Further Reading

Purpose of this page

This page collects further reading for the ALTE Common Corpus SIG project and the European CEFR Vocabulary Atlas pilot. It is organised by topic so that readers can follow the areas most relevant to their role: CEFR, multilingual Reference Level Descriptions, vocabulary profiling, corpus linguistics, language assessment, LLM-assisted classification, NLP and cross-linguistic comparison.

The page deliberately includes resources for multiple languages. The project is not an English vocabulary project. It is a multilingual, cross-linguistic pilot currently focused on English, French, Spanish, German and Czech.

1. CEFR framework and descriptors

These are the core CEFR resources. They provide the overall framework, descriptor base and implementation guidance for learning, teaching and assessment.

Resource	Why it matters	Link
CEFR main page	Central Council of Europe page for the Common European Framework of Reference for Languages.	Open CEFR page
CEFR Companion Volume	Updated CEFR resource with expanded descriptors, including mediation, online interaction and plurilingual/pluricultural competence.	Open Companion Volume resources
Searchable CEFR descriptors	Useful for checking communicative activities, strategies and descriptors by level.	Open descriptor resources
CEFR implementation resources	Useful for course planning, classroom teaching, assessment and action-oriented language education.	Open implementation resources

2. Multilingual Reference Level Descriptions

Reference Level Descriptions are central to this project because they show how CEFR content can be specified language by language. This is directly relevant to the European CEFR Vocabulary Atlas: vocabulary, grammar and constructions cannot simply be translated from English.

Language	Resource type	Why it matters for the Atlas	Link
All languages	Council of Europe RLD overview	Explains the purpose of Reference Level Descriptions and lists available language-specific CEFR descriptions.	Open RLD overview
English	English Profile	Important example of CEFR-related language profiling, including vocabulary and grammar resources for English.	Open English Profile
English	English Vocabulary Profile	Useful comparator for sense-level vocabulary profiling and CEFR vocabulary resources.	Open English Vocabulary Profile
English	English Grammar Profile	Useful comparator for data-informed CEFR grammatical profiling.	Open English Grammar Profile
French	Reference Level Descriptions / CEFR French resources	Relevant for French-specific CEFR content, vocabulary, grammar and communicative needs.	Open RLD overview
French	France Éducation international DELF/DALF	Useful assessment context for CEFR-aligned French certification and level interpretation.	Open France Éducation international
Spanish	Plan curricular del Instituto Cervantes	Major Spanish-specific CEFR reference resource for linguistic content and communicative learning objectives.	Open Plan curricular
Spanish	Instituto Cervantes / DELE	Useful assessment and curriculum context for CEFR-aligned Spanish language certification.	Open Instituto Cervantes exams
German	German RLD / Profile Deutsch context	Relevant for German-specific CEFR vocabulary, grammar and communicative content.	Open RLD overview
German	Goethe-Institut examinations and CEFR levels	Useful assessment context for German CEFR level interpretation and task design.	Open Goethe exams
Czech	Czech RLD / CEFR reference descriptions	Important because Czech morphology, case, aspect and constructions may affect receptive/productive CEFR judgements.	Open RLD overview
Czech	Czech language certification context	Useful for later comparison with Czech CEFR-aligned assessment and learner expectations.	Open Charles University UJOP

3. ALTE, language testing and assessment quality

These resources support the assessment side of the project: test specification, item writing, validation, comparability and quality management.

Resource	Why it matters	Link
ALTE	Main professional association context for multilingual language assessment in Europe.	Open ALTE
ALTE resources	Relevant for test development, alignment, quality management and validation discussions.	Open ALTE resources
Council of Europe assessment resources	Relevant to linking examinations and assessments to the CEFR.	Open CEFR assessment resources
EALTA	Useful wider European context for language testing and assessment research.	Open EALTA
ILTA	International language testing research context.	Open ILTA

4. Vocabulary profiling and lexical resources across languages

The Atlas needs to learn from existing vocabulary profiling work while avoiding an English-only model. These resources are useful either as direct comparators or as examples of language-specific lexical description.

Language / scope	Resource	Project relevance	Link
English	English Vocabulary Profile	Comparator for sense-level CEFR vocabulary profiling.	Open EVP
Spanish	Plan curricular del Instituto Cervantes	Spanish-specific reference for lexical, grammatical and functional content.	Open Plan curricular
German	Profile Deutsch / German RLD context	German-specific model for CEFR-related language description.	Open CoE RLD page
French	French RLD context	French-specific model for CEFR-related vocabulary and grammar description.	Open CoE RLD page
Czech	Czech RLD context	Important for analysing morphology, case government, aspect and constructions in Czech.	Open CoE RLD page
Multilingual	Reference Level Descriptions	Evidence that CEFR language content must be specified language by language.	Open RLD overview

5. Corpus linguistics and corpus-derived vocabulary evidence

Corpus linguistics provides the evidence base for frequency, dispersion, sentence examples, collocation and register. These readings and tools are relevant to all project languages.

