The issue of error correction remains controversial in recent years due to the different positions of interface toward implicit and explicit knowledge of ESL learners. This study looks at the impacts of implicit corrective feedback in the form of recast on implicit and explicit knowledge of adult ESL learners. In an experimental study, lower-intermediate learners first were taught the grammatical features; then they completed communicative tasks during which the experimental group received recast and the control group received no feedback when an error occurred. Acquisition was measured by means of tests designed to measure implicit and explicit knowledge. Results of ANCOVA analysis revealed higher score for the experimental group; and result of t-test revealed that recast has significant effect on implicit knowledge. In line with the weak interface position toward implicit and explicit knowledge, the findings extend empirical support for Schmidt's noticing hypothesis and function of recast in language learning.