I believe pronouns are the oldest words of every language and we can observe that they change during ages not as much as other words do. So are the 3 sentences 'my heart tells me' a prove that there wasn't any ancestor of enni in Common Eldarin? I support the 1st pattern but also Thorsten Renk's theory that Sindarin pronouns which are attested may have evolved from the 'long consonant' stems: nya, cca, mma, lla thats why they result in: in, ech (not eg), em (not ef), and el (?). The ones used by David Salo in the movies: eg and edh may have come from ca, lda. And i suppose that difference is visible in dative too.

I have used "enni" and "annin" both, in my time, but now I use "annin". I think that with the phrase "Guren bêd enni", there's several odd things going that have an elegant solution, one that made me change my mind.

Don't think of it like a normal phrase, with normal grammatical rules. Think of it as a cliche, something repeated so often that it has become worn down into a memorized segment. There are several things that point to this: mutation of "pêd"; vowel shortening of "gúren" to "guren"; and last, I-affection and clipping to "annin". 

This isn't the only phrase Tolkien made like this. There's also "Mae govannen." Looking at what he wrote about trying to justify this phrase in this form, he ended up with "mae ci govannen" getting shortened to "mae c'ovannen" which got shortened to "mae govannen". Basically, he was making something akin to "Howdy!" which comes from the old greeting, "How do you do?"

Therefore, I think that "To/for me" is "annin".

+Fiona Jallings I almost became even more radical, by refusing to believe in any dative pronoun forms. In attested sources, we have le linnathon, le linnon im, le nallon sí, *caro den — all in explicitly dative meanings for me, but expressed in seemingly accusative/oblique way. Comparing to the normal way to express indirect objects, Ónen i•estel edain, where the word order VOI gets the meaning and not any article/preposition, I couldn't but notice that Sindarin doesn't like to over-express grammatically the clear ideas and does not use dative unless ambiguous. Hence, I wanted to banish datives explicitly…

But such constructions cannot explain ammen, popular, widely used, explicitly glossed and translated 'to/for us'. So, datives must exist. Still, I'm unsure about le; what does its usage in only dative contexts mean? Is it *len in accusative with an abnormal dative? Is it fossilized after borrowing and never inflected? Is it the 'simplification' I mentioned above and D. *allen does exist?

And is den in caro den actually a form of de, dhe 'you'?! (I'm imagining things. please restrain me.)

+Александр Запрягаев​​​​ interesting ideas. Absence of dative - that sounds allmost convincing. And the second one - could we have nom. den, gen. lín, dat. le? I've got one more hypotetical explanation for occurrence of final n in dative. If dative is created from nominative maby ones that don't undergo lenition in accucative (ni, le) can't end in -n because it would sound the same. That's why le, len, (al)le, ni, nin, enni but men, ven, ammen But that's only plain guess of the beginner. 

+Fiona Jallings Precious information. What would then ci mean? You?

+Remi Korben Yes, ci 'thou' < kī is really attested, PE17:017. It is also attested that Sindarin borrowed Quenya pronoun le (judging by the form, it means they did it in nominative, but all attested uses are dative in meaning) and it obsoleted earlier native de, dhe of Doriathrin tradition. But neither has any exact conjugated forms stated anywhere.

I actually have some explanations for the dative case for Sindarin pronouns, one that I've posted about on Elfling before, and I include it in my lessons. My post on Elfling went unnoticed though. Anyways, I think that this explanation is more elegant, because there aren't exceptions.

First off, I think that there is a difference between transitive and intransitive verbs. With a transitive verb, when there is a dative pronoun, it is without fail marked with "an". "caro ammen" (do for us) "ai gerir úgerth ammen" (who do sins to us) "guren bêd enni" (my heart says to me) "Naur dan i ngaurhoth ammen" (Fire against the wolf-pack for us). (Note that with regular nouns, it looks like "an" can be dropped, as long as the second noun is the dative one, and the verb is di-transitive) But, if it is an intransitive verb, like in "le nallon" (I cry out to you) or "le linnathon" (I will sing to you), "an" can be dropped, and the pronoun will go in the same place that the pronoun-object of the verb would go, as seen in "hain echant" (made those).

