John L. Austin (1975) _ How to do Things with Words_ [1955], Cambridge, Mass.: Harvard University Press. 

Delivered as a series of lectures (the William James Lectures at Harvard University in 1955, themselves based on lectures at Oxford 'Words and Deeds', 1952-4, and prefiguring the BBC lecture 'Performative Utterances'  in 1959, amongst others; it's interesting in this context that the book is a transcript of speeches and thereby losing some of its performative qualities), Austin examines ordinary linguistic usage, and the contrast between performative and constative utterances. In the first lecture, he establishes that as well as providing descriptions, questions and commands, sentences are performative in that do something as opposed to just saying something. The performative statement does not simply correspond to facts but at a level of abstraction from the facts, not based on whether it is a true or false statement (but what Austin refers to as happy or unhappy, 1975: 133). To utter a sentence is not simply to: '_describe_ my doing of what I should be said in so uttering to be doing or to state that I am doing it: it is to do it' (1975: 6). A sentence or utterance of this type, he calls a 'performative' to indicate 'how the issuing of the utterance is the performing of an action - it is not normally thought of as just saying something' (1975: 6-7). 

In general terms, in saying something we are doing something - a 'speech-act'. Austin introduces the term 'operative' [note: Benjamin's operative writer which suggests a particular kind of doing or producing based on political action] to indicate how an utterance serves to make an effect (he is thinking of law where a statement is instrumental) for his interest is how much saying something can make something happen. The concern is in what ways to say something is to do something, or how in saying something we do something, or even _by_ saying something we do something. He identifies that in saying something, we perform 'locutionary acts' (the meaning of something; that makes a certain reference to something), 'illocutionary acts' (the force in saying something; such as informing, ordering, warning), and 'perlocutionary act' (the achievement of certain effects by saying something; such as persuading someone of something), amongst other types (1975: 109). These interconnected types of actions are all examples of how utterances are the performative issuing of 'doing an action'. The perlocutionary act marks a distinction between the action and its consequences, and in this sense, the consequences almost become the act itself. 

Any spoken utterance is always locutionary in that it is an act of saying certain words, and of making movements with vocal chords and breath. So there is necessarily a connection between saying something and physical action but actions do not simply happen a a consequence of locution or simply produce effects - it takes effect in certain ways depending on the securing uptake, taking effect and inviting responses (1975: 121); depending on the purpose and the context.   

Clearly performative utterances are linked to actions by the speaker in a general sense, but Austin is interested how utterances are an outward expression of an internal act: 'the outward appearance is a description, _true or false_, of the occurrence of the inward performance'. He quotes Hippolytus: 'my tongue swore to, but my heart did not' (1975: 9-10). Indeed saying words is not enough in itself,  it is its relation to ethical action that is important.