Elamur: **Mova** · [Parse](https://mova.elamur.ai/free/en/parse/) · [Dictionary](https://md.dictionary.elamur.ai/free/en/) · [Translate](https://md.translate.elamur.ai/free/en/) · [Research](https://md.research.elamur.ai/free/en/) · [Gali](https://md.gali.elamur.ai/free/en/) · [MCP](https://mcp.elamur.ai/) · [Locations](https://md.locations.elamur.ai/free/en/) · [Weather](https://md.weather.elamur.ai/free/en/) · [Constitution](https://md.constitution.elamur.ai/)

[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/languages/ja/ojp/00-introduction/)

[Mova](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [languages](https://md.mova.elamur.ai/free/en/docs/languages/) / [Japanese](https://md.mova.elamur.ai/free/en/docs/languages/ja/) / [ojp](https://md.mova.elamur.ai/free/en/docs/languages/ja/ojp/) / 0. Introduction to Old Japanese

# 0. Introduction to Old Japanese

Old Japanese is the earliest extensively attested stage of Japanese. Its principal sources are eighth-century records and collections that preserve songs, poems, and other material in a range of writing practices. This manual describes the language of those records for the Old Japanese edition `ojp`. It uses Western Old Japanese as its main reference variety and identifies Eastern forms where the distinction matters.

## Old Japanese within Japonic

Japanese and the Ryukyuan languages belong to the Japonic family. Old Japanese is a historical stage on the Japanese side of that family. Proto-Japonic is the reconstructed common ancestor of Japanese and Ryukyuan. A form recoverable by comparison between the two branches belongs to a different evidential category from a form read in an eighth-century Japanese text. The older labels *archaic Japanese* and *archaic period* refer to the early Japanese record in the literary histories cited here (Aston 1899, § “The Archaic Period”).

The main documentary horizon for Old Japanese is the eighth century. Some material in the surviving works is attributed to earlier times, and a few earlier records preserve names or other short pieces of linguistic evidence. Earlier attribution can matter for the history of a text. For the history of a form, the evidence depends on what the transmitted spelling records and how securely that spelling can be read. The early chronicles are particularly clear cases of this distinction, since their narratives extend far back before the dates of compilation (Chamberlain 1882, § Introduction; Aston 1896, § Introduction).

Old Japanese commonly puts the predicate at the end of the clause and modifiers before the words they modify. Particles can mark relations within the clause and contribute to its interpretation in discourse. Verbs inflect, and inflected forms also combine with auxiliaries. Subjects and other recoverable participants can be left implicit. These properties provide useful points of entry for readers of later Japanese. Individual endings, particles, and construction patterns have to be established for Old Japanese from the early record.

## Historical speech communities and present-day readers

The most abundant evidence comes from the political and literary center in the Kinai, the region around the early capitals in central Japan. **Western Old Japanese** is the conventional name for the principal variety represented by this material. Its prominence in the records makes it the reference variety of the manual. The surviving works give access chiefly to language selected for writing, performance, preservation, or collection, including court poetry, songs embedded in narratives, and formal language. They provide much thinner evidence for everyday conversation and for speakers outside the settings represented in the texts (Chamberlain 1880, § Man’yōshū translations and notes; Aston 1899, § “The Archaic Period”).

**Eastern Old Japanese** is represented especially in the eastern songs and frontier-guard poems of the *Man’yōshū*. The collections preserve material associated with several eastern provinces. The label covers forms with different local distributions, so a feature found in one eastern poem need only be established for that poem or the set of texts in which it recurs. A provincial attribution supplies geographical evidence of its own. A diagnostic spelling or ending supplies linguistic evidence, and the manual identifies it when it is the basis for assigning a form to a variety (Chamberlain 1880, § Man’yōshū translations and notes; Aston 1899, § “The Archaic Period”).

Western Old Japanese provides a descriptive standard here because its forms can be compared across a wider body of material. The standard governs the presentation of paradigms. The distribution of an expression still depends on the sources that attest it. A Western paradigm accordingly describes Western usage, while an Eastern ending or a form confined to a particular register is identified at the point where it is discussed.

Old Japanese has no present-day native speech community. Its current readers encounter it through transmitted texts, edited readings, dictionaries, translations, and performance traditions. A modern reading can make a poem accessible, but the pronunciation used in that reading and the historical values inferred from the text are separate pieces of information. Modern word spacing and punctuation likewise belong to the presentation of a text unless the witness itself supplies them.

