Avallain and CEFR Alignment: Our Position Statement
Avallain recognises the valuable contribution of the Common European Framework of Reference for Languages: Learning, Teaching, Assessment (CEFR) to language education. Since its publication in 2001 and its updates through the Companion Volume (2020), the CEFR has profoundly impacted how students, teachers, institutions, publishers and other organisations across language education view, describe and define language proficiency and progress.
Central to this is the following key paragraph from the CEFR (2001) and Companion Volume (2020), which expresses a fundamental belief about the nature of language use and language learning:
‘Language use, embracing language learning, comprises the actions performed by persons who, as individuals and as social agents, develop a range of competences, both general and in particular communicative language competences.
They draw on the competences at their disposal in various contexts under various conditions and under various constraints to engage in language activities involving language processes to produce and/or receive texts in relation to themes in specific domains, activating those strategies which seem most appropriate for carrying out the tasks to be accomplished. The monitoring of these actions by the participants leads to the reinforcement or modification of their competences.’
(CEFR 2001 Section 2.1 / CEFR CV 2020, p.32)
If you wish to claim alignment, or even reference to the CEFR levels, scales and descriptors, your belief system about the principles of language education must be representative of the concepts in the above paragraph to ensure claims of alignment are valid and credible.
While the CEFR has been widely used in language education, it has also been misused and, in some cases, even abused. Some of the most significant examples involve claims of alignment to CEFR levels made without transparent principles or practices to demonstrate or substantiate the approach or outcomes achieved.
At Avallain, it is our responsibility to state our principles clearly and to demonstrate how we operationalise them in the way we reference the CEFR in our platforms and tools.
We believe it is essential to recognise the key role that language teachers have played in the original genesis of the CEFR descriptors, in their psychometric calibration to levels and in their usefulness for describing learner proficiency and progress. This means that teachers and other education professionals should be empowered by what the CEFR offers, not be submissive to it.
In addition, at Avallain, we acknowledge the following key principles of the CEFR, which organisations often overlook, claiming alignment of texts, tasks, tests and other resources to the CEFR levels and scales:
The CEFR descriptors do not describe ‘text difficulty’; they describe ‘person ability’. This means there’s no single, objective, ‘correct’ CEFR-level which can be assigned to a text, because:
- The difficulty of a text depends not only on the text itself but also on the learner or learners who are engaging with it and their purpose in engaging. The CEFR descriptors invite us to ‘decide’ whether the learners have the language abilities to achieve that purpose successfully.
- Any text under evaluation has both objectively measurable features, such as word count, structural complexity, lexical density, lexical frequency, etc. and subjectively assessable features, such as rhetorical organisation, degree of abstractness, topic familiarity, genre familiarity and so on. This complexity increases further in the case of audio texts.
- The CEFR is a pan-linguistic reference framework, not a framework specific to English. Consequently, all the (undoubtedly useful) projects and tools that computationally assign CEFR levels to English words, phrases, and structures remain experimental, each using different variables and methodologies, and often producing inconsistent results.
- Reputable language education organisations must acknowledge this inherent subjectivity in aligning texts and tasks to the CEFR scales and levels by empowering users to work with the descriptors contextually and critically, not imposing an objective, immutable model on them.
Consequently, the Avallain approach to CEFR alignment includes:
- A combination of computational learning and human-refined algorithms to assign CEFR levels to texts using their objectively quantifiable metrics, and the lexical and structural corpus databases resulting from international alignment projects focusing on words and phrases in English.
- A UX-designed interface allows users to make a priori decisions about text parameters and target-use situations before generating texts.
- An upcoming bespoke self-access training course to empower users (both end-user teachers and internal standard-setting panellists) to make informed, principled and pedagogically driven decisions about the appropriateness of CEFR-aligned outputs, considering text, task and student(s). This training course follows the principles of alignment recommended in the publication ‘Aligning Language Education with the CEFR: A Handbook’ (2022) produced by British Council, UK Association of Language Testing and Assessment (UKALTA), European Association of Language Testing and Assessment (EALTA) and the Association of Language Testers in Europe (ALTE).
This combination, therefore, reflects contemporary best practice in CEFR alignment through computational efficiency and data-processing power; large-scale, language-specific corpora databases; and teacher-led, context-driven decisions about learner proficiency and text/task difficulty.
A cognitively engaged decision-making process from a user of the CEFR, when working with an AI-generated, CEFR-levelled text, or another text which has been CEFR-levelled by a computer programme, could look like:
‘Considering the CEFR descriptor for Reading Correspondence at B1, this text looks a bit challenging for the task I’m thinking of because of X, Y and Z’
From there, a thoughtful user might respond in different ways, such as:
- ‘But I know my learners know a lot of this vocabulary because it’s directly related to their field of study, so I’ll go with it.’
- ‘I’m going to shorten the text by taking out some less-relevant sentences and replacing a few words with more familiar vocabulary that I know they’ve seen before.’
- ‘I’m going to make this an open-ended-response task, rather than a True/False comprehension task, and see what they can get out of the text for themselves first.’
- ‘I’m going to get the AI tool to regenerate a text with different parameters and see what that looks like.’
The underlying competences needed to make these decisions are profoundly human and related directly to the learner group under consideration. This is why we consider the combined processes of human-refined algorithmic analysis, user-informed interface design and training courses outlined in this position statement pedagogically sound. We will continue to uphold the key principles of the Common European Framework for Reference and its vital place in language education.
Developed for Avallain, April 2025, by Dr. Elaine Boyd and Thom Kiddle, Norwich Institute for Language Education (NILE)
