Seaweeds are among the five marine sub-sectors of species cultured in South Africa, with Ulva species cultured at a commercial scale. In South Africa, the annual production of Ulva is approximately 2000 tonnes (wet weight), with the majority of Ulva being grown in landbased paddle raceway systems receiving abalone effluent water. Cultured Ulva is mainly used as abalone feed and for bioremediation of farm effluent water. It is not sold but rather is used either as fresh feed or dried and incorporated into formulated feeds. The main commercial abalone farms growing Ulva in paddle raceway systems in South Africa are Irvin & Johnson (I&J) Cape Abalone, Abagold, Buffeljags Abalone and Diamond Coast Aquaculture in the Western Cape province, and Wild Coast Abalone in the Eastern Cape province. The main aim of this study is to precisely identify the cultured Ulva species as their identity is not clearly understood. This will provide information on the genetic diversity in the cultured material and could enable farmers to select for a species/strain that has the desired traits such as high nutritional value, rapid growth rate, resistance to diseases, and the ability to grow vegetatively, amongst others, so that the best feed is given to abalone. Next, the cultured Ulva species will be compared with closely related seashore Ulva species and with Ulva specimens from the main farming area in Hermanus to see if they are genetically similar. Ulva specimens from these farms, nearby seashores, including the Hermanus abalone farming complex in the New Harbour were identified using morpho-anatomical and molecular methods. The molecular markers employed in this study were the plastid large subunit of Ribulose-1,5-bisphosphate carboxylase/oxygenase (rbcL), Internal Transcribed Spacer of nuclear ribosomal DNA (ITS nrDNA) and the elongation factor tufA. The 12 cultured Ulva specimens belonged in the U. lacinulata clade with weak support value of 0.57 for PP in the rbcL tree, high support value of 0.86 for PP in the ITS tree and high support value of 92% and 0.92 for BP and PP, respectively, in the tufA tree. The seashore U. capensis and farmed Ulva specimens belonged in the same large U. lacinulata clade in the rbcL tree. However, the U. capensis samples and the locally cultivated Ulva samples belonged in separate sister clades with a support value of 70% and 0.75 for BP and PP, respectively, in the ITS phylogenetic tree, and 97% and 1 for BP and PP, respectively, in the tufA phylogenetic tree. Therefore, the identity of the cultivated Ulva samples is U. lacinulata and the clade containing the U. capensis samples has now been labelled as U. uncialis as it is an older available name than U. capensis. Furthermore, the foliose U. lacinulata was also found growing attached near the inlets of the Hermanus abalone farming complex in New Harbour. There was no genetic variation within the farmed Ulva samples as they were collapsed as a single haplotype by the three molecular markers. The genetic distance between the U. uncialis and farmed U. lacinulata samples were 0.16%, 0.76% and 0.92% for the markers, rbcL, ITS and tufA, respectively. Even though the low sequence divergence between the farmed U. lacinulata and U. uncialis specimens fits within the range of variability, these two clades are separate species that are closely related. Incongruences between the molecular and morpho-anatomical identification methods were observed, as the morpho-anatomical identification method identified 9 of the 12 farmed Ulva specimens as U. lactuca and the remaining as U. rigida sensu Stegenga et al. (1997). Fewer Ulva species were resolved morphologically because of the overlap in morphological description within U. lactuca sensu Stegenga et al. (1997) and U. rigida sensu Stegenga et al. (1997). Additionally, three new records of Ulva species (U. ohnoi, U. australis and U. stenophylloides) for South African seashore specimens were molecularly identified in this study, and foliose U. compressa was recorded for the first time in the region. In this study, the molecular marker tufA, was the best marker to delimit species, as its internal clades were better supported compared to the other two markers and it was able to better separate the farmed U. lacinulata samples and the seashore U. uncialis samples into two different clades.

The occurrence of phenotypic and genotypic diversity is the key factor in crop improvement including abiotic stress tolerance. The focal objectives of this study were to evaluate and characterize 74 tropical indica rice breeding lines for phenotypic and genotypic diversity, screening for the most devastating abiotic stresses in rice; drought and salinity at the seedling stage at morpho-physiological and molecular levels. To fulfill these objectives, five studies were conducted in pots; first two experiments aimed at assessing phenotypic and yield variability at seedling and maturity stages respectively; based on several (more than 20) root and shoot traits which exploited a wide range of variability among genotypes for measured traits. Germplasm was then screened for drought stress at two moister regimes, 50%, and 100% moisture levels, under mini-hoop structures. Nine percent of the genotypes exhibited a high tolerance to drought stress, and genotypes IR86638 and IR49830 were identified as the most and least drought tolerant respectively. Germplasm was also screened for salinity tolerance in pure sand pot-culture (a simple, efficient and alternate screening method) at three levels; high salt stress (EC 12 dSm-1), moderate salt stress (EC 6 dSm-1), and control imposed one week after emergence. Thirteen genotypes (17.57%) were identified as highly salt tolerant; genotypes FED 473 and IR85427 were highly salt tolerant and salt sensitive, respectively. Root traits were found more crucial and best descriptors in identifying both salinity and drought tolerant genotypes. Genotypes were further used in Genome-wide Association Study (GWAS) to uncover important SNPs, QTLs or genes related to salinity tolerance. A higher number of significant SNPs were discovered for root traits, indicting the importance of root traits in identifying abiotic stress tolerance in rice. The knowledge gained from this investigation could be useful in breeding for better crop establishment, yield improvement, screening for any abiotic stress tolerance.

