Browse results
Abstract
Although few linguistic corpora are available in Yiddish, there are numerous sources of so-called “found data” that can be adapted for language research, pedagogy, and resource development. We describe the steps taken to create the first speech synthesis (text-to-speech) program in Yiddish. A state-of-the-art TTS model, FastSpeech 2, was trained on a hand-corrected data set consisting of literary texts paired with audio narrations by native speakers of the Polish and Lithuanian dialects. A quantitative evaluation by listeners found that the system produced speech that was both intelligible and natural-sounding. To demonstrate the system’s applications for language pedagogy, we offer a qualitative evaluation of Yiddish phonological features that are present or absent in a sample of synthesized recordings. We hope that the success of speech synthesis in Yiddish will inspire future projects to enable technological support for other minority languages in which transcribed recordings are available.
Abstract
The aim of this essay is to problematize the ontology of Judeo-Spanish qua language. First, I argue that its traditional conceptualization as an autonomous, self-contained language is predicated on a (flawed) classical ontological framework that relies on so-called ‘named languages theory.’ Second, I contend that a more enlightened understanding of Judeo-Spanish as a linguistic phenomenon necessitates a paradigm shift toward a hauntological framework consistent with theoretical models such as translanguaging and revivalistics. I conclude that Judeo-Spanish is best understood as an ensemble of the only partially overlapping idiolects of people who share a common Sephardi cultural/ethnic identity and who manage to communicate with reasonable success. Third, I discuss the momentous implications of this shift in three domains: linguistics, minority rights, and education.
Abstract
In this short article we publish, with translations, a macaronic Hebrew cum Middle Greek religious poem, accompanied by a refrain in Ottoman Turkish, all written in Hebrew characters and fully vocalized. The text comes from a Karaite prayerbook printed in Venice in 1528 on behalf of the Constantinople Karaite community. This poem and its origins played a role in different manipulations of Karaite identities and history during the 19th and 20th centuries.
Abstract
Nine characteristics of discontinuous agreement are shown to flow from a syntax that deals in whole phi structures and a linearisation algorithm that renders these discontinuous when conditions converge. This approach is demonstrably superior to reliance solely on syntax, morphology, or the lexicon, outperforming these across a range morphological and syntactic data from Afroasiatic languages and those of Australia and Papua New Guinea.
Abstract
Semitic agreement is normally discontinuous (i.e. expressed by more than one affix on the verb) only in the second and third persons. However, in restricted cases in particular languages, first person agreement is also discontinuous. I discuss two types of first person discontinuities. The first manifests the hallmarks of a meta split, persisting across paradigms and exponents. I argue that this type of first person discontinuity arises due to postsyntactic Fission which separates antagonistic sets of features prior to insertion and which is driven by markedness constraints on feature coexponence. The second type of first person discontinuity is restricted to a single paradigm and does not evince true discontinuous bleeding effects. Such discontinuities are best captured via morphological Doubling, modeled via Generalized Reduplication. First person discontinuities thus provide strong empirical support for the autonomy of morpheme splitting rules and morpheme copying rules. I demonstrate that each type of rule has a distinct empirical signature and acts as a repair to a different kind of morphotactic constraint. Consequently, there must be more than one route to discontinuous agreement.