Negation in the Elvish languages is a controversial topic, in part because Tolkien kept changing his mind on how it worked. For much of his life, he vacillated between two basic forms of negation, those based on a root √Ū (or √UG or √UM) and those based on a root √LĀ (or syllabic ḷ). Sometimes the two forms of negation coexisted, and other times he preferred one over the other. For a lengthy discussion of this vacillation, see the entry on the Quenya negative. In some respects the picture in Sindarin and its conceptual precursors is simpler, in that Tolkien seems to have clearer roles for the two different negation systems for most (but not all) of his life.

Conceptual Development: For clarity, I’m going to separate the discussion of Tolkien’s use of la-negation (from √la) and u-negation (from √Ū) in Sindarin and its conceptual precursors, followed by a third system of baw-negation that Tolkien did not introduce until 1959.

la-negation: In his earliest writings from the 1910s, derivatives of primitive syllabic ḷ were used as negative prefixes: G. il- < ᴱ✶ḷ- denoted “the opposite, the reversal” (GL/50). It took the form ul- before labials like m, b, w and was therefore sometimes confused with another prefix ul- denoting wrongness or badness. Simple negation was indicated by the prefix u- (see u-negation below).

There are no signs of la-negation in the Early Noldorin of the 1920s, except perhaps for a negative prefix ᴱN. i- mentioned in Early Noldorin word lists (PE13/148) that might have been another iteration of G. il-. In The Etymologies of the 1930s, however, this prefix returned in the form N. al- “no, not”, still developed from syllabic ḷ, but now representing simple negation rather than opposition (Ety/LA).

This prefix made it into The Lord of the Rings in the flower name alfirin (LotR/875), untranslated in LotR but translated “immortal” in a 1969 letter to Amy Ronald (Let/402) and in linguistic notes from this time period (PE22/153, 156). It is very likely that “immortal” (al-firin) was the original intended meaning of this word, but for most of the 1960s Tolkien went through a period where he rejected la-negation. In the early-to-mid 1960s he explored several alternate etymologies for this word, including alph “swan” + irin “?” (PE17/100) and al- “well” + pirin “blinking” (PE17/146).

u-negation: As noted above, u-negation was used for simple negation in Gnomish of the 1910s, most notably in the negative verb G. û- “not to be, not to do” (GL/73). It appeared in the sentence G. û gwilthi – gwilm ’othwenion (GL/45), untranslated but apparently meaning “*it is not peace — [the] cessation of battles” as suggested by the editors of the Gnomish Lexicon. The Gnomish Lexicon mentioned a second negative verb ug- (GL/74), but I suspect this was just the primitive form of û-. U-negation also appeared as the Gnomish negative prefix u- (GL/50, 75) and there was a construction u — u “neither — nor” as in u ’wirn u ’wethrin “not unwelcome or welcome” (GL/47).

U-negation reappeared in Early Noldorin of the 1920s in the phrase ᴱN. {madol û sóg >>} au sóg madol (PE13/128), untranslated but apparently meaning “*he did not drink (while) eating” as suggested by the editors of the Early Noldorin Grammar (PE13/128 note #77). It also appeared in the (rejected) privative prefix ᴱN. um- in Early Noldorin word lists (PE13/155). This was replaced by the privative prefix ur- or or- as in the word ornoth “countless”, a prefix that eventually evolved conceptually into ar- as in Nirnaeth Arnoediad “(the Battle of) Unnumbered Tears”.

In The Etymologies of the 1930s, u-negation appeared again in the negative prefix ú- which in this period had an extra “unpleasant” connotation, both “un-” and “bad” (EtyAC/UGU). This prefix shows up in a couple words from drafts of The Lord of the Rings from the 1940s that did not make it into the published text: ubed “denial” (WR/137), Uvanwaith “Nomenlands” (TI/281), and uluithiad “unquenchable” (SD/62). These negatives from the 1940s seem to be more neutral in meaning.

The best known Sindarin negative is the phrase ú-chebin estel anim “I have kept no hope for myself” from Appendix A of The Lord of the Rings (LotR/1061). Here the particle/prefix ú- is used for verbal negation, and seems to be neutral in tone. It appears as a negative prefix in various words from this period such as Udalraph “Stirrupless” (UT/313), Úmarth “Ill-fate” (S/210), and úgarth “trespass [misdeed]”, the last appearing in plural form úgerth in Sindarin prayers from the 1950s (VT44/22). Some of these words had the additional “bad” connotation of Noldorin from the 1930s.

