Thanks to Jessica Taylor for the discussion that was the stem of this article. The opinions here are all mine and I’m sure she’d disagree with some or all of them.
The Hai’Zu
Imagine the Hai’zu people of Stone Age Langong, using a very simple language that might contain only 6 words:
There (combined with pointing towards a direction)
Do (indicating that the person must do the appropriate action)
Gather (indicating that everyone should come and listen)
Them (indicating another person, combined with pointing)
No (negation)
Sun (indicating the concept of sun, which stands in for both time and weather)
The Hai’zu can coordinate very well using this simple language, Scout comes back to camp and says:
Gather There Do
This obviously means “there is an injured stag wandering around the forest, hunters go hunt it”. If they instead said.
Gather There No Do
This obviously means “there is a tiger around the forest, nobody should go there.
The old shaman can even chip in and say something like:
There No Do No Sun
Indicating that, since it’s a rainy day, so No Sun means “night”, that tigers hunt during the dark too, and nobody should go there at night either, even with torches.
At night a children can ask the Shaman
There No Do Sun?
(the question is implicit from the tone and the children—to—elder relationship)
And the Shaman can say:
There Do Sun
Indicating tigers don’t usually hang out in the same spot for more than a day, so it should be safe tomorrow.
The chieftain can assign tasks:
Them Do
Them Do
Them Do
People get that they are supposed to light the pile of wood, take care of the kid, and cook the dead elk.
Granted, the ancient Hai’zu probably had closer to 2000 words, or they got conquered by a neighboring tribe that did, you can only get so far without names for common things, pointing is inefficient, and only being able to do time and weather keeping on a day-by-day basis has limitations.
Still, the Hai’zu can coordinate much better than any living animal using this relatively simple vocabulary, well enough to synchronize hunting, camp maintenance, and caretaking.
Now, if a Hai’zu philosopher is born they might start saying things like:
There Gather Do
This is a perfectly valid sentence in Hai’zu’ish but it also makes no sense, if enough people take the philosopher seriously they might start having ideas and concepts about the great gathering in the sky, and have conflicting thoughts about whether “Do” in the sky refers to chasing the sun or having transcendental sex or lying in blissful wait. These questions might come up in the effigies they build to “There Gather Do” or in one of them uttering a sentence like:
There Gather Them Do Sun
Symbolizing the great hunt for the sun their chieftain will lead in the afterlife.
This is surely to cause some polarization as yet another scream.
There Gather Them Do Sun
Indicating the spirits of their ancestors are already in a timeless hunt for the Sun
Again, don’t take the Hai’zu very seriously, if you will demand of me to compose an entire prehistoric language and explain it’s flaws and uses in the span of this article, you’d surely get bored. But grant me that you can have a language which is:
a) Much more scarce than modern ones in terms of words and rules in combining them
b) Very useful, able to maintain coordination between humans
c) Able to ask “philosophical” questions, the sort which don’t have obvious answers that everyone agrees on. And make “philosophical” statements, the sort on which people have to ponder at length and might come up with conflicting interpretations.
Language Is Undefined
This is a fundamental problem of any language, no matter how complex. Indeed, the more complex a language the more statements in that language are going to be nonsensical even if correct based on its rules.
*We needn’t go into anything formal like equating language with a poor form of mathematics than pointing to Godel uncertainty proving that any such systems will contain contradictions. Though it’s probably important to keep in mind that **this property of having a contradictory/undefined/nonsensical set of statements, both theoretically and empirically, extends to any conceptual system we have come up with, be it a formalization of mathematical rules, a programming language or an actual language.*
To see that this is the case in language simply think of some sentences which are syntactically and grammatically correct but make no sense. I’d encourage to come up with your own, here are some off the top of my head:
I will swim to comb the hairdryer
They are negating flapping the moon
How do I know if Joel scratched the inside of my one of my dendritic cells?
Photons won’t dance on the windshield of my car
If you can’t imagine some, open the closest book of poetry.
We can analyze each of these sentences and give reasons for why they are false. We can also imagine insane worlds and reach out for far-off meanings to see how maybe they could be true.
If you can’t imagine insane meanings for the above, open the closest book of poetry interpretation.
But there is no rule that says we should be able to come up with a way to refute any nonsensical sentence. Indeed, given the immense ability of humans to stretch language, I’d be silly to think any sentence can be entirely refuted as nonsensical or not, it’s all a spectrum ranging from very nonsensical to very concrete.
