or black coins, or underweight ones, or coins of Tiberias,” or “a thousand less fifty,” or the like, his acknowledgement would be subject to that specific qualification. If an exception constituted a retraction, it would not be accepted in an acknowledgement, since a person may not retract an acknowledgement concerning the rights of others.

Some marginal jurists frequently err in such matters. When asked about a condition stipulated by the founder of an endowment, an oath someone has sworn, or something similar, a jurist sees that the opening words are unrestricted or general, while their ending qualifies them. Sometimes he treats this as a conflict between two pieces of evidence, applying the familiar rules governing conflicting evidence, such as judging them equally strong or giving one precedence. Sometimes he regards the statement as contradictory because its ending differs from its beginning. At other times he flounders in bewilderment, attributing to a capable person the conduct of someone negligent. He may even say, “This is a mistake by the scribe.” All this stems from failing to distinguish between connected speech and separate statements.

Anyone who understands that a speaker’s opening words must not be judged until he has made a pause that definitively ends his statement, and that a writer’s text must not be judged until he has definitively finished writing, will be free of confusion in this matter and recognize the soundness of what scholars say about how discourse conveys meaning.

It is a grave failing to attribute an error to a speaker when his words admit a sound interpretation and follow the finest patterns of ordinary speech, and then to deem one of the two supposedly conflicting passages mistaken rather than the other. For if it were permissible to say, “His words, ‘provided that whoever among them dies without leaving a child,’ are an error,” that would be no more justified than to