If you look back on, for example, video game development you can see the problems that can occur without such walled gardens. The home video game market boomed in the late 70s and early 80s, with families buying new games like hotcakes. A lot of game makers jumped on the bandwagon and pumped the market full of low quality games. Whereas in previous years the total number of games for the Atari 2600, for example, had been in the low dozens in 1982/83 this number ballooned to hundreds. Consumers could no longer have much confidence in which games to buy and so they stopped buying, leading to a massive crash of the video game industry in the US that lasted until Nintendo came along with its own walled garden approach.
However, the video game industry has matured since then, and walled gardens are no longer very helpful (there are far more than sufficient resources these days to determine which games to buy and which to avoid based on individual preferences).
As the mobile app market continues to mature it will strain against its walled garden confines more and more. Increasingly such hand-holding is less necessary and more and more restrictive. Apple has a choice to recognize that the market is changing and to adapt or to ignore the changes and pretend as though it's still 2008 while the world passes them by.
The 2600 was walled-by-obscurity for its first several years, since the video chip was custom designed by Atari and not documented publicly, plus the tools to write and compile and run 6502 assembler were fairly primitive. But by 1982 or so, enough reverse-engineering had been done and the tools had matured and enough Atari expertise was available on the hiring market that the system was essentially fully open.
(I've been there - I wrote an Atari 2600 game my freshman year of college in 1997.)
I'm still having trouble buying the collapse of the walled garden as the reason for the video game crash of 83-84. Pac-man and ET are widely cited as two of the major causes of the crash, and those were in-house Atari productions so the walled garden wouldn't have helped there.
But there were plenty of other management and business causes of the crash too. Atari reportedly produced more ET cartridges than existed 2600 systems, hubristically deciding that ET would drive a wave of system buying. That sort of financial wizardry won't be saved by any amount of ecosystem walls.
I'd say that the openness of the system caused the failure of Atari itself more so than it caused the industrywide crash. Atari couldn't maintain premium game prices and volume against the flood of competition. (If better games weren't available, ET would have sold more in the vacuum.) A similar story played out in the PC market over the next decade: IBM created an open system, then got marginalized out of their own industry in a race to the bottom. But the whole PC industry always thrived.
I also think the crash itself is overblown in latter-day coverage. I was a kid and didn't notice anything at the time; Toys R Us still had shelves full of Atari games right up until the NES caught on. The bankruptcy of invincible titan Atari itself was a big event, but Coleco didn't fold until 1988 and Mattel and Magnavox trucked right on in other fields of business.
So tying all this back to the topic, remember that Apple's interest is Apple's own survivability and profitability. Apple's responsibility is not to maximize the adoption and utility of the mobile phone app market in general. Apple's goal is to capture the biggest slice of the pie for itself, and it perceives that it has the clout to do that by dictating terms. Whether they will go the way of the NES to market domination or the way of IBM's Micro Channel to obscurity is up to the market.
Having said all of that, I too am confused with the argument here -- the videogame console market is very much a walled garden and has been for a very long time.
I guess this argument will be tested by the Mac app store.
i.e. People weren't driven away by the array of choices. They were driven away by bad experiences.
The system broke down, the chances of buying a good game just on chance were low, and thus the chances that any of your friends had happened to buy a good game was also low. It only took a few times of people getting burned for them to stop taking the risk and to curtail their game buying, staying content with the existing library of games they had. Even though there were still many quality games (donkey kong jr, joust, ms. pac man, and pole position were all released in '83), there was so much crap that people simply withdrew from the market. Moreover, the bubble was reliant on an unlikely massive growth spurt in video game buying, the game industry over leveraged itself, it would have crashed even without a collapse of consumer confidence.
The game console industry is still very much based on walled gardens.