[…] The speakers of Proto-Indo-European would not have been misled into using the designation of the sieve (Lat. 'cribrum', OE. 'hrīder' n.) in the ergative '–s' case; a sieve in action would, of course, be used in an instrumental case. IE. neuters are properly the nouns which, on account of their meanings, were never used in the ergative. As some linguists would put it, the absence of an '–s' form must have been 'un fait de parole' before it became 'un fait de langue'.
- Martinet (1962), Pag. 152