In Czechia, teaching L2 English pronunciation is underestimated, which might lead to potential production slips and affect the intelligibility of the learners. This study investigates accuracy in L2 English pronunciation of checked vowels (/ʌ, ɛ, ɪ, ɒ, ʊ, æ/) as produced by Czech students. I applied cross-sectional approach and analysed participants from two distinct school classes who are divided by an 8-year study period. The pronunciation accuracy is analysed with the help of the computer program Praat, which generates frequencies for both Czech and English vowel phonemes produced by the participants reading out loud a set of monosyllabic citation words. The formant frequencies are later compared to a reference set of SSBE frequency values, which represents the targeted native language model. The results of this work show that the L2 English values are influenced by the students’ inventories of L1 Czech, leading to inaccuracy in production of the phonemes /ʌ, ɛ, ɪ, ɒ, ʊ/ and setting constraints to the L2 /æ/ acquisition.

Les programmes informatiques sont rarement écrits d'un seul coup, et sont au contraire composés de changements successifs. Il est également fréquent qu'un logiciel soit mis à jour après sa sortie initiale. De tels changements peuvent avoir lieu pour diverses raisons, comme l'ajout de fonctionnalités ou la correction de bugs. Il est en tout cas important d'être capable de représenter ces changements et de raisonner à leur propos pour s'assurer qu'ils implémentent les changements voulus.En pratique, les différences entre programmes sont très souvent représentées comme des différences textuelles sur le code source, listant les lignes de textes ajoutées, supprimées ou modifiées. Cette représentation, bien qu'exacte, ne dit rien de leurs conséquences sémantiques. Pour cette raison, il existe un besoin pour de meilleures représentations des différences sémantiques entre programmes.Notre première contribution est un algorithme de construction de programmes de corrélation, c'est-à-dire, des programmes entrelaçant les instructions de deux autres programmes de telle sorte qu'ils simulent leur sémantiques. Ces programmes de corrélation peuvent alors être analysés pour calculer une sur-approximation des différences sémantiques entre les deux programmes d'entrée. Ce travail est directement inspiré d'un article de Partush et Yahav, qui décrit un algorithme similaire, mais incorrect en présence de boucles avec des instructions `break` ou `continue`. Pour garantir la correction de notre algorithme, nous l'avons formalisé et prouvé à l'aide de l'assistant de preuve Coq.Notre seconde et plus importante contribution est un cadre formel permettant de décrire précisément et de formellement vérifier des différences sémantiques. Ce cadre, complètement formalisé en Coq, représente la différence entre deux programmes à l'aide d'un troisième programme que nous appelons oracle. Contrairement à un programme de corrélation, un oracle n'entrelace pas nécessairement les instructions des deux programmes à comparer, et peut « sauter » des calculs intermédiaires.Un tel oracle est généralement écrit dans un langage de programmation différent des programmes à comparer, ce qui permet de concevoir des langages d'oracles spécifiques à certaines classes de différences, capables de mettre en relation des programmes qui plantent avec des programmes qui s'exécutent correctement.Nous avons conçu de tels langages d'oracles pour couvrir un large éventail de différences sur un langage impératif jouet. Nous avons également prouvé que notre cadre est au moins aussi expressif que celui de la Logique Relationnelle de Hoare en encodant plusieurs variantes de cette dernière sous forme de langages d'oracles, prouvant leur correction dans la foulée. / Computer programs are rarely written in one fell swoop. Instead, they are written in a series of incremental changes.It is also frequent for software to get updated after its initial release. Such changes can occur for various reasons, such as adding features, fixing bugs, or improving performances for instance. It is therefore important to be able to represent and reason about those changes, making sure that they indeed implement the intended modifications.In practice, program differences are very commonly represented as textual differences between a pair of source files, listing text lines that have been deleted, inserted or modified. This representation, while exact, does not address the semantic implications of those textual changes. Therefore, there is a need for better representations of the semantics of program differences.Our first contribution is an algorithm for the construction of a correlating program, that is, a program interleaving the instructions of two input programs in such a way that it simulates theirsemantics. Further static analysis can be performed on such correlating programs to compute an over-approximation of the semantic differences between the two input programs. This work draws direct inspiration from an article by Partush and Yahav, that describes a correlating program construction algorithm which we show to be unsound on loops that include `break` or `continue`statements. To guarantee its soundness, our alternative algorithm is formalized and mechanically checked within the Coq proof assistant.Our second and most important contribution is a formal framework allowing to precisely describe and formally verify semantic changes.This framework, fully formalized in Coq, represents the difference between two programs by a third program called an oracle.Unlike a correlating program, such an oracle is not required to interleave instructions of the programs under comparison, and may “skip” intermediate computation steps. In fact, such an oracle is typically written in a different programming language than the programs it relates, which allows designing correlating oracle languages specific to certain classes of program differences, andcapable of relating crashing programs with non-crashing ones.We design such oracle languages to cover a wide range of program differences on a toy imperative language. We also prove that our framework is at least as expressive as Relational Hoare Logic by encoding several variants as correlating oracle languages, proving their soundness in the process.

