This dissertation is an analysis of the segmental phonology of the Shangani language as spoken in the South Eastern parts of Zimbabwe. It starts by presenting the language situation in Zimbabwe and comparing the language’s status in Zimbabwe with that of its sister varieties in South Africa where it is referred to as Tsonga and in Mozambique where it is referred to as XiChangana or Changana. The dissertation is based on data collected from the speakers of Shangani using a variety of research techniques. The dissertation identifies and characterizes the language’s distinctive phonemes using the minimal pair and set tests. It presents the language’s consonants, which include aspirated, breathy-voiced, pre-nasalized, labialized and palatalized consonants. It shows that in Shangani, voiceless consonants cannot be pre-nasalized and that there is an incompatibility between that labio-velar glide /w/ and most labial consonants excpt /m/. The phonemes are analysed using Chomsky and Halle’s (1968) distinctive feature theory. The study uses Clements and Keyser’s (1983) CV phonology of the syllable structure to analyse the language’s syllable structure. The language’s canonical syllable structure is CV. It is also shown that consonant clusters are gaining their way into the language through borrowing from English, Afrikaans and other languages that have consonant clusters in their inventories. Onsetless Vs are marginally attested word-initially. In agent nouns, VV sequences are in most cases retained. These sequences are not analysed as diphthongs since they occupy different V slots on the syllable tier. The second vowel in the sequence is the onsetless syllable. Affricates, NCs, Cws and Cjs are presented as unitary segments that occupy a single C slot of the CV tier. Phonological processes that are attested in the language are also presented. Secondary articulation, vowel deletion, feature spreading, vowel coalescence and nasalization are shown to be the most common phonological process in the language. Since Shangani has the CV syllable typology, most of the phonological processes are there to resolve hiatus that would have been induced by suffixation of vowel commencing stems or suffixes to vowel final prefixes or stems. The notion of domains is shown to be a diagnostic tool for identifying a process in a hiatus situation. The study shows that vowel deletion is the least preferred strategy when secondary articulation, feature spreading, vowel coalescence have been blocked by some constraints like syllable structure processes or the language’s phonotactics / African Languages / D. Litt. et Phil. (African Languages)

This dissertation is an analysis of the segmental phonology of the Shangani language as spoken in the South Eastern parts of Zimbabwe. It starts by presenting the language situation in Zimbabwe and comparing the language’s status in Zimbabwe with that of its sister varieties in South Africa where it is referred to as Tsonga and in Mozambique where it is referred to as XiChangana or Changana. The dissertation is based on data collected from the speakers of Shangani using a variety of research techniques. The dissertation identifies and characterizes the language’s distinctive phonemes using the minimal pair and set tests. It presents the language’s consonants, which include aspirated, breathy-voiced, pre-nasalized, labialized and palatalized consonants. It shows that in Shangani, voiceless consonants cannot be pre-nasalized and that there is an incompatibility between that labio-velar glide /w/ and most labial consonants excpt /m/. The phonemes are analysed using Chomsky and Halle’s (1968) distinctive feature theory. The study uses Clements and Keyser’s (1983) CV phonology of the syllable structure to analyse the language’s syllable structure. The language’s canonical syllable structure is CV. It is also shown that consonant clusters are gaining their way into the language through borrowing from English, Afrikaans and other languages that have consonant clusters in their inventories. Onsetless Vs are marginally attested word-initially. In agent nouns, VV sequences are in most cases retained. These sequences are not analysed as diphthongs since they occupy different V slots on the syllable tier. The second vowel in the sequence is the onsetless syllable. Affricates, NCs, Cws and Cjs are presented as unitary segments that occupy a single C slot of the CV tier. Phonological processes that are attested