Northian language: Difference between revisions

Jump to navigation Jump to search
Line 213: Line 213:


As for non-phonemic letters inserted into the text by the Runic writers, the custom has been to insert a dot under the non-phonemic letter to elucidate its nature.  Thus <GAΘRI̯Ā́ỤUŠ> ''gaθriiāouš'' genitive dual "of two genetrices" is for <GATRIAUUS> from *ǵn̥h₁tr̥yéh₂h₁us, where the {{wp|yod}} is a genuine glide but the dotted u is merely added to suggest the second u is not part of a diphthong with the preceding a.  If diacritics were not added, the mistaken impression may (fairly) arise that the word was *gaθriiā-ouš, which would not be correct.
As for non-phonemic letters inserted into the text by the Runic writers, the custom has been to insert a dot under the non-phonemic letter to elucidate its nature.  Thus <GAΘRI̯Ā́ỤUŠ> ''gaθriiāouš'' genitive dual "of two genetrices" is for <GATRIAUUS> from *ǵn̥h₁tr̥yéh₂h₁us, where the {{wp|yod}} is a genuine glide but the dotted u is merely added to suggest the second u is not part of a diphthong with the preceding a.  If diacritics were not added, the mistaken impression may (fairly) arise that the word was *gaθriiā-ouš, which would not be correct.
===Rectified orthography===
Starting from the 1970s, scholastic editions of the Gales began adopting the so-called "rectified orthography" which is a phonological interpretation of the received text.  Such an interpretation is a convenience for scholars not familiar with Runic orthography, which presents many graphical variations and non-phonemic inclusions adduced by recorders.  So, in place of ''hāuuərə'' or ''haōuuarō'', ''hāwṛ'' is printed, as it is the most straightforward form required by the metre and most likely provided the etymology.  Nevertheless there can be ambiguities in the phonologization of a text, which were extensively copied and edited in writing, without reference to the chanting tone; thus, above the phonology, some graphical conventions owe to the chanting tone, while others were editorial.
Unfortunately, the ancient chanting tone has been lost, and modern chanting tone proceeds directly from the text rather than verbal tradition.  Chanting from text began in the 8th century, probably in the Viking settlements, where chanting masters were short but text more widely available, and it became the norm at the expense the verbal tradition by the middle of the 13th century.  Even when the University Liturgy was compiled in the 1300s, the authors could not locate anyone who recited the liturgy from tradition, and all the chanting tone in use by the contemporary clergy simply reflects various manuscripts made recently.  Whatever quantity of verbal tradition that remained was likely eradicated when the liturgy was canonized in 1606, at which point the national priesthood became monolithic and adopted a single, textual tradition.


==Phonology==
==Phonology==

Revision as of 01:46, 3 January 2023

Northian
lāuuozdāguš (civil tongue)
Native toNorthern States
Native speakers
17,420,000 (2010)
Official status
Official language in
Northern States
Language codes
ISO 639-3
This article contains IPA phonetic symbols. Without proper rendering support, you may see question marks, boxes, or other symbols instead of Unicode characters. For a guide to IPA symbols, see Help:IPA.

Northian or Northian language is an Erani-Eracuran language. It is attested first as Galic Northian, through the corpus of ancient Northian cultic formulae.

Genealogy

The genealogical position of Northian in the Erani-Eracuran family is "a known difficulty" according to some scholars. Many scholars consider Northian to be a primary branch of the Erani-Eracuran languages, while the most popular grouping that remains viable is between the Northian and Nordic languages. The reconstructed ancestor of both language is called Proto-Nordic-Northian (PNN) by its proponents.

Nordic Hypothesis

For arguments in favour of this phylogenetic union, the oldest Galic text shows a striking reflex of what appears to be Cowgill's law, an early sound change in the Nordic family. Though Northian does not undergo Grimm's law or Verner's law and furthermore retains all vowels dropped in Nordic, Northian has "very strong reflexes" of Cowgill's law, e.g. in hišuuṓḫ "alive" (with regular changes before *w), compared to Nordic kwikwaz "id." This correspondence via Cogwill's law is nearly exact, and deviations are almost unknown.

Furthermore, Galic has some shared vocabulary with Elder Nordic, of substratic (i.e. non-Erani-Eracuran) origin, that have been argued to be of a bronze-age date, though they cannot be deemed conclusive evidence since they could have been borrowed independently. A few items have been found to have been modified by processes similar to Cowgill's law, though that is also not conclusive evidence of co-evolution, since the phonological shape of those borrowed terms cannot be ascertained; thus, what appears as shared reflexes via Cowgill's law may simply have been unmodified sounds borrowed directly and independently from the unknown source language.

On the other hand, there are strong reasons to think that features putatively connecting Northian and Nordic may be areal phonological changes or chance similarities. The well-noted correspondence via Cowgill's law may be an independent reflex of Erani-Eracuran laryngeals and not a shared sound change. The paucity of shared innovations can but does not necessarily argue against their commonality: since it is hypothesized that the Northians branched off from the Nordics by the Middle Bronze Age, it is admitted that shared innovations should be few. Additionally, the highly-evolved reflexes of PEE *s and *y, contradicting their remainder in situ in Nordic, argue against extensive aerial effects by contact.

The morphology of Northian is perhaps the strongest evidence against a shared ancestry with the Nordic languages. The astonishingly conservative morphology of Galic Northian most starkly contrasts with that of the more innovative Elder Nordic, and the earliest Galic texts are likely to have been composed 2000 – 1500 BCE. If Northian and Nordic languages did share a common ancestor in recent prehistory, it may require Nordic languages to have undergone alarmingly dramatic changes in the time following the Northian split and up to its attestation around 1200 BCE.

Alternatively, the split between Nordic and Northian may have occurred earlier than the (mostly archaeologically-motivated) split hypothesized in 1500 BCE. That chronology allows more time for the Nordic morphology to evolve away from the Erani-Eracuran system but also introduces complications in the chronology of its split from other EE languages, such as the Baltic and Slavic languages. Some authorities also think there is a risk of circular reasoning if a linguistically-motivated chronology is used to rationalize a linguistic chronology.

