=================
Old Elvish Syntax
=================

Old Elvish Syntax Tables
------------------------
(See corresponding info for proto-Elvish. Note that, for the moment, all forms
are rendered directly in IPA. I intend to develop a Romanization standard to replace this.)

Old Elvish pronouns distinguish animacy and inanimacy in the third person only.
They have dual number, and besides the nominative three other cases, namely
accusative, dative, and a comitative/instrumental.

Pronouns
********

Independent forms:

====== ====== ======== ============ ========
Person NOM    ACC      DAT          COM/INS
====== ====== ======== ============ ========
1SG    wɛl:ə* wɛlð     wɛʁ\ |We|    wɛln
2SG    i\:ma  i\:mað   i\:G\ |We|   i:man
3SGAN  dana   danið    danG\ |We|   dan:
3SGIN  danam  damnið   daG\ |We|    damnin
1DU    wɛlz   wɛlzið   wɛʁ\ |We|    wɛlzin
2DU    i\:maz i\:mazið i:mazG\ |We| i:mazin
3DUAN  danaz  danazið  danazG\ |We| danazin
3DUIN  danarm danarmið danaʁ\ |We|  danarmin
1PL    mɑl\:a mɑlð     mɑʁ\ |We|    mɑl:an
2PL    ʃeɔn   ʃeɔð     ʃeɔG\ |We|   ʃeɔn
3PLAN  θi\:l  θi\:lið  θi:ʁ\ |We|   θi:lin
3PLIN  θi\:lm θi\:lmið θi:ʁ\ |We|   θi:lmin
====== ====== ======== ============ ========

.. |We| replace:: :superscript:`w`\ e

\* When the following word begins with a vowel, reduces to wɛl

Polypersonal bound forms (suffixed):

| 1SG: -wɛ/ə\*\*
| 2SG: -i:
| 3SGAN: -da
| 3SGIN: -dam
| 1DU: -wɛz
| 2DU: -i:z
| 3DUAN: -daz
| 3DUIN: -dar
| 1PL: -mɑl
| 2PL: -ʃɔ
| 3PLAN: -i:l
| 3PLIN: -i:m

\*\* when word final

Polypersonal bound forms are compulsory on the verb. By leaving out the subject, object, or both, one can shift the pragmatic focus as a type of noun
incorporation (see below). This is in fact the pragmatically unmarked
strategy, but use of independent pronouns does happen to place emphasis
on any argument in the clause.

Nouns and Articles
******************

Dual marker (rare nouns, including referring to persons): Prefix zu-

Plural marker (used on nouns only):
AN (humanoid, body parts and certain spirits etc.): prenominal plural
word ɛl:ə/ɛl\*
AN (other, eg animals, other animate nouns): prenominal plural word
o:n/o:na\*
IN: circumfix e...(ɑ)m, from proto-Elvish ('also ... thing(s)', meaning
'all').

There are however a few irregular plurals that have other forms, e.g.
ɑm, ɛm ('thing, things', cf. created by analogy θɑm, θɛm, meaning
'dwarf, dwarves'). Note that for small numbers of subject inanimate
nouns (usually 2 or 3, sometimes more in specific contexts) it is
considered literary to use the *singular* agreement form of the verb. This
also applies to animates in diminutive, in particular contexts (little
dogs, little children etc).

Definite article: j(a)-. Acts as a proclitic, either on the noun itself
or on any classifier or demonstrative preceding the noun. Note the
definite article does *not* take stress, so that the stress pattern
shifts one syllable forward. Definite articles are used for ongoing
topicality, referring back to an introduced topic, but *not* for generic
reference ("the world is a big place").

Old Elvish has frequently (especially literarily) used noun classifiers. They
are obligatory when combined with numerals or quantifiers (any, all,
etc.). Otherwise they are optional but literarily preferred, especially
for plurals. It is sometimes in literary style possible to omit the noun
and preserve the classifier, when the noun is implied or obvious.
Classifiers precede the noun but come after the numeral, and are not
inflected in any way. Naturally they only occur with independent nouns,
not incorporated nouns. When they are combined with a definite article,
the definite article (*ja-*) is (pro)cliticized onto the
classifier.

