leap of faith
Democracy is a very poetic word. “Rule of the People” is an evocative phrase, but for this essay it needs to be more rigorously defined. The “rule” in democracy, we may agree, means that, using a standard repeatable process, a particular group of people make decisions, and those decisions are then put into effect. It is a two-step process in which both steps are needed. But another poetic aspect presents itself, because coming to a decision is a personal thing, done by an individual; it is not clear what it means for a group of people to “make a decision”. I suggested in the preceding chapter that, in the context of democracy, a group of people comes to a decision when it reaches consensus, when all its members agree to take a particular action. All the members may not agree it is the best action to take, but they agree it is the action they will take.
That is an odd situation. Why do some people in a democratic collective decide to go along with a decision they think is wrong, or at least sub-optimal? And, as a related question, who in the group takes this awkward position? For the ancient Athenians, I suggested, those who held the minority opinion abandoned their initial position because they knew that, were the matter to be decided violently, they would be on the loosing side of the conflict where every man had a sword. Quite pointedly, the ancient Athenians did not toss a coin, or play rock-paper-scissors, or consult entrails to make their collective decisions. They voted. And the vote tallies showed them who would win and who would loose if the matter were to be settled violently.
There was a hugely important side effect to this behaviour. Avoiding violence mattered, but it wasn’t just that. Those fierce men of long ago assumed a subservient minority position, accepting a decision against their better judgment, because they knew that they might be in the majority come the next decision, and they had faith that those who would be then in the minority would accept the majority’s opinion with the same good grace. It wasn’t just the threat of violence that moved them to acquiesce; it was also trust: trust in their fellow democrats and trust in their democratic process. Foreswearing the use of violence in resolving matters about which they were nevertheless passionate, they poured their energy into arguing with one another. They bared their intellects to their fellow democrats, and that also built mutual trust.
Today we engage in voting in many situations. Families (sometimes), volunteer organizations, local and regional boards, boards of directors of corporations, legislative assemblies and appointed committees, the supreme court, all come to consensus through the argue-together-then-vote-and-majority-rules process. In many cases violence is not right under the surface. Many voters today are not armed male warriors. Children, women, physically handicapped people, prisoners, and many strong adult males who eschew armed violent confrontations, vote in our society. Today, those in the minority adopt the majority opinion because they believe it is more worthwhile for the group (however big or small) to act with unity of purpose than it is to break with the group. By tradition (a tradition rooted in long ago violence) we do the argue-together-then-vote-and-majority-rules process. And, we trust the other people in our groups to follow the process too.
The supreme-court example is illustrative. In supreme-court decisions several judges preside and pass judgment, generally rendering a yes or no answer to a question. The judges usually justify their decisions in a written commentary. Often the decision is not unanimous. Then, as well as a majority opinion commentary, one or more minority opinion commentaries may be issued. It is apparent there was disagreement about the verdict. The disagreement is fairly rigorously detailed in a document of public record. It might be more persuasive and coherent than the majority decision’s commentary. Yet, the dissenting judges, by dissenting and detailing why, do not at all mean to encourage disobedience to the majority decision. The ideas presented in dissenting commentaries are only to offer points of contemplation to the legal community. The majority decision stands as the law of the land in the hearts and minds of all the presiding judges. While they may disagree with one another, they believe in each other and the process by which the decision came about.
These institutions then, families, volunteer organizations, the supreme court, and so on, can be understood to engage in decision-making using the same procedure ancient democracies used. You might choose to call them democracies on a very small scale.
On a large scale, democracy runs into a logistical problem. A big group, engaged in the argue-together-then-vote-and-majority-rules process, is having a referendum. As mentioned in the previous chapter, the problem with referenda, the problem with getting everybody engaged on a regular basis with every issue that the community faces, is that it takes too much time and effort. The other work a society needs to do, to take care of its people, is neglected. Societies (ancient and modern) can only afford referenda every now and then. Referenda cannot be used to decide every issue, and because they cannot, they are rare, and, because they are rare, they do not generally build trust within the community.
