All professions are conspiracies against the laity.
- George Bernard Shaw
One of the problems of specialization, and of living in a literate society of laws, is that specialists use their expertise to game the system at the expense of the majority. This is exacerbated because very important commercial and legal events are big and therefore, for most people, infrequent, and because specialists form effective guilds to exclude nonspecialists (for example, in academia). There are specialists involved in conducting these transactions and because of Pareto-principle-type effects, 20% of the causes (read "merchants") are responsible for 80% of the effects.
Consequently in these transactions, the most important that we conduct in our lives, there is usually a large asymmetry in experience, skill, and confidence. If you're buying a house or condo, you could quite literally be hundreds of times less experienced than the real estate agent (or infinitely, if it's your first time); if you're buying from a developer, the seller is also in a better position. Same going to the dentist; same testifying in court about something; same taking your car to a new shop for a big repair. I've never bought a transmission before, I don't know what I'm doing, and I just want to get out of there with as little damage to my bank account as possible. Because the transactions are infrequent, the chance that I will have repeated encounters with these merchants is low, so from a game theory standpoint they have no reason not to cheat me.
Perhaps the worst such example is legislation. Few citizens of democracy have the time, expertise, or inclination to read laws. Worse, today, no one even makes an effort. The model that modern democracy is based on is Athens, where you could show up to give your two cents, and law and knowledge weren't so specialized as to require two thousand page amendments to prior multi-thousand page laws. That is no longer the world we live in. An oral culture was good enough for the Iroquois constitution but not for more complex Athens. Have there been oral cultures in the past limited in specialization and growth because there's some fundamental upper limit to social complexity of oral cultures? In the same vein, is mere static writing no longer enough for twenty-first century democracies like the U.S?
There is also the problem of separation of gamers, and gamers' interests, from constituents. A large democracy invariably creates a class of legislators and lobbyists (the gamers) who write the laws. The larger the democracy, the smaller the body of decision-makers relative to the population, the greater that the Pareto principle will magnify inefficiencies and bad decisions and holes that in an Athens or Andorra-sized democracy could have been identified and rectified more quickly. The concern is that the legislative class's interests diverge so much from the people they are supposed to represent that the system is no longer representative. (Added later: one possible solution from Conor Friedersdorf here).
The term "broken" is thrown about in political discourse. If size (even if not complexity) differs between democracies, it may be interesting to compare the functioning of the legislature of a small democracy (Athens or Andorra) to that of a large one (the U.S. or Japan), though again what would be interesting would be to compare the functioning of legislatures, although one problem is I don't think we know how to devise a metric unit of measure for legislature function.
The accumulation of these impenetrable strata of the laws which are in theory binding us is a problem in that the current process produces a corpus that is necessarily rife with inconsistencies (especially regarding the use of precedents in the judicial branch); therefore, a worthy goal is a "legislative programming language" that would have to compile successfully on top of the current constitution and set of laws in order to be adopted. It's not compiling? Too bad - root through preceding laws and find the inconsistency; fix it, or fix your new law. We also would have to be less concerned that modern Supreme Court jurisprudence doesn't just amount to some kind of legal sophistry to go through contortions harmonizing what's written in the Constitution against what most of us agree is a reasonable decision, somehow even though we're centuries away from the ideals of the men that first put pen to paper.
But what we're concerned with is not so much having a neat constitution with its i's dotted and t's crossed, but rather one that works predictably. That is the real reason it's so unnerving that legislators do not and in reality cannot read the bills they pass. Robin Hanson has suggested that we hold policy-makers accountable for their policies - that is, have every new law or policy contain its own success criteria, and then follow it up (which he terms futarchy). The penalty for lawmakers - and ultimately, even voters - who are wide of the mark would be decreasing influence (i.e., at 18 we all start out getting 100 votes, but if you vote for a bunch of referenda that don't do what they said they would do, in the next election you only have 50). In essence, it's a prediction market, where the bet is made with future influence - and status, if you're a politician. Hanson describes futarchy as distinct from democracy (in my opinion to be provocative) but I think it's entirely compatible with democracy. Most modern democracies have adjusted the franchise repeatedly in the past to improve themselves. Is there a reason we should stop now?