BAREC-LexRank: A Ranked Arabic Readability Lexicon

Summary

BAREC-LexRank is a ranked readability lexicon of 40K Arabic open-class lemmas (adjectives, nouns, proper nouns, and verbs), ordered from easiest to hardest to read and understand. Instead of assigning lemmas to a small number of coarse levels, it provides a continuous ordering that captures fine-grained readability differences between lemmas.

The lexicon was built using BAREC-LexBench, a benchmark of 36K expert-annotated lemma pairs, and is released together with two variants:

Config	Coverage	Ordering
lexrank (default)	Open-class lemmas	Fully model-ranked
lexcomplete	All lemmas, open- and closed-class	Fully model-ranked
lexsamer	Open-class lemmas	SAMER level order preserved; ranking only within each SAMER level

---

Languages

• Arabic (Modern Standard Arabic)

---

Structure

The structure of this repository is as follows:

.
├── Data/
│   ├── BAREC-LexRank.xlsx
│   ├── BAREC-LexComplete.xlsx
│   ├── BAREC-LexSAMER.xlsx
└── README.md

Metadata

Each file contains the following fields:

• Sort ID: Unique lemma identifier reflecting its rank (lower is easer).
• Lex: The lemma, with diacritics.
• POS: The part of speech of the lemma.

All Lex-POS pairs match the SAMA S31 analyzer.

---

Usage

You can download the files manually using the Hub’s user interface, or use snapshot_download to download all files at once.

from huggingface_hub import snapshot_download

snapshot_download(
    repo_id="CAMeL-Lab/BAREC-LexRank",
    repo_type="dataset",
    local_dir="path/to/local/dir",
    allow_patterns=["Data/*"]
)

---

Citation

If you use BAREC-LexBench in your work, please cite the following paper:

@inproceedings{elmadani-etal-2026-barec-lex,
    title = "A Benchmark for Arabic Lemma-Level Readability Assessment",
    author = "Elmadani, Khalid N. and Saeed, Mostafa and Taha-Thomure, Hanada and Habash, Nizar",
    booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
    month = oct,
    year = "2026",
    address = "Budapest, Hungary",
    publisher = "Association for Computational Linguistics"
}