
Telerin, etc. where lElfwine modelled his transcription rather on the vernacular uses (English
and Irish) that were known to him.'ll
91 In spelling Qenya names in modern English narrative translations, however, it might be
advisable, as giving these a more familiar appearance to use c for k, qu for q, and x for
ks; and n for initial ng.
(3) the uniform use of ht
[xt] to transcribe Feanorian c:l - J£lfwine uses et, cht, beside ht
derived from Anglo-Saxon orthography.
(4) The differentiation of 1:1 [v] and a- [w] gave lElfwine some trouble. He often uses u
initially to transcribe both. Not infrequently he uses the Old English letter p [= w] for 1%' [w],
initially. Medially he also uses u in both values and here seldom employs p; but frequently
writes b to express [v], after the contemporary pronunciation of Latin: a use that was un
ambiguous as Qenya had no stop b between vowels. In this grammar v is uniformly used for
'G. , and w for a:- .
(5) For J£lfwine's use of initial n in words containing Qenya initial &c.r = [:u] and ea: [=
flw], the older and more correct sound is represented in this grammar by ng, ngw. It must be
noted, however, in initial position that this ng = [l]] and not, as always medially, [l]g]. In the
histories and narratives n is used: as in Noldo. See further below p . ...
1 6 This paragraph was added later in pencil at the bottom of the page.

Qenya Spelling f. R. R. Tolkien
Qenya Consonants.
The Quendian grammarians, as reported by JElfwine, analyzed the elementary consonantal
sounds of Qenya and arranged them in approximately phonetic order.'l
9I For such purposes they used the alphabet of Rumil, or the general 'phonetic' form of
the alphabet of Feanor. In the table here given the normal transcription adopted for
Qenya is used in preference to special or more exact phonetic signs.
A similar table, using the transcription as above, may be given thus.
(a) Labial (b) dental (c) front (d) back (e) back-labial
p-series t-series ty-series k-series q-series
1. Voiceless Stops p. t. ty. k. q.
2. Voiced Stops [b] [d] [dy] [g] [gw]
3. Voiceless Spirants f. s. by. h. hw.
4. Voiced Spirant v.
5. Sonants w. 1, r. y. w.
6. Nasal Sonants m. n. ny. ng [= u]
(nasals)
In addition the voiceless breath or 'aspirate' h.
The consonants enclosed in [ ] do not in Qenya occur initially or alone, but only in certain
medial combinations: b in mb; d in Id, rd, nd; dy in ndy; 9 in ng [ = ug]; and gw in ngw [ = !)gw].
The consonants of the ty-series, and q-series are all, except y, w, not simple consonants in
origin, and ty, ny, q, ngw [= uw] are not actually strictly simple consonants in pronunciation
though they can stand initially where Qenya does not tolerate actual consonant groups. See
further under the phonetic description.
The native descriptions of the sounds as reported by JElfwine may be interpreted as follows:
1. The voiceless stops are explosive (as in modern English), but not aspirated; their force is
slightly greater immediately before a stressed vowel. The voiced stops are markedly voiced.
2. The 'dentals' t, d, n, l, r are 'point' consonants, but made as in English just behind the
teeth. 1 is normal between vowels and initially; but is slightly palatalized before y, and
between i and a following consonant. (Thus mult- is sounded as in English, but pilt- has a front l
and the English pronunciation would suggest piult to Qenya-speakers). r is a trill in all
positions (weakest finally). Both l and r are partially unvoiced before p, t, k, ty, q, s.
3. s is not strictly homorganic with the other dentals, but is now whatever its origins
(these are various) the normal s, a voiceless 'blade' consonant, with the blade of the tongue in
contact (and so possessing an i-timbre), but with the point touching the lower teeth. It has no
tendency, even before [i] to palatalization to [s] or [�]. though original sy anciently became
palatalized and yielded [�]. Feanorian ).. transcribed hy.
4. y and w are non-spirantal semivowels, that is i and u respectively in consonantal
function (as in English). After voiceless stops, in q (= kw) , tw, ty, py they are partly unvoiced
(but not spirantal).
5. by is an audibly spirantal voiceless y, that is approximately [�] as eh in German ich. hw is
similarly and audibly spirantal voiceless w, that is [lu] as in English wh in white (in certain
varieties of pronunciation).

6. t in ty is the 'front explosive' [c], as e.g. Hungarian ty; but it is followed by an appreciable
partly unvoiced y-offglide. This is more marked medially (where ty counts as a group); less so
initially.
7. dy was formerly the voiced counterpart [J] followed by a y-offglide. But it only occurred
in the group ndy, which has become simplified to ny. The etymological spelling fh [ndy] is
often retained, but phonetic f'h [ny] is more usual: as e.g. in the adj . Qenya for qendya
'belonging to the Qendi, Quendian'.
8. n in ny is palatal n but followed (cf. ty) by a y-offglide, more marked medially (where ny
counts as a group), less so initially.
9. q (kw) consists of a lip-rounded 1< followed by a partly unvoiced w-offglide (more
marked medially than initially).
10. gw which only occurs in the medial group ngw is the voiced counterpart: a lip-rounded
g followed by a w-offglide.
11. Initial ng [= :n] is a reduction of older [:ng]: it is the simple back-nasal as in English sing.
But in Lindarin, whence the usual 'tarqesta' pronunciation is derived, the sound [:n] is only
retained medially, before k, q, g, where it is conveniently and unambiguously transcribed n.17
Initial [:n] is lost - repeating the treatment of the older Eldarin [:n] : see the phonology. In
pronouncing Qenya, which is primarily a form preserved in writing of a long-past stage of the
Lindarin dialect, the Lindar substitute n for the ancient [:n] represented by the still retained
Feanorian letter OJ:A' . This pronunciation ...
1 7 The beginning of this sentence was emended in the course of composition from: "But it has in Lindarin, and so
in the usual 'tarqesta' pronunciation, the sound ...
"

[Version 1]
JElfwine
In his narrative and translations spelt after Old English fashion. Thus ;sldf ..fal-l gecirfael for
iarvael, Bloodstained. The distinction between voiceless f, p and voiced [v], [o] was not
systematically represented. 1 8
f,
For [f] JElfwine uses initially and occasionally (after Irish and Latin) ph especially in ety
mological relation to p, as e.g. in i-p hinn, the Elves, for i-finn after the normal penn, pl. pinn;19
ff,
medially he often uses and (especially after r, l) ph beside f
For [v] JElfwine usually uses b (ignoring the 'variation') since initially [v] in N. is always a
variation of normal b; but he also occasionally uses u, and (rarely) £.2° Medially he uses f,
occasionally u, and (rarely) b. 2 1
For [p] he uses }> and frequently th, so usually when p is a grammatical variation as in ithail
for ipail 'the feet'. Medially he uses p and often }>}>. For p a is (as in Old English) a variant, but
JElfwine seldom employs it initially.
For [d] he uses d (cf. b above for [v]) but also occasionally a and }>. Medially he uses a and p,
and rarely d.
For [x] he uses h and hh medially, and also eh. eh is normally used by him initially, and
before l, r. 22
[p], [t], [k] are transcribed > p, t, c. b, d, g > b b "5' .
[ A ] he transcribes i and ge.
[lu] " " hw. Voiceless l, r by hi, hr beside lh, rh.
The nasal spirant [v] and back spirant [3] were no longer present in his time. But he knew
the old sign for 'nasal v' and sometimes represents this (espec. where it is an initial variant on
m) by mh.
In his day also voiceless n, m (i'i, m) were obsolete except occasionally initially where he uses
hn, hm.23
1 8 In the left half of the top margin Tolkien added two charts of consonants, probably sounds and their spellings:
[p] [t] [k] p t c
b d g b d g
V 5 3 V d �hl
{bh dh}
f th
f p, s X f th, s {eh »} h
In the second chart there is a brace to the right of "v" and the first "f," showing that these are variant spellings of
the sound [ v]. In the right half of the top margin Tolkien wrote the form yrh, presumably a spelling ofyrch.
1 9 The parenthetical phrase "(after Irish and Latin)" was inserted later.
20 The second clause in this sentence was emended from: "but he also occasionally, on analogy of ph (from Irish
and Latin) uses bh, and (rarely) f."
21 The phrase "occasionally b and u " » "occasionally u, and (rarely) b" in the course of composition.
22 The phrase "and before 1, r" was added later. Originally following this sentence was another, which Tolkien
struck through: "Noldorin has lost [3] in his day."
2 3 This sentence is written in ink over an earlier version hastily written in (now faint) pencil, which mentions both
"obsolete nh [ii], mh [m]" and "v (spirant m)." So probably the second sentence in the previous paragraph was
added at the same time as this one; both appear to be written with a slightly thicker nib than the rest of the text.

But in his notes on the languages he is more careful and original. Here he recommends (and
usually adheres to) the following usage.
[v] f throughout (but proposes b as an 'initial variant') 2 4
[o] a or � " but proposes d or 0 for [d] initially. 2 5
[f] ph but he actually often uses f initially when not a grammatical variant.
[p] th " " " " " p " " " " "
[x] eh " " " " " h, hh medially.
[!] i.
He naturally uses p (vrynn) for Noldorin cc (awae).
JElfwine's use may be modernized thus:-
f
[v] = throughout.
"
[o] = a
[p] = th
[f] = ph
[x] = e h
[!] = i. 2 6
But in Anglicizing names in narrative we use:
f
= [f] except finally where f = v. here we use ff, alff.
v = [ v] " " f
where v is not used in English, we use (as in of) nindalf = nindalv.
hfor eh.* - as Rohan = N Roehan (Horse-land).
d is usually written d but often th. usually d before r, Galadriel = Galaoriel.
[*] Note this corresponds with Toleressean pronunciation. In which eh has become
weakened. 2 7
Also Toleressean has no final o, b since they here became th, f: nindalf.
24 The original phrase "but !; or bh" » "but proposes !;" in ink.

2 5 This phrase was a replacement in ink for: "but o or dh as an initial grammatical variant)."
26 After the second item in this list Tolkien deleted "or dh"; and to the right of the list and pointed to by a large
brace was a note that he also later deleted (all in ink): "But actually though not so regular we employ v or bh for
initial [v] and f for initial f {medial f being written} ph {as a} being used as initial grammatical variant and
medially, so alph 'swan', i-phinn but finn." Below this Tolkien wrote the forms nindalv and nin-dalv, separated by a
large space, the second form being underlined. Following the first form nindalv the original word "or" was deleted
in ink, and above the v he later wrote f in pencil.
2 7 This sentence was originally: "In which eh has become weakened and o has become d wherever[?] preserved."
The following sentence was added in pencil.

[Version 2]
On J£lfwine's Spelling.
J£lfwine, in his own narrative, and in translations of Elvish matter into his own tongue, spelt
the names taken from the Eldarin languages after the manner of the English of his day,
following as nearly as he could the pronunciation that he heard in Eressea. Thus for Iarvael
"the Bloodstained" he usually wrote �'F�el,.
But he gave considerable thought to the transcription of Eldarin letters, and discusses the
system that he proposes to use, and in fact did use, if not with perfect consistency, in his
remarks on languages, or in his translation of linguistic lore and writings. J£lfwine was
acquainted with Latin and, to some degree, with Irish, and shows this in the divergences of his
system from the contemporary usage in Old English. The principal difficulty that faced him
was the representation of the spirantal or hissing sounds, so as to distinguish between the
voiced and voiceless varieties of each, and yet not to produce a system completely unfamiliar
and unintelligible to his countrymen. In English of his time the presence or absence of "voice":
e.g. the difference between [v] and [f]: depended mainly on the position of the sound in a
word. 28 The voiceless form, as [f], as a rule occurred only initially, or finally; and the voiced
form, as [v], only normally occurred medially, except in, say, such combinations as ff, ft where
the sound was naturally voiceless. But this was not the case in Quenya or in Noldorin. In Old
English as J£lfwine knew it f, p, s could do duty for f, p (th in thin), and s, and also for v, a (th in
this) and z, without inconvenience. In Noldorin this would have caused serious confusion. 2 9
J£lfwine therefore proposed to limit f and p to their voiced values as [v] and [o] respectively.
For the voiceless values he proposes ph, th. He suggests that o (in Old English used merely as
variant of p without phonetic distinction) should not be used at all, except initially in Noldorin
where o is a grammatical variation of d. Similarly he suggests that a letter b, "after the manner
of the Old Saxons," should be used initially only instead of f where f [= v] is a grammatical
variation of b. In harmony with ph, th he proposes to use eh for the back-spirant which only
occurred in voiceless form in Quenya and Noldorin. With [s, z] he was less troubled. In
standard Tarquesta, or spoken Quenya, both s and z occurred medially. But the voiceless [s]
medially was derived from older Parmaquesta p (th). J£lfwine represents it as such with[?] th.30
In this doubtless he reveals that his teachers were Noldor. For the Noldor in their own tongue
had no [z], but did possess voiceless [p], they therefore tended to preserve the archaic p, s as
such, though the Lindar represented them as s, z.
For [w, 1J] J£lfwine used the p 'wynn' of Old English, which we here again transcribe as w.31
For U; 1l he uses, especially in linguistic writings or notes, i. But he also often uses t; (g) before
i, e, or � (ge) before a, o, u after the manner of his own language. This causes little confusion
in Quenya, which had no [g], save in the combination [!)g]; but in Noldorin which had back
stop [g] before or after any vowel it is not fortunate.3 2
28 The word "mainly" was inserted later, written hastily in pencil.

29 In these two sentences original "without any serious confusion" was emended to "without inconvenience"; and
"caused great confusion" » "caused serious confusion."
30 The phrase "with[?] th" was inserted later in hastily written pencil.
31 This sentence was originally: "For {[w] »} [w, y] and [j, 11 J£lfwine used p 'wynn' of Old English, which we here
again transcribe as w. " It was revised in ink, presumably before the composition of the next sentence.
32 The phrase "causes no confusion" » "causes little confusion" in ink.

The following is a summary of JElfwine's system and his actual practice, in the matter of the
difficult letters or sounds.
Phonetic value systematic variations
transcript in practice.
V f b occasionally initially in certain cases
described above. Occasionally u.
d l> a " " " But lE. also out of
habit often uses o = p medially.
z s In Quenya only. Here he probably
pronounced [s] in Noldorin fashion.
3 g Only found in Doriathrin.'li33
f ph ff: so usually for Noldorin medial double
or long [f] rather than cumbrous pph. But
also used often after r, l as alff "swan" =
alp h.
th l>l> often rather than tth for Noldorin
long [p] medially; but not otherwise.
X eh Usually hh for N. long medial [x]. Also
usually ht beside rare et and occasional cht
for Quenya ht [xt], which did not occur in N.
s s.
1, j i �, Jf' but this is very rare in actual

transcriptions or linguistic notes.34
y, w. p = w. Occasionally u in tu, cu = tw, cw.
Frequently in gu = gw. Rarely uu initially.35
With regard to c, g which were (especially g) variable in value in English JElfwine states
clearly that in his transcription (of Eldarin Cf. elf) they have "those sounds that we use in
'lj'L-a.,.. � (gylden camb)," that is much the same as in modern English "golden comb."
33 This symbol may refer to the paragraph at the end of the table.
34 The comment was altered in ink from: "but this is hardly ever found medially."
3 5 For this item Tolkien originally wrote simply: "no variations."

[Version 3]
On JElfwine's Spelling
J£lfwine, in his own narrative and in his translations of Elvish matter into his own tongue,
often spelt the names taken from Eldarin languages after the manner of the English of his day,
following as nearly as he could the pronunciation that he heard in Eressea. Thus for Icirvael
"the Bloodstained" he usually wrote S'A:(J"'f'.x.el .
But he gave considerable thought to the transcription of the Eldarin letters, and in his
remarks on these languages he discusses the system that he thought most suitable to use, and
in fact did use, if not with perfect consistency, in his actual transcriptions of passages in Elvish
tongues or in his translations of linguistic lore. J£lfwine was acquainted with Latin, and to
some degree with Irish, and reveals this in the divergences of his system from the
contemporary usage of Old English. The principal difficulty that he found was in the
representation of the spirantal or hissing sounds (such as f, v, s, z) so as to distinguish between
the voiced and voiceless varieties of each, and yet not to employ a system completely
unfamiliar to his own countrymen.36 In the English of his time the presence or absence of
"voice": e.g. the difference between [v] and [f] : was mainly conditioned by the position of the
sound (or letter) in a word. The voiceless forms, such as [f], occurred as a rule only initially
and finally, as in full, h lcif, modern English full, loaf; or in such combinations as ff, ft, as offrian, oft,
modern offer, oft. The voiced forms, such as [v], only occurred as a rule medially between
vowels or other voiced sounds such as [r, 1], as in salfian, hlcifas, modern salve, loaves.37 But this
was not the case with either Quenya or Noldorin. In Old English, as J£lfwine knew it, f, p, s
could do duty both for [f], [p: th in thin], and [s], and for [v], [8: th in this], [z] without
inconvenience. In Noldorin this would have caused serious confusion.
J£lfwine therefore proposed to limit the Old English letters f and p to their voiced values as
[v] and [8] respectively.38 For the voiceless values [f, p] he proposed the use of ph, th. He
suggested also that a, which in Old English was used merely as a variant of p without phonetic
distinction, should not be used at all, except initially in Noldorin when [8] is a "grammatical
variation" of [d] . Similarly he suggested that a letter b, "after the manner of the Old Saxons,"
should be used initially instead off, where this f [= v] was a grammatical variation of b.
He treats N. & Q. differently. Quenya [he] spells as nearly as he can like Latin, Noldorin and
other languages more after English fashion. He avoids the use of specifically[?] English letters
or spelling[?] .39
a, b, b, c, eh, chw, d, a , e, f, g, ge, h, i, hi (Ih), m, n, ng, o, p, ph, qu, r, hr (rh), s, sh, t, th,
u, p = w, hp = hw (wh), x, y, z.40 Not all these signs are used in any one language, and
even[?] in two or more.
36 The list of sounds in this sentence was emended in pencil from: f, th, s, d[?], z.
3 7 The examples at the end of this sentence were added later in ink in the left margin.
3 8 The words "the Old English letters" were inserted later in ink.
39 This paragraph was written very hastily in ink and the reading of much of it is uncertain. There is a torn piece
missing from the top edge of the page where we have supplied the word "he." The word naze had been more
carefully written on the page before the paragraph was composed and now intrudes just before the second
occurrence of the word "English," suggesting that the paragraph was written after the list and chart that follow.
4 0 The list originally included ae, ai following a; the items b , chw, hl (lh), and hp=hw (wh) were inserted later.

labials p b, ph, f( b), hp p m

t d, th p (o) l, hl, r, hr. n
s (sh) s.
e, 9 eh (h), h ng (n)
hi i (ge)
se
cw gw ehw hw p. ngw
qu.
He accordingly transcribes the la of the Parmaquestarin mode as th, but notes that while
the Noldor when using Quenya gave to this letter the same sound (th) as in their own tongue,
the Eldar of V alinor they say pronounce L:t & t alike as s at the beginning of words but C
was[?] in the inner parts[?] of a word softened when[?] . ..41
JElfwine & Latin.42 z, ts, x, qu. He does not use p. He does not require p since he had
combined h when[?] using[?] th or eh or[?] ht[?] a use[?] which did not occur[?] as [?frequently].
The chief difficulty here[?] was in[?] the[?] distinction between w, v. He used[?] Latin f = f, he
used b for v, and[?] u = w. But[?] as[?] a consonant[?] p.
hy, ty, dy, hw, ngw, kw. ke, ki. h� hr.
4 1 The latter half of this uncompleted sentence, and most of the following paragraph, were hastily written and the
reading remains uncertain in places. The beginning was emended a few times as it was composed: "the Noldor"
» "while the learned among the Noldor often gave" » "while the Noldor when using Quenya gave"; and "to this
sound" » "to this letter the same sound."
4 2 Before this phrase (in which the "&" is uncertain) Tolkien deleted a false start: "When quoting Quen[ya]."

[Version 4]
On JElfwine's Spelling.
lElfwine, in his own narrative and his translations of Elvish matter into his own tongue, often
spelt the names of Eldarin origin after the manner of the English of his day, following as nearly
as he could the pronunciation that he heard in Eressea. Thus for Iarvael "the Bloodstained" he
wrote Sfa.rfa.S-L . But he gave considerable thought to the transcription of the Eldarin
letters, and in his remarks on those Elvish languages that he had learned he discusses the
problems and proposes a system which he thinks most suitable. This system he actually
employs, if not with perfect consistency, when he transcribes passages in Eldarin tongues, or
in his translation of Noldorin linguistic lore.
lElfwine was acquainted with Latin, and to some degree with Irish, and reveals this in the
divergences of his system from the contemporary use of the letters in Old English. The
principal difficulty that he met was in the representation of the spirantal or hissing sounds,
such as f, v, s, z, so as to distinguish between the voiced and voiceless varieties of each: e.g.
between v and f. For in the English of his day in general one set of letters only was used: f, s, p :
their varying value as "voiced" (v, z, th in this) or "voiceless" (f, s, th in thin) being determined
by their position in the word.9I But this was not the case in Quenya or in Noldorin. In Noldorin
especially the use of one set of variable letters would have caused great confusion.
'I The "voiced" forms only as a rule occurred in the middle of words, as in salfian, hlcifas =
modern salve, loaves; the "voiceless" forms occurred initially or finally or when doubled
or followed by t, as full, hlaf, offrian, oft = modern full, loaf, offer, oft.43
lElfwine therefore proposed to limit the Old English letters f, }> to their voiced values, as v (in
loaves, very) and th (in other, this) respectively. For the voiceless values, as f{in full, if) and th (in
thin, ether), he proposed the use of ph, th. In harmony with ph, th, he proposed to use eh for
the back spirant (as in Scottish loch, or German bach), which only occurred in voiceless form in
Quenya or Noldorin. With the distinction between s and z he had more trouble.
Noldorin did not possess the sound z, but in the standard Tarquesta, or spoken Quenya, both
s and z occurred in the interior of words.9l44 But this was an alteration of an older (classical or
Parmaquesta) stage in which the sounds were th (as in thin) and s (as in house). For this distinc
tion lElfwine had no available signs: z, if used at all, denoted ts in the usages with which he was
familiar. He suggests the use (in transcribing Quenya only) of sh for the voiceless medial s,
although sh was in his time (as he admits) a novel and not very acceptable combination.45
It had however no such associations as it has in modern English. lElfwine knew quite well
that the voiceless medial s of Tarquesta was derived from older th, and that the Noldor (who
possessed the th-sounds but had no z) often still gave it that sound. This derivation from th
may have directed him to the invention of sh, assisted by the example of ph, th, eh; but he was
not in fact guided by etymological considerations, and does not use sh initially for the Tarquesta
alteration of Parmaquesta (written or classical Quenya) th. In fact he often either omits the h
medially or writes a "Noldorin" th.
4 3 This note is given as a footnote in the manuscript.

44 This symbol suggests that Tolkien intended to write a footnote commenting on this, which he never composed.
45 This sentence was revised in ink from: "He proposes to use (in transcribing Quenya only) sh for the voiceless
medial, although it was in his time (as he admits) a novel and not very acceptable combination."

In the English of JElfwine's day there were two letters p and lJ, the use of which was not
differentiated by phonetics. JElfwine suggests that ):> only should be used in transcribing
Eldarin languages, except that a might be used initially in Noldorin, when [8], that is th as in
this was a "grammatical variation" of d. Similarly he suggested the use of a letter b "after the
manner of the Old Saxons," initially as v, grammatical variation of b.46 However this was an
unnecessary [usage] since in N. voiced th [8] and v only occurred initially as variations of d and
b; and JElfwine in fact usually employs p, f, while he often according to the force[?] of his native
habits often uses o as a mere variant of p in any position.
and used i. In his English the sound was usually represented by g (or ge) beside occasional
and more [?modern] i, but gi, ge were unsuitable in Noldorin. But i was not entirely satis
factory[?], since such words as ior in which it was a vowel beside 1 or (yor) could occur.47
4 6 In this sentence Tolkien originally wrote: "initially as a grammatical variation of b"; later deleted the word "a"
and inserted " v, when it was"; and then subsequently struch through the words "when it was."
4 7 This paragraph is written at the bottom of the manuscript page without any indication of the intended context.
It is possible that the reading of the first phrase should be: "and used l · " The explanatory phrase "in which it was
a vowel" was inserted in ink. The paragraph originally ended with another sentence that Tolkien later struck
through: "1£ . often spelled[?] dissyllabic ior as ihor." This was followed by several false starts on words beginning
with i, all deleted as well. At the end of the text the word iior was underlined and allowed to stand, apparently to
represent the way JElfwine would distinguish the dissyllabic word from the monosyllabic one.

[Version 5]
On JElfwine's spelling
J£lfwine, in his own narrative and in his translations of Elvish matter into his own tongue,
usually spelt the names of Eldarin origin after the manner of the English of his day, following
as nearly as he could the pronunciation that he heard in Eressea. Thus for Iarvael "the
Bloodstained" he wrote �oel. But under the instruction of the Noldor he gave
much thought to the transcription of the Eldarin letters, and in his remarks on the Eldarin
tongues he discusses the difficulties that he had met, and sets out a system which he thinks
most suitable. This system he actually employs, if not with complete consistency, when he
transcribes passages in Noldorin or Quenya, and in his translation of Noldorin linguistic lore.
J£lfwine was acquainted with Latin, and to some extent with Irish, and he reveals this in the
divergences of his system from the use of the letters in contemporary Old English. The prin
cipal difficulty that he found was in the representation of the hissing sounds, or "spirants,"
such as f, v, s, z. For he required to distinguish between the voiced and voiceless varieties of
each: e.g. between v and f. In the English of his day only one set of letters, f, s, p, was used, their
varying values as "voiced" (as v, z, th in this), or "voiceless" (as f, s, th in thin) being sufficiently
determined by their position in any given word.'ll
'll The voiced forms only as a rule occurred in the middle of words, as in wulfas, hlafas =
modern wolves, loaves; the voiceless forms occurred initially and finally, or when
doubled or followed by t, as in full, wulf, hlaf, offrian, oft = modern full, wolf, loaf, offer, oft.
This was not the case in Quenya or Noldorin. In Noldorin especially the use of only one set of
letters would have caused serious confusion.
J£lfwine therefore proposed to limit the Old English letters f, iJ to their voiced values, as v
and th (in this, other) respectively. For the voiceless values, as f and th (in thin, ether), he
proposed the use of ph, th. In harmony with this use of h he proposed to use eh for the "back
spirant" (as in Scottish loch or German bach), which only occurred in voiceless form in Quenya
and Noldorin.
With the distinction between s and z he had more trouble. Noldorin did not possess the
sound z; but in the standard Tarquesta, or spoken Quenya, both s and z occurred in the interior
of words. For this distinction [he] had no signs available: z, if used at all, in the usages with
which he was acquainted denoted ts. He accordingly represents the sign 'b of the Parma
questarin Mode by th, noting that while the Noldor even in using Quenya gave to this letter the
same sound as in their own tongue "the Eldar of Valinor, they say, pronounce h and ' alike as
s, save that in the inner part of a word� is softened as is s in our tongue."48
4 8 In this sentence "transcribes" was changed to "represents" in ink. The whole, written in ink with increasing
haste, was a replacement for the original sentence: "He suggests that sh should be used, in transcribing Quenya
only, for the voiceless interior s, although this {combination} was in his time a novel and not very acceptable
combination." Following this rejected sentence an entire paragraph was also struck out with a single diagonal
stroke:
It had however no such suggestion as sh has for users of modern English. The sound represented now by sh
was, so far as it existed in JElfwine's time, represented in English by se or see. JElfwine knew that the voiceless
interior s of the Tarquesta was derived from older Parmaquesta voiceless th, and that the Noldor often still gave
it that sound. This derivation from th may have directed him to the invention of sh ; but he was nonetheless
primarily concerned to distinguish sounds, and not to indicate etymology: he does not use sh initially in cases

For w (as in modern wood) J£lfwine used the letter p (wynn) of Old English, which is here
transcribed as w; but u often appears instead, notably in cu, gu, tu for initial cw, gw, tw. In
Quenya which he said "is as it were the Latin of the Eldar" he as a rule employed qu, and
avoided the use of specifically English letters such as p, o, p. For y (as in modern yes, beyond) he
employed in transcription or linguistic matter i. But he also uses 'S (g) before e, or ge before a,
o, u, after the manner of his own language. This causes little difficulty in Quenya, which had no
g save in the combination ng; but in Noldorin, which had a back stop g before or after any
vowel, such spellings are inconvenient.
The following is J£lfwine's alphabet, including the special combinations, arranged both in
conventional and in phonetic order.49
These rules apply principally to J£[lfwine's] treatment of N. with which he was most
concerned[?] and[?] most familiar. Q. he[?] treated somewhat differently. It was simpler in
sound structure {and[?] forms[?] also}. J£lfwine [?said] indeed[?] [?it] is[?] "as it were the Elf
latin" as Latin in[?] [?its] [?style]. Quenya is[?] far older[?] and [???] [?greater] [???] not[?] as[?]
the[?] tongues[?] of[?] Men[?] .50
where s of the Tarquesta is derived from Parmaquesta th. The combination sh being unfamiliar, we find that he
often omits the h, or writes th instead of sh (after the Parmaquesta spelling and Noldorin usage).
49 This sentence was emended in ink from: "The following is JElfwine's alphabet, including combinations that
denote a simple sound, arranged both in conventional and in phonetic order."
50 This paragraph was very hastily written (in ink) at the very bottom of the manuscript page, and its readings are
largely uncertain throughout. It includes several more words that we cannot interpret.

[Version 6]
JElfwine, in his own narrative and in his translations of Eldarin legends and chronicles into
English, often spelt the names of Eldarin origin after the manner of the English of his day,
following as nearly as he could the pronunciations that he heard in Eressea. Thus for Yavanna
he wrote �f"D-'1\'1\.CL (geafanna). But this was because he intended such matter to be read
by his contemporaries. He had in fact, under the instruction of the Noldor, given much
thought to the transcription of the Eldarin letters, and in his remarks on the Eldarin tongues
he discusses the difficulties that he had met, and describes the usage which he thinks most
suitable where accuracy is required. This system he actually employs, if not with perfect
consistency, when he transcribes passages in Quenya or in Sindarin, and in his translation of
Noldorin linguistic writings.51 Strictly speaking, he constructed two systems: one for the
transcription of Quenya, and another for Sindarin (and with modifications for other tongues to
which he refers) . In transcribing Quenya he used the letters as nearly as possible according to
the spelling and pronunciation of Latin as he knew it; for he said "the Quenya is indeed the
Latin of the JElfe (Elves), though it be of greater reverence and age than even the language of
Rome or any other tongue among Mortal Men." In transcribing Sindarin he followed rather
the usage of letters in the English of his own time, though he was obliged to introduce many
modifications of English custom; and in the alterations that he made he shows at least some
acquaintance with Irish.
Sindarin. In transcribing Sindarin the principal difficulty that JElfwine encountered was in
the representation of the "spirantal" or hissing sounds, such as are in modern English
represented by f, v, s, z, th. For in the English of his day one set of letters only was employed (f,
s, p), their varying values as "voiced" (modern v, z, th in this) or "voiceless" (modern f, s, th in
thin) being sufficiently determined by their position in any given word.'ll But this was not the
case in Quenya or Sindarin. In Sindarin especially, where either the voiced or voiceless forms
could, as in modern English, occur in any part of a word, initial, interior, or final, the use of one
set of variable letters would have caused great confusion.
'I The voiced forms were as a rule limited to the middle of words, as in wulfas, hlafas =
modern wolves, loaves; the voiceless forms occurred initially (in JElfwine's dialect), and
finally, or when doubled or followed by t as in fulL wulf, hlaf, offrian, oft = modern full,
wolf, loaf, offer, oft. 5 2
JElfwine accordingly proposed that the Old English letters f, ll should only be employed in
their voiced values, as v and th (in other, this). For the voiceless values as f and th (in thin, ether)
he proposed the use of ph, th. In harmony with ph, th he proposed to use eh
for the "back
spirant" (as in German Bach), which only occurred in voiceless form. With the distinction
between s and z he was not concerned in Sindarin, since that language did not possess the
51 The original phrasing "in Quenya or in Noldorin" » "in Quenya or in Sindarin" in ink, using a different pen
from the original composition; and similarly the seven occurrences of "Sindarin" in the rest of the text are all
later emendations. Note that the retention here of the phrase "Noldorin linguistic writings" shows that what is
meant are writings by Noldorin linguists.
52 This is a footnote in the manuscript.

sound z, and all he had to do was to note that in Noldorin words s had always that value which
it had in most men's mouths at the beginning of words of English or in sp, st.53
For w (as in modern will, would) JElfwine used the letter p (wynn) of Old English, which is
here represented by w. For the sound y (as in modern yes, beyond) he had no available sign, and
used simple i. In the English of his time the sound was usually represented by g, ge, but this
usage was unsuitable to Noldorin which had a stop g (as in modern give, good) before all vowels,
and in his more careful spellings and transcriptions JElfwine avoids it.
5 3 That "Noldorin" was allowed to stand here and in the following paragraph was presumably an oversight.

Quenya Verb Structure
Quendian & Common Eldarin Verbal Structure
Quenya Verbal System
Common Eldarin: Verb Structure
Late Notes on Verb Structure
by J. R. R. Tolkien
Edited with an introduction by
Christopher Gilson

Quenya Verb Structure J. R. R. Tolkien
Introduction
). R. R. Tolkien in the 1940s composed a grammatical description of the inflexions and syntax
of Quenya verbs and pronouns and their derivation from Common Eldarin elements and con
structions. The two parts of this Quenya Grammar that describe the formations of the various
tenses and other derivative stems of the verb are given separate titles: "Quendian & Common
Eldarin Verbal Structure"; and "Quenya Verbal System" (which we will abbreviate EVS l and
QVS). The pages of these parts were numbered separately from 1 to 12 and from 1 to 30,
perhaps because Tolkien was unsure that he would leave them together; but the first sheet of
QVS was originally numbered as pages 13 and 14, showing that it was at least begun as a direct
continuation of EVS 1. Both were written with a nib-pen in blue ink using lined candidates'
examination script paper measuring about 7 � by 10� inches.
EVS 1 and QVS were placed in the same box-file as the much earlier Qenya grammar witten
at Leeds in the 1920s/ and the "Quenya Spelling" documents we have presented above.
Following the latter is a piece of cardboard on one side of which Tolkien wrote "CQ CE" in blue
chalk and "CE (Quenderin)" in red ball-point pen, and "Common Eldarin or Quenderin" also in
red on the other side. After this is a sheet with the following heading and list written in ball
point:
Quenya
Accidence &c.
Yet unrevised.
Declensions
Verbs.
Pronouns.
Numerals.
Between this and EVS 1 is a bundle of 16 sheets and half-slips paper-clipped together, which
consists of roughly tabulated declensions for Old Qenya and Parmaqesta and a six-page
description of the Common Eldarin and Old Qenya elements and features underlying the Qenya
declensions. 2 Two of the rough declensions in this bundle are written on the back of the
minutes from a meeting of the General Board of Faculties at Oxford on May 5, 1944.
Tolkien placed a small torn slip behind the manuscript of QVS, on which he wrote in red
ball-point: "Old forms either revised or out of date." Following this slip are eight more loose
sheets and half-slips with rough Old Qenya and Parmaquesta delensions, three of which were
written on the back of a notice informing Tolkien of the appointment of examiners for his
student A. J. Bliss as candidate for the degree of B. Litt., which would have been in 1948. Placed
next to these is a five-page text with the heading Irregular Verbs. This is a revision of what was
originally the final section of QVS, and we have included it in this edition. Behind it Tolkien
placed his treatment of the Quenya "Personal Pronouns" and "Demonstratives, Relative, and
Correlative Stems" from this period, with partial versions and multiple layers of revision. We
intend to publish these texts in a future volume of Parma Eldalamberon.
1 "Early Qenya Grammar" (EQG), ed. earl F. Hostetter and Bill Welden, PE 14, pp. 35-86.
2 Cf. "Notes for Qenya Declensions," ed. Christopher Gilson, PE 21, pp. xvii-xviii, 66-9.

Quendian & Common Eldarin Verbal Structure

Tolkien starts his grammar of the Quendian verb with the observation that only a couple of
features go back to Common Quendian, "the association of an n-suffix or nasal infix with past
time" and of "lengthening or fortification of the stem" with duration. So his survey of older
features of inflexion is concerned mainly with Common Eldarin. This begins with a descrip
tion of the uses of the uninflected verb stem as an impersonal verb, as an infinitive, or with an
explicit subject noun; and then Tolkien describes the word-order of a simple Eldarin sentence
and the relationship of this syntax to the attachment of pronominal affixes to the verb stem.
Next he describes the formations of the tenses of the "simple or strong verbs chiefly of vTAL
form" in five sections on the 1. Aorist, 2. Present, 3. Perfect, 4. Past, and 5 . Future. Tolkien notes
that the Moods like our subjunctive or conditional are expressed by adverbial particles in
Eldarin, and he describes their syntactic interplay with the future and past tenses. The text of
EVS 1 concludes with a description of the formation of other verbal stems from bases of vTALAT
and vTA forms with or without formative or derivative suffixes of various kinds and also with
prefixed adverbial elements.
Much of EVS 1 is a clean manuscript without later emendations and many of its conceptions
can be traced in earlier remarks about the primitive language or are implied in grammatical
features of the Quenya verb. For instance, the Common Quendian association of nasal suffix or
infix with the past tense echoes the Tengwesta Qenderinwa of the late 1930s, which mentions
"the strong past tenses found with either n-infix + suffix e, ye, or with suffix ne, nye, as vKAT pa.t.
*ka-n-te beside vKAR pa.t. *kar-ne."3 The formation of the Common Eldarin aorist stem with
suffix -i, as in tuli 'come', is implied in the Qenya verb conjugations from around 1930, with a
shared stem tuli- in the aorist forms tulin, tulil, tulir, etc.4 Possibly connected is the explanation
in the Tengwesta that "kati-, katu- are variant stems in the conjugation of the verbal stem kat-,"
formed from the base vKAT by varying the vocalic extension in kata- with its repeated sund6ma
or basic vowel (PE 18, p. 34).
In some parts of the text Tolkien appears to be developing the conception in the course of
composition or soon after, making changes in the original ink. For instance, in a note about
the Eldarin suffix -n� used as one way of forming the past tense, which explains that n� was
originally an adverb meaning 'then, ago', Tolkien at first added:
But it is perhaps more likely that n-suffix and nasal infix are related, and that the
choice of which should be employed was originally primarily phonetic - the infix
being normally used before stops.
Subsequently he inserted a sentence before this one saying that Q ne can still "be used both
verbally as 'was' and adverbally as 'ago'." This in turn led him to revise the beginning of the
following sentence to say, in part, that -ne "as a suffix is ultimately distinct from n-infix,"
although they were already closely associated in Common Eldarin. In this edition of EVS 1 we
incorporate all of the revisions Tolkien made in the original ink, while citing prior readings in
the footnotes.
At a later date Tolkien made further revisions to EVS 1 using red ink. These were confined
to the second and third page of the manuscript, and center around a fundamental change in
3 Tengwesta Qenderinwa, ed. Christopher Gilson and Patrick H. Wynne, PE 18, p. 46.
4 "Qenya Conjugations," ed. Christopher Gilson and Carl F. Hostetter, PE 16, p. 127.

Quenya Verb Structure f. R. R. Tolkien
the conception of the basic word-order of the Eldarin sentence. This was originally expressed
briefly as follows:
The normal order in Eldarin was with pronouns as with nouns ... subject: indirect object:
object. "give the man a book" : "give him it." The subject usually preceded the verb, so
that the normal order of a simple sentence was: The man / he, (to) him, gives it.
Although this precise example is not also given in Eldarin, a similar sentence was cited which
illustrates the point about the position of the subject and object with respect to the verb, i.e.
Eldarin ni-kabi antii-t(e), from which comes the Quenya sentence ni·kave antiis = 'I can give it'.
This word-order is similar to that described in EQG:
The natural order in Qenya is (1) subject, (2) verb, (3) object of verb. Nothing usually
separates (3) from (2) except something qualifying either (2) or (3}, or occasionally
some postponed element belonging to (1), as:
ni·hepsine n(mo tanko i·mailinen losselin
'(it was) I myself (that) bound firmly the beautiful flowers' (PE 14, p. 54).
The fact that a pronoun such as ni- 'I' is prefixed to the verb is associated with the subject noun
also normally preceding the verb, and similarly for the object (noun or pronoun) that follows
it. The fundamental change arises from Tolkien's new conception of the position of the verb,
which he explains along with its implications in a marginal insertion written in red:
It seems fairly clear that in a normal simple sentence the verbal form was placed as near
the beginning as possible, though particles, prefixes or simple adverbial forms that
qualified the verbal notion could precede it. The subject normally followed, but if it was
displaced and put earlier for special emphasis the appropriate pronominal element was
often (not always) placed next[?] after the verb, so that the normal order of a simple
sentence was: Gives man it me or Man gives he it me.
And presumably reflecting this new position of the subject pronoun the examples cited above
were also changed in red ink to Eldarin kabinje antasa and Q kavinye antiis, while the translation
'I can give it' was retained, the subject prefix ni- being replaced by the subject suffix, Eldarin
-nje, Q -nye.
This new conception also involves a switch in the relative position of the indirect and direct
obj ects of the verb. Tolkien's treatments of the Common Eldarin nominal inflexions from this
period also show him wavering in his conception of this feature. In the document called Nouns
he originally wrote that "the indirect object (contrary to use in English) usually followed the
direct. [When however the indirect object was animate or personal this would come first.]"
Later he emended this to say that "the indirect or personal[?] object (as in English) usually
preceded the direct," and rej ected the bracketed qualification. Subsequently, in a document
called "Notes for Q. declensions," Tolkien restores the conception with two possible orders, but
a clearer qualification of the difference:
The indirect object, contrary to modern English use, usually followed the direct object,
though this order appears not to have been rigidly fixed, and where the indirect object
was personal (or 'animate') and the direct was 'inanimate', the personal or animate
noun could come first.5
5 Qenya Noun Structure, ed. Christopher Gilson, Patrick H. Wynne, and Arden R. Smith, PE 21, pp. 64, 67.

