Back to Work

After a couple of months of teaching Albanian at Arizona State's Critical Language Institute, I am back on my dissertation. Teaching was a great experience, and it certainly helped me tune up my Albanian, but now it's time to turn my full attention back to writing. First item of business is to finish the writing on the vocabulary chapter (Ch. 3) that I put together for a presentation at the University of Chicago. Here is a link to the handout that I made for that presentation. Right now, I am writing the handout into an article, perhaps to be published as a part of the proceedings from that conference, perhaps as an article somewhere else.

I wish that I could just take this article as a full chapter in the dissertation, but I need to modify it in a couple of ways: expand the scope of the territory covered in the article. The two main sources that I have relied on so far for borrowings from Slavic into Albanian (Gunar Svane, Slavische Lehnwörter im Albanischen, 1992; Xhelal Ylli, Das slavische Lehngut im Albanischen, 1997) disregard contact between Slavic and Albanian in the former Yugoslavia. That is why I will be looking for more work on the matter in the next couple of days.

A second modification (which is also the case with the chapters on morphology and phonology) is that I need to take some of the introductory material about sociolinguistic setting and theoretical discussions out of these individual pieces and combine them with the material in the first two chapters that give those backgrounds to the entire dissertation.

But first, to finish the article at hand...

Turning to vocabulary

Now that I have written my presentation from the Midwest Slavic Conference into my dissertation I am turning to the next presentation: on vocabulary shared between Slavic and Albanian, to be given at the Graduate Slavic Conference in Chicago in two about two weeks.

I have already started some of the research on the topic, as I thought I was going to be presenting on this topic at the Midwest Slavic conference before moving things around. Now it is time to get back into the topic.

Midwest Slavic Conference

Yesterday I presented my research at the Midwest Slavic Conference here in Columbus, OH. This is a fun conference that a lot of the younger scholars present at. I have presented at it for seven years in a row now.

This presentation was "Semantics and Tense in Western Macedonian Dialects." Here is a link to the powerpoint presentation.

Here is a summary of my presentation. (As soon as I finish this post, I will be trying to write up my presentation to incorporate it into the chapter on the perfect for the dissertation.)

My research question is "How do semantic factors influence the choice of tense for speakers in western Macedonia? In particular how are perfect forms selected by speakers?"

I am using data from my questionnaires that I passed out to 58 students in Tetovo and Skopje Macedonia. Using the results from these, I analyze how the results differ from one dialect to another. The conclusions are still tentative, but here is what I found out.

Southwest dialects are innovative in the use of the HAVE + N/T participles, but the Skopje dialects are not far behind. Some other areas are much more conservative: Kičevo and the villages of Tetovo, in particular. The fact that the villages in Tetovo are more conservative than the city itself, combined with the fact that Skopje is quickly incorporating the HAVE perfect form into its speech leads me to believe that the HAVE perfect is being spread more through the influence of the standard language than from contact with dialects or even with other languages. That said, however, there is still more to be learned about its use in the dialects, and its likely that both the standard language and the dialects are exerting some influence on its spread.

One further note on the HAVE perfect is that its use for various semantic contexts may not be as straightforward as Bužarovska and Mitkovska 2010 have claimed. They follow the claim given in Bybee et al, 1995, that anteriors or perfects develop through a set pattern of showing up first in contexts where the subject becomes the possessor of the result of the process involved, as in the example below taken from my questionnaire (modified from the Perfect Questionnaire in Dahl 2000):

1. [I was told that you are writing a book. How many pages you _________________ by now?] (WRITE)

Answer: I _________________________________ fifty pages. (WRITE)

That is, give this context speakers are likely to use a grammatical construct of a perfect using an auxilliary of HAVE plus a participle of the verb WRITE (In English, have [you] written? --> I have written fifty pages). Bybee et. al.'s reasoning is that these type of constructions can start with the HAVE being a main verb - that is something like the idea in: how many pages do you have (that are written) by now. It is just a short shift in meaning from this to the idea behind How many pages have you written? And, as they point out there are many languages that can use a construction with HAVE in these types of contexts, even when they can't be used in other places. Some Slavic languages can be characterized in this way, as in Serbian Koliko stranica imate napisano? My good friend Motoki Nomachi (2006) wrote about this, as have other scholars. And, in general, in Macedonian these are the types of constructions that most consistently give the perfect with the HAVE auxiliary (Bužarovska and Mitkovska 2010, as well as from my results). However, in a couple of the dialects a different type of context gives the HAVE perfect more consistently, that of Experientials (statements of having experienced something):

1. [Question: Can you (= any person) swim in this lake?]

