| Exposure to L2 orthography may help L2 learners distinguish a novel segmental (e.g., Escudero et al., 2008) and tonal contrast (Showalter & Hayes-Harb, 2013). However, little is known about whether learners can associate two phonetic variants with the same lexical entry when the orthographic form of the word is provided during learning. Thus, the goal of this dissertation is to investigate whether exposure to L2 orthography could help learners link two free variants with one lexical entry and whether Taiwanese and American participants will show different degrees of reliance on orthography in word learning. Four experiments were conducted in the current study. Experiment 1 examined whether orthography can help learners learn the [?-u] free variation in an artificial language using a word learning paradigm modified from Hayes-Harb, Nicol, and Barker (2010). No effect of orthography on the learning of vocalic free variation was found in Experiment 1, and both L1 groups showed low detection of the vowel alternations. A follow-up picture naming task investigating learners' encoding and decoding of the words was conducted in Experiment 2. The results from Experiment 2 showed that orthography did help learners score higher in picture naming and also produce more new forms of the free variants. Experiment 3 and 4 paralleled 1 and 2, respectively, but the target alternation was consonantal (i.e., [p-b] or [t-d] counterbalanced). An effect of orthography on the learning of [p-b] free variation (but not [t-d]) was found in Experiment 3. Compared to Experiment 1, only Americans showed better detection of the consonantal alternations but not Taiwanese. Experiment 4, a picture-naming task, was a follow-up of Experiment 3. It was found that, unlike in Experiment 2, orthography in this case did not always help learners score higher in production. In sum, mixed results regarding the effect of orthography on the learning of free variation were found. In general, the effect of orthography was stronger in production than in recognition. The findings reported in this dissertation add support to the role of orthographic information in lexical access. |
