There are two potential ways in which monetary deflation can impact a currency:
1. Usability
Most modern currencies have a minimum unit, such as the penny in the United States, or the yen in Japan. If these currencies existed in only finite quantities, then as the economy grew and/or physical money was lost, there might eventually be too few of these units in circulation to allow normal commerce. Imagine if, for example, the U.S. dollar became so rare that you could buy a car with a single U.S. penny. How would you buy a loaf of bread?
Bitcoin, however, does not have this problem because unlike the physical currencies of most countries, bitcoins are infinitely divisible. Even if only one bitcoin remained, it would be sufficient to run a substantial economy on, based on the ability to use tiny fractions of it like the Satoshi. If even the Satoshi were to become too highly valued/unavailable for commerce, the Bitcoin protocol could be updated to allow even smaller denominations. So the usability of bitcoins is safe from the effects of deflation.
2. Broader economic consequences from deflation
Many economists believe that inflationdeflation is bad for an economy because people will be less likely to invest their money if they can "earn" increased value simply by holding on to it. Why buy something today, the argument goes, if you could buy it for half as much in a few months time? Also, taking out loans becomes very difficult, because to be practical they have to have negative interest rates and then why would anyone loan out money in the first place?
However, there is good reason to believe that deflation might not have such a negative effect on Bitcoin. For one, Bitcoin is unlikely to become the only currency in the world economy, so investments, loans, etc. can simply be denominated in another currency. But even more importantly there are several concrete examples from history of strictly deflationary currencies.
For example, the Iraqi "Swiss" dinar was a currency that ceased being printed after the first Gulf War, and was left in use only by the Kurdish regions of the country. With a finite number of "Swiss" dinars in existence, and some being lost or damaged over time, this left the currency in precisely the same state as Bitcoin. Since it was no longer backed by a government, many economists expected the currency and its economy to collapse. Instead, the Kurdish economy continued largely as normal for 13 years, and the "Swiss" dinar even appreciated against the "Saddam" dinar that was meant to replace it.
Rick Falkvinge also raises another excellent point regarding deflationary economies. If, as economists expect, people would decline to buy something today as opposed to next year for half price, then it should be the case that no one ever buys computers, right? But in fact, the computer industry thrives despite this continued trend of "whatever something costs today, it will cost a fraction of that within a few years".
Ultimately, this is a question for economists to duke out, and you can find more information about how different economic schools approach the problem here. Suffice it to say, this cannot be considered a crippling fault in Bitcoin as a technology.