The mutation of adjectives in Sindarin and its conceptual precursors is a fairly well established feature of the language. For example, in the Gnomish Grammar of the 1910s, Tolkien said “For the rules of adjectival initial mutation see page 1a” (PE11/15), where adjectives following nouns underwent the same mutations seen after the definite article i. Likewise, adjective mutations after nouns was a feature of the Early Noldorin Grammar of the 1920s (PE13/124).
However, the notion of adjective mutation is not obvious from The Lord of the Rings itself. This is because many of the adjectives in The Lord of the Rings do not mutate, and the few adjectives that do mutate such as Parth Galen “Green Sward” (LotR/417) are not obviously mutated forms. In Jim Allan’s 1978 book An Introduction to Elvish, there is no mention of mutation of adjectives mutating after nouns (p. 70). In Allan’s book the words galen “green” was considered distinct from calen seen in words like Calembel, Calenhad, Calenardhon, since the meaning of those words (and their connection to “green”) was not stated in The Lord of the Rings.
I have been unable to determine when the Elvish linguistics community became aware of adjective mutation, but it was a known feature of (Neo) Sindarin systems by the early 2000s. However, it was also known that adjective mutation was sometimes omitted. This fact was noted in both David Salo’s 2004 book A Gateway to Sindarin (p. 102) and Helge Fauskanger’s descriptions of Sindarin on his Ardalambion website. The publication of Parma Eldalamberon #23 in 2024 indicated that Tolkien himself was uncertain of the precise nature of adjective mutation. In a 1969 document on Sindarin Mutations (PE23/142-144), he wrote:
In the case of nouns and adjectives where these were placed together, if they were in any close syntactic relation, the second had a mutated initial according to phonetic development; but this was not always observed in practice [my emphasis]. It was normal when an adjective preceded a noun; since this was the original normal order in CE, and such collocations thus descended from the period during which the mutations developed phonetically. But when either a) the adjective was placed second, or b) two nouns were juxtaposed, there was hesitation; and in the Exilic period the regular usage adopted was not always the phonetic one (PE23/143).
It is not clear what Tolkien meant by “not always observed in practice”. However, in this other contemporaneous documents like the 1969 Common Eldarin Article (PE23/133-140) he seems to be reexamining The Lord of the Rings and trying to reconcile some inconsistencies, and inconsistent adjective mutation may have been one of the things he was trying to reconcile.
It is my contention that in a period from the 1940s into the early 1950s, adjective mutation after nouns was not a regular feature of the Sindarin language, as evidenced by drafts of The Lord of the Rings as well as The Lord of the Rings itself.
It is well-known that Tolkien made a similar reversal in mutational patterns for nouns in juxtaposition, which are used in Sindarin to form genitival relationships. This can be seen by comparing Tolkien’s writings of the 1930s and 40s to his writings in the 1950s and 60s, where Noldorin mutated nouns were frequently revised to Sindarin unmutated nouns:
In the 1950s and 60s, there are very few examples of nouns mutating after other nouns other than in compounds. The only obvious example I can find is {Raith >} Fui ’Ngorthrim “Paths of the Dead” from Tolkien’s “Unfinished Index” of The Lord of the Rings (RC/526). There is also Sarn Gebir “Stone-spikes” which seems to be a hold over from the 1930s-40s conception (LotR/391; RC/327; TI/283; LR/363), and Gil-galad where Tolkien originally considered the second element to be a mutated form of calad “light” < KAL (LR/348; RS/179), but later conceived of it as unmutated galad “radiance” < ÑGAL (PM/347).
This conceptual shift in the mutation of juxtaposed nouns indicates that one way that a mutation “was not always observed in practice”, since juxtaposed nouns was one of the two cases Tolkien mentioned on PE23/143. It is thus possible the same applied to adjectives following nouns, which might likewise (sometimes) not show mutation.
Looking at The Lord of the Rings itself, very few adjective forms are obviously mutated. The only two clear examples are:
There are also a number of examples that are obviously not mutated:
There are some ambiguous examples as well. Dol Baran (LotR/589) was probably intended to be unmutated “Brown Hill” when it was originally written, since Dolbaran was listed with baran “brown” in The Etymologies of the 1930s under the root BARÁN (LR/351). But in later notes Tolkien reconceived of it as “Bare Hill” where its second element was mutated paran “bare” < PAR “peel” (PE17/86). There was also Rath Dínen “Silent Street” (LotR/826) which was probably unmutated dínen when The Lord of the Rings was written given Amon Dîn “*Hill of Silence” (LotR/747). Compare also 1948 Q. lína- “be silent”, probably derived from DIN (PE23/76). However, in his Unfinished Index of The Lord of the Rings Tolkien said both of dínen and dîn were adjectives meaning “silent, quiet”, and mutated from tínen and tîn respectively (RC/551). In the same index he said that Fen Hollen should have been mutated as Fen Chollen (RC/550).