We see this -n show up in Sindarin pronouns. I'm not counting possessive pronouns, but I mean "Tiro nin" (watch over me) "ammen" (for/to us) "caro den" (do it) "hain echant" (made those). Comparing Sindarin to Quenya, in Quenya, the -n is dative, but in Sindarin, looking at these examples, it clearly isn't. Instead, I think that this is like the English Oblique case. But, if it is why is there sometimes no -n?

Looking at times that the proposed oblique -n isn't there, there's a commonality.  "le nallon" (I cry out to you) or "le linnathon" (I will sing to you). Think of Nasal mutation. If the phrases were originally "len nallon" "len linnation" and "hain echant", nasal mutation would delete the -n on len because "linnation" begins with an L and "nallon" begins with an N, but because "echant" begins with a vowel, the -n on "hain" is left untouched.

There is one exception, "gohenam di ai gerir úgerth ammen" (We forgive they who do sins to us), but I think I know why we'd see a mutated nominative pronoun here. It has a relative clause attached, and in that clause, "they" is the subject. So, it's the nominative pronoun, with mutation on it like it's a regular accusative noun because it heads the phrase that is the direct object.

What makes this explanation elegant is that with it, there are no exceptions to the rule. All of the forms aren't just accounted for, they are generated with this system.

So, for clarification, the system I'm proposing is like this:

Nominative: No additions to the pronoun root.
Oblique: add -n to the end of the pronoun, and use the appropriate mutation.
Dative: prefix "an" to an oblique pronoun.
Possessive: Retaining only the initial consonant, delete the vowel of the pronoun and replace it with "-în".

BTW, In PE17 the mutation that "an" uses was called "vocalic", but not quite... Tolkien cited the N>M in "ammen". I think that it's actually using a long-mixed mutation, not a nasal mutation, as was proposed so long ago. Therefore, the dative pronouns would be:

annin, ammen, *angwen, *angin, *anlen, *anden, *anden, *andin, *anhan, *anhain, *anhen, *anhin, anim, *anech, *anest. (and *anhon, *anhyn if you think that there is gender distinction in Sindarin 3rd person singular pronouns, like Thorsten Renk)

+Fiona Jallings Could you please provide the background information on Long Nasal Mutation (original quotes and the reasoning for the remaining reconstruction?) For here the Helge/David/Thorsten pattern suddenly disagrees with yours; is it actually attested that nl should remain without assimilation and ns shifts to nh? Indeed, Tolkien uses the word 'vocalic', but does he mean lenition (anhon) or ordinary intervocalic cluster shifts (third column in my table)?

These are articles written a while back about nasal mutation and mixed mutation, and they are what's been used for a long time:
http://folk.uib.no/hnohf/sindarin.htm#nasal
http://folk.uib.no/hnohf/sindarin.htm#mixed
http://www.elvish.org/gwaith/sindarin_phonetics.htm#mutation

But, the analysis of mutation with "an" was based on data pre-PE17.
PE17/38 "ammen" is broken down into "an+men"
PE17/102 "Aglar 'ni pheriannath" is broken down into "Aglar an|i pheriannath"
PE17/146-7:
"ANA/NÂ [...] S. an, dative chiefly with pronouns or persons. < ana, hence vocalic mutation, but takes form m before m, b.
[...]
AN/NÂ, to, towards. S. an, to, na. an chiefly in forming datives of pronouns, as anim, to myself, ammen to us. 
[...]
S an, to (dative). [...] vocal[ic mutation] (ana), dative with place or person."

From this, I get: use vocalic mutation except with voiced stops and nasals. With those, the N of an assimilates the place of articulation from the voiced stop or nasal following it. Also, once the voiceless stop is a voiced stop, it'd take that place of articulation too. Therefore:

an+p=am b
an+b=am b
an+m=am m
an+c=an g
an+g=an g
an+t=an d
an+d=an d

On the topic of pronouns, something has been bugging me for a while. If the oblique is formed with -n, how do we factor in some of the various 1950s pronouns given in VT50?

