Date of Degree


Document Type


Degree Name



Educational Psychology


Patricia J. Brooks

Committee Members

Bruce Homer

Valerie L. Shafer

Jay Verkuilen

Suzanne V.H. van der Feest

Subject Categories

Cognitive Psychology | Cognitive Science | Educational Psychology | Psycholinguistics and Neurolinguistics


second language learning, testing effect, production effect, retrieval practice, noticing hypothesis, Turkish


Adults generally find it difficult to learn a new language, yet exhibit remarkable individual differences in outcomes. Variation in second language (L2) learning is associated with input conditions (Morgan-Short et al., 2010) as well as learners’ aptitude (Dörnyei, 2005). Recent work has demonstrated benefits of retrieval practice in promoting L2 learning of grammatical patterns and vocabulary in both artificial and natural languages (Hopman & MacDonald, 2018; Keppenne et al., 2021). With that said, when retrieval practice is based on oral recall as opposed to a recognition test, it confounds potential benefits of repeated testing (Rowland, 2014) with those associated with overt articulation (Hintzman, 1976). Hence, the first aim of this dissertation was to disentangle the effects of testing and production on L2 learning.

The second issue I address concerns the relationship between metalinguistic awareness and L2 learning. Schmidt (1990) claimed that encoding of L2 information is not possible if the learner does not attend to and subsequently notice linguistic features of the language. Contrary to these claims, there is research showing that it is possible to learn grammatical features in the absence of awareness, suggesting that adults can learn implicitly (e.g., Grey et al., 2014). Therefore, the second aim of this dissertation was to explore relations between metalinguistic awareness and accurate comprehension of grammatical features. I used a miniature version of Turkish in a computer assisted language learning (CALL) protocol to elucidate the role of speech production in early learners’ comprehension of case and number marking, vocabulary acquisition, and development of metalinguistic awareness.

Undergraduates (N = 156) were presented with Turkish spoken dialogues via a CALL protocol delivered via Zoom, with three learning conditions manipulated between subjects. Learning conditions were as follows: (1) retrieval practice (i.e., generate answers to Turkish questions and produce answer aloud), (2) verbal repetition (i.e., verbally repeat a Turkish inflected noun aloud) and (3) comprehension (i.e., answer forced-choice comprehension questions). Participants completed comprehension pre/posttests assessing comprehension of Turkish number and case marking, training blocks corresponding with their assigned condition, a Turkish vocabulary comprehension test, and open-response questions gauging explicit awareness. Given the significance of individual differences in foreign language acquisition, language learning background and nonverbal ability served as control variables in all models.

Results showed that the retrieval practice group exhibited higher posttest scores overall. For comprehension of number/case marking, the comprehension group performed comparably to the retrieval-practice group; for vocabulary comprehension, the verbal-repetition group performed comparably to the retrieval-practice group. Differential effects of learning conditions on outcomes can be attributed to benefits of articulatory rehearsal for vocabulary learning, transfer-appropriate processing, and the testing effect via retrieval practice. Explicit awareness patterned similarly to case and number comprehension, although the effect of learning condition was not significant. While explicit awareness of number/case marking correlated with comprehension accuracy, some adults demonstrated above-chance comprehension without showing awareness, suggesting the occurrence of implicit learning. Contrary to predictions, Turkish number marking was more difficult than case marking which may be attributed to a serial position effect. That is, the number marker was in the middle of the word and the case marker at the end, which appeared to make it more salient to learners. Contrary to past research on production, posttest comprehension accuracy on ‘old’ and ‘new’ items did not differ, suggesting that learners could readily generalize what they had learned to new inflected word forms. Lastly, nonverbal ability was a robust predictor of comprehension (including pre and posttest), vocabulary, and explicit awareness, highlighting the role of nonverbal ability as a factor in language learning aptitude.

Findings of the dissertation determined specific roles for retrieval practice and production in promoting grammar and vocabulary performance at earliest stages of L2 learning. There are also educational implications concerning the design of L2 pedagogy in classroom-based contexts, commercial language learning applications, and CALL protocols. When the goal is to foster vocabulary comprehension, such platforms should focus on the promotion of production via articulatory rehearsal. For promotion of grammar comprehension, platforms should provide opportunities for repeated testing via recognition and recall tests.