A partir des données contextualisées, extraites d’ESLO, un corpus oral du français contemporain, cette thèse étudie la variation dans l’actualisation des discours représentés (rapportés), avec une double finalité : rendre compte, par un modèle théorique, de leur complexité structurale dans le contexte des interactions orales et analyser leur influence sur la construction des liens interlocutifs et des trajectoires thématiques.Après avoir délimité les frontières externes du phénomène et situé le concept de discours représenté par rapport à plusieurs approches linguistiques et extralinguistiques qui reflètent sa complexité épistémologique, cette étude s’engage dans deux directions. La première, formelle, proposera une définition des discours représentés en tant que constructions grammaticales, définies par l’inséparabilité de trois critères : la structure bipartite, le décalage énonciatif et le fonctionnement métadiscursif. L’apport principal d’une telle analyse se précise dans la réinterprétation des liens entre le « discours citant » et le « discours cité », étudiés dans une perspective macrosyntaxique qui s’applique à toute la variation dans l’actualisation des discours représentés en contexte des interactions orales. La seconde direction, qui se rapporte aux liens entre les discours représentés et la construction d’une relation interlocutive, présente l’étude de plusieurs configurations de discours représentés en contexte, avec deux finalités : l’étude du rapport à autrui, l’interlocuteur, moyennant la mise en scène et l’illusion d’authenticité véhiculées par le discours représenté et l’étude du rapport à soi-même, le locuteur, par la construction d’une image de soi à travers les interventions dans la reconstruction des paroles autres. / Based on a contextualized data extracted from ESLO, an oral corpus of modern French, this PhD thesis examines the variation in the actualization of represented (reported) speech, with a double aim: to account for their structural complexity in oral interaction and to analyze their influence on the construction of the interlocutive relation and thematic paths.After defining the external borders of the phenomenon and placing the concept of represented speech with regard to several linguistic and non-linguistic approaches that reflect its epistemological complexity, this study engages in two directions. The first one, formal, will define represented speech as grammatical constructions, determined by the inseparability of three criteria: the bipartite structure, the enunciation offset and the metadiscursive functioning. The main contribution of this analysis is being illustrated in the reinterpretation of the link between “quoting discourse” and “quoted discourse”, analyzed in the macrosyntactic perspective that applies to all occurrences of represented speech in oral interaction. The second direction, which relates to the links between represented speech and the construction of an interlocutive relation, studies several represented speech configurations in context, with two aims: the study of the relation to the interlocutor, through the staging and the authenticity illusion conveyed by represented speech, and the study of the relation to oneself, the speaker, constructing a self-image through his interventions in the reconstruction of other words.

Notre recherche a été motivée par l’observation personnelle que nous avons pu faire des dispositifs d’enseignement bilingues scolaires en Europe centrale et orientale. Il nous ait paru justifié d’apporter notre réflexion à la problématique générale de l’enseignement bilingue. Nous avons choisi comme entrée la description, au sein du dispositif franco-moldave, des pratiques de classes des professeurs de biologie dispensant leur discipline en utilisant le français. Notre observation se focalise sur la gestion de leurs compétences langagières dans l’alternance entre le roumain et le français. Pour ce faire nous nous appuyons sur les analyses d’un corpus d’entretiens avec les enseignants et d’un corpus important de cours filmés.Nous sommes inscrits dans une démarche ethnographique en convoquant les outils de l’analyse interactionnelle. Notre analyse montre que, dans les classes observées, la transmission des connaissances disciplinaires se fait soit dans un registre monolingue L1 ou L2, soit dans un registre bilingue L1 et L2.La description détaillée que nous en faisons, montre des stratégies d’enseignement appliquées à la discipline avec une intégration de la L2 plus ou moins importante.Paradoxalement, les enseignants disposant moins de compétences langagières en L2 développent davantage un enseignement bilingue, soit un enseignement en deux langues, de leur discipline.En résumé, la gestion raisonnée de l’alternance codique est le coeur du processus de l’enseignement bilingue dont le bénéfice repose sur la maitrise de celle-ci. / Our research was motivated by personal observation realized in bilingual education systems of Central and Eastern Europe. We found it, justified tobring our thoughts to the general issue of bilingual education. We chose to enter in the reflection by the descriptive way of franco-moldovian practica classes of biology, where teachers provide their discipline using French. Our observation focuses on the management of their language skills in the alteration between Romanian and French built on the analysis of the corpus of interviews with the teachers and another filmed one with the lessons.To achieve this, we entered into the ethnographic approach applying the tools of the interactional analysis. Our analysis shows that in the classes underobservation, the transmission of discipline was made either in a monolingualL1/L2 register or in bilingual L1 