Chicago under revision

Discussion
The interaction that was revealed by the ANOVA can be attributed to the fact that the Russian learners made a bigger average gain than the Kazak learners in going from the Listening Only task to the Printed Form task (mean difference = .89 anomaly detections, t[70] = 2.7291, p > .01). This may reflect a bigger difference for the Russian learners than for the Kazak learners between their ability to apply metalinguistic knowledge to the Printed Form task and their tendency to react to the anomalies that present themselves as listening comprehension-cue clashes in the other two tasks. Perhaps due to the pedagogical concern related to this area (Rifkin, 1995), Russian learners would have done a larger amount of pencil and paper activities in some ways similar to the Printed Form task than would Kazak learners. "The endings" are a constant matter of concern (and often, frustrations) to instructed L2 Russian users. The difference may also reflect different levels of reading fluency between the two groups, since extensive reading may be more characteristic of L2 Russian users than L2 Kazak users. In any case, this difference is not particularly relevant to the questions of this paper, since the Printed Form task is likely to heavily invoke nonlinguistic cognitive strategies (Davies and Kaplan, 1998). Given the much longer history of L2 Russian pedagogical methodology, it would not be surprising to find that L2 Russian users tend to be more facile in their L2 metalinguistic reasoning than L2 Kazak learners.
At any rate, prediction 4) in the introduction was not supported. For the L2 Russian participants in the first experiment of Thomson (2000) and the L2 Kazak participants in the present experiment, there is not a significant difference in levels of sensitivities to the particular case-marking anomalies that were investigated. In actual fact, the mean number of anomalies detected in the listening tasks was slightly higher for the L2 Kazak participants than L2 Russian participants, but the difference was nonsignificant (and that remains true when the low groups and high groups are compared separately). In addition, L2 Kazak users showed a substantial advantage over L2 Russian learners with respect to certain items. The two anomalies which were the most detectable for both the Kazak and Russian experiments in the Dual task (instrumental-for-nominative and instrumental-for-locative), were reacted to more often in Kazak than in Russian (37% vs. 15% for the former and 31% vs. 22% for the latter). The possible advantage of L2 Kazak over L2 Russian in those two items is also seen in the Listening Only task (77% vs. 57% and 60% vs. 43%, respectively). This difference could have resulted in part from the greater phonological salience of the Kazak instrumental form in relation to the Russian form. The Kazak suffix receives regular stress marking, while in Russian, stressed syllables were deliberately avoided as locations for anomalous inflectional forms.
The expectation that the simpler form-function relationships of Kazak, in contrast with the more complex form-function relationships of Russian, would result in easier acquisition was based on a tacit assumption of other things being equal. However, were Kazak more difficult to acquire than Russian for other reasons, this could also delay the acquisition of inflectional form-function relationships.
The present negative findings do not, of course, rule out the possibility that the case system is acquired more readily in L2 Kazak than L2 Russian. A more carefully focussed experiment, with a larger number of examples each anomaly type in more precisely controlled contexts, and in a greater variety of contexts, with a larger sample of L2 users who are better matched as to developmental levels, learning histories, and language use profiles, could yet reveal such an advantage of L2 Kazak over L2 Russian. Nevertheless, in view of the current results, it seems unlikely that future research will show that case is acquired radically more readily in Kazak than in Russian. The overall picture of the two L2 groups is too similar, and there is simply no hint that sensitivity to case inflection is readily and robustly acquired in L2 Kazak, as was predicted. In the Kazak experiment, as in the Russian experiment, almost no participants detected a large portion of the anomalies in the Dual Task condition, which strongly distinguishes L2 participants from native controls. As in the Russian experiment, so in the Kazak experiment, there is a significant difference between the high and low groups, indicating that sensitivity to case form and function does increase over time.¹⁹ As in the Russian experiment, so in the Kazak experiment, the increase in reactions under the decreased processing demands that charactere the Listening Only condition further suggest that acquisition may be going on, but that the results of that acquisition are not very robust.
Arguably, part of the reason for the negative findings in relation to hypothesis 4 hypothesis could reside in the very factors that lie behind the positive findings in relation to hypothesis 5. There does appear to be a partially similar overall pattern of development of sensitivity to case anomalies. There was 66% shared variance between the two experiments in the reactions to case-marking anomalies in the Dual Task, and 59% shared variance in the Listening Only task. In both listening tasks in the Kazak experiment, some of the items (instrumental-for-nominative substitution, instrumental-for-locative substitution, nominative-for-locative substitution, nominative-for-instrumental substitution and the subject agreement anomaly) display relative detectability which appears to be analogous to that of the corresponding Russian items in one or both of the listening tasks.²⁰ Overall, then, we are inclined to believe that there are common forces at work in the two languages, accounting for the similarity in the pattern of acquisition, and possibly in turn helping to account for the lack of advantage of one language over the other.