Out of eight KLIP applications that were submitted in response to the call 2013, five
have been selected and are supported by ZüKL with 22'210 CHF.
Below you find a list with short descriptions.
Content:
Linguistic features that are subjectively more noticeable than others are
called salient. Saliency offers explanations
for many linguistic phenomena such as language change, where more salient features
are more prone to change than less salient ones. Project 1 will be a first
investigation of the possibility to establish links between this theoretical concept
and cognitive reality.
To this purpose, the project group will conduct an EEG study using speakers of
Swiss German and pre-recorded Swiss German audio stimuli. Speakers of different
dialects (Zurich, Bern) are presented with recordings from their own and from the
other dialect. The neurological reaction to these stimuli is measured by EEG.
In addition, speakers have to decide whether they are able to localise the
dialect or not.
This pilot study will contribute to a planned application for a SNF project
which will deal with various questions of the theory of saliency in a wider
context.
Content:
Postcards are a well-known genre that forms an important part of modern touristic culture yet so far has received little attention from linguistics. Project 2 would like to change this.
The Department of German Studies hosts a collection of about 6000 picture postcards, which will be converted into a digital corpus in a planned cross-border SNF project. This is a challenging task, for instance, because of the constant switching between Standard and Swiss German in these texts. The finished corpus will provide an important base for investigations into the linguistic particularities of postcards and their place in the history of communication.
In order to make the application possible, much preparatory work remains to be done including literature research and developing research and budget plans. Project 2 will see to these preparations as well as to the writing of the application itself.
Follow up project: SNF and DFG cooperation project "Textsortenentwicklung zwischen Standardisierung und Variation: Das Beispiel der Ansichtskarte. Text- und korpuslinguistische Untersuchungen zur Musterhaftigkeit privater Fern- und Alltagsschriftlichkeit" with Joachim Scharloth (TU Dresden)
Part-of-speech annotation in historical corpora
ID: KLIP-2013-3
Researchers: Marianne Hundt (English Department), Gerold Schneider (Institute of Computational Linguistics and English
Department)
Content:
The English Department has been participating in the compilation of the ARCHER corpus,
a collection of historical texts of about 3.3 mio. words from the period 1650-1999.
In order to make such corpora even more useful for linguists, they are usually
annotated, i.e. additional linguistic information is added to them. One central level
for this are parts of speech.
While the ARCHER corpus has been automatically tagged for parts of speech, taggers
deliver much worse results for historical than for contemporary texts because of the
great number of words that are unknown or look different from their present shape.
Manual correction is therefore crucial for achieving a satisfying result. This is the
aim of a planned SNF project.
The present pilot study will contribute to the application by comparing two taggers
(CLAWS and Tree-Tagger). The main question will be whether manual correction can be
sped up by considering only those cases where the two taggers produced different
outputs.
Inhalt:
Agreement in linguistics refers to grammatical dependencies of the kind where properties of one word in a sentence (e.g. singular/plural of the subject) are reflected by another one (e.g. by the verb). Typological research in this area has traditionally been concentrating on "exotic" non-European languages in order to find unusual or theoretically challenging phenomena.
However, astonishing variation including many unusual agreement systems can also be found in less well-known language varieties of Europe, for instance, in the Italo-Romance dialects. Taking these into account may not only make the typological picture more complete but may also deepen our knowledge on the historical development of agreement systems.
Project 4 will develop an annotation system for agreement phenomena in previously collected texts of six Italo-Romance dialects and will lay the base for annotating this corpus by designing a Filemaker database for this purpose. Annotation and analysis will be carried out later in a planned SNF project.
Follow up project: "The Zurich database of agreement in Italo-Romance" funded by the SNSF (total 666'064 CHF).
Online portal for ZEKO
ID: KLIP-2013-5
Researcher: Elvira Glaser (Department of German Studies)
Content:
The Zurich Electronic Corpus of Language Varieties (ZEKO) is a collection
of transcribed dialect recordings from the German- and Italian-speaking areas
of Switzerland. The corpus has been gradually enhanced, digitised, and prepared
for querying it via a web interface in previous projects.
Project 5 is planning to work on the existing structures in two places.
First, more texts that are stored at the Phonogram Archive will be digitised
via OCR and be added to the corpus. Second, the web interface (presently
not yet open to the public) will be further improved. These steps will also
contribute to a SNF application that will be prepared at the same time and
aims at further increasing ZEKO.