Au sein de la famille des Cyprinidés (Téleostéens), Parachondrostoma toxostoma (le toxostome) et Chondrostoma nasus (le hotu) sont deux espèces (respectivement endémique et invasive) qui se rencontrent dans le sud de la France, formant deux zones hybrides distinctes : la zone de la Durance (un milieu fortement fragmenté) et la zone de l'Ardèche (un milieu non fragmenté). La présence de ces deux zones hybrides nous a donné l'opportunité de caractériser les parts respectives de la sélection exogène (l'environnement) et endogène (compatibilité génomique) permettant d'expliquer les patterns d'hybridation entre les deux espèces. Les travaux présentés dans le cadre de cette thèse illustrent parfaitement la complexité des phénomènes d'hybridation, chaque situation étant fortement dépendante du contexte d'étude et ce à l'échelle même de la station. Nous avons montré dans certaines stations que l'espèce endémique résiste à l'introgression de son génome par l'espèce invasive, dans d'autres cas nous avons des scénarios plus complexe d'admixture qui évoluent au cours du temps. Le potentiel évolutif engendré par les phénomènes d'hybridation est cependant indéniable et nous préconisons de prendre en compte ces processus d'hybridation dans les programmes de gestions et de conservation de la biodiversité. / In the Cyprinidae family (Teleostei), Parachondrostoma toxostoma (the sofie) and Chondrostoma nasus (the nase) are respectively endemic and invasive species which are found in sympatry in the south of France. They form two distinct hybrid zones: the Durance River (a highly fragmented environment) and the Ardèche basin (an unfragmented area). The existence of these two different zones allow us to characterize the respective contributions of exogenous selection (environmental factors) and endogenous selection (genomic compatibility) to explain hybridization patterns between the two species.This PhD thesis highlights the complexity of hybridization phenomena. Each situation is highly dependent of the study context. We showed the resistance of the genome of the endemic species to introgression by the genome of the invasive species in some stations. In other cases, we demonstrated more complex scenarios of admixture that evolve over time. The evolutionary potential generated by hybridization is undeniable, and we recommend to take the hybridization process into account in management programs and conservation of biodiversity.

Le Rhin dans son parcours franco-allemand a été aménagé depuis deux siècles pour la protection contre les inondations, la navigation et la production hydro-électrique. La rectification (19ème siècle), la régularisation (1930) et la construction du Grand Canal d’Alsace (1928-1959) ont profondément altéré le fonctionnement hydro-sédimentaire du « Vieux Rhin », tronçon de 50 km court-circuité entre les barrages de Kembs et de Breisach, conduisant à une réduction de la complexité du tracé en plan, une incision, un pavage du fond du lit et une simplification des habitats aquatiques et riverains. La possibilité de restaurer le transport sédimentaire et la dynamique alluviale du Vieux Rhin est évaluée au sein de projets portés par la Région Alsace et EDF. Dans ce cadre, la présente thèse repose sur deux types d’approches de la géomorphologie fluviale, historique et expérimentale, afin de répondre à la problématique de restauration morpho-écologique du tronçon fluvial. Ce travail, basé sur la collecte de données anciennes et de terrain, ainsi que sur la quantification des évolutions morphologiques à différents niveaux scalaires, a conduit à une meilleure compréhension de la trajectoire temporelle d’ajustement de l’hydrosystème et à une estimation de sa sensibilité aux changements. Le suivi géomorphologique d’un test de recharge sédimentaire conduit sur un site-pilote a servi également à évaluer la pertinence de ce type d’intervention, à travers l’analyse des bénéfices et des risques environnementaux potentiels. L'ensemble des résultats obtenus a permis d’apporter aux gestionnaires des éléments d'aide à la réflexion pour orienter les stratégies de restauration ainsi que des indicateurs physiques d’évaluation des futures actions / The Upper Rhine River has undergone several anthropogenic modifications since the last two centuries for flood protection, navigation and hydropower generation. Channel rectification, groyne fields and lateral “Grand Canal of Alsace” construction heavily altered the hydro-sedimentary functioning of the 50 km-long "Old Rhine" by-passed between Kembs and Breisach: channel pattern simplification, bed incision, armouring and ecological habitats alterations. Two complementary projects have been initiated to restore the sediment transport and the reach alluvial dynamics. This thesis investigated space-time and experimental geomorphic approaches within the morpho-ecological restoration of a modified fluvial system. The archive and field data-based work aimed at understanding the Old Rhine historical trajectory. The geomorphological monitoring of a gravel input experiment also consisted in evaluating environmental benefits and potential risks for society and proposing physical assessment indicators. Results will help river stakeholders to define strategies to restore morphological processes and associated ecological functions in the Old Rhine.