in the language are also presented. Secondary articulation, vowel deletion, feature spreading, vowel coalescence and nasalization are shown to be the most common phonological process in the language. Since Shangani has the CV syllable typology, most of the phonological processes are there to resolve hiatus that would have been induced by suffixation of vowel commencing stems or suffixes to vowel final prefixes or stems. The notion of domains is shown to be a diagnostic tool for identifying a process in a hiatus situation. The study shows that vowel deletion is the least preferred strategy when secondary articulation, feature spreading, vowel coalescence have been blocked by some constraints like syllable structure processes or the language’s phonotactics / African Languages / D. Litt. et Phil. (African Languages)

Στην παρούσα εργασία μελετήθηκε το θέμα της φωνηεντικής ιεραρχίας στην Ελληνική και τις διαλέκτους της και ο ρόλος της στην αντιμετώπιση της χασμωδίας. Η χασμωδία δημιουργείται στην ελληνική λόγω της γειτνίασης φωνηέντων και λύνεται μέσω φωνολογικών διαδικασιών όπως η έκθλιψη, η αφαίρεση, η εισαγωγή / επένθεση, η συνίζηση, η τροπή, η φωνηεντική αρμονία, η αφομοίωση κ.ά.Υπάρχουν όπως αναφέρθηκε μέσω και παραδειγμάτων και οι περιπτώσεις διατήρησης της χασμωδίας. Εξετάσαμε την κλίμακα ηχηρότητας η οποία καθορίζει τον ισχυρό ή ασθενή χαρακτήρα του φωνήεντος ανάλογα με τη θέση που κατέχει στην εν λόγω κλίμακα. Το ποιο από τα φωνήεντα θα φύγει εξαρτάται από τη θέση που κατέχει στην κλίμακα ισχύος ή κλίμακα ηχηρότητας. Η εν λόγω κλίμακα καθορίζεται ως η ακόλουθη: a > o > u > e > i ή i > e. Εξετάσαμε επίσης τον τρόπο με τον οποίο η ισχύς αυτή επηρεάζεται από διάφορους παράγοντες όπως η σύνταξη, το είδος του λόγου, ο τόνος, ο βαθμός ανοίγματος των φωνηέντων, η θέση τους στη λέξη, ο τόπος άρθρωσής τους, το λόγιο στοιχείο, ο μορφολογικός ρόλος, η σημασιολογία καθώς και ο παράγοντας σύνθεση, και επομένως είναι αυτοί οι παράγοντες βάσει της έρευνάς μας και πάντα με βάση τα δεδομένα που έχουμε στη διάθεσή μας που φαίνεται ότι καθορίζουν την ισχύ ή όχι της ιεραρχίας και τη διατήρηση ή όχι του φαινομένου της χασμωδίας. Επίσης αναφέρθηκε ότι και στα αρχαία ελληνικά αλλά και σε διαλέκτους της ελληνικής συμβαίνουν διαδικασίες φωνολογικές με σκοπό την αντιμετώπιση της χασμωδίας - ενώ παρατηρούμε ταυτόχρονα ότι αυτό δεν συμβαίνει πάντα με αποτέλεσμα να έχουμε και διατήρηση της χασμωδίας. Αναφέρονται παραδείγματα - όσο αυτό είναι δυνατό μέσα από την έρευνα και τις πηγές - ούτως ώστε μέσα πάντα από τα ίδια τα γλωσσολογικά δεδομένα, μέσα από τα κείμενα, να αποδεικνύεται κατά πόσον ισχύουν οι κανόνες και οι θεωρητικές θέσεις. Ξεκινούμε λοιπόν από τη μια μεριά από τις θεωρητικές θέσεις, τους κανόνες και από την άλλη όμως ταυτόχρονα αντλούμε πληροφορίες για το βαθμό εφαρμογής τους από τα ίδια τα γλωσσικά δεδομένα της ελληνικής και των διαλέκτων και από δεδομένα σε παλαιότερες χρονικές φάσεις αποδεικνύοντας μέσα από αυτό και μέσα από την κανονικότητα των φαινομένων διαχρονικά, τη συνέχεια της γλώσσας μας. / The present study explores the validity of vowel hierarchy in Greek and its dialects, as well as its role in dealing with hiatus. Hiatus emerges in adjacent vowels and it is resolved through phonological processes such as vowel deletion, vowel epenthesis, fusion, substitution, vowel harmony, assimilation etc., though there are instances in which hiatus is retained. The Sonority Scale (SS) is also investigated given that the (SS) determines the strong or weak status of the vowel, which, in turn, is driven by the position of the vowel in the SS. In other words, vowel loss is governed by its degree of sonority. In most, studies the vowel SS is considered to be the following: a > o > u > e > i ή i > e. Moreover, we examined the extent to which vowel sonority and the SS are circumscribed by other extra-phonological and phonological factors, such as the syntax, the semantics and the morphology of the language, the general linguistic context, the position of the vowels on the vowel space/ triangle, as well as their place of articulation. Additionally, we discuss phonological processes which repair hiatus occuring in Ancient Greek and its dialects and we test the conditions under which this happens. We present interesting data which lead us to adopt specific theoretical approaches regarding the vowel hierarchy. Finally, we evaluate the validity of the vowel hierarchy in the diachrony of Greek.

No description available.