Old Elvish also uses noun incorporation of a moderate type (in terms of the categorization developed by Marianne Mithun and others). An incorporated noun form is possible to make it a general, unspecific, instance of a frequently performed thing, etc., as distinct from a specific or referential case, or
something unusual that needs specifying. Also, there are many words
where the root of the verb and the noun are the same and where only
subsequent morphological inflection determines which word class it falls
into. In such cases, the verb form of the word may imply the noun form
also (e.g. the noun morphology of 'tie' means 'rope' and therefore 'tie'
can also imply 'tie with rope' without needing to specify the noun in
that case). Finally, there are cases where noun incorporation into a
verb changes the lexical meaning of the verb, which represents the
beginning of a new phase of noun incorporation, in this stage not yet
complete. This corresponds to NI classes I & II in Mithun's
classification.

Adjectives
**********

Animate unmarked.
Inanimate marker: -n

Adjectives follow nouns.

Demonstratives, numerals, etc.
******************************

*Demonstratives*:

The pronominal demonstrative is a deictic suffix on 3SG or noun forms. These distinguish nearness of the speaker and the hearer as follows:

| Near speaker: -p\ :superscript:`h`\ a
| Near hearer: -e (*of unknown origin*)
| Distal: -ɑm
| Neutral: -aʃ (*of unknown origin*)


The adnominal demonstrative is formed by a separate word following the noun. It lacks a neutral form:

| Near speaker: jap\ :superscript:`h` (from 'ja p\ :superscript:`h`\ an',  meaning 'the here' in | proto-Elvish)
| Near hearer: imas (from 'i:mas', no longer functional locative of 2SG, ie 'at you')
| Distal: jɑmpɑn (from 'ja ɑmpɑn', 'the there' in proto-Elvish)


There is a distributive suffix (meaning 'all around, all over the place, all about' etc. When used on a noun instead of the verb, it means ‘each’, e.g.
“each of the elves carried...” etc): -um (from proto-Elvish 'um:a', meaning 'around, about')

Demonstratives and numerals precede nouns.

Numerals cannot be pluralized. They require some independent noun,
possibly a placeholder noun, to be associated with. (E.g. 'to three I
gave...' becomes 'to three people I gave...'.)

Adpositions
***********

Adpositions precede nouns. Most, however, appear as derivational
morphology: see `Derivational Morphology`_ below.

Verbs
*****

*Stems*

-i is the stem for the infinitive and imperative, participles and for
passives. This was an old proto-Elvish infinitive that got reanalyzed as
the stem form in certain cases and then extended by analogy.

*Past stems*

Note that all past forms (see above) are based on the past stem, not the
nonpast stem, of the verb. These follow apophony (ablaut) patterns,
based on the stressed (first) vowel, as follows:

a -> u

ɑ -> æ

æ -> o

e -> ɔ

ɛ -> o

i -> ɑ

ɪ -> ɔ

o -> e

ɔ -> ɛ

u -> i

Long vowels, however, keep their form and are unaffected by the ablaut.
The ablaut ignores compounding of any kind, including noun
incorporation.

*Tense*

Present/Future: unmarked.
Hodiernal: -xar (proto-Elvish 'ixaraθ', 'yesterday')
Proximate past: -par (unknown)
General past: -t\ :superscript:`h`\ u (proto-Elvish verb stem 't\ :superscript:`h`\ ul-',
'get, receive')
Remote past: -qa (proto-Elvish verb stem 'qal-', 'say'; evolved from
evidential to tense)

For verb stems ending on a consonant, the -i stem is also inserted in
past tense.

*Aspect*

The perfective aspect is unmarked.
Imperfective marker (follows tense): -o. Imperfective is also used for
habituals.
Strong completive marker ('all the way, completely, totally'): preverbal particle *tul* (from past stem of tali, 'finish' -> tuli)

*Mood*

Infinitive: stem + -ar. (The -ar form comes from the postverbal particle
ar, meaning 'in order to'.)