This gives us three additional examples that were probably unmutated when originally written, but were redefined as mutated forms:
A final example of interest is Iarwain Ben-adar “Oldest and Fatherless” (LotR/265), where Ben-adar “Fatherless” appears to be adjectival. In notes written after the 1st edition of The Lord of the Rings, Tolkien said ben- was mutated pen- “-less” from the root PEN “lack”, which he contrasted with PED “incline, slope” (PE17/171). But in The Etymologies of the 1930s, PEN meant “slope” (LR/380), so it seems possible that Tolkien did not introduce PEN “lack” until after the 1st edition of The Lord of the Rings, so that ben-adar could have had a different etymology when it was first introduced.
As indicated by the above and with the major exception of calen → galen “green”, it seems the majority of adjectives after nouns were originally unmutated in the The Lord of the Rings as first published. However, its seems in notes written by Tolkien leading up to the 2nd edition, he felt that many of these adjectives should have been mutated, and modified the etymologies of some of them to support this. Thus it seems possible that while The Lord of the Rings was being written, adjective mutation after nouns was not the norm.
To test whether the above hypothesis is correct, we should also examine noun-adjective juxtapositions in drafts of The Lord of the Rings, especially the combinations that did not appear in the published version.
One example is Dûn Caron “Red Valley”, the initial name of Nanduhirion, also appearing as Doon-Caron, Carndoom, and Caron-dûn (RS/419; RS/433 note #13), comparable to Carn Dûm, the (untranslated) name of the chief settlement of Angmar from the published The Lord of the Rings (LotR/146). In these instances, it seems that dûn is “valley” and caron is “red” and neither are mutated, but compare tum “valley” from elsewhere in Tolkien’s writings which indicates dûn “valley” might originally have been a mutation (LR/394). There is also Cris-caron “Red Pass” appearing in the same context, again with an unmutated adjective caron (RS/419).
Another example of non-mutation is Ered Myrn “Black Mountains”, the earliest name for Ered Nimrais (TI/124). This example is especially interesting since Tolkien revised this to the compound Eredvyrn, indicating that mutation in compounds was normal but mutation of independent adjectives was not. There is also Kerin-muil, a draft name for Nen Hithoel (TI/364). This muil might be an adjective as it was in later Emyn Muil, but it could also be a noun as in Ilk. muil “twilight” < MUY from The Etymologies (LR/374).
There is Ennyn Dûr “*Dark Gates” [revised to Mornennyn] as the name of the gates to Mordor, as well as Nelig Myrn and Naglath Morn “*Black Teeth” = “Teeth of Mordor” for the gate towers (WR/113, 122). An earlier name of Rammas Echor was Ramas Coren (WR/288), perhaps meaning “*Round Wall” given corn or coron “round, globed” < KOR from The Etymologies (LR/365). There are also various unmutated adjectives that appeared in both the drafts and the published version:
For mutated adjectives, both Parth Galen and Pinnath Gelin also appeared in their mutated forms in the drafts (WR/280, 307, 437). In addition we have Tol Galen “*Green Isle” which had an alternate name Toll-ondren “*Rocky Island” = “Carrock”, also mutated (TI/268, 271). The second name seems to have a mutated adjective gondren “*rocky, of rock”, but it could instead be a compound rather than an adjective mutation. Finally there is mutated Tol Varad “Defended Isle” as the earliest name of Cair Andros (WR/326). However, this could be a loose translation of “*Tower Ilse” since barad is normally translated “tower”, so that it is a noun-noun juxtaposition rather than noun-adjective.
This gives us a fairly long list of unmutated adjectives in Lord of the Rings drafts, the last five of which also appeared in the published version:
For (possibly) mutated adjectives we have:
First two also appear in the published version, and the last two might not represent normal adjective mutation: Tol Varad might be a noun-noun and Toll-ondren might be a compound. Assuming this is true, the main examples of adjectives mutating after nouns in The Lord of the Rings and its drafts all involve the adjective calen “green”.
Taken together, this gives us strong evidence of adjective non-mutation in Lord of the Rings drafts of the 1940s and early 1950s in the lead up to the publication of the 1st edition.