Quenya Verb Structure f. R. R. Tolkien
The late layer of revisions to EVS 1 in red ink was left incomplete. In particular none of the
other Common Eldarin and Quenya example sentences with subject pronouns were emended
to reflect the revised conception. Since incorporating the late changes would not result in a
consistent final text, we have given them in the footnotes.
In addition to the title centered at the top of the first manuscript page of QVS there is a
marginal heading, Quenya, which may reflect Tolkien's earlier intention of making the text a
continuation of EVS 1, as suggested by the originally continuous pagination. The parts of QVS
are arranged in approximately the same way as those of EVS 1 . The first half of the document
treats the "Conjugation of simple or 'strong' verbs," which is divided into six parts with the
headings: Aorist system; The Present (Continuous or Imperfect); The Past & Perfect system; The
Future; Imperative; and The Participles. The rest of the document treats verbal nouns and
adjectives; the conjugation of various other types of verb-stems: the so-called "weak" stems
which include a suffix between the base and the conjugational inflexions; Syntactic notes on the
usage of the infinitives and gerund, subordination of sentences, and the particles used to
express "moods"; certain Irregular Verbs, such as the verb "be"; and The Negative Verb.
Some aspects of this arrangement were redetermined in the course of composition. Thus
after describing the Aorist stem, its inflexion, and the uses of its infinitive and gerund, Tolkien
originally described the Aorist active participle; but his explanation of its usage required a
description of the contrasting usage of the imperfect or "present" participle, expecially where
these were used with other tenses. Consequently he decided to postpone the discussion of the
Participles until after describing all of the tenses. And in the event, sometime after beginning
(and perhaps much later than) the composition of the description of the Participles, which
followed that of the Future tense forms, Tolkien inserted a discussion of the Imperative
between them. This was composed on one side of a separate sheet with a note in the upper
right: "add to p. 9," that being the number of the page where the discussion of the Future
ended and where Tolkien added a notation in brackets to "take in Imperative."
At one point Tolkien's rearrangement of the material involved both EVS 1 and QVS. In the
discussion of Moods in the former text he first noted that the Eldarin languages normally use
adverbial particles of possibility or supposition to express conditionals such as "if he asks me, I
shall come," similarly to English, so that for example in Quenya this could be expressed by qe
e·kestan, ni·tuluva, with the particle qe 'if' in close connexion with the first. verb. Eldarin rarely
makes the distinction that English does with "were he to ask, I should come," where the future
action is subjunctive to indicate that it is suppositional, since for the Eldar the future was felt to
be suppositional in any case; although this could be made explicit by using the particle ai
'suppose', often combined with qe, as in Q aiqe e·kestuvan, ni·tuluva. Tolkien went on to note
that where the conditional is about the past "the possibility of wish or supposition known to
be 'unreal' arises," in which case the particle en, with or without ai, is prefixed to the second
clause; and he gave several examples from Quenya, contrasting in various ways, to express
different degrees of unreality in a wish or condition, including its explicit denial.
In the corresponding section of QVS Tolkien refers back to his remarks on Common Eldarin,
while giving more details of the Qenya expressions of condition and supposition, including the
differentiation of more or less contemporary from markedly subsequent conditioned events,
as well as conditions where there is no real supposition, in which case an adverb meaning
'when, whenever' is used in place of qe. When it came to describing the conditional in past

time, Tolkien apparently felt that he had already covered the details of Quenya thoroughly
enough in what he had written for his discussion of the Eldarin moods. So he incorporated the
original pages 9 and 10 composed for EVS 1 into QVS, renumbering them 29 and 30; and then
he wrote a much briefer explanation of the general point about the "unreal" conditional in the
past, on a fresh sheet of paper incorporated into EVS 1 as page 9, with the verso numbered "10"
but otherwise left blank.
These features of Quenya show that Tolkien had abandoned the conception of distinct
inflexions for the present and past subjunctive and conditional moods that appeared in the
"Qenya Conjugations" (PE 16, pp. 124, 127). The indicative tense-stems that appeared there are
retained: Aorist and Present tuli-; Past and Perfect mlie-; and Future tuluva-. But Tolkien's
further consideration of their historical origins in Common Eldarin has led to some alterations
and elaborations of his conception. Alongside an aorist stem like kari- 'do' with inflected forms
such as karin 'makes me' and infinitive kare 'to make', there is a present stem kara 'be doing',
also used as infinitive '(to) be making', formed with lengthened sund6ma and stem-vowel a
(from older a). There is also a repetitive form kakarra- 'keep on doing', with reduplication of
the first consonant and the sund6ma, doubling of the second consonant, and the same stem
vowel a. Tolkien suggested that in origin the Quenya present forms "probably were imperfect
or continuous without specifying time when." According to EVS 1 the CE present tense usually
implied an action "that began before the immediate moment and is likely to continue for some
time." This is its fundamental difference from the aorist, which "was used of all statements
true at all times," as in proverbs or descriptions of habits.
Regarding the CE perfect and past stems EVS 1 expresses uncertainty whether they were
"completely differentiated" or were simply "competing methods of forming a 'past' tense"
with functions yet to be fixed. In QVS Tolkien originally stated that in Quenya "there is no
distinction" in the significance or use of the CE past and perfect formations, and he tabulated
those that were employed for the past tense in Quenya:
The two types were
*a]manti or *a]mante 'ate' 1 Past
*a]kame 'made' I
*matie "have eaten" 1 Perfect
*karie "have made" I
This would provide an historical explanation for the diversity of past stems that Tolkien had
already devised, such as mlie vs. kame (cf. PE 14, pp. 57-8 ) . But he rejected this conception in
the course of composition, deciding instead that Quenya retained the CE perfect tense as
distinct from the past tense, and among the variants of the perfect the formation amiitiye came
to be preferred. This was a blending of two older formations, one as in amatye 'have eaten'
(which was used where the view was of the past) and the other as in tiiliye 'have come, am
arrived, am here' (which was used where the view was of the present effect of the past action).
The blended stem had lengthened sund6ma, the augment or prefixed repetition of the sund6ma,
and the suffix -iye, which became -ie in Quenya: hence tule 'come', pa.t. tulle, perf. umlie; kare
'do, make', pa.t. kame, perf. akarie; tyave 'taste', pa.t. tyambe or tyave, perf. atyavie, etc.
The emergence of a new conception of the Quenya perfect tense during the composition of
QVS provides a further indirect indication of the date of the text. The evidence suggests that
Tolkien wrote Book VI of The Lord of the Rings in 1948, and in the latest manuscript version of
"The Steward and the King" the first sentence of Aragorn's repetition of the words of Elendil

has the form: Et Earello Endorenna nilendie 'Out of the Great Sea to Middle-earth have I come', in
which the verb had been changed from earlier lendien. In this same manuscript Aragorn's
exclamation when he saw the sapling of the White Tree is: En nittiviet! 'I have found it!' In this
the Quenya had been changed from earlier En a tuvien/6 In the following year Tolkien made a
typescript copy of the manuscript of The Lord of the Rings, which was not finished "until after
Christmas," as he mentioned in a letter to Stanley Unwin in February 1950. In this typescript
Aragorn's Quenya sentences were typed in the revised forms of the latest manuscript, but the
verbs nilendie and niruviet were each subsequently erased and replaced with revisions carefully
written with a nib-pen in the spaces of the erased forms: urulien and utuvienyes, respectively.
These verbs both describe actions completed in the immediate past. If the revision of their
stems is from past-tense formations lendie and tuvie as in Tolkien's earlier conception to the
new perfect-tense stems urulie and utuvie, it suggests that this part of QVS was composed no
earlier than the completion of this chapter of The Lord ofthe Rings in 1948.
QVS presents an assortment of nine different Participles, an active and passive form for the
aorist, imperfect, perfect and future tenses, and a perfective that "was not passive or active but
denoted the completion of the action denoted by the verb." In the "Qenya Conjugations"
Tolkien briefly tabulated four Participles: active tule (tulj), passive trllzna; p. act. tulilya (tulfla),
pass. tulinwa (tulfna). We interpreted "p. act." to mean tulilya was past (or perfect) active, in
contrast with implicitly present active tule, but it is also possible that Tolkien meant tulilya to
be present active in contrast with implicitly aorist active tule (cf. PE 16, pp. 123, 128). In either
case, he seems to have reused the suffixes of at least three of these earlier formations for the
participles in QVS, albeit in more or less altered functions. The suffix -lya is used for three
active participles: imperfect karalya 'doing, making', perfect kamelya (also akarielya) 'having
made', and future karuvalya 'about to make'; the suffix -ina for three passive participles: aorist
karina 'made', imperfect karaina 'being made', and future karuvaina 'going to be made'; and the
suffix -nwa for two participles: perfective karinwa 'made, finished, done, completed', and
perfect passive karienwa 'having been made'. There is also an aorist active karia 'making'.
The imperfect participles describe actions that are taking place at the same time as that of
the main verb in the sentence. The aorist participles differ in that the significance is general,
not referring to a particular time, and thus the forms are adjectival and not used in strictly
verbal expressions. Similarly contrasting with the perfect participles used to describe actions
completed before the main verb, the perfective participle indicates completed action without
reference to time and was also purely adjectival. This led Tolkien eventually to reconsider the
arrangement of the material, when it came to describing the formation of "other substantival
and adjectival derivatives made from verb-stems that though frequent and made on fairly
regular patterns are not strictly part of conjugation." At first he included only a few
formations, such as karima 'able to be made, feasible' and karaite, karamaite 'able to make,
handy, crafty, skilled'. Subsequently he made annotations in red ink indicating that the
descriptions of the perfective and aorist active and passive participles should be moved and
combined with the description of these verbal adjectives.
Later Tolkien rewrote this description in an expanded version, which includes verbal nouns
such as karaste 'making, manufacture, construction', and istare 'knowledge', and additional
adjectival formations such as lirustea 'fit for singing, to be sung' and nyarula 'apt to talk'. He
6 Sauron Defeated, ed. Christopher Tolkien, 1992, pp. 12-13, 56-7. In nituviet the prefixed ni was written in smaller
letters above and to the left of the original form and with an undotted i; hence the uncertain reading [?in]tuviet.

incorporated revised descriptions of the perfective and aorist participles, referring to them as
"general adj ective" karia 'active (in making)', "general 'passive' participle" karina 'made', and
"perfective adjective" karinwa 'fully made, completely'. This was composed on both sides of a
new sheet, each given a number to fit into the existing pagination, and the following pages
were renumbered to accommodate this. The extent of the renumbering, indicating pages that
must already have been present, suggests that the revised description of the verbal nouns and
adj ectives was composed after Tolkien had completed his treatment of the conjugation of the
"weak" or derivative verb-stems. Eventually he deleted the earlier description of the per
fective and aorist participles with a series of vertical strokes and added a cross-reference to
the later description, although he did not remove them from the summary list of "verbal and
participial forms" that precedes the description of verbal nouns and adjectives. In this edition,
because of the complexity of revisions to the text, we have retained the earlier description of
the perfective and aorist participles in its final form before being deleted, with earlier variant
readings given in the footnotes.
QVS originally ended with three pages describing Irregular, Modal, and Auxiliary Verbs. These
were numbered 30 to 32, apparently meant to follow the description of particles used for
expressing wishes or suppositions, the last page of which was 29 (later emended to 30). The
first irregular verb described was the verb 'to be', the base of which was given as ...fYE; or YE ;
followed by a description of The Negative Verb, the base of which is LA./ALA; and the last page
giving some fairly rough notes under the heading "Irregularities," listing various bases whose
phonetic features produced irregular features in Quenya, e.g. KAB 'be able' with future kauva
from kavuva, and ...fN DO, unu-, undu 'down from on high' with derived verb unta 'descend' and
past tense ununte 'came/went down'. This page was eventually filled with forms related both
phonetically and semantically. Later it was deleted with a cross-hatching of ink strokes.
The principal tense formations of the verb 'to be' were Present-Aorist ye, Past yene or ne, and
Future yeva or va. These are comparable to the forms used in The Lost Road in the song sung by
Firiel, e.g. (rima ye Numenor 'lovely is Numenor' and yeva tyel ar i-narqelion 'there will be an end
and the Fading'/ The base was listed in The Etymologies as YE, I and described as the "stem of
verb to be," although no forms from the languages were cited (VT 46, p. 22). According to QVS
"where the subject is a noun the present of the verb to be is rarely expressed," and this was
exemplified in Firiel's song as well, e.g. Ilu vanya 'the World is fair'. A negative verb was given
under the Etym. bases uGu and UMU, described as "negative stems": Q uin and umin 'I do not, am
not', pa.t. lime. The new conception in QVS, with Present lci, Past lane, and Future lava, is
connected with a different Etym. base LA 'no, not', with derivatives such as Q la and lala. In QVS
in the space to the right of the heading The Negative Verb Tolkien added an annotation " no, lci,
lala," apparently alluding to this etymological connection.
Tolkien retained this description of the negative verb, but rejected the conception of the
verb 'to be' derived from a base YE , deleting the entire page with wavy pencil strokes. He had
made some rough annotations in the margins, including a form ye with a gloss that might be
'by, near'; a pair of forms ye- and yen with a gloss 'year'; and a group of forms do, lo, loar and
loyen, with a gloss that might be 'yield' (perhaps in the sense of 'give forth, produce'). These
seem to be connected with his work on the Appendices to The Lord of the Rings in 1949 or 1950.
In the second draft of "The Calendar" he introduced the terms yen, often translated 'year', but
really "a hundred of our years," and loa meaning 'time of growth' (of plants, etc.), a single year
7 The Lost Road, edited by Christopher Tolkien, 1987, p. 72.

in the seasonal sense.8 Perhaps the lack of a satisfactory association with .fyf. contributed to
rejecting this base for the verb 'to be'.
Tolkien composed a new beginning for the section headed Irregular Verbs and devised a new
base .f£ for 'to be', from which a Quenya stem ea is derived. This was used as the Aorist and
Imperfect tense form, with Past ne or eane, and Future va or eva. As with the shorter forms of
these latter two tenses being retained from the earlier conception, Tolkien also retained the
idea that in the aorist/imperfect the verb 'to be' was normally unexpressed in statements of
identity or with a predicate adjective, and he gives several examples, such as Eldar Alphfrimor
'Elves are Immortals'. The form ea was used with the primary sense 'to exist, to have being, to
be found, extant, in the real word', e.g. vahaia n6re ea i a-esta Valinor 'far away (there) is a land
called V alinor'. Location could also be expressed with a locative pronoun and without the
copula, e.g. tasse i·osto ! 'there's the city', but "is, be" with this sense was normally expressed in
Quenya by a more precise verb, and Tolkien gives a list of eight such verbal bases along with
examples of their usage. The resulting treatment filled three manuscript pages which were
combined with the page describing the negative verb and given a new pagination from "a)" to
"d)" in pencil.
Subsequently Tolkien composed a description of the gerund of the verb 'to be' and its usage,
on a half-slip. This originally had the form are and a variant, the reading of which is uncertain,
perhaps asse, the latter meaning 'existence' in the universal sense. These forms do not occur
elsewhere, but are is comparable to the ending of the gerund of certain derivative verbs, such
as farare 'hunting', especially when it was used as an abstract suffix in combination with a noun
or adjective, as in firimare 'mortality, the state of being mortal'. Later this text was revised so
that the forms of the gerund were ie and ere, the former always particular and the latter either
universal or particular. Tolkien hesitated over the forms of the abstracts, for instance revising
firimare to firimere, the giving firimare apparently as an alternative. Finally he composed a fair
version of this description, incorporating all the revisions, on the other side of the half-sheet,
which he annotated "(b) et." and pinned onto page (b).
Tolkien made some additional revisions to the main text of QVS in red ink, probably around
the time he revised the description of the irregular verbs or subsequently. These included the
following insertion into the description of the Future tense:
Verbs whose stem ended in v usually contracted thus .fKAv-, am able, capable, fut. kauva
(< kavuva). In the case of -ov, uv where ovuv, uvuv > 6v, uv so that fut. and imperfect fell
together as ruva (< tuv) the imperfect usually took long form as in weak verbs.
This describes one of the irregularities originally listed at the end of the discussion of irregular
verbs (the base given there as KAB), and makes the added point that this sometimes led to the
use of a variant "long form" of the imperfect. In describing the weak verbs such as fara 'hunt'
where the derivative stem ending in a was used for both the aorist and present imperfect
tenses, Tolkien had noted that the latter could be distinguished by using the imperfect
participle with 'to be' understood, as in ni taltalya 'I am slipping down', which became an
inflected stem as in he·taltalyar 'they are slipping down', referred to as the "long imperfect."
Using red ink he added notes in the margins at the end of the description of the Imperfect
tense of the strong verb about inflected long imperfect forms such as karalyate 'is making it' =
karat, which could be formed for any verb but was ordinarily formed for those "that for any
reason found it impossible or difficult to differentiate the 'imperfect' from aorist (or future)."
8 The Peoples ofMiddle-earth, edited by Christopher Tolkien, 1996, pp. 119, 126.

This new conception was cross-referenced in other parts of QVS with annotations in red
ink, and there were a few other emendations in red probably made at the same time. In this
edition of QVS, as with EVS 1, we have presented the text before these late changes in red ink,
giving these additions and emendations in the footnotes, along with any earlier readings
before revisions made in the original ink. There are also a few late annotations or revisions
indicated in pencil, but not carried through the text consistently, and these have been given in
the footnotes as well.
Common Eldarin : Verb Structure

Probably not long after Tolkien made the latest changes to EVS 1 and QVS he composed a
revised version of the former that also incorporated new conceptions that had emerged in the
latter regarding the Common Eldarin verb. We will refer to this text by the abbreviation EVS 2.
Tolkien gave it a title almost the same as that of EVS 1, written in red ink:
Quendian and Common Eldarin
Verb Structure
To the right of the second line he later wrote the letter "B" in pencil, marking the inclusion of
the text in a comprehensive treatment of the grammatical features of Common Eldarin, the
other two completed parts of which were similarly designated, correlating them with a listing
in pencil on a separate sheet beneath their collective title written in ink:
Common
Eldarin Morphology
A. Noun structure
B. Verb structure
C. Pronouns & Demonstrative stems
& Correlatives
D. Prepositions and basic adverbs
E. Numerals.9
EVS 2 was composed in ink on both sides of eleven sheets of unlined paper each measuring 7
by 8 % inches. The pages are numbered from 1 to 2 1 in pencil; the verso of page 1 1 is marked
"11 et." and contains a paragraph to be inserted near the bottom of page 11, as is also indicated
there in pencil. The text is divided into 24 sections, numbered in the left margin in pencil from
§1 to §25 (Tolkien inadvertently left out a §19). The sections range from a single paragraph to
a page and a half in length. §8 and §20 are each further divided into four subsections
designated by the letters "a)" through "d)" also added later to the text in pencil. All but four of
these sections begins with a heading which is underlined, or includes an underlined key word,
so that a listing of these headings will serve to summarize the contents of EVS 2:
9 Cf. "Common Eldarin: Noun Structure," ed. Christopher Gilson, PE 21, pp. xviii-xxii, 70-86.

§1 [Common Quendian and the Eldarin period.]

§2 Inflexion.
§3 The Verbal Stems.
§4 The verbal stems ofBasic Verbs. 1 . Aorist.
§5 2. Present.
§6 3. Perfect.
§7 4. Past.
§8 5 . Future.
§ 9 - §10 The Augment and Reduplication.
§ 1 1 Formation of derivative verbal stems.
§12 The TALAT dissyllabic stems.
§13 [a, u formative and derivative suffixes.]
§14 a-verbs.
§15 Verbs with consonantal suffix.
§16 u-verbs.
§ 17 Prefixes.
§ 1 8 Adjectival and Substantival forms of Verbal stems.
§20 Noun endings.
§21 - §22 Moods.
§23 Imperative forms.
§24 Negation.
§25 The pronominal affixes.
EVS 2 is a carefully written manuscript throughout, and in the few places where Tolkien
made changes in the course of composition he indicated deletions by careful cross-hatchings
in ink, and for passages that were inserted or reordered after the original writing he drew a
decorative frame and arrow in ink around the text to be added or moved and marking the
point of insertion. §§ 1, 2, 4-8, 1 1-13, 2 1 and 22 were based on the corresponding parts of EVS 1,
the texts sometimes differing only in slight ways, and it seems clear that Tolkien had the
earlier text before him while composing the later one. There are new conceptions introduced
in these sections. For instance in §8, which describes the Future tense of basic stems, Tolkien
derives the Quenya ending -uva from a base vuB 'ponder, have in mind'; so that the primitive
formation underlying Q matuvanye 'I shall eat' was mat-ubii-nje, which signified 'I intend to eat'.
Tolkien seems to have based the other sections of EVS 2 mainly on QVS, using statements
made there about Common Eldarin formations or implied in what is said about Quenya. As an
example we can look at the descriptions of the augment or prefixed sund6ma, which is first
mentioned in EVS 1 as a component of one formation of the Perfect tense of the basic stem:
The so-called "perfect" formations were marked by lengthening of the base (never by
fortification), followed by a suffix ( i)ye These stems were also often augmented. There
.
appears also to have been a variety with compulsory augment and short (aor .) stem
followed by ye (not iye). Thus vMAT: miitiye- "have eaten" I amatye I blended amiitie.
vTUL: tiiliye "have come, am arrived, am here" I utulye I utiiliye.
With slight improvements in wording this same conception is described in EVS 2, §6:
The formations usually regarded as 'perfect' were marked by lengthening of the base
vowel (never by fortification), and addition to the stem of a suffix:je, ije. There was also
a different but allied form with augmented base, with short base vowel and suffix-je.

Thus from v'"MAT: matie or amage; from v'"TuL: tiilie or utulje. There were also possibly
already in Common Eldarin mixed forms, such as amatie, utii lie (but not mage etc.).
In composing QVS Tolkien decided that the blended form of the CE perfect with augment and
lengthened sund6ma became the normal form of the perfect in the Quenya strong verbs. This
led him to consider the form of the augment for bases with phonetic shapes different from the
simple v'"MAT or v'"TUL and to include the following observations in QVS:
The augment was the sund6ma of 'strong' verbs. In case of those verbs beginning with a
vowel since they occurred before a stem already lengthened the augment was usually
omitted or absorbed, as ulle 'poured', ulie, has poured. Such forms as ululie occur only in
verse. For weak verbs see below.
The initial consonant of the stem was usually preserved in its Q. form as far as possible
after the augment, and reduplication. Anciently, of course, many initial groups,
simplified when initial, reappeared after the augment and reduplication. So v'"NDAK- :
nake, andakie; v'"STAR: thare, astarie.
Since these facts about Quenya must have arisen from features of Common Eldarin, Tolkien
included a discussion of these features in EVS 2. Reflecting the conception in the first of these
two paragraphs he writes the following in §10:
In the case of verbs that began with a vowel the augment naturally was absorbed by it,
and since the majority of verbal stems requiring the augment already had a long vowel
this disappeared: as in v'"uL 'pour': u-alye > alie 'has poured'. But forms with virtual
reduplication ul-alie were used in Eldarin also, and survive in the derived languages
especially in the verse-idioms, but also in some forms in regular use as Parmaquesta
ololie 'has become', ororie, has arisen.
This cites explicitly the CE phonetic process that underlies the example of Q ulie, and also
provides parallels to the variant form ululie from the weak verbs. The forms (ol)6lie 'it has
grown up, it has reached its prime' and or6rie 'has arisen' were given later in QVS as the perfect
tenses of the weak verbs ola- 'grow' and ort(a)/ory(a) 'rise'. The conception about the augment
explained in the second paragraph cited above from the QVS is only slightly different in EVS 2:
In languages that considerably changed initial consonants or groups (or treated them
very differently from medial groups) there was of course a constant tendency to
transfer the initial form to position after the augment. But the mb, nd, fig forms were
usually preserved in the older Eldarin tongues in augmentation and reduplication: as Q
nak-, perfect andakie. In Q. however initial groups in 'regular' forms received their initial
treatment after the augment and simple consonants also: as v'"sTAR : thar: atharie; v'"sAL :
asalie (not azalie).
The Quenya perfect tense for the base v'"NDAK is still andakie, but that for v'"sTAR has been revised
from astarie to atharie, with the historical development of initial st > th restored in the perfect
form after the augment.
EVS 2 §§14-18, 20, 23 and 24 are based in a similar way on the descriptions of the various
types of derivative verb stems, verbal noun and adjective suffixes, and the imperative and
negative verb formations as presented in QVS. One new conception is a particular type of
derived verb stem. EVS 1 and QVS describe a class of "formative" suffixes that includes the
consonants -t and -y, which are followed by -a in the aorist-present tense, as in orta/orya 'rise'
and sirya 'flow', and have infixed nasal before the formative consonant in the past, as in ronte

Quenya Verb Structure ]. R. R. Tolkien
or oronte 'rose' and sirinye 'flowed'. These same stems are described in EVS 2, with the slight
difference in that the CE sonant y is represented byj, so that the stems are spelled orta/orja and
sirja; and among the most frequent consonantal stem formants Tolkien lists t, j and s, giving a
further explanation of the last of these:
To this class belonged also the inceptives formed with s . This s was added direct to the
stem of adjectives, but in the case of verbal stems required an 'extension'. Thus taiikiis
'become firm, settle down'; niiikwis-, (begin to) grow pale; but kalas- 'begin to shine, get
light'; talat- > taltas- 'begin to slip down'.
Another revised conception emerged as a modification in the description of the verbal-noun
endings, those derivative suffixes included in EVS 2 §20 because they "had specially verbal
associations." The endings listed are: (a) ije; (b) tl, and strengthened -str, beside te, ste; (c) ndl,
nde; and (d) -le, re, me and extended ale, are, ume. Among these Common Eldarin suffixes iye had
been mentioned in QVS as the ending of the gerund of the basic verb, and also various endings
used to derive nouns of general action: -st�; -r�, or are ; and nde or nde. The suffixes ije and me
are also listed in EVS 2 §2 among "the most usual and probably oldest" suffixes used in the
Eldarin languages to form gerunds, in essentially the same list that was given at the cor
responding point in EVS 1. So the description in §20 is an organization and clarification of the
conceptions developed in the earlier texts. In paragraph (b) Tolkien originally mentioned "the
pronominal affixes t'Uste," noting that in Quenya where tr/str would become the same as these
pronouns (when final 1 > e) the verbal nouns "did not become part of normal conjugation" but
survived in isolated nouns, such as nemeste 'appearance, seeming' and tuluste 'advent, arrival'.
Later Tolkien struck through most of this paragraph in pencil with a marginal note: "no!
conflicts with pronominal suffixes." He deleted the endings te, ste and revised tr and -str to ta
and -sta, and the Quenya examples nemeste and tuluste to nemesta and tulusta.
A number of revisions in pencil, presumably made around the same time, are connected
with the change in Tolkien's historical conception, that the Noldor did not retain their own
language for daily use during the time of their exile in Beleriand, but adopted the Sindarin
language instead. The various mentions of "Noldorin" were emended to "Sindarin" and the
abbreviation "ON" (for Old Noldorin) to "OS" (for Old Sindarin). Tolkien also revised some of
the examples of ON forms, replacing them with the corresponding Common Eldarin forms,
probably because at this point he had not yet worked out all of the details of the ways Old
Sindarin differed from his previous conception of Old Noldorin. For example in §8 the element
marking the future, ON tho is retained as OS tho, but ON matithone 'I am going to eat' is replaced
by CE matithani.
These revisions show that EVS 2 was composed before the change in historical conception,
which emerged in the composition of the Grey Annals around 1951, and that the pencil
revisions to EVS 2 were made subsequently to reflect this change. There are also two short
annotations in red ink in the left margin of §16, providing glosses for two of the listed u-verbs,
liru 'sing gaily' and kheru 'lord it over, be master, own'. All of Tolkien's revisions, those made in
the course of composition in the original ink and those made later in pencil or red ink, have
been included in the present edition, with earlier readings provided in the footnotes.

Quendian
&
Common Eldarin
Verbal Structure
The verbal system was, as far as the evidence goes, hardly elaborated at all in Common
Quendian; though already one or two of the characteristic features of later Eldarin verbs had
appeared: e.g. the association of an n-suffix or nasal infix with past time; the lengthening or
fortification of the stem of strong verbs to indicate duration, with which also an a-suffix or
stem was usually associated.
The full elaboration of the verbal system belongs to the Eldarin period, though many
features were still unfixed at the time of the divergence; and the later languages (even in their
older periods) differ greatly - partly because in addition to choosing differently among vari
ations (of form or function), some of them, such as Noldorin, also abandoned a good deal that
was once common. The best preserved, as well as the one most elaborated by later invention,
was the verb of early classical Quenya.
The following features of the verbal system are old and belong to Common Eldarin.
Inflexion.
Where no subject was expressed; or when the "action" was strictly impersonal, as in "(it)
rains, (it) seems," the bare stem of the verb (or tense) was used uninflected. This was also
(unlike Indo-European) the normal form when the subject was expressed by a full noun or
name, except a person. 1
This bare (tense-)stem functioned not only as the stem to which inflexions (pronominal)
could be added but also as the 'infinitive' or object of another verb. Cf. English give in relation
to I give or gives and in such expressions as I can give. But this "infinitive" could only be used as
an obj ect, it could not be declined as a noun, nor stand as subject. This again is like English I
can give, I can give it but Giving does no good; I have no desire of/idea ofgiving. 2 Differences between
English and Eldarin were: (a) that English usually employs what is virtually a prefixed
infinitive inflexion to : I wish to:give; and in some cases uses the ing-noun such as I avoid giving.
In both these cases Eldarin used the "bare infinitive." In other cases it used a noun with a
suffix comparable to ing: [iye, -be, biye, me, miye were among the commonest].
� NB. As seen below, the pronominal inflexions of Eldarin were objective, not subjective
(as in I-E. languages). In consequence such a form as anta = give, but anta-te "gives" has
sense "give I gives it" or (to) give it. So Eldarin ni-kabi anta-t(e) > Q ni·kave antas = I can
give it.3
1 The phrase "except a person" was added later in pencil.

2 Tolkien originally wrote: "This again is like English I can give, I wish to:give, I can give it but Giving does no good; I
have {to »} no desire of/idea of giving." Later he deleted the second example heavily in ink. Following this was
another sentence, placed in brackets and later struck through in ink: "[In E. 'to give is a mistake' is an alienism.]"
3 This note was revised in red ink to the following:
w NB. As seen below, the pronominal inflexions of Eldarin were both objective and subjective (as in I-E.
languages). In consequence such a form as antii = give, but antii-sa (objective) has sense "gives it" or (to) give
it. So Eldarin kabinje antiisa > Q kavinye antiis = I can give it.

Quendian & Common Eldarin Verba l Structure f. R. R. Tolkien
The normal order in Eldarin was with pronouns as with nouns (see Declension): subject:
indirect object: object. "give the man a book" : "give him it." The subject usually preceded the
verb, so that the normal order of a simple sentence was: The man / he, (to) him, gives it.4
The unemphatic pronominal object thus originally as a rule followed immediately after the
verbal stem. From this already in Common Eldarin grew up a system of inflexions for the
objective pronoun - not subjective (as in Indo-European).5
Q> This has the result that the normal Eldarin verb expresses both the passive and the
active: e.g . ..fMAT "eat": aorist stem mati; 1 pers. matini (agent unspecified) "one eats me; I am
eaten."6
The agglutinated inflexional pronouns expressed normally the direct object in all transitive
verbs. Thus *antll-ni could only mean "gives me" not "gives to me." But in intransitive verbs
(especially the impersonal) - i.e. in verbs only capable of one object - the inflexional form
could be 'indirect': as *nemini "appears/seems to me. m
4 This paragraph was revised in red ink to the following:

The normal order in Eldarin was with pronouns as with nouns (see Declension): subject: object: indirect object.
"give a book the man" : "give it him." It seems fairly clear that in a normal simple sentence the verbal form
was placed as near the beginning as possible, though particles, prefixes or simple adverbial forms that
qualified the verbal notion could precede it. The subject normally followed, but if it was displaced and put
earlier for special emphasis the appropriate pronominal element was often (not always) placed next[?] after
the verb, so that the normal order of a simple sentence was: Gives man it me or Man gives he it me.
The original final example sentence was not deleted, but Tolkien presumably intended to replace it with the new
examples given in the space at the end of the paragraph. They were originally written (in red) as "Gives man it him
or man gives he it him," but the two occurrences of him were subsequently emended to me for clarity.
5 These two sentences were emended in red ink to: "The unemphatic subjective pronominal object thus originally as
a rule followed immediately after the verbal stem. From this already in Common Eldarin grew up a system of
inflexions for the subjective pronoun (as in Indo-European)." The original phrase "pronominal object" in the first
sentence was allowed to stand, though presumably Tolkien intended something like: "unemphatic subjective
pronominal element." Following the second sentence a false start was deleted: "But E. also had ... " This was
originally the last text on the page. In the bottom margin Tolkien added a paragraph written hastily in red ink
without indicating its intended placement:
This order was largely normalized in Eldarin languages, though these may differ in detail. In most it
became more usual for a full noun subject (espec. personal name) to precede verb; and since[?] it[?] would as
pronominal affix in that case come[?] next[?]. Emphatic subjective pronouns normally[?] preceded for their
purpose was full[?] emphasis[?]. In most Eldarin[?] languages [?they] [?have] no pronominal affix.
6 In this note the original phrase "rather the passive than the active" was revised to "both the passive and the
active." Later the note was replaced by the following written in red ink in the upper margin:
But since many words had no expressed subject - e.g. lE impersonals or passives - there was thus a
system also of objective inflexions, which largely corresponded to lE passives, as Q nimin, it seems to me, sakin,
it hurts[?] me, I am hurt[?], {ndakin, I am sl[ain,]} nahtanye, I slay, nahtan, (somebody[?] unspecified) slays me, I
am slain.
In line with the form nimin in this version (which is clearly written) Tolkien may have emended the form *nemini
» *nimini in the following paragraph (which was subsequently deleted).
7 In the first sentence of this paragraph the phrase "active verb[s]" was changed to "transitive verbs" in the course
of composition. The paragraph was later struck through, as well as the replacement for the previous note, and
the following was written in red ink between the lines of original text:
� But since many verbs had no expressed subject (or denoted 'events' rather than 'actions') there also
grew up a system of objective inflexions: e.g. Q nemin 'it seems to me'; nahtan 'slays me = I am slain', which
correspond closely in function if not in origin to the 1-E. middle and passive form. But the relation of
separate (disjunctive) pronominal forms to the affixed inflexions was much closer in Eldarin. In some
branches there arose both reflexive inflexions, and inflexions expressing pronominal subject + object, but
though the habitual grouping that gave rise to these later developments was no doubt already becoming

Quendian & Common Eldarin Verba l Structure J. R. R. Tolkien
The unemphatic subject pronoun always, if present, immediately preceded the verb and did
not in Common Eldarin become agglutinated to it or inflexional, though this happened in
Telerin. In Quenya and Noldorin the subject pronoun was proclitic and only tolerated the
intrusion of an indirect object pronoun (nothing else). In these languages, therefore, there was a
tendency for special coalescent forms of subject + dative pronouns to arise. 8
The unemphatic indirect object ("dative") was freer. It appears never, however, to have
followed a bare stem (of a transitive verb).
[ llli' Even in Quenya verse where both archaisms and poetic latitude of word-order appear
this is the case without exception.] Its normal place was between subject and verb + object (or
object if no verb was expressed).
Q> In Quenya verse it can follow (enclitic but not coalescent) after a pronominal object
inflexion; but this is a literary device (breaking the normal order: subject: dative: object). In
natural language, unless emphatic and in elaborated form, in which case it could be placed at
will, the unemphatic dative pronoun was placed immediately before the verb-stem. In
consequence, if there is only a dative pronoun and the subject is not expressed, the indefinite
subj ect [CE ga, Q a] was used. ni·anta = "I give"; e·ni·anta = "he gives (to) me"; but ga·ni·anta·t(e) =
"one gives me it, I am given it."9
Number in the objects was originally expressed only by their basic stems as ni "I": me "we";
e "it, he, she": khe "they." But number of the subject was expressed by inflexions added to the
verb stem before the addition of object inflexions. Where the subject was plural the inflex
-
ion was r/l : In Quenya r finally; but l medially before inflexions .'I Where the subject was dual
the inflexion was t/s : In Quenya (where final s became same as r) t was final; but s interior as a
rule. Thus: matir "(people) eat"; but "(people) eat something" mati-l·sa = Qu. matilda.
'I In some cases a nasal (assimilated to the pronoun) appears: in Q. before pronominal
elements of the 1st person. This is apparently [n] and not the same as the final -m/n
that sometimes appears at the end of object pronouns in pl. and belongs to them, not to
the subject.
Another notable point was the absence of a reflexive form, but the employment of two third
person pronouns: so called "near" and "remoter." The former were used when only one third
fixed in Eldarin the devising of inflexional forms with more than one pronominal element is a process later
than Common Eldarin.
8 This paragraph was later revised in red ink, mostly written between the lines of the original:
The affixed subject pronoun always, if present, immediately followed the verb or tense-stem, and already in
Common Eldarin became agglutinated to it or inflexional. The object pronoun only became 'inflexional' in
CE in the cases, cited above, where there was no subjective affix. After a subjective verb the object pronouns
followed in order (a) (b) as the one nearer to the verb (or intervening subjective affix) was the direct object, if
only one object was expressed.
�W There was probably no formal distinction but order in CE between accusative and dative of pronouns.
9 This paragraph and the two accompanying notes were partially revised by Tolkien in red, written between the
lines of the original text, which was allowed to stand probably because the revision was never completed:
The unemphatic indirect object ("dative") or 'second object' was freer in placing in CE except only where two
objects were expressed by pronouns they were in CE closely conjoined and the[?] indirect was seldom separated
from the direct. And[?] when[?] the first object was expressed by a noun the pronominal indirect obj . if not
emphatic was also enclitic to the noun in CE. The emphatic forms which grew up alike[?] [in] Eldarin
languages and then tended to oust the older enclitic, such[?] forms[?] were freer in placing; save that in all
Eldarin languages the order (with or without intervention of other words) was "accusative, dative" before or
after the verb. In some languages though[?] not[?] in the case of [?linguists] the habit grew up of placing[?]
an accusative[?] object before the verb and allowing[?] the enclitic form to be objective.

person (or group) was considered, or for the first of such persons. In consequence with a third
person subject it referred always to that subject and was in function "reflexive"; but it was not
reflexive after a 1st or 2nd person subject. See Pronouns.
Thus Common Eldarin ni mati-te = "I eat it (him/her)"; ke mati-te = "thou eatest it" etc.; e
mati-te = "he eats him(selO"; e mati-the or e mati-ste = he eats him (another man); ni mati-te mati
the or ni matitethe/ni matite yu the = I eat it and that.
Verbal Stems.
The richest and most elaborated were those of the simple or strong verbs chiefly of VTAL
form.
Already in CE, the majority of these had developed the following stems :-
1. Aorist. To the simple unmodified stem the suffix -i was added: mati, eat; tuli, come; kari,
make, do.
This "tense" denoted no tense or time. Its uses closely resembled those of the simple so
called present of English: I eat, I do. It was thus used of all statements true at all times (as in
proverbs): "gold glitters"; "the sun rises in the morning"; habits, etc.: "I live in a small house."10
Also as with E. it may be employed as an 'historic present' for narrative; and for the future
when defined by such words as 'tomorrow'.
�& A difference from English is seen in the verbs such as "love, know," which in English
seldom (or never in their proper sense) use the true or analytic present "I am knowing."11 In
verbs of this sort the aorist is usually not constructed in Eldarin; or rather a similar "tense" but
with a sufflx -l- is made: so CE melii-, love, istii-, know.
2. Present. This is a genuine "tense," can only be used of what is happening "now" - or of
course what is regarded as "now" in narrative or supposition; and usually also implies a fairly
continued action (that began before the immediate moment and is likely to continue for some
time).
The suffix of this tense is [i] . The stem was in Eldarin made in various ways. One feature
was lengthening or fortification of the base: so miit-ii; tiilii or taulii = "am eating"; "am coming."
Reduplication often accompanied by lengthening/fortification, or by "dynamic" lengthening of
one of the base consonants, was another feature. This however in Quenya or Noldorin did not
become part of normal conjugation, but served rather to produce new independent verb-stems
as *mammata- "to gobble up, to devour, to gorge." But the "repetitive" form with reduplication
and gemination of the final basic consonant was so common in Q. as to be virtually part of the
system of "strong" verbs; so tutulla-, to keep on coming (and going).12 [In Telerin reduplication
was the normal way of forming continuative presents: mamiita- "is eating."]
3. Perfect : describing an action that was completed in the immediate past, but the effects of
which are still present. It is not certain whether this was completely differentiated from the
next (Past) in C. Eldarin, or whether there were simply two similar competing methods of
10 Tolkien originally wrote: "or habits 'the sun rises in the morning'," but revised this in the course of completing
the sentence.
11 Tolkien originally gave three English examples, "love, think, know," but deleted the second one, presumably
because it does not actually fit the type of verb he is describing.
1 2 These two sentences were elaborated by an insertion in ink to an original single sentence: "This however in
Quenya or Noldorin did not become part of normal conjugation, but served rather to produce new independent
verb-stems as *mammata- 'to gobble up, to devour, to gorge', tutulla-, to keep on coming (and going)."

forming a "past" tense, the functions of which were not yet clearly fixed.13 The so-called
"perfect" formations were marked by lengthening of the base (never by fortification), followed
by a suffix (i)ye. 14 These stems were also often augmented. There appears also to have been a
variety with compulsory augment and short (aor.) stem followed by ye (not iye).15 Thus .fMAT:
miitiye- "have eaten" I amatye I blended amiitie. .fruL: tilliye "have come, am arrived, am here"
I utulye I utilliye.16
so The augment was probably more frequently used in words where the view was of the
past rather than the present: thus rather amatye, but tilliye "I am here.'m
4. Past. Plainly the form amatye described above was also often used in CE as a mere "past.''
But the pure undefined "past" (referring to an action thought of as over) was usually formed
either (a) by nasalizing the aorist (in which case the augment was usual): as manti, amanti "ate";
or by a suffix (added to the unmodified base) -n�:'l .fMBAR: mbame "dwelt" I ambame; or blended
form mante, ambami etc.
'I It is thought (though it is not certain) that this form was in the beginning confined to
intransitive verbs, and that ni (which also functioned alone as a base meaning "was")
was originally a time adverb "then, ago.'' It can in fact still in Q. be used both verbally
as "was" and adverbially as "ago," e.g. a e·ndanne anda ne "Oh! he went back long ago.''
In that case -ne in weak formations or as a suffix, is ultimately distinct from n-infix. But
it is clear that the two forms were already closely associated and regarded as related
already in C. Eldarin - where variation between n-infix and n-suffix was in any case
frequent - and that they were built into one system, in which the choice of which
should be employed was originally primarily phonetic - the infix being normally used
before stops. [With ne cf. en- of similar use as a deictic particle.)18
so In Quenya "past" or "aorist past" was augmentless, and used very commonly nasal
infixion or suffix ni; the perfect was normally augmented and as a rule has lengthened stem +
suffix -ie.19
5. Future. All the Eldarin languages express a simple future inflexionally, but the inflexions
and patterns used are different. This may mean that the device of an inflexional future had
already been achieved in CE but that the later languages all abandoned the older pattern and
separately substituted new ones; or that the future could still in CE be expressed by the aorist
1 3 The phrase "It is not certain" was a later addition to the beginning of this sentence; and the original phrase
"similar but competing methods" was emended to "similar competing methods," both revisions made in ink.
14 Two forms of the perfect ending were originally listed: -ye, iye.
1 5 This sentence was a later replacement in ink for a heavily deleted original: "This[?] especially when the suffix
was -ye and[?] the base was not lengthened."
16 The last form, originally given as utiilie, was altered in ink to utiiliye, although the corresponding form amatie in
the first set of examples was allowed to stand.
1 7 The word "probably" was inserted later in ink.
1 8 This note was expanded later in ink from:
'l[ It is thought (but it is not certain) that this form was in the beginning confined to intransitive verbs, and
that ni (which also functioned alone as a base meaning "was") was originally a time adverb "then, ago."
But it is perhaps more likely that n-suffix and nasal infix are related, and that the choice of which shd. be
employed was originally primarily phonetic - the infix being normally used before stops.
1 9 This note was revised in ink from original:
qo The Quenya "past" was constructed out of a blend of the forms of the perfect and past above described,
the perfect being expressed participially.

with defining adverbs, while there were already a good many other modes of expression in
competitive existence among which the later languages chose different modes to become their
normal pattern. Cf. English "I go home (soon, tomorrow, next year)"; or defined by a future
sentence: I shall go, when he calls; beside "I shall go, will go, am about to go, am going to go
etc."20
In Old Noldorin the future was usually expressed by adding tha to the aorist stem: matithii
-
"will eat." [This thii is probably in origin a defining adverb = then, next, since with time
reference the pronominal stem v'"THA in Noldorin referred forward.]
In Telerin a suffix -ro is added to the base (or the aorist).
� In Quenya the suffix is -ba (Quenya -va-), which is preceded by u (in case of most verbs).
The origin of this formation is not clear; but ba is certainly verbal, not adverbial. Cf. -IBA-, BANA,
go, proceed, as seen in Quenya vanwa "gone, over." Q karuvii- 'will make' thus originally = "is
proceeding to I going to I make". Cf
'li One view of the choice of u as "tense-vowel" before the suffix is that it is phonetic, and
due to the convenience and euphony of uvii - which proceeded from the uva naturally
produced in verbs with a formative -u : as kelu "flow," liru "sing." These verbs in
Quenya usually have future uva (keluva "will flow"); but this is probably secondary (a
device to differentiate them after the uva suffix spread to other verbs) and the -u- of
such verbs was in any case of variable quantity.*
* But is also outside tense associated with the notion of "begin to do something now
(completed later)" so in u-verbs: kelu, begin to flow. Cf. also u-la, adj . = apt to, likely to
do so.21
Moods :-
The notions corresponding to the "subjunctives, optatives, and conditionals" of I-E.
languages were seldom in Eldarin languages expressed inflexionally. The normal method was
to use adverbial particles of "possibility," "remoter possibility," or of "supposition" in close
connexion with the verb: usually immediately before it - in some languages even intruding
between it and the subject pronouns, more rarely before the obj ect pronouns: in which last
case they did in fact become inflexions.
The difference between e.g. "if he asks me, I shall come" and "if he asked me/were he to ask,
I should come" is rarely expressed in Eldarin.Cf
'li Both refer in fact to the future, which is essentially a supposition 'tense'. Contrast the
next.22
The former is expressed as in English, though with permissive use of the future in both clauses:
thus Q qe e·kestan/kestuvan, nHuluva; and similarly in other E. languages. The latter could be
differentiated by ai 'suppose' often combined with qe, if. aiqe e·kestuvan, ni·tuluva : � In this
form future in ifclause is required in Eldarin. 23
20 In this sentence the original example "I shall go, when I wish" was altered in ink to "I shall go, when he calls," in
the course of composition.
2 1 This note was added later, written in ink in the left margin.
22 This note was also a later ink insertion in the margin.
2 3 This paragraph was revised in ink from the following original:
The difference between e.g. "if he asks, I shall come" and "if he asked, I should come" is rarely expressed in
Eldarin, {though the second »} The former is expressed as in English, though with permissive use of the

In the past tense the possibility of wish or supposition, known to be or regarded as "unreal,"
arises. Here the particles were used in both clauses. Those that can be referred to Common
Eldarin are ai of "possibility"; and au of wish or supposition known to be unreal: giving a sim
ilar tone to the use in English of "may" and "might"/or "would have," respectively.
In this pattern some languages, e.g. Old Noldorin, developed inflexional conditionals by
including the particle between the verb-stem and the object pronouns. So ON ni tiilaiyeta "I
should have done it, if -."24
Formation of other Verbal Stems.

