This thesis asks what it is that we are doing when we talk about trust in international politics. It begins by reviewing the recent and growing body of literature on trust and International Relations, locating this more nascent collection of literature within a wider, established body of social science work on trust in disciplines such as psychology, political science, business and management studies. It claims that an implicit but ubiquitous assumption about how words gain meaning underpins the literature, and that this assumption precedes and limits the range of possibilities for the form of the subsequent research. The thesis challenges this way of understanding by deploying Ludwig Wittgenstein's Philosophical Investigations. It then undertakes an alternative study of trust that acts as an ostensive challenge to the literature and thus shows by example how accepting different sites and processes of meaning can add to our understanding of words such as trust in International Relations. It accomplishes this through a 'grammatical investigation' of the uses of trust by President Richard M. Nixon and President Ronald Reagan regarding nuclear weapons and nuclear arms control with the Soviet Union. Using these examples, the thesis then suggests several alternative ways of talking about trust that would provide avenues for further research while avoiding the semantic and methodological difficulties of the dominant social science approaches. The contribution of this work is to challenge prevailing assumptions about words and meaning that exist within the literature and in so doing, to open up a path for alternative ways to talk about words like trust in International Relations.