Research
Research program
Language, learning, and the architecture of access.
A coherent research program across four lines — three theoretical, one empirical — anchored in the question of how education systems open or close access to knowledge for the children and teachers they serve.
The program holds together across registers (theoretical, empirical, applied) because it returns repeatedly to the same site of inquiry: the conditions under which learning becomes possible — cognitively, linguistically, and institutionally — for learners in multilingual, postcolonial, and resource-constrained contexts. I describe each line below, with current status and forthcoming outputs.
Theoretical contributions
A configurational framework for instructional quality
The dominant frameworks for instructional quality were developed in monolingual high-income contexts and treat linguistic accessibility as an accommodation problem — something teachers do for multilingual learners rather than a property of the task itself. I am developing a configurational framework that names three jointly necessary, independently variable constructs.
Authentic learning activity is a design property of the task, with four dimensions: purposeful integration (constitutive), disciplinary authenticity, connection to learners’ worlds, and linguistic accessibility — treated not as accommodation but as a constitutive dimension of authentic task design. This is the central original theoretical move.
Pedagogical mediation is the in-the-moment classroom enactment that couples authentic task structure to learner cognition. It draws on the French didactique tradition (Chevallard, Brousseau), the CEFR Companion Volume’s mediation construct (North & Piccardo), and the CLASS-tradition observation literature (Pianta, Hamre) — positioned as a sharp triangulation rather than a synthesis, with five demarcation flags articulated explicitly.
Meaningful learning is the outcome property: learning that composes, transfers, and holds up under novel conditions. A higher bar than Ausubel’s classic usage, drawing on but not reducible to Ausubel, the transfer literature (Barnett and Ceci), schema construction, and the deep/surface distinction.
The framework generates falsifiable predictions through a 2×2×2 configuration grid, and resolves four empirical puzzles in the existing instructional-quality literature. The first prospective empirical anchor is a within-program configurational test in the MLF/OSUI Executive Master cohort 1 (October 2026), pre-registered before T=0.
Registerial multilingualism
The translanguaging-versus-plurilingualism debate in applied linguistics has been carried out at the level of framework choice. I am developing an account at the level of speakers’ competence: Moroccan multilinguals routinely navigate between two registers — a translanguaging register that is everyday, identity-anchored, and uses Darija as a matrix with in-clause switching; and a plurilingual register that is institutional, performative, and keeps named languages separate. The capacity to navigate between these registers is the competence to educate for.
The construct has three load-bearing components. Architecturally, the multilingual repertoire is organized register-wise (a frame property, not a style property). Interactionally, deployment is conditioned by interlocutor profile and participation framework. As competence, deployment is bounded by named-language capacity — the competence floor that differentiates this position from strong translanguaging accounts (Otheguy, García, Reid) and aligns it with Cummins’s crosslinguistic translanguaging theory.
The construct was first presented in an abstract submitted to the EduLingua Summit in 2026. A full concept paper is in development, with operational implications for ELT curriculum and for teacher education in linguistically complex postcolonial systems.
Language, AI, and learning
Generative AI creates a natural experiment in the dialectic of thought and language. If language is a vehicle for pre-formed thought (the instrumentalist view), AI’s capacity to translate, paraphrase, and reformulate is liberatory — it removes an artificial linguistic barrier for multilingual learners, particularly those navigating high-stakes disciplinary content in a non-dominant language. If language is constitutive of thought (the Vygotskian view), AI bypasses the very struggle through which knowledge becomes knowledge — the writing-to-learn tradition (Bereiter and Scardamalia) and the knowledge-transforming process it names.
I am working toward a resolution that takes the relationship as developmentally contingent: the scaffolding role of AI shifts with the learner’s zone of proximal development, the nature of the task, and the language in which the constitutive struggle is being conducted. The framework is currently in early conceptual stage; an operational instantiation is being developed as part of a LEGO Foundation Fellowship proposal (2026) on an AI-powered teacher coach for multilingual early literacy in Moroccan preschools.
Empirical program
Multilingual teaching and learning in early childhood
My doctoral research at Harvard examined how preschool teachers and administrators in linguistically diverse, English-medium community-based organizations in the United States support Dual Language Learners. Three papers are in publication-stage manuscript preparation: a quantitative study of adult-DLL language match in US preschool classrooms; a qualitative analysis of teachers’ translanguaging shifts during instruction, drawing on the stance/design/shifts framework (García et al., 2017) and the belonging framework (Aleksic and García, 2024); and a study of administrator and teacher perspectives on supporting DLLs in linguistically diverse settings. The shifts paper is currently in revision for the International Journal of Bilingual Education and Bilingualism.
Migrant parents’ preschool experience in Morocco
Phase 1 of a four-phase research program on early childhood education quality and inclusivity in Morocco. The study examines how first-generation immigrant parents — primarily mothers from Sub-Saharan African and MENA backgrounds — in the Rabat–Salé–Témara region experience and conceptualize access to early childhood education. The preliminary abstract has been accepted by the journal editors. The full program (Phases 2–4) extends to rural Amazigh families, children with disabilities, and low-income urban Moroccan families, with a synthesis year in 2029–2030.
Pioneer Schools and the teacher quality crisis in Morocco
A research program built around Morocco’s Pioneer Schools reform and the teacher-quality crisis the reform is implicitly attempting to address. The anchor data point: the 2016 ONDH/World Bank IPSE survey, which found 0% of Moroccan primary teachers met the minimum threshold in French, 3.5% in Arabic, and 67.2% in mathematics — and no SDI-equivalent assessment of the workforce has been conducted since. The system is reforming schools without re-diagnosing the teachers. A first publication, drawing on the TALIS-TKS 2024 public-use microdata (released November 2025; Morocco is one of two MENA systems in TKS), is in development with a quantitative co-author.
Education and climate change
A critical case study of Morocco’s national education response to climate change, structured around the Knowing-Doing-Being framework. The paper triangulates the Morocco evidence base across PISA 2018, TIMSS 2023, and OECD 2022 data, and forms part of a forthcoming book on global comparative climate education.
The pedagogy of collective songwriting
A research-practice partnership examining collective songwriting as a pedagogical practice — its developmental affordances, its relationship to language and literacy development in multilingual contexts, and its potential as a curricular form. The project sits at the intersection of three of my long-standing interests: music as a cognitive and developmental scaffold, embodied and participatory pedagogy, and the question of what kinds of knowledge live in collective practice that do not survive translation into individual instruction.
Funding and partnerships
Active and developing partnerships include Harvard Graduate School of Education, Université de Sherbrooke, Notre Dame, the Muséum National d’Histoire Naturelle (Paris), 88International, the Morocco Innovation and Evaluation Lab (J-PAL), the European Training Foundation, and the Regional Center for Quality and Excellence in Education. Funding pipelines under development with the Spencer Foundation, the Jacobs Foundation, the LEGO Foundation (SSRC-administered), GPE-KIX, and others.
For research collaboration inquiries, please write to mariam.dahbi@um6p.ma. For advisory engagements with international organizations and foundations, see the Advisory page.