Ohalos, Chapter 6, Mishnah 6. Until now the chapter has been dealing with a wall that was put up as part of a house, a "koseil hameshamash es habayis": the builder planned a house, and the wall serves that house. Our mishnah turns the phrase around and speaks of a "bayis hameshamash es hakoseil," a case where the space came first and the wall is only an accidental by-product. Because the wall was never intended as a wall, the halachos of tumah passing through it are entirely different, and the mishnah teaches us how thin a barrier will suffice.
What is a "bayis hameshamash es hakoseil"?
Picture a burial cave. In the times of the Mishnah, people would bury the meis by hollowing out places in the rock of a cave. A person digs one recess into the wall of the cave, and then digs a second recess a little further along. He never set out to build a wall at all, yet once both hollows are dug, the strip of rock left standing between them is now, in effect, a wall. And the hollows themselves are what the mishnah calls "batim," even though nobody built a house here; they are simply the burial spaces on either side of that leftover rock.
Now suppose a k'zayis of flesh from a meis is found inside that accidental wall. The mishnah rules: "yidon," this wall is judged on its own. It is treated as an independent entity, and the tumah inside it does not spread into the burial spaces on either side, even if the layer of rock separating the tumah from the hollow is as thin as "kiklipas hashum," the peel of a garlic. A garlic peel is about as thin as a covering gets, and yet it is enough to hold back the tumah.
Keilim in the wall, tumah in the burial space
The mishnah spells out the cases. "Keitzad? Koseil shebein shnei kukin o bein shtei me'oros." A "kuch" is one of those shelf-like niches dug into the rock of the cave, where the meis was laid; a "me'arah" is a larger area hollowed out of the stone for burial. Either way, one was dug here and one was dug there, and the strip left in the middle became a wall by default.
"V'tumah babatim v'keilim bakoseil, v'aleihen kiklipas hashum, tehorin." The tumah rests in one of the burial hollows, whether the narrow niche or the wider chamber, while keilim sit inside the strip of rock dividing them, with nothing more than a garlic-skin thickness of stone above them. Those keilim stay tahor.
This is precisely the opposite of what we learned about a wall built to serve a house, where a keli embedded in the wall is treated as though it were in the house, or in the closer of the two houses. There the wall was built as a wall for the house, so whatever is inside it is considered part of the house. Here, since the wall arose by itself and was never planned, it is viewed as its own separate domain, and the thinnest covering suffices to keep the keli out of the tumah.
One condition, though: the keli must be entirely inside the wall. The covering over it may be paper thin, kiklipas hashum, but if any part of the keli protrudes out of the wall, that part is inside the ohel hameis and the keli becomes tamei.
Tumah in the wall, keilim in the burial space
Flip the picture and the ruling comes out the same. "Tumah bakoseil v'keilim babatim, v'aleiha kiklipas hashum, tehorin." Now the k'zayis of the meis is lodged inside the unintended partition, whether it separates one niche from another or one chamber from another, and the keilim are lying in the hollows themselves. So long as something covers the tumah, however slight, no thicker than the skin peeled off a clove of garlic, the keilim are unaffected. We do not attribute the tumah to the hollowed space beside it.
Compare this to a partition deliberately raised between one dwelling and the next. There, tumah lodged within it is credited to the dwelling, since the builder put up that partition to serve the house, and it counts as an extension of the house. The strip of rock in our mishnah, by contrast, was nobody's plan, so it stands as a domain unto itself. The earlier stipulation holds here too: no part of the flesh of the meis may stick out past the rock. It has to be sealed in, no matter how thin the layer doing the sealing.
An amud with tumah beneath it
A further scenario is presented: an amud, a pillar, standing inside one of the burial hollows, "v'tumah tachas ha'amud," a fragment of the meis lying below the pillar, with no air space measuring a tefach by a tefach. Tumah squeezed in this fashion, lacking a chalal tefach, carries the name "tumah retzutza," and its route runs vertically only: "bokaas v'olah, bokaas v'yoredes." It pierces its way upward and pierces its way downward, conveying tumah to whatever stands in a straight line above it and whatever lies in a straight line below it, while nothing off to the sides is touched.
In this halachah there is no distinction between a cave and a house. The identical din would apply to a pillar standing in a house, and that is exactly the mishnah's point here.
Suppose instead that a hollow of a tefach by a tefach, or more, opens up under the amud. Then the tumah is no longer confined to that vertical column and travels outward as well. Everything within the burial chamber turns tamei, and were the pillar set up in a dwelling, everything in that dwelling, because the tumah now carries the din of a "kever sasum," a stopped-up grave, a status this perek introduced earlier.
Reviewing the mishnah
When we are dealing with a "bayis hameshamash es hakoseil," where nobody set out to build a partition and two recesses were simply carved into the rock of a cave, leaving whatever remained between them standing as a wall by default, we apply the rule of the garlic peel. Where a wall was put up to serve a dwelling ("koseil hameshamash es habayis"), we measure which half of its width holds the tumah or the keli. That measurement plays no role here; the only question is whether a covering exists, and even a garlic-skin layer qualifies. Both directions come out the same: a k'zayis of the meis lodged in the rock under such a covering while keilim sit in the recesses, or keilim embedded in the rock under such a covering while the tumah lies in a recess. Either way the ruling is taharah, provided that nothing juts out beyond the rock.
And in the case of an amud with tumah underneath it, where cave and house are alike: if there is no chalal of a tefach by a tefach, the tumah pierces upward and downward and defiles only what lies in its direct path. If there is a tefach by a tefach, it becomes a kever sasum, a grave in its own right, and it renders tamei everything around it, so that the entire "bayis," the burial space, becomes tamei, since it functions as an ohel over that grave. Here too, the very same din would apply in a house.