|
Search the site
|
|
Alex Omar Bratkievich(Linguistics major)
“The Existential/Canonical Alternation in Brazilian Portuguese: A Perspective from Optimality Theory”
Sponsor: Professor Line Mikkelsen, Linguistics
Project Description
There are strong indications that the factors influencing the alternation between existential ("There's a book on the table") and canonical ("A book is on the table") constructions might be the same cross-linguistically; for example, existentials across languages exhibit the definiteness effect: indefinite Noun Phrases are preferred in pivot (post-verbal) position. Mikkelsen (2002) proposes that the effect is a consequence of constraints governing the subject position. Since the definiteness effect can be overridden, she suggests an analysis within the Optimality Theory framework to model constraint interaction relative to a hierarchy of constraints. As part of a joint Stanford-Berkeley group research project working on several languages, Alex's project will focus on Brazilian Portuguese in order to establish the distribution of subject and pivot NPs based on analysis of naturally-occurring data, and to propose the ranking of constraints that could explain such a distribution.
Scholar's Photo
 |
|
After a successful journey to Boston, the real work begins as Alex sits back in his chair, pulls out his highlighter and goes into thinking mode.
|
Scholar's Journal
June 20, 2005: The search begins
I’m leaving for Boston in 5 days. I’ll be at MIT for the Summer Institute of the Linguistic Society of America. In the meantime, while I'm still in Berkeley, I've set aside all readings (I can definitely do that in Boston) but I'm working hard on the corpus searches. It's amazing how much I'm learning of Portuguese and of my own language, too, things that I have taken for granted or not given much thought. Some of the things are little assumptions that one makes on the spot, such as assuming that a sequence adverb-adjective-noun is unlikely to happen; an adjective that is too long or another type of adjective-type modifiers usually follow the noun both in Spanish and Portuguese. And although it's true that la más bella muchacha de mi pueblo sounds a little stilted or overly poetical, when I actually looked for such sequences in the corpus of the A Folha de São Paulo, I found quite a few of them –6329, to be exact, although I'm assuming that an important number must be false hits. Half of these are sequences of the type mais (more) + adjective + noun.
This means that I am learning a lot about the structure of noun phrases in Brazilian Portuguese. Of course, the downside is that if I want to be more inclusive and get most of the cases that I can, I need to keep modifying my searches.
Finally, the website through which I can access the corpus uses Perl for its searches. Did I mention before that I'm not computer-savvy? I'm definitely learning on the spot.
August 8, 2005: Back home!
I've just arrived from the Summer Institute at MIT. I took a couple of courses in Optimality Theory, which is the theory I'm going to use in my analysis of the data. The courses were quite different in style, the first one an introduction to solving practical problems that you encounter when ranking constraints. The second course was much more theoretical and basically showed us the power of the theory and how to use its formal tools to improve and verify our analyses.
By the sixth week, I was in desperate need of going back. I wanted to go back to my normal life, my home, my routine. In spite of that, the whole MIT experience was amazing! I got to know students from different universities from all over the world and also got in touch with faculty. For me, it was like seeing "linguistic superstars" everywhere. I would just walk to class or to have lunch and there was Alec Marantz, there was Richard Kayne and so on. I even went to a lecture by Chomsky.
In the meantime, I also finished the main round of corpus searches. After all the searches, I had an initial number for each NP-type, how many occurrences with the verb haver, how many with ter (both verbs are used in existential constructions) and how many with estar (in the canonical constructions). So I started analyzing the sentences and I discovered very interesting results.
After going through the 1064 sentences of the type Tem um/uma/uns/umas X em..., I discovered that only 50 are existential sentences. The problem is this: in Portuguese the same verb is used both for saying "He has a house in the country" (Ele tem uma casa no campo.) and "There is a house in the country" (Tem uma casa no campo.) Of course, this fact is very interesting because it’s possible for a sentence to be ambiguous between the two meanings; in fact, there is a theory out there about how existence and possession are expressed in the same way in many languages. But that's definitely out of the scope of this research.
I still have to go through the other sentences, not only to discard the false hits but also to find out how many of the sentences are within an embedded clause, for example, and other information about the contexts in which the sentences occur. This is information that can tell me more about which structures tend to appear with an existential (in some way, what structures are good hosts for an existential) and which with a canonical sentence. The good thing is that the searches gave me big numbers but it is not impossible to analyze each sentence on its own: only 3 searches gave me more than 1000 hits! So I'm able to look at the details very carefully. I’ve got great work for weeks ahead!
|