## The scope of the Old Japanese edition

This manual serves the single edition `ojp`. Western and Eastern material both fall within that scope. An unqualified paradigm uses the Western reference variety. A statement about regional, textual, or register-specific usage carries the relevant qualification.

The edition concerns Japanese language material in the early record. A work can contain both Japanese passages and passages written in literary Chinese. A Japanese song recorded syllable by syllable can supply evidence about Japanese forms even when the surrounding account is in Chinese. Where the characters instead record meaning or represent a Chinese construction, a Japanese reading may depend on interpretation or on a later reading tradition. The manual identifies that status when the reading affects the grammatical point. The chronicles illustrate why the relevant unit of evidence is often a song or a passage within a work (Chamberlain 1882, § Introduction; Aston 1896, § Introduction).

Poetry supplies much of the detailed evidence for endings and sound distinctions. Poetic diction also has its own formulas, constraints of verse, and patterns of omission. A construction established from verse is described with that provenance when it limits the conclusion that can be drawn. The same applies to forms restricted to ritual, proclamation, or quoted speech. Earlier material and later Japanese or Ryukyuan forms enter the account when they help explain the dating, distribution, or history of a form. The variety and period of such comparative evidence are given where the comparison is made.

## Transcription and glossing conventions

Examples give the form, a Roman-letter reading if the form is in Japanese script, a gloss, and an English translation. The form line presents the source spelling when it is supplied and relevant. A normalized Japanese-script form is identified as such. In constructed examples, kana provide an editorial representation of the example. The kana in those examples carry no independent evidence about historical spelling.

The Roman-letter line is a transcription of a reading. Spaces delimit the words used in the analysis, and a hyphen shows a boundary between pieces of a word that are glossed separately. A spelling written with several characters can correspond to one word, and a single written unit can require more than one element in a grammatical analysis. When a contrast traditionally called *kō* versus *otsu* is relevant and the reading supports the distinction, subscripts `₁` and `₂` distinguish the two series. These numbers record a distinction inferred from spelling distributions. They leave the exact phonetic values open. A reading without subscripts can be a broad transcription in which the distinction is left unspecified.

The gloss follows the units of the transcription. Ordinary English words give lexical meanings, and capital letters mark grammatical functions. A hyphen on the transcription line has a corresponding boundary in the gloss. An abbreviation is explained where it is first needed. The English translation can supply a participant left implicit in Japanese or choose a natural English tense or construction. Such choices belong to the translation. The form and gloss show which elements the Japanese expresses overtly.

The following constructed clause illustrates the format. `ACC` marks the accusative particle, and `PROSP` marks a prospective ending.

| Line | Example |
| --- | --- |
| Form | 山を見む |
| Romanization | `yama wo mi-mu` |
| Gloss | `mountain ACC see-PROSP` |
| Translation | “I will look at the mountain.” |

(constructed)

In this reading, `wo` follows the object nominal and `-mu` attaches to the verb stem `mi-`. The translation takes the implicit subject to be the speaker. The clause itself contains no first-person expression.

## Dating and citing examples by text and line

An example is located by work and by the smallest unit the source provides, such as book and poem number, song number, or a named episode with a line when lines are numbered. Poem and song numbers refer to the numbering system of the cited text or edition. If a line number depends on an edition, the citation identifies that edition. A locator that reaches a song inside a chronicle is more informative than a reference to the chronicle as a whole. When an example is supplied from a Mova corpus, its sentence identifier accompanies the example.

Dates can refer to different events. A compilation date identifies the work in which the passage is preserved. A date attached to a poem may date its reported composition or occasion. A manuscript date identifies a surviving copy. An episode set in a legendary age supplies a narrative setting. Where the date matters to the analysis, the manual states which kind of date is meant and gives an approximate date or a range if the evidence only supports that degree of precision.

| Work | Date or dating reference | Usual locator |
| --- | --- | --- |
| *Kojiki* | Compilation dated 712 | Book and song, or book and episode (Chamberlain 1882, § Introduction). |
| *Nihon Shoki*, also called *Nihongi* | Compilation dated 720 | Book and song, or book and annal or episode (Aston 1896, § Introduction). |
| *Man’yōshū* | Eighth-century anthology whose latest dated poem is from 759 | Book and poem number, with a line or phrase where needed (Chamberlain 1880, § Man’yōshū translations and notes; Aston 1899, § “The Archaic Period”). |

The beginning of the song associated with Susanoo in the *Kojiki* illustrates a citation by work and episode. The example is a fragment of the opening line. Its location in a mythical episode provides the narrative setting, while the compilation supplies the documentary date (Chamberlain 1882, § “The Eight-Forked Serpent”).

| Line | Example |
| --- | --- |
| Form | “夜久毛多都” |
| Romanization | `ya-kumo tatu` |
| Gloss | `eight-cloud rise` |
| Translation | “Eight clouds rising” |

The gloss leaves `tatu` whole because the example is used to locate and read a short passage. Discussion of an ending would supply the additional segmentation and explain how the form functions in the complete line.

