This PhD project aims to investigate the acquisition of L2 articles in European Portuguese by Serbian and Croatian native speakers (whose L1s lacks articles) focussing on their production of less explored complex ±definite DPs involving [±DP ofGEN ±DP] constructions in singular, first-mention contexts, which are assumed to be unambiguously indefinite (specific or non-specific) or definite (unique) in native grammars. We thus consider the effect of uniqueness (through the uniquely identifiable referent determined by world knowledge; ex.: 'the winner of the race') and specificity (determined by the speaker's knowledge of some particular referent; ex.: 'a participant of the race') and compare L1 and L2 speakers' preferences. There has been growing interest in the adult L2 acquisition of the nominal domain in the last decade, partially influenced by a series of studies by Ionin and colleagues about English article acquisition through semantic cross?linguistic comparison among languages that have and those that do not have articles (for other studies, see, e.g., García?Mayo & Hawkins, 2009). These raised a number of interesting questions for L2 research, e.g.: what kind of mental representation for articles will speakers form (if their L1 is articleless)? Is it possible for non?native speakers to converge on the kinds of knowledge of articles that native speakers have? In this study, we test the Fluctuation Hypothesis (FH) (Ionin, Ko & Wexler, 2004). For L2 learners whose L1 lack articles, FH assumes that full convergence is possible and predicts 'fluctuation' between specificity (treated as a semantic notion determined by the noteworthy property) and definiteness patterns until the appropriate value is set. The research is guided by the following questions: do L2 learners initially rely on either the specificity or the definiteness pattern, or do they 'fluctuate' between the two? Does the interpretation of uniqueness [+DP ofGEN +DP] require shorter (reaction) time to process than the interpretation of specificity [-DP ofGEN +DP] in singular first-mention contexts? We administered a fill-in-the-blank task as an on-line behavioural experimental task, testing four [±definite, ±specific] conditions of complex DPs in singular contexts (see example in (1)). Participants (n=79, age mean=23), university students from Belgrade and Zagreb, were distributed across four proficiency levels (A2+, B1, B2, C1), according to the information provided by their teachers; a native group was also included. The stimuli were presented on the screen and consisted of 212 pairs of grammatical sentences: 80 test items, 116 filers and 16 distractors. The experiment was conducted using DMDX (Forster & Forster, 2003). Our exploratory analysis shows that the interpretation of uniqueness requires less processing time in supplying answers than the interpretation of specificity. Preliminary results also challenge the notion that first-mention contexts are unambiguously indefinite or definite, which is not confirmed in specific-indefinite context, where participants supply definite articles in 53% (L1) and 65% (L2) of answers, indicating 'fluctuation' between patterns, even for native speakers. These findings suggest that specificity can alter the speakers' mental representation of (indefinite) articles. Example of test item ([-definite, +specific] condition): (1) Preciso de falar com [uma médica do Hospital Pediátrico]. É minha amiga. Tenho a minha filha adoentada. I need to speak to a doctor from the Children's Hospital. She is my friend. My daughter is ill. References: Forster, Kenneth & Jonathan Forester. 2003. DMDX: A Windows Display Program with Millisecond Accuracy. Behavior Research Methods, Instruments, & Computers 35 (1), 116–24. Ionin, Tania, Heejeong Ko & Ken Wexler. 2004. Article semantics in L2 acquisition: The role of specificity. Language Acquisition 12, 3–70. García Mayo, María del Pilar & Roger Hawkins (eds.). 2009. Second Language Acquisition of Articles: Empirical Findings and Theoretical Implications. Amsterdam: John Benjamins.