Extracting and generalizing rules from a complex input is one of the foundations of learning a language and can be studied using artificial grammar learning tasks. Different results have been found...Show moreExtracting and generalizing rules from a complex input is one of the foundations of learning a language and can be studied using artificial grammar learning tasks. Different results have been found, between artificial grammar learning studies but also within these studies. This thesis tries to explain some of this variance by looking into whether different study backgrounds influence the ability to learn an artificial grammar in adults. The experiment is based on earlier research conducted by Geambaşu (2018), in which she tested adults’ ability to learn an artificial grammar by using clear and vague instructions. She found participants who received clear instructions (explicit instruction condition) to perform better in the experiment than participants who received vague instructions (implicit instruction condition). The participants in the present task were divided into two groups: linguistics students and science students, since they differ most in their courses. They were exposed to the grammar via a passive familiarization and tested with yes/no grammaticality judgment task. The two instruction conditions and the stimuli were similar to those used by Geambaşu (2018). Participants who received clear instructions were expected to perform better in the task, than participants who received vague instructions. Further, participants with a linguistic study background were expected to have more explicit language knowledge and therefore perform better in the task, than participants with a science study background. Results indicate that there is no difference in the ability to extract and generalize rules from the sequences between students with a linguistics study background and students with a science study background. Further, no effect was found based on the different instruction conditions or for the pattern participants were exposed to. However, participants who showed explicit knowledge of the rule, by indicating a strategy based on extracting and generalizing the pattern of the sequences, performed significantly better.Show less
Nowadays, more and more people are raised bilingually, often with English as a second language. A widely accepted hypothesis about how bilinguals choose between languages for production, is the...Show moreNowadays, more and more people are raised bilingually, often with English as a second language. A widely accepted hypothesis about how bilinguals choose between languages for production, is the language non-selective access account, which states that bilinguals always initially activate both languages, and later (partially) suppress the non-target language based on contextual factors. Based on this account, it is expected that a speaker’s first and second language interfere while executing a language task. The present study replicates a covert phoneme monitoring experiment in Dutch, to see whether English (L2) interferes with Dutch (L1). The study focuses on covert lexical production with a task in which the participants have to accept or reject the presence of certain visually-presented phonemes in a picture name. Previous studies did not come to conclusive results (Geambaşu, 2010; Van Hinsberg, 2019), which may be due to the degree of bilingualism of the participants. To test whether participants would show slower RTs for graphemes occurring in their L2 than for unrelated graphemes, and whether this difference would be bigger for balanced than for unbalanced bilinguals, two groups of Dutch high school students were tested on their ability to monitor graphemes: those that are taught in Dutch(unbalanced bilinguals), and those that are taught in English (balanced bilinguals). The results did not show any significant differences in reaction times between the groups, suggesting that the degree of bilingualism does not affect interference. However, evidence suggests that the age of acquisition of English as L2 plays a role: the earlier one becomes actively bilingual, the more interference from the L2 s/he encounters in the L1. The graphemes that occurred only in the L2 English word, but not in the L1 Dutch, were most difficult to monitor for the early English learners, explicitly showing that L2 English interferes with L1 Dutch. However, interference is found not only in stimuli types where interference of English L2 was expected, but also throughout the whole L1 task: people who learned their L2 earlier in life, made more mistakes on the task than people who started learning their L2 later, indicating that an early bilingual speaker needs to put more effort into an L1 task to constantly suppress the L2.Show less
Prosody and structure are important cues for infants when they are learning a language. In this thesis, I investigated which of these two cues infants of seven months old find more salient. A Head...Show moreProsody and structure are important cues for infants when they are learning a language. In this thesis, I investigated which of these two cues infants of seven months old find more salient. A Head-Turn Preference procedure was used in both Experiment 1a and Experiment 1b to see whether infants found an inconsistent prosody pattern or inconsistent structure pattern compared to a familiarized pattern more interesting. Results revealed that infants had a longer looking time for the inconsistent prosody pattern than for the inconsistent structure, which indicates a stronger interest for the inconsistent prosody. If infants have a novelty preference, which is commonly assumed, this would mean that infants rely more on prosodic cues than structural cues. Whether or not this is the case will be examined further in Experiment 2. Also the points of improvement for Experiment 1 and the design of how Experiment 2 is conducted are discussed.Show less