Since the 1950s, the earliest Northians were identified by many archaeologists as the Register Pottery Culture, which is considered a subtype of the Nordic Bronze Age. However, the Retention Hypothesis cannot be resurrected on these grounds because the archaeological Nordic Bronze Age, mainly construed through artifacts, cannot be simply equated with the speakers of Proto-Nordic language. In view of preponderating loanwords both to and from it, it is clear that the Nordic Bronze Age must have consisted of multiple linguistic and very probably ethnic groups as well, and it is equally if not more plausible that the Northians were a non-Nordic population but also of Erani-Eracuran extraction.

The vocabulary dissimilarity between Nordic and Northian languages are also somewhat difficult to account for, though some of this difficulty may be attributed to the generic and topical constraints in received materials.

Retention Hypothesis

The Retention Hypothesis, first advanced by ethnographic historians in 1910, states that the advent of the Iron Age in Nordic-speaking areas was accompanied by rapid cultural changes causing a suitably dramatic alteration of recorded language. In turn, it argues Galic verses consist of even older words and phrases, such that around the time of the hypothetical split at 1500 BCE, Northians and Nordics actually spoke the same language. This argument is founded on some lexical items in Nordic texts claimed to be Northian borrowings that show co-evolutionary traits with Elder Nordic. Galic texts would thus be characterized as a priestly poetic tradition lost by the Nordics with the emergence of an iron-age warrior society and "retained" by the Northians.

In support of the Retention hypothesis, the ethnographers attributed the phonetic development in Northian—described as the opposite of Nordic developments—to the needs of rhythmic chanting, which emphasized vowel quality and quantity over consonants. Long vowels in final position were metrically desirable and accordingly created by deletion of final consonants. The Nordic development from the common ancestor, on the other hand, was supposedly driven by the needs of conquest.

Nearly all linguists reject this hypothesis, as for one it cannot be proved and for another it takes a highly selective view of materials. Though some Galic verses may resemble a nonsensical jumble of words, passed down as a sequence of meaningless syllables, it is also in many places both lucid and fluid and consistently thus praised by contemporaries. Additionally, some Galic verses are demonstrably later than others, strongly suggesting it was actively composed based on contemporary language rather than fossils extracted from a completely unknown source.

Early Branch Hypothesis

In 1972, Kimuti argued that Northian may have been the first language to branch away from the common Erani-Eracuran proto-language some 5,000 years ago. Her arguments are founded on the primitiveness of the aspectual-modal system adduced from Early Galic. While there is no conclusive refutation of the hypothesis, it is considered by too many to be overly aggressive in arguing the primitiveness of various aspects of Northian grammar. While it is difficult to argue that the aspectual-modal system is a result of secondary simplification, its opponents have provided that many of the putative common innovations that would have defined a subsequent period of Erani-Eracuran unity and evolution are to be considered separate and parallel innovations.

This position has also been championed heterodox position that the Nordic languages may have been one of the earlier branches from Erani-Eracuran as well. Rather than regarding the simplified grammar of Elder Nordic as secondary losses, Matto regards them as evidence that such features never fully developed in Nordic as well. His theory has the advantage of unifying the explanations of certain similarities between Nordic and Northian phonologies, namely the prevalence of fricatives and the notorious correspondence via Cowgill's law. However, his formulation is criticized on the grounds that the attested behaviour of various "primitive" features are actually quite different. Tritter argues that Nordic languages strongly clue such features (the subjunctive for instance) once existed and then were lost, while Northian development is generally in the contrary direction, i.e. the subjunctive experienced elaboration later in the language's evolution.

Erani-Eracuran linguistics

Northian is valued in the reconstruction of Proto-Erani-Eracuran, the hypothesized, unattested ancestral language to many major families of languages. On the one hand, it preserves multiple archaic, unparalleled formations, particularly in nominal declension, which have clarified reconstructions of PEE ablaut and accent.

Its verbal system comparatively provides few distinct clues about the ancestral language, since Syaran and Tennite verbs are equally conservative. The most salient feature in historical terms is perhaps the misalignment of the aspectual and modal formations, with some grammars characterizing them as complimentary. Thus, in roots marked for durative aspect, the subjunctive often fails to display the same markers and is thus indistinguishable from the unmarked aorist root; this behaviour contrasts with the optative, which obligatorily uses the same stem in the present. In the aorist, however, the subjunctive and optative usually have the same unmarked stem.

On the other hand, Northian is phonetically very innovative, resulting in plethorae of allomorphs from the same sources, e.g. the nominative plural ending *-es reflected variously as -āḫ / as, -ōḫ / os, -ə̄ / as, and -iš. Allomorphy has contributed to the language's reputation for caprice amongst both ancient neophytes of the Northian religion and modern scholars, but it has likewise been suggested the same prevalence necessitated a system of strict rote memorization that preserved ancient sounds with great fidelity.

Pucter notes that the exceptional Galic conservatism preserved archaic features that are lost in virtually all other branches, such as zero-grade suffixes in nominatives of amphikinetic items, e.g. ϑàm /ϑaGam/ < *dʰéǵʰ-m̥, xom < *ḱóm-t, and xratuš < *krét-us, and the primitiveness of the feminine grammatical gender, which often failed to trigger agreement. In her view, it is "more economical in purely linguistic terms" to consider Northian as the sister of all Erani-Eracuran languages except Hittite, in view of their shared innovations. Under this view, Northian cannot be a Nordic language, and she admits the "very appealing" correspondence via Cowgill's law would have to be interpreted as a distinct preservation of Erani-Eracuran laryngeals.

It was a prevalent position in the mid-19th century that Northian was a Baltic or Slavic language, on the basis of similarity with the Satemized phonology of those languages and the proximity of Silua, which spoke a Baltic language. The sentiment in the Northern States that they were a distinct people from the Nordics, who and whose progeny have been cast in the light of oppressors or brigands, had bolstered this theory. However, Sir Kilby Tapper noted as early as 1877 that, other than a somewhat similar phonetic inventory, shared sound changes between Northian and Siluan cannot be found, and he proposed Northian to be an independent branch of the Erani-Eracura family instead. Additionally, it became clear by 1880 that Northian was a Centum language, and its fricatives were conditioned by phonetic environments rather than genuine reflexes of PEE palatalized consonants. Tapper's theories remain in use and are still defended by researchers who reject a Nordic-Northian grouping.

Forms and stages of development

Proto-Northian

The term "Proto-Northian" is used only loosely by the academic world as there is little to reconstruct outside of the attested Galic material. Notionally it accounts for some aberrant Epic Northian forms that appear to have diverged before the emergence of Galic forms, but none of these absolutely require an early divergence and can also be explained by later processes. More often it refers to Galic texts interpreted phonologically, removing its phonetic and orthographic peculiarities, which are thought to be the products of chanting practice above the underlying language.