Depuis le début des années 1990 on suit l'évolution globale du niveau de la mer grâce aux satellites altimétriques. Ils observent une hausse du niveau moyen global de la mer (GMSL) de 3.4 ± 0.4 mm/an sur la période 1993-2016 (ce qui représente le double de ce qui a été observé au cours du 20ème siècle par les marégraphes, hausse à 1.7 mm/an entre 1900 et 1990). Le GMSL présente aussi des fluctuations interannuelles qui peuvent atteindre quelques millimètres, surtout pendant les épisodes ENSO. Cette hausse n'est pas régionalement uniforme : elle a été 3 fois plus rapide que la hausse moyenne globale dans certaines zones entre 1993 et 2016. Au cours du 21ème siècle, on s'attend à une hausse accrue du GMSL pouvant aller jusqu'à 1 m à l'horizon 2100, avec une forte variabilité régionale. Il est donc important de comprendre l'évolution actuelle du niveau des océans qui constitue une menace sérieuse pour de nombreuses régions côtières basses souvent très peuplées. Cette thèse s'inscrit dans le contexte du projet niveau de la mer CCI (Climate Change Initiative) de l'Agence Spatiale Européenne (ESA) ayant pour objectif de fournir de meilleurs produits du niveau de la mer combinant les missions Topex/Poseidon, Jason-1/2, ERS-1/2 et Envisat. L'objectif premier de cette thèse est de valider ces produits SL_CCI du niveau de la mer en utilisant différentes approches, en particulier par l'étude du bilan (comparaison du GMSL observé avec la somme des différentes contributions : composante stérique, fonte des glaces continentales et transferts d'eau depuis les terres émergées). Un autre objectif est d'estimer les composantes du niveau de la mer mal connues, et tout particulièrement le contenu thermique de l'océan profond non mesurable par le système Argo, et la contribution du stock d'eau sur les continents. Ces travaux ont montré que la contribution de l'océan profond en dessous de 2000m est faible sur la période 2005-2013 et contenue dans la barre d'incertitudes des données (erreurs qui proviennent essentiellement, (1) des produits niveau de la mer altimétriques et des lacunes de la couverture géographique des données Argo dans la région Indonésienne pour la tendance et (2) des produits GRACE et Argo pour la variabilité interannuelle). Nos résultats et la méthode utilisée montrent que le niveau de la mer et ses composantes sont encore entachés d'erreurs importantes. Dans la deuxième partie, nous avons analysé l'influence du phénomène ENSO (El Niño et La Niña) sur les variations interannuelles du GMSL. Nous montrons que lors des évènements La Niña comme celui de 2010-2011, le déficit de précipitations sur l'océan (et l'excès sur les continents) conduit à une baisse temporaire de la masse de l'océan global et donc du niveau de la mer. C'est essentiellement la variation de masse de l'océan qui explique la variabilité interannuelle du niveau de la mer lors des évènements ENSO, et le déficit (La Niña) ou excès (El Niño) de masse se trouve confiné dans l'océan Pacifique tropical Nord. Pour finir, nous analysons l'évolution de la température moyenne de l'air et de l'océan en surface sur la période du "hiatus" (2003-2013). Nous montrons que ce hiatus, c'est à dire le ralentissement récent de la hausse de la température moyenne globale de la Terre est un phénomène quasi global, même si le Pacifique tropical Est s'est fortement refroidi. Cette "supposée" pause récente s'explique par la variabilité naturelle interne du climat. La Terre est toujours en état de déséquilibre énergétique dû à l'accumulation de gaz à effet de serre. Nous mettons en évidence le rôle de la variabilité naturelle à court terme sur les changements à plus long terme associés au réchauffement climatique anthropique. / Since the early 1990s sea level is routinely measured using high-precision altimeter satellites. These observe a rise in global mean sea level (GMSL) of 3.4 ± 0.4 mm/yr over the 1993-2016 period (which is twice what has been observed during the 20th century by the tide gauges, with a rise of 1.7 +/- 0.3 mm/yr). The interannual variability in the GMSL can reach several millimeters, especially during ENSO events. The rate of sea level rise is not regionally uniform. During the altimetry era, it was three times faster than the global mean in some areas. During the 21st century, we expect a greater rise of the GMSL than today, up to 1 m in 2100, with strong regional variability. It is therefore important to understand the current evolution of the sea level, since it represents a serious threat to many low coastal areas, often densely populated of the planet. My thesis research deals with the Sea Level CCI (Climate Change Initiative) project of the European Space Agency (ESA) which objective is to provide improved sea level products combining several altimetry missions, including Topex/Poseidon, Jason-1/2, ERS-1/2 and Envisat. The primary objective of my thesis was to validate the CCI sea level products using different approaches, in particular the sea level budget approach. It consists of comparing the observed GMSL with the sum of different contributions : the steric component, melting of continental ice and transfers of water between the land surface and oceans. Another objective was to estimate the poorly known components to sea level rise, in particular the heat content of the deep ocean not measurable by Argo, and the contribution of water storage on the land. My work has shown that the contribution of the deep ocean below 2000m to the rising sea level is small over the 2005-2013 periods and not significant compared to the data uncertainties. The main uncertainties come from: (1) -in terms of trend- the altimetry sea level products and gaps in the geographical coverage of Argo data in the Indonesian region, and (2) -in terms of interannual variability- the GRACE and Argo products. My results and the method used show that the sea level and its components are still affected by important errors. In the second part, I analyzed the influence of ENSO (El Niño and La Niña) on the interannual variations of the GMSL. I showed that during La Niña events, like that of 2010-2011, the rainfall deficit over the ocean (and excess over the continents) leads to a temporary decrease in the global ocean mass and therefore in the GMSL. This is essentially the ocean mass variation that explains the interannual variability of the GMSL during ENSO events. Furthermore, the deficit (La Niña) or excess (El Niño) ocean mass is confined in the north tropical Pacific Ocean. Finally, I analyzed the evolution of the average temperature of air and ocean surface over the period of the "hiatus" (2003-2013). I showed that this hiatus, i.e. the recent slowdown in the rise of the global mean Earth's temperature is an almost global phenomenon, though cooling of the tropical eastern Pacific has slightly contributed. This recent pause is attributable to natural internal climate variability. The Earth is indeed still in a state of energetic imbalance due to the accumulation of greenhouse gases. I highlighted the role of the natural variability that is superimposed to the anthropogenic global warming.