The best description of u-negation comes from “Definitive Linguistic Notes” written in 1959 (PE17/143-145) in which Tolkien rejected the notion of la-negation entirely:

Delete √AL|LA “not”. Quite unsuitable. AL, LA have already too much to do.

Substitute: Negation. Primitive Eldarin. Negation divided into 1) refusal and negative command (future); and 2) denial of fact (past and established present).

1) √ABA, BĀ- distinct from AWA, WA “away”
2) √Ū, ?UGU - originally expressing privation [the question mark is Tolkien’s] (PE17/143).

Of Sindarin usage Tolkien said:

Sindarin (2) û, adverb or interjection “no, not (of fact)”. Also ú- prefix, applicable to all classes of words, but especially to verbs and adjectives: this simply negates the sense of what follows: as ú-land, ulan(n) “not broad, narrow” ...

Prefixed to verbal stems (without further suffix) ú- forms quasi-participles in aorist mode: as únod “not (ever) counting”, úbed “not saying”, úlal “not laughing, serious”, úgal “not shining, dark”; (úmel) úvel “not loving, enemy, inimical”. If the sense is strictly present (and continuative) the continuative participle -ol may be added, úgarol “not now doing or making, idling” ([vs.] ugar “(generally) idle”) ...

Such negatives of possibility as “countless, innumerable”, “unspeakable” are expressible either (a) by expressions equivalent to our “without count” or (b) by the suff[ix] -ui “having quality”, which with verbal stems has implication of possibility or suitability, úbedui “not fit to say, unspeakable”; únodui “countless” ...

Negation of a sentence is in Sindarin expressed by prefixing ú to verbal forms: as (úmedin) úvedin “I do not eat”, úgar (or ú-gar) “he does not do (make)”. But in archaic Sindarin and verse, as in Quenya uin “I do not”, únen “I did not” could be used with bare verbs, as uin gar “I do not make” (PE17/144-145).

This paints a fairly complete picture for Sindarin verbal negation using the particle/prefix ú- “no, not”, which induces soft mutation on the following verb: ú-vedin “I do not eat”. There is also a “negative verb” with present tense ui, past tense ûn which can take pronominal suffixes. It can be used to negate an (uninflected) verb stem in archaic/poetic speech: †uin vad “I do not eat”. It might also be used as a negative copula “is/am/are not” as in uin adan, ni edhel “I am not a man, I [am] an elf”; únen ennas “I was not there”. Alternately, û by itself might serve as a negative copula after †ui became archaic: ni û adan, ni edhel.

In Quenya verbal notes from 1969, Tolkien changed his mind, rejecting u-negation and restoring la-negation. He wrote:

û will not do. It is not necessary to avoid at all costs similarities with known European languages — Eldarin is deliberately devised to resemble them in style — but here the resemblance either to Greek ou (phon. û) or to the unrelated Norse ú, as a prefix, is too close.

ú should remain, but with the sense “bad, uneasy, hard” — similar to IE *dus, Greek dus-, Gmc. tuz- (tor-). This will leave únótima in GL [Galadriel’s Lament] correct, with meaning “difficult/impossible to count”. But úchebin in Gilraen’s linnod will not fit so well. It must be assumed that in S. ú 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. úchebin will mean not “I do not keep”, but “I cannot keep” ...

The suggested alternative lá, la, ala would be convenient, and the fact that it appears in Semitic would be no objection. I adopt this (PE22/160).

In this new system, ú once again had an “unpleasant” or “emphatic” connotation, the latter especially true in Sindarin. After writing the above, Tolkien outlined how the restored la-negation was used in Quenya, but did not further discuss the implications of this new system in Sindarin. Furthermore, we know that Tolkien reversed himself again a year or so later, writing:

Negation
Back to ú
lá can be beyond ...
ū should be negative particle (VT44/4).

baw-negation: In the 1959 notes mentioned above, Tolkien introduced a new form of negation based on the root √BĀ/ABA “refuse, forbid”. It’s Sindarin usage was:

Sindarin (1) baw! imperious negative imperative “do not, don’t!” avo, adv[erbial] negative with verbs in imperative, often as prefix av, af, as avgaro “do not do it!” Also personalized avon “I won’t”, avam “we won’t”. Cf. avad “refusal”, bauda- “ban, prohibit, refuse” (PE17/143).