There are a few phonetic changes that govern the changes from Proto-Erani-Eracuran to Galic, but not all can be reckoned within a neat chronology.

  • These Erani-Eracuran phonetic laws firmly apply in Northian:
  • Galic is somewhat similar to Tocharian in that it loses most of the distinction between labiovelar and velar consonants, though labiovelars can still be diagnosed by their escape from fricativization in initial position, instead becoming plain velars, and influence on *e. The distinction between velar and palato-velar consonants was lost earlier and cannot be at all detected in attested materials.
  • All plosives in initial position become fricatives, and likewise in all sequences of multiple plosives.
  • *e is coloured by several consonants in fashion not dissimilar to the hypothesized (and well-accepted) laryngeals, namely by /h i y/ to /a/ and by /u w/ and labialized velars to /o/. On the other hand, /w/ also colours *o and turns it into a new vowel written /ə/.
  • Laryngeals in intervocalic positions are quite accurately preserved as hiatuses and circumflex accents, though they are not separately written. *h₂ and *h₃ also devoice any bordering voiced fricatives, but not *h₁. When syllabic, laryngeals become /i/ before coronal consonants and /a/ elsewhere.
  • *s becomes /h/ intervocalically, which is strengthened to /ŋh/ between accented and unaccented like consonants (in practice always /a/). This /h/ colours neighbouring *e to /a/ and was a live phonological rule in Early Galic. *s becomes /z/ when bordering preceding a syllabic nasal or its reflex, even if it is written as a vowel. *s is retroflexed as /š/ when following /i u/ and does not become /h/, and before vowels an epenthentic /t/ is inserted. /z/ also disappears in front of /y/ and causes compensatory lengthening. In final position, *-s is lost after vowels, except before pauses.
  • Syllabic nasals become /ā/ after stops and /ąm ąn/ after nasals. If there is a laryngeal following the syllabic nasal, it is scanned short, and otherwise it is long. Syllabic liquids become /ar/ after stops and /ərə/ after resonants. These are long monosyllables (despite the orthography), unless following a heavy syllable, where they would be short.
  • Galic prohibits long syllables in final position closed by a resonant: all such syllables lose their resonants and terminate with vowels, with or without a change in vowel quality. Epic Northian, however, deals with these syllables by turning their final syllables short, generating the suspicion that the long vowels are a feature of the Galic chanting tradition and not the general language.

Galic

Attested in the eponymous Gales (hymns) themselves in various sources, the False Gales, Sacerdotal Verses, and the Didaskalic Material. These texts were initially studied by the 3rd-century grammarian Himinastainaz in his Words of the Northians, usually in contrast to the Epic language and set apart by the introduction "but in the old language..." This description was modified by Uppapō in his work Āni Himinaštaēnąm "Against Himinastainaz", which pointed out numerous errors Himinastainaz made. Part of Galic grammar is reconstructed by modern linguists, as ancient grammarians did not have a very good understanding of the language that underlay the Gales.

The Gales (G) themselves are divided into several "periods" corresponding to the time they were composed, collated, or redacted. The Period I Gales (G1) were canonized (ceased to be actively edited) around 1500 BCE and mostly composed around or shortly before then, though it also contains set phrases, probably worked into the hymns by poets, called Old Material (OM) that may be considerably older. For other considerations, the bulk of G1 cannot predate about 1800 BCE. The G1 and OM together represent some of the oldest extant Erani-Eracuran literature. Period II (G2) and Period III (G3) Gales were redacted in the 1300s and 1200s BCE, respectively.

The False Gales (FG) appear to be imitative poetry written in the same form but on mundane subjects, later given liturgical functions. Sacerdotal Verse (SV) are small liturgical additions that appears to be instructions for priests, later interpreted as part of the liturgy.

Didaskalic Material (DM) are prose texts written in question-answer forms attributed to the Three Didaskaloi of the 8th to 7th century BCE. This portion accounts for about 80% of extant Galic material. Yet while the DM corpus is grammatically Galic, it is clear that its authors did not use Galic as their first language, and the editors' grasp of the grammar can be questionable at times; thus, for the purpose of linguistic reconstructions of the parent language, the DM is less reliable than the G material. Nevertheless, as G is short and does not attest all grammatical forms, those found in the DM can be referenced to create a more complete, if slightly speculative, Galic grammar.

In all its forms, Galic retains many more characteristics of the Erani-Eracuran parent language than any of the subsequent forms of Northian. Morphologically, this is true of both nouns and verbs, though the simplification in the verb system is much more extensive than in the noun system.

Though it is clear the DM cannot represent anything but an imitation of the Gales, academics seem distant from any conclusion what sort of register the G material synchronically represents. On one extreme, Drs Mittan and Poor argue that the Gales were composed in the common, spoken register, and on the other, the discredited Retention Hypothesis states that the Gales artificially and systematically favours archaic constructions. Though the Hypothesis stands abandoned, a weaker form of its linguistic chronology is still espoused by some academics. The contention rests on the interpretations of relative innovations, which are more or less discernible: if they are contemporary to the poet, then the non-innovated forms are best considered employed archaicisms, and if they represent later emendations, then the non-innovated forms may be contemporary to the poet.

Galic phonology has been the subject of debate. For liturgical usage, the Gales were pronounced with Medieval pronunciation, but this has been known to be anachronistic. Some elements of Galic prosody and orthography strongly suggest that Galic phonology was different from Epic phonology, which is about 1,000 years younger. For instance, Galic reflexes of Erani-Eracuran syllabic resonants, i.e. *r̥, *l̥, *m̥, and *n̥, are very sensitive to the phonetic environment. *m̥ could appear as ā, a, am, əm, ąm, or ą̄m. Some of these are known as conditional allophones, e.g. the nasalization of the vowel following another nasal in a sequence like *-nm̥-. But the variation between -am and -əm is less well-explained this way. Similarly, *r̥ is reflected in final position usually as -ar, but as -ərə following a laryngeal or resonant. Some authorities hold that these resonants were not yet vocalized in Galic, giving rise to a plethora of indeterminate written forms.

