Index of Second Language Learning corpora

These corpora contain the sets of stimuli used in the experimental translation task described in . Julie Van Dyke's M.Sc. Thesis . The stimuli consist of pairs of sentences where the first sentence stands for the context and the second sentence is the sentence to be translated within that context. For example:

Context: There are sometimes scary people in the train station.
Sentence: A criminal was seen in the train station.
There is more than one context for each sentence to translate:
Context2: The bank was robbed by a tall, dark-haired man.
Sentence: The criminal was seen in the train station.
The contexts are set up with two goals in mind:
  1. Require different choice of determiner the English translation (as in the above example).
  2. Require a different word order in the Polish translation (as in the following example).
 
Context: Two boys were sitting in the cafe.
Sentence: A boy left after 5 minutes.
Transl:   Po 5 minutach wyszed= ch=opiec.

Context2: There was a boy sitting in the cafe.
Sentence: The boy left after 5 minutes.
Transl:   Ch=opiec wyszed= po 5 minutach.

There are four groups in the experiment. The long version of the set of stimuli can be viewed by clicking on the underlined words each of the following group descriptions: Reversed sentences means that if a sentence appeared in the first version in English it appears in the second version in Polish.

These were the sets of stimuli I gave to the Polish students learning English. I gave smaller sets of the same stimuli to the English students learning Polish. The difference between these lists is the elimination of context-sentence pairs which seemed confusing to the Polish learners or which did not fit the goals above well. It is the data from these shorter lists that I will focus my analysis on. They can be viewed by clicking on the underlined words in each of the following group descriptions:

The pairs with two numbers show the pair number in this set (left) and the number in the longer version (right).

The actual coverage notes can be found here.

Maintainer: vandyke@cs.cmu.edu (Last updated 95-09-07)