What makes someone real?

I’ve been listening to (and enjoying) the band Savlonic recently, and it’s prompted me to revisit some questions that have been at the back of my mind recently.

See, the interesting thing about Savlonic (other than that they do fun music) is that it’s a band consisting entirely of fictional characters. The band members, Roscoe, Evangeline and Kandi, are not actually real people. They’re animated characters.

But the question, to me, is what does this actually mean? Why aren’t they real?

The band is unquestionably real in that it performs the function of a band: There are songs by Savlonic. AlphaGo may not be a person, but it plays Go and is thus a Go player. Savlonic may not be made of people, but it creates and performs songs and is thus a band.

Two of the individual band members have always mapped pretty closely onto real life people: Roscoe is voiced by Weebl and Evangeline by Sarah Darling. The third, Kandi, started out “life” entirely unvoiced and when singing was just a pitch shifted version of Sarah. These days Kandi is voiced by Katt Wade.

So what makes the members of the band fictional as opposed to just stage names? It’s hardly unusual to go by a stage name, and many (possibly almost all) peoples’ public personas are very different from their private personas. Do we consider stage personas to be fictional characters? What’s the difference between a full blown stage persona and someone just acting a bit different in public?

(Note: This is different from e.g. an actor in a movie. Luke Skywalker does not fulfill the role of an actual Jedi in the same way that Roscoe, Evangeline and Kandi form the role of a real band member).

The big thing that makes them fictional rather than stage personas is that as far as I know, that’s what the actors and characters say they are.

I’m not sure how I feel about self-declaration as the sole determiner of reality though. It’s probably necessary, but is it sufficient? If we stuck a voice box on AlphaGo that shouted “AlphaGo is alive! No disassemble!” would AlphaGo now be a real person? It would certainly simplify the Turing test if so.

The Turing test won’t work as a determiner either: If we wanted to give Evangeline the Turing test it would be pretty easy – we’d just be talking to Sarah, who is probably quite good at convincing you that she’s a real person.

Is the fact that Sarah would be lying when she tells you her name is Evangeline the determiner of whether Evangeline is real? What’s a real name anyway? What if the character were called Evangeline because that’s a nickname Sarah often goes by? (To the best of my knowledge it is not). Is it the fictional biography? If your long interview never really touched on any biographical details would that somehow not make it a valid test of reality? What distinguishes a fictional person from a real person who is just lying?

I don’t really have any conclusions, except to note that “real person” seems to be a fuzzier concept than might naturally be assumed. I do think Savlonic fall on the “not actually real people” side, but I’m not entirely sure why I think that.

Against Dutch Book Arguments

Among my general family of weird niche opinions is that I don’t believe that the common assumption made in decision theory that preferences should be transitive is a valid one. This isn’t a post arguing in favour of my opinion. This is instead a post arguing against one of the most common arguments that preferences should be transitive: The dutch book/money pump argument. As such I think this post should be valid regardless of whether transitive preferences are good or bad.

Transitivity of preference is this: Suppose you prefer A over B and B over C, then you should also prefer A over C.

What “Prefer” mostly means in this context is “If I were to give you a choice between two options, which would you pick?”

(if you disagree with my above definitions, do say. In general please nitpick this post as much as you like. I really do want to understand people’s beliefs on this subject)

The argument goes like this:

Suppose you prefer A to B, B to C and C to A.

Given that these preferences are strict, there must be some (possibly very small) amount of money \(\epsilon\) such that you prefer A to B + \(\epsilon\), B to C + \(\epsilon\) and C to A + \(\epsilon\) (aside: I don’t entirely buy this but don’t have any strong objections to it and will leave it as mostly unquestioned).

Now suppose you have C. I offer you B in exchange for you giving me C + \(\epsilon\). Now you have B. I offer you A in exchange for B + \(\epsilon\). Now you have A. I offer you C in exchange for A + \(\epsilon\). Now you’re back where you started with C but you’ve paid me \(3 \epsilon\) for the privilege.

(Again, if you feel like I have oversimplified this argument or expressed it badly, please say).

I don’t think this argument holds water, for a number of reasons.

The first is that it does not actually model the sort of preference we started with: The question “Given A or B, neither of which you currently have, which would you prefer to acquire?” is fundamentally a different question to “If you already have one of A or B, would you swap it for the other?”.

In particular, you should expect the latter to exhibit a status quo bias: A greater reluctance to switch from B to A than to pick A over B when presented with both options.

And a sufficiently large status quo bias could potentially wipe out intransitivities found in the original preference relation, so it may be that the original preference relation is intransitive while this new sort of “Would you switch?” relation is not.

You can argue that these two relations should be the same, and that a status quo bias is intrinsically irrational, but I’m not sure that’s valid. A status quo has a lot going for it: You know how things work, you’re experienced with working within it, etc. Things can and should be improved, but in general the burden of proof is higher for comparing something with the status quo than when comparing two new things, and I think that’s quite reasonable.

Lets assume for the sake of the argument that we replace our notion of preference with this new one: I prefer A to B if when I have B and someone offers me A for it I will choose to switch.

