Eighteenth-century prescriptive grammars were based for the most part on observable linguistic facts. Discrepancies between the prescriptive rules and actual usage usually indicate changes in progress; eventually the disputed usages either disappear or become standard. A major exception is nonstandard case marking. The same forms that grammarians condemned two hundred years ago are still in use and still marginal. I propose that case marking continues to be problematic because, as Emonds (1986) and Hudson (1995) claim, abstract case does not exist in modern English. I further argue that in opaque structural environments such as conjoined noun phrases, where the appropriate case is not immediately obvious, speakers tend to choose the more stylistically marked form. This will normally be nominative case or <i>whom</i>.