A modern position is that Erani-Eracuran laryngeals had not yet disappeared entirely in Galic, down to the G3 period, and was possibly realized as /ʔ/ intervocalically and /ə/ around syllabic resonants. This would explain the disyllabic scansion of many Galic words that appear to have only one syllable as written, as well as the curious phenomenon that syllabic resonants bordering a laryngeal evolved to be short, whereas a long vowel outcome was the more usual in the broader Erani-Eracuran family. However, when the Gales were committed to writing in the 1st century CE, originally disyllabic words like māʔā "moon" certainly had become monosyllabic (though possibly stretched over three or four morae), as no extant orthography suggests of a second syllable. The second syllable is only inferred by a poetic line rendered either unmetrical or unetymological by the loss of a syllable.

The most powerful argument in favour of laryngeal attestation is its behaviour around *e and in auslaut, which is identical to the allphone /h/ that arose from *s. /h/ had the effect of colouring any adjacent *e to /a/, which is the same as the reconstructed behaviour of *h₂. However, the two sounds must have been either distinct or existed at different times, because /h/ systematically survives as a distinct sound, while *h₂ disappears. One conjecture states that after *h₂ disappeared, intervocalic *s shifted to the same sound as *h₂ and thus continues its colouring effects. In final position following vowels, both *s and *h₂ gave rise to long vowels except at the end of utterances, though before enclitics *s is restored and *h₂ is not.

Epic

The Epic language exists primarily in poetic works that describes the actions of adventurers. Epic poetry is of variable length but often thousands of lines. Epics are always set in the distant past but may obliquely reflect the poet's own times through direct or indirect references to places and events. The identities of individual Epic poets have never been discovered, and it is most likely that Epics were composed by multiple poets, adapted for audiences, and continually evolving as bodies of literature. Nevertheless, scholars consider that some Epics may have been composed by the same poetic schools based on similarities in diction and theme. The Epics are probably an oral tradition prior to their codification around 200 BCE, when Acrean scholars interested in Northian legends committed them to writing. The corpus of Epic poetry is large, with over 50 titles and 150,000 lines.

Medieval

Writing system

Alphabetic texts

The very earliest records of Northian words occur in writings done by Celtic and Syaran travellers who encountered the Northian tribes in the 8th or 7th centuries BCE, though these are typically no longer than a few words, and then most are personal and place names.

Runes

The corpus of early but complete writings in Northian are done in Old Acrean Runes, which is an Alphabet of 24 letters. The phonetic values of the Northian Runes, however, are at variance with those used in the Acrean language, because the languages' phonologies are different. It is possible that they were subsequently regularized at a later date, since the earliest surviving manuscripts do not predate the 3rd century.

Rune
Transcription F U θ A R K G W H N I Y P Z S T B E M L D O

The Runic orthography of Northian presents several challenges to understanding of the text, as the alphabet was not designed for Northian phonology.  At least as far as the early Northian scribes used Runes, they were completely agnostic to both consonant and vowel quantity. Long vowels were not distinguished from short vowels, and geminate consonants were written as single consonants. Only the first of successive vowels, even of different quality, is written. On the other hand, the Northians paid much attention to distinguishing syllabic vowels from glides, to the point of inserting purely orthographical glides to ensure that a separate syllable is not read as a glide closing the preceding vowel. Masters of Galic texts were evidently at great pains to preserve oral traditions in instances the orthography could not convey.

Several letters also present ambiguities that could only be resolved by etymology. <ᚷ> for example could represent /g/ or its allophone /ɣ/ between vowels, but it could also represent the retracted *s following a final nasal, and this was almost certainly nothing more than a glottal stop by historical times. It also represented a /k/ in final position following /i̯/, as in pai̯ḵ "peace, truce". <ᚺ> represented /h/ as well as /ḫ/ and /ḥ/. The latter two were silent except at the end of sentences; the former was a retracted *-s following *a- and *o- that would resurface before enclitics, and the latter represented a vestige of the laryngeal in coda position, which caused the vowel to become short at the end of sentences or when the next word began with a vowel.

<ᛃ ᚹ> represented consonantal /i̯ u̯/ or an orthographic convention to separate vowels separated by hiatus when they were otherwise liable to be combined into a diphthong. The use of orthographic letters like this may have originated in paedagogical contexts, but it has become common in printed Runic texts. Both <ᚨ ᛖ> were used to write epenthentic vowels that represented vocalized, interconsonantal laryngeals that should have been absorbed into preceding syllables separated by a sonorant; their presence suggests that such syllables existed in the period when some texts were composed and were important for the preservation of metre, though synchronically they were meaningless.

Modern printing conventions

Printers up to the 19th century exclusively used "alphabetized runes" to print the Northian canon. This means assigning a single majuscule Alphabet to represent a Rune, regardless what that Rune represents. Thus, by reading the text, the reader can read the received Runic text exactly. By custom, the canon is written majuscule, and commentary, minuscule. Alternative forms of the same text may be included as marginalia at the discretion of the editor.

Midway through the century, enhanced understanding of linguistics and study of the canon prompted printers of academic editions to insert diacritics in the received text. Common diacritics include the macron for long vowels, the schwa, and <Ā̊> for vowels that are written Runically as <A> but are known to differ in quality from normal instances of /a/. A second printing development by the 1880s included geminate consonants where they must be restored metrically or philologically, for example <ĀMƏTĀTĀ̊> would now be printed instead of <AMARATATA> for āmmərətātā̊, accusative plural "immortalities".

As Northian studies further evolved, many quirks of the Galic canon revealed in rhyming or metre were identified as regular reflexes rather than poetic license or errors in transmission. For example, the "O-of-variable-quantity" that is the genitive singular ending of athematic oxytone nouns was identified as the reflex of a final *-s that disappeared in continuous speech but not at the ends of sentences; it also appeared before enclitics. The ancient grammarians failed to connect these facts and instead ascribed them to other causes, and so the phoneme was never written Runically. It is now restored in academic editions, so <HĀTŌḪ>, genitive singular "being", is now printed instead of superficial allomorphs <HATO> and <HATŌ> for the same word.

A similar phenomenon puzzled the ancients where a word-final laryngeal caused vowel lengthening everywhere except at the ends of phrases; however, this did not trigger sandhi before enclitics. This has now been identified too, so <FNAŌMÍÑĪḤ> is printed instead of allomorphs <FNAOMINI> and <FNAOMINĪ> for fnaōmíñīḥ "two lungs".

