significant differences for any of these contrasts. She also found no significant rank order differences between the oral and written versions, despite the fact that performers clearly had more time for the written version. (In Krashen, Sferlazza, Feldman, and Fathman, 1976, we had attempted to run a written version of the SLOPE with our subjects but could not analyze the data due to ceiling effects; we noted a clear rise in III singular rank order, however.) Are Fuller's results consistent with the hypothesis that learning is brought out only by a discrete-point test?
The SLOPE test does ask the subject for an item in a slot, and to that extent it has some of the characteristics of a discrete-point test. A SLOPE item might look like:
Here is a ball. (Picture)
Here are two _______. (Picture)
This differs in a crucial way from the example given above from Larsen-Freeman's test. In the SLOPE the subject contributes the entire item, while for many items in Larsen-Freeman's writing test the subject need only inflect a given form. This may focus the subject more clearly on form. Whether this is the crucial difference remains to be seen and can easily be tested.
Also, these results may not hold for foreign language as well as second language performers. Generally, foreign language students have less access to language acquisition and rely more on learning. It would not be at all surprising if foreign language students show a greater learning effect, manifested by more "unnatural orders". Adult EFL students in other countries might provide the crucial data here.
Objections to the "Natural Order"
1. Is the natural order an artifact of the Bilingual Syntax Measure?
There have been some critiques and objections to this seemingly clear and consistent picture. First, several scholars have suggested that the adult natural order is an artifact of the use of the Bilingual Syntax Measure (Porter, 1977; Hakuta and Cancino, 1977; Rosansky, 1976). This suspicion was based on the following circumstantial evidence: