In order to account for crosslinguistic diversity in the coding of causal:noncausal verb pairs (e.g. seat/sit, scare/fear), various factors can be called into play, such as genetic affiliation, typology (Nichols et al. 2004), frequency (Haspelmath et al. 2014) or, more recently, contact (Bickel 2015). As a contribution of African linguistics, this paper investigates the coding of causal:noncausal alternation in two groups of languages spoken in West Africa. Atlantic and Mande languages belong to the same Niger-Congo phylum but display quite different typological profiles and have long lasting historical contacts in Senegal. Considering their respective typological profiles, two distinct strategies are expected to be favored for the coding of the causal:noncausal alternation: lability for Mande languages which are rather isolating languages with a limited set of derivational suffixes and regularly labile verbs (1); vs. voice opposition by means of nonderived vs. derived verbs (causative, middle or passive) for Atlantic languages which commonly display a large inventory of verbal extensions (2). Mandinka (Mande, Creissels & Bassène 2013) (1) dádaa ‘to repair’ / ‘be repaired’ Wolof (Atlantic) (2) noncausal > causalcausal > noncausal réer‘to be¬ lost’sakk‘seal’ réer-albe_lost-CAUS‘to lose’sakk-useal-MID‘to be sealed’ However, a first survey suggests the possibility of relatively important contrasts between languages belonging to the same family (Creissels et al. 2016), and shows that all possible strategies are actually attested in both families: lability, verbal derivation with all kind of pairing patterns (causal verb derived from noncausal, or conversely, or both verbs are derived) and also lexical alternation. Soninke (Mande, Creissels & Diagne 2013) (3) noncausal > causal bàamì ‘gallop’ bàamì-ndì ‘gallop-CAUS’ Joola-Keeraak (Atlantic) (4) LabilityLexical alternation ja-lɩɩkɛn-aj ‘learn’/’teach’ja-hɛj-aj ‘see’ jə-əs-əj ‘show’ In order to refine the picture and motivate the observed diversity, we decided to conduct fine grained analyses of the same 18 verb-pair meanings (Nichols 2017) across the two groups. First results show that diversity is greater among Atlantic languages than inside the Mande group. Thus, focusing on Atlantic group, we have first retrieved, from the RefLex lexical database (Segerer & Flavier 2011-2018), these same verb-pairs for 65 Atlantic languages, then for the 3 Mande languages in close contact with Atlantic languages. These data are substantiated with available grammars and dictionaries, and analyzed for (i) quantifying the various strategies inside the two language groups in order to measure the gap between the actual strategies and the canonical patterns for each group, (ii) trying to sort out whether the languages deviating from the Atlantic pattern follow areal distribution and can be accounted for by contact induced changes, or are better explained by internal factors. Our preliminary hypothesis is that: (a)suppletive lexicalization should be rare in both Atlantic and Mande languages but for opposite reasons: in Atlantic because of the numerous derivational suffixes, in Mande because of the overall lability; (b)part of the heterogeneity observed in Atlantic might be mainly due to the renewal of verbal extensions, as visible for instance in noncanonical pairings involving two derived verbs.