and L2 one. The detailed description we make shows the teaching strategies applied with more or less important integration of L2 in the discipline.Paradoxically enough, the teachers with less L2 language skills develop moreb ilingual education, meaning more instructions in two languages in their discipline.To summarize, the heart of the process of bilingual education is the rational management of code-switching and the benefit is based on the level of its mastery. / Cercetarea noastră a fost motivată de observația personală pe care amputut sa o facem a sistemelor școlare de învățământ bilingv din EuropaCentrală și de Orientală. Ne pare a fi justificat de a aduce reflecţia noastrăasupra problemei generale a învățământului bilingv. Am ales să abordămreflecţia prin descriere, în cadrul dispozitivului franco-moldovenesc, orepractice a profesorilor de biologie care asigură disciplina utilizînd limbafranceză. Analiza noastră se concentrează pe gestionarea competențelorlingvistice în alternanța între limba romană și limba franceză sprijinindu-ne peanalizele unui corpus de interviuri cu profesori și a unui corpus foarteimportant, a lecţiilor filmate.Pentru a face acest lucru, noi ne-am înregistrat la o abordareetnografică convocînd instrumentele unei analize interacționale. Analizanoastră arată că în clasele observate, transmiterea de cunoștințe disciplinare seefectuiaza fie într-un registru L1 monolingv sau L2 sau într-un registru bilingvL1 și L2. Descrierea detaliată pe care noi o facem, arată strategii de predareaplicate la disciplină cu o integrare a L2 mai mult sau mai puțin importantă.În mod paradoxal, profesorii care dispun de mai puţine competențelingvistice în L2 dezvoltă un învățământ bilingv, fie un învăţămînt în douălimbi, a disciplinei lor.Pe scurt, gestionarea rațională a alternanţei codicale este centrulprocesului învăţămîntului bilingv, a cărui beneficiu se bazează pe stăpânireaacesteia.

La thèse présentée ici a pour but de proposer des systèmes de détection, de composition de constituants sémantiques et d’interprétation dans la compréhension de la langue naturelle parlée. Cette compréhension se base sur un système de reconnaissance automatique de la parole qui traduit les signaux oraux en énoncés utilisables par la machine. Le signal de la parole, ainsi transcrit, comporte un ensemble d’erreurs liées aux erreurs de reconnaissance (bruits, parasites, mauvaise prononciation...). L’interprétation de cet énoncé est d’autant plus difficile qu’il est issu d’un discours parlé, soumis à la disfluence du discours, aux auto-corrections... L’énoncé est de plus agrammatical, car le discours parlé lui-même est agrammatical. L’application de méthodes d’analyses grammaticales ne produit pas de bons résultats d’interprétation, sur des textes issus de transcriptions de la parole. L’utilisation de méthodes d’analyses syntaxiques profondes est à éviter. De ce fait, une analyse superficielle est envisagée. Un des premiers objectifs est de proposer une représentation du sens. Il s’agit de considérer des ontologies afin de conceptualiser le monde que l’on décrit. On peut exprimer les composants sémantiques en logique du premier ordre avec des prédicats. Dans les travaux décrits ici, nous représentons les éléments sémantiques par des frames (FrameNet ). Les structures de frames sont hiérarchisées, et sont des fragments de connaissances auxquels on peut insérer, fusionner ou inférer d’autres fragments de connaissances. Les structures de frames sont dérivables en formules logiques. Nous proposons un système de compréhension de la parole à partir de règles logiques avec le support d’une ontologie, afin de pouvoir créer des liens à partir de composants sémantiques. Puis, nous avons mené une étude sur la découverte des supports syntaxiques des relations sémantiques. Nous proposons une expérience de composition sémantique afin d’enrichir les composants sémantiques de base. Enfin, nous présentons un système de détection de lambda-expression pour mettre en hypothèse les relations à trouver à travers le discours / The thesis presented here is intended to provide detection systems, composition of components and semantic interpretation in the natural spoken language understanding. This understanding is based on an automatic speech recognition system that translates the signals into oral statements used by the machine. The transcribed speech signal, contains a series of errors related to recognition errors (noise, poor pronunciation...). The interpretation of this statement is difficult because it is derived from a spoken discourse, subject to the disfluency of speech, forself-correction... The statement is more ungrammatical, because the spoken discourse itself is ungrammatical. The application of grammatical analysis methods do not produce good results interpretation, on the outcome of speech transcription. The use of deep syntactic analysis methods should be avoided. Thus, a superficial analysis is considered. A primary objective is to provide a representation of meaning. It is considered ontologies to conceptualize the world we describe. We can express the semantic components in first order logic with predicates. In the work described here, we represent the semantic elements by frames (FrameNet ). The frames are hierarchical structures, and are fragments of knowledge which can be inserted, merge or infer other fragments of knowledge. The frames are differentiable structures in logical formulas. We propose a system for speech understanding from logical rules with the support of an ontology in order to create links from semantic components. Then, we conducted a study on the discovery supports syntactic semantic relationships. We propose a compositional semantics experience to enrich the basic semantic components. Finally, we present a detection system for lambda-expression hypothesis to find the relationship through discourse