Notes on The Feanorian Alphabet from the 1930s had the short phrase i vegli vorn “the black bear” (PE22/33) which is a clear example of adjective mutation, similar to the system described in the Early Noldorin Grammar of the 1920s (PE13/124). Silmarillion notes and drafts of the 1930s had a mixture of mutated and non-mutated adjectives:
The final version of The Silmarillion also had a mix. Since The Silmarillion is a collection of edited materials from various periods, assigning precise dates to each name is difficult.
Dor Dínen “Silent Land” is ambiguous for the same reason as Rath Dínen from The Lord of the Rings: it is not certain what Tolkien imagined the unmutated form to be. However, in one note from 1970 Tolkien had the clearly-mutated Dor Dhínen (WJ/333). Talath Rhúnen could be considered ambiguous since it was long assumed that the mutated form of initial voiceless rh- (< sr-) would be thr-, but in Common Eldarin Article written in 1969 and published in 2024, Tolkien made it clear that rh did not mutated in modern Sindarin (PE23/136). Going by its gloss, Laer Cú Beleg could end in an unmutated adjective beleg “strong”, but the actual meaning could instead be “*Song of the Bow of Beleg”, where Beleg is a proper name.
Regardless, there seem to be more instances of mutated rather than unmutated adjectives in The Silmarillion, especially compared to The Lord of the Rings. This is generally true of Tolkien’s writings later in the 1950s and 60s. Tolkien’s Unfinished Index of The Lord of the Rings from between the 1st and 2nd edition had clearly-mutated Cûl Bîn “Little Load” and Cûl Veleg “Bigload” (RC/536). The Quendi and Eldar essay from 1959-60 had Bar-goll “Hollow [coll] Dwelling” (WJ/414). Notes on The Rivers and Beacon-hills of Gondor from the late 1960s had Lô Dhaer “Great Fen” (VT42/14), though it also had Côf gwaeren Bel “Windy Bay of Bel” without mutation (VT42/15). Pauline Bayne’s 1970 map of Middle-earth had Eryn Vorn “Dark Wood” as an addition by Tolkien (RC/lxv).
Silmarillion notes from around 1970 also a mixture of adjective mutations. Athrad Daer “Great Ford” was contradicted by Athrad Dhaer (WJ/335, 338). Duin Daer “Great River” (WJ/336) was contradicted with Duin Dhaer (WJ/191). In the tale of Aldarion and Erendis dated to around 1965, this adjective had the unmutated form daer in Lond Daer “Great Haven” (UT/263) and Lond Daer Enedh “Great Middle Haven” (UT/264). These examples (and Lô Dhaer) indicate ongoing vacillation between mutation and non-mutation of adjectives late in Tolkien’s life, since the unmutated form was definitely daer as in the name Daeron (VT42/11).
There are a few more noun-adjective juxtapositions from the Unfinished Tales, all dating to around or after the publication of the 1st edition of The Lord of the Rings:
All told, there are comparatively few unmutated adjectives in Tolkien’s writings after the 1st edition of The Lord of the Rings. If we exclude the vacillations between daer and dhaer “great” and dínen and dhínen “silent”, there is:
As noted above, the last example is ambiguous because it might be Beleg as a name rather than an adjective. Nan-tathren may be holdover from earlier Nan-tathrin or Nan Tathrin which dates all the way back to the 1910s as the compound Nantathrin (PE11/67).
Before examining possible reasons for non-mutated adjectives, we should first return to the previously-mentioned 1969 notes on Sindarin Mutations:
In the case of nouns and adjectives where these were placed together, if they were in any close syntactic relation, the second had a mutated initial according to phonetic development; but this was not always observed in practice [my emphasis] (PE23/143).
In this document, Tolkien did not indicate why noun-adjectives did not always undergo mutation, so we must seek explanations from elsewhere in his writings.
Non-Mutation of m: In notes associated with the contemporaneous Common Eldarin Article (CEA), Tolkien seems to have noticed the frequent non-mutation of m in The Lord of the Rings, and was perhaps looking for a solution for it:
Great relief if mutation of m could be avoided. It actually occurs v[ery] rarely in LR. Only case I can find is Menelvagor, I 91; silivren [<] silmarin, I 250; but not used in Imloth Melui, III 142, 244! Vagor must then < bakār, bagor. But Menelmagar must be [from] magar [given Q. Menelmacar]. Say m > ṽ, but m retained grammatically after article and in adj. concord to avoid confus[ion] with b (PE23/138 note #12).