Nor is nonsensical language inherently wrong, indeed, the language of experts often appears to be nonsensical, and it often is nonsensical outside of the very restrictive context in which those experts operate. Similarly, the language of lovers, poets, and friends is often nonsensical and that is precisely what gives it value.
I don’t mean to say that the purpose of language is this or that, indeed, that would be a nonsensical question to ask outside of a logical positivist meeting. However, it is important to note that language has a very important role as a coordination mechanism.
Making Sense Of Coordination
Problems often arise when we try to coordinate people around specific linguistical statements, that make no sense. This is expected since the only way to coordinate people around language is to use reality, the thing language is often describing.
We agree on things like:
Fire burns your hand
Salt makes french fries taste better
Letting go of a stone in midair results in it falling towards the ground
Because we can experimentally prove those things to be true, in a way that’s easy and obvious to see by anybody, can be done by anybody.
Also, we have the common sense to understand the limitations of each sentence, “Letting go of a stone in midair results in it falling towards the ground” doesn’t apply on the space station, yet this didn’t generate infinite philosophical debates nor lead to people applying the qualifier “on Earth” when referring to objects that fall. Some people don’t like salty foods, and some “fires” are indeed not that hot.
Out of a long tradition of people more focused on being efficient replication machines than being happy, certain bugs seem to arise. One of them is that we often choose to forget there’s a distinction between language and the real world, that language is both inherently contextual and often says things that are neither true nor false, they are undefined, nonsensical. This doesn’t always happen, but it seems to happen selectively, with reinforcement.
Noticing that language is contextual is especially hard if one always inhibits the same context, and reacts violently to any large changes in context. This violent reaction is often a combination of:
This context is wrong, it’s insane, Why does it exist!?
This context is unpleasant to be in, I want to get back
Parts of this context map onto the use of words in my old context, I’m going to ignore the other parts
Try to coordinate everybody on a meaning of language that is heavily dependent on the context they inhabit, and you have a problem.
Noticing that language often yields largely nonsensical statements is also particularly hard if you’ve been educated to think the opposite, in a system that is purely conceptual and doesn’t often touch reality. The only way to notice most statements are nonsensical is by jumping from linguistic descriptions of reality to reality. Be that reality your phenomenal experience when sitting with your eyes closed, running through a field, or looking through a microscope.
Try to coordinate everybody on a meaning of language that is nonsensical (independent of context), and you will fail.
The biggest problem here is that we hold statements to be true and, most of us, seem to only be able to accept that they are false when they’ve been replaced by other statements.
If I think X is Y, and I’m a particularly nice and reasonable person, I might be able to accept things like:
A is Y so X and A are similar
X is actually Z
X is Z if C and X if B
X is usually Y, but very rarely it’s Z
X is Y for me, but for them X is Z, and that’s fine
But we are unlikely to accept:
X is not a thing that can be other things
Y is not actually a thing that another thing can be
Even less likely to accept:
X doesn’t exist
Y doesn’t exist
And accepting a statement like:
Both X and Y don’t exist
Is reserved for Lao Tzu, Siddhartha, Alan Turing, and other enlightened demigods. Mortals just can’t accept this.
I’m not sure why our minds just can’t “walk back” on language, it seems like an extremely powerful mechanism. It’s one I’ve tried to cultivate but it seems especially hard, to the point where I myself have accepted that often enough it’s easier to “cover” nonsensical linguistical statements with other less nonsensical statements.
I think there might be useful techniques for doing this, like learning a new language and switching all of your conceptual thinking to that language.
I do think there is a middle path where one can’t quite drop the idea that certain sentences “must” make sense, but they can at least understand that they don’t at some rational level, and stop being so attached to them.
I can think “X is Y” and hear somebody say “X is not Y” or “X doesn’t exist” and instead of arguing, I can remember that “both X and Y don’t exist” and internally hug whatever part of my brain has been scarred by the impression that X and Y are indeed things.
Resolution
Don’t expect me to try and disprove any real examples of such undefined statements that people argue about there. Looking from the outside such statements often seem to hinge upon entire mountains of nonsense, and, in most people, trying to unskillfully pull the rug from under the mountain would just result in suffering and a defensive reaction.
I know I have such mountains of nonsense and I can’t even excavate those, I don’t have the hubris to think I could be doing that for the mountains of others. However, I have enough hubris to think that I might be able to gesture at the problem here, and maybe you can see it through your own lense, in whatever way makes sense to you.
