Papers dealing with syntactic evidence in various languages for a higher performative sentence containing information about speaker, addressee and the speech act involved are reviewed and discussed. Arguments for this analysis have the form of showing that overt sentences behave in some way as if they were subordinate to a higher sentence containing a verb of "saying." Some processes related to this analysis depend on the NP's of the higher sentence. They include pronominalization processes which explain the status of overt first and second person pronouns, processes referring to real world facts which are assumed to be coded in the higher sentence, and processes concerning features of independent sentences which show them to be subordinated to a higher verb of saying, including the use of overt complementizers. While it is suggested that some information about the speech act and its participants needs to be expressed in the grammar, it is unclear whether the higher sentence is the best place for all these facts. That pragmatic knowledge affects utterance form in many languages should not necessarily be used as an argument for the universality of higher sentences. Universality is neither mandated nor precluded by the facts now in evidence. (Author/CLK)