In the main text of CEA Tolkien went on to add a marginal note describing how mutation of m did not always occur for grammatically, but the mutation of m still applied as a medial phonetic development including in compounds:
This mutation [of m to ṽ] was preserved in early records (and is reported to have been maintained in Doriath, where however the ṽ remained nasal). But generally it was abandoned as a [grammatical] mutation (it occurred very frequently medially and also even at [the] beginning of second elements of compounds), because of the resulting confusion with mutation of b– (after denasalizing of ṽ) as in morn, black, born, hot, malt, gold, balt, force. The mutation was earliest given up in adjectives, cf. Imloth Melui (PE23/136).
The inclusion of “[grammatical]” is an editorial addition on my part, but I think this is clear from context since he is contrasting non-mutation after the definite article and adjectives following nouns versus mutation in the middle of words include compounds.
It may not be a coincidence that three of the unmutated adjectives in The Lord of the Rings begin with m: Emyn Muil, Ered Mithrin, and Imloth Melui, the last of which was specifically cited in CEA. Tolkien decision that m did not mutate could have been motivated by a need to explain these non-mutations in The Lord of the Rings. As a counterexample, there is na vedui “at last [medui]” (LotR/209). It is possible Tolkien simply missed this in CEA, but it also possible that mutation here is conditioned by the preceding preposition na rather than because it was an adjective.
It is unfortunately difficult to tell whether the above was a 1969 revision to mutational behavior of m or if something like it was a feature of the language while The Lord of the Rings was being written. For example, m did not undergo grammatical mutation in the Gnomish Grammar of the 1910s (PE11/7) or the Early Noldorin Grammar of the 1920s (PE13/121), but m-mutation did appear in phonological charts from the 1930s (PE19/19). If Tolkien reversed himself in the 1940s so that m did not undergo grammatical mutation, this would explain a fair number of adjective non-mutations in The Lord of the Rings.
However, there is the draft inscription on the Gates of Moria where m-mutation did occur: Ennyn Ðurin Aran Vória (TI/182). There is also the mutated of mellon “friend” as mhellyn in the King’s Letter from the unused epilogue to The Lord of the Rings (SD/129), and the revision of unmutated mín to mutated vín “our” in Sindarin prayers from the 1950s (VT44/21-22).
Non-Mutation of d and h: Several of the other adjective non-mutations in The Lord of the Rings involve an initial d: Barad-dûr and Rath Dínen, and in drafts Ennyn Dûr. It is known that Tolkien hesitated to use dh in Lord of the Rings drafts, as with 1940s N. Caradras (RS/433) and N. Fanuidol (TI/306) vs. later S. Caradhras and S. Fanuidhol (LotR/283). This omission was likely because English speakers wouldn’t know how to pronounce dh.
Tolkien’s use of Fen Hollen rather than Fen Chollen might have been similarly motivated, to avoid readers mispronouncing it as English “ch” rather than German “ch”, comparable to Rohan vs. “proper Sindarin” Rochan. If true, then Tolkien might have been motivated by purely orthographic concerns in some cases of apparent non-mutation. Note, however, Tolkien did have ú-chebin “I have kept no” in Lord of the Rings appendices (LotR/1061).
Non-Mutation of Other Initial Consonants: There are a number of other examples like Dol Baran, Cris-caron, and Ramas Coren that do not fit either of the above explanations. Given the preponderance of non-mutating adjectives in The Lord of the Rings, I think the simplest explanation was that, at the time The Lord of the Rings was being written, adjective mutation after nouns was not a feature of the language.
This does beg the question why calen “green” in particular was consistently mutated: Parth Galen, Pinnath Gelin, and in drafts Tol Galen. Given the apparent non-mutation of c in Cris-caron and Ramas Coren from Lord of the Rings drafts, it seems unlikely this was purely a phonological consideration. Perhaps there was something about unmutated calen that Tolkien found aesthetically unpleasing, and he reflexively mutated it wherever it occurred.
As discussed above, I think there is strong evidence that adjective mutation after nouns was not a normal feature of the Noldorin/Sindarin language as Tolkien imagined it in the 1940s while writing drafts of The Lord of the Rings, and this continued into the published form, at least for the 1st edition. He seems to have at least partially reversed himself in the period between the 1st and 2nd edition, contriving new explanations for existing forms to treat them as mutated, though other forms remained uncorrected. He may have made initial m non-mutating in notes from 1969 (CEA) to further explain some of these inconsistencies, though there is also 1970 Eryn Vorn from Pauline Bayne’s map of Middle-earth.
There is a known but currently-unpublished Noldorin grammar and phonology from the 1940s that could shine further light on this matter.
Primary Sources