The present paper analyzes the relationships between vocabulary knowledge, metacognitive monitoring and reading comprehension in L1 and L2 reading in upper elementary school. The relationship between vocabulary knowledge and reading comprehension is presented through the Construction-Integration model of reading developed by Walter Kintsch. The analysis emphasizes vocabulary knowledge as the most important element of L2 reading for this age group. The relationship between vocabulary knowledge and metacognitive monitoring is explored through the measures of resolution and calibration, and the results are fitted into the wider context of research on metacognitive knowledge and control. The results support metacognitive word awareness as an important element of vocabulary learning. Statistically significant differences in bias are found between top and bottom quartile of students according to their vocabulary tests scores. The top quartile shows underconfidence and bottom quartile overconfidence in both the 5th and 8th grades. In the discussion of the relationship between metacognitive monitoring and reading comprehension, error detection is emphasized. The importance of simplifying text language in order for error detection skills to emerge in L2 reading of fifth-grade students is found. A developmental line in comprehension monitoring, according to which these skills first emerge in L1 and then transfer to L2, is also found. In both languages, the first to emerge are local metacomprehension skills, followed by global ones. Finally, a question whether L2 reading is a language problem or a reading problem is discussed, with the results firmly supporting the language threshold hypothesis. Educational implications are discussed throughout the paper.