It is common in intervention research to use measures of working memory either as an explanatory or a control variable. This study examines the contribution of cognitive abilities, including verbal working memory (WM) and attention, to language performance in first- and second-grade children. We assessed children (N = 414) on two forms of verbal WM, one attention, and two standardized assessments of language. Scores from all three measures of cognitive abilities significantly predicted latent language (64% variance). Both verbal WM measures were stronger predictors of a latent language variable compared to attention. Exploratory analyses revealed differences in the role of cognitive variables to language subdomains. The findings deepen our understanding of the relative associations between verbal WM, attention, and language. We conclude that it is important to consider the language demands of tasks when making decisions about assessment of verbal WM, specifically in the context of intervention research in domains that require language.