Instead, a model of the community is made, and that model, with less effort and expense, generates opinions that are treated by the community as though they were majority opinions. The model is called the Assembly. When people practiced what we call direct democracy, the assembly comprised individuals selected randomly from the wider community. In what we call representative democracy, the model is an elected parliamentary assembly.
While the assembly is a pragmatic step, it is also a leap of faith. The community, ancient or modern, believes, without direct evidence, that their model of the community, their Assembly, is generating opinions that are the same as would evolve as majority opinions in a referendum. If you don’t believe the Assembly generates the same opinions the wider community would generate, then you believe the Assembly is generating decisions contrary to the will of the people. How can the people be ruling? How can this be democratic?
And if you do believe the Assembly generates the same opinions the wider community would generate, then, on what grounds? We don’t usually check the model’s results against referenda results. Except for having a referendum right after an assembly opinion is generated, how can anyone know if the assembly got it right or not? We can add to this the irony that the whole point of having an assembly is to avoid having to do referenda. If we believe we are in a democracy, that the people rule, then we believe the assembly’s opinion is the same opinion that the community would come to if the same issue was put to a referendum. Literally, we must believe our assembly is constructed such that each time it generates an opinion, it channels the unknown (and, without a referendum, unknowable) will of the people.
And here, I suggest, is part of our dilemma. In 1992 Canada held a nation wide referendum, and in 1980 and 1995 the province of Québec held province-wide referenda. In all three cases the respective assemblies generated an opinion that proved contrary to the will of the people. Canada’s parliamentary structure failed at the provincial and federal level all three times it was tested. Other times we may have had our personal doubts, but three times we have faced irrefutable inductive evidence that our parliamentary assemblies do not channel the will of the people. We all know this. But we are hooked on the notion that parliamentary assemblies are democratic. After all, we are a modern “representative democracy”! I maintain that it is very much an open question whether parliaments can channel the will of the people they claim to represent, and there are grave consequences when they try. It is the intent of this chapter is to show why that is.
a measure of success
We can use well-established arithmetic formulae to determine, or at least speculate upon, the degree of accuracy with which an elected representative of a riding, or the elected assemblies used in Canada today, or the ancient Athenian assembly, can channel the majority from the wider community.
Let us consider first the Athenians’ case. We wish to calculate the probability that, for any community issue, the majority in an Assembly of randomly chosen participants will have the same opinion as would the majority in the wider democratic community. Say there are N people in the community and A people in the assembly. The pragmatic point of the assembly is that it comprises fewer people than the wider community, and that fact can be expressed as
A ≤ N.
An issue arises. If a referendum were held on that issue, the majority M of the N people would want the issue settled a certain way. Instead, the majority W of the A people in the assembly decide how the issue is handled. We can express the bounds of the community’s majority as
N/2 + 1 < M < N if N is even, and
(N+1)/2 < M < N if N is odd,
and we can express the bounds of the assembly’s majority as
A/2 + 1 < W < A if A is even, and
(A+1)/2 < W < A if A is odd.
Following are some notes about presentation in the probability equations that follow. The factorial sign (!) indicates the product of the all the integers up to and including the antecedent. So
5! = 1 x 2 x 3 x 4 x 5 = 120.
Printing one integer value, j, above and one integer value, k, below, without a dividing line between them, and all within parentheses, indicates the number of different combinations of j things considered k at a time without repetition. So for example, the number of unique sets of 3 things from a group of five things is
There are, then, 10 unique 3 member sets. Consider, for example, the first 5 letters of the alphabet. There are 10 ways to pick 3 of these letters without repetition: abc, abd, abe, acd, ace, ade, bcd, bce, bde, cde. By ‘no repetitions’ I mean we don’t count sets like aaa or abb. By ‘combinations’ I mean the order of the elements in the set does not matter; abc is the same combination of elements as bca.
(As an aside, you may note that the piece of hardware we call a ‘combination lock’ requires the correct order of numbers be dialed in to open the lock; the order is absolutely important and these locks are misnamed. They should be called ‘permutation locks’, but that is another story for another time.)
