Backyard cat fight shows Schelling points preexist language
Two cats fighting for control over my backyard appear to have settled on a particular chain-link fence as the delineation between their territories. This suggests that:
Animals are capable of recognizing Schelling points
Therefore, Schelling points do not depend on language for their Schelling-ness
Therefore, tacit bargaining should be understood not as a special case of bargaining where communication happens to be restricted, but rather as the norm from which the exceptional case of explicit bargaining is derived.
Summary of cat situation
I don’t have any pets, so my backyard is terra nullius according to Cat Law. This situation is unstable, as there are several outdoor cats in the neighborhood who would like to claim it. Our two contenders are Tabby Cat, who lives on the other side of the waist-high chain-link fence marking the back edge of my lot, and Tuxedo Cat, who lives in the place next-door to me.
| |
| Tabby's |
| yard |
| (A) |
------+...........+--------
| (B) |
| | Tuxedo's
| My yard | yard
| |
|
| -- tall wooden fences
.... short chain-link fence
In the first incident, I found the two cats fighting in various locations around my yard. Eventually Tuxedo emerged victorious, and Tabby fled back over the fence into his own yard. Tuxedo looks to be younger and more robust than the comparatively elderly and scrawny Tabby, so this outcome was not surprising.
In the second incident, Tabby and Tuxedo had a staredown through the fence, where they spent almost an hour meowing loudly at each other and refusing to budge from their respective seats (Tabby at A, and Tuxedo at B) a few inches apart, with the fence in between. This appeared to be a symbolic extension of the physical fight from earlier—whichever cat were to retreat first would be taken to concede the territory on the other side of the fence. That is, if Tabby retreats from A while Tuxedo is still at B, then this means Tabby will no longer enter my yard; and if Tuxedo retreats from B while Tabby is still at A, then Tuxedo will not attempt to take the fight into Tabby’s yard.
Scratching is a continuation of meowing by other means. -- Von Clawswitz
(Because Tuxedo had already established a provisional claim to my yard in the first incident, this seemed more plausible than the other possible interpretation of the staredown, i.e. that the retreating cat would be conceding the territory on the same side of the fence.)
As it so happened, Tuxedo backed down first. This surprised me—I was expecting Tabby to retreat, because Tuxedo had already proven that he could win in a fight if it came to that. However, I had not accounted for the home-field advantage, or conversely the importance of leaving your enemy a way out. I now realize that this confrontation had much higher stakes for Tabby than for Tuxedo. Tabby will fight more desperately in his own yard than either cat did in my yard, because if he can’t even defend his own yard, he’ll have nowhere else to go. Therefore, in a fight over Tabby’s yard, Tuxedo is likely to take considerable damage even if he does ultimately prevail, and the risk isn’t worth the reward. And it seems that both cats figured this out by meowing at each other for an hour.
Leave your enemy a litterbox to piss in. -- Sunbeam Tzu
Why the fence?
The settlement seems to have held, and since the second incident I have not seen either cat venturing onto the “wrong” side of the fence. What’s funny about this is that the chain-link fence is a more-or-less arbitrary line—again, it’s only about 3 feet tall, and any cat can easily see through it or climb over it.
But it’s clear that the fence is the only plausible Schelling point in the contested area, and so it sticks. Dividing the territory along any other line would be impermissible in Cat Law.
If animals have Schelling points, then...
This implies that negotiation is conceptually prior to language. When Thomas Schelling introduces the topic in The Strategy of Conflict (ch. 3, pp. 67-9), he explains why the study of tacit bargaining (exemplified in seemingly artificial scenarios where the parties are prevented from communicating with each other) is relevant to the real world:
The concept of “coordination” that has been developed here for tacit bargaining does not seem directly applicable to explicit bargaining. There is no apparent need for intuitive rapport when speech can be used; [...] Yet there is abundant evidence that some such influence is powerfully present even in explicit bargaining. [...] The “obvious” place to compromise frequently seems to win by some kind of default, as if there were simply no rationale for settling anywhere else.
But now it seems that Schelling doth protest too much—in fact, tacit bargaining ought not to be understood as a special edge-case which is relevant only insofar as it approximates the more general case of explicit (verbal) negotiation. Quite the contrary: The fact that animals can negotiate and discern Schelling points, despite not having language, helps us understand what we humans are doing when we talk about things such as “ownership”, “fairness”, and other value-laden concepts, whose lack of obvious real-world referents is otherwise apt to engender no end of confusion among various rival meta-ethical philosophies.
One has to have a reason for digging one’s claws in somewhere; at least, “If not here, where?”—Tortoise Shelling
if you like to think about this kind of thing, I recommend actually reading Schelling’s “the strategy of conflict”; I thought it had a bunch of interesting points that haven’t made their way into the water supply.
