I got to thinking about keilim and started writing. This was the result:
The gemara in chullin states: Dagim shealu b’k’arah, mutar lochlam b’kutach”
Fish which were placed on a (meat) plate, may be eaten with dairy.
The reason for this is that it is a nosen tam bar nosen tam (n”at bar n”at); a flavor which is descended from a flavor. Meaning that the meat flavor which the plate received from the meat and is now transferred to the fish is weakened by its two transferrals to the extent that it no longer is considered to exist. The two transfers are from the meat to the plate, and from the plate to the fish.
There are two caveats. The first is that the counting of two transfers can only begin once the flavor is transferred to a utensil, we do not count a prior transfer into another food item. The second is that only heter is considered to be weakened by the double transfer, issur is considered to remain in force. If the flavor becomes issur before having undergone two full transfers, it is not considered weakened. So we can postulate that n”at bar n”at d’heteira is mutat, and n”at bar n”at d’issura is assur.
The Shulchan Aruch, following this gemara rules that if a parve food is cooked in a meat pot, it may be eaten with milk, and vice versa. Nevertheless, one should not cook a parve food in a meat pot if his intention is to eat it with milk.
The R”ema, however, follows an alternate interpretation of the gemara. The gemara’s statement of “fish which were placed on a [meat] plate…” gives rise to an inference of “davka alu, aval nisbashlu, lo”; only if they were placed but not if they were cooked. Accordingly, the gemara only allows n”at bar n”at in a situation where the transfer was effected by being placed on a plate, but not if it was cooked on a plate. The gemara’s usage of “placed” is certainly intended to refer to a “placing” which transfers taste; a proper case would be where the plate (or the food) was hot from a kli rishon, but is not now on a fire. The rishonim who follow this inference are therefore splitting between cooked on a fire and transferral not on a fire.
The R”ema maintains a compromise where he follows this alternate interpretation for din lechatchila but allows the Shulchan Aruch’s interpretation for din b’dieved. Therefore, the R”ema’s holding is that if food is cooked in a meat pot, it may not be eaten with dairy, but if mixed with dairy, it may be eaten, b’dieved.
It is important to now note the difference between a utensil which was used for meat (or milk) within the past 24 hours, and one which was used for meat only prior to 24 hours. A pot which was used for meat retains the flavor and exudes it into anything which is cooked in it. However, the flavor spoils after 24 hours of being in the pot. Therefore, if one were to cook milk in a meat pot, if the pot was used for meat within 24 hours, the milk is assur; if it was not used for meat within 24 hours, the milk is permitted (the pot is nevertheless assur to use for anything according to the minhag). This follows the rule that any issur which confers bad taste on the food in which it is injected, is permitted b’dieved.
With this in mind, the R”ema sets forth the following rules:
If food is cooked in a meat pot which was used for meat within 24 hours, it may not be eaten with milk, but if combined with milk, it may be eaten, b’dieved.
Even though it may not be eaten with milk, it may be eaten in, and even cooked in, milk utensils. Care should be taken not to pour it hot from the meat pot to the milk utensil, if one does so, the bottom utensil should be considered assur, as well as the food. (In the instance of pouring into the milk pot, it will only be assur if the milk pot was also used for milk within 24 hours.)
If food is cooked in a meat pot which was not used for meat within 24 hours, it may be eaten with milk. (See gr”a 94:10 that one is allowed to do this purposely, that is to cook food in a meat pot which was not used for meat within 24 hours for the purpose of eating with milk. However, it is alleged that some have debated this.) It may also be cooked in a milk pot and poured directly, hot, from the meat pot into the milk pot.
If food is merely placed on a meat plate (which was used for meat within 24 hours), even if one or both are hot from heat of kli rishon, it may be eaten with milk and cooked in a milk pot. In this instance, care should again be taken to avoid pouring directly from the meat plate onto the milk or milk pot; if you do so, the bottom pot and food will be assur- (in the instance of pouring into the milk pot, it will only be assur if the milk pot was also used for milk within 24 hours.)
In the previous case of the food placed on a meat plate, one should not purposely put his food hot onto a meat plate if his intention is to eat it with milk.
There is one more point. Even those who allow food cooked in a meat pot to be eaten with milk, are divided on the issue of food roasted in a meat utensil. We will define roasted to include any cooking without liquid. Accordingly, following the opinion of R”ema, it is a matter of debate whether food roasted in a meat pot (which was used within 24 hours), may be eaten with milk, even b’dieved where it has been already mixed.