A diaspora is a population dispersed from an ancestral or symbolic homeland that maintains a collective identity, a sustained orientation to that homeland, and recognisable boundaries against full assimilation into host societies. The term comes from the Greek diaspeirein, "to scatter," and was originally applied almost exclusively to the Jewish, Armenian, and Greek dispersions before broadening in the late twentieth century to cover migrant communities worldwide.
Brubaker (2005) extracts three core elements common to most academic definitions:
Brubaker himself was sceptical of the term's runaway expansion. He argued that as "diaspora" came to apply to almost any migrant community, the concept lost analytic traction. He proposed treating diaspora not as a bounded entity to be counted but as a category of practice, a stance and a claim that communities and their members make.
Cohen (1997, Global Diasporas) divides diasporas into five types by the cause of dispersion:
The categories are heuristic, not exclusive. Most actual diasporas combine motives.
Diasporic settings produce predictable linguistic outcomes that connect the concept to ELT and SLA. The classic three-generation pattern: the immigrant generation is dominant in the homeland language and acquires the host language as L2; the second generation is bilingual, often dominant in the host language; the third generation is typically monolingual in the host language with at most receptive competence in the heritage language. This pattern produces the heritage speaker population that occupies a distinct space between L1 and L2 acquisition.
Where diasporic communities are large, geographically concentrated, and institutionally supported (schools, religious institutions, media), the heritage language can persist longer and develop distinct varieties. Diaspora Spanish in the United States, diaspora Arabic in Western Europe, and diaspora Vietnamese in Australia all show contact-induced features distinguishing them from the homeland varieties.
Diasporic English itself is significant: communities of English-speaking diasporans (Indian, Nigerian, Filipino) have produced new Englishes with their own conventions. Diasporic settings also produce code-switching norms, distinctive lexical innovations, and translanguaging practices that have become objects of study in their own right.
ELT classrooms in immigrant-receiving countries routinely contain heritage learners whose linguistic profile cannot be assumed. A student of Vietnamese parentage in Australia may have native-like Vietnamese phonology but minimal Vietnamese literacy; a speaker of diaspora Arabic may control a colloquial variety but not Modern Standard Arabic. Treating such learners as monolingual L2 English learners ignores resources that pedagogical translanguaging can deploy. Conversely, treating them as native users of a homeland standard misjudges what they actually control.
The diasporic context also shapes motivation: heritage-language maintenance is often the primary reason a diasporic family invests in their child's bilingualism, and English instruction that reduces the home language to a residue undermines the broader linguistic ecology the family is trying to maintain.