The expression for a probability is written as p followed, in parenthesis, by a brief mathematical expression describing the event under consideration. For example, I can write the probability that a majority W of people in the assembly have the same yes-or-no opinion on a particular issue as a majority M of the people in the wider community as p(W≡M). The symbol ≡ is, properly, congruency. In this essay I am using it instead to mean ‘agrees with’. Since W and M are integer values, using the equals sign could be misleading. So, p(W≡M) can be read as ‘the probability that W people in the assembly agree with M people in the wider community’.
That’s it for symantics. Now we develop the mathematical model. The probability that a particular majority W in the assembly will hold the same opinion on a matter as would a majority M in the wider community, is give as
This is the probability function for what is called the hypergeometric distribution. The mathematical expression was derived originally to help with quality control; if you took a small sample of items from a larger group of mass produced items, and found some in the sample defective, you could derive the probability that a particular number of items were defective in the larger group.
It’s not required that the W and M be proportionally the same majority; any majority in the assembly will do, so
The probability of any particular majority M of all possible majorities of N is
And so, the probability that the majority in the assembly will agree with the majority in the wider community is
Which simplifies to
The two at the front is for symmetry; a majority can hold an opinion for or against an issue.
Now let us consider the parliamentarian elected in a riding. The probability that a single representative from a riding will hold the same opinion on some matter as would be held by a majority within the riding is, by definition
This, combined with p(M) above, gives
which simplifies to
In Graph 3-1 I have plotted the expression for p(1≡N) and p(A≡N) for a range of values of N and for a few cases of A in which A is a constant fraction of N. The probabilities only exist for integer values of N (people may be fractious but not fractional), and so the lines are really a set of discrete points. The scale of N is logarithmic, so those points get very close together as N gets larger towards the right on the graph. I have plotted only the cases of odd values of N and A to avoid a discussion about how to handle ties. The differences in probability between even and odd N and A also becomes very small as the values of N become larger. The selected values of A are A=N/3, A=N/11 and A=N/51. Consequently, only every 3rd, 11th, or 51st N is represented, respectively for these p(A≡N) curves.
We can see the p(1≡N) curve sweeping asymptotically to 50% as N gets large. Sprouting off this curve are the three p(A≡N)curves. The curve p(A≡N) for A=N/3 leave the p(1≡N) curve at N=3 which is the case when an assembly of 1/3 of all the people in the community is an “assembly” of exactly one person. Similarly, the p(A≡N) curve for A=N/11 and A=N/51 leave the p(1≡N) curve at N=11 and N=51. Each of the p(A≡N) curves approaches an asymptote above 50% as N gets large. For the case of A=N/3, the probability limit is about 70%. For A=N/11 and A=N/51, the probability limits are about 60% and 55% respectively.
There may have been at most about 30,000 members of the Athenian free male community, and that number may have been as low as 7,000 during bad times. The number of people in the Athenian assemblies seemed to have ranged between 1,000 and 6,000 people. I have marked on Graph 3-1 a triangular zone representing where democratic Athens might have been. In modern times, there are some 20 million voters in Canada, represented by about 300 members of parliament, and I’ve shown that spot on the graph as well. Finally, the ridings in Canada have a range of voters, from about 15,000 to 70,000. Along the p(1≡N) I’ve marked this range.
The first observation that strikes one about the information in Graph 3-1 is that Canadian parliamentarians have little more than even odds of successfully channeling the will of the people from their respective ridings. For the considerable expense of elections, salaries, and pensions, the results seem little better than tossing a fair coin.
But the Athenian assembly, with its much higher proportion of participants, has a probability of success of only around 65%. That is disappointing. It would seem that practical democracy is a bit of a sham. Direct democracy through an assembly and our parliamentary system, both fail to provide high probabilities of channeling the will of the people. The faith the democratic community (modern or ancient) puts in its model seems misplaced.