In this 1959 system, Tolkien distinguished negation of fact (Ū) from negation of volition (BĀ). One subtle wrinkle in these baw-negations is that they are always negated from the perspective of the speaker. In the first person avon/avam (or *avof [avov] using the later Sindarin 1st plural pronominal suffix) means “I/we won’t” and indicates refusal. However, when addressing another person, the sense is forbiddance or prohibition: baw “don’t”, avgaro or avo garo (see below) “don’t do [that]”. Thus when commanding another person baw/av- means “do not”, but when refusing a command av- means “will not”.

Tolkien went on to describe this same system of baw-negation in the Quendi and Eldar essay written around 1960:

In Sindarin the following forms are found, baw! imperious negative: “No, no! Don’t!” avo negative adverb with verbs, as avo garo! “don’t do it”; sometimes used as a prefix: avgaro (< *aba-kar ā). This could be personalized in the form avon “I won’t”, avam “we won’t”: these were of course not in fact derived from avo, which contained the imperative -o < *ā, but from the verb stem *aba, with inflexions assimilated to the tense stems in -a; but no other parts of the verb survived in use, except the noun avad “refusal, reluctance”. Derived direct from baw! (*bā) was the verb boda- “ban, prohibit” (*bā-ta) (WJ/371-372).

The only difference from the 1959 notes is that avo can be used as an independent negative adverb in commands like avo garo “don’t do [that]”, and that the verb for “prohibit” is boda- rather than bauda-. During the 1969 restoration la-negation, derivatives of √BĀ retained their function as negatives of volition, including another mention of S. boda- “refuse, forbid” (PE22/161). So even as Tolkien vacillated over √Ū vs. √LĀ for the negative of fact, √BĀ retained its function as the negative of volition.

Summary of Conceptual Development: For most of Tolkien’s life, it seems that la-negation in Sindarin and its conceptual precursors was relegated to the role of a negative prefix of varying degrees of strength. The prefix u-/ú- also served as a negative prefix, sometimes with an additional “bad” connotation. For verbal negation, however, the only attested phrases use u-negation pretty consistently. We don’t know how Tolkien would have used la-negation with Sindarin verbs.

One big caveat to this pattern is that there is gap of about 20 years from the 1930s to 1940s where we don’t know anything about verbal negation in Noldorin. It is possible that at some point in this conceptual period la-negation was the norm, as was the case with Quenya in the 1940s. Furthermore, it seems that la-negation was the norm in Sindarin during its brief restoration in 1969, but we don’t know how Tolkien formulated it.

Neo-Sindarin: Older Neo-Sindarin courses recommend u-negation in Sindarin when it comes to negating verbs. This is the case, for example, in David Salo’s Gateway to Sindarin (2004, GS/146-147) and Thorsten Renk’s Pedin Edhellen: a Sindarin Course (2010, PESC/46). With the publication of PE22 in 2015 and its striking rejection of u-negation, several Neo-Sindarin authors began advocate the use of la-negation in Sindarin, including authors for whom I have a great deal of respect such as Fiona Jallings and Elaran. This alternate system is based on a neologism ᴺS. law “no, not”, coined by analogy with S. baw “don’t” and the most likely Sindarin equivalent of Q. lá.

For Neo-Quenya I advocate taking a compromise position and using both la-negation and u-negation for verbal negation, but I find myself unable to make a similar recommendation for Neo-Sindarin. In Quenya, we have numerous examples of lá used with verbs, but we have no such examples for Sindarin or any of its conceptual precursors. On the other hand, we have a well described system of Sindarin u-negation from 1959, which was probably restored to validity after 1969 along with the restoration of u-negation (VT44/4). In general I prefer to use attested over unattested forms, even when the attested forms do not represent Tolkien’s “latest” ideas. In this case Tolkien’s last word on the topic seems to be a restoration of u-negation, making it hard for me to recommend unattested *law over an attested ú.

I would not recommend completely abandoning la-negation in Sindarin, however. In particular, the Sindarin/Noldorin negative prefix al- has always been quite popular, recommended for example by David Salo (GS/139). This prefix is especially useful as a neutral negative prefix, as opposed to u- and ú- which sometimes have an “unpleasant” connotation as described above. This use of al- is in keeping with my recommendations for Neo-Quenya.

Whether you use ú- or *law for verbal negation, it is common Neo-Sindarin practice to limit these exclusively to negations of fact. I recommend using baw! “don’t” or avo (with soft mutation of the following imperative verb as in avo vado “don’t eat”) for negative imperatives. I also recommend avon/avof “I/we won’t” when refusing a command, as described above. Remember these are always negative from the perspective of the speaker. To say something like “you won’t”, you must use a different verb such as boda- “to refuse”, as in: bodol “you refuse”, boda “he/she refuses”.