Cette thèse porte sur la tectonique active et l’analyse de deux nouvelles failles : la faille de Sharkhai et la faille d’Avdar. Les deux failles ont été découvertes, en 2011, à partir des images satellitaires HR, à quelques dizaines de kilomètres de la capitale de la Mongolie, Oulan Bator. Une approche morpho-tectonique, géomorphologique et paléosismologique, à partir des images de haute résolution Pléiades et d’études de terrain, notamment de tranchées, a permis: 1) de cartographier les deux failles en détail; 2) de décrire leur géométrie et segmentation; 3) de contraindre leur cinématique; 4) de documenter leur activité récente et leur comportement sismique (décalages co-sismiques et cumulés, période de retour, taux de glissement). La faille de Sharkhai s’étend sur 46 km selon une direction comprise entre N42°E et N72°E en moyenne, avec un pendage de 66° à 72° vers le sud-est entre E106.31°/N47.352° et E106.53°/N47.485°, et sub-vertical entre E106.474°/N47.473° et E106.75°/N47.57°. Elle est caractérisée par une géométrie rectiligne et simple en surface. C’est une faille senestre avec une composante verticale normale. L’étude paléosismique met en évidence trois séismes : le séisme EQ3 s’est produit avant 3850±120 calBP, le séisme EQ2 entre 2400±70 calBP et 2030±40 calBP, et le séisme le plus récent (MRE) entre 1090±84 calBP et avant l’âge du sol récent. La période de retour minimum des forts séismes sur la faille de Sharkhai est 1195±157 ans, ce qui implique une vitesse de glissement maximum comprise entre 0.6±0.2 et 2.14±0.5 mm/an. Plusieurs scénarios de segmentation de la faille sont proposés indiquant que la faille est capable de produire des séismes de magnitude comprise entre 6 et 7. L’accélération maximale du sol (PGA au rocher) générée à Oulan Bator serait de 0.12 g, au nouvel aéroport de 0.28 g, et à la ville de Zuunmod de 0.17 g. Ces valeurs correspondent à une intensité (MMI) comprise entre VI et X. La faille d’Avdar s’étend sur 47 km selon une direction moyenne comprise entre N26°E et N88°E avec un pendage variant de 40° à 55°. La partie sud-ouest de la faille est caractérisé par une géométrie simple et linéaire. Par contre, l’architecture de la faille est complexe dans la partie nord-est. Les décalages mesurés indiquent que la faille est une faille senestre avec une composante normale. Les tranchées paléosismiques documentent son activité sismique pendant le quaternaire. La faille peut se diviser en plusieurs segments, suggérant des séismes de magnitude comprise entre 5.8 et 7. Les valeurs de PGA (au rocher) les plus importants seraient de 0.1 g pour Oulan Bator, 0.18 g pour le nouvel aéroport et 0.19 g pour la ville de Zuunmod. Ces valeurs indiquent une intensité de l’ordre de VI à X. Les investigations paléosismiques montrent que le dernier séisme est plus récent que 5665±85 calBP. Tous ces résultats doivent être pris en compte dans l’estimation de l’aléa sismique de la région de la capitale Oulan Bator, qui a elle seule comprend plus de la moitié de la population du pays. Située sur un bassin sédimentaire, elle est le centre commercial et industriel du pays et contient divers bâtiments vulnérables. Enfin, elle est en pleine croissance avec un nouvel aéroport en construction à proximité de la faille de Sharkhai, secteur où la ville va très rapidement s’étendre. / This thesis focuses on the active tectonic of two new faults: the Sharkhai fault and the Avdar faults. Both faults were discovered in 2011 from HR satellite images at tens kilometers from the capital of Mongolia, Ulaanbaatar.Morpho-tectonic, geomorphological and paleoseismological approaches using high resolution Pleiades satellites images and field investigations allowed: 1) to map the both faults in details; 2) to describe their geometry and segmentation; 3) to identify their kinematic; 4) to document their recent activity and their seismic behavior (co-seismic and cumulated displacements, the time of the last earthquakes, slip rate). The Sharkhai fault extends 46 km with an average direction between N42°E and N72°E, and a dip between 72° and 66° either to south-east, between E106.31°/N47.352° and E106.53°/N47.485°, or sub-vertical between E106.474°/N47.473° and E106.75°/N47.57°. The fault is characterized by a linear and simple geometry. It is a left lateral strike slip with a normal vertical component. The paleoseismological investigations show evidences for three earthquakes: the EQ3 earthquake occurred before 3850±120 calBP, the EQ2 earthquake between 2400±70 calBP and 2030±40 calBP, and the most recent earthquake (MRE) between 1090±84 calBP and before the age of recent soil. The minimum return period of strong earthquakes on the fault Sharkhai is 1195 ± 157 years, which implies a maximum slip rate between 0.6±0.2 and 2.14±0.5 mm/year. Several segmentation scenarios of the fault where proposed indicating that the fault is capable to produce an earthquake of magnitude between 6 and 7. The maximum peak ground acceleration (PGA at rock) generated at Ulaanbaatar is 0.12 g, 0.28 