Referring to my post “Only One Way to Be ‘President of All Syldavians,’” I gave ChatGPT 4 the following instruction:
To illustrate the article at https://www.econlib.org/only-one-way-to-be-president-of-all-syldavians/, generate an image showing two individuals fighting: one wants to force the other to drink only beer as himself; the other wants to force the first one to drink only wine as himself.
The bot came up with interesting drawings, illustrating the hidden face of democratic politics as we know it, that is, as a method for imposing to all the choices of some—at best, of the majority of the electorate. One of the bot’s images appears on the featured images of this post, and is reproduced below. Our friendly chatbot did not do badly, notwithstanding the little AI glitches in the drawing.
How to democratically choose between beer and wine
READER COMMENTS
Peter
Mar 8 2024 at 10:32am
Given the entire apparent 100% beer drinking electorate, might it secretly been a commentary on tyranny of the minority, or was that the majority. Guess the framing depends on who wins. 🙂
I find the wine syndicatilist like flag amusing as well, likewise the ying-yang beer.
Pierre Lemieux
Mar 8 2024 at 10:48am
Peter: I know your post is humorous, and we need humor. But let me still mention that whether it is the minority (even 1%) or the majority (even 99%) does not really matter. Moreover, “the majority” is a chaotic notion: see my short post on the Condorcet paradox at https://www.econlib.org/electorate-may-want-both-afghan-pullout-and-non-pullout/.
Peter
Mar 8 2024 at 3:01pm
Read that years ago but it’s a good refresher for those that haven’t. On a serious note though, and something that bothered me then as well, is that it seems to break down in a situation where in fact there are only two preferences such as this, i.e. it’s only a workable paradox in a model with 3+ preferences correct? And of which I will easily concede is most cases in reality hence that edge case probably moot.
Craig
Mar 8 2024 at 3:32pm
“On a serious note though, and something that bothered me then as well, is that it seems to break down in a situation where in fact there are only two preferences such as this, i.e. it’s only a workable paradox in a model with 3+ preferences correct?”
I think the two choices model simplification simply works to illustrate a more complex point off course. Often we think of choices like here in a circumstance where we see two different products that have a broader similarity, ie apples aren’t oranges, but both are fruit, but still if you spend money on beer you’re really choosing beer over anything else you could possibly spend money on, whether its wine or a quart of oil or anything really. In microeconomics I often recall indifference curves as between two products as well to illustrate marginal utility/diminishing marginal returns and how you might allocate $10 as between apples and oranges.
I still would suggest Pierre making a cogent argument for consumer sovereignty!
Peter
Mar 8 2024 at 4:56pm
I think you are missing the nuanced point I was making. His article was saying you get paradoxical results with multiple preferences yet only two choices and that seems true but I’m just trying to confirm if I’m correct in case where you have exactly two preferences and two choices, the paradox doesn’t hold.
I’m conceding people smarter and more educated than me have probably thought that through already hence it’s not a novel observation in my part hence curious the answer.
Pierre Lemieux
Mar 9 2024 at 12:52pm
Peter: I am not sure I understand your question. Think in terms of the number of alternatives and the number of voters (each voter with potentially different preferences among the alternatives). With only two voters, any social/collective choice must be unanimous or it will satisfy nobody. With more than two voters and only two alternatives, it is always the same one that will be chosen (assuming of course, as we do in all cases, that individual preferences do not change). The more voters or the more alternatives (>3), the higher the probability that the Condorcet paradox will occur (as more generally demonstrated by Arrow). As I write in my “The Impossibility of Populism“:
steve
Mar 8 2024 at 4:42pm
The guy on the right appears to have 2 left hands and it would be awkward to hold a beer in the right hand as portrayed. Is this the woke people again screwing up the art?
Steve
Peter
Mar 8 2024 at 4:59pm
No, ChatGPT is infamous for its inability to draw hands or arms when asked to draw groups of people resulting in what people amusingly call spaghetti arms so you just have to suspend disbelief on those.
Been a problem since v2 at least, TBH I’m surprised to see it’s still in v4.
Pierre Lemieux
Mar 9 2024 at 1:00pm
Peter and Steve: I was indeed surprised to see how often this (known) sort of glitches appears in this particular drawing. Four-finger hands, double hand, etc. And look how the well-endowed woman on the left is standing on the floor.
Jon Murphy
Mar 9 2024 at 1:10pm
Identifying AI follows the same rules as identifying fae and other supernatural beings:
“Count the fingers, count the knuckles
count the teeth, check the shadows”
john hare
Mar 8 2024 at 7:27pm
To me the problem illustrated is only two choices I don’t care for. Single digit beers per year and I could do without them easily. I don’t care for the idea that others would like to dictate my preferences.
Pierre Lemieux
Mar 9 2024 at 1:14pm
John: You have a point, but I am not sure what it is. Individuals who loved East Germany did not object to the Berlin Wall. Yet, there were certainly exceptions of different sorts. If the people who would have left had “cleansed” society (to speak like Putin ) from those I don’t like, I would have opposed the Berlin Wall. If all cobblers would have escaped absent the Wall, I would have benefited much as the only remaining cobbler (but not as a shoe consumer, of course)…
john hare
Mar 10 2024 at 3:51am
I am not happy with people voting on my choices. If I don’t like either, I shouldn’t be forced to accept what a majority happens to like. If I don’t want to drink either, a vote doesn’t make me accepting of it. Same with Democrat/Republican when I don’t care for either at the moment. Or the high regulations supported by so many that restrict the things I can do that don’t hurt others. I’m not sure exactly where my point is except that I’m often not part of the herd as i see it.
Pierre Lemieux
Mar 10 2024 at 11:15am
John: OK, I see now. Your point becomes clear and very defensible. What misled me was your statement “I don’t care for the idea that others would like to dictate my preferences.” Perhaps “don’t care” can mean “don’t like” (which is what I now understand you meant), but I saw the expression as meaning “don’t mind.” Perhaps it has an ambiguous meaning.
Thomas L Hutcheson
Mar 10 2024 at 9:51am
The real paradox here is to make a choice when none is needed. Those that prefer wine can have wine. Those that prefer beer can have beer. The problem arises when a choice is necessary as when dealing with an externality. Each person cannot chose their own Pigou tax rate.
Jon Murphy
Mar 10 2024 at 10:54am
No? Lots of evidence (starting with Coase’s refutation of Pigou in 1960) shows that people, indeed, can and indeed do.
Your point may hold for public goods, but not for externalities
Pierre Lemieux
Mar 10 2024 at 11:03am
Thomas: What if the real paradox stems from the very concept of externality? Consider the following statements: Each person cannot have his own distribution of income. Each person cannot have his own system of privileges. Each person cannot have his own God (or his own eternal life: there is just one for all). Each cannot have his own dictator. (Note that a system of equal individual liberty does not raise such paradoxes.)
Comments are closed.