Beside the basic type already described there existed already in CE a number of other types:
made either from bases of other shape, as vTALAT; or derivatives from stems of all kinds verbal,
substantival, adverbial.
91 Not all simple bases were in CE naturally "verbal." Such bases as v'"(N)DER "male, man" (of
any speaking people: elvish, human, etc.) could, of course, only form derivative verbs. Bases of
vTA form or those of the invertible type (see Base Structure) oR/Ro "rise" did not form normal
"strong" verbs. In addition many bases of simple vTAL form made verbs of a different sort, em
ploying a fixed vocalic suffix (U, or a) after the base: as VKEL: kelu "flow," "well up";25 VGAL: gala
"grow." These were usually intransitive (not always*). The ii-type described states or actions
necessarily occupying a (continuous) stretch of time, as Q 'alii-, grow; ksarii, yearn, long for.
* e.g. telu- "roof in."
Formative suffixes of various kinds existed in CE of which far the most frequently employed
were :- (a) t, -y; and (b) tcl, ya.
-
These must be distinguished. (a) were employed to complete or define a basic stem, and
though commonly followed by vowel -a could be followed by others. Most important they
were originally regarded as derivative vTALAT-stems, and where nasal infix was used in
conjugation it preceded the affix. So v'"oR: ort-d, ory-a "rise": pa.t. ro-n-t-e or oro-n-t-e; ista
"know": pa.t. (i)sinte; siryQ, flow smoothly: pa.t. sirinye. In the same way were treated the
vTALAT bases as talta "slip down," talante "fell."
The true derivatives did not allow the intrusion of any element, and their ii vowel was
originally always long. ortii "to raise," ortiit(e) "raises it," ortiine "lifted." This -tii was specially
frequent in forming causatives from adjectives: *niiikwi "white, pale": nifikwitii- "to whiten."
Causatives from verbs more often used yii: as tu�ii- "send hither." But after [-u-] tii was used: as
kelutii 'cause to flow out'.
Already in CE certain adverbial/prepositional elements had already become agglutinated to
verbal stems (at their beginning). In that case the conjugation was as a rule the same as that of
the simple verb, as et-tuli "comes out." But this was not necessarily so omata- "eat away,
devour, corrode," or etkeli > ekteli "gush."
future in both clauses: thus Q qe e·kestan/kestuvan, ni·tuluva. The latter could be differentiated by ai 'suppose'
often confused with qe, if. aiqe kestuvan, ni·tuluva : � In this form future in if-clause is required in Quenya.
2 4 These two paragraphs replaced a longer original version, which dealt primarily with the forms used in Quenya.
Tolkien incorporated this into the treatment of the Quenya Verbal System (see below p. 122, footnote 126).
2 5 The second gloss of kelu- "well up" was inserted later in ink.

Conjugation & Syntax
Quenya1
Conjugation of simple or "strong" verbs.
Aorist system: Basic form kan-, do; ole-, become; nemf-, seem etc. 2
Uninflected form: kare, neme, Ole.
Inflected form: this form could only inflect for
(a) number agreeing with that of subject of the action, pl. karir, nemir etc., dual karit, nemit.
(b) the pronominal obj ect, which was then always the direct object, except in the case of
intransitive impersonal verbs where it could be "indirect" or dative (as regarded by European
languages). So karin "makes me"; nemin "seems to me."
"' For the forms of the pronominal inflexions see the Pronouns.
Infinitive and Gerund.
The "infinitive" was the same as the basic form and could be similarly inflected: kare, (to)
make; karin, (to) make me. "' This form was only used following and as the direct object of
another verb. For syntax see further below.
The "gerund" was formed with the suffix [ye] added to the aorist tense-stem: so karie
"making," nemie "seeming," matie "eating," tulie "coming," etc.3
This form was used (a) always as the subject of a sentence, or in other cases where it was not
the direct object of a verb, as when declined (as a noun) or governed by prepositions.4 It could
take singular (only as a genuine verbal form) noun-inflexions, but not pronominal (verbal)
inflexions. When used as the object of a verb the distinction from the "infinitive" was similar
to that between English (to) make and making. But the "gerund" in Quenya was in such cases
genuinely "aorist" or "consuetudinal," and was not used to describe particular actions. E.g. it
could be used in such expressions as "I like eating meat"; but not in "I do not like eating it now"
or "I shall avoid eating it tomorrow." In such cases Quenya uses the "infinitive." See further
under the syntactical notes below. "' The bare stem of every tense could act as an infinitive, but
the Gerund was formed only from the Aorist. 5
1 In the top margin Tolkien added a note: "alter mene throughout in sense 'mean', intend > tele. wish, want is mer
with true R. (root sense). The sense 'feast' is derived." See the section below of Syntatic notes, where e.g. Tolkien
-
altered ni mene tule "I wish to come" » ni tele tule "I intend to come"; but ni·mene i Turo tule "I wish Turo to come"
"
» ni·mere i Turo tule. Note that "with true R probably means that the r in mere is not derived from primitive d.
2 The example "oli , become" was inserted later in ink; and perhaps also the example ole in the following sentence.
-
3 The gerund nemie was originally given two glosses: "seeming," "appearance."
4 This sentence was emended in ink from: "This form was used (a) as the subject of a sentence, or in other cases
where it was not the direct object of a verb, as generally when declined (as a noun) or governed by prepositions."
5 This sentence was added sometime after the following paragraph was begun. This paragraph, from the middle
of the third sentence, replaced an earlier version that was originally followed by a discussion of the participles:
When used as the obj ect of a verb the distinction from the "infinitive" was much the same as in English
between (to) make and making. But note the "gerund" in Quenya was then genuinely "aorist" or general in
significance and is not used for the expression of a particular action.
It could be used in such cases as "I do no like eating meat." But in "I do not like eating this now," "I shall
avoid meeting him tomorrow, if I can," the infinitive must in Quenya be used.
&- See further on uses below.

Quenya Verbal System J. R. R. Tolkien
The Present (Continuous or Imperfect).

Basic form kara [i.e. the ending i only appears as long when final in dissyllables in OQ or in
polysyllables where separated from the initial syllable by an unstressed syllable: see
Phonology].
Q kiira, be doing; ola, is becoming; nema, is appearing; miita, be eating etc. The infinitive was
(as in the Aorist) the bare stem: kara "(to) be making."6
The repetitive form, only made with "strong" verbs, had the stem with reduplication (of first
consonant only), and doubling of last consonant, with stem-vowel -a.'ll
'li ph, th, h > pp, tt, kk. v and r (from d) properly showed mb (< bb), nd (< dd) - as tyatyamba
- but rr often substituted.7
kakarra- "keep on doing"; tutulla "keep on coming"; nenemma "keep on (re-)appearing." This is
virtually a new weak stem with past kakarrane, and fut. -uva.
On the forms taken by the reduplicated syllable see note below.8
In origin these forms probably were imperfect or continuous without specifying time when.
In Quenya however in "strong" verbs these forms are presents (except the participles: see
Participles). Many verbs (as will be seen) describing continuous action like *gal-, grow, only
formed this imperfect and not an aorist in [i] .
The past imperfect is used less than the English form "I was making."9
Participles.
On syntactical uses see below. The passive participles are hardly part of normal "conjugation." Owing to
the "passive" nature of Eldarin verbal inflexion they are seldom required syntactically except as attributive
adjectives.
The Aorist active participle or verbal adjective was formed with -yi: karia, matia, tulia, nemia, etc.
� The "present" time of a participle is that of the main verb or situation. {Deleted in the course of
composition: The "aorist" participle is :. only used when strictly aorist or habitual. As "at night the shining
moon arises" - silia rana [but "there it rises shining!" tasse e·orta silalya].}
This "aorist" participle is the most general equivalent of the English participle in -ing. The "present"
participle (in -alya : see next) is only used when the action is strictly contemporaneous and to some extent
continuous, at least extending parallel with that of the main verb.
So Rana nuresse {orya »} orta silia ; en e·orta silia! The Moon at night rises shining; there it rises shining!
But Rana oronte silalya or silia. The moon rose shining. The latter usual[ly] silia Rana.
But Rana ortalyane silalya (only). The moon was rising, shining (as it rose).
Tolkien drew a line below the sentence about the "Aorist active participle" and deleted the lower half of the page
with diagonal lines in ink. Subsequently he drew a large ink "X" through the entire page, apparently having
decided to delay discussion of the participles until after describing the present, past, and future systems of stems.
6 Following this Tolkien added, but subsequently deleted, in ink: "Also as infin[itive]s: '(to) be doing' etc."
7 This note was added in ink in the bottom margin. Tolkien had previously added a longer note, written in ink in
the left and top margins, with an arrow pointing to the end of the paragraph. He later deleted this note in pencil:
'll In reduplication q-, ty were treated as simple consonants: qeqetta "keep on saying." But by dissimilation
n, h, were used before ny, iiw, hy, hw; before ps, ts, ks s was used: sapsarra "keep on scraping." {Anciently
and so still in early OQ p, t, k were used before ph, th, kh »} In the case of aspirates anciently p, t, k were
used before ph, th, kh. This still occurs in early {OQ »} classical and remained in forms separated from
conjugation as tithilla (V'"THIL) "twinkle." Before ps, ts, ks [s] was used descending from cases where ps, etc.
= sp etc. So sapsarra "keep on rubbing."
8 This sentence was added later in pencil when the note on reduplication (given in the previous footnote) was
deleted; it refers to an expanded treatment of reduplication and the augment given below after describing the
other tense-stems and the participles.
9 This sentence was emended in ink from a longer original: "The past imperfect is used less than the English form 'I
was making', and usually viewed from present standpoint so that it more or less = 'I used to make'."

This is formed in Quenya by the addition of past suffix of "weak" verbs -ne to the bare
"present" stem. Thus karane, was making; ttilane, was coming.10
A "consuetudinal past" only used in the sense "I used to make" can be formed by suffixing
-ne to the imperfect participle (q.v.). Thus karalyane ; tuluryane "I used to come."
* Also occasionally used is a consuetudinal or imperfect future = "I shall be making, I
shall go on making, it will be my habit to make," expressed by adding the future (q.v.)
uva to the participle, karalyuva.11
The Past & Perfect system.

These were derived from those formations described above: see general account of
Quendian and Eldarin. The varieties amatye, (a)manti, miitiye were given up, and for Past the
augmentless formations: manti or kam� : were adopted. For the Perfect the formation amcltiye
was standardized.
Owing to the changes brought about by Quenya phonology the verbal stem became ob
scured in several of the forms made with nasal infix or suffix, especially since the development
of certain consonants, notably b, d, g, s, was very different medially as in the "aorist" from their
developments in conjunction with a nasal. In general among possible or inherited variations
that one was chosen or survived, which did not obscure or least obscured the medial basic
consonant. But there was constant cross influence between past and perfect stems that often
resulted in blending when one or the other became phonetically obscure or unacceptable.
Also verbs with the same apparent medials tended to influence one another: a TQ medial [r] for
instance could refer to CE r (with "past" rr, m) , or d (nd, nn); and form m which had a plainer
relation to the basic forms tended to spread outside its historic limits.12 The current form (not
always same at different periods) has to be noted for each verb; though there were some
general rules or customs: e.g. medials p, t, k normally had nasal-infix.
1 0 These two sentences were revised in ink from: "This is formed in Quenya by the addition of past suffix of 'weak'
verbs -ne either to the imperfect participle (q.v.) or the bare 'present' stem. Thus kdrane, was making; mlane, was
coming, or karalyane, used to dwell [se. make], tularyane, used to come."
11 This note written in ink in the top margin with an arrow indicating where to insert it. Later Tolkien added the
following, written more hastily in red ink in the lower left and bottom margins of the page:
The long imperfect (of same sense as the simple one) was formed by using the Imperfect Participle (q.v.) as a
verb. So karalyate, is making it kdrat. This so called "long imperfect" could {also} be formed by any verb (in
=
verse), but in ordinary language was only formed by verbs that for any reason found it impossible or difficult
to differentiate the "imperfect" from aorist (or future). This is specially common in late classical and TQ in
weak verbs with ending -a, q.v. But certain strong verbs, notably those with stems ending in ov or uv had
contracted futures identical with the imperfect (see Future). Thus YNOV-, future nouva > nova; YTUV, fut. ruva.
For such verbs already in classical the "long imperfects" [were] novalya, mlya (tuvalya).
12 The phonetic example in this sentence was revised in ink from: "a TQ medial [r] for instance could refer to CE r
(with "past" rr, m), d (nd, nn), s (ss, nn), and forms m, ss which had a plainer relation to the basic forms tended to
spread outside their historic limits." This sentence and the previous one were themselves a replacement for the
following, struck through in ink, in the course of composition:
But there were also {two other » three other »} other tendencies at work: a constant influence of past and
perfect stems on one another; a tendency to choose a past form that was same for apparenty identical medials,
as TQ [r]: Eldarin [s, d, r], so that me might be introduced from cases of the basic [r] like kar : kame - and, of
course, in later periods the medial form was likely to be reintroduced after changes had occurred which
obscured the relationship of medial to nasal form: especially if there was some analogy to work on. Thus
after [rn] historic in the case of ancient [r] as in kar : kame might be introduced instead of [ss] < [ns], or nd [<
nd] in such cases as ...

q- NB. Strong pasts "past aorists" were often formed from stems that were weak in the
aorist-present. These were not necessarily accompanied by a corresponding strong perfect.
A few examples are given below :-13
Medials.
1) p, t, k. ..fKAP, leap, ..fMAT, eat, ..fRUK, pluck: past kampe, mante, ruiike; perfect (akcfpie)
akampie, (amatie) amantie, urtikie or uruiikie. The substitution of nasal for lengthening was
already "classical," esp. in the case of medial p, t.
2) b, d, g . ..fKYAB, taste, ..fLAB, lick, ..fKuB, bow; ..fsYAD, cleave through, ..fMED, wish, want, ..fsED,
rest, ..fwAD, err, stray; ..fNDAG, slay, ..fLUG, be heavy.14 More examples are given here since the
developments were more various. Owing to the early pre-classical loss of [3] from initial and
medial g, very few of the older basic verbs with g-medial survived in Quenya in strong form:
thus ..fNDAG appears, as a verb, only in the 'weak' from nahta-, pa.t. nahtane. In case of [v] from b
the historic forms were mn, or mb, but "perfect" stem with lengthening tended to intrude. kave
'can', kambe "could."15 tyave, taste: pa.t. tyambe or later tyave; perf. atyavie. kuve, bow: pa.t.
kumbe, perf. uktivie. lave, lick: pa.t. lave (always so since lambe, substantive = "tongue"), perf.
alclvie. hyare, cleave: pa.t. hyande (non-classical TQ var[iant] hyame), perf. ahyarie. sere, rest:
pa.t. sere, perf. eserie; similarly ware, err.16 mere, wish for, desire: with pa.t. always mende.17 An
example of [g] is the common verb, used impersonally in Q., lue "it is heavy, sad," luin "it is
heavy on me, I am sad": pa.t. lunge, perf. ulungie (the historic form tuluiye is rare and archaic).
3) ph, th, kh. Not many strong verbs with these medials survive. Where they do they
preserve the historic forms with pp etc. < mph etc. with occasional forms derived from phn > pt,
khn > kt, ht (iikh, khn [???]). ..fRAPH, snatch: Q raphe, rafe: pa.t. rappe. ..fLAKH, kick: Q lahe: pa.t.
1 3 The discussion of The Past & Perfect system up to this point is the revision begun on a separate page of an earlier
version that was struck out in blue ink:
{The Past system »} The Past & Perfect systems.
In Quenya {this was a blend of formations: both those of "past" type in CE, and those of "perfect" type »}
these were {formed without [???] »} derived from the formations described above: See description of C.
Eldarin. In Q. there is no distinction at all at any period between the significance or uses of the two different
formations: they are used primarily as true "past tenses," the past of the "aorist."
The two types were
*a]manti or *a]mante 'ate' l Past
*a]kame 'made' f
*matie "have eaten" l Perfect
*karie "have made" f
In Quenya the forms mante-, kame, tulie became recognized all with "past" meaning. The augment is archaic
and only appears in early classical or in verse. The choice of form depends on the phonetic form of the base,
but there were, espec. in the older language and so in verse, many cases where two forms were used of one
verb. The current form has to be noted for each strong verb separately. In general, except for certain
irregular or isolated pasts, the form was chosen which left the basic stem unaltered or least altered according
to Quenya phonology. Thus the nasal-infix was almost universal with bases ending in t, p, k; lengthened
forms were usual with r, l.
� Strong pasts were sometimes formed from stems that had a weak "aorist-present."
A few examples only can be given.

1 4 The original root-form ..fMAD was altered to ..fwAD in ink; and the root 'VMED, wish, want" was added in the left
margin in ink, but this was subsequently struck out in red ink. (See footnotes 1 above and 23 below.)
1 5 This example was written in ink in the left margin without an indication where to insert it.
1 6 The original form mare » ware, in ink.
1 7 This example was inserted in ink; it was subsequently struck out in red.

Quenya Verba l System J. R. R. Tolkien
lahte, lakke. VPITH, spit: Q pise (pithe): pa.t. pitte. The perfects are arafie (araphie), � alahtie,
ip(sie.18
4) m, n. [:n] had become in pre-Eldarin [3] and then vanished, so early that there are
practically no traces of strong verbal forms that can be referred to ancient medial [n]. Possible
cases are the tlinge, ttange "rang, twanged" - descriptive of plucked strings - occurring in
verse as strong variants to the usual lingane, tangane from weak presents linga, tanga. These
may go back to prehistoric liyye, tayye < v'"uN', TAN .
[m] shows in past both mn and mm. v'"KHIM, adhere: hime, pa.t. himne, perf. ih(mie, similarly
v'"KIM, light on, find.19 v'"KHAM, sit down: hame, pa.t. hamme, perf. ahamie (ahammie). [n] usually
shows nn: v'"KEN, see, perceive, pa.t. kenne, perf. ekenie (ekennie); v'"TON- , tap, knock: tone, pa.t.
tonne, perf. otonnie (ot6nie); v'"MEN-, aim at, intend, purpose, with allative "make for, proceed
towards," menne; but v'"MUN, groan, moan (?), in Q. used = to express dislike, not to like: mune,
pa.t. us[ually] mtine (not munne), perf. umtinie.20
5) I, r. [1] usually employs ll (< nl), but ld (< ln) also appears. v'"TUL, come: tule, pa.t. tulle, perf.
uttilie. v'"KHAL (cf. halda, high, tall) in orhale, exalt, pa. t. orhalde (orhalle), perf. orahallie. 21 v'"OL
"grow," olle "became," perf. ol6lie, 6lie.22 Also from weak present: ehtelu-, well, bubble out (< et
kelu), pa.t. ehtelle, perf. ehtelunelye (see below) or etekelie.
[r] usually employs [rn] : KAR : kame, made, perf. akarie. v'"NDUR- "grow, be, dark": nure, nume,
untirie.23
6) y, w. Verbs with these medials rarely survived with full strong conjugations owing to the
weakness, and from point of view of Q. phonetic insufficiency, of medial y, w after a short
vowel. But several strong pasts survive. These always show addition of -ne. So: koita-, live, be
alive, pa.t. koine, perf. okoine. v'"LAw, abound: lauta, pa.t. laune, perf. alaunie used impersonally in
Q. as malta launen "gold abounded to me" = "I had lots of gold."
[3] vanished before Eldarin period, and no strong forms survive.
7) s. This became [z] medially; but ns became prehistorically (common to Noldorin and
Quenya) > ss. [Since later Q. tolerated ns this is sometimes found for pa.t. ss in late classical Q.
and TQ.] This form was adopted since sni > zne > nne was obscure: this nn only survived in
archaic poetic tlanne "heard" [ lasni ], no doubt kept alive by frequent use together with kenne,
saw. So hlasta- "hear"; hlasse (thlanne), perf. ahlcfzie.24 Similarly nusta, smell: nusse, untizie.
1 8 These two sentences and the three example verbs were revised in ink from: "Where they do they preserve the
historic forms with pp etc. < phn etc. with occasional infixed nasals. vRAPH, snatch: Q raphe, rafe; pa.t. rappe (rampe).
vLAKH , kick: Q lahe: pa.t. lakke. VPITH, spit: Q pise (pithe); pa.t. pitte. The perfects are arafie (araphie), alahie, alakkie,
ip(sie." The sign "' was added in the margin to the left of alahtie and the deleted form alahie.
19 The phrase "similarly vKIM, light on, find" was inserted later in ink.
2 0 The second example for [n] was revised in ink from: " ¥TuN-, tap, knock, touch (with the fingers): tune, pa.t.
tunne, perf. utunnie (umnie)." The third example was added in the left margin, without an indication of its in
tended placement; it was subsequently emended in red ink to: " vMEN-, go, proceed, menne" (cf. footnote 1 above).
2 1 The perfect form in this example was revised from orhalie, probably in the course of composition.
22 This example was inserted later in ink.
2 3 The following item was added in red ink: "vMER-, wish for, want, pa.t. meme, perf. emerie."
2 4 This was revised in ink from: "So lasta- "hear"; lasse (tlanne), perf. alazie." The archaic poetic form tlanne was
allowed to stand in the previous sentence, perhaps inadvertently.

Inflexion of Quenya
Past and Perfect forms.
In the past the uninflected forms (including infinitive) always ended in -e even when their stem
was borrowed from perfect as tyave, lave. The forms with long vowel in stem, however (partly
for rhythmic reasons and for the preservation of the characteristic long vowel; partly because
they were historically augmentless perfect stems) use ie before dissyllabic inflexions. So kampe,
kampen, kampelle; lave, laver, lavielle.
q- In verse where extra syllable was frequently convenient the ie forms could be used
before any inflexion.

In the perfect ie was always present in all forms inflected and otherwise. amantie, amantier,
amantielti. lungen "I was sad"; ulungien "I have been sad."
Derivative tenses.
A weak "pluperfect" was made in Quenya, by adding to the perfect participle the past-suffix
L
-ne .
So kamelyane "I was having made = I had made"; ttilielyane, I had come, lasse lyane, I had
heard, etc.
See further under weak verbs. Since these could not *normally distinguish between past
and perfect their perfect was periphrastic: thus ni nahtanelya "I (am) having slain" = "I have
slain"; but also where no pronoun was concerned Orome nahtanelya "0. has slain."25 From the
latter developed the form nahtanelye or by analogy with the strong verbs anahtalye "has slain."
The final vowel -e is on analogy of the normal perfect ending.
[*] se. when they had no strong pa.t.
In verse, where convenient, similar forms were made from strong verbs: akarielye/akamelye
= akarie. But there is frequently a slight difference of sense, the longer form being more nearly
equivalent to English "I have been making."

These are called "long perfects."
q- For karane, karalyane, imperfect and consuetudinal pasts see the Present Imperfect,
above.
The Future. In Quenya this was made with the suffix -vi.
The simple future: this showed -va added to the base + u.
q- On the origin of the vowel [u] see above. It always in Q. appeared before -va in
strong verbs, when added to the aorist base. 26 Also in words with present-aorist
formatives: as ista "know," istuva "shall know." Those with suffix -u already as liru
"sing" usually show lirnva [ -iiva]. The u is absent when -va is added to stems already
ending in an original long vowel: so ortuva "will rise," ortava "will raise"; kameva,
shall have made. 2 7
2 5 This sentence was revised in ink from: "Since these could not normally distinguish between past and perfect
their perfect was periphrastic: thus {ni ye nahtalya »} ni ye (nie) nahtalya 'I (am) having slain' 'I have slain'; but
=
also where no pronoun was concerned Orome nahtalyaye 0 has slain'." At the same time in the next sentence the
' .
original form nahtalya was emended to nahtanelya.

2 6 The phrase "aorist stem" was emended to "aorist base" in the course of composition.
2 7 The gloss of kameva was revised in ink from "was going to have made" » "shall have made."

Quenya Verba l System f. R. R. Tolkien
So: karuva "shall/will, am going to, make"; matuva, nemuva, tuluva, etc. Similarly the
infinitive: "to be about to."28
Derivative tenses. A past-future = "was about to, was going to" is frequently made by suffixing
-ne to the future stem: thus karuvane "was going to make"; later usually accented kanivane.
The form karuvane is to be explained as derived from karuva + ne at a period when the long
vowel of va had not yet been reduced, but ne was no longer stressed (as it had been in C.
Eldarin in all older formations such as kam� "did"); the formation was also probably first made
at time of Prim. Quenya (pre-classical) initial accent.29 So karuva-ne > karuvane. �& Note that
such "weak" futures as ortava, keluva usually make ortavane, keluvane. Now these are more in
accordance with the rhythmic predilections of Quenya (which favour dactylic words or
endings), hence kanivane even in the classical period occasionally appeared for karuvane.30
A future perfect was similarly made by adding vi to the perfect stem: thus akarieva "shall have
made"; similarly urulieva, alavieva etc.31 The augment is often omitted.
A rare form, a past future perfect, "I was going to have made," (a)karievane is sometimes
found, chiefly = "would have gone" in conditional clauses.32 See syntax of verbal forms.
�& It will be observed that the suffixes ne, va must be taken from word end back
towards stem to gain the meaning: uvane, was - going to not will have.
Imperative. 33
There are no special imperative forms. Command (or request) could be indicated by voice
tones. But commands or more imperative requests were often accompanied by i ! - so iisi, s(
"come now, now" - (the Q. particle for calling attention).34 Politer requests were often put as
future questions.
The bare aorist infinitive (or verb stem in weak verbs) was the curt and implied the curt sg. 2
ke. If plurality was desired it could be pluralized.
If i was added to the verbal stem this was imperious. Before the verb it was much less so.35
Thus tul' a! "come" (curtest); tule ! come! (curt). cisi tule, come now! If ke was expressed as
in a ke·tule it was less curt. plurals tulira! etc.
2 8 Below this paragraph Tolkien later added the following in red ink:
Verbs whose stem ended in v usually contracted thus .fKAv-, am able, capable, fut. kauva (< kavuva). In the
case of -ov, uv where ovuv, uvuv > 6v, uv so that fut. and imperfect fell together as ttiva (< tuv) the imperfect
usually took long form as in weak verbs.
2 9 The sentence replaced the following incomplete text in the course of composition: "The form -iivane (not
uvdne) is probably to be explained thus: the u-vowel was originally variable: fi; but in pre-Quenya the past suffix ne
was already bearer, when possible, of the main stress: cf. kameva [???] Thus karuvani ... "
30 The end of this sentence was revised in ink from: "hence kanivane ousted karuvane in the classical period."
31 This sentence was emended in the course of composition from: "A future perfect was similarly made by adding
vi to the past (or perfect) tense stem: thus kameva, lavieva, ttilieva (tulleva) ... " The form alavieva was revised in ink
to olavieva, but subsequently changed back to alavieva in pencil.
3 2 The original variant forms "(a)karievane (or {ak} karielyuvane)" were first revised in ink to the forms: "kamevane
[or (a)karievane]"; which were subsequently emended to just the form (a)karievane.
33 The section on the Imperative was added later on a separate sheet, with the intended position of its insertion in
the original text indicated by a note: "[take in Imperative]."
34 The phrase "so asi, s{ 'come now, now"' was inserted later, written below in ink with a line pointing to where it
should be placed (the dashes are editorial).
35 These two sentences replaced the following in ink: "Thus tule! come! curtest form."

With other persons the pronoun was expressed. le tule! a le·tule : dsi le·kare, come now, do it
(please). le tulir! a le·tulir etc. e·tul' a. he tulir a (imperious). a e tule, dsi e tule/ d(si) he tulir (less
curt).
Polite requests but made less so if necessary, by tone are ke·tuluva? le-tuluva? will you come,
please.
With other than 2nd person a was usually preferred in sense "let him/them" etc. a he
tuluvar, let them come. a he mittar, let them come in. But mitta! mitt' a, come in!
So with the negative verb. ld, ld, don't! a ld, don't do it. pl. lard etc. a le·ldte, don't you do itP6
The Participles.
These are not necessarily formed in the same way, i.e. from the same stems, as the tenses: to
which also they do not exactly correspond.
The perfective participle. This was formed with nwa It originally was not passive or active
-
.
but denoted the completion of the action denoted by the verb. It is found in words outside the
verbal system, as from stem va-, go: vanwa "gone for good," departed (dead, lost). Added to the
aorist stem, or in weak verbs the aorist-present, it retains this function: karinwa "made,
finished, done, completed"; tulinwa "arrived, come, now at hand"; alanwa "full-grown, mature,
adult"; istanwa, known, acknowledged, certain.* [With a perfect stem -nwa forms the perf. part.
passive.)l7 w This is not used as part of verbal system.
* No confusion arises between "weak presents" and similarly formed causatives since
where the forms coalesce so do the senses: ortanwa < 6rt(a) "rise," ortanwa < orta "raise" =
having risen on high/upraised on high = exalted.38
The aorist participle active is just an ordinary substantival adjective made with suffix -i added
to the aorist or aorist-present stem treated as a noun. So karia, making; matia, eating; tulia,
coming. A suffix -yi is also used with similar significance.39
36 The gloss for d le·ldte was originally "now don't you do it!" but the word "now" was deleted in ink. Following
this paragraph a sentence was added later, hastily written in pencil, so the reading of the last five words is very
uncertain:
Wishes were usually expressed with future with or without ti a laituva i·aran, (long[?]) live[?] [the] k[ing?].
37 This sentence was revised in ink from: "With a past stem -nwa forms the past part. passive." The brackets, and
the following note, appear to have been added after the original composition.
3 8 This is a footnote in the manuscript.
39 This paragraph went through several stages of revision, all in ink. In the first sentence the word "substantival"
and the phrase "treated as a noun" were later insertions, the intended placement of the latter being uncertain;
and the original form of the suffix might have been -ii or -vi, which was revised to -yi (with the y obscuring
what was there before), at the same time that another sentence was inserted (later heavily deleted in turn): "A
suffix -i with similar significance is also used." The paragraph originally continued as follows:
So karia, making; matia, eating; tulia, coming; istea, knowing; ortea, rising; keluya, issuing forth (of water). The
verbs with added (originally stressed) a: as ala-, grow (< galii); orta 'raise', should produce aldya, ortdya (alaia,
ortaia); but seldom use these forms; but make one from the gerund orta-verb: ortava.
Subsequently a letter (possibly either y or v) was inserted before the final a of karia, matia and tulia; and later these
inserted letters were all deleted, along with all of the other example verbs, including the following sentence. A
new third sentence was added: "A suffix -yi is also found with similar significance." This was tentatively
replaced by: "A suffix -li with similar significance is also {found »} used"; which was replaced in turn by a
sentence similar to the previous version, in which the suffix might be either -yi or -yi .

B" This participle is purely general and aorist in significance. If the action is thought of
as continuous or parallel to the action of another verb the "present" or rather
imperfect participle must be used.
The passive equivalent is ina. This is not [a] stem form used in strictly verbal expressions,
since the passive was expressed by inflexion in Q.40 karzna, made, matina "eaten"; istana,
known.41
NB. the inflexion is rna in which the [i] is independent of the [ -i] used as aorist stem.
The distinction between te ye matina, it is eaten, and (a)-matis, it is eaten (one eats it), is that
the second expression is the genuinely verbal one pointing to the action, the other is adjecti
val. The suffix is -ina which after a produces aina as ortaina, raised; but there are also many
adjectival forms, isolated from their original bases, in which -na or nasal-infix + a is used with a
past passive significance, so sanga, crowded, packed, as n. pack, throng, vSTAG.42
The imperfect participles.
These denoted (more or less) continuous action contemporary with that of the main verb:
that is "present" from the point of view of the situation envisaged. They are not however made
from the "present" stems, but with basic stem + 6mataima + suffix lyi Thus karalya "doing,
-
.
making"; mataO'a "eating"; himilya "adhering," nurulya "growing dark," etc.43 Where the base
or stem ended in l (lt ld) the suffix was dissimilated to -rya: as in tulurya "coming"; tultarya,
sending hither; olarya "growing"; silirya "shining"; olorya, becoming.44
"Weak presents" seem originally to have often45 made "strong" present or imperfect
participles, as izilya, knowing (ista); sirilya, flowing (sirya); but these (where surviving) were
purely adjectival and outside normal conjugation. The conjugational forms were istalya, ortalya
(from both 6rta 'rise' and orta 'raise'); olarya, growing.
4 0 This was altered in ink from: "The passive equivalent: only used {substantively »} as an adjective since the
passive was expressed by inflexion in Q., is ina."
41 Another sentence was deleted here in ink: "Since the original form was -nii the forms of the causatives etc. are
here -ana : ortana "raised" [< ortanii < ortanii] and a-present is aina : ortaina, raised."
4 2 Following this paragraph Tolkien wrote: "None of the above are used in conjugation"; and above it at the top of
the manuscript page: "These forms are only mentioned for completeness." Both of these notes were deleted in
ink. Subseqently the full discussions of the perfective participle and the aorist participles were outlined with a box in
red ink and the instructions: "Remove to p. 14." In the event Tolkien would incorporate a revised discussion of
these endings in the fuller treatment of verbal adjectives (see below). He then struck out the present sections in
the original ink and added a second sentence to the introductory paragraph on The Participles: "From the {Aorist
or base »} Aorist stem or from the base a number of participial adjectival formations are made that do not enter
into conjugation, or normal verbal system. For these see p. 14." (Cf. footnote 69 below.)
43 These two sentences replaced the following in the course of composition:
They all contained the -a- vowel of "presents"; followed by a suffix -lya; but the stem vowel was not being
stressed as in verbal presents (strong). Thus karalya "doing"; matalya, "eating."
Later the phrase '"present' stems" was emended to '"present' stems of strong verbs" in red ink; and a double
asterisk was inserted following the example nurulya with a note in the left margin, also in red ink:
** In Q. uvu was usually contracted to ii. Hence tuvulya > ttilya, in late Classical & TQ tuvalya with a was
substituted.
44 The original form alarya was revised in ink to olarya. Tolkien added the example olorya 'becoming' in the left
margin without indicating its placement in the list.
45 Tolkien later inserted an asterisk here with a note in the left margin in red ink:
* When derived from verbal bases of ..fTAL form.

"' This is the most used active participial form, often employed adjectivally as well as
verbally. On its use in the derivative tenses: the consuetudinal tuluryane "I used to come": see
above.46
The corresponding passive form was made from the present stem, followed by the suffix ina
described above: karaina, being made; orhalaina "being raised up"; mataina, tyavaina, istaina,
ortaina (only belonging to causative orta) etc.47
This is not much employed; but it is in classical prose the correct form to use in many cases
where English employs "while being (built, taught, etc.)": se. where the subject of the sentence
is also the object of another contemporary action. "While being trained he was often weary."
But it can also be used to qualify objects: "While it was being built people marvelled at the
house": ampanaina i·mar a·tatallanes / lie tatallaner i·mar ampanaina. 48
The perfect participles. Though these are in a sense "perfects" they are only so in that they
naturally normally refer to an action that has been (or was) completed before the main action.
They are formed from the stem (with or without augment) of the past or perfect.49 The active
participle uses the suffix lya as above.
Thus kamelya, akarielya, ..fKAR. ttullerya, uttilielya, ..fTuL. etc.
In early classical Q. there is still observable a distinction in use between the two forms: a
kind of tense concord: kamelya, past part. being used with a main action in the past: mar
kamelya e·tulle = "having built a house he came"; but mar akarielya e utt1lie.50
"' But in later classical Q. and in TQ the past-perfect participle is used (without augment)
made usually from the recognized past stem: kamelya; though the intrusion of the perfect
forms is more frequent. For instance with basic medial [l] the participle has always the form
ttilielya not tullerya.
For the development of cpd. participial tenses like istanelya 'have known'; kamelyane 'had
made'; etc. see above under Past and Perfect system. 51
On the "long perfect" made with the -elya participle see p. 7.52
4 6 The following was added in red ink: "On its use as a 'long imperfect' see above under Imperfect and Future."
47 This was revised in ink from:
The corresponding passive form was made from the present stem, though usually without stem lengthening,
and the suffix -nta: karanta, being made; orhalanta "being raised up"; matanta, tyavanta, istanta, ortanta (only
belonging to causative ortii) etc.
4 8 These examples were emended in ink from: ampananta i·mar a·tatallanes I lie tatallaner i·mar ampananta.
49 The original parenthetical phrase "usually without augment" » "with or without augment" in ink; and in the
following examples kdrielya » akarielya and mlielya » umlielya, both emendations also in ink.
50 In the beginning of this sentence "it is held that" » "there is still observable" in the course of composition. The
examples were first given as: "mar kamelya e·tUve 'he found a house that had been (already) built'; but mar
=
akdrielya e·tuve/ e·umvie 'he finds/or has found a house that is already built'." Subsquently e·mve » e·kimme; e·tuve
» e·kimme » e·kime; and e·umvie » e·k(mie, all in ink. Later changes were made in pencil: e·kimme » e·tulle; the first
gloss emended to "having built a house he came"; e·kime/ e·klmie » e utulie; and the second gloss was deleted.
51 In this sentence the original form istalya was emended to istanelya in pencil.
5 2 This replaced the following unfinished sentence in ink:
A "long perfect" is made espec. in verbs (such as weak presents q.v.) in which it was difficult or impossible to
distinguish perfect and past stems by treating kamelya etc . ...
The cross-reference is to the section above on Derivative tenses under The Past & Perfect system.

