Cross-linguistic regularities and learner biases reflect "core" mechanics.

We suspect that the fact that this difference was marginally significant, as opposed to fully significant, may have been due to a sample size problem in addition to the fact that the measure of the number of characters is somewhat noisy (Experiment 4 below addresses this problem by using a less noisy measure of word length and indeed finds a significant pattern of results for mechanical vs. non-mechanical items in non-Indo-European languages).

also in 26,082 other papers