[bala · home]
[okadenamatī · reviews]
[mesaramatiziye · other writings]
[tedbezī · languages]
lulnez che walnez: torazī tedbez rīb nyer ne rād. epez lūz toraz torazī rād ňer ket.
Morphology and syntax: these are the gates of any language. But what is the gate to these gates?
— Sero lis-Deren, Tedibek che žolezī
“Old Saremite” (tedbez Saremuz) is the name given to oldest well-attested member of the Saremite language family; its descendent, Modern Saremite, is the vernacular language (coexisting with Old Saremite as a literary language) of the Kingdom of Sarem, located in the southern part of Arakhyt. forms of Old Saremite are attested in manuscripts and inscriptions originating throughout the lands between the Hel mountains and the southwest coast; correspondence in Old Saremite is attested as far north as Demesne Soekhis, in the Carun Lands, although as it comes from the pre-unification period it is difficult to date exactly.
Old Saremite is best known through its large body of literary texts, standard forms of many which were established in the libraries at Wel Timas and Emer Ša during the early Kingdom period, especially between the reigns of King Ārel (author of the Olenamatī Ārel Urbeš) and Queen Lār (the earliest manuscript of the standard Semar Tujem, the latest-attested of the major semar cycles, is dated from twelfth year of her reign). there are also scattered and often enigmatic examples of the language, including (mostly fragmentary) literary texts, from the pre-Kingdom period; this grammar will, however, focus on the so-called “standard recension” of the language as represented in the compromise forms developed in the libraries of Wel Timas and Emer Sha.
literary production in Old Saremite continues into the present day, and, indeed, many central works of the Old Saremite literary tradition were produced after the early Kingdom period scholars’ efforts at textual editing; most notable are the emergence of the aboz and other prose genres and the development of literary critical, magic-theoretical, philosophical, and scientific traditions in the language. the majority of these texts — although some of them do display dialectal features that are of significant linguistic interest — are written in the same standard recension of Old Saremite.
āma·ňomūziš beš anat aur rokereš:
cezicereš ikegeneš· ce er·jereneš· ocenet:
—dez azale maudemī eliš tedibet·
[…]He stepped out of the darkness into the light of the sun;
joy filled him, and he cried out, laughing:
“Do you hear the voices of the speaking world?
[...]— Semar Tadān, iv.12-14
Old Saremite’s consonant inventory is of average size and relatively standard distribution — nineteen consonants, including two phonemic affricates, tʃ and dʒ.
p t k b d gm n ɲtʃ dʒs z ʃ ʒl w jɾfor convenience, when phonetic transcriptions are given in the remainder of this grammar, r will be used in place of ɾ unless a strict phonetic transcription is called for.
Old Saremite has eight contrasting vowels, five short and three long.
i iː u uːɛ ɔa aːfor convenience, when phonetic transcriptions are given in the remainder of this grammar, e o will be used in place of ɛ ɔ unless a strict phonetic transcription is called for.
there are, additionally, two phonemic diphthongs, ai and au.
the structure of an Old Saremite syllable is (C)V(C); diphthongs other than ai and au are nonexistent in native Saremite roots, and consonant clusters are uncommon — the presence of one in a word (e.g. tʃaːkten “stranger”) is generally an indication either of syncope in or from Proto-Saremite, or else an indication that the word originates in the otherwise unattested substrate language that influenced the Old Saremite lexicon.
stress in most Old Saremite roots is usually predictable: roots ending in consonants are stressed on their final syllable — except for a large class of verbs ending in -en — and roots ending in vowels are stressed on their penultimate syllable. prepositions and some particles (e.g., dez, the binary interrogative particle) are unstressed.
when affixes — either inflectional or derivational or both — are applied to a root, primary stress will remain in its original position unless this would cause it to fall earlier than the antepenultimate syllable, in which case it will move to fall on the antepenultimate syllable. as such, while roots — typically mono- or disyllabic — are always stressed on either the ultima or penult, words produced with Old Saremite’s highly productive set of derivational affixes may be stressed on any of their final three syllables.
the deverbal suffix -(e)z is — with only one exception (ˈɔlnez “story”, “narrative”, “tale”) — always stressed.
secondary stress is placed on alternating syllables before and after a primary stress; so, for example, ˈlulɛnˌdal “affix”, luˈlɛndaˌliː “affixes”.
the realization of the plosives has lately come into question. the standard analysis holds that the series transcribed here as p t k to be unvoiced and the series transcribed b d g to be voiced; this will be the principle followed in this grammar. however, a growing body of analysis suggests that — as is the case in at least some Modern Saremite dialects — the difference between these series is of aspiration rather than voicing (pʰ tʰ kʰ vs. p t k), or at least that this is closer to being the case than the (presumed) voicing distinction. for convenience, this grammar will follow the standard analysis in its transcriptions.
there is — depending on register and dialect — some reduction of unstressed vowels, especially in rapid or informal speech. typically, unstressed a and ɛ will become [ǝ] and unstressed ɔ will become [ʌ], although the exact realizations will vary.