Les traitements en cancérologie sont soumis au même risque de non-adhérence que les autres traitements ambulatoires au long cours. Ce travail de thèse a pour objectifs : i) d'étudier les facteurs de risque de dis-adhérence ; ii) de développer un programme d'éducation du patient (PEP) permettant d'améliorer l'adhérence des patients. La première étude étudie les facteurs de risque de non adhésion liés à l'interaction avec le système de soin ; les pratiques des oncologues français relatives à la prescription, la surveillance et l'accompagnement des patients ne sont pas suffisantes actuellement pour assurer un bon niveau de sécurité et d'adhérence des patients sous anticancéreux oraux. Les facteurs de risque de non adhésion liés au patient et au traitement sont étudiés dans la deuxième étude, chez des patients traités par capécitabine pour cancer du sein ou du côlon. Les patients actifs avec un niveau éducatif élevé seraient moins adhérents que les 2 autres. L'ensemble des patients sous capécitabine ont des comportements de sur-adhérence qui mettent en péril leur vie à cause des toxicités induites. La troisième étude présente l'évaluation d'un PEP construit selon une méthodologie standardisée. Ce programme démontre une efficacité dans l'amélioration des connaissances des patientes et de la confiance dans le traitement. Cette étude pilote a permis de modifier ce programme pour qu'il soit plus efficace. Identifier les facteurs de risque de non adhérence aux traitements anticancéreux oraux, à l'aide de méthodologies valides et adaptées au contexte, permet de construire des stratégies d'amélioration de l'adhérence ciblées et efficaces, en faisant levier sur ces facteurs / Non-adherence to oral chemotherapies can lead to lowered efficacy and increased risk of adverse events. The objective of this PhD work was twofold : i) to identify dis-adherence risk factors ii) to develop and test the feasibility of a validated, tailored therapeutic educational program with the aim of improving adherence to oral endocrine adjuvant chemotherapy in breast cancer. A survey was carried out to collect information on drug prescription, administration and surveillance, in order to identify non- adherence risk factors related to health professional behaviors : the majority of prescribers followed no standards in prescription writing, safety monitoring, toxicity prevention and patient education. A cohort study was carried out to identify adherence profiles among patients treated with capecitabine, using a mixed method. A profile of low adherence appeared (highly educated patients, with an irregular active life, with occupied relatives) and absolutely all patients showed an over-adherence profile (with a high risk of toxicity). The pilot study assessing the development and the feasibility of an educational program tailored to patients’ needs led to the improvement of the program : an extra session dealing with anxiety was built, and a new recruitment method was developed. Otherwise, the program succeeded in improving knowledge and trust in the treatment. This PhD work succeeded in identifying new dis-adherence risk factors, thanks to qualitative-quantitative methods. Those risk factors were incorporate in the development process of an educational program, in order to tailor it to the targeted population. This method should guarantee the efficacy of the program on patient’s adherence

Au 21ème siècle, communiquer en langues étrangères est désormais un facteur-clé dans un contexte mondialisé. C’est pourquoi, nous consacrons cette recherche à évaluer la compétence orale des apprenants saoudiens. Dans un premier temps, nous étudions les différentes dimensions de la compétence orale tout en mettant l’accent sur les méthodologies visant à développer la compétence orale. De même, nous analysons l’évaluation qui constitue un processus important dans un cours de FLE. Dans le cadre de cette thèse, nous analysons la production orale des apprenants saoudiens qui ont participé à plusieurs situations de communication. Cette analyse concerne, d’une part, la capacité des apprenants de communiquer oralement dans les différentes situations de communication. D’autre part, nous étudions leurs difficultés rencontrées aux niveaux lexicales, grammaticales et phonétiques. Pour aller plus loin dans notre recherche, nous avons recours aux Technologies de l’Information et de la Communication (TIC) en élaborant une formation de FLE à distance. Celle-ci a aidé, malgré certains obstacles, à améliorer la compétence orale chez les apprenants saoudiens. / In the 21st century, communicating in foreign languages has become a key factor in a global context. Therefore, this research is dedicated to assess the oral skills of Saudi learners. First of all, the research studies the various dimensions of the oral skills while focusing on the methodologies of developing those skills. It analyses as well the assessment that constitutes an important process in an FFL course. As part of this thesis, we’ve conducted a study analyzing the oral performance of some Saudi learners who took part in several communication situations. On one hand, this study analyses the ability of learners to communicate orally in various situations of communication. On the other hand, it studies the difficulties encountered at the lexical, grammatical and phonetic levels. To go further in our research, we’ve used the Information and Communication Technologies (ICT) in developing an FFL distance training, which helped, despite some obstacles, improving oral skills of Saudis learners.