The philological importance of vowels held in hiatus was confirmed in the 20th century as further evidence of laryngeals in intervocalic positions, which were often reflected as hiatus. However, Runic orthography did not reflect either the hiatus or the vowel after the hiatus—a major shortcoming if a phonological text is required—though in liturgies missing syllables from otherwise regular metres were often inserted ad hoc by chanters at agreed positions. For this purpose the interpunct <·> is used to denote a laryngeal and a following vowel. Thus, <YA·ƏƏ> is nowadays printed rather than <YAR>, which was known to be iambic and dysyllabic but often (unetymologically) read as yarā before the nature of the missing vowel was known.

As for non-phonemic letters inserted into the text by the Runic writers, the custom has been to insert a dot under the non-phonemic letter to elucidate its nature. Thus <GAΘRI̯Ā́ỤUŠ> gaθriiāouš genitive dual "of two genetrices" is for <GATRIAUUS> from *ǵn̥h₁tr̥yéh₂h₁us, where the yod is a genuine glide but the dotted u is merely added to suggest the second u is not part of a diphthong with the preceding a. If diacritics were not added, the mistaken impression may (fairly) arise that the word was *gaθriiā-ouš, which would not be correct.

Rectified orthography

Starting from the 1970s, scholastic editions of the Gales began adopting the so-called "rectified orthography" which is a phonological interpretation of the received text. Such an interpretation is a convenience for scholars not familiar with Runic orthography, which presents many graphical variations and non-phonemic inclusions adduced by recorders. So, in place of hāuuərə or haōuuarō, hāwṛ is printed, as it is the most straightforward form required by the metre and most likely provided the etymology. Nevertheless there can be ambiguities in the phonologization of a text, which were extensively copied and edited in writing, without reference to the chanting tone; thus, above the phonology, some graphical conventions owe to the chanting tone, while others were editorial.

Unfortunately, the ancient chanting tone has been lost, and modern chanting tone proceeds directly from the text rather than verbal tradition. Chanting from text began in the 8th century, probably in the Viking settlements, where chanting masters were short but text more widely available, and it became the norm at the expense the verbal tradition by the middle of the 13th century. Even when the University Liturgy was compiled in the 1300s, the authors could not locate anyone who recited the liturgy from tradition, and all the chanting tone in use by the contemporary clergy simply reflects various manuscripts made recently. Whatever quantity of verbal tradition that remained was likely eradicated when the liturgy was canonized in 1606, at which point the national priesthood became monolithic and adopted a single, textual tradition.

Phonology

Consonants

Consonants Bilabial Dental Alveolar Palatal Post-Palatal Velar Glottal
Stop Voiceless p t c k ġ
Voiced b d j g
Nasal m n ñ ŋ
Fricative Voiceless f θ s ś š x h
Voiced v δ z ž γ
Liquid r
Approximant w l y (w)

Vowels

Front Mid Back
Close i ī u ū
Mid e ē ə ə̄ o ō
Open ā a ā̊

Galic Northian possessed 13 vowels; in the most common interpretation, they were distributed symmetrically and contrasted closedness and backness. All six vowels could be either long or short, the long version being held for twice as long as the short and assumed to be qualitatively identical.

Grammar

Northian inherited a highly synthetic grammar from its parent language Proto-Erani-Eracuran. In the evolution from this reconstructed ancestral language, there have both been processes of inflection and deflexion.

Ablaut

Northian retains a visible system of Indo-European ablaut, or vowel variations depending on grammatical form. Ablaut is connected to accent, but their precise interrelationships are actively debated. In general, an accented syllable tends to exhibit a vowel, called a full grade (of that vowel), while the same syllable when unaccented tends not have a vowel, called its zero grade. Where no vowel is present, a class of sounds known as sonants (*r, *l, *n, *m, and *H in the Indo-European proto-language) functions like vowels. In nouns, the direct cases (nominative, accusative, vocative, and locative) will tend to have full-grade in the root and zero-grade in the ending, and vice versa in the oblique cases. In verbs, the distinction lies between singular and plural numbers.

Zero grade Full grade Lengthened grade
Ø *e > (variable) *ē > (variable)
*o > (variable) *ō > ō

Also characteristic of PIE morphology is the theme vowel, which interposes between stem and ending and requires a special set of endings in some cases. Because stem and ending tend to blur and fuse over time, the theme vowel generates new surface-forms even when it merely separates familiar forms on either side, but words with the theme vowel eschew ablaut completely and are thus more predictable in their surface behaviour.

Nouns

Nouns in Northian are divided into three genders (masculine, feminine, and neuter) and are marked for four numbers (singular, dual, plural, and collective) and eight cases (nominative, vocative, accusative, locative, genitive, ablative, dative, and instrumental). Adjectives agree with the nouns they modify, and pronouns with the nouns they represent, in gender, number, and case. The genders of most nouns are lexical (i.e. arbitrary), but in some cases they reflect biological gender.  Morphological shape is associated with gender but does not exactly predict it.

The singular number denotes a single member of a countable entity, and the dual number, two of them. Plural number denotes three or more of members or instances of a countable entity, and collective number the category of an uncountable entity. In general, nouns formed either plural or collective forms, but a small group of nouns are capable of forming both plural and collective forms, with contrasting meanings. The plural and collective forms of all neuter nouns agree with singular modifiers and verbs.

The nominative case was used for the subject of sentences, while the vocative served for direct address. Where the nominative had a lengthened strong stem, the vocative had the short stem; where the former ended or once ended in *-s, the vocative lacked *-s. The accusative represents the direct object of most verbs or the location towards. These three cases were called the direct cases. The locative case was used to denote location at, the genitive the possessor, the ablative the source from, the dative the indirect object, beneficiary, or location to, and the instrumental the means whereby; these five cases are termed oblique cases.

Most inherited PEE nouns in Northian are of the athematic type, which could be divided primarily into suffixed and unsuffixed (root) types. Suffixed nouns consist of a root and a morphological suffix and takes declensional endings after the suffix; root nouns take endings directly. Some suffixes have identifiable meanings, such as the -tāt- < *-teh₂t- which gives states of being, while others do not present a coherent meaning considering the nouns that employ it, synchronically or diachronically, as some suffixes are demonstrably generalized in Northian derivations.  Under phonetic evolution, many suffixes have altered the endings that follow them, and paedagogically it it more common to treat the suffix and ending as one element.