Answer: Yes, at least I _________________________________ in it several times. (SWIM)

Specifically, in the results from villages of Tetovo, these types of contexts give results with HAVE perfects than the Possessive ones. I don't know why that is the case; perhaps it is something to look at in the future. But this may be a small but significant piece of information because there is one theory that says that languages develop these categories in the same way everywhere, from Resultatives to also having Possessives to also having Experientials. This theory "Grammaticalization" is very popular and predominates investigations into the history of perfects. The counterargument is that there is not always a unidirectional development of these phenomena. I haven't stepped into the argument, but my work will give credence to one theory or another. After more analysis, I will have to come back to this question...

Two other conclusions emerge from the data, both corroborating the work of previous scholars, first is the observation that the BE + L-perfect is primarily involved in statements involving indirect or nonwitnessed evidentiality (see previous post). My data shows that this is applied fairly consistently across the dialects of western Macedonia. There is some variation in how strictly this is implemented, but it is fairly consistent. Finally, as regards the BE + N.T participle, it appears that his is most commonly found in Resultative contexts, so that as Elliott 2001, etc. have argued, it might be best to consider this not as a perfect, but as a simple resultative. In my data, about 80% of all uses of the BE + N-T perfect are in these contexts, the bulk of the remaining are found in narrative contexts which are prototypically not perfects.

Category 1: Evidentiality

One of the relevant semantic categories I am analyzing perfects with is that of evidentiality.

Evidentiality is concerened with how a speaker knows about a certain event. In some languages, such as Turkish, Bulgarian, Macedonian, Arumanian, Albanian, Georgian, and Estonian (and probably many more!) the grammar requires the speaker to either confirm the truthfulness of what they are saying or at least indicate whether or not they have first-hand knowledge of something.

In Macedonian this distinction is usually made by contrasting two different grammatical explanations. For first hand knowledge (and by consequence the speaker's attestation of the event's actual occurence) a simple past is used as in the following example:

1) Macedonian confirmed past with imperfect

Toj beše vo Skopje.

He was-3sg.imp in Skopje.

‘he was in Skopje’ (I vouch for it)’

(Lunt 1952:93; see also Friedman 1993:272)

Whereas, speakers can distance themselves from the responsibility of vouching for the

occurence by using a perfect form made up of a 'to be' helping verb plus a past participle,

as in this example, below:

3) Macedonian nonaffirmative past with be + l-part

Toj bil vo Skopje.

He was-l-part-masc.sg in Skopje.

“He has been in Skopje.” or “He is/was in Skopje (apparently) / (much to my surprise) / (supposedly).”

(Elliot 2001: 19; see Friedman 1993: 272, Lunt 1953)

This particular distinction shows a lot of variation in Macedonian dialects, so it is of interest in my research. So far, my results show that in contexts where the event in question was not witnessed, this type of construction (the L-Perfect) was much more common, but other verb forms were also used. This may reflect differences in dialect or there may be further subtleties of the semantic sensitivities in the language that other categories also influence.

To cover all of the types of evidentiality in the survey,

I gave four different values for the contexts in the questionnaire: 1-witnessed, 2-not witnessed, 3-not likely to have been witnessed (where the context is not entirely clear as to whether or not there was a visual witness on the part of the speaker) and 4-irrealis (not pictured here), where the event is either hypothetical or in the future.

This analysis is based on only a part of the data so it is still incomplete. However, given the differences between 1 and 2, it shows that this is a really important factor in choosing which grammatical forms to use in Macedonian.

Perfect Semantics

I'll admit it, I use the word perfect way too much in my dissertation, but that is because most of my research is on the grammatical categories talked about as perfects. I have been working on perfects for some time.