Resource / topic	Why it matters	Link
Leipzig Corpora Collection	Relevant because the pilot uses Leipzig-style corpora and multilingual sentence material.	Open Leipzig Corpora Collection
Sketch Engine	Useful corpus platform for concordancing, word sketches, collocations and multilingual corpora.	Open Sketch Engine
Universal Dependencies	Relevant for multilingual POS tagging, syntactic annotation and cross-linguistic comparability.	Open Universal Dependencies
spaCy	NLP toolkit relevant for tokenisation, sentence segmentation, POS tagging and lemmatisation.	Open spaCy
UDPipe	Useful for processing Universal Dependencies models, especially in multilingual workflows.	Open UDPipe

6. Learner corpora and multilingual validation evidence

Learner corpora are important for future validation because they can show what learners actually produce at different CEFR levels. This is especially important for productive CEFR judgements.

Language / scope	Resource	Project relevance	Link
English	Cambridge Learner Corpus / English Profile context	Relevant model for using learner production to support CEFR profiling.	Open English Profile
German / Italian / Czech	MERLIN learner corpus	Important multilingual CEFR-linked learner corpus context for future validation thinking.	Open MERLIN
Swedish	SweLL learner corpus	Useful example of learner corpus design for CEFR-related research.	Open SweLL
Multilingual	Learner corpus research	Useful for validating productive difficulty, errors, collocations and constructional control.	Open CECL learner corpora resources

7. LLM-assisted classification and AI-supported annotation

LLMs are used in this project as classification assistants. The important questions are prompt design, structured outputs, multi-pass validation, expert review and evaluation against human-reviewed data.

Resource / topic	Why it matters	Link
OpenAI structured outputs	Relevant for forcing model outputs into controlled JSON schemas.	Open structured outputs guide
OpenAI prompt engineering guidance	Relevant for writing precise classification prompts and reducing ambiguity.	Open prompt engineering guide
CEFR-based sentence difficulty annotation	Useful example of sentence-level CEFR classification research.	Open CEFR-SP paper
UniversalCEFR	Useful recent multilingual dataset context for CEFR-labelled text classification.	Open UniversalCEFR paper
LLMs for German CEFR classification	Useful example of language-specific CEFR classification using LLMs.	Open German LLM CEFR paper

8. Multilingual NLP and processing tools

NLP processing quality affects the whole workflow. Sentence splitting, tokenisation, POS tagging and lemmatisation must work across languages.

Resource	Why it matters	Link
Universal Dependencies	Shared multilingual annotation framework for morphology and syntax.	Open Universal Dependencies
UDPipe	Relevant for multilingual tokenisation, tagging, lemmatisation and dependency parsing.	Open UDPipe
spaCy language models	Useful for English, French, Spanish and German processing in the pilot.	Open spaCy models
Stanza	Alternative multilingual NLP toolkit, useful for comparison and future QA.	Open Stanza

9. Cross-linguistic semantics and translation equivalence

This project depends on sense-level comparison. The same concept may be expressed as a single word in one language, a phrase in another and a governed construction in a third.

Topic	Why it matters	Useful starting point
Sense alignment	Prevents English headwords from being treated as the real unit of analysis.	Project methodology: concept and sense principles
Translation equivalents	Helps preserve differences between natural forms across languages.	Project rationale: cross-linguistic use cases
Morphological load	Relevant for German and Czech productive difficulty, especially case, gender, aspect and governed patterns.	Project methodology: receptive/productive distinction
Multilingual CEFR comparability	Supports comparison without forcing false equivalence.	Open RLD overview

10. Assessment use cases: text classification and vocabulary load

These readings and resources relate to using Atlas-style evidence for assessment: text selection, vocabulary load checking, item writing, distractor review and cross-language comparability.

Resource / topic	Why it matters	Link
CEFR assessment resources	Relevant for relating tests and examinations to the CEFR.	Open CEFR assessment resources
ALTE resources	Relevant to validation, quality management and assessment standards.	Open ALTE resources
UniversalCEFR	Useful multilingual context for CEFR-labelled text classification.	Open UniversalCEFR paper
CEFR-SP	Useful sentence-level CEFR classification example.	Open CEFR-SP paper

11. Reproducibility, workflow and data governance

The project needs a transparent audit trail. This includes version-controlled scripts, prompt versions, taxonomy versions, model names, output schemas, logs and review decisions.

Resource / topic	Why it matters	Link
GitHub Docs	Useful for version control, branches, commits and GitHub Pages.	Open GitHub Docs
Google Colab	Project execution environment for running notebooks and scripts.	Open Colab
pandas documentation	Useful for CSV, DataFrame and review workbook processing.	Open pandas docs
OpenAI API documentation	Relevant for LLM classification, structured outputs and API-based workflows.	Open OpenAI docs

12. Suggested reading paths

Reader role	Suggested path
SIG member	CEFR main page → RLD overview → project rationale → workflow → methodology.
Reviewer	Methodology → reviewer area → CEFR descriptors → RLD language resources → function-tagging stage page.
Assessment specialist	ALTE resources → CEFR assessment resources → vocabulary load and classification resources → rationale page.
Corpus linguist	Leipzig Corpora → Universal Dependencies → corpus evidence and ARF → sampling stage page.
LLM / NLP contributor	OpenAI structured outputs → prompt design → UniversalCEFR → CEFR-SP → project scripts.
Language-specific reviewer	Relevant RLD or language-specific reference → methodology → reviewer area → sample rows for that language.