The equivalent passive participles were made only from the perfect stem (without* augment)
+ the suffix -nwa.
karienwa "having been made"; matienwa, having been eaten (beside mantienwa) etc. 53
* "without" - because of the 'perfective' significance of the suffix alone. Cf. karinwa and
similar forms: p. 14.
Since nwa is not specifically passive (see above) it is sometimes found with intransitive
verbs in place of -elya : trilienwa "having arrived." This is mostly archaic or t.
The Future participles. These are constructed like "imperfect" participles on the future stem.
Thus: Active: karuvalya "about to make"; matuvalya, tuluvalya etc.
Passive: karuvaina "going to be made"; matuvaina, istuvaina, ortuvaina (ortavaina, alavaina,
keluvaina < ortava, olava, keluva) etc. 54
The following are the verbal and participial forms belonging to a normal "strong" verb in
Quenya.
Aorist: Inf. kare. Infl. karir, karin, etc. (about 56 forms). 55 Parts. karia, karinwa; karina. Gerund.
karie inflected as noun sg. (about 8 forms).
Imperfect: Inf. kara. Infl. karar, etc. Parts. karalya, karaina. 56 Past Imperfect: karane, karaner,
etc. Consuetudinal Past: karalyane etc. Imperf. future: karalyuva, etc.
Repetitive: kakarra etc. Past kakarrane etc. Future kakarruva etc.
Past: Inf. karne. Infl. karner etc. Part. karnelya.
Perfect: Inf. akarie. Infl. akarier etc. Parte. passive karienwa. Long perfect akarnelye etc. Past
perfect. karnelyane etc. Fut. perfect. karieva etc. Fut. perf. in past or conditional
karievane. 57
Future: Inf. karuva. Infl. karuvar etc. Parts. karuvalya, karuvaina. Past future kartivane etc. 58
53 These two sentences were revised in ink from:

The equivalent passive participles were made only from the perfect stem (with or usually without augment)
+ the suffix -nwa.
(a)kdrienwa "having been made"; (a)matienwa, having been eaten (beside amantienwa) etc.
In connection with this revision the asterisked note was added in the top margin, also in ink.
54 This was revised in ink from: "Passive: karuvanta 'going to be made'; matuvanta, istuvanta, ortuvanta (ortavanta,
alavanta, keluvanta < ortava, alava, keluva) etc." Note alava » olava; but the first vowel of alavaina was not changed.
55 The inflexions included 51 endings distinguishing agreement with a singular, dual or plural subject, and 17
possible object pronouns: 1st, 2nd and 3rd person in the singular, dual and plural, with distinctive 1st person
exclusive vs. inclusive in the dual and plural, and 2nd person curt vs. polite and 3rd person nearer vs. remote in
all three numbers. The 5 additional inflexions were for dual and plural subject agreement without an object
pronoun, and 3 reflexives, singular, dual and plural.
56 The original form karanta » karaina in ink; and similarly below under the Future karuvanta » karuvainen.
57 The last two forms were revised in ink from kameva and kamevane, respectively.
58 This tabulation of forms replaced a preliminary version on the other side of the same sheet, struck out in ink:
The full forms of a simple "regular" strong verb may thus be set out for the .fKAR, make.
Aor.: kare "make(s)" - with or without pronominal proclitics. kare, infin. "(to) make." karin, karit, karis etc.
verb or infin. inflected for the object and subjective number (approx[imately] 56 forms). Partic. karia,
karinwa; karina. Gerund. karie.
Imperf. Pres. uninflected form and infin. kdra. Inflected karan, etc. Partes. karalya; karanta. Past imperfect.
karane. Consuetudinal past karalyane.
Past: Infin. kame. Inflected kamen etc. Parte. kamelya. Pluperfect kamelyane etc. [Long perfect: akamelye.]
Perfect akdrie; inflected akarien etc. Parties. [akarielya], {a}kdrienwa.
Future: Inf. karuva. Inflec. karuvan, etc. Parts. karuvalya, karuvanta. Past-fut. karlivane.

If we omit the "repetitive" as being virtually a derived verb,* and the rare "long perfect";
and also omit all the substantival inflections possible for the Gerund and Participles this
amounts to about 694 verbal forms provided by inflection for a fully conjugated Quenya verb.
* This form is not made by all verbs, and [is used] in several developed special senses:
thus vTAL "appraise, esteem, value," tatalla, admire, wonder at (the excellence oO.
There were other substantival and adjectival derivatives made from verb-stems that though
frequent and made on fairly regular patterns are not strictly part of conjugation (like Latin
amabilis, amator, -atio etc.) For some of these see below under derivative suffixes.
The following may be mentioned here since they are so frequently formed from verbal
stems that they may be regarded as part of the verbal system.
The noun of general action (aorist in time): corresponding closely to Latin nouns in tio, sio,
or in English to various verbal forms such as growth (to grow); song (in general not particular
use) beside singing; help; appearance; payment; etc.
The most regular, among many competing forms with different suffixes, was that made with
-st� following in strong verbs after the 6mataima: thus: karaste "making, manufacture,
construction"; meneste, intention; nemeste "appearance"; tuluste "arrival"; faraste "hunting, the
chase"; olaste "growth"; keluste "outflow"; liruste "singing, chant."59 An almost equally common
form was -r�, or air� : see for instance in istare "knowledge," ataltare "collapse," kelure
"fountain." This form in causatives and a-verbs (q.v. below) often acted as a gerund, replacing
the gerundial formations in iyi: ortare, lifting up; nifi.qitare "whitening, whitewashing."60 It was
seldom dissimilated to -le(to avoid confusion with the pronominal suffix), but le is found
where a stem ends in single [r] and the sense or form practically excludes a 2nd person object:
as lirule "merry song."61 Also farale "hunting" (efarale, hunts you).
Very frequent also was -ode,as mitta-, insert, mittande 'insertion', infixion. turyande (turya-,
strengthen), strengthening, fortification. nde was more general and less particular than ste.62
� nde or nde could be added to gerund, so karien, kariende.
The verbal adjectives, not used participially: i.e. not capable of taking direct objects, and not
used in the compound expression of verbal notions, were numerous. Here strictly belong all
the pass ive 'participles', which are not properly participles, but these have for comparison with
the active or true participles been noted above, since they are often made in fact from tense
stems. Here may be noted the most important and frequent adjectival forms that are
independent of the tense-system.
General adjectives, expressing (without reference to time or duration) the notion of the verb
as attached attributively (or predicatively) to a noun, are frequently formed with the suffix yi,
added to aorist stem-form.63 These are frequent with strong verbs: as karia, active (in making);
59 The example "meneste, intention" was added in the left margin and the original form alaste emended to olaste,
both in ink. This paragraph replaced a preliminary version, which was deleted in the course of composition:
The commonest suffix was -li (in strong verbs following the 6mataima). It was seldom confusable with the
pronominal suffix le owing to difference of stem and accent. But in such cases as ...
Note that constructio 'construction, building' < (con)struere 'to heap up, arrange, build' is a Latin noun in tio, as is
amatio 'love, intrigue' < amare 'to love', cited above along with amabilis 'lovely, lovable' and amator 'lover'.
60 The phrase "and a-verbs" was added later, written in the left margin in ink. (See footnotes 96 and 101 below.)
61
The original phrase "the sense practically excludes" » "the sense or form practically excludes" in ink.
62 This paragraph and the following note were added later, written hastily in ink in the top margin.
63 The first letter in the original form of the prefix was heavily deleted in ink; it was probably iyi. The end of the
sentence was revised from "added to simplest[?] stem-form" » "added to aorist stem-form," also in ink.
PARMA ELDALAMBERON No. 22 - Page 1 1 0

nemia, apparent. So also istea 'knowing', ortea 'rising, ascendant', taltea 'insecure'. olea,
growing, living (of plants etc.).64 farea "hunting, of prey" as in farea nasto "hunting animal,
beast of prey" [faraya > faraya > farea].* This was often replaced by a similar formation from
the verbal noun: as nemestea 'apparent'. �W But this was only so in the case of intransitive verbs.
In transitive verbs the force of the forms -astea, area, alea was rather "to be, fit to be, proper
for the action" i.e. passive: farastea nasto is "a beast of the chase (that it is proper to hunt)";
lirustea "fit for singing, to be sung."
* yea is usually reduced > ea. sirea "flowing," liquid.65
The general "passive" participle showed the suffix ini : (cf. those associated with tenses): so
karina "made." This form being outside conjugation usually neglects the base-additions: farina
"fugitive," hunted: as in farino "a hunted man, outlaw, fugitive." Other isolated forms show ni
or nasal-infix + i (similar to past tense): as sanga "packed"/or "crowd"; taii.ka, vTAK, fixed,
firm.66
Adjectives in dla (in which the u is possibly related to the u seen in the futures) express like
lihood, aptitude: similar to Latin -iix : nyarula "apt to talk, or relate"; laivula "flexible, pliant."67
Adj ectives in ite or after omataima or last vowel of stem: aite, ite, oite, uite: are less defined
in sense, the actual significance depending much on the verbal sense. An extension: vowel +
maite is also frequent. Examples: karaite, karamaite, "able to make, handy, craftsmanlike,
skilled"; tulumaite "likely to come, (of future events) probable"; him{te "clinging, able to stick
on."
Adj ectives expressing possibility: "able to be done," like -able, ible, are formed with -imi,
always in simple verbs preceded by a stem strengthened, either by doubling of the final base
medial (usual in case of p, t, k), or by lengthening of base vowel, or by nasal intrusion.68 Thus:
mattima "edible"; qettima "utterable," alqettima "unutterable"; alkarima "impossible to make";
kenima 'visible'; alistima "unknowable, secret." Verbs in -a drop this: alfarima, impossible to
pursue. Those in u usually keep it: l iruima "able to be sung." Weak verbs show ilima [from
'alima]: ortalima "able to be raised"; alaniii.kwitalima "that cannot be made white (again)." So
istalima "knowable, ascertainable."69
64 This example was added in ink after the original composition.

65 This note was written in the left margin in the original ink.
66 This paragraph was inserted later in ink.
67 Examples of Latin adjectives in -iix are: piigniix 'pugnacious' < piigniire 'to fight'; audiix 'bold' < audere 'to dare'.
68 In this sentence the original phrase "usually in simple verbs" » "always in simple verbs" in ink.
69 These two paragraphs replaced a preliminary version on a separate page, struck out in ink:
{Here »} Also here, since every verb is theoretically capable of making these forms, may be noted the active
and passive adjectives of "possibility" or "likelihood," which are frequently used in Quenya.
The {active participle »} passive adjective was formed with "dynamic" doubling of the base medial + suffix
'imi; or in case of those verbs in which the base medial was a continuant (other than s, z) in Quenya: v, r, l, m, n
lengthening of the base vowel + imi. Thus kiirima {"able to make {likely, apt to)" »} "able to be made
(feasible)"; {tulima,} kenima, {able to see »} able to be seen, visible; mattima {"able to eat" »} "edible, eatable";
kUvima 'flexible'.
The active adj ective (less frequently used and less defined in sense): was usually formed with basic stem +
6mataima + maite - in some forms + ite: karaite, karamaite, able to make, handy, crafty, skilled; tuluite,
tulumaite, likely to come, probable (of future events); him(te "clinging, able to stick on"; ...
Sometime before these three paragraphs were replaced Tolkien had added a note in red ink at the top of the page,
which he struck out along with them: "From the Aorist stem or from the Base (of strong verbs) were made
various etc. A take in from pp. 9 and 10" (see footnote 42 above).

There was also a perfective adjective made with -nwa which expressed completion of the
action (transitive or intransitive as the case might be).70 This is seen notably in vanwa (/BA, go,
proceed) = "gone for good, departed (dead, lost)." In ordinary verbs it is usually added to the
aorist stem: karinwa "fully made, completed"; tulinwa "arrived, now at hand"; alanwa "fully
grown, adult, mature"; istanwa, known, generally recognized.71
The reduplicating syllable. Cf. above. Reduplication properly consisted of the first consonant +
repeated sundoma (or in the case of derivative verbs the first vowel of the stem: ai, au yielding
a, etc.). In Quenya q, ty for this purpose counted as single consonants: qeqetta- "repeat, keep on
saying." hy, ny, iiw, hw and also qo- were dissimilated > h, n, n, h, k.
Originally in the case of aspirates p, t, k had been substituted. This remained case in early
classical: papharra, repet[itive] of phara "hunt." In ordinary verbal conjugation after ph, th, kh >
f, p, h, they were repeated: fafarra. But the old forms remained in isolated cases as tithilla [v'"THIL]
"twinkle. "72
Where the stem began with a vowel, or [g] that vanished in Q., such reduplicating forms in
original form were impossible, since the like vowels would have contracted and then been
shortened before the doubled medials. Where repetitive forms were required for such verbs
the basic stem was repeated as ululla "keep on pouring"; ololla.
In the Repetitive form the medials were originally doubled.73 In the case of v (< b), r (< d), ' (<
g) the historic forms were thus bb, dd, gg > mb, nd, fig. So kukumba "keep on bowing = to nod
repeatedly."74
The augment was the sund6ma of 'strong' verbs. In case of those verbs beginning with a vowel
since they occurred before a stem already lengthened the augment was usually omitted or
absorbed, as ulle 'poured', ulie, has poured. Such forms as ululie occur only in verse. For weak
verbs see below.
The initial consonant of the stem was usually preserved in its Q. form as far as possible after
the augment, and reduplication.75 Anciently, of course, many initial groups, simplified when
initial, reappeared after the augment and reduplication. So v'"NDAK- : nake, andakie; v'"sTAR: thare,
astarie. The forms with initial nasal before b, d, g were the last to be simplified initially, in
consequence such forms as andakie are preserved in classical Q. not only in augmentation and
reduplication: nandakka, andakie; but also after the closely proclitic subject pronouns: as
me·ndakilti 'we hate them'. [Later and in TQ the simple form is often reintroduced as anakie,
me·nakilti.]
70 The phrase "perfective participle" » "perfective adjective" in the course of composition.

71 This paragraph was written at the end of the page with a narrower pen-nib than the preceding text, probably
sometime after the original composition.
7 2 Following this a sentence was struck through in ink: "Before ps, ts, ks s- was used: sapsarra- 'keep on rubbing'."
73 "In the Repetitive form" was added later to the beginning of this sentence in the original ink.
74 These two paragraphs replaced a preliminary version struck out in ink before completion:
The medials of the repetitive forms were doubled.
Where the word began {(in Quenya)} with a vowel since the reduplicating vowel was always the same as
sund6ma or {stem »} first vowel of stem there was originally contraction. This occurred also after loss of
initial [g]. But such formations only ...
These two paragraphs were written over two notations in pencil: "Case of vocalic beginning"; and "augment."
75 The phrase "and reduplication" was added in the course of composition.

Quenya Verba l System ]. R. R. Tolkien
But other initial simplifications were too old for any traces of the ancient medial forms to
survive in regular conjugation or outside separated and isolated formations. In the case of ps-,
ks-, which historically could be derived from sp- or ps- etc., the reduplicating consonant
chosen was s- : as in sapsarra "keep on rubbing, fray away."76
Initial voiceless [1], [r]: are transcribed hl, hr but there was no h [h or x] present in writing or
in sound in Quenya.77 The usual source of these initials was [sl-, sr-] since Quenya preserved
few traces of stems with initial [phl, th� kh� phr] etc. The historic forms in such a case as vsuK
"creep" would have been: [sizlikk- ], [izlfk-] > sillikk, illrk, but these being obscure in their relation
to the simple forms do not survive. In Q. reduplication is with 1- I r- while the initial hl, hr is
retained medially (being before the main stress): the only case of medial hl, hr. So hlike
'creeps," lilhikke "sneaks about," ihlfkie "has crept."78
� NB. Initial s remains voiceless in such forms and does not become z. sisilla, glitter
(white), isflie "has shone (white) ."
Forms ofweak verbs, or verbs with weak aorist-present.

A verb is "weak" in Quenya when it receives as an addition to the stem before addition of
vowels or affixes of conjugation and so as a stem-formative some suffix: consonantal as t, y, or
vocalic as -U. -a.
There are various forms and types. Notably the following:
1) The "half-strong" type. These are really verbal vTALAT-stems. (a) The commonest type
is that formed from vTAL-bases followed by a consonantal addition: notably -t(cl), -y(cl), (more
rarely k, n, l, r). (b) genuine vTALAT stems: such as talta- 'slip, fall': these usually have (or had
originally) 6mataima but in Q. have a.79
These are called "half-strong" because they usually formed past tenses either from the
unextended stem, or by intrusion of nasal before the last consonant.
2) -a verbs. Several verbs with an inherently continuative sense, such as gal- "grow," spar
"hunt, pursue"; skor "look out for, expect, wait for," did not anciently form true aorists but
were inflected from an infinitive base golii, sparii, skora etc. Hence Q ola-, grow;9! fara-, hunt;
hora, wait for; mena, be making for, on way to.80
91 cf. ol-, oli- regular = become.
76 The original listing ps-, ts-, ks- » ps-, ks- (i.e. ts- struck out) in ink. Tolkien wrote "reduplicating vowel" in this
sentence, but clearly this was a slip.
77 A listing "lh, rh or hl, hr" was replaced by the phrase "are transcribed hl, hr" in the course of composition.
78 The reading lilhikke is certain. These two paragraphs replaced a preliminary version struck out in ink:
But the nasal was lost before p, t, k; the initials {br[?] »} dr, gl, etc. were simplified {similarly[?]}; and sp, st,
sk- became ph, th, kh (or ps, ts, ks), all so early that there is no trace in Q. of the original initials in ordinary
conjugation. A difficulty was caused by voiceless initial lh, rh (from sl, th� kh� etc. and sr, thr, khr, etc.).
Historical forms would have been, say, from v'"suN- : {sislfn-;} sislinn, islln- > slin : sizlinn, izlln > lhin- : sillin, illln;
-
but VKHLIP > hlip : kihlip, ihlip > lhip : kihyip, ihyip.
79 This was revised in the course of composition from: "(b) genuine v'"TALAT stems: such as talta- 'slip, fall': these
must be divided again into (i) those without true aorist-present forms originally distinct from imperfect present:
so talta-; and [(ii)] those with aorists as tarki- 'to trouble'. The latter are often ... " The revision originally ended
with: "but in Q. have i, u, or a"; then subsequently i, u were deleted.
80 In this paragraph the original forms gala and ala- were emended to gola and ola- in ink; and the note associated
with the latter was added, written also in ink in the left margin. The original form skop was revised in ink
(perhaps in the course of composition) to kop; subsequently this form and the original forms kopa and hopa were
revised to skor, skora and hora, respectively, all in pencil.

Quenya Verbal System f. R. R. Tolkien
3) -u verbs. The sense-difference between these and normal verbs is not so clear, since an
additional -u seems often to have been added as a mere formative: to increase the phonetic
content of the stem, as liru- "sing (gaily)"; or to distinguish between bases of identical (or
similar) form with different senses, as kelu- "spring forth (of water)": vKEL, KYEL, run away
especially downwards or to an end, cf. kelya "(it) sends running down = it rains."
But often as in case kelu- above this suffix marks the beginning (more or less sudden) of an
action that may continue later, and was thus anciently and still in Quenya often conjoined with
the prefix ete/et- 'out': so in the ancient ektelu- (< etkelu) 'gush forth'. It may thus sometimes
correspond to the English up not in the completive sense as in eat up but in such expressions as
wake up, stand up, or 'down' in sit down. It is sometimes combined with the formatives t, y etc.
described above (in Q. the forms -lu, -yu, -ru being favoured).81
Thus Q seru- "settle on, sit or lie down, come to rest on"; hamu- "sit down, take a seat"; tolu
"stand up, get up, leave one's seat." (ek)koiru "come to life"; tuilu, bud, open (of flowers and
leaves), etc.
4) (a) causatives with suffixes -ti, -yi, made from all kinds of stems: as ort&, raise, lift; tulta,
cause to come = send hither, kalt&, cause to shine, light up, or kindle (lamp etc.); nifikwitii, make
[nifikwi] pale, white; tafikata, "make fixed, fix, confirm," karan-yii, make red, redden.82
5) Derivatives of various kinds made from adjectival and noun stems, with or without
suffix, but specially with suffixes 'ta, 'ra, 'la : as ampan6ta, build, erect a (large) building, from
ampano 'building, construction, edifice'. Note also especially those in �nta = to become, grow.
nifiqinta- "to grow pale."
1) "Half-strong": Examples (a) ist(a) "know"; nahta-, slay; ort(a)/ory(a) "rise"; siry(a) "flow";
(b) talta- "slip down," palka "to beat flat"; sulpa- "to lap up, drink greedily."
The number of verbs of class (b) is not large; most of this type have become weak with
weak pasts.
Aorist-Present. These verbs had originally one form for aorist and present. Cf. English "I
know" (not "I am knowing"): Infins. ista, orta (orya), sirya, talta, etc. Inflected forms as for a
present imperfect. Gerunds: istie, ortie, sirie; tal tie, palkie, sulpie. [Gerund also siryare, taltare.]83
9I ie is from 'aye > aye > i1e: showing that the model of this formation is old. oryave, siryave
for old sirie (< siryie) are modeled on Cl[ass] 4 q.v.
Participles: [aorist. istea, ortea (oryava), sirea, taltea, etc.84
complete istanwa etc.
passive istina, nahtina, sulpina, palkina etc.]85
8 1 This sentence was revised in ink from: "It is sometimes combined with the formatives t, y etc. described above.
Thus[?] et-koi-tu: Q ekkoitu 'come to life'; tuiyu 'bud, spring, [???]."
82 A second sub-item was deleted in ink here:
(b) Similarly conjugated the words signifying "to become/or be made" made with suffixes -o , s o , ro , yo ,
tyo . [cf. Telerin future in -ro.]
Q nifikwityo "become pale"; karino-, to get finished/made.
83 The preceding list of Gerund forms originally included sirie/siryave, the second variant of which was revised in
ink to siryare; subsequently it was deleted and the alternative examples in brackets were added, also in ink. The v
in the earlier form siryave is no longer visible but this same form was left unemended in the following note. In the
note "Cl. 4" refers to the causatives described in item (4) below (see especially footnote 101).
84 The original form siryava » sirea in ink.
85 The participles originally listed in this line were: "passive istana, ortana, siryana, taltana etc."

Also the imperfect participles (the only ones used in conjugation) :- Active: istalya, ortalya,
siryalya, taltarya, etc. Passive: nahtaina, being killed, sulpaina etc.86
�& A distinct imperfect where required could be made with the imperfect parte. (to be
understood}, as ni taltalya "I am slipping down." This then became inflected as a verb
(like the long perfect): taltalya "to be slipping down," he·taltalyar, they are etc.87
�& Repetitives are not often made from this type of stem. If so they [are] made usually
only from talt-stems.88

Past. This was originally formed "strong" with n-intrusion before the last consonant: as
sirya : sirinye "flowed"; talta : talante.89 Where the stem was of ..fAT type this past could be made
from 'ta-form: as orta : *ronte "rose"; ista : sinte "knew." But few examples - only sinte "knew" is
common - survive; the usual form is oronte 'rose'.
Where the base was verbal a strong form without present affix frequently appeared: so kenya
"see"; pa.t. always kenne. lasta 'hear'; pa.t. lasse. s(re beside sirinye.
The talat-stems tended to be weak: thus taltane beside talante; or sulpane only (not
*sulumpe).'ll
'I Similarly naktane 'slew' beside analogical nahante. Since the stem is historically ..fNDAG
the historic pasts would have been *ndaiige or *nda'ante. 90

The strong forms are conjugated as for strong verbs described above. But the forms in -nye
which should produce such perfects as *isirinie either form them periphrastically ye sirinyella or
as true strong verbs: is(rie. A curious but classical form (belonging to orya- but also associated
with 6rta) is or6rie 'has arisen' beside more regular but rarer orontie.
The weak forms: taltane etc. are conjugated like weak pasts: see below.
�& In TQ except for a few well-known anomalous forms like sinte "knew" the -ane weak
pasts spread to -ta, ya verbs. And to avoid the confusion with the causatives -ya was
preferred for intransitives: so oryane, rose, ortane, raised.
The participles and derived tenses were as described above: sintelya (sintielya) "having
known"; sintienwa "having been known"; sinteva "shall have known"; sint(i)elyane "had known."
In the Future the vowel of tense-ending became u as in strong verbs. The futures were thus
istuva, nahtuva, ortuva (oryuva), siryuva, taltuva, etc.
Their conjugation and derivatives are the same as for Strong verbs.
86 The forms originally listed were: "act[ive]: istalya, ortalya, {silyarya,} siryalya, taltarya, etc. pass[ive]: istanta, etc."
The headings were rewritten as "Active" and "Passive"; and istanta » istaina, which was placed in brackets and
subsequently deleted, to be replaced by nahtaina and sulpaina, all in ink.
8 7 This note was revised in ink from the following original version:
� A distinct imperfect where required could be made with the imperfect parte. and verb to be, as nie[?]
taltalya "I am slipping down." In verse an agglutinated form like the long perfect is found: taltalye "to
be slipping down."
Later the beginning of the note was revised in red ink to: "A distinct imperfect (the so called 'long imperfect')
where required could be made with the imperfect parte . ... " At the end of the note the following was added, also in
red: "On the form in strong verbs see above."
88 This note was revised in ink from:
� Repetitives are not often made from this type of stem. If so they are of the following forms: ororia, keep
on rising (and falling), sis{ria; or with (b) simple reduplication: tatalta.
89 Tolkien originally began the list of examples with a false start: "as kenya 'see' ... " (cf. the next paragraph).
90 This note was added later in ink in the left margin. In the manuscript the asterisk is at the end of nda'ante.

Quenya Verbal System ]. R. R. Tolkien
2) a-verbs: as fara- "hunt." These have aorist-presents conjugated precisely as the present
imperfects of strong verbs, which they resemble except in having a short base-vowel.91
s- Note that the participial formations faralya, faraina are imperfect. [The adjectival
formations in strong verbs such as karia: are {area as in {area nasto "a hunting animal" (se. one
that is so by nature or habit). But -nwa 'perfective' does not tolerate 'imperfective' a. So
farinwa 'hunted down'. Cf. fama, quarry, prey; olinwa, fully grown, adult.]92
Since it was sometimes desirable to distinguish be[tween] "I hunt" (habitual) and "I am (on
this occasion engaged in) hunting," the latter could be expressed by faralya, the "long
imperfect." Thus Orome tauresse fara: "Orome hunts in forests," general statement; Sinar Orome
i-tauresse faralye "today 0. is hunting in the forest." But some verbs make a strong present, so
ola-, is growing: aldar olar sana nc5resse "trees grow in that land"; but i·aldar hinna 6lar "the tree is
still growing. "93
The past tenses of such verbs could either be weak olane, grew, were growing; or strong 6le,
grew, finished growing, grew up, became. The latter forms (where they existed) could make
the usual perfects, (ol)c5lie "it has grown up, it has reached its prime, become."94
The form olaryane was not much used, since it was scarcely distinct in sense from olane.
The future was made from stem a without intrusion of [u] : so farava "will hunt, be hunting";
olava, will go on growing, will grow; hopava "will continue to wait."95
The Gerund was made with r� (instead of iye): so farare, hunting; olare, growing; hopare etc.
From this or related to it adjectival forms -ea, olarea "growing," could be made.96
3) d-verbs. These could form an aorist, and in normal classical Q. did so.97
In the Aorist there appear the infinitives: liru, kelu as if from [ii]; beside lirue, kelue. But all
inflected forms of aorist were made from stem linl-: lirur, lirut; lirun, lirute, lirus, lirulme &c.98
91 This paragraph originally ended with the sentence: ''The Gerunds are fardve, oldve, etc." The forms were later
emended to farare and oldre, and subsequently the sentence was struck through, all in ink.
92 In this note the original example forms faranta » faraina; farea laman » farea nasto; and "farina or fama" » fama.
all in ink; and the example "olinwa. fully grown, adult" was added later, also in ink. The brackets are Tolkien's.
93 This paragraph was revised in ink from the following original:
Since it was sometimes desirable to distinguish be[tween] "I hunt" (habitual) and "I am (on this occasion
engaged in) hunting," the latter could be expressed by ye faralya or faralye. Thus "trees grow {on earth »} in
the land," aldar {alar »} olar i·n6resse; but "the tree is still growing," i·alda hinna {alarye »} olarye.
The new example sentences were written in the bottom margin with an arrow indicating where to insert them;
the gloss of the final example presumably should have been: "the trees are still growing."
94 In this paragraph the original forms alane » olane; alle » ale » 6le; (al)alie » (ol)6lie; and in the following
paragraphs alaryane » olaryane; alane » olane; and aldva » oldva. Here also the glosses "became" and "become"
were later insertions, all of these changes being made in ink.
95 Next to this paragraph in the left margin Tolkien wrote a large asterisk in pencil, and similarly "bare base"
above "stem a "; a above "[u]"; in the right margin: "which[?] was already -uva "; and below the paragraph another
faint phrase in pencil that mentions uva. The overall intention of these various pencil annotations is unclear.
96 This paragraph was revised in ink from the following ("class I" refers to the half-strong verbs in item (1) above):
The Gerund was made with v� (instead of {ye »} iye): so fardve, hunting; aldve, growing; hopdve etc. From
this or related to it adjectival forms -va, alava "growing," could be made. This form was early introduced
from y-verbs of class I: siryave (siryava) "flowing."
97 This item originally began with two paragraphs, most of which were struck through in ink:
3) d-verbs. These could form an aorist, and in normal classical Q. did so, by again adding aorist -i. So
lirue, inf. 'sing'; kelue "issuing, running away" (of water): conjugated like kare.
Archaically and in verse their aorists were the same as kare (kari-), but with vowel -ii/o. So liro "to sing,"
e·lirute (lirus), he sings it.

[The parallel adjectival forms were lirua beside liruya, pass. liruna; but lirinwa.r
The Gerund was liruye, kaluye.
The present imperfect. llrua, ke1ua etc.
The past was weak linlne, kelilne. Q> Strong forms like kelle, ehtelle (ehtelu)* were not strictly
parts of the u-verb conjugation but parallel forms from defective unextended base-verb.
* Similarly: tolle, stood, hamme, sat.100
The future took form: liniva, keluva.
[past future linivane, keluvane.]
4) The causatives. The suffixes ti, yi (etc.) were originally stressed, except before past
suffix ne, and gerundial -re. Hence infin.: orta, karanya; but infl. ortar etc., ortalme etc., ortate,
karanyate etc. Gerund.* ortare, karanyare.101
These forms have only one form as in Class 11 for aorist and present imperfect; but this is not
due to any peculiarity of sense, only to the form which already possessing stem -ii found it
difficult to form a special imperfect.102 This class was therefore the first to use regular[ly] the
periphrastic participial form ye ortalya/ortalye = "is raising."103
The past is made with -n� : hence ortane, tultane, kaltane. Those with a long stem and short
vowel before ta make past tense in -tane : nifiqitane (analog[ical] after are, ava).104
The future shows regularly -ava : ortava, nifiquitava etc.
The perfects are properly ortanie, (a)karanyanie, (i)nifiqitanie. But at all periods ortanelye,
nifikwitanelye is more usual.105
5) Examples.
(a) nifiqinta-, grow pale; altafikanta "to become unfixed."
Infin. nifiqinta, etc. Ger. ninqintie, etc. Imperf. partciple: nifiqintalya, altankantalya. (No
passive.) Past tense: nifiqintane, altafikantane. Perfects are only formed participially:
nifiqintanelye. The futures are nifiqintuva, altafikantuva.
98 When the previous paragraph was struck through the beginning of this one was emended from original: "These
formed the infinitives" » "In the Aorist there appear the infinitives."
99 This sentence was originally: "The parallel adjectival forms were lirua, liru[???]; lirunwa." The second form was
altered (perhaps from lirunta to liruina) before being heavily deleted and replaced by "pass. liruna"; another form
inserted and then heavily deleted, replaced by "beside liruya"; lirunwa » "but lirinwa," all revisions in ink.
100 This note was inserted later, written in ink in the left margin.
101 In this paragraph the gerundial ending was originally -vi; and the example forms were added later (in ink)
originally as: "Gerund. ortave, {karya »} karanyave." The asterisk refers to a footnote written in ink at the bottom
of the page, which was subsequently struck through also in ink:
* Those with long stems as ninqitii- "whiten" used the infin. ninqita also as gerund; and also used a v[erbal]
noun as {niflqitale » ninqitaste[?] »} ninqitare.
102 "Class 11" refers to the a-verbs described in item (2) above.
103 This sentence was added in ink after the original composition. Following the previous sentence there was
originally a one-sentence paragraph: "Aor. adjs. ortava, karanyava." Tolkien later circled this and indicated it
should be moved to follow the first paragraph of the item, then subsequently it was struck though, all in ink.
104 This paragraph was originally followed by the forms of another verb, which were struck through in ink:
Similarly ninqityo, infin.; ger. niflqity6re. Infl. ninqityor. Parte. ninkwityolya. Past niflkwity6ne.
These are forms of the verb ninkwityo 'become pale' cited above but later deleted (see footnote 82).
105 This paragraph was revised in ink from the following original:
The perfects are properly ortanie, (a)karanyanie, ninqitanie, but rhythmic lengthening (aided by anal[ogy] of
ava etc.) is usual in later classical: ortanie, ninqitanie.

b) Other weak verbs are similarly conjugated. So: ampan6ta, inf. "construct (a building)";
ampan6tie, ger.; infl. ampan6tar etc. parts. ampanotalya, ampanotaina. [ampanotalea, architectur
al; ampanotaste, architecture.] past: ampan6tane, perf. ampanotelye. future: ampan6tuva.
Syntactical notes.
The ordinary verb without subject pronouns acts as infinitive in Quenya: when one verb is
the object of another only.
The following are examples of the types of expression in which the infinitive is used:- ni
tele tule "I intend to come." When the subject of the first verb is plural (or dual) the plurality is
not expressed as a rule in the second verb, unless that verb itself has an object. Thus me telir
tule "we mean to come" (only in verse me·telir tulir). But ni tele karite "I mean to make it"; me
telir karilte "we mean to make it."106
Where the second verb also has a subject that subject can be regarded as the object also of
the first verb. When this is a pronoun such expressions as the following are permissible in
Quenya: ni·merite tule, ni·merite karithe, me·merilte tule (not tulir since the action of tule is sg.
belonging to te), me·merilte karithe, me·merilti karilthi : se. "I wish him to come, I wish him to do
it, we wish him to come, we wish him to do it, we wish them to make them (other things)."107
The same sort of expression can be used with nouns: e·meme ataretta tule "he wished his
father to come"; e·kestane ataretta karithe "he asked his father to do it." But in the latter case
(with nouns) a clause was far more usual, and could be used in all cases where the subject of the
second verb was not the same as that of the first. A clause in such cases is introduced by i,
before vowels in. The tense inside the clause depends on that of the first verb: the time of
which becomes the present of the second verb.
I ni·mere i Ttiro tule, in e·tule, in e·karithe
1 me memer " " " "
I wish I Turo to come, him to come, him to do it.
1 or/ that Turo come(s), that he come(s), that he do(es) it.
We wished that Turo come etc.
Other ex[amples] . me memer i Ttiro karathe : we wished that Turo was making it; he·meruvar i
Ttiro karathe : they will wish that T. is (then) making it; le·meme i ni·kamete : you wished that I
had made (lit. made) it.
�&" With verbs of reporting, saying, stating facts (or as facts) the accusative and infin.
construction can not be used. In Quenya it is felt that the relation between first and second
verb is then quite different. ni qetis tule could only* mean "I tell him to come" se. "say to him
to come, bid him come"; not I say that he comes.
[*] It would however be incorrect since qete does not take a direct personal object.
Correct is nHe·qete tule. 108
106 The first verb was revised in ink in all five of the examples in this paragraph; the original examples were: ni
mene tule "I wish to come"; me menir tule "we wish to come" (me·menir tulir); ni mene karite "I wish to make it"; and
me menir karilte "we wish to make it."
107 In these five examples original menite » merite (twice); menilte » merilte (twice); and menilti » merilti, all in ink;
and similarly in the examples below mende » merne (twice); mene » mere; mender » memer (twice); and menuvar »
meruvar. The first gloss here was tentatively elaborated to: "I wish (mean) him to come," but was subsequently
restored to the original reading, also in ink.
108 This note was added in ink in the left margin.

In reporting or stating a clause is introduced by sa or nil. ni qete (sa) e·tule "I say (that) he is
coming.77109 ni qente (sa) e·tule, I said: "he is coming" or/ I said (that) he was coming. ni qente (sa)
e·urulie, I said he had come. ni qente (sa) e tulle, I said that he came, se. I stated that (at some
time previous to my speaking) he came (but was no longer present).
In all cases where the second verb is the subject of the first (by the first meaning the finite
verb irrespective of the actual word order) the infinitive cannot be used. The Gerund is then
used. matie ye mara "eating is (a) good (thing)"; ksarare psare stile (proverbial), "longing frets the
spirit." karie ye m6lome: "making (things) is hard work." tyavie la tyazie (proverb), tasting is not
necessarily liking.110
The Gerund is also usually employed when the second verb though not the subject of the
first is not the direct object; e.g. when it requires substantival declension.
� The infinitive cannot be declined. The gerund cannot take pronominal affixes.
ni·urulie nyariello ve atarella "I have just come from talking with your father."
ni-la nyara pa matie: "I am not talking about eating."
If this Gerund required pronominal objects or subjects, two methods were open to use: (a)
the most usual especially in long or complex expressions: to convert the whole expression into
clause with ha (the subjective form of the objective sa above).
"Old Turo's eating all the bread was a nuisance to us."111
Yara Turo mante ilqa masta ha me·ne umahtale.
Cf. the objective, me·aller fasta sa yara Turo etc.11 2
(b) In less complex cases: the subject of a Gerund could as in English be put into the
genitive: Turo matie masta. Turo's eating bread; etta matie masta. his eating bread.113
The obj ects followed the gerund. If pronominal they could be enclitic (if quite unemphatic)
but NOT agglutinated. etta matie the ye umahta : "his eating it is a nuisance."
� The inflexional forms are never used since these were purely possessive, matietta
would (if anything) mean "a kind of eating, or style of eating, peculiar to him."
The Gerund can sometimes be followed by [two] pronominal objects: indir. : direct. karie
nye·te "making it for me (would be kind)"; ella karie nye te, making it for me (would be kind) of
you. But ha le·nye karite (a clause) is better.114
� Where a noun precedes a gerund it is not the subj ect since the gerund is not a verb. It
can only be a "genitive of loose composition" (see Declension).115 Distinguish the fol
lowing:
109 The phrase "or nil" in the previous sentence was inserted later in ink; and in this example the parentheses
around sa also look like a later revision. In the subsequent examples, however, the parentheses appear to be part
of the original composition, suggesting that the changes here were made before the paragraph was completed.
n o This example was inserted later in ink.
m In this sentence the original noun "meat" was emended in ink to "bread."
11 2 The first clause in this example was originally: me·ldner fasta (meaning 'we were not pleased'); in which laner »
aller in the original ink; but later the reading was changed back to the original, aller » ldner in red ink.
11 3 The second example was revised in ink from the original: "matietta masta (or etta matie), his eating bread."
Tolkien seems to have altered the first part of this to: matie etta masta, then to etta matie masta; at which point the
variant became superfluous, so he heavily deleted all of the original Quenya and rewrote etta matie masta.
n 4 This paragraph was inserted later, written in the top margin with an arrow indicating its intended position.
Part of the word "pronominal" and the word before it were obscured by an inadvertent blot of ink.
11 5 Following this Tolkien originally wrote: "Note the following"; then deleted this and added a sentence: "Instead
of the ordinary genitive the adjectival form can be used"; and later struck it through as well, all in ink.

kirya karie "ship-making," making a ship/ships (se. is a difficult thing etc.). kiryava
karie, the making of a ship.
karie kirya "making a certain ship." karie kiryali "making some ships."
i nero karie kirya, the man's making of a ship.
R' But the -va adjective is objective not subjective. kiryava karie is practically equal to
kirya karie "nautical construction."

The gerund can be object of another verb. It is then general in sense: see above p. 2. nila
tyaze matie makse "I do not like eating meat" (a statement of general character or taste). But
nlla tyaze matite s4 I don't like eating (to eat) it now.
Distinguish.
m6lome nakin(ye) "hard work kills me."
(ye) m6lome nakie·nye, (it is) hard work to kill me.
ha m6lome le·nakin(ye), it is hard work for you to kill me.116 (rather than ella nakie nye
m6lome).
Expression of "Mood": possibility, condition, supposition, wish.