Imperative: Declined like infinitive but with person prefixes.
Imperatives can also have a softer obligation meaning,
'should/must/ought...', which are the only (English) meaning for
non-second person forms. With a past tense stem it implies epistemic
necessity/inference, ('must have X...')

Ability/Possibility:

- Preverbal marker *t*\ :superscript:`h`\ *ul* (derived from 'to get' in proto-Elvish), indicating physical ability ('can, able to...'), situational possibility, and permission

- Preverbal marker *ælvrin*, derived from 'knowing' in proto-Elvish, indicating mental ability ('can, know how to...');

- Epistemic possibility ('could be, might be') is expressed with the preverbal particle *nɑmjaz*, derived from proto-Elvish 'it suffices', with a normally inflected verb.

Optative:

Auxiliary/modal verb function of inflected *aumi* 'wish' combined with
infinitive acting as optative.

The simple sentence is: 'wish.1SG.NPST go.GER.3SG.NPST' -> "I want him
to go" (I wish his going). Contrast with the optative meaning: "may he
go, I hope he goes, would that he would go": 'wish.NPST.1SG.AN.3SG.DAT
go', with an infinitive of the verb and the object of the wish, if any,
in the dative (I wish for/to him (that) to go).

Frequentative/Iterative:

A preverbal marker *eɑnno*, derived from proto-Elvish ('hea ɑnno',
'stand/stay +'at'' in the imperfective), combined with the infinitive of
the verb and the auxiliary 'to stand' or 'to lie', depending on whether
the so conjugated verb is active or stative respectively. (The auxiliary
can therefore duplicate the original meaning due to reanalysis, and a
finite verb being required.)

E.g. 'stand.NPST.1SG FREQ miss.DETR.INF' -> "I keep missing" or "I miss
all the time". A different word permitted word order is FREQ INF AUX.

Evidential:

There is one evidential marker, for direct sight of an ongoing object
or situation, a main clause initial proclitic *lo:* derived from
proto-Elvish 'I see' ('wɛl:ə o:n')

*Standard Forms*

Reflexive: -su(l) (from proto-Elvish 'sulɪs', 'body, self')

Participle: 'Xing', in the participial phrase sense of 'while doing X',
is indicated by a gerund form given the comitative suffix ("with the
Xing"). Otherwise, the imperfective serves the progressive aspect
purpose ('I am/was walking').

Gerund form: -ar (derived from Proto-Elvish -ara participle). Combines
with person and tense markers. Contrasts with infinitive form (which has
neither), and is used for nominalizations, also in combination with
adpositional derivation. Also an already old fashioned or literary form
of the 'should' clause appears with a gerund, such: 'always
have.2.SG.GER' = 'you should always have, make sure you always have'
etc.

Passive: there are two detransitivizers, which vary by verb. The
detransitivizer combined with an obligatory agent in the instrumental
case creates a passive. They have evolved from a proto-form, resp. a
benefactive and a malefactive. The detransitivizer in question used for
the verb depends on which of the proto-forms was most commonly
associated with it in proto-Elvish, e.g. 'cut' with a malefactive,
'give' or 'help' with a benefactive, etc. This is because in
proto-Elvish these acted a kind of middle between applicatives and
derivation (like in German) and therefore over time came to allow
intransitive usage (eg 'to give-for', or 'to cut-attheexpenseof', no
longer requiring an object). That form then morphed into an intransitive
in Old Elvish. The default form is the one derived from the malefactive.

The benefactive derived form is a suffix -s:a

The malefactive derived form is a suffix -wa (cf. proto-Elvish
*q*\ :superscript:`w`\ *ar* = 'cut', which is thought to be the remote origin of
the malefactive)

Clausal:
********

Negation:

Proclitic on first verb, noun, adjective, demonstrative, or pronoun:
emin- (from proto-Elvish 'e min', 'and not-be')

Qs:

- Polarity enclitic on subject, with SVO word order: -ami:n (from proto-Elvish 'or not-be', derived from negation marker, with vowel lengthening as result of emphatic shift)

- Interrogative enclitic on subject, with SVO word order: -ʃa (from defunct proto-Elvish verb for 'do')

Relative clause markers:

- Interrogative pronoun (inflected for case and noun class): θær (*pl.* θæri) ('who, what'), mainly used in independent pronominal position

- Verbal suffix -daθ (*orig. via 'da+θær'*), forming a relative form of the verb ('the one who died...' -> DEF.man die.PST.3.SG.REL)

- A third form is e.g. 'DEF.house live.GER.PST.1SG' -> "the house where I lived", "the house which I lived in" (lit. "the house of my past living") for subordinate clauses involving literal or figurative locational prepositional complements (in which, against which, to which, etc, contextually to be determined). Note the order change from the usual predicate-complement structure.