ɛ ɔ are also often — again depending on dialect — in free variation with [e o]. as [ɛ ɔ] are the primary stressed values, these are treated as the underlying phonemes.
there is limited evidence for some variation between t and l according to dialect. it is possible that this represents vestiges of some intermediate phoneme — variously possibilities have been suggested, of which *ɬ and *tɬ are the most widely accepted — but there is insufficient evidence (either direct or comparative) to draw any definite conclusion.
it has also been suggested — here on the basis partly of comparative evidence from the other the wider Saremic language family and partly by extrapolation from certain dialects of Modern Saremite — that the Old Saremite graphemes conventionally assigned the values s z and ʃ ʒ actually represented a set of dental and apical sibilants — s̪ z̪ and s̺ z̺ — respectively. this would perhaps account for the development of OS z into θ in some Modern Saremite dialects (typically alongside a collapse of ʒ into ʃ).
this grammar will adopt the monographic romanization recommended by Sauram lis-Laul (4 Sarat Urb.), as follows:
p t k b d g p t k b d gm n ɲ m n ňtʃ dʒ c js z ʃ ʒ s z š žl w j l w yɾ ra ɛ i ɔ u aː iː uː a e i o u ā ī ūelsewhere on this site, including in the lexicon, the common digraphic romanization for consonants is used instead, substituting ny ch sh zh for ň c š ž. personal and place names are transcribed according to their Modern Saremite realizations, substituting h for p and ignoring vowel length. titles of Old Saremite texts follow Sauram’s transcription.
in both systems, where it is noteworthy or where it is being specifically commented on, primary stress is marked by an acute accent on short vowels and a circumflex (replacing the macron) on long vowels — e.g., ólnez “story”, mâze “ghost”.
kināraye yorum be emer· muňaz kaserura er·panerumez: ňerumī onjīz dogenak· ce mešārī karāňum aid žas šur […]
The Kinaites came to the ford where the Kaserites awaited them; they were there for battle, and mešars did not see the new day [...]
— Goran ta-Denem, Olnezī kenaumim aur Tujem
nominals are the first of the two major parts of speech in Old Saremite. the field of Saremite linguistics generally prefers the term “nominals” to “nouns” because the scholarly consensus is that pre-Old Saremite forms of the language (and other Saremite languages, although the attestation of these is so fragmentary that it is difficult to make any categorical claims) nominals and verbals were less distinct than they have come to be in more recent forms of the language. this is supported by comparative evidence from across the wider Saremic family.
the structure of the nominal phrase overall will be dealt with in greater detail in the syntax portion of this grammar. the structure of an individual nominal is simple:
[root] + [collective suffix] + [number suffix]
it is believed that historically the collective could not coexist with a number suffix, but as some collectives became lexicalized or semi-lexicalized (e.g., pasodura “lexicon”, lit. “a group of words”), they began to be able to be marked form number (pasoduraye “lexica”). there are some examples of this process in the canonical semar cycles, and it becomes more common in early prose works like Lires lis-Yahir’s Olam·olenamatī.
we will first address number and then turn our attention to the collective.
Old Saremite distinguishes three numbers: singular, dual, and plural. the singular is the unmarked form of the nominal; the dual and plural are both marked by suffixes, which take two forms depending on whether the root ends in a consonant or a vowel.
as a point of historical interest, there is limited evidence both from pre-Old Saremite inscriptions and from other fragmentary Saremite languages to suggest that -wo and -ye are the oldest forms of these markers and that -ū and -ī result from contraction. the underlying forms have alternately been reconstructed as Proto-Saremic *ǝw *ǝy.
the collective suffix has three functions: first, as a standard collective marker; second, as an associative plural, especially with proper nominals; and third, as a pseudo-definite plural, although definiteness is not otherwise marked (morphologically or otherwise) in Old Saremite. additionally, as noted above, some former collectives have subsequently entered the lexicon as roots in their own right.
