The present experiment was conducted to determine whether vigilance and arousal are different functional entities, even though the two concepts are usually confounded. Our hypothesis is that vigilance is a state specifically associated with attentional availability, whereas arousal is a state independent of attention and based on neuronal activation. If this is true, differential effects on performance can be expected. Namely, we propose that arousal essentially affects speed of information processing, whereas vigilance is more linked to accuracy and/or omissions. The sleep inertia paradigm has been chosen to distinguish the effects of low arousal and hypovigilance. We compared performance in either sleep deprived or non sleep-deprived subjects in the Descending Subtraction Test (DST) presented in a complex and simple version. Speed and error indices were the dependent variables. Results showed that speed was deteriorated in both groups during the first 15 min after awakening, but only in the simple task, suggesting a slowing down of mental processing. By contrast, the error index reflecting accuracy showed, in the complex task only, increased errors during the first and last 15-min period in the sleep-deprived subjects but not in the Control Group. All together, these data are in favor of a functional difference between arousal and vigilance, and these two functions could affect performance in specific ways.