The Interdependence Hypothesis is Cummins's claim that academic and conceptual proficiency in a bilingual's two languages is not stored separately but draws on a single, shared cognitive base. First formulated in the late 1970s and elaborated across his subsequent career, the hypothesis underlies the case for first-language instruction in bilingual education and is one of the most consequential ideas to come out of SLA for school policy.
Cummins (1979, 1981) framed the hypothesis as a developmental prediction: to the extent that instruction in language X is effective in promoting proficiency in language X, transfer of this proficiency to language Y will occur, provided there is adequate exposure to language Y and adequate motivation to learn it. The two conditions matter. Transfer is not automatic; it requires both real exposure to the target language and the will to use it.
The claim cuts directly against the assumption, dominant in mid-twentieth-century bilingual policy, that time spent on the home language was time taken away from English (or whatever the dominant school language might be). If proficiencies share an underlying base, then literacy and conceptual development in the L1 build the very capacities that the L2 will draw on later.
Cummins's expository device for the hypothesis is the dual-iceberg model. Two languages appear as separate peaks above the waterline, distinct in surface form: vocabulary, pronunciation, grammar, and the mechanics of reading and writing. Below the surface, the two peaks merge into a single common underlying proficiency (CUP) that holds the cognitively demanding aspects of language use: literacy strategies, abstract reasoning, conceptual knowledge, and academic registers.
| Layer | Content |
|---|---|
| Above the surface (separate) | Pronunciation, surface grammar, basic vocabulary, accent, the mechanics of decoding |
| Below the surface (shared CUP) | Literacy strategies, conceptual knowledge, academic vocabulary, reasoning, world knowledge |
The contrast Cummins drew was with the separate underlying proficiency (SUP) model, in which the two languages are pictured as two balloons in the head competing for limited capacity. The CUP picture replaces the balloons with a single tank that either language can draw from and add to.
The hypothesis grew partly out of empirical observations that puzzled the SUP picture:
These observations underlie the often-quoted distinction between BICS and CALP (basic interpersonal communicative skills versus cognitive academic language proficiency), which Cummins developed in the same period to explain why immigrant children often sound fluent within two years but take five to seven years to perform on grade level academically.
The pedagogical consequences of the hypothesis have been far-reaching:
Despite the criticisms, the Interdependence Hypothesis remains the dominant theoretical justification for L1-supported bilingual education, and the CUP model has held up better than most alternatives in subsequent neuroimaging and cognitive work showing substantial overlap in how the brain handles bilinguals' two languages. The framework also provides the conceptual anchor for the Threshold Hypothesis, with which it forms a paired account of when and how bilingualism helps or hinders cognitive development.