Unit:
Κατεύθυνση ΓλωσσολογίαLibrary of the School of Philosophy
Author:
Lomis Konstantinos
Supervisors info:
Γεώργιος Μαρκόπουλος, Επίκουρος Καθηγητής Γλωσσολογίας, Τμήμα Φιλολογίας, Φιλοσοφική Σχολή ΕΚΠΑ
Αμαλία Μόζερ, Καθηγήτρια Γλωσσολογίας, Τμήμα Φιλολογίας, Φιλοσοφική Σχολή ΕΚΠΑ
Ελένη Παναρέτου, Αναπληρώτρια Καθηγήτρια Γλωσσολογίας, Τμήμα Φιλολογίας, Φιλοσοφική Σχολή ΕΚΠΑ
Original Title:
Υπολογιστική Ανάλυση της Εσωτερικής Αύξησης στα Σύνθετα Ρήματα με Προθέσεις
Translated title:
Computational Analysis of the Internal Augment of Greek Verbs with Prepositional Prefixes
Summary:
This dissertation focuses on the internal augment of Greek verbs with prepositional prefixes. The main hypothesis is that native speakers of Greek confuse, to a great extent, the cases in which the internal augment is necessary with the cases in which it is not.
In order to test the above hypothesis, a computational model was developed using the Python programming language. This model asks the user to input a sentence which contains a verb with prepositional prefix in one of the following verb forms: either the third person singular of the aorist indicative or the second person singular of the aorist imperative. Afterwards, the model informs the user whether their sentence is grammatically acceptable or not. If it is not, the model provides the user with the correct sentence by replacing the verb form with the correct one (it either removes or adds the internal augment).
Following its development, the model was tested with 20 native speakers of Greek, each of which input 10 sentences to the program giving a total of 200 sentences. The model was successful at a percentage of 73% by presenting the grammatically acceptable sentences as correct and by presenting the grammatically unacceptable sentences as wrong. As for the initial hypothesis, the results support it to some extent since 34,5% of the sentences that the users input was grammatically unacceptable, which shows that there is a significant number of native speakers of Greek who do confuse the cases in which the internal augment is necessary with the cases in which it is not.
Apart from testing the initial hypothesis, the computational model was also developed as a tool to be used for educational purposes. More specifically, it can be used by children in the school environment or by adults who are learning Greek as a foreign language in order to help them distinguish the cases in which the internal augment should be included and the cases in which it should not.
Main subject category:
Language – Literature
Keywords:
computational linguistics, natural language processing, computational model, internal augment, prepositional prefixes, imperative, education, Python
