Cette thèse fournit une description du kakabé, une langue mandée parlée en Guinée, basée sur un corpus et avec un focus sur le système phonologique. Elle contient une brève esquisse grammaticale et deux parties qui portent sur l'analyse phonologique : la phonologie segmentale et la phonologie suprasegmentale. Les processus concernant les segments phonologiques peuvent être conditionnés par des contraintes métriques, par l'interdiction du hiatus, par le découpage de l’énoncé en phrases prosodiques et par le contexte morphologique. Le kakabé applique diverses stratégies d'adaptation des emprunts (principalement, du poular et du français), telles que l'épenthèse vocalique, la simplification d’agglomérations consonantiques. Le kakabé est une langue à ton (H vs. L), avec downdrift, relèvement du ton H, un ton flottant L, et un certain nombre de processus tonals, tels que l'insertion du ton H, la propagation du ton,l'aplatissement du contour HLH. En conséquence, la distance entre les tons lexicaux sous‐jacents et leur réalisation de surface peut être assez importante. Chacun des processus tonals est appliqué dans une unité prosodique particulière.Par conséquent, les processus tonals participent au découpage du discours en unités prosodiques. Le kakabé comporte des tons de frontière qui servent à signaler la force illocutoire de l'énoncé. Les tons lexicaux et les tons de frontièrecoexistent avec des opérations intonatives sur la courbe F0. Les appendices comprennent un dictionnaire kakabé-français, composé de 3400 entrées, et le corpus de 12 heures de textes en kakabé, transcrits, glosés, traduits etaccompagnés des fichiers vidéos et audios. / This thesis provides a corpus‐based description of Kakabe, a Mande language spoken in Guinea, with a focus on phonology. It consists of a short grammatical sketch and two parts dedicated to the analysis of the segmental and the suprasegmental phonology. Segmental phonological processes can be conditioned by metrical constraints, the ban on hiatus, prosodic phrasing and morphological context. Vowel deletion and vowel assimilation which serve to resolve hiatus, apply clause‐internally, as well as across clause boundaries. I also describe various strategies of loanword adaptation used in Kakabe, such as vowel epenthesis and consonant cluster simplification. Kakabe is a terraced‐level tone language (H vs. L), featuring downdrift, downstep, H raising, floating L, and a number of tonal processes, such as OCP style H‐insertion between two L domains, tone spread and leveling of HLH contour. As a result, the distance between the underlying lexical tones and their surface realization can be rather important. Each tonal process is applied within one particular prosodic unit. Therefore, tonal processes participate in phrasing the speech into prosodic units.Kakabe uses a number of boundary tones to signal illocutionary force of the utterance. Lexical tones and boundary tones coexists with intonational operations on the F0 curve. Intonational tone raising is associated with the H% and HL%boundary tones. Apart from that, it affects polarity items, the universal quantifier, and other pragmatically prominent lexemes, such as ideophones and intensifiers. The appendices include a Kakabe‐French dictionary, comprising 3400 entries, and an oral corpus of 12 hours of various genres, transcribed, glossed and time‐aligned with audio and video.

Cette thèse a pour objectif de décrire et d’analyser le rôle de la prosodie dans la structuration de l’énoncé en mooré (langue à ton parlée au Burkina Faso).Le problème majeur que je tente de résoudre dans ce travail est celui de la réalisation de l’intonation indicatrice de la structure prosodique dans les langues tonales. En d’autres termes, comment l’intonation se réalise-t-elle en mooré, une langue où le ton qui utilise les mêmes paramètres acoustiques doit garder son rôle phonologique ? Les explications détaillées du rôle et de la réalisation du ton et de l’intonation me permettent de déterminer une structure prosodique qui peut rendre compte de l’organisation du mooré parlé. D'un point de vue méthodologique, je m'inspire de la théorie de l'intonation de P. Martin, (2009, 2013, 2015), étant consciente qu’une certaine adaptation est nécessaire du fait de la nature de la langue étudiée.Cette thèse comporte deux parties : dans la première partie, je propose une définition des principaux termes de l’analyse et une présentation du mooré ; dans la deuxième, je m’appuie sur des études expérimentales pour analyser les principaux problèmes.Je m'intéresse, dans un premier temps à la réalisation phonétique du ton. Cette étude m'a permis de confirmer que la hauteur relative des tons est le paramètre le plus important pour distinguer les différents types de tons en mooré.Ensuite, l’examen de la relation entre ton et intonation révèle que la réalisation des tons lexicaux est affectée par celle de l’intonation surtout au niveau des frontières prosodiques. De plus dans l’abaissement tonal (downstep ou downdrift), la réalisation des tons est assujettie à la règle tonale, mais le domaine de l’abaissement est délimité par les frontières prosodiques, lieux de manifestation l’intonation.Enfin, l’étude sur la structuration des énoncés dans la parole lue et spontanée montre, à partir de l’identification d’évènements prosodiques, que les indices acoustiques tels que : la durée des pauses, l’allongement des syllabes finales et les variations de la fréquence fondamentale (F0) accompagnent les frontières perçues. Dans la parole lue, le resetting de F0 au début des unités et les variations de durée des syllabes de