本論文以優選理論探討上海話之入聲變調，試圖將文獻中所認為表現不規律之入聲變調納入與舒聲變調相容之分析中。本研究認為入聲變調與舒聲變調皆受制於ANCHOR-L(tσ1, Hd)，因此其首音節之基底聲調在輸出值中皆出現於重讀音節。以此觀之，則入聲變調與舒聲變調的差異主要在於節律重音的位置：在舒聲變調中首音節重讀；在入聲變調中重音則後移──在陰入變調中係移至第二音節，在陽入變調中則移至末音節。 本論文提出兩點假設：上海話中舒聲音節為重量音節，入聲音節為輕量音節；節律結構基本上為一位於左端之雙音節音步。據此，本研究提出COINCIDE (LightHd, FT-final)，認為陰入變調中重音所以後移至第二音節是因為重讀之輕量音節傾向於由音步末音節核可(licensing)。另一方面，由於陽入的單字調為曲拱調，與陰入的水平調有別，是以本研究另外提出COINCIDE(Contour, PrWd-final)，此制約顯示陽入變調中重音的遠距移位是為了遵行「曲拱調須由末音節核可」之普遍現象。 綜述之，本論文提供了一個重量音節、節律重音、曲拱調以及邊際位置等韻律顯著位置間彼此對映的實例。此外，本研究亦顯示上海話的連讀變調涉及聲調與重音的互動，是以所提出之分析或許對相關類型之研究亦有所貢獻。 / This thesis offers an Optimality-theory approach to the tone sandhi of checked tones (TSC) in Shanghai, in an attempt to regulate its surface patterns which have long been considered anomaly as opposed to the tone sandhi of smooth tones (TSS). With a reanalysis in the present study, TSC and TSS arguably have in common that their process of tone mapping is both subject to ANCHOR-L(tσ1, Hd), by which the un- derlying tone of the initial syllable ends up at the stressed syllable in the output. It follows that TSC is different from TSS in the way that metrical head is assigned: all domains undergoing TSS are stressed on their initial syllable; only in domains of TSC does the stress move rightwards, either to the second place in the tone sandhi of Yinru (TSYI), or to the final syllable in the tone sandhi of Yangru (TSYA). Given the assumption that checked syllables and smooth syllables in Shanghai are light and heavy, respectively, in terms of moraicity, and that foot-parsing is binary and left-aligned in general, the one-syllable shift of stress in TSYI can be accounted for by positing COINCIDE(LightHd, FT-final), which sets up the preference for light stress- bearing syllables to be licensed foot-finally. On the other hand, given that Yangru in the citation forms represents a rising contour, different from the level tone of Yinru, a licensing constraint, namely COINCIDE(Contour, PrWd-final), is further posited so that the long-distance movement of metrical head observed in TSYA emerges to satisfy the requirement for the retained rising contour to be licensed word-finally. Taken together, this thesis instantiates a remarkable case of the mapping among multiple prominent positions, including heavy syllables, metrical head, contour tones, and edge positions. Also, the present analysis demonstrates that Shanghai tone sandhi involves an interaction between tone and stress, thus a contribution to the general OT tone-prominence typology literature (cf. Zhang 2001, Barnes 2002, De Lacy 2002).