Nouns may also be divided into classes based on their ablaut pattern. Ablaut is the alteration of vowel quality or quantity in various positions without altering meaning. The old grammarians divide nouns into the oxytone and paroxytone classes, based on the shape of the genitive form, where oxytone nouns have an accented genitive ending, while paroxytone nouns have an accented root or suffix. At the level of the parent language, it is thought that ablaut is conditioned or at least intimately connected to accent: accented syllables had vowels (the full grade) while unaccented syllables had no vowels (zero grade). In the received text, however, vowels and accent correlation is inexact.

The descriptive fact that words can have only one lexical accent (excluding accents lent by enclitics) is held by many to obtain that words must (originally, at a remote level allowed by internal reconstruction) have only one full-grade syllable, co-inciding with the accent. In OX nouns, the accent usually falls on the root syllable in the nominative singular and therefore predicts a sole FG on the root and ZG in the suffix. However, the vast majority of OX nouns in Northian have an o-grade, unaccented suffix instead. The intrusion of the o-grade suffix is attributed by authorities to diverse sources, such as allophony of *o as unaccented variation of *e or the accusative singular where the accent is regularly found on the suffix.

There are two dozen or so OX nouns that have, as the accent predicts, FG root and ZG suffix, notably nouns with primitive meanings like ϑām /ϑaʔam/ "earth" and hazarə "arm". Some scholars take these nouns as very archaic remnants of a once-unified pattern of accent and apophony, stressing the unlikelihood of an analogical introduction of -am < *-m̥ in ϑām from PX, since there are no known m-stem nouns with PX accent. Moreover, the accusative xā̊ appears to reflect directly *ǵʰēm < **dʰǵʰem-m̥, and not *ǵʰōm (probably, cf. *gʷow-m̥ > *gʷōm > bā̊ "bull") > *xō which would be a more likely source of the o-grade. These authorities argue that the heavily-disfigured paradigm suggests that analogical levelling has not affected this noun, increasing the likelihood that the nominative θaɣam has an original zero-grade suffix.

Pronouns

Adjectives

Adverbs

Verbs

The description in this section will be focused on the language of the Didaskalic Material (DM), as the corpus is larger and the language better understood. The language of the Gales is considerably more primitive but more controversial, owing to metrical and other considerations.

The finite Northian verb, in the Didaskalic (or Late Galic) and Epic languages, is highly synthetic and marked for one of two tenses (present or past), three aspects (durative, eventive, and stative), five moods (indicative, injunctive, subjunctive, optative, and imperative), three voices (active, middle, and a defective passive), three persons (first, second, or third), and three numbers (singular, dual, or plural). The verb stem could also form derivatives with special meanings such as inchoatives, intensives, and desideratives. The verb agrees with its subject in person and number.

Structurally, the finite verb is made up of a root, suffix(es), and personal endings. The root gives the basic meaning of the verb and can undergo apophony, reduplication, and suffixation. When a root is combined with a suffix (including the zero suffix), a verbal stem that specifies aspect is formed. To the is further attached the optional modal suffix to specify either the optative or subjunctive, as well as appropriate personal endings, which give the voice, person, and number. The past tense is indicated by the augment e-.

Additionally, verb stems could also form active, middle, and passive participles, infinitives, and gerunds which are closely associated with the verb and can take their own subject and objects.

A verbal root is inherently imperfective or perfective, with the exception of the unique stative verb woida. When this root is used within its inherent aspect, modal and personal desinences may be attached directly without a suffix. When this occurs, the verb is said to be a root verb. To use the verb in the other aspects or to introduce other shades of semantic meaning (such as inchoatives, intensives, and desideratives), the root may undergo modification, and suffixes may be added; these are said to be derived verbs. While it is hypothesized that a strict system of derivational rules existed in the parent language, verbal derivation is prolific in Northian, and consequently many suffixes are used indiscriminately or compounded together, sometimes with overlapping meanings.

Each derivational strategy generally gives rise to an inflectional class in Northian, since the morphology associated with each derivation tended to create through regular phonetic change distinct and non-transparent sets of endings.

Unlike nouns, which often have unaccented syllables in full grade, verbal ablaut is much more in line with accent.

Generally, Northian verbs exhibit a "classical" but unsurprising structure that is confirmed in most aspects by the ancient Tennite and Syaran languages. As with verbs in them, the fininte Northian verb may be marked for aspect, voice, mood, number, and person. Under the canonical structure of verbs, each root had one of three inherent grammatical aspect when used to create a verb, and when the verb is in that inherent aspect, it can accept endings directly and does not need to be suffixed. When used outside of that inherent and implicit aspect, it is obligatorily marked in some wise to denote the altered aspect. Modal suffixes (forming the subjunctive or optative) follow aspectual ones, and personal endings are then attached to form a finite verb, or nominalizing endings for non-finite forms like infinitives, participles, gerunds, and supines.

It is notable that the present tense embraces the more marked primary endings, which are patently the secondary ending plus a hic et nunc particle of some shape. By extension, it is inferred that the secondary endings were unmarked for tense and merely encoded voice, person, and number. This explains the need for the augment e-, of adverbial origin and remaining such in Northian. This particle may not always be included, leaving the form identical to the injunctive (but see below), to be distinguished by context only. Moreover, the augment is only sporadically used outside of the present aspect; aorist and perfect stems are rarely seen with the augment in the earliest literature, though such forms grew in prominence.

The canonical analysis has been regarded as the result of a rationalized system of verbal derivation via suffixes and other markers like reduplication, but the description was much more idiosyncratic than the canonical analysis. There were multiple suffixes that impart the same aspectual value, with or without differing shades of meaning; additionally, same or similar marking strategies were used to obtain different aspects (thus reduplication was found in present, aorist, and perfect stems). This suggests that suffixes formerly with different semantics had fallen together and become indistinguishable markers of grammatical aspect, e.g. reduplication may have connoted iterativity, but by the last stage of the proto-language it merely indicated durative aspect via a fairly straightforward shift in meaning. Once aspect became an obligatory category and roots were assigned an "inherent" aspect, for every other aspect a marker was standardized. Such marked stems became "primary" (i.e. purely aspectual) derivations, from which more specific meanings are then obtained via "secondary" suffixes.  

The resulting situation, present across many Erani-Eracuran languages, was that a root formed a stem for each three aspect (durative/present, eventive/aorist, and stative/perfect), one of which was the bare stem itself, which was morphologically unmarked but then implicitly associated with its inherent aspect.

