1 - 2 of 2
ItemControl in Bangla(University of Hyderabad, 2016-06-01)While exploring the processing differences found in two major types of obligatory empty category (EC) control, this dissertation attempts to substantiate a claim, that while processing any control clause, the language processor in human mind tends to link an argument NP with available event(s) in that sentence, showing an increment in processing load if there is an increment in semantic links between this performer and more given events, within the same sentence. It is claimed here that object control situations provide lesser number of links to process than subject control sentences where there is an increment of such links. This study reports the findings from seventeen selfpaced reading experiments carried out in Bangla (an SOV Language in India), explaining preferences towards object control in terms of increment in the number of AgentEvent Association(s), and its interaction with preferences in the type of nominative or oblique roles that the matrix clause subject (semantically associated directly with a matrix event and indirectly with an embedded event) may get from both these events in Bangla. Increase in number of association links between an agent and any number of events (either from the matrix clause or embedded clause or both) showed a significant effect. Independent preference towards a [+AG] subject was also detected but a variation in agentevent association complexity was found to very strongly override even this preference as well, therefore showing a strong influence of AgentEvent Association load. At the end relevant ungrammatical and implausible sentential constructions are also used as evidence to show how AgentEvent Association load is capable of deciding the grammaticality in a language, clearly indicating that actual, measurable processing constraints have originally shaped certain grammatical restrictions of human language
ItemUnfolding of roads and pipes(University of Hyderabad, 2018-12-01)This thesis reports the study of Non-actual motion descriptions in two sample populations, both of which speak speak Khasi and English as their dominant languages. Non-actual motion refers to the experience of potential motion of static and extended objects such as pipes, roads, wires and fences and the use of motion verbs to describe this property of extension in Language. The thesis explores two aspects of Non-actual motion - the underlying factors that lead to the use of Non-actual motion descriptions and the language resources that are deployed by the sample populations to describe the same, in Khasi and English. The data is obtained through an elicitation experiment, involving a total of 32 images for every participant. The sample populations were tested for the role of enactive motion as a primary motivator for the experience and use of Non-actual motion. The analysis of the behavioural data was done in R Statistical Analysis, using glmer in the lme4 package. The analysis of the linguistic data was done using the Holistic Spatial Semantics framework, along with the the typology of Non-actual motion proposed by Dr. Johan Blomberg. In addition, the motion verbs were also analysed for the kind of information they encode in addition to motion: manner, path and direction. Spatial information as encoded in other parts of speech - adverbs, prepositions, case markers, deictic markers were additionally analysed and compared between the two languages. The results show that the affordance of motion on the objects investigated have a statistical significance in the use of Non-actual motion descriptions, pointing in the direction of having an effect in our experience of potential motion on these objects. The thesis did not find enactive motion to be the primary motivator for the use of Non-actual motion descriptions in both populations, in both the languages. Linguistically, Khasi and English were found to be similar in that they both had very high instances of the use of a bleached motion verb to express motion and to encode path in a satellite. However, Khasi showed a much higher amount of additional information about path as well as body configuration of the objects, which comes under the non-actual movement information, as proposed in Blomberg’s typology. In addition, there is a difference in the most frequent types of verbs used by the participants in their two languages, where they useentry/exit verbs frequently in Khasi but not in English. The thesis proposes further studies in other Monkhmer languages in the future. In addition, it also sees a need for models built for a multilingual population like the ones in South Asia, which would account for language contact, multilingual childhoods and incorporating the notion of South Asia as a Linguistic area, in order to better explain some of the results from the data found in this thesis.