4 ms·

The problem with the ‘no vehicles in the park’ example is that the question is intentionally vague. I feel like we do ourselves a disservice to hold this partic

by dahart 3y ago

The problem with the ‘no vehicles in the park’ example is that the question is intentionally vague. I feel like we do ourselves a disservice to hold this particular example up as a way to conclude that everyone disagrees and it’s impossible to agree and let’s throw up our hands. The way forward is to put more precision into the rule statement. And yes, there might be diminishing returns and ;it might even be impossible for every last person to agree, but that’s an actively harmful take-away if it’s possible to get 90% or 95% or 99% to agree by just being slightly more specific. A huge portion of the ‘no vehicles in the park’ questions are already answered by existing US aviation laws, for example. Another huge swatch of the questions would have been answered if the rule was ‘no motor vehicles in the park’. By intentionally withholding the actual rules, this survey doesn’t strike me as evidence that agreeing is hard, it seems like more of a trick question.

ibejoeb 3y ago
Yes, that's true. But in the act of crafting a deliberately ambiguous rule, it teaches others in the position of crafting real rules that they must be thoughtful in doing so. It is easy to phone it in and act like "do what I mean, not what I say." If you're being thorough, you could iteratively refine the rule and replay those 20-something questions to see how each iteration affects the specificity of the language. A simple rewrite like "The operation of motorized vehicles on park grounds is prohibited" makes a lot of those questions moot, but there are still unhandled cases. I see how it can be a useful exercise, especially for inexperienced rule makers.
- dahart 3y ago
  Yes, exactly! ‘no vehicles in the park’ should be used as a teaching moment, and not as an example of why we can’t or shouldn’t try harder. It would be really interesting if the survey had been A/B tested to find the minimum additional verbiage that brought maximum agreement. Don’t most do people know at some level that laws are hard to make and often have corner-cases, loopholes, or unintended consequences? We don’t conclude that laws are impossible to make perfect, we simply keep refining the laws, right?
  - Hasu 3y ago
    > We don’t conclude that laws are impossible to make perfect, we simply keep refining the laws, right? Uh, we definitely do conclude that it's impossible to make the laws perfect. And we have an entire institution with the responsibility for refining laws (the judiciary), and they disagree with each other all the time. This is very strong evidence that it is, in fact, impossible to get people to agree on what's allowed. With a dedicated institution, you can get close enough for government work (literally) but people are still arguing over the meaning of the Constitution even after hundreds of years of refinement. I think programmer-types think, "Well a detailed-enough specification of all the rules is just like a computer program" and sure, it is. Except that computer programs are written in unambiguous language and interpreted by machines that have perfect understanding, and we still write imperfect programs that don't do what we want all the time.
    - dahart 3y ago
      So you’re saying we should stop refining the laws because it’s impossible?
      - Hasu 3y ago
        I'm not sure how you interpret "It is impossible to make laws perfect and unambiguous for all humans everywhere" as "We should just give up on refining laws because we can't get to perfection". We refine enough to have a working system, which is nowhere close to having no disagreement.
        dahart 3y ago
        What’s the point of saying something’s impossible? Why frame it that way? Should we keep trying, or no? Perfect isn’t the goal, and “impossible to make perfect” isn’t a useful summary, right? It seems like that’s what you saying, and I agree. The only reasonable question is whether it’s possible to improve, and the answer is yes it is possible to improve.
        Dylan16807 3y ago
pixl97 3y ago
> it seems like more of a trick question. Eh, I don't exactly but this may be because I've hung around lawyers too much in my life. Everything is argued at the border between black and white, the size of the grey zone can differ pretty significantly. In addition your upbringing can have a significant impact on what you view as black or white in the first place. There is a reason that in the US alone there are millions of pages defining laws and rules. It's the fact we cannot make simple rules that cover the majority of human behaviors. Humans will always push the grey zone either by intention or complete accident.
- dahart 3y ago
  Right! We keep arguing about the gray areas, and in the mean time we make headway in refining them and defining what is black and what is white, little by little, right? Lawyers don’t usually argue over things that are well defined though, unless they’re trying to make a contrived point. They treat the ‘no vehicles in the park’ example differently than most people. The example invites assumption. It would be entirely different if the example said the park’s pathways are public sidewalks and as such are under the same laws. Then it would be more obvious that the survey results isn’t exhibiting disagreement, it’s exhibiting lack of knowledge and incorrect assumptions of the existing laws. You’re right we have a lot of laws, but I’d argue the laws aren’t actually that complex, framing it as millions of pages might not be accurate for which laws apply to me. Each state has their own laws, the majority of which are similar. For any given topic, say driving in traffic, most of it is covered by a modest number of rules & pages. The large number of total pages in US law is more about the large number of topics and behaviors and technologies. It’s true that a simple shared rule cannot cover driving a car, copying a movie, getting into a fight, trading stocks, and food safety all at the same time. Edit: for fun, I just looked it up and the US federal code is ~60k pages. It’s large, but it’s easy to see why looking at the list of 54 titles https://uscode.house.gov/ https://uscode.house.gov/ I don’t see page counts for any state codes I’d tried to lookup, but I’m not surprised if it’s similar. On one hand, to your point, it’s a lot of laws. On the other hand it’s almost surprising that state and federal laws combined are less than a million pages, considering the breadth and enormity of the topics that is covering hundreds of millions of people.
_xivi 3y ago
> A huge portion of the ‘no vehicles in the park’ questions are already answered by existing US aviation laws It's just an illustration of how a seemingly simple rule at first glance is open to be interpreted differently. Sure you can add nuances and be more specific about edge cases, but every edge case has its own edge cases, you're not solving the problem by this strategy, only digging a bigger hole and increasing overhead for yourself. The author could've used any number of rules and laws and still came up with edge cases where the audience would to fail to reach an unanimous decision. The US law is full of loopholes, adding more rules isn't a solution, it merely change the question. If you commit to that strategy, of eliminating ambiguity by adding more rules, you'd be playing catch-up forever.
- dahart 3y ago
  What I’m objecting to is the binary goalpost of ‘unanimous’. We don’t need unanimous in order to statistically improve the level of agreement. I disagree that covering edge cases makes the hole bigger. The number of edge cases, and the level of agreement are two different things, and it’s possible to improve agreement even if the number of edge cases goes up. > The author could’ve used any number of rules and laws and still came up with edge cases And yet, the author could have stipulated motor vehicles and a 400 ft FAA air ceiling boundary to the park, and most of the trick questions in that survey that generated disagreement would suddenly flip to everyone answering them correctly. I don’t understand what your suggested alternative is. We already continuously refine our laws to cover edge cases. That is the takeaway for the contrived ‘no vehicles’ survey. Are you saying we shouldn’t be improving laws because it’s impossible, or what?
  - _xivi 3y ago
    > We already continuously refine our laws to cover edge cases. Exactly the author's point. There's no set of rules that clear every possibility. You just keep adapting, and the other side will keep involving. > the author could have stipulated motor vehicles and a 400 ft FAA air ceiling boundary to the park Then there'd been new 'trick' questions with new edge cases where people are split too. I think you might have misunderstood the expierement, did you get to the end where the author explain his intention? This set of questions aren't a challenge, just a game to illustrate a point.
boredhedgehog 3y ago
> A huge portion of the ‘no vehicles in the park’ questions are already answered by existing US aviation laws, for example. Which evolved by arguing about rules, because landowners had the reasonable and probably correct expectation that their property rights extended upwards to infinity. > Another huge swatch of the questions would have been answered if the rule was ‘no motor vehicles in the park’. It's trivial that changing the rule to make it narrower changes the outcome. But then people won't agree about whether the change should be made. How annoying a vehicle is isn't determined by its method of propulsion.
- ronald_raygun 3y ago
  This is why the tank example is the most interesting one. Because clearly by any standard it is a vehicle in the park, but clearly its function in the park is completely out of scope of what the rule is trying to achieve.

