Fossilization, the stagnation of second language (L2) learning despite propitious conditions, is an inescapable reality for virtually all L2 learners. The study presented in this dissertation has endeavored to contribute to our current understanding of fossilization by examining, both longitudinally and cross-sectionally, persistent errors in the writing of adult learners of academic English for whom Spanish is a first language (L1). The theoretical framework is the Selective Fossilization Hypothesis (SFH), introduced by Han in 2009, which offers an extrapolative and explanatory framework for analyzing persistent errors in the developing grammars of L2 learners. This research was conducted in two parts. Part I consisted of a cross-sectional investigation of 60 English language learners (ELLs) grouped into three proficiency levels: low intermediate, high intermediate, and advanced. Part II was a longitudinal case study that followed two ELLs over a period of 28 and 56 months, respectively.
For both parts of the study, naturalistic data consisting of college placement, diagnostic, and exit essays were collected at the research site, a large community college in the Northeastern United States, and analyzed quantitatively. Descriptive statistics were computed to identify persistent errors in the participants’ writing. Following that, the longitudinal data were subjected to further analysis, revealing robust evidence of selective fossilization both among and within the target subsystems of English articles, prepositions, and number, and offering empirical support for the SFH. These findings have some implications for second language research and practice. By providing evidence of selective fossilization, they may help challenge earlier conceptualizations of fossilization as a global phenomenon, and, by extension, the myth of the ‘fossilized’ (‘unteachable’) learner. Additionally, they contribute to extant research on the developing academic writing of post-secondary learners, a population and genre largely underrepresented in the L2 research.
Finally, by offering empirical support for selective fossilization and the SFH, they provide L2 practitioners with the means to predict and explain learner errors, enabling them to set more realistic learning goals and achieve more successful outcomes.