That is, en for 1st person, and est, ent, ith, idi/idir for 3rd pl., eth, is for 3rd sg. Apart from idi, none of those seem very accommodating for -n suffixing.

Maybe one of these days we'll see all the Noldorin + Sindarin pronouns and pronominal suffix charts from the 1930s-1960s+ published, and then the situation will be more like Quenya where we're choosing between attested forms, instead of having to reconstruct so much. Though if we're talking priorities, I'd like to see the ET's treatment of the Mágol and Taliska material once it's done, as well as the rest of 1930s Noldorin's historical grammar ... (/offtopic)

That was Noldorin though... Tolkien was changing his mind on Noldorin pronouns a lot, but in later "Sindarin" translations, we see the oblique -n popping up, indicating, I think that he'd abandoned much of the Noldorin pronoun systems.

+Matt Dinse
Off topic:
I'd gladly switch Sindarin 'these four variants are wrong, now deduce the right' to Quenya 'one of these four variants is true, but they are contradictory'! One thing that brings me joy every day is that I ventured into Tolkienian linguistics early; I actually plan to be around when final VT and PE are out. In fact, maybe all of us should postpone any pronoun musings till after PE23 is in print, which by all means should include the long-awaited 'Common Eldarin Pronouns, Demonstratives and Correlatives'?..

+Fiona Jallings
On topic:
To discover the pattern, we need two pieces of information: (1) what does Tolkien mean under 'vocalic': lenition, nasal mutation or intervocalic assimilation and (2) which consonants are affecting the final -n. But I don't believe any Sindarin pattern could lead to such an uncouth (equë Tolkien) combination as nl [attested in Sindarin only in a loose-compound minlammad — in any Gnomish/Noldorin/Sindarin stage!] I believe, his words about ana (they are final, for the quotation after that is 'crossed out', per PE17 editors) suggest the scheme:
1. Ana- is prefixed and causes lenition.
2. Middle a is syncopated.
3. If the remaining cluster is counter-Sindarin, it is assimilated, and it is presumably n of the prefix which assimilates.

Thus we get:
annin, angen, allen (nl > ll), andhen (?! — also a never happening combination, could turn annen more probably), anhon/anhen/anhan and finally anden (plus the plurals).

Indeed the old schemes are Noldorin, but Tolkien was unlucky while pulling Noldorin hain into Moria gate inscription; now we have to accomodate it. The gloss is quite suggestive of sa/te distribution following the lines of 'close/far', 'definite/general' and/or 'mentioned before/impersonal'. But the comparison to Quenya leads to another important distinction: -se and -sa can distinguish animate/inanimate — and that returns us to the 'he/hen/hene' of the Etymologies. The real question is: whether in Sindarin maturity there are only he 'he, she' (pun not intended) and ha 'it' (what is the mechanism that always maked all the pronouns lenited even in the nominative, eh?), or ho still does work (more fun this way for a practising translator such as myself, in +Roman Rausch 's translation of Gil-Galad I cannot unread 'about her the minstrels sadly sing', _o hen_…); but the evidence for this S-cluster of 3rd person pronouns is quite suggestive.

1. If you read page 147 of PE17, you'll see that he refers to nasal mutation and contrasts it with vocalic mutation by comparing na (from nan, "with") and an (from the root ANA, "to").
2. I'll just quote the passage about nasal mutation in PE17/147:
S na, before vowels nan (nasal mutation), means "with" in sense of possessing, provided with, especially of characteristic feature. Orod na Thôn, Mount of the Pine Tree(s). na 'to' and na 'with' are therefore distinct before vowels and b, d, g, p, t, c, m, s but same before h, f, þ, r (rh), l (lh. Late forms as nan-h as for vowels, archaic nath-r, nath-l for nan-rh, nan-lh. nan|sr > nassr > nathr. nað before r (nan-r >naðr).

