The problem of the variation of the truth-status of the propositions over time is one of the favourite topics of the logicians as soon as the end of the XIIth century. The aim of the present article is to acknowledge the various theories of restriction provided in the 12th and 13th centuries, to solve semantic problems by means of contextual determination. Given the texts presently available, up to seven different doctrines are accounted for, depending on whether or not subjects and predicates, on the one hand, substantial and accidental terms, on the other hand, are actually distinguished. Among those doctrines, particular attention is paid to that of the so-called Dialectica Mona-censis. This anonymous treatise, dating from the last decades of the 12th century, introduces two different theories. The first one suffers from a discrepancy between the content of the rules of restriction through present, past and future tense verbs and that of the various conditions laid down in these rules. Thus, though substantial and accidental terms have theoretically the same supposition, whether they be subjects or predicates, in each different tensed form of the verbs, the author practically draws a line between substantial terms like ‘homo’ and accidental ones, like ‘album’. As a matter of fact, the truth of the proposition “homo curret” at instant S (Reichenbach’s ‘point of speech’) necessarily entails that of “homo currit” at instant E (the ‘point of event’), but this is not the case with ‘album’ in “album curret”, since that which is now white (point S) might very well be no longer white at time E. Those difficulties determine a second theory which offers a more satisfying account of the difference between substantial and accidental terms. Finally, a comparison is made between the first theory in the Dialectica and William of Sherwood’s account of the compounded and divided senses of the propositions, and a parallel is suggested with modern paraphrases using A. N. Prior’s tense operators.