The Piaculum

Roman public religion held that a rite performed incorrectly had not been performed. The category of fault was wide: a word misspoken or omitted, a gesture in the wrong order, a defect found in the victim after it was opened, a noise or an interruption at the wrong moment, an ill-omened event during the ceremony. Where the fault was discovered the response was not to apologise but to begin again, repeating the whole ceremony from the start, and to offer a piaculum for the error itself. Public festivals were repeated on this ground, sometimes more than once, which is expensive and was accepted as the cost of doing the thing properly.

This is a different logic from the one that governs the offerings of the Israelite system, where the classes of offering are matched to classes of wrong committed by a person. The Roman fault is procedural. Nothing has been done to anybody; a form has been broken, and the piaculum repairs the form. It is also a different logic from the vow, which promises a future offering in exchange for a future benefit and creates an obligation running forward. A piaculum settles an account rather than opening one, and expects nothing in return.

The offering itself was usually modest, commonly a pig, and its size bore no relation to the gravity of the occasion, because what was owed was an acknowledgement rather than a compensation. That the same word covers both the fault and the offering that repairs it is a piece of Latin that tells the reader where the emphasis lay.

The most quoted piacular formula is agricultural. A farmer needs to thin a sacred grove on his own land, which means cutting trees that belong to a god. Cato tells him to offer a pig and to say a prayer that begs pardon of the deity to whom the grove is sacred, addressing that deity as whichever god or goddess it may be, and asking leave both for himself and for the man who does the cutting. He is then to repeat the offering each day the work continues.

The formula is remarkable for what it concedes. The offerer does not know who owns the grove, does not know whether the owner is male or female, and says so in the prayer rather than guessing. Roman religious language is full of such hedges: alternatives named so that the right one is certainly included, formulae covering a deity by name or by any other name preferred, phrases reserving the possibility that the offerer has got it wrong. This is not vagueness. It is the same exactness that made a misspoken word invalidate a rite, applied to the case where the required information is unavailable, and the solution is to make the uncertainty explicit inside the formula so that the form remains correct.

The prayer also declines to explain itself. It does not say why a grove belongs to a god, what the god will do if unpropitiated, or what the pig accomplishes. Roman religious texts of this kind are procedural throughout, and the theology has to be inferred from the procedure, which is a large part of why the interpretation of Roman religion has been argued over as much as it has.