Material inference opens up an exciting prospect for pluralism in reasoning: because the conceptual content of descriptive language is open-ended and variable, legitimate inferences can take unlimited forms. How many forms of legitimate reasoning are there, and how is legitimacy determined? Wilfrid Sellars and Robert Brandom recognise that practices can be legitimate by virtue of their material forms – that is, with respect to their specific vocabularies. Pluralism also finds a foundation in the work of William James and John Dewey. It will be interesting to examine the diverse styles of reasoning found in the history of science, following the models of A. C. Crombie, Nicholas Jardine, and Ian Hacking. A key observation is that historically generated forms provide the standards for reasoning rightly. It is always a social judgement when something is deemed incorrect, even when prompted by external impulses. Legitimate reasoning is constrained by the rules of a community. Intersubjective assessment provides the required and sufficient objectivity. Unfortunately, Brandom is not fully consistent on this point, as he allows for a user-independent means of determining correctness.