ETD system

Electronic theses and dissertations repository


Tesi etd-02122009-145856

Thesis type
Tesi di laurea specialistica
Creating Italian FrameNet: a Frame Semantic Study on Verbs of Visual Perception
Corso di studi
Relatore Dott. Lenci, Alessandro
Parole chiave
  • cognitive studies
  • Italian verbs of visual perception
  • polysemy
  • frame semantics
  • lexical semantics
Data inizio appello
Data di rilascio
Riassunto analitico
The subject of this work is a study of Italian verbs of visual perception, which I have analyzed using the theory of Frame Semantics developed by Chuck Fillmore .
The main tenet of Frame Semantics is that linguistic entities such as words, idioms, and grammatical constructions (at least in some cases) evoke frames in the mind of language users when they are heard or spoken. A frame is an abstract conceptual schema of a situation or event, which features a series of participants or semantic roles called Frame Elements.
The ideal goal of Frame Semantics is therefore to individuate all the frames that are evoked by the words in a language; this goal then needs to be contained, since frames can be individuated at many different levels of granularity, so it is necessary to determine how much detail about the semantics of the words one wants to represent. This goal has been pursued for English through the FrameNet project at the International Computer Science Institute at Berkeley; the FrameNet research group has by now individuated more than 825 frames, based on more than 10.000 lexical units.
My goal in this project has been to create a sort of “bootstrap” for an Italian FrameNet analogous to the English one, by analyzing a small section of the Italian lexicon, i.e. a small number of verbs of visual perception. Since English FrameNet already features a series of frames related to perception, I decided not to create the frames ex novo for Italian, but rather to conduct a frame semantic analysis of the verbs analogous to the one conducted for English (i.e., analyzing argument structure, semantic roles, and syntactic patterns occurring with these verbs) and then to verify, on the basis of that data, whether the perception-related frames developed for English were applicable to Italian, as well. As will emerge from my discussion, this was not always true; I sometimes had to modify the frames (or add specific Frame Elements) in order to adapt them to Italian verbs.
The verbs I have selected for analysis are avvistare (sight), intravedere (glimpse/make out), notare (notice), osservare (observe/watch), sbirciare (peek), scorgere (glimpse/spot). The reasons why I have chosen not to analyze vedere (see) and guardare (look), the two most important verbs of visual perception in Italian, are the following. First of all, practical considerations led me to choose verbs with a slightly lower frequency than vedere and guardare for my first foray into the intricacies of semantic annotation: perhaps the most delicate part of the annotation process in FrameNet consists in selecting a sample of corpus examples that is representative of the syntactic distribution of the word one is analyzing, and this process is less difficult when the word is not highly frequent. Secondly, vedere in particular is characterized by an extremely intricate polysemy with a high number of possible senses, and the risk in undertaking the semantic analysis of such a verb was that I would be forced to concentrate on it alone, while my main goal was to create a varied sample of annotated lexical units that could serve as a starter for an Italian FrameNet project. Finally, I have noticed that a very large part of the scientific literature on visual perception verbs concentrates on see and look in English, and vedere and guardare in Italian, while the verbs which specify the manner of perception more closely are rarely taken into consideration. I therefore hope to contribute some new information to the literature on verbs of visual perception with this analysis of the semantics of these little studied perception verbs.
The structure of this work is as follows: in Chapter 2, I discuss the methodology I used in order to conduct my frame semantic analysis, and I describe the FrameNet annotation process in detail. I also describe the frames developed for English that are relevant for my analysis, and their semantic relations. In Chapter 3, I present an in-depth discussion of the semantics of the FEs belonging to the three perception-related frames most relevant to my analysis: Perception_experience, Perception_active, and Becoming_aware. In Chapter 4, I present a more specific analysis of the semantics of some visual perception verbs as it emerged from my study of their frame structure. Chapter 5 features some conclusions and proposals for further research.