Why? Think of the second-order effects of what you're saying--I expect the value of ${commodity} to rise indefinitely, so what do I do? Buy ${commodity}. Or, I own a bunch of ${commodity} and expect the price to rise indefinitely--what do I do if someone offers to buy my ${commodity}? Either sell for a higher price, or don't sell at all.

Both of these have the effect of bringing the future price increase into the present. If everyone expects a 10% increase in the next year, the price today will increase by roughly (10% - the risk-free interest rate).

Now, is the picture different for currencies? Yes. Economists call short-term prices "sticky," and sticky prices are why demand-side recessions are possible. If I'm earning $40,000 a year today, and my boss decreases my pay to $39,000 a year, I don't say "well, there was deflation this year, so I actually got a raise!" On the other hand, if I'm earning $40,000 a year, and my boss doesn't increase my pay at all, I don't say "there was inflation this year, so I just got a pay cut!" In the short term, sticky prices make things (like labor) cost more or less than they would in a frictionless market. That stickiness is why countries with deflationary currencies have problems.

Now then--bitcoin may be a currency of sorts, but are bitcoin prices sticky? No. Companies selling products for bitcoins overwhelmingly price their goods in dollars. At whatever time the transaction is conducted, the buyer converts that price into bitcoin and transfers bitcoin to the seller, who sells the bitcoin for dollars. A buyer will never say "yesterday your price was 1BTC, and today it's 1.2BTC--why have you increased your prices?" A seller will never say "yesterday you offered to pay me 1BTC, and today you're only offering 0.8BTC--no sale!" Both sides of the transaction are dealing in dollars, only using bitcoin as a medium to exchange those dollars. In other words, bitcoin is a medium of exchange but not a unit of account.

Is bitcoin likely to become sticky in the future? Well, if a large proportion of people were paid solely in bitcoin, and could live comfortable lives spending bitcoin instead of dollars, it's possible that we could have a cultural shift to thinking "I'm paid 333BTC a year" instead of "I'm paid $60,000 a year." And at that point, bitcoin's deflationary nature would be an issue. How likely are we to reach that point? Not very, not least because it's a self-limiting cycle--if the sticky-price/deflation combo is an issue in the bitcoin economy and not in the dollar economy, people have very little incentive to switch their accounting to bitcoin. But that's an entirely separate issue from whether bitcoin is or will remain a useful medium of exchange.