Children acquiring many different languages have been reported to go through a stage where they use finite clauses and non-finite clauses – so-called Root Infinitives – interchangeably. In this thesis I investigate the Norwegian Root Infinitives, such as Jeg lese denne I read.INF this one , and the Norwegian Root Infinitive stage, using data from 16 files with spontaneous speech from 12 monolingual two-year-olds acquiring Norwegian. The analysis is framed within generative grammar and the minimalist program. I review four different accounts for Root Infinitive in child language (the small-clause hypothesis, the truncation hypothesis, the unique checking constraint hypothesis, and the modal drop hypothesis) in light of the data gathered, and propose a new approach, drawing on the minimalist context-linked grammar of Sigurðsson (2004 and subsequent). Specifically, I propose that Root Infinitives are non-finite clauses, and not finite clauses with something silent or missing. I also argue that there is positive evidence in the child s input that root clauses do not need to be finite, in form of a special Prescriptive Infinitive that exists in the North Germanic, e.g. gjømme seg inni der hide.INF self inside there , which is particularly used in child-directed speech. In my analysis, I give an account for the structure of Root Infinitives and Prescriptive Infinitives, and show that their differences can be derived from their shared structure, making an overgeneralization account for Root Infinitives quite probable. Root Infinitives have also been found in children with specific language impairment (SLI), and verb morphology has been proposed as a diagnostic tool for SLI. I discuss what impact the data and analysis provided in this thesis may have on the use of Root Infinitives as a diagnostic marker of SLI.