| Summary: | This thesis investigates two types of alternate languages: LUDLINGS (also known as language games, speech disguises, etc.), which involve primarily nonconcatenative morphological manipulation of their source languages, and SURROGATE LANGUAGES, which substitute alternative sound-producing mechanisms (whistling or a musical instrument) for the larynx. Chapter 2 explores the autonomy of surrogate systems in relation to both their own modalities and their source language phonologies. After presenting a formal analysis of Akan drum speech, I develop a complete model of the surrogate component. I argue that many properties which distinguish whistle surrogates from instrumental surrogates can only be attributed to the modular organization of this component. The last part of the chapter provides an inventory of the types of processes present in each module of the surrogate component. Chapter 3 presents theoretical treatments of representatives of each of the three major categories of ludlings (templatic, infixing, and reversing), beginning with the katajjait (throat games) of the Canadian Inuit. Although customarily regarded as a form of music, the katajjait are actually a well-developed form of templatic ludling. The implications of an infixing ludling in Tigrinya for tiered and planar geometry are then investigated. The chapter concludes with a detailed analysis of reversing ludlings, based on a parametrized version of the Crossing Constraint. In Chapter 4 I develop an integrated model of alternate linguistic systems, starting with an investigation of where in the grammar the ludling component is located. Drawing on data from more than fifty languages, I propose that there are three conversion modules in this component, each taking a well-defined level of representation as its input. In the last portion of the chapter I explore the possibility that one or more of these modules overlaps with the last module of the surrogate component. I conclude that the similarities exhibited by ludlings and surrogates are not due to a shared conversion module, but rather reflect the interaction of three factors: 1) the salience of certain levels of representation within the grammar; 2) general properties of the domains in which conversion takes place; and 3) membership in a common alternate linguistic component. Arts, Faculty of Linguistics, Department of Graduate
|
|---|