by dahart 3y ago

ibejoeb 3y ago

Yes, that's true. But in the act of crafting a deliberately ambiguous rule, it teaches others in the position of crafting real rules that they must be thoughtful in doing so. It is easy to phone it in and act like "do what I mean, not what I say." If you're being thorough, you could iteratively refine the rule and replay those 20-something questions to see how each iteration affects the specificity of the language. A simple rewrite like "The operation of motorized vehicles on park grounds is prohibited" makes a lot of those questions moot, but there are still unhandled cases. I see how it can be a useful exercise, especially for inexperienced rule makers.

dahart 3y ago
Yes, exactly! ‘no vehicles in the park’ should be used as a teaching moment, and not as an example of why we can’t or shouldn’t try harder. It would be really interesting if the survey had been A/B tested to find the minimum additional verbiage that brought maximum agreement. Don’t most do people know at some level that laws are hard to make and often have corner-cases, loopholes, or unintended consequences? We don’t conclude that laws are impossible to make perfect, we simply keep refining the laws, right?
- Hasu 3y ago
  > We don’t conclude that laws are impossible to make perfect, we simply keep refining the laws, right? Uh, we definitely do conclude that it's impossible to make the laws perfect. And we have an entire institution with the responsibility for refining laws (the judiciary), and they disagree with each other all the time. This is very strong evidence that it is, in fact, impossible to get people to agree on what's allowed. With a dedicated institution, you can get close enough for government work (literally) but people are still arguing over the meaning of the Constitution even after hundreds of years of refinement. I think programmer-types think, "Well a detailed-enough specification of all the rules is just like a computer program" and sure, it is. Except that computer programs are written in unambiguous language and interpreted by machines that have perfect understanding, and we still write imperfect programs that don't do what we want all the time.
  - dahart 3y ago
    So you’re saying we should stop refining the laws because it’s impossible?
    - Hasu 3y ago
      I'm not sure how you interpret "It is impossible to make laws perfect and unambiguous for all humans everywhere" as "We should just give up on refining laws because we can't get to perfection". We refine enough to have a working system, which is nowhere close to having no disagreement.
      - dahart 3y ago
        What’s the point of saying something’s impossible? Why frame it that way? Should we keep trying, or no? Perfect isn’t the goal, and “impossible to make perfect” isn’t a useful summary, right? It seems like that’s what you saying, and I agree. The only reasonable question is whether it’s possible to improve, and the answer is yes it is possible to improve.
      - Dylan16807 3y ago

