Developing a precise argument is not an easy task. In real-world argumentation scenarios, arguments presented in texts (e.g. scientific publications) often constitute the end result of a long and tedious process. A lot of work on computational argumentation has focused on analyzing and aggregating these products of argumentation processes, i.e. argumentative texts. In this project, we adopt a complementary perspective: we aim to develop an argumentation machine that supports users during the argumentation process in a scientific context, enabling them to follow ongoing argumentation in a scientific community and to develop their own arguments. To achieve this ambitious goal, we will focus on a particular phase of the scientific argumentation process, namely the initial phase of claim or hypothesis development. According to Toulmin (1986), the starting point of an argument is a claim, and also data that serves as a basis for the claim. In scientific argumentation, a carefully developed and thought-through hypothesis (which we see as Toulmin's ``claim'' in a scientific context) is often crucial for researchers to be able to conduct a successful study and, in the end, present a new, high-quality finding or argument. Thus, an initial hypothesis needs to be specific enough that a researcher can test it based on data, but, at the same time, it should also relate to previous general claims made in the community. We investigate how argumentation machines can (i) represent concrete and more abstract knowledge on hypotheses and their underlying concepts, (ii)  model the process of hypothesis refinement, including data as a basis of refinement, and (iii) interactively support a user in developing her own hypothesis based on these resources. This project will combine methods from different disciplines: natural language processing, knowledge representation and semantic web, philosophy of science and -- as an example for a scientific domain -- invasion biology. Our starting point is an existing resource in invasion biology that organizes and relates core hypotheses in the field and associates them to meta-data for more than 1000 scientific publications, which was developed over the course of several years based on manual analysis. This network, however, is currently static (i.e. needs substantial manual curation to be extended to incorporate new claims) and, moreover, is not easily accessible for users who miss specific background and domain knowledge in invasion biology. Our goal is to develop (i) a semantic model for representing knowledge on concepts and hypotheses, such that also non-expert users can use the network; (ii) a tool that automatically computes links from publication abstracts (and data) to these hypotheses; and (iii) an interactive system that supports users in refining their initial, potentially underdeveloped hypothesis.