See remarks on Common Eldarin above.
The particles and conjunctions chiefly used in Quenya were qe "if"; ai "maybe, supposing":
the "nearer" particle; and auve, au, "might (have), would (have)": the "remoter" particle. Also
en "in that (future) case."
The difference between the patterns: "if he asks me, I shall come"; "if he asked /were to ask
me, I should come," was not expressed [as] in English, in which the curious use of the past
form with regard to the future is derived from a past optative that Quenya never possessed.
The first simple pattern is expressed as in English: qe e·kestan, ni-tuluva. The aorist is used in
the conditional clause since the aorist has no special time reference and qe alone marks the
clause as hypothetical. The future can, however, in Quenya be used also after qe: qe e·kestuvan,
ni tuluva. The latter is more nearly equivalent to "when he asks me": it is more particular: "if
on some future occasion he asks me, then on that occasion I shall come."
The second, more "remote," pattern is expressed in Quenya (as far as the equivalence goes:
the correspondences are not exact) by ai + qe with the future.
"If he asked me/were to ask me, I should come" aiqe e·kestuvan, ni·tuluva.
The particle en "in that (future) case" can be also inserted: en ni·tuluva.
The conditions with future reference can deal with events regarded as more or less con
temporary (even if the second is conditioned by the first), or those in which the second
conditioned event is markedly subsequent.117 This is not often noted in Quenya since both
events are in the (by nature) hypothetic future, from the point of view of the speaker. But it
can be marked, when desired, by using either the perfect in place of the aorist in the if-clause,
or the composite future perfect. The future perfect is like the future (above) only used after
aiqe where particular events are spoken of with precision.
"If he finishes the boat by Monday, I shall be able to sail on Wednesday" /or "If he
has finished, etc."
11 6 The Quenya example m6lome ha le·nakin(ye) » ha m6lome le·nakin(ye), in ink.

11 7 The end of this sentence was emended in ink from the following: "or those in which the second conditioned
event is markedly (or[?] will be) subsequent."

qe e·karie i kirya aldaryas, ni kauva kiryasta menelyas ; or qe e·karieva i kirya etc.118

Note in general statements where there is no real supposition (although a second event is
conditioned by a previous one), i.e. when "if" is really equivalent to "when, whenever," then in
Quenya qe is seldom used.119 It is possible to say qe (ai) e·tule, (san) inye tule yU; qe (ai) e·tulle, (san)
inye tulle yti, if/when he comes, I come too; if/when he came, I came too. But usually "when,
whenever" are used: {, iqa, illume, iqallume.
(qa ette tule, (san) inye tule
when( ever) he comes, f come.1 2 0
In the past tense the possibility of wish or supposition known to be "unreal" arises; and the
patterns if he asked me, I came; when/if he (had) asked me, I came are naturally differentiated from
if he had asked me, I should have come; I wish he had asked me; would that he had asked me!
The former are, of course, expressed in Quenya qe e·kestanen, ni·tulle; illume e·kestanen /
kestanelyanen, ni·tulle, and similarly in the other languages.121 The "unreal" patterns are in Q.,
which never developed inflexional subjunctives or conditionals, expressed as follows. The
particle ai is used in the conditional clause, usually joined to qe "if," in the latter en, in that
case,* is used with or without ai.1 22
* "future" because the second event is subsequent or future from the point of view of the
first and is in any case subsequent in thought.
It was also possible to add in the conditional clause, immediately after the verbal expression,
the appropriate negative verb (see below). In wishes, or in conditions where the unreality was
specially marked the 'remoter' particle auve, au was substituted for aiqe, ai (and the negative
verb was not used) . We thus get the following possibilities:-
(a) for "if he had asked me, I should have come / made it"
f aiqe e·kestanen, en ni-tulie/karne Ior enai ni rulie/karnet
1 aiqe e·kestanen ela (= he did not)* I en ni rulie/karnet.123
auve e·kestanen, au ni-tulie (= "I might have come").
* This is a form showing no tense since e.g. ela is then an alternative for ela kestane
(or whatever other verb is used).1 2 4
(b) for "would he had asked me !"
auve e·kestanen (ela) ! 9[
Note 'I It was in this pattern that some languages, as Old Noldorin, developed inflexional
"conditionals" by including such particles as ai: as ON ni tiilaiyeta "I should have
done it."
11 8 The Q aldaryas ( < Aldarya 'Trewsday/Tuesday') is a different day from the English "by Monday," but Q menelyas
( < Menelya 'Hevensday/Wednesday') is the same as in the English (cf. The Peoples of Middle-earth, pp. 1 23-4, 129-30).
11 9 The phrase "not used" » "seldom used," in the course of composition.
1 20 The text of this example was changed in the course or composition from: (qa e tule, inye tule » (qa ette tule, inye
tule; and subsequently the optional word san was inserted here, and also in the previous two examples, all in ink.
121 In this sentence the second Quenya example was inserted later and the original verb "is" » "are," both in ink.
1 22 This sentence was revised in ink from origninal: "The particle ai is used in both clauses, in the first usually
joined to qe "if," in the latter en [?only] [?then,] is used with or without ai." The note on the significance of the
particle en was added later, written in ink in the left margin.
12 3 In the course of composing these examples Tolkien added "made it" as an alternative in the English sentence
and kame in the first Quenya example. At the end of both Quenya examples the variant kames was part of the
original composition, but these were subsequently revised to kamet, all changes being made in the original ink.
1 24 This was written as a footnote in the manuscript.

Quenya Verba l System ]. R. R. Tolkien
Quenya developed more precise "tenses": perfect, pluperfect, future perfect, future in past
etc. : in consequence the tense forms in such unreal past statements could be made more exact
if required.
Thus: aiqe e kestanelyanen (ela), en ni-tulie
aiqe e·kestanen (ela), en ni-tuluvane /or uttilievane.125
The compound tenses usually appeared only in one clause, since they were chiefly used to
mark the fact that the supposed action was not only conditional but would have been in time
subsequent. But such expressions as aiqe e·kestanelyanen (ela), en ni tuluvane /or utulievane do
occur.
In the present true "unreal" conditions are rarer. But cf. English: "if I were doing this now
[as I am not], he would be doing it too." This is naturally expressed in Q. as in the past, but
with the present (not aorist) tense: aiqe ni karaza (nila) san ette kara(za) yti.126
Irregular Verbs.
There are a number of verbal elements in Quenya (as in Eldarin generally that are either of
ancient type, of v'"T.A-bases, not of the normal tal-form, or that have by loss of older consonants
in Eldarin [e.g. 3, ii] or in prehistoric Quenya [e.g. g], have assumed this simpler form. They are
often important like the verb 'to be', or as auxiliaries, and modal verbs; but they have several
irregularities, and are usually defective in one or more of the parts found in a full verb.
For instance they seldom use the participial form -O'a. but use rather the older -yi seen in
normal verbs chiefly in "aorist" adjectival formations; and in general they do not distinguish
in function between aorist and imperfect, though the actual forms employed may be of the one
or the other type.
The most important is "BE." The base of this in Q. appears to have the simple form v'"F.,
which is probably a reduction of a primitive *e3e- or *eiie.'ll
'I Because an archaic t form of the pa.t. is found, eiige, from e-ii-3e, or e-ii-iie.127
The present (functionally aorist and imperfect) has the form ea [derived probably from older
e3a. or e3Ja] which in Q. was felt to be the stem of the verb.128
The forms in use are set out below. For their meanings see notes after the conjugation.
"BE." 1. Aorist and Imperfect. Infin. ea; inflected ear, eat. Only in archaic language and t
verse is the dative pronoun object agglutinated: eanye, ean "there is to me," e.g. parka ean, I am
thirsty, classical parka ea nye. The participle is the same: ea "existing, being." The gerund is ie,
or ere.
12 5 In the first example original tuluvane » mlie, in ink; in the second the alternative utulievane was added later,
first written in pencil, and then overwritten in ink; and similarly for the same variant form utulievane in the next
example below.
126 In this example the pronoun ette is a replacement in ink for original e· ; and san replaced another short word
that was heavily deleted and can no longer be read. The brackets in the English example in the previous sentence
are Tolkien's. This whole discussion of the expression of wishes and suppositions known to be "unreal" comprises a
single manuscript sheet; and the original page-numbering shows that it was composed as part of the treatment of
Quendian & Common Eldarin Verbal Structure. There it was later replaced by a briefer summary in two paragraphs
and the original text was incorporated here with emended page numbers (see above p. 98, footnote 24).
12 7 This note was written in ink in the left margin.
12 8 In the margin to the left of this paragraph Tolkien added a note in pencil: "Cf. Ea 'all that is', the World."

2. Past. ne, or, in sense "existed," also eane "was." Perfect ene. The participle is nea, only in
v[erse] tenea.
3. Future va, and in the sense "will exist" (which is not of frequent occurrence) eva.129 A late
classical and TQ form eava also occurs for va or eva. The participle, or future adjective (it is
seldom used verbally) is evea "future," distinguished from tulurya "future, coming" in referring
to remoter time, and to things and persons rather than facts, events, etc.
Other derivatives are easte, being, essence; ere (also used as gerund), existing. enwa, existent,
real, actually found in the primary world, true. enna, n. "a fact." enno (ennie, f[eminine]), a
person.130
The primary sense of this verb was "to exist, to have being, to be found, extant, in the real
world." But it was often weakened to the copula, in statements of identity or predication. This
however in classical Quenya was limited mainly to the past and future.
In the aorist/imperfect in classical Q. ea was not used as a copula, but only where the full
sense of "exists" was present. Even in other tenses in Q. no verb to "be" is expressed where
the context gives sufficient indication of time. Generally also the verb 'be' was omitted in
questions and statements concerning location, or if a verb was required one proper to the
precise circumstances was employed, as "be found, lie, stand, dwell etc." See the list below.
1 2 9 The forms in this sentence were revised in pencil to iiva and euva, but these changes were not carried forward
to the mention of these forms in the next sentence or later in the text.
1 3 0 This description of the forms of the verb "to be" replaced an original version based on a different root:
Irregular, Modal, and Auxiliary Verbs.
The base of the verb to BE is vYE; but it has many peculiarities; notably that (in the spoken or less formal
written language) of becoming enclitic to the subject pronouns, and of actually coalescing with them.
�& Where a pronoun is the subject the verb to be is usually expressed, as in English. But where the subject
is a noun the present of the verb to be is rarely expressed especially in general statements. In fact it might be
said that the "aorist" of the verb to be is nil. {mole »} m6lome mara pold6reain: "hard work (is) good for the
strong-bodied." The subjective [inserted word:] abstract particle ha (see above) often corresponds to "it is" in
English: ha m6lome {a·nake »} a·ndake pold6rear "It is hard work to kill the strong."
YE Present-Aorist. Full form: ye; infl. yer etc. Participle yea, tyelya. Gerund: {ie »} ye. (yeste = being,
existence.)
Coalescent forms: used where neither pronoun nor verb are specially emphasized.
Sg. 1. nie or n{; Za. kie, b. lie ; 3a. ie, b. thie.
Pl. la. mier, b. fiwier; Za. kier, b. lier ; 3a. hier, b. thier.
Du. miet fiwiet kiet liet ( tuyet thuyet).
Past. yene Ior ne. infl. ener, ner etc. participle. enea, nea, tenelya.
Coalescent forms.
Sg. 1. n{ne Za. kene, b. lene. 3a. ene, b. thene.
Pl. la. mener, b. fiwener. Za. kener, b. lener. 3a. hener, b. thener.
Du. menet fiwenet kenet lenet tUnet thunet.
Perfect. {enie »} yenie.
Future. yeva I or va. Parte. {valya, vay ... »} yevalya, vaya.
Coalescent form: n{va, keva, Ieva etc.
�& ye cannot be followed by an infinitive, but requires an adjectival participle. ye karalya, to be making.
A close equivalent of English "I am to make" is seen in ye with allative gerund: nie karienna (I am
for/towards making it), but this is only used for proximate futurity not of obligation or compulsion. I am to
make it/have to can in Q. be expressed: ha inya karite or inya karie·te.
The margins of this page contain some notes witten hastily in ink: at the top "ye = [?by ?near]"; and at the right in
two groupings of forms, sideways to the main text: "do, lo; d... loar, [?yield]; loyen-"; and "ye-, year, yen"; {ye}; {ste}.
The entire page was deleted in pencil. It was followed on the back of the same sheet by the original treatment of
The Negative Verb, which was incorporated as part of the revised text (see below).

Identity was normally expressed by apposition alone. For example: Eldar Alph(rimor: "Elves
are Immortals." Since the normal place of an adjective attributive was in Q. before its noun, an
adjective was marked as predicative by being placed after the noun: Eldar alph(rimar. Such an
adjective was inflected (for number).
Similarly with such adjectives as manima "of what kind," malka "how great?" malkar i·aldar
i·6lar n6ressella? "how great are the trees that grow in your land?" manima i lie? "(and) of what
sort are the people." �& note the distinction between this and malka aldar 6lar etc.'ll
'll A further example: Question: Manar i·ennor/lie i me·kenner? "Who [are or were] the
persons/people that we saw?" Answer: Intin Eldalin = they [are or were] (some) Elves.
Tasse intin nan! There they [are] againP31
Note also the cases of 'simple location': Tasse i·osto! "There's the city." Masse H6tin n6loa?
"Where are the flowers of yesteryear?" In such cases even in past or future "be" can be
omitted, when other indications of time are present. Masse le enta/enar/elloa? "Where (will)
you (be) then (fut.)/tomorrow/next year?" But qciqi me·vcir tasse, we shall all be there; i·ner ne
raiqa ar s( aphastat "the man was angry, but now is in good humour (lit. it pleases him)."
In expressions of direction or origin: whither, whence, the verb to be can be used
instead of go in sense of "be on the way to /or from." Hence it also can be omitted in
such cases. Manna le sinar, whither (are) you (going) this morning? Mallo le, aiano?
whence (come) you, stranger?132
�& The forms eane, eva, tenea are only used in the sense "exist." There is, there's etc. is
usually tasse (yasse) without verbs; but "there is" etc. = there exists is ea, without adverb.
Vahaia n6re ea i a-esta Valinor "Far away (there) is a land called Valinor." 'll':l:
'll 'll or estaina "named" or essenen 'by name'.133
The gerunds ere, ie, are chiefly distinguished by the fact that ie (< eye) is the true gerund and
only used of particular circumstances. Angwion ie uman enwa "The existence of dragons at one
time [is] a fact."134 Whereas ere is really an abstract, and can be used (as ie cannot) in such a
sentence as: ere nairea, ere lunga nairenen, Existence (in this world) is sorrowful, is fraught with
sorrow.
The subject of ere or ie can precede (as if ere was a verb) - which may be parallel to the loose
cpds. with other verbs - kirya·karie "ship-making." Thus instead of as above Angwi ere or angwi
ie uman. Similarly in statements of identity:
aran ie (aran ere) m6lome "being a king is a burden"
Kanya ere kato (or kata tulma) alwara. To be wise afterwards (or - after the event) is
useless.
From such conjunctions (e,* and ere (or are combining with ending -a) developed as abstract
suffixes. aran(e, kingship. flrimcire, mortality (being mortal).
[*] note stress (e.135
1 3 1 This is a footnote in the manuscript. The brackets are all Tolkien's.

13 2 This note was added in the left margin in ink, without an indication of where to insert it.
133 In the last Quenya example and this accompanying note the original forms a-qesta » a-esta; qestaina » estaina;
and qessenen » essenen, all in ink.
134 The brackets in this translation are Tolkien's.
1 35 This footnote, and the stress-marks in {e and aran{e, appear to have been added later, all in ink. The discussion
of the gerunds, written in ink on a slip pinned to the page with the previous text, was the revision of an earlier
version on the other side of the slip. This had been extensively revised in ink before it was deleted in pencil:

�& Where "is, be" refers to location = be placed, found, is to be had; or to stay, remain,
abide, be situated dwell, etc. then various verbs of more precise and particularly applicable
senses can be used, and usually are. These are :-
(i) vKIM, light on, find, come by. akime "one finds." E.g. "where are there trees (to be
found)": masse akime aldar?
(ii) v'"oY-, live = pass one's days. E.g. "For a long while the Elves lived in Valinor, where are
they/do they live now?": andalUme Eldar oianer Valinoresse, masse oiar sf? oia ( < oya) is a
defective verb without an aorist; oiane is the past imperfect. The past/perfect oine can only be
used in sense "lived/ have lived" - se. to the end of their days and so as a euphemism for "then
died, have died."
(iii) vKOY : koita- "live, have life" can only be used of permanent or natural habitat: as atani
koitar endoresse, "men live in Middle-earth"; niksi koitar nenesse "fishes live in water." Of plants
ola "grow" is more usually employed.
(iv) vKHAD : Q har- "dwell, abide, reside (permanently)." Enyasse Manwe hara: "There far
away dwells Manwe [in Valinor] ."
�& Note the use of the imperfect: hara is used always when the "dwelling" of any particular
person or thing is still going on. The aorist would be used in such a sentence as this: "In
summer I live in the hills as a rule, and come down to the plains in the winter." lairesse nihare
to tarassi, yu unta hrivesse landannar.136 Similarly with ham-, sere-, thar- below.
(v) vKHAM : Q ham- "sit" is used often of cities, towns, houses (and also of hills that are not very
high): dwellings, buildings (except high towers) are not said "to stand." Tiristemindon hame*
Ondoresse "Minas Tirith is/stands/is situated in Gondor."
* aorist because M. T. is and was and only could be in Gondor. hama would imply that the
city was once elsewhere.137
When hame is used of persons it refers to " residence" in a built dwelling. Manwe hama
Taniqetildesse implies that on Taniquetil M. has his hall.
�& for KHAM, KHAD, when a house, or home, of persons is meant the old verb m(b)are
(marne) can be used in formal or archaic language.138

(vi) vsED-, Q ser- (properly "to rest, repose") is in Q. frequently used = "stay, tarry, stop, be
for a while, at the moment." Sinan Elessar Aran sera Arkimbelesse "At present King Elessar is at
Rivendell."139
The gerund {are »} ere, ie. {This »} ie differs from {asse[?] »} ere, since the latter is {only »} also used
universally: as in: "Existence (in this world) is fraught with sorrow": {Asse[?] »} Ere lunga nairenen. {are can be
used »} le is only used (and ere can be used) of particular existence in given circumstances as Angwion {are
uman anwa »} ere (ie) uman enwa: "The existence once of dragons (that dragons once existed) is a fact."
From {a use »} this use with endingless genitive in general sense [cf. kirya-karie "ship-making"] {-ire
developed as an abstract suffix »} developed abstract suffixes -�re, ie. {aran-&re »} aran-le > {amare [ ?] »}
aranie "kingship" (the state of being a king ); {flrimare »} flrimere "mortality," the state of being mortal.
{firm ... »} firimare.
I& {This are can »} ere or ie can be used to express identity in which case the subj ect noun or adjective can
precede uninflected.
aran {are m6lome »} ere (ie) m6lome : "kingship, being a king, is a burden"
Kanya are {kata »} kato (kata tulma) alwara : "To be wise after the event/ afterwards is useless."
1 36 In this sentence the original form nuta » unta.
1 37 This note was written in ink in the left margin.
1 38 This note was added later, written in ink in the margin to the left of item (iv).

(vii) ..fTHAR- "stand" is only used - except, of course, with reference to persons or animals
when they are noted especially as "standing" (not sitting or lying) - of mountains, high hills,
towers, pillars etc. Taras Mindolluin thara Ondoresse, Mt. Mindolluin stands in Gondor.140
(vii) ..fKAY, kaita-, to lie is in Q. only used of persons or animals that are lying down (as in
bed, asleep or sick). "Lie" said of land or regions is ..fLAT, be extended, stretch, be situated (of
an area). Masse lante Veleriande, masse lata s(? numen Endorello isse s( Vaia s(ra : "Where did
Beleriand lie, where lies it now? westward of Middle-earth where now Ocean flows."
In Noldorin it replaces LAT- and is used of flat or low ground. So *kaire, espec. used of a
flat isle in a river, opp[osite] of toll-. N caer.141
From ..fLAT distinguish ..fLAD, lie flat, be flat, though the two may be related and are confused
in Noldorin. laden < latrna = flat (and wide). Q lanna or landa, a plain.
The Negative Verb.142
In Quenya when a verb is negatived this is usually expressed by placing the main verb in
the infinitive, the subject pronouns & inflections of number being assumed by the negative:
which thus becomes a "negative verb" signifying "not to be" /or "not to do." Either the infin.
or the negative may assume the pronominal objects. But if the negatived verb is suppressed or
understood the negative must assume these also.143 Thus la can = "is not, does not"; late "does
not do it." (a) late, Don't do it!
The base is LA./ALA.
Present. infin. la. infl. lar etc. Partc. laia. [Imperfect Present t laia ; past impf. laiane.]144
Coalescent forms agglutinated to pronouns:-
Sg. n1la, kela, lela, ela etc. PI. melar etc. Du. melat etc.145
These are often used imperatively, kela! a kela, don't do it, don't be: a kela naikea! "don't be
cruel!" But where the negative is emphasized keld, a keld, don't P46
On the use of the coalescent forms nila, ela etc. as particles in "unreal" conditions, see
above p. 29.147
Past. lane. Part. lanea or lanelya. Perfect: lalie (or lanie).148
Future. lava. Partc. lavalya (tlavea). future in past lavane.149
1 39 In this example original sere » sera, and "is/dwells" » "is," the former probably in the course of composition.
1 40 Above the root form for this item Tolkien later wrote " ToL ?" in pencil.
141 This note was added later in the left margin, at the same time that the phrase "in Q." was inserted in the first
sentence of this item, all in ink.
14 2 To the right of the heading is an annotation in ink: "no, IQ, lala."
1 43 These three sentences were revised in ink from the following original version:
In Quenya when a verb is negatived this is usually expressed by placing the main verb in the uninflected
infinitive, the subject inflections of tense and number being assumed by the negative: which thus becomes a
"negative verb" signifying {"is not ... " »} "not to be" /or "not to do." The infin. assumes the pronominal
objects. But if the {main »} negatived verb is suppressed or understood the negative can assume these also.
1 44 This list of forms was revised from original: "Present. la. infl. lar etc. Parte. { laya »} laia, t larya. [Imperfect
Present { t larye »} t larya.]" In the space above are some tentative additional forms (mostly deleted), all in ink:
{inf. ala.} {Gerund [?alie], not being.} {ala.} aor. &[?] perf. a!-, ala-.
1 45 The singular and plural forms were revised from: "Sg. n(!a, kela, !ela, ela etc. Pl. melar etc."
146 This sentence was inserted later in ink.
1 47 This note was added later in ink, written sideways in the left margin (see above pp. 1 21-2).
14 8 These forms were originally given as: "Past. lane or lane, also alle ? Part. allea (t lanelya). Perfect: lanie or lalie."
149 These forms were originally only: "Future. lava. Partc. lavdya, t lavalya. "

� The tense forms are usually employed only when sense of la is "not to be." In normal
use the tense inflexion is assumed by a following verb: a kela karit "do not make it!" nila karit, I
do not make it (or nilat kare); melalti kara/melar karalti, we are not making them; helalthi menne
kare/helar menne karilthi "they did not mean to make them."150
But in archaic language and frequently in verse the whole apparatus of conjugation could be
assumed by the negative, the actual verb remaining in the aorist or aorist-present infinitive
uninflected. Thus: "We were not going to make them for you," me-le·lavanelti kare.151 This use
was always most common and still survives most often in the case of simple past and future:
helavate mate "they will not eat it"; lelcine tule "you did not come.1115 2
1 50 In this note original phrase "are only normally employed when" » "are usually employed only when"; and the
examples: {kela kare »} a kela karis » a kela karit; nila karis » nila karit; nilas kare » nilat kare; and "helalti menne
kare/helar menne karilti 'they did not need to make them' " » "helalthi menne kare/helar menne karilthi 'they did not
mean to make them'," all revisions in ink.
1 5 1 The original Quenya example me·le·{allavalti »} Ianevalti kare » me·le·{lanevalti »} Iavanelti kare. There was
originally a second example: " 'They will not eat it': helavate mate" (which was included in the next sentence).
1 5 2 On a page numbered to follow the discussion of The Negative Verb there is an assortment of rough notes on
other irregular verbs, the whole of which was crossed out later in ink. Tolkien organized these according to roots
of various types, but otherwise the arrangement here of the text, and most of the punctuation, is editorial.
Irregularities.
KAB-, be able. future kauva (for kavuva).
suB, sink, esp. in water. sumbe. [In the top margin:] sumba, to submerge.
TUB-, to fall low[?], go down, below normal ground-level, esp. to go down (sink, dive) into water. Cf. tumba,
deep, lowlying, tumbo, a deep valley - with hi[gh] sides though often a wide extent as distinct from imbe,
ravine, glen. In N. im(b)lad is a {ravi[ne] »} glen with a long bottom, much longer than its width, that still has
much habitable ground as Imladris Rivendell. Q latimbe. (Latimberista would Imladris{ )} but in fact the Q.
= =
name was Quenyarized from N., Imbelaris.) [In the top margin:] tumba-, to cast down (into a pit[?] etc.[?]).
[Added above the item:] utumno.
VNDO, unu-, undu, down from on high. hence unta, descend, as of sun or of a man from a mountain. (ununte,
came/went down.) [In the left margin:] undo, down, low down.
vsRo, oro, orro, up, rise (from ground). [In the left margin:] orro, av. up in the air, on high.
{Kusu- » KHUSU-.} up from below ground, to ground level. come up, emerge. {kusse[?]} {emerge[?] »} came
up, emerged. {khus.}
EPH-. (especially from water, opp[osite] of "dive") emerge. eph� it emerges. empe. emphe. [Grouped in the
left margin:] ephe, emphe. ethe, enthe. ehe, ehe. enghe; enche. {hache[?]; hacha[?].}
VEK, point, sharp point, thorn. eht� stab. ehtar, spear man. enke, spear point.
EREK, holly. erke, holly.
KHAL-, to erect, lift from ground, (make) stand up.
G]RAD 1 to flatten, lay low/flat, make flat what was standing.
SRAD f
fling, hurl. of something long like arrow, spear, shaft. vRIP. rimpe, hurled.
hurl, cast, send through air, loose from hand but not nec[essarily] fast. KHAT-. hat. through[?] down[?].
climb (as a tree by clinging[?]). RAP-. rampa.
[Further items in the left margin:]
VRASA. stick up. rasse, horn.
VsLIT - LIT, swiftly[?] speed[?]. lhinte, sped.
[Further items in the top margin:]
NORO-, run as of wheels. roll along.
ROKO-, run on foot. rokko, horse. rohta.
run, slip easily[?] as of liquids[?] etc. [run] as of [?big] [?river] SIR!.
go in a line, as a road. RATA. rant� course. [Added above:] usually of animals/or two feet is {SRATA[?]} {ran... }
lenke. lenka[?].

Quendian and Common Eldarin
B. Verb Structure
§11 The verbal system was, as far as the evidence goes, little elaborated in Common Quendian;
though already a few of the characteristics of Eldarin verbs had appeared: e.g. the association
of an n-suffix or nasal infix with past time; the lengthening or 'fortification' of the stem of basic
verbs to indicate duration, with which also an a-suffix or stem was usually associated.
The fuller elaboration of a verbal system belongs to the Eldarin period, though many
features were still unfixed at the time of the divergence of the main branches; for the later
languages, even in their oldest recorded stages, differ considerably: partly because, in addition
to selecting differently among variations of form or function, some of them, such as Telerin,
abandoned much that had once been common. 2 Quenya at once preserves most of the more
ancient features, and has most added to them by later invention or elaboration.
The following are features of the Verbal System that are ancient and belong to the Common
Eldarin period.
§2 Inflexion. Where no subject was expressed, or where the 'action' was strictly imperson
al, as in: "(it) rains, (it) seems," the bare stem of the verb or tense was used without inflexion.
This same form was also the normal form when the subject was expressed by a full noun.
<J[ That is a noun that had a normal independent existence. In Common Eldarin this
probably originally included the 'proper' names of persons, but in accordance with the
strong feeling for personality (discussed under Noun Structure), in the later language
which had or retained personal verbal forms these could be used with all names of
persons.3
This bare tense-stem functioned not only as the stem to which inflexions, mainly pro
nominal affixes, could be added, but also as the 'infinitive' or object of another verb. Cf.
English give in relation to I give or gives, and in such formulas as I can give. But this infinitive, or
undefined tense-stem, could only be used as an object, it could not be declined as a noun, nor
function as a subj ect. This situation, though arrived at by quite a different historical process, is
again like English, I can give (or give it), but giving (or to give) is a mistake; I have no desire to give
or thought ofgiving.
A difference, however, appears between English and Eldarin use in the following cases: (a)
English usually employs what is virtually a prefixed infinitive inflexion to: I wish to give (and I
mean to); (b) and in some cases uses an ing-noun, as e.g. in I avoid giving. In these types of
expression Eldarin always used the bare infinitive: bare, that is, of inflexions defining it in
itself; as will be seen below this infinitive, being genuinely verbal, could receive objective pro
nominal inflexions. The idiom represented above by and I mean to was either dealt with by
complete suppression of the objective infinitive on its second appearance (as in and I will), or
by adding a "neuter" pronoun of reference: and I mean it. In other cases Eldarin used a genuine
1 The section numbers "§1" through "§25" were added later in pencil; and similarly the letters "a)" to "d)" which
head the paragraphs in both §8 and §20.
2 1n this sentence original "Noldorin" » "Telerin" (the latter was written above in pencil as a replacement, though
the former was not actually deleted); and an "X" was written in the margin to the right of this, also in pencil.
3 The original phrase "these were used" » "these could be used" in pencil, with an annotation in the left margin:
"Not so in Galadriel's song," also in pencil. (Cf. Varda ... maryat ... ortane, "Varda ... has uplifted her hands," in lines
9-10 of the song, The Lord ofthe Rings, Vol. I, p. 394.)

Common Eldarin : Verb Structure f. R. R. Tolkien
full verbal noun made with a suffix comparable to the Latin gerund, or to English -ing. [The
suffixes used in the later Eldarin languages for this purpose were various: ije, be, bije, me, mije
were among the most usual and probably oldest.]
On the forms and arrangements of the pronominal affixes or inflexions that can be ascribed
to Eldarin in common: see under Personal Pronouns.
§3 The Verbal Stems.

A primary distinction in Eldarin was between Basic Verbs and Derivative verbs.
Basic verbs were usually (originally probably always) of the monosyllabic KAL-stem-form.
They required no suffix to denote their verbal function, and no extraneous affix intervened
between their stem and the definers of tense.
Derivative verbs were defined by a suffix (such as -ta : the commonest) which was normally
present throughout their conjugation.4 They were derived either from Basic Verbs: as vKAL
"shine," kalta- "cause to shine, kindle"; or from other stems, notably adjectival stems: as tafika
"firm," tafikata "make firm, fix."
There was also an intermediary or mixed class in which a suffix was present in the aorist
and present forms, but was either dropped or treated differently from the true derivative
verbs in other tense-forms. The sense of these verbs was usually not derivative, but the normal
sense of the verbal base. They were usually verbs whose fundamental sense described a state,
or a process or action that could only be exhibited in a period of time: as lie, stand, love, know,
grow.5 But this class was not clearly'H distinguished (in form) from some old "derivative" verbs
made from bases that were not specifically verbal in fundamental senses, but might be classed
as "verbal-adverbial": as v'"oR 'rise/up'; v'"NDU 'sink/down'.
'H Though originally there seems to have been this distinction between the two kinds that
while the 'true verbals' like vis, vMEL, could form, say, a past stem without the present
suffix, the 'adverbials' never abandoned the suffix. Thus CE ista 'know', but inse 'knew';
orta, rise, but (o)ronte 'rose'.6 But already in CE this distinction was blurred and forms
like (i)sinte 'knew' appeared on the one hand, and like ororie 'has arisen' on the other.
§4 The verbal stems ofBasic Verbs.
Already in Common Eldarin the following stems had been developed:-
1. Aorist. To the simple unmodified stem the suffix -i was added: mati, eat, tuli, come, kari,
make, do.
This 'tense' actually defined no tense or time. Its uses most closely resembled those of the
simple "present" of English: 'I eat, I do'. It was thus used of all statements true at all times (as
in proverbs): 'gold glitters'; 'the sun rises in the morning'. It is also used of habitual actions: 'I
walk to my work', 'I sleep badly at night'. It is often employed as an "historic present" in
narrative; in fact this employment is normal in extended narrative. A common narrative
idiom is for the first verb to be in the past tense, and succeeding verbs (with the same subject)
describing subsequent acts or events in a connected series to be in the aorist. This is notably
4 In this sentence the word "normally" was inserted later in pencil.

5 The first example was altered from lie down » lie in the original ink.
6 In the margin to the left of "ista 'know', but inse 'knew'," Tolkien wrote an "X" and to the left of this: "sinte !"; and
in the margin to the right of "orta, rise, but (o)ronte 'rose'," he wrote "ororie, perf[?]," both annotations in pencil.
Note that these two forms are cited in the next sentence, as part of the original composition.

Common Eldarin : Verb Structure J. R. R. Tolkien
the case in swift narrative where the subject (noun or pronoun) is not repeated: as if in English
one wrote: 'he came home, opens the doors, calls, hears no answer, searches all the rooms, etc.'
'I In sentences containing a clear time indicator or adverb the aorist is still
sometimes used. This is for historical reasons only usual in English with future
adverbs: as 'he starts tomorrow'. But the Eldarin languages could also say 'he starts
yesterday'. Neither idiom is frequent in Quenya outside the archaic language of verse.
�& A marked difference between Eldarin and English is found in the case of verbs whose
fundamental sense describes a state, or process of development. Many of these like 'have' (in
sense 'possess' not 'take'), 'love', 'know' in English use the simple present or aorist only, and
avoid the analytic form 'I am loving etc.' Almost the reverse is the case with verbs of this kind
in Eldarin: they usually do not form a normal i-aorist, but make a durative present with tense
vowel I, with or without any other consonantal adjunct. So CE mela 'love', gala 'grow' (of
plants), is-ta 'know'.
§S 2. Present. This is a genuine tense, and can only be used of what is happening now, or of
what is regarded as 'now' in narrative or supposition. This tense also as a rule describes a
fairly continuous action or event or process (that began before the immediate moment and is
likely to go on for some time).
The tense-vowel in this case is i. The stem was in Eldarin constructed in various ways. One
feature was the lengthening or fortification of the base-vowel: thus mata, tiila/taula 'is eating, =
is coming'. 7 Reduplication often, accompanied by lengthening or fortification, or by

'dynamic' lengthening of one of the base-consonants, was another feature.
The latter method did not in Quenya or Sindarin become a part of normal 'conjugation', but
was used rather to produce new independent stems with conjugations of their own.8 Thus
mammata- 'to gobble up', or 'to go on eating, gorge oneself.9 But the "repetitive" form, made
with reduplication; and gemination of the second basic consonant was so frequent in Quenya
as to be virtually part of the system of normal basic verbs: so tutulla- 'to keep on coming (and
going)'. In Telerin reduplication without gemination was the normal way of forming
continuous presents: mamata- 'is eating'.
§6 3. Perfect. This described an action or process that was completed in the immediate past,
but the effects of which are still present. It is not certain whether this tense was completely
differentiated from the next (Past) in Common Eldarin, or there were simply two similar
competing methods of forming a 'past tense', the functions of which were not yet clearly fixed.
The formations usually regarded as 'perfect' were marked by lengthening of the base-vowel
(never by fortification), and addition to the stem of a suffix: je, ije. There was also a different
but allied form with augmented base, with short base vowel and suffix-je. Thus from .fMAT:
matie or amage; from .fTuL: tiilie or utuge. There were also possibly already in Common Eldarin
mixed forms, such as amatie, utiilie (but not mage etc.).
7 In the left margin Tolkien later wrote in pencil: "get rid of a-fortification in tense-system."
8 Original "Noldorin" » "Sindarin" in pencil.
9 This example was moved from its original position following the next sentence (which ended with the phrase
"virtually part of the system of normal basic verbs") where it had been combined with the next example: "Thus
mammata- 'to gobble up', or 'to go on eating, gorge oneself'; tutulla- 'to keep on coming (and going)'." The revision
was indicated in ink.

9[ The augment was probably more frequently used in words where the 'perfect' form
naturally referred rather to the past than to the present: so amatje "has eaten"; but tiilije 'has
come, is here'.
§7 4. Past. The amatje form just described could be used of past time; but the pure
undefined past, referring to an action thought of as over, was usually formed either :-
(a) by nasalizing the aorist: in which case the augment was usual: as amanti, manti 'ate'; or
(b) by the suffix -ne'lf added to the unmodified base: as vMBAR, pa.t. mbame, ambame 'dwelt'.
'lf This suffix was probably, though this is not certain, primitively used only with
intransitive stems. In any case ne, which also functioned alone as a 'base' meaning 'was',
appears to have been in origin a time-adverb = "then, ago": [cf. en in time-function as a
deictic particle.)l0 Thus ne as a suffix was originally distinct from n-infixion. But it is
clear that the two forms were already closely associated in CE. From the beginning n
infixion and n-suffixion was a frequent variation in Eldarin languages, increased by the
tendency towards transposition of consonant combinations (such as tn, dn > nt, nd ). So
that before the separate histories of the branches began n - e, and -ne had become parts
of a unified system, and choice between them depended largely on phonetic con
venience not on significance (mbame but ma-n-te).
In Quenya the past (or aorist-past) was without augment, and used nasal infixion, or
suffix ne (or less frequently lengthening of the base vowel). The perfect was normally
augmented, and as a rule showed lengthened stem with suffix -ie.
§8 5. Future. a) All the Eldarin languages express, or in their older periods expressed, a
simple future inflexionally, but the inflexions and patterns are different in each. It is thus
probable that the development had not been completed at the time of the separation of the
branches. In CE the future could still be expressed evidently by the aorist with defining
adverbs (see above), and some of the habitual collocations were already hardening into fixed
agglutinated groups on the way to becoming inflexional; while there were also probably some
verbal expressions, like English 'will go, am going to say'. Among this competing material the
later languages chose different formulas for their normal patterns.
b) In primitive Sindarin the future was expressed in two ways: (a) by adding tha (> OS tho) to
the aorist stem: as matithani 'I am going to eat', 'the immediate future'; (b) by adding uba to the
bare verbal stem as matubani 'I shall eat', the remoter future.11 The element tha is adverbial,
and meant originally 'then, next'. The pronominal stem the/tha when used in Sindarin with
time-reference pointed forward to the future.1 2 On the element uba see next.
c) In Quenya the adverbial phrases did not harden into inflexional forms, but in older
Quenya the aorist could be used with similar adverbs (en, enya) in a future sense: as AQ en i
matinye 'I shall eat then, soon'.13 The usual formation is with uva, often added to the bare stem:
10 These brackets were added by Tolkien, and he also wrote "X" in the left margin, and a check-mark, all in pencil.
11 Original Noldorin » Sindarin (and similarly in the second following sentence); "ON" » "OS"; matithi5ne »
matithani; and matubone » matubani, all in pencil.
1 2 When Tolkien wrote "Sindarin" to the left of origninal "Noldorin" he also wrote "Telerin" below this, also in
pencil, perhaps as an alternative revision.
13 The original AQ formation en matinye » en i matinye, in pencil (en i being faint and therefore uncertain).

matuvanye 'I shall eat'; but this was used far more constantly than the original CE matubani, and
had become an undefined future without implications of nearness or remoteness.9l14
91 The 'suffix' in this case was -uba, and the u had no connexion with the u-vowel seen
after the stem in a number of old verbs (see below). v'"us was in fact a verbal element
with the sense 'ponder, have in mind', and the form uba is a 'present' formation of the
type mela described above. Thus Prim. N. and Q mat-uba-ni/nje signified 'I intend to eat'.
The stem ub, uv no longer survives in recorded Quenya as a separate verb; but cf. Q
uvie 'pondering, consideration'; ON uba- 'ponder, make up one's mind'.15 In AQ an old
basic past tense (of intransitive form with suffix ne: see above) was still in use to form
analytically the "future in the past": umbeste mate 'he was going to eat'.16 In normal
syntax in Eldarin the subject verb precedes the object verb, but it is not unnatural that
in these early fixed collocations, already virtually tenses before being finally com
pacted, the objective verbal element should be placed first: mat'-uba-rge thus becoming
parallel to mati-nje, mata-nje etc.
d) In Telerin a suffix ro is added to the bare base (usually) or sometimes, especially in verbs
showing a suffix a, u, etc. in the present, to the present tense stem. This ro is originally
adverbial, and appears to derive from *rofio or similar form. Cf. Telerin runga (*rofifia) 'hasten',
Q ron (rongo), rong[a?], soon, ro, the[?] [?next] [???).17 ro had the meaning 'soon, anon'. So T
matroni 'I shall eat'.
§9 The Augment and Reduplication.
As used as parts of organized conjugation in Eldarin the reduplicating syllable normally
was composed of the first consonant of the stem, followed by the repeated (short) sund6ma of
basic verbs, or the vowel of the first syllable of the stem in derivative verbs. In the case of
initial consonant groups the first only was normally repeated, but there was evidently much
variation in Eldarin in this respect.9l
9l Eldarin variations from the norm appear to have been (a) using an unaspirated
consonant before an aspirate: pa-phar; (b) treating m, n, fi (in initial mb etc.), and s in s
groups, as a prefix and repeating the second consonant: as pa-spar, ba-mbar, etc. beside
saspar, mambar. 18
In the case of diphthongs appearing in the stem the first element only was usually repeated:
as ta-taula-. The same rule held good for the augment (below).
� The initial and vocalic changes of the derived languages would have made many of these
forms, if simply phonetically developed from Eldarin, very unlike normal reduplications, often
wholly obscuring the relations of simple and reduplicated forms. The reduplicated syllables
were for the most part re-formed to make them appear normal. Thus in Quenya after the
changes ph > f, and ei > f, a reduplication pheita- 11 pepheita was reformed t o fzta-, fifita. But in
isolated words anomalous forms remained in use in the older Eldarin languages derived from
14 This sentence was revised in pencil from original: "The usual formation is with uva, as in Noldorin added to the
bare stem: matuvanye 'I shall eat'; but this was used far more constantly than the corresponding N matubone, and
had become an undefined future without implications of nearness or remoteness."
1 5 A very faint annotation in pencil to the left of these examples might read: "think over."
1 6 The original form umneste » umbeste, in pencil.
17 These Quenya forms were added in the right margin, written hastily in pencil.
1 8 The phrase "beside saspar, mambar" was added later in pencil.