Possessives:
************

- Possessives are marked by the possessive enclitic -u, which attaches to the first noun of the possessed phrase, followed by the (non-obligatory) pronoun markers.

E.g. 'I get the dog's big white bone' = *t*\ :superscript:`h`\ *ulwə* *eulnda
galau ail yasi:lɑn:il* (get.1SG bone.3SG snow.like big DEF.dog)

Adverbials:
***********

There's some adverbial/conjuncting things that serve primarily temporal
and sequential purposes. These are generally clitics which can be
attached to any word in the clause for relevant emphatic effect, such
as:

- 'Then, and then, subsequently' is the proclitic e:-

- 'Then, at that moment' is the proclitic a:-

.. _`Derivational Morphology`:

Derivational Morphology
***********************

Old Elvish has a great deal of lexically specific derivational suffixes,
which precede TAM markers. These can be attached to the basic roots and
by doing so interact with relevant nominal or verbal morphology,
although some can be attached to either; as well as numerals and other
word classes where relevant ('to three were given rings...').There are
also some more 'general' derivational forms, such as locative ones.

The derivational forms can also appear as parts of predicate
constructions when combined with verbal morphology, e.g.
('shoe.inside.(NONPAST).3SG snail ' = the snail is inside the shoe).

Note that often it is also possible to express the same meaning
nonderivationally. Generally, this will have topicalizing or emphatic
effect.

*Nominal morphology*

- Diminutive: -il

- Augmentative: -ul

(Note that diminutives usually have a positive association and
augmentatives a negative one. Diminutives and augmentatives can also be
used on adjectives to mean 'a little' or 'very' respectively, with a
possible but not obligatory contextual meaning of 'too much' of either.)

- Collective or cluster, belonging together: -mɔn

- Surface, area, room, place for doing: -tal

- Ex-, former: i- (*past root*)

- Food: -ʃal, -ʃah

- -like, akin to: -au

- Time: C/_# -> s (e.g. *nathail* 'moon' -> *nathais* 'month')

- Stative nominalizer (state of being): -lɔn, -ɑn

- Habitual or institutionalized doer or performer of some V (for persons only), or producer or effecter of some V (for objects), or creature identified primarily with some specific action (for animals, spirits etc.; alternative to imperfective habitual verb form as noun): -sul

- Holding, containing (or holder, container): dun-

*Verbal morphology*

- Intermittently, occasionally: -k\ :superscript:`w`\ a

- Intensifier: reduplication of root (usually first V or CV)

- Causative: ɛs-

- Diminutive/de-intensifier: -un or -ɔn directly after the stem, before any verbal morphology. This indicates a semantic form of the verb with reduced intensity, e.g. 'a bit, a little'. Sometimes this has lexically specific meaning. It is not productive on all verbs. Ex.: o:n 'see' -> o:nɔn 'glance at, glimpse'; ɑn: 'stand, stay' -> ɑn:ɔn 'stay briefly'; ɪnd 'sleep' -> ɪndɔn 'nap *(verb*)'

- Together, joining two things: zu\- (cf. dual on certain nouns)

- Again ('re-'): ɪ+duplicate following consonant, else ɪh-

*Locatives*

- At (direction of), toward, onto: -(i)ð

- To, for, into: -G\ :superscript:`w`\ e

- With (com. and ins.), using: -(i)n

- Forward, in a straight line: -kau

- Out, away: o-

- On, at (static), on top of: -efe

- Far (both adj. and adv.): -yɑm (from distal demonst. morphology)

- Near, close (as above): -yap\ :superscript:`h`

- Around, to the back of: -ɪ