Abstract:The study examines several L3/Ln acquisition hypotheses and models to probe into the language acquisition beyond the initial state and factors that influence change of knowledge from one state to another. We focus on several cognitive and experiential factors leading to crosslinguistic influence without being properly considered by existing models:structural complexity, misleading input, deficiency of positive evidence, L3 construction frequency and dominant language activation, etc. After the analysis of previous findings, we hold that wholesale transfer does not occur at the initial stages for lack of necessity. Transfer may derive from either or both previously acquired languages, but not necessarily positive.