Eldarin words or patterns that had once been regular. *For the detail of the treatment of
reduplication in the derived languages see the separate histories.
§10 The augment was originally simply the prefixed sundoma of basic verbs or of verbs
with verbal bases followed by a derivative suffix. 'Weak' verbs, especially denominatives, as a
rule did not employ the augment. The precise arrangements differed in the derived languages.
In the case of verbs that began with a vowel the augment naturally was absorbed by it, and
since the majority of verbal stems requiring the augment already had a long vowel this disap
peared: as in vm 'pour': u-iilije > iilie 'has poured'.19 But forms with virtual reduplication ul-iilie
were used in Eldarin also, and survive in the derived languages especially in the verse-idioms,
but also in some forms in regular use as Parmaquesta olOlie 'has become', ororie, has arisen. 20
� This total reduplication was also the form used instead of the normal pattern as
reduplication. Where the vowel was complex (a diphthong) the first element only was
used in augmentation (cf. above) .
� The augment was not accented. In languages that considerably changed initial con
sonants or groups (or treated them very differently from medial groups) there was of course a
constant tendency to transfer the initial form to position after the augment. But the mb, nd, fig
forms were usually preserved in the older Eldarin tongues in augmentation and reduplication:
as Q nak-, perfect andcikie. In Q. however initial groups in 'regular' forms received their initial
treatment after the augment and simple consonants also: as vSTAR : thar : athcirie; vSAL : ascilie
(not azcilie).21
§ 11 Formation of derivative verbal stems.
Beside the derivative basic type with which the preceding account has been chiefly con
cerned there existed already in Common Eldarin a number of other types, made either from
bases of other shape (e.g. the dissyllabic patterns such as vTALAT), or from stems of various
kinds, verbal, substantival, or adverbial.
Not all simple bases were inherently verbal. Such bases as vNDER/NER 'male person' could, of
course, only form derivative verbs. Bases of vTA-form or those of the "invertible" type (see
Base-structure) such as voR/Ro 'up, rise', did not form verbs without further additions. Also,
many bases of simple vKAL-form made verbs of a kind differing from the 'strong' type in
employing a vocalic suffix (a or u) after the base, either in the 'present' stem or throughout
their conjugation: as vKEL : kelu 'flow, well up'; vGAL : gala 'grow', of plants.
§12 The TALAT dissyllabic stems. These, when they formed basic verbs without further
derivative suffix: not a very large class in Eldarin: followed a rather different pattern from KAL
stems. Their 'aorist' was in kalt[a]-form: thus talta 'slip, slide down, collapse'. Their present
was in kalat- form: talata-, is slipping. 22 Their past was in kalat-form with nasal suffix before the
last consonant. Their perfect was originally in aklat-form with suffix ije: that is it had augment,
but suppression of sund6ma in the next syllable. Thus talanti/e 'fell down', atlatie 'has fallen
19 The phrase "already had a long vowel" was inserted in pencil.

2 0 The example "ororie, has arisen," was added later in pencil.
2 1 In the beginning of this sentence original "In Q. also, though" » "In Q however"; and the phrase "initial groups"
was underlined, both emendations being made in pencil.
22 This sentence was inserted in pencil at the same time that original 'present' was emended to 'aorist' in the
previous sentence. In the inserted sentence the form originally written as talat&- was subsequently revised to
talata-. The brackets in the word "kalt[a]-form" in the previous sentence are Tolkien's.

down'. The latter was usually later remodelled to talcitie or atalatie. The future was made from
the talta form, but before uba had the bare stem talt: OS talthathO-, T taltaro-, Q taltuva-.23
§13 Besides a, u formative and derivative suffixes of various kinds existed in Eldarin of
which those most frequently used, and appearing in more than one language, were (a) t,j, (b)
ta,ja, (c) -lu, ru, nu.24
§14 a-verbs. These have already been referred to above in §§ 3, 4, 5.25 The function of [a]
in Eldarin was originally of two distinct kinds, though these early tended to coalesce (phonet
ically), and the process continued in the later languages.
1). The vowel a was used (a) in basic verbs of dissyllabic kalat-form: as talat-, talta- 'slips
down'; peles-, pelsa- 'fences round'; (b) in verbs that show a formative suffix (such as t, j : see
next): is-ta 'knows', orja/orta 'rises'; some of these have normal KAL-bases: as sirja- 'flows'.
In these verbs the a was not present before other suffixes beginning with a vowel: thus
talanti/e, pa.t.; ist-ubii, future. They could like simple basic (or 'strong') verbs form therefore a
distinct continuous present with � : talatci 'is slipping down, istci 'is aware'. 2 6
2). The vowel a, originally accented, which indicated 'duration' of the action or process. It
thus became the 'tense-vowel' of the 'continuous presents' of verbs not inherently 'durative':
as in matci 'is eating'. But it also frequently appeared as an addition to the base of verbs of
inherently durative or static sense: as gala 'grows', mela 'loves'. The latter is probably its older
function before the elaboration of a tense-system. Verbs of this type, or that preserved this
formation (by no means all verbs of static or durative sense show a in the later languages)
could not form a distinct 'continuous present' - except by analytical devices arising later. 2 7
The same is the case with some English verbs like know, love which have only one form
for aorist and present: as 'I know now, all wise men know', contrasted with 'the sun is
rising now, the sun rises every morning'. 28
Examples of a-verbs that probably go back to Eldarin. gala, grow (of plants); pharii, spara,
hunt, pursue; skora 'look out for, watch for, await'; olii, become, come into being, turn into
(another state); melii, love; ndila, love, be devoted to.
§15 Verbs with consonantal suffix.
These affixes appear always to have been either j or the dentals t, n, r, 1, s . Of these the
most frequent, and the only ones to develop special conjugational forms were t, j, s.29 These
were in Eldarin of two quite distinct classes: formative and derivative.
In the formative class the affix was the consonant. The following vowel was -a (except in
the distinct variety with u : see below); but this, as noted above, only appeared in the 'aorist'.
Though in etymology they were formed sometimes from KAL-bases as sirja, sometimes from
ancient reversible bases, verbal as vis/si : ista, or adverbial as ..foR/Ro 'up', they were regarded
as KALAT-bases in conjugation. Hence where a nasal infix was used in conjugation it preceded the
2 3 In this sentence original "ON" » "OS" in pencil.

24 In item (a) the original letter y was emended toj, probably in the course of composition.
2 5 The cross-reference numbers were added later in pencil in a space left for them in the original composition.
26 Tolkien revised the original example talantci » taliitci in pencil; and he added an annotation to the right of istci 'is
aware': "slta ?" with a box drawn in pencil around the original example and the note.
27 In the margin to the left of this sentence Tolkien added an "X" and a hastily written annotation in pencil: "they
could by length of stem! miitii, giilii ! " (the diacritic over the second a in each of these forms might be a breve).
2 8 The last examples were revised in the course of composition from: 'the moon is rising now, the moon rises at .. . '
2 9 The affix consonant s was added here in pencil to the original pair t,j.

affix. Thus voR/Ro: ort-a, or-ja 'rise', pa.t. ronte, oronte, oronje; vsrR : sirja 'flow', pa.t. sirirye; ista,
pa.t. sinte.
To this class belonged also the inceptives formed with s. This s was added direct to the stem
of adjectives, but in the case of verbal stems required an 'extension'. Thus tafikiis- 'become
firm, settle down'; nifikwis-, (begin to) grow pale; but kalas- 'begin to shine, get light'; talat- >
taltas- 'begin to slip down'. These verbs were originally conjugated like TALAT-bases; but later
their aorist and present forms coalesced in -sj a (nifikwisja); but their past retained the form
-nsi/e (nifikwinse). The majority of these verbs surviving in later Eldarin were from adjectival
stems.30
In the derivative class, which were mostly causative in sense, the suffix was ta, ja, and no
other inflexions could intrude between it and the stem (verbal, adverbial, or substantival) to
which it was added. Thus: orta 'raise', pa.t. ortane; tuija 'cause to come, send for, fetch, sum
mon', pa.t. tuijane.
The suffix was probably originally accented in Eldarin, but when later a past form was
derived with suffix ne (see above) this took the accent: orta, ortan�. The reductions of a
> a when unstressed or before two consonants began in Eldarin, and initiated a process
of the conjugational blending of the formative and derivative classes that became
marked in the later languages.
The suffix ta was specially frequent in forming causative verbs from 'adverbial' stems: as
orta-, raise; ndiita 'cause to sink'; or from adjectives: as nifikwi 'white, pale', nifikwita 'make
white, whiten'. A long vowel was shortened before ta: as tafika 'firm, fixed', tafikata 'make firm,
confirm'. Causative verbs from verbal bases more often employed ja: as tuija 'fetch', matja
'feed'. But those with fixed u-affix employed ta: as kelu-ta 'cause to spring forth, start (water)
flowing, tap'.
§16 u-verbs. The u seems to have been in Eldarin a 'phonetic inflexion': that is, it had no
significance for the defining of tense, mode, or aspect, but served to complete the phonetic
content of the word, which, as far as the particular verb went, was incomplete without it. The
u-addition was thus normally (or very frequently) present throughout the conjugation. Thus
u-addition appears most often after the sonants, 1, r, n, m: as tolu, khamu, and especially after
basic sund6mar e, i: as sedu, liru,t siru (beside sirja), kheru, tt kelu, etc. It most often appears in
intransitive verbs as kelu 'well up', but that is not due to the force of the affix: since there are
old examples of transitive verbs such as telu 'roof in'. In many cases the u-verbs have an incep
tive sense, as in: sedu, go to rest, kelu, well forth, begin to flow, tolu, stand up, khamu, sit down; a
similar sense is often seen in the verbs showing affixes lu, ru, nu: as koiru 'come to life'. It is
probable again that the u is not the bearer of this significance which belongs to the consonant
(in lu etc.), but that the frequent use in Eldarin of inceptive lu, etc. tended to colour the senses
attached to u-verbs. In any case u was frequently used as a differentiator in the cases of bases
with widely developed and divergent senses as vTEL 'close, end, complete': teles 'come at rear,
end a line or series'; telu 'cover in, roof over'.31
t sing gaily. tt lord it over, be master of, own.32
30 This paragraph was inserted later in ink, probably at the same time that the original phrase "mostly causative
or inceptive in sense" was emended to "mostly causative in sense" in the first sentence of the next paragraph.
31 A second example telek 'stop (up)' was emended in pencil to telet 'stop (up)', and later deleted, also in pencil.
32 These two glosses were added later, written in red ink in the left margin.

The suffixes lu, ru, nu, mainly used after bases ending in sonants, and especially after final j
of KAL-bases, had always an inceptive force: as in thillu, thilnu 'shine out, appear (of stars etc.);
vKOJ : et-koiru/lu 'come to life'; vKUJ : et-kuinu 'awake, wake up'; vKAJ : kainu 'lie down'; vTUJ : tuilu
'open (of buds, flowers)'. Contrast: koitcf 'bring to life', kuitcl 'waken, rouse'; and kait-a 'lie, be on
the ground'.33
§17 Prefixes.
Verbs made by manipulation of the base: such reduplicated forms as ¥MAT : mammata 'gobble
up, devour'; vTHIL, vsiL : tithilla, sisilla, 'twinkle, glisten', already noted in §s, have their own
'weak' conjugation.
But already in Eldarin certain adverbial/prepositional elements had become more or less
fixed as prefixes before the stem-forms of simple or basic verbs: so et-tul- 'come out, issue'.
Certain of these combinations had become so firmly agglutinated in Eldarin as to become
inseparable: thus et-kel 'spring out (of water)' > etkel, > ektel. As a rule the conjugation of these
compound or prefixed verbs is the same as that of the simple form: as ettuli- 'comes out', ettilla
'is coming out'. In such cases where the elements were still analysed the augment intervened:
as etutillie 'has come out'. But in some of the older formations this was not so. Thus ekteli,
(water) spurts out; mati, eats; aumata- 'eats away, corrodes'.34
§18 Adjectival and Substantival forms ofVerbal stems.

These in the form of 'participles', 'infinitives', or 'gerunds' were not organized as parts of
conjugation in Eldarin. A great number of different suffixes with varied forms and very varied
functions are found in the later languages, notably in Quenya. The following may be mentioned
here, as being found in more than one derivative language, and probably already in frequent
use in Common Eldarin.
Adjectival endings : 1. -la, -ja, or combined -lja. In general these merely expressed in
adjectival form the sense of the verbal stem, without modifying it. They were therefore
specially suitable for the formation of active participles, and when added to tense-stems
(possibly already in Eldarin) were largely used for that purpose in the later languages.
2. -na, -nwa. Whether this is ancient or due to the influence of the n as infix and in ne,
these suffixes, as applied to verbs had usually an implication of past or complete process. They
were thus specially used to form passive participles; though the fact that they could be applied
anciently to intransitive stems shows that [they] were not in origin specifically passive. So
kalina, bright.35 ni was anciently added to the bare stem, or to the aorist stem (whence came
the very frequent ending ina). Owing to the many consonant changes caused by the addition
of na to other consonants the older forms, without intervening vowel, did not survive in any
recorded language as part of regular conjugation; but they all preserve many isolated adjec
tival forms of this form: as matna 'eaten': Q matta 'food'; quetta, word.36 nwa was added to a
33 The original gloss 'lie, be prone on the ground' was emended to 'lie, be on the ground' in ink.
34 The original stem-form omata- » aumata- in pencil. Tolkien does not say what the augmented forms of the
stems ekteli and aumata- would be. (The forms amatje 'has eaten' and amatie are given above in §3.)
:JS This example was added in pencil in the margin to the left of the previous sentence, without an indication of
where to insert it.

3 6 The original Quenya example manna » matta in pencil; and at the same time the example "qetta, word" was
added in margin to the right of it.

tense-vowel, or 6mataima, except in the case of TA.-stems: as banwa, gone, departed, Q vanwa.37
na was aorist (past), nwa was (present) perfect. As with the aorist-past tense beside suffix -na
was found i with nasal infixion. Though already an Eldarin variation this was probably not an
original one, but due partly to the analogy of the variation ne/n-infix in the past tense, partly
to the tendency in Eldarin towards phonetic transposition of medial groups stop + sonant.
3. Three suffixes of more complex form seem also to have existed in Eldarin: ula, iti, and
imcl/tima.
Uli usually expressed likelihood or aptitude (in what would be the subject of the verb):
matula 'edacious'.
iti, or combined with tense-vowels and other affixes generating more complex forms aiti, fti,
uiti, etc., was not clearly defined in sense, but implied a habit or special association with the
verbal action: khimfti : vKHIM 'adhere': 'clinging, sticking, persistent, chronic'.
imi and other elongated forms expressed possibility: karima, kartima 'able to be done,
feasible'; mat5tima 'edible'. But in intransitive verbs -ima meant much [the] same as as ul[a],
calima, [?likely] to shine, bright[?] .38 [In Q. ima was usually added either to a lengthened stem:
karima; or took the form tima: mastima 'edible'.]
§20 Noun endings.3 9
These were very numerous, and since the verbal bases were also the bases of whole families
of derivatives, their discussion belongs properly under Word-structure in general. Those that
had specially verbal associations, and became (or tended to become) parts of the regular series
of conjugational forms made by normal verbs were the following :-
a) ije : the commonest form of the 'gerund'. It was distinguished from the similar tense
suffix by being added to the bare stem, which showed neither lengthening nor nasal-infixion:
karije 'doing, making'. In origin it was in fact probably the vowel e (most frequent of all end
ings in Eldarin nouns that referred to notions and abstractions not to concrete objects) added
to the aorist tense-vowel. Other suffixes indicating the verbal action or process were :-
b) ta, and strengthened -sti : as Q nemesta 'appearance, seeming'; tulusta 'advent, arrival'.
In Tel. a form -ta was common.40
c) ndi, nde. This was added to base with 6mataima in simple verbs: karandi 'making'; or to
stem with suffix in the case of derivative verbs: as oryande 'arising', tulyande 'fetching'; kelunde
'welling forth'. This form originally denoted a more considerable and durable process than sti,
but this distinction was not always observed. ndi/e could also be added to the ye-gerund: in
this case the sense was durative: karijendi 'a process of manufacture'.
37 In the margin to the left of this sentence Tolkien wrote "wi ?" in pencil, but later struck this out.
38 This sentence was hastily written in pencil at the left. The brackets around the next sentence are Tolkien's.
39 There is no section numbered § 1 9 in the manuscript.
4 0 This item was revised in pencil from the following original version:
ti', and strengthened -sti' (beside te, ste). These were evidently very frequently employed in Eldarin; but in
such a language as Quenya, where ti/sti became the same in form as the pronominal affixes t'i/sti, they did
not become part of normal conjugation, and survived either in isolated forms derived from ti added direct to
a base or in the elongated form in which sti was added to base followed by 6mataima: as Q nemeste 'appearance,
seeming'; tuluste 'advent, arrival'.
Tolkien also wrote an " X " beside this paragraph in the left margin with a note: "no! conflicts with pronominal
suffixes." Possibly connected with this are two forms written in pencil in the top margin: (i)kii, matikii.

d) -le, re, me and extended ale, ire, ume. These like ije above are elaborations of the
'abstract' ending e. They were widely used in Eldarin, but seldom became regular conjuga
tional elements. In Q. re, are (or dissimilated after r > le, ale) was so frequent as to be virtually
formable at will from any verbal stem. Q" NB. These endings may disregard verbal formative
or tense-endings. So Q farale 'hunting' < phdr-ale: farale, you hunt < phar≤ lirale 'merry
singing': lirule 'you sing'.41 In Q. me was usually added direct to bare stem, and was specially
favoured after r, l. As a verbal derivative it was most frequent in compounds: e.g. essekarme
'name-making'. 42
§21 Moods. The modifications of sense, and implication corresponding to the 'moods':
subjunctive, injunctive, optative, etc.: of modern European languages were not in Eldarin
languages expressed inflexionally in the common period. The normal or 'indicative' forms
indeed were used without modification, where the context made it clear that a sentence was
not narrative or statement of fact. Or indeed even when the whole tone and trend of the
narrative, situation, or conversation would not have made it clear to Men.43 To the Eldar an
entire narrative of invention (about things or events not supposed or known in Ea) needed a
"suppositional" form or indicator in its verbs as much, or as little, as the expression of a wish,
or the definition of a possible event contingent on if, or an if-clause, or indeed of a 'statement'
in the 'future' - other than prophecy.
There existed, however, already in Eldarin certain adverbial particles, of possibility, re
moter possibility, supposition, and wish, that could be used for greater precision. These were
placed in close connexion with [the] verb. Most frequently they appear to have immediately
preceded it. But they could follow it; and in cases where (while the primitive collocations of
pronominal affixes and adverbial adjuncts had not yet been agglutinated into inseparable
inflexional forms) they preceded the pronouns the germs of modal inflexion were already to
hand. (These only actually developed in the Noldorin dialects.)44
The difference between e.g. "if he asks me, I shall come," and "if he asked/were to ask me, I
should come" is rarely expressed in Eldarin: both refer in fact to a possible future. The former
is in all Eldarin languages expressed as in English: with the aorist in the if-clause, and the future
in the second clause. The aorist is used because the suppositional or 'future' character of verb
is sufficiently indicated by if. If, however, the future is put also into the if-clause the hypo
thetical nature of the whole expression is emphasized. It might be further emphasized by a
particle. But the second clause remained the same in all cases. Thus in Quenya: [?it/fte]
kestallen, tuluvanye = "if you ask me, I shall come"; [?it/fte] kestuvallen, tuluvanye = (more or less)
"If you were to ask me I should come"; aite[?] kestuvallen etc. would mean: "now supposing you
asked me, a thing very unlikely or ridiculous to suppose" - and such an if-clause would most
usually be followed by 'I should be at a loss, should marvel, should not believe it, etc.'45
4 1 Tolkien wrote an "X" in pencil in the margin to the right of this sentence, and a line that runs partially through
the 2nd person inflected forms, perhaps indicating that they were no longer valid.
4 2 The last two sentences were written with a narrower pen-nib. The final original sentence was later struck
through in pencil: "It corresponded closely as the name of an action to ma indicating the result, as karma, deed."
43 This sentence was emended in pencil from the original: "Or indeed when the whole tone and trend of the
narrative, situation, or conversation did not make it clear."
441 Tolkien wrote an "X" in pencil in the margin to the left of the end of this parenthetical sentence.
45 The original version of this sentence and the accompanying note were as follows:
Thus in Quenya: que kestallen, tuluvanye = "if you ask me, I shall come"; que kestuvallen, tuluvanye = (more or
less) "If you were to ask me I should come"; aique kestuvallen etc. would mean: "now supposing you asked me,

� aite[?] is naturally used with the 'future'. But it can be used with the aorist or
present: in cases where an immediate supposition is being made for argument or jest,
etc., aite[?] kestallen 'now, suppose you ask me'. In colloquial TQ this is sometimes used
as a polite imperative or request formula.
§22 In the past tense the possibility arises of wishes or suppositions known to be "unreal."
For the indication of the requisite tones Common Eldarin possessed such particles as ai
'supposing'; be, aube expressing the fact that a statement is not, or was not, so.46 But the
syntax of conditional clauses and wishes differs in the derived languages. For details see the
separate accounts. It is plain that in Eldarin the normal verb-forms were often used: se. "if he
asked, I came" for "if he had asked, I should have come."47 And that where precision was
required this was obtained analytically as commonly in modern English. For example by the
equivalents of such expressions as these: "Supposing/suppose/let us assume he went, then I
came"; "suppose he came (he did not), then I came." In Quenya the equivalents of 'he did not,
it did not' became crystallized into regular particles placed at the end of an if-clause to
indicate its unreality. See The Negative (below) .48
§23 Imperative forms.
Eldarin had no special imperative verbal forms. But it had some special imperative
arrangements of word order. Command, or request, could of course be expressed by any
aorist, present, or future verbal form with appropriate tones. But where the second person was
concerned the real curt imperative of command was expressed by the aorist form without any
pronominal affix at all: mati, eat! queti, say! orja, rise, get up, up! In conjunction with this the
particle a (common in all Eldarin as a brief call to attention) could be used.
9[ In Q. when this was added with suppression of aorist-present vowel the expression
was very imperious or urgent: tulQ, come at once! orya, get up. Preceding the verb it
a thing very unlikely or ridiculous to suppose" - and such an if-clause would most usually be followed by 'I
should be at a loss, should marvel, should not believe it, etc.'
� aiqe is naturally used with the 'future'. But it can be used with the aorist or present: in cases where an
immediate supposition is being made for argument or jest, etc., aique kestallen 'now, suppose you ask me'.
In colloquial TQ this is sometimes used as a polite imperative or request formula.
Tolkien made various changes and annotations to this in pencil: he crossed out the words que in each of the first
two Quenya examples, and wrote in the margin to their left: "swe > hwe ? or[?] qui, hwl[?]" (with qui underlined),
but subsequently struck this out; to the right of the word aique in the third example he wrote aiqui, but later
struck through both forms; in the bottom margin he wrote: "ai asks an unlikely question," along with the form
aite[?] underlined; he replaced both aiqe and aique in the accompanying note with the same form aite{?]; and in the
left margin he wrote the forms it[?], lte[?], and repeated them above and to the right of the first Quenya sentence.
46 Tolkien emended original au, ve, aube » au, be, aube in ink; and later struck out the form au and wrote "X" in
the margin to the left, both in pencil.
47 The examples were revised in pencil from: "if he went, I came"; and "if he had gone, I should have come.''
4 8 This section replaced an earlier version in the course of composition:
In the past tense the possibility arises of wishes or suppositions known to be, or regarded as, "unreal." Here
particles were often used in both clauses (if at all). Those that can be referred to Common Eldarin are: ai of
possibility or hypothesis; and au of wishes or suppositions known to be 'unreal': these give tones to the verb
similar to those imparted by the English auxiliaries 'may, might' respectively:
Q {(au)que »} (ai)que kestanellen, (ai) tullenye 'if you had asked me, I should have come'
au kestanellen! "If only you had asked me!"
ai tuluvanye; ai tulinye sinar (entan), I may come; I may come today (tomorrow)
au tuluvanye: "I wish/would I were coming.''
Tolkien deleted this with a grid of horizontal and vertical lines in ink.

Common Eldarin : Verb Structure ]. R. R. Tolkien
was less imperious. Thus there was a gradation: a tulel do come!, tule! come, tul&, come
at once.
It is also a special feature of Eldarin imperative expressions that the subject pronoun usually
preceded the verb, and if it followed was separated and in full form and did not become
inflexional. It was, however, sometimes attached to the particle a, especially in other persons
than the second.
Thus, le tuli, tuli le, ale tuli!
se tuli, tuli se, ase tuli!
were probably CE expressions for "come (you)!, now you, come! let him come! etc.49
�& Where the pronoun, expressed or unexpressed, was plural (or dual) the appropriate
number signs were affixed to the verb: le tulir, come ye!

§24 Negation. The negative base in Eldarin was invertible vAL : al/la. As a particle with
verbs the form was la, or la. To judge by the developments in the derived tongues: when la
preceded the verb the subjective pronouns also preceded it and became attached to la, forming
a negative verb = "not to do/be."
This la became fully inflected, when it meant 'not to be', having tense forms of which lane
'was not', (a)lalije 'has not been' are probably to be referred to CE. Similarly when no following
verb was expressed.
When la was followed by a verb then the verbal stem assumed the tense-signs and the
objective pronominal affixes: thus CE lani/lanje kame-sa "I did not do it." But in answer to, say, a
question "Who did it?": CE lane-nje(-sa) "I didn't (do it)."
The imperative was la or ali = "don't."
The verbal expression la-sa, lclsil, la-ssa "not it = it is/was not so," being frequently appended
to unreal wishes or propositions, gave rise to particles such as Q las, [laza] used in conditional
clauses or expressions of wishes: see above. 5° Q au ttilielde [laza] (las)! = "if only you had come
(were here)," or quflas tiileste san inye tiile.51
§25 The pronominal affixes. For these see the description of the Common Eldarin pro
nominal stems. The development of special forms for conjunctive use with verbs had already
developed virtual inflexions before the separation of the branches.
49 The original example iise (iiste) tuli! » iise tuli! in pencil.

50 The original form la-sii » la-sa in pencil; and the brackets around laza were added by Tolkien, also in pencil,
here and in the following Quenya example.
5 1 The second example given here was added later in pencil. The s in quflas is uncertain.

Late Notes on Verb Structure
In 1970 Tolkien began to make revisions to the Common Eldarin Morphology, but these extended
only to the first few pages of the treatment of Noun Structure. Shortly before and probably as
a preliminary to this projected revision he compiled various notes on Eldarin and Quenya
grammar, ranging from rough jottings in ball-point to carefully composed texts written with a
nib-pen and a few typescripts with revisions in fountain-pen or ball-point. We present here
fourteen of these documents that show Tolkien's later conceptions of the tense stems of the
basic verb; the formation of derivative verb stems; verbal nouns and adjectives; irregular
verbs, such as those meaning 'exist', 'can' and 'have'; and expressions of negation.
The evidence for the dating of these documents is mostly indirect, but they all appear to
have been composed around 1969. We designate these texts LVS l through LVS 14 and present
them in their order of composition insofar as that can be determined. The following is a list
with Tolkien's title or heading or a brief description of the topics in each of them:
LVS 1 Some notes for re-ordering and fixing of Quenya. [/EIJE 'exist'.] 147.
LVS 2 Irregular verbs. [IMAGA 'have, possess'. ista.] 148.
LVS 3 Record of Quenya etc. inventions. [v'kiL 'choose'. v'"TEIJ 'indicate'.] 149.
LVS 4 Eldarin. [v'"KuRu 'can, have power'. v'"NAYA 'can, have opportunity'.] 151.
LVS 5 [v'"Aw 'have'. Past tense -ne. ii- negative prefix. Basic tenses.] 151.
LVS 6 Notes on verb and treatment of -w- in Quenya and Sindarin. 152.
LVS 7 Quenya Verbal adjectives - AL(A) negative etc. 153.
LVS 8 [(A)LA negative. ii = 'bad, hard, difficult, unfavourable'.] 156.
LVS 9 The denominative or adjectival verbs, causative & inceptive. 157.
LVS 10 Verbs. [Q. past continuous. ya verbs. -t formative.] 157.
LVS l l Vocabulary notes. [v'"sAB 'believe'. v'"sAN 'think'. KWI 'suppose'.] 158.
LVS 12 Gilraen. [- ta formative, -ta causative.] 159.
LVS 13 Question of the expression of Negation. 160.
LVS 14 WilL shall and the 'future'. 165.
Tolkien left the linguistic documentation of his Eldarin languages in seven box-files, and
Christopher Tolkien has subsequently referred to these as 'Quenya A' through 'Quenya G' for
purposes of indentification. These divisions have no larger significance as such, since Tolkien
left the manuscripts ordered variously by chronology or topic, depending on the circum
stances. For example the 'Gnomish Lexicon', the 'Qenya Lexicon', the late 1920s English-Qenya
dictionary, and The Etymologies were all placed in the box-file later designated 'Quenya E'; while
on the other hand the earlier version of the Tengwesta Qenderinwa, the "Comparative Tables,"
the "Vowels" section of the Outline of Phonetic Development, and the "Declension of Nouns," all
dating from the 1930s, were placed in box-file 'Quenya D'. Within each box-file Tolkien
subdivided the materials in various ways, using pieces of cardboard, folders, or folded papers,
and he often wrote descriptions of the contents on these subdividers, indicating that he was
working on the documents within a particular subdivision - composing, consulting or revising
them - at more or less the same time.
Four of these subdivisions concern us here, three in box-file 'Quenya C' and one in 'Quenya
B'. The first is a brown folder on which Tolkien wrote "Eldarin Linguistics," which was itself
subdivided into three bundles contained within folded papers followed by loose pages divided
into two sections by a piece of cardboard. Two of the folded papers contain notes on Quenya

Late Notes on Verb Structure J. R. R. Tolkien
phonology, and the third contains "Sindarin dialect notes," two of which were dated in 1962.1
Documents LVS l to LVS 4 were placed between these and the cardboard, and following it are
multiple versions of a description of Common Eldarin initial consonants in the midst of which
LVS 5 and LVS 6 were placed. LVS 1 to LVS 3 were composed on Allen and Unwin waste paper
(printed notices dated "7th April 1967"), but the first of these bears a date written by Tolkien
in the upper right corner: "Ap. 6, '69."
Tolkien's use of different writing implements, nib-pen for LVS l, LVS 4 and LVS S, ball-point
for LVS 2 (and part of LVS l), fountain-pen for LVS 3, and pencil for LVS 6, suggests different
occasions for the composition of these documents; but the texts also show interconnections
that suggest he was working on them around the same time. Thus the rejection of the
tentative rough conjugations of the base ..fMAGA 'have, possess' in LVS 2 may be connected with
the new conception in LVS S of ..fAw as a base "for 'have' = possess, own, keep (have in hand, use
or with one)." And the rough conjugations of nii 'is' and ea 'exists' derived from ..fEIJE in LVS 2
are presumably connected with the notes on ..fEIJE and ..fNA in LVS 1. The former includes the
form nai 'let it be that' and consideration of this may have led Tolkien to the new conception in
LVS 4 of a base ..fNAYA 'can' in the sense "have opportunity, chance, or permission," from which
is derived nai 'it may be, there is a chance or possibility, maybe'. There is also a semantic
connection between ..fKIL 'choose, select', which emerged in the composition of LVS 3 and the
base given as ..fKJABA 'taste' in LVS S, but as KJAW 'taste, select, choose' in LVS 6.
The second subdivision is a folder with a label "Q. Linguistics in process," a date "1969," and
a note listing its contents: "ala - la ; verbal adjs.; pronouns of person; reflex., gender." This was
further divided into seven sections using pieces of cardboard, and in one case an envelope
postmarked May 1968. The only indication of a more precise time in 1969 is a letter from
Anthony D. Howlett, dated "19th May 1969," asking for the Sindarin name of Rivendell written
in Elvish script. Tolkien wrote a draft for his reply on the back of the letter and placed it along
with notes on the derivation of various other names in the fourth section of the folder. In the
third section he placed a typescript and associated notes, presented here as LVS 7. Starting
with the flower-name alfirin, this describes the grammar of the negative element ALA. The
formation of verbal infinitives, participles and adjectives is also described, since the expression
of negation depends in part on the function of the word being negated, as in la karita i hamil
mara alasaila (na) 'not to do (in this case) what you judge good (would be) unwise'. Tolkien
composed as he typed, obliterating rejected text with series of typed m's. He later revised and
annotated the typescript in ball-point and this extended to the slips placed with it, one of
which is a narrow piece of paper measuring 2 � by 8 � inches. The final section of the folder
includes a dozen similarly narrow slips containing an assortment of rough notes, two of which
are presented here. LVS 8 has notes on negative (A)LA in contrast with ii 'bad, difficult', which
is connected with LVS S and LVS 7. LVS 9 is a note on the causative and inceptive verbs derived
from adjectives, connected with the description of these derivative suffixes in LVS 12.
The third subdivision is another brown folder with miscellaneous notes on "Vocabulary,"
describing the derivatives of bases or the etymologies of existing words and names. The notes
appear to be sorted primarily by the size of the paper used: there are eleven sheets of Allen
and Unwin waste paper (notices dated "12th jan. 1968"); a dozen narrow slips similar to those
in the second folder mentioned above; three long sheets folded in half; a quarter sheet; and
eight half sheets. Tolkien dated one document in the folder, a derivation of Sindarin mae 'well'
1 cf. "Words, Phrases and Passages in The Lord ofthe Rings," ed. Christopher Gilson, PE 17, pp. 131-5.

from the base ..fMAG 'good', writing "Aug. 23 '69" in the upper left corner. In several of the
notes in these documents the etymological points seem to arise from or lead to consideration
of other features of Tolkien's conception, and this includes two (LVS 10 and LVS 12) that
describe features of different types of verb stems, and one (LVS 1 1) that describes the particles
used for suppositional moods. Again different writing implements were used, ball-point for
LVS 10 and LVS 12, and nib-pen and pencil for LVS 11; but LVS 12 was written on the back of a
letter to Tolkien dated june 5, 1969, which offers some further support for the likelihood that
the documents in this folder were all compiled sometime around the summer of 1969. 2
LVS 10 bears the heading Verbs and was composed on both sides of one of the pieces of Alien
and Unwin paper. It begins with the proposal of a new conception: "Make a Q. past
continuous," a tense form with a meaning such as 'was eating'. Tolkien first suggested the
form matatule, but immediately replaced this with matante, which he supports in part by a
parallel with the Sindarin past ending -ant. Further rough notes seem to suggest other possible
ways of expressing continuous action in the past, but ultimately Tolkien proposes a different
verb stem matane > matane, which he says is better. This is the formation he had previously
described as past imperfect in QVS, derived by applying the "weak" past ending -ne to the
present continuous or imperfect stem, as in karane 'was making', ttilane 'was coming'. The
second side of LVS 10 begins: "ya verbs must all be intransitive, and be ya," and the examples
siryane and oryane are cited. In EVS 2 §15 Common Eldarin sirja 'flow' and orja 'rise' had short a
(added to formative j) which appeared only in the aorist, and they had past-tense forms or01ye
and sirinje. Here they have the past-tense forms characteristic of the derivative class, where
according to EVS 2 the CE suffixes such as ta,ja always come between the stem and any other
inflexions, as in orta 'raise' pa.t. ortane, tuija 'cause to come' pa.t. tuijane. The change is that the
meaning of the suffixja (here spelledya) is intransitive rather than causative.
LVS ll with the heading "Vocabulary notes" was composed on both sides of another Alien
and Unwin sheet. It begins with a description of the Quenya derivatives of the bases ..fsAB
'believe (that statements, reports, traditions, etc. are) true, accept as fact' and ..fsAN 'think, use
mind; ponder, consider (in thought)'. This led to a consideration of the bases KWI 'suppose' and
KE 'may (be)', the former being the source of Q qui 'if'. In EVS 2 §21, when describing the
expression of 'moods' in Eldarin, Tolkien at first gave the adverbial particle for marking the
conditional clause in Quenya as either que 'if' or, for an unlikely supposition, aique 'now
supposing'. Later in pencil annotations he proposed changing these to quf and aiqui, and sub
sequently he rejected these and suggested it, fte and aite. LVS 1 1 presents a conception close to
the interim quf/aiqui with the modification that qur is unemphatic 'if' "which can be replaced
by 'when'," while quf or qufta is used for a supposition "known to be not in accord with fact."
In addition the "unreality can be more explicitly expressed" by including a second particle ke
'maybe' or ke1a 'maybe not', as in qu(ta kela tuldes, which was never translated, but probably
something like 'I suppose he did not come but if he had' is the intended meaning.
Tolkien subsequently struck out this discussion of quf/ qufta and ke/ke1a, although retaining
the base forms and description of the use of Q qui 'if'; and on the other side of the sheet he
cites kwi/ qui, iqui 'if' and an extended form of the base KWIS 'imagine, suppose' used as a verb in
Quenya. In LVS 7 each of the particles qu( (spelled kw() and ke is used by itself to modify a verb
which expresses an opinion about some future action: kare mara kwf (or kwita) tyare naxa 'doing
2 One of the half-sheets in the folder contains the first three versions of "The Ambidexters Sentence," published as
part of "Eldarin Hands, Fingers & Numerals and Related Writings," ed. Patrick H. Wynne, VT 49, pp. 3-27.

good may cause evil'; la karitas alasaila lee nauva 'not doing this may be/prove unwise'. Both
express uncertainty about the consequences, but ke is used where the hypothetical is 'unreal'
in the sense that it is the opposite of what the speaker is advising. The particle ke (or ke) is also
used in the Ambidexters Sentence, e.g. in the final version: an ke mo queme kendele ntimenna, ve
senya, i hyarma tentane Melcorello 'for if one turned the face westwards as was usual, the left
hand pointed away from Melkor'. In this case the suppositional or 'unreal' force of ke may be
due to the context, which is an explanation of what the Eldar thought about the left hand "in
their imaginations." The accompanying glossarial notes on the back of the half-sheet with the
first three versions of the sentence and on one of the narrow slips placed next to it in the
folder (all hastily written in ball-point) include the following notes on the adverbial particle:
if. ke. kenai, if it be that ...
related t[o] KEYE. [?being] in CQ kfta-. {keit »} kitan, I suppose. {final stress[ed] ei > e.}
kenasit. *kenaszta. kenasta, if it be so, may be. perhaps[?].
ke usually with aorist. ke mo qete ulca. ke tu lis, {tan} tanome nauvan. {inw[?] »} nf nauva
tanome.3
LVS 12 is a folded sheet with an etymology for the name Gilraen on one half and a descrip
tion on the other of three suffixes to "distinguish": (a) -ta as a "necessary" formative added to
elements that were not verbal; (b) -ta as a formative not necessarily appearing outside the
aorist; and (c) -ti as a causative suffix. Tolkien notes examples of tense stems for these types,
citing forms of four verbs: (a) orta ('rise'); (b) caita ('lie') and ista ('know'); and (c) orta 'raise';
and he presents these in a rough chart, written upside down below the etymology of Gilraen,
where he also includes the verb nahta ('slay'). The tenses are not stated explicitly, but are
readily identified and we have added labels at the top of the chart. Thus for the causative orta
the stems are aorist orta, present (continuous) ortea, future ortauva or orttiva, past ortane, and
perfect orortie. The perfect form emerged in the composition of LVS 10 replacing orortanie.
In the notes Tolkien cites what appears to be a Common Eldarin perfect form orortaiye, with
some hastily written notes on the development of its ending that are difficult to interpret, and
an alternate form orotaye apparently from a similar stem without the repeated r. There is also
a conceptual change in the present tense of the type (a) verb orta, where the form ortia that
replaced ortea in LVS 10 is here first described as an "old" variant of ortea, and then revised to
ortya, which is the form that appears in the chart. For the type (b) verb caita the present tense
caitea is labelled CQ (probably for 'Classial Quenya'), and perfect acciye is given with an alter
native "late acaitie," so that the only form without the formative t(a) is past tense caine. Tolkien
adds a list of the other Old Quenya forms, aorist caia, cea, present cciya, future cayuva, and
perfect acciye. These variations suggest that the formative t(a) has spread by analogy to pro
duce alternate forms where the basic syllable cai- has become obscured, as in aorist caita for
cea, or perfect acaitie for acciye. In the chart Tolkien gives two rows for caita, tcea including
these variants, and a third row where all of the tenses incorporate the formative t(a).
The fourth subdivision consists of three bundles of documents and one loose sheet, folded in
a newspaper dated October 9, 1969, and placed between two cardboard dividers in box-file
'Quenya B'. On the newspaper Tolkien wrote "Negatives ii, la, va I Fate I Will & Shall future."
The paper he used is all discarded copies of an Allen and Unwin note dated " 12th Jan. 1968."
The first of the bundles is nine sheets clipped together, comprising two documents which we
3 For an analysis of these notes in connection with the Ambidexters Sentence, cf. VT 49, pp. 19-20.

