This paper investigates sentence-level text reuse in multilingual journalism, analyzing where reused content occurs within articles. We present a weakly supervised method for detecting sentence-level cross-lingual reuse without requiring full translations, designed to support automated pre-selection to reduce information overload for journalists (Holyst et al., 2024). The study compares English-language articles from the Slovenian Press Agency (STA) with reports from 15 foreign agencies (FA) in seven languages, using publication timestamps to retain the earliest likely foreign source for each reused sentence. We analyze 1,037 STA and 237,551 FA articles from two time windows (October 7-November 2, 2023; February 1-28, 2025) and identify 1,087 aligned sentence pairs after filtering to the earliest sources. Reuse occurs in 52% of STA articles and 1.6% of FA articles and is predominantly non-literal, involving paraphrase and compositional reuse from multiple sources. Reused content tends to appear in the middle and end of English articles, while leads are more often original, indicating that simple lexical matching overlooks substantial editorial reuse. Compared with prior work focused on monolingual overlap, we (i) detect reuse across languages without requiring full translation, (ii) use publication timing to identify likely sources, and (iii) analyze where reused material is situated within articles. Dataset and code: https://github.com/kunturs/lrec2026-rewrite-news.
Figures & tables
Articles
Sentences
STA
FA
STA
FA
True matches
387
2,070
989
2,731
The earliest matches
—
698
—
875
False positives
419
2,376
1,117
3,288
Table 1: Number of sentences and articles matched using the similarity approach in the STA and FA datasets. An em dash (—) denotes “not applicable”. The earliest matches filtering only applies to the FA data.
Figure 1: Distribution of reused sentences over time: (A) STA (sentences taken from FA) and (B) FA articles (sentences reused by STA). The plots show how reuse varies by sentence position (with a granularity of: beginning, middle, end of the article) within each article. The color scale on the right indicates the number of reused sentences: zero – blue, through white up to red – ten reused sentences. The dotted line marks the boundary between the years 2023 and 2025 in the data distribution.
Figure 2: Word cloud illustrating recurrent lexical patterns in reused sentences, highlighting the shared editorial templates underlying coverage during major news cycles: corresponding to the most frequently reused sentences in (A) 2023 and in (B) 2025.
Appendix figures & tables3 assets
Supplementary material from the paper’s appendix.
Appendix
source text
target text
According to the United States Census Bureau, the CDP has a total area of , of which, of it is land and of it (0.05 ) is water.
According to the United States Census Bureau, the village has a total area of 2.8 square miles (7.1 km ), of which, 2.4 square miles (6.2 km ) of it is land and 0.3 square miles (0.9 km ) of it (12.32 ) is water.
Wiseman has a continental subarctic climate (Köppen Dfc).
Babylon Village is bordered to the west by West Babylon, to the north by North Babylon, to the east by West Islip, and to the south by the Great South Bay.
As of the census of 2000, there were 21 people, 7 households, and 3 families residing in the CDP.
As of the census of 2010, there were 12,166 people and 4,585 households in the village, with 2.65 persons per household.
The population density was 0.3 people per square mile (0.1 km ).
The population density was 4,975.9 people per square mile (2,864.3 km ).
There were 30 housing units at an average density of 0.4 sq mi (0.1 km ).
There were 4,768 housing units.
Appendix
Table 2: Example of the sentence alignment in the Webis-Wikipedia-Text-Reuse-18 corpus.
language
full text
different sentences
similar sentences
support
English
0.72
0.37
0.63
500
German
0.73
0.35
0.64
99
French
0.69
0.35
0.59
99
Croatian
0.71
0.37
0.63
99
Italian
0.74
0.37
0.69
100
Polish
0.69
0.35
0.63
99
Appendix
Table 3: Similarity scores for the Webis-Wikipedia-Text-Reuse-18 corpus.