In the current study, data from A Corpus of English Dialogues (1560-1760) are used to consider contexts with the have-perfect and temporal adverbs of the definite past time such as yesterday, last night, ago. Data analysis is conducted within the framework of a usage-based approach, which gives evidence to the hypothesis that in Early Modern English the have-perfect in spoken register was gradually developing perfective semantics and that it followed the stages of generalization of meaning depending on the degree of event remoteness. Investigation of the instances where the have-perfect is used in narrative passages shows that the have-perfect in such contexts does not lose its pragmatic component of current relevance but is employed to highlight a crucial event out of a chain of past events. The paper proposes the hypothesis that the main mechanism preventing the have-perfect from further aoristicization is the operation of syntactic analogy within the syntactic paradigm of the present perfect, which had already fully developed by the time of Early Modern English.