According to the Mishna (43b) there is a difference of opinion between Beit Shammai and Beit Hillel with regard to someone who is watching something that belongs to his friend, and decides to be shole’ah yad – to make use of it for himself. According to Beit Shammai, from the moment that he thinks that he will be shole’ah yad, he is already responsible; according to Beit Hillel, he would not be responsible unless he actually took the object and used it.
The Gemara explains that both Beit Shammai and Beit Hillel base their positions on passages in the Torah. Beit Shammai understands the pasuk (Shemot 22:8) to mean that any type of intention of using the item for personal use would make the guard liable. Beit Hillel points to a qualifying pasuk (Shemot 22:7) as teaching that an act must take place before a person will be held responsible.
Some of the rishonim believe that Beit Shammai‘s position holds true even if the man simply thought that he would be shole’ah yad (clearly this will only be true if he admits that that was his intention; otherwise we could not possibly know what he was thinking). The Ritva suggests that Beit Shammai could not obligate the individual unless he minimally stated his intention – which is, in fact, the simple meaning of the passage that Beit Shammai quotes as his source – al kol devar pesha (Shemot 22:8) – for every matter of trespass – with the word devar meaning a spoken word. Rashi goes so far as to require that the statement be made before two reliable witnesses, if we are to hold him responsible.
Most of the rishonim agree that our concern is only if the man stated his intent aloud. This forces them to explain why the Mishna used the terminology ha-hoshev li-shlo’ah yad – someone who intends to (or “thinks that he will”) make personal use of the object. One explanation is that even if he says that he will do it, he must intend to do it, as well. If his statement was made in jest or if he was bragging, but he did not intend to use the object, he will not be held responsible.