This shows what we used to think would be a case of "long nasal mutation" as just regular nasal mutation. Plural "the" in mutates the same way that nan does. Tolkien refers to the mutation used with an as vocalic mutation consistently, through both versions of the 'dative/allative base' entries in PE17. Therefore, the mutation used in an is not nasal mutation. It's "vocalic", but not quite fitting into the traditional vocalic system, but instead, behaving a lot like the preposition en (of the).

Tolkien only mentioned the spelling of an changing before a b or m, not before any other letters. Though, frustratingly, he gives so little to go on it's hard to say "it is THIS way" with any certainty. We'll just have to wait for that mysterious manuscript that demystifies Sindarin pronouns to come forth.

I take for the demonstratives, se-close, si-pl.close, sa-far, sai-pl.far. (edro hi ammen and sí nef aear and i thîw hin and im Narvi hain echant) and the ten and ti from Ae Adar Nín as 3rd person pronouns. (note that with all of these, the plural version is a typical Sindarin plural of the singular version) Sindarin and Quenya can have related pronouns without them being the same. I think this is a case of Tolkien re-purposing an old word into something new.

+Fiona Jallings I have a feeling we are speaking almost the same things, only in different terminology… I prefer not to apply the word 'demonstrative' for something that is not an adjective and used only substantively. We treat Noldorin information differently: I rather attempt to re-use and fit as much older information as possible while not having more actual one. Indeed they are not exactly alike Quenya ones, with its ending/independent distribution, but Sindarin also does not (as we know yet) possess the KHY-pronoun, and so the actual usage of 3rd person pronouns becomes strange. I don't believe that Tolkien intentionally created such a perfectly gender-equal language, for, especially without the KHY-pronoun for 'another one over there', I don't quite get the idea of the complete impossibility to distinguish 'he' from 'she'. I believe, that with his ordinary carelessness to produce Sindarin grammar compared to Quenya, Tolkien would never do something like that if confronted with conscious desire to complete an inflection table. Thus, Noldorin glosses with their definite gender/animate distribution (and mirroring the -sse/-ssa of Quenya, importantly!) fit very nicely with the fact that the known uses of te are indefinite/far-demonstrative/gender-irrelevant!
The same with an- ones: Tolkien never elaborated his mutations to include all the possible initial consonants. He skipped the rare (but trickiest, as it later turns) rh, lh, h, hw quite often (cf. his description of S-assimilation after ah and the original gloss of stop mutation where he only states such form's existence!), and I do not really feel it a violation of Tolkien's word when I extrapolate beyond glossed nm-sitiuation (remember, Tolkien needed it for annin and ammen only!). If he actually tried to employ his neatly described ana-rule to the actual pronouns, he'd stumble into the most violent nl and a much worse ndh which we have no idea how to deal with for it never appears in a native word; there is NO mechanism leading to this cluster beyond the employed here: a compounding or prefixation of something ending with a vowel with its later syncopation! A stop to anden may be expected after Salo's rule 4.117 of un-lenition on morpheme boundary after n, but annen, adhen, even athen are not something to be surprised of!

I agree - I also use older bits to fill the gaps... Like in the charts on my website, I hade the inclusive 1st person plural out of the Quenya one, which is where I got Gwe, Gwen, Angwen, Gwîn.

Mostly we disagree on the topic of how the mutation would work with an, and I see no reason to look for masculine and feminine pronouns to fill that gap, because I don't perceive it as a gap.

Your reduplication idea may be something that happened to the reflexive pronouns, est and ech. It's something worth looking into.

