Browsing by Author "Duran, Daniel"

Now showing 1 - 1 of 1

Open Access
Computer simulation experiments in phonetics and phonology : simulation technology in linguistic research on human speech
(2013) Duran, Daniel; Dogil, Grzegorz (Prof. Dr. phil. habil.)
One goal of this thesis is to give a representative overview of computer simulation experiments in phonetics and phonology. A number of research disciplines are identified as being relevant for the subject of this thesis which are concerned with human speech perception and production. Among these are in particular computational psychology and cognitive sciences. Important methods and research approaches using simulation technology for the study of human speech can also be found in natural language processing and speech signal processing as well as in areas of computer science such as artificial intelligence and machine learning. Due to this interdisciplinary breadth of the topic, this thesis comprises a comprehensive overview of representative publications. Chapter 2 provides a short historical and a detailed thematic overview of the literature and presents the various research questions which have been studied with computer simulation experiments. First, works are addressed which focus on sound systems, their emergence and evolution. Various models of speech perception and production are presented. Computer simulation experiments allow the investigation of phenomena which are empirically hard to capture or even totally inaccessible. Additionally, works on language change and historical linguistics are discussed. The development of a language system over many generations is only hard to capture empirically. The origin of human language seems completely inaccessible to empirical inquiry. The overview of existing work shows how different research questions can be addressed by the same methods and how the same research questions can be addressed by a variety of different methods. An overview of often applied methods and basic concepts is therefore given separately in chapter 3. These are especially methods and foundations of machine learning, artificial neural networks and principles of self-organisation. The second part of this thesis presents a series of my own computer simulation experiments and discusses relevant work. Chapter 4 presents experiments on speech segmentation and addresses the question how humans can achieve this faculty based on the available information. A central problem in this context is that of boundariness. This term is used to denote the continuous nature of segmental information at any position in the speech stream. An important problem of speech segmentation models which is emphasised is the evaluation of computer simulations. Different objective quality measures are compared and their properties are investigated. Apart from conventional quality measures, I propose two new measures and additionally investigate two quality measures which so far have been applied to text segmentation. Chapter 5 addresses the question of how speech units can be grouped into categories once they are extracted from the segmented speech stream. Methodologically, the presented experiments are based on cluster analysis. The focus of the experiments in chapter 5 is the investigation of acoustic speech recordings in comparison to corresponding articulographic recordings. In the first experiment, the linguistic information contained in different segment classes is investigated. Vowels and consonants show different results for acoustic and articulatory data. Vowels in different syllabic contexts are investigated in the second experiment. The experiment shows that contextual information can be found across syllable boundaries. Chapter 6 addresses the question of how the Context Sequence Model can be expanded by adding an articulatory modality. Relevant work on exemplar theory and speech perception and production models are discussed. The experiments take on a new approach in which a production model treats articulatory recordings in the same way as acoustic recordings. This dissertation offers a comprehensive overview of a promising approach to the study of spoken language. The present thesis is as such the first of this kind in phonetics and phonology. Computer simulation experiments can not only complement laborious empirical studies and replace them to some extent. Simulation technology also offers the potential to directly study theoretical models and the interactions of their components. This allows for testing of existing hypotheses and the generation of new ones.