A perhaps more inspiring picture emerges when we consider the distributions across the range of majorities. For large numbers of N there are very large numbers of cases of slim majorities, but the number of wide majorities grows less quickly, anchored by the fact that no matter how large N gets there is only one possible unanimous majority. In Graph 3-2 I’ve plotted the probability from the expression p(A≡M) against M/N for particular cases of N and A.
The cases of N = 7,035 and A = N/7 =1005 and the case of N = 29,145 with A = N/29 = 1005 and with A = N/5 = 5,829 are the three points that form the triangle bounding the range of the Athenian assembly in Graph 3-1. The case of N = 19,935,045 and A = 307 represents the Canadian Parliament. For the special case in which A=1, the case of the parliamentarian and the riding, the expression for p(A≡M) simplifies to p(1≡M) and is independent of N.
Graph 3-2 shows that for both the Canadian parliament and the Athenian assembly, the probability of the assembly channeling the majority opinion of the wider community is quite good for strong and even modest majorities. Both kinds of assemblies only fail to channel slim majorities well.
Of course, the slim majorities are the most numerous and so Graph 3-2 is, in a sense, focusing on the less probable higher majorities. Graph 3-3 plots the probabilities of correct channeling for the same particular cases of N and A against the probability of a majority M, rather than M/N as in Graph 3-2.
Compare the respective Canadian parliament curves on Graph 3-2 and Graph 3-3. Graph 3-2 shows that the Canadian parliament, with only 307 people representing nearly 20 million, channels well for all majorities greater than about 55%, and, though trailing, places respectably when compared to the various Athenian assemblies. By contrast, Graph 3-3 illustrates how many more slim majorities there are with increasing values of N. The areas under the curves in Graph 3-3 are the probabilities (y values) plotted for the respective assemblies in Graph 3-1.
What Graphs 3-2 and 3-3 show is what assemblies can and cannot do well. As mentioned above, when the majority opinion with the whole community would be moderate or strong, the assembly has a high probability of correctly channeling that opinion. What assemblies cannot do is channel the opinions of narrow majorities with high probability. Sometimes this may not be a problem. If it is generally understood the majority is narrow and no one knows which side of the issue is favoured by the majority, it probably doesn’t matter to the majority which side wins. But, if there is a narrow majority division on an issue that arouses intense passion on both sides, this can be a disaster for a community. The assembly cannot save the community by providing leadership, for, in these instances, the assembly’s opinion is rightly suspect. It has a low probability of being right. That is why faith is so important. If people believe in the Assembly, it may be able to navigate the democratic society through such dangerous passage. Without that belief, the issue must be settled either by a fair referendum or civil war, and in either case faith in the assembly will have been shaken, for it has just demonstrably failed.
Graphs 3-1, 3-2 and 3-3 also show how strikingly poor are the odds a parliamentarian has of holding the majority opinion of the riding he (and now sometimes she) represents. This level of representation is the foundation of representative democracy. By the preceding analysis, faith in ‘representative democracy’ as democracy seems wholly unjustified. But the analysis does not take elections into account. Now we need to consider if winning an election gives an elected representative the ability and inclination to channel majority opinion in the riding.
the party begins
At first glance this does not seem likely. Mathematically, there is no individual community member who is in the majority all the time. Such a person cannot exist. An elected representative might claim to be that way, by offering to change their opinion as required to that of the majority in their community. But how would that work? There is no mechanism that requires an elected representative keep their promises once elected. Beyond that, no one, not the representative, and not anyone in the community, would know the majority opinion. Even if the representative were sincere about the promise, there is still no way of telling if they were keeping the promise or not.
In fact, sometimes a poll is commissioned and the actions of the representative are compared with the results of the poll. This is “representative democracy” using an approximation of “direct democracy” to check its results. The approximation is rough; the poll sample is not an assembly. People in an Assembly study a question and discuss and research the matter until each is satisfied they have made the right decision. People in a poll sample are asked for their knee-jerk reaction to the question. One could argue that a free press and an engaged populace would reasonably simulate the assembly’s research phase. But clearly there are strong pressures on the greater community that contradict that.