La tolérance orale permet la modulation de la réponse immunitaire à l’égard des antigènes exogènes présents dans la lumière intestinale. Essentiels à l’établissement d’une relation symbiotique entre le système immunitaire et la flore intestinale, l’induction et le maintien de la tolérance orale reposent sur différents mécanismes immunologiques. Parmi eux, l’induction de cellules T régulatrices par les cellules dendritiques et de mécanismes apoptotiques. Or, la glycoprotéine membranaire CD47 est impliquée, en périphérie, dans ces mécanismes. Cependant, le rôle de CD47 dans la tolérance orale n’est pas connu. À l’aide d’un modèle murin déficient en CD47, nous avons démontré principalement, que l’absence de CD47 est associée à une diminution de 50 % de la proportion de cellules dendrites myéloïdes CD11b+CD103- retrouvées dans les ganglions mésentériques. Suite au transfert adoptif de cellules T antigènes spécifiques dans nos différents modèles expérimentaux, on a, aussi, observé une diminution de 45 % de leur niveau d’activation dans les ganglions mésentériques. Malgré les effets observés, le CD47 n’est pas impliqué dans l’induction d’une réaction de tolérance orale secondaire à l’administration intragastrique de fortes doses d’ovalbumine. Cependant, nous avons démontré que CD47 est impliquée au niveau de la migration des cellules dendritiques de la peau et de certaines sous-populations retrouvées dans les ganglions mésentériques. / Oral tolerance allows the modulation of the immune response against exogenous antigens present in the intestinal lumen. Essential to establish a symbiotic relationship between the immune system and intestinal flora, the induction and maintenance of oral tolerance rests on different immunological mechanisms. Among them, induction of regulatory T cells by dendritic cells and apoptotic mechanisms. However, the membrane glycoprotein CD47 is involved in the periphery of these mechanisms. However, the role of CD47 in oral tolerance is unknown. With a mouse model deficient in CD47, we showed mainly that the absence of CD47 is associated with a decrease of 50% in the proportion of myeloid dendritic cells CD11b+ CD103- found in the mesenteric lymph nodes. Following the adoptive transfer of antigen specific T cells in our experimental models, we also observed a decrease of 45% of their level of activation in mesenteric lymph nodes. Despite the observed effects, CD47 is not involved in the induction of oral tolerance response secondary to intragastric administration of high doses of ovalbumin. However, we have shown that CD47 is involved in the migration of dendritic cells of the skin and some sub-populations found in mesenteric lymph nodes.

