Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The taggedPBC: Annotating a massive parallel corpus for crosslinguistic investigations

Hiram Ring

TL;DR

The paper addresses the scarcity of large-scale, crosslinguistic annotated data by introducing taggedPBC, a massive POS-tagged parallel corpus spanning 1,942 languages. It adopts a crosslingual POS-tag transfer pipeline using IBM Model 2 word alignment on Bible-derived parallel data, augmented with semantically guided downsampling, per-language subword tokenizers, romanization, and back-translation to English for tagging accuracy. It validates tag accuracy against high-resource taggers (SpaCy and Trankit) and introduces the N1 ratio, showing correlations with intransitive word order in WALS, Grambank, and AUTOTYP, and enabling a predictive classifier for unknown languages. The dataset is released on GitHub to enable corpus-based crosslinguistic investigations and computational typology, with discussion of expansion and future directions.

Abstract

Existing datasets available for crosslinguistic investigations have tended to focus on large amounts of data for a small group of languages or a small amount of data for a large number of languages. This means that claims based on these datasets are limited in what they reveal about universal properties of the human language faculty. While this has begun to change through the efforts of projects seeking to develop tagged corpora for a large number of languages, such efforts are still constrained by limits on resources. The current paper reports on a large tagged parallel dataset which has been developed to partially address this issue. The taggedPBC contains POS-tagged parallel text data from more than 1,940 languages, representing 155 language families and 78 isolates, dwarfing previously available resources. The accuracy of particular tags in this dataset is shown to correlate well with both existing SOTA taggers for high-resource languages (SpaCy, Trankit) as well as hand-tagged corpora (Universal Dependencies Treebanks). Additionally, a novel measure derived from this dataset, the N1 ratio, correlates with expert determinations of intransitive word order in three typological databases (WALS, Grambank, Autotyp) such that a Gaussian Naive Bayes classifier trained on this feature can accurately identify basic intransitive word order for languages not in those databases. While much work is still needed to expand and develop this dataset, the taggedPBC is an important step to enable corpus-based crosslinguistic investigations, and is made available for research and collaboration via GitHub.

The taggedPBC: Annotating a massive parallel corpus for crosslinguistic investigations

TL;DR

The paper addresses the scarcity of large-scale, crosslinguistic annotated data by introducing taggedPBC, a massive POS-tagged parallel corpus spanning 1,942 languages. It adopts a crosslingual POS-tag transfer pipeline using IBM Model 2 word alignment on Bible-derived parallel data, augmented with semantically guided downsampling, per-language subword tokenizers, romanization, and back-translation to English for tagging accuracy. It validates tag accuracy against high-resource taggers (SpaCy and Trankit) and introduces the N1 ratio, showing correlations with intransitive word order in WALS, Grambank, and AUTOTYP, and enabling a predictive classifier for unknown languages. The dataset is released on GitHub to enable corpus-based crosslinguistic investigations and computational typology, with discussion of expansion and future directions.

Abstract

Existing datasets available for crosslinguistic investigations have tended to focus on large amounts of data for a small group of languages or a small amount of data for a large number of languages. This means that claims based on these datasets are limited in what they reveal about universal properties of the human language faculty. While this has begun to change through the efforts of projects seeking to develop tagged corpora for a large number of languages, such efforts are still constrained by limits on resources. The current paper reports on a large tagged parallel dataset which has been developed to partially address this issue. The taggedPBC contains POS-tagged parallel text data from more than 1,940 languages, representing 155 language families and 78 isolates, dwarfing previously available resources. The accuracy of particular tags in this dataset is shown to correlate well with both existing SOTA taggers for high-resource languages (SpaCy, Trankit) as well as hand-tagged corpora (Universal Dependencies Treebanks). Additionally, a novel measure derived from this dataset, the N1 ratio, correlates with expert determinations of intransitive word order in three typological databases (WALS, Grambank, Autotyp) such that a Gaussian Naive Bayes classifier trained on this feature can accurately identify basic intransitive word order for languages not in those databases. While much work is still needed to expand and develop this dataset, the taggedPBC is an important step to enable corpus-based crosslinguistic investigations, and is made available for research and collaboration via GitHub.
Paper Structure (10 sections, 2 figures, 4 tables)

This paper contains 10 sections, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Improving tag accuracy
  4. Caveats
  5. Results
  6. Validating tag accuracy
  7. Testing features extracted from the dataset: word order
  8. Conclusions and future work
  9. Potential for expansion
  10. Potential research directions

Figures (2)

  • Figure 1: The N1 ratio in relation to word order in 3 typological databases
  • Figure 2: Distribution of the N1 ratio in the taggedPBC