The next question is then: How is this preference actually being elicited? I think when people talk about this sort of preference process they imagine some sort of perfect copy of you is asked for its preferences and then wiped out of existence, so you are left with no memory of the question. This might be thought of as a “static preference”: It is your preference as captured at some snapshot point in time.

But that’s not what’s happening in the dutch book scenario: In this scenario you do have memory. This is a set of dynamic preferences being elicited by a series of questions where you know what has happened before. So it could be that your static preferences are intransitive but you will never exhibit intransitive dynamic preferences because you have a rule that if you notice you’re being dutch booked you stop trading in the resource.

This can be quite rational. I’ve previously argued that any agent for whom solving NP hard problems isn’t free will engage in this sort of path dependence where the exhibited set of preferences depends on the order in which you ask the questions (not even which questions you ask! Just the order).

Finally, supposing that for whatever reason we’ve decided to actually exhibit intransitivity even in this dynamic preferences case. Then, yes, we do get dutch booked and can be used as a money pump.

This is an argument against intransitive preferences in the same way that the possibility of being attacked is an argument for never going outside.

When you argue for something, you have to argue not only that it has benefits but also that the costs outweigh the benefits, and it’s not clear to me that they do.

This is both because insisting on transitive preferences in the first place is potentially very high cost (it requires essentially global knowledge and solving NP hard problems), and also because it’s unclear what the real cost of money pumps is to an agent: Money pumps are a problem if the agent eliciting your opinions is intrinsically adversarial, but most cases for preference elicitation are going to be neutral or cooperative. The universe doesn’t hate you, and your fellow agents are probably more or less out to cooperate (life is an iterated prisoners dilemma, not a one-shot one), so the actual instances of deliberate money pumping in the wild are relatively low. If you’re playing the investment markets you should probably remove your intransitive dynamic preferences as you notice them, but it’s not really clear that they cause much harm in broader contexts and I don’t think the argument that they do has been convincingly made (I’m not sure I’ve even seen it made unconvincingly).

In conclusion, I think dutch book arguments prove much less than they are usually claimed to, and I don’t think they should be taken as having much significance for how actual agents should reason.

Why I’m in favour of proportional representation

Suppose I offer a group two options I like and most of the group hates. They vote on them, and you get the result that the majority hates the least and we go with that one.

This process is “democratic” – people voted, and got the result that they voted for – but hopefully it’s clear that I’m the one with the power here.

This is, of course, the trick to getting what you want out of democracy: Give people a heads I win, tails you lose choice and it doesn’t matter how they vote. All power rests with the people presenting the choices.

And, at least for now, we don’t have any good way of coming up with those choices except a bunch of people in a room talking to each other. There is no algorithm or voting system for taking a population with a diverse set of opinions, preferences and knowledge and automatically turning that into a coherent and concrete set of policies.

And I’m not sure there ever could be: Many of these choices are complex and require days or weeks of study to understand the ramifications, which is hard to do if you also have to hold down a full time job, and your opinion without taking that time might be very different from your opinion if you had taken it. Which should we respect?

So we’re back to talking. This is quite hard to scale up to millions of people, so this is what we have parliaments and other houses of representatives for: It reduces the problem of millions of people without enough time trying to have a conversation and come up with the choices to one which is small enough that… well at least there’s the possibility of a productive conversation.

But in order, as an individual, to have any sort of say in what policies parliament creates, you need to have someone with a voice close to yours. It doesn’t matter if you elected someone who will vote more or less the same way you want (it doesn’t even matter if you live in a direct democracy and parliament will put the bill to a referendum!): If a voice close to yours is not part of the political debate, your interests are not represented. You can vote as much as you like, but it won’t help much – you’ll just be getting the less bad option of the two someone has handed to you. You don’t just need votes, you need a voice in parliament.

And this is true at the group level too: Even if a strong voting bloc manages to consistently push for the less bad option for then, if that bloc does not have actual representation of their views in parliament, they’re never going to do better than that.

And existing non-proportional systems have consistently failed to produce that representation. Instead they take us right back to the beginning: Which of these two parties that you don’t like would you prefer to have in power?

A move to a proportional system gives you that, by giving significant groups actual representation in parliament. Instead of just getting to voting between a few choices, you get to be part of the process for creating the choices instead, and you get a voice in government as well as a vote.

Ain’t nobody here but us survivors

I think people aren’t aware enough of survivorship bias. Or, at least, even amongst people who are aware of it, I think it is hard to notice.

It comes up a lot in group situations: Companies, communities, cultures, any other assemblage of people. This is because any group of people is composed solely of the ones who did not leave the group, so any observation you make about the group is only made about the survivors.

This is particularly important when people tell you that you’ll just get used to something.

There’s a great explanation from Julia Evans (via Dan Luu) of normalization of deviance:

new person joins
new person: WTF WTF WTF WTF WTF
old hands: yeah we know we’re concerned about it
new person: WTF WTF wTF wtf wtf w…
new person gets used to it
new person #2 joins
new person #2: WTF WTF WTF WTF
new person: yeah we know. we’re concerned about it.

