Laboratory for the Computational Studies of Language


Welcome to TUWPA:
A tool for experimenting with
Turkish Word Production and Analysis


This is a tool for analyzing and producing Turkish word forms.

There are two kinds of exercises:

This is a self-paced learning tool without supervision. We suggest that you look at the solutions proposed by TUWPA after sufficiently many trials.


What you can do

In A-exercises, you can test your knowledge of affixes and allomorphy. You enter a word, and you are asked to write an analysis of the root and affixes. Your proposed analysis must include abstract representation of affixes. For instance,
you type the word :evlerimde
you type the analysis:ev-PLU-POSS1s-LOC
When you ``evaluate'', TUWPA responds: CORRECT
You may have noticed the use of dashes (-) for affixes, for example -PLU for the plural affix. You don't have to memorize these names; they are provided with the click of a button. Once you learn the names, you can type them (with the dash) yourself.

TUWPA analyzes any permissible root-affix sequence as CORRECT. This includes non-words like zamircek.


In P-exercises, you can check your knowledge of affix ordering and sound alternations in affixes and roots. You select a root and a sequence of affixes (in their abstract form), and try to write a word form for it. For instance,
you type the root-affix order: ev-PLU-POSS1s-LOC
you write the corresponding word : evlerimde
When you ``evaluate'', TUWPA responds: CORRECT


Before you start,

See the notation for typing affixes and words.
The help lines contain brief descriptions of options and some terminology.

Click the button (available in all pages) if the system's response makes no sense or it does not respond at all. This will let you send mail to us; please specify what you typed (and in what context), how the system responded (if at all) etc.


Here are the exercises:


TUWPA credits:
Design team: Izzet Pembeci, Cem Bozsahin, Deniz Zeyrek
Implemented by: Izzet Pembeci
Thanks to: Onur Sehitoglu (implementation)
Financial support: TU-LANGUAGE project
Laboratory for the Computational Studies of Language
Department of Computer Engineering
Middle East Technical University

Number of visitors since August 19, 1998 is Counter