frontières prosodiques constituent les principaux indices de démarcation. Dans la parole spontanée, les résultats montrent qu’il y a une relation d’échange entre la durée des syllabes de frontières et celle des pauses et qu’elle s’associe aux variations de F0 pour permettre la structuration des énoncés en mooré / This dissertation describes the role of prosody in the organization of oral speech in Moore (a tone language spoken in Burkina Faso). It investigates the realization of intonation as a sign of prosodic structure in an African tone language. The main problem dealt in this study is: how does intonation work in moore, a tone language in which tone has already an important phonological role?The aim of the analysis is to explain the realization of tone and intonation in order to identify a prosodic structure which will account for the prosodic organization in moore. The theoretical framework is based on P. Martin’s theory of intonation in romance languages (2009, 2013, and 2015). Of course amendments were necessary to adapt it to moore.This dissertation is made up of two sections; the first one is dedicated to defining some important notions of the study and presenting general characteristics of moore. In the second section, experiments were carried out to address the main issues.The phonetic realization of tone was first examined. This study confirms that the relative height of the pitch is the main acoustic parameter used to distinguish the two types of tones in moore.Then the analysis of the relation between tone and intonation reveals that the realization of tones is affected by intonation especially at prosodic boundaries. When downstep is applied, tonal rules determine the realization of tones. However, the domain of the downstep is also marked out by prosodic boundaries (place of the realization of intonation).Finally, based on the identification of prosodic events, the study of moore oral speech (both spontaneous and reading speech) organization shows that acoustic parameters such as: duration of pauses, final syllable lengthening and F0 variations are found with perceived boundaries. The analysis of reading speech reveals that F0 resetting and variations of boundary syllables duration are the major indices which mark the limits of prosodic phrases. In spontaneous speech, the results suggest the existence of a trading relationship between pauses duration and boundary syllables duration. Therefore, the prosodic organization of utterances in moore derives from the combination of F0 variations and the trading relationship.

With the changing perceptions in coastal engineering in recent time, creation of bay beach for recreation by combining artificial headlands with nourishment has become one of the favorable options for mitigating erosion and shore protection in several foreign countries advanced in coastal engineering. To achieve this goal, hydraulic model tests and numerical simulations have been applied for the planning and design of stable bay beaches. While using the latter approach, numerical computation includes four major components, these being the waves, topography changes, flow field and sediment transport. This study utilizes the so-called Modified Hsu3D model developed by Serizawa et al. (1996) and Kumada et al. (2002) in Japan for the bathymetry within a static equilibrium bay beach defined by Hsu and Evans (1989). This model enables the direct calculation of bottom bathymetry within a static bay without using not only iterative numerical steps for wave transformation and current distribution, but also the continuity equation for total sediment transport. The results of this simplified approach can be used to estimate the distribution of erosion and accretion within a static bay, hence, suitable for pre-assessment of an artificial beach nourishment project. Prior to applying the Modified Hsu3D model to a bay beach undertaken in this study, sensitivity tests are performed on the setting of several key parameters associated with this model, such as limiting slope on land , limiting slope in the water , height of the berm , alignment angle at downdrift of the bay beach, and wave incident angle . The verification results are then adopted to compare with that reported in Serizawa et al. (2000), as well as to investigate the effects of each parameters on the accuracy of the modeling, in order to enhance the reliability of this model and the setting of the parameters. Finally, the Modified Hsu3D model is applied to simulate the changes in the shoreline and bathymetry for the Sizihwan Bay in Kaohsiung, under the action of normal incident waves during summer monsoon. This study also takes the advantage of the Modified Hsu3D model to explore the effect of oblique wave incidence on the deviation of downdrift control point of a static bay, in order to assist the prediction of downdrift control point for beach changes on an artificially nourished bay beach. The results reveal that the downdrift control point does not shift, under normal incident waves with different for the breaker. On the other hand, under oblique wave action within from the external boundary line of 20 m depth offshore of a bay beach comprising sediment 0.2~0.5 mm, the offset of point (i.e., and coordinates from the original origin) versus wave incident angle can be established. By establishing a regression equation between and versus , the result can be used to assist the determination, more precisely without guess work, on locating the downdrift control point of a bay beach, while working on the MEPBAY (Klein, et al., 2003) on computer screen to assess its stability.