The problem in itself, remember, is non-conceptual. You can’t come at it linguistically, the proof that statements are nonsensical must come from observing the real world, and it must come often come from seeing absence, not from seeing existence.
A great book that has helped me notice just how wide this problem was, is Meaningness.
David Chapman is, I think, highly skilled at doing this excavation work that I point towards. He tries to do it with subjects that might be very touchy but goes about it in a very roundabout, slow, and deliberate way. The book itself is, I think, intentionally written to be “obvious”, nothing it says is remotely controversial so you might find yourself nodding along and thinking “why the fuck am I reading this, it’s just a book of obvious statements” — The next step one should take is taking those obvious statements, using them to look at reality, and see what conflicting (nonsensical) statements arise in your mind during those moments.
I wish that this final section could be something more than a plug for this book, I certainly think that many other works, from ancient religious scriptures to modern textbooks on computational complexity point towards it, but they don’t seem to be nearly as efficient. If you know of any better techniques I would be eternally grateful for pointers to the works describing them.
It might be the case that some hard problems are nonsensical statements, but it doesn’t follow that all hard problems are: hardness can come from computational intractibility, lack of clinching evidence, etc.
“Looking from the outside such statements often seem to hinge upon entire mountains of nonsense”
I think it’s all nonsense, and that some nonsense is just more useful than other nonsense.
But it doesn’t confuse me why we don’t attempt to identify or reduce nonsense, nor does it confuse me why we don’t agree on what’s useful. We can thank our stupidity for this, and we’re stupid because we’re healthy. Hear me out here.
All this nonsense is actually ourselves, and “useful” means “useful for me”. So any truly objective thinking would require that we were without ego and sense of self, because only in this state would we stop prioritizing our own existences and preferences. Whenever you attack a specific persons nonsense, that person will feel as he is under attack, because he is.
Being objective, correct, logical, etc. is simply preferring that which isn’t human above that which is human. It’s no wonder that most people aren’t capable of this. We evolved specifically to stay alive, only people who are sick in the same manner that we are, are capable of being so “rational” that they might destroy themselves.
What’s the best we can do here? Perhaps it’s considering all human beings to be valuable. It’s the least egoistic perspective which still has subjective preferences (and thus considers life to be worth more than death). But even here, we can’t consider everything human to be equal, for there are human aspects that we find more desirable than others, e.g. I would like to reduce aggression, and then convice myself that this egoistic preference of mine is somehow objectively correct, rather than the desire for the growth of people who think and value like I do. Only by deceiving myself like this can I bring myself to prevent the existence of people who offend me or who are otherwise a danger to myself (and this includes people who are better than me!), and still believe that I’m a good person acting selflessly.
Of course, this comment is all nonsense. I’m just learned in similar nonsense to everyone else, so there’s a sort of coherence between people which is easily confused for some sort of objective correctness (rather than the shared subjectivity that it is). Of course, even logic and mathematics has zero relation to reality. Math is a set of paths from assumptions to their conclusion, and in the same way, answers relate to questions pair-wise. And these pairs rely solely on eachother, and it’s the same whenever you have neither or both. They’re zeros which has been expanded to some plus and minus which add up to said zero. This is why “solving” anything destroys it.
The difference in proxies for brain activity when speaking or doing nothing is marginal.
Non-conceptual acts from parachuting and sex to running and climbing likely result in more activation than speaking, they also tend to be more memorable and most people find them more desirable than talking to others, let alone to themselves.
Ultimately, this is subjective, of course, and if you tell me that your conceptual maps and word ladened narratives are important for who you are, I will of course grant that to you.
My personal impression is that they aren’t, they are quite important for what I do to be useful in society but that is like saying that, as a taxi driver, my car is a core part of who I am, which is also fair, but you know what they say, it’s sometimes good to just get out of the car.
For reference, this post was motivated by failure modes I see among capital R rationalists, scientists, philosophers, and other “many word many symbol” people—I don’t think the issues I point to would affect “normal” people as much. Normal people have reasonably good mechanisms to decouple from conceptual thinking, they might have a conceptual issue or two, but usually nothing that affects their quality of life that much.
Of course, you can interpret it the other way around too
That sounds about right, words and language are just communication of something, it’s stored quite differently in the brain.
My point was that we’re communicating things which we identify with and which we have made part of ourselves.
You know how LLMs sometimes dream or hallucinate? I like to think of people as more consistent versions of that.