Late Notes on Verb Structure ]. R. R. Tolkien
present here as LVS 13 and LVS 14. The second bundle consists of two versions of a note on
"Fate and Free Will," treated as both linguistic and philosophical concepts among the Eldar.4
The loose sheet and third bundle contain notes on demonstrative pronouns and adverbs.
LVS 13 is a three-page typescript bearing the title "Question of the expression of Negation."
This begins with the statement: "u will not do." Tolkien felt that the resemblance to the Greek
negative particle ou and the Norse negative prefix u- "is too close," referring to the conception
of u- as a Quenya negative prefix and the stem of a negative verb, as in umare 'not good' or uin
kare 'I do not make', which he had outlined in LVS 5. He had previously described this at length
in certain Definitive Linguistic notes from around 1959 or 1960, where Eldarin .fABA, BA is used for
refusal and negative command, and .fu is used for denial of fact.5 Tolkien expressed a similar
rejection of this conception in LVS 7: "u is not a suitable negative element. Too obviously
derived from Germanic un, Norse u." In LVS 13 he says "u should remain" but with a more
suitable meaning 'bad, uneasy, hard' derived from .fuG 'dislike', so that Q un6tima means
'difficult/impossible to count'. A similar meaning was given for u in the rough notes in LVS 8,
and in earlier notes on Galadriel's Lament (cf. PE 17, pp. 62-3).
In LVS 7 the rejection of negative u served to introduce the alternative conception of .fALA as
the negative element from which Tolkien derived a negative prefix al(a)- and "quasi-verb" or
interjection la (unstressed la) 'no, not', lanye 'I do not, am not', etc. In LVS 13 he says of this
alternative: "I adopt this" and refers to "a separate study of its forms," presumably LVS 7. A
marginal annotation in the earlier text that "conjugated la fell out of use, since the mere
negative lanye, I don't, not I, was sufficient in answers," is expanded here to a proposal that the
stem la "should not form a negative verb or take pronominal affixes, unless the verb is not
expressed," as in: 'Can you come?' - '(No) I can't'; or 'He says so, but I don't'.
Most of the second and third pages of LVS 13 describe another Quenya negative interjection
va, initially derived from a stem BA which "expressed the will or wish of the speaker in refusal
or prohibition," as in 'I won't'; 'don't, you shan't'; or 'he shan't, he mustn't'. The meaning
depended on the context, but could be clarified with pronominal affixes, as in vanye 'I won't'.
This is a development of the conception from around 1959 or 1960 of .fABA, BA (mentioned
above), which had a derived interjection va! that meant 'I will not, shall not' or 'don't!' and was
"often personalized in Quenya" as in van/vanye 'I won't'. In the first version of this description
in LVS 13, when Tolkien reached the point of explaining the distinction between itas la
tuluvanye and itas va tuluvanye, both of which mean 'in that case I shan't come' (the former in
the sense that something will prevent me, but the latter in the sense that something will occur
to make me refuse), he typed the meanings as "in that case he won't come." He subsequently
corrected these in ball-point, but the initial typed version ends at this point, and later he
retyped the second page of the text starting it as before but revising it as he typed.
It appears that Tolkien had paused to consider that with a third person pronoun the use of
va should mean that the speaker will prohibit him from coming: va tuluvase 'he is not to come,
he shan't come'; so there would be a different way to say "he won't come" in the sense that
something will occur to make him refuse. In ball-point notes on the earlier typed page Tolkien
proposed a verbal base .fABA and in the revised typescript va has the same meaning and usage
as before, but it is derived from the base ABA 'refuse (an order, request, petition); prohibit,
discountenance another's proposed or likely action'. After Tolkien completed the description
4 Cf. "Fate and Free Will," ed. earl F. Hostetter, Tolkien Studies, Vol. VI, 2009, pp. 184-8.
5 Cf. Bill Welden, "Negation in Quenya," VT 42, pp. 32-4, and the list of "Eldarin Roots and Stems " appended to
"Words, Phrases and Passages," PE 17, pp. 143-5. '

of va and its usage he cited "the full verb" Q ava- 'refuse, forbid', which can be used to express
"the refusal or disapproval of other persons than the speaker." This is an a-verb, with stem
vowel a instead of i in the aorist, comparable to ala- 'grow' and fara- 'hunt', the Eldarin stems of
which are cited in EVS 2 as gala and phara/spara. The revised typescript does not reach the
point of citing the inflected forms of the verb ava-, but from the earlier ball-point notes
comparing its various forms to those of other verbs we can reconstruct his conception: avan ('I
refuse'); avan, tavean ('I am refusing'); auvan ('I will refuse'); tavanen, aunen ('I refused') and
avavien ('I have refused'). Tolkien was in the middle of a sentence explaining the significance
of the lack of a "verbalizing consonant" in ava- and was inserting a note about verbs that did
have a formative -t- or -y- before the stem vowel: orya- 'rise'; nuya- 'descend'; anya- 'arrive at,
reach'. Probably connected with this are rough notes in ball-point on the back of the rejected
typescript developing further the tense forms of " -ta/ya causatives" and "-t/y intransitives or
formatives" with tabulations of the same verbs treated in LVS 7, LVS 9, LVS 10 and LVS 12.
LVS 14 is a five-page manuscript with the heading "Wil� shall and the 'future'." The main
text was written in ink with a nib-pen and notes were inserted in the course of composition
written in ball-point, while further annotations were added in the margins, some in ball-point
and some in pencil. The text begins with a description of two bases and their derivatives that
are connected with the semantic notions of 'will' and 'shall'. The first notion is that of "a fully
conscious purpose (to the accomplishment of which a 'person' would apply himself with
vigour of mind and body)," expressed by derivatives of the base v'"NID 'force, press(ure), thrust',
notably the noun indo 'the mind in its purposing faculty, the will'.
The second notion is "to be willing, consent, agree," expressed by the Eldarin element
I>A /APA with basic sense 'be helpful; be willing to assist, in any work etc., agree, consent'. This
has a derived interjection, as in Q pa 'very good/all right/I will', or more forcible apa 'I agree of
course', and a fully inflected verb apanye 'I will (do so)'. In Sindarin the verb atha is used to
answer questions as in athon "yes, I will," but is otherwise agglutinated to the verb it governs
to form "a kind of 'future', expressing the intention of the subject," as in linnathon 'I will sing, I
intend to sing'. That the Eldarin element originally meant 'be helpful' is shown by derivatives
such as the name of the healing herb, Q asea, S athelas, and Tolkien describes the history of this
word. But later in hastily written annotations he says: "This needs correction," and he derives
the words having to do with healing from a base HATHA 'treat kindly/make easy, (help to) cure'.
LVS 14 concludes with a description of the various forms of the future tense in the Eldarin
languages. As in EVS 2 there is no CE future suffix and Quenya, Telerin, and Sindarin form the
future tense in different ways. Among the suffixes used -uba is formally the same as in the
earlier conception, but in EVS 2 the meaning of the base v'"us was 'ponder, have in mind' while
in LVS 14 v'"us has a derived verb uba- 'impend, be imminent, approach, draw near'. Contrasting
with the personal wish or intention expressed by the Sindarin future with suffixed -atha- there
is a 'future simple' using v'"TUL as auxiliary verb: tolen cared 'I am coming, drawing near to doing,
I am going to do/shall do'. Tolkien hesitated between two explanations for Telerin, at first
indicating it used the form that was originally a 'present continuous' (Q karanye) to express the
future; but in a later marginal annotation (which was left uncompleted) he says that the aorist
karin was 'undefined' and could be used with adverbs indicating future or past time. Rough
notes in the margin above the earlier text and on the back of the same sheet propose adverbial
stems, such as epe 'after' and awa 'before, ago', different from the adverbs of similar meaning
proposed in LVS 1.
PARMA ELDALAMBERON N o . 2 2 - Page 146

[LVS l]
Some notes for re-ordering and fixing of Q.6
Stem of verb 'exist' (have being in primary world of history) was vEfJE, distinct from vNA
joining adjs./nouns/pronouns in statements (or wishes) asserting (or desiring) a thing to have
a certain quality, or to be the same as another. 7
eue is not followed by any adj . or noun but only by a[n] adverb (or negated adverb)
mainly[?] of time.
Present continuous of vEfJE : is ea, It exists: used as noun = the total created universe.8
Properly cannot be used of God since ea refers only to all things created by Eru directly or
mediately.
S{nome maruvan tenn' ambarmetta, 5mm m�lim pm· 1'pr.>mp
arhildinyar, 1r.>A Sl?.='r.>.9
lan arkanwar Valaron tauvar
lan i Valaron arkanwar tauvar.10
Q si, now, also s{r, hence sinome, sfmen. S sfr, sf.11
sfr, now. sf, here.
before oftime. PHA, APHA. enge. 1 2
before, in front of place. KHYA.13
after {later than) of time. KA, KATA. (e)uva.14
after, behind of place. PA./PATA.15
[Added above:] afar, khyar; katar, patar.16
[Added below:] fai; afea.17
Eru fiaz,. sz, euva. 1 8
"
6 The primary text is in ink written with a nib pen. The page was dated April 6, 1969, in the upper right corner.
7 Tolkien started to gloss vNA as "be of ... " but struck this out; and the second occurrence of the word "a" was
inserted later. This paragraph and the second following one were cited in an Appendix to Eldarin Hands, Fingers &
Numerals (edited by Patrick H. Wynne, 2005-7): "Late Writings on vNA. 'to be'," VT 49, p. 28.
8 "Present/continuous" » "Present continuous"; and originally following the gloss "It exists" was the beginning
of an example: Eru ea ar . . (struck out in the course of composition).
.
9 These phrases in Quenya and the following two were written with a ball-point pen. In the spaces remaining on
the page Tolkien transcribed lines 1-7 and 9-10 of Galadriel's Lament into tengwar, also in ball-point pen, but
without a transcribed version.
1 0 These Quenya phrases are not translated, but both apparently mean 'while the thrones of the Valar endure' (cf.
The Lord of the Rings, Book 6, Chapter V). On the back of the page, which is an unused publication notice from Alien
& Unwin, Tolkien wrote various rough notes in the blank areas, mostly relating to the etymology of the word
arkanwar and certain related compounds, which we have given below.
11 Tolkien apparently first wrote: "si,
si, now. Q, S sfr, sf," then emended the language designations and added the
other forms. The repetition of sfr and sf with glosses probably refers to them as specifically Sindarin forms.
1 2 EBE, B E » APA/PA » PHA.
13 AKA, I<A » AK ... » KHYA. To the right of this is a deleted form ede.
14 L. >> RA, ARA >> APA, PA >> EBE, BE >> KA, KATA; and euva >> (e)uva.
••
1 5 PAT » PAT(A)/PATA » PA/PATA.

16 To the right of afar, khyar the form sfr was added and then struck through.
1 7 Above this was earlier: "p� pai, at a prior time; ka, kai ... "; and below: fan, afra » afayi[?], afea » fai; afea.
1 8 This sentence was not translated, but seems to mean: 'Eru (was) before, (is) now, will be (after)'.

[Rough etymological notes on the verso:]19

polo. {m6lu.} pol.
khado. haro. hazo.
khadma. xanma > xanwa > hanwa. tarhanwa 'throne', high seat.
ara. handal, handar. arhand. khandla/ra. handala. handara.
arhandal. arahaom. arahaaw, archaf. chaow. haaw.
voihandal. {buyor.} v'"Bo! o. voivo. bu > v. voihadal[?].
S sm > ph. dm > v. gm > yw. dagma, dau(v). [?haul].
S bm > mm > m. nm > nw > n. tm > dm > df. km > gf.
ha<Jma. ha<Jwa, to seat. haawad. arahaawad. archap[?].
akh. aktha. akhada.
ahta. hao. khadatawar. hazataur. harator. [?khatha] toro. hadatorma. haratorma.
artorn. khazo > khaa > hauo. toro. toron. hartor(n). harator.
arthorof. artom[?]. tarthom.
[LVS 2]
Irreg. verbs
VMAGA, have, possess.20
main(ye), maitye, mailye, mai (maise)
mai(l)me, mail(de), mainte
mai(l)we. 2 1
ai is preserved[?] when [?hiatus] earlier, cf. na'in.22
plural[?] not[?] maiwe = bird. w is from [?dual]. mai(l)ve.23
past mange, manne, maine. fut. mauva. amaye, perf[ect] . part[iciples] maita. manna.
isile.
ista. short form isin, isit, iselis. {isi} isilme, isilve, is[...]. ista, long[?].
isil lSl irme ille.
{is.} {isil,} is(e), inf. isil.24
1 9 The arrangement and punctuation of these notes is editorial. They were written in the spaces of a publication
notice, but the order of composition can only be inferred from the notes themselves. Scattered among them are
also various words written in Tolkien's New English Alphabet, mostly transliterations of words in the notice.
20 Tolkien originally gave the forms main I manye under this base; but then he struck through the first form and
began again with the second, adding a statement and conjugating the present stem (all later struck out):
..fMAGA, have, possess. has only aorist {/} imperfect pres[ent] stem miigii-.
man(ye), matye, mal(ye), mase [mal
mame, mave, male, mate
{ma »} malme, malve, malde, mante, mar.
21 Tolkien originally wrote maiwe, mailwe; then he struck through the second form and altered the first to mai(l)we.
22 This note was added in the margin to the right of the present-tense forms. The reading is very uncertain: the
word that looks like "hiatus" might conceivably be "length"; and the referenced form could perhaps be nain.
2 3 This note was added in the margin to the left of mai(l)we, the final form mai(l)ve perhaps intended to replace it.
2 4 The entire notes on ..fMAGA and istii were later deleted. Below these on the manuscript page are the notes on nii
'is' and ea 'exists', which were presented with an analysis in "Late Writings on ..fNii. 'to be'," VT 49, pp. 28-30.

[LVS 3]
Record of Q. etc. inventions. 25
v'"KIM-, edge, brink of. kimba, edge, brink (espec. that of [?cliff ?or] fall in rock or stone).
kfma, edge of a cutting weapon / tool. kirmakin (im-), knife-edge.
v'"KIL-, choose, select. kilme, choice. tengwakilme, spelling.26
tengwa is not from tekma, since -ma is not passive but instrumental, and prob[ably] by
rev[ised] phonology -km- does not > Dm- > ngw.27 but v'"T EIJ, indicate, mark, signify. teD ma >
tengwa ? sign, token.
tengwar should then NOT be used simply of the Feanorian System. This should be Feanorian
Alphabet as in Q Feanoreva Tengwasse. (Certhas Daeron.)28
TEK = write. a written letter should therefore be tekda > Q[?] tekta. In early forms b, d, g (not
3) were nasalized before nasals and followed At. In later tB forms after change d > o > f, d > rm,
m; 9 > 0 with length, gm > a_ m etc.; b > lJV > U.
Given as certa & tengwa. 29 t{w ? in LR. certh. tfw. cf. Moria gate[?].
teD ma > s* teDwa > tewa > tfw.30
At31 -nm > nw > nw. {dm > nm >}
-nn > nn {dn >}
[-nD > DD > ng.]
D m = nw > ngw.
D n = D n. remaining, but later > nn.
[DD > D9· ]
mn > mn.
mm > mm (mb).
[m D does not occur ?]
S nm > nw
nn > nn
[nD > D9]
D n > D n > -n
D m > DW > -w
[rm > ng]
mn > b n (vn)
mm > mm (m)
[m D ?]
2 5 These notes were all written and emended in ink with a fountain-pen.
2 6 These etymologies repaced the following original:
v'"KIL, + KILIM. kilma 'edge, brink'. killa, edge cutting-edge, kirmakil.
2 7 The original reading here was: "-km- does not in S. > gm > rym- > ngw."
28 Tolkien deleted two preliminary forms Nairon and Dairon, the first perhaps intended to be the Quenya form.
2 9 Original "certar & tengwar" » "certa & tengwa" in the course of composition.
30 This etymology went through several versions: "teryma > S* tirym- > tf" » "teryma > early S* tema > tfw, tff ? cf[?]
tfw." >> "teryma > S* terywa > tewa > tz" >> "teryma > S* terywa > tewa > tfw."
31 This list was written to the left of the paragraph about tengwar, but was later labelled "At" and cross-referenced
in the discussion of TEK. (No further explanation of "tB" is given.) There is an arrow connecting this list to the
following list of Sindarin developments written at the bottom of the page to the left of the final etymology of s
tfw. The brackets and question marks in these lists are all Tolkien's.

Q k, t, p were not voiced before medial nasals. 32

gm > vm > ngw km > km > k1v > kw
�>�>M �>�>�
dm > nw tn > tt
dn > nn tm > tm [>] t1v > tw
bm > mm, mb pm > pp
bn > mn. pn > pp > pt (ft)33
In later forms km > kum
kn > kan
tm > tum
tn > tan
pm > pum
pn > pan.
certar, LR Ill 395, could be emended in text to cirtar, and Certhas be held a late formation as it
was.34
e > i.3s
Phonology cannot be altered, since we have elenf > elin not ilin.36
KIR. rather a mess here. LR has certar = cirth, and we have Angerthas, therefore v'"KER is in
dicated, & comparison with kirya is not in point of fact likely! Easy to say v'"KER = cut with
tool/weapon, but v'"KIR = ? shave; skim (surface), slip along, glide quickly, and kirya is really an
adj . = swift gliding. But what of Cirith.37 e > i only before vanishing f/j.
3 2 This note and the lists of consonant developments are written in pencil on the verso of the same page {a notice
from Alien & Unwin), and a notation on the recto "PTO" {also in pencil) refers to them.
33 This is a variant of another version that was allowed to stand: "pn > pt > ft."
34 This sentence is in ink; but the following notes are in pencil.
35 Above this Tolkien wrote a brief allusion to the "phonol[ogy]" of "Middle Welsh" but later struck this out.
3 6 In this sentence "elen > elin" » "elenf > elin" in the course of composition. {Note that the Sindarin plural form
elin was published in The Road Goes Ever On, 1967, p. 65.)
3 7 Following this Tolkien wrote: "Here[?] we could say -itt," but then he struck this out.

Late Notes on Verb Structure f. R. R. Tolkien
[LVS 4]
Eldarin
Points needing clarification, and explanation of some names.
No ..f bases appear to have been devised for "can, may" etc. Foll[owing] are here suggested.
"can" = have power, strength, ability inherent physically or mentally. ..fKuRu. Cf. *kurwe
'power, ability', S curu in curun(r 'wizard', us[ually] applied to exceptional powers espec. of
mind, ability to make one's will effective.38 It thus approaches some uses of our 'magic', esp.
when applied to powers not understood by the speaker, but it does not even then (except
perhaps when the word was used by Men) connote any alteration or disturbance of the
'natural order', which to the Eldar were either 'miracles' performed by agents of the One or
counterfeits by delusion (or by means other than miraculous which impressed the un
instructed as supernatural).
* distinguish ..fTuR, power of domination or dominion: control of other wills, legit
imate or illegitimate "mastery"; also used of the effect of kurwe upon other persons or
things in making them conform to the kurwe of the user.
"can" [=] have opportunity, chance, or permission - be allowed by circumstances, 'way of
the world;. YNAYA.
Hence nai ( < nayi) 'it may be, there is a chance or possibility', maybe. This mostly used
where the opportunity or permission is certain or very probable ('may well be'). It is not used
personally but either with dative pronoun or with an undefined clause [in which case nai is
virtually adverbial in our syntax] .39 nai nin h(res "it may well chance for me to find it" or nai
hirinyes.40 In this case the verb may be future: usually is [nai hirinyes being used rather in gen
eral statements, I have chances of finding it, almost equivalent to our "I sometimes find it."]41
"can" of mere possibility according to likelihood, natural probability etc. ebe- .
..fNDAB- endeavor, try. seek opportunity. nauva tiile, he will try to come.42
[LVS S]
Finding a suitable ..f for 'have' = possess, own, keep (have in hand, use or with one) difficult.
Suggest ..fAw (distinct from ..fAB, make it go back, refuse). This to be an -a verb (as in other
verbs ref[erring] to continuous processes or states even when temporally 'aorist'). This could
have pres. (aor.) auta, pa.t. ayante (oante, vante) or oa (ihya) beside ava, pa.t. aune. noun auvie
'possession' abstract, auda > aura, a possession or thing owned. fut. ayuva > auva.
Kif'would it not be best to eliminate inserted nasal from pa.t., the rare cases being trans
position of stops (p, t, k + n). So that e.g. vKJABA, taste, LABA, lick, would not have pa.t. tyambe or
lambe, and insertion be used for other purposes (as already in lambe 'tongue'), but tyamne, lamne
or later tyave, lcive.
38 The original phrase "ability to make effective" » "ability to make one's will effective."
39 The brackets in this sentence are Tolkien's; and so are those in the last sentence of this paragraph.
4 0 In the first example hires replaced another form, perhaps hirie[s] (the brackets in the original); and in the second
hirinyes replaced a false start, perhaps hiruv...
41 In the margin to the left of this sentence Tolkien added the example: nai hiruvalyes » nai hiruvanyes.
4 2 This etymology was added in pencil in the top margin of the page.

-ne is chiefly added to w, y, r, � n, m, !J, s > z (r).

aune, koine, kame, tulde, panne, nemne, (engne) enne.43
n used mainly in stems with sund6ma a, o, e. i, u used ai, au.
KOLo-, bear, kolde or kole > k9le.44
ii- negative prefix. cannot be used when it leaves more than one positive possibility. It
only denies the stem joined with it and asserts its opposite. timare, not good = evil (not any
diminished grade of goodness as fair, mediocre, middling[?]). Hence never ticame 'not red'.
ii- is not prefixed to verbs but itself conjugated and joined with 'bare infinitives' se. kar- >
kare, karin 'I make', uin kare, I do not make.
kar-e, do, make. aor. karin. pres. cont[inuous] kdranye. fut. karuvan(ye). perf. akarienye. past
kamen. [? later karen to avoid clash with karani > came (cami-) 'red'.)45
bare infin. kare. general noun karie, doing. passive n. karda, deed. instr[umental] karma,
tool, implement, means. aorist passive adj . karina. also used with intrans[itives] as kwal- 'die',
qualina 'dead'. aorist act[ive] karil(y)a. kdrielya (kamela). karuvalya.
kdriela. karuvaila. lya clashes with -lya, thine.
m(riel. 46 mirieil, mirieail.
S -iel is not precisely same as Q. being gerundial[?].
�
[LVS 6]
Notes on verb and treatment of -w- in Q. & S.
In Q. -w- disappeared after y with modification of the stressed vowel, as &wa > oa, but it did
not disappear before i. After long vowel w remained & so event[ually] fell with v < b. w lost
before ii.
y = vowel. x = cons.47
In S. -w- did not vanish or bec[ome] absorbed while foil[owing] vowel remained, but became
lJ and formed diphthongs when the foil[owing] vowel was lost.
Thus KJAW-, taste, select, choose. accord[ing] to regular [???]
kyawr > tyawe > tyave. {

Q *kyawine (so by assimil[ation] of nr > other pronouns) > tyawin > tyavin.
lgawa > tyava
kyawuba > tyauva. algawie > aty&vie. imper[ative] kyaw-a > kyawa > tyoa.
but LAB-, lick, exactly corresponded in all but imper[ative] lab& > lava, & lab uva but lavuva.
In [???] verbs with medial b, aw coalesced. The future usually be[ing] tyauva, lauva. The
[?imperative] tyava, lava.
s kyawini > kawin(e) > cewin. but Igawr > caw. future lgawathani [>] cawathon (? cauthon).
Imper[ative] cawo. past akjawe-n (lgatJne > cawn) agauwen > agowen.
��
4 3 These examples were added later in the right half of the space before the next paragraph.
44 This note was written at the foot of the page without an indication to what it refers in the text.
4 5 These brackets and the question mark are Tolkien's.
46 After this form Tolkien originally wrote: "Q miriela," but then struck this out.
4 7 This symbolic key was written in the upper right corner of the page; but in the event Tolkien only used the
symbol y in the first sentence, where j stands for 'stressed short vowel'.

[LVS 7]
Q. Verbal adj ectives - AL(A) negative &c.
[Rough notes on a narrow slip:]
The neg. element should be like (NA) only in form LA to avoid clash with GAL.
la- before verbal adj ective.
la bef. nouns. lcikare.
?
ala-, un- only as a prefix so no clash.
[Typescript:]
alfirin evid[ently] intended to mean 'immortal' as a flower name; j oined with mallos ([LR] Ill
151), both golden bell-shaped flowers. Immortal, as being of texture like immortelles, or as a
symbol? Can use be made of AL/LA as a negative element? How differing from u?
Yes. (J is not a suitable negative element. Too obviously derived from Gmc. un, Norse ti.
Also it is intended that Eldarin should express negation by a negative verb (as Finnish) or
rather that the personal and tense affixes should be attached to the negative and the verb be
expressed by an indeclinable stem-form, and ti is not suitable.
The negative element is to be vALA (of the class in which the sund6ma can appear on either
side (or both) of the basic consonant). As a quasi-verb its form is to be la. aor. lanye (I do not,
am not); pres. *laja > laia, past lane; perf. alaie; fut. lauva. The imperative ala, ala = don't! The
simple form la = no, not (it isn't, it doesn't, not it). lanye, I don't, not 1.48
With nouns and adjs. (not regularly attached to verbal stems, such as -ima, -ite) the form
was al before vowels* and usually (in all later and free formations) before initial consonants
that in Q. were permitted after l : all except hy, hw, n, r* before which ala- was used.
* In old formations al before g- which disapp[eared] initially became all-; sim[ilarly] ln
became ld ; so also l before r/l from original initial d.
[a]la was also normally used later before [s], h, f.49
Older formations were those in which al functioned as E. '-less' or 'no, non-, without'. as
alahen, pl. alahendi, eyeless. 5°
As a prefix (in-, un-) the derived forms were ala, a� la. The form la had a short vowel because
originally the stress was on the following verbal stem. ala was an intensified and stressed form
with stress on the first a, and consequent ancient shortening of *ala ; al- was a reduction of this
according to the Q. tendency to syncopate a short vowel after a stressed vowel of the same
quality when the contact of consonants so produced was permitted in Q.
4 8 The gloss "not it" and the example "lanye, I don't, not I" were added in ball-point. In the upper margin of the
page a note, possibly connected with this example, was added later (also in ball-point):
But in colloq[uial] language the {declined} conjugated la fell out of use, since the mere negative lanye, I
don't, not I, was sufficient in answers.
49 This note was typed in the left margin at the very edge of the page, so that some letters are uncertain.
50 This paragraph replaced an original, which was circled and deleted in ball-point, in the course of composition:
As a prefix the reduced from la (short a) appears. This is used with verbal derivatives, nouns or adjectives,
such as la-karima, not feasible, impossible; laken{te, unseeing (general), as e.g. laken{te kuimar, creatures that
cannot/do not (by nature) see, se. 'blind' (equals lomba (S dom, [p]l. dym) from v'DOM, dark, as in l6me, night);
laista, ignorance (laistea, laistala, ignorant).

la.51 This was used with verbal stems, nouns or adjectives formed with suffixes regularly
employed and regarded in Q. as part of 'conjugation': e.g. the general (aorist) 'infinitive'
formed by added -i (not as such capable of any further suffixion; with pronominal affixes it was
the stem of the aorist tense); the particular infin. with -ita differing in use from the preceding
mainly in being able to receive pronominal object affixes;* the verbal adjs. or participles
formed with -(i)lii and capable of expressing tense;5 2 -ima expressing possibility or per
missibility; -iti added to the 'sund6ma' forming general adjs. of aorist reference, and often
functioning as an active form of passive -ima, with sense 'capable of doing, generally (and
naturally) doing.
Examples. kuita'r pare 'live and learn'; kare mara kw( (or kw(ta) tyare naxa 'doing good may
cause evil'; la navin karitalya(s) mara, 'I do not advise you to do so', lit. 'I don't judge your doing
(it) good'. 53 lakare, not-doing, inaction (general). Not doing anything in a particular case, is not
expressed by a prefix since the negative was normally stressed; hence la karita i hamil mara
alasaila (na), "not to do (in this case) what you judge good (would be) unwise." Q. needs no verb
before alasaila but can add na 'is'.54 English normally says 'would be' because the whole
expression is equivalent to 'if you think this action right, it would be unwise not to take it' and
because it is plainly a piece of advice that will be acted on, or not, in the future. If this
uncertainty is emphasized Q. can say nauva 'will be'. Uncertainty in the advice must be other
wise expressed in E. and Q. 'not doing this would be (I think) unwise' or 'not doing this may
be/prove unwise'; la karitas, navin, alasaila na or la karitas alasaila ke nauva.55 (On ke see notes on
translation of 'can, may etc.')
Note that though a verbal derivative such formations as lakare are nouns and not 'infini
tives'; they cannot take an object any more than E. 'inaction'. To express the advice in general
'aorist' terms one must use the separate negative: alasaila na la kare tai mo nave (or navilwe) mara
'it is unwise not to do what one judges (or we judge) good'.56
mo, indefinite personal pronoun 'somebody, one'.
ma, neuter personal pronoun 'something, a thing'.57
For the indefinite mo the inclusive l.pl. = 'we, you and I (and others associated)' was often used,
espec. colloquially, like E. 'you'.58
5 1 In the margin to the left of the previous paragraph and the beginning of this one is an annotation written in
ball-point and sideways to the main text: " *Abandon conception of la except when verb is not expressed."
5 2 In this clause the original ending -la » -ila (typed in the left margin) » -(i)la (in ball-point).
53 These three Quenya examples were revised in ball-point from the typed original versions:
kuita'r neme 'live and learn'; kare mara ke tyare naxa 'doing good may cause evil'; lanye {neme »} hame
karitalya(s) mara, 'I do not advise you to do so', lit. 'I don't judge your doing (it) good'.
In the right margin Tolkien wrote a note about the third example: "NDAB, to judge, Navo for Namo."
54 This sentence and the preceding example were revised in ball-point from the typed original:
hence la karita i nemil mara (na) alasaila, not to do {what you judge good »} (in this case) what you judge good
(would be) unwise. Q. needs no verb before alasaila but can insert na 'is'.
In the left margin is an explanatory annotation, also in ball-point: "na usually follows emphatically an adj ."
55 This example was revised in ball-point from: "la karitas nil, {nemin »} hamin, alasaila or la karitas ke nauva alasaila."
5 6 This example was similarly revised from: "alasaila na la kare sai e {neme »} hame (or {nemilme »} hamilme) mara."
5 7 These notes written in ball-point replaced an original typed note:
e, neuter, is an indefinite pronominal element corresponding to E. 'one' (neut[er] 'it, a thing'); a later
development, to make the purely impersonal verb form more precise and dist[inguish] it from the bare
infinitive, as is shown by its preceding the verb.
58 In this sentence the original typed phrase "indefinite e" » "indefinite mo" in ball-point. Note that the bulk of
these two paragraphs was cited by Bill Welden in "Negation in Quenya," VT 42, pp. 33-4.

Further examples: laista, not-knowing, ignorance (deriv[atives] laistila, laistea, ignorant);

lamate, fasting, a fast. Adjectival. 1. The verbal participles (capable like the definite infin. ita of
taking pronominal affixes) in -ila.59 Thus from vKAR were made: a) aor. indef[inite] karila,
doing; perfect kariela; future karuvaila. *
* Only these 3 forms were made: the aor. was used as a present; the perfect functioned as
'past' because this participle, the least frequently used, was only employed of a past im-
. mediately preceding the time of a subsequent and often consequential action or event.
2. -iti and -ima. These could be added to any verbal stems suitable in sense (and so were
regarded as part of ' conjugation') but were not capable of tense or of verbal constructions. -iti
was added to the 'sund6ma' or basic vowel of the stem, forming originally {ti, eiti, aiti, oiti, uiti.
In classical Q. these appear in forms changed by normal sound-changes, or by adaptation, as
ite (from iti, eiti), ete (from eiti after i, y), aite (from aiti, and oiti after w), oite (from oiti and
from uiti after v, by adaptation, since the sequence uvu was not favoured and unless altered
became u).*
* cf. wo to wa ; also the futures such as tuvuva 'will find' which became tUva or since this
was the same as the pres. continuous) tuvua.
Ex. tirite, ken{te, xiete, karaite, koloite, yuluite, kuvoite from roots TIR, look at (turn eyes to, keep
eyes on, watch), KEN, see, SKEY, pass, KAR, do, KOL, bear, carry, YUL, drink, KUB, hide, secrete. The
senses differed from the participle in la in describing some general, natural or usual charac
teristic: thus tirila, watching (something), tirite, watchful, vigilant; similarly xiete, passing,
impermanent; karaite, active, busy (karaitie, activity, business); koloite (KOL referring to the
ability to support weight or a burden, physical or mental, not necessarily to transporting it),
capable of bearing, tolerant (of), enduring (koloitie, endurance, staunchness, fortitude); yuluite,
drinking (as a habit), as in yuluiti kuimar, fishes etc., contrasted with nefiti k., air-breathing
animals (yulunefite, amphibious ); kuvoite, hiding, secretive, used both of hoarding things, and
not revealing one's mind or purposes (cf. pekkuvo 'nut-hider', squirrel), sometimes in a sinister
sense 'treacherous' (cf. S gurgof, traitor, f[rom] *gor(i) ku b a ): Q 6rikuvoite, uncommunicative,
keeping one's own counsel (not necessarily sinister).*
* The original accent was on the ai, oi etc. and the verbal stem is thus usually short. In
the next (ima) the suffix was added direct and the verbal is usually long.
Ex. of ima. t{rima, able to be watched, observable; kenima, visible; karima, feasible, possible
(to do), ankarima, 'very doable', easy; matima, edible; n6tima, countable, limited in number
(weight or measure), frequently used espec. colloquially in sense 'moderate (in amount), some
few, several' as in tas kennen n6time eldali, I saw a few elves there; k6lima, bearable, light (of
burdens and things comparable, troubles, labour, afflictions); pulima, pourable, liquid; yulima,
drinkable; ruvima, discoverable; kUvima, concealable. *These adj ectives were made chiefly from
transitive verbal stems.6° Forms in iti after intransitive stems were often virtually active
equivalents of the passive ima : noroite, (capable of) running.
[*] When formed from intransitive stems as kalima 'luminous' they differed from the verbal
participles in -ila (a) as being more intensive, (b) as being always aorist[?] and without
special reference to a present or particular occasion. Also when combined with la- they
functioned as equivalent of la + the causative (whether actually existing or not) as
59 The ending originally typed as -la was revised in ball-point to -ila, and probably in connection with this the
original example laistala » laistila (above) and future participle karuvala » karuvaila (in the following sentence).
60 This was revised in ball-point from: "These adjs. were naturally made only from transitive verbal stems."

lakalima, not possible to be kindled (made to shine), over[?] a long[?] time, rather than
·
lakaltaima. 61
These adjs. could also be formed from derivative verbs of which the most usual suffixes
were ta, ya, t, y, u. 62 ta, ya were causatives, whether from verbal stems or adjs., as vTUL, come,
*tulta- 'send' (from point of view of receiver) - "he will send it to me," tultauvas sa nin. Their a
ran through full[?] conjug[ation], hence tultaila, tultaima. t, y were 'formatives' only normally
used after stems ending in w, y, as KUY, kuita, live, KAY, lie, kaita, LAW, flourish (green), grow,
lauya. Also after stems that were not primarily verbal, as OR, up(wards), orya, to rise, intr., orta,
raise, trans., (a)n(d)tiya, descend, andtita, nuta, lower. t was also used as (originally) an intensive
or differentiator as in vNDAK, hew, ndakta, slay, to[?] Q nakin, nanke/nahtan, nakante, slay, slew.
The a in these verbs was only used in present-tense, thus[?] istaila, nahtaila, kuitaila, kaitaila,
oryaila, but also before zti, espec. after s, kh, th, tht (> st), ph, pht (ft). 63
not ima, which when add[ed] to pre[sent?] stem had [?none] as in ntitaila, nutaite, but
nutima.64
[Note on a half-slip:]
-ima adjs. Only the adjs. in -ima found with long stem vowel were adjs. of possibility. -ima
was frequently used with stems (verbal, adj [ectival] or nominal) with a short vowel, and the
sense possessing to a high degree (at all times & by nature) the property mention[ed]. So
kalima 'luminous' (by nature always), vanima, beautiful, norima, running, swiftly a course[?],
kelima, fluent, istima, wise (in sense of knowing much), knowledgeable, very well informed,
melima, loving, very affectionate - but lame1ima, unlovable.
The 'short' -ima adjs. were not combined with la-prefix. But they could be joined with ii- by
a sort of [?recourse] . uvanima, hard[?] to call beautiful = hideous. ukalima, dim, murky.
[LVS 8. Rough notes on a narrow slip:]

(A)LA should be negative. as alamanyar.
verb la-. lain, lanye; lail, lalye. laia. lauva. lane, alaie.
ii = Ger[manic] tor- (tus), [Greek] Ou<; , bad, hard, difficult, unfavourable.
un6tima = hard to count. ukarima = hard to do, difficult.
lanotoite, innumerable. l&karaite, impossible.65
with neg[ative] ala. alamanyar.66
lakarina, undone.
aleldarin, unelvish. alfirin, immortal.
6 1 This note (along with the asterisk in the text) was added in ball-point, written sideways in the left margin.
62 The typescript ends here, and the remainder of the text is written in ball-point.
63 The forms in this sentence were emended from original: istila, nahtila, kuitila, kaitila, and orila; and the phrase
"but also before lti " replaced another preliminary list: "but {istai... » } ist{te, nahtaite, kuitaite[?], kait{te, oriyoit . . .
"
64 This note was added in the left margin, written sideways.

65 These forms were originally: {alanot.. . » } alanotoite and alakiiraite, in both of which the initial vowel was deleted,
the accent struck through in the first, and a breve added in the second. Similarly lakarina » liikarina below.
66 following this Tolkien originally gave the example: "areldarinwa, unElvish," but later struck this out.

[LVS 9. Note on another narrow slip:]

The denominative or adjectival verbs, causative & inceptive.
The transitive causatives were usually form[ed] with -ta, as were the causatives from verbal
stems (as tulta-, send < TUL, come). So *tankata-, to make firm, fix. > Q tankata.67
The inceptives (become, grow, be made to be) used most frequently -ta, as *ninkwita-, grow
white, whiten. > Q ninquitan, pa.t. ninquinte, fut. -itiiva (u < tau > tau > tii), perf. ininquitie.
no difference between aorist and pres.
[LVS lO]
Verbs
Make a Q. past continuous. was eating. matante. istante, oryante, ortante. 68
[?More] like [?aorist] in syntax. and also helps to explain wide[?] use of -ant as past in S.
better matane > matane. istane. [?pret.] is[?] fse, sfne.69
augmented[?] verb often with nasal infix + a.
vnG. tinga-, go (for a long while) [???]. 70
lenda, go free. f...
already. still, up to now, here. tensi tenta.
it is still here it was still waiting[?] .
is it here yet has it come yet.71
anya = go to, reach. an.
ya verbs must all be intransitive, and be ya. siryane. oryane.
the[?] present contin[uous] us[es?] ea . sirya, s{rea. siryane.
'
-t formative or diff[erentiator] .
kaitla, kaitea [written above:] kaya. keante/kaine. akainie, akaine[?], akaitie.72
ortcl, ortia, oronte, or6rie. orta, ortea, ortane, orortie. orort{anie}. 73
no verb stems from[?] -t- and ta forms. 74
tultane. utultanie.
nahta, nahtia. [nanke] nakante. anakie, anahtie/5
orya, rise. oryan, 6reanye. oryane (6reane). oryuva. or6rie. or-oryaye, or-orye1 e. or6rie.76
�
67 This paragraph was struck through with a single diagonal stroke.

68 These are all rough notes written in ball-point. In this first line the original form matatule » miitante.
69 This was written in the right margin, circled, with an arrow pointing to it from discussion of past continuous.
70 This etymology has "X X " written against it in the left margin.
71 The phrases in these two sets were written in the reverse order, but they were associated with tensi and tenta by
means of arrows pointing from the forms to the first phrase in both columns.
72 The original form akayie » akainie, akaitie (each of which was circled), and the uncertain form akaine added later.
73 In this group of forms original ortea » ortia; or > or6{rtie} » or6rie; ortiine » ortane.
...
74 This note was added to the left of the conjugations of ortii and ortii, and seems to suggest rejection of the former.
75 The brackets around nanke are Tolkien's. This group was written at the bottom of the page.
76 This group of forms beginning with orya was written with a different ball-point pen. The two forms oryan,
6reanye were connected by a brace; the first was revised from original oryanye.

[LVS U]
Vocabulary notes.
vsAB believe (that statements, reports, traditions, etc. are) true, accept as fact.77 This in Q.
does not take as direct object a person - in sense 'believe he speaks truly'. Q savin when it has
a noun or name or pronoun as direct object means I believe he/she/it really exists/existed: as
savin Elessar ar i ncfne aran Ond6reo 'I believe that E. really existed and that he was a King of
Gondor'.
To trust a man/woman, as not lying (or twisting words, or concealing something relevant),
is expressed by: 'I believe (that) the words of A (are true)': savin Elesamo quetie /or i E. quetie
naite. Or by a different verb, to trust, rely on (a person):78
vSAN think, use mind; or tr[ansitive] ponder, consider (in thought). sanwe, (act or process
of) thinking, thought.
Cf. yosanwe of pl. congruence, or more often osanwe of dual: 'thinking together', com
municating in thought without words or other signs.
Some further monocons[onantal] bases.
KWI- suppose. KE may (be).79
kwi- seen in Q qui 'if'. qur when simple and unemphatic, of a simple supposition or con
sequence, which can be replaced by 'when' - laqui, unless, or qui ... la.
quf or qufta is used when (a) supposition is known to be not in accord with fact, as in "if he
had come (as he did not), I should have welcomed him," or "if you had not returned (as you
did), I should have been angry." quf(ta) la tuldes, nane marie (nin), [if he had not come,] it was
well to me (I was glad), or quenten tulil marie (nin), I said 'you come happily (for me)'.80 The
unreality can be more explicitly expressed by insertion of ke, may be, into a negative protasis
or kela, maybe not, into a positive. qu(ta ke1a tuldes, qu(ta. 81
kwi, iqui, if. possibly older forms of extend[ed] KWIS-, inquire, suppose. - espec. opposed to
ista, know. as in quistanyes : la istan, I suppose so, I don't know.
ER single etc. ere. re, single item taken out of a (long) series, hence Q re, day in calendar.
rea, single : rie, only.82
In Gondorian Quenya. The Rohirrim were called by an adaptation of their own name
Eorlingas - Erulingar, which was assoc[iated] with Eru. as if Erulingar was cpd. like Eruh(ni, sen
(children of God).83
No - for this would suppose an actual contact between Quenya and Germanic in the
"Third Age."84
77 Before "accept" another word (perhaps "Thus") has been deleted.