pixl97 3y ago

> it seems like more of a trick question. Eh, I don't exactly but this may be because I've hung around lawyers too much in my life. Everything is argued at the border between black and white, the size of the grey zone can differ pretty significantly. In addition your upbringing can have a significant impact on what you view as black or white in the first place. There is a reason that in the US alone there are millions of pages defining laws and rules. It's the fact we cannot make simple rules that cover the majority of human behaviors. Humans will always push the grey zone either by intention or complete accident.

dahart 3y ago
Right! We keep arguing about the gray areas, and in the mean time we make headway in refining them and defining what is black and what is white, little by little, right? Lawyers don’t usually argue over things that are well defined though, unless they’re trying to make a contrived point. They treat the ‘no vehicles in the park’ example differently than most people. The example invites assumption. It would be entirely different if the example said the park’s pathways are public sidewalks and as such are under the same laws. Then it would be more obvious that the survey results isn’t exhibiting disagreement, it’s exhibiting lack of knowledge and incorrect assumptions of the existing laws. You’re right we have a lot of laws, but I’d argue the laws aren’t actually that complex, framing it as millions of pages might not be accurate for which laws apply to me. Each state has their own laws, the majority of which are similar. For any given topic, say driving in traffic, most of it is covered by a modest number of rules & pages. The large number of total pages in US law is more about the large number of topics and behaviors and technologies. It’s true that a simple shared rule cannot cover driving a car, copying a movie, getting into a fight, trading stocks, and food safety all at the same time. Edit: for fun, I just looked it up and the US federal code is ~60k pages. It’s large, but it’s easy to see why looking at the list of 54 titles https://uscode.house.gov/ https://uscode.house.gov/ I don’t see page counts for any state codes I’d tried to lookup, but I’m not surprised if it’s similar. On one hand, to your point, it’s a lot of laws. On the other hand it’s almost surprising that state and federal laws combined are less than a million pages, considering the breadth and enormity of the topics that is covering hundreds of millions of people.

_xivi 3y ago

> A huge portion of the ‘no vehicles in the park’ questions are already answered by existing US aviation laws It's just an illustration of how a seemingly simple rule at first glance is open to be interpreted differently. Sure you can add nuances and be more specific about edge cases, but every edge case has its own edge cases, you're not solving the problem by this strategy, only digging a bigger hole and increasing overhead for yourself. The author could've used any number of rules and laws and still came up with edge cases where the audience would to fail to reach an unanimous decision. The US law is full of loopholes, adding more rules isn't a solution, it merely change the question. If you commit to that strategy, of eliminating ambiguity by adding more rules, you'd be playing catch-up forever.

dahart 3y ago
What I’m objecting to is the binary goalpost of ‘unanimous’. We don’t need unanimous in order to statistically improve the level of agreement. I disagree that covering edge cases makes the hole bigger. The number of edge cases, and the level of agreement are two different things, and it’s possible to improve agreement even if the number of edge cases goes up. > The author could’ve used any number of rules and laws and still came up with edge cases And yet, the author could have stipulated motor vehicles and a 400 ft FAA air ceiling boundary to the park, and most of the trick questions in that survey that generated disagreement would suddenly flip to everyone answering them correctly. I don’t understand what your suggested alternative is. We already continuously refine our laws to cover edge cases. That is the takeaway for the contrived ‘no vehicles’ survey. Are you saying we shouldn’t be improving laws because it’s impossible, or what?
- _xivi 3y ago
  > We already continuously refine our laws to cover edge cases. Exactly the author's point. There's no set of rules that clear every possibility. You just keep adapting, and the other side will keep involving. > the author could have stipulated motor vehicles and a 400 ft FAA air ceiling boundary to the park Then there'd been new 'trick' questions with new edge cases where people are split too. I think you might have misunderstood the expierement, did you get to the end where the author explain his intention? This set of questions aren't a challenge, just a game to illustrate a point.

boredhedgehog 3y ago

> A huge portion of the ‘no vehicles in the park’ questions are already answered by existing US aviation laws, for example. Which evolved by arguing about rules, because landowners had the reasonable and probably correct expectation that their property rights extended upwards to infinity. > Another huge swatch of the questions would have been answered if the rule was ‘no motor vehicles in the park’. It's trivial that changing the rule to make it narrower changes the outcome. But then people won't agree about whether the change should be made. How annoying a vehicle is isn't determined by its method of propulsion.

ronald_raygun 3y ago
This is why the tank example is the most interesting one. Because clearly by any standard it is a vehicle in the park, but clearly its function in the park is completely out of scope of what the rule is trying to achieve.