g at the new airport, and 0.17 g at the city of Zuunmod. These values correspond to intensity between VI and X at rock sites. The Avdar fault, 47 km length, have an average orientation between N26°E and N88°E and a dip between 40° and 55°. The southwestern part of the fault is characterized by a simple and linear geometry while the architecture of the surface rupture is complex in the northeastern part. The measured offsets indicate that the fault is left lateral with a normal component. The paleoseismic trenches attest its seismic activity during the Quaternary. The fault can be divided into several segments suggesting earthquakes of a magnitude ranging from 5.8 to 7. The most important values of PGA (at rock site) would be 0.1 g at Ulaanbaatar, 0.18 g for the new airport and 0.19 g for Zuunmod city. These values indicate an intensity between VI and X. Paleoseismic trench shows that the last earthquake occurred on the fault since 5665± 85calBP. All these results should be considered in estimating the seismic hazard in the region of the capital Ulaanbaatar especially as, alone, it contains more than the half of the country population. Situated in a sedimentary basin, it is the commercial and industrial center of Mongolia and contains many vulnerable buildings. Also, the city is growing with a new airport construction near the fault Sharkhai area where the city will extend very quickly.

Cette thèse porte sur la description morpho-syntaxique du nââ numèè, langue de l’extrême-Sud de la Nouvelle-Calédonie, parlée par moins de 1 000 locuteurs dans les tribus de Touaourou et de Goro dans la commune de Yaté et de l’île Ouen dans la commune du Mont-Dore.Dans la première partie, l’auteur aborde les faits sociaux qui ont fait de la région du Grand Sud l’une des plus impactés par les guerres traditionnelles, la colonisation, l’évangélisation, la construction du barrage de Yaté et plus récemment, l’exploitation minière.Dans la deuxième partie, l’auteur procède à l’identification des phonèmes de la langue. Ensuite, il effectue une comparaison entre le système consonantique, vocalique et tonal ainsi établi et celui réalisé par Rivierre dans les années 70. Cette analyse d’écart lui permet de mettre en exergue les évolutions phonologiques subies par la langue en presque quarante ans. Pour définir les différents morphèmes de la langue, l’auteur utilise la méthode traditionnelle dite « distributionnelle ». Il propose ensuite une catégorisation des lexèmes.Dans la troisième partie, l’auteur traite des procédés morphologiques utilisaient en dérivation et en composition.Dans la quatrième partie consacrée à la syntaxe, l’auteur expose le syntagme nominal avec les différents modes et types de détermination. Cette partie comprend également un exposé des substituts du nom, des déictiques et du système numéral. Puis, l’auteur propose une description du syntagme verbal avec en point d’orgue, l’analyse des marqueurs aspecto-temporels et les constructions sérielles. Dans la partie dédiée aux circonstants, l’auteur s’intéresse au système d’orientation développé et les différents cadres référentiels utilisaient (absolus et relatifs).Dans la dernière partie, l’auteur expose les différents types d’énoncés (simples, complexes et marqués). / This thesis deals with the morpho-syntactic description of Nââ Numèè, a language of the far south of New Caledonia, spoken by fewer than 1,000 speakers in the tribes of Touaourou and Goro in the town of Yate and the isle Ouen in the town of Mont-Dore.In the first part, the author explains the social facts which led to the great southern region being one of the most affected by traditional wars, colonization, evangelization, the construction of the dam at Yate and more recently, mining.In the second part, the author will identify the phonemes of the language. Then a comparison will be made between the consonant, vowel and tonal system thus established and made by Rivierre in the 1970s. This gap analysis allows the author to highlight the phonological changes in the language over almost forty years. The author then uses the traditional "distributional" method to define the different morphemes of language before proposing a categorization of lexemes.In the third part, the author deals with morphological processes used in derivation and in composition. In this chapter, it is interesting to see the increased use of verbal classifiers.In the fourth section dealing with syntax, the author offers a description of the nominal phrase with the modes and types of determination. This part also includes a presentation of the nominal substitutes, deictic and numeral systems. Then, the author describes the verbal phrase with special focus on the analysis of the aspect and temporal markers and serial verbs. In the part dedicated to the circonstants, the author focuses on the orientation system developed and the various repositories spaces used (absolute and relative).In the last part, the author exposes the different types of clauses (simple, complex and marked).

