I, like others, have long taught that one of the ceteris paribus assumptions is that expectations remain constant. This allows for the usual interpretation of the kind of situations Doug listed -- that the market demand curve is sloping downward -- sigh of relief since this means the "law of demand" remains in force -- put it is shifting out due to a change in expectations about price. But I have also long found this explanation too slick, rather like sophistry, and the kind of reasoning that renders the law of demand a tautology that is impossible to disprove. The important thing is that there ARE many situations in which the demand for something increases when price goes up, rather than the reverse; and the supply of something decreases when price goes up, rather than the reverse. This is true because in the real world a significant number of actors in many markets interpret the information that the market price has risen as a more important indicator of where the price is heading -- as a signal about the direction of future prices -- than as an indication of what the price of the item is currently. The important point is for people to understand that demand may well rise and supply fall when price rises, and this should not be seen as practically impossible or the result of an aberation involving some kind of irrational behavior. For nothing is more "rational" than buying when you think the price of an item is rising -- whether you want the item or not! My complaint with the usual interpretation is that it makes it less likely that people -- people like my students but also most who teach them -- will understand this. But my quibble could be handled by emphasizing that expectations are frequently not constant, that any price movement in any market has an element -- its implications for the likely direction of future price changes -- that is likely to violate the ceteris paribus assumption of constant expectations; and therefore, that demand curves are very likely to be shifting due to changed expectations. But I close with an innocent question: If demand curves are wont to shift whenever market price changes, what good is it to know that they obey the law of demand? For that matter, what good are they period?