Perceptive-auditory perception of the material of different kinds: experimental research

The article describes experiments carried out at the Department of Applied and Experimental Linguistics for a prolonged period under scientific supervision of Ph.D., Sc.D., prof. Rodmonga Potapova. Problems solved in the course of the research as well as the most conspicuous results of auditory perceptual experiments are presented. An annotated database of Russian spoken language containing phonograms of reading and speaking (monologues, dialiogues, and polylogues) formed in MS Access DBMS was used, which ensured the most efficient data representation and logic of the research. The database consists of several tables containing information about the speakers (N=72), namely, biographical data, physioanthropological characteristics, information on individual speech peculiarities, and lists of phonograms containing the speech of each speaker. Auditory perception of the said materials under different listening conditions (white and pink noise, different signal to noise ratios) was analyzed. The experiments involved different groups of auditors, all of them filled special questionnaire before taking part in the research, so that all peculiarities of their perception could be taken into account. Difference in perception of phonograms that included reading and speaking, monologues, dialogues and trilogues were analyzed, probability of perception of different spoken language units was estimated. Methods of analysis and assessment of the data obtained are proposed, interpretation of the results is given. Special attention is paid to semantic perception of texts that the auditors recovered basing on the speech units perceived, which included also the detection of the topics of the texts. For the assessment of the latter data a special scale was devised which included overall correctness evaluation, possible modifications and substitutions. Alternative topic formulations proposed by the auditors were carefully analyzed. While interpreting the data on speech perception in white noise, topic formulations obtained were grouped into clusters, analyzed from the point of view of their formal structure and parts of speech used. The assessment of semantic perception of the recovered texts was performed using the approach proposed by T.M. Dridze, which proved efficient with regard to the specific material used in this research and allowed to estimate what the auditors perceive as the most significant information.