+Fiona Jallings You're a scholar and a teacher of language, I'm generally a translator, that's all… I rarely need personal pronouns, especially dative ones, but in my actual practice the complicated gender distinction is one of the most frustrating omissions.
Angwen? Why not ana-gwen > ana-wen > *anwen, to follow Tolkien's words exactly? We have lots of nw in words… Though isn't is rather angwen > amben > ammen clashing with the known one? I believed initial and intervocalic ñw turn into ñwgw > mb in Sindarin, if not all Lindarin tongues. (Though in fact we know exactly from Carl Hostetter that these pronominal complications are somehow resolved in Tolkien's writings, and resolved to Quenya's degree of overabundance…)

I'm actually more of a translator than a scholar. I mostly leave the scholarly stuff to others, and pick out the theories I like. I do have some pet theories that i'm fond of, like my pronouns theory, and I am a teacher, trying to make the theories understandable and usable to non-linguists... but I'm primarily a translator. I specialize in making translations that fit characters and stories, which is why I focus so much on dialects. When I translate a poem, I usually invent the mouth that would have uttered it too.

Anwen is actually equally acceptable to me as angwen, but ammen as a 1st person inclusive... I don't think that the compound is old enough for the ngw>mm to have happened. 

+Fiona Jallings I'm really no fanfictioner, and my translations normally involve the 'primary world' works, hence I ordinarily employ a kind of 'Received Pronunciation', stabilized as told by Pengolodh to Ælfwine, Sindarin of Nth Age (I did my sonnet as a thing composed by Noldorin poet knowing the accepted norm but making yet mistaked such as mutating beleg > veleg); though the matter of recognizability makes me altering the spelling (I always spell lth to fit the Etymologies, though it is doubtful it was pronounced that way outside Noldorin influence).

I guess this is unfortunately the way Sindarin study normally works: each and every one of us has one's own personal comfortable headcanon, but when we start showing our ideas to each other we cannot understand half of the tricks we use! Especially in such minor yet important language zones: athar or athan? (I'm all for the first: the only reasoning for athan in mature Sindarin is comparison to late Quenya han of similar meaning — but what is the prefix? Ad/at- 'two'? No meaning! Ath-? The re-derivation of athrad in PE17 as a gerund banished ath stem in favour of thar 'across' < A-THAR! In fact, I'm not doubting the possibility of actual Noldorin athan in SD; only even if it existed, said re-derivation refurbishes into into athar < athār from sundóma reduplication or maybe the 'exactifying' a-prefixation/dynamic lengthening a-ttar, which this time coincide!) I wonder if anyone here has his own ideas to say 'he went'! Event? (Thorsten's) Evenn? (Yours) Evín? (David's) I believe that an analogical past tense 'I drank' is Ŷngin ( ûnc < augm. u-hunc [cf. Etym. Nold.sunc,also auv < a-havw] + in); does anybody agree here at all?

More to the original point: I decided to search the newer materials for any pronoun discussion. I found in PE21:75: "_Thus in unemphatic pronouns (which are archaic in form and largely escaped the later inflexional elaborations), where two such occurred in a sentence, the one nearer to the verb (or most closely agglutinated to it) was taken as the direct or nearer object; the second was in function usually what we should describe as 'dative'. There was in Eldarin no distinction felt or marked between "I taught K. music" and "I gave K. a gift." In such cases in Eldarin, and some of the derived tongues, it remained possible to express both by uninflected forms." Well, in Quenya these cases are explicitly different: by accusative with null and by dative with _-n — then which are 'derived tongues' if not Lindarin branch? A proof it can be, I guess, for a universal 'objective'/oblique form in S. with only a couple fossilized dative elements (annin(?), ammen, enni and the reflexives which are obviously late and different from the rest)!

The problem with Sindarin is that there is a lot of information missing. If Tolkien ever made charts for mutation - they haven't been published. The Sindarin past tense is still mostly unknown. Pronouns are... let's see, how many are actually attested? We have le, dhe, im, ech, est, ammen, nin, ci, den, di, han, hain, anim, hen, hin, hi, sí, nîn, lîn, vîn, dîn, în, enni.... there's very little to go on. Til we have more definitive information, we'll just have to wait and play with our theories.

+Fiona Jallings I hope Pronouns, Demonstratives and Correlatives give us lots of relevant information. I'm only afraid it would be primarily Noldorin, Etymologies-compatible and not late enough…

Any insight is better than none, in my opinion.