First, people are busy. Indeed, the whole point of having assemblies is that the rest of us don’t have to be engaged in the decision-making. More important, arguing effectively has always been, and is especially today, an arcane discipline. It is our culture to be friends with people with whom we agree, and demonize those with whom we do not. If there is disagreement within the family, then in the interests of family peace, controversial topics are not addressed. Read internet comments on any political web page and you will see the level of constructive argument of which, as a polity, we are capable.
The free press cannot help. It is a business community funded by advertising. Any controversial ideas that news media business owners fear might cause people to not read or listen or watch will not be put forth. Any ideas contrary to the desires of the advertisers will be actively attacked. There is also pressure on media firms, as businesses, to have a good working relationship with government. It is simply less expensive to co-operate with government, and to promote government agendas, than it is to take a watchdog role, to challenge government statements, and bear the hardships and forego the benefits that governments hand out. And, when the interests of business and government align, the fourth estate has every incentive to be on side as well. We may believe what we wish, but the free press today does not encourage speculative debate.
And, finally, polls are usually funded by people who want a certain answer. The poll questions are leading (the term for this kind of survey is a push-poll).
With all that, elected representatives are not even bound by poll results. If a representative’s actions coincide with poll results, the representative trumpets this fact as proof they channel community majority opinion. If a representative’s actions do not coincide with a poll result, the representative will declare they do not believe the poll results reflect majority opinion (not an unreasonable hypothesis, though self-serving) and if challenged further on the matter, will declare they are showing leadership, which is to say, acting contrary to the will of the people (or what people think is the will of the people) for the people’s own good. Rarely will a representative change position because of poll results. Yet if they do, still no one really knows if the majority opinion was truly channeled.
This whole swirling flux of parliamentary uncertainty was recognized in the late 1600’s when England became a kingdom again in name at least. Parliament, out of need, had recalled the king, but parliament was, and would forever after be, the dominant partner. Yet parliament’s decisions were no more likely to channel the wider community’s opinion than would a coin toss. So, the parliamentarians of the time found a way to give their voters certainty about how the parliamentarians would behave. They invented political parties.
The problem was that Parliamentarians were free agents, free to do as they liked from the time they were elected till the next general election. Sure, they could be kicked out several years later, but it was fun till then. Voters could not discipline parliamentarians. But other parliamentarians could. Groups of parliamentarians would present voters with an agenda that comprised a set of principles and initiatives, and they promised to stick to that platform, and ride herd on any of their number who might think to break ranks. Any parliamentarian who crossed the party would have his privileges revoked, his pet projects voted down, his voice booed when he rose to speak. Life in parliament would be miserable. As enforcers, political parties brought predictability to the chaos of parliament.
Voters voted for the platform they liked, and the most popular platform determined who would win the election in each riding. The platform with a majority of supporters in parliament ruled parliament. It is the process we follow to this day. Supposedly, this process of platform-selection-by-election makes up for the comparatively poor channeling characteristics parliament displays in Graphs 3-1, 3-2, and 3-3, making ‘representative democracy’ more or less as democratic as ‘direct democracy’. But this may not be the case.
A voter looks at the platforms of the parties and chooses a platform, and hence a party, with which they most agree. But a voter might not agree about everything in the platform. There may be some elements from the platforms of other parties, on some specific issues, that may appeal to a voter more than how that voter’s chosen party’s platform handles those issues. Voter preference for a revision to a party platform is not expressible in the parliamentary system. Voters must take the bad with the good. That was a compromise that ancient Athenians never had to make.
In an effort to maximize votes for themselves, parties do what they can to dissuade voters from supporting the platforms of other parties. A platform is associated with a party, a group of people. Therefore, a valid way to discredit a platform is to discredit the people who promote it. A party’s platform, then, can have a vicarious weakness if some of the people in the party are susceptible to criticism through an unrelated front. The platform may not get enacted because the people who would enact it are not popular enough, even though the platform itself is good. If a policy was judged good in ancient Athens, there was no one faction that also had to be judged good before the policy could be carried out.
