ICLE-RC: International Corpus of Learner English for Relative Clauses
Debopam Das & Izabela Czerniak

ICLE-RC is a corpus of learner English texts annotated for relative clauses (RCs) and related phenomena (called 'other constructions' or OCs, representing four types: it-clefts, pseudo-clefts, there-relatives, and fused relatives). The corpus contains a collection of 144 academic essays from the International Corpus of Learner English (ICLE; Granger et al., 2020), representing six L1 backgrounds – Finnish, Italian, Polish, Swedish, Turkish, and Urdu. These texts are annotated for 940 RCs and 366 OCs, with respect to seven broad features (as listed below) and subsequent sub-features.

1. relative marker (RM)
2. grammatical function of referent
3. grammatical function of RM
4. embedding of RC
5. extraposition of RC
6. type of referent
7. restrictiveness

The corpus is further annotated for three additional meta-features: native language (L1 background), institution (the source institution and also the country), and gender (of the writer; male or female).

For more information about the corpus, see Das et al. (2025): https://aclanthology.org/2025.law-1.16/.
For more information about the annotation scheme, see the annotation manual included with the corpus.

The annotation in ICLE-RC was performed and can be accessible using the 2.8.16 version (or later versions) of UAM CorpusTool (http://www.corpustool.com/). 

===============================================================

A description of the directories, sub-directories, and data is given below:

The root directory, ICLE-RC-version-1.1, includes four sub-directories: (1) Analyses, (2) Corpus, (3) Results, and (4) Schemes. The root directory also includes a UAM CorpusTool project file, called ICLE-RC.ctpr, which can be used to open, view, and edit the annotations of the RCs and OCs. 

(1) Analyses: This directory includes a sub-directory, (1.1) All_files, which further includes 144 sub-directories containing the annotation for the all (sub-)features and additional meta-features. Each of these sub-directories begins with the name <docno>.txt, in which <docno> represents the number of the source text in ICLE-RC for which the annotation is provided. 

Each <docno>.txt directory includes three main files: (1.1.1) Metadata.xml, (1.1.2) relative-clause.xml, and (1.1.3) other-constructions.xml.

(1.1.1)	Metadata.xml: This includes information about the metadata of the annotation (language, encoding format, font type, and font size).

(1.1.2) relative-clause.xml: This includes the annotation of the target RCs.

(1.1.3) other-constructions.xml: This includes the annotation of the target OCs.

Each <docno>.txt directory also contain two additional files: (1.1.4) relative-clause.xml.old and (1.1.5) other-constructions.xml.old. These files represent earlier versions of the annotations for the RCs and PCs for the same <docno>.txt. 


(2) Corpus: This directory contains the source corpus for the annotation. This directory includes a sub-directory, (2.1) All_files, which further includes 144 .txt files, each with the name <docno>.txt.


(3) Results: This directory is empty, but it can be used to store different search results and statistics for ICLE-RC produced by UAM CorpusTool.


(4) Schemes: This directory includes three files, (4.1) ACRuleList.xml, (4.2) relative-clause.xml, and (4.3) other-constructions.xml.

(4.1) ACRuleList: This XML file is automatically produced, and contains only some meta-information about the annotation. 

(4.2) relative-clause.xml: This XML file contains the annotation scheme for RCs.

(4.2) other-constructions.xml: This XML file contains the annotation scheme for OCs.