En France, le cancer du sein est le cancer le plus fréquent chez les femmes et la deuxième cause de décès par cancer. Si la mortalité relative du cancer du sein a diminué ces dernières années, de nombreux progrès restent à faire pour les 20 à 25 % de patientes qui évoluent vers un stade métastatique. Cette étude s’appuie sur la réalisation de 103 entretiens semi-directifs réalisés auprès d’acteurs de la prise en charge en oncologie et de patientes atteintes de cancer du sein métastatique. L’objectif de ce travail de thèse était d’étudier : l’impact des nouvelles thérapies ciblées sur l’organisation de la prise en charge et leur diffusion, le vécu des patientes avec les traitements et leur adhésion thérapeutique. Nous avons montré, dans un premier temps, que les nouvelles thérapies ciblées orales modifient la prise en charge traditionnelle des patientes qui restent moins dans la sphère d’influence de la structure de soins et qu’il existe de nombreux freins à la diffusion de ces traitements chez les oncologues. Dans un deuxième temps, nous avons mis en avant les représentations positives des patientes sur les thérapies ciblées orales qui sont sources d’espoir et améliorent la qualité de vie. Pour autant, le vécu au quotidien vient parfois nuancer ces représentations et peut révéler des freins à l’adhésion thérapeutique tels que l’apparition d’effets secondaires ou les responsabilités engendrées par la demande d’autonomie. Enfin, dans un troisième temps, nous avons mis en exergue le manque de connaissances des patientes sur les innovations thérapeutiques et l’importance de la décision partagée avec l’oncologue tout particulièrement face à la chronicisation progressive de certains cancers du sein métastatiques. / In France, breast cancer is the most common cancer amongst women and the second cause of cancer death. Although breast cancer relative mortality has decreased significantly in recent years, there is still a lot to be done for the remaining 20% to 25% of patients who have metastasized. This study is based on 103 semi-structured interviews carried out with oncology healthcare professionals and metastatic breast cancer patients. The aim of this work was to study the impact that new targeted therapies can have on managing the patient’s care and their diffusion, the patients’ experiences with these treatments, their social representations and their therapeutic adherence. First, we showed that new oral targeted therapies disrupt the traditional steps in a patient care. Patients arefar less in the hospital’s sphere of influence and there are many barriers to the dissemination of these treatments for oncologists. Second, we highlighted the positive representations of oral targeted therapies that patients have, which is bringing them hope and improving their quality of life. However, the day-to-day experience with the treatment sometimes moderates these representations and may reveal obstacles to their therapeutic adherence such as side effects or the responsibilities created by the demand for autonomy. Finally, we showed the lack of patients’ knowledge about therapeutic innovations and the importance of the implementation of shared decision making with the oncologist, especially in the face of the progressive chronicisation of a number of metastatic breast cancers.

Background: To assess the long-term course of the feasibility and safety of orally administered heroin [diacetylmorphine (DAM)] tablets in substitution treatment of severely addicted opioid users. Design: Open-label, prospective cohort study with 2 non-randomly assigned treatment arms: DAM tablets only (n = 128) or DAM tablets combined with injected DAM and/or other opioids (n = 237). The average duration of the observation period was 62 months. Study endpoints were the time to discharge from treatment and the number of serious adverse events. Results: Both patient groups had a higher than 70% retention rate after the first 48 months of treatment, with similar long-term retention rates (after 8 years both groups had retention over 50%). The physician-verified rate of serious adverse events was 0.01 events per application year among the exclusively oral substitution group (intention-to-treat analysis) during the last year of observation, and 0.005 events per application year in the other group. Conclusions: Because of their feasibility and safety over years, DAM tablets may be a valuable long-term therapeutic alternative. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.

