English periphrastic causative constructions, i.e. constructions where a causative verb like make or get controls a non-finite complement clause, have been the subject of many studies representing different theoretical frameworks, among which generative grammar (e.g. Kastovsky 1973), the universal-typological theory (e.g. Wierzbicka 1998), cognitive linguistics (e.g. Hollmann 2006) and construction grammar (e.g. Stefanowitsch 2001). Most of the time these studies have focused on the way periphrastic causative constructions are used (or should be used) by native speakers of English. Fewer studies have considered the use of these constructions by non-native speakers of English (cf. Ziegeler & Lee 2009, Gilquin 2012). In this presentation, I adopt a constructionist approach to investigate the use of periphrastic causative constructions in two non-native varieties of English, namely English as a Foreign Language (EFL) and English as a Second Language (ESL). While both of these varieties correspond to L2s that are acquired in addition to the L1, the settings of acquisition are different (mainly an instructional setting for EFL and mainly a natural one for ESL), which could lead to some differences in the way the causative construction behaves in the two varieties. The present study is based on corpus data coming from the International Corpus of Learner English for EFL and from the International Corpus of English for ESL, and representing different L1 populations among the two varieties. Relying on a corpus of native English as a reference, I examine the well-formedness of causative constructions in EFL and ESL, but also their idiomaticity, which is measured through a collostructional analysis (Stefanowitsch & Gries 2003) of the lexemes occurring in the non-finite verb slot. For EFL, this investigation reveals, among others, that learners sometimes use non-standard patterns like [X cause Y Vprp] or [X make Y Vto-inf], and that they tend to produce certain infelicitous constructions, which display lexical preferences different from those of native speakers (e.g. make their norms legalised). These findings are compared with the results of the ESL corpus analysis. This study provides insights into the impact of the acquisitional setting on the behaviour of causative constructions, and hence helps to bridge the paradigm gap that exists between EFL and ESL (cf. Sridhar & Sridhar 1986, Mukherjee & Hundt 2011). More generally, it demonstrates the viability of construction grammar as a theoretical framework to conduct a corpus-based study of interlanguage since, given the right level of abstraction, this framework provides a tertium comparationis for the contrastive analysis of varieties that may not necessarily follow the same norms. The study also underlines the relevance of the collostructional method to perform a contrastive interlanguage analysis, by showing that in both native and non-native varieties words interact with constructions (though sometimes in different ways). Such considerations, hopefully, will contribute to a rapprochement between the constructionist approaches and second language acquisition.