78 There is a " 11 " below this paragraph where Tolkien may have intended to insert the different verb for 'to trust'.
7 9 These base forms were added later above and to the right of the heading, and the glosses were added even later.
80 In the first Quenya sentence another word was heavily deleted before tuldes in the course of composition, then
la inserted later, and nane nin marie » nane marie nin. This was perhaps meant to translate the first English
example. (The bracketed part of the gloss is editorial.) The second clause was given a more literal gloss and the
second Quenya sentence provided an alternative way to express the idea in this second cause of the first sentence.
8 1 The original phrase "negative sentence" » "negative protasis" (a protasis is the first clause of a sentence). This
entire paragraph about quf/ qufta and ke was later struck through with a single diagonal stroke.
8 2 This etymology was added in the top margin of the page.
83 The form sen (or perhaps seni) was added below Eruh(ni, and may represent an alternative compound Erusen(i).
84 An "X" was written to the left of the etymology of Erulingar, and this note and the following added in pencil.

Eorlingas must be an interpretation (of the 20th C[entury]) of some other name in Rohanese.
The real word was puron (rel[ated] to Elvish tur- ?), porunahim. Sindarized Thoron1 on, Thoron
hfn[?]. Q Soronildi, assoc[iated] in Q. with soron, porn, eagle.85
�
[LVS 12]
Gilraen.86
gil-, star. vRAY, net: knit (contrive a network) - catch, involve (in a net). Q raita, pa.t. raine,
pres. nfya, trea, to net.
Q raima, a net. raina, netted, caught in net. raita rembe. S raef. rai. karrai. carrea.
[Added below:] raye. raiya > rae. cathrae. [Added above:] rea, raya. rae. rean. {raie} rayan.
distinguish.
(a) -ta as a nec[essary] formative to elements that were not verbal, as orta, which must in
clude the t element throughout conjugation, except perfect.
(b) -ta as a formative (that spread in CQ {added} especially to verbal bases with medial cons.
Y, w, and original GjN) which {did not appear »} did not nec[essarily] appear in other parts.
(c) -tii causative.
(a) orta, ortea [inserted:] {old ortia »} ortya, ortuva, oronte, or6rie later orortie.
(b) caita, CQ caitea, caine, caituva, acaye lat[ er] acaitie. OQ caia, cea, cciya, cayuva, acaye.
[Added above this line:] ista, sia, {siuva »} istuva, sfne, isfsie.
[Added in the top margin:] ista, [lsa], istuva, inse, isfsie, isistie.87
(c) orta, raise, {ort{e}a lat[er] ortea »} ortea, ortane, ortauva, orortaiye. orotaye > ee > ie.
aya > ea properly had p[erfect] ayai > aye, but ie [?replaced] from[?] zya > e1a > ea.
[Aorist] [Present] [Future] [Past] [Perfect]
orta ortea orttiva ortane orortie
orta ortya ortuva oronte [or6rie].
,
caita caya cayuva caine acaye
tcea caitya caitilva caine acaitie
eaita caitea caittiva caitane acaitie.
. , .
ista [istea] istuva tinse zszsze
isinte isistie.
nahta nahtea nahtuva nakante anakie
tnanke anahtie.
��
8 5 Original forms turs » tursernil » J:mron; J:mrnachim » j:JOrunahim; "pl. Thoryni, Thoryn" » Thoron-hfn (or -hfm).
With the tentative forms turs(ernil) compare tiiran » tiirac- 'king' and Lohtiir 'Horse-folk' in "The Appendix on
Languages," The Peoples ofMiddle-earth, ed. Christopher Tolkien, 1996, pp. 53, 60 (note 39).
86 These notes in ball-point are written on the back of a letter to Tolkien dated ]une 5, 1969. He folded the sheet in
half, composing the notes on the name Gilraen in the top of the left half, and beginning the notes on verb-stem
forms in the right half, continuing these on the left half turned upside-down with respect to the earlier notes.
The notes on Gilraen are probably earlier workings for a similar etymology of its second element, S raen, in
corporated by Tolkien into The Rivers and Beacon-hills of Gondor, ed. Carl F. Hostetter, VT 42, pp. 1 1-12.
87 The brackets around (sa are Tolkien's, and so also are those around or6rie and istea in the chart below.

[LVS 13]
Question of the expression of Negation.88
t1 will not do. It is not necessary to avoid at all costs similarities with known European lan
guages - Eldarin is deliberately devised to resemble them in style - but here the resemblance
either to Greek ou (phon. u) or to the unrelated Norse U. as a prefix, is too close.
u should remain, but with the sense 'bad, uneasy, hard' - similar to lE *dus, Greek dus-, Gmc.
tuz- (tor-). This will leave un6tima in GL correct, with meaning 'difficult/impossible to count'.
But uchebin in Gilraen's linnod will not fit so well. It must be assumed that in S. u was used as a
verbal as well as an adjectival prefix, with a meaning intensified to 'impossible' so that it came
near to a negative. The nuance will remain important. uchebin will mean not 'I do not keep',
but 'I cannot keep'.
u should be from a .fuG 'dislike' with varying degrees of intensity, with other derivatives,
such as Q uhta, dislike, feel disgust with, avoid as painful or nasty; S oer (< ogra < ugnl), nasty, Q
ura, owing to contact with .fuR, heat, 'torrid, sultry, unpleasantly hot'. Probably the spider
word ungu- would be related.
but -gr- > rr in urra ?89
The suggested alternative lc:i, la, ala would be convenient, and the fact that it appears in
Semitic would be no objection. I adopt this. See separate study of its forms: ala as prefix,
reduced to al- before vowel, or by syncope when resulting contact of l with following con
sonant was suitable; la as stressed negative 'no, not', and la unstressed. The latter was also
used as a negative prefix before certain verbal adjectives. (There is no other base AL, and
confusion with base GAL 'grow, flourish', as in Q alda, tree, is negligible.)
This stem should not form a negative verb or take pronominal affixes, unless the verb is not
expressed. In that case lanye and other similar verbalized forms become equivalent to E. 'I
don't' or 'I'm not' etc. This occurs principally in answers to questions, such as: 'Is it nice?'
'(No) it isn't'; 'Can you come?' - '(No) I can't'.90 But it can occur in other cases, such as: 'He
says so, but I don't'; 'I was afraid, but he wasn't'. Note: the la does not express difference of
tenses, normally unnecessary: the tense of la plus pronominal affix is always that of the pre
vious verb, now negatived. (la is usually shortened to la before 2 consonants, according to the
usual Q. procedure, but the long vowel can be retained, especially for additional emphasis, as
in other cases where pronominal affixes follow a long vowel, as in marya 'his hand'.)*
* Questions appear in Q. to have had in colloquial speech the same form as statements,
distinguished in tone only, not by word order. But all questions could be preceded (and
usually were in writing) by the particle ma (a always short). In indirect questions ma
was always present (equivalent of E if, whether). This ma could similarly take pro
nominal affixes when the previous verb was not repeated. But note that a previous
negation was also held to be repeated. So that in such a case as 'you are to come/ are
not to come' or 'you must say this/must not say this' or 'he did it, but you did not/ he
did not do it, but you did', the rejoinder manye "what? me?'' would be equivalent to 'am
I/ aren't I' or 'must 1/mustn't I' or 'did I/ didn't 1?'91
88 This is a typescript clipped together with the manuscript of "Wil� shall and the future" (see below) and enclosed
in a newspaper dated October 9, 1969. Excerpts were cited in "Negation in Quenya" (see footnote 58 above).
89 This note was written in pencil in the margin to the left of the preceding paragraph.
90 The second example was emended in ink from: 'Will you come?' - '(No) I won't'.
9 1 The literal gloss "what? me?" was inserted later in pencil.

But in GL man is a pronoun = who? If Varda was meant the verb should have
been enquantuvas. 92
Where ma was used it had to be placed immediately before the subject or a verb with
a pronominal subj ect inflexion, and this group normally came first. If some other
element was given special emphasis it could precede ma, but not intrude between it and
the subject. Thus if Galadriel had asked: 'Will Varda now refill the cup for me?', the
normal order would have been: Ma Varda enquantuva i yulma nin, with sz placed either at
the end or before i yulma; more emphatic for s( would be to place it after Varda; quite
exceptional emphasis would allow s( to be placed first of all. Where the sentence
contained an interrogative pronoun as subject the order was freer, but the order in GL
is not normal, since the object is usually placed after the verb, unless great emphasis is
placed on it (not the case here), and even so that is more frequently shown by tone
only. The emphasis is in GL on s{, now, since Galadriel is thinking of her present state as
a 'rebel' who refused the pardon and recall of the Exiles to Valinor.93
The negative la simply denies that the positive statement is true in fact. Q. possesses
another negative derived from the base ABA 'refuse (an order, request, petition); prohibit,
discountenance another's proposed or likely action'. In Q. the derived negative takes the form
va, often reiterated as va vci, vava.* This as a negative interjection, standing alone, is a usual
equivalent of E. 1 I won't; 2 don't, you shan't; 3 he shan't, he mustn't.
* Cf. S boda < bata- 'refuse, forbid'.
In this simple use va always expressed the will or wish of the speaker, and referred to the
immediate present or to the future; it could only be fully understood by one who had heard the
words (order, request, or statement of intention) to which it was a rejoinder. Like la it could be
combined with pronominal affixes, assimilated to those used in verbs; but the long vowel was
retained before two consonants, since va was always emphatic.*
* For the 3 pers. dual and plural only the number-inflexions -t, -r were normally used. In
the sg. only the long inflexions: -nye, lye, tye, se were used. Since va was naturally
peremptory and not used where courtesy was required the second person singular
usually had the form -tye. In the first person dual and plural only the exclusive forms
were naturally used; since the rare cases where a speaker could associate those
addressed with himself in a peremptory refusal or command he would be excluding
others than 'we' (present or not) to whom the reply va was to be given.
Thus: vanye, I won't; vatye, don't! you are not to; vase, he mustn't, he is not to; valme, we won't,
we say no; valde, you are not to (do so); var, they are not to, etc. In the first person plural and
dual some hesitation occurred. The exclusive form was naturally correct in most cases: the
speaker associating himself with others of the same mind, not those addressed. But if va was
used in answer to a request or order using the inclusive 1st person, as e.g. '(I say, they say) we
are to go', valwe might be used, including the person addressed. This was naturally most likely
when matters were in debate, and questions were asked rather than orders issued or decisions
announced. Thus a member of an assembly or council might say in answer to a question ma
karuvalwe ohta? shall we make war? - valwe or in full va karuvalwe (ohta), we will not (make
war), asserting as it were that there would be a unanimity of decision when the matter had
9 2 This note was added in pencil in the margin to the left of the discussion of ma at the beginning of questions.
93 The end of this paragraph (from "unless great emphasis is placed on it ... ") was added in the second typed
version of the text (see footnote 98 below).

been fully considered.94 But if there was a recognized division of opinion, the rejoinder to such
a question by a member of the war-party might be vtflme or vtfnye ar vtflme = "we or I and those
with me (of the peace-party) are against it."
The use of vtf with any other affix than that of the 1 pers. sg. was, however, infrequent. In
all circumstances vtf was abrubt and emphatic, and expressed the will of the speaker. Associ
ating other persons with this will needed special justification.
Where the reference of va required definition, it could be used as a negative particle in a full
sentence, always preceding the verb.95 This was in the future normally: as e.g. itas vtf tuluvanye
= in that case I won't come; vtf tuluvc:fse = he is not to come, he shan't come; vtf matuvatye
mastanya = you are not to eat my bread, you shan't eat my bread.96 The aorist could be used
when the time was indefinite, that is general (including the immediate present). Thus vtf karis i
'he is not to do it (now or ever)'; va meninye 6 le 'I won't come with you'* in a case where
immediate departure is concerned. 97
* E. says 'come' in such cases where there is a question of accompanying others. Q. uses
'go' of movement to or towards any point other than the 'here' of the speaker, actual or
reported within a narrative.
In such a case the present may sometimes be used to represent immediate urgency: vtf menanye
'I am not coming (I say)'. The past tense or perfect are never used, since in reported speech the
tenses of verbs were not shifted in Q. In any case va was not used in reported speech, except of
course in oratio recta when the exact words of a speaker are repeated or represented: this was
at all times more usual in Q. than oratio obliqua. 98
94 Typed original: ma ohta karuvalwe? » ma karuvalwe ohta? and ohta vd karuvalwe » vd karuvalwe (ohta), in pencil.
95 The use of va in this sentence is probably a slip for vd, and similarly at three subsequent occurrences below.
96 In the word mastanya in the third example in this sentence Tolkien placed pencil strokes before and after the
letter t, and below he wrote "se ?" in pencil, apparently indicating an alternative form massenya for 'my bread'.
97 The typed original sentence: va meninye yo le » va meninye o le.
98 The central section of the text, after the long note about questions and up to this point, is a typed replacement
for an earlier version typed on a separate page:
The negative Id simply denies that the statement is (was, or will be) true in fact. Q. had another negative
with stem BA. This expressed the will or wish of the speaker in refusal or prohibition. Thus as a negative
interjection: in the forms va, a va and reiterated vd vd, vdva it was equivalent to 1 pers. I won't; 2. don't, you
shan't; 3. he shan't, he mustn't. To make the reference clearer vd could like Id take pronominal affixes, but
owing to its emphatic and emotional force always retained the long vowel, though as with Id in the sg. the
long forms are always used.
Thus vdnye, I won't; vdlye or vdtye, don't, you are not to (do that); vdse, he must not, he is not to. As a
negative particle with a fully expressed verb vd is naturally less used than Id; but it can be used in such cases
as these: 'in any case (whatever may be said, desired, or may occur to the contrary) he is not to do it': vd karis
i. In such cases as these, and in any where the refusal or prohibition does not apply to an immediate situ
ation, the verb is usually in the future: vd karuvas i. Other examples: vd tuluvanye, I won't come (I am not
coming, going to come - where these are clearly expressions of refusal); vd tulinye would only be used in
some situation where an immediate start was necessary, and others were probably waiting.*
* In the latter case, if there was any emphasis on the pronoun, va inye or vdnye with uninflected verb
would be used: 'as for me, I'm not coming'. In such cases both la and va prefer the attachment of the
pronoun to the use of the separable emphatic pronouns inye, elye, isse, etc.
la tuluvanye/tulinye would only be used where foresight or knowledge of some reason or obstacle preventing
my coming, independent of my personal wishes, was in mind: itas la tuluvanye, in that case {he won't »} I
shan't come (something will prevent {him »} me); itas vd tuluvanye, in that case {he won't »} I shan't come
(something will have been said or have occurred to make him refuse).
At the top of this page is a note in ball-point: ".fABA, verb. avan, avan (aveant), auvan, avanent (aunen), avdvien."

When the refusal or disapproval of other persons than the speaker were [expressed] the
form va could not be used. It was then necessary to use the full verb derived from base ABA.
This was an a-verb,* of a type which originally appears only to have differed from the normal
in having a stem ending in -a (not -i) in the aorist.
* Those in which the vowel of the aorist was short a and not i. For the probable origin
and original significance of this difference between the normal 'strong' verbs with i
aorists and those with -a, or -u in the aorist, see Quenya Verbs.
The type occurred especially where the stem was not in origin verbal; though an additional
consonant t- or y- was in later formations usually added* before the -a.
* As in oro/ro, (high) above: orya-, rise; undu /ndu, down: *ndiiya, Q ntiya-, descend; ana/na,
to (of movement that reached its object): anya-, arrive at, reach. The Q. causative anta
(*anta-) 'cause a thing/person to go to an object, send' usually has the sense 'give', but
this is due to blending AN with the unrelated verbal stem HAN 'give', which in S. lost the
initial breath h of CE, that in Q., T. remained h- (T hanin 'I give').99 From this is derived
Q anna and anwa 'gift', S anw and strong verb (lost in Q.) aned, pa.t. 6nen, T hanna. HAN
properly had the sense 'enhance, enrich, add to'; its direct obj ect was therefore the
recipient, and the thing given was instrumental, expressed either by a preposition or
an inflexion. This construction (like E. present one with) was retained in T., but lost in
S., but retained in Q. when referring to great presentations (as e.g. endowments, or
gifts of land etc.).100
The verb *aba-, Q ava-, refuse, forbid, was thus derived from a primitive exclamatory ba /aba;
and was probably an early formation, though the absence of a verbalizing consonant ...
99 This sentence was originally typed as: "(The Q. causative anta- (*anta-) 'cause a thing/person to go to an object,
send' usually has the sense 'give', but this is due to blending AN with the unrelated verbal stem 3AN 'give', which in
Q. and S. lost the initial spirant 3, that in T. became h- (T hanin 'I give')." In connection with this Tolkien added the
following annotation in pencil, below the typed text:
no. S. retained 3 {but lost} and strengthened it to g.
Q. lost 3 - as also did T.
The stem, to add to, to [?enhance] is han with primitive {back} breath h retained only in Aman (Q., T.) not s .
Tolkien revised the sentence to the version given above, and also emended 3AN » HAN in the subsequent text.
100 The statement about Quenya in this sentence was added later in ball-point. The typescript ends here;
the last
sentence in the main paragraph (typed before the note) was never completed .

[Rough notes following the rejected typescript:]

1 only to disappear after e, i. aya > ea. w remains [?after] i, e. y [?after] o, u.
tJ [only to disappear after] o, u. awa > oa.
ABA. verb is[?] a-verb. avan, avean, {avauva, avavien »} auva.
avan, {avean, avan,} avuva > auva, {aune »} avane (aune), {ava... »} avavie.
galan, {galean, ga » galan,} galuva, tgalane, {gale,} agalie, gagalie.
Q alan, {alean, alan,} aluva, talane (alde), alie or alalie.
ta causative & ta when not made from ver[bal] stem.101
tulta, fetch. tultan, tulteanye, tultauvan, tultanen, utultien. ' taye. tayen [>] te1 en > tien.
{talta, fall. talt...}
ista, know. istan, istianye (istyanye), istuvan, (i)sinte, isfsie.
Make[?] Q. ea as present[?] [?tense] invaded[?] other[?] [?forms].
alan, alan [aleant], aluvan, alanent [talden], alalien [alien] .
faran, faran [-ean], faruvan, faranen ({amen), afarien.
avan, avan (avean), avuvan > auvan, avanen (aunen), avavien.
auta, oan, go away. oa. oa. .. autya. oante. awawe, oavie.102
nahatan.
[Rough notes on the back of the rejected typescript:]
-ta/ya causatives origin[ally] had -ta, ya.
-t/y intransitives or formatives had -t, y addition but used a in aorist-present.
aor. pres. cont.103 future past aorist[?] perfece04
thus raise, orta ortan orotan ortauva ortane orortaye
orta-, rise ortan ortian < -tiyan ortuva oronte orortie
oryan oryean oryauva oryane (or)oryanie
oryan orian oryuva oronye ororie
talat taltan talatan taltuva talante atalantie
nahta nahta nahtia nahtuva nakante anahtie
kaita kaitia kaituva kayante, keante akaitie.
fall is taltaya > taltea. talatta, talatya. talasta, talatya.
ista, {sf. .. »} istia, istuva. {oryia > o... } oriy(y)a. istya. nah...
tulya, tulyea, tulyauva, tulyane, {utulye »} utulya.
ista istya istuva (i)sinte isistie, istie is{sie
orta ortya ortuva oronte {ortie »} (or)ortie. or6rie.
sirya sfria siryuva sirinye is{rie
orya oria oryuva oronye ororie.
�
1 01 This note was added on the right, and the three following sets of forms were listed in the left margin.
102 Several tentative forms were deleted here: o... autya, awatya, auta-, avoant...
1 03 Above the heading "pres[ent] cont[inuous]" is a form ortean, and beside this aya (presumably the source of -ea-).
1 04 Several of these perfect forms were revised from originals: orotaye » orortaye; orortaye » orortie; ororyaye »
(or)oryanie; atalatie » atalantie; anakie » anahtie.

[LVS 14]
Will, shall and the 'future'.
'will' in the full sense: a fully conscious purpose (to the accomplishment of which a 'person'
would apply himself with vigour of mind and body), was derived from a base vNID 'force,
press(ure), thrust'.
Q indo < inido/in'do : the mind in its purposing faculty, the will.105 The basic stem NID-,
however, was represented by verb nirin, nfran, niruvan, tninden, nimen, in{rien, which meant
press, thrust, force (in a given direction) and though applicable to the pressure of a person on
others, by mind and 'will' as well as by physical strength, could also be used of physical pres
sures exerted by inanimates: n{re was the general word for 'force', from which was derived
n{r{te, forceful, exerting great thrust or pressure, driving. In the special application to rational
will the derivative verb (from indo) indu-, pa.t. indune was used: e.g. indunenyes, I willed it, I did
it on purpose.106
cf. induinen, n. purpose. cf. the expr[ession] Eru-indonen, by the will of God.
cf. turindo, purposeful mind, strong-will, as name Turin. turindura, done necessarily.10 7
In T. and S. NID became used exclusively of 'rational' purpose and will. S nfo meant 'full
vigorous purpose, or the exertion of will' (= in form Q n{re); and from it a weak verb n{oa was
formed and applied to exertions of strong will in major purposes (am determined to). The
older strong verb niOin, pa.t. enioen was weaker, and sometimes became no stronger than 'will'
in E. (when that is used [to] imply the wish or intention of the subject): I will do it, I mean to do
it, & so could operate sometimes almost as a future auxiliary: nioin mened, I have a mind to go, I
intend to go. The older *indu becoming ind > inn, was used of some particular purpose or inten
tion of an individual (not necessarily of great strength or in a maj or matter): as in inn da v'im
"there is an 'inn' in me" = "I have a good mind to (do so)."108 T nfde = S nza in form and sense;
but the verb nid- was as in Q. used as 'urge, press, force' generally without limitation to opera
tions of mind. indo, however, as S ind, inn was used of a particular will or intention of a person.
2. to be willing, consent, agree: positive of vABA 'refuse' but naturally less often so em
phatic. It was, however, similarly constructed. The element found in Eldarin appears to have
been an old one, a 'mono-consonantal reversible' r.A. /At>A.
Its basic sense was probably 'be helpful',* be willing to assist, in any work etc., agree, con
sent.
* Though this sense is obscured in the uses of apa- with other verbs, it may be supposed
to be the earliest sense, because of derivatives made from it, such as Q apea ( < apaya)
name (as = 'beneficial' ?) of a herb - not now known or at least identifiable, useful in
healing both (as an infusion of its leaves) in 'easing the pain' and hastening the healing
of heavy blows or shocks, and as an invigorating odour in reviving the sick from de
pression, or even unconsciousness, after wounds or shock. 109
105 The original forms inido/in'do » inidu/in'du, and perhaps in connection with this the notes on induinen and Eru
indunen were added (see below); later the original forms inido/in'do were restored, and perhaps a this point Eru
indunen » Eru-indonen.
106 Excerpts from this paragraph were cited in "Notes on 6re," ed. earl F. Hostetter, VT 41, p. 17 (n. 10).
107 These three notes were added in the margins, the first to the right and the second and third to the left of the
preceding paragraph. In the second note the expression was originally Eru-indunen (see footnote 105).
1 08 In this sentence the original Sindarin phrase: inn no v'im » inn da v'im, in pencil.
109 In the space around this paragraph Tolkien later added an annotation, using a different ball-point pen:

The Sindarin form athelas is probably derived from Q athe-a + las 'leaf'. If so, it must
have been derived from medical lore, the uses of the plant being known or having been
discovered in M-E. only by the Noldor: apea had become asea in Exilic Quenya, but p was
restored in the specialist learned language of lore-masters. The S. form may even have
been derived from a written form 1J51 (not 1') l). In S. the simple verb stem apa > atha
was no longer used as an independent verb; but related to it must be S natha- (probably
aphetic < an-apa), bring help to, save, rescue. In T. apa- was a full verb, but had devel
oped the sense > side with, ally oneself with, assist, serve. (aparo was an ally, assistant
or satellite. Cf. Q aJ:mmo [< *athmo] 'good companion', friend at need, friend with shared
interests, colleague.)110
The interjectional form as in Q pci, apa may (as probably in case of va, etc.) have been original
and a verb later developed from it; but more probably apa, pa developed from a verbal-stem
before the pronominal elements were affixed as inflexions.*
* Note, unlike la (no), va (I won't) pa did not allow pronominal affixes of verbal form: in
place of lanye 'I don't'; vanye 'I won't', only pa or uninflected apa appear - unless a fully
inflected verb is used, as apanye, I will.
In Q. the forms used were ):>a, a):>a in response to requests, orders, or expressions of intentiont
expecting agreement or disagreement.111 The latter was the more forcible, but was most used
when the speaker regarded the question as unnecessary. 'Close the door, please! (a pak' i fende,
mekin)'; 'very good/all right/I will (pa)'.112
t not opinion[s] of fact which are answered by nci, nasa, it is (so): yes, as in rokkor i
Erulingaron mare nar (ma naite)? The horses of the Rohirrim are good (are they not [lit. is
it true]) . To which the answer is nci, nasa, it is a fact (= yes), or nar (naitie), they are
(indeed = truly).113 But if mara tyen, good to you = our "like," the question would mean:
'Do you like their horses - you like them, don't you?' The answer might then be apa "I
agree of course. Of course I do."114
But the full verb was naturally often used as in a):>anye, I will (do so) - as you ask, or order.
The full verb was used with other verbs freely, in the senses I will, I am willing, I agree (to do
so). The negative was ava (va), but only in strong or curt refusal. More polite was la apanye, I
don't wish to, I don't agree.
This needs correction. ..fATHA = be willing, agree, consent, grant - assist, join with others[?] . It is HATHA, Q
hapa- that meant treat kindly/make easy, (help to) cure.
110 This note was added in ball-point at the bottom of the page and continuing on the back; following the end of
the note are various hastily written additional forms (cf. the annotation cited in the previous footnote):
hapa. Q., T hapa-, treat (medically). {hathro, hath... »} haoro, hadro.
Q {asea » } apea. treating[?]. medically [?use], caring[?] . - {apumo » } apar(o), doctor, leech.
S {gathra. gathur[?] »} athra. athur[?]. Q. to treat medically. leech, doctor. aparta, asartar[?].
111 Originally three forms were given, }>a, a}>a, a}>a (the third later deleted heavily); and the phrase "expressions of
opinion" » "expressions of intention," both revisions being made in the original ink.
11 2 The Quenya sentence a pak' i phende, mekin was emended to a pak' i fende, me kin in the original ink.
11 3 Tolkien struck through original "nar 'they are' (naitie, truly, indeed)" and in the left margin he wrote: "nar
(naitie), are (indeed = truly)." Below this in the margin he added a further annotation:
Verb nii- is used to assert qualities etc. of separate things in the Universe, verb eva (ea) to assert their actual
real existence extra-mentally.
114 This note was written in ball-point in a space left in the original composition. The bracketed phrase in the
translation of the first question is Tolkien's.

In S. the verb apa, atha had become agglutinated to the verb stem, and formed a kind of
'future', expressing the intention of the subject, closely resembling in sense and uses English
will (when not mere future): I will (I'll) go, he will (he'll) go, espec. in the 1st and 3rd persons.
In the second person the implication of 'will' of the subject is clearest in questions or negation.
Cf. song in LR, linnathon 'I will sing, I intend to sing'.115 This was a fairly late development, as
is shown by the fact that apa, apon etc. could still be used with ellipse of the verb stem, as e.g. in
linnathol ? will you sing (please), answer apon, I will. Apart from this athon "yes, I will" (or in
plur. athof, athab) the verb atha- was no longer 'free'.
In T. as noted above apa- had not been connected with the verbal system, and apane meant 'I
help, assist, support' and so often = I agree.116
A future 'tense' (with inflexional element added) had evidently not been devised in CE.
S. had no pure future tense, but used the verbal vTUL as an auxiliary of the future (unrelated
to personal wish or intention): see below.
T. used the karan form as present (continuous) & consuetudinal. The aorist form karin was
strictly 'undefined' & so was normally used when any other indication of time was expressed
for future and past (e.g. words = soon, ago, tomorrow, yesterday, now, later, in time, 'to come'.)
[Colloquially often when such words were absent.] Where a need to express the future ... 11 7
Q. developed a pure future of fact or eventuality, with a stem uba suffixed to the bare base
(without 6mataima): as *karuba-, will do, is going to do. This base uB provided a verbal stem (an
a-verb) uba- in original sense: impend, be imminent, approach, draw near.118 In this form it
was at an early date affixed to the simple verbal stem, but with a weakened and generalized
sense: as e.g. *kar/uba/ni = I draw near to doing, I am about to do (it) > CQ caruvan(ye), I am
going to do/shall do (it). The sense of imminence was lost; and no special sense of unwelcome
approach was developed, such as appeared in Q. partly perhaps owing to influence of the 'bad'
sense of a-prefix, and its derivatives.
As an independent word the verb uba was lost. But a strengthened form iiva was used
'impend, be imminent' nearly always in a bad sense: 'threaten (to come)'. Thus ulo ava, rain
(unwelcome) is coming/threatens; but uluva, it is going to rain, it will rain;119 hrive ava vena,
winter is drawing near (to us), which would not be said of laire 'summer' - unless for some
11 5 The second gloss was originally 'I mean to sing', which was emended in ink to 'I intend to sing'. (Cf. LR I 250.)
116 The phrase "and so often I agree" was added later in pencil. In the margin above the beginning of the
=
discussion of pci, atha Tolkien compiled the following tabulation in ball-point pen:
ba-kwet Q vaquet-, refuse, prohibit. say, no. ava-, v[erb] refuse, forbid.
laquet Q. say 'it is not', deny fact, accusation etc. ala, deny.
paquet Q. say 'yes' (saquet), I will, agree (to do). apa, be willing.
naquet say 'it is', assent, agree (on fact). but na, it is.
11 7 This paragraph added in the left margin was never completed. (In it the bracketed sentence is Tolkien's.) He
composed the paragraph partly in pencil and partly in ball-point pen to replace an original, against which he
wrote an "X" in pencil, although he did not delete it:
T. used the form, which in Q. and probably in its origin was a 'present continuous': with lengthened base
vowel, and sund6ma [se. tense-vowel ?] -a, as in Q karanye, I am doing. T. used the 'present aorist' (Q karin) as
the only inflexional present; aorist, consuetudinal, or immediate; but also auxiliaries [added in ball-point:]
not participles.
Perhaps in connection with the revised paragraph, the following adverbial bases were noted in the top margin:
EPE, after of time, or[?] after of place. AWA, before of time, OPO, before of place.
11 8 The order of these meanings was revised in ink from original: "approach, draw near, impend, be imminent."
11 9 The original example ulla iiva » ulo iiva, in pencil.

special reason its approach was unwelcome. Derived also from this stem is Q umbe (< ubme)
'imminence/threat', which however in Q. usually referred to the expectation experienced by a
person: "premonition, fear, unwelcome expectation" as e.g. in umbe nin i hrive nauva urra, I
have a feeling that winter will be bad (si loa, this year).120
uva when conjugated had only the forms uva aorist, present and future, and the pa.t. uvane.
The limitation in Q. usage to approach is probably the original CE meaning.121
The stem or derivatives are not found in 5.122
S. lost the stem UB altogether, and used vTUL, come, approach, for both coming in space and
in time. The 'future simple' was expressed by using vTUL as an auxiliary, as in tolen cared 'I am
coming, drawing near to eating, I am going to eat/ shall eat' .123
T. the stem UB only survived in a bad, ominous sense like Q umbe. uba was usually im
personal, "it is on my mind that it will happen, I feel sure it will happen," mostly of things
unwelcome.
[Rough notes in ball-point:]

{6.} ho. hamen. lomentie. with, y6(m). won't, u. omentie.
{epe, before of time.} {awa.}
a/na, to(wards).
o f ha, from. [?being] {near, behind, past[?] »} as described[?] by a [?genitive] [???]
l lo, from. a place within[?] the [?world] [???]
epe, 'after' {before} of time. after, behind of place.
awa, before, ago of time. opo, before, ahead, in front of place.
auta, ago. autas, a former occur[rence] .
to, on, [???] above, on [???]
tosse, [???]
nu. {6a.} yo. with.
1 20 The original example umbe nin i hrive nauva unca » umbe nin i hrive nauva urra, in pencil.
121
This sentence was revised in ink from the original: ''The limitation in Q. usage to approach in time is not un
natural since approach in space normally occupies measurable time, so that 'he is on the way/is coming /is
moving towards us and getting nearer in space' is nearly [the] same as 'he is going to come/arrive, he will arrive."
1 22 An original paragraph following this sentence was struck through in ink:
In {Q. » } T. uba is usually applied to space: it is closely near at hand in {in space); ubie, nearby region{s),
neighborhood, district, ubra, adj . near, closeby, ubrie, nearly/almost.
In the top margin Tolkien added a note in pencil:
Q. limited to time owing to fact that approach in place takes time, but {Q. » } T. applies to space.
1 2 3 Tolkien deleted m before cared, presumably a false start on the Sindarin infinitive 'to eat', and in haste he used
"eat" in the English glosses, which should be 'I am coming, drawing near to doing, I am going to do/shall do' .

The Feanorian Alphabet and Quenya Verb Structure ]. R. R. Tolkien
List of Abbreviations
adj., adjs. = adjective, adjectives. LVS 1 ... LVS 14 = "Late Notes on Verb
1£. = J£lfwine. Structure" (see p. 141).
J£S 1 ... J£S 6 = "On lElfwine's Spelling" M-E. = Middle-earth.
(Versions 1 to 6 - see pp. 56-7). n. = noun.
aor. = aorist. N. = Noldorin.
AQ = Ancient Quenya. NB. = nota bene, note well.
av. = adverb. no., nos. = number, numbers.
bilab. = bilabial. N.W. = Northwestern.
C. = Common. OE = Old English.
c. = circa. ON = Old Noldorin.
CE = Common Eldarin. OQ = Old Quenya.
cf. = refer to. orig. = originally.
cons. = consonant. OS = Old Sindarin.
cpd., cpds. = compound, compounds. p., pp. = page, pages.
CQ = Common Quendian; Classical Quenya. part(s)., parte., partic(s). = participle(s).
du. = dual. pass. = passive.
E. = English; Eldarin. pa.t. = past tense.
ed. = edited by; edition. PE = Parma Eldalamberon.
e.g. = for example. perf. = perfect.
EQG = "Early Qenya Grammar." pers. = person.
esp., espec. = especially. pl. = plural.
etc. = et cetera. PQ = Primitive Quendian; Parmaquesta.
Etym. = The Etymologies (see p. 57). prec. = preceding.
EVS 1 = "Quendian & Common Eldarin pres. = present.
Verbal Structure." Prim. = Primitive.
EVS 2 = "Common Eldarin: Verb Q., Qu. = Qenya, Quenya.
Structure." QS = "Qenya Spelling."
fut. = future. q.v. = quod vide, which see.
G., Gond. = Gondolic. QVS = "Quenya Verbal System."
ger. = gerund. R. = Rumilian.
GL = "Galadriel's Lament" (LR I 394). S. = Sindarin.
i.e. = that is. sg. = singular.
lE, 1-E. = Indo-European. se. = scilicet, to wit, namely.
impf., imperf. = imperfect. T., Tel. = Telerin.
inf., infin. = infinitive. TQ = Tarqesta, Tarquesta.
infl., inflec. = inflected. vs. = versus.
length. = lengthening. vol. = volume.
lit. = literally. VT = Vinyar Tengwar.
LR = The Lord ofthe Rings. Y.S. = Year of the Sun.

1, 2, 3 = 1st, 2nd, 3rd person.

& = and.
&c. = et cetera.
* = unattested.
t = archaic or poetic.
v = base or root.
§, §§ = section, sections.
< = derived from (historically).
> = shifted to (historically).

« = changed from (in the manuscript).
» = changed to (in the manuscript).
{ ... } = deleted (in the manuscript).

[ ... ] = phonetic spelling; added editorially.
Tolkien occasionally used brackets to mark text
he was apparently considering for removal.
Editorial brackets are placed around missing
words needed for the sense; expansions of
uncertain abbreviations; and words that are
unclear in the manuscript (the last being
indicated by a question mark) .
*, t, tt, 91, 9191, �& = signs used by Tolkien to

mark his own notes in the manuscripts.
italic or bold text (the difference is editorial)

is used to represent Tolkien's underlining for
emphasis, and to highlight sounds, words and
phrases of Elvish and other languages. (Also
see above, p. 8.)

J7
,, J
'hu<c'•
J
'tttttlalcU<.' •
u. U
tf
,o
I
l&.
• '"l:lvJa "-r• "'"'"'" � t'lu Jl,..,.,u -"""''"'9"' •,.,_ ' _... .... � ak-'
"'•" clt..•wr"'" •f ""- V"•l•�e. , f....hl
.., ,\o "" e�w v.;"wal.o
•
Vowe.l.S . .h· wil.l be. -h4\ Ho,.,.t- t""-C ..c. "' ""' � 2 vn •nl o'C
t. �.,.. ' CA".(o1 e��... d. • o �1 · � .. ;....., 1,,._ .IJ """"'... ,. .. -.ce
...1\w c'k 1111 .,.c \c--1:" (!c.c"'"'" �,,...,. ,. �•., "" c'" ._,...,... •h-vnc.) Jkcw(- \nlw&kt
�""c.'l\y ) I�U •l.o,.t o .. ....;.. u . rau.· c• � �, ..,,\1 ""''
Q.,,ruut\ t,y a .· .. �\o!.·u r'•c.t d. .. � ..,... 1 ., M.,. .s\ul,.� ' � .. l.cov•
- .. 1 ..
� le·� � ....1:"'('
.. "" '" (\,.. o. t'\oih'c �Aut"• �r"'ss (a., a 1 ;
�,.)(• c\ ; ) �ftt»e..A: te ......,., � � «ttp'i'•U (t.,'"£1 i ��

,.,..""'- • " { {
t""y ..,.r.,... , (c , r J . "r1.... f"U. �• .., ..t. 4·,4- c. ��
' J (" 0 '1 J J fB 0 . ,"" ll..s VA1.\l. Ca.1 ,

l:\. i 8 o 1,.1. a. C '- o t!
i) f.'ot"'•"'•lly �"eh..- t' 1<owo'\oi'o.tly , ,. ...(\ 1 .. 'tw W.l\L'I (ctl J la
rw,.. "'�tr 'W't'i�'" J

.. � -""·..t' .

PE22

Uploaded by

Copyright:

Available Formats

PE22

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PE22

Uploaded by

Copyright:

Available Formats

Sec.u;� C\"11\A. l,..n,'\SC't"tp'h'tw' .

'""' "'"'� -coAAf'"c .. .... of »t& f'&ftW'o�.. .... .,.'1,.,..,,

OftCo'K c�•..,.c.h""'-'"'11" t'h.c A1f'ka1nts. .H••• c•"''Y n..a __,.,.,

'""C'he. �&'tt•�• t.Ucc\ ...,.,.1 �g '""�"'"-'bet" , .-ec.."tct\ .- ... t'h�

f-""""-·� ht \&low. "l':'h.c � of t'k.o \.tt:t- c:; t'-4 �"

�'"" �l\\& J .., � •(-.,1 . o .. tlo& vu.hi.C •f 1 "a.,.' ,lu. vauJ. .

a gn .. 6 11., ..-m ..J;r.; .. po •Y

pARCDA elDALACDBeRON 1\All

Lhe �eanoRian Aiphabec ··

Quenya VeRb StRuccuRe

Lhe �eanoRian Aiphaber

Quenya VeRb SmuccuRe

Copyright© 2015 The Tolkien Trust

MOUNTAIN VIEW, CALIFORNIA

We would like to thank our colleagues Carl F. Hostetter,

We are indebted to the Tolkien Trust for permission to

PARMA ELDALAMBERON No. 22 - Page 2

The Feanorian Alphabet, Part 1

Qenya Grammar- Spelling and Transcription 61.

On J£lfwine 's Spelling

Quenya Verb Structure

Quendian & Common Eldarin Verbal Structure 92.

PARMA ELDALAMBERON No. 22- Page 3

Quenya Verbal System- Conjugation & Syntax 99.

Forms of weak verbs, or verbs with weak aorist-present 113.

Irregular Verbs 122.

Common Eldarin: Verb Structure 128.

The Verbal Stems 129.

Adjectival and Substantival forms of Verbal stems 136.

Imperative forms 139.

Late Notes on Verb Structure 141.

List of Abbreviations 169.

PARMA ELDALAMBERON No. 22 -Page 4

Edited with introduction and commentary

PARMA ELDALAMBERON No. 22 - Page 5

PARMA ELDALAMBERON No. 2 2 - Page 6

PARMA ELDALAMBERON No . 2 2 - Page 7

Date of the Manuscripts

PARMA ELDALAMBERON No. 22 - Page 8

5.* ln oa b da cd 3a t:d 3wa

PARMA ELDALAMBERON No. 22 - Page 9

7. Ul. ma Q:) na c..c..r ga a::r gwa

PARMA ELDALAMBERON No. 22 - Page 10

The Lindarin Use.

PARMA ELDALAMBERON No. 22 - Page 11

The Lindarin Use.

(a) p. (b) t. (c) ty. (d) k (e) q.

mba nda ndya nga16 iigwa1 7

PARMA ELDALAMBERON No. 22 - Page 12

The old s-sign 6

1 7 ngwa » iigwa in red ink.

PARMA ELDALAMBERON No. 22 - Page 1 3

and change of):> > s) cl b were used for ht, st."

29 Tolkien later added "see below" at the end of this sentence.

PARMA ELDALAMBERON No. 22 - Page 15

Table of the Parmaqestarin letters & values

similarity of the letters in form & function.

A. Parmatema, p-series: 1. fL p. 2. b.. f. 3. t1. v. 4. 1%1 m. 5. ln. mp. 6. F mb.

E. Napannar, additional letters: 1. � hy. 3. J short vowel carrier. 4.

PARMA ELDALAMBERON No. 22- Page 16