Problématique: En immunothérapie orale (ITO), le manque de variété et l’aversion envers les doses d’allergènes peuvent compromettre l’adhérence au traitement, toutefois essentielle pour maintenir la désensibilisation. Le but de l’étude était de développer et de valider une nouvelle intervention nutritionnelle pour l’utilisation d’options d’équivalences à domicile. Méthodes: L’intervention a été développée selon les besoins de familles déjà en ITO, exprimés lors d’entrevues préliminaires. De nouveaux patients débutant l’ITO ont ensuite été invités dans un essai contrôlé randomisé pour évaluer l’impact de l’intervention. Les participants (n = 30) ont été randomisés en 3 groupes : A) Consultation nutritionnelle avec outils d’options d’équivalences (intervention complète); B) Consultation nutritionnelle sans les outils (intervention partielle) et C) Groupe contrôle avec l’intervention complète retardée de 4 semaines. La compétence des parents pour le calcul de doses d’équivalences était suivie de façon longitudinale par une série d’exercices pratiques. Résultats: Les résultats aux exercices étaient en moyenne supérieurs avec l’intervention complète (93,3% ± 3,1), comparés au groupe contrôle sans intervention (1,7% ± 1,7, p<0,001). La compétence était maintenue 12 semaines plus tard (résultats de 88,9% ± 4,7). Sans les outils, l'acquisition initiale (résultats de 46,7% ± 7,3) et la rétention après 4 semaines (résultats de 26,7% ± 5,1) étaient inférieures, mais augmentaient après l’ajout des outils (résultats de 83,3% ± 7,5). La satisfaction et la diversité des doses ont également augmenté avec l’intervention complète. Conclusion: Cette étude démontre l'efficacité d'un programme d'intervention nutritionnelle pour accompagner la gestion des doses d'allergènes à domicile. L'utilisation de documents écrits est essentielle pour en obtenir tout le bénéfice. / Background: During oral immunotherapy (OIT), lack of palatability or diversity in daily allergen doses can compromise treatment adherence, which is essential to maintain benefit. The aim of the study was to develop and validate a nutritional intervention program on the use of whole food alternatives for allergen daily dosing during OIT. Methods: The program was initially developed based on preliminary interviews with families already on OIT. Patients beginning OIT were then invited to participate to an open-label randomized controlled trial to assess the impact of the intervention. Participants (n=30) were randomized into 3 arms when they transferred to whole foods: A) Dietitian counselling with supporting documents (full intervention); B) Dietitian counselling without document; C) Control group where full intervention was delayed by 4 weeks. Parent competency was followed longitudinally using a series of practical food dose calculation exercises. Results: Results of exercises at week 4 were in average higher in the full intervention group (93.3% ± 3.1) compared to reference group without intervention (1.7% ± 1.7, p<0.0001). Competency was maintained 12 weeks after intervention (results of 88.9% ±4.7). Without written documents, the initial acquisition (results of 46.7% ±7.3) and retention of competency at 4 weeks (results of 26.7% ±5.1) were lower, but competency was rescued by adding written documents (results of 83.3% ±7.5). Patient satisfaction and food diversity also increased with full intervention. Conclusion: This study demonstrates the efficacy of a nutritional intervention program to help patients and their parents manage their OIT allergen doses. The use of written documents is essential to achieve the full benefit.

Negli ultimi 70 anni, c'è stato un incremento degli studi e ricerche inglesi sulle collocazioni (Firth 1957; Hoey, 2005; Moon, 1998b; Sinclair 1991; 2004; Stubbs, 1996; 2001), i quali hanno evidenziato che la fraseologia è pervasiva alla lingua (Altenberg, 1998; Biber et al., 1999; Cowie, 1991; 1992; Howarth; 1998). Questo indica anche che una buona padronanza delle collocazioni è necessaria se i discenti mirano a raggiungere una fluidità simile a quella di un nativo nella L2. Infatti, la ricerca sulla produzione di linguaggio formulaico da parte degli apprendenti ha dimostrato che le collocazioni sono essenziali nell'acquisizione della lingua seconda (Cowie, 1998; Pawley & Syder, 1983; Peters, 1983) e sono una componente chiave per lo sviluppo della "fluency" (Ellis, 2002; 2003; Ellis et al., 2015; Howarth, 1998). Nonostante il maggior numero di studi sulle collocazioni, la maggior parte degli studiosi si è concentrata su dati scritti e su un insieme ristretto di combinazioni, come le collocazioni verbo + sostantivo. La scarsa disponibilità di corpora orali di discenti e la maggiore attenzione per le sequenze formulaiche più soggette a errori hanno portato i ricercatori a trascurare collocazioni come avverbio + aggettivo. L'intensificazione è una parte intricata dell'apprendimento delle lingue straniere (Lorenz, 1999) e merita ulteriore attenzione, soprattutto per quanto riguarda i dati parlati, che riflettono meglio il linguaggio dei discenti (Myles, 2015). Il presente lavoro indaga le collocazioni di avverbi + aggettivi in un corpus parlato di recente compilazione di studenti italiani avanzati di inglese L2. La tesi adotta un approccio di Analisi Interlinguistica Contrastiva (Granger, 1998) per verificare se: a) ci sono differenze tra la produzione di collocazioni degli studenti italiani di inglese rispetto ai coetanei madrelingua; b) ci sono differenze tra le collocazioni prodotti dagli studenti italiani e quelle dei madrelingua in termini di modelli sintattici e significato lessicale; c) la congruenza della L1 ha un effetto di trasferimento sulla produzione da parte dei discenti di collocazioni poco frequenti e/o non attestate. Per rispondere alle tre domande di ricerca, sono state condotte analisi quantitative e qualitative sull'Italian Spoken Learner Corpus (ISLC) e sul corpus gemello di LINDSEI, LOCNEC. LOCNEC è stato