لخّصلي

As the field of language acquisition advances, and particularly as it advances ahead of the field of bilingualism study, questions remain as to how people successfully acquire a third or additional language (henceforth, L3A; additionally, the language being acquired is referred to as the target language) in adulthood. Specifically, it is unclear exactly what role, if at all, Cross-linguistic Influence (CLI) plays in the very early stages of adult acquisition of a language. It is clearly observable that adults are able to become incredibly proficient speakers of an additional language - one must simply look at global migrant populations to find people who learn a new language to a high degree of fluency, even in advanced age. One must ask, then, that if people who are already bilingual learn a third language, what is the degree to which their existing bilingualism assists or hinders their language acquisition? How is it affected in the earliest stages of language learning?

The field devoted to the study of L3A is relatively nascent, particularly in the realm of

morphosyntax, and inherits a peculiar problem from the field of bilingualism study: that of multiple potential sources of CLI. As detailed in Rothman et al. (2019), this problem is accounted for in a variety of prior works and proposed models, which will be discussed in more detail in Chapter 2. These models typically fall into one of two categories: order of acquisition models, where order of acquisition is the key factor in determining a source of influence, and structural/typological models, which assert that the source of influence ultimately is derived from the typological or structural similarity between the third language and one of the two previously acquired languages. The usage of an artificial or constructed language as the research medium allows for a

researcher to conduct studies while controlling for prior knowledge and exposure of the target language, as well as offering the unique ability to control the variables of the third language being learned. This amount of flexibility allows for a surgical focus on aspects of CLI which would otherwise have too many variables to study in a succinct manner and with repeatable, verifiable results. Therefore, because language is a human phenomenon and thus is variable, the clinical ability to control for the variables is an invaluable tool available to the language research community.

Page 1 of 86

In this thesis, a Miniature Artificial Language (MAL) called Englike is used to facilitate research on CLI in the earliest stages of L3A by Norwegian-English bilinguals in a

quantitative study of 77 individuals. All participants speak Norwegian as their L1 and English as their L2, and by utilizing a MAL the current study is equipped to explore the absolute initial stage of L3A and how CLI interfaces with the acquisition of morphosyntax.

In the following chapter I describe the background which informs the current study, list the models of CLI in L3A, and discuss the relevant languages – Norwegian and English. In Chapter 3, I outline the design philosophy and research questions guiding the present study, as well as describe the study and the results of the pilot. In Chapter 4 both the present study and its results are analyzed, while in Chapter 5 I discuss how the results of the present study interface with the research questions and hypotheses outlined in Chapter 3; Chapter 6 concludes this thesis.