<p>Predmet istrživanja ove doktorske disertacije predstavljaju morfo-funkcionalni i motorički status fudbalera, odnosno specijalno programirani sistem treninga u letnjem pripremnom periodu kod fudbalera seniorskog uzrasta različitih rangova takmičenja. U okviru rada na disertaciji sprovedeno je inicijalno i finalno merenje morfo-funkcionalnih i motoričkih sposobnosti fudbalera kao i efikasnost efekata tretmana. Imajući u vidu da se na fudbalskim utakmicama ostvaruju mnogobrojni zadaci, a koji se odnose na različite segmente psihosomatskog statusa, u centar ovog istraživanja postavljena je problematika koja je vezana za iznalaženje puteva za efikasniju realizaciju fudbalske igre.<br />U metodologiji fudbalskog treninga postoji veliki broj vežbi. Važno je odabrati najpotrebnije od njih. Još je važnije da se te vežbe međusobno kombinuju i povežu na takav naćin koji će omogućiti da se kompleksno rešavaju potrebni zadaci, odnosno da uče igrače komponentama na kojima se temelji fudbal. Ona adaptacija koja se stvara kod fudbalera u pripremnom periodu pomoću određenog modeliranja tokom treninga, treba da se prenese na utakmicu i da se u njoj ispolji. Dakle treba težiti da fizički trening koji se primenjuje u pripremnom periodu daje momentalne i trajnije efekte, a strateški cilj celokupnog trenažnog procesa je uspešnost u takmičenju. Neposredni značaj ovog istraživanja ogleda se u mogućnosti stručne, teorijske i praktične nadgradnje fudbalskih trenera u pogledu modifikacije treninga. Dalje značaj rada ogleda se u traganju za novim sredstvima i metodama kako bi inovirali trenažni proces, u cilju ostvarenja što boljih vrhunskih rezultata.</p> / <p>The topic of this PhD thesis are morpho -functional and motoric status of football players specially programmed system of training in summer preparation period with professional football players in different levels of competition. Initial and final measuring of morpho-functional and motoric capabilities of football players are given as a part of this thesis, also as the efficiency of treatment effects. Considering the fact that many tasks are completed during football matches and many of them are related to different segments of psycho somatic status, the care of this research are problems connected to finding solutions to more efficient football play. There are many exercises in methodology of football training. It is very It important to choose those that are the most needed. The more important is to combine and connect those exercises in the way that will ensure complex solution of needed tasks and therefore to teach players all components of<br />football. The adaptation achieved with football players in training period using specific modeling should be transferred to the match and shown during it. The aim of physical training in preparation period is to give longer effects. The strategic goal of training process is success in competition. Immediate significance of this research is in it&acute;s ability of professional, theoretical and practical upgrade of football couches concerning modification of training process. Further significance of the research is in the search for mew methods and assets for innovation of training process in order to achieve best results.</p>

Cette thèse porte sur l'intégration de ressources lexicales et syntaxiques du français dans deux tâches fondamentales du Traitement Automatique des Langues [TAL] que sont l'étiquetage morpho-syntaxique probabiliste et l'analyse syntaxique probabiliste. Dans ce mémoire, nous utilisons des données lexicales et syntaxiques créées par des processus automatiques ou par des linguistes afin de donner une réponse à deux problématiques que nous décrivons succinctement ci-dessous : la dispersion des données et la segmentation automatique des textes. Grâce à des algorithmes d'analyse syntaxique de plus en plus évolués, les performances actuelles des analyseurs sont de plus en plus élevées, et ce pour de nombreuses langues dont le français. Cependant, il existe plusieurs problèmes inhérents aux formalismes mathématiques permettant de modéliser statistiquement cette tâche (grammaire, modèles discriminants,...). La dispersion des données est l'un de ces problèmes, et est causée principalement par la faible taille des corpus annotés disponibles pour la langue. La dispersion représente la difficulté d'estimer la probabilité de phénomènes syntaxiques apparaissant dans les textes à analyser mais qui sont rares ou absents du corpus ayant servi à l'apprentissage des analyseurs. De plus, il est prouvé que la dispersion est en partie un problème lexical, car plus la flexion d'une langue est importante, moins les phénomènes lexicaux sont représentés dans les corpus annotés. Notre première problématique repose donc sur l'atténuation de l'effet négatif de la dispersion lexicale des données sur les performances des analyseurs. Dans cette optique, nous nous sommes intéressé à une méthode appelée regroupement lexical, et qui consiste à regrouper les mots du corpus et des textes en classes. Ces classes réduisent le nombre de mots inconnus et donc le nombre de phénomènes syntaxiques rares ou inconnus, liés au lexique, des textes à analyser. Notre objectif est donc de proposer des regroupements lexicaux