Promoting constructiveness in online comment sections is an essential step to make the internet a more productive place. On online marketplaces, customers often have the opportunity to voice their opinion and relate their experience with a given product. In this thesis, we investigate the possibility to model constructiveness in product review in order to promote the most informative and argumentative customer feedback. We develop a new constructiveness 4-class scale taxonomy based on heuristics and specific categorical criteria. We use this taxonomy to annotate 4000 Amazon customer reviews as our training set, referred to as the Corpus for Review Constructiveness (CRC). In addition to the 4-class constructiveness tag, we include a binary tag to compare modeling performance with previous work. We train and test several computational models such as Bidirectional Encoder Representations from Transformers (BERT), a Stacked Bidirectional LSTM and a Gradient Boosting Machine. We demonstrate our annotation scheme’s reliability with a set of inter-annotator agreement experiments, and show that good levels of performance can be reached in both multiclass setting (0.69 F1 and 57% error reduction over the baseline) and binary setting (0.85 F1 and 71% error reduction). Different features are evaluated individually and in combination. Moreover, we compare the advantages, downsides and performance of both feature-based and neural network models. Finally, these models trained on CRC are tested on out-of-domain data (news article comments) and shown to be nearly as proficient as on in-domain data. This work allows the extension of constuctiveness modeling to a new type of data and provides a new non-binary taxonomy for data labeling.

Ce travail propose un ensemble d’analyses sociogénétiques et psychogénétiques, fondées sur les sociologies de la connaissance de Norbert Elias et Günter Dux, portant sur les conséquences qui découlent de l’utilisation de causes premières dans l’explication des phénomènes du monde. L’enjeu de l’analyse, en termes généraux, a consisté en rendre visible le fait que la structure de la logique, en tant que produit des processus psychogénétiques, et lorsqu’elle n’a pas été réorganisée en termes systémiques par les transformations sociogénétiques de la modernité, a conduit dans le passé de nos sociétés à revêtir les causes premières de caractéristiques subjectives. Le concept de « nature » dans l’Allemagne du XVIIIe siècle, forgé et utilisé par la philosophie et le naturalisme, en est un exemple clair. Nous illustrons à travers nos analyses le fait que, grâce à la structure de la logique, la « nature » était « créatrice » dans la vision du monde des membres de la société allemande de l’époque, c’est-à-dire, une entité dotée d’intentionnalité et de capacité d’action.À la croisée de la sociologie de la connaissance et de l’épistémologie historico-génétique, l’enquête porte, dans un premier temps, sur la place occupée par le concept de « nature » dans les systèmes philosophiques d’Emmanuel Kant et de Johann Herder, afin de retrouver son fondement subjectiviste et les limitations qu’il imposait, à l’époque, à la construction de l’explication séculaire de l’esprit. Puis, dans un deuxième temps, nous illustrons quelques transformations importantes de la structure de la logique, tant dans la théorie biologique et psychologique du XIXe siècle que dans la psychologie et la sociologie du XXe siècle, ayant conduit à l’exploration systémique de la cognition à travers le concept de « psychogenèse ». Ces éclaircissements instaurent la possibilité, selon nous, d’envisager la construction de modèles scientifiques ayant comme but, entre autres, de surmonter l’opposition « nature/culture » qui affecte de nos jours la théorie de la connaissance. / This work proposes a set of socio and psychogenetic analyzes, based on the sociologies of knowledge of Norbert Elias and Günter Dux, on the consequences that arise from the use of primary causes in the explanation of the phenomena of the world. Our challenge, in general terms, has been to make visible that the structure of logic, as a product of psychogenetic processes, has led in the past of our societies to invest primary causes of subjective characteristics, when it has not been reorganized into systemic terms by modernity’s sociogenetic transformations. The concept of "nature", forged and used by philosophy and naturalism in eighteenth-century Germany, is a clear example. We illustrate through our analyzes that, due to the structure of logic, "nature" was "creative" in the world view of the members of the German society of the time, that is to say, an entity with intentionality and ability to act.At the crossroads of historico-genetical epistemology and the sociology of knowledge, the inquiry begins with the place of the concept of "nature" in the philosophical systems of Emmanuel Kant and Johann Herder, in order to find its subjectivist foundation and the limitations it imposed, at the time, to the construction of a secular explanation of the mind. Then, in a second step, we illustrate some important transformations in the structure of logic, as much in the nineteenth-century biological and psychological theory as in the psychology and sociology of the twentieth century, leading to the systemic exploration of cognition through the concept of "psychogenesis". These clarifications establish the possibility, in our opinion, of considering the construction of scientific models whose aim, among other things, is to overcome the opposition "nature / culture" which affects today the theory of knowledge.