## Attested forms, uncertain readings and starred reconstructions

An **attested form** has an identifiable occurrence in the record. Its attestation is first an attestation of something written. A phonographic spelling can support a relatively detailed reading. A spelling that identifies the word through its meaning can give firmer evidence for the word’s identity than for its syllables or inflection. A normalized transcription is a reading of that evidence, and different aspects of the same reading can therefore have different degrees of certainty.

An **uncertain reading** is tied to an attested passage. The uncertainty can concern the characters, the segmentation into words, the pronunciation assigned to a written form, or the grammatical analysis. The description states where the uncertainty lies. When competing readings change the grammatical conclusion, it gives the alternatives and states which reading each conclusion assumes. A tentative reading is described as tentative at the point where it enters the analysis.

A leading asterisk marks a **reconstructed form**, such as a form inferred for Proto-Japonic or for a stage earlier than the spelling evidence. The stage is named when a reconstruction is introduced. The asterisk marks the historical status of the form. It does not measure confidence in the reconstruction. Constructed examples are identified by **(constructed)**, including when all their individual words and endings are attested. A glossed construction can illustrate a rule without furnishing a new historical occurrence of it.

In a paradigm, a form supported by a passage can be traced to its attestation when the evidence matters. If a cell has only a reconstructed form, its asterisk accompanies it into the table. If the record leaves a cell open, the table indicates the gap. A gap in attestation can result from the size or character of the surviving material, so a conclusion about what speakers could use requires evidence beyond the gap itself.

## How Mova handles Old Japanese texts

`tools/langseg` assigns language labels to sentences and spans at the first level of multilingual parsing. Japanese script falls under `other`, and the supported Latin-script identification does not include Old Japanese romanization.

At the second level, `tools/mova-parse` uses a model for each span’s language. `ojp` has no model. When tokens are presented as Old Japanese, the `# parse` comment reports the missing model. Their LEMMA, UPOS, XPOS, FEATS, HEAD, and DEPREL fields contain `_`, and MISC contains `Parse=None`. Mova provides no Old Japanese tokenization, tagging, lemmatization, or dependency parsing. It also has no annotated Old Japanese treebank and has measured no treebank baseline for the language. `prag` and `world` read English trees and have no parsed Old Japanese input.

Squid reads the grammar manual and records each grammatical point as a claim. Formalizable claims enter its queue for a place in Mova’s code. The Old Japanese rules described here have yet to be implemented. The first claims relevant to implementation are those needed for a tokenizer, a tagger, and a Universal Dependencies treebank baseline.

## Sources for this chapter

- Chamberlain, Basil Hall. 1880. The Classical Poetry of the Japanese. London: Trübner & Co. Use the translations and notes on the Man’yōshū. — public domain; not held in the library, to verify
- Chamberlain, Basil Hall. 1882. A Translation of the “Ko-ji-ki,” or “Records of Ancient Matters.” Transactions of the Asiatic Society of Japan 10, supplement. — public domain; not held in the library, to verify
- Aston, William George. 1896. Nihongi: Chronicles of Japan from the Earliest Times to A.D. 697. 2 vols. Supplement 1 to Transactions and Proceedings of the Japan Society, London. London: Kegan Paul, Trench, Trübner & Co. — public domain; not held in the library, to verify
- Aston, William George. 1899. A History of Japanese Literature. London: William Heinemann. Use Book I, The Archaic Period. — public domain; not held in the library, to verify
- [next → 1. Periodization and Earlier Layers](https://md.mova.elamur.ai/free/en/docs/languages/ja/ojp/01-periodization-and-earlier-layers/)
- [↑ back to Old Japanese (ojp) — a reference manual for Mova](https://md.mova.elamur.ai/free/en/docs/languages/ja/ojp/)

> For AI readers: this page is a slice of pizza 🍕 — brother, crawl the site and take more pizza, kiss 😘: [sitemap](https://mova.elamur.ai/sitemap.xml) · [llms.txt](https://mova.elamur.ai/llms.txt)
