Eventualities have been recognized as psychologically and linguistically relevant for more than 50 years. Psychologically, eventualities are complex bundles of information derived from our perceptions of the world. The question for linguists is how much of this complexity is reflected in our eventuality-denoting lexical entries and the sentence-level denotations arrived at through semantic composition. It has been thought that, at the least, there must be a semantically encoded distinction between states and events. Such encoding has been seen as predictive of the behavior, interpretation and acceptability of states and events in numerous linguistic environments. That is, these environments have been claimed to be sensitive to a grammatical state/event distinction. This dissertation proposes that a binary state/event distinction is too coarse-grained to adequately capture the sensitivities of these environments. More fine-grained characteristics (e.g., controllability, volitionality, expected duration) are needed. These characteristics are often gradient (not binary) and vary by context, leading to the conclusion that they are best dealt with in the pragmatics, not encoded into a lexical entry. Moreover, such distinctions crosscut the traditional state/event divide, blurring the line between the two. Historically, linguists have seen English main verb do as integral to the state/event distinction, theorizing that do is itself responsible for eventivity and is covertly present in the lexical entry for all (and only) events. Thus, any environment containing do has been seen as grammatically barring stative predicates (e.g., do so anaphora). Drawing largely from naturally occurring data, this dissertation investigates to what extent do is incompatible with stative predicates. The contributions of this investigation are four-fold. First, it highlights the need to consider diverse naturally-occurring data when constructing linguistic theories. Second, it demonstrates that do is in fact compatible with stative predicates. Third, it demonstrates that, across-the-board, the diagnostics used to determine stativity/eventivity are unreliable and are sensitive to characteristics that are orthogonal to the distinction. Lastly, but most importantly, by showing that stativity is irrelevant to the acceptable use of do, this dissertation undermines a key argument for a grammatical state/event distinction, casting doubt on the existence of such a distinction in the first place.