It is evident that the grammaticalization of aspect as a separate and obligatory category for all finite verbs was incomplete in Galic Northian. Many modal formations, though considered marked for aspect, are built from the root directly rather than a suffixed stem, as though the suffix retained a meaning more than mere aspect that was either not proper or not necessary to a given formation. The "split aorist" occurs with sigmatic aorists not showing the suffixal -s when used in the optative and injunctive moods. The s-suffix evidently conflicted with the optative suffix. Even more revelatory, subjunctives in the G1 period were made exclusively from the e-grade root and not from any suffixed stem, and it could be used where any aspect is implied; only later were subjunctive endings attached to aspect-marked stems.

In very early texts, the nasal-suffixed verbs (e.g. -naō ~ nu-) also eschew the suffix when used in the imperative and injunctive (which did not have a modal suffix). Linguists debate how this behaviour came to be and whether some underlying condition synchronically explains this divergence. The injunctive thus is not merely an unaugmented imperfect and has a contrasting stem, since the imperfect always has the same stem as the present. Considering the fact that many s-stem aorists lose their aspectual markers when used in the subjunctive and optative and suffixed presents of a greater variety lose theirs in the subjunctive and imperative, it has been surmised by some authorities, at a stage earlier than Galic, all non-indicative forms may have been unmarked for aspect. This leaves the optative built to present-marked verbs, which require the present marker, as the exception rather than the norm, and some authorities attribute it to a late emergence of the optative.

Some grammars of the Period I or Early Galic (such as Revised Handbook of Early and Pre-Galic of 1998) have attempted to abolish the distinction between markers for aspect and mood, describing them to be, at least morphologically, in complementary distribution: that is, a verb could either display aspect (and consequently be in the "indicative" mood, which is otherwise unmarked), or display mood and not aspect. The primitive finite verb is thought to surface as the (marginally attested) root injunctive. From this sprouts the aspectual injunctives and the modal forms, and the latest to emerge are the present, imperfect, and aorist tenses, which are built on the present injunctive with the hic et nunc particle and augment, respectively. The root injunctive survives, mainly in prohibitive constructions and secondary sequence, because modal function is respectivelly expressed adverbially and defined by the primary sequence, so synthetic morphology would have been redundant; yet it is also attested in cases where a root did not form a root stem—in the few cases like this, it is possible to argue the root forms were unmarked for aspect.

However, other linguists hold that it is difficult to show that aspectual meaning is absent in modal formations not explicitly marked for aspect, and proximal adverbs often elucidate "aspectual value" when it was not synthetically marked on the finite verbs. In their views, it would be more prudent to default to the later material where aspectual meanings were definitely present even in the unmarked bare root, and thus in any modal formations built upon them. They assert these "non-aspectual" forms are a small minority even in Period I Galic material, and metre prohibits the addition of suffixes to an hypothetical original with unmarked stems, so the "aspectual and modal forms" must have been original to the text. Furthermore, this behaviour of dropping affixes is often observed in only some affixes, leaving (notably) reduplicatives to keep their markers regardless of any modal function.

Furthermore, not only did this aspect-mood system develop and mature in Northian, it was a general tendency in all Erani-Eracuran languages to grammaticize aspects and moods in basically the same way. They assert that the a full-fledged verbal system, like the ones found in Tennite and Syaran, was inherited in Northian and not developed independently. Thus, a few forms that elude the developed grammar are to be treated as relics of genuinely great antiquity preserved by happenstance, but should not prevent the provision of a stereotyped paradigm of the "conjugational possibilities" from a single verbal root (which largely recapitulates what is known from the Tennite and Syranan languages).

Vocabulary

Galicisms

During the Epic and Late Canon periods, much literature shows departures from the standardized grammar of the Epic language then used as a lingua franca amongst the Northian tribes. Early grammarians attributed most, if not all, of these departures to imitation of the earlier Galic language, which was spoken about 1,500 years prior. Galicisms may have been employed, in the most direct form, as quotations from the actual Gales, which are the most ancient and hallowed part of the Northian corpus; however, they "are more often encountered as later concepts cloaked in the morphology of Galic language," according to Marmer.

An exhaustive concordance of galicisms in Epic literature was first attempted in 1505, and subsequent analyses have showed that, aside from direct quotations, there was a general trend for galicisms to become more ungrammatical and arbitrary as time passed, most logically explained by the opacity of Galic grammar in view of the contemporary language. For example:

  1. in verbs:
    • The form of present active third person plural endings, i.e. usually primary -énθi and secondary -áõṯ, contains an initial e that interacts with a stem-final laryngeal and can surface as a or o; in Galic this variation is fairly strictly reserved, but in galicisms it is often employed for stems which have no final laryngeal.
    • The second person plural ending of the perfect stem is similar, i.e. has an initial e, and by regular sound change should have the same coloration as the present third plural; in galicisms the coloration is often unetymological under the same writer.
    • Long endings, in Galic language a reflection of interconsonantal laryngeal in stem-final position, is employed where short ending is regular, e.g. -aēmi for -mi.
    • Thematic verbs are often given athematic endings.
  2. in nouns and adjectives:
    • The dative singular of feminine nouns often has a short stem vowel where a long vowel is expected; this alteration is usually attributed to a particular feminine stem in Galic which has been analogized to other feminine nouns on the basis of the gender and not the shape of the stem.
    • Proparoxytone and paroxytone endings are employed for stems which have oxytone accent, e.g. piθri for piθroi "to the father".
    • Creation of unetymological strong and weak stems.

After the establishment of philological studies of ancient literature in Northian academia, aberrant galicisms were considered evidence of the decay of linguistic purity. But this view was more recently challenged by scholars arguing that "correct" Galic grammar was available and simply not employed by writers using galicisms. They point out that for a grammatical treatise like Mino's to exist, knowledge of Galic grammar had to be extant and available, and grammatical correctitude was not the object of galicisms:

Most writers did not employ galicisms (mis)lead the reader into thinking that the text is actually from the Gales—if that were the case, to be remotely convincing, the entire text would have to be galicized and written in Galic metre rather than prose. Instead, we find galicisms employed emphatically and sporatically, always to significant nouns and verbs and never to grammatical particles and pronouns. A grammatical quirk sourced somewhere from the Gales is enough to create the general allusory effect.

It is emphasized in recent studies that many galicisms may not be consciously employed and represent dialectal variations retained in individual writers' vocabularies. The Didaskalic text, dated to the very beginning of the Epic age, c. 650 BCE, are cited as examples of legitimate survivals of Galic grammar.

