The storehouse/correspondence partition in memory research: Promises and perils