This is absolutely a thing that happens, but I think there’s another form:

new person joins
new person: WTF WTF WTF WTF WTF
old hands: yeah we know we’re concerned about it
new person: WTF WTF WTF WT- actually bugger this for a game of soldiers
new person leaves
new person #2 joins
new person #2: WTF WTF WTF WTF
old hands: yeah we know. we’re concerned about it.

A big problem is that if you are new person #2 you cannot tell which scenario you are in, because of survivorship bias: Either way you are going to be surrounded by people who got used to it.

It’s hard to know what to do about this, but I offer two small pieces of advice:

If you’re the new person: It’s OK to leave. The fact that however many people seem to have got used to it doesn’t mean that it’s actually sensible to try to get used to it.

If you’re the old hands: Consider what aspects of your culture arise this way. Are most people really getting used to it, or is it just that the people who didn’t get used to it left?

Trading influence: A half-baked parliamentary design

Advance warning: I’m pretty sure this is a bad idea. I’m sharing it because I think it might be an interestingly bad idea. Or maybe it makes a better game mechanic than democratic system, I don’t know.

The starting question from this was: How would you design a voting system for a parliamentary government (that is, the system that the parliament itself uses, not how they are elected) so that it encourages compromise and finding acceptable middle grounds rather than just doing straight up majority rule?

The key idea is to make in an explicit system of trading influence (to replace or augment existing implicit systems of trading influence). On questions you are mostly indifferent to, you can trade your vote away for influence on later votes.

In order to do this, we introduce a system of virtual currency representing influence. The currency is given value through an annual tithe, designed such that the amount of influence in circulation is only sufficient for, say, 80% of parliament to pay the tithe. Anyone who cannot pay the annual tithe drops out and comes up for re-election (thus creating a system of rolling elections). Some sort of additional term limit is probably required too.

Votes for or against a bill are then cast as follows:

Each representative casts one of two types of votes:

a firm vote, which is a vote for or against plus an amount of influence the representative is willing to spend to ensure that vote.
a tentative vote, which is an abstention plus an amount of influence the representative would require to cast their vote for or against.

Once the vote is cast, the auction begins. I’m a little unsure on the optimal design of the auction here. I believe the following is reasonable, but I’m pretty sure it’s not optimal (for any sense of optimality). The design is basically that of a Vickrey Auction, but with votes instead of influence.

For each side, we determine the number of purchasable votes: That is, counting from the smallest offer upwards we buy votes until we would require more influence than we have allocated left to spend. However many we bought is our set of reachable votes.

For each side we add up the number of purchasable votes with the number of firm votes cast for that side. This gives the number of accessible votes.

If both sides have the same number of accessible votes, that’s a tie. Move to tie breaking procedure (e.g. default against the bill) and nobody pays anything.

Otherwise, whichever side has the larger number of accessible votes wins the vote but must purchase votes to the tune of one more than the other side’s accessible votes (this may be zero votes) using their influence. This works by buying up votes starting from the cheapest until you’ve purchased enough votes. The cost of these votes is then distributed amongst the people paying for them using a system I described previously that basically splits the bill evenly amongst everyone who can afford it, and everyone who can’t afford the even split pays their whole budget.

I don’t think this encourages entirely honest play unfortunately – although the vickrey mechanism forces you to be honest about the number of votes you need, it’s a bit too disconnected from the actual bids (and I don’t think people are encouraged to be honest in the prices they set their vote at at all?) but it’s probably not too bad. Insert better auction design here.

Problems

There are a number of obvious problems with this and probably some more non-obvious problems with this.

The fundamental problem with using any sort of currency mechanism for voting is what you do when someone submits the same bill twice. I think this is fixable here by having a “parliamentary behaviour” notion that when you’re bought you stay bought. In order to enforce this it may be useful for MPs to declare blacklists where they refuse to buy votes from certain candidates. This complicates the calculation of how many votes are accessible significantly (I think it may turn it NP-hard. It’s certainly solvable through a mixed-integer linear program, but that starts to sound like something you don’t want in your voting procedure).

This is vulnerable to bad actors who just always put their vote up for sale, taking currency out of circulation and thus increasing turnover in the parliament. By doing this preferentially only when the makeup of the parliament is not to your liking it might be possible to stack the parliament that way.

If you have either term limits or are obviously not going to be re-elected because it’s coming up to tithe time and you don’t have even close to enough influence it creates an incentive to just go wild on your spending. Of course term limits already do that.

It’s unclear how this interacts with party systems and party whips. I’d like to avoid that by having more fluid parties and no party whips, but I think reality may work against me on that one…

Advantages

Well I think it’s interesting, and interesting is obviously good, right?

But the concrete advantage of it is that by creating a system of trading influence fairly explicitly it creates a much firmer platform for small parties who are really mostly interested in a core set of issues to get influence over those issues while achieving a sort of “compromise liquidity” for the larger more influential parties.

Moreover, in systems with small parties you already have this sort of influence trading happening, only it’s behind the scenes and thus illegible to the public. By making it official and part of the system you get greater public oversight into what is going on.

Over all, I think the disadvantages beat the advantages by some margin, but maybe that will give the undecided middle some interesting influence that they can spend on a better system later…