utilizzato come corpus di riferimento di madrelingua per il suo alto livello di comparabilità con ISLC. Per le analisi quantitative, è stato seguito l'approccio di Durrant e Schmitt (2009) per il calcolo dei punteggi delle misure di associazione delle collocazioni (t-score e MI) sulla base del corpus di riferimento BNC e le collocazioni sono state poi divise in tre categorie in base al loro punteggio: collocazioni (t-score e MI maggiore o uguale a 2 e 3 rispettivamente), collocazioni infrequenti/non attestate (t-score e MI non disponibili a causa dell'infrequenza), collocazioni in area grigia (t-score e MI inferiore a 2 e 3 rispettivamente). I test T-test e Wilcoxon rank sum test sono stati utilizzati sulle collocazioni estratte da ISLC e LOCNEC e sono state calcolate le dimensioni degli effetti. Inoltre, i test sono stati impiegati per valutare i valori medi individuali di t-score e MI degli studenti e dei madrelingua. Per quanto riguarda le analisi qualitative, è stato impiegato uno schema a tre livelli per analizzare due serie di collocazioni: la prima serie comprende 11 collocazioni con t-score e MI maggiore uguale a 2 e 3 rispettivamente e una frequenza di 5 nell'ISLC; la seconda serie comprende 9 collocazioni infrequenti/non attestate con una frequenza maggiore o uguale a 2 nell'ISLC. Seguendo lo schema, i due set di collocazioni estratti sia dall'ISLC sia dal LOCNEC sono stati analizzati tenendo conto del loro background collocativo (etimologia, livello CEFR, congruenza L1), delle variabili del discente (sesso, esperienza di soggiorno all'estero, corso universitario, altre lingue), e delle variabili testuali (funzione attributiva vs predicativa dell'aggettivo, pronomi vs it-sentences, tempo verbale, affermativo vs negativo, connotazione positiva vs negativa). I risultati dei test statistici sono stati tutti significativi con effect size medio-grandi e, insieme alle analisi qualitative, hanno indicato che: gli studenti italiani di inglese producono un minor numero di collocazioni; un maggior numero di non-collocazioni; le loro combinazioni sono meno collocative di quelle dei madrelingua (ovvero, i loro punteggi di misura delle associazioni sono in media più bassi di quelli dei nativi); non ci sono differenze marcate in termini di modelli lessico-grammaticali tra le collocazioni degli studenti e quelle dei madrelingua, ma gli studenti tendono ad assegnare alle loro collocazioni funzioni più creative dal punto di vista pragmatico; non è stata trovata alcuna prova di trasferimento L1 (negativo) in relazione alla produzione da parte dei discenti di collocazioni infrequenti/non attestate, sostenendo così ulteriormente la conclusione precedente. I risultati corroborano ulteriormente la letteratura sulle collocazioni prodotte dai discenti e aggiungono un altro tassello al puzzle della lingua parlata: il ritardo collocazionale, cioè lo sviluppo più lento delle prestazioni di produzione di collocazioni, può essere trovato anche nei dati parlati e i discenti sembrano anche produrre meno collocazioni identificate da punteggio t-score. Questo ha due importanti, anche se semplici, implicazioni: che gli studenti dovrebbero probabilmente essere esposti a più input di lingua parlata, e che le teorie di acquisizione della lingua seconda potrebbero utilmente rivedere i processi di acquisizione fraseologica degli studenti nel contesto EFL. Un'altra scoperta è relativa ai modelli lessico-grammaticali delle collocazioni degli studenti non erano marcatamente diversi da quelli dei madrelingua, ma erano meno vari e mostravano una creatività pragmatica. Questo potrebbe informare gli studiosi sui potenziali processi di fossilizzazione (Selinker, 1972) nella fraseologia e/o sulle strategie di semplificazione o di evitamento (Farghal & Obiedat, 1995). Infine, anche se gli studi tradizionali hanno trovato che la congruenza L1 gioca un ruolo chiave nella produzione di collocazioni (cfr. Bahns, 1993; Granger, 1998b; Nesselhauf, 2005; Wang, 2016), nessuna prova di congruenza L1 è stata trovata per quanto riguarda i dati parlati, il che è un'interessante controprova. Nel complesso, questa tesi ha sottolineato che la produzione di collocazioni, sia quantitativamente sia pragmaticamente, è in ritardo rispetto alla competenza collocazionale e, sebbene questa linea possa essere molto sottile e non significativa nei testi scritti, il divario si allarga nella lingua parlata. / In the last 70 years, there has been an increase in English studies on collocations (Firth 1957; Hoey, 2005; Moon, 1998; Sinclair 1991; 2004; Stubbs, 1996; 2001) and research which have documented that phraseology is pervasive to language (Altenberg, 1998; Biber et al., 1999; Cowie, 1991; 1992; Howarth; 1998). This also indicates that a good command of collocations is needed if learners aim to achieve native-like fluency in the L2. Indeed, research on learner production of formulaic language has shown that collocations are essential in the acquisition of second language (Cowie, 1998; Pawley & Syder, 1983; Peters, 1983) and are a key component for the development of fluency (Ellis, 2002; 2003; Ellis et al., 2015; Howarth, 1998). Despite the surge in studies on collocations, the majority of scholars have focused on written data and