When using propositional logic, each atom is often associated with a sentence (a string of words) that gives the intended meaning of the atom, and so a propositional formula can be associated with a Boolean combination of these associated sentences. In this paper, we investigate the use of sentence embeddings to determine the similarity between pairs of sentences. Since different sentences can provide very similar information, we can enrich logical reasoning by treating atoms as “equal” when their associated sentences are very similar. We then use similarity between sentences to identify conflict. We can see some sets of formulae that are syntactically consistent, but via a similarity measure, are inconsistent. This means that we can use sentence embeddings to get a more intuitive understanding of conflict in real-world applications when a sentence is associated with each atom. We investigate notions of conflict based on this idea in this paper.