README.md

metadata

dataset_info:
  features:
    - name: text_id
      dtype: string
    - name: sent_id
      dtype: string
    - name: volume
      dtype: string
    - name: sentence_idx
      dtype: int32
    - name: text
      dtype: string
    - name: tokens
      struct:
        - name: id
          list: string
        - name: form
          list: string
        - name: lemma
          list: string
        - name: upos
          list: string
        - name: xpos
          list: string
        - name: feats
          list: string
        - name: head
          list: string
        - name: deprel
          list: string
    - name: designation
      dtype: string
    - name: ancient_author
      dtype: string
    - name: ancient_recipient
      dtype: string
    - name: archive
      dtype: string
    - name: collection
      dtype: string
    - name: provenience
      dtype: string
    - name: genre
      dtype: string
    - name: dialect
      dtype: string
    - name: language
      dtype: string
    - name: period
      dtype: string
    - name: ruler
      dtype: string
    - name: script
      dtype: string
    - name: object_type
      dtype: string
    - name: material
      dtype: string
    - name: museum_no
      dtype: string
    - name: date_of_origin
      dtype: string
    - name: has_date
      dtype: string
    - name: vol_title
      dtype: string
    - name: primary_publication
      dtype: string
  splits:
    - name: train
      num_bytes: 23684474
      num_examples: 7355
  download_size: 20030844
  dataset_size: 23684474
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
language:
  - akk
task_categories:
  - token-classification
tags:
  - akkadian
  - assyriology
  - cuneiform
  - ancient-near-east
  - dependency-parsing
  - conllu
  - glam
  - cultural-heritage
size_categories:
  - 1K<n<10K

State Archives of Assyria — Morphosyntactic Letter Corpus

7,355 fully annotated sentences from 2,586 Neo-Assyrian letters — the royal correspondence of the Assyrian empire, 8th–7th century BCE.

Every sentence carries full morphosyntactic annotation (lemma, UPOS, morphological features, dependency head and relation) and the letter's archival context: who wrote it, who received it, which archive it was excavated from, and where.

Row 0 is Adda-hati writing to the king, from the Northwest Palace at Nimrud:

ana šarri bēlīya urdaka Adda-hati — "To the king, my lord: your servant Adda-hati."

Fields

• tokens — {id, form, lemma, upos, xpos, feats, head, deprel} (CoNLL-U)
• text — the normalised sentence
• text_id (ORACC/CDLI P-number), sent_id, volume (saa01–saa21), sentence_idx
• Letter metadata joined from the ORACC catalogue: ancient_author, ancient_recipient, archive, collection, provenience, genre, dialect, language, period, ruler, script, object_type, material, museum_no, date_of_origin, designation, vol_title, primary_publication

Covers 10 of the 21 SAA volumes. All 2,586 letters join to a catalogue record.

Notes on the conversion

Two quirks of the deposit are handled here, and are worth knowing if you go back to the source:

• The deposit ships a concatenated/ directory holding 10 per-volume merged files — the same letters again, in UD-release form. They are excluded; including them would double the corpus.
• The *Renumbered files open with a provenance pseudo-sentence (saao / saa01 / P224395, tokenised but unlemmatised). These are dropped — they are not sentences of the letters.

Filenames in the source are not uniform (P<id>lookup, P<id>Renumbered, P<id>lookupRenumbered, P<id>_lookup, P<id>_intermediate, plus a few X/Q ids), so text ids are extracted by pattern rather than by stripping suffixes.

Caveats

• Akkadian is a broken-text tradition. Damaged/illegible signs appear as x runs in the text; they are preserved, not cleaned.
• Annotation quality descends from ORACC's lemmatisation; xpos and feats are sparse in some volumes.

Load

from datasets import load_dataset

ds = load_dataset("biglam/state-archives-assyria-letters", split="train")

# letters from one excavated archive, with their ancient sender
nimrud = ds.filter(lambda r: r["provenience"] and "Nimrud" in r["provenience"])
print(nimrud[0]["ancient_author"], "->", nimrud[0]["ancient_recipient"])

Source & credit

Matthew Ong. A Full Morphosyntactic Annotation of the State Archives of Assyria Letter Corpus. Zenodo, 2024-02-06. https://zenodo.org/records/10622984 — deposit CC-BY-4.0.

Underlying text and lemmatisation from ORACC (http://oracc.org/saao/), released CC0; adapted from the State Archives of Assyria print volumes. Please cite both.

@dataset{ong_2024_saa_morphosyntax,
  author    = {Ong, Matthew},
  title     = {{A Full Morphosyntactic Annotation of the State Archives of Assyria Letter Corpus}},
  year      = {2024},
  version   = {1.0.0},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.10622984}
}

Underlying text and lemmatisation are from ORACC, released CC0.