A schelling point is an arbitrary default choice converged upon without communication when agreement is needed more than correctness. A territorial border between animals is an extremely non-arbitrary result of often very thorough tests of relative strength and communications of will. Animal borders are opposite to schelling points.
Borders between human territories are pretty arbitrary, we don’t really have the kind of bounded conflict that can produce a relative strength estimate any more (some of us used to), and most of us engage in antinductive commitment races by propagandising mythic histories about the legitimacy of our land claims (I don’t believe in that shit though, personally). The present order truly seems to be satisfied with schelling points for borders, it doesn’t matter what you choose, as long as we can agree, and never disagree, and whatever we agree about is the true border.
But animal borders aren’t arbitrary, they’re constantly renegotiated. The negotiations may be partly tacit, but there’s nothing whimsical or symbolic about the outcomes. The animals know where the resources are, they know how much they want them, they know their neighbors, they know how often the neighbors come to check their border so they can estimate the amount of pressure they face, and they know the risks of getting into a fight with their neighbor, so they’re able to make really pretty rational calculations to decide where the borders are.
If that’s the case, then isn’t it an astonishing coincidence that the negotiated boundary just so happens to coincide with the fence?
Some experiments that would more clearly test the use of Schelling points:
If instead of a fence we have an even more trivial boundary marker, like a line of 4-inch-wide curb stones, do the animals still use that as their boundary? (Prediction: Yes)
If the boundary marker is surreptitiously moved slightly to one side, do the animals respect the new location, or do they continue using the old location? (Prediction: New location, until the deviation from the original line becomes significant, at which point the disfavored party will breach it.)
If the boundary marker is removed entirely, do the animals start fighting again? (Prediction: Yes)
Not a coincidence, there are practical reasons borders end up on thresholds. A sort of quantization that happens in the relative strength calculation. Two models:
Simple model: You could say the true border can be defined in terms of the amount of time or effort it takes to get there from the cat’s houses. It takes a certain amount of time and effort to get over a fence, so if the true border is (from tuxedo’s house) between distance tud + yd and tud + yd + fd, then the border in practice will end up being exactly on the fence, because you can’t put a border halfway up the fence, or the situation would look the same if you did.
A more accurate model: The border is measured in terms of how hard it is to defend a space from being used by the other side. I’d guess that cats become vulnerable to attack when they mount a fence (same as humans crossing a river), either coming or going, so extending your territory beyond the fence is difficult. If your strength is higher than the amount of strength it takes to defend everything before the threshold, but lower than the amount it takes to cross the threshold and then defend some on the other side, then the border will be exactly on the threshold.
I’m pretty sure I’d predict no for 1. Cats don’t seem to care about that stuff.
For 2, I’m not sure, if there were a hole in the fence, I’d expect confrontations to happen there because that’s a chokepoint where a cat could get through safely if the other one wasn’t standing on the other side, and maybe the chokepoint is a vulnerability threshold, too. Chokepoints are thresholds for projectile combat (because when you come through the defender sees you immediately but you don’t spot them until they start shooting), cats may be partly characterizable as stealth projectiles.
Also worth noting is that, eg, dogs, they engage in “boundary aggression” at things like fences, but experiments show that they’re doing it for the love of the game. If you remove the fence, hostilities cease. Cats may have some of this going on as well. They may on some level enjoy yelling and acting tough while being in no risk of having to actually fight.
3: Yeah, but because it makes the relative strength calculation harder. A fence is a blessed device that allows cats to get a good look at each other without engaging. I wish humans had something like that. (A hole in a fence may also be a good device for this)
Also with eg dog territory, the boundary markers aren’t arbitrary—presumably the reason dogs piss on trees & lampposts, which are not physical thresholds, is (a) they provide some protection for the scent against being removed eg by rain; (b) they are (hence) standard locations for rival dogs to check for scent, rather than having to sniff vast areas of ground; ie they are (evolved) Schelling points for potential boundary markers.
(Walls are different as they are both potential boundary markers and physical thresholds.)
Do we know whether wolves really treat scent marks as boundary markers.
Some confusing things about wolf territoriality is they frequently honestly signal their locations through howling while trying (and imo failing?) to obfuscate their number in the way that they howl.
I don’t know; I had assumed so but maybe not
Most of the time people say “Schelling point” they mean this. Maybe it would be better to call it a Schelling fence, but even that post claims that it is a Schelling point. I suspect that you can reframe it to make it true Schelling point, such as the participants coordinating to approximate the real game by a smaller tractable game, but I’m not sure.