à partir d'informations tirées des lexiques syntaxiques du français, et d'observer leur impact sur les performances d'analyseurs syntaxiques. Par ailleurs, la plupart des évaluations concernant l'étiquetage morpho-syntaxique probabiliste et l'analyse syntaxique probabiliste ont été réalisées avec une segmentation parfaite du texte, car identique à celle du corpus évalué. Or, dans les cas réels d'application, la segmentation d'un texte est très rarement disponible et les segmenteurs automatiques actuels sont loin de proposer une segmentation de bonne qualité, et ce, à cause de la présence de nombreuses unités multi-mots (mots composés, entités nommées,...). Dans ce mémoire, nous nous focalisons sur les unités multi-mots dites continues qui forment des unités lexicales auxquelles on peut associer une étiquette morpho-syntaxique, et que nous appelons mots composés. Par exemple, cordon bleu est un nom composé, et tout à fait un adverbe composé. Nous pouvons assimiler la tâche de repérage des mots composés à celle de la segmentation du texte. Notre deuxième problématique portera donc sur la segmentation automatique des textes français et son impact sur les performances des processus automatiques. Pour ce faire, nous nous sommes penché sur une approche consistant à coupler, dans un même modèle probabiliste, la reconnaissance des mots composés et une autre tâche automatique. Dans notre cas, il peut s'agir de l'analyse syntaxique ou de l'étiquetage morpho-syntaxique. La reconnaissance des mots composés est donc réalisée au sein du processus probabiliste et non plus dans une phase préalable. Notre objectif est donc de proposer des stratégies innovantes permettant d'intégrer des ressources de mots composés dans deux processus probabilistes combinant l'étiquetage ou l'analyse à la segmentation du texte / This thesis focuses on the integration of lexical and syntactic resources of French in two fundamental tasks of Natural Language Processing [NLP], that are probabilistic part-of-speech tagging and probabilistic parsing. In the case of French, there are a lot of lexical and syntactic data created by automatic processes or by linguists. In addition, a number of experiments have shown interest to use such resources in processes such as tagging or parsing, since they can significantly improve system performances. In this paper, we use these resources to give an answer to two problems that we describe briefly below : data sparseness and automatic segmentation of texts. Through more and more sophisticated parsing algorithms, parsing accuracy is becoming higher for many languages including French. However, there are several problems inherent in mathematical formalisms that statistically model the task (grammar, discriminant models,...). Data sparseness is one of those problems, and is mainly caused by the small size of annotated corpora available for the language. Data sparseness is the difficulty of estimating the probability of syntactic phenomena, appearing in the texts to be analyzed, that are rare or absent from the corpus used for learning parsers. Moreover, it is proved that spars ness is partly a lexical problem, because the richer the morphology of a language is, the sparser the lexicons built from a Treebank will be for that language. Our first problem is therefore based on mitigating the negative impact of lexical data sparseness on parsing performance. To this end, we were interested in a method called word clustering that consists in grouping words of corpus and texts into clusters. These clusters reduce the number of unknown words, and therefore the number of rare or unknown syntactic phenomena, related to the lexicon, in texts to be analyzed. Our goal is to propose word clustering methods based on syntactic information from French lexicons, and observe their impact on parsers accuracy. Furthermore, most evaluations about probabilistic tagging and parsing were performed with a perfect segmentation of the text, as identical to the evaluated corpus. But in real cases of application, the segmentation of a text is rarely available and automatic segmentation tools fall short of proposing a high quality segmentation, because of the presence of many multi-word units (compound words, named entities,...). In this paper, we focus on continuous multi-word units, called compound words, that form lexical units which we can associate a part-of-speech tag. We may see the task of searching compound words as text segmentation. Our second issue will therefore focus on automatic segmentation of French texts and its impact on the performance of automatic processes. In order to do this, we focused on an approach of coupling, in a unique probabilistic model, the recognition of compound words and another task. In our case, it may be parsing or tagging. Recognition of compound words is performed within the probabilistic process rather than in a preliminary phase. Our goal is to propose innovative strategies for integrating resources of compound words in both processes combining probabilistic tagging, or parsing, and text segmentation