Language is like a nonsense protocol which people can make sense of simply because we share it. Isn’t DNA the same? DNA is useful to itself, but whatever is encoded in DNA doesn’t really say anything about reality external from the DNA. But none of this is a problem unless two different contexts conflict.
And yeah, I was mostly agreeing. I just think the idea can be taken much further. Possibly so far that it cancels itself out again.
I suppose that some intelligent people tend to confuse their mental model with reality itself. This is the type which are a little too logical for their own good, and they tend to have difficulties with whatever isn’t rigid and unambiguous, relationships included.
I think that there’s fewer problems for intuitive types, since intuitive thinking is flexible and quite a lot faster. They realize that language is a tool for doing various things, rather than something valuable in itself. To demand “correct language” to the extent that it hinders communication would be an example of Goodhart’s law.
But even “correct” language is just that, language.
I suppose the best way I can explain this is, and I should have said this to begin with: Human life is itself a context. Most are under the impression that individual and subjective things are nonsense, while everything formal is objective and correct. I propose that everything only covers a limited context, and that some of these contexts merely cover a larger scope. I also propose that it’s not possible to create anything with an infinite scope, so at best we can create something universal in the case that it applies to our universe, but this is still a finite context.
If you think about the world in this manner, avoiding concepts like ‘infinite’, ‘true’ and ‘real’, making sense of the world becomes a lot easier.
And contexts conflict with (contradict) eachother, and higher consistency can be reached by solving contradictions. But every time you do this, something specific is destroyed, and everything is some degree of specific, so it’s entirely possible to destroy everything. With every generalization, you cover a larger scope but your generalization gets thinner and less rich.
If you increase depth, you decrease width and vice versa.
The nonsense of people is that which is unique to themselves and conflicts with other people, and the nonsense of humanity is our shared errors which conflict with logic, and logic is just a limited axiomatic system when viewed from the outside. But resolve these conflicts, and you’ll be left with nothing. The person is no more, humanity is no more, and if math requires universality, then math would have refuted itself as well.
It’s not just us which look silly from an outside perspective, everything does, and I don’t recommend destroying any of it for that reason.
Sorry for being verbose, it’s difficult to explain my view (the lessons I learned the hard way by being stupid) in just a few words, haha. If you want a saner engagement with your post I’m afraid you have to wait for other people to post
Ok, I think I better see your point, one thing I’m not sure you are on board with is that unique language can come from:
Direct experience of the world—or linguistic thinking about a memory (sky here has green~ish tint before a storm)
Linguistic thinking -- with several nested layers of recursion (what is <abstract concept>)
Ultimately there’s no hard line between the two but in practice, this seems quite relevant. Statements that come from 1) seem usually ok , and statements that come from 2) usually seem maladaptive.
Isn’t that unique composition of language? If you consider descriptions of the world to be something like a linear combination of words (or the impression they map to in your mind), then language can be said to span/cover some intrepretation of experiences.
I agree with you, and linguistic thinking is certainly the more reductive one.
But aren’t they both mistaken? As soon as anything is put into words, or encoded in logic for that matter, it has already been severely reduced and lost its connection to reality.
Difficult questions are basically syntax errors, loaded questions, self-contradicting statements, or statements with mistaken assumptions baked into them. But more importantly, I don’t think that even correct language has much in common with reality. We have no reasons to assume that the words we’ve come up with have the expressive power needed to align with reality, and it might be the case that no such words can’t exist.
You might be arguing that the meaning or concept behind a word is valid, and that manipulating these underlying things can allow us to understand the world. I suppose that this understanding of the world is correct enough to be useful, but I doubt that there’s any formal equivalence between even our intuitions and “reality”. As I see it, “getting out of the car” allows one to live a genuine human life, but it still doesn’t allow one to move outside the scope of human.
Languages used in daily life indeed have no formal, logically consistent, definitions, nor will they likely ever have one.
Pure math, if you count it, might have this, depending on how sparse and ‘elegant’ the minimal universal axioms turn out to be.
What do you consider the definition of a ‘language’ for the purposes of this post?
Just a gradient, math is not logically consistent either, nor are there any formal languages which are (again, one might claim, even provably so) -- besides maybe a very limited set of languages that have so many constraints as to be irrelevant (e.g. the logic required for a finite PD game might be self-consistent) -- regardless, any system of any real use isn’t so I don’t see much point in differentiating (as mentioned even in the post)
Maths is incomplete. Inconsistency isn’t proven.
Is this wrong?
Would love to see some false diagrams. Flow charts or circuits etc
Why the “actually”?
corrected to
aktually