the collective suffix has two forms; generally:
while the collective is always realized as -ra after a vowel root, some consonant roots may also take -ra. as a general rule, a multisyllabic consonant root may take either -ura or -ra if its stress does not full on its last syllable. thus ólnez-ra “literature” (alongside the occasionally attested olnéz-ura), but only ever pasód-ura (never **pasód-ra).
use-wise, let us begin with the collective as collective. the use of the collective as such was common in early Old Saremite texts but became less so as time went on; the pseudo-definite use of the collective came to be more prominent in later periods of Old Saremite. in its most productive phase, the collective could be used either as a derivational suffix to produce collective nouns — for example, ritegura “galaxy”, literally “an assortment of stars” — or as effectively a number that could not coexist with the dual or plural (although this eventually became possible). collective nouns like ritegura and gejegura “an assortment of mountains, a mountain range” thus have a somewhat ambiguous status, especially since for Old Saremite speakers there was only one “assortment of stars” and only one “assortment of mountains” that either could refer to.
the collective may also be used as an associative plural, a usage that is especially common with proper nominals. this is attested in some pre-Kingdom writing in the names of political groups, and it appears notably in the Semar Tujem during the legendary conflict between the Kināra, the followers of queen-to-be Kinā, and the followers of the usurper Kaser, the Kaserura — literally the “Kinā-and-company” and the “Kaser-and-company,” respectively. the associative may also appear with common nominals; depending on context, buratura (from burat “aristocrat”) could mean either “assorted aristocrats” or “a lord and their retinue”. (in later renderings, especially in prose, of stories from the Semar Tujem, we see the application of number suffixes to these: the Kināraye are all the disparate groups of Kinā’s followers that gathered on the banks of the Sha for the final battle.)
finally, the collective may be used to form pseudo-definite plurals — this is the origin of a further subset of lexicalized collectives. pasodura, for example, etymologically means “a collection of words”, but in grammatical texts it comes to mean something more like “all of the [specific] words” — that is, the Old Saremite lexicon — and subsequently it refers to any language’s lexicon. the pseudo-definite collective sometimes (particularly when the nominal in question is a group of living things) takes plural verbal agreement; this is believed to be part of the origin of the neo-Old Saremite stylistic technique known as dīdnez, asymmetric verbal agreement where number is marked only on the verbal and not on the subject nominal (see the section on verbals for further commentary on number agreement).
Old Saremite pronouns are — with two exceptions — morphologically just a group of nominals. four of them behave entirely as normal nominals:
the collective forms of these pronouns typically mean something like “all of us”, “all of you”, “all of them”, and “all of whom/what?”.
in addition to these, there are two pronouns that behave syntactically as nominals but are not — at least in the standard recension of Old Saremite — marked for number, dūr and ne.
dūr is an anaphoric pronoun used in relative clauses to indicate the syntactic role of the nominal the clause is dependent on in the relative clause itself:
lūz bosa la deparak ket· yaz ījanez aid rap dūr.
lūz bosa-Ø-Ø-Ø la depar-ak ket-Ø yaz ījan-Ø-ez-Ø aid rap dūr
INTER be.able-PRES-PERF-SG PURPOSE say-VN what-SG REL know-PRES-IMPF-SG NEG already RELANAPH
What can I say that you don’t already know?
here, for example, dūr is the direct object of ījanez “[you] know”, indicating that the antecedent of the relative clause, ket, should be read also as the direct object of ījanez. by the same token:
nidal· yaz dūr karāňeš burat
nidal-Ø yaz dūr karāň-eš-Ø-Ø burat-Ø
person-SG REL RELANAPH see-PAST-PERF-SG aristocrat-SG
the person who saw the lord
here dūr is the subject of karāňeš “[they] saw”, indicating that nidal should also be read as the subject of karāňeš.
ne is an indefinite pronoun with a variety of uses:
in dialectally variant, later, and neo-Old Saremite writings both dūr and ne are sometimes marked for number (as in fact is standard in most modern Saremite varieties); in these cases, dūr will agree in number with its antecedent (although collective *dūrura is unattested) and ne in any of the first three senses will be marked appropriately for the context; for example, neye· yaz dūrī karāňeš burat “the ones who saw the lord” or newo rād “these two”.
(forthcoming)
(forthcoming)
(forthcoming)