Este trabalho propõe a análise da morfologia verbal do lembaama, que é uma língua do subgrupo banto (da floresta), B.62 (Guthrie, 1971), do grupo Benuê-Congo, do tronco Nigero-congolês. Como esta língua não apresenta nenhum estudo deste gênero, espera-se que esta primeira análise possibilite estudos posteriores neste e em outros campos linguísticos. A análise morfológica dos verbos revelou, primeiramente, que a estrutura verbal dessa língua é composta dos seguintes elementos: sujeito, índice do sujeito, negação1, marca de Tempo, raiz do verbo, extensão, vogal final ou marca de Aspecto, objeto(s) ou índice do objeto, negação2 que se seguem numa ordem fixa na oração. Constatamos que o índice do objeto, que é geralmente anteposto à raiz do verbo na maioria das línguas bantas, é posposto ao verbo em lembaama. A negação, por sua vez, é representada pelo morfema descontínuo composto por dois elementos: ka- (negação1) e -ní (negação2). Sendo que na estrutura verbal, ka- aparece depois do índice do sujeito (à esquerda da raiz do verbo) e -ni aparece depois do índice do objeto (à direita da raiz do verbo), é o último elemento da estrutura do verbo. Ao estudar a derivação verbal em lembaama, constatamos que a estrutura das extensões desta língua difere daquela encontrada nas outras línguas bantas pelo fato do lembaama ter acrescentado à estrutura -VC- preexistente no protobanto, estruturas do tipo -C-, -CV- e -CVC-. A extensão mais comum e mais fácil de ser reconhecida nessa língua é aquela que marca o habitual, tanto no presente (-ag-) como no passado (-ig-). O estudo dos sufixos derivativos do lembaama revelou também a existência de uma correlação entre o valor gramatical e o valor semântico das extensões. Vimos que seres humanos e animados são envolvidos em orações que apresentam uma transitividade alta. O estudo da transitividade mostrou-se, ainda, como uma noção fundamental nessa língua porque distingue ações e estados, por exemplo. Quanto à análise das categorias Tempo, Aspecto e Modo (TAM), essa língua evidenciou três tempos: um presente que não tem marcas e que se confunde com o presente pontual, o progressivo e o futuro. Dois futuros: um mais próximo, sem marcas, que se confunde com o presente pontual e o progressivo; outro, mais distante, marcado pelo verbo auxiliar odze \"ir\". Três passados: um recente, marcado pelo morfema mí- de tom alto e anteposto ao radical do verbo; um distante, marcado pelo morfema máá- de tom alto e anteposto ao radical do verbo; um remoto, expresso pela junção do verbo auxiliar -ki \"estar\" e o morfema do passado mí-. Assim, contrariamente aos morfemas de Aspectos, os morfemas de Tempo são sempre antepostos ao radical do verbo. É o caso dos dois morfemas que marcam os passados recente e distante. Há dois aspectos: o perfectivo, marcado pela vogal final -í; o imperfectivo, sem marca específica. A análise dos tempos e dos aspectos revelou que o Aspecto é mais fundamental em lembaama. O estudo dos modos permitiu identificar três: o imperativo, o condicional e o indicativo. / This work proposes an analysis of the verbal morphology of Lembaama (B62) according to Guthrie (1971). Officially called Obamba in Gabon, Lembaama is a Bantu language of the forest, from the Benue-Congo group and Niger-Congo phylum. As far as we know, this language has yet to receive a detailed study. It should be noted that Lembaama shows some interesting features. Indeed, the verbal morphology analysis shows that a single inflected verb contains the following elements: subject, subject marker, negative1, Tense marker, root, extension, Final Vowel or Aspect marker, negative2 occurring in a fix order in a sentence. The object marker comes after the root. Negation consists of a discontinuous morpheme: kaní. Ka- is placed before the root (by the left) and -ní occurs after the root (by the right) being the last element of this structure. The verbal derivation study reveals the following structure of Lembaama extensions: -C-, -CV-, -VC- and -CVC-.The habitual marker -ag- being the commonest extension. This study also highlights the existence of a correlation between the grammatical value and the semantic value of extensions. Hence, human and animate beings are evoked in clauses with higher transitivity than things. Therefore, transitivity is fundamental in Lembaama, as it can distinguish actions from states, for example. Tense, Aspect and Mood study defines three Tenses. First, there is a present that, because it is used without tense marker, can be merged with near future or with progressive. Then, we note two future tenses: a near future (F1) occurring without mark and a distant future (F2) marked by the auxiliary verb odze go. Finally, we count three past tenses: a recent (P1) marked by mí- with a high tone; a distant (P2) marked by máá- with a high tone too, and a remote past (P3) marked by kí be, an auxiliary verb with high tone and mí-, the recent past marker, both coming before the root (by the left). Thus, Tense marker morphemes are always placed before the root (by the left) and Aspect markers occur after the root (by the right). This analysis highlighted two Aspects: a perfective marked by the Final Vowel -í; an imperfective without a specific mark. Tense and Aspect analysis allows to conclude that Aspect is more fundamental than Tense in this language. Concerning Mood, the analysis revealed three: imperative, conditional and indicative.