A political party may “move” its platform. In moving its platform, the party adopts some of the agenda of a fairly successful rival party. The intent of this maneuver is to woo some voters away from the rival. Every vote brought from the rival party decreases the rival’s total by one and increases the maneuvering party’s total by one. If some of the voters who originally support the maneuvering party do not like the revised platform, they may choose not to vote, or to vote for a less successful party, in either case costing the maneuvering party no more than it gains. This common betrayal was a frustration that ancient Athenian democrats never faced.
But building restrictive platforms, denouncing rivals, and raiding other parties for votes, these activities carried on by successful parties should be seen for what they are. This is the ‘democracy’ part of ‘representative democracy’. The parties strategically construct and strategically move their agendas to maximize the number of seats won in elections, which, in general, involves maximizing votes. They “rally the troops” by portraying their rivals as a fools or agents of evil, in either case dangerous to the well-being of the community if they came to power, a threat to our society that must be stopped (by voting for the right party). It’s all about maximizing votes. It is our democracy.
To study the trends, analyze all the poll data, and weigh the move, then take the risk and embrace the hyperbole, this is channeling the hard way. Maybe it works as well as, or nearly as well as, having an assembly with enough randomly selected citizens to be statistically significant. If we believe we are practicing democracy, then we believe it does do about as well (the results of those three referenda notwithstanding). In some sense, we have to believe it, because a Canadian assembly with enough randomly selected citizens to be statistically significant would be an assembly of about 20 million/50 = 400,000 people. That would put us on the yellow line in Graph 3-1, about as effective at channeling as the ancient Athenians at their worst, at the bottom of the triangle in Graph 3-1. On the face of it, this proposal seems too absurd to consider. Our present system is, presumably, the only affordable option, so, making a virtue of necessity, we close our eyes to the evidence and believe that it works well.
All in all, representative democracy is a much less certain way to channel. It is also much more expensive than it at first seems. A lot of the effort party members go to is hidden from the rest of us, true, but there is also a cost we see clearly and do not count. As today’s adversaries can be tomorrow’s allies, the practice of real democracy builds trust within the democratic community. Parliamentary assemblies do not. A successful party must defend itself from the vote raids of rival parties strategically moving their platforms. Equally, a successful party must be able to undertake similar vote raids upon its rivals. To prevent supporters being drawn away and to keep supporters from drifting away, a successful party needs brand loyalty. A party needs passionate supporters, and an easy path to that voter passion, a path that no successful party can ignore, is to build in the minds of the party’s supporters a fear and loathing of the other parties, and of the other parties’ supporters. It is a sad paradox that parties must do this as they strive to maximize the level of democracy that can be attained with a parliamentary system.
On December 4, 2008, then Governor General Michaëlle Jean granted the Prime Minister’s request to prorogue Parliament so the minority Conservative government would not face a likely defeat on a motion of confidence in Parliament. Many people who supported the Conservative Party of Canada cheered this action. They were convinced that those who had not voted for the Conservative Party of Canada were a danger to our society. That those other parties comprised a majority in parliament, that those people who had voted for the other parties made up the majority in Canadian society, held no weight. Passing the reins of government to a majority coallition was portrayed as a coup by the mainstream media, who enthusiastically endorsed the closing of parliament and condemned as fools or traitors those who argued otherwise. Thus we have another paradox of Parliamentary representative democracy, that on the path to maximizing democracy we took a step towards fascism.
Here is one last twist. Although parliamentarians may meet fairly frequently, elections are held in modern times only about once every four years. In the wider community, there is not the constant testing of the democratic deal; if I loose a vote, I accept that result, and if I win, you accept that result, and then we do it all again for the next issue. Instead, the community votes once every four years, and after the vote, the winners are the winners on every issue until the next election. For the voter in the community, the cost of one loss is huge; it is a loss of four years. With first-past-the-post ridings, and especially with more than two parties, generally a minority of voters elects the majority party in parliament. How is that for irony? In Canada’s ‘representative democracy’ most voters lose.