It seems to me, it is advantageous for the animals to fight for the territory up to a certain degree, where plus-munis delta territory does not justify more fighting, so they sort of agree on some point, be it a clear small fence in this case, or, what I would imagine happens in nature, a somewhat broad imaginary line.
So both of your points stand.
This doesn’t even necessarily need cat-level intelligence. Cicada-killer wasps will adopt effectively arbitrary landmarks set out by humans as territorial markers: https://www.sciencedirect.com/science/article/abs/pii/S0003347299911338 . Here, it lowered fighting as much as 80%.
A more readable version perhaps: https://www.theatlantic.com/national/archive/2013/07/the-cicada-killers-are-coming/277688/
Sorry about all the paywalls. Of course you could argue that in the case of wasps there might be a hard-coded instinct to find a random line in the sand (so to speak) and use it as a territory definition.
Great post, and great cat philosopher quotes :)
Great text!
I think the generalized version is also true:
- Game Theory is a spirit close to our world, eager to be summoned. Situations with short game theoretic description are likely on priors once evolution created anything pseudo-agentic.
- Language is cool, but is much more narrow than communication. What it brings to the table is subtle, and is rarely necessary for the game theoretic scenario in question.
I would love to see a plot where the dots are geographic regions and the x axis is density of rivers and the y axis is number of wars since 1700
Relevant to this is that reinforcement learning models can ‘negotiate’ with each other even in scenarios where communication channels are restricted and the opponent is unknown (preventing a pseudo-language of goal-irrelevant low-cost actions from emerging).
As an example, a model trained on no-press Diplomacy, which was in desperate need of an alliance with another power to its West, took an action that destroyed one of its armies to no gain, when said army was in a position to damage the power it was attempting to ally with.
Interesting! What’s the source for the second paragraph?
I think I read it in one of the papers Cicero cited, but that was a few years back, so I unfortunately don’t have a link.
Can you elaborate on what led you to this conclusion from the non-verbal bargaining demonstrated by the cats? I don’t quite follow the logic that “apparently effective non-verbal bargaining” necessarily leads us to “non-verbal bargaining is the default from which explicit bargaining emerges when it’s insufficient”.
My mental model for this is more like tool use. As a crude analogy, I could technically try to drive an object like a nail into an object like wood with my hand, but given a tool and my knowledge of the tool’s efficacy, I’m going to use that instead to avoid injury and make the job easier.
It could be that I don’t quite understand your connection of this observation to the “value-laden concepts” you conclude with. Is there an example you can share between humans who otherwise could engage in explicit bargaining that reinforces the non-verbal bargaining you observed in the cats?
The point I’m gesturing at is more abstract than this—it’s not that tacit bargaining is literally prior to explicit bargaining, in the sense that everyone always starts tacit and then resorts to explicit if that fails. I’m saying that tacit bargaining is “conceptually” prior to explicit bargaining; i.e.:
We can observe a world where tacit bargaining takes place without even the possibility of explicit bargaining (i.e. the world of cats).
However, (I claim) we cannot imagine the opposite. Explicit bargaining cannot take place without the background possibility of tacit bargaining, because otherwise the words used in the attempted explicit bargaining would have no meaning.
I haven’t actually read this, but there’s a book on this subject, The Territorial Imperative, whose Wikipedia summary is interesting and provocative. The author argues that human territorial behavior is essentially the same as that of animals, and so the verbiage we pile on top of it is less relevant to understanding what goes on in human conflicts than observations of animals.
Thank you for clarifying, that helps.
I think this is true, but if this is true, then that hints that the claim may not be falsifiable.
I say that not to diminish the claim in any way, but I’m not sure how to apply something you can’t test.
Assuming the claim is true, where do you go from here?
Potential answers:
This can motivate us to look more closely at animal behavior and evolutionary history to help us understand how humans behave in territorial conflicts (as alluded to in that book I mentioned) - because if human territorial behavior is not fundamentally linguistic in nature, then there’s no reason to suppose it’s unique to humans.
By contrast, in an area of human behavior that clearly is linguistic, e.g. storytelling, studying animals won’t tell us much.
The problems of philosophy are iatrogenic [1] . So maybe if you haven’t spent a lot of time getting yourself confused by debating metaethics, this conclusion seems trite to you; but for some of us it helps dissolve the confusions we acquired earlier. (For example, I used to favor the position that moral statements are purely logical, analytic truths like those of mathematics, and so any attempt to justify them by empirical observation or “human nature” is immediately suspect.)
More speculatively: People have been talking recently about LLMs engaging in negotiations. But LLMs are not animals, but purely linguistic entities. So what they’re doing may differ from what humans do in some important way.
See also Aaronson: “the world needs good philosophers, if for no other reason than to refute bad philosophers!”
Those are really clarifying! Thanks for taking the time to draw those out.