Este trabalho propõe a análise da morfologia verbal do lembaama, que é uma língua do subgrupo banto (da floresta), B.62 (Guthrie, 1971), do grupo Benuê-Congo, do tronco Nigero-congolês. Como esta língua não apresenta nenhum estudo deste gênero, espera-se que esta primeira análise possibilite estudos posteriores neste e em outros campos linguísticos. A análise morfológica dos verbos revelou, primeiramente, que a estrutura verbal dessa língua é composta dos seguintes elementos: sujeito, índice do sujeito, negação1, marca de Tempo, raiz do verbo, extensão, vogal final ou marca de Aspecto, objeto(s) ou índice do objeto, negação2 que se seguem numa ordem fixa na oração. Constatamos que o índice do objeto, que é geralmente anteposto à raiz do verbo na maioria das línguas bantas, é posposto ao verbo em lembaama. A negação, por sua vez, é representada pelo morfema descontínuo composto por dois elementos: ka- (negação1) e -ní (negação2). Sendo que na estrutura verbal, ka- aparece depois do índice do sujeito (à esquerda da raiz do verbo) e -ni aparece depois do índice do objeto (à direita da raiz do verbo), é o último elemento da estrutura do verbo. Ao estudar a derivação verbal em lembaama, constatamos que a estrutura das extensões desta língua difere daquela encontrada nas outras línguas bantas pelo fato do lembaama ter acrescentado à estrutura -VC- preexistente no protobanto, estruturas do tipo -C-, -CV- e -CVC-. A extensão mais comum e mais fácil de ser reconhecida nessa língua é aquela que marca o habitual, tanto no presente (-ag-) como no passado (-ig-). O estudo dos sufixos derivativos do lembaama revelou também a existência de uma correlação entre o valor gramatical e o valor semântico das extensões. Vimos que seres humanos e animados são envolvidos em orações que apresentam uma transitividade alta. O estudo da transitividade mostrou-se, ainda, como uma noção fundamental nessa língua porque distingue ações e estados, por exemplo. Quanto à análise das categorias Tempo, Aspecto e Modo (TAM), essa língua evidenciou três tempos: um presente que não tem marcas e que se confunde com o presente pontual, o progressivo e o futuro. Dois futuros: um mais próximo, sem marcas, que se confunde com o presente pontual e o progressivo; outro, mais distante, marcado pelo verbo auxiliar odze \"ir\". Três passados: um recente, marcado pelo morfema mí- de tom alto e anteposto ao radical do verbo; um distante, marcado pelo morfema máá- de tom alto e anteposto ao radical do verbo; um remoto, expresso pela junção do verbo auxiliar -ki \"estar\" e o morfema do passado mí-. Assim, contrariamente aos morfemas de Aspectos, os morfemas de Tempo são sempre antepostos ao radical do verbo. É o caso dos dois morfemas que marcam os passados recente e distante. Há dois aspectos: o perfectivo, marcado pela vogal final -í; o imperfectivo, sem marca específica. A análise dos tempos e dos aspectos revelou que o Aspecto é mais fundamental em lembaama. O estudo dos modos permitiu identificar três: o imperativo, o condicional e o indicativo. / This work proposes an analysis of the verbal morphology of Lembaama (B62) according to Guthrie (1971). Officially called Obamba in Gabon, Lembaama is a Bantu language of the forest, from the Benue-Congo group and Niger-Congo phylum. As far as we know, this language has yet to receive a detailed study. It should be noted that Lembaama shows some interesting features. Indeed, the verbal morphology analysis shows that a single inflected verb contains the following elements: subject, subject marker, negative1, Tense marker, root, extension, Final Vowel or Aspect marker, negative2 occurring in a fix order in a sentence. The object marker comes after the root. Negation consists of a discontinuous morpheme: kaní. Ka- is placed before the root (by the left) and -ní occurs after the root (by the right) being the last element of this structure. The verbal derivation study reveals the following structure of Lembaama extensions: -C-, -CV-, -VC- and -CVC-.The habitual marker -ag- being the commonest extension. This study also highlights the existence of a correlation between the grammatical value and the semantic value of extensions. Hence, human and animate beings are evoked in clauses with higher transitivity than things. Therefore, transitivity is fundamental in Lembaama, as it can distinguish actions from states, for example. Tense, Aspect and Mood study defines three Tenses. First, there is a present that, because it is used without tense marker, can be merged with near future or with progressive. Then, we note two future tenses: a near future (F1) occurring without mark and a distant future (F2) marked by the auxiliary verb odze go. Finally, we count three past tenses: a recent (P1) marked by mí- with a high tone; a distant (P2) marked by máá- with a high tone too, and a remote past (P3) marked by kí be, an auxiliary verb with high tone and mí-, the recent past marker, both coming before the root (by the left). Thus, Tense marker morphemes are always placed before the root (by the left) and Aspect markers occur after the root (by the right). This analysis highlighted two Aspects: a perfective marked by the Final Vowel -í; an imperfective without a specific mark. Tense and Aspect analysis allows to conclude that Aspect is more fundamental than Tense in this language. Concerning Mood, the analysis revealed three: imperative, conditional and indicative.