Syntax

Number agreement

In the Galic corpus, only plural animate nouns (masculine or feminine) are capable of agreeing with plural adjectives and verbs. As a rule, neuter nouns of any number can always agree with singular adjectives and verbs and, if dual, with dual adjectives and verbs. This is attributed to the low animacy typical of neuter nouns, wherein the attribute of number is less salient and fails to trigger agreement. Additionally, many neuter nouns are not capable of forming genuine plurals, instead forming a "collective" stem. This is distinguished in the nominative by the long o-grade in the suffix, and in the oblique cases by zero-grade suffix and plural endings. For example:

  • wədā̊ "body of water, waters" from wodār "water"
  • oštō "bones, esp. skeleton" from ošti "bone"
  • θenō "firs" from θonū "fir".

Adjectives in Galic do not have collective forms, since adjectives are not entities that can aggregate into a semantically distinct set or collection. As Donny said, "a collection of good things are just a plural number of good things, while a collection of bones can mean a skeleton—contrasted with a number of arbitrarily selected bones." Instead, adjectives modifying neuter collectives follow the fairly strict rule that in attributive position, the neuter plural form is used and, in predicative position, the neuter singular is used. However, in the Didaskalic text, collective forms are innovated for i-, u-, n-, and s- stem adjectives and co-exist with neuter plural forms; in predicative position, the neuter singular is still used. Into the Epic age, the adjective agreement rule effectively breaks down.

Gender agreement

In Northian, adjectives agree in gender with the nouns they modify, as a additional strategy to clarify semantic connection between them. In this way, if multiple nouns can agree with an adjective via the same case and number, a specific referent can be identified though gender. Gender agreement between adjectives and the nouns they modify is obligatory in all forms of Northian.

Of a group of a single grammatical gender, the adjective agrees with their common gender. But Northian stands in contrast with the Nordic languages and in agreement with most other Erani-Eracuran languages that a group of mixed-gendered entities are referred to in the masculine. The propotion of the referents' genders does not matter; that is, in the statement maxraṇhā̊ nauuan hŋaa-ska xāti hān-ka nō "nine women and one man are tall", the adjective "tall" is grammatically masculine. Early grammarians comparing Northian to Nordic languages have coined the term "Nine Women Rule" to describe this rule, and the name may have originated in jest.

Culture

Other than being the civil tongue of the southern Northian states, various forms of Northian are used for religious and cultural proceedings. These forms of Northian are often petrified in the sense that their vocabulary and grammar are prescribed since the time of the Acrean Empire and can thus present challenges to native and secondary speakers of Northian trying to become familiar with the more sublimated Northian culture.

Galic pedagogy

Possibly dating to the creation of the Galic tradition itself, the training of Galic priests was centred on the memorization of the liturgical canon, which included the Gales. Ascribing great importance to the accurate reproduction of the Gales, trainee priests were given exceedingly rigorous drills to prevent wrong words or inaccurate enunciation from encumbering their delivery. Indeed, prior to learning the Gales themselves, the trainees were given preparatory courses.

In the first stage, trainees were required to undergo two years of vocal lessons, which consisted exclusively of the Galic vowels /a ā e ē i ī u ū o ō ā̊ ə/. During this stage, the trainees would sing long sequences of pure vowels or vowels interspersed with every consonant, such as /eəuiōiū/ or /nomokaraludubā̊wā̊/. The object of this lesson was to inculcate the ability to create the exact vowel timbre and duration required at quarter, half, full, double, and triple chanting pace, and whatever the phonetic enviornment, as it was a cardinal error for allow the quality and quantity of vowels to fluctuate.

There is some dispute when this phonetic tradition arose, since it is not attested until the 7th century, but it was widely adopted, perhaps in defence against the changing vowel system of the spoken language, which by that time had lost the distinction between /o/ and /ā̊/. The comparison with chanting traditions in Tennite culture, however, has attracted the most academic interest. The Gales are not connected at the literary level with the Tennite canons, but the training methods to master canons in the two cultures are strikingly similar. As the Northians and Tennites ultimately speak languages derived from the same ancestor, it has been suggested these traditions may have arisen from a common prototype.

Alphaism

The term "language of alpha" was coined by the 12th century scholar Vatingas of Puro to describe the proliferation of a and its varieties in the endings of nouns. a often arise when Northian regularly deletes final resonants and sibilants following long vowels and changes the preceding vowel to -ā, alters -um- and -un- internally and initially to -ā-, and changes -e following -i- to -a. Additionally, the thematic feminine, genitive singular and accusative plural of paroxytone and proparoxytone nouns, and the plurals of neuter nouns also frequently ended in -ā or its varieties. After the final consonant is deleted, the nominative (and frequently genitive) form does not disclose the actual stem of the word and can lead unsuspecting speakers to make grammatical errors where the full stem (such as the nominative plural) is required.

Alphaism was reportedly very troublesome amongst the Viking settlers who colonized the northern reaches of DNS in the 8th to 12th centuries. Texts produced by these speakers often transfer nouns ending in -r, -n, and -m to the thematic -ā declension and others with genitive ending -ō to the thematic -ō declension; the number of athematic nouns were thus sharply reduced, and only athematic nouns in the Nordic languages with identifiable cognates were preserved. Citizens of the southern cities joked at the expense of speakers employing innovated forms. Vatingas of Puro, who originated in the north, stressed to his countrymen that the language should be learned "carefully and without assumptions".

It should be remembered that letters like <ḥ> and <ḫ> which can be very helpful to distinguish different grammatical endings did not appear until the 19th century. Additionally, vowel length and nasalization were not indicated, so endings like -a, -ą̄, -ā, -ā̊, -āḥ and -āḫ were all written and printed as <a>.

There is some literature that examines the experience of Late Old Nordic speakers attempting to acquire a working knowledge of Northian in the 7th through 9th centuries. On the one hand, some nominatives would have been recognizable to the Late Old Nordic speakers, such as kinship terms mātar "mother" and things like faụuərə "fire", but their oblique forms would have been more distant, e.g. genitives mātūš and fivaṇġ. On the other hand, there were some oblique forms that were more recognizable, e.g. doruš "tree's" due to the loss of the strong ablaut forms in Nordic. Thus, there may have been pressure to reshape words according to their more recognizable stems and in a transparent paradigm such as the thematic declensions.

See also