on a restricted set of combinations, such as verb + noun collocations. The poor availability of spoken learner corpora and the more error-prone formulaic sequences have led researchers to neglect collocations such as adverb + adjective. Intensification is an intricate part of foreign language learning (Lorenz, 1999) and deserves further attention, especially as regards spoken data, which is a better reflection of learner language (Myles, 2015). The present work investigates adverb + adjective collocations in a newly compiled spoken learner corpus of advanced Italian learners of English L2. The thesis adopts a Contrastive Interlanguage Analysis (Granger, 1998) approach to verify whether: a) there are any differences between the collocation production of Italian learners of English compared to native-speaker peers; b) there are any differences between the Italian learners’ collocations and the native speakers’ in terms of syntactic patterns and lexical meaning; c) L1 congruency has a transfer effect on the learner production of infrequent and/or unattested collocations. In order to address the three overarching research questions, quantitative and qualitative analyses were carried out on the Italian Spoken Learner Corpus (ISLC) and the sister corpus of LINDSEI, LOCNEC. LOCNEC was used as the native-speaker reference corpus for its high level of comparability to ISLC. For the quantitative analyses, Durrant and Schmitt’s (2009) approach was followed for the calculation of the collocation’s association measure scores (t-score and MI) based on the large reference corpus BNC and the collocations were then divided into three categories based on their score: collocations (t-score and MI equal or greater than 2 and 3 respectively), infrequent/unattested collocations (t-score and MI scores unavailable due to infrequency), grey area collocations (t-score and MI lower than 2 and 3 respectively). T-tests and Wilcoxon rank sum tests were computed on the collocations extracted from ISLC and LOCNEC and effect sizes were calculated. In addition, the tests were employed to assess the average individual t-score and MI values of learners and native speakers. As regards the qualitative analyses, a three-fold scheme was employed to analyse two sets of collocations: the first set comprises 11 collocations with t-score and MI equal or greater than 2 and 3 respectively and a frequency of equal or greater than 5 in the ISLC; the second set includes 9 infrequent/unattested collocations with a frequency equal or greater than 2 in ISLC. Following the scheme, the two sets of collocations extracted from both ISLC and LOCNEC were analysed by taking into account their collocational background (etymology, CEFR level, L1 congruence), the learner variables (gender, stay-abroad experience, university course, other languages), and the text variables (attributive vs predicative function of the adjective, pronouns vs it-sentences, tense, affirmative vs negative, positive vs negative connotation). The results of the statistical tests were all significant with medium to large effect sizes and, together with the qualitative analyses, indicated that: Italian learners of English produce a fewer number of collocations; a higher number of non-collocations; their combinations are less collocational than native speakers’ (i.e., their association measure scores as on average lower than the natives’); there are no marked differences in terms of lexico-grammatical patterns between the learners’ collocations and the native speakers’, but the learners tend to assign more pragmatically creative functions to their collocations; no evidence of L1 (negative) transfer was found in relation to the learners’ production of infrequent/unattested collocations, thus further supporting the previous finding. The findings further corroborate the literature on learners’ collocations and add another piece to the puzzle of spoken language: collocational lag, that is the slower development of collocation performance, can also be found in spoken data and learners also seem to produce fewer t-score collocations. This has two important, though simple, implications: that learners should probably be exposed to more spoken language input, and that second language acquisition theories might usefully review phraseological acquisition processes of EFL learners. Another finding is that the lexico-grammatical patterns of learners’ collocations were not markedly different from native speakers’, but they were less varied and displayed pragmatic creativity. This could inform scholars about potential fossilisation processes (Selinker, 1972) in phraseology and/or simplification or avoidance strategies (Farghal & Obiedat, 1995). Lastly, although mainstream studies have found that L1 congruency plays a role in the production of collocations (cf. Bahns, 1993; Granger, 1998b; Nesselhauf, 2005; Wang, 2016), no evidence of L1 congruency was found as regards spoken data, which is an interesting counter-finding. Overall, this thesis has underlined that collocation production, both quantitatively and pragmatically, lags behind collocation competence and, although this line may be very thin and not significant in written texts, the gap widens in spoken language.