Semantics, on the other hand, is something that I have been starting to focus in on, just recently. In this post, I want to talk about some semantic variables that are important for perfects, and are the basis for the analysis I will be talking about next week at the Midwest Slavic Conference.

This post will also comprise some of what I will say on that occasion, as I am finding it a little hard to put my thoughts together anywhere else.

Perfects are grammatical categories that typically show some relation between a current situation and something that has occurred in the past, as in I have read that book. Some scholars talk about the perfect in terms of "Current Relevance" like Bernard Comrie (1976); that is to say that the perfect is used when a speaker wants to make a past action relevant to the situation being spoken about, such as the context similar to one given in a questionnaire that I used in my research:

A conversation between a friend and me. My friend (not knowing my sister, who actually is a much more accomplished reader than me - and most of my friends) says "It seems to me that your sister never finishes the books she reads." So I say (defending my sister's reputation) "That's not true, she has read that one."

Another way scholars have talked about the perfect is as a form that expresses an "Extended Now" that is, things that have been valid about a particular past event are still valid, so in the example above, it is not the fact that we are talking about reading that influences me saying has read instead of read, but rather the fact that her having read the book is still true.

Talking about perfects gets quite technical quite quickly, so I'll stop there for this post. Now on to something else that gets really technical really quickly: semantics.

Semantics, when talked about technically in linguistics has the meaning of talking about meaning. One part of semantics that is relevant in talking about perfects is the characteristics of individual verbs, such as whether they show a change of state (Telicity), a change of location (Motion verbs), whether they are repeatable (Iterative), as well as whether they require grammatical subjects and objects (transitive vs. intransitive & active vs. non-active or passive). One other part of semantics is shown in the context in the sentence. Is it something that is repeated (Iterated) is it something that has occurred in the past or is ongoing (continuous), Is it something that happened at a particular point in time (Punctuality, like at 3:00, or on January 16th, 1987), or lasted for some time (Durativity, like for three days now). A couple of other distinctions will be used in my particular research, drawing on characterizations that have been made, both in Linguistics in general, and in literature about these languages specifically.

Macedonian has three different forms that look like perfects. Graves 2000 is one of the best attempts to show how they function differently in expressing semantic differences in Macedonian. Her work is a great start in exploring the question of how the perfect forms are used in response to a variety of semantic contexts. However, there is more to be studies, as she only used a handful of surveys, and had only one speaker from Western Macedonia where there is more variation as regards the perfect constructions.

Using a very similar survey and including responses from about 60 respondents from Western and Central Macedonia, I hope to give an even more complete picture f the use of these perfects in Western Macedonia. Ultimately, I will compare this to patterns found in Albanian and Serbian dialects, but in and of it self, the temporal effects of semantics in Western Macedonia alone is worth pursuing.

The following are the semantic criteria that I am using in this analysis:

1) Evidentiality

2) Pragmatic purpose of sentence

3) Telicity

4) Repeatability

5) Iterativity

6) Criteria from the Perfect Questionnaire

7) Time reference

8) Sentence type

(I may rearrange these to make sense of what I have!)

First R graphs made

Well, it pretty much took the whole working day to do it, but I have produced the first graphs from using R. I'll have to work on the statistics a little later, but here is what the graphs look like at this point.

Obviously, gray isn't the most captivating color. I still need to work on the presentation, too. Maybe when they're completed I'll put more up.

April 1, 2011

It has been a couple of days since I have written, so I will briefly catch up with where I am at in my research. I finally made it through the introduction of Gunnar Svane's book on Slavic Loanwords in Albanian. That is a great book that I plan on using profusely as I continue writing about the vocabulary borrowed across Slavic and Albanian.

Because I don't know for sure whether I will be presenting on vocabulary or semantics of perfects at the Midwest Slavic Conference, I am trying to work on a little bit of both. Yesterday I was able to figure out how to load my table onto R and have started compiling the numbers from that. Hopefully by the end of today I can put together numbers and a graph or two from one of the categories: time referencing adverbials. If i can get that up, I'll try to post about what I find out next time.