This study investigates the role of previously acquired linguistic systems, Mazandarani and Persian, in the acquisition of third language (L3) English at the initial stages. The data have been obtained from 31 students (age 13-14 years), testing the placement of attributive possessives in a grammaticality judgment task, an element rearrangement task and an elicited oral imitation task. The participants consist of three groups: The first two groups have Mazandarani as the first language (L1) and Persian as the second language (L2), but differ from each other with respect to the language of communication, Mazandarani and Persian, respectively. The third group has Persian as the L1 and Mazandarani as the L2, with Persian as the language of communication. English and Mazandarani pattern similarly in the target structures. That is to say, possessors precede possessed nouns and possessive adjectives come before nouns. In contrast, in Persian, possessives occur post-nominally. The results of this study reveal that none of the proposals tested (e.g. the L1 Factor, Hermas, 2010, 2014a, 2014b; the L2 Status Factor, Bardel and Falk, 2007; Falk and Bardel, 2011; the Cumulative Enhancement Model (CEM), Flynn et al., 2004; the Typological Proximity Model (TPM), Rothman, 2010, 2011, 2013, 2015) could account for the results obtained. This study provides support that at the initial stages of L3 acquisition, syntactic transfer originates from the language of communication, irrespective of order of acquisition.